text stringlengths 1.36k 1.27M |
|---|
Structural perfection of InGaAs/InP strained-layer superlattices grown by gas source molecular-beam epitaxy: A high-resolution x-ray diffraction study
J. M. Vandenberge, D. Gershoni, R. A. Hamm, M. B. Parish, and H. Temkin
AT&T Bell Laboratories, Murray Hill, New Jersey 07974
(Received 20 April 1989; accepted for publication 23 June 1989)
High-resolution x-ray diffraction (HRXRD) studies have been carried out to determine the structural perfection and periodicity for a number of high-quality InGaAs/InP strained-layer superlattices grown by gas source molecular-beam epitaxy. X-ray scans were carried out with a compact four-crystal monochromator resulting in a resolution of one molecular layer ($\sim 3$ Å), which enables one to observe very small variations in the periodic structure. Sharp and strong higher-order satellite reflections in the XRD profiles were observed indicating smooth interfaces with well-defined modulated structures. Excellent computer simulated fits of the x-ray satellite pattern could be generated based on a kinematical XRD step model which assumes ideally sharp interfaces. Our results demonstrate that HRXRD in conjunction with the kinematical step model provides a powerful tool to evaluate the structural perfection of InGaAs/InP strained-layer superlattices.
INTRODUCTION
Recent advances in the gas source molecular-beam epitaxy (GSMBE) have led to the growth of InGaAs/InP quantum well structures of very high crystal quality. Superlattices (SLs) have been grown with very thin wells, sharp interfaces, and well-defined periodic structures along with promising electronic and optical properties for device applications. Characteristic for the GSMBE is its capability for excellent control of composition, lattice match, and layer thickness, which makes it possible to grow very closely matched SLs as well as strained-layer superlattices (SLSs) having positive or negative strain. In previous work on lattice-matched InGaAs/InP SLs grown by GSMBE on [100] InP, high-resolution x-ray diffraction (HRXRD) has proven to be an effective tool to evaluate their structural integrity. The HRXRD spectra are analyzed with kinematical simulation based on a step model which assumes alternating layers with ideally sharp interfaces. In essence, the x-ray studies on GSMBE SLs demonstrated that kinematical simulation of HRXRD curves allows for very accurate determination of the periodic structure, including small structural modifications in the interfaces. This can be done with a precision of one molecular layer. With the thickness of the well small enough to accommodate the mismatch strain coherently, it is possible to grow either positively or negatively strained In$_x$Ga$_{1-x}$As/InP SLSs by GSMBE spanning a wide range of the In concentration $x$. In the present study, kinematical simulation of the HRXRD curves has been carried out in order to determine the periodic structural parameters of a number of InGaAs/InP strained-layer superlattices which were grown on [100] InP with various degrees of strain in the InGaAs well. This method appears to be very useful in order to quantitatively determine the strain with the aim of exploring the relationship between strain and the optical and electronic properties of the quantum wells.
RESULTS AND DISCUSSION
HRXRD was carried out using the x-ray geometry proposed by Bartels and consisting of a compact four-crystal monochromator, the sample as the fifth crystal and an open-end detector. X-ray scans were taken along the [h00] direction in the vicinity of the (400) reflection with the diffractometer run in the $\theta-2\theta$ mode. The x-ray scan of the (400) reflection is shown in Fig. 1 for three 10-period superlattices:
a negatively strained ($x = 0.43$), a lattice-matched ($x = 0.53$), and a positively strained ($x = 0.64$) SL. The presence of very sharp and strong satellite intensities indicates that these SLSs maintain their structural integrity throughout the whole superlattice. This result is consistent with the low-temperature photoluminescence (PL) spectra (Fig. 1) which show sharp PL peaks up to room temperature.\textsuperscript{6} While for the lattice matched SL the satellite intensities are symmetric ($+n$ vs $-n$) around the main (400) SL peak ($n = 0$), a strong asymmetry in the x-ray satellites around the (400) SL peak is characteristic for the SLSs, because of the strain modulation introduced by the accommodation of the large mismatch of the InGaAs wells with respect to the InP barriers. For negative strain as is the case for $x = 0.43$, the ($+n$) satellites become stronger than the ($-n$) satellites [Fig. 1(a)] while for positive strain this asymmetry is reversed as is demonstrated in Fig. 1(c) for the SLS with $x = 0.64$. Normally for zero-strained or weakly strained superlattices the period $\Lambda$ and lattice mismatch can be directly determined from the x-ray scan\textsuperscript{4,8} because the satellite intensities are fairly symmetric around the (400) SL peak. However, for the strongly strained superlattices this strain modulation causes the (400) x-ray satellite pattern to become asymmetric to the extent that the position of the (400) zero-order peak ($n = 0$) of the superlattice can no longer be identified unambiguously.
In order to identify the (400) SL peak and simultaneously determine the structural parameters, the x-ray satellite scans were computer simulated using a previously described kinematical step model which assumes ideally sharp interfaces. This model\textsuperscript{9,10} permits calculation of the diffracted amplitude $F$ of the structural periodicity of the superlattice along the [100] growth direction. The variable input parameters of the step model are the lattice spacings $d_w$ and $d_B$ between subsequent atomic layers and the number of molecular layers $N_w$ and $N_B$ in the InGaAs well (W) and InP barrier (B), respectively, where a molecular layer consists of one group III and one group V atomic layer. For alternating wells and barriers the diffracted amplitude $F$ in the vicinity of the (400) reflection ($h = 4$) is given by:
$$F(h) = \sum_{N_{SL}} \exp \left(2\pi ihN_{SL}\left(f_{WIII} + f_{WV}e^{iQd_w}\right)\right) \times \sum_{n=0}^{N_w-1} e^{iQ2nd_w} + \left(f_{BIII} + f_{BV}e^{iQd_B}\right)e^{iQ2Nd_B} \times \sum_{n=0}^{N_B-1} e^{iQ2nd_B},$$
where $Q = 2\pi h/d_{SL} = 4\pi \sin \theta/\lambda$, and $\theta$ is the diffraction angle; $f_{WIII}, f_{WV}$, etc. are the scattering factors for the group III and group V elements, respectively, calculated at the (400) diffraction angle $\theta$, $N_{SL}$ is the number of periods $\Lambda$ in the superlattice and $N := N_w + N_B$. Once the variable parameters are extracted from the step model, the thickness of the well and barrier can then be calculated from the equations $t_w = 2d_wN_w$ and $t_B = 2d_BN_B$, and the corresponding lattices parameters from $a_w^R = 4d_w$ and $a_B^R = 4d_B$. The strain in the well is then given by the equation
$$\epsilon_w^R = (a_w^R - a_w^0)/a_w^0,$$
where $a_w^0 = a_{\text{InP}} = 5.8687$ Å. From the strained-lattice parameter $a_w^R$, the relaxed value can be calculated by multiplying the strain $\epsilon_w^R$ with the quantity $c_{11}/(c_{11} + 2c_{12})$ using elastic theory\textsuperscript{11}; from Vegard’s law, the In concentration $x$ can then be extracted. The results of the kinematical simulation of a number of SLSs are presented, together with the structural parameters obtained from this model, in Table I. Also listed are the period $\Lambda$ which can be determined from the position of the satellite peaks and the perpendicular lattice mismatch $\Delta a_z/a_z$ which can be calculated from the difference of the main SL peak ($n = 0$) and the InP peak.\textsuperscript{7} As an illustration, the simulated fit is shown in Fig. 2 for one of the three SLSs of Fig. 1($x = 0.43$) with the structural input parameters listed in Table I, which also lists the fitted parameters of the other two SLSs.
With the precision afforded by GSBME it is possible to grow SLSs with very low In concentrations with the thickness of the well small enough to accommodate the large mismatch coherently. Two such examples of SLSs are shown in
| A (Å) | $\Delta a_z/a_z$ (%) | $N_w$ | $N_B$ | $\epsilon_w^R$ (%) | $t_w$ (Å) | $t_B$ (Å) | $x(\text{In})$ |
|-------|----------------------|-------|-------|-------------------|-----------|-----------|-------------|
| 528 | $-0.2$ | 26 | 154 | $-1.4$ | 75 | 452 | 0.43 |
| 540 | $0$ | 27 | 157 | $0$ | 79 | 461 | 0.53 |
| 517 | $+0.25$ | 26 | 150 | $+1.5$ | 77 | 440 | 0.64 |
| 316 | $-0.3$ | 7 | 101 | $-5.0$ | 30 | 203 | 0.15 |
| 326 | $-0.32$ | 7 | 105 | $-5.7$ | 19 | 308 | 0.10 |
| 429 | $+0.66$ | 34 | 112 | $+2.7$ | 102 | 328 | 0.72 |
| 573 | $-0.63$ | 59 | 137 | $-2.2$ | 169 | 405 | 0.37 |
FIG. 2. X-ray scan of the (400) reflection of a 10-well strained-layer In$_{0.43}$Ga$_{0.57}$As/InP superlattice with simulation of fit in lower part.
Fig. 3 for nominal In concentrations $x = 0.13$ and 0.075 and a well thickness $\sim 20 \text{ Å}$. The very strong asymmetry in the x-ray satellite patterns reveals a very large negative strain $\varepsilon_{\perp}^W$ as a result of the much lower In concentration in the well. Despite the large lattice mismatch, these SLSs retain their structural perfection, as judged by the sharpness and intensity of satellites and the excellent simulated fits. The fact that excellent computer-simulated fits can be obtained with the kinematical step model, which assumes ideally sharp interfaces, implies that the interfaces in the GSMBE SLSs are indeed extremely smooth and their periodicity is very well defined. Previous work on single quantum-well structures grown by the same method and studied by photoluminescence (PL) show single well-defined PL peaks for layers as thin as 5 Å.\textsuperscript{12} The results of this work are consistent with the PL results in that layers of varying thickness, such as may occur with other methods,\textsuperscript{13} are not seen.
During a recent study\textsuperscript{14} on the relationship between critical layer thickness and electrical characteristics in strained In$_x$Ga$_{1-x}$As/InP quantum wells, SLSs were grown by GSMBE with the nominal well thickness varying from 40 to as high as 200 Å, while the nominal In composition was kept at $x = 0.37$ and 0.74. Sharp and intense x-ray satellite spectra were measured as shown in Fig. 4 for a low
and high In concentration SLS. Both SLSs were at least partly relaxed by the networks of misfit dislocations as judged from the electrical measurements.\textsuperscript{14} Since the electrical measurements are sensitive to a very small amount of relaxation, on the order of 4%, these SLSs maintained most of their structural perfection. It should be noted that the large strain in the relatively thick wells results in very pronounced maxima and minima in the x-ray satellite structure of these two SLSs. In the other SLSs the large strain is accommodated in much thinner wells resulting in a more gradual decrease of the satellite intensities; this is, in particular, so for the two SLSs shown in Fig. 3, where a very large strain of $\sim 5\%$ is accommodated in very thin wells of $\sim 20$ Å. In either case, despite the large strain or thick wells, these SLSs retain their structural integrity, as judged by the sharpness and intensity of the satellites and the excellent simulated fits.
**CONCLUSION**
Our results demonstrate that HRXRD curves of GSMBE superlattices can be computer simulated in great detail by using a kinematical step model which assumes a series of alternating layers with ideally sharp interfaces. The successful application of this x-ray method is strongly enhanced by the fact that GSMBE superlattices can be grown with extreme precision and a large degree of structural perfection resulting in a very accurate assessment of the periodic structure and the strain in strained-layer superlattices.
\textsuperscript{1}M. B. Panish, H. Temkin, and S. Sumski, J. Vac. Sci. Technol. B \textbf{3}, 657 (1985).
\textsuperscript{2}H. Temkin, M. B. Panish, and S. N. G. Chu, Appl. Phys. Lett. \textbf{49}, 859 (1986).
\textsuperscript{3}H. Temkin, D. Gershoni, and M. B. Panish, Appl. Phys. Lett. \textbf{50}, 1776 (1987).
\textsuperscript{4}J. M. Vandenberg, S. N. G. Chu, R. A. Hamm, M. B. Panish, and H. Temkin, Appl. Phys. Lett. \textbf{49}, 1302 (1986).
\textsuperscript{5}J. M. Vandenberg, R. A. Hamm, M. B. Panish, and H. Temkin, J. Appl. Phys. \textbf{62}, 1278 (1987).
\textsuperscript{6}D. Gershoni, J. M. Vandenberg, R. A. Hamm, H. Temkin, and M. B. Panish, Phys. Rev. B \textbf{36}, 1320 (1987).
\textsuperscript{7}D. Gershoni, H. Temkin, J. M. Vandenberg, S. N. G. Chu, R. A. Hamm, and M. B. Panish, Phys. Rev. Lett. \textbf{60}, 448 (1988).
\textsuperscript{8}W. J. Bartels, J. Vac. Sci. Technol. B \textbf{1}, 338 (1983).
\textsuperscript{9}A. Segmüller and A. E. Blakeslee, J. Appl. Crystallogr. \textbf{6}, 19 (1973).
\textsuperscript{10}D. B. McWhan, M. Gurvitch, J. M. Rowell, and L. R. Walker, J. Appl. Phys. \textbf{54}, 3886 (1983).
\textsuperscript{11}J. Hornstra and W. J. Bartels, J. Cryst. Growth \textbf{44}, 513 (1978).
\textsuperscript{12}M. B. Panish, H. Temkin, R. A. Hamm, and S. N. G. Chu, Appl. Phys. Lett. \textbf{49}, 164 (1986).
\textsuperscript{13}P. C. Mohanty, H. M. Cox, P. L. Bastos, D. M. Hwang, J. M. Worlock, and E. Yablonovitch, Appl. Phys. Lett. \textbf{54}, 442 (1989).
\textsuperscript{14}H. Temkin, D. Gershoni, S. N. G. Chu, J. M. Vandenberg, R. A. Hamm, and M. B. Panish, Appl. Phys. Lett. \textbf{55}, 1668 (1989). |
Web Search Results Exploration via Cluster-Based Views and Zoom-Based Navigation
Karol Rástočný
(Slovak University of Technology in Bratislava, Slovakia
firstname.lastname@example.org)
Michal Tvarožek
(Slovak University of Technology in Bratislava, Slovakia
email@example.com)
Maria Bielikova
(Slovak University of Technology in Bratislava, Slovakia
firstname.lastname@example.org)
Abstract: Information seeking on the Web has become day-to-day routine for more than two billion human beings most of who use traditional keyword-based search engines. Developers of these search engines stress personalization, prediction of users’ next actions and mistake correction. But they are still struggling with results presentation and support for users, who make atypical queries or who do not exactly know what they are looking for. We address these issues via a novel approach for exploring web repositories, which naturally combines user search activities – look up, learning and investigation. We achieve this via view-based navigation in hierarchical clusters and two-dimensional graphs of search results.
Keywords: Navigation, Exploratory Search, Adaptive Views, Results Clustering, Graph Visualization, Web, Semantic Web
Categories: H.1.2, H.3.3, H.5.1, H.5.2, H.5.4
1 Introduction
In the last decade, we have witnessed significant advances in web information retrieval mostly related to query formulation, disambiguation and search results relevance. However, the improvement of search results exploration by end users has been somewhat limited. For example, while there is much improved autocomplete support and even limited faceted filtering in current web search engines, visualization and navigation in search results remains largely based on a linear list of results and the corresponding snippets.
In 2002 Broder already referenced a three step information seeking process for information retrieval consisting of a query, result selection and (optional) query refinement [Broder 02]. At the time it was assumed that users were largely able to define what they were looking for (i.e., formulate a query) and that retrieving a set of relevant search results was sufficient to satisfy informational user needs, which was known as “known item search” in classical information retrieval.
Present day user requirements however go beyond this assumption in several ways:
– *Users are not able to exactly formulate what they are looking for in advance* (i.e., when formulating their query). This has been shown by the number of short and ambiguous queries and the many query modifications needed to actually find satisfactory results [Jansen et al. 00];
– *Finding a set of relevant results (e.g., web pages) is by far not sufficient* to accomplish most present day tasks. Specifically for informational queries, it is necessary for users to learn and/or understand the discovered information. This shift in end-user perception of search was coined as *Exploratory search* by Gary Marchionini in 2006 [Marchionini 06].
Users typically focus on three aspects (steps) when exploring search results returned by a search engine:
1. *They identify whether a query modification is necessary and what modification it should be.* For this, users quickly glance over the search results to verify whether they correspond to their expected information need or not;
2. *They gain a broader understanding of the discovered results space, its scope and contents.* This usually occurs once search results seem to satisfy the user’s information need and need to be examined more closely to either gain an understanding of what is out there (e.g., user wants just an overview of the ‘domain’) or select specific results for further exploration (e.g., user needs to know the details about something);
3. *They gain a deeper understanding of something specific.* Once users select a specific search result, they may want to learn as much about it as possible, or just look for a specific bit of information contained in the result description. This typically results in a navigational session where a user explores a single search result and information associated with it.
Work has been done on all three aspects of search results exploration; most real-world advances however focus on query modification support, disambiguation and better search results relevance [Carpinetto and Romano 12]. We aim to improve search results exploration support for acquiring an overview of the results space by presenting the discovered relations and properties of individual search results.
In this paper we propose a novel approach to search results exploration that utilizes a hierarchy of adaptive views on search results, among which users can navigate by zooming via the mouse wheel and double-clicking. This way, we naturally interconnect user search and browsing activities while supporting all three of the aforementioned results exploration goals.
We describe current search results overview navigation approaches and details of exploration approaches in section 2. Our approach is described in section 3 along with its application in the image exploration domain in section 4. We present our evaluation in section 5 and discuss findings and outline future work directions in section 6.
## 2 Search Results Overview and Exploration
The creation of and interaction with search results overviews are crucial parts of the information retrieval process, since such processes present the majority of information
to users during a search session. After identifying suitable search results in a results overview, users typically proceed to explore individual search result details. Existing results overview and result detail view approaches can be viewed from these perspectives:
- **Generation** – how are data and metadata about results organized to generate the overview;
- **Visualization** – how are the data and metadata visualized to provide users with maximum information while minimizing information overload;
- **Interaction** – how, if at all, can users interact with the results overview.
### 2.1 Generation of Results Overview and Result Views
The capabilities of search results overviews depend on the amount of metadata available for their generation. These can range from large heterogeneous spaces, like the Web, to closed, highly structured relational databases in digital libraries with strong semantics. Based on these properties we distinguish three main types of data:
- **Structured data** consist of a thorough interlinked metadata structure associated with primary resources (i.e., images, multimedia, or even web pages). They are most often found in digital libraries or relational databases (that can have a web-based frontend) or in the Semantic Web. Structured data can be visualized by attractive hype graphs [Lanzenberger et al. 10], used for results overview personalization [Tvarožek and Bielikova 10], etc.;
- **Semi-structured data** usually consist of unstructured data for which some metadata are available that provide a makeshift structure or simple organization of the data based on some common attributes. Semi-structured data are often found in the open web (e.g., in form of folksonomies, tags, ratings, etc.) or in less organized relational databases. Examples of metadata for web pages include tags, recency and file type; for images attributes such as resolution, content type and creation date/time/place. One possibility how to use web page metadata (e.g., clicks streams logs) is personalized re-ranking search results [Ahmed et al. 09];
- **Unstructured data** typically include *generic web pages* and/or *plain multimedia content* (images, audio or video) where little to no metadata is available. Without metadata, plain content can be presented only in its raw form. This problem is addressed via multiple metadata mining techniques e.g., ontology learning [Hazman et al. 11].
Similarly to the generation of a results overview, the generation of views for exploring individual search result details depends on their content type and the amount of available metadata:
- Individual results in *unstructured data* are typically explored via content type specific viewers (e.g., web browser, image viewer, video player). Exploration of results usually corresponds to next/previous style navigation or in case of interconnected web pages to browsing hyperlinks;
Browsers of *semi-structured data* take advantage of limited metadata availability to provide improved visualization and browsing of content. This often includes rendering of the actual content and an additional table-based visualization of result attributes via label/value pairs. Where attribute semantics are known, specialized visualizations are used to provide more intuitive interaction (e.g., map-based visualization of GPS coordinates) or links to other related content (e.g., other images at that location) [Berners-lee et al. 06];
Exploiting metadata available in *structured data* browsers often initially shifts focus from the actual content (text/images/etc.) to exploration of the associated metadata and switching to specialized content browsers once the desired content has been located. This enables the creation of advanced browsers taking advantage of the metadata to create for example cluster-based, graph-based or map-based views of result properties and relations.
Since our interest is the use of metadata to improve search result exploration, we do not go into detail of content type specific image/video/audio and web browsers for unstructured data and focus on approaches for semi-structured and structured data.
One must note that any kind of interlinked structured data is essentially a graph structure usually with nodes corresponding to entities and edges corresponding to relations respectively. An example relationship in the audio domain might be “Ludwig van Beethoven” *isComposerOf* “The ninth symphony”. This is analogous to a triplet in the Semantic Web, where “Ludwig van Beethoven” corresponds to the subject, *isComposerOf* to the predicate and “The ninth symphony” to the object. Moreover, the Semantic Web often explicitly defines a meta-layer above the actual data in terms of classes (e.g., *Composer* or *Composition*) and properties (e.g., *isComposerOf*).
### 2.2 Visualization of Results Overview and Result Views
Currently, most search engines employ text-based visualization of search results where text or other (meta)data corresponding to individual results are rendered in lists or tables. Advanced or experimental search modes sometimes use graphical visualization of search results via maps, graphs, histograms etc.
The most used text-based search result overview visualizations include:
- *List-based result overviews* used by most keyword-based web search engines like Google or Bing. These are displayed as linear paginated lists of search results usually containing short snippets of text from target web sites. The ranking of results mostly defaults to ‘relevance’ and cannot be changed;
- *Table-based result overviews* used by many database driven search engines (e.g., in online shops) where results are shown in a paginated table of search result attributes. Ranking is usually based on popularity or other attributes of the results (e.g., price, availability);
- *Facet-based results overviews* used by advanced faceted browsers that in addition to traditional list/table-based search results overviews include a set of facets that describe the attributes of the entire results set and not just the first page of results as in previous approaches. The facets correspond to a faceted (multidimensional hierarchical) classification of the information space and
enable users to gain an overview of the entire results set and further refine the query simultaneously [schraefel et al. 06].
Graphical search results overview visualizations include:
- **Map-based results overviews** used for visualization of geographical data (e.g., hotel locations in Booking.com, photo geo-tagged in Panoramio) or to show aggregated data or distributions for entire regions;
- **Histogram-based results overviews** used to visualize distributions of search results attributes or time-based data (e.g., publication download counts over time). A special type of histogram-based results overview is provided by faceted browser – Elastic lists [1], which visualize weight proportions of facet values [Stefaner et al. 08];
- **Cluster-based results overviews** show groups and hierarchies of results graphically (similarly to facets, which do the same via text). These groups are obviously predefined by domain expert (e.g., in case of Pivot project [2]) or built by clustering algorithm from results metadata (e.g., from keywords extracted from snippets provided by search engines in case of IGroup [Jing et al. 06], or Yippy [3]).
Different results overviews are often combined together to give users different views on results. The VisGets system [4] displays these visualizations (list-based, tag-based, map-based and histogram-based results overviews) in one view. VisGets coordinates user actions in all overviews, while it visualizes dependencies among overviews by highlighting of affected results in the other overviews [Dörk et al. 08]. Bozzon, et al. choose a different approach and present different results overviews in separate specialized tabs [Bozzon et al. 11]. This visualization hides some details from users, but decreases information overload allowing users to select the type of information they want to see.
While results overviews visualization can be customized for specific domains, an exploration of a single search result is most often performed in one of these ways:
- **Table-based visualization**, which renders flat textual label/value pairs of result attributes. Extensions to this approach include the generation of links to related resources in values where applicable or the expansion of values into visualizations of associated resources effectively resulting in a nested table of label/value pairs. Information overload occurs for results with many attributes or relations or when too many values are expanded;
- **Graph-based visualization**, which natively renders the aforementioned graph structure as a set of nodes and edges. The view usually starts with the explored search result in the center showing only its attributes (either as edges to literal nodes, inside the center node itself or as a tooltip) and relations (as edges to other unexpanded nodes). Users can expand additional nodes, compare resources and pan the view to shift focus to other resources. However, graph-
---
[1] http://well-formed-data.net/archives/54/elastic-lists
[2] http://www.microsoft.com/presspass/features/2010/feb10/02-11pivot.mspx
[3] http://search.yippy.com
[4] http://innovis.cpsc.ucalgary.ca/Research/VisGets
based views often quickly become too complex due to information overload and node layout issues. Moreover, orientation support, albeit crucial, is often limited as the view can rearrange itself to accommodate new nodes confusing users. To counteract information overload, nodes are sometimes hierarchically clustered resulting in *cluster-based visualization* that groups similar nodes at different levels of abstraction [Herman et al. 00].
### 2.3 Interaction with Results Overviews and Result Detail Views
Historically, the interaction with search results was virtually non-existent and limited to selecting a result for further exploration (i.e., clicking it). Exploratory search approaches as well as interactivity of Web 2.0 and social applications extended the interaction options to include:
- **Selection** – the generic interaction option to select a result for further exploration;
- **Rating** – the possibility to rate a result thus giving feedback about its perceived value to the user and ideally usefulness for others. This interaction is most often implemented in browsers of homogenous content such as e-shops or video portals;
- **Previewing** – the option to preview more (relevant) information about the result without actually having to select and explore it manually. Previewing of a result is usually mapped to the ‘mouse over’ action, when a control with detailed information about a result is displayed (e.g., Google Images);
- **Refinement** – the option in faceted browsers to further refine the current results set via facets. For example, some keyword-based search engines also include faceted refinement of results via additional metadata (e.g., freshness) or based on search type (e.g., image, shopping). In more database-driven applications (e.g., online shops), faceted refinement of results is often possible based on attributes of search results as well as other metadata.
Each of search results visualizations can optionally be augmented with previews of the actual content such as image thumbnails or video screenshots, and data type specific visualizations of location (maps) and time (calendars/timelines). Typical interaction options include:
- **Scrolling and panning** – horizontal movement in the visualized information space at a constant abstraction level to explore other related resources;
- **Node/cluster expansion/collapsing** – vertical movement in the visualized information space either providing more or less details about particular resources;
- **Zooming** – changing the scope of the visualized information to get a better overview of the entire visualized information space typically without changing the level of abstraction.
User interaction with the aforementioned search results visualizations consists of successive steps comprising a combination of scrolling/panning, expansion/collapsing, and zooming in/out until users satisfy their information need. But most of these visualizations are independent, narrowly focused on specific types of information so their usage for seeking different types of information is ineffective and uncomfortable.
for users. If visualizations support adaptive interoperability and natural switching of different views on explored information, one tool will almost completely cover all possible types of desired information and almost all user specific seeking strategies.
These visualizations could be naturally interconnected by zooming, whereby users’ zooming activities lead to new views with a different granularity of the presented information. Zooming activities are already often used in different information systems, but still for switching or enrichment of views aimed to same goals. E.g., by zooming into a place on the map, continents view is firstly switched to more detailed landscape view with added smaller rivers and roads and after next zoom in, photos from the zoomed on place are displayed.
3 Zoom-Based Results Navigation
We combine the aforementioned visualization and interaction approaches into a single navigation paradigm to improve overall user experience during search results exploration. We integrate two core visualizations into the proposed navigation paradigm which reflects users’ actions during search session:
1. *Visualization of search results* – helps user orientation in a set of results provided by a search engine;
2. *Visualization of related results* – allows users to explore detailed information.
Users seamlessly navigate between these views via zoom which is performed via the mouse wheel and zoom-in/out buttons like in map or photo viewers, which are already familiar to many users. This makes it easy for users to identify the desired results in a set of results provided by a search engine and find out more detailed information about the identified result and about results related to it.
The main idea of interconnecting search results overview and exploration into one paradigm, in which users use the same interaction principles, is better supporting domain exploration and learning. Thus users find the initial result whose features partially satisfy the user’s requirements via the visualization of search results. Thereafter the visualization of related resources allows users to naturally learn and understand its unknown or more obscure features and also helps users to find more satisfying results.
3.1 Visualization of Search Results
Users often refine their queries after quick inspection of the provided results in the first page of a search engine. They repeat this refinement until they discover a suitable result or they give up on the search session [Levene 10]. Users’ success and spent time depend on their understanding of the results, which is markedly influenced by the form of results presentation. Since it is not possible to present all results to users, we present results clustered into hierarchical clusters. To preserve readability, we allow up to nine clusters in one level based on the general rule – seven, plus or minus two items [Miller 56].
Clusters are built by standard clustering algorithms (e.g., c-means algorithm). First, properties of results used for computing the similarity of results and for labeling of clusters must be specified. For unstructured data keywords are often used and can
be extracted from snippets [see section 2.2]. Our clustering approach for structured data is described in section 4.1.2. It works reasonably well also for unstructured data after their enrichment by lightweight semantics (e.g., tags or keywords).
We visualize the created clusters as labeled boxes, in which the most prominent sub-clusters with representative results are displayed [Figure 1]. This visualization allows users to easily and quickly understand clusters without the necessity of looking into clusters. Users can browse results in clusters by zooming into clusters or by double clicking their labels, which shows clusters from next level. If the zoomed in cluster does not have any sub-clusters, all results from the cluster are displayed.
**Image**
**Computer Graphics**
- Visualization (computer graphics) - Wikipedia, the free encyclopedia
en.wikipedia.org/wiki/Visualization_(computer_graphics)
Visualization is any technique for creating images, diagrams, or animations ...
- Visualization and Graphics at UC Davis
www.graphics.stanford.edu
Stanford University. Research areas include volume rendering, rendering algorithms and ...
**Visualization tools**
- ITT Visual Information Solutions, Image Processing & Data Analysis
www.ittvis.com/
ENVI, IDL, and IAS are the premier software solutions for image processing, analysis and
- VTK - The Visualization Toolkit
www.vtk.org/
The Visualization ToolKit (VTK) is an open source, freely available software system for 3D
**Figure 1: Examples of a cluster and its sub-clusters visualization generated from**
a) website results; b) image results of the search query “Visualization”
### 3.2 Visualization of Related Results
Our cluster-view shows basic information about results, which we deemed insufficient. We address this via the visualization of related results which allows users to explore related results and therefore supports users who seek more details about the discovered result or answers to new questions which arise.
One of the greatest concerns during result exploration is becoming "lost in hyperspace", i.e. losing track of the original purpose of the exploration sessions. To prevent this problem, a user has to know answers to the following questions at each phase of the exploration process [Nielsen 00]:
1. Where am I relative to my initial result?
2. Where did I come from and how can I return back?
3. Where can I go to now and how can I achieve my goal?
We support exploration of related results via visualization of graphs of related results. At the beginning of an exploration session, the initial result (found within the original set of search results) is displayed in the middle of the graph and related results are displayed around it. As related results we consider e.g. websites linked from the initial web site or results with common properties. Users interact with this graph by zooming into or double clicking on results, while each zoom in action adds related results to the zoomed in result [Figure 2]. The history of users’ actions stays visible in this form of navigation, which answers the first two questions. Navigation by graph visualization also displays relations between results thus helping users to better understand the explored information space and answering to the third question.
*Figure 2: Graph visualization of image results. Oriented edges represent sequence of navigation actions, while gray edges represent possible ways how to continue in exploration and black edges represent the user’s previous path through the graph*
Visualizations of graphs often struggle with readability and clarity. These problems are mainly caused by a quick increase of the number of results presented in the graph. Too many displayed results complicate future navigation decisions. We address this by extending simple graph visualization with support of tools that decrease the amount of presented information, make the graph more concise, and help users to decide how to continue in navigation:
- **Node (results) clustering** – connects nodes with common incident nodes to one cluster node, what markedly decrease the number of nodes in the graph;
- **Node marking** – users can mark nodes with marks that specify whether newly added nodes (results) have to or may not be incident with the marked nodes;
- **Personalization** – results, which are likely uninteresting for users in the context of the current navigation session, are automatically filtered out;
- **Next actions recommendation** – nodes likely leading to the desired information, are highlighted.
4 Case Study: Image Exploratory Browser
We describe an open information space case study for a multimedia ontology. For this case study we employed our faceted browser Factic [Tvarožek and Bielikova 10], which allows users to search in a repository via facets generated from an ontology. We have extended Factic with search results exploration via hierarchical clusters generated from a faceted search results set and with zoom-based navigation [Rástočný et al. 11].
4.1 Cluster-Based Navigation
Factic provides two types of cluster-based navigation. The first type is navigation via facet-based clusters which correspond to facets generated from properties of objects in the ontology. This type of navigation allows users to create and refine facet-based search queries. The second type is navigation via hierarchical clusters of search results, which simplifies search results exploration and understanding for end users.
4.1.1 Facets Generated from the Ontology
Facets can be generated by fully autonomous approaches based on knowledge extraction from ontologies [Teufl and Lackner 11]. But these approaches often have high time and memory complexity and they need to be re-executed after each modification in ontologies. Therefore Factic provides facet generation as a semiautomatic algorithm, which has to be initialized by a human expert with facets specifications consisting of four templates [Tvarožek 11]:
- A *facet template* specifies the overall type and behavior of the facet and its corresponding pattern in the domain ontology (i.e., how its label is generated and whether its restriction values are ontology instances, classes or a mix of both);
- A *restriction template* defines how individual restrictions in the facet are constructed and mapped onto the domain ontology (i.e., how restriction labels are created and whether a flat enumeration of values or a hierarchical tree of restrictions is to be generated);
- A *query template* describes the mapping of facet restrictions to database queries which are used to filter results to a specific cluster based on the facet (i.e., which domain properties link search results to restrictions and how SPARQL queries are constructed to find results);
- A *visualization and interaction template* binds the facet to the graphical user interface (i.e., defines which GUI controls are used for facet visualization and user interaction; while our prototype employs lists of textual values, other interaction styles include sliders, calendars, maps, tag clouds, etc.).
For example, a simplified informal facet specification might look like this:
- Facet template: *instance facet* based on instances of the class *Author* using the instance property *rdfs:Label* as the user interface label of the facet;
– Restriction template: flat instance enumeration of Author instances (resources associated via rdf:type with the class Author), use instance property rdfs:Label as the user interface label for facet values (restrictions);
– Query template: direct relation between search results and restriction via the property hasAuthor (i.e., corresponds to the SPARQL query pattern {search result} property {restriction}, e.g. {search result} hasAuthor {michalTvarozek});
– Visualization and interaction template: list of textual values (this corresponds to a user control “template” that is fed with data and used in the GUI to interact with the user).
After human initialization (with the abovementioned templates), facets are automatically generated from the underlying ontology and displayed to users. In the case of Factic, we present facets as named lists of values [see Figure 3]. When a user selects a facet value, a new database query is built from the facet’s query template by application of the selected value and new set of results is retrieved from the repository. After that, facet value ranges are restricted to values, which are applicable to the new results set.
This search approach always displays relevant search results properties (facets with possible values) and effective search results exploration by selecting facet values [Fagan 10]. But it does not show relations between facets (e.g., how a selection of specific value in one facet affects value in another facet) and examples of results to users.
### 4.1.2 Hierarchical Clusters of Search Results
Navigation issues of faceted browsing are solved by search results classification into hierarchical clusters, which uses semantic properties of search results to generate clusters and thus does not have to be predefined by domain specialists. We visualize the generated clusters along with results as named boxes with results examples [see
Figure 4]. Cluster names are composed from common semantic properties of results in a cluster and most representative results are used as examples. To improve overall readability, we also present four most valuable clusters from the next level. When users double click on a cluster, clusters for the next two levels are displayed.
*Figure 4: Visualization of hierarchical clusters. Each box represents one cluster. Nested boxes represent clusters from lower levels of hierarchy*
We generate the cluster hierarchy via a customized algorithm based on the hierarchical-hyperspherical divisive fuzzy c-means algorithm (H2D-FCM) [Bordogna and Pasi 09]. We chose the hierarchical fuzzy algorithm for two main reasons:
- It can quickly classify results into clusters in the first levels, which are presented to users and calculate clusters for deeper levels, while users are exploring the presented clusters (facets);
- Fuzzy clustering creates more natural results classification into clusters. In addition, the selection of a proper threshold value ($\leq 0.5$) leads to classification of results on the clusters borders into multiple clusters, so users do not lose information if an optimal number of clusters on the first level has not been selected (e.g., as a consequence of computational costs).
Our clustering algorithm works in six steps:
1. **Building of a feature vector** – we use a feature vector of semantic properties of the clustered search results. Because these properties have discreet Boolean values (e.g., a photo has or has not been created by a Nikon camera), the number of feature vector dimensions is growing rapidly with the number of results (e.g., in our dataset we obtain more than 700 properties for 100 results). This negatively affects the effectiveness of the clustering algorithm and the quality of clusters. For these reasons we optimize the feature vector by removing properties associated with almost all or only few results (these properties would lead to one giant or many tiny clusters) and properties, which are uninteresting for users (clusters based on these properties have no added value for users);
2. **Determine the number of clusters in the first level** – the number of clusters in the first level is calculated by (1), where $K_{top}$ is the number of clusters and $K_{prop}$ is
the dimension of the features vector. We limit the number of clusters to 9 in order to overloading users with information [Miller 56]:
\[
K_{top} = \min \left( 9, \text{round} \left( \sqrt{\frac{K_{prop}}{2}} \right) \right)
\]
3. **Initialize cluster prototypes** – cluster prototypes are feature vectors with one active dimension (a semantic property). During selection of these vectors, the number of search results with the selected property and distance of the prototyped clusters are considered;
4. **Construct clusters via fuzzy c-means clustering**;
5. **Generate clusters labels** – labels of clusters are generated from semantic properties in their prototype feature vectors;
6. **Create clusters on next levels** – nested clusters are built by the original H2DFCM algorithm, with a maximum of 9 clusters per level.
Once users discover the desired information, they ask new questions (e.g., when a user has found a photo of a horse, he often wants to explore more photos of horses). To satisfy evolving exploration goals, users can start a new search session and or continue in the current session by exploring similar or related results to a previously identified result. We propose the second possibility as more effective and comfortable for users, because new exploration goals usually are closely related to previously identified results. We address this by zoom-based navigation in the graph of related results, which visualizes relations between results.
### 4.2 Zoom-Based Navigation
Zoom-based navigation in graphs simplifies the understanding of the abstract information space by end users. Additionally, graph visualization is an attractive way to visualize relations. But tools which utilize graph views struggle with graph readability and clarity especially for large information spaces. To address these issues, we employ a force-based layouting algorithm, which allows us to smoothly visualize modifications in graphs. It also generates symmetric graphs, which are more acceptable for users [van Ham and Rogowitz 08]. We also address information overload via leveled graph visualization for different degrees of detail and extensions that decrease the number of nodes in the graph and improve user orientation based on results clustering, attributes marking, next action recommendation and user adaptation.
#### 4.2.1 Graph Visualizations
We propose two graph visualizations as levels of abstractions, which can be perceived by end users as zooming, in order to improve end-user usability. To increase this effect, we subordinate the most of user’s actions to zooming via the mouse wheel and zoom-buttons, similarly to other zoom-based tools such as maps or image viewers. So a transition between them is natural and easily understandable.
The first proposed visualization is the *attributes graph*, which is shown after zooming onto an identified result. At the beginning, the graph consists of one node
representing the result and several nodes around it, which represent the result’s attributes. Users can explore new results by expanding an attribute node, which adds new results with this attribute to the graph (represented by results thumbnails). Attribute nodes connect results’ properties together with properties’ values [Figure 5]. This way, we allow users to compare results directly through one node, what makes the graph more readable and understandable.
This visualization does not allow straight interaction with property values. For example, if a user has found a photo, which was taken by *Adam*, this person is joined with the property *Author*, so the user cannot find photos of *Adam*, i.e. associated with the same person via a different property.
We address this with our second visualization via *RDF graph* – the graph of RDF triples from the multimedia ontology. Users switch view to this visualization by naturally zooming into an already expanded attribute node. This transition is visualized to users via division of attribute nodes into two nodes – property node and value node [see Figure 5].
RDF graph visualization tools traditionally display types of properties directly on edges between subjects and objects (e.g., IsaViz [4]) or they use different visualizations of edges for properties, while users see details about properties after selection of subject or object (e.g., RDF Gravity [5]). Both these RDF graph visualizations add some additional meaning to edges, so it is not enough for users to read only nodes; users also have to understand the meaning of edges. In addition, the second visualization divides information about properties into two parallel views where users have to compare a graph with a list of results’ properties. Our approach with properties visualized as nodes does not change the meaning of edges and gives better interaction possibilities that are the same as in the attribute graph.
[4] http://www.w3.org/2001/11/IsaViz/
[5] http://semweb.salzburgresearch.at/apps/rdf-gravity/
This graph has more user interaction options, but the results have to be perceived through three nodes (the property node of the first result, the value node and the property node of the second result), which conventional users are not used to.
Besides node expansion, which is performed by zooming into or double clicking on a node, we provide three additional interactions with nodes via controls on the node itself [see Figure 6]:
- **Node hiding** – the node and its connections are removed from view;
- **Node locking** – prevents change of the node’s position by layouting algorithm;
- **Node marking** – marks the node via a three-state check box [see section 4.2.2].
We devised user interaction with the graph in such way that users can perform all their actions right in the graph only with a computer mouse. This decreases cognitive user load and the probability of losing navigation context by looking around the interface in order to find the necessary tool button.
### 4.2.2 Attributes Marking
Node expansion can introduce a lot of new nodes into a graph, yet only a subset of these is usually of interest to the user. To allow users to directly affect properties of newly added results, we propose the concept of attributes marking, which allows users to pre-filter new results by marking attribute nodes with one of tree marks, which set up additional restrictions for new results:
- **Wanted** – newly added results must have all attributes with this mark;
- **Unwanted** – newly added results must not have any attribute with this mark;
- **Allowed** – newly added results are not restricted with these attributes.
### 4.2.3 Personalized Filtering of Nodes
In addition to attributes marking that provides manual node filtering, we also perform automatic pre-filtering of potentially uninteresting results and attributes nodes, which are displayed after node expansion. We base user adaptation on a model of users’ interests, which stores interest weights of resources (results and attributes’ values) and predicates (attributes’ types, properties). Updates to weights of interests in the model are caused by users’ interaction with results and their attributes in graph visualization. Changes of these weights occur, when users perform these actions:
- **Hide a result node** – the weight of the result is decreased by one;
- **Hide attribute node** – weights of the attribute’s type and attribute’s value are decreased by one;
- **Result node expansion** – the weight of the result is increased by one;
– **Attribute node expansion** – weights of the attribute’s type and attribute’s value are increased by one. The same action is performed for all marked attributes.
To include changes to user interests over time, we perform degradation of interest weights at regular intervals, which means that all negative interest weights are increased and all positive interests weights are decreased by one.
### 4.2.4 Results Clustering
Another way to decrease the number of displayed results in the graph is results clustering, when newly added results are not displayed separately, but are wrapped by cluster nodes. We group results with common attributes displayed in a graph to one cluster node. This allows users to gain an overview of results categories in the context of known (displayed) information, choose an interesting results subset and interact with it instead of overloading with the sheer number of result nodes. Clustering results also increases performance of the layouting algorithm by lowering overall node count.
We visualize cluster nodes as a matrix of thumbnails of randomly selected results from clusters in the attribute graph and as ellipses labeled with results types in the RDF graph [Figure 5]. To give better visual information about the size of a results set covered by clusters, we logarithmically size them based on the number of results. Interaction with cluster nodes is similar to interaction with other nodes with one exception. Zooming into a cluster node does not expand this node, but displays results in the cluster by means of facet-based navigation. Users can explore results in this view, easily identify results which are interesting and extract them from the cluster by double clicking [see Figure 7].
### 4.2.5 Next Action Recommendation
In addition to the previously described extensions focused on increasing readability by removing information (nodes) from the graph, we also propose next action recommendation to provide user guidance in the graph by highlighting nodes, interaction with which will likely lead to the desired information. Next action
recommendation is based on query flow graphs [Boldi et al. 08], from which we derive an activity flow graph by substituting queries by sequences of actions.
Activity flow graphs are built separately for each session, whereby we collect four base activities: node expansion and collapsing, marking and hiding. Each node in the flow graph contains a normalized action sequence, so that all action sequences that generate equivalent navigation graphs have the same representation. If a situation where a new actions sequence has the same representation as a previous one occurs, the previous action sequence is set as the active node (active action sequence). Edges in activity flow graphs represent transitions between action sequences by application of one action. Edge weights are set to 1 by default. When the last action sequence in an active node of activity flow graph has led to details about a result, all edge weights on the path to the root of the activity flow graph are changed to 2. An example of the construction of an activity flow graph is presented in Table 1.
Recommendation based on query flow graphs employs graph comparison by determining of their distances [Boldi et al. 09], [Bordino et al. 10]. We utilize this in our action flow graphs, where we compare action flow graphs (that are built from actions of all users) only by paths from active nodes to their roots. Our recommendation based on action flow graphs is based on the premise that users, who at same point performed similar actions, have similar goals or at least the same sub-goals of their exploratory activities. So if we find paths in the history of all users’ activities, that are similar to paths of active users, we can help them make decisions about the next navigation step. While this recommendation could be used to recommend the final results directly, it would also prevent users from understanding relations between the currently explored results and the proposed final results. This happens because users would not have explored the results on the path to the final results and thus would be unable to acquire the necessary knowledge about the explored domain.
| No | Action | Normalized action sequence |
|----|---------------------------------------------|----------------------------|
| 1. | Expansion of the node $N_1$ | $a_1 = [+N_1]$ |
| 2. | Expansion of the node $N_2$ | $a_2 = [+N_1+N_2]$ |
| 3. | Expansion of the node $N_3$ | $a_3 = [+N_1+N_2+N_3]$ |
| 4. | Roll up of the node $N_2$ | $a_4 = [+N_1+N_2+N_3-N_2] = [+N_1+N_3]$ |
| 5. | Roll up of the node $N_3$ | $a_5 = [+N_1+N_3-N_3] = [+N_1] = a_1$ |
| 6. | Expansion of the node $N_4$ | $a_6 = [+N_1+N_4]$ |
| 7. | Display of details of the node $N_5$ | $a_7 = [\text{display}(N_5)]$ |
Table 1: Example of activity flow graph construction showing the final graph
Based on this, we count weights of possible next actions by equation (2), where \( w(a_i, a_j, g_k) \) is the weight of the action \( a_i \) in the transition from the action \( a_j \) for the graph \( g_k \) (the graph, for which is counted weight of the action \( a_i \)), \( G \) is the set of all possible action flow graphs, \( c(a_j, g_k, g_l) \) defines the number of nodes with same action sequences in graphs \( g_k \) and \( g_l \) from the action \( a_j \) to graph’s root, \( h(a_j, g_k) \) is the depth of the action \( a_j \) in the graph \( g_k \) and value \( e(a_j, a_i, g_l) \) is the weight of transition edge from the action \( a_j \) to the action \( a_i \) in the graph \( g_l \).
\[
w(a_i, a_j, g_k) = \sum_{g_l \in G / \{g_k\}} \frac{c(a_j, g_k, g_l)e(a_j, a_i, g_l)}{\max(h(a_j, g_k), h(a_j, g_l))}
\]
(2)
Computed weights of possible actions following the last user action (\( a_j \) in equation(2)) are sorted and the first three actions with the highest weights are selected and nodes, with which they interact, are highlighted.
### 4.3 Scenario
In the real world, users start their information seeking session mostly with a more or less concrete goal. In this scenario,
- a user wants to see photos from an event in Beijing.
Therefore he selects “Place Beijing” via facets. After that he notices a group of photos with the tag “Great Wall of China” [Figure 8].
*Figure 6: Clusters generated over photos from Beijing with examples*
This subset of results holds his interest, so he previews these photos and notices that his colleagues Pavol Navrat and Michal Tvarozek are together on one photo. He decides
- to find photos with Pavol Navrat without Michal Tvarozek at Great Wall of China.
So he zooms onto this photo. As a reaction to this action, the view is switched to the attributes graph view [Figure 9]. After that the user marks the attribute node “Shows Michal Tvarozek” as unwanted and node “Shows Great Wall of China” as
wanted and zooms onto the attribute node “Shows Pavol Navrat”. This action adds one new result to the graph [Figure 10].
*Figure 7: Attributes graph of the photo with Pavol Navrat and Michal Tvarozek at Great Wall of China*
*Figure 8: Searching for a photo with Pavol Navrat without Michal Tvarozek*
He sees that photos were shot by “Mária Bieliková”, who he does not know. So he tries
- to find some photos of her.
This can be accomplished by switching to detailed view, which is realized by the RDF graph. So he zooms onto the attribute node “author Mária Bieliková” [Figure 11]. As a reaction to switching views, attribute nodes are split into value and property nodes and values with same relations to remaining nodes are clustered to nodes titled
by types of values (e.g., values “Great Wall of China” and “2008\2008-04-19…” are hidden in the cluster “Resource 2 items”).
*Figure 9: RDF graph with details about resource “Mária Bieliková”*
In the RDF graph view users see one node with nine photos, which is connected with the resource node “Mária Bieliková” through the predicate node “Shows”. By zooming onto this node, photos of Mária Bieliková are shown [Figure 12].
*Figure 10: Photos of prof. Bieliková, which were discovered via RDF graph grouped into two clusters – one shows the person Jakub Simko during the event ICupParis, the other shows photos from the place Pittsburgh with people showing facial expressions*
5 Evaluation
Our evaluation of the proposed method of web search results exploration is based on the hypothesis, that interconnection of cluster-based navigation and zoom-based navigation decreases the user effort required to identify desired results and increases understanding of the explored domain. For these reasons we implemented a prototype of the faceted browser Factic described in the case study.
5.1 Prototype and Dataset
The faceted browser Factic works with an ontological repository of multimedia content (digital images and videos), which is based on Kanzaki Exif data description vocabulary ([see 6]). The ontology contains about 11,000 digital photos with both manually and automatically acquired annotations, including authors, tags, locations, topics and EXIF metadata. In total, the dataset consists of about 37 classes, 58 predicates and more than 42,000 individuals and 150,000 facts.
The server side services provide access to the ontology via a Sesame endpoint and SPARQL and SerQL queries [see Figure 13], while the Data access layer converts results to objects and caches often accessed objects. The faceted search engine generates facets from the ontology and builds SPARQL queries from faceted queries, a clustering service performs incremental clustering of results.
We implemented the client side of the prototype as a Silverlight application which has two main parts: Faceted query builder and Result explorer. Faceted query builder visualizes facets to users and builds faceted queries. Results explorer visualizes results in Cluster or Graph view and loads and caches data from server.
5.2 Experiment Description
We performed the experiment with ten volunteers that had different computer skills. These volunteers were between 20 to 53 years old, while their age average was 30 years. We selected volunteers with the goal to cover web users with different jobs (e.g., student of economics, accountant, professor of automation). The experiment had the form of a qualitative test and was performed at volunteers’ home computers to minimize influence of an unknown environment on experiment results.
Each volunteer received a brief instruction about the basic functionality of our tool for results exploration. During the experiment, volunteers were working with a questionnaire which contained two tasks, whereby the user’s goal was to find specific information. Questions were focused on the prototype’s usability, its graphical user interface, provided functionality and quality of clusters (facets as search results categories). The expected answers to these questions were values from 1 to 5, 1=best, 5=worst, and a short commentary.
The main goal of the first task was the familiarization of volunteers with the faceted browser Factic. This task contained small subtasks (e.g., find photos with [tag] foals and then filter out photos with a [tag] mare) with questions (e.g., How many photos have you found?), which helped users to understand the controls and to
[6] http://www.kanzaki.com/ns/exif
reach their goal. Hereby we have acquired feedback about the proposed style of results exploration and the prototype’s usability for inexperienced users.
The second task was focused on collecting feedback on the proposed method’s usability for reaching unclear goals, did not provide any guidance and only contained a poor description of the desired results (e.g., How many photos show running foxes on a snowy parcel without trees?). Descriptions of these goals do not contain identified tags or facets. Users had to identify features and types of desired photos in context of the given dataset and create their own strategy to solve the task.
**Figure 11: Architecture of the prototype**
### 5.3 Results
During the experiment, we have collected logs about the volunteers’ activities. We have compared the collected logs with reference logs obtained by simulation of the most probable scenario of the tasks’ solutions [see Table 2]. The first task has straightly described steps, thus we used the minimal number of necessary actions as the reference value. Multiple possible strategies exist for the second task, which together with the exploratory nature of the task exclude the use of the shortest strategy as a suitable reference value. We thus evaluated several strategies and selected the most assumed strategy (the strategy with short exploratory steps followed with direct steps to the goal).
During the first task, volunteers have performed on average about 76.88% more actions than the reference value. This difference was caused by the users’ exploration of our tool’s features. We have obtained better and more substantial results during the second task. To find the solution of the second task, participants have needed only about 22.5% more actions on average than the reference value. We consider these results as really good, because we selected a group of people who did not know the
used dataset and they had to explore and understand it on the fly. In addition, from users’ actions and comments we conclude, that all unnecessary user actions had exploratory background and that these actions did not lead to loss of users, but they helped users understand the domain and to fulfill users’ sub-goals. Thus we conclude that the proposed visualization and navigation method is easily understandable and useable for real world users even in previously unknown information domains.
Based on user answers and the analysis of users’ behavior we determine that ([see Table 3]):
| | Actions in the 1st task | Actions in 2nd the task |
|----------------------|------------------------|-------------------------|
| Minimal value | 18 | 5 |
| Maximal value | 43 | 27 |
| Average value | 28.3 | 14.7 |
| Standard deviation | 7.9 | 6.4 |
| Reference value | 16 | 12 |
| Difference | +76.88% | +22.5% |
Table 2: The number of actions performed by volunteers during the experiment
- users did not have any problems with interaction with the implemented tool (grade $2.2 \pm 0.63$), but it was less understandable for participants that the attribute marking checkbox [see Figure 6] had three states (wanted, unwanted and allowed) and most of participants used only unwanted and allowed state (participants simply replaced the wanted state by double clicking on attribute nodes);
- graph navigation has been reasonably natural for them (grade $1.6 \pm 0.84$) and participants correctly predicted consequences of their actions;
- graph visualization is readable and clear for users (grade $1.8 \pm 0.79$) – participants were able to correctly describe how they understand to graphs;
- result clusters were correct (grade $1.2 \pm 0.42$), but participants noticed that some results had wrong classification to clusters. This was caused by errors in the dataset and by unbalanced descriptions of photos in their details.
In addition, we noticed that participants naturally filtered out nodes with a distance above two edges from their actual center of interest. This could be utilized in graph visualization by automatic hiding of outlying nodes thus increasing the efficiency of user navigation.
6 Conclusions
We proposed a novel approach for web search results exploration and navigation, in which we combine several search results overview and individual results exploration techniques. We take advantage of user familiarity with zooming and seamlessly transition between our views based on different levels of search results details either via zooming onto results or zooming out to see the big picture.
This natural interconnection of different views gives us opportunity to support all users’ search activities and to join them into one environment with homogenous graphics and control, thus making information seeking more comfortable and attractive for end users.
We also proposed a lightweight user model and next action recommendation method as a part of the Semantic Web exploration approach described in our case study. These methods are directly focused on graph-based navigation and suitable for experiments, but they do not make provisions for global user interests and actions. For real world implementations several existing approaches based on globally collected data [Barla 11] can be employed.
| Volunteer | Easiness of interactivity | Usability | GUI | Quality of clusters | Graph navigation | Clarity of a graph | Readability of a graph | Overall satisfaction |
|-----------|--------------------------|-----------|-----|---------------------|------------------|--------------------|------------------------|----------------------|
| 1 | 2 | 3 | 2 | 1 | 1 | 2 | 1 | 1 |
| 2 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| 3 | 2 | 4 | 2 | 1 | 1 | 3 | 1 | 1 |
| 4 | 2 | 3 | 2 | 1 | 1 | 1 | 1 | 2 |
| 5 | 2 | 1 | 2 | 1 | 1 | 2 | 1 | 2 |
| 6 | 3 | 1 | 3 | 2 | 3 | 3 | 1 | 3 |
| 7 | 3 | 2 | 2 | 1 | 2 | 2 | 2 | 2 |
| 8 | 3 | 2 | 2 | 2 | 1 | 1 | 2 | 2 |
| 9 | 2 | 2 | 2 | 1 | 1 | 1 | 1 | 1 |
| 10 | 2 | 2 | 2 | 1 | 2 | 2 | 1 | 2 |
| Min. | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| Max. | 3 | 4 | 3 | 2 | 3 | 3 | 2 | 3 |
| Average | 2.2 | 2.1 | 2 | 1.2 | 1.6 | 1.8 | 1.2 | 1.7 |
| Std. dev. | 0.63 | 0.99 | 0.47| 0.42 | 0.84 | 0.79 | 0.42 | 0.67 |
Table 3: Volunteers’ ratings of Factic obtained from the questionnaire
We realized a prototype of the proposed approach as an exploratory image browser for a multimedia ontology which we used for automatic facet generation, hierarchical results clustering, multiple graph view generation and personalization based on user’s graph interactions.
We evaluated our approach in several experiments with the exploratory image browser Factic with real world users, and discovered these findings based on user activity logs and observations:
- The interconnection of different views via zooming is natural for users and users quickly mastered controlling the prototype via zooming – users naturally perceive the zoom in action as looking on an item in more detail (as looking via a larger magnifying glass);
- Real world users are able to effectively use the graph visualization – correct answers of the experiment participants and their comments confirm that users
can understand the graph, they compare the results in the graph and they are also able to use the graph for navigation to desired results;
– Users typically read only nodes whose *distance from the current point of interest is smaller than three*. This observation can be utilized to increase graph clarity by directly hiding more distant nodes;
– Users consider the proposed approach *attractive*. This fact is mostly caused by the more interactive results view, compared to the traditional list-based result view, with which users can interact and make it more understandable.
Our approach is mainly focused on content, which can also be viewed via small thumbnails (e.g., photos, videos) and is thus less useful for purely textual content, whose title, topics and keywords are main representatives. But the main idea of zooming is also applicable in text-based domains. Graph visualization may be used as a map of similar articles, in which users follow their trail of navigation and also explore other possibilities. The main focus stays in text, while zooming into a text area switches the view to clustered articles with similar content to the zoomed on area.
While the evaluation of our approach was performed with a strong-typed ontology, we have not taken advantage of its full semantics. One good candidate for our visualization approach are also regular unstructured web search results. While few of these results are already annotated with semantic annotations (e.g., written in RDFa), the rest can easily be enriched with lightweight semantics sufficient for our approach (e.g., tags, keywords, bag of words). This can be done via techniques of lightweight ontology learning approaches that have already achieved reasonably good results [Wong et al. 12], [Gahrib et al. 12]. We have indirectly proved this in the first task of our experiment in which users manipulated only tags assigned to images. In the experiment, we assumed that if users are only given one type of properties (relations), it will be easier to understand the new form of navigation. We have not observed any problems with our approach and users have not had any problems in locating the expected information during the experiment.
Observations from the evaluation and results of research in ontology learning and matching [Martinez-Gil et al. 12] indicate that our approach can be used as an easily learnable form of web search result exploration. This approach based on visualization of clusters and graphs can also be attractive for modern touch-based devices that can utilize it for exploring web search results, users’ content such as photos, videos or other content that can be viewed via thumbnails.
**Acknowledgements**
This work was partially supported by the grants Scientific Grant Agency of Slovak Republic, grant No. VG1/0675/11 and the Slovak Research and Development Agency under the contract No. APVV-0208-10 and by the Ministry of Education, Science, Research and Sport of the Slovak Republic within the Research and Development Operational Programme for the project “University Science Park of STU Bratislava”, ITMS 26240220084, co-funded by the European Regional Development Fund.
References
[Ahmed et al. 09] Ahmed, S.; Park, S.; Jung, J.J.; Kang, S.: “A Personalized URL Re-ranking Method using Psychological User Browsing Characteristics”; Journal of Universal Computer Science, 15, 4 (2009), 926-940.
[Barla 11] Barla, M.: “Towards Social-based User Modeling and Personalization”; Inf. Scien. and Technologies Bul. of the ACM Slovakia, 3, 1 (2011), 52-60.
[Berners-lee et al. 06] Berners-lee, T.; Chen, Y.; Chilton, L.; Connolly, D.; Dhanaraj, R.; Hollenbach, J.; Lerer, A.; Sheets, D.: “Tabulator: Exploring and Analyzing Linked Data on the Semantic Web”; Proc. 3rd Int. Sem. Web User Interaction Workshop, (2006), 6.
[Boldi et al. 08] Boldi, P.; Bonchi, F.; Castillo, C.; Donato, D.; Gionis, A.; Vigna, S.: “The Query-Flow Graph: Model and Applications”; Proc. 17th ACM Conf. on Inf. and Knowledge Mining - CIKM’08, ACM Press, New York (2008), 609-618.
[Boldi et al. 09] Boldi, P.; Bonchi, F.; Castillo, C.; Donato, D.; Vigna, S.: “Query Suggestions Using Query-Flow Graphs”; Proc. 2009 Workshop on Web Search Click Data - WSCD’09, ACM Press, New York (2009), 56-63.
[Bordino et al. 10] Bordino, I.; Castillo, C.; Donato, D.; Gionis, A.: “Query Similarity by Projecting the Query-Flow Graph”; Proc. 33rd Int. ACM SIGIR Conf. on Res. and Dev. in Inf. Ret. - SIGIR’10, ACM Press, New York (2010), 515-522.
[Bordogna and Pasi 09] Bordogna, G.; Pasi, G.: “Hierarchical-Hyperspherical Divisive Fuzzy C-Means (H2D-FCM) Clustering for Information Retrieval”; Proc. 2009 IEEE/WIC/ACM Int. Joint Conf. on Web Intelligence and Intelligent Agent Technology, IEEE, Washington, DC (2009), 614-621.
[Bozzon et al. 11] Bozzon, A.; Brambilla, M.; Ceri, S.; Fraternali, P.; Vadacca, S.: “Exploratory Search in Multi-domain Information Spaces with Liquid Query”; Proc. 20th Int. Conf. Companion on World Wide Web - WWW’11, ACM Press, New York (2011), 189-192.
[Broder 02] Broder, A.: “A Taxonomy of Web Search”; ACM SIGIR Forum, 36, 2 (2002), 3-10.
[Carpineto and Romano 12] Carpineto, C.; Romano, G.: “A Survey of Automatic Query Expansion in Information Retrieval”; ACM Computing Surveys, 44, 1 (2012), 1-50.
[Dörk et al. 08] Dörk, M.; Carpendale, S.; Collins, C.; Williamson, C.: “VisGets: Coordinated Visualizations for Web-Based Information Exploration and Discovery”; IEEE Trans. on Visualization and Computer Graphics, 14, 6 (2008), 1205-1212.
[Fagan 10] Fagan, J.C.: “Usability Studies of Faceted Browsing: A Literature Review”; Information Technology and Libraries (ITAL), 29, 2 (2010), 58-66.
[Gahrib et al. 12] Gharib, T.F.; Badr, N.; Haridy, S.; Abraham, A.: “Enriching Ontology Concepts Based on Texts from WWW and Corpus”; Journal of Universal Computer Science, 18, 16 (2012), 2234-2251.
[Hazman et al. 11] Hazman, M.R.; El-Beltagy, S.; Rafea, A.: “A Survey of Ontology Learning Approaches”; Int. Jour. of Computer Applications, 22, 9 (2011), 36-43.
[Herman et al. 00] Herman, I.; Melancon, G.; Marshall, M.S.: “Graph Visualization and Navigation in Information Visualization: A survey”; IEEE Transactions on Visualization and Computer Graphics, 6, 1 (2000), 24-43.
[Jansen et al. 00] Jansen, B.J.; Spink, A.; Saracevic, T.: “Real Life, Real Users, and Real Needs: a Study and Analysis of User Queries on the Web”; Information Processing & Management, 36, 2 (2000), 207-227.
[Jing et al. 06] Jing, F.; Wang, C.; Yao, Y.; Deng, K.; Zhang, L.; Ma, W.: “IGroup: Web Image Search Results Clustering”; Proc. 14th Annual ACM Int. Conf. on Multimedia - MULTIMEDIA’06, ACM Press, New York (2006), 377-384.
[Lanzenberger et al. 10] Lanzenberger, M.; Sampson, J.; Rester, M.: “Ontology Visualization: Tools and Techniques for Visual Representation of Semi-Structured Meta-Data”; Journal of Universal Computer Science, 16, 7 (2010), 1036-1054.
[Levene 10] Levene, M.: “An Introduction to Search Engines and Web Navigation”; Wiley-Blackwell, John Wiley & Sons, Inc., New Jersey (2010).
[Miller 56] Miller, G.A.: “The magical number seven, plus or minus two: some limits on our capacity for processing information”; Psych. Review, 63 (1956), 81-97.
[Marchionini 06] Marchionini, G.: “Exploratory Search: From Finding to Understanding”; Communications of the ACM, 49, 4 (2006), 41-46.
[Martinez-Gil et al. 12] Martinez-Gil, J.; Navas-Delgado, I.; Aldana-Montes, J. F.: “MaF: An Ontology Matching Framework” ; Journal of Universal Computer Science, 18, 2 (2012) 194-217.
[Nielsen 00] Nielsen, J.: “Designing web usability: the practice of simplicity”; New Riders Publishing, Indianapolis (2000).
[Rástočný et al. 11] Rástočný, K.; Tvarožek, M.; Bielikova, M.: “Supporting Search Result Browsing and Exploration via Cluster-Based Views and Zoom-Based Navigation”; Proc. IEEE/WIC/ACM Int. Conf. on Web Intelligence and Intelligent Agent Technology (WI-IAT), IEEE CS, Washington (2011), 297-300.
[schraefel et al. 06] schraefel, m. c.; Wilson, M.; Russell, A.; Smith, D.A.: “mSpace: Improving Information Access to Multimedia Domains with MultiModal Exploratory Search”; Communications of the ACM, 49, 4 (2006), 47-49.
[Stefaner et al. 08] Stefaner, M.; Urban, T.; Seefelder, M.: “Elastic Lists for Facet Browsing and Resource Analysis in the Enterprise.” Proc. 19th Int. Conf. on Database and Expert Sys. Applications, IEEE, Washington, DC (2008), 397-401.
[Teufl and Lackner 11] Teufl, P.; Lackner, G.: “Knowledge Extraction from RDF Data with Activation Patterns”; Journal of Universal Computer Science, 17, 7 (2011) 983-1004.
[Tvarožek 11] Tvarožek, M.: “Exploratory Search in the Adaptive Social Semantic Web”; Inf. Scien. and Technologies Bul. of the ACM Slovakia, 3, 1 (2011) 42-51.
[Tvarožek and Bielikova 10] Tvarožek, M.; Bielikova, M.: “Factic: Personalized Exploratory Search in the Semantic Web”; Proc. Web Engineering, 10th Int. Conf., ICWE 2010, Springer, Berlin (2010), 527-530.
[van Ham and Rogowitz 08] van Ham, F.; Rogowitz, B.E.: “Perceptual Organization in User-Generated Graph Layouts”; IEEE Trans. on Visualization and Computer Graphics, 14, 6 (2008), 1333-1339.
[Wong et al. 12] Wong, W.; Liu, W.; Bennamoun, M.: “Ontology learning from text”; ACM Computing Surveys, 44, 4 (2012), 1–36. |
Origins of $k \cdot p$ errors for [001] GaAs/AlAs heterostructures
This article has been downloaded from IOPscience. Please scroll down to see the full text article.
1996 Europhys. Lett. 33 383
(http://iopscience.iop.org/0295-5075/33/5/383)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 126.96.36.199
The article was downloaded on 23/08/2013 at 07:53
Please note that terms and conditions apply.
Origins of $k \cdot p$ errors for [001] GaAs/AlAs heterostructures
D. M. Wood\textsuperscript{1}(*), A. Zunger\textsuperscript{1}(**) and D. Gershoni\textsuperscript{2}
\textsuperscript{1} National Renewable Energy Laboratory - Golden, CO 80401, USA
\textsuperscript{2} Department of Physics, Technion-Israel Institute of Technology - Haifa, 32000, Israel
(received 5 September 1995; accepted in final form 3 January 1996)
PACS. 73.20Dx – Electron states in low-dimensional structures (including quantum wells, superlattices, layer structures, and intercalation compounds).
PACS. 71.10+x – General theories and computational techniques (including many-body perturbation theory, density-functional theory, atomic sphere approximation methods, Fourier decomposition methods, etc.).
Abstract. The $k \cdot p$ method + envelope function combination used for semiconductor heterostructures is based on approximations dubious under some conditions. We directly compare 8-band $k \cdot p$ with pseudopotential results for [001] GaAs/AlAs superlattices and quantum wells with $k \cdot p$ input parameters directly computed from bulk GaAs and AlAs pseudopotential bands. We find generally i) quantitative agreement for the lowest hole states within 10 meV of the GaAs valence band maximum, but ii) systematic errors deeper in the valence band and ii) qualitative errors for even the lowest conduction bands with appreciable contributions from off-$L$ zinc-blende states. We trace these errors to inadequate $k \cdot p$ description of \textit{bulk} GaAs and AlAs band dispersion away from the zone center.
Nanostructures $\gtrsim 100 \text{\AA}$ in size were until recently [1] beyond reach of the atomistic electronic structure methods used for bulk crystals, \textit{i.e.} direct solution of the Schrödinger equation
$$\left[ -\frac{\hbar^2}{2m} \nabla^2 + \sum_{i:R_i} v_i(\mathbf{r} - \mathbf{R}_i) \right] \psi(\mathbf{r}) = \varepsilon \ \psi(\mathbf{r}), \quad (1)$$
with the crystal potential $V(\mathbf{r})$ here written as a superposition of screened atomic pseudopotentials $v_i$ for atom species $i$. The spectroscopy of A/B heterostructures was instead interpreted [2] using an approach so common we term it the ‘standard model’ (SM): the $k \cdot p$ method combined with the envelope function approximation (EFA). Although the SM has been eminently successful [3], even for ultrathin systems [4], approximations on which it is based compromise its description of heterostructures. Their impact has been partially masked by fitting of its parameters to experimental data, as described below. On general grounds one expects the SM to fail for short-period superlattices but would like to know when (for what thicknesses) and how (for which states) it fails. While detailed analyses of potential
(*) Permanent address: Department of Physics, Colorado School of Mines, Golden, CO 80401 USA; e-mail: email@example.com
(**) E-mail: firstname.lastname@example.org
© Les Editions de Physique
pitfalls of the SM have appeared [5], the SM cannot assess its own validity. Evaluations of actual SM errors via non-SM methods, e.g., tight-binding [6], generally used input data (e.g., effective masses) from different sources and did not necessarily reflect SM deficiencies. Only by comparing direct solutions of the fully atomistic Schrödinger equation as in eq. (1) can such errors be systematically assessed. This approach is free of approximations made in the standard model, includes full-zone dispersion of all bands, and can predict complete Bloch wave functions. We use [901] (AlAs)$_n$(GaAs)$_n$ superlattices for $n \leq 20$ and (GaAs)$_n$/AlAs quantum wells to test the standard model against a direct pseudopotential approach (eq. (1)), with the former’s input parameters computed from the latter to guarantee meaningful comparison. We trace systematic errors in the SM for heterostructures to inadequate description of dispersion of bulk bands in the Brillouin zone region where coupling to off-$\Gamma$ states is important.
For periodic systems the cell-periodic part $u$ of the Bloch function $\psi_{n\mathbf{k}} = \exp[i\mathbf{k} \cdot \mathbf{r}] u_{n\mathbf{k}}(\mathbf{r})$ may be expanded [7] about a reference point $\mathbf{k}_0$:
$$u_{n\mathbf{k}}(\mathbf{r}) = \sum_{n'}^{N} b_{n'} u_{n'\mathbf{k}_0}(\mathbf{r}). \quad (2)$$
Choosing $\mathbf{k}_0 \equiv \Gamma$, the $u_{n\mathbf{k}}(\mathbf{r})$ obey [8]
$$\sum_{n'}^{N} \left\{ \left[ \varepsilon_n(\Gamma) - \varepsilon_{n'}(\mathbf{k}) + \frac{\hbar^2 k^2}{2m} \right] \delta_{n,n'} + \frac{\hbar}{m} \mathbf{k} \cdot \mathbf{p}_{n,n'} \right\} b_{n'} = 0; \quad (3)$$
the effects of the microscopic crystal potential are now encoded in the $\mathbf{p}_{n,n'} \equiv \langle u_{n,\Gamma} | \hat{\mathbf{p}} | u_{n',\Gamma} \rangle$. Diagonalized with large enough $N$, eq. (3) would predict full (non-parabolic) bands throughout the zone, equivalent to direct solution of the Schrödinger equation for Bloch electrons [9]. The standard model [8], instead uses: i) Degenerate perturbation theory: For bulk semiconductors, the three valence states degenerate at $\Gamma$ are usually augmented by the first conduction state. Including spin-orbit defines the 8-band $k \cdot p$ model used below; ii) Fitting: The small $N$ values ($\lesssim 20$) in eq. (3) used in most $k \cdot p$ calculations poorly describe band dispersion away from $\Gamma$ if the $\{\mathbf{p}_{n,n'}\}$ are actually computed from exact Bloch states $u_{n,\Gamma}(\mathbf{r})$. The SM uses instead ‘effective’ matrix elements found [10] from measured gaps and band effective masses at $\Gamma$, mitigating the errors of i); iii) Envelope functions: To treat A/B heterostructures, the standard model generally uses the Luttinger-Kohn formalism [7] for response of a homogeneous crystal to a weak, slowly-varying external perturbation. The wave function in, e.g., material A, takes the form
$$\psi(\mathbf{r}) = \sum_{n=1}^{N} F_n^A(\mathbf{r}) u_{n,\Gamma}^A(\mathbf{r}), \quad (4)$$
where the $F_n^A(\mathbf{r})$ are ‘envelope functions’. The $\{u_{n,\Gamma}\}$ formally differ in A and B but virtually all EFA’s assume the same set in both except insofar as they affect boundary conditions, via materials properties (e.g., gaps) which differ in A and B. Winkler and Rössler [11] have developed an alternate approach for solving the multiband problem.
Direct solution of eq. (1) makes unnecessary all the approximations above. To compare on an equal footing this ‘direct’ approach with the ‘standard model’, 1) We use recent [12] empirical pseudopotentials $v_i(\mathbf{r})$ (including spin-orbit) in eq. (1) to compute band structures for bulk zinc-blende (ZB) GaAs and AlAs; the small GaAs/AlAs lattice mismatch is neglected. These pseudopotentials fit to measured band structures of GaAs and AlAs (so that LDA errors do not appear), also closely reproduce important symmetry-related trends in short-period superlattices. 2) We then equate numerically computed effective masses to their formal
Fig. 1. — AlAs and GaAs near-edge bands between $\Gamma$ and $X$ ([001] direction). Relativistic labels at $\Gamma$ and $X$ and conventional names for valence bands are used. Gray circles show points which for [001] (AlAs)$_5$(GaAs)$_5$ fold to the SL zone center.
expressions [10] in the $8 \times 8$ Kane model to extract [13] Luttinger parameters and all relevant matrix elements of eq. (3). 3) We use these as input to $k \cdot p +$ EFA calculations for [001] GaAs/AlAs superlattices and quantum wells and compare results with direct solutions of eq. (1) for the same structures. The direct ‘all band pseudopotential’ (ABP) calculations use a conjugate gradient program with a plane wave basis and a 5 Ry kinetic energy cutoff. ‘Standard model’ (SM) calculations use the $8 \times 8$ $k \cdot p +$ EFA formulation of Baraff and Gershoni [10]. Envelope functions are expanded in 75 Fourier components along the unit cell.
The procedure above permits clean evaluation of approximations made in the ‘standard model’. Figure 1 contrasts pseudopotential and $k \cdot p$ bands for bulk AlAs and GaAs. For GaAs, $k \cdot p$ and pseudopotential bands agree to within 50 meV only up to 12%, 18%, 14%, and 14% of the distance toward $X$ for the electron (CB1) heavy hole (hh), light hole (lh), and split-off (s-o) bands, respectively. The SM GaAs $X_{6c}$ conduction state is 26 eV higher than the direct pseudopotential value. This gross error in the bulk is important for heterostructures since zinc-blende $X$ states fold to the zone center and interact with zinc-blende $\Gamma$-derived states. Since the (pseudopotential) GaAs $X_{6c} - \Gamma_{6c}$ conduction band splitting is only 0.4 eV this interaction is strong, but in the SM is unphysically negligible because the $X_{6c}$ state is 26 eV too high. Superlattice conduction bands in the SM will thus always be spuriously
Fig. 2. – Band energies for [001] (AlAs)$_n$(GaAs)$_n$ superlattices; $n_c$ indicates point of transition from indirect to direct gap system. Overbars indicate SL states which derive mainly from the ZB state in parentheses. Bulk ABP AlAs and GaAs band energies are given at right; the energy zero for a) and b) are bulk GaAs $\Gamma_{6c}$ and $\Gamma_{8v}$ states, respectively. Dashed lines show band connectivity near crossings, since $\Gamma$-$X$ mixing is practically omitted.
$\Gamma$-like [14]. While the SM $X_T$ GaAs valence state is almost 10 eV too low with respect to the pseudopotential value, the resulting error for heterostructure states is small because the pseudopotential $X_T - \Gamma_{8v}$ valence band splitting is large (2.4 eV), so that interaction between heterostructure $\Gamma$- and $X$-derived valence states which fold to the heterostructure zone center is relatively weak.
Figure 2 compares zone-center ABP and SM band energies for [001] (AlAs)$_n$(GaAs)$_n$ superlattices (SL) as a function of $n$. We label SL states via an overbar, with the ZB Brillouin zone point from which they derive in parentheses. $\overline{\Gamma}(\Gamma)$ states derive principally from ZB $\Gamma$ states, while $\overline{\Gamma}(X_z)$ states derive mostly from folded-in zinc-blende $X_z$ states. Only the lowest $\overline{\Gamma}(\Gamma)$ and $\overline{\Gamma}(X_z)$ conduction and near-edge valence bands are shown. The extremely high energy of the SM bulk GaAs $X_{6c}$ state (fig. 1) has important consequences: i) several additional folded-in $\overline{\Gamma}(\Delta_{6c})$ ABP conduction states (not shown) in this energy window are completely missing in the SM; ii) the SM thus misses the transition evident in ABP results (circle in a)) from $\overline{\Gamma}(X_z)$ to $\overline{\Gamma}(\Gamma_{6c})$ as the lowest conduction band; iii) non-monotonicity in the ABP $\overline{\Gamma}(\Gamma)$ conduction band for small $n$ (also present for other points in the SL zone [13] and in first-principles calculations [15]) is absent in the SM, which iv) also overestimates its energy (shading in panel a)). For valence bands, i) for states with a binding energy $\lesssim 200$ meV, SM and ABP agree very well, though ii) deeper into the valence band, SM curves reproduce
ABP trends but place them *too deep* in energy; iii) for systems lacking inversion symmetry, lifting of the spin degeneracy away from the zone center is permitted in some directions. This spin splitting —absent in the SM— is significant ($\gtrsim 30$ meV for the first heavy-hole state for $\frac{q_{z,a}}{2\pi} > 0.1$) for ‘in-plane’ dispersion in ABP calculations. SM band dispersion (not shown) agrees with ABP results only relatively near the SL zone center [13]. For $n = 5$, ABP values of $m_\downarrow/m_\uparrow$ at $\overline{T}$ are, *e.g.*, $\simeq 3.4$ for the hh1 state and $\simeq 0.95$ for the $\overline{T}(I_{bc})$ electron state, while SM values are 4.4 and 1.3, respectively; the anisotropy of effective masses is thus exaggerated within the SM. (GaAs)$_n$/AlAs ($1 \leq n \leq 10$) quantum wells show [13] $n_c \simeq 9$ (cf. fig. 2), differences in valence band dispersion, and $1/n^2$ $\overline{T}(\Gamma)$ conduction band behavior to smaller $n$ than for superlattices (fig. 2).
Figure 3 contrasts square moduli of ABP wave functions (full lines) averaged over transverse dimensions of the primitive cell to facilitate comparisons with SM envelope functions (dashed lines), for the $\overline{T}(\Gamma)$ electron state and the third hole state at $\overline{T}$. Envelope functions for states whose energy (fig. 2) is well described by the SM closely average inner and outer envelopes of Bloch states. The $n = 5$ $\overline{T}(\Gamma)$ ABP electron wave function shows interfacial peaks absent in the SM.
Projections of SL states onto zinc-blende states provide insight into why and where the SM fails. [001] (AlAs)$_n$(GaAs)$_n$ superlattice states at $\overline{T}$ derive from ZB states at the SL reciprocal lattice vectors $G_j = \frac{n-1}{n\ell}$ for $j = 0, 1, 2, \ldots, n$ along the [001] $\Gamma$-$X$ ($\Delta$) line. We may thus expand a $\overline{T}$ SL state in a complete set of ZB Bloch states at these $G_j$:
$$|\psi_{\overline{T}}^{\text{SL}}\rangle = \sum_s^\infty \sum_{G_j=\Gamma}^X \alpha_{s,G_j} |\psi_{s,G_j}^{\text{ZB}}\rangle;$$
(5)
the $G_j \neq 0$ (gray circles for (AlAs)$_5$(GaAs)$_5$ in fig. 1) fold to $\overline{T}$ in the SL geometry. The projection of a specified SL state onto zinc-blende band $s$ at $G_j$ is thus $P_{s,G_j} \equiv |\langle \psi_{\overline{T}}^{\text{SL}} | \psi_{s,G_j}^{\text{ZB}} \rangle|^2 =$
Table I — Projections $P_{s,G_j}$ of ABP near-edge (GaAs)$_5$(AlAs)$_5$ states at $\overline{\Gamma}$ ('State') onto GaAs band $s$ labeled 'On' at the $G_j = \frac{G_j}{a}$, and net projection $P_s$ on band $s$. $G_0$ and $G_5$ correspond to ZB $\Gamma$ and $X$ points, respectively.
| State | On | $G_0$ | $G_1$ | $G_2$ | $G_3$ | $G_4$ | $G_5$ | $P_s$ |
|-----------|-----|-------|-------|-------|-------|-------|-------|------|
| $\overline{\Gamma}(\Gamma)$ | CB1 | 0.80 | 0.08 | 0 | 0.05 | 0.02 | 0 | 0.94 |
| $\overline{\Gamma}(X_5)$ | CB1 | 0 | 0 | 0.00 | 0.00 | 0.03 | 0.95 | 0.99 |
| 'hh1' | hh | 0.84 | 0.15 | 0.00 | 0.00 | 0 | 0 | 1.00 |
| 'lh1' | lh | 0.93 | 0.06 | 0 | 0 | 0 | 0 | 0.98 |
| 's-o1' | s-o | 0.80 | 0.01 | 0 | 0 | 0 | 0 | 0.81 |
| 'hh2-a' | hh | 0 | 0.89 | 0.04 | 0 | 0 | 0 | 0.93 |
| 'hh2-b' | hh | 0.14 | 0.81 | 0.02 | 0.00 | 0 | 0 | 0.97 |
| 'lh2-a' | lh | 0 | 0.93 | 0.02 | 0 | 0 | 0 | 0.95 |
| 'lh2-b' | lh | 0.04 | 0.78 | 0.01 | 0 | 0 | 0 | 0.83 |
$= |\alpha_{s,G_j}|^2$. The net contribution $P_s \equiv \sum_{j=0}^{n} P_{s,G_j}$ measures how completely the SL state derives from ZB band $s$. Similarly, the quantity $P \equiv \sum_{s=1}^{N_b} P_s$ measures how nearly the finite set of $N_b$ zinc-blende bands used is 'complete' ($P \equiv 1$), i.e. adequately describes the specified SL state.
For near-edge ABP (AlAs)$_5$(GaAs)$_5$ superlattice states table I shows the projections $P_{s,G_j}$ onto the (spin-split) first conduction band CB1 and hh, lh, and s-o valence bands (1)—the same set used by the $8 \times 8$ Kane model. We see that i) superlattice hh1, lh1, and s-o1 states derive mostly from the ZB $\Gamma$ point ($G_0$), hh2 and lh2 states mostly from $G_1$, etc. ii) hh1, lh1, and s-o1 SL states, for example, have significant but monotonically decreasing contributions from $G_j \neq 0$, since GaAs valence bands disperse monotonically from $\Gamma$ (fig. 1b); iii) the $\overline{\Gamma}(\Gamma)$ conduction state has only a $\approx 80\%$ projection on the GaAs $\Gamma$-point ($G_0$), with contributions $\approx 20\%$ as large from $G_j \neq G_5$ which depend non-monotonically on $j$, reflecting the dispersion. (fig. 1) of the GaAs CB1 band. We have also evaluated (not shown) $P_s$, describing mixing of different zinc-blende bands in a SL state, and $P$, measuring basis set completeness. For the two SL conduction states in table I the CB1 and CB2 bands of fig. 1 are quasi-complete, i.e. $1 - P < 0.004$. Only the $\overline{\Gamma}(\Gamma)$ state—ostensibly described by the SM—has an appreciable ($P_s \approx 5.3\%$) contribution from a state ($s=$CB2) outside the set used in the 8-band $k \cdot p$ approach. For SL valence states there is appreciable mixing of ZB states only for s-o1 and deeper valence states, and the hh, lh, s-o set is also quasi-complete.
This shows that provided their dispersion is accurately described (as for ABP bands in fig. 1), a small number of near-edge zinc-blende bands suffice, in terms of projections, to quantitatively describe SL states. If so, why is the SM unsatisfactory for SL conduction and deep hole bands (figs. 2, 3)? Superlattice hh1, lh1, and s-o1 hole bands, which derive mostly from the zinc-blende $G_0$ ($\equiv \Gamma$) point, will be well described by the standard model since $k \cdot p$ bands are fit there (fig. 1). Deeper SL hole bands (lh2, lh2, etc.) derive mostly (table I) from $G_k \neq \Gamma$ points outside the quadratic region of good ABP/SM agreement (fig. 1) and will be found too deep in the valence band (fig. 2). Since the quadratic region is largest for the GaAs hh band, however, the SM tracks the ABP hh2 band quite well for most $n$ in fig. 2. For the $\overline{\Gamma}(\Gamma)$ conduction band, however, contributions from $G_j \neq \Gamma$ (where SM values are too
(1) Due to the great similarity of GaAs and AlAs Bloch states, projections would be very similar had AlAs states been used.
high in energy: fig. 1) are important (table I) and SM predictions must be too high (fig. 2) until the shortest $G_j$ move into the CB1 quadratic region. As the period $a$ increases, points along the zinc-blende $\Delta$ direction which fold to $\Gamma$ move into the region where $k \cdot p$ adequately represents the bulk zinc-blende band structures and all near-edge superlattice states will be well described by the ‘standard model’.
We have thus traced errors in the $k \cdot p +$ EFA approach to poor $k \cdot p$ description of dispersion of bulk bands which are mixed in heterostructure states. The 8-band $k \cdot p$ approach correctly focuses on four spin-split bands, but fails to keep enough ($N$ in eq. (3)) zone-center states to adequately describe their dispersion for thin heterostructures. For Ga$_{0.8}$In$_{1-x}$P ordered alloys [16] zinc-blende states along the [111] $\Gamma - L$ direction fold and couple, so the SMs inadequate description of the bulk $L$ point will cause errors similar to those for [001] superlattices. The central issue is not the heterostructure thickness per se, but whether off-$\Gamma$ bulk states poorly described by the ‘standard model’ are significantly mixed (as determined by the proximity in energy of bulk zinc-blende $L$, $\Gamma$, and $X$ states) in heterostructure bands.
***
We thank K. Mäder and R. Maurer for useful discussions and suggestions, A. Franceschetti and L.-W. Wang, and S. Froyen for suggestions about heterostructure state projections and the conjugate gradient program. This work was supported by the Office of Energy Research, Materials Science Division, US DOE grant DE-AC36-83CH10093, and the Israel Science Foundation, administered by the Israel Academy of Sciences and Humanities.
REFERENCES
[1] Wang L.-W. and Zunger A., J. Phys. Chem., 98 (1994) 2158; Phys. Rev. Lett., 73 (1994) 1039; J. Chem. Phys., 100 (1994) 2394; Mäder K. M., Wang L.-W. and Zunger A., Phys. Rev. Lett., 74 (1995) 2555.
[2] Fasol G., Fasolino A. and Lugli P. (Editors), Spectroscopy of Semiconductor Microstructures, NATO ASI Ser. B., Vol. 206 (Plenum Press, New York, N.Y.) 1989.
[3] See, e.g., Bastard G., Brum J. A. and Ferreira R., in Solid State Physics, edited by D. Turnbull and H. Ehrenreich, Vol. 44 (Academic, New York, N.Y.) 1991, p. 229.
[4] See, e.g., Alonso M.-L., Ilg M. and Ploog K. H., Phys. Rev. B, 50 (1994) 1628.
[5] Burt M. G., J. Phys. Condens. Matter, 4 (1992) 6651.
[6] Smith D. L. and Mailhiot C., Rev. Mod. Phys., 62 (1990) 173.
[7] Luttinger J. M. and Kohn W., Phys. Rev., 97 (1955) 869.
[8] Kane E. O., in Handbook on Semiconductors, edited by T. S. Moss, vol. 1 (North Holland, Amsterdam) 1982 Chapt. 4A, pp. 193-217.
[9] Cardona M. and Pollak F. H., Phys. Rev., 142 (1966) 530; Cardona M., Christensen N. E. and Fasol G., Phys. Rev. B, 38 (1988) 1806.
[10] Baraff G. A. and Gershoni D., Phys. Rev. B, 43 (1991) 4011. See also Gershoni D., Henry C. H. and Baraff G. A., IEEE J. Quantum Electron., 29 (1993) 2433.
[11] Winkler R. W. and Rössler U., Phys. Rev. B, 48 (1993) 8918.
[12] Mäder K. A. and Zunger A., Phys. Rev. B, 50 (1995) 17393.
[13] Wood D. M. and Zunger A., Phys. Rev. B, 53, 15 March 1996.
[14] Edwards G. and Inkson J. C., Solid State Commun., 89 (1994) 595 review $\Gamma$-$X$ mixing in heterostructures and other deficiencies of the EFA.
[15] Wei S.-H. and Zunger A., J. Appl. Phys., 63 (1988) 5794.
[16] Franceschetti A., Wei S.-H. and Zunger A., Phys. Rev. B, 52 (1995) 13992. |
A MODEL FOR PARTICLE LADEN FLOW IN A SPIRAL CONCENTRATOR
Sungyon Lee\textsuperscript{a)}, Yvonne Stokes\textsuperscript{**} & Andrea L. Bertozzi*
\textsuperscript{*}Dept. of Mathematics and Applied Mathematics Laboratory, University of California, Los Angeles, CA 90095
\textsuperscript{**}School of Mathematical Sciences, The University of Adelaide, South Australia, 5005, Australia
\textit{Summary} We apply recent results for gravity driven slurries to the model of a spiral concentrator in the case of a monodisperse particle slurry. We use a thin film approximation to derive an equilibrium profile for the particle concentration and fluid thickness. Our results explain observations seen in laboratory experiments and commercial products.
INTRODUCTION
Spiral concentrators are used in the mining industry to separate particles of different size or density. The existing modeling literature considers the flow as a background fluid carrying non-neutrally buoyant particles. However recent work on modeling of slurries on inclines shows that at relatively modest volume fractions of particles, the presence of the particles affects the flow and, moreover, interparticle interactions such as hindered settling and shear-induced migration can quantitatively explain the dynamics of the separation of particle mixtures under gravity. We incorporate this physics into a model for particle segregation in a spiral concentrator.
We consider a mixture of heavy particles and a viscous liquid flowing down a helical channel of radius $A$ and pitch $2\pi P$, with a rectangular cross section, as shown Figure 1(a). The particle and liquid mass densities obey $\rho_p > \rho_\ell$ and the liquid bulk viscosity is $\mu_\ell$. The volume fraction of particles is denoted by $\phi$. Individual particles as well as the liquid phase are incompressible. All lengths have been scaled by the half width of the channel, so that the channel has width 2. The slope of the channel is $\tan \alpha = P/A = \tau/\epsilon$, where $\epsilon = A/(A^2 + P^2)$ and $\tau = P/(A^2 + P^2)$ are the helix curvature and torsion, respectively. For a spiral concentrator, $\epsilon \sim 0.5$ and $\tau \sim 0.1$, so that we assume $\tau \ll 1$ and, for the coordinate system adopted, $\epsilon < 1$. Furthermore the flow is, typically, of small depth relative to the channel width, characterised by the dimensionless parameter $\delta \ll 1$.
\textbf{Figure 1.} (a) A helical channel of rectangular cross section. (b) Particle concentrations in the scaled $\tilde{z}$ direction ($\tilde{z} = z/h$) for various total concentrations, $\phi_0$, at the inclination angle, $\alpha = 25^\circ$. There is a unique $\phi_c$, denoted by the horizontal line, such that the particles are uniformly distributed in $\tilde{z}$ and the net flux in $y$ is zero. (c) Streamlines in a helical channel of rectangular cross section for parameters $\alpha = 25^\circ$, $\epsilon = 0.5$, $\tau = 0.1$, $\delta = 0.1$, $Q = 0.2$, $Q_p = 0.0015$. The domain is separated into two regions at $y = y_c$: the particle-heavy region $-1 \leq y < y_c$ ($\phi(y, z) = \phi_c$), and the clear fluid region $y_c < y \leq 1$ ($\phi = 0$). We note that the transverse flow shown here is secondary to the dominant downstream flow; however, it plays an important role in particle separation.
EQUILIBRIUM THIN-FILM MODEL
We seek a steady-state solution that is independent of distance along the channel, reducing the problem to a two-dimensional domain. We simplify the Navier-Stokes and particle transport equations by a perturbation expansion in terms of the torsion $\tau$ and $\delta$. Let $x$ be the direction of the primary flow down the channel and $y$ and $z$ be the axes in the channel cross section, with $z$ being directed upwards. In this $(x, y, z)$ coordinate system the (dimensionless) velocity is $(u, v, w)$ and $p$ is pressure. To leading order in $\tau$ and $\delta$ the model is
$$\frac{\partial v}{\partial y} + \frac{\epsilon v}{1 + \epsilon y} + \frac{\partial w}{\partial z} = 0,$$
$$\frac{\partial \sigma}{\partial z} + \rho \sin \alpha = 0, \quad \sigma = \mu(\phi) \frac{\partial u}{\partial z},$$
\textsuperscript{a)} Corresponding author. E-mail: email@example.com
\[-\frac{\partial p}{\partial z} - \rho \cos \alpha = 0, \quad -\frac{\partial p}{\partial y} + \frac{\partial}{\partial z} \left( \mu \frac{\partial v}{\partial z} \right) + \frac{\rho_X}{1 + \epsilon y} u^2 = 0, \quad \chi = \frac{\delta K}{2R}, \tag{1c}\]
\[-\frac{2 \cos \alpha}{9K_{coll}} \rho_s (1 - \phi) - \phi \frac{\partial \sigma}{\partial z} = \left[ 1 + 2 \left( \frac{K_{visc} - K_{coll}}{K_{coll}} \right) \left( \frac{\phi}{\phi_{max} - \phi} \right) \right] \sigma \frac{\partial \phi}{\partial z} = 0, \tag{1d}\]
where we have set $\delta R / F^2 = 1$, and $R = \rho U a / \mu_\ell$ and $F = U / \sqrt{ga}$ are the Reynolds and Froude numbers, respectively; $K = 2 \epsilon R^2$ is the Dean number. The viscosity $\mu$ and density $\rho$ of the fluid-particle mixture depend on the particle volume fraction $\phi$ as follows:
\[\mu(\phi) = \left( 1 - \frac{\phi}{\phi_{max}} \right)^{-2}, \quad \rho = 1 + \rho_s \phi, \quad \rho_s = \frac{\rho_p - \rho_\ell}{\rho_\ell}. \]
The constant $\phi_{max}$ is the maximum possible volume fraction of particles. Boundary conditions are $u = v = w = 0$ on the channel wall $z = 0$ and, on the free surface $z = h(y)$: $\partial u / \partial z = 0$, $\partial v / \partial z = 0$, $p = 0$, and $w = v (dh / dy)$. In the particle-transport equation (1d) we have fluxes due to gravitational sedimentation and shear-induced migration as in [1, 2, 3]. In particular, shear-induced migration includes the effects of particle collisions and gradient in the effective suspension viscosity, $\mu(\phi)$, and it plays an essential role in describing key properties of the suspension flow [1]. Neglect of Brownian motion is justified by a large Péclet number. For further details on the derivation of the flow model see [6, 7] and for the particle transport model see [1, 2].
**RESULTS**
Equations (1b) and (1d) are solved simultaneously at each horizontal position $y$ for a given inclination angle $\alpha$ and total volume fraction of particles $\phi_0(y) = \int_0^h \phi(z, y) dz$. For each value of $y$, (1d) may be written in terms of a new variable $\tilde{z} = z / h(y)$, such that a solution may be found over $0 \leq \tilde{z} \leq 1$ that is independent of $h(y)$. The particle transport problem in this form is exactly the inclined plane problem of [3, 5], and Figure 1(b) shows examples of $\phi(\tilde{z})$ for different values of $\phi_0(y)$. *Note that there is a specific value of $\phi_0(y)$ such that the particle concentration is constant (i.e. $\phi = \phi_c$) in the $\tilde{z}$ direction.* The steady state solution for the particle laden flow requires zero net flux in the $y$ direction, both for fluid and for total fluid-particle mixture: $\int_0^{h(y)} v(y, z) dz = 0$, $\int_0^{h(y)} v(y, z) \phi(y, z) dz = 0$, respectively. Given $v(y, z)$ satisfying the first of these conditions, the only possible $\phi(y, z)$ from the full set of curves (Figure 1(b)) to satisfy the second is the constant case, $\phi = \phi_c$. If $\phi$ is monotone decreasing in $z$ (lower portion of Figure 1(b)), then the net flux in $y$ is negative (towards the center of the spiral). Likewise if $\phi$ is monotone increasing in $z$ (upper portion of Figure 1b), the net flux is positive (away from the center of the spiral). Based on this simple argument, we assume a form of steady state solution where the cross-sectional domain is divided into two distinct regions at a critical position $y = y_c$: region A, $-1 \leq y < y_c$, where the particle concentration is constant everywhere, ($\phi(y, z) = \phi_c$), and region B, $y_c < y \leq 1$ of clear fluid ($\phi = 0$). For given total flux $Q$ (fluid and particles), flux of particles $Q_p$ and channel angle $\alpha$, and requiring the free surface height at $y_c$ to be continuous, we are able to solve for the position $y_c$, free surface $h(y)$, and the flow; the resultant streamlines are shown in Figure 1(c). The clear separation of particle-laden and clear fluid, in the case of a monodisperse slurry flow, has been observed in our own preliminary laboratory experiments and in commercial applications [8].
**CONCLUSIONS**
In summary we propose an equilibrium solution to the particle flow problem in a helical channel, in which the particles all collect near the center of the spiral, consistent with observations of heavy particles in spiral concentrator geometry. In the future, we will model particles of different densities to study the separation of a polydisperse slurry. This will allow us to identify the parameter regime that optimizes the separation in the helical channel, which will be important in the mining and various industry applications.
**References**
[1] Murisic N., Ho J., Hu V., Latterman P., Koch T., Lin K., Mata M., and Bertozzi A.L., Particle-laden viscous thin-film flows on an incline: experiments compared with an equilibrium theory based on shear-induced migration and particle settling, *Physica D*, **240**(20):1661-1673, 2011.
[2] Cook B.P, Bertozzi A.L., and Hosoi A.E., Shock Solutions for Particle-Laden Thin Films, *SIAM J. Appl. Math.*, **68**(3):760-783, 2008.
[3] Cook B.P, Theory for particle settling and shear-induced migration in thin-film liquid flow, *Phys. Rev. E*, **78**:045303, 2008.
[4] Holland-Batt A.B., A method for the prediction of the primary flow on large diameter spiral troughs, *Miner. Eng.*, **22**:352-356, 2009.
[5] Murisic N., Pausader B., Peschka D., and Bertozzi A.L., Dynamics of particle settling and resuspension in viscous liquids, preprint, 2012.
[6] Stokes Y.M., Wilson S.R., Duffy B.R.: Thin-film flow in open helically-wound channels. In M. Behnia, W. Lin, G.D. McBain (eds), *15th Australian Fluid Mechanics Conference (CSROM)*, 2004, Paper AFMC00187.
[7] Stokes Y.M., Duffy B.R., Wilson S.R., and Tronaolone H., Flow down a rectangular helical channel, preprint, 2012.
[8] http://www.alicoco.com/index.php |
Minutes of the regular meeting of the city council of the City of Kingsford, County of Dickinson, and State of Michigan.
A regular meeting of the Kingsford city council was held Monday, September 17, 2018 at 6:30 p.m. in the council room in City Hall.
Roll Call: Present: Councilmember Baldinelli, Dixon-Miller, Groeneveld, Flaminio and Erickson
Absent: None
Also present were City Manager Anthony Edlebeck, City Attorney Bruce Brouillette, Public Safety Director Brian Metras, Holly Palmer, Maggie Lanther [The Daily News], Ken Clawson [Dickinson County News], Mark Shanks and Floyd Lindholm.
The pledge of allegiance was stated.
A motion was made by Councilmember Flaminio and supported by Councilmember Groeneveld to approve the September 4, 2018 City Council Meeting minutes as presented.
ALL AYES, NO NAYS, MOTION CARRIED.
A motion was made by Councilmember Dixon-Miller to approve the invoices for September 5, 2018 through September 17, 2018 in the amount of $270,568.69.
ALL AYES, NO NAYS, MOTION CARRIED.
The consent agenda consisted of the following:
Public Safety Report
August Treasurer’s Report
A motion was made by Councilmember Baldinelli and supported by Councilmember Dixon-Miller to receive and place on file the Public Safety Report and to approve all KPS requests included in the packet.
ALL AYES, NO NAYS, MOTION CARRIED.
There was no Public Comment.
The only item under Communications was a letter dated September 7, 2018 from the County Treasurer re: 125 Birch Street. This property was foreclosed upon by the Dickinson County Treasurer, Lorna Carey, the foreclosing governmental unit (FGU) for Dickinson County. The property was offered at two separate public auctions in the summer and fall of 2017; it went unsold. Ms. Carey is looking for any insight or recommendations from the City regarding demolition of the property in hopes of entertaining offers to purchase the property. She also asks if there would be any interest in cost-sharing of the project. A brief discussion took place.
A motion was made by Councilmember Baldinelli and supported by Councilmember Dixon-Miller to table the issue re: 125 Birch Street at this time to allow for review and additional discussion.
ALL AYES, NO NAYS, MOTION CARRIED.
Copies of all communications are affixed to and made a part of the permanent record.
The Manager’s Report was presented by Edlebeck.
- Bacco Construction Co. is nearing completion of the preliminary work on Woodward Avenue. The current labor dispute lockout has stopped all paving work in the State. An update on this project will be provided once available.
- The pavilion at Menominee River Park has been reroofed, the displaced decking repaired and the soffits finished with aluminum.
- An order approving application for replacing current street lighting to light emitting diode (LED) street lighting was approved at the August 28, 2018 Public Service Commission meeting. Wisconsin Electric Company will now be completing application to submit for cost of conversion and operation for consideration. A copy of the Order Approving Application was included in the packet.
- The Youth Football League has completed the work on the scoreboard at Lodal Park, and it is operational.
o City Manager Edlebeck has been in communication with GEI Consulting Engineers in regard to the backup power source for the Public Safety building. The design and cost estimates are under review.
o The preliminary information on the potential replacement of the operating valve for well no. 7 to a variable frequency drive will be reviewed. There is potential for improved control of the operation of our largest capacity well, as well as an electricity energy savings. Edlebeck met with a representative from Energy United and discussed a potential rebate for the project.
o The deer archery permits have been approved by the DNR; the City was approved for 65 permits. This has been advertised, and the information has been provided to the staff at City Hall and KPS. Hardcore Pursuit is once again providing the archers with certification.
A motion was made by Councilmember Dixon-Miller and supported by Councilmember Groeneveld to adjourn as the regular City Council and reconvene as the Zoning Board of Appeals.
ALL AYES. NO NAYS. MOTION CARRIED.
At this time, the Public Hearing scheduled for 6:45pm began addressing 190 North Boulevard Rear Set-Back submitted by Mr. Mark Shanks. City Manager Edlebeck read the notice as it was published in the Daily News on August 29, 2018. Edlebeck had not received any comments, written or verbal. Mr. Mark Shanks attended the meeting.
Ayes: Baldinelli, Erickson, Dixon-Miller, Flaminio and Groeneveld
Nays: None
MOTION CARRIED.
The Mayor asked three times if there was any public comment; there was none.
A motion was made by Councilmember Baldinelli and supported by Councilmember Dixon-Miller to adjourn as the Zoning Board of Appeals and reconvene as the Regular City Council.
ALL AYES. NO NAYS. MOTION CARRIED.
A motion was made by Councilmember Groeneveld and supported by Councilmember Baldinelli to receive and place on file the September 17, 2018 Manager’s Report.
ALL AYES. NO NAYS. MOTION CARRIED.
The first item under Old Business was the Recommendation of Award for a 4x2 Plow/Dump Truck from the September 4th bid opening. Following the bid opening, City Managers Edlebeck and Tim Pollock, the City’s Mechanic reviewed each bid. A memo was provided in the packet stating the two low bidders did not meet specifications in two areas. The next bidder also did not meet specifications. The fourth low bid was submitted by UP Truck Center in Quinnesec, MI and did meet all specifications. Edlebeck and Pollock’s recommendation was to award the bid to UP Truck Center of Quinnesec, MI in the amount of $147,986.00 which includes the trade allowance of $2500.00 for Unit 57 [1981 IHC Dump/Plow].
A motion was made by Councilmember Flaminio and supported by Councilmember Baldinelli to award the 4x2 Plow/Dump Truck bid to UP Truck Center in the amount of $147,986.00 with the trade allowance of $2500.00 for Unit 57 [1981 IHC Dump/Plow].
ALL AYES. NO NAYS. MOTION CARRIED.
The last item under Old Business was the Bid Opening for 22A Road Gravel. Quantity requested was approximately 600 tons. Three bids were received:
- Midwest Asphalt & Gravel – Iron Mountain, MI – $8.40/ton
- Morin Excavating – Niagara, WI – $9.23/ton
- Jim Wender & Sons Trucking – Iron Mountain, MI – $8.00/ton
A motion was made by Councilmember Baldinelli and supported by Councilmember Dixon-Miller to award the 22A Road Gravel Bid to the low bidder, Jim Wender & Sons Trucking, in the amount of $8.00/ton provided all specifications are met.
ALL AYES. NO NAYS. MOTION CARRIED.
The first item under New Business was a Request for Front Yard Set-Back Variance for 649 Edsel Street Addition. A request was submitted by Mr. Michael Romback to add an addition onto the east-side of his house, consideration due to the topography in his rear yard and drainage concerns. In order to do so, he would need a front set-back variance of 21’ in lieu of the required 25’.
A motion was made by Councilmember Baldinelli and supported by Councilmember Flaminio to schedule a public hearing for October 15th at 6:45pm re: Front Yard Set-Back Variance Request for 649 Edsel Street.
ALL AYES. NO NAYS. MOTION CARRIED.
The second item under New Business was Application to Display Fireworks at KHS Football Game on Friday, October 12, 2018.
A motion was made by Councilmember Groeneveld and supported by Councilmember Flaminio to approve the Request for Fireworks Display at the KHS Football Game on Friday, October 12th provided liability insurance coverage is provided.
ALL AYES, NO NAYS. MOTION CARRIED.
The third item under New Business was Resolution Recognizing 50-Year Anniversary of TRICO Opportunities. TRICO Opportunities celebrated their 50th Anniversary on September 16, 2018, and this resolution is for the council’s consideration to recognize their work and dedication to their clients and customers.
A motion was made by Councilmember Baldinelli and supported by Councilmember Dixon-Miller to approve Resolution No. 2018/9/17.1 re: 50-Year Anniversary of TRICO Opportunities.
ROLL CALL:
Ayes: Baldinelli, Erickson, Dixon-Miller; Flaminio and Groeneveld
Nays: None
MOTION CARRIED.
The fourth item under New Business was Resolution requesting release of reverter clause on city-owned lot on Crest Street. Edlebeck was approached by the Michigan State Land Bank, who owns Lot 10 of Block 6, Replat of Blocks 1 and 3 of the First Addition to Kingsford Heights whether the City of Kingsford would be interested in acquiring 10 to join with the city-owned Lot 11. Both lots contain a portion of alley.
A motion was made by Councilmember Baldinelli and supported by Councilmember Dixon-Miller to approve Resolution No. 2018/9/17.2 re: Release of Reverter Clause on city-owned Lot 11 on Crest Street as prepared by the City Attorney.
ROLL CALL:
Ayes: Baldinelli, Erickson, Dixon-Miller, Flaminio and Groeneveld
Nays: None
MOTION CARRIED.
The fifth item under New Business was Resolution requesting interest in acquiring a State of Michigan-owned lot on Crest Street. The City has had discussions with representatives of the Michigan Land Bank in regard to acquiring Lot 10 of Block 6, Replat Block 1 and 3 of the First Addition to Kingsford Heights, and a tentative sales price for the described parcel of $100.00.
A motion was made by Councilmember Groeneveld and supported by Councilmember Baldinelli to approve Resolution No. 2018/9/17.3 re: Interest in Acquiring a State of Michigan-owned Lot 11 of Block 6 of the First Addition of Kingsford Heights on Crest Street in the amount of $100.00 as prepared by the City Attorney.
ROLL CALL:
Ayes: Baldinelli, Erickson, Dixon-Miller, Flaminio and Groeneveld
Nays: None
MOTION CARRIED.
The last item under New Business was the MML Board of Directors Election Ballot. Two incumbents have agreed to seek re-election. A brief biography of each candidate was provided for review.
A motion was made by Councilmember Baldinelli and supported by Councilmember Dixon-Miller to vote for the two incumbents listed on the ballot.
ALL AYES, NO NAYS. MOTION CARRIED.
There was no Public Comment.
Under Councilmember’s Privilege, Councilmember Groeneveld expressed his appreciation to the City Manager for his attention and forethought re: Lodal Park fields (soil, etc.) and mentioned the new pavilion roof at Menominee Park looks nice.
There being no further business, a motion was made by Councilmember Dixon-Miller and supported by Councilmember Groeneveld to adjourn the meeting.
ALL AYES, NO NAYS. MOTION CARRIED. |
Monolithic Fabrication of EWOD Chips for Picoliter Droplets
Wyatt C. Nelson and Chang-Jin “CJ” Kim, Member, IEEE, Fellow, ASME
Abstract—We report monolithic fabrication of parallel-plate electrowetting-on-dielectric (EWOD) chips for digital microfluidics of picoliter droplets. Instead of assembling a second substrate to form a top plate—the common practice with all previous parallel-plate EWOD chips—the top plate is surface micromachined as a transparent thin-film membrane that forms a monolithic cavity having a gap height on the order of micrometers with excellent accuracy and uniformity. The membrane is embedded with EWOD driving electrodes and confines droplets against the device substrate to perform digital microfluidic operations. Two main attributes of the monolithic architecture that distinguish it from traditional methods are: (i) it enables excellent control of droplet dimensions down to the micrometer scale, and (ii) it does not require the typical alignment and assembly steps. Basic device functions such as creation and splitting are verified by EWOD actuation of $\sim 100$ picoliter droplets surrounded by air or oil inside a 10 $\mu$m-high cavity. Additionally, flow focusing of droplets containing 5.3 $\mu$m beads demonstrates one example of the utilities afforded by monolithic fabrication.
Index Terms—Digital microfluidics, electrowetting, electrowetting-on-dielectric (EWOD), monolithic fabrication, picoliter droplets.
I. INTRODUCTION
For various reasons [1], we have yet to see the broad implementation of microfluidic systems as automated bioanalytical tools. For example, the benefits of miniaturization through microfluidics (e.g., fast reaction times, low waste, and high sensitivity detection) are usually not attained without trained technicians and a specialized assortment of benchtop components, often including mechanical pumps and valve manifolds. In this light, digital microfluidic [2], [3] systems based on electrowetting are attractive because they enable fluid actuation by electrical signals alone, thereby reducing the complexity of the chip as well as the overall system. Such electrically driven microfluidics, particularly using the chip configuration of electrowetting-on-dielectric (EWOD) [4], have shown promise as “generic platforms,” because they can be reprogrammed on-the-fly using software to accommodate various biochemical protocols. Further, because EWOD chips have low power requirements, it is conceivable to build a handheld (e.g., cellphone-like) system that runs on batteries [5].
Most EWOD-based digital microfluidic devices have one of the following generic architectures: parallel-plate [6], [7] or open-planar [8], [9]. Of these, the former is often preferred, despite the additional steps in fabrication, for its utility and reliability in applications. For example, it is far easier to generate droplets and split them apart when a liquid is squeezed between two plates. Also, controlling the device gap (between the plates) is a reliable way to scale the working fluid. In comparison, on an open-planar device, droplet dimensions are determined by the contact angle, which is nearly impossible to control or predict with accuracy. Third, if the gap is much smaller than the capillary length (e.g., 2–3 mm for water), large droplets are insensitive to inertial forces. The same cannot be said for open-planar devices unless the droplet volume is small, i.e., the diameter is much smaller than the capillary length.
Traditional parallel-plate EWOD device manufacturing begins with thin film deposition and patterning of the two plates—typically a bottom substrate patterned with EWOD electrodes and a top plate with a blank conductive layer—and ends with assembly. Overall, the plates are simple to fabricate in the clean room, for example, using one step each of metallization, photolithography, and dielectric deposition. The assembly process, which can be performed using a variety of methods, involves alignment of the two plates (which is not critical for most applications because the ground plate is unpatterned), fabricating and positioning spacers, and bonding of the plates; the last two steps determine the gap between the plates and therefore the thickness of the droplets. In research laboratories, it is common to align and affix cutouts of an adhesive spacer, e.g., double-sided tape, to the plates for spacing and bonding. This practice is acceptable for fabricating parallel-plate EWOD devices with relatively large gaps (i.e., $> 50 \mu$m). When smaller gaps are desired, a thick photoresist (e.g., SU-8) can be coated and lithographically defined as a spacer, and additional provisions, e.g., external alignment and clamping mechanisms, are employed to hold the plates together. Song et al. [10], Lin et al. [11], and Welch et al. [12] have demonstrated excellent performance using such devices for handling droplets as small as 35 pL in oil. Welch et al. [12] have also demonstrated how problems associated with slight irregularities in...
the thickness of spin-on spacers can be alleviated by using an additional polymer layer on the top plate. Note, however, that a spin-coated polymeric spacer would not provide the levels of uniformity and repeatability in the gap dimension that are readily achievable using surface micromachining methods based on common integrated circuit processes, e.g., oxidation of silicon, chemical vapor deposition.
This paper explains our recent progress in developing a concept initially proposed in [13], in which we demonstrated EWOD actuation of $\sim 100$ pL droplets immersed in oil in a monolithically fabricated parallel-plate device. Having refined the design and fabrication processes, our newest results include repeatable EWOD-driven creation, transportation, splitting, and merging of $\sim 100$ pL droplets in air as well as oil. Our purpose in developing a surface-micromachined parallel-plate EWOD chip is not to replace traditional chips; instead, we are interested in exploring a new process because it brings new opportunities. In fact, since the proposed micromachining processes is much more involved than typical EWOD device fabrication, the latter will undoubtedly be preferred in applications for which it makes sense, e.g., large-gap devices. The main attributes of our surface micromachining processes that distinguish it from traditional methods are: (i) it enables excellent control of droplet dimensions down to the micrometer, scale and (ii) it does not require alignment and assembly steps.
Since it requires special fabrication techniques to repeatedly form accurate gaps smaller than $\sim 50 \mu m$, most reported parallel-plate EWOD chips are designed to handle droplets greater than $\sim 100$ nL (i.e., $1000 \mu m \times 1000 \mu m \times 100 \mu m$; note that a water droplet needs to be at least five times wider than its thickness to allow for all the basic digital microfluidic operations in air, i.e., without filling the gap with oil [7]). The gap heights of the reported monolithic chips are set by the thickness of the sacrificial layer, which is the device layer of a silicon-on-insulator (SOI) wafer and guaranteed by the manufacturer (Ultrasil) to $\pm 0.5 \mu m$ across the wafer. We chose to use SOI wafers to obtain a relatively thick sacrificial layer (i.e., $10 \mu m$ used in this study); one would deposit a thin film for a thin sacrificial layer (e.g., $1 \mu m$). With such accuracy in the gap, the minimum droplet volume in the monolithic chip is determined by the lithography of the EWOD electrodes. Using electrodes $100 \mu m$ on a side with a $10 \mu m$ distance between them and a $10 \mu m$-thick sacrificial layer, as a conservative example, the new architecture can handle droplets of $\sim 100$ pL. This represents a 1000-fold reduction in volume compared to droplets in traditional assembled chips. Although it is not experimentally demonstrated in this paper, one could use common micromachining facilities to fabricate electrodes $10 \mu m$ on a side with a $2 \mu m$ distance between them and a $2 \mu m$-thick sacrificial layer. This would result in a monolithic EWOD chip capable of handling femtoliter droplets.
The ability to handle much smaller droplets will make new applications, such as single-cell analysis, accessible to digital EWOD. Motivated by the recent push from biologists to study heterogeneity within cell populations, various microfluidic platforms have been developed for trapping (e.g., physical microtraps [14], dielectrophoretic (including light induced [15]) traps [16]), and monitoring (e.g., optical [17] or electrical [18]) of single cells and organelles under various stimuli [19]–[23], etc. A promising approach is to create, manipulate, and examine discrete nano- and picoliter liquid volumes containing one or a few cells each [24], [25]. In such platforms, stimuli (e.g., chemical, mechanical, thermal, and electrical) and responses (e.g., secreted chemicals, division, apoptosis) are localized to each droplet, i.e., in an individual reactor, and thus rare cell-type responses are detected—this is in contrast to conventional techniques that mash together cell populations before analyzing the contents, giving output that is spatially and temporally averaged. Cai et al. [26] reported isolation and long-term monitoring of protein expression of single cells in 100 pL of liquid. Their chip effectively solved the “efflux” problem, i.e., loss of analyte signal due to rapid diffusion away from the cell, by physically confining the cell and its secretions. He et al. [24] integrated optical trapping with a microfluidic T-channel droplet generator, thereby demonstrating a different way to encapsulate single cells and molecules of interest inside pico- and femto-liter droplets. Even among the highly versatile pressure-driven droplet microfluidic systems of today that are capable of splitting, mixing, and merging thousands of droplets in parallel [27]; digital microfluidic chips (such as the one reported here) bring a distinct set of capabilities such as independent and reconfigurable droplet routing and liquid-in-gas operation, both of which are important for long-term cell culture and monitoring.
Contamination can be an issue when operating EWOD chips, and the processes used for removing residues prior to or between experiments are different for assembled and monolithic devices. For example, it is a common practice in the research setting to assemble chips outside the clean room and risk contamination from the environment. It should be noted that this is less of a concern when oil is used as a filler medium. During experiments, disassembly is often required to clean or recondition the substrate. This can be complicated by the fact that adhesive spacers are generally dissolved by cleaning agents, e.g., organic solvents and sulfuric acid, and need to be replaced. These complications, however, can be avoided by using chemically resistant spacers, e.g., SU-8 or polyimide, that do not need to be replaced during the cleaning process. In any case, the reported monolithic chip fabrication is not subject to contamination from the ambient environment because the chip cavity is created in the clean room. This procedure ensures that there is no initial contamination. Also, the device is relatively impervious to cleaning agents; the only polymers that exist on the chip are Teflon or Cytop, which are exceptionally chemically stable. To clean the monolithic chip between experiments, it was soaked in a solvent (e.g., acetone) and rinsed with water.
Fig. 1 provides a visual comparison of assembled (1a) and monolithic (1b) parallel-plate EWOD devices. With assembled devices intended for actuating droplets in a gas medium, issues associated with assembly (i.e., poor dimensional control, contamination, spacer degradation) are more detrimental to device performance as the gap height is reduced. By eliminating the need for device assembly, monolithic chips can be scaled down very accurately to handle picoliter droplets in air immediately following clean room fabrication. After designing
and fabricating the monolithic EWOD chip, we confirmed its successful operation by demonstrating core digital microfluidic functions: droplet creation, transport, splitting, and merging. We also explored additional microfluidic functions involving droplet flow focusing and particle manipulation.
II. DEVICE DESIGN, ARCHITECTURE, AND FABRICATION
A. Fabrication
The monolithic EWOD chip is fabricated by surface micromachining, as shown in Fig. 2. (1) The starting material is a 4" SOI wafer with a 10 μm-thick device layer, 0.25 μm-thick buried SiO₂, and 525 μm-thick handle layer (resistivity = 0.01 – 0.02 Ω·cm). (2) The process begins with photolithographic patterning and deep reactive-ion-etching of the silicon device layer. The recesses etched in this step define the anchors of the eventual membrane. (3) Wet thermal oxidation is used to grow 100 nm of SiO₂ on all exposed silicon. This oxide layer is necessary to protect the membrane from the release etch, discussed later. (4) A 1 μm low-stress silicon nitride (Si₃N₄) layer is deposited by plasma-enhanced chemical vapor deposition (PECVD). (5) Metals (7/100 nm Ti/Au) are deposited onto the substrate by electron-beam evaporation. (6) The metals are photolithographically patterned and wet etched (Au was etched by commercial gold etchant; Ti was etched by a commercial buffered oxide etchant containing HF) to form electrodes. (7) Another photolithography and wet etch are used to selectively remove Au from electrodes on the membrane, leaving behind the transparent Ti electrodes. Even though the process flow does not show it, Au remains on contact pads and lead wires on the chip. We note that the 7 nm of Ti served as an adhesion layer for the Au as well as a highly transparent and sufficiently conductive layer for EWOD actuation. It was therefore unnecessary to add another layer, e.g., ITO, to form transparent actuation pads. (8) Another low-stress Si₃N₄ layer (3 μm thick this time) is deposited by PECVD. (9) The next step (not shown as it is outside the view) is photolithographic patterning and CF₄ plasma etching of the uppermost Si₃N₄ layer to uncover electrical contact pads at the edge of each chip on the wafer. (10) Release holes are patterned and etched using CF₄ plasma. (11) The membranes are released by XeF₂ vapor-phase etching, and the cavities are formed. (12) Finally, the chips are spin coated (6000 rpm for 120 s) with a hydrophobic polymer solution (1 wt% Cytop) and baked in an oven at 150 °C for 10 min. The following procedure was used to measure the thickness of the Cytop layer inside the EWOD cavity: (i) membranes were peeled off (using adhesive tape) of experimentally verified chips, (ii) the bottom surface of the cavity was scratched several times using a stainless steel probe tip, (iii) and the area was scanned using a stylus profilometer. For three devices that were spin coated using the same process, the thickness was 38 ± 6.8 nm. For each device, two spots were...
measured inside the cavity—one near the center of the cavity and one near the entrance from the reservoir. We did not observe any tendency in the thickness variation related to measurement location.
In a previous report [13], we characterized the gap height variation attributable to membrane deflections caused by intrinsic stress in the films. We found that, for membranes identical to those reported here, the maximum deflection of a membrane spanning 300 μm was less than 200 nm. This represents an extreme case because anchors need not be so far apart. Even so, such a minor deflection would lead to only a 1.3% uncertainty in droplet volume of a 100 pL droplet (assuming the volume is 100 μm × 100 μm × gap height).
B. Architecture
Fig. 3 shows schematic top views of a completed chip consisting of three basic components: anchors, electrodes, and membranes. Anchors are the structural linkages between the substrate and the suspended membrane. Also, anchors serve as etch stops (in the direction parallel to the wafer surface) during the membrane release etch and define the cavity areas. Note that the EWOD electrodes are embedded within the membrane, i.e., within the top plate. Typical assembled parallel-plate EWOD chips have the actuation electrodes on the bottom substrate. Whether the electrodes are in the “top” or “bottom” is not an important distinction given that they are relative terms—the key distinction of the monolithic chip from the assembled chips is that the EWOD dielectric is deposited before the electrodes instead of after. This layer configuration, which happens to produce a perfectly flat actuation surface, was a natural choice after our decision to use the SOI device layer as a sacrificial layer. Release openings in the membrane allow access to the sacrificial layer etchant during the formation of the cavity beneath the membrane. Also, release openings, if placed near or on top of the EWOD electrodes, can serve as liquid source or waste reservoirs during device operation. For example, two large circular release openings (diameter = 1 mm) were filled with liquid (~1 μL each) for most of the device demonstrations described in later sections.
The monolithic architecture used in this study (Fig. 1) utilizes parallel-plate actuation because the handle wafer is electrically grounded, and actuation force is delivered via electrodes embedded in the membrane. As in most parallel-plate systems, the droplet is not actually grounded, but the dielectric covering the grounding plate (the substrate in our case) is very thin compared to that covering the actuation electrodes, so the latter represents the dominant capacitance in the circuit.
C. Scaling of EWOD Forces With Thicknesses of Layers
Surface micromachining enables accurate control of cavity heights in the micrometer range. Digital EWOD devices having gap heights below 100 μm have demonstrated droplet manipulations in oil using an assembled chip with an SU-8 spacer [10]–[12] and in air using a chip mounted to a micrometer to control the gap [28]. Using a similar device configuration, but with a photo-conductive layer rather than patterned electrodes, Chiou et al. [29] demonstrated “optoelectrowetting” manipulation of 10 pL droplets. Importantly, in the parallel-plate configuration, all layers of the EWOD device (i.e., liquid, vapor, dielectric, hydrophobic, etc.) have at least some contribution to the overall electrical driving force [30]. In this way, EWOD differs from driving mechanisms based on thermally or chemically induced wettability gradients, which have also been employed in digital microfluidic systems [31]–[33]. As a reference for surface force-based droplet actuation, we recommend the review of Darhuber and Troian [34]. To illustrate the importance of the entire EWOD circuit, in which electrodes, insulators, and fluids are represented by electrical components, we use the electro-mechanical model developed by Jones et al. [35]. The following simplified analysis shows how the variation of EWOD forces with actuation frequency is dependent on the ratio of cavity height $h$ to dielectric thickness $\sigma$. It is important to note that in most cases, the electrowetting force equation is developed with the assumption that $h \gg d$, leading to the result that the time-averaged force parallel to the direction of motion is $F = 0.5eV^2$, where the specific capacitance $c$ is only a function of $d$. The following analysis, however, does not assume $h \gg d$ because our intention is to show how the total electrical force changes when $h$ is scaled down to dimensions similar to $d$. Using the circuit model shown in Fig. 4(a), we express the dimensionless voltage drops (divided by the applied voltage $V$) across the liquid (subscript “l”) and dielectric (subscript “d”) layers as
$$\frac{V_{1,\text{liquid}}}{V} = \left| \frac{1}{1 + \frac{d}{h} \left( \frac{\varepsilon_1}{\varepsilon_2} - j \frac{\sigma}{\omega \varepsilon_2} \right)} \right|; \quad \frac{V_{1,\text{vapor}}}{V} = \frac{1}{1 + \frac{\varepsilon_0}{\varepsilon_2} \frac{d}{h}} \tag{1a}$$
$$\frac{V_{2,\text{wet}}}{V} = \left| \frac{d}{h} \left( \frac{\varepsilon_1}{\varepsilon_2} - j \frac{\sigma}{\omega \varepsilon_2} \right) \right|; \quad \frac{V_{2,\text{dry}}}{V} = \frac{1}{1 + \frac{\varepsilon_0}{\varepsilon_2} \frac{h}{d}} \tag{1b}$$
where $\varepsilon_1$ and $\varepsilon_2$ are dielectric constants of liquid and dielectric, respectively, and $\sigma$ is the liquid conductivity. Referring to Fig. 4(a): Equation (1a) correspond to the voltage across the liquid, which is modeled as a resistor and capacitor with $\varepsilon_1 (=80$ for water$)$ in parallel and the voltage drop across the vapor layer, which is modeled as a capacitor with $\varepsilon_v (=1$, assuming similar to vacuum$)$; Equation (1b) correspond to the voltage drop across the dielectric beneath the liquid and the vapor, which is modeled as a capacitor with $\varepsilon_2 (=4$ for silicon nitride$)$.
Using lumped elements, the time-averaged force per unit width (into the page) moving the liquid from left to right is [36]
$$\langle F_x \rangle = -\frac{1}{2h} \left( \varepsilon_1 V_{1,\text{liquid}}^2 - \varepsilon_v V_{1,\text{vapor}}^2 \right) + \frac{\varepsilon_2}{2d} \left( V_{2,\text{wet}}^2 - V_{2,\text{dry}}^2 \right). \tag{2}$$
In Fig. 4(b), the above force is plotted in terms of $\langle F_x \rangle$ versus $f/\sigma$, where $f$ is the actuation frequency. Four curves corresponding to different values of $h/d$ ($d=1 \ \mu m$ assumed in all cases) show that this ratio affects the actuation force in three ways: (i) decreasing $h/d$ decreases the force at low frequency, (ii) decreasing $h/d$ delays the transition from the low- to high-frequency limits. In order words, if we define a crossover frequency as the inflection point of the force curve, then decreasing $h/d$ increases the crossover frequency, and (iii) decreasing $h/d$ increases the force at high frequency. From Fig. 4(b) and (ii) and (iii) above, we can make a useful conclusion in designing EWOD chips: scaling down of $h/d$ can be a way to lower the actuation voltage when working with low-conductivity liquids (e.g., oils) or at high actuation frequencies.
To keep the model simple, we have left several dielectric layers (e.g., hydrophobic coatings and thermal oxides) out of the circuit. Refer to Chatterjee et al. [30] for an excellent demonstration of how to incorporate all layers on the parallel-plate EWOD chip. The model used here is sufficient to show the scaling behavior of $h/d$, and even without the thin dielectrics, it is reasonably accurate because the capacitances of the thick dielectric and fluid layers are dominant.
We conclude our discussion on the scaling of EWOD forces with a few comments on the fluid dynamics. First, the Reynolds number scales directly with the gap $h$. A high-speed camera (Phantom v7.2) was used to measure a droplet speed of 100 mm/s at 60 V$_{\text{rms}}$ (20 kHz) in the current monolithic chip, which has $h \sim 10 \ \mu m$. This is similar to the droplet speed in a typical assembled chip of $h \sim 100 \ \mu m$ (e.g., [37]), except that the associated Reynolds number is 10 times lower in the monolithic chip. Secondly, if particles are suspended within the liquid, and the gap $h$ is only slightly larger than the particles, then they can move in only one or two dimensions. The second point is demonstrated by our flow focusing experiment discussed in the next section. The ability to scale the microfluidic chip with biological particles such as cells can greatly simplify processes such as inspection and sorting of the particles.
III. EXPERIMENTAL SETUP AND PROCEDURE
A. Experimental Setup
The experimental setup is shown in Fig. 5. A function generator (Agilent 32200A) outputs a sinusoidal waveform to the voltage amplifier (TREK PZD700A), which outputs the actuation voltages to a homemade electrical multiplexer having a manually operated array of 10 SPDT Submini lever switches (RadioShack). The multiplexer routes voltages to ten addressable channels via a ribbon cable ([3M round conductor flat cable with 1 mm spacing between the wires (28 AWG)]) connected to the 34-pin clip holding the EWOD chip. The 34-pin clip, a modified parallel port plug, provides both electrical connections and mechanical clamping to the chip during operation, so no other clamps are necessary. The manual multiplexer was suitable in this study because the tested EWOD devices had ten or less electrodes, many more of which would have required an automated multiplexer. Videos of EWOD operations were captured using two different cameras, depending on the required frame rate (note that the camera in Fig. 5 represents either a normal USB CMOS camera or a high-speed camera). For most observations, we recorded at 30 frames per second using a USB digital camera (Edmund Optics CMOS EO-0413C, 752 × 480 pixels) mounted to an optical microscope (20× objective) on a vibration isolation table. For measuring droplet speeds and tracking the movements of beads inside of droplets, we recorded at 4000 frames per second using a Phantom v7.2 high speed camera (Vision Research). The setup shown in Fig. 5 allowed open access to the chip for pipette sample loading and extraction. In the picture in Fig. 5, the microcavity with
EWOD electrodes (too small to see) is located between the two reservoir droplets, which are about 1 mm in diameter.
B. Evaporation of Droplets
Evaporation is an important issue for the monolithic EWOD device because droplets have extremely small volumes. In preparation for EWOD testing, we evaluated the evaporation of water droplets and explored conditions that minimized or eliminated the volume loss. In the current chip designs (similar layouts as Fig. 3), the cavity area is relatively small, and the EWOD electrodes are close to the cavity openings, so the droplets are expected to evaporate nearly as fast as outside the cavity. With all the release openings clear, for instance, a 100 pL water droplet inside the cavity was observed to evaporate completely in about 10 s. However, with full reservoirs, a 100 pL water droplet lasted more than 5 min (versus 10 s) in different devices, where the cavity was designed much deeper and the release openings smaller. In the same condition, a 100 pL 50:50 mix of water and glycerol showed no signs of evaporation in 10 min. When a large (∼10 μL) liquid droplet was placed on the device to block all the openings, a 100 pL water droplet inside the cavity showed no signs of evaporation over several hours, suggesting the cavity was saturated with water vapor. The latter condition was used to verify droplet-in-vapor operations shown in Figs. 7–10. As expected, when the cavity was filled with oil [8], no volume loss of water droplet was observed over hours of experimentation.
C. Initial Loading
In a preliminary report in which droplets were actuated only in oil [13], we demonstrated that monolithic EWOD chips could be loaded with liquids via NanoPorts (Upchurch Scientific), which are used commonly in microfluidic devices as micro-to-macro interfaces, i.e., mating microchip inlets/outlets with millimeter-scale tubing. In [13], we used a syringe to pump liquids through tubes connected to the NanoPorts, which were bonded to the chip. The use of NanoPorts with the monolithic device, however, was only meant to prove the EWOD core functions (creation, transport, and splitting) in the surface-micromachined cavity. In this report, the device was designed for loading without NanoPorts. Specifically, with the current chip, liquids are pipetted directly onto the reservoirs, which are defined by release openings next to the EWOD cavity. This method enables straightforward loading of nonwetting liquids (contact angle > 90°), because they could be deposited onto the reservoirs without flooding the entire chip. Controlled loading of wetting fluids, e.g., organic solvents or oils, on the hydrophobic Cytop surfaces require another method, e.g., the NanoPort.
Fig. 6 shows a subtle, yet important, detail of chip loading: filling the rim around the reservoir with a nonwetting liquid, e.g., water. For the liquid to be accessible to the EWOD electrodes, it must first occupy the space beneath a rim of membrane that surrounds the reservoir, as shown in Fig. 6(a). When a droplet is first loaded [Fig. 6(b)], the region beneath the rim is dry. To fill, or prime, the reservoir [Fig. 6(c) and (d)], we deposited a warm water droplet onto the chip. This led to condensation within the rim region and complete filling in several seconds. Once primed [Fig. 6(e)], the chip could be filled with any liquid and operated. Various design modifications will eliminate this priming step in the future, such as using an EWOD electrode that extends into the rim of the membrane.
Fig. 7. Video frames (seen through the top membrane) of introducing water into the cavity and creating a 100 $\mu$L droplet (100 $\mu$m $\times$ 100 $\mu$m $\times$ 10 $\mu$m) using 60 $V_{rms}$ at 20 kHz. The liquid is under the electrodes, which are imbedded in the transparent membrane. Actuated electrodes are labeled “ON,” and all other electrodes as well as the substrate silicon are grounded. The two small droplets on the electrode wires above the actuation electrodes are satellite droplets formed unintentionally during previous actuations. Anchors appear yellow because gold remained (harmlessly) in the 10 $\mu$m-deep recesses after the lithography.
Fig. 8. Video frames (seen through the top membrane) of incremental droplet splitting using 30, 40, and 50 $V_{rms}$ at 20 kHz. Each voltage was applied continuously and raised by 10 V every 3–4 s. The droplet immediately broke apart at the application of 50 $V_{rms}$.
and/or reducing the width of the rim itself such that EWOD forces pull liquid directly from the reservoir.
IV. RESULTS AND DISCUSSION
A. Creation and Splitting of Droplets
The core digital microfluidic functions of droplet creation, transport, splitting, and merging [4] were demonstrated using the monolithic EWOD chip. Fig. 7 is a sequence of video frames showing creation of deionized water droplets in air (saturated with water vapor) inside the microcavity using 60 $V_{rms}$ at 20 kHz for EWOD actuation. While droplet transport was repeatable at voltages down to 30 $V_{rms}$, creation was not repeatable below 60 $V_{rms}$. Similarly, droplet splitting required voltages greater than 50 $V_{rms}$ to be repeatable. Fig. 8 is a sequence of video frames showing the shape of a droplet at various voltages too low to cause splitting. The voltage was applied continuously and increased by 10 V every 3–4 s. At 30 $V_{rms}$ the droplet begins to neck, at 40 $V_{rms}$ the neck becomes much thinner, and the droplet breaks immediately upon increasing the voltage to 50 $V_{rms}$. We tested a range of actuation frequencies (up to 50 kHz) and found that actuation was most reliable at 10 to 20 kHz. Also, the transportation of droplets with 30 $V_{rms}$ was successful at $f < 30$ kHz but not at higher frequencies (tested up to 50 kHz). This behavior could be a result of the redistribution of electric field and concomitant decrease in actuation force [as seen in [30] and indicated by the curves in Fig. 4(b)]. For example, in Fig. 4(b), the curve for $h/d = 10$, roughly representing the current fabrication, sustains its maximum value up to $f/\sigma \sim 5 \times 10^8$ before decreasing. Using the conductivity of deionized water after exposure to air reported in [38] ($\sigma \sim 7 \times 10^{-5}$ S/m), we calculate the critical frequency to be $f = 35$ kHz, which agrees with our observation.
B. Flow Focusing and Particle Manipulation
In exploring the utilities of the surface-micromachined EWOD chip, we have tested flow focusing of droplets suspended with particles. With the reported fabrication process, it is simple to incorporate arbitrarily shaped posts or walls in the EWOD cavity at many locations. It is also important to note that such structures are perfectly sealed to the upper and lower inner surfaces of the cavity. This is a consequence of the fact that the vertical structures (anchors) and membrane(s) are made of the same structural layer. It is possible to build constrictions through which droplets can be squeezed by EWOD actuation, thereby accomplishing flow focusing. By flow focusing, we are referring the process by which liquid and suspended particles are forced through an orifice to increase the sensitivity of a measurement, e.g., electrical impedance [39]. Note that flow focusing also refers to the process by which liquid streams are pinched such that they break up into droplets [40]. Fig. 9 is a sequence of video frames showing the creation, transport, and flow focusing of a droplet through a 40 $\mu$m wide constriction defined by the two constriction anchors. Flow focusing is often used in microchannels either by physical constrictions or sheath flows to analyze particles one-by-one, e.g., microfluidic cytometry [41].
Dhindsa et al. [42] recently reported an interesting EWOD configuration in which liquid is actuated through arrays of microposts, thereby creating $\sim 20 \times 20 \mu$m virtual microchannels of water in oil. This type of flow constriction is similar to that shown here, except the monolithic chip has digitized electrodes for handling individual droplets. Fig. 10 shows examples of focusing droplets containing twelve [Fig. 10(a)] and one [Fig. 10(b)] polystyrene beads (Invitrogen, Surfactant-Free Fluorescent Nile Red CML Polystyrene Latex, Product No: 2-FN-5000, diameter = 5.3 $\mu$m). Note the bead diameter is more than a half of the cavity height, rendering particle movements nearly 2-D. To track the path of the particles through the constriction, it was necessary to use a high-speed camera, recording at 4000 frames per second. Fig. 10(c) is a composite image showing the initial meniscus of a droplet (broken line) and final meniscus after 2.5 ms into the focusing actuation (solid line). Also, shown are the locations of the particle at 0.25 ms intervals, which are connected by a blue line to show its path through the constriction.
Fig. 9. Video frame sequence of droplet flow focusing. Between steps 6 and 7, the water droplet is squeezed between the 40 μm wide constriction formed by two triangular anchors located between two EWOD electrodes.
Fig. 10. Video frames of droplet flow focusing and particle handling. Large black arrows indicate the direction of droplet motion. (a) A droplet containing twelve 5.3 μm beads before (1) and after (2) moving through the constriction. (b) A droplet containing one 5.3 μm bead before (1) and after (2) moving through the constriction. (c) A composite image created using high-speed video frames of focusing a droplet containing one 5.3 μm bead. Lines were drawn over menisci at t = 0 (dashed black line) and t = 2.5 ms (solid black line). Bead locations at 0.25 ms intervals are shown by black circles, which are connected by a blue line to show the bead path.
In the reported chip, a 100 pL (100 μm × 100 μm × 10 μm) droplet containing a single 5.3 μm bead was created as follows. First, a droplet was created from a reservoir of liquid containing the beads. If the initial droplet happened to contain several beads, a series of splitting and merging operations were performed to render a droplet containing only one bead. The number of beads within the initial droplet was dependent upon the concentration of beads in the reservoir. We have not fully characterized this relationship, but with roughly $3 \times 10^5$ beads/mL, initial droplets contained one to five beads.
V. CONCLUSION
Surface micromachining has been developed to fabricate monolithic EWOD chip on a single wafer, requiring no post-processing or assembly to obtain the parallel-plate configuration. The monolithic architecture used a thin-film membrane with embedded actuation electrodes as the top plate instead of a second substrate, as is necessary for assembled EWOD chips. The new process enables unprecedented dimensional control in the fabrication of parallel-plate EWOD devices. Reported devices have gap heights of 10 μm with 0.5 μm accuracy across the entire wafer. The core digital microfluidic functions (creation, transport, and splitting) as well as advanced functions (flow focusing of droplets with suspended particles) have been demonstrated using 100 pL water droplets in air inside a 10 μm-high microcavity.
REFERENCES
[1] G. M. Whitesides, “The origins and future of microfluidics,” Nature, vol. 442, no. 7101, pp. 368–373, Jul. 2006.
[2] R. B. Fair, “Digital microfluidics: Is a true lab-on-a-chip possible?” Microfluid. Nanofluid., vol. 3, no. 3, pp. 245–281, Jun. 2007.
[3] J. Lee, C. Nguyen, and C.-J. Kim, “Droplet actuation by electrowetting-on-dielectric (EWOD): A review,” J. Adhes. Sci. Technol.—Special Issue on Electrowetting, to be published.
[4] J. Lee, “Microactuation by continuous electrowetting and electrowetting: Theory, fabrication, and demonstration,” Ph.D. dissertation, Univ. California, Berkeley, CA, 2009.
[5] J. Gong, S.-K. Fan, and C.-J. Kim, “Portable digital microfluidic platform with active but disposable lab-on-chip,” in Proc. IEEE Int. Conf. MEMS, Maastricht, The Netherlands, Jan. 2004, pp. 355–358.
[6] M. G. Pollack, R. B. Fair, and A. D. Sherderov, “Electrowetting-based actuation of liquid droplets for microfluidic applications,” Appl. Phys. Lett., vol. 81, no. 11, pp. 2725–2726, Sep. 2002.
[7] S. K. Cho, H. Moon, and C.-J. Kim, “Creating transporting, cutting and merging liquid droplets by electrowetting-based actuation for digital microfluidic circuits,” J. Microelectromech. Syst., vol. 12, no. 1, pp. 70–80, Feb. 2003.
[8] U.-C. Yi and C.-J. Kim, “Characterization of electrowetting actuation on addressable single-side coplanar electrodes,” J. Micromech. Microeng., vol. 16, no. 10, pp. 2053–2059, Oct. 2006.
[9] C. G. Cooney, C. Y. Chen, M. R. Emmerling, A. Nadim, and J. Sterling, “Electrowetting droplet microfluidics on a single planar surface,” *Microfluid. Nanofluid.*, vol. 2, no. 5, pp. 435–446, 2006.
[10] J. H. Song, R. Evans, Y.-Y. Lin, B.-S. Hsu, and R. B. Fair, “A scaling model for electrowetting-on-dielectric microfluidic actuators,” *Microfluid. Nanofluid.*, vol. 7, no. 1, pp. 75–89, Jul. 2009.
[11] Y.-Y. Lin, R. D. Evans, E. Welch, B.-N. Hsu, A. C. Madison, and R. B. Fair, “Low voltage electrowetting-on-dielectric platform using multi-layer insulators,” *Sens. Actuators B, Chem.*, vol. 150, no. 1, pp. 465–470, Sep. 2010.
[12] E. L. E. Welch, Y.-Y. Lin, A. Madison, and R. B. Fair, “Picoliter DNA sequencing chemistry on an electrowetting-based digital microfluidic platform,” *Biosci. Tech.*, vol. 6, no. 2, pp. 165–176, Feb. 2011.
[13] W. C. Nelson and C.-J. Kim, “A monolithic EWOD chip by surface micromachining,” in *Proc. Solid-State Sens. Actuator Microsys. Workshop, Tech. Dig.*, Hilton Head Island, SC, Jun. 2000, pp. 158–161.
[14] D. DiCarlo, L. Y. Wu, and L. P. Lee, “Dynamic single cell culture array,” *Lab Chip*, vol. 6, no. 11, pp. 1445–1449, 2006.
[15] P. Y. Chion, A. T. Ohta, and M. C. Wu, “Massively parallel manipulation of single cells and microrotating using optical images,” *Nature*, vol. 436, no. 7050, pp. 371–372, Jul. 2005.
[16] J. Voldman, M. A. Gray, M. Tozer, and M. A. Schmidt, “A microfabrication-based dynamic array cytometer,” *Anal. Chem.*, vol. 74, no. 16, pp. 3984–3990, Aug. 2002.
[17] A. R. Wheeler, W. R. Thordissen, R. J. Whelan, A. M. Leach, R. K. Prasad, H. M. K. Farrell, I. D. Manger, and A. Daridon, “Microfluidic device for single-cell analysis,” *Anal. Chem.*, vol. 75, no. 14, pp. 3581–3586, Jul. 2003.
[18] A. A. Werdich, E. A. Lima, B. Ivanov, I. Ges, M. E. Anderson, J. P. Wikswo, and F. J. Baudenbacher, “A microfluidic device to confine a single cardiac myocyte in a sub-nanoliter volume on planar microl electrodes for extracellular potential recordings,” *Lab Chip*, vol. 4, pp. 357–362, 2004.
[19] D. DiCarlo and L. P. Lee, “Dynamic single-cell analysis for quantitative biology,” *Anal. Chem.*, vol. 78, no. 23, pp. 7918–7925, Dec. 2006.
[20] M. Leslie, “The power of one,” *Science*, vol. 331, no. 6013, pp. 24–26, Jan. 2011.
[21] D. T. Chiu and R. M. Lorenz, “Chemistry and biology in femtoliter and picoliter volume droplets,” *Acc. Chem. Res.*, vol. 42, no. 5, pp. 649–658, May 2009.
[22] R. N. Zare and S. Kim, “Microfluidic platforms for single-cell analysis,” *Annu. Rev. Biomed. Eng.*, vol. 12, pp. 201–224, 2010.
[23] H. Wu, A. R. Wheeler, and R. N. Zare, “Chemical cytometry on a picoliter-scale integrated microfluidic chip,” *Proc. Nat. Acad. Sci. U.S.A.*, vol. 101, no. 35, pp. 12 809–12 813, Aug. 2004.
[24] M. He, J. S. Edgell, G. D. M. Jeffries, R. M. Lorenz, J. P. Shelly, and D. T. Chiu, “Electrically induced splitting of single cells and subcellular organelles into picoliter- and femtoliter-volume droplets,” *Anal. Chem.*, vol. 77, no. 6, pp. 1539–1544, Mar. 2005.
[25] D. Irimia, R. G. Tompkins, and M. Tozer, “Single-cell chemical lysis in picoliter-scale closed volumes using a microfabricated device,” *Anal. Chem.*, vol. 76, no. 20, pp. 6137–6143, Oct. 2004.
[26] L. Cai, J. Friedman, and S. Xie, “Stochastic protein expression in individual cells at the single molecule level,” *Nature*, vol. 440, no. 7082, pp. 358–362, Mar. 2006.
[27] S.-Y. Teh, R. Lin, L.-H. Hung, and A. P. Lee, “Droplet microfluidics,” *Lab Chip*, vol. 8, no. 2, pp. 198–220, 2008.
[28] B.-S. Song, D. Chatterjee, and R. L. Garrell, “Electrically-induced splitting and generation of sub-picoliter droplets in air on a digital microfluidic device,” in *Proc. Int. Conf. pTAS*, Tokyo, Japan, 2006, pp. 128–130.
[29] P. Y. Chion, S.-Y. Park, and M. C. Wu, “Continuous optoelectrowetting for picoliter droplet manipulation,” *Appl. Phys. Lett.*, vol. 93, no. 22, pp. 221101–221110, Dec. 2008.
[30] D. Chatterjee, H. Shepherd, and R. L. Garrell, “Electromechanical model for actuating liquids in a two-plate droplet microfluidic device,” *Lab Chip*, vol. 9, no. 9, pp. 1219–1229, 2009.
[31] T. S. Sammarco and M. A. Burns, “Thermocapillary pumping of discrete drops: A distributed-analysis devices,” *AIChE J.*, vol. 45, no. 2, pp. 350–366, Feb. 1999.
[32] M. K. Chaudhury and G. M. Whitesides, “How to make water run uphill,” *Science*, vol. 256, no. 5063, pp. 1539–1541, Jun. 1992.
[33] N. L. Abbott and G. M. Whitesides, “Potential-dependent wetting of aqueous solutions on self-assembled monolayers formed from 13-(ferrocenylcarbonyl)pentadecanethiol on gold,” *Langmuir*, vol. 10, pp. 1493–1497, 1994.
[34] A. A. Darhuber and S. M. Troian, “Principles of microfluidic actuation by modulation of surface stresses,” *Annu. Rev. Fluid Mech.*, vol. 37, pp. 425–455, 2005.
[35] T. B. Jones, J. D. Fowler, Y. S. Chang, and C.-J. Kim, “Frequency-based real-time drop of electrowetting and dielectrophoretic liquid microactuation,” *Langmuir*, vol. 19, pp. 7646–7651, 2003.
[36] T. B. Jones and K.-L. Wang, “Frequency-dependent electromechanics of aqueous liquids: Electrowetting and dielectrophoresis,” *Langmuir*, vol. 20, pp. 2813–2818, 2004.
[37] T. B. Jones, R. B. Fair, M. G. Pollack, and E. J. Shaughnessy, “Dynamics of electrowetting and droplet transport,” *Sens. Actuators B, Chem.*, vol. 87, no. 1, pp. 201–206, Nov. 2002.
[38] R. M. Pashley, M. Rzechorzewicz, L. R. Pashley, and M. J. Francis, “Degassed water is a better cleaning agent,” *J. Phys. Chem. B*, vol. 109, no. 3, pp. 1251–1254, Jan. 2005.
[39] T. S. Sammarco, H. Morinaga, “Single-cell microfluidic impedance cytometry: A review,” *Microfluid. Nanofluid.*, vol. 8, no. 4, pp. 423–443, Apr. 2010.
[40] T. Ward, M. Faivre, M. Abkarian, and H. A. Stone, “Microfluidic flow focusing: Drop size and scaling in pressure versus flow-rate-driven pumping,” *Electrophoresis*, vol. 26, no. 19, pp. 3718–3724, Oct. 2005.
[41] M. A. McQuain, C. T. Callison, S. C. Jacobson, and J. M. Ramsey, “Flow cytometry of *Escherichia coli* on microfluidic devices,” *Anal. Chem.*, vol. 73, no. 21, pp. 5334–5338, Nov. 2001.
[42] M. Dhindua, J. Heikenfeld, S. Kwon, S. Park, J. Park, P. D. Rack, and I. Papautsky, “Virtual electrowetting channels: Electronic liquid transport with continuous channel functionality,” *Lab Chip*, vol. 10, no. 7, pp. 852–836, 2010.
**Wyatt C. Nelson** was born in Seattle, WA, in 1983. He received the B.S. degree in mechanical engineering from Cornell University, Ithaca, NY, in 2006, and the Ph.D. degree in mechanical engineering from the University of California at Los Angeles (UCLA), in 2011.
He is currently with the Micro and Nano Manufacturing Laboratory at UCLA. His research interests include digital microfluidics and electrowetting-on-dielectric.
**Chang-Jin “CJ” Kim** (S’89–M’91) received the B.S. degree from Seoul National University, Seoul, Korea, the M.S. degree from Iowa State University, Ames, and the Ph.D. degree from the University of California, Berkeley, in 1991, all in mechanical engineering.
Since joining the faculty at the University of California at Los Angeles (UCLA), in 1993, he has developed several microelectromechanical systems (MEMS) devices and methodologies. MEMS is a major field in the Mechanical and Aerospace Engineering Department in 1997. Directing the Micro and Nano Manufacturing Laboratory, he is also a founding member of the California NanoSystems Institute at UCLA. His research interests are in MEMS and nanotechnology, including design and fabrication of micro/nano structures, actuators, and systems, with a focus on the use of piezoelectric materials.
Prof. Kim has served on numerous technical committees and panels, including Transducers, the IEEE International Conference on MEMS, and the National Academies Panel on Benchmarking the Research Competitiveness of the U.S. in Micro and Nanotechnology. He is currently serving on the Editorial Advisory Board for the *IEEE Transactions on Electronics and Electronic Engineering* and on the Editorial Board for the *Journal of Microelectromechanical Systems*. An ASME Fellow, he was the recipient of the Graduate Research Excellence Award from Iowa State University, the 1995 TRW Outstanding Young Teacher Award, a 1997 NSF CAREER Award, the 2002 ALA Achievement Award, and the 2008 Samuell Outstanding Teaching Award. He has also been active in the commercial sector as a board member, scientific advisor, consultant, and founder of start-up companies. |
Analyzing Schema.org
Peter F. Patel-Schneider
Nuance Communications
firstname.lastname@example.org
Abstract. Schema.org is a way to add machine-understandable information to web pages that is processed by the major search engines to improve search performance. The definition of schema.org is provided as a set of web pages plus a partial mapping into RDF triples with unusual properties, and is incomplete in a number of places. This analysis of and formal semantics for schema.org provides a complete basis for a plausible version of what schema.org should be.
Schema.org\textsuperscript{1} “provides a collection of schemas, i.e., html tags, that webmasters can use to [mark up] their pages in ways recognized by major search providers.”\textsuperscript{2} The major search engine providers, including Bing, Google, Yahoo!, and Yandex use schema.org markup to improve the display of search results and schema.org has been designed by and is controlled by these organizations. This makes schema.org markup an important kind of machine-understandable data in the web. Not only are there many web pages with schema.org information, but this information is used in important ways.
Aside from being a collection of schemas, schema.org is a language for representing information on the Web, different from other languages used for this purpose, such as RDF \cite{1, 2}, OWL \cite{3, 4}, and the language underlying Freebase \cite{5}. Using this language, the schema.org schemas are organized into a simple taxonomy by generalization relationships and other ontological aspects of schema.org information are specified.
The publicly available definition of schema.org is, however, incomplete and contradictory. It is only provided as English text on various web pages in \textit{schema.org}, plus mappings of the collection of schemas\textsuperscript{3} into RDF (\url{http://schema.org/docs/full_md.html}) and OWL (\url{http://schema.org/docs/schemaorg.owl}). The RDF mapping centrally uses non-RDFS properties, such as \url{http://schema.org/domainIncludes}, so it is not possible to determine the meaning of schema.org constructs from the RDF mapping. The OWL mapping is somewhat better, as domains and ranges employ OWL unions, but the mapping is only a translation of part of what defines schema.org. The lack of a complete definition of schema.org limits the possibility of extracting the correct information from web pages that have schema.org markup.
This paper provides a full basis for schema.org as it should be, filling in the holes in the available descriptions of schema.org and fixing up discrepancies. The paper provides both a pre-theoretic analysis of schema.org and an abstract syntax and formal model-theoretic semantics for schema.org. This paper does not, however, draw on
\textsuperscript{1} Throughout this paper schema.org refers to the general idea and \textit{schema.org} refers to the collection of documents available at the \url{https://schema.org} web site.
\textsuperscript{2} From \url{https://schema.org}, as of 1 April 2014.
\textsuperscript{3} See \url{http://schema.org/docs/datamodel.html}.
the use of schema.org on web pages. Researchers can use the basis provided here to further investigate the properties of schema.org and schema.org markup. Providers of schema.org data can use this basis to reliably determine the meaning of the schema.org data they create. Developers can use this basis to build software that uses schema.org markup as information in a way that is compatible with the description of schema.org.
**Description of Schema.org at schema.org**
The description of schema.org in this section of the paper is taken from information on the web pages in *schema.org*, as of 1 April 2014. It ignores most of the surface syntax aspects of schema.org, concentrating on the underlying concepts and their intent.
Schema.org information is about items, e.g., the movie *Avatar*. Items can have types, e.g., the type identified by the URL `http://schema.org/Movie`. Items can have associated property-value pairs, e.g., the property identified by `http://schema.org/director` with value “*James Cameron*”. The value in a property value pair can be text, i.e., a Unicode string; a literal, e.g., a number or date; a URL, which identifies an item; or another item. There is no requirement that properties have only a single value for an item. Items can have associated URLs, e.g., `http://www.avatarmovie.com/index.html` and `http://en.wikipedia.org/wiki/Avatar_(2009_film)`, each of which identifies the item.
Schema.org provides a collection of types, via pages in *schema.org*, organized in a multi-parent generalization hierarchy. Each type is identified by the URL of the page that provides its definition. Each type has a set of parents, i.e., more-general types. Each type, except for datatypes, has a set of allowable properties for the type.
The types that are more specific than `http://schema.org/Enumeration` are enumeration types that also specify a set of URLs identifying all the items that are instances of the type. Datatypes are the types more specific than `http://schema.org/Datatype` and implicitly provide a set of non-item data values for them and a mapping from text to these values.
Schema.org also provides a collection of properties, again from *schema.org*, which may be also organized in a multi-parent generalization hierarchy.\(^4\) Each property is identified by the URL of the page that provides its definition. Each property may have one or more types as domains, and can be used on items belonging to any of these types. Each property has one or more types as ranges, and values for the property belong to one or more of these types. However, property values can always be provided as just text.
There is a description of an extension mechanisms for schema.org, which only permits very simple extensions. It appears that the extension mechanism exists only to further subdivide existing schema.org properties, classes, and enumeration items and that these extensions are ignored within schema.org.
The translations of the type and property definitions of schema.org into RDF and OWL abide by the above description, except that there is no translation for the property hierarchy. These translations provide no extra information beyond what is given here.
\(^4\)At the time of writing of this paper there was no general notation of the property hierarchy. While this paper was in review, the property hierarchy was officially announced ([http://lists.w3.org/Archives/Public/public-vocabs/2014Jun/0095.html](http://lists.w3.org/Archives/Public/public-vocabs/2014Jun/0095.html)).
Analysis of schema.org as a Description of Schema.org
There are quite a number of aspects of schema.org and schema.org markup that are left unspecified in schema.org, are unclear, or raise issues. This section describes these aspects and provides extra assumptions that will be used in the account for schema.org presented here. The extra assumptions have been made in a way that is congruent with the information on schema.org, that make sense in an environment where there are large central consumers of large amounts of data, and that generate a reasonable representation formalism. (In several places, the comments in schema.org do not match the actual class or property, for example, instances of http://schema.org/StructuredValue are not strings, but this sort of mismatch is not the subject of this paper.)
It is unclear whether types and properties can also be items. However, items work quite differently from types and properties, and having arbitrary web pages being able to modify the types and properties of schema.org leads to difficulties, such as not being able to determine when a property is valid for an item until after all item information has been processed, so this account treats types and properties as being different from items. In particular, in this account different URLs that identify the same item do not identify the same type or the same property. Data values also act differently from items, so this account treats them as being disjoint from types, properties, and items. The identifiers of types and properties are different in schema.org, as URLs for types have initial capitals and URLs for properties do not, so it is fairly obvious that types are disjoint from properties.
Schema.org uses URLs as identifiers. URLs can be used to retrieve web pages, and this aspect of URLs is a main basis of schema.org. URLs officially can include fragment ids, and such URLs then identify parts of web pages. Although fragment identifiers are not currently used for any types and properties in schema.org, there is nothing technical preventing their use, and so they will be allowed in the account herein for types, properties, and items.
It is unclear whether schema.org types and properties must be identified by URLs in schema.org, but all current schema.org types and properties are so identified. This account does not formally make the assumption that types and properties must be identified by URLs in schema.org, but some of the pragmatic analysis does make the assumption that type and property definitions change infrequently, as is the case for types and properties identified by URLs in schema.org.
The mechanisms for working with datatypes are underspecified in schema.org. This account adds in a formal mechanism for determining the set of values for a datatype and a formal method for determining the data value corresponding to a text string for the datatype.
The name of the most general datatype in schema.org is http://schema.org/Datatype. This is an unfortunate name—http://schema.org/Literal would be much better—but the schema.org name will be used in this account. The name of the datatype for floating point numbers in schema.org is http://schema.org/Float, http://schema.org/Float and http://schema.org/Integer both have generalization http://schema.org/Number. This can lead to problems because floating point numbers are imprecise whereas integers are precise. This account, however, does not address the issue.
It is unclear whether the instances of an enumeration have to be items, or can also be data values. This account assumes that the instances of an enumeration are given as URLs, as is the case for all examples currently in schema.org, and thus that instances of an enumeration are items, not data values.
Some examples in schema.org only make sense if different but similar URLs identify different items. This is particularly the case for URLs that make up enumerations. This account assumes that different URLs in an enumeration identify different items, but does not otherwise assume that different URLs in the same namespace, e.g., different Wikipedia URLs, or in the same document identify different items. This extra assumption would be easy to add.
The domains of a property are specified both as part of types and as part of properties in schema.org. In all the examples there is no divergence between the two specifications, but the possibility of divergence is not ruled out. This account treats the specification in the type as the actual specification, as that seems to make more sense for disjunctive domains.
Because several properties indicate that they are subproperties of other properties, this account incorporates a multi-parent property hierarchy. There are some additions to the account herein that have to be made to support the property hierarchy.
Both domains and ranges of properties are disjunctive. This is different from most other representation formalisms, such as description logics [6] and RDF [2]. The stated rationale for this decision is that it reduces the need for general types that exist only to be domains or ranges. However, disjunctive domains and ranges mean that additions to a collection of schema.org information can be non-monotonic. The disjunctive nature of domains and ranges is fully explored in this account, including how it interacts with the property hierarchy.
Several aspects of the predominant syntaxes for schema.org markup obscure the workings of schema.org. This account transforms these aspects of surface syntax into a different abstract syntax.
Several types and properties are used as part of the foundations of schema.org in schema.org. Nearly all uses of these types and properties as general types or properties undermines the foundations of schema.org, so their use is disallowed in this account. The extension mechanism for schema.org is of very limited utility and appears to not have any effect on the processing of schema.org markup, so it is ignored in this account.
**Description of Schema.org as It Should Be**
This section contains a pre-theoretic description of schema.org and schema.org content as it should be, consonant with the discussion in the previous section. This description is designed to say how schema.org could work in a way that can be easily turned into a formal definition of schema.org, as is done in the following section of this paper.
Throughout this account, a URL is a uniform resource locator, optionally including a fragment part. The document (fragment) at that URL is (the appropriate fragment of) the document obtained by the usual web mechanisms for retrieving a document given a URL. URLs will be generally written as CURIES [7], with the prefixes s expanding
to http://schema.org/ and w expanding to http://en.wikipedia.org/wiki/, and the prefixes rdf, rdfs, and owl expanding to their usual expansions. The constituents of schema.org information are types, properties, data values, and items.
There is a collection of types, in a multi-parent generalization taxonomy, with two roots, s:Thing and s:Datatype. Each type is identified by a unique URL. The document (fragment) at that URL defines the type, listing:
1. some types that are more general than it (its parents), and
2. for non-datatypes, its properties (see below).
Parents and properties, and information about instances where appropriate, are the only information about a type obtainable from its defining document (fragment).
Each type has as a generalization (not necessarily directly specified in its defining document) either s:Thing or s:Datatype, but not both.
The types with strict generalization s:Datatype are datatypes. All the data values belonging to the datatype are described in the datatype’s defining document (fragment), as is a way of transforming text strings into these data values. The datatypes are s:Boolean, s:Date, s:DateTime, s:Number, s:Float, s:Integer, s:Text (Unicode strings), s:URL, and s:Time. The details of these datatypes do not matter for this account, except for s:Text, and are not described here.
The type s:Enumeration has s:Thing as a parent.\(^5\) Those types with strict generalization s:Enumeration are enumeration types. All those items with the enumeration type as a direct type are listed in the type’s defining document (fragment). Different URLs identify different items in an enumeration.
The type s:Thing has properties s:description and s:name.\(^6\)
There is a collection of properties, disjoint from types, in a multiple-parent generalization taxonomy with multiple roots. Each property is identified by a unique URL. The document (fragment) at that URL defines the property, providing:
1. types that its values belong to (its ranges), and
2. some properties that are more general than it (its parents).
Ranges and parents are the only information about a property obtainable from its defining document (fragment).
For each parent of the property for each range of the property the parent must have a range that is the same as or a generalization of the range. This condition on property ranges means that the validity of a property value can be checked by looking only at the range types of the property itself.
The properties s:description and s:name both have range s:Text.
Data values belong to one or more datatypes, and are disjoint from types and properties. Data values are written as a combination of a URL identifying a datatype and a
\(^5\) Enumeration actually has a different supertype on schema.org but this account removes the unneeded supertype.
\(^6\) There are several other properties for s:Thing on schema.org, but these do not play a role in this account and are ignored here.
text string. The mapping in the datatype turns the text string into a value of the datatype. Every data value belongs to \( s:\text{Datatype} \). If a data value belongs to a datatype then it belongs to the parents of the datatype.
Items are things in the world, including information things, and are disjoint from types, properties, and data values. Items belong to (one or more) non-datatype types. Items have zero or more URLs identifying them. Items are associated with (other) items and data values via properties. Every item belongs to \( s:\text{Thing} \). If an item belongs to a type then it belongs to the parents of the type.
If an item or data value is associated with an item via a property then the item or data value is also associated with the item via each parent of the property. For each item or data value associated with an item via a property,
1. one of the item’s types has the property as one of its properties, and
2. the item or data value belongs to one of the ranges of the property.
The documents (document fragments) at the URLs identifying an item provide information about the item, including types for the item as well as items and data values associated with the item via properties.
Bare text can be used as if it was the value for any property. If the property does not have \( s:\text{Text} \) or \( s:\text{Datatype} \) as one of its ranges, but does have one or more datatypes as a range that have a data value that can be written as the bare text then the actual value for the property is one of these data values. If the property does not have \( s:\text{Text} \) or \( s:\text{Datatype} \) as one of its ranges, and does not have any suitable datatypes as a range, but does have one or more non-datatypes as a range, then the actual value for the property is some item that has a type that is one of these ranges and this item has the text as a value of its \( s:\text{description} \) property. (The property \( s:\text{description} \) is used instead of \( s:\text{name} \), as the text might not truly be a name for the value.) Otherwise the actual value for the property is the bare text itself.
Any surface syntax must provide ways to write all possible data values (as long as they are not too big). Any surface syntax must have ways to provide items with any number of types, including none, and values for any property of any of the provided types or their generalizations or \( s:\text{Thing} \), including allowing multiple values for a property. Any surface syntax must provide ways for writing items with no identifying URLs. Any surface syntax must specially process syntax that would otherwise produce values for \( s:\text{additionalType} \), turning the values into types; and \( s:\text{url} \) and \( s:\text{sameAs} \), turning the values into identifying URLs.
The following URLs are not used to identify types or properties: \( s:\text{Class} \), \( s:\text{Property} \), \( s:\text{domainIncludes} \), \( s:\text{rangeIncludes} \), \( \text{rdfs:subClassOf} \), \( \text{rdfs:subPropertyOf} \), \( \text{rdfs:domain} \), \( \text{rdfs:range} \), \( \text{rdfs:type} \), \( \text{rdfs:Class} \), \( \text{rdf:Property} \), and \( \text{owl:Class} \). If they are used in a surface syntax to provide information about an item they and their values must be ignored. The following URLs are not used to identify properties: \( s:\text{additionalType} \), \( s:\text{url} \), and \( s:\text{sameAs} \).
Formal Definition for Schema.org
This definition for schema.org defines an abstract syntax for schema.org, abstracting away from the details of the various surface syntaxes, and a model-theoretic semantics, that provides a formal meaning for schema.org. It conforms to the pre-theoretic description above.
Abstract Syntax
Surface syntaxes for schema.org are transformed into an abstract syntax, in a process not fully described here. The abstract syntax plays a similar role as triples do for RDF [2], but is more complicated, as it makes distinctions between definitions and information about items. The abstract syntax removes artifacts of the surface syntaxes that make a formal account difficult, but transforming schema.org surface syntaxes into this abstract syntax is simple.
The gathering of information from Web documents is performed when building the abstract syntax. Constructs in the abstract syntax that start with a URL (or set of URLs) may be constructed from the document at the URL (or documents at one or more of the URLs), although they need not be.
**Definition 1.** A URL in this document is a URL with optional fragment identifier, as defined in the W3C Working Draft on URLs [8]. A text string is a sequence of Unicode characters [9]. A literal is a pair consisting of a URL, the datatype identifier of the literal, and a text string.
One part of schema.org information consists of definitions—of regular types, of enumerated types, of datatypes, and of properties.
**Definition 2.** A (regular) type definition is a triple, \( \langle U, S, P \rangle \), where \( U \) is a URL, the identifier of the type; \( S \) is a set of URLs, the supertypes of the type; and \( P \) is another set of URLs, the properties of the type.
For example, the type definition for movies could be\(^7\)
\[
\langle s:Movie, \{ s:CreativeWork \}, \\
\{ s:actor, s:director, s:producer, s:duration, s:musicBy, \\
s:productionCompany, s:trailers:author, s:copyrightYear \} \rangle
\]
indicating that movie is a subtype of creative work and has eight locally-specified properties.
The type of actions that update a collection would be
\[
\langle s:UpdateAction, \{ s:Action \}, \{ s:collection \} \rangle
\]
indicating that update actions are actions and have only one locally-specified property.
---
\(^7\) This definition of movies ignores the legacy properties for movies in schema.org, and adds in some properties from \( s:CreativeWork \) for illustrative purposes.
**Definition 3.** An enumerated type definition is a quadruple, \( \langle U, S, P, E \rangle \), where \( U \), \( S \), and \( P \) are as in a regular type definition, and \( E \) is yet another set of URLs, the direct instances of the enumerated type.
For example, the enumerated type definition for book formats would be
\[
\langle s:BookFormatType, \{s:Enumeration\}, \{\},
\{s:EBook, s:Hardcover, s:Paperback\} \rangle
\]
indicating that there are three different book formats.
**Definition 4.** A datatype definition is a quadruple, \( \langle U, S, W, I \rangle \), where \( U \) and \( S \) are as in a regular type definition, \( W \) is a set of values for the datatype, and \( I \) is a partial mapping from text strings into \( W \).
For example, the datatype for URLs would be
\[
\langle s:URL, \{s:Text\}, U, id^U \rangle
\]
where \( U \) is the set of Unicode strings that are valid URLs and \( id^U \) is the string identity function restricted to valid URLs, indicating that URL values are those text strings that are valid URLs.
It is a bit unusual to include formal datatype definitions in an abstract syntax. It would be more usual to pull these out into some sort of side definition. However, this way of defining datatypes puts all aspects of the definition in one place, as is done for other type and property definitions.
**Definition 5.** A property definition is a triple, \( \langle U, S, R \rangle \), where \( U \) is a URL, the identifier of the property; \( S \) is a set of URLs, the superproperties of the property; and \( R \) is a non-empty set of URLs, the ranges of the property.
For example, the \( s:collection \) property (used in update actions), a subproperty of the \( s:object \) property with range \( s:Thing \), would be written as the first of the following property definitions:
\[
\langle s:collection, \{s:object\}, \{s:Thing\}\rangle,
\langle s:actor, \{\}, \{s:Person\}\rangle,
\langle s:director, \{\}, \{s:Person\}\rangle,
\langle s:copyrightYear, \{\}, \{s:Number\}\rangle,
\langle s:author, \{\}, \{s:Organization, s:Person\}\rangle.
\]
**Definition 6.** A type definition is either a regular type definition, an enumerated type definition, or a datatype definition. A definition is either a type definition or a property definition. A definition is said to be a definition for its identifier.
Type and property definitions define separate generalization partial orders on types and properties, building up from the supertypes and superproperties of type definitions and property definitions.
**Definition 7.** A type (property) definition, $D$, is a child of another, $D'$, if $D'$ is one of the supertypes (superproperties) of $D$. A type (property) definition, $D$, is a descendant of another, $D'$, in a set of definitions, if there is a chain of child relationships from $D$ to $D'$, in the set of definitions.
**Definition 8.** A URL, $U$, is a type (property) descendant of another, $U'$, in a set of definitions, written $U < U'$, if $U$ is the identifier of a type (property) definition that is a descendant of a type (property) definition that is identified by $U'$.
The other part of schema.org information consists of information about items, providing types and property values for items. Note that this information is not called definitions, as there is no requirement that different items in the abstract syntax provide information about different resources.
**Definition 9.** A property value pair, $\langle U, V \rangle$, consists of $U$, a URL identifying the property, and $V$, a text string or a literal or a URL or an item, indicating the value.
**Definition 10.** An item is a triple, $\langle N, T, PV \rangle$, where $N$ is a (possibly empty) set of URLs, identifiers of the item; $T$ is a (possibly empty) set of URLs, identifiers of types of the item; and $PV$ is a (possibly empty) set of property value pairs.
For example, an item for a particular movie could be
\[
\langle \{ http://www.avatarmovie.com/index.html, w:Avatar_(2009_film) \}, \\
\{ s:Movie \}, \\
\{ \langle s:name, "Avatar" \rangle , \\
\langle s:director, \langle \{ \}, \{ s:Person \}, \{ \langle s:name, "James_Cameron" \rangle \} \rangle \rangle , \\
\langle s:actor, "Sam_Worthington" \rangle , \\
\langle s:actor, w:Sigourney_Weaver \rangle \} \\
\langle s:year, "2009" \rangle , \\
\langle s:author, "James_Cameron" \rangle \rangle
\]
This item has two identifiers and six property-value pairs, one providing a text value for a name of the movie, one providing an in-line item for a director of the movie, one providing a text value for an actor in the movie, one providing a URL identifying an actor in the movie, one providing a text value for a copyright year of the movie, and another providing a text value for an author of the movie.
A collection of definitions and items is a knowledge base, the overall way of collecting schema.org information together. There are many side conditions on schema.org knowledge bases to provide an overall structure of the generalization hierarchies for types and properties, to account for the built-in types and properties, and to ensure that literals are well behaved.
**Definition 11.** A schema.org knowledge base is a triple, $\langle T, P, I \rangle$, where $D$ is a set of type definitions, $P$ is a set of property definitions, and $I$ is a set of items that satisfies the following conditions:
1. Each URL is the identifier of at most one definition in $T$, and similarly for $P$. There is at most one definition for any URL in $T$ and $P$.
2. The descendant relationship for types (properties) in $T$ ($P$) is a strict partial order.
3. $T$ contains the following regular type definitions:
\[
\langle s:\text{Thing}, \{\}, \{s:\text{description}, s:\text{name}\} \rangle,
\]
\[
\langle s:\text{Datatype}, \{\}, \{\} \rangle, \text{ and}
\]
\[
\langle s:\text{Enumeration}, \{ s:\text{Thing}\}, \{\} \rangle.
\]
4. $T$ contains the following datatype definition, where $S$ is the set of text strings and $id$ is the identity mapping on text strings:
\[
\langle s:\text{Text}, \{s:\text{Datatype}\}, S, id \rangle
\]
5. $P$ contains the following property definitions:
\[
\langle s:\text{description}, \{\}, \{s:\text{Text}\} \rangle \text{ and } \langle s:\text{name}, \{\}, \{s:\text{Text}\} \rangle.
\]
6. For each literal $(U, V)$ in the knowledge base there is a datatype definition $\langle U, S, W, I \rangle$ in $T$ such that $I$ is defined on $V$.
7. $T$ has a datatype definition for $U$ iff $U < s:\text{Datatype}$ in $T$.
8. $T$ has an enumerated type definition for $U$ iff $U < s:\text{Enumeration}$ in $T$.
9. $T$ has a regular type definition for $U$ iff $U$ is $s:\text{Datatype}$ or $s:\text{Thing}$ or $U < s:\text{Thing}$ but not $U < s:\text{Enumeration}$ in $T$.
There is nothing about web document retrieval in the abstract syntax (nor in the formal semantics immediately following). The intent should be clear, however—that type and property definitions come from the document (fragment) obtained by dereferencing the URL identifying the type or property and that item information often comes from the document (fragment) obtained by dereferencing a URL identifying the item.\footnote{Note that types and properties have a unique identifier whereas items may have multiple identifiers, or none.}
Building a schema.org knowledge base then generally starts with a collection of web documents that have schema.org markup about items. However, first the pages in \textit{schema.org} that define types and properties are parsed to produce type and property definitions for the knowledge base. The document (fragments) accessible from URLs identifying items encountered during this parsing are added to the initial collection of web documents. Then this collection of documents is parsed to produce items for the knowledge base.
URLs for items that are encountered during the parsing may be used to direct that the web document (fragment) at that URL also be parsed to produce items for the knowledge base. Whether this “follow your nose” behavior is actually performed during the construction of a particular knowledge base depends on many factors that are outside the purview of this account.
\subsection*{Model-theoretic Semantics}
The semantics for schema.org here is built up in the standard way from interpretations, which provide formal meanings for all URLs as identifying types or properties, and items and URLs as identifying resources. Items are mapped into sets of resources, not resources, as the resource corresponding to an item is indeterminate unless there is a URL that identifies the item. A single URL can independently identify a type, a property, and a resource, but these do not have any formal connection between them.
**Definition 12.** An interpretation is a sextuple, \( \langle I_R, I_V, I_T, I_P, I_U, I_I \rangle \), where \( I_R \) is a set of resources; \( I_V \) is a set of data values, disjoint from \( I_R \); and
\[
\begin{align*}
I_T : U & \rightarrow 2^{I_R} \cup 2^{I_V} \\
I_P : U & \rightarrow 2^{I_R \times (I_R \cup I_V)} \\
I_U : U & \rightarrow I_R \\
I_I : I & \rightarrow 2^{I_R} \\
I_T(\text{s:Thing}) &= I_R \\
I_T(\text{s:Datatype}) &= I_V
\end{align*}
\]
where \( U \) is the set of URLs with optional fragments and \( I \) is the set of items.
\( I_T \) maps types into their extensions, a set of resources or a set of data values. \( I_P \) maps properties into their extensions, sets of pairs whose first element is a resource and whose second element is a resource or a data value. \( I_I \) maps items into their extensions, a set of resources. \( I_U \) maps URLs into their extensions as item identifiers, a resource;
Although the mappings above are infinite, in any knowledge base the only part of the mappings that are relevant are for the URLs and items that occur in the knowledge base (or query, in entailment and querying situations).
**Definition 13.** An interpretation, \( \langle I_R, I_V, I_T, I_P, I_U, I_I \rangle \), satisfies a knowledge base, \( \langle K_T, K_P, K_I \rangle \) iff
1. for \( \langle U, \{S_1, ..., S_n\}, P \rangle \) a regular type definition in \( K_T \), \( I_T(U) \subseteq I_T(S_i) \);
2. for \( \langle U, \{S1, ..., Sn\}, P, \{E1, ...Em\} \rangle \) an enumerated type definition in \( K_T \),
- \( I_T(U) \subseteq I_T(S_i) \),
- \( I_T(U) = \{I_U(e) \mid e \in E\} \), and
- \( \forall e_i \neq e_j \in E \quad I_U(e_i) \neq I_U(e_j) \),
where \( E = \bigcup \{E' \mid (U', S', P', E') \in K_T \land (U' = U \lor U' < U)\} \);
3. for \( \langle U, \{S_1, ..., S_n\}, W, I \rangle \) a datatype definition in \( K_T \),
- \( I_T(U) \subseteq I_T(S_i) \) and
- \( I_T(U) = W \);
4. for \( \langle U, \{S1, ..., Sn\}, R \rangle \) a property definition in \( K_P \), \( I_P(U) \subseteq I_P(S_i) \);
5. for \( I = \langle \{U_1, ..., U_n\}, \{T_1, ..., T_m\}, \{(P_1, V_1), ..., (P_l, V_l)\} \rangle \), an item in \( K_I \),
- \( I_I(I) = \{I_U(U_i)\}, \text{for } 1 \leq i \leq n \),
- \( I_I(J) \subseteq I_T(T_i), 1 \leq i \leq m \),
- for \( x \in I_I(I), \text{for } 1 \leq i \leq l, \text{there exists } \langle P_i, S, R \rangle \rangle \) a property definition and there exists \( y \) such that \( \langle x, y \rangle \in I_P(P_i) \) and
- if \( V_i \) is an item then \( y \in I_I(V_i) \)
- if \( V_i \) is a URL then \( y = I_U(V_i) \)
- if \( V_i \) is a literal \( \langle D, T \rangle \) then there exists \( \langle D, S, W, ID \rangle \) a datatype definition and \( y = I_D(T) \)
- if \( V_i \) is a text string then
- if \( \text{s:Text} \in R \) or \( \text{s:Datatype} \in R \) then \( y = V_i \)
- otherwise if there exists \( \langle D, S, W, I \rangle \) a datatype with \( D \) in \( R \) and \( V_i \) mapped by \( I \) then \( y = I_T(V_i) \) for some one of these datatypes
- otherwise if there exists \( \langle T, S, P \rangle \) a type with \( T \) in \( R \) then \( y \in I_T(T) \) for some one of these types, and \( \langle y, V_i \rangle \in I_P(\text{s:description}) \)
• otherwise $y = V_i$;
6. for each $U$, for each $\langle x, y \rangle \in I_P(U)$,
(a) there exists $\langle T, S, P \rangle$ a type definition or $\langle T, S, P, E \rangle$ an enumerated type definition in $K_T$ such that $U \in P$ and $x \in I_T(U)$, and
(b) there exists $\langle U, S, R \rangle$ a property definition in $K_P$ such that either $R = \{\}$ or there exists $R' \in R$ with $y \in I_T(R')$.
From this basis the standard notions of entailment and inference and simple querying can be defined in the usual way.
The first clause in the satisfaction definition above provides the basis for the type generalization hierarchy, saying that the extension of a regular type is a subset of the extensions of each of its parent types. This is repeated for enumerated types and datatypes in the first part of the second and third clauses. Because regular and enumerated types are all descendants of $s$:Thing, their extensions are subsets of the set of resources.
The second part of the second clause states that the extension of an enumeration is just the set of all the items that are stated to belong to it and its subtypes. The third part of the second clause states that all these items are different.
The third clause states that the extension of a datatype is its set of values. Because datatypes are all descendants of $s$:Datatype, their extensions must be subsets of the set of data values.
The fourth clause enforces the property generalization hierarchy.
The fifth clause handles all the parts of item syntax. The first part of this clause says that the extension of an item is the same as the extension of its identifiers, if any. Note that if there are no identifiers for an item, then the extension of the item need not be a singleton set. The second part says that the extension of an item is in the extension of the extensions of its types. The third part provides meaning for the property-value pairs in the item. There must be a property relationship from the item to a value that for item values is in the extension of the item, for URL values is the extension of the URL as an item identifier, and for literals is the correct data value. For values that are text there is a determination of what the most suitable ranges are with text datatypes the most suitable, other compatible datatypes next, and other types least suitable. Then one of these types is chosen and a data value or item is chosen to belong to this type.
So the movie item above would be a resource that is the same as the extensions of http://www.avatarmovie.com/index.html and w:Avatar_(2009_film) and is in the extension of $s$:Movie. This resource would be related to the string "Avatar" via (the property extension of) $s$:name, because $s$:name has $s$:Text as its sole range; to a resource in the extension of $s$:Person that has name "James Cameron" via $s$:director, to a resource in the extension of $s$:Person that has description "Sam Worthington" via $s$:actor, because $s$:director has $s$:Person as its sole range; to the extension of w:Sigourney_Weaver via $s$:actor, to the number 2009 via $s$:copyrightYear, because the supplied text is compatible with the $s$:Number datatype; and via $s$:author to a resource that is either in the extension of $s$:Organization or $s$:Person and that has description "James Cameron", because $s$:author has ranges $s$:Organization and $s$:Person, and no datatype ranges.
The sixth clause enforces domain and range restrictions. The first part says that for each property, for each relationship between an item and a value in that property, there is a regular or enumerated type that has that property and contains the item. The second
part similarly says that if the definition of the property states ranges, then the value belongs to one of the ranges. Because parent properties have to have a range that is an ancestor of this range, this also satisfies the range restriction for each ancestor property.
**Discussion**
The above formal semantics is quite dense, particularly the definition of satisfaction. However, there is nothing particularly sophisticated going on, it is just that there are quite a few bits of schema.org markup to take into account.
The formal semantics is actually more standard than the formal semantics of RDF [10], as there are no resources for types and properties. It is easy to see that nothing about items can affect the relationships within and between types and properties (except that, as usual, inconsistencies in the information about items cause the semantics to collapse). If two URLs identify the same item, it is not necessarily the case that the two URLs define the same type or define the same property. This stands in stark contrast to RDF, but means that the only source of information for a type or property is its definition (which would come from the appropriate *schema.org* page). Thus consumers of schema.org information do not need to process any items to understand types and properties.
If two items share an identifying URL, then their extension is the same. If an item does not have a URL, but has a type that is an enumeration, then the item is one of finite, enumerated instances of the enumeration. This provides a weak form of disjunction for schema.org. The distinctness of the extensions of the URLs in an enumeration provides inequality for arbitrary resources in schema.org.
As schema.org has weak disjunction and inequality for resources, its expressive power is considerably above that of RDFS, even though there is a translation provided from schema.org types and properties into RDFS.
Schema.org can, however, be translated into OWL, but the translation is not into a simple variant of OWL. For some parts of schema.org it is easy to see the translation. The special types $s$:Thing and $s$:Datatype are translated into $owl$:Thing and $owl$:Literal, respectively. All other non-datatype types become OWL classes. Supertypes where the parent type is a regular type translate into subtype axioms.
Property ranges for a property translate into a disjunctive property range axiom. However, a property can have both regular types and datatypes as ranges. A property with such a range cannot be categorized as either an object property or a data property, and so cannot be translated into OWL 2 DL. However, OWL 2 Full [11] permits uncategorized properties. Superproperties for properties then translate into subproperty axioms.
For some parts, the translation needs to take into account more than just one part of the knowledge base. Domains for a property are constructed by taking all the types that mention the property and producing a disjunctive property domain for the property from them. Enumerations are constructed by finding all the item URLs belonging to the enumeration type and its subtypes and constructing an axiom stating that the enumeration type is equivalent to the object one-of containing all these objects. This construction handles the supertype relationship where the supertype is an enumeration
type. As well, a different individuals axiom is added stating that all the distinct URLs belonging to the enumeration type and its subtypes are different.
The translation for datatypes requires the construction of a datatype map that has the same effect as the datatype definitions. The datatypes in schema.org fit within what can be done for OWL datatypes, so this is possible. Supertypes for datatypes are either true, because the datatype’s value spaces are in the correct relationship, or false. The first case can be ignored, as it has no effect. The second case can be translated into an inconsistent OWL assertion, as it produces an inconsistency.
The translation for items is a bit tricky, to allow for items that do not have any associated URLs. Anonymous individuals are employed in the translation to avoid the need for extra URLs, for each item there is generated a different anonymous individual.
For each identifier of the item, there is a same-as assertion between the anonymous individual and the identifier. For each type of the item, there is a class assertion stating that the anonymous individual belongs to the type.
Property value pairs are treated as follows.
1. If the value is a URL, then it identifies an item, and the URL is used directly as the value of a property assertion from the anonymous individual to the value via the property.
2. If the value is a literal, then the corresponding OWL literal is used instead in the property assertion.
3. If the value is an item, then the item is translated, and the anonymous individual for the item is used instead.
4. If the value is text, the situation is more complex. First then the ranges for the property are determined.
(a) If $s:Text$ or $s:Datatype$ is one of the ranges then a string literal is constructed from the text and used as the value as above.
(b) If there is a range that is a datatype with the text in the domain of its literal-to-value mapping, then those datatypes with the text in the domain of their literal-to-value mappings are used to construct one or more literals. The anonymous individual is then asserted to belong to a data some-value-from with the property to a data one-of constructed from these literals.
(c) If there are no suitable datatype ranges then if there are any non-datatype ranges then the translation of the property-value pair is a property assertion from the anonymous individual via the property to a fresh anonymous individual. This fresh anonymous individual is asserted to belong to the disjunction of the non-datatype ranges and is also asserted to have the text as the value of the $s:description$.
The correctness of this mapping is not hard to verify, but a full proof of the correctness would be long and tedious and so is omitted here.
The translation into OWL does not determine how hard or easy reasoning is in schema.org because reasoning in OWL Full is undecidable. There are no inverse properties in schema.org, so not making the division between object and data properties does not appreciably affect reasoning. As this is the only part of the mapping that is
not in OWL 2 DL, the mapping into OWL shows that the reasoning in schema.org is decidable.
It is easy to show that reasoning here is in PSpace, as reasoners need not introduce new types, properties, or items. Showing the precise complexity of reasoning is more difficult, as enumerations and the disjunctive nature of domain and range includes need to be addressed. For example, if this account of schema.org were modified to use the underlying semantics of RDF and lift the restrictions on the use of certain URLs the set cover problem could be encoded, introducing a new source of hardness to reasoning.
The intent of schema.org appears to be that all the types and properties defined in schema.org will remain in the schema.org namespace and thus under the control of the owners of schema.org and will change only infrequently. Other web pages will not be allowed to make changes or additions to these types and properties. This limits the effect of the non-monotonic nature of the disjunctive domains and ranges of schema.org.
The model-theoretic account here is a standard one, based on inference instead of constraints. If an item has a value for a property, then the item is inferred to belong to one of the domains of the property. The constraint reading [12] would instead require that the item be stated to belong to a domain for the property before a value could be provided. There are benefits to the constraint account, as it is closer to the database situation, but it is less flexible [13].
It is possible to get the effect of the constraint approach in a surface syntax. A surface syntax can have constructs that require that property-value pairs for an item be only for items that mention a type or subtype of one of the domains of the property. In this way most benefits of both approaches can be obtained.
**Conclusion**
This paper has provided an analysis of what schema.org should be, leading up to a complete formal treatment of schema.org including an abstract syntax and a model-theoretic semantics. It fills in voids in the publicly-available description of schema.org, including whether types and properties and items are disjoint, whether enumerations are distinct, whether properties can have generalizations, and how to handle text values. This may not exactly correspond to intent of the schema.org members, but it is consistent with the available information about schema.org, and uses only a reasonable set of additional assumptions.
This paper shows that even the unusual parts of schema.org can be translated into OWL. Although schema.org cannot be translated into OWL 2 DL, because schema.org properties cannot be categorized into object and data properties, the extensions are cosmetic, and schema.org reasoning is no harder than reasoning in OWL 2 DL. Determining just how hard schema.org reasoning is remains as further work.
Schema.org does not provide local ranges for properties, such as saying that the author of a movie is a person even though in general authors can be either people or organizations. This lack of expressive power limits what can be said about property values and is especially problematic as quite a few roles in schema.org could benefit from local ranges (e.g., a season of a TV series should be a TV season, but can only be
a general season, as Radio Series also have seasons). Adding this feature to schema.org would usefully improve its expressive power.
The account of schema.org here should provide a starting point for further formal analysis of schema.org and a firm foundation for systems that consume schema.org information. Web pages that contain schema.org information can be checked against this account to provide a formal account of what the information conveys, thus reducing the possibility of mismatches between providers and consumers of schema.org information.
The obvious next step is to gather large amounts of schema.org information to see how schema.org is used in practice. This usage should then be analyzed to see how well the various aspects of schema.org are used and how the account here helps to better provide meaning for actual information that uses schema.org.
References
1. Schreiber, G., Raimond, Y.: RDF 1.1 primer. W3C Working Group Note, http://www.w3.org/TR/rdf11-primer/ (25 February 2014)
2. Cyganiak, R., Wood, D., Lanthaler, M.: RDF 1.1 concepts and abstract syntax. W3C Recommendation, http://www.w3.org/TR/rdf-concepts/ (25 February 2014)
3. W3C OWL Working Group: OWL 2 Web Ontology Language: Document overview. W3C Recommendation, http://www.w3.org/TR/owl2-overview (11 December 2012)
4. Motik, B., Patel-Schneider, P.F., Parsia, B.: OWL 2 web ontology language: Structural specification and functional-style syntax. W3C Recommendation, http://www.w3.org/TR/owl2-syntax/ (11 December 2012)
5. Bollacker, K., Tufts, P., Pierce, T., Cook, R.: A platform for scalable, collaborative, structured information integration. In: Sixth International Workshop on Information Integration on the Web, Vancouver, British Columbia, AAAI Press (July 2007)
6. Baader, F., Calvanese, D., McGuinness, D.L., Nardi, D., Patel-Schneider, P.F., eds.: The Description Logic Handbook: Theory, implementation, and applications. 2nd edn. Cambridge University Press (2007)
7. Birbeck, M., McCarron, S.: CURIE syntax 1.0: A syntax for expressing compact URIs. W3C Working Group Note, http://www.w3.org/TR/curie/ (2007)
8. Arvidsson, E., Smith, M.: URL. W3C Working Draft, http://www.w3.org/TR/URL/ (24 May 2012)
9. The Unicode Consortium: The unicode standard. http://www.unicode.org/versions/latest/ (2013)
10. Hayes, P., Patel-Schneider, P.F.: RDF 1.1 semantics. W3C Recommendation, http://www.w3.org/TR/rdf11-mt/ (25 February 2014)
11. Schneider, M.: OWL 2 web ontology language: RDF-based semantics (second edition). W3C Recommendation, http://www.w3.org/TR/owl-rdf-based-semantics/ (11 December 2012)
12. de Bruijn, J., Polleres, A., Lara, R., Fensel, D.: OWL DL vs. OWL Flight: Conceptual modeling and reasoning for the semantic web. In: Proceedings of the 14th World Wide Web Conference, Japan (May 2005)
13. Patel-Schneider, P.F., Horrocks, I.: Position paper: A comparison of two modelling paradigms in the semantic web. In: Proceedings of the 15th International Conference on the World Wide Web (WWW2006), New York, NY, ACM Press (May 2006) |
OPPONENT MODELING IN POKER: LEARNING AND ACTING IN A HOSTILE AND UNCERTAIN ENVIRONMENT
by
John Aaron Davidson
A thesis submitted to the Faculty of Graduate Studies and Research in partial fulfillment of the requirements for the degree of Master of Science.
Department of Computing Science
Edmonton, Alberta
Fall 2002
The author has granted a non-exclusive licence allowing the National Library of Canada to reproduce, loan, distribute or sell copies of this thesis in microform, paper or electronic formats.
L’auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la forme de microfiche/film, de reproduction sur papier ou sur format électronique.
The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author’s permission.
L’auteur conserve la propriété du droit d’auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.
Name of Author: John Aaron Davidson
Title of Thesis: Opponent Modeling in Poker: Learning and Acting in a Hostile and Uncertain Environment
Degree: Master of Science
Year this Degree Granted: 2002
Permission is hereby granted to the University of Alberta Library to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only.
The author reserves all other publication and other rights in association with the copyright in the thesis, and except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatever without the author's prior written permission.
John Aaron Davidson
502-11111 87th Ave
Edmonton, AB, Canada, T6G 0X9
Date: Aug 20th 2002
The undersigned certify that they have read, and recommend to the Faculty of Graduate Studies and Research for acceptance, a thesis entitled **Opponent Modeling in Poker: Learning and Acting in a Hostile and Uncertain Environment** submitted by Aaron Davidson in partial fulfillment of the requirements for the degree of Master of Science.
---
Jonathan Schaeffer
Supervisor
Duane Szafron
Supervisor
Robert Holte
Witold Pedrycz
External Examiner
Date: Aug 20th 2002
all your base are belong to us
Artificial intelligence research has had great success in many classic games such as chess, checkers, and othello. In these perfect-information domains, alpha-beta search is sufficient to achieve a high level of play. However Artificial intelligence research has long been criticized for focusing on deterministic domains of perfect information – many problems in the real world exhibit properties of imperfect or incomplete information and non-determinism.
Poker, the archetypal game studied by game-theorists, is a challenging domain containing both imperfect information and non-deterministic elements. The random shuffling of cards provides a great deal of uncertainty. The cards held by an opponent or in the remaining deck provide the imperfect information. A poker player has only a small incomplete description of the entire game-state. Furthermore, opponents play deceptively in order to hide information about the cards they hold. These properties prevent traditional game-tree search methods from working. In order to build a high-performance poker playing agent, a multitude of difficult problems must be solved in machine learning, search and simulation, and belief management.
Poki, a state-of-the-art poker playing program is discussed, with focus on the problem of opponent modeling. Several methods of opponent modeling are evaluated, including neural networks and decision trees. Poki can currently play poker at the level of an average human player.
We introduce a new method of directly searching imperfect information game trees, called miximax. Like minimax search, we do a search of the game-tree. Since the objective best moves of the opponent cannot be determined (we do not know their cards) we cannot choose a best move in the search. Instead, we use our knowledge of the opponent to choose their mixed strategy. Instead of using a min, we use a mix of their choices. The result is a reliable estimate of the expected values of each of our choices.
Acknowledgements
There are so many people to thank, so let’s get this party started!
First of all, Darse Billings – if a thesis could have a co-author, it would be you. Your support, ideas, insight and friendship has been invaluable. Thanks of course, to my supervisors Jonathan Schaeffer and Duane Szafron. Thanks for hiring me to work with the Poker Group, three years ago – it was an excellent opportunity, a challenging project, and a terrific journey. Thanks to Lourdes Peña and Denis Papp for their blood sweat and tears laying the ground work in *Loki*.
Thank you Christine Thesen, for supporting me lovingly, through all of this hard work. And thank you Vega for being such a cute kitty cat. Thanks Jason Spencer for being a great friend and a good guy to bounce ideas off of. Thanks to my parents Bill & Maion Davidson, for being great, supportive parents.
And to everyone else that has put up with me – friends and family and colleagues, thanks. Oh! And thanks to all the people who helped test *Poki* by playing it online for countless hours.
Did I mention thanks?
## Contents
1 Introduction 1
1.1 Poker as a Testbed for AI 2
1.2 Texas Hold’em Poker 3
1.2.1 An Example Hand of Texas Hold’em 4
1.3 Opponents 6
1.4 Research Goals / Thesis Contributions 7
2 Games in Theory and in Practice 9
2.1 RoShamBo 10
2.2 Prisoner’s Dilemma 11
2.3 Nash Equilibria 12
2.4 Optimal Poker Strategies 13
2.5 Other Poker Research 14
2.5.1 Evolutionary Methods 14
3 Computer Poker Basics 16
3.1 Poki’s High-Level Architecture 17
3.2 Pre-flop Betting Strategy 19
3.3 Hand Evaluation 20
3.3.1 Hand Strength 21
3.3.2 Hand Potential 22
3.3.3 Weighting the Enumerations 24
3.3.4 Pot Odds 27
3.3.5 Multi-player Considerations 28
3.3.6 Effective Hand Strength 29
3.4 Basic Betting Strategy 30
3.5 Selective Sampling and Simulation Betting Strategy 31
3.6 Summary 34
4 Opponent Modeling 35
4.1 Methods of Prediction 37
4.1.1 Expert Systems 37
4.1.2 Statistics 38
4.1.3 Neural Networks 38
4.1.4 Decision Trees 43
4.2 Confusion Matrices 44
4.3 Multi-Predictor Voting 45
4.4 Summary ........................................ 46
5 Experiments ..................................... 48
5.1 Evaluation of Reweighting Methods ............ 49
5.1.1 Reweighting System A ...................... 49
5.1.2 Reweighting System B ...................... 50
5.1.3 Experimental Design and Results .......... 52
5.2 Evaluating Predictors .......................... 53
5.2.1 Self-Play Results ........................... 55
5.3 Performance Results from Online Games ......... 55
5.3.1 IRC Poker .................................. 56
5.3.2 Results From the Poki Poker Server ...... 58
5.4 Summary ....................................... 61
6 Simulation Based Betting Strategies Revisited .. 63
6.1 Problems with Simulations ..................... 63
6.1.1 Negative Reinforcement Through Self-Prediction 64
6.1.2 Biased Weight Tables ....................... 64
6.1.3 Patches to Fix Simulations ................. 65
6.2 A New Approach – ‘Miximax’ Search ............. 65
6.2.1 Miximax and Miximix ......................... 68
6.2.2 Related Work ................................ 69
6.2.3 Estimating The Probability of Winning at Leaf Nodes 69
6.2.4 Pruning Large Trees ......................... 70
6.2.5 Current Implementation w/ Context-Trees .... 71
6.2.6 Results .................................... 72
6.2.7 Future Work ................................ 72
7 Conclusions and Future Work ..................... 74
7.1 Thesis Contributions ........................... 74
7.2 Future Work ................................... 75
A Miscellaneous ................................... 77
A.1 Hardware for Experiments ...................... 77
A.2 Internet Links ................................ 77
B Backpropagation Example .......................... 78
C Sample Confusion Matrices ........................ 80
D Glossary of Poker Terms .......................... 83
Bibliography ....................................... 87
List of Figures
1.1 A snapshot of the example hand on the turn. It is *Poki*'s turn to act. 6
3.1 High level architecture. 18
3.2 High level player architecture. 18
3.3 Hand strength calculation. 21
3.4 Hand potential calculation. 23
3.5 The algorithm for updating a Weight Table. The probability of the observed action (A) is taken from an opponent model’s (OM) prediction of the opponent (PT) based on the hand (E) and the current game context (GC). This probability (PT[A]) is used to weight the table entry (WT[E]). 25
3.6 An example weight table after a pre-flop fold (left), call (middle), or raise (right). 26
3.7 Two scenarios for calling pre-flop (left) and raising pre-flop (right), followed by betting after a flop of J♠-8♣-4♠ (See Chapter 1.2.1). 26
3.8 The difference between classic game-tree search and simulations. 32
3.9 Pseudo-code for poker simulations. 33
4.1 A simple neural network. 40
4.2 A neural network predicting an opponent’s future action. 42
4.3 Part of a basic decision tree, with probability triples at the leaf nodes. 44
5.1 A histogram of over 100,000 effective hand strengths (EHS) taken from *Poki* in ten-player online poker games (combined from all post-flop betting rounds). 50
5.2 Histograms of roughly 55,000 effective hand strengths (EHS) taken from *Poki* in two-player online poker games. Each betting round is shown separately. 51
5.3 Simulation using roll-outs (v.s. FBS-Poki) with different noise factors. All of the SBS programs lose to FBS-Poki, so shorter bars are better (a smaller loss). 52
5.4 Performance of three different simulation predictors against FBS-Poki. Shorter bars are better (a smaller loss). 56
5.5 Collected from IRC poker #holdem1 in early 2001. 57
5.6 Poki poker server results. 59
6.1 A game tree with opponent modeling and hand evaluation for a situation on the river. 66
6.2 A partial calculation of the EV for checking and betting. 67
6.3 Miximax v.s. three different opponents. .......................... 73
B.1 A simple neural network. ........................................... 79
| Table | Description | Page |
|-------|-----------------------------------------------------------------------------|------|
| 1.1 | Examples of four different types of games. | 2 |
| 1.2 | Poker hand type frequencies. There are $\binom{52}{5} = 2,598,960$ different five-card poker hands. The probability of being dealt a particular type of hand is shown above. | 4 |
| 1.3 | Betting history of example hand (actions proceed one column at a time). | 5 |
| 2.1 | The payoff matrix for a Prisoner’s Dilemma game. Each cell contains the penalty for Prisoner A, then Prisoner B. | 11 |
| 3.1 | Iterated income rate (profitable hands). | 20 |
| 3.2 | Breakdown of cases in a hand potential for the hand $7\heartsuit-9\heartsuit$ with a flop of $8\heartsuit-6\spadesuit-4\heartsuit$. | 24 |
| 4.1 | Neural network inputs. | 42 |
| 4.2 | A typical confusion matrix showing the errors made in predicting an opponent.| 45 |
| 5.1 | Average predictor accuracy in online games. | 53 |
| 5.2 | Average predictor accuracy. | 55 |
| 5.3 | Poki Poker heads-up contest results (January 2002). The results are for all 500 hands played per room. | 61 |
| 5.4 | Poki poker heads-up contest results summary. | 61 |
| C.1 | Statistics-Loki2 (Human Opponent) | 81 |
| C.2 | Statistics-II (Human Opponent) | 81 |
| C.3 | Expert Formula (Human Opponent) | 81 |
| C.4 | Neural Network (Human Opponent) | 81 |
| C.5 | Meta-Predictor (Human Opponent) | 81 |
| C.6 | Statistics-Loki2 (FBS-Poki) | 82 |
| C.7 | Statistics-II (FBS-Poki) | 82 |
| C.8 | Expert Formula (FBS-Poki) | 82 |
| C.9 | Neural Network (FBS-Poki) | 82 |
| C.10 | Meta-Predictor (FBS-Poki) | 82 |
Chapter 1
Introduction
My work is a game, a very serious game.
– M. C. Escher (1898-1972)
Games are a natural choice for artificial intelligence (AI) research. They present unique challenges that require complicated problem solving to overcome. Games typically have well defined rules and definitions. Consequently, game states and transitions are usually easy to represent in a symbolic computer program. By definition, games have a built-in performance measure. This facilitates the quantification of a player’s strength. Because of these features, games have been a main focus of artificial intelligence research since its inception [50]. The example that would naturally come to mind for most people is that of chess – a game that has had a long history in AI research [28]. Another classic example is Samuel’s checkers program that used machine learning to achieve an average level of play [39].
Artificial intelligence research has had great success in many classic games such as chess [10], checkers [40], and Othello [9]. AI for these games has matured to the point that they can play at the level of strong grandmasters, and beyond. Using game tree search algorithms, programs have challenged and defeated human world champions in these games. These are deterministic domains of perfect information and their difficulty lies in the huge size of the search space. Computers excel at searching deeply into these game trees, which have a moderate branching factor. Recursive minimax search is ideally suited to these well-defined, noise-free domains. In Table 1.1, these types of games fall into a single category – deterministic games of perfect information. Until recently, games in the other three categories have remained largely unresearched.
The late 1990’s have seen research progress in several games of imperfect information. Tesauro’s backgammon program, $TD$-$Gammon$, which uses temporal difference learning, has had success at overcoming the non-deterministic aspect of
| | deterministic | non-deterministic |
|----------------|---------------------|-------------------|
| perfect information | chess, checkers, go | backgammon |
| imperfect information | battleship, roshambo | poker, bridge, scrabble |
Table 1.1: Examples of four different types of games.
the game [48, 49]. Sheppard’s Maven program for playing Scrabble is stronger than the best humans [43]. Ginsberg has produced a bridge program (GIB) that is world-champion caliber at declarer play (but not at bidding or defense) [20]. All of these programs have had to use new, innovative techniques to achieve high-performance in these imperfect-information or non-deterministic domains.
Presumably, the ultimate goal for much of artificial intelligence research is to develop useful systems that can adaptively make intelligent decisions in large uncertain domains, much like the world we inhabit. Artificial intelligence research has often been criticized for being too narrowly focused on systems that are only of use in a ‘perfect’ world. Techniques that work in small discrete domains often flounder in real-world situations. Reality is inherently noisy and unpredictable. There are no oracles. The world does not give away information for free.
1.1 Poker as a Testbed for AI
Continuing in the tradition of using games as a testbed for artificial intelligence, poker is an excellent game to explore decision making under conditions of imperfect information [3]. The poker domain contains a wide variety of challenges to standard learning and decision-making algorithms. The domain is inherently noisy and uncertain. Poker has high variance in the outcomes due to the non-determinism (the random shuffling of the deck), and imperfect information (the unknown cards held by the opponent). Furthermore, if a player folds, we never know what cards they were holding. These factors impose a large hurdle for reinforcement learning algorithms, since it is often difficult to determine if an action was positive or negative.
To complicate matters, poker games are normally played at a fast pace (typically lasting only a few minutes per hand), so players must make each decision in at most a few seconds.
All reasoning in poker must be probabilistic, as things are rarely ever certain. The cumulative sum of a series of games matter more than any individual game. In addition to the probabilistic aspects of poker, there are also psychological elements to address in observing and determining how our opponents play the game. Opponents are often deceptive, deliberately playing in ways which are difficult to learn and model.
In order to play strong poker, one must be able to understand the mathematics and strategy of the game, as well as understand the opponent’s understanding of the
game (and further, the opponent’s perception of us, and so on...). A large part of the allure of using poker as a testbed for artificial intelligence research is the human aspect of the game. A crucial part of a poker game is modeling the opponent – trying to learn how they think, how they believe we think, and determining what tactics we can use to take advantage of this knowledge.
Artificial intelligence has long been criticized for working in small toy domains of perfect knowledge [14]. Poker has properties that are ideal for exploring many areas of AI that have traditionally been avoided or ignored.
### 1.2 Texas Hold’em Poker
Texas Hold’em is a common variant of poker played in casinos and championship poker tournaments. It has a high skill-to-luck ratio, compared to other types of poker, and it is known to be one of the most strategically rich variants of poker. The game is played with a standard deck of fifty two unique playing cards with four suits (clubs (♣), diamonds (♦), hearts (♥), spades (♠)), and thirteen ranks (in increasing value: deuce (2), trey (3), four (4), five (5), six (6), seven (7), eight (8), nine (9), ten (T), jack (J), queen (Q), king (K), ace (A)).
Anywhere from two to twelve players can play (ten is most common in casinos), making for a dynamic game with a wide range of situations. Each player is dealt a hand consisting of two private *hole cards*. Once each player is dealt their cards, a round of betting ensues. The first two players after the dealer each pay a forced bet called the *blind* (instead of an *ante*) to start the betting. The first blind is half the minimum bet size, and is referred to as the *small blind*. The next player, the *big blind*, pays a full forced bet. During a round of betting players may either fold their hand, call the current bet, or raise the bet.
After the first betting round, called the *pre-flop* stage, three community *board* cards are dealt. This stage is called the *flop*. After another round of betting, a fourth board card is dealt called the *turn*, and is followed by another round of betting. Finally a fifth board card is dealt, the *river*, and we enter the last betting round.
Any players remaining in the hand when the betting is finished must go to a *showdown*. In the showdown, players must show their hole cards to the table. The player with the best five card poker hand (see Table 1.2) made from the best combination of the two hole cards and the five board cards wins the money in the *pot* (if there is a tie the money is divided amongst the winners). A player may win the pot uncontested if all of the opponents fold before the showdown, in which case the winner is not required to reveal his or her cards.
In *limit* Hold’em, the bet increment sizes are pre-determined. The limit is defined by a bet increment size called a *small bet*. During the pre-flop and flop betting
| Hand Type | Example | Frequency |
|-------------------|---------------|------------|
| Straight Flush | Q♦ J♦ T♦ 9♦ 8♦ | 0.0015% |
| Four of a Kind | 5♥ 5♣ 5♦ 5♠ K♣ | 0.0240% |
| Full House | 5♥ 5♣ 5♦ K♠ K♣ | 0.1441% |
| Flush | 2♣ 3♣ A♣ J♣ 7♣ | 0.1965% |
| Straight | 5♦ 6♠ 7♦ 8♥ 9♦ | 0.3925% |
| Three of a Kind | 9♦ 9♣ 9♥ 2♦ 6♦ | 2.1129% |
| Two Pair | A♥ A♣ J♥ J♠ 4♦ | 4.7539% |
| One Pair | J♣ J♠ 3♥ 6♠ K♦ | 42.257% |
| High Card | Q♣ 6♥ 3♥ T♦ 4♦ | 49.882% |
Table 1.2: Poker hand type frequencies. There are \( \binom{52}{5} = 2,598,960 \) different five-card poker hands. The probability of being dealt a particular type of hand is shown above.
rounds, all bets are equal to one small bet. For the last two betting rounds this bet size doubles to a big bet (two small bets). For example, a game with a small bet of $10 would be called a $10/$20 Hold’em game.
A more complex version of Texas Hold’em is no-limit, where players may bet as large an amount as they wish. Due to the high variability of betting and increased strategy, we do not currently consider no-limit Texas Hold’em. Limit Hold’em should be well understood before working on this more difficult version of the game.
### 1.2.1 An Example Hand of Texas Hold’em
In order to give a clear example of how the game is played, and to illustrate some of the strategy, the following example hand of Texas Hold’em is given as a guide. While no single hand of poker can exemplify all of the strategy and situations that occur in the game, this example gives a flavour for how it is played, and is intended for readers who are unfamiliar with the game.\(^1\)
Suppose we (\textit{Poki}) are playing in a five player game of Texas Hold’em, with a $10/$20 limit. We are the button, which means we are the last to act in each round of betting. Being last is advantageous, since we get to see the actions of all the other players before making our betting decision. Conversely, the player first to act, a position known as under the gun, has a disadvantage. \textit{Poki} will play against four opponents, who we will refer to as Alice, Bill, Chris, and Dave. Alice is the small blind.
Alice is the player after the button, so she posts the $5 small blind and Bill posts the $10 big blind. Each player is dealt two private hole cards. \textit{Poki} is dealt T♠-9♠. The first player to act. Chris, calls the big blind for $10, and the next player, Dave, folds his hand. It is now our turn. Since we have a suited-connector hand (two cards
\(^1\)Throughout this thesis, poker terminology printed in italics is described in the glossary.
of the same suit and neighbors in rank), we feel we have a positive expectation and call. Alice raises, putting another $15 into the pot (she has now put a total of $20 into the pot), and Bill, the big-blind folds (forfeiting his $10 big-blind). Chris calls Alice’s raise, and then it is back to us. We have already committed $10 to the pot, so we call Alice’s raise as well. This ends the first betting round. There is $70 in the pot and there are three players left.
The flop comes down J♠-8♣-4♦. Although this does not pair either of our cards, we have a very high potential to make a straight or a flush on the turn or river. Alice is first to act and bets $10. Chris is next and calls the bet. We figure we have more than a 30% chance that the turn card will give us a flush or straight (a very strong hand) and the pot is $90, so we are getting 9-to-1 odds for a call.\footnote{Suppose we played this situation the same way 100 times. In 30 of these instances, we would make a strong hand and win $90, for a total of $90 \times 30 = \$2700$. The 70 times we called $10 and did not make our draw, we lose $10 each time, for a total loss of $700$. By calling, we would net a total of $2000 in 100 hands, or $20 per hand.} We decide that with our huge potential, it is well worth the $10 and we call. With such a high potential it would even be prudent to raise. The pot is now $100.
The turn card is a J♦ which still does not help us. Alice bets $20, and Chris folds. It is known from past experience that Chris is a fairly tight player who doesn’t like to take many risks. He might have had a hand like a pair of nines, and he now feels that he is beat because of the presence of the two Jacks on the board. Since Alice has bet, she may have a Jack, giving her three-of-a-kind, or perhaps a big pocket-pair such as two Queens. We suspect this might also be the case, but we have observed Alice to be an aggressive player and she frequently bluffs. The pot is now $120, so our odds are now 6-to-1, which are more than ample to justify calling the $20. Figure 1.1 shows a snapshot of the game at this point.
The river card is the T♥ which gives us two pair. Again, Alice bets $20. We don’t think we necessarily have the best hand, but the pot is large enough that we should call in case Alice is bluffing (to consider folding with a pot this large, we must be more than 80% certain of having the worst hand). We call and Alice shows A♥-J♥, giving her three-of-a-kind. We toss our hand in the muck and let Alice take the pot of $180. Table 1.3 summarizes the history of this game.
\begin{table}[h]
\centering
\begin{tabular}{|c|c|c|c|c|c|}
\hline
Player & Preflop & Flop & Turn & River & Net Gain \\
\hline
Alice & Blind $5 & Raise $10 & Bet $10 & Bet $20 & +$110 \\
Bill & Blind $10 & Fold & & & -$10 \\
Chris & Call $10 & Call $10 & Call $10 & & -$30 \\
Dave & Fold & & & & -$0 \\
Poki & Call $10 & Call $10 & Call $10 & Call $20 & -$70 \\
\hline
\end{tabular}
\caption{Betting history of example hand (actions proceed one column at a time).}
\end{table}
1.3 Opponents
Players come in all types imaginable. Part of modeling an opponent is classifying where they lie on the various spectrums of playing styles. One must try to discover from observation the general strategy an opponent employs in their play. For example this might involve determining that a player is *tight* (playing only strong hands), or *loose* (playing many marginal hands). We can measure how *passive* or *aggressive* the player is, by observing how often they bet and raise hands they consider to be strong. We can also measure how observant a player is, and how quickly they adapt to other player’s strategies.
For the most part, many of these expert-defined values are easy to calculate. From a scientific standpoint, it would be far more interesting to have a system learn these relevant features of opponents from scratch, rather than having expert knowledge of these features directly encoded. Computing a measure of tightness or aggressiveness is easy to do for an overall measure of a player’s style, but a good player mixes or frequently switches styles depending on the particular context of a hand. In this case, overall measures do not provide enough useful information.
Many tactics are deployed by opponents in order to make themselves difficult to model. For instance, if a player has only a mediocre hand, they may bluff, representing a strong hand, hoping their opponent will fold. Or a player may under-represent a very strong hand early in the game to sucker more players into staying in the hand, which later leads to a higher pay-off from the pot.
It is not how many games are won that is important, but how much money is won over all hands played. In order to cause opponents to build flawed models, a player may purposefully play a few relatively unimportant hands terribly. Later,
this *table image* can be exploited.
Opponent modeling is the task of learning how an opponent plays poker. A complete opponent model can tell us several things. It can tell us the general styles that the opponent plays. It can infer from an opponent’s actions a probability distribution over the cards that they may hold. Given a hand and a game situation, the model should also be able to mimic playing poker the same way the opponent would. This, in essence, is the goal of this thesis research.
### 1.4 Research Goals / Thesis Contributions
The generalized aim of this thesis is to develop ways of modeling uncertain environments and how to make intelligent decisions within them.
This thesis builds heavily upon the past thesis research by Darse Billings, Denis Papp, and Lourdes Peña. They have all participated in building the foundation for this computer poker research – the poker-playing agent *Loki*. Darse Billings, who is both a researcher and a poker expert, laid the foundations of our computer poker AI [3]. Denis Papp brought many of these ideas to life in the first incarnation of *Loki* [34]. Lourdes Peña extended *Loki* with a new adaptive, simulation-based betting strategy [35].
The organization of this thesis is as follows:
- Chapter 2 gives an overview of the game theoretic properties of poker, and demonstrates some of these elements using simpler games. Other poker research is also discussed.
- Chapter 3 explains basic algorithms used for computer poker. This includes hand evaluation, betting strategies, and the overall architecture of our poker playing program. The ideas presented have been embodied in a poker playing agent called *Poki*, a Java-based descendant of the C++ poker program *Loki*. *Poki* was a complete reimplementation of *Loki*, in a much cleaner object-oriented architecture. Numerous improvements were made, most notably in the opponent modeling systems.
- Chapter 4 explores the opponent modeling problem in poker. The opponent modeling in *Loki* was fairly crude. Many machine learning techniques are discussed – leading to major improvements in *Poki*’s opponent modeling abilities. This chapter represents a large portion of this thesis’ contributions.
- Chapter 5 discusses the experiments and results obtained from self-play experiments, offline opponent modeling analysis, and online games against human players.
• In Chapter 6 the results of the simulation based betting strategy are discussed. A new search method for imperfect-information game trees is introduced and evaluated.
• Chapter 7 presents conclusions, provides a summary of the thesis contributions, and presents suggested avenues for future research.
Research on imperfect information games can be divided into two camps. The older of the two is *game theory*, pioneered by theoreticians such as John Von Neumann and John Nash [52]. Game theory is concerned with the mathematical and logical structure of non-cooperative games, and how to make rational choices within them. Game theory is often applied in economics, politics, and classic games.
The other camp uses a more empirical approach with the goal of studying and creating high-performance systems for playing games. This thesis research belongs in this latter category. However, in building high-performance game playing systems, one must be aware of the theory behind the game. Game theory can tell us what to expect from various situations in the game, as well as which strategies may be strong or weak. This section will deal with the theoretical issues surrounding poker. To begin, we will look at some simpler games that share some of the game-theoretic properties of poker.
A *non-cooperative game* is one where the interests of each player are strictly opposed, and a *zero sum game* is scored such that all players’ scores sum to zero. If one player gains, it is by the equivalent loss of the other players. Games of *imperfect information* require a player to make decisions without knowing the complete state of the game.
The following discussion will present two simple games. As the main topic of this thesis is opponent modeling, there will be particular focus on the importance of identifying your opponent’s strategy.
2.1 RoShamBo
RoShamBo, also known as Rock, Paper, Scissors, is a game with very simple rules (it is often played by children). Two opponents choose one of three actions (paper, rock, or scissors) simultaneously. Paper beats rock, rock beats scissors, and scissors beats paper. For instance, if player A chooses rock, and player B chooses scissors, then A wins one point, while B loses one point. If both players choose the same thing, they draw and score no points. RoShamBo is zero-sum, and contains hidden information (the simultaneous choice of actions can be viewed as each player choosing their action secretly, before the both players choices are revealed). Its simplicity, however, is deceptive when it comes to playing well against an arbitrary opponent.
Game theory tells us that the optimal solution is trivial. A player who chooses an action uniformly at random (an equal probability of choosing each of the three actions) cannot be exploited in any way. No matter what strategy an opponent employs against this optimal strategy, the expected value of a match is a draw. Of course, while this strategy cannot lose in the long run against the opponent, neither can it win.
In 1999 and 2000, Darse Billings hosted an international RoShamBo programming competition that was modeled after the iterated prisoner’s dilemma tournaments (discussed in the following section) [5, 4]. In a tournament setting, each pair of programs played a match of 1000 games against one another. The cumulative score of each opponent dictated its final place in the rankings. Hence, if the field contained at least two non-optimal players that did not draw against one another, an optimal player could not place first.
The tournaments drew many high quality submissions. It was observed that the success of many strategies was significantly dependent on the composition of the other strategies. The authors of the top entries, including some well-known AI researchers, have commented that writing a strong RoShamBo program was much more challenging than they initially expected. The best programs do a sophisticated analysis of the full history of the current match in order to predict the opponent’s next action, while avoiding being predictable itself. Programs that used a simple rule-based model for making their decisions consistently finished near the bottom of the standings.
A program named Iocaine Powder stood out from the crowd [15]. Although deterministic, it outplayed the other exploitable strategies by out-guessing them. Iocaine Powder does not play an optimal strategy. Since it is deterministic, a program could be written to beat it.\footnote{For instance, one could copy the source code to Iocaine Powder, and play the winning move against what Iocaine Powder would have done. This would play perfectly against Iocaine Powder, but would lose to other strategies.} With the knowledge that the other players are
playing non-optimally, Iocaine Powder attempted to play a strategy that maximized winnings by exploiting these program’s predictability.
The example of RoShamBo demonstrates the importance of opponent modeling in imperfect information games. The variety of strategies and the composition of the field determine the success of each non-optimal strategy. The contrast is clear between the optimal strategy and a context-dependent *maximizing* one. The optimal strategy chooses randomly, with no regard for the opponent’s strategy, while maximal strategies explicitly observe and outguess their opponents.
### 2.2 Prisoner’s Dilemma
The classic example of an imperfect information game studied by theorists is *Prisoner’s Dilemma*.\(^2\) In Prisoner’s Dilemma, the two players have just been arrested for a crime and imprisoned in separate cells. The players may not communicate with one another (this gives us the element of missing information). If both players deny their guilt, there will not be sufficient evidence to keep them imprisoned for more than two years. If one player squeals on the other, while the other denies his guilt, the squealer will go free, and the other will be imprisoned for ten years. However, if both squeal on each other, they will both be imprisoned for five years.\(^3\)
The dilemma is obvious. If both prisoners cooperate and stay quiet, they will only serve two years. However, it is tempting to defect and squeal on the other prisoner in an effort to walk free. But if both prisoners decide to defect, the penalty is severe. A typical representation of this kind of game is a *payoff matrix*, shown in Table 2.1.
If the game is played just once, the best strategy is to defect, since it dominates the other options (defecting gets a worst case of five years in prison, while cooperating can lead to ten years in prison). The game can be made even more interesting by playing a series of games, and keeping score. This is called *Iterated Prisoner’s Dilemma*. In the iterated case, cooperation becomes a viable option. After a few trials, one can begin to analyze the choices made by the opponent,
---
\(^2\)It is important to note that this game is not zero-sum, nor is it non-cooperative. Hence, it does not fall into the category of proper game theory. However, it is a widely researched and discussed game of imperfect information, where modeling the opponent is a critical aspect.
\(^3\)These numbers can be arbitrary, provided the relative orderings are preserved.
and develop a specific strategy to thwart it. In 1979, Axelrod conducted an *iterated* Prisoner’s Dilemma programming contest. Participants were invited to submit programs, which played one another in a tournament [1].
A wide variety of different strategies competed in the tournament. One of the fundamental lessons learned from these competitions (which is why this is being discussed here) was that the best strategy is dependent on the composition of the field (the other competitors).
For instance, one of the most successful strategies was called *tit-for-tat*. The strategy was extraordinarily simple. It starts by cooperating on the first trial. Afterwards, it does whatever the opponent did in the previous trial. The reason this works so well is not so easy to see. In short, it rewards cooperation with cooperation, and punishes defection with defection, but is quick to forgive. If the opponent cooperates, so will *tit-for-tat*. If the opponent ever defects, *tit-for-tat* will be burned once, but then will defect in the next round. If the opponent defects again, both will take a penalty, and as long as the opponent continues defecting, so will *tit-for-tat*. Once the opponent realizes defecting is not working (it is taking a penalty each time), it may decide to cooperate again. But the instant it switches from defecting to cooperating, *tit-for-tat* will defect since that was the opponent’s last choice. In the next trial, *tit-for-tat* will go back to cooperating, and the entire pattern may play out again. In the long run, the entire sequence ends up as (or close to) a draw, because *tit-for-tat* plays so defensively. While *Tit-for-tat* can draw, it also can’t win an individual match – a draw is the best it can do. However, it can lose every single match and still win a tournament with a high overall score.
### 2.3 Nash Equilibria
Nash showed that all zero-sum imperfect information games have an equilibrium (or saddle point) in which every player can ensure the optimal outcome with an appropriate randomized mixed strategy [31]. A game-theoretic *optimal strategy* is a strategy that maximizes gain and minimizes loss against *any* possible opponent strategy. An optimal strategy is the safest strategy one can use when the opponent’s strategy is unknown. In other words, if your opponent is using an optimal strategy, the best you can do against them is to play optimally as well.
However, an *optimal* strategy is not necessarily a *maximal* strategy. If your opponent is not playing optimally, there may exist a non-optimal strategy which will maximize winnings against their non-optimal strategy. In order to exploit a non-optimal opponent, one must know what kinds of mistakes they are making. The goal of opponent modeling is to learn how an opponent plays so that weaknesses may be exploited, and strengths may be avoided.
Iterated Prisoner’s Dilemma is not a zero-sum game, so Nash Equilibria do not
necessarily exist. The *tit-for-tat* strategy is not optimal, but it is robust (it will never lose by more than a few points, no matter what the opponent’s strategy). While it is robust, it never initiates defection, which means it cannot maximize winnings against weak opponents. For instance, if it were playing an opponent that always cooperated, *tit-for-tat* would never attempt to exploit it by defecting.
### 2.4 Optimal Poker Strategies
Some of the previous academic research involving poker has been focused on game theory and optimal solutions. In fact, poker was used for many examples in early game theory explorations by John von Neumann and John Nash [52, 32, 31].
While full-scale poker is still much too large to solve, smaller variants of poker have been analyzed and solved. For example, a smaller deck of cards may be used, along with fewer, less complicated betting rounds. Typically, only two-player variants are examined, and not full ring games (usually about ten players). One of the smallest variants is Kuhn’s two-player, three-card, poker game [27]. Each player is dealt one card, and there is a single round of betting. This game is simple enough that it can be dissected and solved by hand.
More recently, simplified poker games have been solved computationally. Koller and Pfeffer have worked on *Gala*, a generalized system for finding optimal randomized strategies for two-player imperfect information games [25]. The standard method of computing optimal strategies for small games is to convert the game tree into *normal form*. However, this representation is exponential in the size of the game tree, rendering the technique useless for all but the simplest of games.
Instead, Koller and Pfeffer’s approach uses a new method that is linear in the size of the game tree [24]. Using this technique, they have found optimal strategies for several scaled-down poker-like games. However, the size of the game tree for full-scale two-player Texas Hold’em is still many orders of magnitude higher than the largest problem the Gala system can solve.
Ken Takusagawa, a former student of Daphne Koller, has extended this work by combining the method with abstraction, to produce some near-optimal solutions for particular scenarios of Texas Hold’em [47].
A subset of Texas Hold’em, two-player *Pre-flop Hold’em* has been solved optimally by Alex Selby [41]. In Pre-flop Hold’em, once the pre-flop round is finished, all five board cards are dealt and the players proceed directly to the showdown.
Jiefu Shi and Michael Littman have worked on a smaller version of Texas Hold’em, humorously dubbed *Rhode Island Hold’em* [44]. While much smaller than Texas Hold’em, the game is still too large to solve optimally. Littman collects similar groups of cards into equivalence classes in order to reduce the size of the problem to a manageable size. The resulting solution, while not perfect, is *near-optimal*. Since
poker has such high variance, a near-optimal strategy is likely to perform extremely close to an optimal strategy against other non-optimal opponents.
2.5 Other Poker Research
Besides Texas Hold’em, there has been work on other Poker variants such as Five-Card Draw. Five-Card Draw poker is not as strategically complex as Texas Hold’em or 7-card Stud.
Nicolas Findler did early work on this game, intermittently over a twenty year period [18]. His main objective was not to achieve a high level of play, but to model human cognitive processes. His program was able to learn, and was moderately successful.
In the 1970’s and 1980’s some research was done in exploring production systems for Five-Card Draw strategies. Waterman developed a program which was able to learn the production rules through three different methods – advice (expert rules), analysis (post-game), and adaptation (opponent modeling) [53]. Smith later developed a different production system that learned rule sets by using a genetic algorithm [46]. Smith’s program was able to win against Waterman’s program, which was considered to play at the level of an experienced human player.
Five-Card Draw poker has been recently studied using Bayesian models with mixed results [26, 12].
2.5.1 Evolutionary Methods
Recently, evolutionary algorithms have been applied to game theory and to poker. Genetic algorithms have been suggested as a way to discover local, and possibly global, Nash-equilibria in game-theoretic domains [16]. A population of co-evolving strategies will eventually converge to a local optimum with respect to the population of competing strategies.
Because of poker’s game-theoretic underpinnings, a similar approach has been applied to evolve strategies for ten-player Texas Hold’em [33]. Simplified strategies are described by a small array of values. A regular co-evolutionary genetic algorithm (GA) starts with a population of randomly generated strategies. The strategies are then evaluated by having them compete against one another. The highest scoring strategies are reproduced with some random variation (usually by small perturbations of the values in the strategy). Once a new population has been constructed, the process repeats. The selection pressure steers the random search towards higher scoring strategies, until the population converges on a local optimum.
The method of selecting the surviving strategies from a population determines how the population will evolve, and towards which local equilibrium it will converge. Using ideas from game theory, a more robust method of selection in co-evolving
populations can be used. Pareto co-evolution [17] is a new method of selection based on the concept of Pareto optimality. A strategy Pareto-dominates another if it is at least as good as the other strategy over all dimensions, provided it is strictly better in at least one dimension. A Pareto-optimal strategy would be a strategy that Pareto-dominates all other existing strategies. This idea is applied to co-evolution as a way to select for Pareto-dominating strategies. The dimensions of evaluation become the strategy’s performance against the other strategies in the population itself. The other strategies are, in a sense, the dimensions for evaluation.
This changes the selection pressure substantially. It selects the more robust strategies, that do well against a larger set of the opponents. The early research indicates that the Pareto co-evolution outperforms the standard selection method of choosing the $n$ strategies with the highest gross scores.
Currently, however, no evolutionary computation methods have been applied to generate high-performance betting strategies for poker. It is likely that the complexity of Texas Hold’em is too great for a naive evolutionary algorithm to tackle effectively.
Chapter 3
Computer Poker Basics
Do not expect to arrive at certainty in every subject which you pursue. There are a hundred things wherein we mortals... must be content with probability, where our best light and reasoning will reach no farther.
– Isaac Watts (1674-1748)
The fundamentals of poker strategy are determined by the probabilistic nature of the game. The way a hand is played depends on the probability of both playing the hand profitably, and how it affects the profitability of future hands (by changing your table image). Winning an individual hand is not as important as winning in the long run. As mentioned earlier, poker strategy requires both mathematical understanding of the game and psychological understanding of the opponent. To play well, the concrete mathematical foundations must be in place. The psychological or intuitive aspect of poker can be used to enhance the accuracy of this foundation.
There are several skills required to play strong poker. These will all be required for a world class poker agent:
**Hand Evaluation:**
Accurate assessment of the current strength of our hand, and the potential strength of our hand when the future board cards are dealt, is crucial. We need the probability that our hand is the best, given the context of the game and the opponents we are up against.
**Unpredictability:**
Our actions must not give away the strength of our hand. We must play deceptively by *bluffing*, *slow-playing*, and *check-raising*. It is important that our betting strategy be non-deterministic, and mixes strategies in order to conceal information about our hand.
Opponent Modeling:
We must understand how our opponent plays in order to both exploit their weaknesses and defend against their strengths. We must be able to deduce the strength of hands our opponents hold by interpreting their actions, and we must be able to anticipate how they will play.
This chapter discusses the pre-flop play, post-flop hand evaluation, and betting strategies used by Poki. Many of the methods in this chapter are described more thoroughly in [34].
3.1 Poki’s High-Level Architecture
A poker game consists of a dealer together with multiple players that represent either human players or computer players. In our Java implementation, these players are defined as objects. The dealer handles the addition and removal of players from the game, deals the cards to each player at the start of a new hand, prompts each player for an appropriate action when it is their turn, broadcasts player actions to other players, and updates a public game context as the game progresses. The game context contains all of the public information about the game, including the names and relative locations of the players, and the board cards.
Each player extends an abstract Player class, that has access to all the public game information, as well as the player’s private information (their cards). This class is extended to add the player’s personal strategy. A player can be a bot (software program), or an interface for a human to play in the game, just by extending this class appropriately.
Likewise, the Dealer is a generalization for game controllers. We have implemented several different dealer interfaces: an IRC-Dealer for playing against other players on the IRC poker server, a Tournament-Dealer for self-play experiments, and a TCP/IP-Dealer that allows Poki to play against humans using a web browser, and against other programs using a published protocol (see Appendix B).
Each player is responsible for responding to a prompt by choosing an action: fold, call, or raise. In addition to the public game context, each player knows its private hand and the previous actions of all of the players. Each player is defined by an interface, and each version of Poki is a different implementation of that player interface.
Figure 3.1 shows the high level architecture of our Java-based poker system. The arrows depict the flow of information. The public game state and history is comprised of a GameInfo object, also containing several PlayerInfo objects which track public information for each player. The Player objects are each given cards by the dealer, and are then prompted for each action in the game. The players may
Figure 3.1: High level architecture.
Figure 3.2: High level player architecture.
access the GameInfo and PlayerInfo objects to make their decisions. The dealer then modifies the GameInfo and PlayerInfo objects according to the rules of the game.
Although each version of Poki represents and uses the available information in a different way, all versions share a common high-level architecture. In addition to the public game context, Poki has two private components: its hand and a collection of persistent opponent models. Poki maintains statistical models describing each player participating in the game, including itself. When the dealer broadcasts a player action, the corresponding opponent model is responsible for processing that information.
### 3.2 Pre-flop Betting Strategy
The strategy of Texas Hold’em in the pre-flop stage (before any board cards are dealt) is different than in the post-flop stages. Because there are no public cards available, the state in the pre-flop is far simpler than later in the game. In total, there are \( \binom{52}{2} = 1326 \) different hands, but many of these hands are equivalent before the flop. Without board cards, the particular suit of the cards is irrelevant. Both Q\(\heartsuit\)–3\(\heartsuit\) and Q\(\spadesuit\)–3\(\spadesuit\) have the same probability of making a flush, and in all other respects they are equal. If the flop contains any diamonds or clubs, then the two hands can no longer be treated identically. Using this knowledge, there are only 169 distinct hand types in pre-flop Hold’em.
Poki’s pre-flop strategy is a simple expert system built on top of a set of machine-learned tables containing an expected income rate for each hand. These tables were calculated off-line using a simple technique called roll-out simulations. Several million hands were played where each player calls the blind and then checks to the showdown. The cumulative wins and losses for each hand type are kept in the table. Although based on unrealistic play, these values give a good baseline for the relative profitability of each of the 169 hand types.
A refinement to roll-out simulation is to use repeated iterations of the technique, where the previous results govern the betting decision for each player. For instance, in a ten player simulation, a net negative value in the previous iteration would dictate that the hand be folded, rather than calling the big blind. This introduces some hand selection into the roll-outs, which reduces the number of active players in each hand. This generates a more realistic distribution of opponents and probable hands. Many mediocre hands that were profitable to play against players with no hand selection become negative in this more realistic simulation.
After each round of simulations has reached a reasonable degree of stability, another iteration is performed. This process eventually reaches an equilibrium, defining a set of hands that can be played profitably against the blinds and the other unknown hands. The results are most applicable to the “play or don’t play”
| Hand | IR-10 | Iterated | Hand | IR-10 | Iterated | Hand | IR-10 | Iterated |
|-------|-------|----------|-------|-------|----------|-------|-------|----------|
| AA* | +2112 | +2920 | ATs | +736 | +640 | KQo | +515 | +310 |
| KK* | +1615 | +2180 | 99* | +553 | +630 | QTs | +655 | +280 |
| QQ* | +1224 | +1700 | KQs | +858 | +620 | QJs | +720 | +270 |
| JJ* | +935 | +1270 | AQo | +555 | +560 | A9s | +538 | +220 |
| TT* | +714 | +920 | KJs | +767 | +480 | ATo | +335 | +200 |
| AKs | +1071 | +860 | 88* | +450 | +450 | KTs | +694 | +190 |
| AKo | +718 | +850 | 77* | +364 | +390 | KJo | +391 | +160 |
| AQs | +915 | +780 | AJo | +430 | +380 | A8s | +469 | +110 |
| AJs | +813 | +680 | JTs | +657 | +360 | 66* | +304 | +40 |
Table 3.1: Iterated income rate (profitable hands).
decision for each player. Although much better than a simple roll-out simulation, this technique is still far from perfect, because other important considerations such as betting position and known opponent actions have not been accounted for.
In our experiments, each iteration lasted for 50,000 trials. A diminishing noise factor was added to each income rate, analogous to the cooling factor used in simulated annealing. This gives negative expectation hands a chance to recover as the prevailing context changes. After ten generations, the remaining positive expectation hands were played for another 500,000 trials, to ensure stability. The resulting set of profitable hands, shown in Table 3.1, is in strong agreement with expert opinion on this matter [45]. The table shows a comparison of the income rates for 10-player roll-out simulations (IR-10) and the results refined by iterating (Iterated). The values shown are in milli-bets (eg. a hand with an income rate of +1000 should win an average of one small bet each time it is played). The iterated values are reasonable estimates of actual income rates, unlike the simple roll-out values, which are only used as relative measures.
We can generate near-optimal solutions for the two-player game of pre-flop Hold’em with iterated roll-out simulations. The results closely match Selby’s results [41] for the optimal two-player pre-flop Hold’em game.
### 3.3 Hand Evaluation
The post-flop betting strategy is far more difficult than the pre-flop strategy. There are several different strategies for the post-flop that will be described later. Before we do that, we must first describe how hands are evaluated. Later, the betting strategies will show how all of the evaluation information can be tied together into a rational betting decision.
3.3.1 Hand Strength
A crucial piece of information needed to make good decisions is the strength of our hand. A player must have a good estimate of the probability that they currently have the winning hand. A simple calculation of hand strength (HS) is to enumerate all possible other hands and compare them to ours. For instance, after the flop there are only 47 unknown cards remaining (52 cards minus our hole cards and the three board cards). Thus there are $\binom{47}{2} = 1081$ possible two card combinations we could be up against.
By counting the number of these other hands that win, tie, and lose to our hand, we can compute the probability of having the best hand, against one random hand. Figure 3.3 gives the algorithm for a simple hand strength calculation [34]. Essentially, this is just a calculation of the proportion of possible scenarios in which we have the best hand.
For example, suppose our hand is A♦-Q♣ and the flop is J♥-4♣-3♥. There are 1081 possible hands an opponent could be holding. Of these possible hands there are 444 that are currently better than our hand, 9 that would tie, and 628 that are worse. Counting ties as one half, this corresponds to a percentile ranking, or hand strength, of 0.585. In other words, there is a 58.5% chance that A♦-Q♣ is better than a random hand.\footnote{Hand Strength calculations for more than one opponent will be discussed below in Section 3.3.5.}
This value, called the \textit{hand rank} (HR) or \textit{raw hand strength} (RHS), is a good first-order approximation of the strength of our hand. However, for a real game of poker, this value is too crude. If our opponent has called one or more bets, then we are most certainly not up against a random hand anymore. Rational opponents will not play most weak hands they are dealt. It is typical for Hold’em players to fold over 50\% of their hands before the flop. In most situations, if an opponent
has not folded and has voluntarily put money into the pot, there is a much higher probability they are holding a hand of higher quality than average. Hands that are clear losers would normally be folded. It is too simplistic to count unlikely hands as having the same probability as likely hands.
To make the measurement more realistic, the enumeration can be weighted by the probability of each hand being held. Instead of adding one to each counter for *ahead*, *tied*, and *behind*, a weight representing the relative probability that an opponent has a specific hand can be used instead. Weighting the enumeration will be described in more detail in Section 3.3.3.
### 3.3.2 Hand Potential
Hand strength is a measure of the current strength of our hand. However, after the flop, the turn and river cards are yet to be dealt. These future community cards can significantly affect our hand strength. For instance, suppose Alice has K♣-K♠, and Bob has 7♥-9♥, with a flop of 8♥-6♣-4♥. At this point, Alice has a high hand strength (RHS = 0.945), while Bob has a low hand strength (RHS = 0.104). However, Bob’s hand has several *outs* (cards that improve the hand) that would give him a winning hand (any of 9 remaining hearts to make a flush, plus the T♠, T♦, T♣, 5♠, 5♦, and 5♣ for 6 ways to make a straight). The probability that any one of these cards will come down on the turn or river is high (15/47 unknown cards gives a 32% chance the next card will improve Bob’s situation). Despite the weak hand strength, Bob should stay in the hand because of the high potential.
The *Positive Potential*, or *PPot*, is the probability that our hand will *improve* after a new card is dealt. In general, improvement is measured by counting each situation where we are currently *behind* and will end up *ahead*, or *tied*. Likewise, *Negative Potential*, or *NPot*, is the count of hands where we are currently ahead, but will end up behind.
The actual potentials of a hand can be computed by enumeration of all the possibilities. On the flop, there are still two cards to be dealt, so a full calculation requires a two-card look ahead. There are 1,070,190 combinations in a two card look ahead (1081 possible opponent hands, multiplied by the 990 possible turn and river cards). On the turn, there is only one card to come, making for only 43,560 combinations. The algorithm for hand potential is given in Figure 3.4 [34]. A breakdown of cases from the example hand above is shown in Table 3.2.
\[
PPot = P(ahead \mid behind) + \frac{P(tied \mid behind)}{2} + \frac{P(ahead \mid tied)}{2}
\]
(3.1)
Using this equation we can compute the one and two card look ahead potentials for this example from Table 3.2. PPot1, the one-card potential is 0.406, while the two card potential, PPot2, is 0.620.
HandPotential(ourcards,boardcards) {
/* Hand potential array, each index represents ahead, tied, and behind. */
integer array HP[3][3] /* initialize to 0 */
integer array HPTotal[3] /* initialize to 0 */
ourrank = Rank(ourcards,boardcards)
/* Consider all two card combinations of the remaining cards for the opponent. */
for each case(oppcards) {
opprank = Rank(oppcards,boardcards)
if(ourrank>opprank) index = ahead
else if(ourrank==opprank) index = tied
else /* < */ index = behind
HPTotal[index] += 1
/* All possible board cards to come. */
for each case(turn) {
for each case(river) {
/* Final 5-card board */
board = [boardcards,turn,river]
ourbest = Rank(ourcards,board)
oppbest = Rank(oppcards,board)
if(ourbest>oppbest) HP[index][ahead ]+=1
else if(ourbest==oppbest)HP[index][tied ]+=1
else /* < */ HP[index][behind]+=1
}
}
}
/* PPot: were behind but moved ahead. */
PPot = (HP[behind][ahead] + HP[behind][tied]/2 + HP[tied][ahead]/2)
/ (HPTotal[behind]+HPTotal[tied]/2)
/* NPot: were ahead but fell behind. */
NPot = (HP[ahead][behind] + HP[tied][behind]/2 + HP[ahead][tied]/2)
/ (HPTotal[ahead]+HPTotal[tied]/2)
return(PPot,NPot)
}
Figure 3.4: Hand potential calculation.
| Before | After | Turn Cases | % | River Cases |
|--------|--------|------------|-----|-------------|
| Ahead | Alcad | 4380 | 9% | 89836 |
| Ahead | Tied | 0 | 0% | 4449 |
| Ahead | Behind | 480 | 0.9%| 12635 |
| Tied | Ahead | 81 | 0.17%| 3240 |
| Tied | Tied | 324 | 0.67%| 5625 |
| Tied | Behind | 0 | 0% | 45 |
| Behind | Ahead | 17663 | 36% | 590955 |
| Behind | Tied | 0 | 0% | 3010 |
| Behind | Behind | 25717 | 53% | 360395 |
Table 3.2: Breakdown of cases in a hand potential for the hand 7♥-9♥ with a flop of 8♥-6♣-4♥.
Just as hand strength can be computed by enumeration, so can hand potential. Figure 3.4 shows the enumeration algorithm for computing PPot and NPot. One card look ahead is computable on today’s hardware (see Appendix A) in roughly 280 milliseconds. Two card look ahead can be done in 2000 milliseconds, which may limit its usage in practical algorithms for real-time play.
### 3.3.3 Weighting the Enumerations
The calculations of hand strength and hand potential in Figures 3.3 and 3.4 assume that all two card combinations are equally likely. However, the probability of each hand being played to a particular point in the game will vary. For example, the probability that the opponent holds Ace-King is much higher than 7-2 after the flop, because most players will fold 7-2 before the flop.
During the course of each hand, a *weight table* is maintained for each opponent. This is simply an array of values that represent the probability of the opponent holding each possible combination of hole cards. For each of the 1326 entries (one for each two-card combination)\(^2\), we store a weight in the range 0.0 to 1.0. This weighted distribution is updated after each opponent action to reflect the hands that are consistent with the betting decisions observed throughout the current hand. The precise details of this *re-weighting* process depends on our method of modeling each opponent (described in Chapter 4).
The strength of each possible hand is assessed, and a probable distribution of actions is determined by a formula-based betting strategy. These values are then used to update the weight table after each opponent action. The algorithm is shown in Figure 3.5. For each possible opponent hand, the opponent model returns a probability distribution over the likely actions with that hand, given the current
---
\(^2\)We use a 52 by 52 table for indexing convenience. Weights for all card combinations containing known cards (in our hand and on the board) are set to zero.
UpdateWeightTable(Action A, WeightTable WT, GameContext GC, OpponentModel OM) {
foreach (entry E in WT) {
ProbabilityDistribution PT[FOLD, CALL, RAISE]
PT = PredictOpponentAction(OM, E, GC)
WT[E] = WT[E] * PT[A]
}
}
Figure 3.5: The algorithm for updating a Weight Table. The probability of the observed action (A) is taken from an opponent model’s (OM) prediction of the opponent (PT) based on the hand (E) and the current game context (GC). This probability (PT[A]) is used to weight the table entry (WT[E]).
game context. From this distribution, the probability of the actual observed action is used to multiply the weight table entry.
For example, suppose an opponent calls before the flop. The updated weight for the hand 7-2 might be 0.01, since it would normally be folded. The probability of Ace-King might be 0.40, since it would seldom be folded before the flop, but is often raised. The relative value for each hand is increased or decreased to be consistent with every opponent action.
Figure 3.6 shows the reweightings of an opponent’s weight table. The images are constructed to show the 1326 different hands in a 13 by 13 table. The axes represent the rank of each of the cards. Suited hands are in the upper-right half of the table, unsuited hands are in the lower-left half, and pairs are along the diagonal. The shading in the cell represents the maximum value for the hands grouped in that cell. The darker the cell, the higher its weight. For instance, for the Ace-King Suited cell (in the upper-right half of the grid), its value is determined by the maximum weight of the four different Ace-King suited hands (Ace-King of Spades, Ace-King of Hearts, etc...).
From Figure 3.6, we can see the difference in the weights for an opponent that folds, calls or raises in the pre-flop betting round. In the case of folding, the strong and mediocre hands are all shaded lightly or not at all. The weak hands (low cards, unsuited, unconnected) are more likely to be folded and thus have high (dark) weights. For the calling scenario the high weights are in the middle – hands that are consistent with being called. Weak hands would likely have been folded and strong hands would likely have been raised, so their weights are low. Likewise, the raising example shows strong weights only for the very good hands, and low weights for everything else.
To better understand how the weights are updated, consider the weight tables in Figure 3.7, generated from the sample hand in Chapter 1.2.1. The weight tables for Alice on the flop (J♠-8♣-4♠), after she bets $10, might look something like in
Figure 3.6: An example weight table after a pre-flop fold (left), call (middle), or raise (right).
Figure 3.7: Two scenarios for calling pre-flop (left) and raising pre-flop (right), followed by betting after a flop of J♠-8♠-4♠ (See Chapter 1.2.1).
Figure 3.7. On the left is a table for the scenario where she calls pre-flop, and on the right is how the table would appear if she had raised pre-flop (as she did in the example).
The table for the pre-flop call scenario (left) represents hands such as any hand with a Jack (except for a pocket pair of Jacks, because that hand would have been raised pre-flop). Hands that make a pair of Eights are likely, if they have a large kicker (for example, King-Eight has a high weight, but Eight-Three does not). Hands that make two-pair (like Jack-Eight) also have relatively high weights. The upper-right triangle of the table represents suited hands, and because there are two spades on the board, the suited half of table is darker than the unsuited half -- Alice could be betting a strong flush draw.
In the pre-flop raise scenario (right), There are several differences. Jack-Jack now has a high weight, since it would have been raised pre-flop, then bet on the flop. Other hands like Ace-Ace, and Ace-King are much higher than they were in the other scenario, since they too are hands that are more likely to be raised pre-flop rather than called. Hands like Four-Four are weaker than in the calling scenario, since Four-Four is less likely to be raised in the pre-flop round.
Noise-Factors
Each action a player makes conveys some form of information, however little, about the hand they are holding. The reweighting process attempts to modify the table based on this information. In order to do this accurately, we need to know how much information the observed action gives us, and how much this information should alter the contents of the weight table.
A noise-factor is used to dampen the amount of reweighting done to a table with each action. The noise-factor is a value between zero and one, that is extracted from the probability distribution \{Pr(fold), Pr(call), Pr(raise)\} used in reweighting, and then smeared into the distribution equally. For each action \(a\) in the probability distribution \(D\), \(a\) is updated by \(a \leftarrow a - a * \beta + \frac{\beta}{|D|}\), where \(\beta\) is the given noise factor.
A noise-factor of 0.0 leaves the probability triple untouched, whereas a noise-factor of 1.0 would flatten the triple into a uniform distribution. For instance, suppose we have a triple of \{0.0, 0.2, 0.8\} and adjust it with a noise-factor of 0.5. The resulting triple would become roughly \{0.167, 0.267, 0.567\}. That is, half of the distribution has been proportionally smeared toward uniformity (half of one-third is 0.167).
The noise-factor represents the amount of uncertainty we have about the opponent’s actions reflecting information about the cards they hold. If an opponent always plays without deception, using a noise-factor would be detrimental. However, if an opponent plays deceptively, some uncertainty must be kept in the weight table.
3.3.4 Pot Odds
‘A new car built by my company leaves somewhere traveling at 60 mph. The rear differential locks up. The car crashes and burns with everyone trapped inside. Now: should we initiate a recall?
You take the number of vehicles in the field (A) and multiply it by the probable rate of failure (B), multiply the result by the average out-of-court settlement (C). A times B times C equals X. If X is less than the cost of a recall, we don’t do one.’
‘Are there a lot of these kinds of accidents?’
‘You wouldn’t believe.’
‘... Which car company do you work for?’
‘A major one.’
-- Fight Club (1999)
Returning to our earlier example where Player A has K♦-K♠, and player B has 7♥-9♥, with a flop of 8♥-6♣-4♥, recall that B’s hand has high potential, but weak strength. Should B call A’s bet? Player B has a 32% chance of his draw coming in on the turn. If the pot size is currently $50, with $10 bet by player B, then it is correct for B to call the $10. B is getting a 32% chance to win with a 5-to-1 bet.
For a one bet investment, we will win 5 bets 32% of the time. The remaining 68% of the time we miss our draw – at worst we can fold and have lost only 1 bet each of these times. The return on our investment is ($50 \times 0.32) - ($10 \times 0.68) = +$9.20.
There are several different ways to estimate pot odds. For instance, the immediate pot-odds (the current odds for calling a bet), is formulated as:
\[
\text{immediate\_pot\_odds} = \frac{\text{amount\_to\_call}}{\text{pot\_size} + \text{amount\_to\_call}}
\]
(3.2)
Using this equation, our pot odds are $(10/($50 + $10)) \approx 0.167$. If we call in situations where we have more than a 16.7% chance of making a winning draw, then the expected payoff is positive. For instance, if we only had a 15% chance of winning, we would not have sufficient odds to call. We would win 5 bets only 15% of the time, and lose 1 bet 85% of the time for a net loss of $(50 \times 0.15) - (10 \times 0.85) = -$1.00.
Pot odds give the return on investment for calling a bet. We must have a chance of winning greater than the pot odds in order to win money.
Pot odds calculations should also take future betting rounds into account. If we make our hand on the next round, there is a good chance we will be able to win a pot much larger than the current size. The implied pot odds account for the payoff from winning future bets as well as the current pot. There is no simple formula to measure implied pot odds – players rely on judgment to calculate these odds.
Conversely, the reverse implied odds are used in situations where we are playing with a made-hand, but against an opponent who may be on a draw or may already have us beat. In the next betting round the opponent may miss their draw and fold, or they may make the draw or already have you beat in which case it will cost you more to see the showdown – increasing the cost of your investment to win the pot.
### 3.3.5 Multi-player Considerations
The Hand Strength computation gives the strength against one random hand. If the enumeration is weighted, it gives the strength against a player with the hand distribution reflected by the weight table. If we are up against more than one opponent then we must, of course, beat all of them. If a weight table is kept for each opponent, then a good approximation is to multiply the hand strengths for each weight table together. This is the probability that we have a hand better than player one, and player two, through to player \( n \).
This is not an exact value, because the hand strengths of the opponents are not independent of each other. For instance, the weight tables do not reflect the fact that two players cannot hold the same cards. However, the full computation is too complicated for a real-time system. Tests comparing the two methods in [34] showed that the error from Equation 3.3 was small (less than 3%).
Another data structure that is used for multi-player reasoning is the field array. This is another weight table that represents the average of all opponent weight tables. This allows us to reason about the whole field of opponents at once, and is a further approximation that reduces accuracy in exchange for speed. For instance, a hand strength against the field array, raised to the power of the number of opponents, can be a reasonable estimate of strength.
\[
HS_n = (HS_{field})^n \tag{3.4}
\]
The field array has been found to be a large source of error in Loki. Recent versions of Poki have eliminated most of the dependence on the field array, using it only when the regular calculations are too expensive.
A fundamental problem with the field array in Loki is that it is an average of the opponent weight tables. This seems to be the wrong abstraction. Suppose we are up against four opponents, and one has a very high probability of holding a strong hand such as a pair of Aces. Further, suppose the three other opponents have very low probabilities of having such a hand, but instead, they all have mediocre hands represented in their weight tables. If the field array is constructed from the average of the four tables, hands like a pair of Aces will have a low value (say, 0.25), while the mediocre hands held by the majority of opponents will have relatively high weights (say, 0.75). This biases the field array in favor of having the weaker hands. What matters most, of course, is whether we can beat the best hand in the field. If we cannot beat the pair of aces held by one opponent, it does not matter that we can beat the three other opponents.
Poki has been enhanced by constructing the field array from the maximum of each entry in the table, instead of the average, which is a more realistic approximation. Also, in many places where the field array was used in the past, it has been replaced with the more expensive, but more accurate Hand Strength calculation (3.3).
### 3.3.6 Effective Hand Strength
For purposes of simplification it can be useful to have a single value for the overall strength of a hand that includes potential. What we want is an estimation of the
probability that we will hold the best hand after all the cards have been dealt. We call this the *effective hand strength* (EHS). Equation 3.5 combines the hand strength, negative potential and positive potential, into a single value.
\[
EHS = HS \times (1 - NPot) + (1 - HS) \times PPot \tag{3.5}
\]
However, in practice, we remove the negative potential from the measure and use Equation 3.6. At first this may seem counter-intuitive. Negative potential is the chance that we are currently ahead, but further board cards will improve the opponent’s hand enough to beat ours. When the negative potential is high it is usually necessary to bet anyways since the *free card* danger is high. We must protect our hand and force drawing hands to pay for their draw.
\[
EHS = HS + (1 - HS) \times PPot \tag{3.6}
\]
### 3.4 Basic Betting Strategy
The basic betting strategy is a fairly simple formula-based system that uses the hand evaluation described above as input and outputs an appropriate betting action.
First we evaluate our Hand Strength, Potential, and Pot Odds. The formula generates a *probability triple* over the available actions, \(\{Pr(fold), Pr(call), Pr(raise)\}\). For instance, a very strong hand might generate a triple such as \(\{0.0, 0.05, 0.95\}\). This means that in this situation we should call five percent of the time, raise ninety-five percent of the time, and we should never fold. To make a betting decision, a uniform random number is generated between 0 and 1. This value is used to pick an action from the probability distribution.
Randomization is important to make the strategy less predictable [35]. If we always raised with strong hands, folded weak hands, and called drawing hands, we would quickly get taken advantage of by any perceptive opponent. When we raised, opponents would learn to fold, giving us only small pots. When we don’t bet, opponents would know to raise, forcing us to fold our weak or drawing hand. By using a properly mixed strategy, we can hide the strength of our hands.
The formula-based strategy makes an attempt to do this, but it is far from perfect. It is also sensitive to the accuracy of the input. If our opponent modeling is poor, the values generated for strength and potential will be bad estimates.
For many hands in Hold’em, the obvious strategy is sufficient. If we have a extremely strong hand (\(EHS > 0.85\)), we should almost always raise, with some percentage of check-raises. If we have a very weak hand (\(EHS < 0.25\)), we should almost always fold (unless we have a high potential that gives us correct pot-odds to call, or we think a bluff is in order). Somewhere in the middle, we should mix
our strategy to add unpredictability. Some moderate hands should occasionally be raised - this way, the opponent cannot be sure that we have a strong hand when we raise.
The formula is based around two thresholds for calling and raising. For hands that are well above the raise threshold, we raise nearly all of the time, and for hands well below the calling threshold, we almost always fold. Likewise, hands well in the mid-range are almost always called. The hands that are near the thresholds are treated differently. These borderline hands are mixed between calling and raising, or folding and calling, in proportion to their strengths.
This simple betting strategy plays a fairly solid and safe game, but it is too predictable. Any fixed formula will eventually show exploitable weaknesses (unless it is optimal). The formula could be developed further and further, patching weakness and adding more intelligence. This, however, is of minor scientific interest, as it would be little more than an expert system. Not much work has been done to improve the formula based system since *Loki-2* [35]. The formula was only ever intended as a first-cut approximation. It should have been revisited and improved to incorporate more knowledge and information hiding, but it never was. A fully adaptive knowledge-free system is of much more interest as the techniques could be applied more broadly to imperfect information domains.
### 3.5 Selective Sampling and Simulation Betting Strategy
In perfect information games like chess, minimax game-tree search algorithms such as alpha-beta [38] provide a dynamic method for discovering good moves based on minimal knowledge. The formula based betting strategy described in the previous section can be thought of as the equivalent of a static evaluation function in deterministic perfect information games. Since game-tree search works so well in these domains, the natural step is to test something analogous to game-tree search in the poker domain.
Unfortunately standard game-tree search algorithms like alpha-beta do not work for poker. The nature of the imperfect information game tree makes traditional search of little utility.
In most two-player perfect information games it is effective to search the game-tree several moves ahead, and then call an evaluation function. In these games it is possible to recursively search to find the opponent’s best response to our moves. In poker the best response for the opponent is unknown because we do not know what cards they are holding. The minimax principle does not apply because nodes of the game tree are not independent. It is insufficient to call an evaluation function part-way through the tree, and the game-tree is generally too large to search completely
to the leaves.\footnote{For example, Darse Billings has estimated the two-player Texas Hold'em game-tree to be $3 \times 10^{17}$ in size (counting the number of terminating nodes).} The leaf nodes of a poker game tree can be evaluated exactly, since the game-state is fully known. However, little can be inferred about the value of internal nodes without searching first to the leaves.
A technique known as \textit{Selective Sampling} can be used to assign probable values to the unknown variables (in poker, the unknown cards). Instead of doing a comprehensive but shallow search, a deep but sparse simulation can probe numerous times to the leaves. Each trial of the simulation selectively samples the search space. With enough trials the cumulative selective sample converges to a statistically confident result.
Selective sampling has been used successfully in other games such as Scrabble [43], backgammon [48], and bridge [20]. \textit{Likelihood weighting}, a similar technique used for Bayesian inferencing algorithms, is another way to bias stochastic simulations [19, 42].
Figure 3.8 shows the difference between standard game-tree search algorithms like alpha-beta (left), compared to the selective sampling simulation method. Alpha-beta will search only an upper region of the tree as deeply as possible, but there is essentially full coverage of the game-tree in the upper region. An evaluation is then applied to the nodes on the frontier. Simulations, however, search a selection of lines to the leaf nodes of the tree. Each line searched samples the game-tree in order to get a statistically relevant expected value for our moves [7].
When \textit{Poki} is faced with a betting decision, it can simulate each of its possible actions, and then simulate the game to the end of the hand. Since we don't know the complete game-state (the cards of our opponents, and the future board cards remain unknown), we must assign probable cards to each player, in each \textit{trial} of the simulation. The weight table of each opponent can be used to selectively bias the chosen cards to match the hands they are likely to hold [35, 8].
After simulating several hundred trials, the average amount won or lost by folding, calling, and raising will typically converge to a stable estimate of the \textit{Expected Values} (EVs) of each action. The action with the largest EV can be chosen as the
maximal action to take. This, of course, is not a proper mixed-strategy (unless the opponent modeling used in the simulations is sophisticated enough to take the effect of mixed-strategies into account). In practice, the problem of converting the EVs to produce an appropriate mixed strategy remains.
A simulation-based betting strategy for poker proceeds as follows. For each trial, probable hands must be assigned to each opponent to hold. These hands can be picked from the distribution of hands in each opponent’s weight table. We also choose hands for players that have already folded. This effectively changes the composition of the remaining deck by removing cards that are consistent with folded hands. Once each simulated player has a hand, the unknown board cards are chosen randomly from the remaining cards in the deck.
For each trial, the hand is simulated twice. In the first simulation, Poki’s first action will be a call (or check if there is no bet). In the second simulation, Poki will raise (or bet) for its first action. Once the first action is made, the rest of the hand is simulated to completion. When a simulated opponent must act, their opponent model is asked to predict what the opponent would likely do in the current context, with the hand we have assigned to them. Once the simulated hand is finished, the net amount won or lost is recorded.
After many of these trials the average EV for calling, and raising becomes fairly stable. This must be accomplished within the time constraints of a poker game (usually one or two seconds), limiting the number of trials (typically 200-300 trials for flop decisions).
This technique has several benefits over the formula-based system. It is not a fixed strategy – depending on the opponent models, it can shift dramatically, and adapt to any mix of different opponents. Computing properties like implied pot-odds and draw-odds can be difficult and expensive. Simulations uncover this information naturally, and produce a simple EV for each action. Furthermore, simulations can uncover complex strategies without any specific expert knowledge. For example, tactics such as check-raising, slowplaying, and bluffing, are discovered.
as an emergent property of the simulations.
The mix of possible cards throughout the trials helps provide realistic estimates of the value of our hand and actions. This process can work effectively, but it is also sensitive to errors and biases both in the simulation, and in the opponent models. Consider the situation where an opponent model incorrectly predicts that the opponent will fold 30% of the time in a particular context, but in actuality the opponent folds only 20% of the time. 10% of the simulations will erroneously predict that by raising, *Poki* will immediately win the money in the pot uncontested. This bias will greatly inflate the EV for raising and *Poki’s* resulting behaviour will be far too aggressive. Or consider a situation where the opponent model’s weight table is representing mostly mediocre hands, when in fact the opponent has *slowplayed* a very good hand. The simulations will consistently pick the wrong set of cards for that opponent, making the EVs unreliable and resulting in a greater loss.
However, given a correct opponent model, the simulation-based betting strategy works well. In general, improvements in the opponent models yield improvements in the accuracy of the simulations. Simulations are sound in principle, but they have not been robust in practice. Chapter 6 will revisit the problems with simulations and present a more robust formulation based on our new *Miximax* algorithm.
### 3.6 Summary
*Poki* is a complete reimplementation of the *Loki* AI. The fundamental algorithms for determining hand strength and hand potential involve weighted enumerations of the unknown cards. Weight tables are maintained for each opponent. These tables reflect the hands consistent with their actions throughout the hand. Two different betting strategies have been used to choose actions based on our hand evaluation. The formula based strategy is simple and well tested. The simulation based strategy is a much more adaptive strategy, but it is dependant on quality opponent modeling to get good results. The opponent modeling components of *Poki* will be discussed in the next chapter.
Chapter 4
Opponent Modeling
Knowledge is power, if you know it about the right person.
– Erastus Flavel Beadle (1821-1894)
No poker strategy is complete without a good opponent modeling system. A strong poker player must develop a dynamically changing (adaptive) model of each opponent, to characterize their style and identify potential weaknesses.
In traditional games, such as chess, this aspect of strategy is not essential to achieve a world-class level of play. In perfect information games, it is sufficient to play an objectively best move, without special regard for the opponent. If the opponent plays sub-optimally, then continuing to play objectively good moves will naturally exploit those errors. Opponent modeling has been studied in the context of two-player perfect information games, but the research has not led to significant performance benefits [11, 21, 22].
The situation with poker is much different. For example, one opponent may bluff too much, the other too little. We adjust by calling more frequently against the former, and less frequently against the latter. To simply call with the optimal frequency would decline an opportunity for increased profit, which is how the game is scored. Even very strong players can employ radically different styles, so it is necessary to deduce each opponent’s basic approach to the game, regardless of how well they play.
In poker, opponent modeling is used in at least two different ways. First, we want a general method of deducing the strength of the opponent’s hand based on their actions. For example, we would like to know if an opponent will bluff repeatedly on each round. Secondly, we would like to predict their specific action in a given situation. For example, if we are considering a bluff on the river, we will want to know the probability that the opponent will fold or call us down.
At the heart of an opponent modeling system is a predictor. The predictor’s job
is to map any given game context into a probability distribution over the opponent’s potential actions. In limit poker, this distribution can be represented by a *probability triple* \{Pr(fold), Pr(call), Pr(raise)\}. Guessing the next action is useful for planning advanced betting strategies, such as a *check-raise*, and is also used in each trial of a full-information simulation.
The difficulty of this task may not be apparent at first. Learning how an opponent plays poker, and how they will play poker in the future, is a formidable challenge for a multitude of reasons:
**Uncertainty:**
The numerous unknown cards in a game adds a great deal of noise to the game, making it extremely difficult to extract a signal. Each hand of poker can be completely different than the previous ones, simply from the variance of the cards dealt in the hands and on the board. A large number of hands must be played before even the most common situations are encountered several times.
**Missing Information:**
We only get to see a player’s cards if there is a showdown (usually only a small percentage of hands are revealed in a showdown). This means we can only fully verify our model on the few hands that our opponents take to the showdown (which is a heavily biased sample of the total number of hands).
**Unknown Dimensions:**
There is an unknown number of variables that affect how a player will play a hand. Some factors are much more important than others, and some factors are only relevant to particular players. For one player, their position in the betting round may be a strong criteria for their betting decisions, while position may not be a strong consideration for another player. A particular opponent may favour heart flush draws over flush draws with other suits. The day of the week could even be a factor for some players. One must always be on the lookout for significant correlations.
**Intuition:**
After a small number of observations, humans form elaborate theories about an opponent, based mostly on intuition and past experience. If a theory holds in future observations, it becomes integrated and used in future betting decisions. A human may even alter their play in order to test a theory. In contrast, machine learning methods typically need a large number of observations, and build a model from statistical inferencing. These statistical methods are passive and slow to adapt.
**Multiple Levels:**
If we observe an opponent in hands against another player, it may be wrong to predict that they will play similarly towards us. Assuming an opponent models his opponents, they will adjust their play to match each specific opponent. Not only does one need to model the opponent in general, but how the opponent models others (and we can regress further to how the opponent models how we model them, and so on...).
**Moving Targets:**
We can only learn about an opponent from past observations. A good opponent will change strategies over time. It may be the case that our model is obsolete by the time we have made enough observations to make good inferences about the opponent.
Opponent modeling in poker appears to have many of the characteristics of the most difficult problems in machine learning – noise, uncertainty, an unbounded number of dimensions to explore, and a need to learn quickly and generalize from a relatively small number of training examples, often with missing or misleading data. Furthermore, the real-time nature of poker (a few seconds per betting decision) limits the effectiveness of some popular learning algorithms.
### 4.1 Methods of Prediction
There are several ways of looking at the opponent modeling problem in poker. From a machine learning point of view, it is convenient to think of an opponent model as a predictor of the opponent. Given the knowledge of the opponent’s cards, a perfect model would be able to correctly predict what action the player would choose at each decision. The next best thing would be to predict an accurate probability distribution over the actions they might choose.
#### 4.1.1 Expert Systems
One way to predict an opponent action would be to use our own betting strategy, or some other set of rules, to make a rational choice on behalf of the opponent. When we use this type of fixed strategy as a predictor, we are assuming the player will play in a particular “reasonable” manner, and we refer to it as *generic opponent modeling*. This is not a very effective method of modeling, but it provides a decent baseline. It is a useful starting point for modeling when playing an opponent for the first time.
4.1.2 Statistics
Another obvious method for predicting opponent actions is to expect them to continue to behave as they have done in the past. For example, if an opponent is observed to bet 40% of the time immediately after the flop, we might infer that they will normally bet with the top 40% of their hands in that situation (perhaps including a certain percentage of weak hands that have a good draw). When we use an opponent’s personal history of actions to make predictions, we call it *specific opponent modeling*. As one would expect, specific opponent modeling outperforms generic opponent modeling [35].
Our first opponent modeling effort was based on the collection of simple statistical information, primarily on the betting frequencies in a variety of contexts. For example, a basic system distinguishes twelve contexts, based on the betting round (pre-flop, flop, turn, or river), and the betting level (zero, one, or two or more bets).
This history table is essentially a set of conditional action probabilities such as \( P(Call \mid River \& OneBetToCall) \). However, this is a very limited definition of distinct contexts, since it does not account for many relevant properties, such as the number of active opponents, the relative betting position, or the texture of the board cards (*eg.* whether or not many draws are possible). Establishing a suitable set of conditions for defining the various situations is not an easy task. There are important trade-offs that determine how quickly the algorithm can learn and apply its empirically discovered knowledge. If a context is defined too broadly, it will fail to capture relevant information from very different circumstances. If it is too narrow, it will take too long to experience enough examples for each scenario, and spotting general trends becomes increasingly difficult. Equally important to deciding how many equivalence classes to use is knowing what kinds of contextual information are most relevant in practice.
Furthermore, there are many considerations that are specific to each player. For example, some players will have a strong affinity for flush draws, and will raise or re-raise on the flop with only a draw. Knowing these kinds of personality-specific characteristics can certainly improve the program’s performance against typical human players, but this type of modeling has not yet been fully explored.
4.1.3 Neural Networks
To create a more general system for opponent modeling, we implemented a *neural network* for predicting the opponent’s next action in any given context. A neural network is a popular machine learning data structure, loosely inspired by biological neural structures.
Neural networks have several nice qualities, but of course there are also many caveats. They are easy to build and train without any domain-specific knowledge.
They typically provide reasonable accuracy, especially in noisy domains. However, they rarely can produce better results than a more formal system built from specific domain knowledge. Furthermore, it can be difficult to extract the knowledge a neural net has learned from a training set.
Neural Networks have been used with success in backgammon programs [49]. The networks in *TD-Gammon* were trained by playing over a million games against itself. The neural network’s ability to handle noisy domains was well suited to backgammon, which has stochastic elements. *TD-Gammon* was able to achieve a world-class level of play using this approach.
**Structure**
A neural network consists of neurons, or nodes, which are connected to each other by weighted directed edges [29, 30]. Input nodes (containing the problem data) are connected to internal (or ‘hidden’) nodes, and those internal nodes are connected to a set of output nodes. The weighted connections, which can be both negative and positive, transform the signal from the input nodes, through the internal nodes, and to the output nodes. The internal nodes take the sum of all connections entering them, and use a threshold function to decide the output sent to all outgoing connections. For instance, if the sum of inputs is above a threshold, it may output the value ‘1’, otherwise it outputs ‘0’. More gradual functions can be used such as a sigmoid function which gives a smoother output.
Figure 4.1 shows a simple neural network consisting of four inputs, two internal neurons, and one output neuron. The strength of the inputs are represented by the amount of shading in the input neuron. Connections are shown between each node, where the thickness represents the weight of the connection. Black connections have positive weights, while grey connections have negative weights. In this example, the internal neurons use a threshold function to decide their outputs, and the output neuron uses a sigmoid function to decide its output. Note that in this example, the two input neurons have different thresholds for firing. This is known as a *bias*, and each neuron can have a positive or negative bias affecting it.
Depending on the input values, and the relative values of the weights, each of the internal neurons may or may not fire, which in turn will effect the value of the output neuron. The first internal neuron may only fire when a particular pattern of inputs is detected. The trick is building the correct set of weighted connections that cause the desired behaviour of the network.
**Backpropagation**
Fortunately, there is a gradient-descent learning algorithm for neural networks called *backpropagation* [37]. Backpropagation is a method of working backwards through
the neural network, starting from the output and ending back at the input. For a given training example the output on node $i$, $O_i$, is compared to the correct answer to determine the network’s error $E_i$. Starting with the weights at the output nodes, the weights are adjusted in small increments so that the error is reduced. A weight is adjusted proportionally to its contribution to the error.
To assess the blame for a weighted connection $(W_{j,i})$ from an internal node $j$ to an output node $i$, the error at the output node ($E_i$) is multiplied by the derivative of the activation function $g$ along the input value ($I_i$) to the node.
$$\Delta_i = E_i * g'(I_i) \quad (4.1)$$
This determines both the direction of the error, as well as the share of the blame this weight had on the error realized at the output. The weight is updated by adding the $\Delta_i$ (i.e. its share of the error) times the activation level ($O_j$). A learning rate $\alpha$ is also used to increase or decrease the amount of change made to the weight for each training example. A high learning rate will allow the network to converge more quickly, but it may overstep the correct configuration.
$$W_{j,i} \leftarrow W_{j,i} + \alpha * O_j * \Delta_i \quad (4.2)$$
After updating the weights in the output layer, the blame for the errors at the output layer is propagated into the internal layer, since it is likely to share in causing the error. The blame for error at an internal node connection is determined as:
$$\Delta_j = g'(I_j) \sum_i W_{j,i} \Delta_i \quad (4.3)$$
The new delta is computed by propagating the deltas from the output layer. The reweighting rule is essentially the same as the output layer reweighting equation [38].
With each training example, the error in the network is determined and the weights are modified using this backpropagation method. The learning rate $\alpha$ determines the magnitude of the changes made during reweighting. As a network is trained on each sample, the weights are adjusted to minimize the error. All of the weights in the network are modified proportionally with respect to their relative contribution to the total error. After several training iterations, the network usually converges to a set of weights that minimizes the error over all examples in the training set.
**Poker Neural Nets**
A standard feed-forward neural net (also known as a *multi-layer perceptron*) was trained on contextual data collected from online games against human opponents. The networks contain a set of eighteen inputs corresponding to properties of the game context, such as the number of active players, *texture* of the board, opponent’s position, and so on (Table 4.1). The output layer consists of three nodes corresponding to the fold, call, and raise probabilities. By graphically displaying the relative connection strengths, we are able to determine which input parameters have the largest effects on the output. After observing networks trained on many different opponents, it is clear that certain factors are dominant in predicting the actions of most opponents, while other variables are almost completely irrelevant.
The inputs encode public knowledge of the game context. All inputs range from values of zero to one. For instance the current betting round is encoded using two Boolean inputs. If it is the flop, then both input #5 and #6 are zero.\footnote{Pre-flop actions are not modeled by the neural networks. We are only concerned with post-flop prediction here.} If it is the turn, then input #5 is set to one. If it is the river, input #6 is activated. Inputs #3 and #4 signal if there are bets to call, and inputs #9-12 encode information about the number of players and positions.
Inputs #13 and #14 are the estimated hand strength and potential of the cards the opponent holds. Inputs #15 and #16 indicate what the formula based system would do in this context. In cases where the opponent’s cards are unknown, a representative hand is chosen from the opponent’s weight table. This hand is used to generate the estimated hand strength and potential.
Figure 4.2 shows a typical neural network after being trained on a few hundred hands played by a particular opponent. The inputs are on the top row, with the activation level ranging from zero (fully white) to one (fully black). The thickness
\footnote{For a supplementary explanation of backpropagation, see the Appendix B.}
| # | Type | Description |
|----|----------|-----------------------------------------------------------------------------|
| 0 | real | immediate pot odds |
| 1 | real | bet ratio: $bets/(bets + calls)$ |
| 2 | boolean | committed (has put money in the pot this round) |
| 3 | boolean | one bet to call |
| 4 | boolean | two or more bets to call |
| 5 | boolean | betting round = turn |
| 6 | boolean | betting round = river |
| 7 | boolean | last bets called by player > 0 |
| 8 | boolean | player’s last action was a bet or raise |
| 9 | real | $0.1 * numPlayers$ |
| 10 | boolean | active players is 2 (heads-up) |
| 11 | boolean | player is first to act |
| 12 | boolean | player is last to act |
| 13 | real | estimated Hand Strength for opponent |
| 14 | real | estimated Hand Potential for opponent |
| 15 | boolean | expert predictor says they would call |
| 16 | boolean | expert predictor says they would raise |
| 17 | boolean | Poki is in the hand |
Table 4.1: Neural network inputs.
Figure 4.2: A neural network predicting an opponent’s future action.
of the lines represent the magnitude of the weights (black being positive, grey being negative). In this example, the connections from input node number eight (true if the opponent’s last action was a raise) are very strong, indicating that it is highly correlated with what the opponent will do next. The bottom row shows the network predicting that the opponent will probably bet, with a small chance of checking.
By examining the graphical representations of neural nets trained on an opponent, the public game features that have the highest correlations with the opponent’s actions can be spotted. Inputs that are strongly correlated with the actions, will have relatively stronger weights (\textit{i.e.} line thickness). Early experiments helped reveal strong indicators for detecting how an opponent will play. It also became clear that many strong factors could differ significantly between each specific opponent.
The first result of this study was the identification of new features to focus on when modeling common opponents. This produced a relatively small set of context equivalence classes which significantly improved the statistical opponent modeling reported previously [13]. It was discovered that the player’s last action is highly correlated with their future actions. This makes intuitive sense – if a player has just bet, there is a very small chance they will fold, and a much larger chance that they will call or raise. Likewise, if their last action was to check, then there is a much higher chance they will fold, rather than raise. As a result, the player’s last action was added to the context of the action frequency statistics. This improved the accuracy of the statistics by 10-to-20% on average.
Perhaps the biggest drawback of neural networks, for our purposes, is that they do not output a proper probability distribution over the likely outputs because they are not trained on a probability distribution, but the actual events. This skews the output to represent the most likely action rather than the most accurate distribution over the possible actions.
### 4.1.4 Decision Trees
Decision trees are another good way to solve classification problems such as the poker action prediction problem. A decision tree is navigated by asking a question at each node, and depending on the answer, traversing down a particular branch, towards a leaf node. At a leaf, there is a count of the frequencies of choices in the training set that would fall under that leaf node.
Figure 4.3 shows an example of a decision tree. Given a set of training data, a tree can be created to classify the data based on some simple rules. A typical method is to start with a single node, then split the node on one of the features such that the feature used results in a split that maximizes the information gain.
In a past feasibility study, we used Paul Utgoff’s \textit{Incremental Decision Tree Induction} software [51] to test decision trees on action prediction. Due to the high
amount of noise in poker data, all training was done with tree pruning. Pruning after training helps decrease the possibility of over-training and generally improves performance as a result. Preliminary tests on smaller data sets clearly indicated that the decision trees with pruning returned more accurate predictions of the opponent’s next move.
While not as robust or noise-tolerant as neural networks, decision trees have several benefits. They can output accurate probability distributions over the available choices, which neural networks cannot do. A decision tree’s representation is also a more human-understandable format, so determining what knowledge has been encoded in the tree is much easier.
In preliminary offline testing, decision trees achieved accuracies slightly below the neural networks. While Poki does not have any decision-tree software built in, it is a potential avenue for exploration in the future.
### 4.2 Confusion Matrices
In order to examine what types of classification errors the predictors are making, we construct *confusion matrices*. A confusion matrix displays the distribution of classifications across the combinations of predicted actions and the correct actions. With this matrix, one can see what percentage of test examples were expected to be *folds* when they were actually *calls*, and so on. The diagonal cells are correct classifications and the totals are shown in the margins.
An important use of these matrices is to see what the predictor has learned. For
instance, by examining a matrix one may find that a predictor getting 67% accuracy is merely choosing call 100% of the time, which the opponent happens to do 67% of the time. In this case we can see that all that has been learned is to pick the most frequent action overall. It has not learned how to distinguish the situations in which the opponent folds or raises from that in which the opponent calls.
We may also observe the tendency to make certain specific types of errors with an opponent. For instance, we may see that when the predictor predicts a fold, it is always correct, but ten percent of the time it incorrectly predicts a raise as a call. Table 4.2 shows an example confusion matrix. Certain types of errors are non-existent – the predictor in question never predicts a fold when the opponent actually calls, and it never predicts a raise when the opponent actually folds. It does make errors in three other cases. For example, it sometimes mistakes a raise for a call – that is, when it predicts a call, $\frac{3.2}{30.4} = 10.5\%$ of those times it is actually a raise. Certain types of errors can be more serious than others. For instance, incorrectly predicting folding can lead to massive errors in EV calculations, while confusing a call with a raise will have a much smaller impact.
Knowing which situations confuse our predictors, a program can make a more informed decision. As mentioned above, it is typical to see an opponent with a high bias towards one of the three actions. Some players have been observed to make a particular action as often as 75% of the time. If we imagine using the simplest predictor imaginable – one that simply chooses the most frequent overall action – it is trivial to get better than 50% accuracy for most opponents. One must keep this in mind when evaluating the more advanced learners. The first 50%-60% accuracy is trivial. The next 10-20% can be far harder to reach.
### 4.3 Multi-Predictor Voting
With all of these different methods of predicting opponents, it can be difficult to know which to use. An elegant way of solving this problem is to use all of the methods, and let them vote. Each vote can be weighted by the current accuracy.
of the predictor. When one predictor makes a mistake, there is a good chance that some of the others will pick correctly. This helps to reduce errors, and make the system more robust. It has been observed that the multi-predictor voter system is often slightly more accurate than any of the individual predictors.
In *Poki* the *meta-predictor* system dynamically tracks the accuracy of the predictors. The accuracy of the last $n$ actions of a player are used (where $n$ is typically set to a recent range of the last 25–100 actions) to dynamically evaluate the predictors.
Each predictor votes for an action, and each vote is weighted based on that predictor’s accuracy. To estimate accuracy, a confusion matrix is computed over the last $n$ actions. If the predictor is predicting a fold, for instance, we can examine the confusion matrix to see how accurately it predicts folding. If this predictor is often wrong about the opponent folding, its vote will not count for much. For the $m$ predictors, each probability triple $\Phi_i$ is multiplied by the accuracies for that predictor:
$$\Phi = \sum_{i=1}^{m} \Phi_i \cdot a_i$$
For instance, if $\Phi_i = [0.1, 0.85, 0.05]$ and the accuracies $a_i = [0.9, 0.5, 0.4]$, this means there is a 90% chance that the 10% fold prediction in $\Phi_i$ is correct, and so on. We treat the probability triple and accuracies as vectors and take the dot product to get a weighted distribution of $[0.09, 0.425, 0.02]$ put towards the final vote.
The resulting weighted sum of predictions, $\Phi$, is then normalized to a proper probability distribution, and used as the final prediction.
This interesting technique adds a second level of prediction, in that it takes both the predictions and knowledge about the prediction quality into account. The *meta-predictor* is sometimes more accurate (by a few percentage points) than any of the individual predictors. It allows for the seamless addition of any number of different prediction systems, and dynamically adjusts its use of each one to adapt to the current game. The system could perhaps be improved with a better, nonlinear weighting function. For instance, the weights given to a predictor with 90% accuracy and to one with 80% accuracy are quite similar, but it might be beneficial to give a much higher relative weighting to the first one.
### 4.4 Summary
Opponent modeling is a critical component for a world-class poker program. The problem of opponent modeling can be tackled with numerous methods including statistics, neural networks, decision trees, and other novel approaches. *Poki’s* architecture defines opponent modeling as an action prediction task. Any method of action prediction may be implemented and plugged into *Poki*. The *Meta-Predictor*
system will automatically evaluate each predictor for each opponent and use each predictor accordingly. Action predictions can be used in reweighting to assess the likelihood that an opponent holds a given hand. The predictions may also be used in simulation based betting strategies in order to anticipate future outcomes of a game.
Chapter 5
Experiments
The art of drawing conclusions from experiments and observations consists in evaluating probabilities and in estimating whether they are sufficiently great or numerous enough to constitute proofs. This kind of calculation is more complicated and more difficult than it is commonly thought to be...
— Antoine Laurent Lavoisier (1743-1794)
Measuring the performance of a poker-playing program is difficult. Poki is a complex system of interacting components. Changing a single component often has cascading effects, leading to unpredictable and unforeseen behavior. A variety of methods for assessing the program can be employed, but none of them is completely adequate in isolation.
Poker is a game of high variance. The element of luck strongly effects the outcome of each individual hand. Even two players of equal strength may see many anomalous games and long runs of good or bad luck. Measurements are susceptible to large amounts of noise and many thousands of hands are required to gain confidence in any result. To complicate matters, perceptive players are not static – through the playing of several hundred or thousand hands of poker these players will adapt their play. In addition to the challenge of measuring a noisy result, our target is constantly moving as we measure it.
Self-play experiments are commonly used to compare different features, variations or parameters of a strategy. The games in a self-play experiment can be controlled to minimize some of the variance. For instance to reduce the variance from lucky cards, the same hands can be dealt many times, but to a different player each time. The seating order can also be controlled to reduce the effects of positional advantage.
A major caveat of self-play experiments is that the games lack a wide variety of different styles and conditions that human players provide. Other researchers have previously commented on the “myopia” of self-play games in chess [2]. The problem is far more of a hindrance in poker. Poker styles can vary greatly, and the style of the opponent is a key element in playing correctly.
A self-play experiment that proves one strategy dominates another lends little ability to generalize the results beyond the experiment itself. While a player may be able to beat a specific opponent, it may flounder against another. The results are non-transitive: player A beats player B, player B beats player C, yet player C beats player A. This problem is commonly observed when testing and building computer strategies for poker. It is easy to build a program that takes advantage of a strategic weakness in another program. However, such a program will more often than not prove to play weakly against a strategy for which it has not been tuned to beat (recall the RoShamBo example about the ability to easily build an Iocaine Powder beater that still cannot win a tournament). A program that does very well against normal opponents may be vulnerable to a particular type of erratic or irrational player, even if their play is objectively worse.
Numerous different experiments will be presented in the following sections. First, two different methods of reweighting shall be examined in Section 5.1. In Section 5.2 several tests are conducted to evaluate the opponent modeling methods. The tests include self-play experiments, online games with humans, and offline tests of opponent models. Section 5.3 will discuss the actual performance of the different programs and strategies from online games.
5.1 Evaluation of Reweighting Methods
The accuracy of a weight table is paramount to the success of simulation-based betting strategies and to obtaining accurate hand evaluations for the formula-based strategy. The following experiment was designed to test two different reweighting methods, as well as to find the best table noise levels for heads-up games.
5.1.1 Reweighting System A
In both Loki and older versions of Poki reweighting used an error-prone method of opponent modeling to predict the action probability distribution for each entry in the weight table. The model would return the general action frequencies for the current game context – for example, a triple of \{0.2, 0.5, 0.3\} on the flop with one bet to call. This triple would then be used to generate the estimated call and raise thresholds for the player by directly mapping the frequencies onto effective hand strengths. Since the opponent raises 30% of the time, our model would state that the opponent will generally raise with hands that have a EHS of 0.7 or greater. The problem with this
Figure 5.1: A histogram of over 100,000 effective hand strengths (EHS) taken from Poki in ten-player online poker games (combined from all post-flop betting rounds).
mapping is that it incorrectly assumes that 30% of the opponent’s possible hands have an EHS of 0.7 or greater. Typically, this is not true – the distribution of EHS is not linear, nor is the distribution the same for each board.
**EHS Distribution Histograms**
Each board has a unique texture determined by the number and types of draws, and the selection of high and low cards. Figure 5.1 shows a histogram of the EHS distribution over a wide variety of boards (the EHS values were collected from thousands of online games). It is clearly a non-uniform distribution. A majority of hands have either low or high strengths, with relatively few hands in between. Intuitively, most hands will either miss-match the board cards, or match to make a strong hand. Figure 5.2 shows a separate histogram for each betting round. Many hands are folded on the flop and turn, so the overall number of instances drops in each progressive round. The overall shape of the distribution remains fairly constant between betting rounds.
### 5.1.2 Reweighting System B
A new approach was devised to pick more correct thresholds for reweighting. Before reweighting, the EHS of every entry is computed and placed into a sorted list. If the opponent raises 30% of the time, the EHS 30% from the top of the list is chosen as the raise threshold. Likewise, if they fold 20% of the time, the EHS 20% from the bottom of the list is chosen as their folding threshold. This allows the thresholds to be dynamically computed for any EHS distribution.
Figure 5.2: Histograms of roughly 55,000 effective hand strengths (EHS) taken from *Poki* in two-player online poker games. Each betting round is shown separately.
Figure 5.3: Simulation using roll-outs (v.s. FBS-Poki) with different noise factors. All of the SBS programs lose to FBS-Poki, so shorter bars are better (a smaller loss).
5.1.3 Experimental Design and Results
A random number seed was chosen to generate a fixed set of 2000 random hands. A match to evaluate two strategies consists of playing the set of 2000 hands twice. The second time, the two players switch seats so that each player gets the other player’s cards from the previous run. This eliminates much of the variance from the experiment, since both players get the same opportunities from any lucky and unlucky cards. For all of the following matches the same set of 2000 hands was used, so that comparisons between different strategies is possible.
Ten levels of reweighting noise-factors were tested in increments of 0.1. A simulation-based strategy (SBS) was tested against a standard formula-based strategy (FBS). The simulation-based player uses no action prediction, but instead always predicts a call, such that the simulation just becomes a roll-out to the showdown. This emphasizes the quality of the weight tables in the simulations as the primary factor in determining expected value. Since it does not simulate future bets and raises it cannot discover tactics like slow-playing or check-raising.
There is a clear disadvantage in using no noise in the reweighting. Without some uncertainty added, the weight table becomes far too committed to a small set of hands. Some noise is needed, otherwise the opponent’s actions are given far too
much credit. A noise-factor of 0.4 seems to be (in this experiment) the best level for heads-up poker, given the current system. The newer reweighting system is also clearly better than the previous version, especially with low noise-factors (with high noise-factors, the reweighting is rendered increasingly irrelevant).
### 5.2 Evaluating Predictors
In online games played on our poker server, there are both humans and computer programs ('bots') playing in full ten-player games. After each hand, the recent accuracy of each prediction method was logged. The average accuracy of each method is displayed in Table 5.1. Average accuracies (taken from several thousand logged games) for both bots and humans are listed.
The *Statistics I* uses the basic action frequency tracking that was used in *Loki-2* and in older versions of *Poki*, but with an enhanced context [13]. *Statistics II* uses a more sensitive, but similar technique that does not keep statistics for all observed actions, but only the last 40 in each context class. The *Expert Formula* uses *Poki*'s own formula based betting strategy to predict the opponent. The only difference between *Neural Network I* and *Neural Network II* is the training data used. In the former, the network is trained on all actions ever observed by the opponent. In the latter, a new network is created and trained from random weights with each new session played with the opponent. That is, if the opponent plays for a few hours one day, and returns to play two days later, NN-I will remember the past sessions, and NN-II is restarted from scratch. The *Meta-Predictor* is using all five of these predictors to form its own predictions.
Note that as expected, *Meta-Predictor* is the best method, faring slightly above the best sub-prediction method that it uses. The standard deviation of the prediction accuracy is also shown for humans. Note that the *Meta-Predictor* has a lower standard deviation than most of the other predictors. This means that it is able to consistently maintain its high accuracy.
The experiment above shows the average, stable prediction accuracies. Another data point is needed to see how quickly the predictors can model a new opponent.
| Predictor | Bots | Humans | $\sigma$ |
|--------------------|--------|--------|----------|
| Statistics I | 60.94% | 66.76% | 0.023 |
| Statistics II | 63.94% | 69.04% | 0.115 |
| Expert Formula | 67.87% | 55.98% | 0.061 |
| Neural Network I | 64.67% | 79.34% | 0.086 |
| Neural Network II | 73.49% | 79.91% | 0.054 |
| Meta-Predictor | 75.11% | 80.49% | 0.023 |
**Table 5.1:** Average predictor accuracy in online games.
From a set of training data collected from online games, an offline experiment can be done to test the different predictors.
Given a set of training data (containing past game situations, and actions the opponent took) we can use any kind of machine learning technique to try and classify the situations where the opponent took each different kind of action. Once a model has been trained, a second set of test data is used to evaluate the model’s accuracy of predicting the actions the opponent will take. This testing data should not be in the training set, otherwise the evaluation may be biased in favour of the model. Testing data should always be on instances of unseen data. This is the classic way of evaluating a machine learning system. However, there are characteristics of our chosen domain that allow us to perform a more realistic evaluation.
Actions in poker are temporally ordered. Opponents may significantly change their methods of play over time, or play particular hands in different ways. In an applied poker system, opponent modeling data is collected one instance at a time. To mimic this chronological ordering, a more precise evaluation is needed. When predicting an opponent’s next action, the predictor may draw upon all past data for the opponent. The classic method of dividing the past data into two large sets of training and testing data does not accurately represent the way poker data is available in a real game. Iterating through the data set, we keep count of how many actions we can correctly predict, when all prior items in the set are known, but all future instances are unknown.
This method calculates two types of information. By graphing the cumulative number of correct predictions over time, we can observe how quickly the system learns the opponent’s style of play, and we can see how it maintains accuracy as the opponent changes styles. We also have the overall percentage of accurate predictions. This temporal ordering method directly mimics opponent modeling constraints in a live game situation.
In this test, opponent actions from several thousand online games were collected for 24 different opponents. Two of these opponents were bots – one using FBS, the other SBS. The rest were regular human players who had played a sufficient number of hands. In Table 5.2 five methods were used to predict these opponent’s actions. In the classic evaluation, the data set for an opponent was split into two equal portions. The first half was used to train the models and the second half was used to test the accuracy. In the incremental test, the predictor was asked to predict each item in the data set one at a time, and told the correct answer after guessing. The difference between the two results was surprisingly similar. On average, the Meta-Predictor performed the best, by a small margin, followed by the Neural Networks.
The ‘always call’ predictor, as the name implies, always predicts a call. Since this is, by far, the most common action made in the game, it sets a good base-line.
| Predictor | Classic | Incremental |
|--------------------|---------|-------------|
| Always Call | 57.4% | 57.4% |
| Statistics Loki-2 | 57.5% | 57.5% |
| Statistics I | 60.4% | 60.6% |
| Statistics II | 60.7% | 60.9% |
| Expert Formula | 54.7% | 54.9% |
| Neural Network | 68.4% | 66.1% |
| Meta-Predictor | 70.4% | 69.5% |
Table 5.2: Average predictor accuracy.
We can see that it is trivial to attain an average accuracy of 57.4%. It takes much more effort to gain any additional accuracy.
5.2.1 Self-Play Results
A self-play experiment was conducted to assess different predictors for the simulation based strategy. The conditions were exactly the same as the experiment above where two versions of the SBS played against $FBS\text{-}Poki$. The SBS players all used the newer reweighting method discussed in Section 5.1. The three different prediction methods were *Always Call*, the *Expert Formula* (*i.e.* the FBS was used as a predictor), and the *Meta-Predictor* which encapsulated several predictors (statistics, expert formula, neural networks).
The results are shown in Figure 5.4. Again, the SBS players all lost to $FBS\text{-}Poki$ in these self-play matches. Overall, the *Meta-Predictor* SBS lost the least. The *Expert-Formula* SBS was not too far behind, although one would expect this predictor to perform well against $FBS\text{-}Poki$ since the predictor is nearly identical to $FBS\text{-}Poki$’s actual strategy. Against other opponents, *Expert Formula* would not perform as well.
5.3 Performance Results from Online Games
Perhaps the best way to test a program is by playing games against a variety of strong human players. A poker program can participate in a real game with willing participants, using a laptop computer on the table. This turns out to be surprisingly difficult, due to the fast pace of a real game and the amount of information to be entered. Even with numerous single-character accelerators, text entry is a bottleneck to the process. A well-designed graphical interface might help considerably, but it is currently more practical to have human players participate in a virtual game, rather than having programs compete in the physical world. Online games with humans have been our best method of empirically assessing $Poki$.
Figure 5.4: Performance of three different simulation predictors against FBS-Poki. Shorter bars are better (a smaller loss).
5.3.1 IRC Poker
Poki played frequently on Internet Relay Chat (IRC) poker against humans and the occasional other ‘bots’ (as online AI programs are named). A special IRC server allows humans and bots to connect and play poker games against each other online. No real money is at stake, but the accumulated bank-roll for each player is preserved between sessions, and a variety of statistics are maintained. There is a hierarchy of games for limit Hold’em, and a player must win a specified amount in the introductory level games to qualify for the higher tiered games.
The lowest level games (open to everyone) vary from wild to fairly normal, offering a wide variety of game conditions to test the program. The second and third tier games resemble typical games in a casino or card room. Most of these players take the game seriously, and some are very strong (including some professionals). Since Poki has been a consistent winner in these higher tiered games (and is in the top 10% of all players on the server), we believe the program plays better than the average player in a low-limit casino game.
Performance is measured in small bets won per hand (sb/hand). This measure is independent of both the speed of the game (hands per hour) and the bet sizes. This enables comparison between many different games. Since no variance reduction methods are available for online games, we generally test new algorithms for a
minimum of 20,000 hands before interpreting the results. On this scale, the trends are usually clear and stable amid the noise. Unfortunately, it can take several weeks to accumulate this data, depending on the popularity of the online game in question.
In recent play on the IRC poker server, *Poki* has consistently performed between +0.10 and +0.20 sb/hand in the lowest level games, and between +0.07 and +0.10 sb/hand in the higher tiered games against stronger opposition. In a game of $10/$20 Hold’em with 40 hands per hour, an income rate of +0.10 sb/hand translates into $40 per hour.
The results of simulation-based betting strategies have so far been inconsistent. Despite some programming errors that were discovered later, the earliest (1998) versions of simulation-based *Loki* outperformed the regular formula-based version in both self-play experiments (+0.10 ± 0.04 sb/hand), and in the introductory level games of IRC (+0.13 sb/hand vs +0.08 sb/hand). However, it lost slowly in the more advanced IRC games, whereas the regular version would at least break even.
Figure 5.5 shows that both the regular FBS betting strategy (labeled “FBS-Poki”) and the simulation-based betting strategy (labeled “SBS-Expf”) win at about +0.20 sb/hand in the introductory level games on the IRC poker server. However, there is a clear difference in the more advanced games, where the regular betting strategy routinely wins at about +0.09 sb/hand, but the simulation-based version could only break even (peaking at +0.01 sb/hand after 5000 hands, but returning to zero after 10000 hands).
As the quality of human players on IRC has deteriorated recently and the games have become less frequent, (many of the better players have left the play-money games for the more lucrative real-money online poker rooms) the most recent versions of *Poki* have not been tested on IRC.
### 5.3.2 Results From the Poki Poker Server
For nearly two years, we have hosted our own online poker server. Although not as popular as IRC Poker, the fact that we can control the server allows for greater flexibility in testing and evaluating different strategies. Using a free Java web applet, users can play at a table with poker programs and other people. Computer programs can support our TCP/IP based poker communication protocol and join in the different hosted games. The mix of humans and different poker programs provides an interesting hybrid between self-play experiments and games against humans.\footnote{See \url{http://www.cs.ualberta.ca/~games}.}
This game consists of several computer players, and at least one human opponent at all times. Since the artificial players have a style that is quite conservative, this game is quite a bit *tighter* than most IRC games. As a result of the tougher field, the income rates are much lower than in the IRC \#holdem1 games. The relative performance of the different players against the same field is more important than the absolute scores.
Figure 5.6 shows some recent results from our server. The top graph is from games played on the server in early-to-mid 2001. The bottom graph is for late 2001 to early 2002. Again, it is important to note that the field grew substantially harder in the second time-period. Some of the weaker bots were replaced with stronger ones, and the site’s popularity grew substantially as well, bringing many more strong human players to the game.
The performance of *FBS-Poki* dropped from +0.12 sb/hand to +0.066 sb/hand between 2001 and 2002. Stronger humans find the *FBS-Poki’s* play to be fairly predictable after playing a few hundred hands. *SBS-Meta* is a SBS player that uses the meta-predictor in its simulations. *SBS-Expf* uses the FBS as a predictor (i.e the Expert Formula). Not surprisingly, *SBS-Meta* outperforms *SBS-Expf*, earning +0.11 sb/hand, for a 40% overall improvement in this particular game. Although *SBS-Meta’s* income rate dropped between the two time periods (+0.11 sb/hand to +0.097 sb/hand), its relative performance to the *FBS-Poki* improved significantly. This is likely due to some small bug fixes and minor enhancements.
The *FBS/SBS* player was designed to be less predictable – for each hand it flipped a coin. If the coin is heads it plays using FBS; tails, it plays as *SBS-Meta*. The *FBS/SBS* program performed, interestingly, in between the FBS and SBS programs with an income rate of +0.075 sb/hand.
Figure 5.6: Poki poker server results.
One of the strongest human players that plays regularly is also shown in the bottom graph. In 17500 hands, this player has achieved an income rate of +0.18 sb/hand. As one can see, there is much room for improvement in the play of the bots. Very strong human players could likely double or triple this rate.
**Poki Poker Heads-up Contest Results**
There are many problems with the experiments from IRC and the 10-player room on our Poker Server. For one, the field is always changing. Players come and go, some play terribly, while others are quite strong. Over time, the conditions can change substantially. A more controlled experiment against humans was desired. It was decided that a controlled contest would be a good way to gather results.
In order to test three different versions of *Poki*, a special online contest was conducted. Each contestant played 500 hands of $10-$20 Limit Texas Hold’em against each different program. Of the nearly fifty players that attempted to finish the contest, only 11 completed the full 1500 hands required. It is important to note that many of the weaker players became discouraged by their performance and quit before completing the contest. As a result, the eleven players that finished represent some of the better players from our server.
In each room (each version of *Poki* played in its own room), the first 150 hands were random, and the last 350 were the same set of hands. This was done to reduce the variance from one player getting luckier cards than other players. Humans are notoriously good at noticing repeated hands, so the 150 random hands at the start were introduced to make it difficult to notice. Unfortunately, each room had to have a different set of 350 hands, otherwise the participants could unfairly recall hands in the subsequent rooms. This means that while the results between players within a room are comparable, the results between rooms are subject to much higher variance. Designing an experiment in which humans could not cheat or detect repeated cards proved to be quite difficult.
The first bot, *Poki-1*, used the formula based betting strategy, and the new reweighting method. *Poki-2* also used the formula based betting strategy but used the old reweighting system instead. *Poki-3* was our best simulation based betting strategy (at the time). It used the meta-predictor for simulations and also used the newer reweighting system. The overall scores (in small bets won or lost) are shown in Table 5.3. A summary of each bot’s performance is shown in Table 5.4. Average winning rates are given for both the total 500 hands and for the last 350 hands.
The results are somewhat surprising, but any conclusions must be taken lightly due to the small sample size and potentially high variance in the cards dealt in each room. The FBS bot with the old reweighting did far better than the other two bots. In fact, it was the only bot to finish with a positive score. There are several
Table 5.3: Poki Poker heads-up contest results (January 2002). The results are for all 500 hands played per room.
| Rank | Poki-1 | Poki-2 | Poki-3 | Total |
|------|--------|--------|--------|-------|
| 1 | 279.0 | 102.5 | 321.5 | 703.0 |
| 2 | 265.0 | -42.5 | 302.0 | 524.5 |
| 3 | 196.0 | -22.5 | 193.5 | 367.0 |
| 4 | 12.5 | -6.0 | 115.5 | 230.0 |
| 5 | 245.0 | -103.5 | 33.0 | 174.5 |
| 6 | 5.0 | -80.0 | 234.0 | 159.0 |
| 7 | -150.0 | 51.5 | 180.5 | 82.0 |
| 8 | 21.0 | -62.5 | 40.5 | -1.0 |
| 9 | -45.5 | -138.0 | 155.0 | -28.5 |
| 10 | -34.0 | -151.5 | 38.5 | -147.0|
| 11 | -13.0 | -267.5 | 54.5 | -226.0|
Table 5.4: Poki poker heads-up contest results summary.
| Name | Strategy | Reweighting | SB/Hand (500) | SB/Hand (350) |
|------|--------------|-------------|---------------|---------------|
| Poki-1 | Formula (FBS) | New | -0.162 | -0.132 |
| Poki-2 | Formula (FBS) | Old | +0.131 | +0.110 |
| Poki-3 | Simulation (SBS) | New | -0.303 | -0.208 |
possible explanations for this. It is quite possible that the set of 350 cards used for the *Poki-2* games may have been significantly biased against the human players. Another possibility is that the formula based system, written several years ago, has been tuned for the old reweighting system. If this is the case, introducing a newer system, despite being more correct, could have a negative effect on its performance. Of course, this is all conjecture – more detailed experiments should be conducted to investigate these possibilities.
The (then) state-of-the-art simulation-based bot did terribly, relative to the FBS bots. In our opinion, the variance in cards was not the cause of this result, but genuinely poor play. It was far too tight, folding too often and only betting aggressively if it had a very strong hand. These problems will be addressed in the following chapter.
### 5.4 Summary
This chapter evaluated the performance of several program components. Problems with reweighting were investigated and a newer method was shown to improve performance. Numerous opponent modeling techniques were evaluated. The *metapredictor* successfully encapsulates the other methods and empirically outperforms the other methods in isolation. Neural networks are shown to be the most successful
of the individual prediction methods. Finally the performance of our programs in online games is shown. Both FBS and SBS players prove to be successful against humans, however, the relative performance of FBS vs SBS is inconclusive. It was expected that the SBS would be far superior to the FBS system, but to date, this has not been demonstrated. The following chapter will discuss this in detail.
Chapter 6
Simulation Based Betting Strategies Revisited
There are many methods for predicting the future. For example, you can read horoscopes, tea leaves, tarot cards, or crystal balls. Collectively, these methods are known as ‘nutty methods.’ Or you can put well-researched facts into sophisticated computer models, more commonly referred to as ‘a complete waste of time.’
– Scott Adams, ‘The Dilbert Future’
There are numerous problems and difficulties with Poki’s simulation based betting strategy. The results clearly indicate that it does not outperform the simpler formula-based betting strategy. While the simulation based strategy (SBS) does outperform the formula-based strategy (FBS) in certain situations, there are many occasions where it does far worse. The FBS is much more robust, whereas the SBS can vary widely in performance for numerous subtle and not so subtle reasons. The following section will discuss some of these problems in detail.
6.1 Problems with Simulations
Some strange things are observed when watching the SBS programs playing. One of the problems exhibited by SBS in full ten-player games is over-aggressiveness. SBS players will check-raise and re-raise with moderate strength hands. In a two-player game, the opposite problem occurs; SBS becomes extremely pessimistic and folds far too often. Even when evaluating strong hands, the EV’s returned from the simulations will often be negative. Situations where bluffing is clearly profitable are not discovered, despite the fact that they should be. There are several reasons for this behavior.
6.1.1 Negative Reinforcement Through Self-Prediction
A major problem in simulations is caused by choosing our own future actions within a simulation. This can result in driving the simulations towards a self-fulfilling prophecy of playing badly. For instance, if we predict ourselves folding within a simulation, this decreases the simulated EV. If we predict ourselves folding too often, the simulated EVs will drop below zero (despite the fact that they often should be positive). With negative EV’s Poki will fold. Since Poki then uses this action to update its own model of itself, this increases the chance that it will predict itself folding in future simulations. This drives the simulations into a downward spiral of predicting future folds, which in turn causes the probability of actually folding to increase.
In a ten-player game, the SBS is too aggressive. It again, falls into a self-fulfilling prophecy of raising in the future, and the opponents calling, more often than reflects reality. With all of the extra bets going into the simulated pot, the pot grows so large that even a moderate-strength hand has the pot-odds to call or raise. The EV’s come back unnaturally high in this case.
6.1.2 Biased Weight Tables
Another problem is that the frequency of the simulated actions do not reflect the actual frequencies observed. Hands are chosen from the opponent’s weight table which reflects hands that are more consistent with their play – but the weight table does not actually reflect the true probability distribution of cards an opponent holds (ideally it would, but our current approach to reweighting is far away from being this accurate). Instead, the weight table is biased to favour the stronger set of hands that match the board cards. Hands chosen from the weight table distribution are biased towards better hands. Since the simulated actions of opponents are predicted based on the strength of their hand, the actions themselves will be biased to reflect these better cards.
Take for instance, the following situation. We are heads-up and we are first to act on the river. The pot has 6 big bets in it and we have missed our draw. Our model tells us the opponent will fold 20% of the time to a bet on the river. That means that if we bet, we will immediately win 6 big bets 20% of the time, for \((0.2 \times 6) = 1.2\) big bets. If we are called (and we lose), or we are raised (we will fold our hand to a raise) we will lose 1 big bet. This will happen 80% of the time, so we will lose \((0.8 \times 1) = 0.8\) big bets. This gives a net EV of \((1.2 - 0.8) = +0.4\) big bets for betting on the river. Despite having a weak hand, it is profitable to bet.
In a simulation, we may bias the cards given to the opponent enough that we simulate them folding only 10% of the time, instead of 20% of the time. We fail to give them weak enough cards to fold 20% of the time. This changes the odds we are
getting from the pot to -0.3 big bets for betting, which is clearly incorrect. If we had a moderately strong hand, despite missing the draw (for instance, a small pair), the biased opponent hands would also win the showdown more often than they should. This all compounds to cause simulated EV’s to be far too low.
6.1.3 Patches to Fix Simulations
There are numerous other ways that small biases in the simulation can cause the EVs generated to be problematic. The problems all compound to make the current SBS a very volatile system. There are cases where it performs well, but far more where it swings wildly in the wrong direction. There are many kludgey ways to patch the specific problems. Allowing ourselves to fold or raise too often in the simulation can be artificially dampened. Noise can be added to the weight tables to reduce the bias towards strong hands. However, with these patches, one does not know the right amount of noise to add to the simulations in any given situation. Furthermore, noise merely dampens the bias, but does not correct it. An ideal solution would naturally eliminate bias in order to create a more robust simulation.
6.2 A New Approach – ‘Miximax’ Search
A new method of game-tree search w/ selective sampling was developed to explore more robust methods of computing the expected values of our actions. For simplicity, only the two player game will be discussed.
This new method performs a full search to the leaf nodes of the imperfect-information game tree. At the leaf nodes where show-downs occur, cards for the opponent are chosen that are consistent with the path to the leaf. The EVs from the leaf nodes are propagated back to the decision point, and weighted by the estimated probability of each branch point. We call this new search Miximax. Like minimax search in perfect information domains, we choose the maximal action for ourselves. But instead of choosing the minimal action for the opponent (or ‘min’ player) we choose the weighted sum of their actions, where the weights are taken from the estimated probability distribution over their actions.
The old simulation method chooses the cards first and then generates the actions. The miximax method generates the actions, and then estimates the unknown cards last. This mirrors the actual viewpoint of a poker player in a poker game. The players act in turn, and only until the end does a showdown reveal the cards of the opponent. The majority of hidden-information is placed at the leaves of the search. The assumption is that the action information is more robust than the card information.
For instance, imagine we are on the river with a strong hand, and we are contemplating our next action. Our model tells us that if we check the opponent will
Figure 6.1: A game tree with opponent modeling and hand evaluation for a situation on the river.
check 20% of the time and bet 80% of the time. If we bet, the opponent will fold 20%, call 60%, and raise 20% of the time. We can construct the full betting tree, weighted by our knowledge of the opponent. Figure 6.1 shows the game tree from our decision point and the historical frequencies for the opponent’s actions. Circular nodes show our actions, and square nodes represent the opponent’s actions. Probabilities for opponent actions are given in each node. The leaves are EV’s representing an estimate of the net expected value earned between our original decision point and the leaf node. Since we have a very strong hand in this example, the EV’s are large positives. Note that in the betting sequences where our opponent raises, our estimate is less optimistic, since we are more likely to be up against a stronger hand in this case. If we were to recompute the EV’s with a weaker hand, the EV’s at the leaves would drop substantially, since the probability of winning a showdown would decrease.
By backing up the values at the leaves, we can get good estimates of the EV of the actions available to us. For instance, in Figure 6.2 we can compute that the EV of checking is equal to 0.2 times the EV of the subtree for the situation where we check, and the opponent checks, plus 0.8 times the EV for the subtree for the situation in which we check and the opponent bets.
To summarize this process:
- To compute the EV for one of our actions we take the weighted sum of its children’s EV’s, where the weights are the probability of each child’s occurrence. Let $Pr(A_i)$ be the estimated probability that the opponent will choose child move $i$ of node $A$, and let $n$ be the number of children. The EV of our action $A$ becomes:
$$EV(A) = \sum_{1 \leq i \leq n} Pr(A_i) \times EV(A_i)$$ \hspace{1cm} (6.1)
- To compute the EV of an opponent’s action we take the maximum EV of the children. The maximum is used for simplicity, but this can be modified to a more accurate function (\textit{i.e.} a mixed strategy). We assume we would always choose the action with the highest EV.
$$EV(B) = \max(EV(B_{fold}), EV(B_{call}), EV(B_{raise}))$$ \hspace{1cm} (6.2)
Or in general, for $n$ choices:
$$EV(B) = \max(EV(B_1), EV(B_2), \ldots, EV(B_n))$$ \hspace{1cm} (6.3)
• To compute the EV of a chance board card (for the turn or river) the subtree below the chance card is expanded multiple times with each of the different possible cards. Since these outcomes occur uniformly at random, the average EV from the expansions is used as the final EV. Let $(\star)$ be a chance node for a board card and $n$ be the number of possible cards left in the deck.
$$EV(\star) = \frac{\sum_{i=1}^{n} EV(\star_i)}{n} \quad (6.4)$$
• The EV at a leaf node is the probability of winning the pot times the size of the pot, minus the cost of reaching the leaf node. If the leaf node is due to a fold then the probability of winning is 1.0 if the opponent folded, and 0.0 if we folded. If the leaf is a showdown, the probability of winning is estimated as best as possible. Methods of estimating this are discussed later. Let $L$ be a leaf node, $P_{win}$ the probability of winning the pot, $L_{\$pot}$ be the size of the pot, and $L_{\$cost}$ be the cost of reaching the leaf node.
$$EV(L) = (P_{win} \times L_{\$pot}) - L_{\$cost} \quad (6.5)$$
### 6.2.1 Miximax and Miximix
Although the algorithm described in the previous section was developed as an improved method for doing poker simulations, it is in fact a fully autonomous algorithm for directly computing an estimate of the value of an imperfect game tree.
The *Miximax* strategy computes the maximal EV’s of our actions. However, suppose we are presented with two positive EV’s for calling and raising. Typically, one would want to choose the action with the highest EV. However, in poker, we do not necessarily always want to do this. Consistently choosing the maximum may lead to play that is too exploitable. We sometimes want to pick an action with a smaller (but non-negative) EV.
The *Miximax* algorithm predicts a probability distribution for the opponent’s actions, and considers the maximum EV made at each of our decision points. The *Miximix* algorithm is more general, considering mixed strategies for both players.
Let us define a function $\Phi$ that takes the EV of each action and outputs a randomized mixed strategy over these actions (*i.e.* a probability triple). Instead of using the max function as our strategy within the search, we will instead compute the EV as the weighted sum of our child EV’s. Let V be a vector containing the EV of each of our actions: $V = [EV(B_1), EV(B_2), \ldots, EV(B_n)]$
Then the EV can be computed as a dot product of the mixed strategy and the EV’s:
Essentially this models our own mixed strategy within the search itself. Instead of a *Miximax* search, we now have a *Miximix* search. However, defining a reasonable function $\Phi$ is still an open problem.
### 6.2.2 Related Work
*Miximax* has some high-level similarities to perfect-information game search algorithms such as *-Min* [36] and *ProbiMax* [23, 22]. Both Jansen’s *ProbiMax* and Reibman and Ballard’s *-Min* algorithm change the min nodes in minimax into chance nodes. Instead of returning the minimum successor of a min node, these algorithms return the weighted sum of the min node’s children. In perfect-information games such as Chess, the goal of these predictive searches is to model a fallible opponent within the search – the min nodes become chance nodes that model the opponent choosing sub-optimal moves with a certain probability. This type of search is especially useful when a program is playing a losing position – if the opponent plays optimally, there is nothing that can be done. However, if the opponent is fallible they may make a mistake that can lead to a draw or even a win for the program. Using a predictive opponent model search, the program can select optimal (or even slightly sub-optimal) moves which lead to a position that increases the probability that the opponent will make a mistake.
Much of the similarity of *ProbiMax* to our method is superficial. The high level structure of the algorithm is quite similar, however the application to an imperfect information domain changes many things. The problem being solved, the ‘purpose’ of the search, is quite different. In Poker, minimax is not possible because we cannot compute the values for an opponent’s choice without first knowing their cards. We are not modeling the objective errors of the opponent, but instead we are modeling the hidden information possessed by our opponent.
It is possible to have a good model of the probability with which an opponent will choose each action. Essentially we can model their mixed strategy within our search. Another big difference is seen in the game-tree – in poker there are also real chance nodes (for future board cards) that must be searched. Finally, at the leaf nodes of our tree we must estimate our probability of winning, whereas in perfect information games, there is an objective, correct value to assign a leaf. The estimation at the leaf nodes is loosely equivalent to an evaluation function in a classic search.
### 6.2.3 Estimating The Probability of Winning at Leaf Nodes
At a leaf node the probability that the pot is won must be determined. There are many ways this could be accomplished. Ideally the weighted set of possible hands
that the opponent might take to the showdown, at a particular leaf node, could be used to compute our Hand Strength. In practice, getting an accurate set of weighted hands is difficult.
Ideally, for each leaf node encountered, the hands shown by the opponent could be recorded. The hand’s strength and potential at each betting decision could also be computed and stored. When revisiting the leaf node in a simulation, we could then choose hands that match the same distribution of strength and potential through the hand. This, of course, is a prohibitively expensive computation on current hardware. For now, an approximation is desired.
Currently, the *Hand Rank* (HR) of the opponent’s hand on the river is stored. When estimating our probability of beating the opponent, we compute our hand’s hand rank as well. Recall that a hand rank is the proportion of opponent hands we can beat or tie. A hand with a lower hand rank must be beaten by a hand with a higher rank. Our probability of winning is calculated as the historical percentage of opponent hand ranks that are lower than our current hand rank.
To see, intuitively, why this works, consider the following example. Suppose we are at a leaf node with a betting sequence of $Cc/Kk/Bc/Bc$,\(^1\) and a 0.8 HR. Historically, with this betting sequence the opponent has shown us a hand rank distribution where 90% of the hands were less than our hand rank of 0.8. Thus we can assume that with a 0.8 HR, we have a 90% probability of winning the pot at this leaf. However, when examining the leaf node prefixed by $Cc/KbC/BrC/BrRrC$, only 20% of the hands the opponent has shown have a HR less than 0.8. In the latter sequence the opponent is betting and raising – representing a much stronger hand than in the first sequence. In this case we only estimate our chance of winning the pot with a 0.8 hand rank at 20%. Given enough observations, this turns out to be a decent measure of our actual chances.
### 6.2.4 Pruning Large Trees
The game tree for Texas Hold’em, particularly during the pre-flop and flop, can be too large to search in real-time. There are several ways to cut down the amount of search needed. The primary explosion in tree size is due to the number of cards that must be dealt on the turn and river. It is possible to get a reasonable estimate of EV without expanding every single unseen card. As long as a smaller, representative sample of turn and river cards are expanded, the EVs should generally be close to that which is generated from the full expansion.
A second consideration is to prune away highly improbable branches of the tree that have a negligible effect on the overall EV of the parent. For instance, if the
\(^1\)This string represents a betting sequence. Capital letters are our actions, while lowercase are the opponent’s actions. ‘/’ = the start of a new betting round. ‘K’ = check, ‘C’ = call, ‘B’ = bet, and ‘R’ = raise.
compounded probability of reaching a leaf-node is so small that its contribution to the final EV is miniscule compared to other leaf nodes, then it can be pruned without fear of radically changing the EVs. There are many common sequences in poker, and many sequences that are extremely rare (such as call-raising, or raise-folding) and will not significantly effect the outcome.
This pruning method will be necessary when using this method on games of more than two players. With added players, the size of the game tree becomes exceptionally large. In this case, it will be intractable to search the full tree in real time.
6.2.5 Current Implementation w/ Context-Trees
This new method is only in the beginning stages of research. Only a preliminary implementation has been done to verify its validity. Currently, the implementation can only play heads-up games. The search is also tied to a pure form of opponent modeling that is extremely slow at learning.
All of the previous different prediction methods attempt to consolidate and generalize from correlations in the data. An alternative is to store and use the data in its complete granularity, without any generalization. Every observed instance is stored with the complete betting sequence as the context of the action. A prefix-tree or trie can be used to store and group the observed betting sequences. This is a tree where the nodes are labelled by characters and a path from the root to a leaf is concatenated to form a string. If we find our current state in this prefix-tree (or as we call it in this case, a Context-Tree) the frequency of the opponent’s actions (in the past) can be measured exactly. All public context information is implicitly represented in the Context-Tree.
Currently, no generalizations are made for each observed instance. One observed hand will add only one new instance to the tree. Given sufficient observations, the tree becomes a very accurate model – assuming the opponent plays a fixed strategy. Unfortunately it can easily take 20000 hands to sufficiently fill the tree.
To improve the learning speed of a Context-Tree, we need to generalize observed hands to other, similar paths through the tree. For instance, an observation of Bc/Bc/BrC/KbC could also be placed into a nearly identical path of Ck/Bc/BrC/KbC, where only the pre-flop actions have changed slightly. However, this problem has been left for future work.
The reason for using a Context-Tree at this stage of development is to have as near a ‘perfect’ model as possible. Generalizing data is an error prone process. Given an error-free model (i.e. the model makes predictions based only on perfect matches from past observations) allows us to test the Minimax search under ideal conditions.
6.2.6 Results
Our miximax-based betting strategy was tested against three different programs: FBS-Poki, SBS-Poki, and a recently developed heads-up player called *ArtBot* which was written by Arthur Choung.
*ArtBot* is a very loose and fairly passive heads-up player. This style of play makes it difficult to read hands. It also plays quite defensively, which makes it difficult to win large pots. Against FBS-Poki (who is currently a fairly weak heads-up player), ArtBot does very well – winning at a rate of around +0.35 sb/hand.
There is a wide difference in playing styles between FBS-Poki, SBS-Poki, and ArtBot – a good thing to have when testing a new strategy. The programs were pitted against each other for at least 15000 hands. The Miximax player was either started with no tree at all, or a default tree built from playing a different player.\(^2\)
Technically, the Miximax player tested was actually a Miximix player – but the ad hoc mixing function used was very close to a max function. If the EV’s are positive and close (within one small bet of each other), the maximum EV action is chosen 70% of the time, the other 30%. If the two EV’s differed more than this the maximum is chosen 95% of the time, and the lesser action 5% of the time. This was done to encourage some exploration – if the maximum action was always chosen, the other alternatives might never be tested.
Miximax’s results against these three opponents are shown in Figure 6.3. Typically during the first 5000-10000 hands, Miximax doesn’t do so well as it is still lacking a sufficient model of the opponent. However, once the model is sufficient, Miximax is very successful. After learning, the winning rates against FBS and SBS strategies are as high as +0.4 to +0.5 sb/hand. Against ArtBot, Miximax also wins at a rate that fluctuates between +0.1 and +0.2 sb/hand. ArtBot changes its style of play fast enough that the weaknesses of Context-Trees are exposed.
6.2.7 Future Work
As mentioned earlier, this is a recent idea and only preliminary research has been done. There are many ideas for future development. First, the modeling must be made to learn faster. Either the Context-Tree idea must be improved to learn faster, or the past opponent models must be used instead. Another important test would be to extend the Miximax method to handle multi-player games. The size of the game tree grows exponentially with added players. Scalability of the Miximax search might be hard to achieve without substantial pruning. Finally, the method must be tested on more than just other programmed strategies. The acid test will be measuring its performance against strong human players.
\(^2\)The FBS-Poki match was started with an empty tree. The other two matches were started with a tree built from games against FBS-Poki.
Figure 6.3: Miximax v.s. three different opponents.
Chapter 7
Conclusions and Future Work
*Life consists not in holding good cards, but in playing well those you do hold.*
— Josh Billings (1818-1885)
The game of poker has been a challenging domain to apply artificial intelligence techniques. It contains a variety of difficult problems in machine learning, search, simulation, and game theory. Solving these problems is a difficult but worthwhile endeavor. The real-world is full of situations that contain both non-determinism and imperfect-information properties. This research may provide a better understanding of how to model an uncertain environment, and how to use that model to make effective decisions.
The *Poki* project has developed over the last five years from its beginning as a C++ program called *Loki* into its current incarnation as a large and complex Java program. *Poki* has shown incremental improvements in performance over past versions of *Loki*. It is a winning player in online poker games against human players of average strength.
### 7.1 Thesis Contributions
This thesis has contributed many new enhancements to the *Poki* system. A great deal of analysis has been done to discover flaws and bugs within the older programs. As well, numerous opponent modeling enhancements were made. To summarize the major accomplishments:
- *Poki* is a complete object-oriented rewrite of *Loki*.
- *Loki* had a very simplistic 12-bin action frequency collection system for opponent modeling. This was shown to have low accuracy [13].
• The task of opponent modeling was recast as a machine learning classification/prediction problem.
• Neural networks were introduced to improve the opponent modeling abilities of Poki. The neural networks were successful at finding useful patterns in the noisy poker data.
• A Meta-Prediction system was added to encapsulate any number of prediction methods.
• A way to graphically plot the weight tables improved our ability to visualize the consequences of different reweighting methods.
• The old method of reweighting was shown to be based on the faulty assumption that EHS is uniformly distributed. A new method was designed to address this problem, and was shown to improve performance.
• Analysis of the simulation methods showed that the expected values (EV’s) generated are unreliable for numerous reasons.
• Miximax and Miximix, a new, more robust method of performing a search on an imperfect-information game-tree was introduced and evaluated.
7.2 Future Work
Although Poki is consistently a winning player in online games against humans and other programs, Poki is still far from being a world-class player. There are numerous possibilities for future work:
• The opponent modeling is an ongoing challenge. There is likely to be much room for improvement. One of the biggest problems is that current machine learning methods of modeling are slow to learn. None of the methods are capable of intuitive generalization after only a few observations.
• A good understanding of game-theoretic optimal mixed strategies is needed. With knowledge of the proper optimal or near-optimal betting frequencies, Poki’s strength could grow by leaps and bounds. Currently, our randomization is extremely ad hoc.
• Research into the power of Miximax and Miximix search is only just beginning. Much more research is needed in this area.
• While weight tables have been a successful technique, the implementation needs to be rethought. There are currently many flaws and biases in our weight tables – a much more sophisticated implementation is needed.
• Context-Trees are an interesting new method of modeling that stores opponent model data in a pure unadulterated form. Generalization from the data is done at run-time.
God does not play dice with the universe; He plays an ineffable game of his own devising, which might be compared, from the perspective of any of the other players [i.e., everybody], to being involved in an obscure and complex version of poker in a pitch-dark room, with blank cards, for infinite stakes, with a Dealer who won’t tell you the rules, and who smiles all the time.
– Terry Pratchett and Neil Gaiman, *Good Omens*, 1991
Appendix A
Miscellaneous
A.1 Hardware for Experiments
All reported experiments were conducted on a computer with dual 933MHz Intel processors, and 512mb of RAM. The operating system used was RedHat Linux 6.2. Most of Poki was implemented in Java, and was tested under the JDK 1.4.0 using Sun’s JVM.
A.2 Internet Links
The following links are of sites directly related to this thesis.
The Author’s Home Page:
http://www.cs.ualberta.ca/~davidson/
The University of Alberta GAMES Group:
http://www.cs.ualberta.ca/~games/
The Poker Research Group:
http://www.cs.ualberta.ca/~games/poker/
A protocol for online poker games:
http://www.cs.ualberta.ca/~davidson/poker/opp.html
Source Code for Poki:
http://www.cs.ualberta.ca/~davidson/poker/src/
Appendix B
Backpropagation Example
The following is a supplementary explanation of backpropagation:
Suppose we have the neural network shown in Figure B.1. We are testing the network on the input vector \([1.0, 0.0, 0.77, 0.0]\). In this feed-forward network, evaluation is fairly simple. An input value is weighted by the strength of each connection – so the first input value of 1.0 is connected to the first internal node (left) by a weight of 0.2 (for example), and is connected to the second input neuron by a weight of -0.25. Hence, the value the first internal node sees of the first input feature is \(1.0 \times 0.2 = 0.2\).
A neuron’s output is computed by passing the sum of its inputs through a threshold function – in the case of our output neuron a sigmoid function. Essentially, if the sum of the inputs is above a threshold – the neuron fires. The neuron now acts as a feature detector and can excite or inhibit other neurons below it, eventually effecting the final output value of the network.
Backpropagation is a way of training the network on examples by assessing the blame of each weight in the network. It starts at the output node(s) and works back towards the inputs.
Suppose, in our example, the correct output was 1.0, but our network’s actual output was 0.6. It has made an error of \(1.0 - 0.6 = +0.4\) (the sign tells us which direction the error was in. To determine a share of the error for the output node we compute the derivative of the threshold (or activation) function \(gx\) on the input. If our activation function is sigmoidal: \(g(x) = \frac{1}{1+e^{-x}}\) and the derivative \(g'(x) = g(x) \times (1 - g(x))\).
If the input was 0.8, then the node’s share of the error is \(+0.4 \times [g(0.8) \times (1 - g(0.8))]\) (Equation 4.1).
Now that we know the portion of blame to assign the output node, we can adjust the weights leading into that node (Equation 4.2). A connection weight is adjusted by the blame at the node multiplied by the connection’s input. multiplied by \(\alpha\), the learning rate which controls the magnitude of the adjustment.
There is a slight difference in computing the blame for nodes further up in the network (Equation 4.3). The activation derivative is multiplied by the sum of all blames for the output connections of that node. That is to say, the blame at output nodes is dependent on the blame at internal nodes.
During backpropagation, each weight’s blame is computed and then adjusted accordingly. If a weight is to blame for the error, that weight will be adjusted so that it will contribute slightly to making the error closer to zero.
Appendix C
Sample Confusion Matrices
The following tables are sample confusion matrices for specific opponents for each of the specific modeling methods. The variance from player to player is enormous. Note these are not averages, but single data points intended for supplementary information.
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 12.8 | 8.3 | 0.0 | 21.0%|
| call | 14.5 | 43.8 | 0.0 | 58.3%|
| raise | 4.6 | 16.1 | 0.0 | 20.7%|
| % | 31.9%| 68.1%| 0.0% | 56.51%|
Table C.1: Statistics-Loki2 (Human Opponent)
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 12.5 | 8.6 | 0.0 | 21.0%|
| call | 12.5 | 44.6 | 1.1 | 58.3%|
| raise | 2.9 | 13.8 | 4.0 | 20.7%|
| % | 27.9%| 67.0%| 5.2% | 61.14%|
Table C.2: Statistics-II (Human Opponent)
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 17.6 | 3.4 | 0.1 | 21.0%|
| call | 5.2 | 19.5 | 33.5 | 58.3%|
| raise | 0.4 | 2.8 | 17.5 | 20.7%|
| % | 23.2%| 25.7%| 51.1% | 54.5% |
Table C.3: Expert Formula (Human Opponent)
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 15.7 | 5.3 | 0.0 | 21.0%|
| call | 3.8 | 54.5 | 0.0 | 58.3%|
| raise | 0.1 | 20.6 | 0.0 | 20.7%|
| % | 19.6%| 80.4%| 0.0% | 70.22%|
Table C.4: Neural Network (Human Opponent)
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 14.8 | 6.2 | 0.0 | 21.0%|
| call | 3.8 | 54.1 | 0.4 | 58.3%|
| raise | 0.2 | 18.6 | 1.9 | 20.7%|
| % | 18.8%| 78.9%| 2.4% | 70.83%|
Table C.5: Meta-Predictor (Human Opponent)
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 2.6 | 11.3 | 0.0 | 13.8%|
| call | 1.8 | 26.8 | 8.8 | 37.3%|
| raise | 0.3 | 10.5 | 38.2 | 48.9%|
| % | 4.6% | 48.4%| 47.0% | 67.5%|
Table C.6: Statistics-Loki2 (FBS-Poki)
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 3.8 | 9.8 | 0.3 | 13.8%|
| call | 1.5 | 26.8 | 9.0 | 37.3%|
| raise | 0.9 | 10.1 | 38.0 | 48.9%|
| % | 6.1% | 46.7%| 47.3% | 68.55%|
Table C.7: Statistics-II (FBS-Poki)
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 8.6 | 4.9 | 0.4 | 13.8%|
| call | 2.0 | 17.7 | 17.6 | 37.3%|
| raise | 0.3 | 1.9 | 46.8 | 48.9%|
| % | 10.9%| 24.4%| 64.8% | 73.1%|
Table C.8: Expert Formula (FBS-Poki)
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 7.4 | 2.3 | 4.2 | 13.8%|
| call | 1.6 | 13.0 | 22.8 | 37.3%|
| raise | 0.3 | 13.0 | 35.7 | 48.9%|
| % | 9.2% | 28.2%| 62.6% | 56.0%|
Table C.9: Neural Network (FBS-Poki)
| Actual | Prediction |
|--------|------------|
| | fold | call | raise | % |
| fold | 8.6 | 5.0 | 0.2 | 13.8%|
| call | 2.0 | 22.3 | 13.1 | 37.3%|
| raise | 0.3 | 4.7 | 43.9 | 48.9%|
| % | 11.0%| 31.9%| 57.2% | 74.8%|
Table C.10: Meta-Predictor (FBS-Poki)
Appendix D
Glossary of Poker Terms
This appendix contains definitions of common poker terms used in this thesis.\footnote{Adapted from [6].} More extensive poker glossaries are available on the world wide web, such as \url{http://www.kimberg.com/poker/dictionary.html}, or \url{http://conjelco.com/pokglossary.html}.
- **Aggressive Player** A player who is characterised by frequently betting and raising.
- **All-in** To have one’s entire stake committed to the current pot. Action continues toward a side pot, with the all-in player being eligible to win only the main pot.
- **All-in Equity** The expected income if the current hand was permitted to go to the showdown with no further betting.
- **Ante** A mandatory fee placed into the pot from each player at the start of a game (in some types of Poker).
- **Bet** To make the first wager of a betting round (compare raise).
- **Big Bet** The largest bet size in limit poker. $20 in $10-$20 Hold’em.
- **Big Blind** A forced bet made before the deal of the cards. $10 in $10-$20 Hold’em, posted by the second player to the left of the button.
- **Blind** A forced bet made before the deal of the cards (see small blind and big blind).
- **Bluff** To bet with the expectation of losing if called.
- **Board** The community cards shared by all players.
• **Button** The last player to act in each betting round in Texas Hold’em.
• **Call** To match the current level of betting. If the current level of betting is zero, the term **check** is preferred.
• **Check** To decline to make the first wager of a betting round (compare **call**).
• **Check-Raise** To **check** on the first action and then **raise** in the same betting round after someone else has **bet**.
• **Community Cards** The public cards shared by all players.
• **Connectors** Two cards differing by one in rank, such as 7-6. More likely to make a straight than other combinations.
• **Draw** A hand with good potential to make a strong hand, such as a *straight draw* or a *flush draw* (compare **made hand**).
• **Dry** Lacking possible draws or betting action, as in a *dry board* or a *dry game*.
• **Flop** The first three **community cards** dealt in Hold’em, followed by the second betting round (compare **board**).
• **Fold** To discard a hand instead of matching the outstanding **bet**, thereby losing any chance of winning the pot.
• **Free-Card Raise** To **raise** on the **flop** intending to **check** on the **turn**.
• **Hand** (a) A player’s private cards (two **hole cards** in Hold’em). (b) One complete game, from the dealing of the cards to the **showdown** (or until one player wins uncontested).
• **Hole Card** A private card in Hold’em.
• **Implied Odds** The **pot odds** based on the probable future size of the pot instead of the current size of the pot.
• **Income rate** The expected amount a hand will win.
• **Kicker** A side card, often deciding the winner when two hands are otherwise tied (*eg.* a player holding Q-J when the **board** is Q-7-4 has **top pair** with a Jack kicker).
• **Loose Game** A game having several **loose players**.
• **Loose Player** A player who does not fold often (*eg.* one who plays most hands at least to the **flop** in Hold’em).
• **Made hand** A hand with a good chance of currently being the best, such as top pair on the **flop** in Hold’em (compare **draw**).
• **Mixed strategy** Handling a particular situation in more than one way, such as to sometimes **call**, and sometimes **raise**.
• **Near-optimal** A good approximation of a game-theoretic optimal solution.
• **Offsuit** Two cards of different suits (compare **suited**).
• **Open-Ended Draw** A **draw** to a straight with eight cards to make the straight, such as 6-5 with a **board** of Q-7-4 in Hold’em.
• **Outs** Cards that will improve a hand to a probable winner (compare **draw**).
• **Passive Player** A player who is characterised by frequently checking and calling his hand rather than betting or raising.
• **Pocket Pair** Two cards of the same rank, such as 6-6. More likely to make three of a kind than other combinations (see **set**).
• **Pot Odds** The ratio of the size of the pot to the size of the outstanding bet, used to determine if a **draw** will have a positive expected value.
• **Pre-flop** In Hold’em, the first betting round after the deal of the cards and before the **flop**.
• **Raise** To increase the current level of betting. If the current level of betting is zero, the term **bet** is preferred.
• **River** The fifth community card dealt in Hold’em, followed by the fourth (and final) betting round.
• **Second pair** Matching the second highest community card in Hold’em, such as having 7-6 with a **board** of Q-7-4.
• **Set** Three of a kind, formed with a **pocket pair** and one card of matching rank on the **Board**. A powerful well-disguised hand (compare **trips**).
• **Showdown** The revealing of cards at the end of a **hand** to determine the winner and award the pot.
• **Side-pot** A second pot for the bets made by active players after another player is all-in.
• **Slow-play** To **call** with a strong hand, and then **raise** in a later betting round, for purposes of deception.
• **Small Bet** The smallest bet size in limit poker. $10 in $10-$20 Hold’em.
• **Small Blind** A forced bet made before the deal of the cards. $5 in $10-$20 Hold’em, posted by the first player to the left of the button.
• **Suited** Two cards of the same suit, such as both Hearts. More likely to make a flush than other combinations (compare offsuit).
• **Suited-Connector** A hand that is both Suited and a Connector (i.e. Q♥-J♥).
• **Table Image** The general perception other players have of one’s play.
• **Table Stakes** A poker rule allowing a player who cannot match the outstanding bet to go all-in with his remaining money, and proceed to the showdown (also see side pot).
• **Texture of the Board** Classification of the type of board, such as having lots of high cards, or not having many draws (see dry).
• **Tight Player** A player who usually folds unless the situation is clearly profitable (*eg.* one who folds most hands before the flop in Hold’em).
• **Top pair** Matching the highest community card in Hold’em, such as having Q-J with a board of Q-7-4.
• **Trips** Three of a kind, formed with one hole card and two cards of matching rank on the board. A strong hand, but not well-disguised (compare set).
• **Turn** The fourth community card dealt in Hold’em, followed by the third betting round.
• **Under-the-gun** The position of the player who is first to act in the betting round.
• **Wild Game** A game with a lot of raising and re-raising. Also called an action game.
[1] R. Axelrod. *The Evolution of Cooperation*. Basic Books, Inc., 1984.
[2] H. Berliner, G. Goetsch, and M. Campbell. Measuring the performance potential of chess programs. *Artificial Intelligence*, 43(1):7–20, 1990.
[3] D. Billings. Computer poker. Master’s thesis, University of Alberta, 1995.
[4] D. Billings. The first international RoShamBo programming competition. *International Computer Games Association Journal*, 23(1):42–50, 2000.
[5] D. Billings. Thoughts on RoShamBo. *International Computer Games Association Journal*, 23(1):3–8, 2000.
[6] D. Billings, A. Davidson, J. Schaeffer, and D. Szafron. The challenge of poker. *Artificial Intelligence*, 134(1-2):201–240, 2002.
[7] D. Billings, D. Papp, L. Pena, J. Schaeffer, and D. Szafron. Using selective-sampling simulations in poker. In *AAAI Spring Symposium on Search Techniques for Problem Solving under Uncertainty and Incomplete Information*, pages 13–18, 1999.
[8] D. Billings, L. Pena, J. Schaeffer, and D. Szafron. Using probabilistic knowledge and simulation to play poker. In *AAAI National Conference*, pages 697–703, 1999.
[9] M. Buro. The Othello match of the year: Takeshi Murakami vs. Logistello. *International Computer Chess Association Journal*, 20(3):189–193, 1997.
[10] M. Campbell, A. J. Hoane, and F h. Hsu. Deep Blue. *Artificial Intelligence*, 134(1-2):57–83, 2002.
[11] D. Carmel and S. Markovitch. Incorporating opponent models into adversary search. In *AAAI National Conference*, pages 120–125, 1995.
[12] F.A. Dahl. A reinforcement learning algorithm applied to simplified two-player Texas Hold’em poker. In *12th European Conference on Machine Learning (ECML’01)*, pages 85–96, 2001.
[13] A. Davidson, D. Billings, J. Schaeffer, and D. Szafron. Improved opponent modeling in poker. In *International Conference on Artificial Intelligence (ICAI’2000)*, pages 1467–1473, 2000.
[14] H. Dreyfus. *What Computers Still Can’t Do: A Critique of Artificial Reason*. MIT Press, 1993.
[15] D. Egnor. Iocaine powder. *International Computer Games Association Journal*, 23(1):33–35, 2000.
[16] S. G. Ficici and J. B. Pollack. A game-theoretic approach to the simple coevolutionary algorithm. In *Parallel Problem Solving from Nature - PPSN VI 6th International Conference*, Paris, France, 16-20 2000. Springer Verlag.
[17] S. G. Ficici and J. B. Pollack. Pareto optimality in coevolutionary learning. In *European Conference on Artificial Life*, pages 316–325, 2001.
[18] N. Findler. Studies in machine cognition using the game of poker. *Communications of the ACM*, 20(4):230–245, 1977.
[19] R. Fung and K. Chang. Weighting and integrating evidence for stochastic simulation in bayesian networks. In *Uncertainty in Artificial Intelligence*, pages 209–220. Morgan Kaufmann, 1989.
[20] M. Ginsberg. GIB: Steps toward an expert-level bridge-playing program. In *International Joint Conference on Artificial Intelligence*, pages 584–589, 1999.
[21] H. Iida, J. Uiterwijk, J. van den Herik, and I. Herschberg. *Thoughts on the Application of Opponent-Model Search*, pages 61–78. Univ. of Limburg, Maastricht, 1994.
[22] P. Jansen. *Using Knowledge About the Opponent in Game-Tree Search*. PhD thesis, Carnegie-Mellon University, 1992.
[23] P. Jansen. KQKR: Speculatively thwarting a human opponent. *International Computer Chess Association Journal*, 16(1):3–17, 1993.
[24] D. Koller, N. Megiddo, and B. von Stengel. Fast algorithms for finding randomized strategies in game trees. In *26th Annual ACM Symposium on the Theory of Computing*, pages 750–759, 1994.
[25] D. Koller and A. Pfeffer. Representations and solutions for game-theoretic problems. *Artificial Intelligence*, 94(1):167–215, 1997.
[26] K. Korb and A. Nicholson. Bayesian poker. In *Uncertainty in Artificial Intelligence*, pages 343–350, 1999.
[27] H. W. Kuhn. A simplified two-person poker. *Contributions to the Theory of Games*, 1:97–103, 1950.
[28] T. A. Marsland and J. Schaeffer. *Computers, Chess, and Cognition*. Springer-Verlag, 1990.
[29] M. L. Minsky. *Theory of Neural-Analog Reinforcement Systems and its Application to the Brain-Model Problem*. PhD thesis, Princeton University, 1954.
[30] M. L. Minsky and S. Papert. *Perceptrons: An Introduction to Computational Geometry*. MIT Press, Cambridge, Mass., 1969.
[31] J. F. Nash. Non-cooperative games. *Annals of Mathematics*, 54:286–295, 1951.
[32] J. F. Nash and L. S. Shapley. A simple three-person poker game. *Contributions to the Theory of Games*, 1:105–116, 1950.
[33] J. Noble and R. A. Watson. Pareto coevolution: Using performance against coevolved opponents in a game as dimensions for pareto selection. In *Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2001)*, pages 493–500, 2001.
[34] D. Papp. Dealing with imperfect information in poker. Master’s thesis, University of Alberta. 1998.
[35] L. Pena. Probabilities and simulations in poker. Master’s thesis, University of Alberta, 1999.
[36] A. L. Reibman and B. W. Ballard. Non-minimax search strategies for use against fallible opponents. In *AAAI National Conference*, pages 338–342, 1983.
[37] D. E. Rumelhart, G. E. Hinton, and R. J. Williams. Learning representations by back-propagating errors. *Nature*, 323:533–536, 1986.
[38] S. Russell and P. Norvig. *Artificial Intelligence: A Modern Approach*. Prentice Hall, 1995.
[39] A. L. Samuel. Some studies in machine learning using the game of checkers. *IBM Journal of Research and Development*, 3:211–229, 1959.
[40] J. Schaeffer, J. Culberson, N Treloar, B. Knight, P. Lu, and D. Szafron. A world championship caliber checkers program. *Artificial Intelligence*, 53(2-3):273–290, 1992.
[41] A. Selby. Optimal heads-up preflop poker. website, 1999. http://www.archduke.demon.co.uk/simplex/.
[42] R. Shacter and M. Peot. Simulation approaches to general probabilistic inference on belief networks. In *Uncertainty in Artificial Intelligence*, pages 311–318. Morgan Kaufmann, 1989.
[43] B. Sheppard. Computer Scrabble. *Artificial Intelligence*, 134(1-2):241–275, 2002.
[44] J. Shi and M. Littman. Abstraction models for game theoretic poker. In *Computer Games’00*, pages 333–345. Springer-Verlag, 2001.
[45] D. Sklansky and M. Malmuth. *Hold’em Poker for Advanced Players*. Two Plus Two Publishing, 2nd edition, 1994.
[46] S. Smith. Flexible learning of problem solving heuristics through adaptive search. In *IJCAI*, pages 422–425, 1983.
[47] K. Takusagawa. *Nash Equilibrium of Texas Hold’em Poker*. Undergraduate thesis, Stanford University, 2000.
[48] G. Tesauro. Temporal difference learning and TD-Gammon. *Communications of the ACM*, 38(3):58–68, 1995.
[49] G. Tesauro. Programming backgammon using self-teaching neural nets. *Artificial Intelligence*, 134(1-2):181–199, 2002.
[50] A. M. Turing, C. Strachey, M. A. Bates, and B. V. Bowden. Digital computers applied to games. In B. V. Bowden, editor, *Faster Than Thought*, pages 286–310. Pitman, London, 1953.
[51] P. E. Utgoff. Incremental induction of decision trees. *Machine Learning*, 4:161–186, 1989.
[52] J. von Neumann and O. Morgenstern. *The Theory of Games and Economic Behavior*. Princeton University Press, 2nd edition, 1947.
[53] D. Waterman. A generalization learning technique for automating the learning of heuristics. *Artificial Intelligence*. 1:121–170. 1970. |
Vibrational Circular Dichroism Spectroscopy Study of Paroxetine and Femoxetine Precursors
MARIE URBANOVÁ,1 VLADIMÍR SETNIČKA,2 PETR BOUŘ,2,3 HANA NAVRÁTILOVÁ,4 KAREL VOLKA2
1 Department of Physics and Measurements, Institute of Chemical Technology, Technická 5, 166 28 Praha 6, Czech Republic
2 Department of Analytical Chemistry, Institute of Chemical Technology, Technická 5, 166 28 Praha 6, Czech Republic
3 Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic, Flemingovo nám. 2, 166 10 Praha 6, Czech Republic
4 Department of Organic Chemistry, Masaryk University, Kotlářská 2, 611 37 Brno, Czech Republic
Received 8 September 2001; revised 31 October 2001; accepted 11 November 2001
Published online 25 April 2002 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/bip.10104
ABSTRACT: The solution structures of (3R,4S)- and (3S,4R)-4-(4-fluorophenyl)-3-hydroxymethyl-1-methylpiperidine, which are intermediates in the synthesis of the two pharmaceuticals paroxetine and femoxetine, were studied by vibrational circular dichroism (VCD) spectroscopy. In addition, six derivatives with different substituents attached to the C3 atom were prepared and their VCD and absorption spectra discussed with the aid of ab initio simulations. The VCD spectra were found to be sensitive to the geometry changes. In addition, a subtle variation caused by intermolecular aggregation was apparent in the spectra. The VCD technique can be applied for structural analysis of chiral pharmaceuticals in solutions. © 2002 Wiley Periodicals, Inc. Biopolymers (Biospectroscopy) 67: 298–301, 2002
Keywords: vibrational circular dichroism; conformation; paroxetine; femoxetine
INTRODUCTION
Chiral drugs marketed as single enantiomers represent examples where vibrational circular dichroism (VCD)1,2 spectroscopy can be conveniently applied. Ab initio calculations of VCD spectra are no longer restricted to small molecules3,4 because they can be used for the interpretation of experimental results for medium sized systems5,6 as well as biopolymers.7,8 The piperidine derivatives (3S,4R)- and (3R,4S)-4-(4-fluorophenyl)-3-hydroxymethyl-1-methylpiperidine are the key intermediates in the synthesis of the antidepressive drugs paroxetine and femoxetine, respectively. Our previous VCD and NMR study9 confirmed the absolute configuration of the (3R,4S)-enantiomer in solution, indicated its prevalent conformation, and implied that the hydroxyl group mediates intermolecular aggregation in concentrated solutions. In this work we test the influence of hydrogen bond formation on the VCD spectra in detail. We also found it interesting for pharmaceutical purposes to monitor the...
Figure 1. The structures of the (3S,4R)-4-(4-fluorophenyl)-3-hydroxymethyl-1-methylpiperidine derivatives.
Substitutions of the hydroxyl group by other functional groups and to follow possible conformation changes using VCD.
MATERIALS AND METHODS
Samples (Fig. 1) of (3S,4R)-3 (100% ee)\(^9\) and (3R,4S)-3 (96% ee) were kind gifts from Synthon BV, The Netherlands. Compounds 4–8 were prepared as described elsewhere.\(^{10,11}\) The solutions in CCl\(_4\) and CDCl\(_3\) within 0.64–0.064 mol L\(^{-1}\) were used for the VCD measurements. The VCD and absorption spectra were scanned with a resolution of 4 cm\(^{-1}\) using a Bruker FTIR IFS 66/S spectrometer equipped with the VCD/IRRAS module PMA 37 as previously described in detail.\(^{12}\) A demountable cell separated by a 50- or 200-\(\mu\)m Teflon spacer was used.
A simulation of the VCD intensities based on the Gaussian program package\(^{13}\) was performed,\(^{6,7}\) for geometries optimized with the aid of the conformer searching routine implemented in the Spartan program.\(^{14}\) The vibrational frequencies and intensities were calculated at the BPW91/6-31+G*\(^{15}\) density functional theory\(^{15}\) level at the harmonic approximation using the MFP/GIAO theory\(^{16}\) for VCD. The spectra were simulated using Lorentzian profiles with a 5 cm\(^{-1}\) bandwidth. Normal mode assignment is based on a visual inspection of the dynamic displacements.
RESULTS AND DISCUSSION
Figure 2 shows the experimental VCD spectra of derivatives 3, 4, and 8 defined in Figure 1. All spectra were recorded with a high S/N ratio. Recording of the VCD spectra of (3R,4S)-3 and (3S,4R)-3 reduces the risk of artifacts and confirms the quality of the experimental data: the opposite enantiomers exhibit mirror-image VCD spectra as shown for (3S,4R)-3 and (3R,4S)-3 (Fig. 2). For the sake of brevity, we deal only with the (3R,4S) enantiomers in the following text.
As an example, the agreement between the experiment and calculation is shown for derivative 7 in Figure 3. The numbers indicate the corresponding vibrational modes. Such a comparison of the absorption and VCD spectra (Fig. 2) for all the derivatives reveals the common features summarized in Table I. The mode numbers of compound 3 are used also for the other compounds, because the ordering of the normal modes differs only slightly. Comparing the VCD spectra in Figure 2, we find that the individual derivatives differ in certain regions where the vibration modes involve the chiral centers C3 and C4 and the atoms in their near vicinity (modes 33–39, 47–49, and 58–59).
The VCD measurements require rather high concentrations of the compounds and enable only a narrow interval for concentration variations. In our case we recorded the VCD of 3 in CDCl\(_3\) and CCl\(_4\) solutions with a reasonable S/N ratio in the range of 0.64–0.064 mol L\(^{-1}\) [Fig. 4(B)]. The observed decrease of the absorption of the free OH
Figure 2. The experimental VCD spectra of derivatives 3, 4, and 8 (0.64 mol L\(^{-1}\) in CDCl\(_3\)) and the typical noise spectrum (N).
Figure 3. The (A) VCD and (B) absorption spectra of compound 7. The experimental spectra of the CDCl$_3$ solution ($c = 0.64$ mol L$^{-1}$; spectra a) and the simulated spectra (spectra b).
group at 3627 cm$^{-1}$ and its increase at 3350–3150 cm$^{-1}$ for aggregated OH groups with increasing concentration [Fig. 4(A)] confirmed the hydrogen bond formation. Although the intensity of the free $\nu$(OH) increases about 2 times as the concentration decreases from 0.64 to 0.064 mol L$^{-1}$, only slight changes were observed in the VCD in the mid-IR region (cf. Fig. 4). This is in accord with our detailed analysis, which revealed that only limited parts of the spectral region used are affected by the OH vibrations, for example, the VCD features at 1250–1200 and $\sim$1000 cm$^{-1}$ (the gray areas in Fig. 4), which originate in skeletal deformation coupled to or including the C—O bond. Also, based on the concentration dependence of the VCD spectra, we can conclude that the aggregation does not have a significant influence on the conformation.
CONCLUSIONS
The simulated and experimental spectra are in a very good agreement over the entire region of recorded frequencies. The observed variations in the regions specific to the particular substituents could be identified and reasonably explained using the normal mode assignment. In all of the derivatives studied, the substituents do not significantly influence the spectral response of the piperidine skeleton. The theoretical and experimental results suggest that the phenyl and methyl groups are in equatorial positions and the torsion angle C(substituent)-C3-C4-C(piperidine) is close to 180° (except 8). The concentration dependence of the VCD and absorption reveals strong intermolecular interactions through hy-
Table I. VCD and Absorption Bands Common to All Derivatives
| Mode | VCD and Absorption Frequencies (cm$^{-1}$) | Assignment |
|------|------------------------------------------|------------|
| | Experimental | Calculated | |
| 19–21| 1604, 1510 | 1604, 1496–1498 | Phenyl deformation |
| 27 | 1439–1440 | 1437–1447 | C—H bend (in neighborhood of C3) |
| 30 | 1379–1384 | 1369–1372 | C—H bend on phenyl |
| 41 | 1281–1289 | 1263–1271 | CH$_2$ twist and CH$_3$ deformation |
| 43 | 1221–1225 | 1211–1214 | C—F and phenyl deformation |
| 52 | 1096 | 1092–1097 | No specific skeletal vibration |
| 55 | 1060–1070 | 1052–1060 | Skeletal vibration including C—O |
Figure 4. The concentration dependence of (A) the normalized absorption spectra in the ν(OH) region and (B) the normalized VCD spectra of compound 3 in CDCl₃ at concentrations of 0.64 (spectrum a), 0.128 (spectrum b), and 0.064 mol L⁻¹ (spectrum c) and the noise (N) spectrum for a concentration of 0.128 mol L⁻¹.
Hydrogen bridges with a minor effect on the molecular conformation.
REFERENCES
1. Keiderling, T. A. In Practical Fourier Transform Infrared Spectroscopy; Ferraro, J. R., Krishnan, K., Eds.; Academic: San Diego, CA, 1990; pp 203–284.
2. Nafie, L. A.; Freedman, T. B. In Circular Dichroism: Principles and Applications; Berova, N., Nakanishi, K., Eds.; Wiley: New York, 2000; pp 97–132.
3. Bouř, P.; McCann, J.; Wieser, H. J Phys Chem A 1998, 102, 102–110.
4. Tam, C. N.; Bour, P.; Keiderling, T. A. J Am Chem Soc 1996, 118, 10285–10293.
5. Bour, P.; Navrátilová, H.; Setnička, V.; Urbanová, M.; Volka, K., J Org Chem 2002, 67, 161–168.
6. Setnička, V.; Urbanová, M.; Bour, P.; Král, V.; Volka, K. J Phys Chem A 2001, 105.
7. Bour, P.; Záruba, K.; Urbanová, M.; Setnička, V.; Matejka, P.; Fiedler, Z.; Volka, K. Chirality 2000, 12, 191–198.
8. Silva, R. A. G. D.; Kubelka, J.; Bour, P.; Decatur, S. M.; Keiderling, T. A. Proc Natl Acad Sci USA 2000, 97, 8318–8323.
9. Christensen, J. A.; Squires, R. F. U.S. Pat. 4,067,196, 1977.
10. Navrátilová, H. Ph.D. Thesis, Masaryk University, 2001.
11. Navrátilová, H. Chirality 2001, 13.
12. Urbanová, M.; Setnička, V.; Volka, K. Chirality 2000, 12, 199–203.
13. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98. Revision A.3 and A.7; Gaussian Inc.: Pittsburgh, PA, 1998.
14. Spartan, P. C. PC Spartan Pro 2000, Pro 1.0.5; Wavefunction Inc.: Irvine, CA, 2000.
15. Becke, A. D. J Chem Phys 1993, 98, 5648–5652.
16. Devlin, F. J.; Stephens, P. J.; Cheeseman, J. R.; Frisch, M. J. J Phys Chem A 1997, 101, 9912–9924. |
POSTAL PRINTS
MONEY ORDERS
The largest money order you may purchase is for $100.00. The money order fees are:
- $1 to $10.00 - 25¢
- $10.01 to $50.00 - 35¢
- $50.01 to $100.00 - 40¢
Collecting Payment:
United States currency, military payment certificates (MPC with which you are paid while here in Vietnam) are authorized for postal money orders.
Cashing Money Orders:
Postal money orders issued on or after July 1, 1951, are valid for 20 years from the last day of the month of original issue, (i.e. if you buy a money order this month, it will be good until 31 Dec., 1986). Money orders issued by Navy Post Offices may be cashed at any military post office, a military banking facility, or at any post office within the U.S., its possessions, or territories.
Authorized Persons:
Only THREE persons are authorized to cash a money order. The first of these is the purchaser, whose name appears on the "from line" with his address below it. The purchaser may also include his name on the payee line if he intends to cash the order himself at a later date. The payee is the person whose name appears on the line marked "pay to". He is the person whom the purchaser (you) intends to cash the money order. If the payee wishes to transfer ownership, he may do so by filling in the name of the new owner on the top line of the reverse side of the order marked "pay to", and signing the order himself on the second line marked "payee".
R TIPS R
BANGKOK
CURRENCY: Thai currency called "Baht" is roughly equal to the U.S. nickel. The current rate is $1.00 U.S. equal 20.62 Baht.
HOTELS AND TRANSPORTATION: Upon arrival, your R&R reception center will arrange for hotel accommodations. Generally, the hotel rates are $6.00 to $12.00 single; and $6.00 to $8.00 per person for a double room. Bus services operate in Bangkok, but taxis (5 Baht for each mile; minimum fare is 5 Baht) and motor tricycles called "Sam Laws" (3 Baht for tricycle) are usually favored by visitors.
SIGHTSEEING: Floating Markets - every morning from 7 to 8:30 vendors in sampans congregate on the canal. Anything from fresh fruit to exotic herbs and live chickens to shellfish may be bought from this floating market. Thai Dance Show - every Wednesday and Saturday at 9 a.m. this show is held at the National Museum. Thai dances are also performed at the Cultural Auditorium, Sanam Sunphah. Photographs may be taken at both shows.
SHOPPING: Bargaining is a must at nearly all Bangkok shops; especially jewelry stores. Thai silk may be purchased in three different weights-----lightweights sells for 60-65 Baht ($3-$3.50), and medium and heavyweight up to 100 Baht ($5.00) per yard. The best selections may be seen at The Star of Siam Shops and the Thai House. Other interesting items for the shopper in Bangkok are: Thai cotton, jewelry, bronzeware, Nielloware, Thai celadon (stoneware), lacquerware, Thai Dolls, and wood carvings.
(cont. page 6)
CAPTAIN'S CORNER
With the Christmas Holiday fast approaching, I would encourage you all to visit the Military Affiliate Radio Station (MARS) located over in MCB-9's camp. This is an excellent opportunity to talk with your loved ones during this time of the year. Telegrams as well as telephone calls may be made to any place in the states.
Certain procedures have been established by the battalion concerning the use of the MARS station.
Individuals desiring use of the MARS Station should submit a special request chit to their Company Commanders for approval. The request should include date and time. The approved chit will be presented to the SOG before leaving the area. Upon return from the station, the individual will check back in with the SOG.
But let's not limit ourselves to just phoning, we can also send telegrams. The sender can write his own message of not more than 30 words or use the ARRL or N'AT pre-written radiogram abbreviations which appear on the back of the message forms. These forms can be picked up from the Special Services office.
The desire to keep in contact with those we love back home is one we all have in common. Making use of the MARS station is an outstanding way to keep this contact strong.
XO'S NOTES
This issue I would like to brief you on your gift privileges while serving in Vietnam.
Public Law 89-368 authorizes United States Armed Forces personnel serving in a combat zone to send gifts tax and duty free. In general, the law provides that personnel serving in Vietnam may send as many gift shipments as desired when each parcel does not exceed $50, and provided shipments are not parts of a complete set, such as china, silverware, etc. Shipments over $50 will be subject to duties for the amount in retail value in excess of $50. Gifts must have been purchased in or through authorized agencies of the Armed Forces.
(CONT PAGE 5)
CHAPLAIN'S COLUMN
Where are we today in Civic Action? A great many of you men have been asking me this question. Many of you have a deep interest in doing something beneficial for the Vietnamese. May I offer you a few thoughts on this subject?
First of all, there can be no doubt that civic action is the most important activity that Americans can engage in while here. The goal of the Vietnamese government is to build a free nation at peace with its neighbors, along with a prosperous economy and strong educational system. Let us be assured that the responsibility for accomplishing this goal rests with the Vietnamese government. But we can help. Our military effort will help to end the terror and violence that has disrupted the country. We should also assist in the business of nation-building: encouraging the democratic process, priming the economic pump, and providing educational assistance.
We must certainly avoid making gross errors which usually stem from our native enthusiasm. Take, for example, the matter of Christmas. At Christmas time we like to give gifts, have a big dinner and make a big fuss over children. We come to Vietnam and try to overwhelm these people without Christian generosity. But the situation is that these people are not Christians, care nothing about Christmas and think that adults are a lot more important than children. For them the important celebration is at "Tet" which comes next year in early February. During this four day holiday the Lunar New Year is celebrated. The fact that American Servicemen try to be helpful to Vietnamese during Christmas does not abate the fact that we have not tried to understand these people. This year we have been told to de-emphasize Christmas celebrations in Vietnam except in Christian institutions.
We have hopes of having a Christmas party for Vietnamese who are Christians. Nearby us is the Catholic Refugee Camp of Phuoc Thun whose chief is Nguyen Huu Phu. Their priest is Nguyen Xuan Van. The population of this camp numbers 1500 people. We will ask them if they wish to celebrate Christmas with us. If they do not, we will not impose ourselves on them. We will also ask them about holding a
(CONT PAGE 5)
CHARLIES' CAPERS
by Geronimo Weitzman
As I begin this article, I would like to thank all the crews for their cooperation in giving me information on happenings around the camp and the jobs. Without the interesting little insights from the men on the job, I couldn't write the outstanding articles that I write.
As time flies by, all the men of Charlie Company are becoming accustomed to the work and beginning to be real professionals. As some of the jobs that we have built are coming to an end, the results are quite evident that we can do the job and do it well.
TAYLOR's crew is working hard at the last minute jobs at the open storage warehouse. They will finish ahead of schedule. ANDREWS and his illustrious man are nearing the finish of the Data Processing Building. They have learned a lot on how to be precise, especially in laying the raised deck. This building will house the most expensive electronic computers which will aid in the cataloging of supplies and materials for the Marines. PEDERSEN's crew is certainly taking a long time to complete the covered passageway at Building 78. It may take them five weeks to finish. Could the secretaries have anything to do with it? You are going to have to keep your mind on your work, men! The maintenance building at the DaNang Airfield for NSA is expected to be finished by Christmas, POLEY tells me. He says that it will be the Charlie Company present to NSA for Christmas.
(CONT PAGE 6)
THE KEYSTONE COMPANY
by M.D. NOLLSCH
Name me one person who can hitch-hike 1200 miles a week. As far as I know there's only one person who is able to do it. This man hiked from Davisville, Rhode Island, to Verona, Pennsylvania, every weekend during our stay there. Naturally there was a reason for it, his wife. I have to give her credit, she's one of the few women around who don't care if her husband gets picked up. It must be pretty hard to hitch-hike when you've promised the little lady you won't accept any rides with women, sort of takes the fun out of it.
Naturally I'm talking about Ronald McMASTERS, 26 year old Surveyor from Verona, Pa., where he attended high school, and I believe no graduated last year. Prior to coming into the service, he worked for Auburn Engineering in Pittsburgh, Pa. He never participated in sports during his high school days—I guess he was saving his legs for those long walks home later on.
Here's a person that the Seabees imported from Export, Pennsylvania, Ronald HOMCE, EAD2. He is 26 years of age, attended high school in Export, participated in basketball, baseball, sticking his tongue out at the teacher, etc. HOMCE worked for Swindell-Mueller Engineering Company in Pittsburgh, which is a rival firm of Auburn Engineering also in Pittsburgh. I think we've got a spy vs spy case going on over in drafting.
Paul RITS, EAD3, also comes from the Pittsburgh area. He attended high (CONT PAGE 7)
DELTA DEMONS
Here comes news of the 4.0 platoon. Since last talking to you, we've heard that HERBELL is putting on lots of muscle from taking his daily rations of Fizzies. Herby's new dazzling statistics are: wrist 8½; toes 5" each; elbow joints 12"; and ankles 2½. Gads, is there no end to his fabulous improvement!
Bob MARTIN's crew did a fine job on the showers with the help of part of OFF's crew. Both showers were completed at approximately the same time and all the men are to be commended for a real fine job. Now they are putting up strongbacks over at FLSO. Over at the DaNang Airbase there is only a couple hundred more feet of ditch to be placed. The reason SIMEK's Raiders got this job is because they are all qualified swimmers, so with snorkels, fins, and diving gear, they have been pouring concrete underwater. ERHARDT says he likes to pour underwater because that's how they do it in New York. GOLIO was angry because he's losing a few pounds. Most people would be happy to lose weight, but it seems that his wife likes him the way he used to be----(Tons of fun). While snorkeling through the area we overheard these comments: (scene in D Co. jeep) Mr. MARTIN: "Slow down Chief TARDY." Chief SCHWIDT: "Let me out." Mr. PACIFICO: "Hey, wait for me." MOSS: (Pouring concrete) "Why didn't I stay company clerk?" LISBY: "That's my flagpole, don't touch it."
Oh yes, before I sign off, we just had a late report that MAGFE has a (CONT PAGE 5)
In this article I will discuss Article 112 of the Uniform Code of Military Justice, Drunk on Duty. This is another of the more serious offenses against the UCMJ that can be violated while we are deployed in Vietnam. Drunkenness on or off duty, is a very common form of misconduct and frequently leads to very serious crimes. We will discuss what constitutes being drunk generally, and then we will consider its application in different situations. The word drunk is defined as any intoxication which is sufficient sensibly to impair the rational and full exercise of the mental and physical faculties. Article 112, UCMJ reads, "any person subject to this code, other than a sentinel or lookout, who is found drunk on duty, shall be punished as a court martial may direct." This article applies to all persons subject to the code except sentinels or lookouts, who are covered by similar provisions in Article 113, UCMJ. First of all, let's discuss being on duty. This includes every duty which an officer or enlisted person may legally be required by superior authority to execute. It also includes duties of an anticipatory nature. This includes a ready or standby status. To constitute this offense, a person must have undertaken the responsibility or entered upon the duty. The commencement of the duty status requires an affirmative act such as relieving someone of the duty or mustering with the duty section to which you are assigned. The duty status may be terminated by being relieved, dismissal, expiration of the period of duty, or abandonment of the duty. The second element to be proved is that the accused was drunk while on the duty alleged. It must appear that he was drunk while on duty. The drunkenness need not be the result of alcohol or drugs consumed while on duty, but may be, of course. Drunkenness may be proven by fact or opinion testimony of an expert or non-expert witness who witnessed the accused's condition via his own eyes, ears and or nose. It need not be proven that the accused's drunkenness was intentional or resulted from culpable or ordinary negligence. However, it must be the result of a voluntary act. A valid defense for this article is intoxication as a result of an accidental overdose administered for medicinal purposes. As I mentioned at the outset of this article, drunkenness can lead to very serious crimes. Drunkenness may lead to anything from disorderly conduct to involuntary manslaughter. If you drink while in Vietnam, be moderate in your consumption and be sure you can adequately hold the amount you drink.
This week I am starting off by introducing you to our latest addition in Alpha Company. He is Robert BARBER, SWF2, brother to our own Wayne BARBER, EOM2. Bob has been transferred to us from MCB-5. Another addition to the battalion, in the near future, will be William TRAILL's brother, Tom. Tom will be coming to us as a radioman, from the Sub Base at New London, Connecticut.
Besides those, Alpha Company has had many more additions lately, but they are in the equipment category. We have received new graders, pick-ups, cargo trucks, tank trucks, lube trailers, wood working trailers, and many others. The pick-ups will be replacing our weapons carriers, which have been on the job since 1953.
Our motor pool has been real busy hauling concrete lately. They have hauled 18,000 out of a total of 58,000 bags of cement here, from the piers, 30th NCR, and East DaNang. They have also been hauling mixed concrete to MCB-9 and for use here on our base for shower pads and other projects. Another large item being hauled constantly is our bulldozers. It appears that they're using more fuel idling on the way between job sites, than on the job sites. One last word from the motor pool. EOL WILSON, our license examiner, has this message: The written and driving tests, that go along with all new license applications, will be administered at 1300 on Tuesdays and Thursdays only, so plan accordingly.
(cont. page 8)
(CHAPlLIN CONT)
"fair" in their camp. In such a "fair", we would hope to sit down to a meal with them, sharing food, and participate in sports competition such as soccer and volleyball. We must certainly try to understand one another better and yet respect one another's customs and differences.
In the same way we will make friends in the Buddhist hamlets. At the present time we are working in Trung Nghia hamlet. Each week our MEDCAP team visits this hamlet providing medical and dental care. The Chief, Chau Iun, has asked us for help in building a dam across the river that waters his rice paddies, and a school house for the children. We have told him that he must work through his government's channels to get approval and materials for these projects. It may be that the government would not support them.
So, in our civic action program we are mostly trying to understand these good people and to do the things which are going to be of long-range benefit. We are deliberately going slow in order not to offend the Vietnamese.
In our particular area of responsibility there are no hospitals, orphanages, or institutions that need immediate support. There are only small hamlets where the greatest need is for employment in steady jobs. We simply are not able to meet this need. As we continue to visit the hamlets and understand these people we will find small ways to help them meet the goals they have set for themselves.
This is how it stands right now in Civic Action. By the way, Phuoc Than could use an English teacher. Any volunteers?
Don't buy Chinese-type goods in any country unless the shopkeeper can supply a special certificate of origin acceptable to the U.S. Treasury. Currently they are available only in Hong Kong. (AFNB)
(XO CONT)
of the United States or in accordance with regulations prescribed by the Secretary of Defense. Personnel sending gifts from an area other than Vietnam must be able to submit evidence they were in the area from which gifts were sent; or R&R or hospitalized as result of service in Vietnam. Alcoholic beverages and tobacco products are prohibited. Personnel entitled to this privilege will inscribe clearly on the outer wrapping of all gifts the following: "Public Law 89-368 Value not more than fifty dollars."
(D CONT)
girl who used to be a real square, but now we hear she's grown up into a full blossomed rectangle. You think that's bad, I used to know a comma who married a decimal point. Phew! must be the weather.
We all hope that Harold THOMAS who was admitted to Danang Hospital with tonsillitis, will be coming back real soon. At last report, he was looking fine.
Is it true that Speedy ESTRONZA has seen his share of action? While on a dangerous bolt-tightening mission he was ambushed by a crescent wrench. He escaped with a broken tooth while the wrench got away clean. Paul ROGERS attacked a peavy hook and suffered a slight nose injury while his adversary was broken in two.
Next week we'll hear more from the second platoon and by then we should have some interesting stories from our R&R boys.
There will be a muster of all "F" Troop personnel (Advance Party to Camp Lojeune and RWN) today, Sunday, at 1400 on the beach behind the EM club.
SGT. KING
PLAYBOY LOUNGE
by H. LOCKETT
Due to the shortage of beer at the Air Force Exchange, we have been forced to buy beer in broken case lots. The beer when received is loose. It has been exposed to the weather. This is the cause for the rust and uncleanliness of the cans. However, they are being washed prior to being cooled. It would be advisable to purchase a mug or glass of some sort to prevent drinking from the rusty cans. Don't give the bartender a rough time. There is nothing else to be had. "Help him to help you". Your cooperation will be appreciated.
Two pool tables are the latest addition to the lounge. These are for your enjoyment, so take good care of them. The lighting above the tables is inadequate at the present, but additional lights will be installed in the near future. Please do not set drinks on the tables. This can easily ruin the felt. These tables are leveled and balanced, therefore if you miss a shot, don't take it out on the table.
We do not have the keys to the Pin Ball machines or Juke box. When they are out of order, your patience and cooperation would be appreciated. A maintenance man from DaNang will come out and service the machines.
On the 15th, the club will be decorated for the holidays. Any one desiring to assist in the decorating contact H. LOCKETT, BUL; or K. CAPLISHER, BUL. Any ideas or suggestions are welcome.
(TEPS CONT.)
RESTAURANTS: The Drop-In Tavern on Oriental Avenue offers both a wide selection of food and entertainment. The Trolley on Silom Road offers the style of the "roaring twenties". For Chinese food, both The Golden Dragon on Sukhumvit road, and The Hiramai in the Rajprasong Shopping Center, offer the best in food and comfort.
NIGHT LIFE: Many of Bangkok's night clubs have a 50-60 Baht per person cover charge. For the entertainment and floor shows, there are the New Champagne Room in Custom House Lane; the Lido at Rajadamri Road; the Crown Club on Sukhumvit Road; and the Keynote Club on Pat Pong Road. Dancing partners at the Oasis Night Club on Mew Road cost from 30 to 100 baht per hour. For those who like to patronize the bars, the recommended ones are Max's Place, Rome Bar and the Can Can.
SPORTS QUIZ
1. In six seasons of play an American League Football player has missed only four extra points. Can you name him?
2. Notre Dame has made only one postseason Bowl appearance. Where and when was it made?
3. Big Dill Tilden was the top-ranked American Amateur tennis player for 10 years, 1920-29. He won several U.S. National Singles Championships. How many?
4. The United Nations sponsors teams in several sports and events. Can you name them?
5. What is the shortest touchdown pass on record?
(ANSWERS PAGE 8)
DaNang Deep Freeze built by Seabees and now operational is believed to be one of the largest cold storage areas in the world. It is capable of preserving meats, vegetables, and other perishable foods to feed more than 40,000 servicemen for several months. Prior to construction, several Navy refrigerator ships were required to supply food for Navy and Marine Corps personnel in I Corps area of the Republic of Vietnam.
(THE MASTHEAD)
The BLAZE is published weekly in accordance with NAVSO P-1035 (Rev. 12-65) on government equipment and is financed with appropriated funds. No material published herein is to be considered as representing the official opinion of the Defense Department or of the Commanding Officer unless otherwise indicated. The BLAZE uses AFNB material. All correspondence should be addressed to: Editor, FIFTY-EIGHT BLAZE, c/o FPO N.Y. 09501
Commanding Officer:
CDR W.W. DeGroot III
Executive Officer:
CDR L.D. LAWSON
Public Affairs Officer:
LTJG R.A. DERR
Editor:
JO3 R.D. BURNS
Layout:
SN J.L. HOLLIS
15,000 PATIENTS FLOWN HOME
The Military Airlift Command has returned more than 15,000 patients from the Pacific area so far this year, Col. Hubert H. Kerr, director of medical material at the Defense Personnel Support Center in Philadelphia, told delegates attending a meeting of the Association of Military Surgeons.
He noted that aeromedical evacuation of wounded serviceman from Vietnam has resulted in the highest medical save rate in combat history.
Statistically, Col. Kerr explained of wounded servicemen who get substantial medical attention--more than battlefield first aid--only about 1 per cent die.
The same type casualty in the Korean War was more than double the rate in Vietnam. In World War II the rate was about 1.5 per cent and about 8 per cent in World War I.
A wounded American in Vietnam who reaches "definitive medical aid will survive 99 times out of 100," the Colonel said.
(VFNB)
VIETNAMESE LESSON
The lesson this week will be a continuation of the common everyday greetings that was introduced in a previous issue. The following pertain to this week's lesson:
How are you?
Ông mạnh giỏi không?
Bà mạnh giỏi không?
Cô mạnh giỏi không?
Em mạnh giỏi không?
giỏi (voy)
Thank you, I'm fine.
Cảm ơn ông. Tôi mạnh giỏi.
Average age of U.S. sailors and Marines today is 23 years. There are about 250,000 enlisted men in the Marine Corps and approximately 660,000 in the Navy. (AFNB)
(ALPH. CONT)
Our well drilling crew, consisting of EOL CASTONGUAY, EOR2 FORTUNE, EOM3 MORGAN, and CN THOMPSON has finally gotten under way. They are at MAG 11, where they have drilled over seventy-five feet. When finished there, they plan to begin drilling by the NSA Maintenance building.
Meanwhile, down on Route 4, over 1300 yards of 4 inch rock has been placed on only the first five hundred feet of this road. There's a pretty deep hole there, and Chief BENDZLOWICZ has had almost 500 yards of large rock put in it so far, and they still haven't found the bottom.
We've had all kinds of interesting things going on in our shops. If you've been near the lube pit, you'd have noticed their pretty pink coffee mess. The rumor is that they think that pot is filled with instant "sex". It seems that CLARK and KENYON made themselves a cuss box, of all things. It only costs them 10¢ a word. The idea has already spread to Hut 17, the battery shop, and spare parts. I'm afraid there's going to be quite a few needy mechanics around in a short while.
We had a bit of excitement here last week, in the tire shop. URBANIK and DAVIS were inflating a tire there, when the safety rim came off, and it exploded. Both men received minor cuts.
The fellows in the battery shop are curious to know whether or not LIGHTFOOT has been testing any more switches lately. It seems that he's had a shocking experience with one already.
I was considering sending an instruction manual for installing exhaust systems in one particular jeep, they spent the rest of the night trying to get it started. After tearing apart the carburetor and the distributor to no avail, they finally decided to remove that plug that was left in the new exhaust system. How about this, TRUITT?
If any of you should step outside your hut to find a boardwalk there, that was not there a while ago, don't be surprised. My crew, consisting of DOANE, ABEAUX, and RÜCKE have been busy this week making improvements to the Alpha Company area. Besides building boardwalks, we've been patching up the hards, reshielding the urinals, making improvements to the Company C.P., and painting new signs for each hut.
My last note comes from our Accident Investigator, Mr. REWOLINSKY. He is thinking of having several sets of overhead power lines and an underground water line installed near here. These won't be real, of course, but they will enable the spotters (UT's and CE's) to become more acquainted with spotting these lines for our operators who are working in the areas where these lines are located. Lately, the spotting hasn't been quite up to par with evidence to prove it.
(CHARLIE CONT)
RUMBLINGS ABOUT THE CAMPUS:
-- SPINDLER, what's wrong with your boots? Been walking all over DUKE's concrete? Shame on you.
-- I hear HODGES is doing good with the sledge hammer. He only missed once but hit DUKE in the head. It looks like DUKE is getting it from all sides.
-- I hear DAWKINS loves those Marines to walk on his forms.
-- Men, does GRONKE like to dig those ditches.
-- RUNYAN works so fast they call him "Lightning".
-- What's this about GRISSE and STONE doing a little Slippin' and Slidin' on rooftops lately.
From Rupert, Idaho, comes Ken VILHAUER. VILHAUER was born on April 16, 1945. A country slicker, VILHAUER has worked on farms all his life. He worked in a potato plant before joining the Navy in 1965. A fine artist, VILHAUER has done many portrait sketches for his friends. When he gets out, he plans to attend college in California and major in Art.
Ronald NICOLSSON was born in December 1946, in Long Beach, California, and was raised in Sonora. While attending Sonora High School, he was very active in sports. He has worked for J. S. West, installing appliances. He has also worked in a hardware store as a salesmen.
One of the swinging Soul Brothers, L. E. BROWN, hails from Sioux City, Iowa. He was born in March of 1946. L. E. was a telephone installer before joining the Navy last year. Rad-Man BROWN has distinguished himself in the battalion, by being one of the stars on the Great 58 Drill Team.
ANSWERS
1. George Blanda of Houston
2. It was the 1925 Rose Bowl game. Notre Dame beat Stanford 27-10.
3. Seven
4. The UN sponsors teams in polo, golf, cricket, table tennis, bridge and karate.
5. The shortest touchdown pass on record went from Eddie LeBaron to Dick Bielski of the Dallas Cowboys. It traveled two inches. |
ON WEAKLY ORDERED SYSTEMS
M. RICHARDSON
1. Introduction. By a weakly ordered system we mean a system $D$ of elements $x, y, \cdots$ with a binary relation $\succ$ such that
(1) \[ x \succ y \text{ implies } x \neq y \]
and
(2) \[ x \succ y \text{ implies } y \succ x \text{ is false.} \]
The statement "$x \succ y$" may be read "x dominates y." Transitivity is not assumed; a transitive weakly ordered system is a partially ordered system. By a solution of a weakly ordered system is meant a set $V$ of elements of $D$ such that (a) $x \in V$ and $y \in V$ implies $x \prec y$ is false and (b) $x \in D - V$ implies $y \succ x$ for some $y \in V$. The concept of solution was introduced in J. von Neumann and O. Morgenstern, *Theory of games and economic behavior*, Princeton, 1944, where it is proved that a weakly ordered system which is strictly acyclic\(^1\) possesses a solution which is unique, and for which a construction is given. This result suggests the problem of finding conditions for the existence and uniqueness of solutions of weakly ordered systems in general. The simplest examples show that if cycles exist neither the existence nor the uniqueness of solutions can be expected in all cases. For example, the system of three elements $a \succ b \succ c \succ a$ has no solution, while the system of four elements $a \succ b \succ c \succ d \succ a$ has the two solutions $(a, c)$ and $(b, d)$. The purpose of this note is to prove the existence of solutions for certain non-acyclic systems. The proof will itself provide a method of construction for the solutions. Zermelo's axiom of choice, the well-ordering theorem, and transfinite induction will be used. The result presented below is a contribution to the general problem suggested above rather than to the theory of games. For the hypothesis of the theorem below precludes transitivity completely; that is, it precludes the existence of three elements $a, b, c$, such that $a \succ b, b \succ c,$ and $a \succ c$. This restriction is too severe for the theory of games, just as is the assumption of transitivity. The problem remains open for weakly ordered systems which are not strictly acyclic but also do not satisfy the hypothesis of the theorem below.
\(^1\) See loc. cit. pp. 590–600 for definitions and proof.
Presented to the Society, September 17, 1945; received by the editors August 10, 1945.
2. The theorem. We shall find it convenient to represent the weakly ordered system $D$ by an oriented 1-complex or linear graph whose vertices are in 1-1 correspondence with the elements of $D$ and such that two vertices $x$ and $y$ are joined by an oriented 1-cell $xy$, oriented from $x$ toward $y$ if and only if the element corresponding to $x$ dominates the element corresponding to $y$. This oriented graph will also be denoted by $D$, as well as its set of vertices, and we shall denote indiscriminately by $x$, $y$, $\cdots$ either the elements of $D$ or the vertices of the graph with which they are identified. The graph $D$ of a weakly ordered system is not the most general oriented graph; it contains no loops consisting of an oriented 1-cell whose initial and terminal vertices are the same (condition (1) above), and no two vertices are joined by more than one 1-cell (condition (2) above). A graph is called even if all its unoriented\(^2\) 1-cycles contain an even number of 1-cells (or vertices). We shall prove the following theorem.
**Theorem.** If the graph $D$ is even, then solutions exist.
Of course, the hypothesis of this theorem implies that all cycles, in the sense of von Neumann and Morgenstern, loc. cit., are even, but not conversely.
3. A lemma on graphs. We shall use the following known lemma whose proof we include for the sake of completeness and to facilitate the construction of examples by the reader.
**Lemma (König).\(^3\)** The set of vertices of an even graph $D$ can be decomposed into two disjoint sets $P$ and $Q$ such that every 1-cell of $D$ joins a vertex of $P$ to one of $Q$.
**Proof.** Select a definite vertex $x$ of any (connected) component of $D$. Put a vertex $y$ of the same component into $P$ or $Q$ according as the unoriented 1-chains joining $x$ and $y$ have an odd or even number of 1-cells. If a vertex $y$ could be joined to $x$ by two 1-chains, one with an odd and the other with an even number of 1-cells, then there would exist an unoriented 1-cycle with an odd number of 1-cells, contrary to hypothesis. This may be done for each component of $D$. This proves the lemma.
4. Proof of the theorem for finite systems. For simplicity, we give the proof first for finite systems.
---
\(^2\) That is, modulo 2.
\(^3\) D. König, *Über Graphen und ihre Anwendung auf Determinantentheorie und Mengenlehre*, Math. Ann. vol. 77 (1916) pp. 453–465. Cf. also D. König, *Theorie der endlichen und unendlichen Graphen*, Leipzig, 1936.
Let $P$ and $Q$ be two classes selected according to the lemma. Let the elements $p$ of $P$ be ordered by means of a subscript index ranging over a lower segment of (finite) ordinal numbers. Let $P_0 = 0$ and $Q_0 = Q$. We now define sets $N_k$, $P_{k+1}$, $R_k$, $Q_{k+1}$ inductively for $k = 0, 1, 2, 3, \ldots$. Let $N_k$ be the set of all elements of $P - P_k$ which are not dominated by any element of $Q_k$. If $N_k = 0$, let $V = P_k \cup Q_k$. If $N_k \neq 0$, let $p_{\alpha_k}$ be the first element of $N_k$, that is, the element with the lowest ordinal subscript. Let $P_{k+1} = P_k \cup [p_{\alpha_k}]$. Let $R_k$ be the set of all elements of $Q_k$ which are dominated by $p_{\alpha_k}$. Let $Q_{k+1} = Q_k - R_k$.
Since the set $P$ is finite, it is clear that there is a least ordinal $\lambda$ for which $N_\lambda = 0$. We shall prove that
$$V = P_\lambda \cup Q_\lambda$$
is a solution.
Notice the obvious relations
$$P_0 \subseteq P_1 \subseteq P_2 \subseteq \cdots \subseteq P_\lambda \subseteq P,$$
$$Q = Q_0 \supseteq Q_1 \supseteq Q_2 \supseteq \cdots \supseteq Q_\lambda,$$
$$(6) \quad Q = (\bigcup_{\beta < \lambda} R_\beta) \cup Q_\lambda \text{ where the sets } (\bigcup_{\beta < \lambda} R_\beta) \text{ and } Q_\lambda \text{ are disjoint.}$$
We must prove: (a) if $x \in V$, $y \in V$, then $x \prec y$ is impossible; (b) if $x \in D - V$, then there exists an element $y \in V$ such that $x \prec y$.
**Proof of (a).** Suppose $x \prec y$.
(i) Suppose $x$ and $y$ belonged both to $P_\lambda$ or both to $Q_\lambda$. Since $P_\lambda \subseteq P$ and $Q_\lambda \subseteq Q$, this would contradict the property of $P$ and $Q$ given in the lemma of König.
(ii) If $x \in P_\lambda$ and $y \in Q_\lambda$, then $x = p_{\alpha_\gamma} \in N_\gamma$ for some $\gamma < \lambda$ and hence is not dominated by any element of $Q_\gamma$, and, a fortiori, of $Q_\lambda$. This is a contradiction.
(iii) If $x \in Q_\lambda$ and $y \in P_\lambda$, then $y = p_{\alpha_\gamma}$ for some $\gamma < \lambda$. Then $x \in R_\gamma$ and cannot be an element of $Q_{\gamma+1} = Q_\gamma - R_\gamma$, nor, a fortiori, of $Q_\lambda$. This is a contradiction.
**Proof of (b).** Suppose no such $y$ existed.
(iv) If $x \in P \cap (D - V)$, then $x \notin P - P_\lambda$ since $N_\lambda = 0$. Hence $x \in P_\lambda$. Therefore $x \in V$. This is a contradiction.
(v) If $x \in Q \cap (D - V)$, then, by (6), $x \in R_\gamma$ for some $\gamma < \lambda$. Then there exists a $y$ such that $x \prec y \in P_{\gamma+1} \subseteq P_\lambda \subseteq V$. This is a contradiction.
This completes the proof.
5. **Proof of the theorem for infinite systems.** We indicate the modifications necessary in the infinite case.
We assume that all finite unoriented cycles have an even number of 1-cells. By the component to which a vertex belongs we mean the
set of all vertices which can be joined to it by finite unoriented 1-chains. Then the lemma of König holds as before. Now, let the elements $p$ of $P$ be well-ordered by means of a subscript index ranging over a lower segment of ordinal numbers. The construction of the sets $N_k$, $P_{k+1}$, $R_k$, $Q_{k+1}$ will be as before for finite $k$. But now it may happen that $N_k = 0$ for no finite $k$.
Consider any ordinal number $\gamma$. We shall construct the sets $N_\gamma$, $P_\gamma$, $Q_\gamma$, $R_\gamma$, and so on, by transfinite induction.
**Case I.** If $\gamma$ is a limit ordinal and $N_\beta = 0$ for no $\beta < \gamma$ then let $P_\gamma = \bigcup_{\beta < \gamma} P_\beta$ and $Q_\gamma = \bigcap_{\beta < \gamma} Q_\beta$. Let $N_\gamma$ be the set of those elements of $P - P_\gamma$ not dominated by any element of $Q_\gamma$. Let $p_{\alpha_\gamma}$ be the first element of $N_\gamma$, and let $P_{\gamma+1} = P_\gamma \cup [p_{\alpha_\gamma}]$. Let $R_\gamma$ be the set of those elements of $Q_\gamma$ dominated by $p_{\alpha_\gamma}$, and let $Q_{\gamma+1} = Q_\gamma - R_\gamma$.
**Case II.** If $\gamma$ is not a limit ordinal, it has an immediate predecessor $\gamma - 1$. If $N_\beta = 0$ for no $\beta < \gamma$, let $p_{\alpha_{\gamma-1}}$ be the first element of $N_{\gamma-1}$. Let $P_\gamma = P_{\gamma-1} \cup [p_{\alpha_{\gamma-1}}]$. Let $R_{\gamma-1}$ be the set of those elements of $Q_{\gamma-1}$ dominated by $p_{\alpha_{\gamma-1}}$. Let $Q_\gamma = Q_{\gamma-1} - R_{\gamma-1}$. Let $N_\gamma$ be the set of those elements of $P - P_\gamma$ not dominated by elements of $Q_\gamma$.
Let $\pi$ be the ordinal number of the set $P$ in the chosen well-ordering. Note that $\beta < \gamma$ implies $\alpha_\beta < \alpha_\gamma$. There must exist a least ordinal $\lambda$ such that $N_\lambda = 0$, $\lambda \leq \pi$. Let $V = P_\lambda \cup Q_\lambda$. It is readily seen that (4), (5), and (6) hold whether $\lambda$ is a limit ordinal or not. The proofs of (a) and (b) can now be seen to proceed as in the infinite case.
Brooklyn College |
Friends’ Reference Library, 1901-1959
Part of an Address given by Muriel A. Hicks to the Friends’ Historical Society on 31st March, 1960.
ON the early issues of the Journal of the Friends’ Historical Society was an emblem on which were the words, “the roots of the present lie deep in the past.” Anna L. Littleboy in her Presidential address to the Friends’ Historical Society, 1920, “A History of the Friends’ Reference Library,” covers the years from 1673-1901 and demonstrates, through the reading of early minutes, the conscientious care in the collecting of all records, which is still the foundation of this library. This paper is an attempt to survey the developments of the library through its Committee since the appointment of Norman Penney, the first librarian, in 1901, and of M. Ethel Crawshaw, first assistant, in 1903, a record of approximately 60 years.
When the librarian was appointed there was no especial room known as “The Library” at Devonshire House. There were, however, three strong-rooms under his care, one on the first floor opposite the “Meeting for Sufferings Room,” the second on the ground floor in the yard, tucked away behind a small cloakroom, and the third below the Recording Clerk’s Office, at the bottom of a flight of steep stone steps. The manuscripts and 17th century printed books were in one or other of these rooms, but bundles of manuscripts were also found in cupboards in the “Meeting for Sufferings Room.”
The early days of the librarian were spent in the first strong-room upstairs, about 8 ft. × 20 ft. in size. The “Meeting for Sufferings Room,” across the landing, had cupboards containing books published after 1700, but it was full of forms, with a long refectory table (now in the canteen at Friends House) and was reserved for other Committees besides Meeting for Sufferings, such as the Friends’ Foreign Mission Association, the Home Mission Committee, and the Six Weeks Meeting.
Pressure of space for the librarian, working in a comparatively small room with poor lighting and ventilation, and the need for suitable accommodation for readers,
led the Committee in 1902 to send forward a recommendation to the Meeting for Sufferings "that owing to the more frequent use of the Reference Library (meaning the collection) the want of a suitable room in which Friends may read is increasingly felt. The observance of the rule that Friends shall not be left alone in either of the strong-rooms is often a serious drawback, involving inconvenience and loss of time to the librarian. The Committee feels that the time has come when it may ask the Meeting for Sufferings to allow the use of the Meeting for Sufferings room for the purpose of research and for the conduct of the ordinary work of the librarian, except when the room is especially required as a Committee Room. The Committee therefore suggests that in future the room is called the Library."
The next month the librarian was installed in what henceforth was to be known as the "Library" of the Society of Friends. The minutes of the Committee for some years onwards convey a feeling of excitement and enthusiasm.
**THE CATALOGUES**
Joseph Smith's *Descriptive catalogue of Friends' books*, published in 1867, was the basis for the catalogue of the library. An interleaved copy, which had already been used by Joseph Smith himself for additional entries, was continued and as items were found the press-mark was inserted in the margin of the page. In 1903 the Committee proposed there should be a card catalogue, "Smith's Manuscript interleaves being less useful," and in the following year, "one oak cabinet (meaning one drawer) for the catalogue" was purchased, to be followed by a second cabinet. This continued until two large trolleys on rubber wheels were built, to contain about seventy drawers. They were taken daily from the upper strong-room across into the library and back again at night and were the foundation of the present catalogue. Smith's catalogue was taken each day to a safe in the strong-room in the yard. Incidentally, the ventilation of this strong-room was also attended to daily by a member of the library staff. Fire precautions at Devonshire House were limited, open fires were in all the Committee-rooms and there were two in the library; the building in general was of course without any of the modern fireproof materials or safeguards, hence the necessity of taking such care of the catalogues.
In 1923, it was reported that there were 70,000 cards and some 200,000 author, title and subject entries.
In the lower strong-room, below the Central Office, wooden boxes contained the great mass of 17th century loose manuscripts, which were for the time being only numbered and placed in portfolios. "Gone through a collection of old manuscripts to be kept in a box labelled A.R.B.," again in 1902, "Bundles of manuscripts concerning Finland Famine 1857, placed in a box marked 'Finland Distress'," and so on.
Mr. Francis B. Bickley, of the British Museum, was invited to Devonshire House to see the library and its contents. He advised that the binding should be in "½ morocco and buckram" for all the portfolios of manuscripts, and as for the printed collection, "avoid binding tracts of different sizes together; the dust collects." Advice on the durability of paper and typewriting inks was sought, an enquiry which was made again about forty years later. A small room tucked away was put at the disposal of a binder, a Friend, William Crump, who was a craftsman at his trade, and was set up with the necessary equipment to bind the portfolios of manuscripts and volumes of tracts, as his time allowed, a work which covered many years.
On the purchase in 1903 of the Thompson Collection, collected by Thomas Thompson of Liverpool, father of Silvanus Thompson, which contained manuscripts, printed volumes and a large number of 17th century pamphlets, it was necessary to provide boxes for their safety. The first seventy-five boxes, each numbered, were purchased especially for a pamphlet collection. These were in red cloth. Twenty-eight boxes, in green cloth, for the "Adverse Tracts," each lettered in alphabetical order, were also provided, and guard books for the larger pamphlets and publications were used. The first shelf catalogue for all printed books was completed by 1907. This gave a record, shelf by shelf, of what was kept there.
**LOANS OF BOOKS**
Just how early books were loaned to readers is not quite clear, but in 1901 it is reported that there were "more books gradually borrowed." Cards stating the conditions of loan were printed. It was then necessary that each borrower should have a member of the Meeting for Sufferings as a
guarantor, and the loan was for three weeks. After a few years this rule was dropped in the case of books but not pamphlets; by 1925 the loans reached to about 350 a year. In 1905 Rufus M. Jones requested the loan of books to America to write an article on Robert Barclay for the *Friends' Historical Journal*, "but the Committee does not feel its way to send books abroad," was the reply to that request.
With the gradual collecting of manuscripts, books and pamphlets, came a policy of thorough completeness, even retrospectively, and the necessity to consider the limits of space and expense did not arise for some years.
Every effort was made to interest Friends in the library and the care of records and to encourage Friends who had papers needed, or books, to present them to the library.
A list of books still required was printed on three different occasions and circulated to preparative meetings, in case copies should be found in a Meeting House library. "At Homes" were held in the library, inviting public librarians, members of the Society of Genealogists, and Fellows of the Society of Antiquaries, as well as Friends. A special one was held to "celebrate the binding of the 500th volume of tracts." A pamphlet, "Preserve our History," was published in 1907, and in the same year a Meeting was held in the large Meeting House, where Thomas Hodgkin, Silvanus Thompson and others addressed the Meeting on "Quaker Literature and its Aims." Many exhibitions were also arranged, as opportunity occurred.
This determined effort to collect all material was extended to America. In 1901 Norman Penney wrote to Haverford College for back issues of American Yearly Meeting proceedings. Parcels of books, pamphlets and periodicals were sent each way across the ocean to supply the needs of the respective librarians.
In 1911 Norman Penney visited America and gave a course of lectures on Quaker history and naturally made it his business to contact Quaker libraries. In 1914 and in 1923, the then assistant librarians, M. Ethel Crawshaw and John L. Nickalls, also paid visits to the States under the auspices of a Young Friends visit, but again contacts were made with libraries.
Visitors recorded in the early visitors' book were at the rate of about one a month, sometimes not that, but the
subjects were very much like those today. "Welsh Quakers in America," "History of Adult Schools," "Music a Part of Worship," "Friends in Holland," "Quaker Women," "American Separations." Work on the "Rowntree Series" of Quaker histories was begun and naturally the Library was the place where a great deal of research was done, which lasted for some years. In 1910 Dr. Auguste Jorns, the first visitor from the Continent, of Freiburg University, Germany, arrived; she was sent by her Professor to get material for a thesis on the "Social Work of Friends." The direction he gave her was, "to go to Liverpool Street Station, and having crossed the road to enquire for Devonshire House; if she went through a long tunnel and up some stairs, she would find a door with 'Library' on it. If she went in, some people would help her." The committee was still cautious as to what source-materials Friends could see for their research work. An American Friend came over, to work on a "Life of William Penn." He was allowed access to all Penn material, "but in case he should desire to use hostile literature, the request to be referred to the Committee."
A great deal of genealogical work was done. The registers were housed in the Recording Clerk's office, and had to be carried along the tunnel and up the stairs to the library, which was sometimes quite a heavy task.
As the years went by, the need for further equipment and space was evident. More forms were removed from the library, and book cases erected under the windows around the room. Collections of manuscripts were presented or purchased; for example, "The Thirnbeck Manuscripts," a collection of "Swarthmore Manuscripts" now described in the library as Vol. 7 of that collection, and most notably of all, the purchase, jointly by American and English Friends, in 1920, of the three volumes of Spence MSS., the first two of which contain the manuscript Journal of George Fox, the third a volume of letters. Meeting records from many Quarterly Meetings were sent to Devonshire House under the care of the library.
In 1910 the committee reported, "that notices of the library and some of its principal possessions have appeared in such books as the 'Guide to Manuscript Materials for United States History . . .' 'The Libraries of London' by the
¹ A. Jorns, Studien über die Sozialpolitik der Quäker, Karlsruhe, 1912.
Goldsmiths’ Librarian, University of London, and the ‘Literary Year Book’; the work of this Committee has thus been brought under the notice of a wider circle.”
A newscutting service was subscribed to for £3 3s. od. per year. This service gradually extended and became a shared interest for other committees. The picture collection began to take shape, along with the other collections; portraits, pictures of meeting houses, schools, etc., were all pasted into albums. The well-known collection of Spence etchings from George Fox’s Journal began by the purchase, by subscribers, of thirty-seven of the etchings. In 1901 the Meeting for Sufferings had to give permission for a photograph to be made for publication of an item in the library! Friends were still very cautious in those days!
The committee, then known as the Library and Printing Committee, had under its care not only the cataloguing of books and manuscripts, but the preparation and seeing through the press of the Society’s publications, *Yearly Meeting Documents in Advance*, *Yearly Meeting Proceedings*, and the *Epistle, Book of Meetings*, revisions of the *Books of Discipline*, among others.
The Committee was also responsible for the “Free Grants List of the Meetings for Sufferings.” In 1905 there were fifty-five books on the list. They were sent to Meetings in England and abroad, at their opening, to enquirers and to public libraries. The librarian was in touch with the “Continental Committee,” and suitable books were translated into French and German. Copies of *Christian Practice* were given to Quaker Chaplains, for distribution to C.O.s in prison during the first World War.
Friends’ books for sale in the bookshop (not then owned by the Society), were also part of the committee’s work, and great thought was given to the supply of Friends’ books that would meet the demand of the public in that way.
By 1918 the shelves in the library were full. There was no space for further book-cases and it was difficult to find space for books as they were added to the library. The files of periodicals and reports of Friends’ schools and committees had also increased.
In that year a proposal for co-ordinating the literary work of the Society was brought before Yearly Meeting, and in the following year the Central Literature Council was
established. This actually relieved the library of the oversight of the Society’s publications already mentioned. Some of it went to the new Council, and some to the Central Office. The Free Grants list went to the Literature Committee, which later became a sub-committee of the Home Service Committee.
The Library and Printing Committee remained a committee of the Meeting for Sufferings until 1925, when it became a sub-committee of a co-ordinating Literature Committee. Four years later it reverted and again came directly under the Meeting for Sufferings. Meanwhile the name was changed and henceforth it was to be known as the Library Committee.
**THE MOVE TO FRIENDS HOUSE**
With the possibility of the move from Devonshire House, the expectations and plans for a new library were great. Hitherto there was no classification of books; to meet this, a decimal classification scheme, based on Dewey, was worked out by John Nickalls, the Assistant Librarian, and work was begun on the books a good year before the move in 1925.
A proposal was also made by him for a complete catalogue of printed books, based on the existing entries on the card catalogue, by the incorporation in it of the entries in Joseph Smith’s *Descriptive catalogue of Friends’ books* (1867). This was accepted, and a beginning made by each Smith entry being pasted on a card.
A new system for recording loans was introduced by John Nickalls in 1924; the necessity for a borrower to obtain a member of Meetings for Sufferings as guarantor was removed, and the books were available to all Friends. Moreover non-Friends, for the first time, could borrow books provided the necessary guarantee was made, and special arrangements were provided for borrowers abroad. This was encouraged largely through the establishment of the work of the Council for International Service, later merged with the Friends’ Foreign Mission Association to form the Friends Service Council.
Weeks before the move to Friends House, we began packing in the strong-rooms. Waterproofed lined wooden boxes were provided, and the volumes of manuscripts, and seventeenth century printed books were carefully packed.
Each box contained a duplicate list of its contents. The main library was packed by the movers, and as we left that old room, the empty book-cases were being pulled down around us, bringing with them, it seemed, the dust of ages. Friends House was an uncompleted building when we moved, the library a mere shell, and most of the boxes were taken upstairs into the Institute, whilst the staff found temporary accommodation at the Penn Club (then in Tavistock Square), and later in the Institute. The precious interleaved volumes of Smith’s Catalogue, used for reference work, were daily carried from the safe somewhere among the scaffolding at Friends House, to the Penn Club, and back again in the evening.
The actual instalment in the library was about Easter time 1926, and the thrill of placing books for the first time in a classified order, in the new book-cases, was not easily forgotten by us. Naturally, there were alterations and adjustments to be made, but gradually the staff settled down to their new surroundings. The move to Friends House brought with it the strong desire that the work of the Society should progress, and that the new building should be put to the fullest and best possible use, to make known the principles of Friends. Despite the great affection most people had for our old historic site at Devonshire House, it was uneconomic to run, and had many limitations. Now, there was a sense of freedom, even to experiment with new ideas. The Library Committee shared in this, and desired to make more easily available the use of the printed word, especially in a Religious Society which has no paid ministry. Suggestions were made that the library should broaden its scope and take more religious, mystical, devotional and social books. One result of this was that the Industrial and Social Order Committee’s library was housed in the library, and later, that of the Penal Reform Committee, and the Marriage and Parenthood Committee, to be followed by others, including a section for “Enquirers” to borrow. Mutual arrangements for loan of books between this library and Woodbrooke were made.
Since coming to Friends House, the library had one strong-room, and shared a second, for deposited minute books. Although still some distance from the library, the convenience of everything being in one room was greatly appreciated. Easy access to these documents and economy
of space was essential, and many schemes were adopted to enhance this, including the refurnishing with steel cupboards.
After the move, the large picture collection from the London Friends’ Institute at Devonshire House came under the care of the library. This added considerably to the collection, but still more important was the purchase of certain Quaker pictures as they came to the notice of the library. The care to establish the authenticity of any doubtful picture was the librarian’s concern. No longer did the Committee apply to the Meeting for Sufferings for permission for an item to be photographed; indeed, if that were so, much precious time of that Meeting would be used for that purpose, so large has the demand grown. In all this development of the library, the Committee believed it was its main duty to “accumulate and make available material which may be mediated by students for the good of general readers, rather than appeal to these.”
The rapid growth of the library during these years, plus correspondence and attention to visitors, with a small staff, was at times overwhelming. For some time the genealogical work was not done by the staff, but someone interested from outside was called in to do it and paid accordingly. Not for many years had the library reminded local Monthly and Quarterly Meetings of the duty to care for their local records, but now an informative document was circulated, enquiring and advising about the physical conditions under which these records were kept—the lack of ventilation and dampness and need of repairs, also the advisability of depositing the records in the local County Record Office, where their own local conditions were not suitable. This led to a great many documents going to the local archives office, by whom a full and detailed list is provided both to the particular Monthly or Quarterly Meeting concerned and to the library.
**THE WAR AND AFTER**
The outbreak of World War II found the library prepared for action in the case of manuscripts and valuable printed books. It was the policy of the Committee that they should be placed in safe keeping, and four different places in the country were found. Some five hundred volumes were placed in that way, and some 1,000 books from the library taken to
the strong-room. Through the generosity of a Friend, a microfilm of the card catalogue and of certain irreplaceable manuscripts was made and sent to Haverford College. Once again Friends in America came to our aid and collected books, pamphlets and periodicals, to be preserved for the library when the war was over. The work of the library continued very much as usual, with fewer readers. Despite certain difficulties an exhibition was arranged in 1944, open to the public for four days, celebrating the tercentenary of the birth of William Penn. The demand for devotional and mystical books by borrowers increased during the war, and a member of the Committee provided a sum of money for several years which enabled such books to be provided, the Committee feeling that they did not come within their budget policy.
After the war was over, and the books and manuscripts restored to their places, the Committee was faced with the cry of all librarians for space and yet more space! The library was now half as large again as it was in 1924, and the loans increased from 400 to 2,000 a year.
With the sanction of Meeting for Sufferings, second copies of many editions of books preserved by the Committee under its original Minute, "two of all books written by Friends and one against them" were disposed of; this released some 2,000 volumes. Ephemeral periodicals and publications were discontinued, and a policy of careful scrutiny regarding the intake for all the collections, bearing in mind the permanent value of the material, was followed.
WIDER CONTACTS AND CO-OPERATION
In 1947 an exchange of the Assistant staff of Haverford, Swarthmore and Friends House libraries was arranged: each American visitor worked two months here, and I worked two months in each of the American libraries. This valuable intervisitation was a stimulating experience, apart from the knowledge gained of the contents of each library and the methods used.
Mention has already been made of the microfilm. One of the outcomes of this exchange was the much wider sharing of the various Manuscript Collections of each library through the microfilm. To enhance this a microfilm reader was presented to the library by American Friends, "to the
Memory of Rufus Jones." American Friend colleges were not the only concerns interested in this process, but large numbers of manuscripts have been microfilmed for South Africa, Nigeria, Virginia and for many individuals and libraries. Moreover the loan of films to libraries has made it possible for readers to use the manuscripts in that way locally, rather than come to London.
The possession of a film not only safeguards against the complete loss of a manuscript, but has justified the loan of the original manuscript on occasions of particular importance. Thus, after 1949, as the European countries recovered after the war and contacts reopened, the British Council arranged an Anglo-Dutch exhibition in Amsterdam, to which a letter from Margaret Fell to William III was loaned. In this country, during the Festival of Britain, Volume I of the Manuscript Journal of George Fox was loaned for eight months to the Victoria and Albert Museum, being one of two manuscripts, in the field of religious literature, exhibited.
In 1928 the library became an "outlier library," to the National Central Library. This last opened the door to the outside world of libraries, and books were borrowed each way. It later became a member of Dr. Williams's Library, and joined the scheme for inter-library loans to reference and university libraries for special books and certain manuscripts, as required. In 1939 it became affiliated to the British Records Association and was also one of the original members of the Standing Conference of Theological and Philosophical Libraries.
A further contribution, in the more national sense, was the library's contribution of a list of well over 400 periodicals held here to the published *British Union Catalogue of Periodicals*.
In the last few years the proposal made to combine Smith's Catalogue with the main card catalogue, has been achieved. By the co-operation of the library, Wing's *Short-Title Catalogue* of English Books, 1641-1700 includes entries for works to be found at Friends House.
Various other schemes for the publicity of the library to Friends and the public were constantly brought forward by the librarian. The routine work with which we were engaged; the constant attention to an increased number of
long-term research workers, some of whom stayed for months, even years; the help given to borrowers; along with the constant planning and adjustment, made the work absorbing and full of interest. Days went far too quickly for the work that had to be done.
Two wars and a growth of interest in social questions has greatly increased the number of students doing research work for advanced degrees and diplomas. For Friends the library has provided the material for their especial interests as they have arisen, through the housing of the particular Committee’s library and obtaining on loan books from other libraries. Throughout the years, however, Quaker biographies have been more borrowed than any other class of books.
Despite the amount of material now available and made so easily accessible, the output of Quaker books in this country has declined, for various reasons. The “pamphleteering” age seems to have returned, there are times when the library does not appear to be used either for historical purposes, or for borrowing books, to the extent that it might be, by Friends. In the early part of this century, the library was mainly used by members of the Society, now the far greater use is made by non-Friends, for research purposes.
It has appeared that America, with its larger number of Friends, has supplied a greater proportion of recent books on Quakerism. It may be that in their membership there are a greater number of people more qualified than in this country. It surely behoves more Friends to make enquiries and to see for themselves what the library holds and the need it can supply, for there is much that can be used and put to great service.
**Librarians:**
- 1901-21 Norman Penney.
- 1921-26 M. Ethel Crawshaw.
- 1927-57 John L. Nickalls.
- 1957- Edward H. Milligan.
**Clerks to Committee:**
- 1895-1906 Frank Dymond.
- 1906-47 Anna L. Littleboy.
- 1948-50 Reginald H. Robson.
- 1950- Richenda C. Scott.
**Consulting Librarian:**
- 1921-33 Norman Penney |
A physicochemical framework for interpreting the biological calcification response to CO$_2$-induced ocean acidification
Justin B. Ries*
Department of Marine Sciences, University of North Carolina—Chapel Hill, 4202G Venable Hall, Campus Box 3300, Chapel Hill, NC 27599, USA
Department of Geology and Geophysics, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA
Received 16 June 2010; accepted in revised form 22 April 2011; available online 1 May 2011
Abstract
A generalized physicochemical model of the response of marine organisms’ calcifying fluids to CO$_2$-induced ocean acidification is proposed. The model is based upon the hypothesis that some marine calcifiers induce calcification by elevating pH, and thus $\Omega_A$, of their calcifying fluid by removing protons (H$^+$). The model is explored through two end-member scenarios: one in which a fixed number of H$^+$ is removed from the calcifying fluid, regardless of atmospheric $pCO_2$, and another in which a fixed external–internal H$^+$ ratio ([H$^+)_E]/[H^+]_I$) is maintained. The model is able to generate the full range of calcification response patterns observed in prior ocean acidification experiments and is consistent with the assertion that organisms’ calcification response to ocean acidification is more negative for marine calcifiers that exert weaker control over their calcifying fluid pH. The model is empirically evaluated for the temperate scleractinian coral *Astrangia poculata* with in situ pH microelectrode measurements of the coral’s calcifying fluid under control and acidified conditions. These measurements reveal that (1) the pH of the coral’s calcifying fluid is substantially elevated relative to its external seawater under both control and acidified conditions, (2) the coral’s [H$^+)_E]/[H^+]_I$ is approximately the same under control and acidified conditions, and (3) the coral removes fewer H$^+$ from its calcifying fluid under acidified conditions than under control conditions. Thus, the carbonate system dynamics of *A. poculata*’s calcifying fluid appear to be most consistent with the fixed [H$^+)_E]/[H^+]_I$ end-member scenario. Similar microelectrode experiments performed on additional taxa are required to assess the model’s general applicability.
© 2011 Elsevier Ltd. All rights reserved.
1. INTRODUCTION
Recent experiments have revealed that marine calcifiers exhibit a broad spectrum of calcification responses to CO$_2$-induced ocean acidification, including negative, threshold-negative, neutral, threshold-positive, parabolic, and positive (e.g., Fabry et al., 2008; Iglesias-Rodriguez et al., 2008; Wood et al., 2008; Doney et al., 2009; Ries et al., 2009; Rodolfo-Metalpa et al., 2010; Fig. 1). There is growing evidence that one of the primary factors controlling the response of individual species to CO$_2$-induced ocean acidification is their control over carbonate chemistry at their site of calcification (e.g., Cohen and Holcomb, 2009; Ries et al., 2009). Here, two general end-member scenarios of calcification site chemistry (fixed H$^+$-removal vs. fixed external–internal H$^+$ ratio) are presented and explored under four atmospheric $pCO_2$ scenarios: 400, 600, 900, and 2850 ppm. The applicability of these model scenarios to the calcifying fluid of the temperate coral *Astrangia poculata* was empirically assessed by measuring the pH of the coral’s calcifying fluid with H$^+$-sensitive microelectrodes under control (non-acidified) and acidified conditions.
2. BIOLOGICAL CONTROL OVER CALCIFICATION SITE CHEMISTRY
Many calcifying marine organisms, including scleractinian corals (Al-Horani et al., 2003; Cohen and
chemistry at the site of calcification can be summarized by the following equilibrium reactions:
\[
H_2CO_3 \leftrightarrow HCO_3^- + H^+ \quad \text{and} \quad (1)
\]
\[
HCO_3^- \leftrightarrow H^+ + CO_3^{2-} \quad (2)
\]
which are, respectively, governed by the following stoichiometric dissociation constants:
\[
K'_1 = [HCO_3^-][H^+]/[H_2CO_3] \quad \text{and} \quad (3)
\]
\[
K'_2 = [CO_3^{2-}][H^+]/[HCO_3^-]. \quad (4)
\]
Thus, reducing $[H^+]$ at the site of calcification shifts the carbonic acid system towards elevated $[CO_3^{2-}]$, thereby increasing CaCO$_3$ saturation state ($\Omega_{CaCO_3}$):
\[
\Omega_{CaCO_3} = [Ca^{2+}][CO_3^{2-}]/K'_{sp}; \quad (5)
\]
where $K'_{sp}$ is the stoichiometric solubility product of the appropriate CaCO$_3$ polymorph (e.g., calcite, aragonite, vaterite).
Various mechanisms have been proposed for elevating pH at the site of calcification, including conventional H$^+$-channeling (McConnaughey and Falk, 1991), Ca$^{2+}$–H$^+$ exchanging ATPase (McConnaughey and Falk, 1991; McConnaughey and Whelan, 1997; Cohen and McConnaughey, 2003), light-induced H$^+$-pumping (De Beer and Larkum, 2001), transcellular symporter and co-transporter H$^+$-solute shuttling (McConnaughey and Whelan, 1997), cellular extrusion of hydroxyl ions (OH$^-$) into the calcifying medium, and CO$_2$-consumption via photosynthesis (Borowitzka and Larkum, 1976).
The decrease in seawater pH that will accompany the forecasted rise in anthropogenic atmospheric $pCO_2$ will reduce seawater $[CO_3^{2-}]$, which has been shown to inhibit biological deposition of CaCO$_3$, or even promote its dissolution (cf. Langdon, 2000; Kleypas et al., 2006; Fabry et al., 2008; Doney et al., 2009; Ries et al., 2009). However, if seawater is the source of an organism’s calcifying fluid (e.g., Gaetani and Cohen, 2006), then the concentration of dissolved inorganic carbon (DIC) in this fluid will increase as atmospheric $pCO_2$ increases. Organisms able to maintain their calcifying fluid at sufficiently elevated pH, despite reduced external pH, will convert much of this increased DIC, occurring primarily as HCO$_3^-$, to CO$_3^{2-}$. Thus, when such organisms are exposed to elevated $pCO_2$, they may experience a $[CO_3^{2-}]$ in their calcifying fluid that is only slightly less than, or possibly equal to or greater than that attained under present-day atmospheric $pCO_2$—depending upon the strength of their H$^+$-regulating mechanism (Ries et al., 2009).
The removal of H$^+$ from an organism’s calcifying fluid requires energy. Two factors that influence the amount of energy required to regulate calcification site pH are the quantity of H$^+$ removed from a given volume of the calcifying fluid and the H$^+$-gradient across the membrane(s) that bounds the calcifying fluid (Goldman, 1943; Hodgkin and Katz, 1949; Kitasato, 1968; Saddler, 1970; Richards and Hope, 1974; McConnaughey and Falk, 1991). The energy required to maintain a H$^+$-gradient across a membrane, known as the Nernst potential ($E$), is defined as:
\[
E = \Delta G/(-nF); \quad (6)
\]
where \( n \) is the valence charge of \( \text{H}^+ \) (1), \( F \) is the Faraday constant (96485.3 C mol\(^{-1}\)), and \( \Delta G \) is the Gibbs free energy of the system. The Gibbs free energy of the system can be defined as:
\[
\Delta G = (-RT) \times \ln([\text{H}^+]_E/[\text{H}^+]_I)
\]
(7)
where \( R \) is the universal gas constant (8.31451 J K\(^{-1}\) mol\(^{-1}\)), \( T \) is absolute temperature (298.16 at 25 °C), and \([ \text{H}^+]_E\) and \([ \text{H}^+]_I\) are \( \text{H}^+ \) concentrations of the external seawater and of the organism’s calcifying fluid, respectively.
The Nernst potential (\( E \)) can therefore be defined as:
\[
E = (RT)/(nF) \times \ln([\text{H}^+]_E/[\text{H}^+]_I)
\]
(8)
Because \( R \), \( T \), \( n \), and \( F \) are constants in the described \( \text{H}^+ \)-membrane system, the magnitude of the Nernst potential, or the energetic cost of maintaining a \( \text{H}^+ \)-gradient between external seawater and an organism’s membrane-bound calcifying fluid, is proportional to the ratio of the external and internal \( \text{H}^+ \) concentrations ([\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I]).
### 3. MODELING CARBONATE SYSTEM CHEMISTRY AT THE SITE OF CALCIFICATION
To explore the effects of \( \text{H}^+ \)-pumping on the carbonate chemistry of calcifying fluids under various levels of atmospheric \( pCO_2 \), two end-member scenarios of the calcifying fluid proton-pumping model are presented: one that assumes a fixed number of \( \text{H}^+ \) are removed from the calcifying fluid (i.e., variable [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I]) under each atmospheric \( pCO_2 \) level and one that assumes only enough \( \text{H}^+ \) are removed from the calcifying fluid to maintain a constant [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I] (i.e., variable \( \text{H}^+ \)-removal) under each atmospheric \( pCO_2 \) level. It should be noted that these two conditions were not selected for their mechanistic accuracy, but because they represent end-member scenarios for investigating the energetic requirements and carbonate system dynamics of calcification via \( \text{H}^+ \)-removal from the calcifying fluid.
Carbonate system parameters of the model were calculated for the site of calcification and for the external seawater with the program CO2SYS (Lewis and Wallace, 1998) using Roy et al. (1993) values for the stoichiometric carbonic acid constants \( K'_1 \) and \( K'_2 \), the Mucci (1983) value for the stoichiometric aragonite solubility product, \( T = 25 \) °C, \( S = 32 \), \( P = 1.015 \) atm, TA = 2000, 5545 \( \mu \)mol/kg-SW, DIC = 1756, 1824, 1884, 2029 \( \mu \)mol/kg-SW, and \( pCO_2 = 400, 600, 900, \) and 2850 ppm (see Table 1). Total alkalinity at the site of calcification was calculated as the external seawater alkalinity (2000 \( \mu \)mol/kg-SW) plus the number of \( \mu \)mol of \( \text{H}^+ \) removed from 1 kg of the organism’s calcifying fluid, as defined by Scenario 1 (fixed \( \text{H}^+ \)-removal) or Scenario 2 (fixed [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I])—since removing \( \text{H}^+ \) from the organism’s calcifying fluid will, by definition, cause an equal-magnitude increase in its alkalinity. In the absence of DIC measurements of organisms’ calcifying fluid, it is assumed in the model calculations to be equivalent to that of the external seawater. Nevertheless, in some organisms it is possible that additional DIC is respired and transported across cell membranes into the calcifying fluid or, alternatively, removed from the calcifying fluid of plants, algae and/or organisms harboring photosynthesizing symbionts (e.g., zooxanthellate corals) via photosynthesis.
#### 3.1. Scenario 1: Fixed \( \text{H}^+ \)-removal with variable [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I]
Scenario 1 (Fig. 2A and Table 1) demonstrates how [\( \text{CO}_3^{2-} \)] at the calcification site varies with atmospheric \( pCO_2 \) (400, 600, 900, 2850 ppm) for three hypothetical organisms that differ only in the number of \( \text{H}^+ \) that they are able to remove from a given volume of their calcifying fluid [external seawater (white diamonds); weak \( \text{H}^+ \)-pump (black circles) = 600 \( \mu \)mol \( \text{H}^+ \)/kg-SW; moderate \( \text{H}^+ \)-pump (gray circles) = 2000 \( \mu \)mol \( \text{H}^+ \)/kg-SW; and strong \( \text{H}^+ \)-pump (white circles) = 2900 \( \mu \)mol \( \text{H}^+ \)/kg-SW]. Assuming that an organism removes a fixed number of \( \text{H}^+ \) from a given volume of its calcifying fluid under the four \( pCO_2 \) levels, [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I] (and therefore the energetic cost of removing each \( \text{H}^+ \)) will increase with increasing atmospheric \( pCO_2 \) (Table 1).
#### 3.2. Scenario 2: Fixed [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I] with variable \( \text{H}^+ \)-removal
Scenario 2 (Fig. 2B and Table 1) demonstrates how [\( \text{CO}_3^{2-} \)] at the calcification site varies with atmospheric \( pCO_2 \) (400, 600, 900, 2850 ppm) for three hypothetical organisms that differ only in the \( \text{H}^+ \)-gradient ([\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I]) that they maintain across the membrane(s) that bounds their calcifying fluid [external seawater (white diamonds); weak [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I] (black circles) = 7:1; moderate [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I] (gray circles) = 45:1; strong [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I] (white circles) = 300:1]. Assuming that an organism maintains a fixed [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I], the number of \( \text{H}^+ \) removed from a given volume of its calcifying fluid (and therefore the energetic cost of removing \( \text{H}^+ \)) will decrease with increasing atmospheric \( pCO_2 \) (Table 1).
#### 3.3. Comparison of model scenarios
Both scenarios demonstrate that the response of calcification site [\( \text{CO}_3^{2-} \)] to elevated \( pCO_2 \) becomes more negative as the strength of the \( \text{H}^+ \)-pump decreases (i.e., response is more negative for organisms that remove fewer \( \text{H}^+ \) from a given volume of calcifying fluid and/or maintain a lower [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I]; Fig. 2 and Table 1). Both scenarios can also generate the full range of calcification response patterns observed in prior ocean acidification experiments (Figs. 1 and 2; e.g., Fabry et al., 2008; Doney et al., 2009; Ries et al., 2009). However, the energetic requirements of the two model scenarios are different. Scenario 1 (Fig. 2A) defines three hypothetical organisms that differ only in the number of \( \text{H}^+ \) that they remove from their calcifying fluid. The energetic requirements of \( \text{H}^+ \)-removal will increase with atmospheric \( pCO_2 \) for a given type of organism (i.e., weak, moderate, or strong \( \text{H}^+ \)-pump) because the given organism’s [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I] will increase with atmospheric \( pCO_2 \) (Table 1), even though a fixed quantity of \( \text{H}^+ \) is removed. For example, the [\( \text{H}^+ \)_E]/[\( \text{H}^+ \)_I] for an organism with a “moderate” \( \text{H}^+ \)-pump (2000 \( \mu \)mol \( \text{H}^+ \)/kg-SW) is approximately 140% greater at a \( pCO_2 \) of 2850 ppm than at a \( pCO_2 \) of 400 ppm (Table 1).
Table 1
Carbonate system constraints and energetic requirements of the two end-member calcifying fluid models.
| $pCO_2^a$ | $[H^+]-removed^b$ | $[H^+]_f/[H^+]_h^d$ | $\Delta$-Energy$^e$ (%) | TA$^f$ | DIC$^g$ | pH | $[HCO_3^-]^h$ | $[CO_3^{2-}]^i$ | $\Omega_A^j$ |
|-----------|-------------------|---------------------|--------------------------|--------|---------|----|-------------|-------------|-------|
| **Fixed $H^+$-pump** | | | | | | | | | |
| External SW$^c$ | 400 | — | — | — | 2000 | 1756 | 8.14 | 1570 | 175 | 2.8 |
| | 600 | — | — | — | 2000 | 1824 | 7.99 | 1674 | 133 | 2.1 |
| | 900 | — | — | — | 2000 | 1884 | 7.84 | 1760 | 98 | 1.6 |
| | 2850 | — | — | — | 2000 | 2029 | 7.37 | 1911 | 36 | 0.6 |
| Weak $H^+$-pump | 400 | 600 | 4.8 | 0 | 2600 | 1756 | 8.82 | 1142 | 613 | 9.9 |
| | 600 | 600 | 5.7 | 18 | 2600 | 1824 | 8.75 | 1256 | 566 | 9.1 |
| | 900 | 600 | 7.0 | 45 | 2600 | 1884 | 8.68 | 1356 | 525 | 8.5 |
| | 2850 | 600 | 13.9 | 188 | 2600 | 2029 | 8.52 | 1601 | 423 | 6.8 |
| Moderate $H^+$-pump | 400 | 2000 | 62 | 0 | 4000 | 1756 | 9.93 | 222 | 1534 | 24.8 |
| | 600 | 2000 | 69 | 12 | 4000 | 1824 | 9.83 | 281 | 1543 | 24.9 |
| | 900 | 2000 | 81 | 31 | 4000 | 1884 | 9.75 | 343 | 1541 | 24.9 |
| | 2850 | 2000 | 148 | 140 | 4000 | 2029 | 9.54 | 530 | 1499 | 24.2 |
| Strong $H^+$-pump | 400 | 2900 | 191 | 0 | 4900 | 1756 | 10.42 | 79 | 1678 | 27.1 |
| | 600 | 2900 | 239 | 25 | 4900 | 1824 | 10.37 | 91 | 1733 | 28.0 |
| | 900 | 2900 | 304 | 59 | 4900 | 1884 | 10.32 | 105 | 1779 | 28.7 |
| | 2850 | 2900 | 643 | 236 | 4900 | 2029 | 10.18 | 153 | 1876 | 30.3 |
| **Fixed $[H^+]_f/[H^+]_h$** | | | | | | | | | |
| External SW$^c$ | 400 | — | — | — | 2000 | 1756 | 8.14 | 1570 | 175 | 2.8 |
| | 600 | — | — | — | 2000 | 1824 | 7.99 | 1674 | 133 | 2.1 |
| | 900 | — | — | — | 2000 | 1884 | 7.84 | 1760 | 98 | 1.6 |
| | 2850 | — | — | — | 2000 | 2029 | 7.37 | 1911 | 36 | 0.6 |
| Weak $[H^+]_f/[H^+]_h$ | 400 | 805 | 7 | 0 | 2804 | 1756 | 8.99 | 986 | 770 | 12.4 |
| | 600 | 709 | 7 | −12 | 2709 | 1824 | 8.84 | 1173 | 650 | 10.5 |
| | 900 | 601 | 7 | −25 | 2602 | 1884 | 8.68 | 1355 | 526 | 8.5 |
| | 2850 | 348 | 7 | −57 | 2348 | 2029 | 8.22 | 1781 | 237 | 3.8 |
| Moderate $[H^+]_f/[H^+]_h$ | 400 | 1825 | 45 | 0 | 3825 | 1756 | 9.79 | 292 | 1464 | 23.6 |
| | 600 | 1768 | 45 | −3 | 3768 | 1824 | 9.65 | 400 | 1474 | 23.0 |
| | 900 | 1678 | 45 | −8 | 3678 | 1884 | 9.49 | 539 | 1345 | 21.7 |
| | 2850 | 1256 | 45 | −31 | 3256 | 2029 | 9.03 | 1093 | 936 | 15.1 |
| Strong $[H^+]_f/[H^+]_h$ | 400 | 3545 | 300 | 0 | 5545 | 1756 | 10.62 | 51 | 1705 | 27.5 |
| | 600 | 3163 | 300 | −11 | 5163 | 1824 | 10.47 | 74 | 1751 | 28.3 |
| | 900 | 2883 | 300 | −19 | 4883 | 1884 | 10.32 | 107 | 1777 | 28.7 |
| | 2850 | 2400 | 300 | −32 | 4400 | 2029 | 9.85 | 302 | 1727 | 27.9 |
---
$a$ “$pCO_2$” = CO$_2$ partial pressure of air in equilibrium with external seawater (ppm)
$b$ “[H$^+$]-removed” = [H$^+$]-removed from calcifying fluid (µmol/kg-SW) by proton pump
$c$ “SW” = seawater
$d$ “[H$^+$]$_f$/[H$^+$]$_h$” = ratio of external seawater [H$^+$] to calcifying fluid [H$^+$]
$e$ “$\Delta$-Energy” = change in energy required to maintain given level of proton regulation relative to 400 ppm $pCO_2$ scenario, assuming that energy required varies linearly with [H$^+$]-removed or [H$^+$]$_f$/[H$^+$]$_h$
$f$ “TA” = total alkalinity of fluid (µmol/kg-SW), calculated as external seawater TA plus [H$^+$]-removed from calcifying fluid (assuming that seawater is the ultimate source of calcifying fluid—see text for discussion)
$g$ “DIC” = dissolved inorganic carbon of fluid (µmol/kg-SW)—calcifying fluid DIC is treated as equivalent to seawater DIC (assuming that seawater is the ultimate source of calcifying fluid—see text for discussion)
$h$ “[HCO$_3^-$]$^h$” = bicarbonate ion concentration of fluid (µmol/kg-SW)
$i$ “[CO$_3^{2-}$]$^i$” = carbonate ion concentration of fluid (µmol/kg-SW)
$j$ “$\Omega_A$” = saturation state of fluid with respect to aragonite
Fig. 2. Generalized proton-pumping model of marine organisms’ calcifying fluids showing relationship between atmospheric $pCO_2$ (400, 600, 900, 2850 ppm), calcification site pH ($x$-axis), $[CO_3^{2-}]$ (circles), and $[HCO_3^-]$ for three hypothetical organisms differing only in the strength of their H$^+$-regulating mechanism (“INITIAL SEAWATER” = white diamonds; “WEAK” = black circles, “MODERATE” = gray circles, “STRONG” = white circles; Table 1). (A) Scenario 1 assumes that an organism removes a fixed number of H$^+$ from a given volume of its calcifying fluid (variable [H$^+)_e$/H$^+_i$), regardless of atmospheric $pCO_2$ (“WEAK” H$^+$-pump = 600 μmol H$^+$/kg-SW; “MODERATE” H$^+$-pump = 2000 μmol-H$^+$/kg-SW; “STRONG” H$^+$-pump = 2900 μmol H$^+$/kg-SW). (B) Scenario 2 assumes than an organism maintains a fixed [H$^+)_e$/H$^+_i$] (variable number of H$^+$ removed from a given volume of calcifying fluid) regardless of atmospheric $pCO_2$ (“WEAK” [H$^+)_e$/H$^+_i$] = 7:1; “MODERATE” [H$^+)_e$/H$^+_i$] = 45:1; “STRONG” [H$^+)_e$/H$^+_i$] = 300:1). Both model scenarios are able to predict the full range of calcification responses observed in prior ocean acidification experiments (Fig. 1) and demonstrate that the calcification response to CO$_2$-induced ocean acidification should be more negative for organisms that exert weaker control over their calcifying fluid pH (Ries et al., 2009). Microelectrode measurements of the calcifying fluid pH of the temperate coral *Astrangia poculata* (Table 2) suggest that the physicochemical dynamics of its calcifying fluid are most consistent with the model scenario that assumes a fixed [H$^+)_e$/H$^+_i$] (B).
The increased energetic cost of pumping the same number of H$^+$ against a steeper H$^+$-gradient could cause energy to be diverted from other critical life processes under elevated $pCO_2$, such as tissue growth and/or reproduction (e.g., Wood et al., 2008).
Conversely, the energetic cost of maintaining a fixed [H$^+)_e$/H$^+_i$] (Scenario 2; Fig. 2B) should decrease with increasing atmospheric $pCO_2$ because the organism (i.e., weak, moderate, or strong [H$^+)_e$/H$^+_i$]) would be required to pump fewer H$^+$ from a given volume of calcifying fluid in order to maintain a fixed [H$^+)_e$/H$^+_i$] under increasing atmospheric $pCO_2$ (Table 1). For example, an organism that maintains a “moderate” [H$^+)_e$/H$^+_i$] (45:1) would be required to remove approximately 31% fewer H$^+$ at a $pCO_2$ of 2850 ppm than at a $pCO_2$ of 400 ppm (Table 1). It should be noted that for approximately equivalent ranges of calcification site pH (e.g., black circles in Scenarios 1 and 2); Fig. 2 the response of $[CO_3^{2-}]$ at the site of calcification to elevated $pCO_2$ is generally more negative for hypothetical organisms that maintain a fixed [H$^+)_e$/H$^+_i$] (Scenario 2) than for those that remove a fixed number of H$^+$ from a given volume of their calcifying fluid (Scenario 1). However, as described above, the energetic demands of removing a fixed number of H$^+$ from a given volume of calcifying fluid (Scenario 1) will increase with increasing atmospheric $pCO_2$, while the energetic demands of maintaining a fixed [H$^+)_e$/H$^+_i$] (Scenario 2) will decrease with increasing atmospheric $pCO_2$.
4. MODEL EXPLORATION WITH THE TEMPERATE CORAL *ASTRANGIA POCULATA*
The calcifying fluid pH of the temperate coral *A. poculata* was measured with H$^+$-sensitive microelectrodes under control and acidified conditions to investigate the applicability of the proposed model scenarios to calcification site pH-regulation within a temperate scleractinian coral.
This section should be prefaced with a brief overview of the two general models of coral calcification: the physicochemical model and the biological model (which are by
no means mutually exclusive). The physicochemical model of coral calcification (e.g., Constantz, 1986) asserts that corals accrete their skeletons by elevating the $\Omega_A$ of a calcifying fluid that is initially similar in composition to that of their external seawater, potentially via Ca$^{2+}$-ATPase-driven exchange of Ca$^{2+}$ and H$^+$ ions across the coral tissue (e.g., Cohen and McConnaughey, 2003 and references therein; Al-Horani et al., 2003; McConnaughey, 2003; Cohen and Holcomb, 2009). The assertion in this model that the calcifying fluid of corals is initially similar to that of their external seawater is supported by measurements of elemental ratios within coral skeletons that are consistent with a Rayleigh-type distillation of discrete batches of a seawater-type fluid (Gaetani and Cohen, 2006; Cohen and Gaetani, 2010) and by observations that a decrease in the molar Mg/Ca ratio of corals’ external seawater from the modern value of 5.2 [favoring the abiotic precipitation of aragonite and high-Mg calcite (calcite molar Mg/Ca > 0.04)] to 1.5 [favoring the abiotic precipitation of low-Mg calcite (calcite molar Mg/Ca < 0.04)] causes corals to accrete a portion of their skeleton as low-Mg calcite rather than aragonite (Ries et al., 2006; Ries, 2010a). The assertion in this model that corals accrete their skeletons by elevating the $\Omega_A$ of their calcifying fluid is supported by measurements of isotopic ratios within coral skeletons that are consistent with a calcifying fluid characterized by a strong pH gradient (e.g., Adkins et al., 2003) and by measurements of elevated pH (via pH microelectrodes; Al-Horani et al., 2003) and $\Omega_A$ (inferred from aragonite crystal aspect ratio; Cohen and Holcomb, 2009; Holcomb et al., 2009) of the coral calcifying fluid.
The biological model of coral calcification (cf. Cuif and Dauphin, 2005) asserts that calicoblastic cells and organic matrices exert strong control over the coral mineralization process. This model is supported by direct observations of the tissue-skeleton interface (e.g., Clode and Marshall, 2002; Cuif and Dauphin, 2005; Tambutté et al., 2007), by the existence of desmocytes in the calicoblastic epithelium of corals (e.g., Goldberg, 2001), by differential rates of calcification across the coral skeleton (e.g., Houlbreque et al., 2009), and by compositional variations in the coral skeleton that appear to be independent of thermo-chemical changes in the coral’s external seawater (Meibom et al., 2008).
Regardless of which model most accurately describes the mechanism(s) by which corals calcify, the coral’s ability to control pH and, thus, the carbonate chemistry at the site of calcification should play an important role in determining its calcification response to CO$_2$-induced ocean acidification.
### 4.1. Materials and methods
#### 4.1.1. Preparation of experimental seawaters
Two 10-L batches of 0.2 µm-filtered seawater were obtained from Vineyard Sound, Massachusetts. One 10-L batch was reserved as the control and the other 10-L batch was acidified with 2700 µL of a 1 M HCl solution. The experimental seawaters were maintained at $25 \pm 0.1^\circ$C (partial-immersion mercury-glass thermometer) using a 50-watt electric heater, salinity was $32.0 \pm 0.1$ (Autosal conductivity meter), total alkalinity of the normal and acidified seawaters was $2001 \pm 5$ and $1732 \pm 5$ µmol/kg-SW, respectively (Gran titration calibrated with certified reference material provided by Prof. Andrew Dickson of the Scripps Institute of Oceanography, La Jolla, USA), and seawater pH of the normal and acidified seawaters was $8.16 \pm 0.01$ and $7.47 \pm 0.01$, respectively ($Orion$ pH electrode/meter calibrated with certified NBS pH buffers of 4.00, 7.00, 10.00; see discussion below). Total DIC, [CO$_3^{2-}$], [HCO$_3^-$], dissolved CO$_2$ ([CO$_2$]$_{\text{aq}}$), $\Omega_A$, and $p\text{CO}_2$ of the gas in equilibrium with the experimental seawaters were calculated for the control and acidified treatments (Table 2) from measured values of temperature, salinity, total alkalinity, and pH using the program CO$_2$SYS (Lewis and Wallace, 1998) with Roy et al. (1993) values for the stoichiometric carbonic acid constants $K_1^{-1}$ and $K_2^{-1}$, the Mucci (1983) value for the stoichiometric aragonite solubility product, and pressure of 1.015 atm (Table 2).
Ideally, ocean acidification experiments should be performed by varying atmospheric $p\text{CO}_2$, rather than through acid addition. However, the microelectrode approach employed in the present study was not compatible with the bubbling of CO$_2$-enriched gas or with the flux of seawater previously bubbled with CO$_2$-enriched gas into the vessel containing the coral, as the resulting turbulence destabilized the microelectrode. Thus, acidification was achieved by adding small aliquots of HCl to the seawater. It should be noted that the addition of HCl to lower the pH of the experimental seawater will also lower its total alkalinity and DIC (as the seawater was open to the atmosphere; Table 2), whereas lowering its pH via CO$_2$ bubbling will raise its DIC and not alter its total alkalinity. Nevertheless, the carbonate system parameters calculated for the experimental seawaters account for the effects of acid addition instead of CO$_2$ bubbling, as they were based upon measured pH and total alkalinity, which would have both been affected by (thereby accounting for) the acid addition.
#### 4.1.2. Construction of pH microelectrodes
One-mm-diameter silanized glass capillary tubes were heated with a heating coil and machine-pulled to produce a tip diameter of approximately 1 µm. The 1-µm tip was sealed with a 3-mm membrane column of Hydrogen Ionophore II—Cocktail A (Sigma–Aldrich). The ionophore membrane was then backfilled with a 20–30-mm column of 0.2-µm-filtered 1 M KCl solution. The 1 M KCl electrolyte was connected to the reference with a Ag/AgCl wire. The microelectrodes were then mounted onto a micromanipulator that enabled their precise manipulation in three dimensions. The microelectrodes were calibrated with NBS pH buffers of 6.00, 7.00, 8.00, 9.00, 10.00, and 11.00 before its deployment and immediately after its removal from the calcifying fluid.
The use of NBS buffers to calibrate an electrode used to measure the pH of seawater (or calcifying fluid) generates an unknown and systematic error arising from the difference in liquid junction potentials between the two media. Comparison of pH calculated from the measured total alkalinity and DIC of a seawater sample with its pH measured using the NBS-calibrated microelectrodes suggests that this error was less than approximately 0.05 pH units,
Table 2
Carbonate system parameters of seawater and the calcifying fluid of the coral *Astrangia poculata* determined via pH microelectrode.
| Non-acidified seawater (control) | Sal<sup>a</sup> | T<sup>d</sup> | pH | TA<sup>c</sup> | DIC<sup>f</sup> | pCO<sub>2</sub><sup>g</sup> | [HCO<sub>3</sub><sup>−</sup>]<sup>i</sup> | [CO<sub>3</sub><sup>2−</sup>]<sup>j</sup> | [CO<sub>2</sub>]<sub>aq</sub><sup>k</sup> | Ω<sub>A</sub><sup>k</sup> | [H<sup>+</sup>]<sup>l</sup> | [H<sup>+</sup>]-removed<sup>m</sup> | [H<sup>+</sup>]<sub>cf</sub>/[H<sup>+</sup>]<sub>sw</sub><sup>n</sup> |
|---------------------------------|----------------|-------------|------|----------------|----------------|----------------|----------------|----------------|----------------|----------|----------------|----------------|----------------|
| Trial 1 | SW<sup>a</sup> | 32 | 25 | 8.16 | 2001 | 1745 | 375 | 1552 | 183 | 11 | 3.0 | 6.8 × 10<sup>−3</sup> | 2253 | 90.9 |
| | CF<sup>b</sup> | 32 | 25 | 10.12 | 4254 | 1745 | 375 | 149 | 1596 | 0 | 25.8 | 7.5 × 10<sup>−5</sup> | | |
| Trial 2 | SW | 32 | 25 | 8.18 | 2001 | 1739 | 362 | 1542 | 187 | 10 | 3.0 | 6.7 × 10<sup>−3</sup> | 2359 | 103.8 |
| | CF | 32 | 25 | 10.19 | 4360 | 1739 | 362 | 128 | 1611 | 0 | 26.0 | 6.4 × 10<sup>−5</sup> | | |
| Trial 3 | SW | 32 | 25 | 8.15 | 2001 | 1755 | 396 | 1567 | 177 | 11 | 2.9 | 7.2 × 10<sup>−3</sup> | 2062 | 68.5 |
| | CF | 32 | 25 | 9.98 | 4063 | 1755 | 396 | 201 | 1554 | 0 | 25.1 | 1.0 × 10<sup>−4</sup> | | |
| AVG (±SD) | SW | 32 ± 0 | 25 ± 0 | 8.16 ± 0.02 | 2001 ± 2 | 1747 ± 8 | 377 ± 17 | 1554 ± 13 | 182 ± 5 | 11 ± 1 | 2.9 ± 0.1 | 6.9(±0.3) × 10<sup>−3</sup> | 2225 ± 150 | 87.7 ± 17.8 |
| | CF | 32 ± 0 | 25 ± 0 | 10.10 ± 0.11 | 4226 ± 150 | 1747 ± 8 | 377 ± 17 | 160 ± 38 | 1587 ± 30 | 0 ± 0 | 25.6 ± 0.5 | 8.1(±2.1) × 10<sup>−3</sup> | | |
| Acidified seawater | Sal<sup>a</sup> | T<sup>d</sup> | pH | TA<sup>c</sup> | DIC<sup>f</sup> | pCO<sub>2</sub><sup>g</sup> | [HCO<sub>3</sub><sup>−</sup>]<sup>i</sup> | [CO<sub>3</sub><sup>2−</sup>]<sup>j</sup> | [CO<sub>2</sub>]<sub>aq</sub><sup>k</sup> | Ω<sub>A</sub><sup>k</sup> | [H<sup>+</sup>]<sup>l</sup> | [H<sup>+</sup>]-removed<sup>m</sup> | [H<sup>+</sup>]<sub>cf</sub>/[H<sup>+</sup>]<sub>sw</sub><sup>n</sup> |
|---------------------------------|----------------|-------------|------|----------------|----------------|----------------|----------------|----------------|----------------|----------|----------------|----------------|----------------|
| Trial 1 | SW | 32 | 25 | 7.47 | 1732 | 1729 | 1943 | 1634 | 39 | 56 | 0.6 | 3.4 × 10<sup>−2</sup> | 1603 | 90.4 |
| | CF | 32 | 25 | 9.43 | 3335 | 1729 | 1943 | 547 | 1182 | 0 | 19.1 | 3.7 × 10<sup>−4</sup> | | |
| Trial 2 | SW | 32 | 25 | 7.48 | 1732 | 1726 | 1902 | 1632 | 40 | 55 | 0.6 | 3.3 × 10<sup>−2</sup> | 1619 | 92.0 |
| | CF | 32 | 25 | 9.44 | 3351 | 1726 | 1902 | 532 | 1194 | 0 | 19.3 | 3.6 × 10<sup>−4</sup> | | |
| Trial 3 | SW | 32 | 25 | 7.46 | 1732 | 1731 | 1979 | 1635 | 38 | 57 | 0.6 | 3.4 × 10<sup>−2</sup> | 1487 | 74.0 |
| | CF | 32 | 25 | 9.33 | 3219 | 1731 | 1979 | 632 | 1099 | 0 | 17.7 | 4.6 × 10<sup>−4</sup> | | |
| AVG (±SD) | SW | 32 ± 0 | 25 ± 0 | 7.47 ± 0.01 | 1732 ± 2 | 1729 ± 2 | 1941 ± 39 | 1634 ± 2 | 39 ± 1 | 56 ± 1 | 0.6 ± 0.0 | 3.4(±0.1) × 10<sup>−2</sup> | 1570 ± 72 | 85.5 ± 9.9 |
<sup>f</sup>–<sup>k</sup>DIC, pCO<sub>2</sub>, [HCO<sub>3</sub><sup>−</sup>], [CO<sub>3</sub><sup>2−</sup>], [CO<sub>2</sub>]<sub>aq</sub>, and Ω<sub>A</sub> of SW were calculated from measured sal, T, pH, and TA; TA, [HCO<sub>3</sub><sup>−</sup>], [CO<sub>3</sub><sup>2−</sup>], [CO<sub>2</sub>]<sub>aq</sub>, and Ω<sub>A</sub> of CF were calculated from pH of the CF measured via microelectrode and from measured sal and T and calculated DIC of the external SW, each assumed to be equivalent for the CF on the premise that SW is ultimate source of CF (see text for discussion)
<sup>a</sup> “SW” = seawater
<sup>b</sup> “CF” = calcifying fluid
<sup>c</sup> “Sal” = salinity
<sup>d</sup> “T” = temperature (°C)
<sup>e</sup> “TA” = total alkalinity (µmol/kg-SW)
<sup>f</sup> “DIC” = dissolved inorganic carbon (µmol/kg-SW).
<sup>g</sup> “pCO<sub>2</sub>” = partial pressure of CO<sub>2</sub> of air in equilibrium with fluid (ppm); pCO<sub>2</sub> of SW and CF are assumed to be equivalent on the premise that SW is the ultimate source of CF (see text for discussion)
<sup>h</sup> “[HCO<sub>3</sub><sup>−</sup>]<sup>i</sup>” = bicarbonate ion concentration of fluid (µmol/kg-SW)
<sup>i</sup> “[CO<sub>3</sub><sup>2−</sup>]<sup>j</sup>” = carbonate ion concentration of fluid (µmol/kg-SW)
<sup>j</sup> “[CO<sub>2</sub>]<sub>aq</sub><sup>k</sup>” = CO<sub>2</sub> concentration of fluid (µmol/kg-SW)
<sup>k</sup> “Ω<sub>A</sub>” = saturation state of fluid with respect to aragonite
<sup>l</sup> “[H<sup>+</sup>]<sup>l</sup>” = proton concentration of fluid (µmol/kg-SW)
<sup>m</sup> “[H<sup>+</sup>]-removed” = [H<sup>+</sup>]-removed from calcifying fluid (µmol/kg-SW), calculated from difference between total alkalinity of SW and CF after H<sup>+</sup>-removal (assuming composition of CF before H<sup>+</sup>-removal is equivalent to that of external SW)
<sup>n</sup> “[H<sup>+</sup>]<sub>cf</sub>/[H<sup>+</sup>]<sub>sw</sub>” = ratio of external seawater [H<sup>+</sup>] to calcifying fluid [H<sup>+</sup>]
4059
which is less than the standard deviation in calcifying fluid pH amongst the three trials for both the acidified and control treatments. Thus, the error arising from the use of NBS buffers should not materially affect the reported results nor the interpretation of these results.
4.1.3. Measurement of calcifying fluid pH
Six ca. 4 cm-diameter coral colonies of *A. poculata* were harvested from Great Harbor of Vineyard Sound, Massachusetts, and maintained in an aquarium filled with Vineyard Sound seawater (25 °C) for approximately 7 days. One colony was immersed in a polypropylene container filled with 500 mL of the control seawater and allowed to acclimate for 1 h. The container of seawater and the coral were placed beneath a dissecting stereomicroscope illuminated with a ring lamp. A micromanipulator was used to insert an empty microelectrode tube through the coral’s epithelial tissue layers at locations where the tissue layers were stretched taught across adjacent septal ridges of the coralite (Fig. 3). Once the empty microelectrode tube had contacted the coral skeleton, it was completely withdrawn from the coral, leaving behind an approximately 2–4 μm wide hole through the coral’s tissue layers. A calibrated H\(^+\)-selective microelectrode of similar dimensions to the empty microelectrode tube was then slowly inserted with the micromanipulator back through this hole to the point of contact with the skeleton, and then moved away from the skeleton by approximately 1–2 μm (Fig. 3; McConnaughey and Whelan, 1997; Al-Horani et al., 2003). Although the author is not aware of any direct measurement of the thickness of coral calcifying fluid, the observation that the aragonite needles that make up the coral skeleton are typically longer than 3 μm (e.g., Cohen and McConnaughey, 2003) suggests that the calcifying fluid is at least that thick. Thus, the positioning of the microelectrode tip within 1 μm of the coral skeleton surface, combined with the consistently elevated pH values that the microelectrode registered at this location, suggests that the tip of the microelectrode was indeed positioned within the coral’s calcifying fluid.
Approximately 10 min after insertion of the microelectrode, the coral’s epithelial tissue closed around the microelectrode, effectively sealing off its calcifying fluid from the external seawater. The microelectrodes were deployed until a minimum of 20 min of stable pH data were obtained (at 20-s intervals). Calcifying fluid pH (Table 2) was calculated from the microelectrode voltage over the measured interval, pursuant to the calibration of the microelectrode that was conducted immediately after its removal from the calcifying fluid. External seawater pH was also monitored at 20-s intervals throughout the experiment. This process was repeated for two additional coral colonies immersed in 500 mL aliquots of the control seawater and for three additional coral colonies immersed in 500 mL aliquots of the acidified seawater (Table 2).
4.1.4. Carbonate system calculations
Total alkalinity, $[\text{CO}_3^{2-}]$, $[\text{HCO}_3^-]$, dissolved CO$_2$ ($[\text{CO}_2_{\text{aq}}]$), and $\Omega_A$ of the corals’ calcifying fluid under the control and acidified treatments (Table 2) were calculated using the program CO$_2$SYS (as described above) from the microelectrode pH measurements and from temperature, salinity, and DIC of the external seawater, which the model assumes are transferred to the coral’s calcifying fluid (i.e., model assumes that seawater is the ultimate source of the calcifying fluid; e.g., Gaetani and Cohen, 2006). Again, it is possible that DIC is added to the calcifying fluid via respiration or, alternatively, removed via photosynthesis, causing calcifying fluid DIC to deviate from that of the external seawater—its putative source. Yet in the absence of direct measurements of calcifying fluid DIC, it is assumed to be equivalent to DIC of the external seawater. H$^+$-removal was calculated as the difference between the measured total alkalinity of the external seawater (assumed to be the initial total alkalinity of the calcifying fluid) and the total alkalinity calculated for the calcifying fluid (Table 2).
4.2. Carbonate system chemistry of the calcifying fluid of *A. poculata*
The pH and $\Omega_A$ of the coral’s calcifying fluid under the control treatment was $10.10 \pm 0.11$ and $25.6 \pm 0.5$, elevated by approximately 1.9 and 23 units relative to external seawater pH ($8.16 \pm 0.02$) and $\Omega_A$ ($2.9 \pm 0.1$; Table 2). These measurements reveal that under non-acidified conditions, *A. poculata* promotes calcification by maintaining a substantially elevated pH and $\Omega_A$ within its calcifying fluid. These results are generally consistent with a set of microelectrode measurements that show that pH of the calcifying fluid of the tropical scleractinian coral *Galaxea fascicularis* is substantially elevated (up to pH 9.3) relative to the coral’s external seawater pH of 8.2 (Al-Horani et al., 2003). These results are also generally consistent with previous studies that used the crystal aspect ratio of coral skeletal aragonite to deduce that the calcifying fluids of the temperate scleractinian coral *Astrangia poculata* (Cohen and Holcomb, 2009) and the tropical scleractinian corals *Diploria labyrinthiformis*, *Porites lutea*, *Porites solida* (Holcomb et al., 2009), and *Favia fragum* (Cohen and Holcomb, 2009) are substantially elevated relative to the corals’ external seawater.
The microelectrode measurements revealed that *A. poculata* maintains its calcifying fluid at elevated pH ($9.40 \pm 0.06$) and elevated $\Omega_A$ ($18.7 \pm 0.8$) under acidified conditions, as well, although at lower values than those maintained under the control treatment. Although there are no published microelectrode measurements of calcifying fluid pH for corals reared under acidified conditions to which these observations can be compared, a decrease in the calcifying fluid $\Omega_A$ of *F. fragum* to approximately 2–3 in seawater acidified to a $\Omega_A$ of 0.6 has been inferred from the crystal aspect ratio of the coral’s skeletal aragonite (Cohen and Holcomb, 2009). This is inconsistent with the observation that *A. poculata* is able to maintain a calcifying fluid $\Omega_A$ of approximately 19 as the $\Omega_A$ of its external seawater is reduced to 0.6. This discrepancy may reflect species-specific differences between *A. poculata* and *F. fragum*, differences in the health and/or nutritional status of the coral specimens that were investigated, and/or differences in the developmental stages of the corals studied (e.g., *A. poculata* specimens were colonial adults while *F. fragum* specimens were solitary new-recruits).
4.3. *Astrangia poculata* maintains a fixed external-internal H$^+$-ratio and removes fewer H$^+$ under elevated pCO$_2$
Critically, the $[\text{H}^+]_E/[\text{H}^+]_I$ maintained by *A. poculata* under the control treatment ($[\text{H}^+]_E/[\text{H}^+]_I = 88 \pm 18$) was not significantly ($p < 0.05$) different than that maintained under the acidified treatment ($[\text{H}^+]_E/[\text{H}^+]_I = 86 \pm 10$; Table 2). However, the corals removed significantly ($p < 0.05$) more H$^+$ from their calcifying fluid under the control treatment (2225 ± 150 μmol H$^+$/kg-SW removed) than under the acidified treatment (1570 ± 72 μmol H$^+$/kg-SW removed; Table 2). This observation is consistent with previous studies showing that the $[\text{H}^+]_E/[\text{H}^+]_I$ of marine algal cells remains nearly constant when seawater pH is reduced from 8 to 6 (Kitasato, 1968; Richards and Hope, 1974). This is surprising, nonetheless, because it reveals that *A. poculata* does not attempt to offset the reduction in seawater $[\text{CO}_3^{2-}]$ under acidified conditions by pumping more H$^+$ from its calcifying fluid. Instead, it appears to remove fewer H$^+$ from its calcifying fluid under acidified conditions, resulting in even lower $[\text{CO}_3^{2-}]$ at the site of calcification and, presumably, slower rates of calcification.
Although the end-member scenarios presented here were intended only to serve as a tool for exploring the proposed H$^+$-pumping model of biological calcification under conditions of elevated pCO$_2$, the microelectrode measurements suggest that H$^+$-regulation at the site of calcification of the temperate coral *A. poculata* is indeed consistent with the fixed $[\text{H}^+]_E/[\text{H}^+]_I$ scenario (Scenario 2). However, as is the case for most experimental perturbation studies, the corals were exposed relatively rapidly to the experimental conditions, and their responses measured over comparably short intervals of time (on the order of hours). Therefore, the measured changes in pH and, thus, carbonate chemistry at the site of calcification represent the corals’ plastic (i.e., phenotypic) responses to CO$_2$-induced ocean acidification, and provide little insight into their potential evolutionary (i.e., genotypic) responses. Furthermore, there was minimal water circulation in the incubation chambers, which may have influenced the corals’ ability to exchange ions between their calcifying fluid and their external seawater. Thus, future experiments of this type would ideally be conducted over longer durations—potentially across multiple generations of organisms—under experimental conditions that more closely mimic natural conditions.
4.4. Implications for the response of *A. poculata* to ocean acidification
The microelectrode measurements suggest that $\Omega_A$ at the site of calcification of *A. poculata* will decrease from approximately 26 to 19 as external seawater is acidified from a pH of 8.16 to 7.47, which corresponds to an increase
in imputed $pCO_2$ from 377 to 1941 ppm. If a reduction in calcifying fluid $\Omega_A$ causes a reduction in the rate of coral calcification, then this suggests that calcification rates would decline for *A. poculata* as seawater acidifies due to rising atmospheric $pCO_2$. This prediction is consistent with experimental work (Holcomb et al., 2010) showing that calcification rates of *A. poculata* decline by more than 50% as atmospheric $pCO_2$ increases from 380 ($\Omega_A = 3.0$) to 760 ppm ($\Omega_A = 1.8$). Yet Holcomb et al. (2010) also showed that calcification rates of *A. poculata* are unimpaired by this increase in atmospheric $pCO_2$ when reared under nutrient-enriched conditions. It is possible that the increased nutrient levels promoted higher rates of photosynthesis by the corals’ endosymbionts, which provided additional energy (as translocated photosynthate) for H$^+$-regulation at the site of calcification (i.e., shifted the coral’s response to the right in Fig. 2B), enabling it to maintain a higher [H$^+$]$_E$/[H$^+$]$_I$ and calcification site $\Omega_A$ under conditions of elevated nutrients (Holcomb et al., 2010). However, interpretation of Holcomb et al.’s (2010) results in the context of the present microelectrode experiment is limited by the fact that they investigated the calcification response of *A. poculata* to increased $pCO_2$ in seawater of similar alkalinites, while the present study investigated the coral’s response to increased $pCO_2$ in seawater of decreased alkalinites.
5. APPLICABILITY OF THE PROPOSED MODEL TO UNDERSTANDING AND PREDICTING MARINE CALCIFIERS’ RESPONSES TO CO$_2$-INDUCED OCEAN ACIDIFICATION
The two model scenarios of calcifying fluid chemistry proposed here, which assume that marine calcifiers induce calcification either by removing a fixed number of H$^+$ from their calcifying fluid (i.e., variable [H$^+$]$_E$/[H$^+$]$_I$; Scenario 1) or by maintaining a fixed [H$^+$]$_E$/[H$^+$]$_I$ (i.e., variable H$^+$-removal; Scenario 2), represent only end-member scenarios of a theoretical physicochemical framework for interpreting the response of marine calcifiers to CO$_2$-induced ocean acidification. They are not intended, in isolation, to describe or predict the response of any individual species. Nor should the ability of Scenario 1 and 2 to predict the disparate calcification responses observed in prior ocean acidification experiments (e.g., Fabry et al., 2008; Doney et al., 2009; Ries et al., 2009; Fig. 1) necessarily be interpreted as evidence of the model’s mechanistic accuracy. Additional microelectrode investigations of the calcifying fluids of a range of marine calcifiers are required to rigorously evaluate the viability of the model and its various iterations.
It should also be noted that there is considerable variability amongst calcifying taxa, as well as uncertainty amongst workers in the field of biomineralization, as to the nature of the medium from which marine calcifiers produce their shells and skeletons. Corals are thought to accrete CaCO$_3$ directly from a discrete calcifying fluid (e.g., Cohen and McConnaughey, 2003 and references therein; Al-Horani et al., 2003; Gaetani and Cohen, 2006; Cohen and Holcomb, 2009), with mineralization sites and crystal orientations being influenced by organic templates and/or calicoblastic cells (e.g., Goldberg, 2001; Cuif and Dauphin, 2005; Tambutté et al., 2007; Meibom et al., 2008). It has also been argued that mollusks secrete their shells from a discrete calcifying fluid known as the pallial fluid (e.g., Crenshaw, 1972), although hemocytes have also been implicated in molluscan shell formation (Mount et al., 2004). Calcifying bryopsidalean algae, such as halimeda (Borowitzka and Larkum, 1976; De Beer and Larkum, 2001; Ries, 2009), are also thought to precipitate their aragonite needles from a discrete calcifying fluid—with little or no biological control over crystal orientation. Likewise, crustacea are thought to nucleate their shells from a discrete calcifying fluid (Cameron, 1985)—a process that must occur rapidly within these molting organisms. Although the general distribution of mineralized structures within crustacea appears to be guided by organic templates, the fine-scale crystal organization within these structures is highly complex and does not appear to be governed by the gross morphology of the organic templates (Roer and Dillaman, 1984). Echinoids, in contrast, are thought to initiate calcification on Ca$^{2+}$-binding organic matrices within cellular vacuoles (Ameye et al., 1998).
Regardless of the exact composition (e.g., seawater vs. modified seawater) or nature (e.g., fluid vs. gel) of their calcifying media, or the mechanisms by which they accrete their CaCO$_3$ (e.g., organic templates vs. cellular mediation vs. proton-pumps vs. Ca$^{2+}$-ATPase), organisms’ ability to control the pH of their calcifying media should strongly influence their ability to convert DIC into CO$_3^{2-}$, thereby impacting their specific calcification response to CO$_2$-induced ocean acidification.
A potential weakness of the proton-pumping model is that the relative changes in [CO$_3^{2-}$] at the site of calcification predicted in response to elevated $pCO_2$ (Fig. 2) are in some cases substantially less than the relative changes in net calcification rates observed in prior ocean acidification experiments (e.g., Ries et al., 2009; Fig. 1). These discrepancies may be partly attributable to a non-linear relationship between calcification site [CO$_3^{2-}$] and net calcification rate—e.g., calcification may only commence once a threshold [CO$_3^{2-}$] of the calcifying fluid is crossed. These discrepancies may also result from the existence of other factors influencing rates of biological calcification that are not incorporated in the proposed model, including the fertilization of photosynthesis under elevated $pCO_2$, which could provide additional energy for H$^+$-regulation at the site of calcification, and/or dissolution of unprotected shell/skeleton in undersaturated conditions, which could convert a “positive” or “neutral” response in [CO$_3^{2-}$] at the site of calcification (e.g., white circles in Fig. 2A and B) into a “parabolic” or “threshold-negative” response in net calcification (e.g., Fig. 1D and E). Elevated [H$^+$] of seawater may also affect other aspects of organisms’ physiology (e.g., Langdon et al., 2003; Michaelidis et al., 2005; Wood et al., 2008) that may impact calcification indirectly, yet which are not incorporated into the physicochemical model proposed here.
6. CONCLUSION
A physicochemical model of organisms’ calcifying fluid is proposed and explored through two end-member scenarios:
one that assumes organisms remove a fixed number of H\(^+\) from a given volume of their calcifying fluid (variable \([H^+]_e/[H^+]_i\) regardless of atmospheric pCO\(_2\), and another that assumes that organisms maintain a fixed \([H^+]_e/[H^+]_i\) (variable H\(^+\)-removal). Both model scenarios are able to generate the full range of calcification responses observed in prior ocean acidification experiments and are consistent with the assertion that the calcification response to CO\(_2\)-induced ocean acidification will be more negative for marine calcifiers that exert weaker control over the pH of their calcifying fluid (Ries et al., 2009). Microelectrode measurements reveal that pH (and therefore \(Q_A\)) of the calcifying fluid of the temperate coral *A. poculata* is substantially elevated relative to external seawater pH under both control and acidified conditions, which is consistent with the principal assumption of the generalized H\(^+\)-pumping model of calcification. The microelectrode measurements also revealed that *A. poculata* maintained the same \([H^+]_e/[H^+]_i\) under the control and acidified treatments, yet removed fewer H\(^+\) from a given volume of its calcifying fluid under the acidified treatment. These results suggest that the fixed \([H^+]_e/[H^+]_i\) scenario is most applicable to this particular coral species. These models and measurements are among the first attempts to explore physicochemical changes in the calcifying fluids of marine calcifiers in response to CO\(_2\)-induced ocean acidification. Future experiments performed on a range of marine calcifiers that simultaneously monitor the impact of CO\(_2\)-induced ocean acidification on the composition of their calcifying fluids and on their net rates of calcification are required to evaluate and further constrain the theoretical framework presented here.
**ACKNOWLEDGMENTS**
The author acknowledges P. Smith and D. Borgdorff for assistance with the microelectrode measurements, and A. Cohen, D. McCorkle, M. Holcomb, J. Erez, A. Szmant, R. Whitehead, A. Taylor, and D. Allemand for their stimulating feedback on the ideas presented herein. The author also acknowledges three anonymous referees and A.E. Mucci, whose reviews improved the quality of this manuscript. This research was supported by a WHOI Ocean and Climate Change Postdoctoral Fellowship, NSF award #1031995, and a UNC faculty development grant.
**REFERENCES**
Adkins J. F., Boyle E. A., Curry W. B. and Lutringer A. (2003) Stable isotopes in deep-sea corals and a new mechanism for ‘vital effects’. *Geochim. Cosmochim. Acta* **67**, 1129–1143.
Al-Horani F. A., Al-Maghraibi S. M. and DeBeer D. (2003) The mechanism of calcification and its relation to photosynthesis and respiration in the scleractinian coral *Galaxea fascicularis*. *Mar. Biol.* **142**, 419–426.
Ameye L., Compere P., Dille J. and Dubois P. (1998) Ultrastructure and cytochemistry of the early calcification site and of its mineralization organic matrix in *Paracentrotus lividus* (Echinodermata: Echinoidea). *Histochem. Cell Biol.* **110**, 285–294.
Borowitzka M. A. (1987) Calcification in algae—mechanisms and the role of metabolism. *Crit. Rev. Plant Sci.* **6**, 1–45.
Borowitzka M. A. and Larkum A. W. D. (1976) Calcification in the green alga Halimeda. III: The sources of inorganic carbon for photosynthesis and calcification and a model of the mechanism of calcification. *J. Exp. Biol.* **27**, 879–893.
Cameron J. N. (1985) Post-moulт calcification in the blue crab (*Callinectes sapidus*): relationships between apparent net H\(^+\) excretion, calcium and bicarbonate. *J. Exp. Biol.* **119**, 275–285.
Clode P. L. and Marshall A. T. (2002) Low temperature FESEM of the calcifying interface of a scleractinian coral. *Tissue Cell* **34**, 187–198.
Cohen A. L. and Gaetani G. A. (2010) Ion partitioning and the geochemistry of coral skeletons: solving the mystery of the vital effect. *EMU Notes Mineral.* **11**, 377–397.
Cohen A. L. and Holcomb M. (2009) Why corals care about ocean acidification: uncovering the mechanism. *Oceanography* **22**, 118–127.
Cohen A. L. and McConnaughey T. A. (2003) A geochemical perspective on coral mineralization, In *Biomineralization* (eds. P. M. Dove, S. Weiner and J. J. De Yoreo), pp. 151–187. Reviews in Mineralogy and Geochemistry. The Mineralogical Society of America.
Constantz B. R. (1986) Coral skeleton construction: a physiochemically dominated process. *Palaios* **1**, 152–157.
Crenshaw M. A. (1972) The inorganic composition of molluscan extrapallial fluid. *Biol. Bull.* **143**, 506–512.
Cuif J.-P. and Dauphin Y. (2005) The environmental recording unit in coral skeletons—a synthesis of structural and chemical evidences for a biochemically driven, stepping-growth process in fibres. *Biogeosciences* **2**, 61–73.
De Beer D. and Larkum A. W. D. (2001) Photosynthesis and calcification in the calcifying algae *Halimeda discoidea* studied with microsensors. *Plant Cell Environ.* **24**, 1209–1217.
Doney S. C., Fabry V. J., Feely R. A. and Kleypas J. A. (2009) Ocean acidification: the other CO\(_2\) problem. *Ann. Rev. Mar. Sci.* **1**, 169–192.
Fabry V. J., Seibel B. A., Feely R. A. and Orr J. C. (2008) Impacts of ocean acidification on marine fauna and ecosystem processes. *ICES J. Mar. Sci.* **65**, 414–432.
Gaetani G. A. and Cohen A. L. (2006) Element partitioning during precipitation of aragonite from seawater: a framework for understanding paleoproxies. *Geochim. Cosmochim. Acta* **70**, 4617–4634.
Goldberg W. M. (2001) Desmocytes in the calicoblastic epithelium of the stony coral *Mycetophyllia reesi* and their attachment to the skeleton. *Tissue Cell* **33**, 388–394.
Goldman D. E. (1943) Potential, impedance, and rectification in membranes. *J. Gen. Physiol.* **27**, 37–60.
Hodgkin A. L. and Katz B. (1949) The effect of sodium ions on the electrical activity of the giant axon of the squid. *J. Physiol.* **108**, 37–77.
Holcomb M., Cohen A. L., Gabitov R. I. and Hutter J. L. (2009) Compositional and morphological features of aragonite precipitated experimentally from seawater and biogenically by corals. *Geochim. Cosmochim. Acta* **17**, 4166–4179.
Holcomb M., McCorkle D. C. and Cohen A. L. (2010) Long-term effects of nutrient and CO\(_2\) enrichment on the temperate coral *Astrangia poculata*. *J. Exp. Mar. Biol. Ecol.* **386**, 27–33.
Houbreque F., Meibom A., Cuif J.-P., Stolarski J., Marroccoli J., Ferrier-Page C., Domart-Coulon I. and Dunbar R. B. (2009) Strontium-86 labeling experiments show spatially heterogeneous skeletal formation in the scleractinian coral *Porites porites*. *Geophys. Res. Lett.* **36**, L04604. doi:10.1029/2008GL036782.
Iglesias-Rodriguez M. D., Halloran P. R., Rickaby R. E. M., Hall I. R., Colmenero-Hidalgo E., Gittins J. R., Green D. R. H., Tyrrell T., Gibbs S. J., von Dassow P., Rehm E., Armbrust E. V. and Boessenkool K. P. (2008) Phytoplankton calcification in a high-CO\(_2\) world. *Science* **320**, 336–340.
Kitasato H. (1968) The influence of H\(^+\) on the membrane potential and ion fluxes of *Nitella*. *J. Gen. Physiol.* **52**, 60–87.
Kleypas J. A., Feely R. A., Fabry V. J., Langdon C., Sabine C. S. and Robbins L. L. (2006) Impacts of Ocean Acidification on Coral Reefs and Other Marine Calcifiers: A Guide for Future Research. Report of a workshop held 18–20 April 2005, St. Petersburg, FL, 88 p.
Langdon C. (2000) Review of experimental evidence for effects of CO$_2$ on calcification of reef-builders. In *Proceedings of the 9th International Coral Reef Symposium*. pp. 1091–1098.
Langdon C., Broecker W. S., Hammond D. E., Glenn E., Fitzsimmons K., Nelson S. G., Peng T. H., Hajdas I. and Bonani G. (2003) Effect of elevated CO$_2$ on the community metabolism of an experimental coral reef. *Global Biogeochem. Cycles* **17**, 1101–1114.
Lewis E. and Wallace D. W. R. (1998) Program developed for CO$_2$ system calculations. ORNL/CDIAC-105. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, US Department of Energy, Oak Ridge, Tennessee.
Meibom A., Cuif J.-P., Houlbreque F., Mostefaouia S., Dauphin Y., Meibom K. L. and Dunbar R. (2008) Chemical variations at ultra-structural lengthscales in coral skeleton. *Geochim. Cosmochim. Acta* **72**, 1555–1569.
McConnaughey T. A. (2003) Sub-equilibrium oxygen-18 and carbon-13 levels in biological carbonates and kinetic models. *Coral Reefs* **22**, 316–327.
McConnaughey T. A. and Falk R. H. (1991) Calcium-proton exchange during algal calcification. *Biol. Bull.* **180**, 185–195.
McConnaughey T. A. and Whelan J. F. (1997) Calcification generates protons for nutrient and bicarbonate uptake. *Earth Sci. Rev.* **42**, 95–117.
Michaelidis B., Ouzounis C., Paleras A. and Poertner H. O. (2005) Effects of long-term moderate hypercapnia on acid-base balance and growth rate in marine mussels *Mytilus galloprovincialis*. *Mar. Ecol. Prog. Ser.* **293**, 109–118.
Mount A. S., Wheeler A. P., Paradkar R. P. and Snider D. (2004) Hemocoele-mediated shell mineralization in the Eastern Oyster. *Science* **304**, 297–300.
Mucci A. (1983) The solubility of calcite and aragonite in seawater at various salinities, temperatures, and one atmosphere total pressure. *Am. J. Sci.* **283**, 780–799.
Richards J. L. and Hope A. B. (1974) The role of protons in determining membrane electrical characteristics in *Chara corallina*. *J. Membr. Biol.* **16**, 121–144.
Ries J. B. (2010a) Geological and experimental evidence for secular variation in seawater Mg/Ca (calcite–aragonite seas) and its effects on marine biological calcification. *Biogeosciences* **7**, 2795–2849.
Ries J. B. (2010b) Shell-shocked: how different creatures deal with an acidifying ocean. *Earth* **55**, 46–53.
Ries J. B. (2009) Effects of secular variation in seawater Mg/Ca ratio (calcite–aragonite seas) on CaCO$_3$ sediment production by the calcareous algae Halimeda, Penicillus and Udotea—evidence from recent experiments and the geological record. *Terra Nova* **21**, 323–339.
Ries J. B., Cohen A. L. and McCorkle D. C. (2009) Marine calcifiers exhibit mixed responses to CO$_2$-induced ocean acidification. *Geology* **37**, 1131–1134.
Ries J. B., Stanley S. M. and Hardie L. A. (2006) Scleractinian corals produce calcite, and grow more slowly, in artificial Cretaceous seawater. *Geology* **34**, 525–528.
Rink S., Kuhl M., Bijma J. and Spero H. J. (1998) Microsensor studies of photosynthesis and respiration in the symbiotic foraminifer *Orbulina universa*. *Mar. Biol.* **131**, 583–595.
Rodolfo-Metalpa R., Martin S., Ferrier-Pages C. andGattuso J. P. (2010) Response of the temperate coral *Cladocora caespitosa* to mid-and long-term exposure to pCO$_2$ and temperature levels projected for the year 2100 AD. *Biogeosciences* **7**, 289–300.
Roer R. and Dillaman R. (1984) The structure and calcification of the crustacean cuticle. *Integr. Comp. Biol.* **24**, 893–909.
Roy R. N., Roy L. N., Vogel K. M., Porter-Moore C., Pearson T., Good C. E., Millero F. J. and Campbell D. M. (1993) The dissociation constants of carbonic acid in seawater at salinities 5 to 45 and temperatures 0 to 45 °C. *Mar. Chem.* **44**, 249–267.
Saddler H. D. W. (1970) The membrane potential of *Acetabularia mediterranea*. *J. Gen. Physiol.* **55**, 802–821.
Tambutté E., Allemand D., Zoccola D., Meibom A., Lotto S., Caminiti N. and Tambutté S. (2007) Observations of the tissue-skeleton interface in the scleractinian coral *Stylophora pistillata*. *Coral Reefs* **26**, 517–529.
Wood H. L., Spicer J. I. and Widdicombe S. (2008) Ocean acidification may increase calcification rates, but at a cost. *Proc. R. Soc. B Biol. Sci.* **275**, 1767–1773.
*Associate editor:* Alfonso Mucci |
Extending Hybrid CSP with Probability and Stochasticity
Yu Peng, Shuling Wang\(^{(\otimes)}\), Naijun Zhan, and Lijun Zhang
State Key Laboratory of Computer Science, Institute of Software,
Chinese Academy of Sciences, Beijing, China
email@example.com
Abstract. Probabilistic and stochastic behavior are omnipresent in computer controlled systems, in particular, so-called safety-critical hybrid systems, because of fundamental properties of nature, uncertain environments, or simplifications to overcome complexity. Tightly intertwining discrete, continuous and stochastic dynamics complicates modelling, analysis and verification of stochastic hybrid systems (SHSs). In the literature, this issue has been extensively investigated, but unfortunately it still remains challenging as no promising general solutions are available yet. In this paper, we give our effort by proposing a general compositional approach for modelling and verification of SHSs. First, we extend Hybrid CSP (HCSP), a very expressive and process algebra-like formal modeling language for hybrid systems, by introducing probability and stochasticity to model SHSs, which is called stochastic HCSP (SHCSP). To this end, ordinary differential equations (ODEs) are generalized by stochastic differential equations (SDEs) and non-deterministic choice is replaced by probabilistic choice. Then, we extend Hybrid Hoare Logic (HHL) to specify and reason about SHCSP processes. We demonstrate our approach by an example from real-world.
1 Introduction
Probabilistic and stochastic behavior are omnipresent in computer controlled systems, such as safety-critical hybrid systems, because of uncertain environments, or simplifications to overcome complexity. For example, the movement of aircrafts could be influenced by wind; in networked control systems, message loss and other random effects (e.g., node placement, node failure, battery drain, measurement imprecision) may happen.
Stochastic hybrid systems (SHSs) are systems in which discrete, continuous and stochastic dynamics tightly intertwine. As many of SHSs are safety-critical, a thorough validation and verification activity is necessary to enhance the quality of SHSs and, in particular, to fulfill the quality criteria mandated by the relevant standards. But modeling, analysis and verification of SHSs is difficult and challenging. An obvious research line is to extend hybrid automata [10], which is the most popular model for traditional hybrid systems, by adding probability and
stochasticity. Then, verification of SHSs can be done naturally through reachability analysis, either by probabilistic model-checking [1–3,6,8,20,21], or by simulation i.e., statistical model-checking [15,23]. Along this line, several different notions of *stochastic hybrid automata* have been proposed [1–3,6,8,20,21], with the difference on where to introduce randomness. One option is to replace deterministic jumps by probability distribution over deterministic jumps. Another option is to generalize differential equations inside a mode by stochastic differential equations. Stochastic hybrid systems comprising stochastic differential equations have been investigated in [1,5,13]. More general models can be obtained by mixing the above two choices, and by combining them with memoryless timed probabilistic jumps [4], with a random reset function for each discrete jump [6]. An overview of this line can be found in [4].
To model complex systems, some compositional modelling formalisms have been proposed, e.g., HMODEST [7] and stochastic hybrid programs [18]. HCSP due to He, Zhou, et al [9,22] is an extension of CSP [12] by introducing differential equations to model continuous evolution and three types of interruptions (i.e., communication interruption, timeout and boundary condition) to model interactions between continuous evolutions and discrete jumps in HSs. The extension of CSP to probabilistic setting has been investigated by Morgan et al. [16]. In this paper, we propose a compositional approach for modelling and verification of stochastic hybrid systems. First, we extend Hybrid CSP (HCSP), a very expressive and process algebra-like modeling language for hybrid systems by introducing probability and stochasticity, called stochastic HCSP (SHCSP), to model SHSs. In SHCSP, ordinary differential equations (ODEs) are generalized to stochastic differential equations (SDEs), and non-deterministic choice is replaced by probabilistic choice. Different from Platzer’s work [18], SHCSP provides more expressive constructs for describing hybrid systems, including communication, parallelism, interruption, and so on.
Probabilistic model-checking of SHSs does not scale, in particular, taking SDEs into account. For example, it is not clear how to approximate the reachable sets of a simple linear SDEs with more than two variables. Therefore, existing verification techniques based on reachability analysis for SHSs are inadequate, and new approaches are expected. As an alternative, in [18], Platzer for the first time investigated how to extend deductive verification to SHSs. Inspired by Platzer’s work, for specifying and reasoning about SHCSP process, we extend Hybrid Hoare Logic [14], which is an extension of Hoare logic [11] to HSs, to SHSs. Comparing with Platzer’s work, more computation features of SHSs, and more expressive constructs such as concurrency, communication and interruption, can be well handled in our setting. We demonstrate our approach by modeling and verification of the example of aircraft planning problem from the real-world.
## 2 Background and Notations
Assume that $\mathcal{F}$ is a $\sigma$-algebra on set $\Omega$ and $P$ is a probability measure on $(\Omega, \mathcal{F})$, then $(\Omega, \mathcal{F}, P)$ is called a *probability space*. We here assume that every
subset of a null set (i.e., $P(A) = 0$) with probability 0 is measurable. A property which holds with probability 1 is said to hold \textit{almost surely} (\textit{a.s.}). A \textit{filtration} is a sequence of $\sigma$-algebras $\{\mathcal{F}_t\}_{t \geq 0}$ with $\mathcal{F}_{t_1} \subseteq \mathcal{F}_{t_2}$ for all $t_1 < t_2$. We always assume that a filtration $\{\mathcal{F}_t\}_{t \geq 0}$ has been completed to include all null sets and is right-continuous.
Let $\mathcal{B}$ represent the Borel $\sigma$-algebra on $\mathbb{R}^n$, i.e. the $\sigma$-algebra generated by all open subsets. A mapping $X : \Omega \rightarrow \mathbb{R}^n$ is called $\mathbb{R}^n$-valued \textit{random variable} if for each $B \in \mathcal{B}$, we have $X^{-1}(B) \in \mathcal{F}$, i.e. $X$ is $\mathcal{F}$-\textit{measurable}. A \textit{stochastic process} $X$ is a function $X : T \times \Omega \rightarrow \mathbb{R}^n$ such that for each $t \in T$, $X(t, \cdot) : \Omega \rightarrow \mathbb{R}^n$ is a random variable, and for each $\omega \in \Omega$, $X(\cdot, \omega) : T \rightarrow \mathbb{R}^n$ corresponds to a \textit{sample path}. A stochastic process $X$ is \textit{adapted} to a filtration $\{\mathcal{F}_t\}_{t \geq 0}$ if $X_t$ is $\mathcal{F}_t$-measurable. Intuitively, a filtration represents all available historical information of a stochastic process, but nothing related to its future. A \textit{càdlàg} function defined on $\mathbb{R}$ is \textit{right continuous} and has \textit{left limit}. A stochastic process $X$ is \textit{càdlàg} iff all of its paths $t \rightarrow X_t(\omega)$ (for each $\omega \in \Omega$) are \textit{càdlàg}. A $d$-dimensional \textit{Brownian motion} $W$ is a stochastic process with $W_0 = 0$ that is continuous almost surely everywhere and has independent increments with time, i.e. $W_t - W_s \sim N(0, t - s)$ (for $0 \leq s < t$), where $N(0, t - s)$ denotes the normal distribution with mean 0 and variance $t - s$. Brownian motion is mathematically extremely complex. Its path is almost surely continuous everywhere but differentiable nowhere. Intuitively, $W$ can be understood as the limit of a random walk. A \textit{Markov time} with respect to a stochastic process $X$ is a random variable $\tau$ such that for any $t \geq 0$, the event $\{\tau \leq t\}$ is determined by (at most) the information up to time $t$, i.e. $\{\tau \leq t\} \in \mathcal{F}_t$.
We use \textit{stochastic differential equation} (SDE) to model stochastic continuous evolution, which is of the form $dX_t = b(X_t)dt + \sigma(X_t)dW_t$, where $W_t$ is a Brownian motion. In which, the drift coefficient $b(X_t)$ determines how the deterministic part of $X_t$ changes with respect to time and the diffusion coefficient $\sigma(X_t)$ determines the stochastic influence to $X_t$ with respect to the Brownian motion $W_t$. Obviously, any solution to an SDE is a stochastic process.
### 3 Stochastic HCSP
A system in Stochastic HCSP (SHCSP) consists of a finite set of sequential processes in parallel which communicate via channels synchronously. Each sequential process is represented as a collection of stochastic processes, each of which arises from the interaction of discrete computation and stochastic continuous dynamics modeled by stochastic differential equations.
Let $Proc$ represent the set of SHCSP processes, $\Sigma$ the set of channel names. The syntax of SHCSP is given as follows:
$$P ::= \text{skip} \mid x := e \mid ch?x \mid ch!e \mid P; Q \mid B \rightarrow P \mid P^* \mid P \sqcup_p Q \mid \langle ds = bdt + \sigma dW & B \rangle \mid \langle ds = bdt + \sigma dW & B \rangle \triangleright \parallel_{i \in I} (\omega_i \cdot ch_i * \rightarrow Q_i)$$
$$S ::= P \mid S \| S$$
Here \( ch, ch_i \in \Sigma \), \( ch_i* \) stands for a communication event, e.g. \( ch?x \) or \( ch!e \), \( x \) is a variable, \( B \) and \( e \) are Boolean and arithmetic expressions, \( P, Q, Q_i \in Proc \) are sequential processes, \( p \in [0, 1] \) stands for the probability of the choice between \( P \) and \( Q \), \( s \) for a vector of continuous variables, \( b \) and \( \sigma \) for functions of \( s \), \( W \) for the Brownian motion process. At the end, \( S \) stands for a system, i.e., a SHCSP process.
As defined in the syntax of \( P \), the processes in the first line are original from HCSP, while the last two lines are new for SHCSP. The individual constructs can be understood intuitively as follows:
- **skip**, the assignment \( x := e \), the sequential composition \( P; Q \), and the alternative statement \( B \rightarrow P \) are defined as usual.
- \( ch?x \) receives a value along channel \( ch \) and assigns it to \( x \).
- \( ch!e \) sends the value of \( e \) along channel \( ch \). A communication takes place when both the sending and the receiving parties are ready, and may cause one side to wait.
- The repetition \( P^* \) executes \( P \) for some finite number of times.
- \( P \sqcup_p Q \) denotes probabilistic choice. It behaves as \( P \) with probability \( p \) and as \( Q \) with probability \( 1 - p \).
- \( \langle ds = bdt + \sigma dW & B \rangle \) specifies that the system evolves according to the stochastic process defined by the stochastic differential equation \( ds = bdt + \sigma dW \). As long as the boolean expression \( B \), which defines the domain of \( s \), turns false, it terminates. We will later use \( d(s) \) to return the dimension of \( s \).
- \( \langle ds = bdt + \sigma dW & B \rangle \sqsupset \bigoplus_{i \in I} (\omega_i; ch_i* \rightarrow Q_i) \) behaves like \( \langle ds = bdt + \sigma dW & B \rangle \), except that the stochastic evolution is preempted as soon as one of the communications \( ch_i* \) takes place, after that the respective \( Q_i \) is executed. \( I \) is supposed to be finite and for each \( i \in I \), \( \omega_i \in \mathbb{Q}^+ \) represents the weight of \( ch_i* \). If one or more communications are ready at the same time, say they are \( \{ch_j*\}_{j \in J} \) with \( J \subseteq I \) and \( |J| \geq 1 \), then \( ch_j \) is chosen with the probability \( \frac{\omega_j}{\sum_{j \in J} \omega_j} \), for each \( j \in J \). If the stochastic dynamics terminates before a communication among \( \{ch_i*\}_I \) occurring, then the process terminates without communicating.
- \( S_1 \| S_2 \) behaves as if \( S_1 \) and \( S_2 \) run independently except that all communications along the common channels connecting \( S_1 \) and \( S_2 \) are to be synchronized. The processes \( S_1 \) and \( S_2 \) in parallel can neither share variables, nor input nor output channels.
### 3.1 A Running Example
We use SHCSP to model the aircraft position during the flight, which is inspired from [19]. Consider an aircraft that is following a flight path consisting of a sequence of line segments at a fixed altitude. Ideally, the aircraft should fly at a constant velocity \( v \) along the nominal path, but due to the wind or cloud disturbance, the deviation of the aircraft from the path may occur. For safety, the aircraft should follow a correction heading to get back to the nominal path as quickly as possible. On one hand, the correction heading should be orthogonal to
the nominal path for the shortest way back, but on the other hand, it should also go ahead to meet the destination. Considering these two objectives, we assume the correction heading always an acute angle with the nominal path.
Here we model the behavior of the aircraft along one line segment. Without loss of generality, we assume the segment is along $x$-axis, with $(x_s, 0)$ as the starting point and $(x_e, 0)$ as the ending point. When the aircraft deviates from the segment with a vertical distance greater than $\lambda$, we consider it enters a dangerous state. Let $(x_s, y_0)$ be the initial position of the aircraft in this segment, then the future position of the aircraft $(x(t), y(t))$ is governed by the following SDE:
$$\begin{pmatrix} dx(t) \\ dy(t) \end{pmatrix} = v \begin{pmatrix} \cos(\theta(t)) \\ \sin(\theta(t)) \end{pmatrix} dt + dW(t)$$
where $\theta(t)$ is the correction heading and is defined with a constant degree $\frac{\pi}{4}$ when the aircraft deviates from the nominal path:
$$\theta(t) = \begin{cases} -\frac{\pi}{4} & \text{if } y(t) > 0 \\ 0 & \text{if } y(t) = 0 \\ \frac{\pi}{4} & \text{if } y(t) < 0 \end{cases}$$
Let $B$ be $x_s \leq x \leq x_e$, the movement of the aircraft described above can be modelled by the following SHCSP process $P_{Air}$:
$$x = x_s; y = y_0; \langle [dx, dy]^T = v[\cos(\theta(t)), \sin(\theta(t))]^T dt + dW(t) \& B \rangle$$
## 4 Operational Semantics
Before giving operational semantics, we introduce some notations first.
**System Variables.** In order to interpret SHCSP processes, we use non-negative reals $\mathbb{R}^+$ to model time, and introduce a global clock $now$ as a system variable to record the time in the execution of a process. A *timed communication* is of the form $\langle ch, c, b \rangle$, where $ch \in \Sigma$, $c \in \mathbb{R}$ and $b \in \mathbb{R}^+$, representing that a communication along channel $ch$ occurs at time $b$ with value $c$ transmitted. The set $\Sigma \times \mathbb{R} \times \mathbb{R}^+$ of all timed communications is denoted by $T\Sigma$. The set of all timed traces is
$$T\Sigma^*_\leq = \{ \gamma \in T\Sigma^* \mid \text{if } \langle ch_1, c_1, b_1 \rangle \text{ precedes } \langle ch_2, c_2, b_2 \rangle \text{ in } \gamma, \text{ then } b_1 \leq b_2 \}.$$
If $C \subseteq \Sigma$, $\gamma \upharpoonright_C$ is the projection of $\gamma$ onto $C$ such that only the timed communications along channels of $C$ in $\gamma$ are preserved. Given two timed traces $\gamma_1, \gamma_2$, and $X \subseteq \Sigma$, the *alphabetized parallel* of $\gamma_1$ and $\gamma_2$ over $X$, denoted by $\gamma_1 \gamma_2$, results in the following set of timed traces
$$\{ \gamma \mid \gamma \upharpoonright_{\Sigma - (\Sigma(\gamma_1) \cup \Sigma(\gamma_2))} = \epsilon, \gamma \upharpoonright_{\Sigma(\gamma_1)} = \gamma_1, \gamma \upharpoonright_{\Sigma(\gamma_2)} = \gamma_2 \text{ and } \gamma \upharpoonright_X = \gamma_1 \upharpoonright_X = \gamma_2 \upharpoonright_X \},$$
where $\Sigma(\gamma)$ stands for the set of channels that occur in $\gamma$.
To model synchronization of communication events, we need to describe their readiness. Because a communication itself takes no time when both parties get
ready, thus, at a time point, multiple communications may occur. In order to record the execution order of communications occurring at the same time point, we prefix each communication readiness a timed trace that happened before the ready communication event. Formally, each *communication readiness* has the form of $\gamma.ch?$ or $\gamma.ch!$, where $\gamma \in T\Sigma^*_\leq$. We denote by $RDY$ the set of communication readiness in the sequel.
Finally, we introduce two system variables, $rdy$ and $tr$, to represent the ready set of communication events and the timed trace accumulated at the considered time, respectively. In what follows, we use $Var(P)$ to represent the set of process variables of $P$, plus the system variables $\{rdy, tr, now\}$ introduced above, which take values respectively from $\mathbb{R} \cup RDY \cup T\Sigma^*_\leq \cup \mathbb{R}^+$, denoted by $Val$.
**States and Functions.** To interpret a process $P \in Proc$, we define a state $ds$ as a mapping from $Var(P)$ to $Val$, and denote by $D$ the set of such states. Because of stochasticity, we introduce a random variable $\rho : \Omega \to D$ to describe a distribution of all possible states. In addition, we introduce a stochastic process $H : Intv \times \Omega \to D$ to represent the continuous flow of process $P$ over the time interval $Intv$, i.e., state distributions on the interval. In what follows, we will abuse state distribution as state if not stated otherwise.
Given two states $\rho_1$ and $\rho_2$, we say $\rho_1$ and $\rho_2$ are parallelable iff for each $\omega \in \Omega$, $Dom(\rho_1(\omega)) \cap Dom(\rho_2(\omega)) = \{rdy, tr, now\}$ and $\rho_1(\omega)(now) = \rho_2(\omega)(now)$. Given two parallelable states $\rho_1$ and $\rho_2$, paralleling them over $X \subseteq \Sigma$ results in a set of new states, denoted by $\rho_1 \oplus \rho_2$, any of which $\rho$ is given by
$$\rho(\omega)(v) \overset{\text{def}}{=} \begin{cases}
\rho_1(\omega)(v) & \text{if } v \in Dom(\rho_1(\omega)) \setminus Dom(\rho_2(\omega)), \\
\rho_2(\omega)(v) & \text{if } v \in Dom(\rho_2(\omega)) \setminus Dom(\rho_1(\omega)), \\
\rho_1(\omega)(now) & \text{if } v = now, \\
\gamma, \text{ where } \gamma \in \rho_1(\omega)(tr)\rho_2(\omega)(tr) & \text{if } v = tr, \\
\rho_1(\omega)(rdy) \cup \rho_2(\omega)(rdy) & \text{if } v = rdy.
\end{cases}$$
It makes no sense to distinguish any two states in $\rho_1 \oplus \rho_2$, so hereafter we abuse $\rho_1 \oplus \rho_2$ to represent any of its elements. $\rho_1 \oplus \rho_2$ will be used to represent states of parallel processes.
Given a random variable $\rho$, the update $\rho[v \to e]$ represents a new random variable such that for any $\omega \in \Omega$ and $x \in Var$, $\rho[v \to e](\omega)(x)$ is defined as the value of $e$ if $x$ is $v$, and $\rho(\omega)(x)$ otherwise. Given a stochastic process $X : [0, d) \times \Omega \to R^{d(s)}$, for any $t$ in the domain, $\rho[s \to X_t]$ is a new random variable such that for any $\omega \in \Omega$ and $x \in Var$, $\rho[s \to X_t](\omega)(x)$ is defined as $X(t, w)$ if $x$ is $s$, and $\rho(\omega)(x)$ otherwise.
At last, we define $H^\rho_d$ as the stochastic process over interval $[\rho(now), \rho(now) + d]$ such that for any $t \in [\rho(now), \rho(now) + d]$ and any $\omega$, $H^\rho_d(t, \omega) = \rho[now \mapsto t](\omega)$, and moreover, $H^\rho_{d,s,X}$ as the stochastic process over interval $[\rho(now), \rho(now) + d]$ such that for any $t \in [\rho(now), \rho(now) + d]$ and any $\omega$, $H^\rho_{d,s,X}(t, \omega) = \rho[now \mapsto t, rdy \mapsto \emptyset, s \mapsto X_t](\omega)$.
4.1 Operational Semantics
Each transition relation has the form of \((P, \rho) \xrightarrow{\alpha} (P', \rho', H)\), where \(P\) and \(P'\) are processes, \(\alpha\) is an event, \(\rho, \rho'\) are states, \(H\) is a stochastic process. It expresses that starting from initial state \(\rho\), \(P\) evolves into \(P'\) by performing event \(\alpha\), and ends in state \(\rho'\) and the execution history of \(\alpha\) is recorded by continuous flow \(H\). When the transition is discrete and thus produces a flow on a point interval (i.e. current time now), we will write \((P, \rho) \xrightarrow{\alpha} (P', \rho')\) instead of \((P, \rho) \xrightarrow{\alpha} (P', \rho', \{\rho(\text{now}) \mapsto \rho'\})\). The label \(\alpha\) represents events, which can be an internal event like skip, assignment, or a termination of a continuous etc, uniformly denoted by \(\tau\), or an external communication event \(ch!c\) or \(ch?c\), or an internal communication \(ch.c\), or a time delay \(d\) that is a positive real number. We call the events but the time delay discrete events, and will use \(\beta\) to range over them. We define the dual of \(ch?c\) (denoted by \(\overline{ch?c}\)) as \(ch!c\), and vice versa, and define \(comm(ch!c, ch?c)\) or \(comm(ch?c, ch!c)\) as the communication \(ch.c\). In the operational semantics, besides the timed communications, we will also record the internal events that have occurred till now in \(tr\).
For page limit, we present the semantics for the new constructs of SHCSP in the paper in Table 1. The semantics for the rest is same to HCSP, which can be found at [17]. The semantics for probabilistic choice is given by rules (PCho-1) and (PCho-2): it is defined with respect to a random variable \(U\) which distributes uniformly in \([0, 1]\), such that for any sample \(\omega\), if \(U(\omega) \leq p\), then \(P\) is taken, otherwise, \(Q\) is taken. In either case, it is assumed that an internal action happened. A stochastic dynamics can continuously evolve for \(d\) time units if \(B\) always holds during this period, see (Cont-1). In (Cont-1), the variable \(X\) solves the stochastic process and the ready set keeps unchanged, reflected by the flow \(H^{\rho, s, X}_d\). The stochastic dynamics terminates at a point whenever \(B\) turns out false at a neighborhood of the point (Cont-2). Communication interrupt evolves for \(d\) time units if none of the communications \(ch_i*\) is ready (IntP-1), or is interrupted to execute \(ch_{i_j}*\) whenever \(ch_{i_j}*\) occurs first (IntP-2), or terminates immediately in case the continuous terminates before any communication happening (IntP-3).
The following theorem indicates that the semantics of SHCSP is well defined.
**Theorem 1.** For each transition \((P, \rho) \xrightarrow{\alpha} (P', \rho', H)\), \(H\) is an almost surely càdlàg process and adapted to the completed filtration \((F_t)_{t \geq 0}\) (generated by \(\rho\), the Brownian motion \((B_s)_{s \leq t}\), the weights \(\{\omega_i\}_{i \in I}\) and uniform \(U\) process) and the evolving time from \(P\) to \(P'\), denoted by \(\Delta(P, P')\), is a Markov time.
**Proof.** The proof of this theorem can be found at [17].
5 Assertions and Specifications
In this section, we define a specification logic for reasoning about SHCSP programs. We will first present the assertions including syntax and semantics, and then the specifications based on Hoare triples. The proof system will be given in next section.
5.1 Assertion Language
The assertion language is essentially defined by a first-order logic with emphasis on the notion of explicit time and the addition of several specific predicates on occurrence of communication traces and events. Before giving the syntax of assertions, we introduce three kinds of expressions first.
\[
h ::= \varepsilon \mid \langle ch.E,T \rangle \mid h \cdot h \mid h^*
\]
\[
E ::= c \mid x \mid f^k(E_1,...,E_k)
\]
\[
T ::= o \mid now \mid u^l(T_1,...,T_l)
\]
$h$ defines trace expressions, among which $\langle ch.E,T \rangle$ represents that there is a value $E$ transmitted along channel $ch$ at time $T$. $E$ defines value expressions,
**Table 1.** The semantics of new constructs of SHCSP
\[
U \text{ is a random variable distributed uniformly in } [0,1], U(\omega) \leq p
\] (PCho-1)
\[
(P \sqcup_p Q, \rho) \xrightarrow{\tau} (P, \rho[tr \mapsto tr \cdot \langle \tau, now \rangle])
\]
\[
U \text{ is a random variable distributed uniformly in } [0,1], U(\omega) > p
\] (PCho-2)
\[
(P \sqcup_p Q, \rho) \xrightarrow{\tau} (Q, \rho[tr \mapsto tr \cdot \langle \tau, now \rangle])
\]
$X : [0,d) \times \Omega \to \mathbb{R}^{d(s)}$ is the solution of
\[
ds = bdt + \sigma dW \land \forall t \in [0,d), \forall \omega.\rho[now \mapsto now + t, s \mapsto X_t](\omega)(B) = T
\] (Cont-1)
\[
(\langle ds = bdt + \sigma dW & B \rangle, \rho) \xrightarrow{d} \left( \langle ds = bdt + \sigma dW & B \rangle, \rho[now \mapsto now + d, s \mapsto X_d], H_d^{\rho,s,X} \right)
\]
$\exists \omega.(\rho(\omega)(B) = F)$ or $(X : [0,d) \times \Omega \to \mathbb{R}^{d(s)}$ is the solution of $ds = bdt + \sigma dW,$
$\exists \varepsilon > 0 \forall t \in (0,\varepsilon) \exists \omega.\rho[now \mapsto now + t, s \mapsto X_t](\omega)(B) = F)$
\[
(\langle ds = bdt + \sigma dW & B \rangle, \rho) \xrightarrow{\tau} (\epsilon, \rho[tr \mapsto tr \cdot \langle \tau, now \rangle])
\] (Cont-2)
\[
(ch_i*;Q_i, \rho) \xrightarrow{d} (ch_i*;Q_i, \rho', H_i), \quad \forall i \in I
\]
\[
(\langle ds = bdt + \sigma dW & B \rangle, \rho) \xrightarrow{d} ((\langle ds = bdt + \sigma dW & B \rangle, \rho', H)
\] (IntP-1)
$\{ch_{i_k*}\}_{1 \leq k \leq n}$ get ready simultaneously while others not
$U$ is a random variable distributed uniformly in $[0,1]$, and for $1 \leq j \leq n$
\[
\sum_{k=1}^{j-1} \omega_{i_k} \leq U(\omega) < \sum_{k=1}^{j} \omega_{i_k} \text{ and } (ch_{i_j*};Q_{i_j}, \rho) \xrightarrow{ch_{i_j*}} (Q_{i_j}, \rho')
\]
\[
(\langle ds = bdt + \sigma dW & B \rangle \supseteq \llbracket_{i \in I}(\omega_i \cdot ch_i* \to Q_i), \rho) \xrightarrow{ch_{i_j*}} (Q_{i_j}, \rho')
\] (IntP-2)
\[
(\langle ds = bdt + \sigma dW & B \rangle, \rho) \xrightarrow{\tau} (\epsilon, \rho')
\]
\[
(\langle ds = bdt + \sigma dW & B \rangle \supseteq \llbracket_{i \in I}(\omega_i \cdot ch_i* \to Q_i), \rho) \xrightarrow{\tau} (\epsilon, \rho')
\] (IntP-3)
including a value constant \( c \), a variable \( x \), or arithmetic value expressions. \( T \) defines time expressions, including a time constant \( o \), system variable \( now \), or arithmetic time expressions.
The categories of the assertion language include terms, denoted by \( \theta, \theta_1 \) etc., state formulas, denoted by \( S, S_1 \) etc., formulas, denoted by \( \varphi, \varphi_1 \) etc., and probability formulas, denoted by \( P \) etc., which are given by the following BNFs:
\[
\begin{align*}
\theta & ::= E \mid T \mid h \mid tr \\
S & ::= \bot \mid R^n(\theta_1, ..., \theta_n) \mid h.ch? \mid h.ch! \mid \neg S \mid S_1 \lor S_2 \\
\varphi & ::= \bot \mid S \text{ at } T \mid \neg \varphi \mid \varphi_1 \lor \varphi_2 \mid \forall v.\varphi \mid \forall t.\varphi \\
P & ::= P(\varphi) \bowtie p \mid \neg P \mid P \lor P
\end{align*}
\]
The terms \( \theta \) include value, time and trace expressions, plus trace variable \( tr \). The state expressions \( S \) include false (denoted by \( \bot \)), truth-valued relation \( R^n \) on terms, readiness, and logical combinations of state formulas. In particular, the readiness \( h.ch? \) or \( h.ch! \) represents that the communication event \( ch? \) or \( ch! \) is enabled, and prior to it, the sequence of communications recorded in \( h \) has occurred. The formulas \( \varphi \) include false, a primitive \( S \) at \( T \) representing that \( S \) holds at time \( T \); and logical combinations of formulas (\( v, t \) represent logical variables for values and time resp.). For time primitive, we have an axiom that \( (S_1 \text{ at } T \land S_2 \text{ at } T) \Leftrightarrow (S_1 \land S_2) \) at \( T \). We omit all the other axiom and inference rules for the formulas, that are same to first-order logic. The probability formula \( P \) has the form \( P(\varphi) \bowtie p \), where \( \bowtie \in \{<, \leq, >, \geq\} \), \( p \in \mathbb{Q} \cap [0, 1] \), or the logical composition of probability formulas free of quantifiers. In particular, \( P(\varphi) \bowtie p \) means that \( \varphi \) is true with probability \( \bowtie p \). For the special case \( P(\varphi) = 1 \), we write \( \varphi \) for short.
In the sequel, we use the standard logical abbreviations, as well as
\[
\begin{align*}
\varphi \text{ dr } [T_1, T_2] & \overset{\text{def}}{=} \forall t.(T_1 \leq t \leq T_2) \Rightarrow \varphi \text{ at } t \\
\varphi \text{ in } [T_1, T_2] & \overset{\text{def}}{=} \exists t.(T_1 \leq t \leq T_2) \land \varphi \text{ at } t
\end{align*}
\]
**Interpretation.** In the following, we will use a random variable \( Z : \Omega \rightarrow (Var \rightarrow Val) \) to describe the current state and a stochastic process \( H : [0, +\infty) \times \Omega \rightarrow (Var \rightarrow Val) \) to represent the whole evolution. The semantics of a term \( \theta \) is a function \( [\![\theta]\!] : (\Omega \rightarrow (Var \rightarrow Val)) \rightarrow (\Omega \rightarrow Val) \) that maps any random variable \( Z \) to a random variable \( [\![\theta]\!]^Z \), defined as follows:
\[
\begin{align*}
[\![c]\!]^Z &= c \\
[\![x]\!]^Z &= Y \text{ where } Y(\omega) = Z(\omega)(x) \text{ for } \omega \in \Omega \\
[\![f^k(E_1, ..., E_k)]\!]^Z &= f^k([\![E_1]\!]^Z, ..., [\![E_k]\!]^Z) \\
[\![o]\!]^Z &= o \\
[\![now]\!]^Z &= Y \text{ where } Y(\omega) = Z(\omega)(now) \text{ for } \omega \in \Omega \\
[\![u^l(T_1, ..., T_l)]\!]^Z &= u^l([\![T_1]\!]^Z, ..., [\![T_l]\!]^Z) \\
[\![\varepsilon]\!]^Z &= \varepsilon \\
[\![\langle ch.E, T \rangle]\!]^Z &= \langle ch. [\![E]\!]^Z, [\![T]\!]^Z \rangle \\
[\![h_1 \cdot h_2]\!]^Z &= [\![h_1]\!]^Z \cdot [\![h_2]\!]^Z \\
[\![h^*]\!]^Z &= ([\![h]\!]^Z)^*
\end{align*}
\]
The semantics of state formula $S$ is a function $\llbracket S \rrbracket : (\Omega \rightarrow (Var \rightarrow Val)) \rightarrow (\Omega \rightarrow \{0, 1\})$ that maps any random variable $Z$ describing the current state to a boolean random variable $\llbracket S \rrbracket^Z$, defined as follows:
\[
\begin{align*}
\llbracket \bot \rrbracket^Z &= 0 \\
\llbracket R^n(\theta_1, \ldots, \theta_n) \rrbracket^Z &= R^n(\llbracket \theta_1 \rrbracket^Z, \ldots, \llbracket \theta_n \rrbracket^Z) \\
\text{where } R^n(\llbracket \theta_1 \rrbracket^Z, \ldots, \llbracket \theta_n \rrbracket^Z)(\omega) &= R^n(\llbracket \theta_1 \rrbracket^Z(\omega), \ldots, \llbracket \theta_n \rrbracket^Z(\omega)) \\
\llbracket h.ch? \rrbracket^Z &= I_{\{\omega \in \Omega | \llbracket h \rrbracket^Z(\omega).ch? \in Z(\omega)(rdy)\}} \\
\llbracket h.ch! \rrbracket^Z &= I_{\{\omega \in \Omega | \llbracket h \rrbracket^Z(\omega).ch! \in Z(\omega)(rdy)\}} \\
\llbracket \neg S \rrbracket^Z &= 1 - \llbracket S \rrbracket^Z \\
\llbracket S_1 \lor S_2 \rrbracket^Z &= \llbracket S_1 \rrbracket^Z + \llbracket S_2 \rrbracket^Z - \llbracket S_1 \rrbracket^Z * \llbracket S_2 \rrbracket^Z
\end{align*}
\]
where given a set $S$, the characteristic function $I_S$ is defined such that $I_S(w) = 1$ if $w \in S$ and $I_S(w) = 0$ otherwise. The semantics of formula $\varphi$ is interpreted over a stochastic process and an initial random variable. More precisely, it’s a function $\llbracket \varphi \rrbracket : ([0, +\infty) \times \Omega \rightarrow (Var \rightarrow Val)) \rightarrow (\Omega \rightarrow (Var \rightarrow Val)) \rightarrow (\Omega \rightarrow \{0, 1\})$ that maps a stochastic process $\mathcal{H}$ with initial state $Z$ to a boolean random variable $\llbracket \varphi \rrbracket^{\mathcal{H}, Z}$. The definition is given below:
\[
\begin{align*}
\llbracket \bot \rrbracket^{\mathcal{H}, Z} &= 0 \\
\llbracket S \text{ at } T \rrbracket^{\mathcal{H}, Z} &= \llbracket S \rrbracket^{\mathcal{H}}(\llbracket T \rrbracket^Z) \\
\llbracket \neg \varphi \rrbracket^{\mathcal{H}, Z} &= 1 - \llbracket \varphi \rrbracket^{\mathcal{H}, Z} \\
\llbracket \varphi_1 \lor \varphi_2 \rrbracket^{\mathcal{H}, Z} &= \llbracket \varphi_1 \rrbracket^{\mathcal{H}, Z} + \llbracket \varphi_2 \rrbracket^{\mathcal{H}, Z} - \llbracket \varphi_1 \rrbracket^{\mathcal{H}, Z} * \llbracket \varphi_2 \rrbracket^{\mathcal{H}, Z} \\
\llbracket \forall v. \varphi \rrbracket^{\mathcal{H}, Z} &= \inf\{\llbracket \varphi[b/v] \rrbracket^{\mathcal{H}, Z} : b \in \mathbb{R}\} \\
\llbracket \forall t. \varphi \rrbracket^{\mathcal{H}, Z} &= \inf\{\llbracket \varphi[b/t] \rrbracket^{\mathcal{H}, Z} : b \in \mathbb{R}^+\}
\end{align*}
\]
The semantics of probability formula $\mathcal{P}$ is defined by function $\llbracket \mathcal{P} \rrbracket : ([0, +\infty) \times \Omega \rightarrow (Var \rightarrow Val)) \rightarrow (\Omega \rightarrow (Var \rightarrow Val)) \rightarrow \{0, 1\}$ that maps a stochastic process $\mathcal{H}$ with initial state $Z$ to a boolean variable $\llbracket \mathcal{P} \rrbracket^{\mathcal{H}, Z}$. Formally,
\[
\llbracket P(\varphi) \bowtie p \rrbracket^{\mathcal{H}, Z} = (P(\llbracket \varphi \rrbracket^{\mathcal{H}, Z} = 1) = P(\{\omega \in \Omega : \llbracket \varphi \rrbracket^{\mathcal{H}, Z}(\omega) = 1\}) \bowtie p)
\]
The semantics for $\neg$ and $\lor$ can be defined as usual.
We have proved that the terms and formulas of the assertion language are measurable, stated by the following theorem:
**Theorem 2 (Measurability).** For any random variable $Z$ and any stochastic process $\mathcal{H}$, the semantics of $\llbracket \theta \rrbracket^Z$, $\llbracket S \rrbracket^Z$ and $\llbracket \varphi \rrbracket^{\mathcal{H}, Z}$ are random variables (i.e. measurable).
**Proof.** The proof of this theorem can be found at [17].
### 5.2 Specifications
Based on the assertion language, the specification for a SHCSP process $P$ is defined as a Hoare triple of the form $\{A; E\} P \{R; C\}$, where $A, E, R, C$ are probability formulas. $A$ and $R$ are precondition and postcondition, which specify
the initial state and the terminating state of $P$ respectively. For both of them, the formulas $\varphi$ occurring in them have the special form $S$ at now, and we will write $S$ for short. $E$ is called an assumption of $P$, which expresses the timed occurrence of the dual of communication events provided by the environment. $C$ is called a commitment of $P$, which expresses the timed occurrence of communication events, and the real-time properties of $P$.
**Definition 1 (Validity).** We say a Hoare triple $\{A; E\} P \{R; C\}$ is valid, denoted by $\models \{A; E\} P \{R; C\}$, iff for any process $Q$, any initial states $\rho_1$ and $\rho_2$, if $P$ terminates, i.e. $(P \| Q, \rho_1 \oplus \rho_2) \xrightarrow{\alpha^*} (\epsilon \| Q', \rho'_1 \oplus \rho'_2, \mathcal{H})$ then $\llbracket A \rrbracket^{\rho_1}$ and $\llbracket E \rrbracket^{\mathcal{H}, \rho_2}$ imply $\llbracket R \rrbracket^{\rho'_1}$ and $\llbracket C \rrbracket^{\mathcal{H}, \rho'_1}$, where $\mathcal{H}$ is the stochastic process of the evolution.
## 6 Proof System
We present a proof system for reasoning about all valid Hoare triples for SHCSP processes. First we axiomatize SHCSP language by defining the axioms and inference rules for all the primitive and compound constructs, and then the general rules and axioms that are applicable to all processes.
**Skip.** The rule for skip is very simple. Indicated by $\top$, the skip process requires nothing from the environment for it to execute, and guarantees nothing during its execution.
$$\{A; \top\} \text{skip} \{A; \top\}$$
**Assignment.** The assignment $x := e$ changes nothing but assigns $x$ to $e$ in the final state, taking no time to complete.
$$\{A[e/x]; \top\} x := e \{A; \top\}$$
**Input.** For input $ch?x$, we use logical variables $o$ to denote the starting time, $h$ the initial trace, and $v$ the initial value of $x$ respectively, in the precondition. The assumption indicates that the compatible output event is not ready during $[o, o_1)$, and at time $o_1$, it becomes ready. As a consequence of the assumption, during the whole interval $[o, o_1]$, the input event keeps waiting and ready, as indicated by the commitment. At time $o_1$, the communication occurs and terminates immediately. As indicated by the postcondition, $x$ is assigned by some value $v'$ received, the trace is augmented by the new pair $\langle ch.v', o_1 \rangle$, and now is increased to $o_1$. Assume $A$ does not contain $tr$ and $o_1$ is finite (and this assumption will be adopted for the rest of the paper). Let $h'$ be $h[v/x, o/now] \cdot \langle ch.v', o_1 \rangle$, the rule is presented as follows:
$$\begin{align*}
&A \land \text{now} = o \land tr = h \land x = v; \neg h.ch! \text{ dr } [o, o_1) \land h.ch! \text{ at } o_1 \} ch?x \\
&A[o/now] \land \text{now} = o_1 \land \exists v'.(x = v' \land tr = h'); h.ch? \text{ dr } [o, o_1]
\end{align*}$$
A communication event is equivalent to a sequential composition of a wait statement and an assignment, both of which are deterministic. Thus, as shown above, the formulas related to traces and readiness hold with probability 1.
If such finite $o_1$ does not exist, i.e., the compatible output event will never become available. As a consequence, the input event will keep waiting forever, as shown by the following rule:
\[
\{A \land now = o \land tr = h; -h.ch! \ dr [o, \infty)\} ch?x \\
\{A[o/now] \land now = \infty; h.ch? \ dr [o, \infty)\}
\]
**Output.** Similarly, for output $ch!e$, we have one rule for the case when the compatible input event becomes ready in finite time. Thus the communication occurs successfully.
\[
\{A \land now = o \land tr = h; -h.ch? \ dr [o, o_1) \land h.ch? \ at \ o_1\} ch!e \\
\{A[o/now] \land now = o_1 \land tr = h[o/now] \cdot \langle ch.e, o_1 \rangle, h.ch! \ dr [o, o_1]\}
\]
We also have another rule for the case when the compatible input event will never get ready.
\[
\{A \land now = o \land tr = h; (-h.ch?) \ dr [o, \infty)\} ch!e \\
\{A[o/now] \land now = \infty; h.ch! \ dr [o, \infty)\}
\]
**Stochastic Differential Equation.** Let $f$ be a function, and $\lambda > 0, p \geq 0$ are real values. We have the following rule for $\langle ds = bdt + \sigma dW & B \rangle$.
\[
f(s) \in C^2(\mathbb{R}^n, \mathbb{R}) \text{ has compact support on } B, \lambda, p > 0 \text{ and } \\
A \rightarrow B \rightarrow (f \leq \lambda p) \quad B \rightarrow (f \geq 0) \land (Lf \leq 0)
\]
\[
\{A \land s = s_0 \land now = o; \top\} \langle ds = bdt + \sigma dW & B \rangle \{P(f(s) \geq \lambda) \leq p \land A[s_0/s, o/now] \\
\land now = o + d \land cl(B); B \land P(f(s) \geq \lambda \ dr [o, o + d]) \leq p\}
\]
where $o, s_0$ are logical variables denoting the starting time and the initial value of $s$ resp., $d$ is the execution time of the SDE, and $cl(B)$ returns the closure of $B$, e.g. $cl(x < 2) = x \leq 2$; and the Lie derivative $Lf(s)$ is defined as $\sum_i b_i(s) \frac{\partial f}{\partial s_i}(s) + \frac{1}{2} \sum_{i,j} (\sigma(s) \sigma(s)^T)_{i,j} \frac{\partial^2 f}{\partial s_i \partial s_j}(s)$. The rule states that, if the initial state of the SDE satisfies $f \leq \lambda p$, and in the domain $B$, $f$ is always non-negative and $Lf$ is non-positive, then during the whole evolution of the SDE, the probability of $f(s) \geq \lambda$ is less than or equal to $p$; on the other hand, during the evolution, the domain $B$ holds almost surely, while at the end, the closure of $B$ holds almost surely.
**Sequential Composition.** For $P; Q$, we use $o$ to denote the starting time, and $o_1$ the termination time of $P$, if $P$ terminates, which is also the starting time of $Q$. The first rule is for the case when $P$ terminates.
\[
\{A \land now = o; E\} P \{R_1 \land now = o_1; C_1\} \quad \{R_1 \land now = o_1; C_1\} Q \{R; C\}
\]
\[
\{A; E\} P; Q \{R; C\}
\]
On the other hand, if $P$ does not terminate, the effect of executing $P; Q$ is same to that of executing $P$ itself.
\[
\{A \land now = o; E\} P \{R \land now = \infty; C\}
\]
\[
\{A \land now = o; E\} P; Q \{R \land now = \infty; C\}
\]
**Conditional.** There are two rules depending on whether $B$ holds or not initially.
\[
\frac{A \Rightarrow B}{\{A; E\} P \{R; C\}} \quad \text{and} \quad \frac{A \Rightarrow \neg B}{\{A; \top\} B \rightarrow P \{A; \top\}}
\]
**Probabilistic Choice.** The rule for $P \sqcup_p Q$ is defined as follows:
\[
\frac{\{A \land \text{now} = o; E\} P \{P(S) \bowtie_1 p_1; P(\varphi) \bowtie_2 p_2\} \quad \{A \land \text{now} = o; E\} Q \{P(S) \bowtie_1 q_1; P(\varphi) \bowtie_2 q_2\}}{\{A \land \text{now} = o; E\} P \sqcup_p Q \{P(S) \bowtie_1 pp_1 + (1 - p)q_1; P(\varphi) \bowtie_2 pp_2 + (1 - p)q_2\}}
\]
where $\bowtie_1, \bowtie_2$ are two relational operators. The final postcondition indicates that, if after $P$ executes $S$ holds with probability $\bowtie_1 p_1$, and after $Q$ executes $S$ holds with probability $\bowtie_1 q_1$, then after $P \sqcup_p Q$ executes, $S$ holds with probability $\bowtie_1 pp_1 + (1 - p)q_1$; The history formula can be understood similarly.
**Communication Interrupt.** We define the rule for the special case $\langle ds = bdt + \sigma dW & B \rangle \supseteq (ch?x \rightarrow Q)$ for simplicity, which can be generalized to general case without any difficulty. We use $o_F$ to denote the execution time of the SDE. The premise of the first rule indicates that the compatible event (i.e. $h.ch!$) is not ready after the continuous terminates. For this case, the effect of executing the whole process is thus equivalent to that of executing the SDE.
\[
\frac{\{A \land \text{now} = o; E\} \langle ds = bdt + \sigma dW & B \rangle \{R \land \text{now} = o + o_F; C\} \quad A \land \text{now} = o \land E \Rightarrow (tr = h \land \neg h.ch! \ dr [o, o + o_F])}{\{A \land \text{now} = o; E\} \langle ds = bdt + \sigma dW & B \rangle \supseteq (ch?x \rightarrow Q) \{R \land \text{now} = o + o_F; C\}}
\]
In contrary, when the compatible event gets ready before the continuous terminates, the continuous will be interrupted by the communication, which is then followed by $Q$. Thus, as shown in the following rule, the effect of executing the whole process is equivalent to that of executing $ch?x; Q$, plus that of executing the $SDE$ before the communication occurs, i.e. in the first $o_1$ time units.
\[
\frac{\{A \land \text{now} = o; E\} \langle ds = bdt + \sigma dW & B \rangle \{R \land \text{now} = o + o_F; C\} \quad (A \land \text{now} = o \land E) \Rightarrow (tr = h \land h.ch! \ at (o + o_1) \land o_1 \leq o_F) \quad \{A \land B \land \text{now} = o; E\} ch?x; Q \{R_1; C_1\}}{\{A \land \text{now} = o; E\} \langle ds = bdt + \sigma dW & B \rangle \supseteq (ch?x \rightarrow Q) \{R_1; R|_{[o, o + o_1]} \land C_1\}}
\]
where $R|_{[o, o + o_1]}$ extracts from $R$ the formulas before $o + o_1$, e.g., $(P(S \text{ at } T) \bowtie p)|_{[o, o + o_1]}$ is equal to $P(S \text{ at } T) \bowtie p$ if $T$ is less or equal to $o + o_1$, and true otherwise.
**Parallel Composition**
For $P \| Q$, let $X$ be $X_1 \cap X_2$ where $X_1 = \Sigma(P)$ and $X_2 = \Sigma(Q)$, then
\[
A \Rightarrow A_1 \land A_2, \quad \{A_1 \land \text{now} = o; E_1\} P \{R_1 \land tr = \gamma_1 \land \text{now} = o_1; C_1\} \quad \{A_2 \land \text{now} = o; E_2\} Q \{R_2 \land tr = \gamma_2 \land \text{now} = o_2; C_2\} \quad \forall ch \in X.(C_1[o_1/\text{now}] \upharpoonright_{ch} \Rightarrow E_2 \upharpoonright_{ch}) \land (C_2[o_2/\text{now}] \upharpoonright_{ch} \Rightarrow E_1 \upharpoonright_{ch}) \quad \forall dh \in X_1 \setminus X.E \upharpoonright_{dh} \Rightarrow E_1 \upharpoonright_{dh} \quad \forall dh' \in X_2 \setminus X.E \upharpoonright_{dh'} \Rightarrow E_2 \upharpoonright_{dh'}
\]
\[
\{A \land \text{now} = o; E\} P \| Q \{R; C'_1 \land C'_2\}
\]
where $A_1$ is a property of $P$ (i.e., it only contains variables of $P$), $A_2$ a property of $Q$, and $o_1$ and $o_2$, $\gamma_1$ and $\gamma_2$ logical variables representing the time and trace at termination of $P$ and $Q$ respectively. Let $o_m$ be $\max\{o_1, o_2\}$, $R$, $C'_1$ and $C'_2$ are defined as follows:
$$
\begin{align*}
R & \overset{\text{def}}{=} R_1[\gamma_1 / tr, o_1 / now] \land R_2[\gamma_2 / tr, o_2 / now] \land now = o_m \land \gamma_1 \upharpoonright_X = \gamma_2 \upharpoonright_X \land tr = \gamma_1 \gamma_2 \\
C'_i & \overset{\text{def}}{=} C_i[o_i / now] \land R'_i[o_i / now] \ dr [o_i, o_m] \text{ for } i = 1, 2
\end{align*}
$$
where for $i = 1, 2$, $R_i \Rightarrow R'_i$ but $tr \notin R'_i$. At termination of $P \| Q$, the time will be the maximum of $o_1$ and $o_2$, and the trace will be the alphabetized parallel of the traces of $P$ and $Q$, i.e. $\gamma_1, \gamma_2$. In $C'_1$ and $C'_2$, we specify that none of variables of $P$ and $Q$ except for $now$ and $tr$ will change after their termination.
**Repetition.** For $P^*$, let $k$ be an arbitrary non-negative integer, then $(tr \notin A)$
$$
\begin{align*}
\{A \land now = o + k * t \land tr = (h \cdot \alpha^k); E[o / now]\} & P \\
\{A \land now = o + (k + 1) * t \land tr = (h \cdot \alpha^{k+1}); C\}
\end{align*}
$$
$$
\{A \land now = o \land tr = h; E\} \ P^* \{A \land now = o' \land tr = (h \cdot \alpha^*) + \tau; C \lor (o = o' \text{ at now})\}
$$
$t$ and $\alpha$ are logical variables representing the time elapsed and trace accumulated respectively by each execution of $P$, and $o$ and $o'$ denote the starting and termination time of the loop ($o'$ could be infinite).
The general rules that are applicable to all processes, such as Monotonicity, Case Analysis, and so on, are similar to the traditional Hoare Logic. We will not list them here for page limit.
**Theorem 3 (Soundness).** If $\vdash \{A; E\} P \{R; C\}$, then $\models \{A; E\} P \{R; C\}$, i.e. every theorem of the proof system is valid.
**Proof.** The proof of this theorem can be found at [17].
**Example 1.** For the aircraft example, define $f(x, y)$ as $|y|$, assume $f(x_s, y_0) = |y_0| \leq \lambda p$, where $p \in [0, 1]$. Obviously, $B \rightarrow (f \geq 0) \land (Lf \leq 0)$ holds. By applying the inference rule of SDE, we have the following result:
$$
\{now = o; True\} P_{Air} \{ \exists d. now = o + d \land B \land P(f \geq \lambda) \leq p; \\
B \land P(f \geq \lambda) \ dr [o, o + d]) \leq p \}
$$
which shows that, the probability of the aircraft entering the dangerous state is always less than or equal to $p$ during the flight. Thus, to guarantee the safety of the aircraft, $p$ should be as little as possible. For instance, if the safety factor of the aircraft is required to be 99.98%, then $p$ should be less than or equal to 0.0002, and in correspondence, $|y_0| \leq \frac{\lambda}{5000}$ should be satisfied.
## Conclusion
This paper presents stochastic HCSP (SHCSP) for modelling hybrid systems with probability and stochasticity. SHCSP is expressive but complicated with interacting discrete, continuous and stochastic dynamics. We have defined the semantics of stochastic HCSP and proved that it is well-defined with respect to stochasticity. We propose an assertion language for specifying time-related and
probability-related properties of SHCSP, and have proved the measurability of it. Based on the assertion language, we define a compositional Hoare Logic for specifying and verifying SHCSP processes. The logic is an extension of traditional Hoare Logic, and can be used to reason about how the probability of a property changes with respect to the execution of a process. To illustrate our approach, we model and verify a case study on a flight planning problem at the end.
References
1. Abate, A., Prandini, M., Lygeros, J., Sastry, S.: Probabilistic reachability and safety for controlled discrete time stochastic hybrid systems. Automatica 44(11), 2724–2734 (2008)
2. Altman, E., Gaitsgory, V.: Asymptotic optimization of a nonlinear hybrid system governed by a Markov decision process. SIAM Journal of Control and Optimization 35(6), 2070–2085 (1997)
3. Bujorianu, M.L.: Extended stochastic hybrid systems and their reachability problem. In: Alur, R., Pappas, G.J. (eds.) HSCC 2004. LNCS, vol. 2993, pp. 234–249. Springer, Heidelberg (2004)
4. Bujorianu, M.L., Lygeros, J.: Toward a general theory of stochastic hybrid systems. In: Blom, H.A.P., Lygeros, J. (eds.) Stochastic Hybrid Systems. LNCIS, vol. 337, pp. 3–30. Springer, Heidelberg (2006)
5. Bujorianu, M.L., Lygeros, J., Bujorianu, M.C.: Bisimulation for general stochastic hybrid systems. In: Morari, M., Thiele, L. (eds.) HSCC 2005. LNCS, vol. 3414, pp. 198–214. Springer, Heidelberg (2005)
6. Fränzle, M., Hahn, E.M., Hermanns, H., Wolovick, N., Zhang, L.: Measurability and safety verification for stochastic hybrid systems. In: HSCC 2011, pp. 43–52. ACM (2011)
7. Hahn, E.M., Hartmanns, A., Hermanns, H., Katoen, J.: A compositional modelling and analysis framework for stochastic hybrid systems. Formal Methods in System Design 43(2), 191–232 (2013)
8. Hahn, E.M., Hermanns, H., Wachter, B., Zhang, L.: PASS: abstraction refinement for infinite probabilistic models. In: Esparza, J., Majumdar, R. (eds.) TACAS 2010. LNCS, vol. 6015, pp. 353–357. Springer, Heidelberg (2010)
9. He, J.: From CSP to hybrid systems. In: A Classical Mind, Essays in Honour of C.A.R. Hoare, pp. 171–189. Prentice Hall International (UK) Ltd. (1994)
10. Henzinger, T.A.: The theory of hybrid automata. In: LICS 1996, pp. 278–292, July 1996
11. Hoare, C.A.R.: An axiomatic basis for computer programming. Commun. ACM 12(10), 576–580 (1969)
12. Hoare, C.A.R.: Communicating Sequential Processes. Prentice-Hall (1985)
13. Hu, J., Lygeros, J., Sastry, S.S.: Towards a theory of stochastic hybrid systems. In: Lynch, N.A., Krogh, B.H. (eds.) HSCC 2000. LNCS, vol. 1790, p. 160. Springer, Heidelberg (2000)
14. Liu, J., Lv, J., Quan, Z., Zhan, N., Zhao, H., Zhou, C., Zou, L.: A calculus for hybrid CSP. In: Ueda, K. (ed.) APLAS 2010. LNCS, vol. 6461, pp. 1–15. Springer, Heidelberg (2010)
15. Meseguer, J., Sharykin, R.: Specification and analysis of distributed object-based stochastic hybrid systems. In: Hespanha, J.P., Tiwari, A. (eds.) HSCC 2006. LNCS, vol. 3927, pp. 460–475. Springer, Heidelberg (2006)
16. Morgan, C., McIver, A., Seidel, K., Sanders, J.W.: Refinement-oriented probability for CSP. Formal Asp. Comput. 8(6), 617–647 (1996)
17. Peng, Y., Wang, S., Zhan, N., Zhang, L.: Extending hybrid CSP with probability and stochasticity. Technical report, Institute of Software, Chinese Academy of Sciences (2015). http://arxiv.org/abs/1509.01660
18. Platzer, A.: Stochastic differential dynamic logic for stochastic hybrid programs. In: Bjørner, N., Sofronie-Stokkermans, V. (eds.) CADE 2011. LNCS, vol. 6803, pp. 446–460. Springer, Heidelberg (2011)
19. Prandini, M., Hu, J.: Application of reachability analysis for stochastic hybrid systems to aircraft conflict prediction. In: 47th IEEE Conference on Decision and Control (CDC), pp. 4036–4041. IEEE (2008)
20. Sproston, J.: Decidable model checking of probabilistic hybrid automata. In: Joseph, M. (ed.) FTRTFT 2000. LNCS, vol. 1926, p. 31. Springer, Heidelberg (2000)
21. Zhang, L., She, Z., Ratschan, S., Hermanns, H., Hahn, E.M.: Safety verification for probabilistic hybrid systems. In: Touili, T., Cook, B., Jackson, P. (eds.) CAV 2010. LNCS, vol. 6174, pp. 196–211. Springer, Heidelberg (2010)
22. Zhou, C., Wang, J., Ravn, A.P.: A formal description of hybrid systems. In: Alur, R., Sontag, E.D., Henzinger, T.A. (eds.) HS 1995. LNCS, vol. 1066. Springer, Heidelberg (1996)
23. Zuliani, P., Platzer, A., Clarke, E.M.: Bayesian statistical model checking with application to stateflow/simulink verification. Formal Methods in System Design 43(2), 338–367 (2013) |
Incorporating Prior Knowledge in Fuzzy Model Identification
J. Abonyi†, R. Babuška, H.B. Verbruggen, F. Szeifert*
Delft University of Technology, Department of Information Technology and Systems
Control Engineering Laboratory, P.O.Box 5031 2600 GA Delft, The Netherlands
*University of Veszprem Department of Chemical Engineering Cybernetics
P.O. Box 158, H-8201, Hungary
Abstract
This paper presents an algorithm for incorporating a priori knowledge into data-driven identification of dynamic fuzzy models of the Takagi-Sugeno type. Knowledge about the modelled process such as its stability, minimal or maximal static gain, or the settling time of its step response can be translated into inequality constraints on the consequent parameters. By using input-output data, optimal parameter values are then found by means of quadratic programming. The proposed approach has been applied to the identification of a laboratory liquid level process. The obtained fuzzy model has been used in model-based predictive control. Real-time control results show that when the proposed identification algorithm is applied, not only physically justified models are obtained, but also the performance of the model-based controller improves with regard to the case where no prior knowledge is involved.
1 Introduction
Recent years have witnessed a rapid growth in the use of fuzzy controllers for complex and poorly defined processes. Most fuzzy controllers developed until now are of the rule-based type, where the rules in the controller model the operator’s response in particular process situations.
An alternative approach to the design of fuzzy controllers is the use of more advanced model-based design methods, including the system modeling and identification steps. When the process under control is nonlinear and cannot be described by first principles with sufficient accuracy, it is advantageous to use
†On leave from the University of Veszprem, Department of Chemical Engineering Cybernetics, P.O. Box 158, H-8201, Hungary, sponsored by the Hungarian Ministry of Culture and Education (MKM) Eötvos Foundation.
fuzzy modeling as a way of combining first-principle knowledge, linguistic rules describing the system, and process data.
Fuzzy identification is an effective tool for the approximation of uncertain nonlinear systems on the basis of measured data (Hellendoorn and Driankov, 1997). Data-driven identification techniques alone, however, sometimes yield unrealistic models in terms of steady-state characteristics, local linear behavior or physically impossible parameter values. This is typically due to insufficient information content of the identification data and due to over-parameterization of the models.
The Takagi-Sugeno (TS) fuzzy model is often used to represent nonlinear dynamic systems, by interpolating between local linear, time-invariant (LTI) ARX models. The TS fuzzy model is over-parameterized and when data-driven identification is used, the model can exhibit regimes which are not found in the original system (Babuška, 1998). It is demonstrated in this paper that this problem can be remedied by incorporating prior knowledge into the identification method.
Recently, combinations of a priori knowledge with black-box modeling techniques have been gaining considerable interest. Two different approaches can be distinguished: gray-box modeling and semi-mechanistic modeling. In gray-box modeling, a priori information enters a black-box model, for instance, as constraints on the model parameters or variables, smoothness of the system behavior, or open-loop stability (Tulleken, 1993; Johansen, 1996). One can also start with deriving a model based on first principles and include black-box elements as parts of the white-box model. This approach is usually denoted as hybrid-modeling or semi-physical modeling (Schubert, 1994; Thompson and Kramer, 1994; Psichogios and Ungar, 1992; van Can, et al., 1997).
The main contribution of this article is the development of a gray-box modeling approach for data-driven identification of dynamic Takagi–Sugeno (TS) fuzzy models. The main idea is to constrain the candidate model parameters of the rules in the TS fuzzy model. Knowledge about the process such as stability, minimal or maximal gain, or the settling time are translated into inequality constraints on the parameters. The fuzzy model then can be identified from input-output data by quadratic programming.
The proposed approach is applied to a laboratory liquid level process. A fuzzy model is first obtained from input-output measurements by using the proposed identification technique. The model is then used in model-based predictive control. Real-time control results show that when the gray-box identification algorithm is used, not only physically justified model is obtained, but also the performance of the model-based controller is improved with regard to the case where no prior knowledge is used.
The paper is organized as follows. In Section 2, the applied TS fuzzy model. Section 3 describes
how prior knowledge can be implemented as constraints in the data-driven generation of a fuzzy model. In Section 4, the identification technique is detailed and Section 5 presents the application example. Conclusions are given in Section 6.
2 The Takagi-Sugeno fuzzy model
This paper addresses the identification of fuzzy models with the structure proposed by Takagi and Sugeno (1985). This fuzzy model consists of a set of rules of the following form:
\[ R_{i_1,\ldots,i_n} : \text{If } z_1 \text{ is } A_{1,i_1} \text{ and } \ldots \text{ and } z_n \text{ is } A_{n,i_n} \text{ then } y = f_{i_1,\ldots,i_n}(z_1, \ldots, z_n), \]
(1)
where \( n \) is the number of inputs, \( z = [z_1, \ldots, z_n] \) is a vector containing all the inputs of the fuzzy model and \( A_{j,i_j}(z_j) \) is the \( i_j \)th antecedent fuzzy set for the \( j \)th input. The same symbol is used to denote a fuzzy set and its membership function. \( M_j \) is the number of the fuzzy sets on the \( j \)th input domain.
\( f_{i_1,\ldots,i_n}(z) \) is a (crisp) consequent function. For a given input, \( z \), the output of the fuzzy model, \( y \), is inferred by computing the weighted average:
\[ y = \frac{\sum_{i_1=1}^{M_1} \cdots \sum_{i_n=1}^{M_n} \beta_{i_1,\ldots,i_n} f_{i_1,\ldots,i_n}(z_1, \ldots, z_n)}{\sum_{i_1=1}^{M_1} \cdots \sum_{i_n=1}^{M_n} \beta_{i_1,\ldots,i_n}}, \]
(2)
where the weight, \( \beta_{i_1,\ldots,i_n} > 0 \), is the overall truth value of the \( i_1, \ldots, i_n \)th rule calculated as:
\[ \beta_{i_1,\ldots,i_n} = \prod_{j=1}^{n} A_{j,i_j}(z_j). \]
(3)
Triangular membership functions are used in this article to define the antecedent fuzzy sets as shown in Figure 1, where \( a_{j,i_j} \) denotes the cores of fuzzy sets defined by:
\[ a_{j,i_j} = \text{core}(A_{j,i_j}(z_j)) = \{ z_j | A_{j,i_j}(z_j) = 1 \}. \]
(4)
The support of a set is determined by the cores of the adjacent fuzzy sets:
\[ A_{j,i_j}(z_j) = \frac{z_j - a_{j,i_j-1}}{a_{j,i_j} - a_{j,i_j-1}}, \quad a_{j,i_j-1} \leq z_j < a_{j,i_j} \]
\[ A_{j,i_j}(z_j) = \frac{a_{j,i_j+1} - z_j}{a_{j,i_j+1} - a_{j,i_j}}, \quad a_{j,i_j} \leq z_j < a_{j,i_j+1} \]
(5)
This ensures that the sum of the membership functions is equal to one. Constraints such as this one help to obtain an interpretable, grid-type partitioning rule-bases. The consequent-estimation method presented in this paper is, however, independent of the membership functions used.
As the product operator (3) is used for the ‘and’ connective, the overall truth values of the rules fulfill:
\[
\sum_{i_1=1}^{M_1} \cdots \sum_{i_n=1}^{M_n} \beta_{i_1,\ldots,i_n} = 1 .
\]
(6)
Therefore, (2) can be simplified to:
\[
y = \sum_{i_1=1}^{M_1} \cdots \sum_{i_n=1}^{M_n} \left[ \left( \prod_{j=1}^{n} A_{j,i_j}(z_j) \right) f_{i_1,\ldots,i_n}(z_1,\ldots,z_n) \right] .
\]
(7)
### 3 Fuzzy identification of nonlinear dynamic systems using a priori knowledge
The Nonlinear AutoRegressive with eXogenous input (NARX) model is frequently used within many nonlinear identification methods, such as neural network models and fuzzy models. This model establishes a nonlinear relation between the past inputs and outputs and the predicted output:
\[
y(k+1) = F(y(k),\ldots,y(k-n_y+1),u(k-n_d),\ldots,u(k-n_u-n_d+1)) .
\]
(8)
Here, \( n_y \) and \( n_u \) are the maximum lags considered for the output, and input terms, respectively, \( n_d \) is the discrete dead time, and \( F \) represents the mapping of the fuzzy model.
The TS fuzzy model of the NARX type interpolates between local linear, time-invariant (LTI) ARX models as follows:
\[
R_{i_1,\ldots,i_n} : \text{ If } z_1(k) \text{ is } A_{1,i_1} \text{ and } \ldots \text{ and } z_n(k) \text{ is } A_{n,i_n} \text{ then }
\]
\[
y^{i_1,\ldots,i_n}(k+1) = \sum_{i=1}^{n_y} a_i^{i_1,\ldots,i_n} y(k-i+1) + \sum_{i=1}^{n_u} b_i^{i_1,\ldots,i_n} u(k-i-n_d+1) + c^{i_1,\ldots,i_n} .
\]
(9)
Here, \( z(k) \) is usually a subset of \( \{y(k), \ldots, y(k - n_y + 1), u(k - n_d), \ldots, u(k - n_u - n_d + 1)\} \).
This fuzzy model can be regarded as a linear parameter varying (LPV) system:
\[
\sum_{i=0}^{n_y} a_i y(k - i + 1) = \sum_{i=1}^{n_u} b_i u(k - i - n_d + 1) + c
\]
(10)
with
\[
a_0 = 1
\]
(11)
\[
a_i = - \sum_{i_1=1}^{M_1} \cdots \sum_{i_n=1}^{M_n} \left( \prod_{j=1}^{n} A_{j,i_j}(z_j) \right) a_i^{i_1,\ldots,i_n}, \quad i = 1, \ldots, n_y
\]
(12)
\[
b_i = \sum_{i_1=1}^{M_1} \cdots \sum_{i_n=1}^{M_n} \left( \prod_{j=1}^{n} A_{j,i_j}(z_j) \right) b_i^{i_1,\ldots,i_n}, \quad i = 1, \ldots, n_u
\]
(13)
\[
c = \sum_{i_1=1}^{M_1} \cdots \sum_{i_n=1}^{M_n} \left( \prod_{j=1}^{n} A_{j,i_j}(z_j) \right) c^{i_1,\ldots,i_n}.
\]
(14)
The parameters \( a_i, b_i \) and \( c_i \) are always bounded within the convex sets (polytopes) delimited by the individual rules’ parameters. This follows from equations (11) to (14) and the fact that the membership degrees add up to one, see (6).
It has been shown, that certain types of a priori knowledge about an LTI model can be expressed in the form of linear inequality constraints (Tulleken, 1993; Karny, et al., 1995; Timmons, et al., 1997):
\[
\Lambda_{LTI} \theta_{LTI} \leq \omega_{LTI},
\]
(15)
where \( \theta_{LTI} = [a_1, a_2, \ldots, a_{n_y}, b_1, \ldots, b_{n_u}, c] \) denotes the parameters of the LTI model. These constraints on the LTI model parameters define a convex parameter set \( \Omega \):
\[
\Omega = \{\theta_{LTI} | \Lambda_{LTI} \theta_{LTI} \leq \omega_{LTI}\}.
\]
(16)
The aim of this paper to incorporate important prior knowledge into the fuzzy model (9). The parameter set, \( \theta_{LTI} \), of the LTI model (10) has to be a subset of \( \Omega \). Because of the convexity of \( \Omega \) and the convexity of the applied fuzzy inference method (11)–(14), it is sufficient to check the constraints for the rule consequents. This means that the constraints can easily be adapted to the TS fuzzy model:
\[
\Lambda^* \theta_{i_1,\ldots,i_n} \leq \omega^*,
\]
(17)
where \( \theta_{i_1,\ldots,i_n} = [a_1^{i_1,\ldots,i_n}, a_2^{i_1,\ldots,i_n}, \ldots, a_{n_y}^{i_1,\ldots,i_n}, b_1^{i_1,\ldots,i_n}, \ldots, b_{n_u}^{i_1,\ldots,i_n}, c^{i_1,\ldots,i_n}] \) denotes the parameters of the \( i_1, \ldots, i_n \)th local model, \( \Lambda^* \) and \( \omega^* \) represent global constraints on the fuzzy model. In the following, it will be shown how prior knowledge about the stability, stationary gains and the settling time of the process can be transformed into the linear inequalities (17).
3.1 Prior knowledge about process stability
It is well known that the poles of a stable discrete-time model that emerge from a correctly sampled, continuous-time system cannot be situated in $\mathbb{C}^-$, the left half of the complex plane. This knowledge on sampling can be translated into inequality constraints on the parameters on the model parameters:
\[
(-1)^i a_i \geq 0, \quad 1 \leq i \leq n_y,
\] (18)
Additional constraints can be derived from stability considerations. Let $C(m, R)$ denote the set of complex numbers within or at a circle with a real-valued center $m$ and radius $R$,
\[
C(m, R) = \{ z \in \mathbb{C} \mid \|z - m\| \leq R; \ m \in \mathbb{R}; \ R \in \mathbb{R}^+ \}.
\] (19)
Tulleken (1993) has shown that for the poles of a linear discrete-time system to be in $C(m, R)$, the following linear constraints on the model parameters must be satisfied:
\[
\Re \cdot [1, a_1, a_2, \ldots, a_{n_y}]^T \geq 0,
\] (20)
\[
\Im \cdot \Im \cdot [1, a_1, a_2, \ldots, a_{n_y}]^T \geq 0,
\] (21)
where the non-zero elements of the right and left triangular matrices $\Re$ and $\Im$ are defined by
\[
[\Re]_{ij} = (-1)^i \binom{j}{i}, \quad i = 0, 1, \ldots, n_y, \ j = 0, 1, \ldots, n_y,
\] (22)
\[
[\Im]_{ij} = (m - R)^{i-j} \binom{n_y - j}{i - j} (2R)^{n_y - i}.
\] (23)
If $m = 0$ and $R = 1$, the previous equation constitutes the smallest convex hull of the admissible parameter region corresponding to the stable system model having all poles in the unit circle, $C(0, 1)$, in the complex plane. Therefore, these are necessary conditions for asymptotic stability. More details about the derivation of the above constraints can be found in (Tulleken, 1993).
Example 1. Let us consider a second-order system with $n_y = 2$ and $n_u = 1$: $y(k + 1) = a_1 y(k) + a_2 y(k - 1) + b_1 u(k) + c$. In this case $\Re \cdot \Im$ becomes:
\[
\begin{bmatrix}
1 & 1 & 1 \\
0 & -1 & -2 \\
0 & 0 & 1
\end{bmatrix}
\begin{bmatrix}
4 & 0 & 0 \\
-4 & 2 & 0 \\
1 & -1 & 1
\end{bmatrix}
=
\begin{bmatrix}
1 & 1 & 1 \\
2 & 0 & -2 \\
1 & -1 & 1
\end{bmatrix}.
\] (24)
It can be seen from the coefficients of the above matrices and equation (20) that for the given second-order system the following holds:
\[
1 + a_1 + a_2 \geq 0,
\]
\[
-a_1 - 2a_2 \geq 0,
\]
\[
a_2 \geq 0.
\]
From (21) and (24), the following inequality constraints are obtained:
\[
1 + a_1 + a_2 \geq 0,
\]
\[
2 - 2a_2 \geq 0,
\]
\[
1 - a_1 + a_2 \geq 0.
\]
### 3.2 Knowledge about process gain
Often, not only the stability of the plant is known beforehand, but also the admissible intervals or at least the signs of the static gains are known. Since the steady-state gain is only defined for stable systems, open-loop stability must also be imposed on the estimates (see the Section 3.1). Given the minimal and maximal gains \( K_{\text{min}} \) and \( K_{\text{max}} \), for the steady-state gain of a linear system, the following holds:
\[
K_{\text{min}} \leq \frac{\sum_{i=1}^{n_b} b_i}{1 + \sum_{i=1}^{n_a} a_i} \leq K_{\text{max}}.
\]
(25)
This knowledge can be represented by the following linear inequality constrains:
\[
K_{\text{min}} \left( 1 + \sum_{i=1}^{n_a} a_i \right) - \sum_{i=1}^{n_b} b_i \leq 0,
\]
(26)
\[
-K_{\text{max}} \left( 1 + \sum_{i=1}^{n_a} a_i \right) + \sum_{i=1}^{n_b} b_i \leq 0.
\]
(27)
### 3.3 Knowledge about the settling time
The approximate settling time (the time required for a step response to settle within a band around its final value) is usually known. For second-order discrete-time systems, a known maximum settling time \( \tau_s \) approximately translates to all poles lying within a circle centered at the origin in the \( z \)-plane with the radius \( r = \exp \left( -4.6 \frac{T_s}{T} \right) \), where \( T_s \) denotes the sampling time. Based on this consideration, the following inequality constraints can be developed, based on the method presented by Timmons, et al. (1997):
\[
r_{\text{max}}^2 a_1 + (r_{\text{min}} - 2r_{\text{max}})(-a_2) \leq -r_{\text{max}}^2 r_{\text{min}},
\]
(28)
\[-r_{\text{max}} a_1 - a_2 \leq r_{\text{max}}^2,\]
\[-a_2 \leq 0.\]
### 3.4 Knowledge about the nonlinearity
For nonlinear processes, besides the bounds on the gain and/or the settling time, the tendency of the change of these parameters may be known as well.
**Example 2.** Let us consider a nonlinear liquid level process (Posthlethwaite, 1994):
\[
A \frac{dy}{dy} = u - \alpha \sqrt{y},
\]
(29)
where \(A\) is the cross-sectional area of the tank, \(y\) is the liquid level in the tank, \(u\) is the inlet flow-rate, and \(\alpha\) is the outflow coefficient. The gain and the time constant of the process can be obtained by linearizing the process model: \(K = \frac{2\sqrt{\pi}}{\alpha}, \tau = \frac{2A\sqrt{\pi}}{\alpha}\). The gain and the time constant thus increase when the level of the liquid in the tank increases.
This kind of a priori knowledge can be included in the TS fuzzy model identification by using local constraints instead of the global ones given in equation (17):
\[
\Lambda_{i_1, \ldots, i_n} \theta_{i_1, \ldots, i_n} \leq \omega_{i_1, \ldots, i_n}.
\]
(30)
Here, \(\Lambda_{i_1, \ldots, i_n}\) and \(\omega_{i_1, \ldots, i_n}\) are the constraints of the \(i_1, \ldots, i_n\)th rule parameter vector \(\theta_{i_1, \ldots, i_n}\). By using local constraints, the prior information on the nonlinear behaviour around a given operating point can be used in the identification procedure, while the global constraints (17) represent knowledge that is independent of the operating point.
### 4 Constrained identification of TS fuzzy models
The output of the Takagi-Sugeno fuzzy model is linear in the consequent parameters. In the unconstrained case (no prior knowledge used), these parameters can be estimated by linear least-squares techniques. A global or a local approach can be followed. With the global approach the parameters of all rule consequents are estimated within one least-squares problem, yielding an optimal predictor. Due to its cubic complexity, global parameter estimation becomes computationally expensive for fuzzy systems with many rules (Fischer and Nelles, 1998). The local parameter estimation approach does not estimate all parameters simultaneously. It rather divides this task into \(\prod_{j=1}^{n} M_j\) weighted least-squares problems. This method forces the local linear models to fit the data locally, but it does not give an optimal TS
model in terms of a minimal global prediction error. As the identified dynamic fuzzy model will be used in model-based predictive control, it is important that it gives the least possible prediction error. Therefore, the global estimation approach is pursued in this paper.
The $N$ data pairs and the truth values of the fuzzy rules are arranged in the following matrices.
$$\mathbf{Z} = \begin{bmatrix} \mathbf{z}^1 \\ \mathbf{z}^2 \\ \vdots \\ \mathbf{z}^N \end{bmatrix}, \quad \mathbf{y} = \begin{bmatrix} y^1 \\ y^2 \\ \vdots \\ y^N \end{bmatrix}, \quad \mathcal{B}_{i_1,\ldots,i_n} = \begin{bmatrix} \beta_{i_1,\ldots,i_n}^1 & 0 & \cdots & 0 \\ 0 & \beta_{i_1,\ldots,i_n}^2 & \cdots & 0 \\ \vdots & \vdots & \ddots & \vdots \\ 0 & 0 & \cdots & \beta_{i_1,\ldots,i_n}^N \end{bmatrix}. \tag{31}$$
Appending a unitary column to $\mathbf{Z}$ gives the extended regression matrix $\mathbf{Z}_e = [\mathbf{Z} \ 1]$. Denote $\mathbf{Z}'$ the matrix composed of matrices $\mathcal{B}_{i_1,\ldots,i_n}$ and $\mathbf{Z}_e$ as follows:
$$\mathbf{Z}' = [\mathcal{B}_{1,\ldots,1,1}, \mathbf{Z}_e, \mathcal{B}_{1,\ldots,1,2}, \mathbf{Z}_e, \ldots, \mathcal{B}_{M_1,\ldots,M_n}, \mathbf{Z}_e] \tag{32}$$
Denote $\theta'$ the vector given by $\theta' = [\theta_{1,\ldots,1,1}, \theta_{1,\ldots,1,2}, \ldots, \theta_{M_1,\ldots,M_n}]$. The least-squares method can be applied to solve the underlying regression problem $\mathbf{y} = \mathbf{Z}_e \theta' + \varepsilon$:
$$\theta' = \left[ (\mathbf{Z}')^T \mathbf{Z}' \right]^{-1} (\mathbf{Z}')^T \mathbf{y}. \tag{33}$$
In order to be able to use the constraints derived in the previous section, quadratic programming (QP) has to be used instead of the least-squares method.
The constrained optimization problem can be formulated as follows:
$$\min_{\theta'} \left\{ \frac{1}{2} (\theta')^T \mathbf{H} \theta' + \mathbf{c}^T \theta' \right\} \tag{34}$$
with $\mathbf{H} = 2 (\mathbf{Z}')^T \mathbf{Z}'$, $\mathbf{c} = -2 (\mathbf{Z}')^T \mathbf{y}$ and the constraints on $\theta'$:
$$\Lambda' \theta' \leq \omega' \tag{35}$$
where $\Lambda' = [\Lambda_{1,\ldots,1,1}, \Lambda_{1,\ldots,1,2}, \ldots, \Lambda_{M_1,\ldots,M_n}]$ and $\omega' = [\omega_{1,\ldots,1,1}, \omega_{1,\ldots,1,2}, \ldots, \omega_{M_1,\ldots,M_n}]$
### 5 Example: identification for model predictive control of a liquid level process
This real-time control example is used to illustrate the advantages of the proposed identification method. A laboratory process consisting of two cascaded tanks is considered. Fuzzy models are constructed from input-output data and from different pieces of prior knowledge. The obtained models are used in a predictive control scheme. It is shown that by using prior knowledge a more reliable control model is obtained.
5.1 Process description
The laboratory process consists of two cascaded tanks depicted in Figure 2. Water is supplied into the upper tank through a controlled peristaltic pump. A pressure transmitter attached to the bottom of the lower tank measures the level of the liquid in this tank. The process is connected to a personal computer through a data acquisition board. The identification and control software is written under MATLAB, using the Real-Time Toolbox to collect data and to control the process. The sampling time is 2.5 s.
5.2 Fuzzy model identification
The TS model was constructed from process measurements. The input variables of the fuzzy model were selected on the basis of prior knowledge about of the process. Since the system can be modelled approximately as a second-order system with a time delay, and the source of the nonlinearity is the level-dependent outflow from the tank, the antecedent variable of the fuzzy model is chosen to be $z(k) = y(k)$. This results in the following TS fuzzy model structure:
$$R_i : \text{If } y(k) \text{ is } A_{1,i} \text{ then } y^i(k + 1) = a_1^i y(k) + a_2^i y(k - 1) + b_1^i u(k - 2) + c^i.$$ \hspace{1cm} (36)
Six triangular fuzzy sets were used on the antecedent universe $y(k)$. Based on the range of the liquid level and after some manual optimization, the cores of the fuzzy sets were selected to be $\{0.0781 \ 0.2000 \ 0.4000 \ 0.5000 \ 0.6000 \ 0.8081\}$. The identification data set contains $N = 1400$ samples, using an input signal shown in Figure 3.\footnote{In this example, the process input and output signals are plotted in percentage of their full range.} The signal was designed to contain the important frequencies in the \textit{expected} range of the process dynamics. Later on, we will see that the frequency content of this signal is not ideal, which is typical for real-life experiments.
Three different models were identified:
Model 1 No prior knowledge was used during the identification.
Model 2 Prior knowledge on process stability was used and the minimal and maximal steady-state gain were assumed: $K_{\min} = 0$ and $K_{\max} = 2.5$.
Model 3 The second case was extended by the following constraints on the settling time of each submodel (Table 1).
The developed fuzzy models were validated by means of (recurrent) simulation on a separate validation data set. The simulated and the measured liquid levels are shown in Figure 4. The two curves can hardly be distinguished from each other.
The performance of the obtained models was measured by the variance accounted for (VAF) index given by:
\[
VAF = 100\% \left[ 1 - \frac{var(y - y_m)}{var(y)} \right],
\]
where \(y\) is the true process output and \(y_m\) is the model output obtained by simulation of the fuzzy model. The VAF of two equal signals is 100 %, if the signals differ, the VAF value is lower.
Models 1 and 2 both achieve the performance VAF =99.89 and Model 3 has VAF=99.9. Clearly, there is no significant difference between the performance of the fuzzy models identified with and without the use of a priori knowledge. It is interesting to note that even though the modeling performance is
almost unchanged, the consequent parameters of the three models differ considerably (Table 2).
**Table 2. Parameters of the local models obtained by different identification method.**
| | Model 1 | | | | | |
|----------------|---------|-------|-------|-------|-------|-------|
| | 1 | 2 | 3 | 4 | 5 | 6 |
| $a^i_1$ | -0.2419 | -0.0398 | -0.0217 | -0.0238 | 0.0375 | 0 |
| $a^i_2$ | -0.4630 | -0.6574 | -0.6499 | -0.6470 | -0.7032 | -0.6680 |
| $b^i_1$ | 0.0217 | 0.0691 | 0.0350 | 0.0392 | 0.0325 | 0.0384 |
| $c^i$ | 0.1315 | 0.3112 | 0.6487 | 0.8094 | 0.9753 | 1.3141 |
| | Model 2 | | | | | |
|----------------|---------|-------|-------|-------|-------|-------|
| | 1 | 2 | 3 | 4 | 5 | 6 |
| $a^i_1$ | 1.1271 | 1.3291 | 1.3472 | 1.3452 | 1.4064 | 1.3689 |
| $a^i_2$ | -0.4630 | -0.6574 | -0.6499 | -0.6470 | -0.7032 | -0.6680 |
| $b^i_1$ | 0.0217 | 0.0691 | 0.0350 | 0.0392 | 0.0325 | 0.0384 |
| $c^i$ | 0.0246 | 0.0374 | 0.1011 | 0.1250 | 0.1539 | 0.2079 |
| | Model 3 | | | | | |
|----------------|---------|-------|-------|-------|-------|-------|
| | 1 | 2 | 3 | 4 | 5 | 6 |
| $a^i_1$ | 1.3632 | 1.5495 | 1.6560 | 1.6134 | 1.6827 | 1.6422 |
| $a^i_2$ | -0.4646 | -0.6047 | -0.6815 | -0.6407 | -0.7049 | -0.6668 |
| $b^i_1$ | 0.0297 | 0.0718 | 0.0333 | 0.0394 | 0.0324 | 0.0384 |
| $c^i$ | 0.0055 | -0.0182 | -0.0088 | -0.0124 | -0.0107 | -0.0139 |
This effect is caused by the fact that the unconstrained global least-squares estimation method biases the estimate of the local model parameters, that may hamper the local interpretation of the TS fuzzy model, while the prior-knowledge-based constraints regularize the model parameters. If the incorporated constraints are adequate, this regularization does not result in worse modeling performance. Moreover, if the identification data do not contain all information about the process (which is the usual case), the constraints can even improve the modeling performance.
Figure 5 shows three measured step responses of the process and the step responses of the fuzzy
model at these operating points.
**Figure 5.** Step responses of the fuzzy models (--- model 1, - - model 2, -- model 3) and the measured step response at the corresponding operating points (dotted curve).
The generated step responses clearly show the above-mentioned regularization effect of the constraints. One can see that by using more prior knowledge, more accurate local models can be identified from the global process data. The large deviation in the step response and the small difference in the modeling performance is mainly caused by the offset parameters, $c^i$, of the rules. By using the proposed constrained identification method, the offset parameters influence the steady-state behavior of the plant only (as one would normally expect). The reason why the constraints on the settling time give the most significant improvement is that the identification data do not contain enough information on low-frequency and steady-state behavior of the system. This is a typical real-life situation, as it is often difficult to design identification experiments that yield data with an appropriate balance between low and high frequencies. An advantage of the proposed method is that the information on low-frequency behavior can originate from other sources or separate experiments (step responses) and it can easily be integrated with the information coming from the dynamic data.
5.3 Fuzzy model-based predictive controller
A generalized predictive controller (GPC) has been designed for the considered process, using the fuzzy models obtained with the proposed identification method. The GPC algorithm computes the control sequence \(\{u(k + j)\}\) such that the following quadratic cost function (Clarke, et al., 1989) is minimized:
\[
J(H_{p1}, H_{p2}, H_c, \lambda) = \sum_{j=H_{p1}}^{H_{p2}} (w(k+j) - \hat{y}(k+j))^2 + \lambda \sum_{j=1}^{H_c} \Delta u^2(k+j-1).
\]
(38)
Here, \(\hat{y}(k+j)\) denotes the predicted process output, \(H_{p1}\) is the minimum costing horizon, \(H_{p2}\) is the maximum costing or prediction horizon, \(H_c\) is the control horizon, and \(\lambda\) is a weighting coefficient. The predicted outputs are calculated by using the following approximate step-response model:
\[
\hat{y}(k+j) = \sum_{i=1}^{j} g_i \Delta u(k+j-1) + p_j,
\]
(39)
where \(g_i\) represent the plant step-response coefficients, obtained from the fuzzy model at the current operating point, and \(p_j\) is the response of the fuzzy model at the \(k+j\)th step assuming that the future control signal remains constant. By combining all predicted outputs into the vector \(\hat{y} = [\hat{y}(k+H_{p1}), \ldots, \hat{y}(k+H_{p2})]\), the key equation of the GPC algorithm can be formulated:
\[
\hat{y} = G \Delta \pi + p
\]
(40)
Here, \(\Delta \pi = [\Delta u(k), \ldots, \Delta u(k+H_c)]\), and \(p = [p_1, p_2, \ldots, p_{H_{ps}}]\), where the elements of \(p\) are obtained by using the fuzzy model as follows:
\[
\begin{align*}
p_1 &= f(g(k), g(k-1), u(k-2)), \\
p_2 &= f(p_1, y(k), u(k-1)), \\
p_j &= f(p_{j-1}, p_{j-2}, u(k-1)), \quad \text{for } j > 2.
\end{align*}
\]
(41)
\(G\) is a \((H_{p2} - H_{p1} + 1) \times H_c\) matrix with zero entries \(g_{ij}\) for \(j-i > H_{p1}\):
\[
G = \begin{bmatrix}
g_{H_{p1}} & g_{H_{p1}-1} & \cdots & 0 \\
g_{H_{p1}+1} & g_{H_{p1}} & g_{H_{p1}-1} & \\
\vdots & & \ddots & \\
g_{H_{p2}} & g_{H_{p2}-1} & \cdots & g_{H_{p2}-H_c}
\end{bmatrix}.
\]
(42)
The elements of this matrix are the step-response coefficients obtained from the fuzzy model at the current operating point:
\[
g_j = 0, \forall j \leq n_d,
\]
(43)
\[ g_j = -\sum_{i=1}^{j} a_i g_{j-i} + \sum_{i=1}^{j} b_i, \quad \text{for } j > n_d. \]
When constraints are considered, the minimum of (38) can be found by quadratic optimization with linear constraints:
\[
\min_{\overline{u}} \left\{ (G \Delta \overline{u} + p - w)^T (G \Delta \overline{u} + p - w) + \lambda \Delta \overline{u}^T \Delta \overline{u} \right\} = \min_{\overline{u}} \left\{ \frac{1}{2} \Delta \overline{u}^T H \Delta \overline{u} + d \Delta \overline{u} \right\}
\]
(44)
with \( H = 2 \left( G^T G + \lambda I \right), \ d = -2 \left( G^T (w - p) \right) \), where \( I \) is an \((H_c \times H_c)\) unity matrix.
The constraints on \( u \) and \( \Delta u \) can be formulated with the following inequality:
\[
\begin{pmatrix}
I_{\Delta \overline{u}} \\
-I_{\Delta \overline{u}} \\
I_{H_c} \\
-I_{H_c}
\end{pmatrix} \Delta \overline{u} \leq
\begin{pmatrix}
u_{\max} - I_{\overline{u}} u(k-1) \\
-u_{\min} + I_{\overline{u}} u(k-1) \\
\Delta u_{\max} \\
-\Delta u_{\min}
\end{pmatrix}
\]
(45)
where \( I_{H_c} \) and \( I_{\overline{u}} \) is an \( H_c \times H_c \) unity matrix, \( I_{\Delta \overline{u}} \) is an \( H_c \times H_c \) lower triangular matrix with all elements equal to one, and \( \Delta u_{\min}, \Delta u_{\max}, u_{\min}, u_{\max} \) are \( H_c \)-vectors, with the constraints \( \Delta u_{\min}, \Delta u_{\max}, u_{\min}, u_{\max} \) respectively.
### 5.4 Control results
In order to compare the control performance of the different fuzzy models, the following quadratic performance index is used:
\[
PI = \sum_{k=1}^{T} \left\{ (w(k) - y(k))^2 + \lambda \Delta u^2(k) \right\},
\]
(46)
where \( T \) is the number of time steps in a control experiment.
The prediction and the control horizons are selected to be \( H_{p1} = 3, \ H_{p2} = 10, \) and \( H_c = 2 \). The move suppression coefficient, \( \lambda = 0.15 \). The constraints are set to \( \Delta u_{\max} = -\Delta u_{\min} = 0.25, \ u_{\max} = 1, \ u_{\min} = 0 \).
In order to cope with the model-plant mismatch and unmeasured disturbances, the GPC was implemented within an internal model control (IMC) scheme (Garcia and Morari, 1982), see Figure 6.
The inputs of the controller are the reference \( w \), the predicted liquid level, \( \hat{y} \), and the filtered modeling error signal \( e_{mf} \). The following first-order low-pass Butterworth filter was used:
\[
e_{mf}(k) = b_{f1} e_m(k) + b_{f2} e_m(k-1) - a_{f2} e_{mf}(k-1)
\]
(47)
with $b_{f1} = 0.2929$, $b_{f2} = 0.2929$, and $a_{f2} = -0.4142$. The filter parameters were designed based on the desired cut-off frequency in order to reliably filter out the measurement noise and to provide a fast response.
The control performance of the predictive controller based on model 1 (no prior knowledge involved) is shown in Figure 7. The corresponding performance is $PI = 1.2$. Note the oscillations of the control signal, caused by the fact that an inaccurate step-response model is obtained from the fuzzy model (44).
By using prior knowledge, the control performance improves considerably. For model 2, which was identified by using the knowledge about stability and the minimal and maximal steady-state gains, the performance index is $PI = 0.98$. If the knowledge about the approximately settling time is also used, the performance improves even further, yielding $PI = 0.73$. The reason is that by using prior knowledge in the identification, the local linear ARX models in the fuzzy model become more realistic. Therefore, the performance of the model-based controller is much better. The corresponding control result is shown in Figure 8. Note that the control signal becomes less oscillatory. The remaining small fluctuations at low liquid levels are caused by the measurement noise of the level sensor.
6 Conclusions
A new approach to data-driven identification of Takagi–Sugeno fuzzy models has been presented. It allows to translate prior knowledge about the process (including stability, minimal or maximal static gain and settling time) into constraints on the model parameters. This procedure allows the development of TS models also in cases where little experimental data are available. Experimental results have been obtained for a laboratory setup consisting of two cascaded tanks. It has been shown that fuzzy models
Figure 7. Performance of the fuzzy-model-based controller when the fuzzy model is identified without using prior knowledge.
Figure 8. Performance of the fuzzy-model-based controller when the fuzzy model is identified with the use of prior knowledge.
built on the basis of data combined with prior knowledge perform better in control than models obtained from data only. This is because the prior-knowledge-based constraints result in better local modeling of the plant.
References
Babuška, R. (1998). *Fuzzy Modeling for Control*. Boston, USA: Kluwer Academic Publishers.
Can, van, H.J.L., H.A.B. te Braake, C. Hellinga, K.Ch.A.M. Luyben and J.J. Heijnen (1997). An efficient model development strategy for bioprocesses based on neural networks in macroscopic balances. *Biotechnology and Bioengineering* 54, 549–566.
Clarke, D. W., C. Mothadi and P.S.C. Tuffs (1989). Generalized predictive control — part I. the basic algorithm. *Automatica* 23, 137–148.
Fischer, M. and O. Nelles (1998). Predictive control based on local linear fuzzy models. *International Journal of Systems Science* 29, 679–697.
Garcia, C.E. and M. Morari (1982). Internal model control: 1. A unifying review and some new results. *Ind. Eng. Chem. Process Res. Dev.* 21, 308–323.
Hellendoorn, H. and D. Driankov (Eds.) (1997). *Fuzzy Model Identification: Selected Approaches*. Berlin, Germany: Springer.
Johansen, T.A. (1996). Identification of non-linear systems using empirical data and a priori knowledge – an optimisation approach. *Automatica* 32, 337–356.
Karny, M., A. Haluskova and P. Nedoma (1995). Recursive approximation by ARX model: A tool for grey-box modelling. *Int. J. Adaptive Control and Signal Processing* 9, 525–546.
Posthlethwaite, B. (1994). A model-based fuzzy controller. *Chem. Eng. Res. Des. (Trans IChemE)* 72, 39–48.
Psichogios, D.C. and L.H. Ungar (1992). A hybrid neural network – first principles approach to process modeling. *AIChE J.* 38, 1499–1511.
Schubert, J. (1994). Bioprocess optimization and control: Application of hybrid modelling. *Journal of Biotechnology* 35, 51–68.
Takagi, T. and M. Sugeno (1985). Fuzzy identification of systems and its application to modeling and control. *IEEE Trans. Systems, Man and Cybernetics* 15(1), 116–132.
Thompson, M.L. and M.A. Kramer (1994). Modeling chemical processes using prior knowledge and neural networks. *AIChE J.* 40, 1328–1340.
Timmons, W.D., H.J. Chizeck and P.G. Katona (1997). Parameter-constrained adaptive control. *Ind. Eng. Chem. Res.* 36, 4894–4905.
Tulleken, H.J.A.F (1993). Gray-box modelling and identification using physical knowledge and Bayesian techniques. *Automatica* 29, 285–308. |
Intermediate-Band Photometry in the Open Cluster NGC 6664
Edward G. Schmidt
University of Nebraska at Lincoln, firstname.lastname@example.org
Follow this and additional works at: http://digitalcommons.unl.edu/physicsschmidt
Part of the Physics Commons
Schmidt, Edward G., "Intermediate-Band Photometry in the Open Cluster NGC 6664" (1982). Edward Schmidt Publications. 39.
http://digitalcommons.unl.edu/physicsschmidt/39
This Article is brought to you for free and open access by the Research Papers in Physics and Astronomy at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Edward Schmidt Publications by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln.
INTERMEDIATE-BAND PHOTOMETRY IN THE OPEN CLUSTER NGC 6664
EDWARD G. SCHMIDT\textsuperscript{a)}
Behlen Observatory, Department of Physics and Astronomy, University of Nebraska, Lincoln, Nebraska 68588-0111
Received 15 March 1982; revised 12 April 1982
ABSTRACT
Four-color and H$\beta$ photometry is given for early-type stars in the open cluster NGC 6664. From these data the true distance modulus is found to be 10.7 and the reddening is found to range from $E(b-y) = 0.48$ to 0.63 across the cluster. The distance modulus agrees well with some values from the literature, but the present color excesses are about 0.1 mag larger than published values. There are two or perhaps three blue stragglers in the cluster and its age is estimated to be $46 \times 10^6$ yr.
I. INTRODUCTION
This is the fifth in a series of papers giving intermediate-band photometry for B stars in clusters containing classical Cepheids. The motivation for this project is discussed in the first paper (Schmidt 1980). The present paper is concerned with the cluster NGC 6664, which contains the 3-day Cepheid EV Scti.
Arp (1958) obtained $UBV$ photometry of stars in NGC 6664 while Kraft (1958) classified the spectra and measured radial velocities for 13 stars in the cluster field. The photometry showed that this cluster is at a distance of approximately 1.5 kpc and that the foreground reddening is highly variable. The radial velocities indicated that the Cepheid EV Sct is a likely member. Its absolute magnitude, as inferred from the cluster distance modulus, also supports its membership. Arp concluded from his $UBV$ photometry that NGC 6664 is slightly older than the Pleiades. However, Harris (1976) places this cluster in her age group II, which makes it significantly younger than the Pleiades.
In addition to the Cepheid, NGC 6664 contains about seven red giants. Of these, two are members according to their radial velocities and spectral types. It is likely that at least some of the other five are also members. Additionally, Arp remarked that there are two stars, F and L in his lettering scheme, that seem to be too close to the main sequence for the evolutionary age of the cluster. Both are possible members on spectral grounds but for neither is the evidence strong. It is possible that these stars are blue stragglers and the present photometry should clarify this point.
II. THE OBSERVATIONS
The photometric study of Arp was used to select stars for the present program. All of the stars he listed with magnitudes brighter than $V = 13.2$ and with $B - V$ less than 0.75 have been included with only two exceptions, stars A and K. The magnitude limit was imposed by the availability of large-telescope time and corresponds to an absolute magnitude of about 0.0. The color limit includes all the stars that appear near the cluster main sequence but excludes apparent red giants for which the intermediate-band photometry is less useful. Of the two stars that were omitted, star A is unlikely to be a member on the basis of its spectroscopic parallax and its radial velocity. The spectroscopic parallax of star K indicates that it is a probable member. However, Kraft indicated that it was double lined and thus probably a spectroscopic binary.
There are 15 stars in our sample. Although it would obviously be desirable to have a larger number, this would require observing fainter stars. Given the rather large reddening, the availability of telescope time, and the requirements for accuracy, this was not practicable. However, the sample we have is large enough to yield accurate estimates of the distance modulus and color excess.
The stars that were observed are listed in Table I. The second column gives the spectral type from Kraft (1958). The visual magnitudes are means of those from Arp and from the present photometry weighted by the number of nights on which they were observed. The mean indices from the present observations are listed in columns 4–6 and 8, while columns 7 and 9 list the number of nights on which a star was observed with the four-color filters and with the H$\beta$ filters, respectively.
The observations were made at Cerro Tololo Inter-American Observatory between April 1979 and May 1981. The 0.9-m and the 1.5-m telescopes were used with a two-channel photometer. One channel served as a monitor of the transparency while all the measurements were made through the other channel. Standard four-color and H$\beta$ filters were used. The observations were referred to the $uvby\beta$ system through the use of secondary standard stars located along the equator. These have in turn been referred to the primary four-color and H$\beta$ standards (Crawford and Mander 1966; Crawford and Barnes 1970). A complete discussion of these secondary standards is in preparation and will be published elsewhere.
Because this cluster is heavily reddened, we must be concerned about the question of reducing the photoTABLE I. Photometric indices of NGC 6664 stars.
| Star | Spectral type | $V$ | $b - y$ | $m_1$ | $c_1$ | $n$ | $\beta$ | $n$ | $E(b - y)$ | $V_0 - M_V$ |
|------|---------------|-------|---------|-------|-------|-----|--------|-----|------------|-------------|
| F | B3IV | 10.97 | 0.517 | −0.101| 0.196 | 3 | 2.644 | 4 | 0.626 | 10.51 |
| G | A0III | 11.02 | 0.436 | −0.052| 0.669 | 7 | 2.688 | 5 | 0.487 | 10.56 |
| I | | 11.79 | 0.564 | +0.043| 0.581 | 3 | 2.703 | 3 | 0.425 | 10.92 |
| J | | 11.84 | 0.472 | −0.068| 0.466 | 3 | 2.692 | 5 | 0.548 | 10.49 |
| L | B3V | 11.94 | 0.495 | −0.058| 0.202 | 2 | 2.648 | 5 | 0.603 | 11.44 |
| M | B9V | 11.99 | 0.466 | −0.067| 0.724 | 3 | 2.734 | 4 | 0.512 | 10.21 |
| N | | 12.42 | 0.523 | −0.063| 0.585 | 3 | 2.705 | 5 | 0.586 | 10.76 |
| O | | 12.44:| 0.551 | −0.059| 0.662 | 3 | 2.723 | 5 | 0.606 | 10.35 |
| P | | 12.65 | 0.524 | −0.052| 0.725 | 3 | 2.731 | 4 | 0.570 | 10.66 |
| Q | | 12.98 | 0.548 | +0.045| 0.706 | 3 | 2.694 | 4 | 0.529 | 12.21 |
| R | | 13.02 | 0.507 | −0.069| 0.709 | 3 | 2.727 | 4 | 0.564 | 11.18 |
| S | | 13.06 | 0.562 | −0.056| 0.808 | 3 | 2.834 | 4 | 0.605 | 9.09 |
| T | | 13.18 | 0.435 | −0.034| 0.732 | 3 | 2.741 | 4 | 0.480 | 11.41 |
| 10 | | 12.54 | 0.552 | −0.062| 0.612 | 2 | 2.696 | 3 | 0.613 | 11.09 |
| 19 | | 12.99 | 0.485 | −0.057| 0.684 | 3 | 2.757 | 4 | 0.535 | 10.46 |
metry to the standard system. In the case of the H$\beta$ index, Schmidt and Taylor (1979) have discussed the effects of color on the transformations. For the 1979 observations, we included a color term as they recommend, but for the more recent observations the filter set was well matched and this was not necessary. Among our secondary standards there are a number of highly reddened B stars. Those which were used in the course of the present observations are listed in Table II. The mean indices obtained from the nights on which NGC 6664 was observed are also tabulated, together with the number of nights for which each standard was used. These values fit the standard values well and no trends with reddening are apparent. However, the establishment of the standard indices poses some difficulty because the primary standards of the four-color system contain no stars that are reddened nearly as much as most of the stars in Table II. The question of how close the indices for these stars are to the uvby system will be discussed fully in the paper describing the standards. However, we can perform a preliminary test by deriving their color excesses and comparing them with independently determined values. In column 6 of the table we give the color excesses from Crawford's calibration. Spectral types and $B - V$ colors can be used to obtain another estimate of the reddening. These values have been multiplied by 0.74 to put them on the scale of the $b - y$ colors that are tabulated in column 8 of the table. It can be seen that the differences [column 9] between the two completely independent estimates of the color excesses of these stars are small and, more importantly show no trend with reddening. We therefore conclude that the four-color photometry of these stars is accurately on the uvby system.
Because some of the stars in this cluster are relatively faint for intermediate-band photometry and because the substantial reddening reduces the ultraviolet flux considerably, the internal errors are somewhat higher than has been the case in some other four-color studies. However, the bulk of the stars were observed at least three times so the mean values are accurate enough to be useful. The standard deviations for a single observation are as follows: 0.013 for $b - y$; 0.020 for $m_1$; 0.030 for $c_1$; and 0.018 for $\beta$.
III. THE DISTANCE AND REDDENING
In order to apply the four-color and H$\beta$ calibrations to the photometric data it is necessary to determine the spectral range in which each star falls. Using the dia-
TABLE II. Reddened standard stars.
| Star (SAO) | $b - y$ | $m_1$ | $c_1$ | $n$ | $E(b - y)$ | Sp. | 0.74$E(B - V)$ | $\Delta$ |
|-----------|---------|-------|-------|-----|------------|-----|----------------|--------|
| 113973 | −0.050 | 0.076 | 0.210 | 3 | 0.045 | B2V | 0.07 | 0.02 |
| 122716 | 0.050 | 0.062 | 0.434 | 4 | 0.119 | B6V | 0.09 | −0.03 |
| 103439 | 0.023 | 0.053 | 0.036 | 1 | 0.138 | B1.5III | 0.15 | 0.01 |
| 114001 | 0.171 | −0.017| −0.010| 3 | 0.291 | B0.5V | 0.31 | 0.02 |
| 114112 | 0.232 | −0.033| 0.045 | 3 | 0.349 | B0.5V | 0.37 | 0.02 |
| 124467 | 0.259 | −0.035| 0.031 | 3 | 0.378 | B1.5III | 0.35 | −0.03 |
| 114058 | 0.315 | −0.052| −0.007| 3 | 0.437 | B1V | 0.46 | 0.02 |
| 123966 | 0.522 | −0.172| 0.289 | 1 | 0.621 | B3V | 0.56 | −0.06 |
| 142589 | 0.631 | −0.080| −0.066| 1 | 0.766 | B0.5Ia | 0.78 | 0.01 |
| 161476 | 0.824 | −0.151| 0.011 | 1 | 0.934 | B0.3Ia | 0.97 | 0.02 |
| 113125 | 0.870 | −0.110| 0.194 | 2 | 0.986 | B1V | 1.10 | 0.11 |
gram of \([m_v]\) vs \([c_1]\), it appears that all of the stars in Table I are B stars except one, star I. This agrees with the spectral types given for three of the stars. However, the spectral type given for star G, A0III, is in conflict with the photometric indices. From its value of \(m_v\) and \(\beta\) it appears that star G should have a spectral type between B4 and B6. We will therefore treat it as a B star in applying the photometric calibration.
For all the stars except I in the table the distance moduli and color excesses have been obtained from the B-star calibration of Crawford (1978). For star I, his F-star calibration (Crawford 1975) was used. These quantities are listed in the last two columns of Table I. A value of \(A_V/E(b-y) = 4.28\) was used (Crawford and Mandewela 1976) in correcting the distance moduli for interstellar absorption.
Figures 1 and 2 show several plots of the data from Table I. These diagrams serve to assist in detecting any nonmembers among our sample and to allow the reader to form an impression of the cluster's main characteristics. In Fig. 1(a) the apparent magnitude \(V\) is plotted against the \(c_1\), while Fig. 1(b) shows the plot of \(\beta\) against \(c_1\). Figure 2 is a plot of the color excess against the true distance modulus.
It is quite obvious that star I is a foreground star on the basis of its distance and its reddening (Fig. 2). Stars F and J are nearer the main sequence than should be the case based on the remaining stars in Figs. 1(a) and 1(b). However, they are in locations that are appropriate for blue stragglers. Since such stars appear to give the correct distance modulus from four-color photometry (see Schmidt 1982a for a discussion of this point) we will retain them in our sample. Star L poses some difficulty in that it is nearly a magnitude fainter than star F, even though its other photometric indices match well. The lower apparent magnitude has caused the distance modulus to lie among the largest of those in Fig. 2 that appear to form the cluster group. Since there is no obvious explanation for its lower magnitude, we will reject it from membership. Star S is located away from the other stars in all three samples and is probably a foreground star. Its rather high reddening is not unreasonable in view of the patchiness of obscuration in this field. Finally, star Q appears to be among the cluster stars in both Figs. 1(a) and 1(b) but has a higher distance modulus than the cluster stars. We will therefore regard it as a background star.
The mean distance modulus for NGC 6664, with stars I, L, S, and Q omitted, is \(V_0 - M_V = 10.69 \pm 0.38\) [standard deviation of one star]. The rejection of stars from the cluster is not crucial in determining the distance modulus. For example, if we include all the stars except star I, the mean distance modulus we obtain is 10.74. Therefore, we will adopt as our mean distance modulus 10.7 \pm 0.1\) (standard error of the mean).
The mean color excess of the cluster stars is \(E(b-y) = 0.556\). However, there is a large range from 0.48 to 0.63. The color excesses are correlated with position in the way found by Arp (1958); the largest color excesses are found in the western part of the cluster and lines of constant color excess run nearly north-south.
IV. DISCUSSION
Previous values of the distance modulus of this cluster have all been based on the same \(UBV\) data, those of Arp. However, they have covered a large range. Becker and Fenkart (1971) give a true distance modulus of 10.33, while Arp obtained 10.8. Sandage and Tammann (1969) rediscussed the distance of this cluster and quote 11.03. These values were all based on the old value of the Hyades distance. They should perhaps be increased by about 0.3 mag in light of recent studies of the Hyades and thus represent a range from 10.6 to 11.3. In his discussion of Cepheid luminosities, de Vaucouleurs (1978) used a more recent distance for the Hyades and obtained a modulus for NGC 6664 of 10.89. Thus it can be seen that the smaller of these moduli are in reasonable agreement with the present value but the larger ones are not.
Our mean color excess for the cluster, \(E(b-y) = 0.556\), corresponds to \(E(B-V) = 0.75\).
This is considerably higher than the value derived by Arp of about 0.60. Some of this discrepancy can be attributed to selection in a small sample with a large range of reddening. However, when the individual color excesses are compared it is found that the present values are larger by an average of 0.12 mag in $E(B-V)$. To further investigate this discrepancy we have plotted the present $b-y$ colors against the $B-V$ colors of Arp and the $u-v$ colors derived from the indices of Table I against the $U-B$ colors of Arp. This is shown in Fig. 3. Also shown for comparison are points representing some of the stars in M25, which has a similar reddening to NGC 6664. The $UBV$ data were taken from Wampler et al. (1961), while the four-color data came from Schmidt (1982a). It can be seen that both the diagrams exhibit discrepancies from M25 but in opposite senses. Any discrepancy in $U-B$ or $u-v$ will have a relatively small effect on the derived reddenings, but the differences in $B-V$ and $b-y$ between the present values and those for M25 fully account for the different derived color excess.
Unfortunately, there is very little photometry of this cluster beyond that of Arp and the present investigation. Roslund (1963) obtained $UBV$ photometry of two of the stars and his results are in reasonable agreement with Arp. However, neither of these stars was observed in the present program so it is not clear how strong this point is.
For the three B stars in Table I with spectral types, we can derive spectroscopic estimates of the color excesses. These are shown in Table III. The standard relations between spectral type ($B-V$ and $U-B$) listed by Johnson (1966) were used to obtain the reddenings given in columns 1 and 2. The mean indices for various spectral types listed by Crawford (1978) were used to obtain the reddening from $b-y$ and $c_1$ (columns 4 and 5). It can be seen that both $b-y$ and $c_1$ give spectroscopic color excesses in agreement with the photometric values. On the other hand, while the spectroscopic color excesses from $B-V$ agree with the photometric values, those from $U-B$ show serious discrepancies for two of the stars. In other words, the $UBV$ colors are not consistent with the spectral types.
One final point regarding the difference between the $UBV$ and the $ubvy$ photometry is that the $ubvy$ data were obtained during four different observing seasons and on some of the same nights as the M25 four-color data. A few were also obtained from Behlen Observatory. A search of the original data does not indicate any systematic differences among the data from the various observing runs. It is difficult to see how a systematic error of the order of 0.1 mag could occur between this cluster and M25 under these circumstances.
It is clear that it would be worthwhile to reobserve some of the stars in this cluster in the four-color system, the $UBV$ system, or preferably both, to check on these discrepancies. Until such an investigation is carried out, we will have to regard the reddening of this cluster as somewhat uncertain.
As Arp pointed out, although there is a large range in the color excess in this cluster, the variation is rather smooth across the face of the cluster and it is possible to interpolate reasonably well to the location of EV Sct. When we do this we obtain a color excess of $E(b-y) = 0.55$, which corresponds to $E(B-V) = 0.75$ (for a B star). The scatter about the interpolating line is 0.015 mag and this is likely to be the uncertainty in the color excess of EV Sct.
It appears that this cluster contains two and perhaps three blue stragglers. Stars F and J are clearly hotter
Table III. Color excesses from spectral types.
| Star | $E(B - V)$ | $E(b - y)$ |
|------|------------|------------|
| | (1) | (2) | (3) | (4) | (5) | (6) |
| F | 0.73 | 0.63 | 0.75 | 0.60 | 0.62 | 0.63 |
| L | 0.72 | 0.69 | 0.72 | 0.58 | 0.59 | 0.60 |
| M | 0.56 | 0.39 | 0.57 | 0.50 | 0.53 | 0.51 |
Explanation of the columns
1. From spectral type and $B - V$.
2. From spectral type and $b - y$.
3. From spectral type and $(U - B)$ diagram.
4. From spectral type and $b - y$.
5. From spectral type and $c_1$.
6. From $b - y$ and $c_1$.
Fig. 3. Plots of colors from the four-color photometry against the corresponding colors from the $UBV$ photometry. Solid circles represent the stars in NGC 6664 while the crosses indicate the stars from M25, which are shown for comparison. The arrows show the direction of the reddening trajectory in these diagrams.
than the turnoff point of the cluster in Figs. 1(a) and 1(b). Additionally, their distance moduli place them within the cluster. Star L is also a blue straggler if it is a cluster member. However, as discussed above, it is possibly a background star. Further studies of this star to more clearly establish its membership might, however, prove useful.
As discussed in a previous paper (Schmidt 1982a) the diagram of $\beta$ against $c_0$ is sensitive to age for clusters with the turnoff among the B stars. In Fig. 1(b) it can be seen that the turnoff occurs at about $c_1 = 0.70$, which corresponds to $c_0 = 0.59$. Referring to Fig. 5 of Schmidt (1982a), we see that this makes NGC 6664 slightly younger than the Pleiades but significantly older than the $\alpha$ Per cluster. We can estimate its age (assuming from Harris 1976, that the $\alpha$ Per cluster is about $20 \times 10^6$ yr old and the Pleiades is about $60 \times 10^6$ yr old) to be about $46 \times 10^6$ yr. This conflicts with the data given by Harris for NGC 6664, which imply an age of about $30 \times 10^6$ yr. However, she based her estimates on the bluest stars on the main sequence and it is possible that she was misled by the blue stragglers in this case.
The observations described in this paper were obtained at Cerro Tololo Inter-American Observatory. The author wishes to express his appreciation to the director and staff for the use of their facilities and for aid in obtaining the observations. This research was supported by a Cottrell Research Grant from the Research Corporation and by the National Science Foundation under Grant No. AST 77-17520.
REFERENCES
Arp, H. A. (1958). Astrophys. J. 128, 166.
Becker, W., and Fenkart, R. (1971). Astron. Astrophys. Suppl. 4, 241.
Crawford, D. L. (1975). Astron. J. 80, 955.
Crawford, D. L. (1978). Astron. J. 83, 48.
Crawford, D. L., and Barnes, J. V. (1970). Astron. J. 75, 978.
Crawford, D. L., and Mander, J. (1966). Astron. J. 71, 114.
Crawford, D. L., and Mandewala, N. (1976). Publ. Astron. Soc. Pac. 88, 917.
de Vaucouleurs, G. (1978). Astrophys. J. 224, 351.
Harris, G. L. H. (1976). Astrophys. J. Suppl. 30, 451.
Johnson, H. L. (1966). Annu. Rev. Astron. Astrophys. 1, 193.
Kraft, R. P. (1958). Astrophys. J. 128, 161.
Roslund, C. (1963). Ark. Astron. 3, 97.
Sandage, A., and Tammann, G. A. (1969). Astrophys. J. 157, 683.
Schmidt, E. G. (1980). Astron. J. 85, 158.
Schmidt, E. G. (1982a). Publ. Astron. Soc. Pac. 94, 232.
Schmidt, E. G. (1982b). Astron. J. 87, 650.
Schmidt, E. G., and Taylor, D. J. (1979). Astron. J. 84, 1193.
Wampler, J., Pesch, P., Miltner, W. A., and Kraft, R. P. (1961). Astrophys. J. 133, 895. |
Privacy as a Public Good: A Case for Electronic Cash*
Rodney J. Garratt†
University of California at Santa Barbara
Maarten R.C. van Oordt‡
Bank of Canada
December 17, 2019
Abstract
Privacy is a feature inherent to the use of cash for payments. With steadily increasing market shares of commercial digital payments platforms, privacy in payments may no longer be attainable in the future. In this paper, we explore the potential welfare impact of reductions in privacy in payments in a dynamic framework. In our framework, firms may use data collected through payments to price discriminate among future customers. A public good aspect of privacy in payments arises because individual customers do not bear the full costs of failing to protect their privacy. As a consequence, they may suboptimally choose not to preserve their privacy in payments. When left to market forces alone, the use of privacy-preserving means of payments, such as cash, may decline faster than is optimal.
Keywords: Money, privacy, social costs, overlapping generations model
JEL Classification Numbers: E42, G28
*We are grateful to Charles M. Kahn, Yu Zhu and participants in the Summer Workshop on Money, Banking, Payments and Finance (Ottawa, 2019), the TSE Sustainable Finance Inaugural Conference (Toulouse, 2019), and seminars at the Bank of Canada, Erasmus University Rotterdam and the Dutch Central Bank for helpful comments and suggestions. The views expressed in this paper do not necessarily reflect those of the Bank of Canada.
†Corresponding author. Email: email@example.com; Tel. +1(805)893-2849.
‡Email: firstname.lastname@example.org
Privacy in payments is a feature inherent to the use of cash, but transactional usage of cash is in decline. Cash cannot be used for the increasing share of online transactions, and the share of cash payments at the point of sale is also gradually decreasing.\(^1\) Going forward, transaction privacy may no longer be attainable due to steadily growing market shares of commercial payments platforms that generate datasets with payment histories at the user level.
Commercial payments platforms can monetize user data in multiple ways. They can share data with third parties, who may combine it with data from other sources and use the resulting dataset for marketing purposes or to generate predictions on creditworthiness.\(^2\)\(^3\) Alternatively, without sharing user-generated data, commercial payments providers can employ payments data to cluster users into profiles that, in turn, can be shared with third parties for marketing purposes or could be used for marketing purposes on behalf of those third parties.\(^4\) The effectiveness and profitability of such methods is expected to grow with advances in prediction tools (e.g., machine learning) and increases in computational power and the scale and scope of data gathering.
---
\(^1\)Khiaonarong and Humphrey (2019) report this trend in cash shares for various countries based on aggregate statistics. For national estimates of downward trends based on survey data, see, e.g., Kumar and O’Brien (2019, United States), Deutsche Bundesbank (2017, Germany) and Henry et al. (2018, Canada).
\(^2\)See, e.g., the Bloomberg (2018) news report, which reads “For the past year, select Google advertisers have had access to a potent new tool to track whether the ads they ran online led to a sale at a physical store in the U.S. That insight came thanks in part to a stockpile of Mastercard transactions that Google paid for.”
\(^3\)Zhima (aka Sesame) credit, an affiliate of the Chinese Alibaba Group, reportedly uses data from Alipay to determine credit scores (Financial Times, 2018).
\(^4\)Patent applications provide evidence of ample research and development investments of major commercial payment providers in this area; see the excerpts provided in Appendix A.
Lack of privacy in payments can also lead to consequences beyond marketing and credit provision. Credit scores generated in part from payments data are used by governments in “Orwellian” ways. In China, Zhima credit scores determine eligibility for discounts and deposit waivers, while low scores can reportedly lead to punishments such as career restrictions and penalties on asset holdings (CNBC, 2017). Some have even argued that payments data might someday play a role in China’s proposed new “Social Credit System” (WSJ, 2019; Wired, 2017). Finally, concern that some countries may deny entry to individuals if they know they have purchased cannabis in Canada – even though cannabis is legal in Canada – led the Office of the Privacy Commissioner of Canada (2018) to advise people to use cash for such transactions.
This paper shows that the failure of individuals to preserve their privacy in payments by using privacy-enhancing techniques, such as cash, can lead to socially suboptimal outcomes. In the environments we consider, firms use information extracted during payments of one consumer without privacy-enhancing techniques to cluster potential future customers into groups with different reservation prices. Hence, failure on the part of one individual to preserve his or her information imposes a negative externality on others. In that sense, our approach to modelling privacy in payments is complementary to, but distinct from, the seminal work of Kahn et al. (2005) on the private cost of information disclosure. In the Kahn et al. model, individuals incur a private cost when information is disclosed during payment (modeled as the possibility of the purchase being stolen). Consumers are not given a choice to pay privately in the Kahn et al. framework; however, if they were, they would choose to use a privacy-enhancing technique if the private benefit exceeds the cost. In our set-up, information disclosure during the payment does not affect the terms and conditions of the current transaction, and the customer does not face any negative consequences from disclosing information during the payment. Our model emphasizes the possibility that individuals may suboptimally choose not to use privacy-enhancing techniques because the cost of foregoing privacy in payments is primarily borne by the rest of society.
Our paper connects to the so-called privacy paradox, which refers to the mismatch between stated preferences for privacy and actual behaviour to preserve privacy (Norberg et al., 2007; Athey et al., 2017). The literature offers several explanations such as unawareness of the extent of data collection and behavioural targeting and unfamiliarity with privacy-enhancing measures (Acquisti et al., 2016). Our paper adds an additional explanation by providing a mechanism that leads to differences between the private and social cost of information disclosure: Information disclosed by others allows inference about you. Hence, inference about you occurs regardless of your own choices to protect your information. Moreover, you do not bear the full or any cost of the information you inadvertently reveal about others. Both aspects induce a low willingness to take costly measures to protect privacy.
The notion that there may be a public good aspect to privacy has been noticed by legal scholars (MacCarthy, 2010; Fairfield and Engel, 2015). Choi et al. (2019) formulate an economic model where a ‘nuisance cost’ of foregoing privacy appears in the utility function. This cost is assumed to be an increasing function in the number of agents who reveal information, which imposes a negative externality that may lead to too little privacy in equilibrium. Similarly, Kahn et al. (2000) consider the case where a third party derives a fixed utility benefit from learning about a transaction. In both cases, the privacy externality itself is imposed as a model assumption rather than an economic outcome. Our model provides an explicit mechanism that generates a negative externality in the context of privacy in payments. Data revealed by one customer can be used to price discriminate among future customers. Hence, information disclosed during the payment is associated with a negative externality on future customers.
We seek to make these points in a model in which money is essential (Wallace, 2001). However, the analysis of privacy in payments in a model where money is essential is complicated by the fact that such a model requires a dynamic structure, so that intrinsically worthless money can have a positive price. The existence of monetary equilibria requires balanced modelling that ensures that agents form rational expectations regarding endogeAchieving this feature makes it harder to see the tension between private and social benefits and costs to maintaining privacy in payments. As such, in Section 2, we describe a one-period, three-cohort model that can be used to derive simplified versions of the main results. In this model, agents only live one period, and we give money value by including it in the utility function. The money in this model could be interpreted as a commodity money. The results for the one-period, three-cohort model are described in Section 3. Section 4 describes an infinite-horizon, overlapping-generations model, which builds off of the one-period model and justifies a number of assumptions made in the one-period model. Most notably, an assumption in the one-period model that ensures the entire market is served in the absence of price discrimination emerges endogenously in the full monetary equilibrium. In this alternative, dynamic model, money no longer appears directly in the utility function. Agents live multiple periods, and the incentive to acquire money when they are young is to be able to purchase items when they are middle-aged or old (Wallace, 1980; Balasko and Shell, 1981). Section 5 provides a discussion of policy issues and the potential role of government regulation. Section 6 discusses implications of some alternative modelling assumptions. Concluding remarks appear in Section 7.
2 One-Period, Three-Cohort Model
The economy consists of equal numbers of three cohorts of agents: young merchants \((y)\), middle-aged consumers \((m)\) and old consumers \((o)\). Young merchants are endowed with two units of an indivisible good that they do not wish to consume themselves but would like to sell for money to the middle-aged and old consumers. Middle-aged and old consumers are each endowed with at least \(r_H > 0\) units of money, where \(r_H\) denotes the highest possible reservation price consumers have for the good. The aggregate endowment of money is denoted by \(M\).
---
5See Kovenock and De Vries (2002, pp. 147–149) for an overview of different modelling approaches.
First, each merchant is randomly matched with a middle-aged consumer and makes a take-it-or-leave-it offer that the consumer can either accept or refuse. Then, each merchant is randomly matched with an old consumer and makes a take-it-or-leave-it offer that the consumer can either accept or refuse.
Middle-aged and old consumers can either have a $H[igh]$ or $L[ow]$ value for the good. A consumer’s type $s \in \{H, L\}$ determines her reservation price $r_s$ for the indivisible good.\footnote{In the dynamic model, reservation prices are determined by consumers’ endogenously determined money holdings.} The known quantities of types $H$ and $L$ are the same in each age cohort and are given by $zn$ and $(1 - z)n$, respectively, where $n$ is the population size of each of the three cohorts of agents and $0 < z < 1$. The utility of consumer $i$ in cohort $a \in \{m, o\}$ depends on her consumption $c_{ia} \in \{0, 1\}$ of the indivisible good and her terminal money holdings $m_{ia} \geq 0$.\footnote{We use $a$ (short for age) to denote cohorts in the one-period model. In the dynamic model, cohorts will be identified by the generation $t$ (for time) in which they were born.} A consumer of type $s_{ia} \in \{H, L\}$ weighs consumption of the indivisible good by the reservation price $r_s$, where $zr_H < r_L$. Consumer $ia$’s utility is augmented if she chooses to take an action $e_{ia} \in \{0, 1\}$ to protect her privacy during payments. The action involves a private benefit $\beta > 0$ and a cost of effort $\delta > 0$. The utility function of consumer $i$ in cohort $a \in \{m, o\}$ of type $s_{ia} \in \{H, L\}$ is given by
$$u_a(c_{ia}, m_{ia}, e_{ia}; s) = c_{ia}r_s + m_{ia} + e_{ia}(\beta - \delta).$$
Merchants only care about the amount of money they end up with from selling goods. The utility function of young merchant $i$ is given by $u_y(m_{iy}) = m_{iy}$. To maximize this amount, merchants seek to price discriminate. Ideally, a merchant would like to charge a price of $r_H$ to consumers of type $H$ and a price $r_L$ to consumers of type $L$. However, consumer type is not directly observable. Merchants can, however, profile middle-aged and old customers based on an observable characteristic $h_{ia} \in \{X, Y\}$, for $a = m, o$, respectively, that is related to consumer type. Observable characteristics refer to consumer traits that
are observable prior to the transaction. These can be things that are physically observed in the case of in-person transactions, either because they relate to the consumer’s appearance or because the consumer uses coupons or vouchers that were sent to targeted areas. For an online store, examples of observable characteristics include online profiles or other customer information that can be revealed when customers access websites.
Consider the interaction between the merchant and the middle-aged consumer. The observable characteristic $X$ is associated with all middle-aged consumers of a particular type, and the observable characteristic $Y$ is associated with all middle-aged consumers of the other type. The merchant’s problem is that, before making an offer to the middle-aged consumer, the merchant does not know which characteristic is associated with each type, and there is no credible way for consumers to signal their type. The merchant only knows that either of the two associations is equally likely.
The merchant cannot price discriminate against middle-aged consumers based on observable types. However, her interaction with the middle-aged consumers may help her to price discriminate against old consumers. This can happen in two ways. First, by charging the high reservation price to middle-aged consumers, she can learn which observable characteristic is possessed by old customers that are likely to accept a high offer. Second, the merchant can observe the consumer’s type, regardless of the amount they pay, if consumers do not act to preserve their privacy in payments when they pay.\footnote{Privacy in payments is not protected in reality when people use payment cards or mobile payment services that are linked to their broader payment history and other financial metrics, and this information is fed back to the retailer in some form.} Either of these methods allows the merchant to discover the relationship between the observed characteristic of the consumer and her type.
Knowledge of the correlation between the observable characteristic and type of the middle-age consumer allows the merchant to imperfectly price discriminate against old consumers. Price discrimination against old consumers is imperfect because we assume that only a fraction $1 - \varepsilon$, where $0 < \varepsilon < 1/2$, of old consumers has the same correlation between the
observable characteristic and type as the middle-aged consumers. The remaining fraction $\varepsilon$ has the opposite correlation.\footnote{The perfect correlation between the observable characteristic and type of the middle-aged consumers is not essential to our results. The same results hold true when we assume the same imperfect correlation for middle-aged and old consumers, although the expressions for $\theta^C$ and $\theta^W$ will be more complex.} Profiling of old consumers becomes increasingly uninformative as $\varepsilon$ approaches 1/2 and closer to perfect as $\varepsilon$ approaches 0.
We can now specify social welfare in terms of the total utility of the three generations of agents. Utility is assumed to be linear in money for both merchants and consumers so that the “consumption” of money is irrelevant for social welfare: the money in the utility function of consumers balances out the presence of money in the utility function of the merchants. As a consequence, we can ignore the distribution of money holdings and write social welfare as
\begin{align*}
W = & M + zn r_H \Pr(c_{im} = 1|H) + (1 - z)n r_L \Pr(c_{im} = 1|L) + \\
& + zn r_H \Pr(c_{io} = 1|H) + (1 - z)n r_L \Pr(c_{io} = 1|L) + 2n \mathbb{E} e_{ia}(\beta - \delta),
\end{align*}
(1)
where $\Pr(c_{im} = 1|H)$ and $\Pr(c_{im} = 1|L)$ denote the probabilities that a middle-aged consumer of, respectively, type $H$ and $L$ accepts the offer from the merchant and where $\Pr(c_{io} = 1|H)$ and $\Pr(c_{io} = 1|L)$ reflect these probabilities for the merchant’s encounter with the old consumer.
3 Results
Middle-aged consumers will not act to preserve their privacy if $\beta < \delta$ – that is, whenever the private benefit to preserving privacy is less than the private cost. However, this outcome may not be socially optimal. Actions taken by middle-aged consumers to preserve privacy benefit old consumers by reducing the ability of merchants to price discriminate against them. So the social benefit to the middle-aged consumers of acting to preserve privacy is
greater than the private benefit. Because middle-aged consumers do not fully internalize the social benefit, they may make socially suboptimal decisions.
Suboptimal outcomes occur if the equilibrium that arises when middle-aged consumers take action to preserve their privacy has higher welfare than the equilibrium that occurs when they do not. Below we characterize equilibrium for the two relevant cases, $\beta > \delta$ and $\beta < \delta$. The difference in social welfare across these two cases can be used to define the social benefit of a subsidy designed to induce the socially optimal outcome.
### 3.1 Equilibrium with $\beta > \delta$
First, we consider the strategy where merchants quote the same price to all consumers in both encounters. It is optimal for merchants to quote customers a price just below or at their reservation price. So merchants will either quote a price $r_H$ or $r_L$. When merchants quote the higher price $r_H$ to all consumers, then only type $H$ consumers will accept the offer, and the expected revenue of each merchant equals $2zr_H$. When quoting the lower price $r_L$ to all consumers, then all the offers will be accepted. This will generate an expected revenue of $2r_L$ for each merchant, which is strictly higher since we assumed $zr_H < r_L$.
The expected utility of the young merchant when quoting the same price $r_L$ to all consumers equals
$$\mathbb{E}u_y^* = 2r_L.$$
Moreover, in the absence of price discrimination, social welfare equals
$$W^* = M + 2znr_H + 2(1 - z)nr_L + 2en(\beta - \delta), \tag{2}$$
where $e = 1$ if $\beta > \delta$ and $e = 0$ otherwise. Note that this strategy maximizes social welfare, because it is equivalent to setting all probabilities in Eq. (1) equal to one.
Second, we consider the strategy where merchants try to differentiate their price quotes in the second encounter after learning the relationship between the observable characteristic and consumer type based on observable behaviour in the first encounter.
If the merchant quotes the lower price $r_L$ in the first encounter, then the merchant cannot infer the relationship between the observable characteristics and consumer type from consumer behaviour, because both consumer types react in the same way. So, the merchant can only learn the relationship between the observable characteristics and consumer type when quoting a price larger than $r_L$ in the first encounter. If the merchant quotes a price $r_H$ in the first encounter, then only type $H$ customers will take the offer. If the customer accepts the offer, then the merchant knows that consumers with the same observable characteristic are of type $H$. If the customer declines the offer, then the merchant knows that consumers with the same observable characteristic are type $L$ customers.
In the second encounter, the merchant can use this information to quote the higher price $r_H$ to consumers with the observable characteristic associated with type $H$ customers, while quoting price $r_L$ to consumers with the characteristic associated with type $L$. The only consumers who will decline the offer in the second encounter are the fraction $\varepsilon$ of type $L$ consumers who randomly adopted the characteristic associated with type $H$ customers. Moreover, a fraction $\varepsilon$ of type $H$ consumers, who randomly adopted the characteristic associated with type $L$ customers, will be paying the low price $r_L$. Hence, the merchant’s expected utility from price discriminating (D) while consumers act to preserve universal privacy (U) in payments equals
$$\mathbb{E}u_y^{D,U} = z(2 - \varepsilon)r_H + [z\varepsilon + (1 - z)(1 - \varepsilon)]r_L.$$
Comparing $\mathbb{E}u_y^*$ and $\mathbb{E}u_y^{D,U}$ gives that, in a situation of privacy in payments, the merchant will price discriminate only if the profiling of consumers is sufficiently precise, i.e., only if
$\varepsilon < \theta^U$, where
$$\theta^U = \frac{zr_H - zr_L - (r_L - zr_H)}{r_L + zr_H}.$$ \hspace{1cm} (3)
If merchants choose to price discriminate in a situation with privacy in payments, i.e., if $\varepsilon < \theta^U$, then social welfare will be lower and equals
$$W^{D,U} = W^* - n(1 + \varepsilon)(1 - z)r_L.$$ \hspace{1cm} (4)
A social welfare loss occurs as a consequence of price discrimination with privacy in payments for two reasons. First, to learn about the link between consumer profiles and their behaviour, the merchant needs to quote prices that some customers will not accept in the first meeting. Second, the merchant may quote high prices that some customers will not accept in the second meeting due to imprecise profiling ($\varepsilon > 0$).
Finally, note that the threshold $\theta^U$ will be negative when the reservation price $r_H$ is sufficiently low such that $r_H < r_L(1+z)/2z$. If so, then price discrimination will never occur when transaction privacy is given, regardless of the precision of the profiling technique as indicated by $\varepsilon$. This implies that the minimum level of $r_H$ at which price discrimination occurs is lower for higher values of $z$. In other words, price discrimination may occur with a lower dispersion in reservation prices if a larger fraction of consumers is willing to pay a high reservation price.
### 3.2 Equilibrium with $\beta < \delta$
Without universal transactions privacy, the merchant receives information on the consumer type in the first encounter once the payment for the consumption good is settled when the consumer makes no effort to protect her privacy in payments. Otherwise, the merchant does not receive this information. The merchant can use the information to infer the relationship between the observable characteristic and consumer type. This information may be used by the merchant to price discriminate in the second encounter.
Would the consumers in the first encounter make an effort to protect their privacy in payments? The information is only disclosed to the merchant once the payment for the consumption good is settled. Hence, the disclosure of information to the merchant cannot adversely affect the outcome for the consumer who makes the payment, although subsequent consumers could be worse off. Moreover, there is a net utility cost of maintaining privacy in payments when $\beta < \delta$. In other words, individual consumers in the first encounter have no incentives to maintain privacy in payments whenever $\beta < \delta$, because doing so strictly reduces their utility.
Merchants can follow a different strategy that leads to price discrimination when consumers do not maintain privacy in payments. First, they can quote the low price $r_L$ to all customers in the first encounter. They collect information on the consumer type during this first transaction since there is no privacy in payments. In the second encounter, the merchant uses this information to quote the higher price $r_H$ to consumers with the observable characteristic associated with type $H$ customers, while quoting price $r_L$ to consumers with the characteristic associated with type $L$. The merchant’s expected utility from following this strategy in an environment without universal privacy ($W$) is
$$\mathbb{E}u_y^{D,W} = r_L + z(1 - \varepsilon)r_H + [z\varepsilon + (1 - z)(1 - \varepsilon)]r_L.$$
The merchant’s expected utility from relying on this method to price discriminate, $\mathbb{E}u_y^{D,U}$, strictly exceeds $\mathbb{E}u_y^{D,U}$, since $r_L > zr_H$. Moreover, comparing $\mathbb{E}u_y^{D,W}$ to $\mathbb{E}u_y^*$ gives that, without privacy in payments, the merchant will price discriminate only if the profiling technique is sufficiently precise such that $\varepsilon < \theta^W$, where
$$\theta^W = \frac{zr_H - zr_L}{r_L + zr_H}. \tag{5}$$
If the merchant chooses to price discriminate without privacy in payments, i.e., if the profiling of consumers is sufficiently precise such that $\varepsilon < \theta^W$, then social welfare equals
$$W^{D,W} = W^* - n\varepsilon(1-z)r_L.$$ \hspace{1cm} (6)
Price discrimination in the absence of universal transaction privacy leads to a welfare loss whenever the profiling of consumers is imperfect ($\varepsilon > 0$).
### 3.3 Price discrimination and profiling errors
Whether price discrimination occurs in equilibrium depends on the accuracy of consumer profiling and whether consumers choose to protect their privacy in payments. Figure 1 summarizes the relationship between price discrimination and the probability of profiling errors $\varepsilon$.
Starting at the right of Figure 1 is the nonempty region $\theta^W \leq \varepsilon < 1/2$, where the social optimum is attained regardless of whether consumers protect their privacy in payments.\footnote{The region is nonempty because $r_L > zr_H$ implies $\theta^W < 1/2$.} In this region, price discrimination does not occur in equilibrium because the profiling of consumers is too imprecise. Price discrimination is not profitable because the expected
loss in sales from erroneously quoting high prices to consumers with low reservation prices outweighs the potential benefit of earning more on some transactions.
In the middle region, where $\theta^U \leq \varepsilon < \theta^W$, price discrimination does not occur when consumers protect their privacy in payments, and hence the social optimum is achieved in equilibrium whenever $\delta \leq \beta$. In contrast, whenever $\varepsilon$ lies in this middle region and $\delta > \beta$, consumers do not protect their privacy, and merchants find it profitable to price discriminate. In this case, the equilibrium is not socially optimal because, under price discrimination, not all potential win-win situations lead to transactions between merchants and consumers.
In cases where $\theta^U \leq \varepsilon < \theta^W$ and $\delta > \beta$, the social optimum could be achieved if consumer efforts to protect their privacy were subsidized by an amount $S$ such that $\delta - S - \beta = 0$. This would be a socially optimal policy if the net private cost of protecting privacy is smaller than the per-person social benefit doing so, that is, when
$$\delta - \beta < \varepsilon(1 - z)r_L.$$
If this condition holds, than society would be better off if the government imposed a lump-sum tax on each consumer equal to $S = \delta - \beta$ and returned this amount to consumers who act to protect their privacy when purchasing from the merchant.
Finally, there may be a third region to the left where $0 \leq \varepsilon < \theta^U$. This region only exists when the reservation price $r_H$ is sufficiently high such that $r_L(1/2 + 1/(2z)) < r_H < r_L(1/z)$. In this region, the social optimum will not be attained regardless of whether there is universal transaction privacy.
4 Overlapping Generations Model
Time is discrete and continues without end. A new generation of $n$ identical three-period lived agents starts at each date $t$. “Generation $t$” is young at $t$, middle-aged at $t + 1$, and old at $t + 2$ (the three cohorts from the one-period model). Agents maximize their life-time
utility, which depends negatively on the amount of labour to produce consumable goods when they are young and positively on their consumption pattern when they are middle-aged and old. Moreover, as before, agents face a the net private cost (or benefit) of preserving privacy in payments. Agent $i$ from generation $t$ can produce $Q_{it}$ perishable indivisible consumption goods at a utility cost $Q_{it}f$, but derives no utility from consuming the good.\footnote{Throughout this section we use the subscript position to identify a particular individual ($i$) born at a particular date ($t$).} Middle-aged and old agents cannot produce goods but derive utility from consuming them. Each agent has three opportunities to consume a good: one opportunity when middle-aged (referred to as meeting $A$), and an early and late opportunity to consume when old (referred to as meetings $B$ and $C$, respectively). Their utility function reads
\begin{equation}
u(c^A_{it}, c^B_{it}, c^C_{it}, Q_{it}, e_{it}) = c^A_{it} + c^B_{it}b + c^C_{it}c - Q_{it}f + e_{it}(\beta - \delta).
\end{equation}
Preferences of the agents are such that they enjoy consumption more at certain points in their lives. This is indicated by the parameters in the utility function: $\frac{1}{3} < c < 1 < b$. The utility cost of production is assumed to be low but positive, and satisfies $0 < f < \min\{\frac{1}{3}, (b-c)/2\}$. The assumption $f < \frac{1}{3}$ implies that the benefit of consumption in meeting $C$ exceeds the utility cost of producing the consumption good. The assumption $f < (b-c)/2$ places a restriction on the minimum level of impatience agents have when old.
There is no technology for record-keeping. However, there is a durable, indivisible, non-consumable good and a transaction technology that allows exchanging the durable asset for the consumable good whenever an offer is made and accepted. We refer to the durable asset as “money.” The total money stock is fixed at $4n$. At all times, individual agents cannot carry more than three units of money, and the amount agents carry is not observable.\footnote{Lagos and Wright (2005, p. 464) set out the complexities of solving monetary equilibria without constraints on money holdings. Their solution is to make assumptions that lead to a degenerate distribution of money holdings, so all agents hold the same amount of money. An alternative is to resort to finding numerical solutions for the distribution of money holdings (e.g., Molico, 2006). We opt for an approach where agents can hold up to three units of indivisible money, which allows us to derive analytical solutions in an environment where agents can hold different amounts of money in equilibrium.}
Figure 2 shows a timeline of the only interactions between agents in the economy. Meeting $A$ involves a random matching between a young agent and a middle-aged agent. Meetings $B$ and $C$ involve random matchings between young agents and old agents (each old agent has both an early and a late encounter with two different young agents). During each of these meetings, the young agent may make a take-it-or-leave-it offer regarding the number of units of money at which the young agent is willing to produce a unit of the consumable good for the older agent.
The price offers that are received by agent $it$ during the meetings are denoted by $p^A_{it}$, $p^B_{it}$ and $p^C_{it}$. Hence, the offer made by young agent $it$ to middle-aged agent $j,t-1$ is denoted as $p^A_{j,t-1}$, etc. Similarly, the amount of money that agent $it$ carries into the meetings when middle-aged and old is denoted as $m^A_{it}$, $m^B_{it}$ and $m^C_{it}$, respectively. If an agent accepts an offer, then the young agent will produce the good, a transaction occurs and the older agent consumes the good.
After the meetings, consumable goods and old agents perish. Young agents may inherit one or more units of money from old agents. Money left by old agents is allocated in rounds.
Young agents who own less than three units of money may receive a unit of money in each round. Whenever there are insufficient funds to allocate a unit of money to every eligible young agent, the agents who do not inherit a unit of money are randomly selected. Another allocation round occurs whenever there are funds left. The inheritance that each young agent $it$ receives is denoted as $m^I_{it}$.
The amount of money each agent of generation $t$ accumulates when young determines their lifetime status $s_{it} \in \{H, L\}$. All agents who accumulate three units of money attain the lifetime status $s_{it} = H$, and agents who accumulate less than three units of money have a lifetime status $s_{it} = L$. An agent’s lifetime status is not directly observable to other agents. Instead, at the start of each period $t$, generations born in periods $\tau \in \{t-1, t-2\}$ take on an observable characteristic $h^t_{i\tau} \in \{X, Y\}$, which is assigned to all middle-aged and old agents depending on their lifetime status. The correlation between $s_{i\tau}$ and $h^t_{i\tau}$ is perfect for middle-aged agents ($\tau = t - 1$) but imperfect for old agents ($\tau = t - 2$) due to $\varepsilon$. The young generation does not know the sign of the correlation, which may be different in every period $t$.
From the model it follows that a few equalities must hold true in the aggregate. First, the total amount of money that agent $it$ earns and inherits should equal the amount that the agent brings to meeting $A$ when middle-aged. As a direct consequence, it must hold true for any young agent $it$ who meets a random middle-aged agent $j, t-1$ and two different random old agents indexed $k, t-2$ and $l, t-2$, that
$$\Pr \left[ c^A_{j,t-1} p^A_{j,t-1} + c^B_{k,t-2} p^B_{k,t-2} + c^C_{l,t-2} p^C_{l,t-2} + m^I_{it} = w \right] = \Pr \left[ m^A_{it} = w \right] =: g_t(w) \quad (8)$$
for $w = 0, 1, 2, 3$. Function $g_t(w)$ is the density of money holdings for young agents (the wealth distribution). Moreover, since the aggregate money holdings must equal the total quantity of money, and given that young agents are born without any money, it holds true
at the start of every period $t$ that
$$\sum_i m_{i,t-1}^A + \sum_i m_{i,t-2}^B = 4n.$$
(9)
Agents need to decide on their strategies regarding which offers to make during the meetings when young. These strategies may depend on whether offers in previous meetings were accepted as well as on the observable characteristics of the agents they encounter, and which offers they expect and accept when middle-aged and old.
**Definition 1** A steady-state Nash equilibrium is defined as the state of the economy where none of the agents can increase their expected utility in Eq. (7) by deviating from their current strategy given the strategies of all other agents and rational expectations regarding the steady-state wealth distribution in Eq. (8), and where the restriction on aggregate money holdings in Eq. (9) holds true. A monetary equilibrium is defined as a steady-state Nash equilibrium in which meetings occur where the consumable good is exchanged for the durable asset.
### 4.1 Non-monetary equilibrium
There exists a steady-state Nash equilibrium where no exchanges of consumable goods for money occur. Young agents do not make any offer to exchange consumable goods for money given that they rationally don’t expect any offers when being middle-aged and old. Middle-aged and old agents have no opportunities to spend their money holdings and keep their money until they die. Young agents inherit those money holdings. The steady-state distribution of money holdings depends on the initial wealth distribution.
### 4.2 Optimal consumption decisions
In a monetary equilibrium, agents need to decide which offers to accept and which to decline when middle-aged and old. The decision for old agents is simple. Since old agents
have no further use for money after their last meeting, they will accept any offer that they can afford in meeting $C$. So, agent $it$ accepts any offer where $p_{it}^C \leq m_{it}^C$ when old. Moreover, since agents prefer early consumption when old ($c < b$), they will also accept any offer that they can afford in meeting $B$. Hence, agent $it$ accepts any offer where $p_{it}^B \leq m_{it}^B$ when old. Therefore, the reservation price of an old agent depends on the agent’s money holdings, where larger money holdings are associated with a higher reservation price.
The decision to consume is more complicated for middle-aged agents. They prefer early consumption when old above consumption when middle-aged ($b > 1$), but accepting an offer when middle-aged may affect the ability to pay for consumption when old. The expected utility from consumption when old depends on which offers an old agent with a given amount of money and given status can afford (early, late, or both). For an old agent $it$ with money $m_{it}^B$ and status $s_{it}$, this can be written as
\begin{align*}
V^o(m_{it}^B; s_{it}) &= c \Pr \left[ p_{it}^C \leq m_{it}^B < p_{it}^B | s_{it} \right] + b \Pr \left[ p_{it}^B \leq m_{it}^B < p_{it}^B + p_{it}^C | s_{it} \right] \\
&+ (b + c) \Pr \left[ p_{it}^B + p_{it}^C \leq m_{it}^B | s_{it} \right].
\end{align*}
Accepting offers when middle-aged may reduce the probabilities in Eq. (10) by lowering money holdings $m_{it}^B$. Agent $it$ will accept an offer of $p_{it}^A$ when middle-aged if and only if the utility of consuming when middle-aged outweighs the expected loss in consumption when old, i.e., if and only if
\[ 1 \geq V^o(m_{it}^A; s_{it}) - V^o(m_{it}^A - p_{it}^A; s_{it}). \]
Whether this holds true depends on the agent’s money holdings $m_{it}^A$, the offer under consideration $p_{it}^A$ and the anticipated distribution of future offers $p_{it}^B$ and $p_{it}^C$ given the agent’s status $s_{it}$.
4.3 Best social outcome without free goods
The ability to benefit from offers received when middle-aged and old increases in the amount of money collected when young. Young agents face a positive cost of producing the consumable good. Since there is no record of whether an agent has lead a productive life while young, except for the agent’s money holdings, young agents will only propose offers where the potential benefits from additional monetary earnings exceed the cost of producing the consumable good. As a consequence, free goods cannot exist in a monetary equilibrium. Or conversely, any transfer of the consumable good to the middle-aged and old consumers will require at least one unit of money in a monetary equilibrium.
One can ask what the best feasible economic outcome could look like given the restriction that each transfer of the consumable good involves at least one unit of money, while further ignoring the choices made by optimizing agents. Obviously, more goods can be consumed when the price offers received by agents are as low as possible, i.e., if $p_{it}^A = p_{it}^B = p_{it}^C = 1$ for all $it$. Moreover, more goods can be consumed when they are consumed early in life: When all agents earn one unit of money when young and directly spend that unit when middle-aged, then this requires $n$ units of money in total. By contrast, when all agents earn one unit of money when young, save that unit when middle-aged, and spend that unit on a consumable good when old, then this requires $2n$ units of money in total. Finally, exchanges in meeting $C$ are the least attractive, because they require a relatively large amount of money and yield the lowest utility.
Altogether, it requires $3n$ units of money in total to allow all young agents to earn two units of money and to spend that money in meetings $A$ and $B$. Given that there are $4n$ units of money, there is still room to facilitate exchanges in meeting $C$ for half of the agents. The following lemma describes this economic situation:
**Lemma 1** The best achievable social outcome in the absence of free goods is that where all young agents offer the consumable good at a price of 1 in each meeting. Half of the agents earn 2 units of money when young and accept offers in meetings $A$ and $B$ when middleaged and old. Half of the agents earn 3 units of money when young and accept offers in all meetings. The steady-state wealth distribution is \( g(2) = g(3) = 1/2 \). Expected utility of new agents is
\[
\mathbb{E}u^* = 1 + b - 2f + \frac{1}{2}(c - f).
\]
Social welfare, in terms of the aggregate per-period utility, equals \( W^* = n\mathbb{E}u^* \).
### 4.4 Monetary equilibrium
In what follows, we show that the best achievable social outcome in the absence of free goods in Lemma 1 is a monetary equilibrium conditional on \( \varepsilon \) being sufficiently large. As in the single-period model, price discrimination may become optimal for a low probability of profiling errors \( \varepsilon \). Moreover, the threshold for \( \varepsilon \) at which price discrimination becomes optimal for young agents will depend on whether agents take action to preserve their privacy in payments in a similar manner. That is, the decision will depend on the private rather than the social costs and benefits of privacy in payments.
#### 4.4.1 Offers without considering price discrimination
We start by showing that it is optimal for a young agent \( it \), who meets a random middle-aged agent \( j, t-1 \) and two random old agents indexed \( k, t-2 \) and \( l, t-2 \), to make offers \( p_{j,t-1}^A = p_{k,t-2}^B = p_{l,t-2}^C = 1 \) when young agents do not consider price discrimination based on observable consumer characteristics. For convenience, we will drop the subscript of the prices when discussing the optimal offers made by the young agent \( it \) (i.e., \( p^A \) refers to \( p_{j,t-1}^A \), etc.) since the index of the matched agent is irrelevant to the pricing decision.
Start with the optimal choice for the last offer (meeting \( C \)). Could it be optimal to offer a price \( p_{l,t-2}^C > 1 \)? Given the strategies of the other agents, half of the old agents will carry no money and half of the old agents will carry one unit of money at the start of meeting \( C \). So, none of the old agents can accept an offer \( p_{l,t-2}^C > 1 \). Would it be optimal for the
young agents to make offers when their money holdings are smaller than three? Producing the consumable good costs $f$. Given the strategies of other agents, one unit of money will facilitate buying a consumable good when middle-aged or old, which yields at least utility $c > f$ (there is no chance of receiving an inheritance because all agents spend all their money during their lifetime). Hence, it is optimal to offer a price $p_{l,t-2}^C = 1$ in meeting $C$ whenever cumulative earnings from meetings $A$ and $B$ are less than three.
Now consider the optimal choice for the second offer (in meeting $B$). Could it be optimal to offer a price $p_{k,t-2}^B > 1$? Again, since all agents have spent one unit of money in meeting $A$, no agent is expected to have sufficient funds to accept an offer $p_{k,t-2}^B = 3$. Moreover, agents who started with three units of money would accept an offer of $p_{k,t-2}^B = 2$ because $b > c$.
Suppose that the young agent earned one unit of money in meeting $A$. If the offer $p_{k,t-2}^B = 2$ is accepted, which occurs with 50 percent probability, then the agent ends with three units of money by producing only two goods. If the offer is not accepted, then it would be optimal to offer a price $p_{l,t-2}^C = 1$ in the next meeting, which would lead to a 25 percent probability of earning two units of money when young (facilitating consumption in meetings $A$ and $B$) and a 25 percent probability of ending up with one unit of money (facilitating consumption in meeting $B$ only). Altogether, the expected utility derived from this strategy is
$$\frac{1}{2} [b + 1 + c - 2f] + \frac{1}{2} \left[ \frac{1}{2}(b + 1 - 2f) + \frac{1}{2}(b - f) \right] = \mathbb{E}u^* - \frac{1}{4} + \frac{3}{4}f,$$
which is less than $\mathbb{E}u^*$ since $f < 1/3$.
Moreover, suppose that the agent earned no unit of money in meeting $A$. In this situation, there is a 50 percent probability that the offer $p_{k,t-2}^B = 2$ will be accepted and a 50 percent probability that the offer $p_{l,t-2}^C = 1$ will be accepted. The expected utility of this strategy is
$$\frac{1}{2} \left[ \frac{1}{2}(b + 1 + c - 2f) + \frac{1}{2}(b + 1 - f) \right] + \frac{1}{2} \left[ \frac{1}{2}(b - f) \right] = \mathbb{E}u^* - \frac{1}{4}(b + c) - \frac{1}{2} + \frac{3}{2}f,$$
which is also less than \( \mathbb{E}u^* \) since \( f < \frac{1}{2}(b - c) \), this is also strictly less than the expected utility of earning no money in meeting \( A \) and subsequently offering \( p_{k,t-2}^B = 1 \) and \( p_{l,t-2}^C = 1 \), which gives \( (b - f) + \frac{1}{2}(1 - f) < \mathbb{E}u^* \). In summary, regardless of how much the agent earned in meeting \( A \), it is optimal to offer \( p_{k,t-2}^B = 1 \) in the second meeting.
Finally, we consider the optimal choice for the first offer (in meeting \( A \)). Could it be optimal to offer a price \( p_{j,t-1}^A > 1 \)? No middle-aged agent would accept an offer \( p_{j,t-1}^A = 3 \) because agents prefer to consume when old in meeting \( B \). A middle-aged agent with two units of money would not accept an offer \( p_{j,t-1}^A = 2 \) for the same reason, but an agent with three units of money would accept an offer \( p_{j,t-1}^A = 2 \) because they prefer to consume in meetings \( A \) and \( B \) rather than in meetings \( B \) and \( C \) since \( 1 + b > b + c \). If the offer \( p_{j,t-1}^A = 2 \) is not accepted, then it would be optimal for the young agent to offer prices \( p_{k,t-2}^B = 1 \) and \( p_{l,t-2}^C = 1 \), which leads to a 50 percent probability of earning one unit of money and a 50 percent probability of earning two units of money as discussed above. If the offer \( p_{j,t-1}^A = 2 \) is accepted, then it is optimal to offer \( p_{k,t-2}^B = 1 \), which ensures earning three units of money while producing only two goods. Altogether, the expected utility of offering \( p_{j,t-1}^A = 2 \) is the same as that in expression (12), which is less than than \( \mathbb{E}u^* \). Hence, an offer of \( p_{j,t-1}^A = 2 \) is not optimal. Could it be optimal to make no offer in meeting \( A \)? As we have seen, it is then subsequently optimal to have \( p_{k,t-2}^B = 1 \) and \( p_{l,t-2}^C = 1 \), which yields a lower utility than \( \mathbb{E}u^* \). In summary, it is also optimal to offer \( p_{j,t-1}^A = 1 \) in the first meeting.
This concludes the argument that it is optimal for young agents to set \( p_{j,t-1}^A = p_{k,t-2}^B = p_{l,t-2}^C = 1 \) if young agents do not consider price discrimination based on the observable characteristic \( h_{i\tau}^t \).
### 4.4.2 Price discrimination with privacy in payments
All agents act to preserve their privacy when \( \beta > \delta \). The status of an agent \( i\tau \), \( s_{i\tau} \), is therefore not revealed when making a payment. However, it is still possible for young agents to learn the status of customers based on their purchasing behaviour.
Young agent $i_t$ observes the characteristic of the middle-aged agent, $h^t_{j,t-1}$, in meeting $A$, but they do not know how it is correlated with that agent’s status, $s_{j,t-1}$, at date $t$. As derived above, the middle-aged agent will accept an offer $p^A_{j,t-1} = 2$ when $m^A_{j,t-1} = 3$, and, hence, $s_{j,t-1} = H$, but not when $m^A_{j,t-1} < 3$, and, hence, $s_{j,t-1} = L$. So, young agents could try to learn the relationship between the observable characteristic $h^t_{j,t-1}$ and status $s_{j,t-1}$ by making an offer $p^A_{j,t-1} = 2$. Is it optimal for young agents to do so given that they may benefit from price discrimination in subsequent encounters?
We established that none of the agents has funds to accept an offer $p^C_{k,t-2} > 1$. So, a young agent may only benefit from price discrimination in meeting $B$ in cases where those with initial balances $m^A_{k,t-2} = 3$ (i.e., $s_{k,t-2} = H$) still have two units of money and those with initial balances $m^A_{k,t-2} = 2$ (i.e., $s_{k,t-2} = L$) still have one unit of money. Young agents could aim at offering a high price $p^B_{k,t-2} = 2$ to old agents profiled as bearing status $s_{k,t-2} = H$.
If the young agent decides to offer $p^A_{j,t-1} = 2$, two potential outcomes may occur, each with probability $1/2$. First, the offer of $p^A_{j,t-1} = 2$ may be accepted when the young agent meets a middle-aged agent with status $s_{j,t-1} = H$ in meeting $A$. In this situation, it is optimal for the young agent to make an offer of $p^B_{k,t-2} = 1$ in meeting $B$ – which will certainly be accepted – and no offer in the third encounter.
Second, the offer of $p^A_{j,t-1} = 2$ may be rejected because the young agent meets a middle-aged agent with status $s_{j,t-1} = L$. The young agent can then infer which observable characteristic $h^t_{k,t-2}$ tends to be associated with status $s_{k,t-2} = L$. If the old agent encountered in meeting $B$ bears that same observable characteristic, then the young agent makes an offer $p^B_{k,t-2} = 1$. This offer is accepted not only by old agents with status $s_{k,t-2} = L$, but also by the fraction $\varepsilon$ of old agents with status $s_{k,t-2} = H$ who are incorrectly profiled as poor customers based on the observable characteristic. If the old agent encountered in meeting $B$ bears the other observable characteristic, then the young agent infers that the old agent is likely to have status $s_{k,t-2} = H$ and makes an offer $p^B_{k,t-2} = 2$. This offer will be accepted only if the old agent’s status is indeed $s_{k,t-2} = H$, which, given the wealth distribution,
occurs with probability $1 - \varepsilon$. Regardless of the outcome in meeting $B$ it will be optimal for the young agent to make an offer $p^C_{l,t-2} = 1$ in meeting $C$.
Altogether, the expected utility of deviating to the price discrimination strategy with privacy of payments is
$$\mathbb{E}u^{D,U} = \frac{1}{2} [1 + b + c - 2f] + \frac{1}{4} \left[ b - f + \frac{1}{2}(1 - f) \right] + \frac{1-\varepsilon}{4} \left[ 1 + b - f + \frac{1}{2}(c - f) \right] + \frac{\varepsilon}{4} \left[ \frac{1}{2}(b - f) \right].$$
Comparing the levels of $\mathbb{E}u^{D,U}$ and $\mathbb{E}u^*$ gives that price discriminating in a situation with privacy of payments is only optimal whenever the profiling error $\varepsilon$ is sufficiently small; i.e., $\varepsilon < \theta^U$, where
$$\theta^U = \frac{6f - (1 - c)}{2 - 2f + b + c}.$$
This result is summarized in the following proposition:
**Proposition 1** The economic situation described in Lemma 1 is a monetary equilibrium whenever it is privately optimal for agents to preserve privacy in payments (i.e., $\beta > \delta$) and profiling errors are sufficiently large (i.e., $\varepsilon \geq \theta^U$).
### 4.4.3 Price discrimination without privacy in payments
When $\beta < \delta$, none of the agents acts to preserve their privacy. As a consequence, it is possible for young agents to learn the status of the agents in meeting $A$ whenever a payment occurs.
This allows young agents to deviate to a new strategy where they offer $p^A_{j,t-1} = 1$, learn the period $t$ relationship between $s_{i\tau}$ and $h^t_{i\tau}$ from the payment of the middle-aged agent in meeting $A$, and then use this information to price discriminate against old agents in meeting $B$ at the risk of making a profiling error because of the imperfect correlation due to $\varepsilon$. As before, practicing price discrimination can only result in additional money for a young agent in meeting $B$ because no agent will accept an offer $p^C_{l,t-2} > 1$.
The offer of $p^B_{k,t-2} = 1$ to agents with the observable characteristic associated with $s_{k,t-2} = L$ will be accepted by old agents who have status $L$ and the fraction $\varepsilon$ of old agents who have status $H$ and are incorrectly profiled as having status $s_{k,t-2} = L$. The offer of $p^B_{k,t-2} = 2$ to an agent who is profiled as having status $s_{k,t-2} = H$ is only accepted by old agents who truly have status $s_{k,t-2} = H$. The young agent benefits from collecting a total of three units of money with effort $2f$ if an offer of $p^B_{k,t-2} = 2$ is accepted. If the offer is declined, then there might still be a possibility to sell a second good to an old agent in the third encounter. Hence, the expected utility of this strategy is
$$\mathbb{E}u^{D,W} = \left[\frac{1}{2}(1 - \varepsilon)\right](b + 1 + c - 2f) + \left[\frac{1}{2}\right](b + 1 - 2f + \frac{1}{2}(c - f)) + \left[\frac{1}{2}\varepsilon\right](b - f + \frac{1}{2}(1 - f)).$$
Comparing the levels of $\mathbb{E}u^{D,W}$ and $\mathbb{E}u^*$ reveals that price discriminating without privacy of payments is only optimal when the profiling error is sufficiently small; i.e., $\varepsilon < \theta^W$, where
$$\theta^W = \frac{c + f}{2c + 1 - f}.$$
This result is summarized in our second proposition:
**Proposition 2** The economic situation described in Lemma 1 is a monetary equilibrium whenever it is privately optimal for agents not to act to preserve privacy in payments (i.e., $\beta < \delta$) and profiling errors are sufficiently large (i.e., $\varepsilon \geq \theta^W$).
### 4.5 Welfare and price discrimination
Lemma 1 shows that any monetary equilibrium that does not involve young agents offering $p^A_{j,t-1} = p^B_{k,t-2} = p^C_{l,t-2} = 1$ must be inferior from a welfare perspective. Moreover, any form of price discrimination must involve potential transactions at a price of more than one unit of money. Hence, any equilibrium that involves price discrimination is inferior from a welfare perspective relative to the economic situation described in Lemma 1. Whether a
socially inferior outcome occurs depends on whether agents act to preserve privacy in payments and on whether the probability of profiling errors $\varepsilon$ is sufficiently small. In particular, price discrimination will occur when consumer profiling is less precise if there is no privacy in payments. This result is expressed in our third proposition:
**Proposition 3** The range of $\varepsilon$ for which the best achievable social outcome corresponds to a monetary equilibrium is smaller when agents do not act to preserve their privacy.
**Proof.** Proposition 1 shows that the economic outcome described in Lemma 1 is a monetary equilibrium when $\varepsilon \geq \theta^U$ and agents act to preserve their privacy in payments. Proposition 1 shows that the economic outcome described by Lemma 1 is a monetary equilibrium when $\varepsilon \geq \theta^W$ and agents do not act to preserve their privacy in payments. Appendix B shows that $\theta^U < \theta^W$. Hence, the best achievable social outcome in the absence of free goods is a monetary equilibrium for a smaller range of $\varepsilon$ when agents do not act to preserve their privacy. ■
## 5 Policy
Figure 3 summarizes the results regarding the feasibility of the economic outcome described in Lemma 1 for two key parameters in the model: the probability of profiling errors based on observable data, $\varepsilon$, on the horizontal axis and the net private benefit of privacy in payments, $\beta - \delta$, on the vertical axis.\footnote{As $\beta - \delta$ becomes an increasingly small negative number there eventually reaches a point where it is no longer socially optimal to protect privacy in payments because the benefits in terms of avoiding price discrimination are capped by the consumption loss. We do not show this region in Figure 3.} Moving along the line from past to present illustrates two trends over the past decades that are likely to continue going forward.
The first trend is the increasing cost of maintaining privacy in payments that has resulted from the expansion of ecommerce (in general, one cannot order goods online using cash), improvements in digital payment methods, such as contactless payments, which reduce processing time at point of sale, and reduced acceptance of cash. These factors have led to a
The second trend is the enhanced ability to profile consumers. This enhanced ability results from the progress in prediction techniques and computational power, the increase in the scale and scope of data gathering and the enhanced ability to manage and combine big datasets. These developments reduce the amount of errors in attempts to profile consumers, leading to reductions in $\varepsilon$ and a movement to the left of the past situation.
Continuation of these trends may lead the economy into a situation where the optimum level of consumption may no longer be feasible, as shown in Figure 3. Such an outcome would involve price discrimination where firms use payments data to predict consumers’ willingness to pay from observable consumer data, such as address information and online profiles.
There are several policies that could promote privacy in payments among consumers and hence avoid undesirable outcomes. One option that could, at least temporally, reduce
the cost of privacy in payments is the promotion or subsidization of cash use. Naturally, such policies would require a careful balancing of all social costs and benefits of cash usage (Rogoff, 2016; McAndrews, 2017). However, such a policy leaves the net private benefit or cost of maintaining privacy unchanged for the increasing share of online purchases.\footnote{Sometimes, one has to forego anonymity in online transactions. For example, providing an address may be necessary in order to have goods delivered. However, foregoing anonymity is different from foregoing privacy in payments: foregoing anonymity means that the counterparty can identify you but does not allow insight into information related to a person’s past payment history with third parties.}
Another option is the promotion of \textit{electronic cash}. Transactions using electronic cash would differ from payments using methods facilitated by today’s mainstream commercial payments providers in the sense that it would be impossible for any commercial parties to track payments of individuals. Electronic cash could rely on privacy-enhancing techniques where information such as the identity and characteristics of the payer remain shielded from commercial parties such as the payee, the providers of the payments infrastructure and financial institutions. Arguably, electronic cash could be issued by a commercial party (e.g., the Digicash initiative of Chaum, 1985; 1992), in a decentralized collective (e.g., cryptocurrencies such as Bitcoin of Nakamoto, 2008) or by a central bank. Each of these variations has its own costs and benefits in terms of providing privacy in payments. Of course, a prerequisite for electronic cash to have an impact is that it has the ability to preserve privacy in payments while also allowing consumers to trust the technique (Kahn, 2018). Moreover, for electronic cash to reduce the costs of privacy in payments, $\delta$, it needs to be widely adopted and accepted.
In discussions on electronic cash, it is often conjectured that the government will need some possibility to access payments data in order to protect society and enforce criminal law (e.g., after the court issues a warrant). Others focus predominantly on the risks of foregoing privacy in payments with respect to the government, as has been raised in relation to China’s Social Credit System. It is possible that both these concerns can be addressed without government or the private sector having full information about payments. If commercial parties were to provide the front-end to government-issued electronic cash, then the processing of
payments and storage of payments data could be designed such that the reconstruction of someone’s payments history would require data disclosure by both the government entity and the commercial entity.
A third policy option is to counter the declining trend in $\varepsilon$. A government could limit the availability of data on observable characteristics by carefully weighing privacy concerns when releasing populational statistics (Abowd and Schmutte, 2019) and by imposing privacy regulations (e.g., the General Data Protection Directive in the European Union). However, the path to effective privacy regulation is filled with obstacles, and it is not clear how long the journey will take. For example, since privacy is a public good, one cannot simply assume that a consent-based approach will lead to socially optimal choices of privacy. Even with good privacy regulations there can be security breaches that undermine their effectiveness. Ultimately, slowing the trend in $\varepsilon$ may prove difficult, because success requires countering advances made in the areas of computing and prediction techniques that are constantly evolving.
The mechanism in the present paper illustrates a positive externality of privacy in payments. More data-sharing may in certain situations also lead to more efficient outcomes. One could think, for example, about situations where more data is used to construct better credit scores that could potentially improve the allocation of credit (Pagano and Jappelli, 1993). Another example is the situation where additional information about consumers allows producers to help consumers make better choices (Ichihashi, 2019), although it is not guaranteed that consumers benefit from revealing information in such a situation (Board and Lu, 2018).
6 Discussion
6.1 Bargaining power for consumers
Market power is essential for price discrimination. Privacy in payments, and privacy more generally, prevents young merchants in the model from exploiting situations in which they have market power. Since the young agents have all the bargaining power in the model, one may consider the economic framework in the present paper as a relatively extreme case where all retailers in the economy have monopoly pricing power.
The main results are unchanged in a more balanced specification of the model where bargaining power is distributed more evenly between merchants and customers. In particular, we explored a setting where a coin toss decides whether a new agent $i$ has all bargaining power when young, or whether it will be the middle-aged and old customers who will make a take-it-or-leave-it offer to the young agent $i$. This alternative specification does not change the equilibrium conditions for the economic situation in described Lemma 1. In particular $p_{j,t-1}^A = p_{k,t-2}^B = p_{l,t-2}^C = 1$ will still be an equilibrium whenever $\varepsilon$ is sufficiently large with exactly the same thresholds $\theta^U$ and $\theta^W$. The bargaining power does not allow middle-aged and old consumers to improve upon the unit price, so they will offer to pay a unit price for the consumable good in each meeting. The sole difference is that only the proportion of young merchants who received bargaining power based on the coin toss will be inclined to deviate from the equilibrium strategy by price discriminating if the probability of profiling errors is sufficiently small.
6.2 Increasing the quantity of money
One may be tempted to believe that the reason exchanges do not occur in every meeting in the monetary equilibrium described in Lemma 1 is a lack of money. This is not the case. Rather, some scarcity of money is needed to achieve the consumption gains that are realized in the monetary equilibrium without price discrimination. Rational behaviour requires the
young merchant to be sufficiently uncertain about the money holdings of old agents in meeting $B$ in order for any consumption to occur as a consequence of encounters in meeting $C$. If all middle-aged agents were able purchase a good for one unit of money and were to enter meeting $B$ with two units of money when old (this would be a feasible outcome with an aggregate quantity of money of $5n$), then each of the young merchants could simply make a take-it-or-leave-it offer of two in meeting $B$. All old agents would optimally choose to accept this offer. As a consequence, old agents would run out of money after accepting the offer in meeting $B$, and no consumption would occur in meeting $C$. Hence, increasing the quantity of money does not necessarily lead to an outcome with higher levels of consumption.
### 6.3 Credit arrangements
Credit arrangements cannot improve the allocation in the model. Middle-aged and old agents have nothing to offer to young agents besides money, since agents in the model can produce and sell goods only when young. As a consequence, there is no way that they could return a favour that was granted on credit unless they already had money to pay for the favour in the first place.
### 7 Concluding Remarks
Privacy is a public good in our model because it protects consumers from price discrimination. Consumers do not have to contribute to *provide* this public good. It is not like national defence or a public park. Rather, they have to take costly actions to *preserve* it. In that sense privacy is like clean air. This analogy is not perfect in our model because the actions that agents in our model take to preserve the protection against price discrimination only help others. That is, if an agent in our model fails to protect her privacy, she herself is not price discriminated against. Rather, price discrimination occurs to other customers based on observable characteristics. This is not an essential feature of our model, however.
We could have allowed individuals to make repeated transactions with the same merchant. It simply was not necessary to make our general point, which is that information collected from one consumer may reveal something about other consumers with similar observable characteristics that can be exploited by retailers. The private costs of individuals failing to protect their privacy will be less than the social costs as long as some of the cost associated with each individual’s failure to preserve their privacy is borne by others.
We contend that the increased speed and convenience of commercially provided digital payment options, which reduce the private incentives to protect privacy in payments, together with advancements in big data processing, not only promote price discrimination, but also lead to a reduction in social welfare. The notion that price discrimination decreases social welfare may seem surprising. This result arises in our model for two reasons. First, due to unavoidable errors in customer profiling, retailers will mistakenly quote high prices to some customers who have low reservation prices. As a consequence, some transactions that would increase social welfare do not occur. This first effect particularly hurts the agents in the model who have less money. Second, price discrimination reduces the purchasing power of agents who earn more money, which results in a lower overall level of consumption and also weakens incentives for young agents to earn money in the first place.
One remedy to the current trend in declining privacy in payments would be the widespread adoption of a digital cash substitute that offers users a similar level of privacy in payments as physical cash. Decentralized cryptocurrencies like Bitcoin meet some of the privacy requirements but have other features, such as costly proof-of-work transaction validation or scalability issues, that may limit their usefulness as a universal cash substitute. Other solutions include the addition of privacy features to conventional private-sector payment platforms. Tokenization (e.g., Apple Pay) is a step in this direction, but it only provides privacy from the retailer and the provider of the payment infrastructure. It does not provide privacy from the credit card company. As such, the risk that payments data might be exploited remains. The epigraph to this article includes an assertion by IMF Managing
Director Christine Lagarde that privacy in payments cannot be provided by the private sector. A compelling reason for a public solution is that the government has no profit incentive to exploit payments data. This unique feature of payment instruments provided by central banks may become increasingly important as private-sector alternatives, such as Facebook’s initiative to launch an electronic payment instrument called Libra (Bloomberg, 2019), begin to gain momentum.
## Appendix A Patents and Patent Applications
**Mastercard** (2011): “Systems and methods for analyzing and segregating payment card account profiles into clusters and targeting offers to cardholders. (...) Customers who have no transaction history with a merchant may be selected for offers based on similarities with respect to other customers of the merchant.”
**Mastercard** (2018): “Disclosed herein are systems and methods of individual level learning that include receiving purchase event data from a merchant device that indicates that a purchase event occurred by a user on a user device, and transmitting the purchase event data to an analytics server. The methods may also include processing the purchase event data. (...) When the purchase hazard probability is above a threshold, the system may push a message to the user device.”
**American Express** (2014): “End consumer-facing business entities tend to have limited data relating to their customers. While a business entity may have a customer list containing demographic information about a customer and/or a set of prior transactions conducted by the business entity and the customer, many business entities may not have additional information about their customers. (...) It would thus be useful for a business entity to enhance the value of this limited dataset to gain additional insights into its
consumer base through, for example, a cooperative data exchange.”
“In various embodiments, a node comprises a private data store and/or a public data store that comprises internal data. Internal data may be gathered from a transaction system (…) Consumer transactional data may include any data pertaining to the particular transactions in which a consumer engages during any given time period.”
“Node 204 may supplement received data (…) with third party data sources. For example, a third party data source may provide customer credit scores, social network histories (which include any information a social network may gather regarding a consumer, for example, posted messages, approximate age and gender of spouse, children and other members of household, pictures, past consumer geographic locations, patterns of past consumer geographic locations, propensity to engage in risky behaviors and the frequency of engaging in the same, marital status, substance use history, dating history, education level, present and past health status including disease status), public records, consumer transactions conducted using alternate payment systems, consumer health status, and any other data relating to consumers who may appear in the internal data.”
**Visa (2018):** “Systems, apparatus, and methods for determining incentives based on consumer history. When, how, and to whom incentives are sent can be determined. For example, an incentive can be sent to a consumer to encourage a transaction at a time when the particular consumer is predisposed to initiate the transaction. Also, an incentive for a transaction can be sent to a consumer when that transaction has a high likelihood of leading to other transactions.”
Appendix B Proofs
Proof that $\theta^U < \theta^W$ in the dynamic model
From $f < (b - c)/2$, which implies $b > 2f + c$, we have that
$$\theta^U < \frac{6f - (1 - c)}{2 + 2c}.$$
Proving that this quantity is smaller than $\theta^W$ is sufficient to prove $\theta^U < \theta^W$. Using the expression for $\theta^W$, this is equivalent to proving
$$(6f + c - 1)(1 + 2c - f) < (c + f)(2 + 2c).$$
This expression holds true, because rewriting gives
$$0 < 3c - 9cf + 1 - 5f + 6f^2,$$
$$0 < 3c(1 - 3f) + (1 - 2f)(1 - 3f). \quad \square$$
References
J.M. Abowd and I.M. Schmutte. An Economic Analysis of Privacy Protection and Statistical Accuracy as Social Choices. *American Economic Review*, 109(1):171–202, 2019.
A. Acquisti, C. Taylor, and L. Wagman. The Economics of Privacy. *Journal of Economic Literature*, 54(2):442–492, 2016.
American Express. Systems and Methods for Cooperative Data Exchange. Inventors: Bonalle et al. *U.S. Patent*, 8,676,992 B2, 2014.
S. Athey, C. Catalini, and C. Tucker. The Digital Privacy Paradox: Small Money, Small Costs, Small Talk. *NBER Working Paper*, 23488, 2017.
Y. Balasko and K. Shell. The Overlapping-Generations Model. II. The Case of Pure Exchange with Money. *Journal of Economic Theory*, 24(1):112–142, 1981.
Bloomberg. Google and Mastercard Cut a Secret Ad Deal to Track Retail Sales. *News Report (30 August)*, 2018. URL https://www.bloomberg.com/news/articles/2018-08-30/google-and-mastercard-cut-a-secret-ad-deal-to-track-retail-sales.
Bloomberg. Facebook’s Libra Gambit Forces Washington’s Hand on Crypto Policy. *News Report (20 June)*, 2019. URL https://www.bloomberg.com/news/articles/2019-06-20/facebook-s-bold-move-spurs-d-c-to-confront-its-crypto-dithering.
S. Board and J. Lu. Competitive Information Disclosure in Search Markets. *Journal of Political Economy*, 126(5):1965–2010, 2018.
D. Chaum. Security without Identification: Card Computers to Make Big Brother Obsolete. *Communications of the ACM*, 28(10):96–101, 1985.
D. Chaum. Achieving Electronic Privacy. *Scientific American (August)*, pages 96–101, 1992.
J.P. Choi, D.-S. Jeon, and B.-C. Kim. Privacy and Personal Data Collection with Information Externalities. *Journal of Public Economics*, 173:113–124, 2019.
CNBC. FICO with Chinese Characteristics: Nice Rewards, but Punishing Penalties. *News Report (March 16)*, 2017. URL https://www.cnbc.com/2017/03/16/china-social-credit-system-ant-financials-sesame-credit-and-others-give-scores-that-go-beyond-fico.html.
Deutsche Bundesbank. Payment Behaviour in Germany in 2017. *Statistical Report*, 2017.
J.A.T. Fairfield and C. Engel. Privacy as a Public Good. *Duke Law Journal*, 65:385, 2015.
Financial Times. Alibaba’s Social Credit Rating Is a Risky Game. *News Report (February 21)*, 2018. URL https://www.ft.com/content/99165d7a-1646-11e8-9376-4a6390addb44.
C.S. Henry, K.P.H. Huynh, and A. Welte. 2017 Methods-of-Payment Survey Report. *Bank of Canada Staff Discussion Paper*, 2018-17, 2018.
S. Ichihashi. Online Privacy and Information Disclosure by Consumers. *American Economic Review*, (forthcoming), 2019.
C.M. Kahn. The Threat of Privacy. *Journal of Financial Market Infrastructures*, 6(2/3):1–10, 2018.
C.M. Kahn, J. McAndrews, and W. Roberds. A Theory of Transactions Privacy. *Federal Reserve Bank of Atlanta Working Paper*, 2000-22, 2000.
C.M. Kahn, J. McAndrews, and W. Roberds. Money Is Privacy. *International Economic Review*, 46(2):377–399, 2005.
T. Khiaonarong and D. Humphrey. Cash Use Across Countries and the Demand for Central Bank Digital Currency. *IMF Working Paper*, 2019-46, 2019.
D. Kovenock and C.G. De Vries. Fiat Exchange in Finite Economies. *Economic Inquiry*, 40(2):147–157, 2002.
R. Kumar and S. O’Brien. 2019 Findings from the Diary of Consumer Payment Choice. *Statistical Report*, 2019.
C. Lagarde. Winds of Change: The Case for New Digital Currency. *Speech*, 2018. URL https://www.imf.org/en/News/Articles/2018/11/13/sp111418-winds-of-change-the-case-for-new-digital-currency.
R. Lagos and R. Wright. A Unified Framework for Monetary Theory and Policy Analysis. *Journal of Political Economy*, 113(3):463–484, 2005.
M. MacCarthy. New Directions in Privacy: Disclosure, Unfairness and Externalities. *I/S Journal of Law and Policy for the Information Society*, 6:425–512, 2010.
Mastercard. Techniques for Targeted Offers. Inventors: Alfonso et al. *U.S. Patent Application*, 2011/0106607 A1, 2011.
Mastercard. Individual Level Learning Mechanism. Inventors: J.P. Gerard, S.X. Meng and P. Hu. *U.S. Patent Application*, 2018/0211267 A1, 2018.
J. McAndrews. The Case for Cash. *ADBI Working Paper*, 679, 2017.
M. Molico. The Distribution of Money and Prices in Search Equilibrium. *International Economic Review*, 47(3):701–722, 2006.
S. Nakamoto. Bitcoin: A Peer-to-Peer Electronic Cash System. *White Paper*, 2008.
P.A. Norberg, D.R. Horne, and D.A. Horne. The Privacy Paradox: Personal Information Disclosure Intentions versus Behaviors. *Journal of Consumer Affairs*, 41(1):100–126, 2007.
Office of the Privacy Commissioner of Canada. Protecting Personal Information: Cannabis Transactions. *Guidance Document (December 17)*, 2018.
M. Pagano and T. Jappelli. Information Sharing in Credit Markets. *Journal of Finance*, 48(5):1693–1718, 1993.
K.S. Rogoff. *The Curse of Cash*. Princeton University Press, 2016.
Visa. Determining Targeted Incentives Based on Consumer Transaction History. Inventors: P. Faith and K.P. Siegel. *U.S. Patent*, 9,984,379 B2, 2018.
N. Wallace. The Overlapping Generations Model of Fiat Money. In J.H. Kareken and N. Wallace, editors, *Models of Monetary Economies*, pages 49–82. Federal Reserve Bank of Minneapolis, 1980.
N. Wallace. Whither Monetary Economics? *International Economic Review*, 42(4):847–869, 2001.
Wired. Inside China’s Vast New Experiment in Social Ranking. *News Report (December 14)*, 2017. URL https://www.wired.com/story/age-of-social-credit/.
WSJ. China’s New Tool for Social Control: A Credit Rating for Everything. *News Report (17 February)*, 2019. URL https://www.wsj.com/articles/chinas-new-tool-for-social-control-a-credit-rating-for-everything-1480351590. |
Double Horizontal Branches in NGC 6440 and NGC 6569 unveiled by the VVV Survey
Francesco Mauro\textsuperscript{a}, Christian Moni Bidin\textsuperscript{a}, Roger Cohen\textsuperscript{a}, Doug Geisler\textsuperscript{a}, Dante Minniti\textsuperscript{b,c}, Marcio Catelan\textsuperscript{b,c}, André-Nicolas Chené\textsuperscript{a,d}, Sandro Villanova\textsuperscript{a}
\textbf{Abstract}
We report the discovery of a peculiar horizontal branch (HB) in NGC 6440 and NGC 6569, two massive and metal-rich Galactic globular clusters (GGCs) located in the Galactic bulge, within 4 kpc from the Galactic Center. In both clusters, two distinct clumps are detected at the level of the cluster HB, separated by only $\sim 0.1$ magnitudes in the $K_s$ band. They were detected with IR photometric data collected with the “VISTA Variables in the Vía Láctea” (VVV) Survey, and confirmed in independent IR catalogs available in the literature, and HST optical photometry. Our analysis demonstrates that these clumps are real cluster features, not a product of field contamination or interstellar reddening. The observed split HBs could be a signature of two stellar sub-populations with different chemical composition and/or age, as recently found in Terzan 5, but it cannot be excluded that they are caused by evolutionary effects, in particular for NGC 6440. This interpretation, however, requires an anomalously high helium content ($Y > 0.30$). Our discovery suggests that such a peculiar HB morphology could be a common feature of massive, metal-rich bulge GGCs.
\textit{Subject headings:} globular clusters: general
1. Introduction
Our understanding of the complexity of Galactic Globular Clusters (GGCs) has impressively expanded in the last decade, propelled by the discovery that they can host multiple populations of stars with a different chemical enrichment history (Piotto et al. 2005). The classical text-book definition of GGCs as prototypes of a simple stellar population, i.e. a chemically homogeneous aggregate of coeval stars, is now out-dated. While a certain degree of inhomogeneity of light chemical elements is observed in nearly all GGCs (Carretta et al. 2009), a spread in iron content is a characteristic restricted to only a few very massive objects (Freeman & Rodgers 1975; Yong & Grundahl 2008; Cohen et al. 2010). Ferraro et al. (2009) discovered two horizontal branches (HBs) in the Bulge GGC Terzan 5, separated by 0.3 magnitudes in the $K_s$ band. The existence of multi-modality in the morphology of HBs has been known for nearly four decades (Harris 1975), and has been associated with the presence of multiple stellar populations since shortly thereafter (Rood & Crocker 1985). However, to date, Terzan 5 is the only GC known to have two distinct HBs. The two features in Terzan 5 have a different spatial distribution, the brighter one being more centrally concentrated, more metal rich ($\Delta \text{[Fe/H]} \sim 0.5$ dex, Origlia et al. 2011), and possibly helium enhanced (D’Antona et al. 2010) and/or younger (Ferraro et al. 2009). Lanzoni et al. (2010) confirmed that Terzan 5 is...
more massive than previously thought, and it could be the relic of a Bulge building block.
In this Letter, we show evidence that the Bulge GGCs NGC 6440 and NGC 6569 host split HBs, similar to that of Terzan 5. NGC 6440 is a high-metallicity ([Fe/H] $\approx -0.5$, Origlia et al. 2008) cluster, located 8.5 kpc from the Sun and only 1.3 kpc from the Galactic center (Harris 1996, 2010 edition, H10). NGC 6569 is slightly less metal-rich ([Fe/H] $\approx -0.79$, Valenti et al. 2011) and is found at a distance of 10.9 kpc from the Sun and 3.1 kpc from the Galactic center (H10). Both NGC 6440 and NGC 6569 are among the ten most luminous of the 64 GGCs located within 4 kpc from the Galactic center.
2. Observations and reductions
The “VISTA Variables in the Vía Láctea” (VVV) Survey (Minniti et al. 2010) is one of the six ESO Public Surveys operating on the 4-meter Visible and Infrared Survey Telescope for Astronomy (VISTA). VVV is scanning the Galactic bulge and the adjacent part of the southern disk ($-65 \leq l \leq -10$, $-2 \leq b \leq +2$), in five near-IR bands ($YZJHK_s$) with the VIRCAM camera (Emerson et al. 2010), an array of sixteen 2048$\times$2048 pixel detectors with a pixel scale of 0''.341/pix. VVV images extend four magnitudes deeper and exhibit increased spatial resolution (Saito et al. 2010) versus Two Micron All Sky Survey (2MASS, Skrutskie et al. 2006), which is particularly important for mitigating contaminated photometry in crowded regions near the Galactic center.
We retrieved from the Vista Science Archive website\footnote{http://horus.roe.ac.uk/vsa/} the VVV images of the two GGCs, pre-reduced at the Cambridge Astronomical Survey Unit (CASU)\footnote{http://casu.ast.cam.ac.uk/} with the VIRCAM pipeline (Irwin et al. 2004). The selected data consist of four frames, sampling twice each point in an area of 17' $\times$ 22' around the GGCs, in each of the $ZYJHK_s$ filters, plus 17 and 11 additional epochs in the $K_s$ passband (Saito et al. 2012) for NGC 6440 and NGC 6569, respectively. The VVV images of the two clusters, extracted from a single $K_s$ frame, are shown in Figure 1.
The PSF-fitting photometry was obtained with the VVV-SkZ-pipeline (VSp, Mauro et al. submitted), code based on DAOPHOT and ALLFRAME (Stetson 1987, 1994) procedures, optimized for the VVV data. The photometry was tied to 2MASS JHKs standards, as described in Moni Bidin et al. (2011) and Chené et al. (2012). Combining all the 36 and 24 $K_s$ measurements, the final photometric errors were 0.003 and 0.008 mag respectively at the brightness level of the cluster HB. The $K_s$ errors for NGC 6569 are constant with distance, while in NGC 6440 they increase up to 0.005 mag for $r < 0''.7$. The completeness of our photometry is heavily affected by crowding in the inner 0''.7 of both clusters, as can be appreciated in Figure 1 and 4. Due to incompleteness, $\sim 80\%$ of the detected HB stars stay outside this problematic inner region, where crowding is not a significant issue.
3. Results
The $(J-K_s; K_s)$ Hess diagrams (HD) for the stars detected within 1''.83 and 1''.66 from the cluster center, respectively, are shown in Figure 2 and 3. They were obtained calculating the number of stars in a bin of width 0.06 mag in $J-K_s$ (0.04 mag in the lower panels) and 0.04 mag in $K_s$, moved along the axis with steps of 0.015 mag in color (0.01 mag in the lower diagram) and 0.01 mag in magnitude. The HDs of both GGCs reveal a peculiar HB morphology, with two distinct clumps separated by $\sim 0.1$ mag. They will be referred to as HB-A and HB-B (lower panels of Figures 2 and 3) for the brighter and fainter one, respectively. The overdensity observed at redder color is the red giant branch (RGB) bump: in fact, Valenti et al. (2005, V05) found $K_{s,bump}=14.08$ in NGC 6569, and their $K_{s,bump}$-[Fe/H] relation predicts $K_{s,bump} \approx 14.1$ in NGC 6440, in good agreement with our data.
To verify if both HBs are real and belong to their host cluster, we checked the data for stochastic fluctuations as a cause of the overdensity, and analyzed their spatial distribution as a function of the central distance. Furthermore, we compared our data with the IR photometry of Valenti et al. (2004, V04) and V05, and with the optical HST data from Piotto et al. (2002).
Dereddening. We used the maps from Gonzalez et al. (2011) to correct for reddening. They reveal that $E(J-K_s) \sim 0.5 - 0.7$ in the $r = 1\farcs8$ field of NGC6440 under analysis. The case of NGC6569 is much less extreme, with $E(J-K_s) = 0.20 - 0.24$. The HD of NGC6440 shows a clear improvement (see Figure 2), with the two features less blurred and HB-B still presenting a slope. For NGC6569 the dereddened HD is approximately similar to the raw one, as expected.
Checking for stochastic variation. We reran the procedure on four subsets of the original data, each one containing the $ZY$ and $JHK_s$ data, but different $K_s$ epochs: one subset included only the first epoch, while a unique set of three epochs were used in each of the three following subsets. The declared photometric errors in $K_s$ passband vary from $0.007 - 0.009$ mag to $0.003 - 0.005$ mag at the level of the HB. Comparing the $(J-K_s, K_s)$ HDs, obtained with the previous spacial selection and sampling procedure, both GGC's always exhibit a split HB, with only negligible differences in their
morphology. As an additional test, we checked the HBs of other GGCs in the VVV, namely NGC 6380, NGC 6441, NGC 6528 and NGC 6553, finding no evidence of a split or peculiar HB.
**Field Contamination and Spatial Distribution.** We checked the field contribution to the HDs, selecting an annular region with the same area of the previous selection, but just outside the tidal radius. The HDs of the field are barely populated at the HB location, and the field contamination is negligible. This results is evident even in Figure 4, where the number counts drop to near-zero levels at large distances from the cluster centers.
The behavior of the stellar densities (SDs) with distance $r$ from the center is shown in Figure 4 for the two features highlighted in the lower panels of Figures 2 and 3. The SDs (stars per arcmin$^2$) were calculated with a bin width of 10'' moved at steps of 2'' for NGC 6440, while for NGC 6569 we used the values of 15'' and 6'', respectively. The SDs of the two groups steeply decay at increasing radii, in both GGCs, and their members are distributed on the CCD with circular symmetry. The radial profile of the two features in NGC 6569 is identical. The HB-B group in NGC 6440 is more populated than the brighter HB-A by a factor of two, but a Kolmogorov-Smirnov test reveals that their radial behavior coincides also in this case. The stellar counts in the inner 0'7 $\simeq 5r_c$ of NGC 6440 (where $r_c$ is the core radius from H10) are incomplete because of crowding. The photometry of NGC 6569 is also incomplete for $r < 0'7 \simeq 2r_c$.
The radial profile of NGC 6569 was fit with a King (1962) profile of the form
$$f(r) = k \left\{ \left[ 1 + \left( \frac{r}{r_c} \right)^2 \right]^{-\frac{1}{2}} - \left[ 1 + \left( \frac{r_t}{r_c} \right)^2 \right]^{-\frac{1}{2}} \right\}^2 + F,$$
where $k$ is a scale parameter, $r_t$ is the tidal radius and $F$ the field contribution. For NGC 6440, we used the approximation for $r \gg r_c$
$$f(r) = kr_c^2 \left( \frac{1}{r} - \frac{1}{r_t} \right)^2 + F.$$
For NGC 6440, the fit leads to $r_{t,A} = 5'.1 \pm 0'.7$ and $r_{t,B} = 5'.2 \pm 0'.5$, consistent with $r_t = 5'.84$ quoted by H10. The core radius and the scale parameter cannot be separated and estimated individually. The SD of the two features in NGC 6569 are compatible with $r_c = 0'.35$ and $r_t = 7.15$ (H10).
**Comparison with previous photometry.** We matched our VVV photometry of NGC 6440 and NGC 6569 with the catalogs of V04 and V05, respectively. The photometry of V04 is based on observations with the near-IR camera IRAC2@ESO/MPI 2.2m, covering a 250'' $\times$ 250'' field centered on the cluster. Similarly, the photometry of V05 was performed on data collected with the near-IR camera SOFI@ESO/NTT, covering a 300'' $\times$ 300'' field centered on the cluster. The estimated internal photometric errors are lower than 0.03 mag. Both photometries were calibrated onto the 2MASS photometric system and astrometrically corrected by using the 2MASS catalog.
For both GGCs, the luminosity distributions in the $K_s$ magnitudes of the V04 and V05 catalogs for the matched stars do not show a clear bimodal distribution. However, when the stars belonging to the HB-A and HB-B groups are identified, their luminosity distributions are different, as shown in Figure 5. For NGC6440, the Gaussian fits of the two distributions are centered at $K_{s,V04} = 13.55$ and $K_{s,V04} = 13.66$ for HB-A and HB-B, respectively, with a dispersion of $\sigma = 0.12$, while in our VVV photometry the values are $K_{s,VV} = 13.55$ and 13.67, respectively, with a dispersion of $\sigma = 0.03$. Analogously, we find $K_{s,V05} = 14.26$ and $K_{s,V05} = 14.36$, respectively, for the HB-A and HB-B clumps in NGC6569, with a dispersion of $\sigma = 0.07$, and $K_{s,VV} = 14.26$ and 14.35 for the same features in our photometry, with a dispersion of $\sigma = 0.02$. Thus, the mean magnitude of the clumps of both GGCs is identical in VVV and Valenti et al.’s catalogs, but the separation is four times more statistically significant in the VVV data. This result proves that the HBs of both GGCs are intrinsically split in magnitude, with a brighter and a fainter part that remain separated even in V04 and V05 photometry, respectively, once the stars are identified.
The strong differential reddening affecting the region of NGC6440 causes the HB to be strongly tilted at optical wavelengths, hence a simple luminosity distribution does not show a bimodal behavior. For this reason, we analyzed the HST optical data of Piotto et al. (2002) projecting the position of each HB star along the HB slope, according to the equation
\begin{align}
F555W'_{nr} &= F555W_{nr} - a[(F439W_{nr} - F555W_{nr}) \\
& \quad -(F439W_{nr} - F555W_{nr})_0],
\end{align}
where $a = 3.7$ is the slope of the HB and $(F439W_{nr} - F555W_{nr})_0 = 2.2$ is the HB mean color. As advised by the authors, the magnitudes adopted were those not corrected for reddening ($nr$). To avoid contamination from RGB stars, we selected only the sources with $18.2 \leq F555W_{nr} \leq 19.1$ and $2 \leq (F439W_{nr} - F555W_{nr}) \leq 2.4$. The distribution of $F555W'_{nr}$, calculated with bin width of 0.08 mag and step of 0.04 mag (see Figure 5), reveals a clear double peak separated by $\sim 0.23$ mag, with the fainter peak 1.6-1.7 times more populated. A similar procedure was performed for NGC6569 also, but we were not able to disentangle any bimodality.
### 4. Discussions and Conclusions
The analysis of VVV data reveals that the HB of the GGCs NGC6440 and NGC6569 is split into two distinct clumps. This behavior is not introduced by stochastic fluctuations of the density in the CMD, or induced by photometric errors, as it is found to be identical in four independent subsets of data. Field contamination is not the cause
either, because the members of both the HBs are distributed with spherical symmetry with respect to the cluster center, and their density steeply decays with distance. The separation in NGC 6440 is even cleaner after applying a differential reddening correction, while in NGC 6569 it remains similar, presenting lower differential reddening. This HB split is, however, not found in the same VVV data of four less massive Bulge GGCs.
The magnitude difference between the two HB clumps is only $\sim 0.08 - 0.1$ magnitudes in $K_s$, smaller than in Terzan 5 by a factor of three. It is thus not surprising that this HB split passed unnoticed in previous investigations, also considering the strong differential reddening affecting the NGC 6440 field ($\Delta E(J-K_s) = 0.2$, Gonzalez et al. 2011). We find that the dichotomy is blurred by observational errors in the IR photometry of V04 and V05, but the two features are clearly separated even in their data, once their members are identified in their catalog. The HST optical data of NGC 6440 from Piotto et al. (2002) show two peaks separated by 0.23 mag in the corrected magnitude $F555W'_{nr}$ defined in Equation 3.
The fainter HB of NGC 6440 is bluer than the brighter one. This resembles what was previously found in Terzan 5 (Ferraro et al. 2009; Lanzoni et al. 2010). In addition, the fainter HB of NGC 6440 is more populated than the brighter one by about a factor of two, slightly higher than what was found in the central regions of Terzan 5 ($\sim 1.6$, Ferraro et al. 2009). However, these results are not directly comparable, because our photometry is incomplete in the inner 0\farcs7 of NGC 6440. HST data suggest that in the central region this ratio could be lower ($\sim 1.6$), as expected if, analogously to Terzan 5, the brighter HB was more centrally concentrated. On the other hand, we did not detect any difference in the radial density profile of the two clumps, so the issue remains open.
The two HBs of Terzan 5 are associated with two sub-populations of different metallicity, with the brighter HB being richer in iron by $\sim 0.5$ dex (Ferraro et al. 2009; Origlia et al. 2011). According to Salaris & Girardi (2002), a difference of $\sim 0.3$ dex would be expected, if the observed HB splits are interpreted only in terms of metallicity. Nevertheless, this is only a rough upper
limit, because differences in helium content and age also can contribute to cause the same split. Origlia et al. (2008) measured the metallicity of ten stars in NGC 6440, finding a dispersion of only 0.06 dex, compatible with observational errors. However, their targets are mainly located within 0\arcmin7 of the center, where the HB-A members could be few if the population ratio is constant at all radial distances. Hence, Origlia et al.’s sample likely contains only a small quantity of HB-A stars, and their results are insufficient to exclude a metallicity spread in this cluster.
Contrary to the case of NGC 6440 and Terzan 5, the two groups identified in the HB of NGC 6569 have approximately the same color and the same radial distribution, the fainter HB B being 1.3 times more populated than the other clump. Valenti et al. (2011) measured the metallicity of six stars in this cluster, finding a bimodal distribution with two groups separated by $\sim 0.08$ dex.
It is possible that the peculiar HB morphology discovered in NGC 6440 is a pure evolutionary effect, and not a signature of the presence of sub-populations. In fact, wedge-shaped HBs are predicted under special circumstances, as depicted in Fig. 4 of Catelan & de Freitas Pacheco (1996) and Fig. 1 of Dorman et al. (1989), and statistical effects could lead to the actual bimodal distribution of HB magnitudes. These features are found in the luminosity-temperature plane, but the simulated optical CMDs reveal only a clump at the bluer (and brighter) end of the sloped HB, at variance with what is observed in HST data, while the behavior of these features in the IR bands has not been simulated. Hence, this interpretation seems unlikely, but it cannot be excluded and represents an intriguing possibility. In fact, the high metallicity alone cannot explain the formation of a wedge-like HB, and a very high helium abundance ($Y > 0.30$) is required. Such a He-enriched field population has been recently suggested in the bulge (e.g., Nataf & Gould 2012), although at higher metallicities. Hence, interpreting the HB morphology of NGC 6440 as an evolutionary effect implies that the helium content of this cluster must be anomalously high, actually higher than what is predicted at $[\text{Fe/H}] = -0.5$ by the models of Bulge chemical enrichment (e.g., Catelan & de Freitas Pacheco 1996). The split HB observed in NGC 6569, whose components are well separated and with a narrow color spread, is very different to the simulated HBs of Catelan & de Freitas Pacheco (1996) and Dorman et al. (1989). The interpretation of these features as an evolutionary effect induced by a high helium content is unlikely.
The fainter HB of NGC 6440 (HB-B) is tilted, the brightness of its stars increasing at bluer colors. This is clearly visible in Figure 2, where we indicate the direction of the reddening in the IR bands from Catelan et. al (2011) for comparison. Dereddening the photometry, the slope is still present, but the map resolution of 1\arcmin does not permit strong claims. The tilt is more pronounced in the optical HST photometry of Piotto et al. (2002), where we measured a slope $\Delta(V)/\Delta(B-V) \approx 3.7$, which is higher than the standard reddening law $R_V = A_V/E(B-V) \approx 3.1$. Moreover, the optical extinction is non-standard toward the Galactic bulge, and $R_V$ can be as low as $\sim 2.5$ (Nataf et al. 2012). In conclusion, the slope of the HB in NGC 6440 is directed approximately aligned with the reddening vector, but it is steeper than the expectations of interstellar reddening. This behavior was already observed in NGC 6388 and NGC 6441 (Sweigart & Catelan 1998; Busso et al. 2007), two other massive, metal-rich Bulge GGCs, and it was attributed to an anomalously high helium content (Caloi & D’Antona 2007). Very interestingly, the HB stars of NGC 6388 could show the same peculiar properties observed in $\omega$ Centauri (Moehler & Sweigart 2006; Moni Bidin et al. 2011), the most famous cluster hosting a complex mix of sub-populations with different chemical enrichment histories.
Our results indicate that Terzan 5 is not a unique object. A complex HB morphology could be a relatively common feature among metal-rich, massive Bulge GGCs. This is not detected in less massive objects (e.g. NGC 6528, NGC 6553), nor in equally massive but metal-poor Bulge GGCs, such as M 22 and M 28, whose HB is very extended toward the blue. The large metallicity spread observed in Terzan 5 (Origlia et al. 2011) is also present in M22 (Marino et al. 2009). Further investigations are needed to unveil if the HB splits reflect the presence of two stellar populations with different chemical composition and/or age.
We gratefully acknowledge support from the Chilean Centro de Astrofísica FONDAP No.15010003 and the Chilean Centro de Excelencia en Astrofísica y Tecnologías Afines (CATA) BASAL PFB-06/2007. ANC also gratefully acknowledges support from Comite Mixto ESO-Gobierno de Chile. This project is supported by the Chilean Ministry for the Economy, Development, and Tourism’s Programa Iniciativa Científica Milenio through grant P07-021-F, awarded to The Milky Way Millennium Nucleus; by Proyecto Fondecyt Regular #1110326; and by Proyecto Anillo ACT-86.
REFERENCES
Busso, G., Cassisi, S., Piotto, G., et al. 2007, A&A, 474, 105
Harris, W. E. 1975, ApJS, 29, 397
Carretta, E., Bragaglia, A., Gratton, R. G., et al. 2009, A&A, 505, 117
Caloi, V., & D’Antona, F. 2007, A&A, 463, 949
Catelan, M., & de Freitas Pacheco, J. A. 1996, PASP, 108, 166
Catelan, M., Minniti, D., Lucas, P. W., et al. 2011, in RR Lyrae Stars, Metal-Poor Stars, and the Galaxy, ed. A. McWilliam, 145
Chené A.-N., Borissova, J., Clarke, J. R. A., et al. 2012, ArXiv e-prints, arXiv:1206.6104
Cohen, J. G., Kirby, E. N., Simon, J. D., & Geha, M. 2010, ApJ, 725, 288
D’Antona, F., Ventura, P., Caloi, V., D’Ercole, A., Vesperini, E., Carini, R., & Di Criscienzo, M. 2010, ApJ, 715, L63
Dorman, B., VandenBerg, D. A., & Laskarides, P. G. 1989, ApJ, 343, 750
Emerson, J. & Sutherland, W. 2010, The Messenger, 139, 2
Ferraro, F. R., Dalessandro, E., Mucciarelli, A., et al. 2009, Nature, 462, 483
Freeman, K. C., & Rodgers, A. W. 1975, ApJ, 201, L71
Gonzalez, O. A., Rejkuba, M., Zoccali, M., Valenti, E., & Minniti, D. 2011, A&A, 534, A3
Harris, W. E. 1996, AJ, 112, 1487
Irwin, M. J., Lewis, J., Hodgkin, S., et al. 2004, SPIE, 5493, 411
King, I. 1962, AJ, 67, 471
Lanzoni, B., Ferraro, F. R., Dalessandro, E., et al. 2010, ApJ, 717, 653
Marino, A. F., Milone, A. P., Piotto, G., et al. 2009, A&A, 505, 1099
Minniti, D., Lucas, P. W., Emerson, J. P., et al. 2010, New A, 15, 433
Moehler, S., & Schweigart, A. V. 2006, A&A, 455, 943
Moni Bidin, C., Villanova, S., Piotto, G., Moehler, S., & D’Antona, F. 2011, ApJ, 738, L10
Moni Bidin, C., Mauro, F., Geisler, D., et al. 2011, A&A, 535, A33
Nataf, D. M., Gould, A., Fouqué, P., et al. 2012, ApJ, submitted
Nataf, D. M. & Gould, A. 2012, ApJ, 751L, 39
Origlia, L., Valenti, E., & Rich, R. M. 2008, MNRAS, 388, 1419
Origlia, L., Rich, R. M., Ferraro, F. R., et al. 2011, ApJ, 726, L20
Piotto, G., King, I. R., Djorgovski, S. G., et al. 2002, A&A, 391, 945
Piotto, G., Villanova, S., Bedin, L. R., et al. 2005, ApJ, 621, 777
Rood, R. T., & Crocker, D. A. 1985, hbuv proc., 99
Saito, R., Hempel, M., Alonso-García, J., et al. 2010, The Messenger, 141, 24
Saito, R. K., Hempel, M., Minniti, D., et al. 2012, A&A, 537, A107
Salaris, M., & Girardi, L. 2002, MNRAS, 337, 332
Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131, 1163
Stetson, P. B. 1987, PASP, 99, 191
Stetson, P. B. 1994, PASP, 106, 250
Sweigart, A. V., & Catelan, M. 1998, ApJ, 501, L63
Valenti, E., Ferraro, F. R., & Origlia, L. 2004, MNRAS, 351, 1204
Valenti, E., Ferraro, F. R., & Origlia, L. 2005, MNRAS, 361, 272
Valenti, E., Origlia, L., & Rich, R. M. 2011, MNRAS, 414, 2690
Yong, D., & Grundahl, F. 2008, ApJ, 672, L29 |
Quantum states with Einstein-Podolsky-Rosen correlations admitting a hidden-variable model
Reinhard F. Werner*
Dublin Institute for Advanced Studies, 10 Burlington Road, Dublin 4, Ireland
(Received 1 May 1989)
A state of a composite quantum system is called classically correlated if it can be approximated by convex combinations of product states, and Einstein-Podolsky-Rosen correlated otherwise. Any classically correlated state can be modeled by a hidden-variable theory and hence satisfies all generalized Bell's inequalities. It is shown by an explicit example that the converse of this statement is false.
I. INTRODUCTION
Consider a composite quantum system described in a Hilbert space \( \mathcal{H} = \mathcal{H}_1 \otimes \mathcal{H}_2 \). An uncorrelated state of this system is given by a density matrix \( W \) [i.e., an operator \( W \in B(\mathcal{H}) \) with \( W \geq 0 \) and \( \text{tr } W = 1 \)] in \( \mathcal{H} \) of the form \( W = W^1 \otimes W^2 \) for two density matrices \( W^i \in B(\mathcal{H}_i) \). This is equivalent to saying that the expectation value \( \text{tr}(WA_1 \otimes A_2) \) for the joint measurement of observables \( A^i \in B(\mathcal{H}^i) \) (\( i = 1, 2 \)) on the respective subsystems always factorizes, i.e.,
\[
\text{tr}(WA^1 \otimes A^2) = \text{tr}(W \cdot A^1 \otimes 1)\text{tr}(W \cdot 1 \otimes A^2)
\]
\[
= \text{tr}(W^1A^1)\text{tr}(W^2A^2).
\]
Such uncorrelated states can be prepared very easily by using two preparing devices for systems 1 and 2, which function independently and yield the states \( W^1 \) and \( W^2 \), respectively. Then the factorization property means that if the measuring devices described by \( A^1 \) and \( A^2 \) also operate independently, we are simply conducting two separate experiments at the same time and the classical multiplication rule for probabilities applies.
One way of preparing correlated states is the following: Suppose that each of the two preparing devices has a switch with settings \( r = 1, \ldots, n \), say, and that with setting \( r \) the device \( i \) produces systems in the state \( W^i_r \). Suppose we have also a random generator, which produces numbers \( r = 1, \ldots, n \) with probability \( p_r \). We can combine these three devices into a new preparing apparatus by the following prescription: In each individual experiment one first draws a random number \( r \in \{1, \ldots, n\} \). The switches of the two preparing devices are then set according to the result. Clearly then, the expectation of a measurement of observables \( A^1 \) and \( A^2 \) will be
\[
\sum_{r=1}^{n} p_r \text{tr}(W^1_r A^1)\text{tr}(W^2_r A^2) = \text{tr}(W \cdot A^1 \otimes A^2)
\]
with the density matrix \( W = \sum_{r=1}^{n} p_r W^1_r \otimes W^2_r \), i.e., \( W \) is a convex combination of product states. Expectation values for this state \( W \) no longer factorize. The physical "source" of these correlations is the random generator, which can be chosen as a purely classical device. Therefore, we shall call a density matrix classically correlated if it can be approximated (e.g., in trace norm) by density matrices of the form (1). States that are not classically correlated have been called EPR correlated\(^1\) to emphasize the crucial role of such states in the Einstein-Podolsky-Rosen paradox, and for the violations of Bell's inequalities (see below). EPR correlation and classical correlation are defined as a property of the density matrix \( W \). Since there are usually very different ways of preparing the same state \( W \), classical correlation does not mean that the state has actually been prepared in the manner described, but only that its statistical properties can be reproduced by a classical mechanism.
The terminology "classically correlated" is further justified by the observation that in classical probability theory all states have this property. States in probability theory are given by probability measures, and the state of a composite system is given by a probability measure on a product space. Like every probability measure this can be represented as a limit of convex combinations of measures concentrated on a single point. And since the point measures on a product space are product measures, we conclude that any probability measure on a product space can be represented as a limit of convex combinations of product measures, i.e., is classically correlated in the above sense. In the wider context of \( C^* \)-algebraic quantum theory,\(^2\) it is known\(^3\) that if one of two subsystems of a composite system is classical (i.e., has a commutative observable algebra), all states of the composite system are classically correlated.
For any set of correlations determined in an experiment one can raise the question whether these correlations can be described within a purely classical "hidden-variable" theory. Such a theory is based on some probability space \( (\Omega, \Sigma, M) \), called the space of hidden variables, consisting of a \( \sigma \)-algebra \( \Sigma \) of subsets of \( \Omega \) and a \( \sigma \)-additive normalized measure \( M \) on \( \Sigma \). For any measuring device \( A \) with possible outcomes \( v \), one demands the existence of a measurable response function \( \omega \mapsto F_A(v, \omega) \in \mathbb{R} \) interpreted as the probability that the outcome \( v \) is obtained in an experiment with known value \( \omega \in \Omega \) of the hidden variables. Therefore the response functions must satisfy \( F_A(v, \omega) \geq 0 \) and \( \sum_v F_A(v, \omega) = 1 \) for every \( \omega \in \Omega \), where the sum is over all possible outcomes of the measurement of \( A \). A hidden-variable model for some set of correlations is then given by a probability space and a collection of response functions such that the probability in an experiment with a measuring device $A$ on system 1 and a device $B$ on system 2 for obtaining the result $\nu$ on $A$ and simultaneously the result $\mu$ on $B$ is given by the expression $\int M(d\omega)F_A(\nu,\omega)F_B(\mu,\omega)$.
Any hidden-variable model can be extended to measurements with continuous outcome parameters, and can be modified to a "deterministic model" in which the response functions take only the values 0 and 1. Therefore we shall stay with the above simple definition. The existence of a hidden-variable model is exactly the hypothesis of the usual derivations of Bell's inequalities.\textsuperscript{4} The "locality" of the theory being expressed by the fact that the response function for $A$ is independent of $B$ and vice versa. It is known that while these inequalities are necessary conditions for the existence of a hidden-variable model, they are not sufficient.\textsuperscript{5} On the other hand, the set of correlations admitting hidden-variable models is a convex set, and as such is completely described by some set of linear inequalities. We shall refer to any one of these inequalities as a (generalized) Bell's inequality. Despite some partial results in this direction,\textsuperscript{6} no efficient procedure for obtaining all generalized Bell's inequalities is known.
An interesting question is, then, whether or not the correlations described by a quantum state of a composite system admit a hidden-variable model. The measuring devices of system 1 are then represented by observables, i.e., by Hermitian operators $A \in B(\mathcal{H}^1)$ with spectral resolution $A = \sum_\alpha \alpha P_\alpha$, i.e., eigenvalues $\alpha_\nu$ and eigenprojections $P_\nu$. We then say that a state $W \in B(\mathcal{H}^1 \otimes \mathcal{H}^2)$ admits a hidden-variable model if there are a probability space $(\Omega, \Sigma, M)$ and response functions, defined for all Hermitian $A = \sum_\alpha \alpha P_\alpha \in B(\mathcal{H}^1)$ and $B = \sum_\mu \beta Q_\mu \in B(\mathcal{H}^2)$ with discrete spectrum, such that for all $A$, $B$, $\nu$, and $\mu$,
$$\int M(d\omega)F_A(\nu,\omega)F_B(\mu,\omega) = \text{tr}(W \cdot P_\nu \otimes Q_\mu). \quad (2)$$
We claim that all classically correlated states admit hidden-variable models, and hence satisfy all Bell inequalities. This can be proven quite simply for convex combinations of products as in (1). We then take $\Omega = \{1, \ldots, n\}$, $M(\{\nu\}) = p_r$, $F_A(\nu, r) = \text{tr}(W_r \cdot P_\nu)$, and define $F_B$ analogously. Then Eq. (1) implies Eq. (2). We omit the somewhat technical, but straightforward approximation arguments needed to extend this result to all classically correlated states.
We conclude that every state violating some generalized Bell's inequality, i.e., any state not admitting a hidden-variable model, cannot be classically correlated, i.e., is EPR correlated. The well-known experiments demonstrating a violation of Bell's inequalities can thus be taken as direct experimental evidence for the existence of EPR correlated states. The vital importance of such states for quantum theory is further underlined by the fact that they are automatically generated by an interacting time evolution. To be precise, any unitary time evolution, which takes all classically correlated initial states again to classically correlated states, necessarily factorizes into the product of two separate time evolutions. Consequently, the ground state of an interacting system, which is often especially easy to prepare, is usually not classically correlated. The states of a relativistic quantum-field theory are even more universally EPR correlated, since any state of finite energy (in particular, the vacuum) violates Bell's inequalities for suitable spacelike localized observables.\textsuperscript{7}
Since any state violating some generalized Bell's inequality is EPR correlated, one might conjecture that the converse holds, i.e., that every state admitting a hidden-variable model is classically correlated. This conjecture is indeed true for pure states of a composite quantum system, given by unit vectors in $\mathcal{H}^1 \otimes \mathcal{H}^2$. (This can be demonstrated by constructing violations of Bell's inequalities using techniques from Ref. 8). The purpose of this paper is to show that this conjecture is false for general mixed states. We shall do this by explicitly constructing hidden-variable models for a family of quantum states, which are not classically correlated.
II. CONSTRUCTION OF THE EXAMPLE
There are two main difficulties in constructing an example of a state with nonclassical correlations, which nevertheless admits a hidden-variable model. The first is to prove that some explicitly given density matrix $W \in B(\mathcal{H}^1 \otimes \mathcal{H}^2)$ is not classically correlated, i.e., that there is no way in which it can be represented as a convex combination of product states. The second difficulty is to verify Eq. (2), which is an infinite system of equations, indexed by the set of all observables. We shall circumvent both difficulties by considering only states of very high symmetry. To be specific, we take $\mathcal{H}^1 \cong \mathcal{H}^2 \cong \mathbb{C}^d$ as Hilbert spaces of equal finite dimension $d$. The states considered are those given by density matrices $W$ satisfying $(U \otimes U)W(U^* \otimes U^*) = W$ for all unitary matrices $U$ on $\mathbb{C}^d$. We shall call such states $U \otimes U$ invariant for short.
Before embarking on the construction of the hidden-variable model, it is useful to establish some facts about such states. The first is that any operator $A$ commuting with all unitaries of the form $U \otimes U$ is a linear combination of 1 and the "flip" $V$ defined by $V \varphi \otimes \psi = \psi \otimes \varphi$. This is seen most easily by considering matrix elements $\langle \varphi_n \otimes \varphi_m, A \varphi_p \otimes \varphi_q \rangle$ for some basis $\varphi_1, \ldots, \varphi_d$. Since for all $r$ there is a unitary operator taking $\varphi_r$ to $-\varphi_r$ leaving the other basis elements fixed, the matrix element vanishes unless $\{n, m\} = \{p, q\}$. Since any permutation of the basis can also be realized unitarily, $A$ depends only on the three matrix elements with $n = m = p = q$, $n = p \neq q = m$, and $n = q \neq p = m$. These are linked further by considering for $U$ rotations in a two-dimensional plane, making $A$ depend on just two complex parameters, which can be taken as the coefficients of 1 and $V$, or as $\text{tr}A$ and $\text{tr}(AV)$. For density matrices $W$ tr$W = 1$ is fixed, so that the set of $U \otimes U$ invariant states is parametrized by the single parameter $\Phi = \text{tr}(WV)$. Since $V^2 = 1$ and $V^* = V$ we have $-1 \leq \Phi \leq 1$. Note that $\Phi = 1$ corresponds to a state with Bose symmetry and $\Phi = -1$ corresponds to a state with Fermi symmetry. In terms of $\Phi$ the state is described by
\[ W = (d^3 - d)^{-1} \left[ (d - \Phi) \cdot 1 + (d\Phi - 1) \cdot V \right], \tag{3a} \]
\[ \text{tr}(WA \otimes B) = (d^3 - d)^{-1} \left[ (d - \Phi)(\text{tr}A)(\text{tr}B) \right. \]
\[ + (d\Phi - 1)\text{tr}(AB) \left] , \tag{3b} \]
where we have used the formula \( \text{tr}(V \cdot A \otimes B) = \text{tr}(A \cdot B) \).
There is a linear projection \( P \) mapping arbitrary density matrices \( W \) to \((U \otimes U)\)-invariant ones by \( P(W) = \int dU(U \otimes U)W(U^* \otimes U^*) \), where \( dU \) denotes the Haar measure of the unitary group of \( \mathbb{C}^d \). This integral does not have to be calculated explicitly, since its value depends only on the two parameters \( \text{tr}P W = \text{tr}W \), and
\[ \text{tr}[(PW)V] = \int dU \text{tr}[W(U^* \otimes U^*)V(U \otimes U)] = \text{tr}(WV) . \]
Suppose now that \( W \) is classically correlated. Then so is each \((U \otimes U)W(U^* \otimes U^*)\), and hence, by approximating the integral by suitable Riemann sums, \( PW \) is represented as the limit of convex combinations of classically correlated states, and is hence itself classically correlated. Thus in order to determine the set of all \((U \otimes U)\)-invariant, classically correlated states we only have to compute the possible values of \( \Phi = \text{tr}(WV) \) for classically correlated \( W \). For a product state \( W = W^1 \otimes W^2 \) we find that \( \Phi = \text{tr}(WV) = \text{tr}(W^1 \cdot W^2) \) is positive. Hence the same is true for convex combinations and for norm limits of such states, i.e., for all classically correlated states. Moreover the extremal value \( \Phi = 0 \) (respectively, \( \Phi = 1 \)) is attained by taking \( W^1 \) and \( W^2 \) to be orthogonal (respectively equal) pure states. In short, a \((U \otimes U)\)-invariant density matrix is classically correlated if and only if \( \Phi = \text{tr}(WV) \) is positive.
For constructing the hidden-variable model we shall utilize the \( U \otimes U \) symmetry by choosing the space \( \Omega \) as the unit sphere \( \{\omega \in \mathbb{C}^d ||\omega|| = 1\} \) and \( M \) as the unique probability measure on \( \Omega \) invariant under all unitary rotations of the sphere. The response functions \( F_A \) of an observable \( A = \sum_\nu \alpha_\nu P_\nu \) will be taken to depend only on the family \( \{P_\nu\} \) of orthogonal projections, but not on the eigenvalues \( \alpha_\nu \), or the labeling of the projections \( P_\nu \). Moreover, it suffices to consider only the case where each \( P_\nu \) is one dimensional, since for projections of higher dimension the response function can be chosen as a sum of response functions of one-dimensional projections. The symmetry under unitaries of \( \mathbb{C}^d \) will be imposed by the relation
\[ F_{U^*AU}(\nu, \omega) = F_A(\nu, U\omega) . \tag{4} \]
The simplest choice for such functions is the one we shall adopt for system 2:
\[ F_B(\mu, \omega) = \langle \omega, Q_\mu \omega \rangle \quad \text{where} \quad B = \sum_\mu \beta_\mu Q_\mu . \tag{5} \]
Note that for any positive integrable function \( \rho: \Omega \to \mathbb{R} \) there is a unique positive operator \( \hat{\rho} = \int M(d\omega)|\omega\rangle\langle\omega| \) such that
\[ \int M(d\omega)\rho(\omega)F_B(\mu, \omega) = \text{tr}(\hat{\rho} \cdot Q_\mu) . \]
This will allow us to compute the integrals (2). A characteristic feature of the response functions \( F_B \) is that \( F_B(\mu, \omega) \) depends only on \( Q_\mu \), but not on the remaining projections in the spectral family of \( B \). It is vital for our construction that the choice of response functions \( F_A \) for system 1 does not have this property. Suppose to the contrary, that \( F_A(\nu, \omega) = f(P_\nu, \omega) \) depends only on \( P_\nu \). Then for each fixed \( \omega \) the map \( P \mapsto f(P, \omega) \) is additive on families of mutually orthogonal projections. Hence (assuming \( d \geq 3 \)) by Gleason’s theorem there is a density matrix \( W_\omega \) such that \( f(P, \omega) = \text{tr}(W_\omega P) \). Then Eq. (2) holds for the density matrix \( W = \int M(d\omega)W_\omega \otimes |\omega\rangle\langle\omega| \), but this state \( W \) is explicitly represented as an average of product states and is thus classically correlated. We note in passing that the reason not to settle for the simplest case \( d = 2 \) in the present paper was exactly to exclude the suspicion that the result is a spurious effect due to the failure of Gleason’s theorem for \( d = 2 \).
Using only Eqs. (4) and (5), we can reduce the calculation of all the integrals in Eq. (2) to the computation of a single integral. By the choice of \( F_B \) there exists, for each \( A \) and \( \nu \) a positive operator \( \hat{F}_A(\nu) \) such that
\[ \int M(d\omega)F_A(\nu, \omega)F_B(\mu, \omega) = \text{tr}[\hat{F}_A(\nu) \cdot Q_\mu] . \tag{6} \]
Using the \( U_d \) invariance of \( M \), we find
\[ \text{tr}(U\hat{F}_A(\nu)U^* \cdot Q_\mu) = \int M(d\omega)F_A(\nu, \omega)\langle\omega, U^*Q_\mu U\omega \rangle \]
\[ = \int M(d\omega)F_A(\nu, U^*\omega)\langle\omega, Q_\mu \omega \rangle \]
\[ = \int M(d\omega)F_{UAU^*}(\nu, \omega)F_B(\mu, \omega) \]
\[ = \text{tr}[\hat{F}_{UAU^*}(\nu) \cdot Q_\mu] . \]
Since this holds for all one-dimensional projections \( Q_\mu \), we find \( U\hat{F}_A(\nu)U^* = \hat{F}_{UAU^*}(\nu) \). In particular, \( \hat{F}_A(\nu) \) commutes with all unitaries commuting with \( A \), that is to say it has a representation \( \hat{F}_A(\nu) = \sum_\lambda \lambda(\nu, \mu)P_\mu \). Since \( F_A \) is not to depend on the labeling of the projections \( P_\mu \), we conclude that \( \lambda(\nu, \mu) \) depends only on whether or not \( \nu = \mu \). Hence \( \hat{F}_A(\nu) = \lambda_1 P_\nu + \lambda_2 1 \) for some \( \lambda_1, \lambda_2 \in \mathbb{R} \), which are independent of \( A \). Since \( \sum_\nu \hat{F}_A(\nu) = 1 \) we must have \( \lambda_1 + d\lambda_2 = 1 \). Hence for computing all \( \hat{F}_A(\nu) \) it suffices to compute the integral (6) with \( Q_\mu = P_\nu \), for which the right-hand side of (6) is \( \lambda_1 + \lambda_2 \). Comparing Eqs. (6) and (3) we find that Eq. (2) holds with \( \Phi \) determined by
\[ \Phi = -1 + (d^2 + d) \int M(d\omega)F_A(\nu, \omega)\langle\omega, P_\nu \omega \rangle . \tag{7} \]
It remains to construct a family of response functions \( F_A(\nu, \omega) \) satisfying (4), for which \( \Phi \) becomes negative. Thus we have to make the integral (7) as small as possible under the constraints \( F_A(\nu, \omega) \geq 0 \) and \( \sum_\nu F_A(\nu, \omega) = 1 \) for all \( \omega \) and all \( A \). Since for every \( \omega \) and every \( A \) these constraints single out a convex set in \( \mathbb{R}^d \), we expect that the smallest values of \( \Phi \) is attained for response functions taking only the values 0 and 1. This suggests the following choice:
\[ F_A(\nu, \omega) = \begin{cases}
1 & \text{if } \langle \omega, P_\nu \omega \rangle < \langle \omega, P_\mu \omega \rangle \text{ for all } \nu \neq \mu \\
0 & \text{if } \langle \omega, P_\nu \omega \rangle > \langle \omega, P_\mu \omega \rangle \text{ for some } \mu \neq \nu ,
\end{cases} \tag{8} \]
Note that we have left $F_A(v, \omega)$ unspecified at all points where $\langle \omega, P_v \omega \rangle$ is the minimum of the numbers $\langle \omega, P_\mu \omega \rangle$, but not the unique minimum. However, since this set is of measure zero, it will not contribute to the integral (7) anyway, and we may choose on this subset any measurable function satisfying the constraint. We have written Eq. (8) in such a form that the covariance property postulated in Eq. (4) is manifest. Moreover, $F_A$ is independent of the labeling of the $P_v$ in the sense that it only depends on $\langle \omega, P_v \omega \rangle$ and the set of numbers $\langle \omega, P_\mu \omega \rangle$, but not on their ordering.
Since with Eq.(8) the integrand in Eq. (7) is a function of the $d$ numbers $\langle \omega, P_\mu \omega \rangle$, we need to compute the probability distribution of these special random variables with respect to the measure $M$. Choosing a basis $\psi_\mu$, in which the $P_\mu$ are diagonal, we have $\langle \omega, P_\mu \omega \rangle = x_\mu^2 + y_\mu^2$, where $\omega = \sum_\mu (x_\mu + iy_\mu) \psi_\mu$. A function $f$ depending only on $u_\mu(\omega) = x_\mu^2 + y_\mu^2$ has expectation value
$$\int M(d\omega) f(u_1(\omega), \ldots, u_d(\omega))$$
$$= N \int dx_1 dy_1 \cdots dy_d \delta \left[ 1 - \sum_\mu u_\mu(\omega) \right]$$
$$\times f(u_1(\omega), \ldots, u_d(\omega))$$
$$= \pi^d N \int du_1 \cdots du_d \delta \left[ 1 - \sum_\mu u_\mu(\omega) \right] f(u_1, \ldots, u_d),$$
where $N$ denotes a normalizing factor, and we have substituted $dx_\mu dy_\mu = \frac{1}{2} du_\mu d\varphi_\mu$ and have integrated out the angle variables $\varphi_\mu$. We thus obtain a probability measure on the simplex (generalized tetrahedron) $\{ (u_1, \ldots, u_d) | \sum_\mu u_\mu = 1 \}$, which is just the restriction of the Lebesgue measure on the hyperplane $\sum_\mu u_\mu = 1$ to the simplex. Remarkably, this measure depends on the choice of the complex number field for the Hilbert spaces $\mathcal{H}^1, \mathcal{H}^2$. For the real and quaternion fields, which axiomatic quantum mechanics has been forced to consider, as well, we would obtain an additional factor $(\prod_\mu u_\mu)^{\alpha}$, with $\alpha = -\frac{1}{2}$ for the reals, emphasizing the corners of the simplex, and $\alpha = 1$ for the quaternions, emphasizing the center.
With this probability distribution of $\langle \omega, P_\mu \omega \rangle$, the computation of the integral in (7) can now be made without pencil and paper. After the above-mentioned substitution, we have to compute an integral over a regular simplex $S$ with $d$ vertices, embedded into $\mathbb{R}^{d-1}$, and which is best imagined to “stand” on the face with $u_v = 0$, so that $u_v$ represents the “vertical” coordinate. By definition of the measure, the total volume of $S$ is 1. The subset in which $u_v \leq u_\mu$ for all $\mu$ is again a simplex $S'$ with the same base, whose apex is the barycenter of $S$, which is at height $u_v = d^{-1}$. The volume of $S'$ is $d^{-1}$ since $S$ is the disjoint union of $d$ pieces congruent to $S'$. We have to compute the integral of $u_v$ over $S'$. This is equal to the height of the barycenter of $S'$, which is $d^{-1}$ times the height of $S'$, which is equal to $d^{-2}$, multiplied by the volume of $S'$, which is $d^{-1}$. Hence the integral is $d^{-3}$, and we obtain our final result
$$\Phi = -1 + d^{-2}(d+1).$$
The positive zero of this expression, considered as a polynomial of $d^{-1}$, is at the inverse golden ratio, hence $\Phi < 0$ and, in fact, $\Phi \leq -\frac{1}{4}$ for all $d \geq 2$. Hence we have constructed a hidden-variable model for a $U_d$-invariant state [see Eq. (3) with $\Phi$ given by Eq. (9)], which is not classically correlated since $\Phi < 0$. A surprising aspect of Eq. (9) is that for large $d$, $\Phi$ approaches $-1$, which is as far removed from the classically correlated states $\Phi \geq 0$ as is compatible with positivity of $W$. Note also that there is no *a priori* reason why a hidden-variable model should give correlations compatible with a quantum-mechanical positivity condition. This is shown by a strange phenomenon happening for large values of $\Phi$: If one chooses $F_A(v, \omega) = \langle \omega, P_v \omega \rangle$ instead of (8), one obtains a hidden-variable model for the $U_d$-invariant state $\mathbb{P}W$, where $W$ is a pure state with wave vector $\varphi \otimes \varphi$. As shown above this is the classically correlated state with $\Phi = +1$. Suppose, however, that we define $F_A$ as in Eq. (8), but with reversed inequality signs. Then it is easy to see that the integral in Eq. (7) is larger than for $F_A(v, \omega) = \langle \omega, P_v \omega \rangle$. Hence we obtain a hidden-variable model for a “state” with $\Phi > 1$ (e.g., $\Phi = \frac{5}{4}$ for $d = 2$). Thus the correlations are given by (positive) expressions of the form $\text{tr}(W \cdot P_v \otimes Q_\mu)$ as in Eq. (2), but $W$ is not a positive operator. Thus there must be some positive operator $C \in B(\mathcal{H}^1 \otimes \mathcal{H}^2)$ with $\text{tr}(WC) < 0$, and the only “paradox” here is that by the positivity of all $\text{tr}(W \cdot P_v \otimes Q_\mu)$ this $C$ cannot be a positive linear combination of products of positive operators, i.e., $C$ cannot be classically correlated.
Finally, we would like to mention a possible extension of this result. Up to now we have identified quantum-mechanical observables with decompositions $1 = \sum_v P_v$ of the identity into mutually orthogonal projections $P_v$. However, for many purposes this concept turns out to be too narrow, and it is useful to allow, instead of projections, arbitrary positive operators $G_v$ with $1 = \sum_v G_v$, where $\text{tr}(WG_v)$ is interpreted as the probability of the outcome $v$ of some measuring device described by the decomposition $\{ G_v \}$. In order to also represent observables of this type in the hidden-variable model one would have to define the corresponding response functions $F_G$, and compute the integrals of Eq. (6) also for these. Due to the lack of symmetry, it is a much more difficult task than for the projection valued case to do this explicitly. Even the case $d = 2$ seems to involve some nontrivial geometrical estimates. We conjecture, however, that the existence of hidden-variable models for some nonclassically correlated states can be demonstrated also in this wider setting.
**ACKNOWLEDGMENTS**
The author would like to thank Guido Raggio for stimulating discussions. The support by the Alexander von Humboldt Foundation is gratefully acknowledged.
*On leave from Fachbereich Physik, Universität Osnabrück, Postfach 4469, D-4500 Osnabrück, West Germany.
1H. Primas, in *Moderne Naturphilosophie*, edited by B. Kanitschneider (Königshausen und Neumann, Würzburg, 1984).
2R. Haag and D. Kastler, J. Math. Phys. 5, 848 (1964).
3G. Wittstock, in *Foundations of Quantum Mechanics and Ordered Linear Spaces*, Vol. 29 of *Lecture Notes in Physics*, edited by A. Hartkämper and H. Neumann (Springer-Verlag, Berlin, 1974).
4J. F. Clauser and A. Shimony, Rep. Prog. Phys. 41, 1881 (1978).
5A. Garg and N. D. Mermin, Phys. Rev. Lett. 49, 1220 (1982).
6A. Garg and N. D. Mermin, Found. Phys. 14, 1 (1984).
7S. J. Summers and R. Werner, Ann. Inst. Henri Poincare 49, 215 (1988).
8S. J. Summers and R. Werner, J. Math. Phys. 28, 2440 (1987).
9A. M. Gleason, J. Ration. Mech. Anal. 6, 885 (1957).
10C. Piron, *Foundations of Quantum Physics* (Benjamin, Reading, MA, 1976).
11G. Ludwig, *Foundations of Quantum Mechanics I* (Springer, Berlin, 1983). |
Power Conservative Multi-Attribute Queries on Data Broadcast
Qinglong Hu*
Aleph Computer System, Inc.
1700 Shattuck Ave., Suite 1
Berkeley, CA 94709, USA
firstname.lastname@example.org
Wang-Chien Lee
GTE Laboratories Inc.
40 Sylvan Road
Waltham, MA 02451, USA
email@example.com
Dik Lun Lee
Dept. of Computer Science
Univ. of Science and Technology
Clear Water Bay, Hong Kong
firstname.lastname@example.org
Abstract
In this paper, we study power conservation techniques for multi-attribute queries on wireless data broadcast channels. Indexing data on broadcast channels can improve client filtering capability, while clustering and scheduling can reduce both access time and tune-in time. Thus, indexing techniques should be coupled with clustering and scheduling methods to reduce the battery power consumption of mobile computers. In this study, three indexing schemes for multi-attribute queries, namely, index tree, signature, and hybrid index, are discussed. We develop cost models for these three indexing schemes and evaluate their performance based on multi-attribute queries on wireless data broadcast channels.
1. Introduction
Wireless communication and mobile computing have gained much attention from the computer and communication research community. Among the various issues under study, efficient utilization of limited wireless bandwidth and battery power is critical for disseminating information to a large user population.
Wireless broadcast is an attractive approach for data dissemination, because it allows simultaneous information access by an arbitrary number of users, and thus facilitates efficient bandwidth usage. Meanwhile, mobile computers accessing data by monitoring broadcast channels consume less battery power than those accessing data by sending requests through traditional point-to-point communications. A lot of studies on data broadcast have appeared in the literature [1, 2, 9, 10, 3].
Power conservation techniques for broadcast data management include indexing methods and data organization methods (i.e., clustering and scheduling). Among the proposed indexing techniques, signature[8] and index-tree [7] methods represent two major classes\(^1\) of indexing techniques for data broadcast. Evaluation has been conducted on each of the indexing technique. In recent papers [4, 5], we examined these two indexing techniques based on clustering and scheduling, and proposed a hybrid indexing method which combined the strength of the signature and index-tree methods.
All of the above studies focused on data management techniques for single-attribute based queries. In the real world applications, data items usually contain multiple attributes. Thus, multi-attribute based queries may provide more precise information to the users. In this paper, we investigate the issues of multi-attribute based queries on wireless data broadcast channels. We study three different power conservative indexing techniques, namely index tree, signature, and hybrid, and their supporting data organizations for broadcast of multi-attribute data items. For each method, query processing algorithms for multi-attribute queries are presented. We develop cost models and evaluate the access time\(^2\) and tune-in time\(^3\) for the multiple attribute queries. Finally, empirical comparisons of the query performance are conducted.
The rest of the paper is organized as follows. Section 2 gives an informal introduction of the broadcast channels. In Section 3, indexing techniques and access methods for multiple attributes are introduced. Section 4 evaluates the indexing techniques and query efficiency for multiple attributes. Finally, Section 5 concludes the paper.
2. Wireless data broadcast
In a wireless data broadcast environment, a base station maintains a set of records for formatted data and multime-
\(^1\)In [6], hashing and flexible indexing methods were explored.
\(^2\)The average time elapsed from the moment a query is issued to the moment when all the requested data frames are received.
\(^3\)The period of time spent by a mobile computer staying active in order to obtain the requested data.
dia such as text, image, audio and video. To facilitate our discussion, we assume that each record consists of a number of common attributes. To support broadcast data delivery, the base station periodically delivers these records to its clients as a series of *data frames* through one *shared* broadcast channel. A data frame, the logical unit of information broadcasted on the air, consists of *packets*\(^4\) which are the physical units of broadcast. A frame contains a header for synchronization and meta-information that indicates the type and length of the frame. Since data frames are periodically broadcast, a complete broadcast of data frames is called a *broadcast cycle*. From the user’s viewpoint, broadcast data are perceived as a stream of frames flowing along the time axis. Logically, there is no specific start and end frame for a broadcast cycle. A broadcast cycle starts with any frame and ends when the frame appears again. Updates to the frames are reflected between successive broadcast cycles. To receive the frames from the air, mobile applications or users issue queries which specify values for one or multiple common attributes to their mobile computers. As a result, the mobile computers monitor the broadcast channels and return the frames that satisfy these queries.
**Indexing techniques** is based on the idea that by interleaving auxiliary information with data frames on broadcast channels, mobile computers are able to predict the arrival of requested data frames. Hence, by staying in doze mode most of the time and only waking up to receive the data when they arrive (also called *selected tune-in*), power consumption of the mobile clients can be conserved. Since only the qualified data are retrieved into the mobile clients, CPU time and cache space are also saved.
In addition to indexing, broadcast data organization and data access methods are two other major factors determining the access time and the tune-in time of mobile computers. In the following, we first discuss the assumptions about scheduling and clustering we make in this paper. The intention is to fix the related parameters about data organization on broadcast channel, so that we can focus on the indexing techniques and access methods in the next section.
**Data scheduling** determines the contents of the broadcast cycle and the broadcast frequency of the data frames [1, 9, 10] (i.e., the hot data set for broadcast and the relative broadcast frequency of each frame). A simple broadcast schedule, called *flat broadcast*, is to broadcast each data frame once in every cycle [1]. Since the client access pattern for an attribute is usually skewed (i.e., some data are accessed more frequently than others), another scheduling method, called *broadcast disks*, was proposed in [1]. Broadcast disks broadcast important data more often than the other data to reduce the average access time for queries based on that attribute value. However, broadcast disks increase the length of a broadcast cycle and the client has to spend more time to retrieve the less commonly requested data. For multi-attribute data records, a cycle can be organized in broadcast disks based on one of attributes. Due to data frame duplication, queries other than that attribute may take longer to answer. Therefore, broadcast disks are not efficient for data records with multiple attributes. In this paper we assume that flat broadcast scheduling is used.
**Data clustering** refers to arranging the data items with the same value for a specific attribute to appear consecutively in a broadcast cycle [5, 6, 7]. By monitoring the arrival of the first data item with the desired attribute value, the client can retrieve all of the successive data items with the same attribute value. However, similar to data scheduling, a broadcast cycle can only be clustered based on one attribute. We call this attribute the *clustered attribute* and the other attributes are *non-clustered attributes*. Although the other attributes are non-clustered in the cycle, a second attribute can be chosen to cluster the data items within a data cluster of the clustered attribute. In turns, a third attribute can be chosen to cluster the data items within a data cluster of the second attribute.
For each non-clustered attribute, the broadcast cycle can be partitioned into a number of segments called *meta segments* [7], each of which holding a sequence of frames with non-decreasing (or non-increasing) values of that attribute. Thus, when we look at each individual meta segment, the data frames are clustered on that attribute and indexing techniques designed for clustered data broadcast can still be applied within the meta segment. To facilitate our study, we define the number of meta segments in the broadcast cycle for an attribute as the *scattering factor*\(^5\) of the attribute. For multiple attributes, the broadcast cycle is partitioned into meta segments for each attribute in the order of decreasing access frequency. The scattering factors of an attribute increases as the importance of the attribute decreases. Organizing data broadcast with the clustering structure discussed above can improve retrieval efficiency. Thus, in this paper, we assume that the data frames within a broadcast cycle are partitioned into meta segments based on the first \(r\) attributes.
### 3. Multiple attribute indexing techniques
In this section, we investigate the application of the index tree, the signature and the hybrid techniques to broadcast data with multiple attributes. The cost models of access time and tune-in time for the investigated indexing techniques are derived. The comparisons among these three techniques based on access time and tune-in time are presented in the next section. To save space, we do not repeat
\(^4\)In our analysis, we assume that the size of a data frame is a multiple of the packet size.
\(^5\)To simplify our discussion, we neglect the variance of the meta segment size.
the basic index tree, signature and hybrid techniques in the paper. Reader can refer to [5, 7, 8] for detail.
Before we discuss the multi-attribute indexing techniques and their associated access methods, we first describe the system parameters used in the study. We assume that there are $m$ common attributes in each data frame and the attributes are sorted based on their access frequency. Let the ordered attributes be $a_1, a_2, ..., a_m$ and their probabilities contained in a query be $p_1, p_2, ..., p_m$, where $p_i \geq p_{i+1}$ $(1 \leq i < m)$. $a_1$, called *major attribute*, is the most frequently accessed attribute and all other attributes are called *minor attributes*. The minor attribute $a_m$ is the least accessed attribute. Indexes are built on multiple attributes (i.e. the frequently accessed attributes, $a_1, a_2, ..., a_r$ where $1 \leq r \leq m$). The rest of the system parameters are defined in Table 1.
**Table 1. System Parameter**
| $F$ | Information frame number in a broadcast cycle |
|-----|-----------------------------------------------|
| $P$ | Average packet number in an information frame |
| $m$ | Attribute number in a frame |
| $r$ | Indexed attribute number in a frame $(0 \leq r \leq m)$ |
| $q$ | Attribute number in a query $(1 \leq q \leq m)$ |
| $S_i$ | percentage of frames with required attribute value |
| $p_i$ | Probability of queries based on $a_i$ |
A multi-attribute query usually contains multiple attributes and consists of many combinations of Boolean operators, such as conjunction ($\land$) and disjunction ($\lor$). Since it is difficult to consider all combinations in a study, we choose the queries with either all conjunction or all disjunction operators as the representative multiple attribute queries for evaluation.
Let $Q\{a'_1, \ldots, a'_q\}$ denote a query with $q$ attributes, where the attributes $a'_i$ $(1 \leq i \leq q \leq m)$ are sorted according to the order they are indexed. Without loss of generality, we assume that $q$ attributes are the first $q$ frequently accessed attributes, $a_1, ..., a_q$. $Q\{a_1 \land \ldots \land a_q\}$ and $Q\{a_1 \lor \ldots \lor a_q\}$ are the only two query expressions. This is the minimal configuration to cover the common queries involved.
To provide the comparison baseline, we first introduce the non index approach (denoted *non-index*). Since every arriving frame must be retrieved into the client cache to check against the attribute values specified in the query, the tune-in time is very long and is equal to the access time. The estimated access time and tune-in time for $Q\{a_1 \land \ldots \land a_q\}$ is as follows.
$$A(a_1 \land \ldots \land a_q) = T(a_1 \land \ldots \land a_q) = P \cdot F \cdot (1/2 + S_1)$$
where $a_1$ is clustered. $A$ and $T$, hereafter, are used to denote the access time and tune-in time of the queries. For $Q\{a_1 \lor \ldots \lor a_q\}$, the client must scan the entire broadcast cycle to retrieve the desired frames:
$$A(a_1 \lor \ldots \lor a_q) = T(a_1 \lor \ldots \lor a_q) = P \cdot F$$
### 3.1. The index tree method
For a multiple-attribute data set, we can build index trees for each attribute separately. Several factors influence the order of attribute indexes in a cycle. To reduce index tree overhead for multiple attributes, the index tree should be built in the order that the attribute with the highest selectivity is indexed first. In this way, the clustered data frames can be received successively. On the other hand, to achieve the best average system performance, index should be built in the order of access frequency such that the higher the access frequency the higher the priority in choosing an attribute for indexing. From the index point of view, an index built on an attribute with a lower selectivity is better, because low selectivity attribute has high partitioning power and consequently high filtering capability. Obviously, the factors of selectivity and access probability have contradicting effects. We have to balance the access probability and the attribute selectivity to determine the order of attributes to be indexed.
Based on research results in [5, 7], data access is more efficient for clustered attribute than for non-clustered attribute. For example, both access time and tune-in time are improved by clustering. In addition, index built on clustered attribute requires less overhead than that built on non-clustered attributes. Since data frames can be clustered on at most one attribute, we choose to cluster data frames in a broadcast cycle based on major attribute (i.e., $a_1$). Thus, the minor attributes are non-clustered. Let $I = \{a_1, \ldots, a_r\}$ be the set of indexed attributes. The broadcast cycle is partitioned separately for each attribute in $I$ ordered from $a_1$ to $a_r$. Hence, the scattering factor for $a_i$, denoted as $M_i$, increases with $i$. The index tree overhead of $a_i$ depends on $M_i$. Hence, the smaller the subscript of the attribute in $I$, the lower the index overhead.
Since the broadcast channel is a linear medium and the index information increases the length of the cycle, index for an attribute influences the performance of queries based on not only that attribute but of those based on other attributes as well. When $M_i$ (where $i > r$) is too large, constructing a distributed indexing tree on $a_i$ does not help improve tune-in time much but results in a high index overhead. For example, the broadcast cycle can become very long and the access time for queries based on any attribute increases considerably. In this case, non-index is a better approach.
The value of $M_j$ depends on $S_i$ and the inter-relation among $a_i$, where $1 \leq i \leq j$. However, the dependency
among the attributes is difficult to determine because it is most likely semantic-based and it may vary from application to application. To simplify the cost model, we assume that the attributes are random and independent. Based on this assumption, a simple estimate on $M_i$ is $M_i = 1 / \prod_{j=1}^{(i-1)} S_j$.
**Table 2. Parameter for Index Tree**
| $h_i$ | Height of the whole index tree built for $a_i$ |
|-------|-----------------------------------------------|
| $t_i$ | Height of the replicated tree part for $a_i$ |
| $N$ | Number of packets in an index tree node |
| $n$ | Search-key plus pointer number a node holds |
Table 2 describes the parameter settings for index tree cost models. In addition, we let $X[h]$ and $X[t]$ be the total number of nodes of the full index tree and the replicated part of the index tree, respectively. Also, the number of nodes in the $i$-th level of the index tree is denoted as $L[i]$. For multi-attribute indexing, extra index information increases the cycle length to $P \cdot F + N \cdot \sum_{j=1}^{r} E_j$ packets, where $E_j$ is the total number of index nodes allocated to $a_j$. For $a_1$ ($j \neq 1$), there is an index tree constructed for each meta segment. Therefore, there are $M_j$ index trees corresponding to $a_j$. Since each index tree has $X[h_j] + L[t_j + 1] - 1$ index nodes allocated for $a_j$ within a meta segment, the total overhead incurred in building index trees for $a_j$ in a broadcast cycle is $E_j = M_j \cdot (X[h_j] + L[t_j + 1] - 1)$ nodes.
To simplify the cost models, we average the index tree overhead to each data frame so that the size of a frame is considered to consist of a data part and an index overhead part. Of course, the actual index tree overhead for each data frame is different, but from the statistical point of view we can assume that all data frames have the same average index tree overhead which is $TREE = N \cdot \sum_{j=1}^{r} E_j / F$. The replicated index tree part for $a_i$ is broadcast every $1 / L[t_i + 1]$ of each meta segment. The data frames are divided into $M_i \cdot L[t_i + 1]$ data segments with replicated index nodes at the beginning of each data segment. The length of each data segment is $(P \cdot F + N \cdot \sum_{j=1}^{r} E_j) / (M_i \cdot L[t_i + 1])$. Hence, the initial probe time for the index tree built for $a_i$ can be estimated as follows [7],
$$PROBE^{idx} = \frac{P \cdot F + N \cdot \sum_{j=1}^{r} E_j}{2 \cdot M_i \cdot L[t_i + 1]}$$
In the following paragraph, for conjunction operators ($\land$) and disjunction operators ($\lor$), their access methods are described and then the cost formulae are derived. The general access method for $Q\{a_1 \land \ldots \land a_q\}$ is:
**Search:** FOR each query attribute $a_i$ DO
- Initial probe for the index tree built based on $a_i$ within the data segment qualified for $a_{i-1}$.
- Search the index tree built based on $a_i$: the client follows a list of pointers to find out the arrival time of the desired data frame.
**Retrieval:** Scan the current meta segment for the desired data frames. At the end of the meta segment a jump is made to the other meta segments where the client should examine whether that meta segment is qualified for the attributes $a_1, \ldots, a_q$ contained in the query.
The access time for $Q\{a_1 \land \ldots \land a_q\}$ is upper bounded by:
$$A^{idx}(a_1 \land \ldots \land a_q)$$
$$= \text{initial probe time for major attribute index} + \text{waiting time for first desired data segment} + \text{access time for desired data within the segment}$$
$$= PROBE^{idx} + \frac{P \cdot F + N \cdot \sum_{j=1}^{r} E_j}{2} + P \cdot S_1 \cdot F$$
The tune-in time for $Q\{a_1 \land \ldots \land a_q\}$ depends on the number of levels in the index trees for $a_i$, $1 \leq i \leq r$ and the selectivity of the conjunction query. The initial probe of the client is to find the occurrence of the control index for $a_1$. The control index directs the client to the required higher level index tree nodes for $a_1$. A search in the index tree is conducted and the last pointer in the index tree points to the occurrence of the desired data segment. Within that data segment, a probe for index tree for $a_2$ is conducted. Once the index tree for $a_2$ is obtained, a search in the index tree is carried out to find the data segment which matches $a_2$. The filtering process continues with all other attributes involved with the query. Finally, the data frames of interests are received. The tune-in time is obtained as follows:
$$T^{idx}(a_1 \land \ldots \land a_q)$$
$$= (1 + \sum_{i=1}^{l} h_i) \cdot N + (1 + F \cdot \prod_{i=1}^{l} S_i) \cdot P$$
where $l = min(r, q)$ and $\prod_{i=1}^{l} S_i$ is the number of matched items for $Q\{a_1 \land \ldots \land a_l\}$.
For $Q\{a_1 \lor \ldots \lor a_q\}$, the client monitors the channel for every index tree built on the attributes in the query as if the client queries each of the $q$ attributes simultaneously, but within one scan. The following is the access method:
**Initial probe:** For any index trees built on $a_1, \ldots, a_q$, the client tunes into the broadcast channel and determines when the next index tree nodes for the data frames are broadcast.
**Search:** For each index tree built on $a_1, \ldots, a_q$, the client follows a list of pointers to find out the arrival time of the desired data frame.
**Retrieval:** The client tunes into the broadcast channel to download all the qualified frames.
The access time for $Q\{a_1 \lor \ldots \lor a_q\}$ is,
$$A^{idx}(a_1 \lor \ldots \lor a_q) = \text{initial probe time} + \text{waiting time for retrieving all desired frames} = PROBE_i + P \cdot F + N \cdot \sum_{j=1}^{r} E_j$$
and the tune-in time for $Q\{a_1 \lor \ldots \lor a_q\}$ is as follows:
if $r < q$
$$T^{idx}(a_1 \lor \ldots \lor a_q) = r \cdot N + P \cdot F$$
else (i.e., $q \leq r$)
$$T^{idx}(a_1 \lor \ldots \lor a_q) \leq (1 + \sum_{i=1}^{q} h_i) \cdot N + P \cdot \sum_{i=1}^{q} M_i + F \cdot P \cdot \sum_{i=1}^{q} (-1)^{i+1} S^i \binom{q}{i}$$
where the number of true matches is no greater than $P \cdot F \cdot \sum_{i=1}^{q} (-1)^{i+1} S^i \binom{q}{i}$ and $S = \max\{S_i : 1 \leq i \leq q\}$.
### 3.2. Signature for multiple attributes
| $P_i$ | a frame false drop probability for $a_i$ |
|-------|------------------------------------------|
| $P_S^f$ | a simple signature false drop probability |
| $P_I^f$ | an integrated sig. false drop probability |
| $k$ | number of frames indexed by an integrated sig. |
| $l_i$ | average number of true matches in a frame group |
| $p$ | number of bits in a packet |
| $R_S$ | number of packets a simple signature takes |
| $R_I$ | number of packets of an integrated sig. takes |
| $s$ | number of bit strings superimposed into a sig. |
For the index tree method, there is an index tree for each indexed attribute. Therefore, the index overhead is directly proportional to the number of indexed attributes. For the signature method, one signature can provide index information to all attributes. The index overhead is much less influenced by the number of attributes and the access method is simpler than the index tree method.
Table 3 defines the parameters for signature cost models. For multi-attribute indexing, the multi-level signature method is the best approach [8, 5] and is used in this paper for data indexing. We assume that every attribute is indexed in the signature (i.e., $r = m$). The the signature false drop probability for either simple signature or integrated signature can be estimated as follows [8]:
**LEMMA 1 (Optimal false drop probability)** Given the size of a signature, $R$, hereafter denotes either $R_S$ or $R_I$, and the number of bit strings superimposed into the signature, $s$, the false drop probability for the signature is:
$$P_S^f = P_I^f = 2^{-R \cdot (p \ln 2)/s}$$
We define $P_i$ as the false drop probability of a data frame for queries based on $a_i$. $P_i$ has two components $P_I^f$ and $P_S^f$, which, respectively, reflects the false drops introduced by the integrated signature and the simple signature. Based on [8], we derive the estimate of $P_i$ as follow:
$$P_i = \left( (1/k - \lceil S_i/l_i \rceil) \cdot P_I^f \cdot (k \cdot R_S/P + k \cdot P_S^f) + \lceil S_i/l_i \rceil \cdot (k - l_i) \cdot P_I^f \right)/(1 - S_i)$$
To simplify our discussion, we assume that frames with the same attribute value for $a_i$ ($1 \leq i \leq m$) are evenly distributed in each meta segment. Consequently, the average number of frames with the same value for $a_i$ in each meta segment is $\lceil F \cdot S_i/M_i \rceil$. Let $k$ be the number of data frames grouped in an integrated signature. The number of distinct attribute values in $k$ data frames can be estimated as $\lceil k/\lceil F \cdot S_i/M_i \rceil \rceil$. For frames in a meta segment partitioned for $a_i$, the average number of qualified frames corresponding to a matched integrated signature, called *locality* of true matches $l_i$ ($1 \leq l_i \leq k$ for $1 \leq i \leq r$), can be estimated as, $l_i = k/\lceil k/\lceil F \cdot S_i/M_i \rceil \rceil$.
For randomly distributed frames, the locality of true matches corresponding to $a_i$ ($r < i \leq m$), $l_i$, equals to 1. Since each data frame contains $m$ attribute values\footnote{We assume that the hashed attribute values for different attributes are not the same (i.e., independent event). Therefore, the number of distinct bit strings superimposed into a simple signature for $m$ attributes is $m$.} corresponding to $a_1, \ldots, a_m$, the expected distinct superimposed bit strings, $s$, for an integrated signature can be estimated as, $s = \sum_{i=1}^{m} \lceil k/l_i \rceil$, where $\lceil k/l_i \rceil$ is the expected number of distinct bit strings superimposed for $a_i$.
To simplify the estimates of the tune-in time, the initial probe for the integrated signature is a true match. The average waiting time for retrieving one data frame is $SIG_S + SIG_I + P$, where $SIG_S$ and $SIG_I$, the average simple and integrated signature overheads for a data frame, can be calculated as $SIG_S = R_S$ and $SIG_I = R_I/k$.
Generally speaking the length of a signature ($R_S$ or $R_I$) is very small compared to a data frame (i.e., $P >> R$). Otherwise, both the access time and the tune-in time would be large. Therefore, the time for the client to filter out the partial signature and the partial data frame, called initial probe time, can be be approximated as half of the data frame size.
$PROBE = P/2$. The tune-in time in the initial probe period is the expected sum of the time when the client is active for filtering a partial simple signature, a partial integrated signature, and a partial data frame. For similar reasons, the tune-in time for filtering a partial signature is negligible and the initial tune-in time can be approximated as $P/2$.
One major difference between the index tree and the signature method is that an index tree node is good for only one indexed attribute, while a signature contains the information for all indexed attributes. Therefore, the client first retrieves the signature into the cache, by scanning that signature, the client knows whether the associated data frame matches more than one attribute. Only when the signature matches all query attributes, is the corresponding data frame fetched into the cache.
We develop the access method for queries $Q\{a_1 \land \ldots \land a_q\}$ as follows. Signature match for $a_1$ is conducted first. Matched signatures (either true match or false drops) are subjected to further signature matches based on the other query attribute values.
**Initial probe:** The client tunes into the broadcast channel for the first received signature
**Filtering:** For each signature received in a broadcast cycle do the following:
- for each query attribute $a_i$,
- IF the signature does not match the query signature based on attribute $a_i$
- THEN repeat the filtering step
**Retrieval and checking:** The client retrieves frames corresponding to the matched signature for further checking to eliminate false drops. Repeat filtering step.
For $Q\{a_1 \land \ldots \land a_q\}$, the access time is:
$$A^{sig}(a_1 \land \ldots \land a_q)$$
$$= \text{initial probe time} + \text{filtering time for the first desired frame} + \text{retrieving time for all the desired frames}$$
$$= PROBE + CYCLE/2 + S_1 \cdot F$$
$$= P/2 + F \cdot (SIG_I + SIG_S + P)/2 + P \cdot S_1 \cdot F$$
The tune-in time for $Q\{a_1 \land \ldots \land a_q\}$ is:
$$T^{sig}(a_1 \land \ldots \land a_q)$$
$$= \text{the tune-in time in the initial probe period} + \text{true match data frames in the broadcast} + \text{every integrated signature in half the cycle} + \text{simple sig associated with qualified integrated sig} + \text{false drop data frames in half the cycle}$$
$$= \frac{P}{2} + P \cdot F \cdot \prod_{i=1}^{q} S_i + \frac{F}{2} \cdot (R_I/k + \lceil S_1/l_1 \rceil k \cdot R_S)$$
+ false drop data frames in half the broadcast
$$\leq \frac{P}{2} + P \cdot F \cdot \prod_{i=1}^{q} S_i + \frac{F}{2} \cdot (R_I/k + \lceil S_1/l_1 \rceil k \cdot R_S)$$
$$+ P \cdot F \cdot \sum_{i=1}^{q} (q \cdot S^{i-1} \cdot D^{q+1-i})/2$$
where the number of true matches is estimated as $F \cdot \prod_{i=1}^{q} S_i$, $S = \max\{S_i : 1 \leq i \leq q\}$, $P_i$ is given in Equation (1), $P_i \cdot (1 - S_i)$ is the false drop probability for $a_i$, $D = \max\{P_i \cdot (1 - S_i) : 1 \leq i \leq q\}$.
For a query $Q\{a_1 \lor \ldots \lor a_q\}$, the access method is:
**Initial probe:** The client tunes into the broadcast channel for the frame signature in a broadcast cycle.
**Filtering:** The client tunes into retrieve the signatures arriving and matches the frame signatures with the signatures for any attribute values in $a_1, \cdots, a_q$.
**Retrieval and checking:** For the matched signatures, the client tunes into get the corresponding data frames from the channel for further checking to eliminate false drops.
The access time for $Q\{a_1 \lor \ldots \lor a_q\}$ is:
$$A^{sig}(a_1 \lor \ldots \lor a_q)$$
$$= \text{initial probe time} + \text{a broadcast cycle}$$
$$= P/2 + F \cdot (SIG_I + SIG_S + P)$$
and the tune-in time for $Q\{a_1 \lor \ldots \lor a_q\}$ is upper bounded by:
$$T^{sig}(a_1 \lor \ldots \lor a_q)$$
$$= \text{tune-in time in the initial probe period} + \text{true match data frames in the cycle} + \text{every integrated signature in the cycle} + \text{simple sig associated with qualified integrated sig} + \text{false drop data frames in the cycle}$$
$$\leq P/2 + F \cdot P \cdot \sum_{i=1}^{q} (-1)^{i+1} S^i \left( \begin{array}{c} q \\ i \end{array} \right) + F \cdot (R_I/k + \lceil S_1/l_1 \rceil k \cdot R_S) + \text{false drop data frames in the whole broadcast}$$
$$\leq P/2 + F \cdot P \cdot \sum_{i=1}^{q} (-1)^{i+1} S^i \left( \begin{array}{c} q \\ i \end{array} \right) + F \cdot (R_I/k + \lceil S_1/l_1 \rceil k \cdot R_S) + P \cdot F \cdot \sum_{i=1}^{q} ((1 - S_i) \cdot P_i)$$
where the number of true matches is no greater than $P \cdot F \cdot \sum_{i=1}^{q} (-1)^{i+1} S^i \binom{q}{i}$ and $S = \max\{S_i : 1 \leq i \leq q\}$.
### 3.3. The hybrid methods
The hybrid index consists of two parts: sparse index tree and signature. The sparse tree is built on the major attribute and used for global filtering. The multi-level signature is used to carry out the local filtering based on attribute values specified in the queries. In this way, the index tree part helps improve tune-in time of the client. The average waiting time for retrieving one data frame from the broadcast cycle can be expressed as:
$$TREE + SIG + P = N \cdot (X[t_1+1] - 1) / F + R \cdot (1 + 1/k) + P$$
where $TREE$ and $SIG$ respectively are the index overheads of the index tree part and the signature part of a frame. The average number of data frames in one data block, $F[B]$, can be calculated in a similar way as in the index tree method, which is $F' L[t_1 + 1]$. Thus, the overheads of the index tree and the signatures in a data block are $F'[B] \cdot TREE$ and $F'[B] \cdot SIG$, respectively. Hence, the average initial probe time for the signature and the length of a cycle is given by:
$$PROBE^{hyb} = (TREE + SIG + P)/2$$
$$CYCLE^{hyb} = (TREE + SIG + P) \cdot F$$
The access method for a query $Q\{a_1 \land \ldots \land a_q\}$ is as follows:
**If** $a_1$ is the major attribute
- based on the single index tree access protocol to retrieve the frame block which contains all the frames satisfying the value for $a_1$
- successive signature matches based on $\{a_2, \ldots, a_q\}$ are conducted to filter out the desired frames
**Else** based on the signature access method to retrieve the qualified frames based on $\{a_1, \ldots, a_q\}$. The sparse index tree is skipped.
Therefore, the access time for $Q\{a_1 \land \ldots \land a_q\}$ is:
$$A^{hyb}(a_1 \land \ldots \land a_q)$$
$$= PROBE^{hyb} + CYCLE^{hyb}/2 + F \cdot S_1 \cdot P$$
$$= (TREE + SIG + P) \cdot (1/F')/2 + F \cdot S_1 \cdot P$$
The tune-in time primarily depends on the initial probe of the client to determine the next occurrence of the sparse index, the access time for the index tree part which equals to the number of levels $t_1$ of the sparse index tree, the tune-in time for the data block $B$, and the selectivity of the query $Q\{a_1 \land \ldots \land a_q\}$. Therefore, it is upper bounded by:
$$T^{hyb}(a_1 \land \ldots \land a_q)$$
$$= (1 + t_1) \cdot N + T(B) + (1 + F \cdot \prod_{j=1}^{q} S_j) \cdot P$$
where $T(B)$ is the tune-in time for filtering the data block $B$ with the signature method, which can be estimated as:
$$T(B) = \text{every sig in the data block B} + \text{false drop data frames in data block B}$$
For $Q\{a_1 \lor \ldots \lor a_q\}$, the sparse index tree part does not help the retrieval of multiple attributes and the access method is similar to the multi-level signature. Thus, the access time for $Q\{a_1 \lor \ldots \lor a_q\}$ is:
$$A^{hyb}(a_1 \lor \ldots \lor a_q) = PROBE^{hyb} + CYCLE^{hyb}$$
$$= (TREE + SIG + P) \cdot (1/2 + F)$$
In order to skip each of the index tree, we assume that the client needs to retrieve one tree node to get the control information such as the size of the sparse index tree and the size of the data block. Therefore, the tune-in time for $Q\{a_1 \lor \ldots \lor a_q\}$ is:
$$T^{hyb}(a_1 \lor \ldots \lor a_q) = N + T^{sig}(a_1 \lor \ldots \lor a_q)$$
where, $T^{sig}(a_1 \lor \ldots \lor a_q)$ represents the tune-in time for the corresponding signature scheme used (i.e., multi-level signature).
### 4. Evaluation of query performance
In this section, we make analytical comparisons of the performance for the multi-attribute indexes discussed in the previous sections. The comparisons are based on the formulae developed in Section 3. The access time and the tune-in time for the index tree, signature, and hybrid methods, are evaluated. The study addresses two kinds of Boolean query expressions: conjunction and disjunction.
Table 4 lists the parameter settings used in the comparisons. We assume that the broadcast data file contains 10 attributes (i.e., $a_1, \ldots, a_{10}$). For multiple attribute queries, three most frequently accessed attributes (i.e., $a_i, 1 \leq i \leq 3$) are included in the queries. Flat broadcast schedule is used for data broadcasting. A data frame consists of 1000 packets and a packet contains 64 bits. For the index tree, each tree node, which consists of 100 packets, contains 128 search-key plus pointers. For the signature method, the frame group size is 4. To simplify the discussion, the selectivity of all attributes is set to the same value (i.e., 0.01).
Table 4. Parameters of the cost models
| $F = 10^4 \sim 10^6$ | $P = 10^3$ | $S = 0.01$ | $p = 64$ | $m = 10$ |
|----------------------|------------|------------|---------|---------|
| $n = 128$ | $N = 100$ | $k = 4$ | $r = 6$ | $q = 3$ |
For the index tree method, the index overhead is proportional to the number of attributes. For the signature and the hybrid methods, the index overhead is much less influenced by the number of attributes. Therefore, among the 10 attributes, the first six frequently accessed attributes are indexed in the index tree method and all attributes are indexed for the signature and the hybrid methods. The index tree is balanced (all leaves are on the same level), and each node has the same number of children.
To show the experimental results in a clear way, access time overhead is computed with respect to non-index approach. ON the other hand, tune-in time for non-index approach is included in the figures as a baseline for tune-in time comparisons. We first investigate multi-attribute conjunction queries. Then, we evaluate the multi-attribute disjunction queries.
4.1. Conjunction queries
In this set of experiments, we assume the conjunction queries are used and the queries contain three most frequently accessed attributes $a_1$, $a_2$, and $a_3$. Figures 1 and 2 illustrate the access time and the tune-in time for $Q\{a_1 \land a_2 \land a_3\}$ when the cycle length is varied. Obviously, the access time of the index tree method is much worse than the other two index methods. This is because the index tree overhead for non-clustered attributes (i.e., $a_i, i > 1$) is very large. Based on the construction of the multiple-attribute index tree, the less the attribute accessed, the higher its index overhead. On the other hand, for the signature and the hybrid methods, the index overhead is almost independent of the number of attributes and is much less than that of the index tree. The sparse index tree in the hybrid method has a very low overhead. As a result, the access time for the signature and the hybrid methods is similar (Figure 1).
When the tune-in time is considered (refer to Figure 2), the signature method performs worse than the other two index methods and the index tree shows a marginally better performance than the hybrid method. This is because the tune-in time for signature false drop is greater than that for searching in the index tree. All three index methods give a considerably better performance than the non-index approach.
Figure 3 shows the tune-in time as a function of the query selectivity for $Q\{a_1 \land a_2 \land a_3\}$ where the cycle length is $10^5$. For any query selectivity, the index tree always has the similar performance as the hybrid method. For small query seselectivity, the signature does worse than the other two. However, as the query selectivity is increased, all three methods give similar performance, because for large selectivity the tune-in time for the retrieval of the qualified data dominates the tune-in time of the queries.
Intuitively, a conjunction query with more attributes in the predicates should result in a smaller set of qualified data items. This is demonstrated in Figure 4 where the number of attributes in a query is varied while the cycle length is fixed to $10^5$. As the number of attributes in a query increases, the tune-in time for all index methods decreases as a result of the decrease in the number of qualified data. The tune-in time for the signature method is similar to that of the index tree and the hybrid methods only for single attribute query. For queries with more than one attribute, the signature has an obvious poor performance than the other two. That is because the tune-in time for eliminating false matches may become too significant even though the tune-in time for receiving qualified frames decreases. For queries with few attributes (i.e., less than three), the index tree and the hybrid methods have similar performance. However as the number of attributes in a query is increased, the hybrid method outperforms the index tree method. Due to the tune-in time for each index trees, the curve representing the index tree method goes up a little bit after the number of attributes is greater than three.
### 4.2. Disjunction queries
This section investigates the performance of disjunction queries. In Figures 5 and 6, the access time and the tune-in time for $Q\{a_1 \lor a_2 \lor a_3\}$ are shown. Similar to conjunction queries, the index tree method has the worst access time among the three index methods and the other two have similar access time. Figure 6 shows all three index methods have much better tune-in time performance than the non-index method. The hybrid and the signature have similar tune-in time performance which is better than that of the index tree.
The query selectivity is varied in Figure 7 while other parameters are fixed. The tune-in time for the signature and the hybrid methods is proportional to the selectivity of queries, while their performance are about the same. For the index tree method, the tune-in time is similar to that of other index methods when query selectivity is greater 6%. It has an odd performance when query selectivity is less than 6%. For example, the tune-in time first decreases and then increases. This probably is due to two contradicting factors: 1) a larger selectivity may reduce the size of an index tree and thus introduce lower index overhead; 2) a larger selectivity may result in more qualified frames and thus higher tune-in time for retrieving these frames. After the query selectivity is greater than 6%, the retrieval factor dominates the tune-in time of the query.
In contrast to conjunction queries, the more attributes involved in a disjunction query may result in the more qualified data retrieved. Thus, Figure 8 shows a reversed performance compared to Figure 4. For queries with many attributes, the index tree does worse than the other two. This is due to the retrieval of each meta segment in each index tree. The number of meta segments retrieved increases rapidly when the number of attributes involved in a query.
5. Conclusions and future work
In this paper, we examined the problem of multi-attribute queries for wireless data broadcast. We studied the index tree, the signature, and the hybrid index methods for multi-attribute queries. To facilitate our evaluation, we assumed a query involves all of the attributes either conjunctively or disjunctively.
We provided cost models of the three multi-attribute indexing methods for the estimates of access time and tune-in time and compared their performance. We found that the index tree method, while performing well for single-attribute queries results in large access time overhead. This is due to the creation and replication of index trees for the indexed attributes. Moreover, the index method has an update constraint, i.e., updates of a data frame are not reflected until the next broadcast cycle. The comparisons revealed that all index methods introduced certain access time overhead while improving power conservation by reducing their tune-in time. We conclude from our study that the hybrid is the best choice for multi-attribute queries due to its good access time and tune-in time. The signature method has the similar performance as the hybrid method except conjunction queries. The index tree method is poor in access time for any types of multi-attribute queries. It gives similar tune-in time as the hybrid method for the conjunction queries.
In this paper, we only studied two simple multi-attribute queries: conjunction and disjunction. In the future, we will investigate more complicated ones such as, mixed conjunction and disjunction queries, join queries, and range queries etc. In a mobile computing environment, index information is frequently accessed data. Caching this index information in the clients may reduce both access time and tune-in time considerably. One future research would incorporate the index schemes with data caching algorithms to achieve a better system performance.
References
[1] S. Acharya, R. Alonso, M. Franklin, and S. Zdonik. Broadcast disks: Data management for asymmetric communications environments. In *Proceedings of the ACM SIGMOD Conference on Management of Data*, pages 199–210, San Jose, California, May 1995.
[2] S. Acharya and S. Muthukrishnan. Scheduling on-demand broadcasts: New metrics and algorithms. In *Proceedings of the Fourth Annual ACM/IEEE International Conference on Mobile Computing and Networking (MobiCom'98)*, pages 43–54, Dallas, TX, October 1998.
[3] D. Ailkooy and M. Franklin. Scheduling for large-scale on-demand data broadcasting. In *Proceedings of IEEE INFOCOM '98*, San Francisco, CA, March 1998.
[4] Q. L. Hu, W.-C. Lee, and D. L. Lee. Indexing techniques for wireless data broadcast under data clustering and scheduling. In *Proceedings of the Eighth ACM International Conference on Information and Knowledge Management*, pages 351–358, Kansas City, Missouri, November 1999.
[5] Q. L. Hu, W.-C. Lee, and D. L. Lee. A hybrid index technique for power efficient data broadcast. *Distributed and Parallel Databases Journal*, to appear.
[6] T. Imielinski, S. Viswanathan, and B. R. Badrinath. Power efficient filtering of data on air. In *Proceedings of the International Conference on Extending Database Technology*, pages 245–258, 1994.
[7] T. Imielinski, S. Viswanathan, and B. R. Badrinath. Data on the air - organization and access. *IEEE Transactions of Data and Knowledge Engineering*, June 1997.
[8] W.-C. Lee and D. L. Lee. Using signature techniques for information filtering in wireless and mobile environments. *Distributed and Parallel Databases Journal: Special Issue on Database and Mobile Computing*, 4(3):205–227, July 1996.
[9] K. Stathatos, N. Roussopoulos, and J. S. Baras. Adaptive data broadcast in hybrid networks. In *Proceedings of the 23rd VLDB Conference*, pages 326–335, Athens, Greece, August 1997.
[10] C. Su and L. Tassiulas. Broadcast scheduling for information distribution. In *Proceedings of IEEE INFOCOM'97*, Kobe, Japan, April 1997. |
NAME OF PROJECT: EQUITY EMPOWERS: iCOUNT REDEFINES LIBRARY SERVICE
DATE OF IMPLEMENTATION/ADOPTION: AUGUST 2016
PROJECT STATUS: Ongoing
EXECUTIVE SUMMARY: Describe the project in 15 lines or less using Arial 12 point font. State clearly and concisely what difference the project has made.
LA County Library’s equity initiative, iCount, ensures the Library makes a conscious effort in designing services and programs that address the complex needs of the diverse community it serves, including customers of different ages, genders, sexual orientations, ethnicities, socioeconomic statuses, physical abilities, nationalities, legal statuses, and more. With 85 unique locations, the Library has always designed programs with local communities’ needs in mind, but addressing equity head-on required a new strategy. In 2016, iCount set out to develop a cohesive plan for uniting staff in the intentional development of more equitable service delivery models to reduce barriers and increase opportunities for under-resourced populations. Staff trainings were developed to help supervisors and managers educate staff on equity, explore and discuss related issues like implicit bias, and create and implement annual Library Equity Action Plans (LEAPs). Each LEAP gives Library staff the opportunity to create or redesign a service or program at the local level to better serve a chosen demographic, resulting in the development of over 100 programs, annually. Please consider this submission for the Changemaker & Community Inclusion Awards.
BENEFITS TO THE COUNTY
| (1) ACTUAL/ESTIMATED ANNUAL COST AVOIDANCE $ | (2) ACTUAL/ESTIMATED ANNUAL COST SAVINGS $ | (3) ACTUAL/ESTIMATED ANNUAL REVENUE $ | (1) + (2) + (3) = TOTAL ANNUAL ACTUAL/ESTIMATED BENEFIT $ |
|---------------------------------------------|------------------------------------------|--------------------------------------|--------------------------------------------------------|
| | | | |
ANNUAL = 12 MONTHS ONLY
SUBMITTING DEPARTMENT NAME AND COMPLETE ADDRESS
LA County Library, 7400 E Imperial Hwy, Downey, CA 90242
PROGRAM MANAGER’S NAME
Deborah Anderson
EMAIL email@example.com
PRODUCTIVITY MANAGER’S NAME AND SIGNATURE
Samangi Skinner
DATE 6/23/2021
TELEPHONE NUMBER (562) 940-4106
EMAIL firstname.lastname@example.org
DEPARTMENT HEAD’S NAME AND SIGNATURE
Skye Patrick
DATE 6/23/2021
TELEPHONE NUMBER (562) 940-8400
**ELECTRONIC, WET, OR SCANNED SIGNATURES ARE ACCEPTABLE**
**CHALLENGE:** Many Los Angeles County residents are desperate for life-altering opportunities. Measure of America’s study, *A Portrait of Los Angeles County*, used three key indicators in its Human Development Index (HDI)—Health, Education, and Economy—to assess the well-being of County residents. Scores varied so widely by community that the report divides LA into 5 different counties: Glittering LA, Elite Enclave LA, Main Street LA, Struggling LA, and Precarious LA. The last two categories are home to more than half of County residents—over 5 million people. These residents do not enjoy the same access to opportunity or levels of well-being as inhabitants of the first three areas; they lack high school diplomas, struggle for jobs, often live in poverty, and are reliant on public transit. These areas have the highest concentration of youth 18 and under, who often attend under-resourced schools and have a greater likelihood of developing learning disabilities.
They also happen to comprise of the communities where the majority of LA County Library locations are based, indicating an opportunity for greater depth of service, helping to bridge some of these equity gaps. In addition to profiling high-needs areas, the report also identifies countywide trends, including quickly-growing populations—those experiencing homelessness, older adults, and foreign-born residents. With 85 library locations covering a 300,000 square mile service area, it is no question that communities served by LA County Library have varying needs. While libraries have always designed offerings with local communities in mind, the Library wanted to find a centralized, systemwide strategy for addressing equity. As public servants, librarians are constantly adapting their traditional role to bridge service gaps and meet changing needs, but this was big—how could a library system take this on?
**SOLUTION:** The *iCount* equity initiative was conceived in 2016, after Skye Patrick became Library Director. Committing the department to breaking down barriers and increasing access for all, Patrick aimed to more equitably serve the County’s vast and increasingly diverse population—not just in terms of race and ethnicity, but also age, sexual orientation, gender identification, socioeconomic status, physical and mental ability, and more. Equity extends beyond equality (fairness and universal access) to deliberate, intentional efforts to create service delivery models that assure community members have the resources they need. To accomplish this, the Library needed a planning team. The voluntary *iCount Leadership Committee* is comprised of 20 staff across divisions, working alongside Patrick to develop *iCount*’s three key objectives:
1. **Train staff.** Inequity and implicit bias are tricky issues—often sensitive and emotionally-charged, they require ground rules and safe spaces. After establishing a foundational understanding of core topics, the Library would train its trainers, relying on supervisors to facilitate smaller conversations throughout the system.
2. **Break down a complex issue into bite-sized pieces.** The Library would leave it up to each library or division and their local expertise, to identify an inequity and create or revise a program to better serve that identified group.
3. **Reevaluate systemwide initiatives.** While many programs already sought to reach underserved demographics, leadership would refocus using an equity lens.
*iCount* was first introduced to staff at the Library’s Staff Development Day in August 2016. Simran Noor, Vice President of Policy & Programs for the Center of Social Inclusion, delivered the keynote. Emily Weak from Oakland Public Library addressed the perspective of “white librarianship” and the profession’s staggering statistics. Breakout sessions with multiethnic and multigenerational professionals encouraged staff to reflect on what more they could do to serve underrepresented communities.
In May 2017, a train-the-trainers event was held for the Library’s 200 supervisors, who would go on to become facilitators for discussions within their divisions. Speakers included Arnold Chandler of Forward Change Consulting who spoke on a Life Course Framework for Improving the Lives of Young Men of Color, and representatives of the Los Angeles LGBT Center’s RISE Project, who spoke on Gender Equity as well as Creating Safe, Supportive Libraries. An Equity Training Toolkit provided a facilitator’s guide with tips for effective small group communication on cultural competency, making the most of “hot moments” in conversations, and addressing sensitive topics (e.g., exploration vs. fault-finding.) The Toolkit also laid out the timeline for the year:
- **Jul 2017:** Training 1 – Laying the Ground Rules for Discussion
- **Aug 2017:** Training 2 – What is our Current Experience in the Library?
- **Sep 2017:** Training 3 – What Can We Do to Support Equity?
- **Oct – Nov 2017:** Training 4 – Library Equity Action Plan (LEAP). Staff picks one equity issue to address and formalizes a plan to address it, defining the issue, creating a timeline, and developing anticipated outcomes & measurements.
- **Dec 2017 – Jan 2018:** LEAP implementation period.
- **Feb – Jun 2018:** Submit LEAP progress report (Feb) and final report (Jun).
After a successful first year and the creation of 103 LEAPs, the Library kept a similar framework/timeline for 2018 and beyond. Each year would begin with reviewing the previous year’s LEAP—*How well did it address an equity issue? What were some of the successes and challenges? What impact did the LEAP have on the community? Did we meet our anticipated outcomes? What will the long-term impact be?* Implicit bias became a key theme. Staff were invited to take a private, anonymous implicit association test from Harvard University’s Project Implicit. LEAP highlights were featured in case study videos, produced by Library Marketing, to inspire future plans.
**BENEFITS:** Equity is a difficult concept to measure, especially in terms of gradual improvements. But in terms of meeting program objectives, *iCount* has been a resounding success. Over 100 LEAPs have been conducted annually, representing 85 libraries plus all support staff teams (e.g., facilities, marketing, human resources) and localized equity improvements are thriving. Because each plan is unique, quantitative
data like program attendance or customers served cannot be totaled comprehensively. However, annual reports on LEAP successes allow the Library to develop case studies for replication, which are shared with staff. Highlights of populations served include:
- **Older adults (25% of LEAPs):** Libraries added workshops on computers and eBook checkouts; expanded large-print collections; hosted social mixers, movie screenings, and board game play; and expanded outreach to senior centers.
- **Customers with intellectual or physical disabilities (15% of LEAPs):** Art activities, Storytimes, and programs for all ages were developed, including a *Differently Abled Adults Reading Club* and an *All Abilities Children’s Concert* series. IT made plans to improve accessibility hardware and software in libraries.
- **Non-native English speakers (15% of LEAPs):** Libraries added computer literacy courses, book clubs, and ESL conversation clubs in Spanish and Chinese. West Hollywood Library expanded its Russian language collection. The Marketing team launched a Spanish-language Awareness Campaign, after learning many immigrants are unaware library services are free; it featured key services targeted to Spanish-speaking neighborhoods using billboards, bus shelters, radio, and TV.
- **At-risk youth (14% of LEAPs):** Systemwide programs like the *Turns the Tables DJ Lab* and *Microsoft DigiCamps* have helped tweens and teens in select communities learn about STEAM and career opportunities. The Library’s *My Brother’s Keeper Peer Advocate* program hires young people of color with lived experience to serve as Library advisors and liaisons between staff and youth in select communities, while providing opportunities to gain valuable life and work skills, helping the Peer Advocates to reach their career goals.
- **People experiencing homelessness (8% of LEAPs):** Libraries created resource lists for showers, meals, and local shelters, and hosted wellness events.
Making a dent in an issue as sweeping as inequity can seem like a daunting task, but by empowering staff as change agents knowledgeable of local community needs, the Library can take over 100 steps toward addressing inequities each year. *iCount* was developed from the ground up by Library staff, who recognize that perception cannot be changed overnight; it must be achieved through continual education. In introducing *iCount* trainings, the Library initiated an essential conversation—one that encourages both frontline and support staff to think of programs and services as adaptable and inclusive, in order to be truly impactful. By filling service gaps and improving resource offerings based on identified needs, *iCount* helps all County residents feel welcome in the library.
Within the library industry, *iCount* has been recognized as a replicable and sustainable model, bringing greater visibility to the County’s public services. The Library was invited to present on *iCount* during a *Library Journal* equity webinar, the New England Library Association Diversity Summit, the Urban Library Council Equity Convening, the California Library Association Conference, the American Library Association Conference, and the Joint Conference of Librarians of Color, to name a few.
NAME OF PROJECT: EQUITY EMPOWERS: iCOUNT REDEFINES LIBRARY SERVICE
Linkage to the County Strategic Plan – 1 page only. Which County Strategic Plan goal(s) does this project address? Explain how. Use Arial 12-point font.
GOAL I. MAKE INVESTMENTS THAT TRANSFORM LIVES
Strategy I.1 – Increase Our Focus on Prevention Initiatives and Strategy I.2 – Enhance Our Delivery of Comprehensive Interventions
In intentionally designing programs and services to meet the needs of under-resourced County communities that experience inequities, iCount addresses many Strategic Plan Strategy I goals, including: Promote Supportive Parenting (I.1.1), Support Educational Outcomes for Systems Involved Youth (I.1.4), Support Library Efforts to Engage Children and Families (I.1.7), Support Job Readiness and Increase Employment Opportunities for Youth Served by the County (I.2.4), Implement My Brother’s Keeper (MBK) (I.2.5), Women and Girls’ Initiative (I.2.6), Plan for Age-Friendly Communities for Older Adults (I.2.7), Support the Long Term Success of Transitional Aged Youth (I.2.9). Hundreds of localized programs at libraries and in Library support divisions focus their design on these audiences, based on community needs.
GOAL II. FOSTER VIBRANT AND RESILIENT COMMUNITIES
Strategy II.2 – Support the Wellness of our Communities
Systemwide Library programs developed as part of iCount focus on serving at-risk and systems involved youth, parents and caregivers lacking access, unemployed and job-seeking adults, potentially isolated older adults, and more. For example, these programs have helped address Strategy II.1.3 (Coordinate Workforce Development) by helping connect priority populations to workforce development opportunities as well as careers in the creative economy. The Spanish-language marketing campaign that resulted from a LEAP helped address Strategy II.2.2 (Expand Access to Recreational and Cultural Opportunities) by informing high-needs communities of free library programs and services available to them, thereby expanding access to populations that underutilized these services simply out of a lack of awareness.
GOAL III. REALIZE TOMORROW’S GOVERNMENT TODAY
Strategy III.1 - Continually Pursue Development of Our Workforce
As an initiative developed by staff and highly reliant on a train-the-trainer model for regular staff discussions, iCount supports Strategy III’s goal III.1.1 (Develop Staff Through High Quality Multi-Disciplinary Approaches to Training.) iCount training has included Staff Development Days, guest speakers from a variety of institutions, video content, and internal presentations. The rotating iCount Leadership Committee and regular iCount trainings give management staff the opportunity to hone leadership skills while facilitating sometimes difficult conversations, and give all staff the opportunity to explore topics they may have been unaware of, which will inevitably have an expanding impact on workplaces and public service.
Title of Project (Limited to 50 characters, including spaces, using Arial 12-point font):
NAME OF PROJECT: EQUITY EMPOWERS: iCOUNT REDEFINES LIBRARY SERVICE
COST AVOIDANCE, COST SAVINGS, AND REVENUE GENERATED (ESTIMATED BENEFITS TO THE COUNTY): If you are claiming cost benefits, include a calculation on this page. Please indicate whether these benefits apply in total or on a per unit basis, e.g., per capita, per transaction, per case, etc. You must include an explanation of the County cost savings, cost avoidance or new revenue that matches the numbers in the box. Remember to keep your supporting documentation. Use Arial 12-point font.
Cost Avoidance: Costs that are eliminated or not incurred as a result of program outcomes. Please indicate whether these are costs to the County or to other entities.
Cost Savings: A reduction or lessening of expenditures as a result of program outcomes. Please indicate whether these were expenditures by the County or by other entities.
Revenue: Increases in existing revenue streams or new revenue sources to the County as a result of program outcomes.
| (1) ACTUAL/ESTIMATED ANNUAL COST AVOIDANCE | (2) ACTUAL/ESTIMATED ANNUAL COST SAVINGS | (3) ACTUAL/ESTIMATED ANNUAL REVENUE | (1) + (2) + (3) TOTAL ANNUAL ACTUAL/ESTIMATED BENEFIT | SERVICE ENHANCEMENT PROJECT |
|-------------------------------------------|------------------------------------------|-----------------------------------|-------------------------------------------------|-----------------------------|
| $ | $ | $ | $ | ☑ |
ANNUAL = 12 MONTHS ONLY |
Ziad Zaky Adwan
Abu Dhabi, UAE, Tel: +971509995562, Email: email@example.com
Quick Info.
Nationality: United States of America.
Residence: Abu Dhabi/UAE, Texas/USA & Saigon/Vietnam.
Education:
B.Sc, Mathematics, American University of Beirut, Lebanon, 1999.
M. Sc, Mathematics, American University of Beirut, Lebanon, 2001.
M. Sc, Mathematics, Temple University, Philadelphia, 2003.
PhD, Mathematics, Temple University, Philadelphia, 2006.
Post Doctoral Research Assistant, Mathematics, Temple University, Philadelphia, 2006-2007.
Employment:
Title: Associate Professor of Mathematics (Zayed University) (January 2019 to Present)
Title: Associate Professor of Mathematics (American University in Dubai) (September 2016 to July 2018).
Title: Assistant Professor of Mathematics (American University in Dubai) (September 2012 to August 2016)
Title: Assistant Professor of Mathematics (University of Texas, USA) (Tenure-Track Position) (September 2007 to August 2012)
Title: Post-Doc. Research Assistant (Temple University, Philadelphia, PA, USA) (September 2006 to August 2007)
Title: Graduate Assistant (Fellowship) (Temple University, Philadelphia, PA, USA) (September 2001 to August 2006)
Dear Sir/Madam,
My Name is Ziad Adwan, and I am interested in applying for a Mathematics position in your university. I have a PhD in Mathematics from Temple University in 2006, two MSc. Degrees in mathematics and a BS degree from the American University of Beirut.
Currently, I am a full-time Associate Professor of Mathematics at Zayed University. Since finishing my PhD, I have been involved in a lot of projects pertaining to quality teaching in mathematics and popularizing mathematical ideas.
I have taught almost all undergraduate mathematics courses and masters level courses (see my CV) and my research focuses on studying analytical properties of solutions of Complex Vector Fields in n-dimensional Complex Space. Here is my personal link on Research Gate: (https://www.researchgate.net/profile/Ziad_Adwan).
I have created a YouTube Channel (https://www.youtube.com/@ziadadwan) to record online lectures with the highest quality pedagogical experience. This was very successful with my students as evidenced by their evaluation of my courses. Here is one sample lecture on Statistics with Population Proportions that had a lot of views on my YouTube Channel; it also shows my teaching techniques:
- **Statistics with Population Proportions** [https://youtu.be/fmt4jGm1iV8](https://youtu.be/fmt4jGm1iV8)
Artificial Intelligence will eventually replace human beings in most sectors of education and there will be less demand for teachers and more demand for content producers of highest quality educational lectures and tools. We explored the possibility of teaching full classes using VR technology and we made several videos on this topic. Here is a Sample video that was made on how we see the use of VR in education in the future:
- **An introduction to Teaching in Virtual Reality** [https://youtu.be/QZfgod8cc_0](https://youtu.be/QZfgod8cc_0)
I am a team player and easy-going person who can adjust quickly to new demands and adventures and will be successful in fulfilling all my duties and more once I join your university. Thank you for considering my application.
Ziad Adwan
firstname.lastname@example.org
Teaching Statement
My teaching has been constantly changing in the past two decades. I have taught all levels of mathematics from Developmental mathematics [College Algebra, Precalculus, etc.] up to master’s level Pure/Applied mathematics courses (See my CV).
I believe that to teach any topic, one must find an interesting and relative example for the student audience that can facilitate the understanding of the topic being presented. Usually, this example should contain all the main ideas of the section so that once we are done with the example, going through the topic can become much more comprehensible to the student. Finding such an example for each of the topics is not a fixed exercise; it is always changing and improving throughout the years of experience. To give you a concrete example of what this means, here is a link to one of my lectures on Statistics that describes the meaning of the P-Value (Short for Probability-Value) using a Coin that was flipped 8 times. Notice that I try to relate things to the UAE: https://youtu.be/G2PpUy2JK1I?t=585.
I also believe that making instructional videos to teach the main topics of the course would be an excellent practice for students’ out of class experience. Below are 2 of my YouTube video lectures showing my teaching approach. They were recorded at a time when we had to teach remotely because of Covid-19, and I made sure that they will be enough for a student to understand the entire textbook without a teacher who is available face-to-face. Please take some time to watch their content as this is the best way that I can think of to show my teaching philosophy and teaching style. The first video is on the topic of Statistical Hypothesis Testing and the second is on the topic of Population Proportions:
1. Fundamentals of Hypothesis Testing: https://youtu.be/G2PpUy2JK1I
2. Statistics with Population Proportions: https://youtu.be/fmt4jGm1iV8
To prevent my own teaching from becoming stagnant, I encourage students to provide feedback, either one-on-one, or in writing, including anonymous e-mails. I also encourage them to share their opinions with classmates alone or in my presence. Occasionally, I have been approached with methods not discussed in the text. For example, one student recently inquired about an integration method called the “ladder method”. Although the method only works in special cases, it illustrates how student participation leads to greater understanding through research. Another strategy that deserves mention is that of presentation. It is easy to teach calculus as a series of rules and procedures. If this is done, however, mathematical insight and the practical value of the concepts discussed disappear under the weight of procedural lists. To avoid this, I give equal emphasis to both concepts and procedures by giving additional problems not covered in the text. Such problems are designed to direct attention to both procedures and concepts. I have also incorporated an essay assignment which has been useful in eliciting critical thinking. One student, who was struggling with integration techniques, told me that he liked the idea of essay writing because it forced him read his text thoroughly in addition to investigating online calculus sources. As a result of the extra effort, he found that the techniques that had originally been overwhelming lost their power to defeat him and he felt more in control of his learning.
Courses Taught:
**Undergraduate Courses:** (These were taught both at the University of Texas & the American University in Dubai)
1. *College Algebra* (*Developmental Mathematics*),
2. *Pre-Calculus, Differential and Integral Calculus*,
3. *Calculus for Business*,
4. *Mathematics for the Arts*,
5. *Calculus I, II & III*,
6. *Linear Algebra*,
7. *Probability Theory*,
8. *Discrete Mathematics*,
9. *Ordinary Differential Equations*,
10. *Foundations of Mathematics*,
11. *Real Analysis I & II*,
12. *Complex Analysis*.
**Graduate Courses:** (These were taught at the University of Texas)
1. *Measure & Integration*,
2. *Abstract Algebra*,
3. *Algebraic Topology*,
4. *Introduction to PDEs*.
Curriculum From the Fall of 2007 until the Spring of 2009, I have been the Chair of the Mathematics Undergraduate Curriculum Committee at the University of Texas. In the American University in Dubai, I served as the Coordinator of the Mathematics Division for three years and I served on the College of Arts & Sciences Curriculum Committee, on the Promotion Committee and on the Leadership Committee.
Research My research focuses on Theory and Applications of Partial Differential Equations and Complex Analysis. More precisely on Boundary Values of Solutions of Complex Vector Fields. So far, I have 8 published research papers in refereed mathematical journals, some of which are high ranking (see Publications below) and 1 paper submitted for publication.
Mathematical Reviewer
- Referee for the Journal "Complex Variables and Elliptic Equations".
- Referee for the journal "Geometry and Physics" which is published by ELSEVIER.
- Referee for the journal “Mathematical Reviews” which is published by the University of Michigan, Ann Arbor.
Publications
(1) Z. Adwan, G. Hoepfner and Andrew Raich: Global $L^q$ -Gevrey Functions and Their Applications, *Journal of Geometric Analysis*, DOI 10.1007/s12220-016-9743-6, **2017**.
(2) Z. Adwan and G. Hoepfner: Denjoy-Carleman Classes: Boundary Values, Approximate Solutions and Applications, *Journal of Geometric Analysis*, DOI 10.1007/s12220-014-9491-4, **April 2014**.
(3) Z. Adwan and S. Berhanu: On microlocal analyticity and smoothness of solutions of first-order nonlinear pdes, *Mathematische Annalen*, Vol. 352 (**2012**), 239-258.
(4) Z. Adwan and G. Hoepfner: Approximate Solutions and Micro-Regularity in the Denjoy-Carleman Classes, *J. Differential Equations* 249 (**2010**) 2269–2286.
(5) Z. Adwan and G. Hoepfner: *A Generalization of Borel’s Theorem and Microlocal Gevrey Regularity in Involutive Structures*, *J. Differential Equations* 245 (**2008**), 2846-2870.
(6) Z. Adwan and G. Hoepfner: *On the $C^\infty$ Wave-front Set of Traces of CR Functions on Totally Real Submanifolds*, *Proceedings of the American Mathematical Society* 136 (**2008**), no. 3, 999-1008.
(7) Z. Adwan and S. Berhanu: *Edge-of-the-Wedge Theory in Involutuve Structures*, Asian Journal of Mathematics, 11 (2007), no. 1, 1-17.
(8) Z. Adwan: *The Exceptional Lie Group $E_8$*, Advances in Applied Clifford Algebras, Vol. 14, (2), 253-284, 2004.
**Professional Development**
June 27-July 1, 2016: **Invited Speaker**: CR Geometry and PDE’s VII. Dedicated to the memory of Giuseppe Zampieri. Levico T. (Trento) Italy.
April 14, 2015: **Invited Speaker**: Weekly Mathematics Seminar at AUS, American University of Sharjah, Sharjah, UAE.
April 02-05, 2015: **Participant**: Second International Conference on Mathematics and Statistics, American University of Sharjah, Sharjah, UAE.
March 14, 2015: **Speaker**, UAE Math Day, Paris Sorbonne University, Abu Dhabi.
January 04-08, 2015: **Invited Speaker**: Analysis and Geometry in Several Complex Variables, Texas A&M University.
January 09, 2014: **Invited Participant**: Student Evaluation and Grading: Challenges and Best Practices, American University in Dubai, Dubai, UAE.
August 29, 2013: **Invited Participant**: New Teaching and Learning Technologies, American University in Dubai, Dubai, UAE.
August 01-05, 2011: **Invited Participant**: Geometric Analysis of PDE and Several Complex Variables, Serra Negra, SP, Brazil.
January 13-14, 2011: **Invited Participant**: NSF Analysis Workshop "Active Learning Materials for Critical Thinking in a First Course in Real Analysis", University of Texas at Arlington, Arlington, Texas.
August 12-13, 2010: **Invited Speaker**: NSF Analysis Workshop "Active Learning Materials for Critical Thinking in a First Course in Real Analysis", University of Texas at Arlington, Arlington, Texas.
May 10-14, 2010: **Invited Speaker**: Geometric Analysis of Several Complex Variables and related topics, Marrakech, Morocco.
August 3-7, 2009: **Invited Speaker**: Workshop on recent developments
in several complex variables and partial differential equations, Serra Negra, SP, Brazil.
July 6-10, 2009: **Invited Speaker**: *International Conference on Several Complex Variables, Complex Geometry and Partial Differential Equations*, Wuhan University, Wuhan, China. Title: “Microlocal Analysis in $C^M$ Involutive Structures”.
June 06, 2008: **Invited Speaker**: *First Joint Meeting of the American Mathematical Society and the Brazilian Mathematical Society*, June 4th – 7th, Rio De Janeiro, Brazil. Title: “Microlocal Gevrey Regularity in Involutive Structures”.
April 30, 2007: **Invited Speaker**: *Analysis Seminar at Temple University*. Title: “On Gevrey Wave-front Sets in Gevrey Involutive Structures”.
June 11-16, 2006: **Invited Speaker**: *International Conference in PDE, Complex Analysis, and Differential Geometry*, University of Notre Dame, IN, USA. **Title**: *Edge of the Wedge Theory in Involutive Structures*.
January 23-27, 2006: **Invited Participant**: *II Escola Brasiliera de Equações Diferenciais*, IMPA, Rio de Janeiro, RJ, Brazil
August 01-05, 2005: **Invited Participant**: *III Workshop on Geometric Analysis of PDE and Several Complex Variables*, Serra Negra, SP, Brazil.
April 08-10, 2005: **Invited Participant**: *A Conference Celebrating the Mathematics and 75th Birthday of Francois Treves*, Rutgers University, NJ, USA.
1. Shiferaw Berhanu, Professor of Mathematics, University of Maryland, College Park, MD.
Phone: (301) 405 5047.
E-mail: email@example.com
2. Hamid Meziani, Chair and Professor of Mathematics, Florida International University, Miami, Fl.
Phone: (305)-348-2957.
E-mail: firstname.lastname@example.org.
3. Haitham Solh, Professor of Mathematics, American University in Dubai, Dubai, UAE.
Phone: +97143183498.
E-mail: email@example.com
4. Taeil Yi, Professor of Mathematics, University of Texas Rio Grande Valley, Brownsville, TX.
Phone: (956) 665-3535.
E-mail: firstname.lastname@example.org. |
Topographic Processing of Relational Data
Barbara Hammer\textsuperscript{1}, Alexander Hasenfuss\textsuperscript{1}, Fabrice Rossi\textsuperscript{2}, Marc Strickert\textsuperscript{3}
1 - Clausthal University of Technology, Department of Informatics
\{hammer,hasenfuss\}@in.tu-clausthal.de
2 - Projet AxIS, INRIA, Le Chesnay Cedex, France
email@example.com
3 - IPK Gatersleben, Pattern Recognition Group
firstname.lastname@example.org
Keywords: relational data, batch clustering, visualization, SOM, NG
Abstract — Recently, batch optimization schemes of the self-organizing map (SOM) and neural gas (NG) have been modified to so-called median variants which allow a transfer of these methods to arbitrary distance measures beyond standard euclidean metrics. This principle is particularly suitable for complex applications where data are compared by means of problem-specific, possibly discrete metrics such as protein sequences which are compared by FASTA or BLAST. However, median variants do not allow a continuous update of prototype locations and their capacity is thus restricted in particular for small data sets. In this contribution, we consider the relational dual of batch optimization which can be formulated in terms of pairwise distances only such that an application to distance matrices without known euclidean embedding becomes possible. For SOM, a direct visualization of data is given by means of the underlying (euclidean or hyperbolic) lattice structure. For NG, pairwise distances of prototypes can be computed based on a given data matrix only, such that subsequent mapping by means of multidimensional scaling can be applied. These algorithms are evaluated in several experiments.
1 Introduction
The self organizing map as proposed by Kohonen [17] constitutes one of the most popular methods for visual data inspection and clustering. It is highly flexible and can be easily applied to data of interest. Alternative methods exist such as neural gas as proposed by Martinetz [20] which focuses on the quality of clustering and which can serve as inspection tool in combination with subsequent visualization e.g. using multidimensional scaling [16, 27].
The original online variants of SOM and NG, however, have only been proposed for euclidean data. Extensions to more general situations are inevitable if the euclidean distance measure is not appropriate or not applicable such as the case of discrete data. A variety of extensions of SOM has been proposed to deal with this situation including statistical variants [28], variants which built on generative models [30], and encoder-decoder frameworks [4, 5, 26]. These proposals have in common that the original intuitive version of SOM is not directly obtained in the case of standard euclidean data but modifications are necessary. Recently, batch optimization schemes have been extended to general proximity data by means of the generalized median [3, 18]. This restricts prototype locations to data points, thereby maintaining the standard batch optimization principle as far as possible. A theoretical background for this method as well as a convergence proof is provided in [3]: median variants are shown to be equivalent to optimization of the cost function of NG resp. SOM with respect to assignments and prototype locations. Thus, median clustering is very similar to standard batch variants since it relies on the same cost function. However, it restricts prototype locations to the data space such that a severe loss of accuracy can be expected in particular for an only sparsely covered input space since no continuous update takes place.
Here we propose relational variants of SOM and NG which allow a continuous update of prototypes also in the case of given distance matrix. These versions optimize the same cost function as the standard batch variants such that convergence is guaranteed. Relational clustering is quite well known for simple k-means clustering and fuzzy variants thereof [9, 10]. The article [12], for example, provides a reliable (but costly) solution to optimize the cost function of Relational k-means by means of statistical physics. For SOM and NG, online kernelized variants have been proposed e.g. in [23, 32] which are related to relational clustering since every kernel induces a general metric (but not vice versa). However, these proposals do not provide a fast batch optimization scheme and they require similarities instead of dissimilarities. We embed relational clustering into a general framework by means of a cost function such that convergence is guaranteed and extensions such as supervision [8] can be easily integrated.
The focus of this article lies on the usefulness of these methods for inspection and low-dimensional visualization of relational data characterized by pairwise distances. For this purpose, standard SOM with either euclidean or hyperbolic lattice as proposed in [24] can be used. Alternatively, we derive formulas which describe pairwise distances of prototypes in the relational models, such that also NG structures without fixed prior lattice can be embedded in low dimensional euclidean or hyperbolic space [16, 27, 31]. Thereby, depending on the complexity of the data, hyperbolic space might be better suited to cluster and visualize hierarchical structures and close connections of the data [1].
Now, we first introduce standard batch optimization of NG and SOM and extend this models to relational data. We discuss the possibility to incorporate supervision and visualization, and demonstrate the applicability in several experiments from bioinformatics.
2 Relational topographic maps
Classical neural topographic maps consider vectorial data $\vec{x} \in \mathbb{R}^n$ which are distributed according to an underlying distribution $P$ in the euclidean plane. The goal of clustering is to distribute prototypes $\vec{w}^i \in \mathbb{R}^n$, $i = 1, \ldots, N$ faithfully among the data. A new data point $\vec{x}$ is assigned to the winner $\vec{w}^{I^*(\vec{x})}$ which is the prototype with smallest distance $\|\vec{x} - \vec{w}^i\|^2$. This clusters the data space into the receptive fields of the prototypes.
Different popular variants of neural clustering have been proposed to learn prototype locations from given training data [17]. Assume the number of prototypes is fixed to $N$. Neural gas (NG) [20] has been introduced based on the cost function
$$E_{NG}(\vec{w}) = \frac{1}{2C(\lambda)} \sum_{i=1}^{N} \int h_\lambda(k_i(\vec{x})) \cdot \|\vec{x} - \vec{w}^i\|^2 P(d\vec{x})$$
where
$$k_i(\vec{x}) = |\{\vec{w}^j : \|\vec{x} - \vec{w}^j\|^2 < \|\vec{x} - \vec{w}^i\|^2\}|$$
is the rank of the prototypes sorted according to the distances, $h_\lambda(t) = \exp(-t/\lambda)$ scales the neighborhood cooperation with neighborhood range $\lambda > 0$, and $C(\lambda)$ is the constant $\sum_{i=1}^{N} h_\lambda(k_i(\vec{x}))$ which we neglect in the following for simplicity. Classical NG is optimized online by means of a stochastic gradient descent. For a fixed training set, an alternative fast batch optimization scheme is offered by the following algorithm, which in turn computes ranks, which are treated as hidden variables of the cost function, and optimum prototype locations [3]:
init $\vec{w}^i$
repeat
compute winner assignments $I^*(\vec{x}^j)$ minimizing
$$\sum_k h_\lambda(n(I^*(\vec{x}^j), k)) \|\vec{x} - \vec{w}^k\|^2$$
compute new prototype locations $\vec{w}^i =$
$$\sum_j h_\lambda(n(I^*(\vec{x}^j), i) \cdot \vec{x}^j / \sum_j h_\lambda(n(I^*(\vec{x}^j), i)$$
It has been shown in e.g. [3] that these batch optimization schemes converge in a finite number of steps towards a (local) optimum of the cost function, provided the data points are not located at borders of receptive fields of the final prototype locations. In the latter case, convergence can still be guaranteed but the final solution can lie at the border of basins of attraction.
2.1 Relational data
Relational data $x^i$ are not embedded in a vector space, rather, pairwise similarities or dissimilarities are available. Assume training data $x^1, \ldots, x^m$ are given by means of pairwise distances $d_{ij} = d(x^i, x^j)^2$. We assume that this stems from an unknown distance measure such that we can find (possibly high dimensional) euclidean points $\vec{x}^i$ such that $d_{ij} = \|\vec{x}^i - \vec{x}^j\|^2$. Note that this notation includes a possibly nonlinear mapping (feature map) $x^i \mapsto \vec{x}^i$ corresponding to the embedding in a euclidean space. However, this embedding is not known, such that we cannot directly optimize the above cost functions in the embedding space.
Median clustering restricts prototype locations to given data points and determines the prototype locations in each iterative step in such a way that the corresponding part of the cost function (assumed fixed assignments) becomes minimum. These values are determined by extensive search, turning the linear complexity to a quadratic one w.r.t. the number of training data. This procedure has the severe drawback that only discrete adaptation steps can be performed and the result is usually worse compared to standard SOM or NG in the euclidean setting.
Relational learning overcomes this problem. The key observation consists in the fact that optimum prototype locations $\vec{w}^j$ can be expressed as linear combination of data points. Therefore, the unknown distances $\|\vec{x}^j - \vec{w}^i\|^2$ can be expressed in terms of known values $d_{ij}$.
More precisely, assume there exist points $\vec{x}^j$ such that $d_{ij} = \|\vec{x} - \vec{x}^j\|^2$. Assume the prototypes can be expressed in terms of data points $\vec{w}^i = \sum_j \alpha_{ij} \vec{x}^j$ where $\sum_j \alpha_{ij} = 1$. Then
$$\|\vec{w}^i - \vec{w}^j\|^2 = (D \cdot \alpha_i)_j - 1/2 \cdot \alpha_i^t \cdot D \cdot \alpha_i$$
where $D = (d_{ij})_{ij}$ constitutes the distance matrix and $\alpha_i = (\alpha_{ij})_j$ the coefficients.
Because of this fact, we can substitute all terms $\|\vec{x} - \vec{w}^i\|^2$ in batch optimization schemes. For optimum solutions we find the equality $\vec{w}^i = \sum_j \alpha_{ij} \vec{x}^j$ for batch NG and batch SOM as introduced above, whereby
1. $\alpha_{ij} = h_\lambda(k_i(\vec{x}^j))/\sum_j h_\lambda(k_i(\vec{x}^j))$ for NG, and
2. $\alpha_{ij} = h_\lambda(n(I^*(\vec{x}^j), i))/\sum_j h_\lambda(n(I^*(\vec{x}^j), i)$ for SOM.
This allows to reformulate the batch optimization schemes in terms of relational data. We obtain the algorithm
init $\alpha_{ij}$ with $\sum_j \alpha_{ij} = 1$
repeat
compute $\|\vec{x}^j - \vec{w}^i\|^2$ as $(D \cdot \alpha_i)_j - 1/2 \cdot \alpha_i^t \cdot D \cdot \alpha_i$
compute optimum assignments based on these values
$\hat{\alpha}_{ij} = h_\lambda(k_i(\vec{x}^j))$ (for NG)
$\hat{\alpha}_{ij} = h_\lambda(n(I^*(\vec{x}^j), i))/\sum_j h_\lambda(n(I^*(\vec{x}^j), i)$ (for SOM)
compute $\alpha_{ij} = \hat{\alpha}_{ij}/\sum_j \hat{\alpha}_{ij}$
Hence, prototype locations are computed only indirectly by means of the coefficients $\alpha_{ij}$. Initialization can be done e.g. setting initial prototype locations to random data points, which is realized by a random selection of $N$ rows from the given distance matrix.
Given a new data point $x$ which can isometrically be embedded in euclidean space as $\vec{x}$, and pairwise distances $d_{ij} = d(x, x^j)^2$ corresponding to the distance from $x^j$, the winner can be determined by using the equality
$$\|\vec{x} - \vec{w}^i\|^2 = (D(x)^t \cdot \alpha_i) - 1/2 \cdot \alpha_i^t \cdot D \cdot \alpha_i$$
where $D(x)$ denotes the vector of distances $D(x) = (d_{ij})_j = (d(x, x^j)^2)_j$.
The quantization error can be expressed in terms of the given values $d_{ij}$ by substituting $\|\vec{x} - \vec{w}^i\|^2$ by $(D \cdot \alpha_i)_j - 1/2 \cdot \alpha_i^t \cdot D \cdot \alpha_i$. Interestingly, using the formula for optimum assignments of batch optimization, one can also derive relational dual cost functions for the algorithms. The relational dual of NG is
$$\sum_i \frac{\sum_{kl} h_\lambda(k_l(\vec{x}^l))h_\lambda(k_k(\vec{x}^k))d_{lk}}{4 \sum_j h_\lambda(k_j(\vec{x}^j))}.$$
For SOM, we obtain
$$\sum_i \frac{\sum_{kl} h_\lambda(n(I^*(\vec{x}^l), i))h_\lambda(n(I^*(\vec{x}^k), i))d_{lk}}{4 \sum_j h_\lambda(n(I^*(\vec{x}^j), i)}.$$
Note that this relational learning gives exactly the same results as standard batch optimization provided the given relations stem from a euclidean metric. Hence convergence is guaranteed in this case since it holds for the standard batch versions. If the given distance matrix does not come from a euclidean metric, this equality does no longer hold and the terms $(D \cdot \alpha_i)_j - 1/2 \cdot \alpha_i^t \cdot D \cdot \alpha_i$ can become negative. In this case, one can correct the distance matrix by the $\gamma$-spread transform $D_\gamma = D + \gamma(1 - I)$ for sufficiently large $\gamma$ where $I$ equals 1 for each entry and $I$ is the identity. Alternatively, one can directly apply the optimization scheme since it converges also for general symmetric and nonsingular matrix $D$ towards a value of the dual cost function as shown in [7] for relational NG.
### 2.2 Supervision
The possibility to include further information, if available, is very important to get meaningful results for unsupervised learning. This can help to prevent the ‘garbage in – garbage out’ problem of unsupervised learning, as discussed e.g. in [14, 15]. We assume that additional label information is available which should be accounted for. Thereby, labels are embedded in $\mathbb{R}^L$. We assume that the label attached to $x^j$ is denoted by $\vec{y}^j$. We equip a prototype $\vec{w}^i$ with a label $\vec{Y}^i \in \mathbb{R}^L$ which is adapted during learning. For the euclidean case, the basic idea consists in a substitution of the standard euclidean distance $\|\vec{x}^j - \vec{w}^i\|^2$ by a mixture
$$(1 - \beta) \cdot \|\vec{x}^j - \vec{w}^i\|^2 + \beta \cdot \|\vec{y}^j - \vec{Y}^i\|^2$$
which takes the similarity of labels into account. $\beta \in [0, 1]$ controls the influence of the labels. This procedure has been proposed in [8] for euclidean clustering. One can use the same principles for relational clustering.
For discrete euclidean settings $\vec{x}^1, \ldots, \vec{x}^{m}$ cost functions and related batch optimization is as follows: $E_{NG}(\vec{w}, \vec{Y}) =$
$$\sum_{ij} h_\lambda(k_i(\vec{x}^j)) \left((1 - \beta) \cdot \|\vec{x}^j - \vec{w}^i\|^2 + \beta \cdot \|\vec{y}^j - \vec{Y}^i\|^2\right)$$
where $k_i(\vec{x}^j)$ denotes the rank of neuron $i$ measured according to the distances $(1 - \beta) \cdot \|\vec{x}^j - \vec{w}^i\|^2 + \beta \cdot \|\vec{y}^j - \vec{Y}^i\|^2$. This change in the computation of the rank is accompanied by the adaptation $\vec{Y}^i = \sum_j h_\lambda(\vec{x}^j)\vec{y}^j/\sum_j h_\lambda(\vec{x}^j)$ for the prototype labels for batch optimization. In the same way, the cost function of SOM becomes
$$E_{SOM}(\vec{w}, \vec{Y}) = \delta_{i, I^*(\vec{x}^j)} \sum_{k} h_\lambda(n(i, k))$$
$$\left((1 - \beta) \cdot \|\vec{x}^j - \vec{w}^k\|^2 + \beta \cdot \|\vec{y}^j - \vec{Y}^k\|^2\right)$$
where $I^*(\vec{x}^j)$ denotes the generalization of the winner notation proposed by Heskes to the supervised setting, i.e. it minimizes $\sum_k h_\lambda(n(I^*(\vec{x}^j), k)) \left((1 - \beta) \cdot \|\vec{x}^j - \vec{w}^k\|^2 + \beta \cdot \|\vec{y}^j - \vec{Y}^k\|^2\right)$. Batch optimization uses this winner and extends the updates by $\vec{Y}^i = \sum_j h_\lambda(n(i, I^*(\vec{x}^j))\vec{y}^j/\sum_j h_\lambda(n(i, I^*(\vec{x}^j))$. It has been shown in [8] that these procedures converge in a finite number of steps.
Relational learning becomes possible by substituting the distances $\|\vec{x}^j - \vec{w}^i\|^2$ using the identity $\vec{w}^i = \sum_j \alpha_{ij} \vec{x}^j$ for optimum assignments which still holds for these extensions. The same computation as beforehand yields to the algorithm for dissimilarity data:
init $\alpha_{ij}$ with $\sum_j \alpha_{ij} = 1$
repeat
compute the distances as $(1 - \beta) \cdot ((D \cdot \alpha_i)_j - 1/2 \cdot \alpha_i^T D \alpha_i) + \beta \cdot \| \vec{Y}^i - \vec{y}^j \|^2$
compute assignments $\tilde{\alpha}_{ij}$ based on these values
compute $\alpha_{ij} = \tilde{\alpha}_{ij} / \sum_j \tilde{\alpha}_{ij}$
compute prototype labels $\vec{Y}^i = \sum_j \alpha_{ij} \vec{y}^j$
Since this version is identical to the euclidean version for an euclidean distance matrix this procedure converges in a finite number of steps. Note that, for vanishing neighborhood size, the final prototype labels coincide with the averaged label taken over the receptive field of a prototype. For rapid learning, one can improve the classification result by setting the prototype labels to the averaged label of the receptive fields after training. Note that, still, the prototype locations are affected by the label information, unlike a pure unsupervised learning with posterior labeling. For the supervised setting, the prototypes are forced to follow the borders given by the class information.
### 2.3 Visualization
As discussed before, data can be clustered using trained prototypes. But it is not obvious whether we are able to create a visualization of such extracted information. For low-dimensional euclidean SOM, a direct visualization is given by an embedding of the data points to the respective positions of their winner in the lattice space. Also the hyperbolic SOM allows an embedding of the points by means of the Poincaré disk model [1] which embeds the hyperbolic space non-isometrically into the unit disk such that the focus of attention is put onto the point which is mapped into the center of the disk and an overall fish-eye effect results. Moving this focus allows to browse through the map.
For NG no direct visualization is given, but it can be easily reached by a subsequent embedding of the prototypes into the two-dimensional euclidean plane by means of distance preserving projection techniques of the prototypes such as multidimensional scaling. Given pairwise distances $\delta_{ij} := \| \vec{w}^i - \vec{w}^j \|$ of the prototypes (possibly nonlinearly preprocessed, i.e. $\delta_{ij} = f(\| \vec{w}^i - \vec{w}^j \|)$ where $f$ is an appropriate weighting function), this model finds two-dimensional projections $p(\vec{w}^i)$ with pairwise distances $\Delta_{ij} := \| p(\vec{w}^i) - p(\vec{w}^j) \|$ such that the stress-function
$$\left( \frac{\sum_{i<j} (\delta_{ij} - \Delta_{ij})^2}{\sum_{i<j} \Delta_{ij}^2} \right)^2$$
or a similar objective function is minimized.
To apply these techniques, the pairwise distance of prototypes needs to be computed. As beforehand, we assume the identity $\vec{w}^i = \sum_l \alpha_{il} \vec{x}^l$ for optimum prototypes $\vec{w}^i$. We assume that pairwise distances $d_{ij} = \| \vec{x}^i - \vec{x}^j \|^2$ are given, accumulated in a matrix $D$ as beforehand. Then we find
$$\| \vec{w}^i - \vec{w}^j \|^2 = \alpha_i^T D \alpha_i - \frac{1}{2} \alpha_i^T D \alpha_j - \frac{1}{2} \alpha_j^T D \alpha_i$$
where $\alpha_i$ denotes the vector $(\alpha_{il})_l$.
### 3 Experiments
#### Protein Classification
The evolutionary distance of 226 globin proteins is determined by alignment as described in [21]. These samples originate from different protein families: hemoglobin-$\alpha$, hemoglobin-$\beta$, myoglobin, etc. Here, we distinguish five classes as proposed in [6]: HA, HB, MY, GG/GP, and others. Table 1 shows the class distribution of the dataset.
For training we use 29 neurons for Relational Batch NG and Relational Hyperbolic SOM, and a 5x5 grid for standard Relational SOM, respectively. The number of neurons is derived from the hyperbolic grid of depth three (cf. figure 1). The neighborhood range is annealed starting from $N/2$ to 0, $N$ being the number of neurons, in all experiments. The results reported in Table 2 are gained from repeated
10-fold stratified cross-validation averaged over 100 repetitions and 150 epochs per run. Supervision is included in the cost function with mixing parameter 0.5.
Unlike the results reported in [6] for SVM which use one-versus-rest encoding, our setting gives one integrated clustering model. Depending on the choice of the kernel, [6] reports errors which approximately add up to 4% for the leave-one-out error. This result, however, is not comparable to our results due to the different error measure. A 1-nearest neighbor classifier yields an accuracy 91.6 for our setting (k-nearest neighbor for larger k is worse, [6]) which is comparable to our results. Thus, thereby continuous updates improve the performance beyond mere clustering by 3%.
The projections of a Relational SOM with hyperbolic grid structure and of Relational BNG with non-metric multidimensional scaling using Kruskal’s normalized stress1 criterion are shown in figure 1 and 2. The neurons are depicted according to majority vote. Obviously, the neurons arrange according to the associated class and a very clear two-dimensional representation of the data set is obtained.
**Chromosome Images**
The Copenhagen chromosomes database is a benchmark from cytogenetics [19]. A set of 4200 human nuclear chromosomes from 22 classes (the X resp. Y sex chromosome is not considered) are represented by the grey levels of their images and transferred to strings representing the profile of the chromosome by the thickness of their silhouettes. The edit distance is a typical distance measure for two strings of different length, as described in [13, 22]. In our application, distances of two strings are computed using the standard edit distance whereby substitution costs are given by the signed difference of the entries and insertion/deletion costs are given by 4.5 [22].
The algorithms are tested in repeated 2-fold stratified cross-validation using 85 neurons for Relational BNG and Relational HSOM (corresponding two three rings), and a 9x9 grid for the standard Relational SOM. The results presented are the mean accuracy over 10 repetitions per method and 100 epochs per run (cf. [3]). Supervision is incorporated using the mixing parameter 0.9. Again, an improvement of 2% can be observed.
**Macroarray Data**
Barley is an important crop plant, thus, gene expression analysis of the temporal development of barley seeds is an important issue for the derivation of key metabolic processes during different stages of growth. Extensive gene expression measurements at 14 time points from day zero after flowering to day 26 in steps of two days were carried out using cDNA macroarray technology. High signal to noise ratios and high reproducibility between two independently taken experimental series led to a selection of 4824 genes out of 11786 genes available on the 12k macroarrays. Thus, a data matrix of 4824 (genes) by 14 (time points) is considered. As common in gene expression analysis, log2-transformed final expression values are considered. Compression analysis of the available 12k gene expression time series is required to identify groups of commonly regulated genes that may have temporal impact on each other. Such a clustering helps to extract candidate genes responsible for triggering later events like, for example, the influence of cell wall degradation factors for lateral tissue nutrition or subsequent starch accumulation processes. Fig. 3 shows the arrangement which is obtained when training a HSOM with 85 neurons for 150 epochs on these data, thereby using a transformation of Pearson correlation of the expression patterns as distance measure which better accounts for the overall principled shape as described in [29]. Obviously, the map organizes the data according to the evolution of up- and down-regulation of the genes over time. Due to the hyperbolic structure, the map clearly sep-
| Class No. | Count | Percentage |
|-----------|-------|------------|
| HA | 72 | 31.86% |
| HB | 72 | 31.86% |
| MY | 39 | 17.36% |
| GGI/GP | 30 | 13.27% |
| Others | 13 | 5.75% |
Table 1: Class Statistics of the Protein Dataset
| Median Batch | Relational Batch | Standard Relational | Relational Hyperbolic |
|--------------|------------------|---------------------|-----------------------|
| NG | NG | SOM | SOM |
| Accuracy Protein Data set |
|---------------------------|
| Mean | 89.5 |
| StdDev | 1.0 |
| Accuracy Copenhagen Chromosome Database |
|------------------------------------------|
| Mean | 88.8 |
| StdDev | 1.2 |
Table 2: Classification accuracy on the Protein Data Set and Copenhagen Chromosome Database, respectively, for posterior labeling.
4 Conclusions
The extension of SOM and NG to relational data offers a very intuitive and simple possibility to project data which are given by a distance matrix only into a low-dimensional space, thereby preserving the data topology. The method is directly based on the lost functions SOM and NG, respectively, thus opening the way towards further extensions such as supervision as demonstrated in our experiments or hyperbolic MDS, presented for euclidean data in [13, 1]. Unlike alternative proposals such as [5, 12], relational visualization maintains the simplicity of original SOM and NG. The only drawback of the proposed method is given by the complexity of training, which scales with $m^2$, $m$ being the number of data, instead of $m$ such as original SOM and NG. This complexity can be drastically reduced using either exact methods such as introduced in the article [2], or low-dimensional approximations of the vectors which encode the prototypes as proposed in the approach [25].
References
[1] James W. Anderson (2005). Hyperbolic Geometry, second edition, Springer.
[2] B. Comon-Guez, F. Rossi, and A. El Golli (2006), Fast Algorithm and Implementation of Dissimilarity Self-Organizing Maps, *Neural Networks* 19:855-863.
[3] M. Cottrell, B. Hammer, A. Hasenfuss, and T. Villmann (2006). Batch and median neural gas, *Neural Networks* 19:762–771.
[4] T. Graepel, R. Herbrich, P. Bollmann-Sdorra and K. Obermayer (1999). Classification on pairwise proximity data. In M. I. Jordan, M. J. Kearns, and S. A. Solla (Eds.), NIPS, vol. 11, MIT Press, p. 438-444.
[5] T. Graepel and K. Obermayer (1999). A stochastic self-organizing map for proximity data, *Neural Computation* 11:139–155.
[6] B. Haasdonk and C. Bahlmann (2004), Learning with distance transformation kernels, in *Pattern Recognition - Proc. of the 26th DAGM Symposium*.
[7] B. Hammer, A. Hasenfuss (2007). Relational neural gas, to appear at *KI 2007*.
[8] B. Hammer, A. Hasenfuss, F.-M. Schleif, and T. Villmann (2006). Supervised batch neural gas, In *Proceedings of Conference Artificial Neural Networks in Pattern Recognition (ANNPR)*, F. Schwenker (ed.), Springer, pages 33-45.
[9] R. J. Hathaway and J. C. Bezdek. Nerf c-means: Non-euclidean relational fuzzy clustering. *Pattern Recognition* 27(3):429-437, 1994.
[10] R. J. Hathaway, J. W. Davenport, and J. C. Bezdek. Relational duals of the c-means algorithms. *Pattern Recognition* 22:205-212, 1989.
[11] T. Heskes (2001). Self-organizing maps, vector quantization, and mixture modeling. *IEEE Transactions on Neural Networks* 12:1289-1305.
[12] H. Hofmann and J. Buhmann Multidimensional Scaling and Data Clustering (1995), in *Advances in Neural Information Processing Systems* 7:459-466.
[13] A. Juan and E. Vidal (2000). On the use of normalized edit distances and an efficient k-NN search technique (k-AESA) for fast and accurate string classification, in *ICPR 2000*, vol.2, p. 680-683.
[14] S. Kaski, J. Nikkilä, E. Savin, and C. Reiss (2005). Discriminative clustering of yeast stress response, in *Bioinformatics using Computational Intelligence Paradigms*. U. Seiffert, L. Jain, and P. Schweizer (eds.), pages 75-92, Springer.
[15] S. Kaski, J. Nikkilä, M. Oja, J. Vennu, P. Törönen, and E. Castrén (2003). Trustworthiness and metrics in visualizing similarity of gene expression, *BMC Bioinformatics*, 4:48.
[16] J. B. Kruskal (1964). Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis. *Psychometrika* 29:1–27.
[17] T. Kohonen (1982), Self-Organized formation of topologically correct feature maps, *Biological Cybernetics*, 43:59-69.
[18] T. Kohonen and P. Somervuo (2002), How to make large self-organizing maps for nonvectorial data, *Neural Networks* 15:945-952.
[19] C. Lundsteen, J. Phillip, and E. Granum (1980), Quantitative analysis of 6985 digitized trypsin G-banded human metaphase chromosomes, *Clinical Genetics* 18:355-370.
[20] T. Martinetz, S.G. Berkovich, and K.J. Schulten (1993), ‘Neural-gas’ network for vector quantization and its application to time-series prediction, *IEEE Transactions on Neural Networks* 4:558-569.
[21] H. Meyer and M. Vingron (1996), Quantifying the local reliability of a sequence alignment, *Protein Engineering* 9:127–132.
[22] M. Neuhaus and H. Bunke (2006), Edit distance based kernel functions for structural pattern classification *Pattern Recognition* 39(10):1852-1863.
[23] A.K. Qin and P.N. Suganthan (2004), Kernel neural gas algorithm, with application to cluster analysis, *ICPR 2004* vol.4, pp.617-620.
[24] H. Ritter (1999), Self-organizing Maps in non-euclidean Spaces, *Kohonen Maps*, 97-108, Eds.: E. Oja and S. Kaski.
[25] F. Rossi, A. Hasenfuss, B. Hammer. Accelerating relational clustering algorithms with sparse prototype representation, submitted to *ICDM 07*.
[26] M. Schaalb, T. Villmann, A. Wismuller, J. Ontrup, H. Ritter, T. W. Nattkemper (2005). A Hyperbolic Topographic Mapping for Proximity Data, *Proc. of the IASTED International Conference on Artificial Intelligence and Applications*, 106-111, ACTA Press.
[27] J. W. Sammon Jr. (1969). A nonlinear mapping for data structure analysis. *IEEE Transactions on Computers* 18:401–409.
[28] S. Sava and K. Obermayer (2004). Self-organizing maps and clustering methods for matrix data, *Neural Networks* 17:1211–1230.
[29] M. Strickert, U. Seiffert, N. Sreenuvaiulu, W. Weschke, T. Villmann, B. Hammer (2006), Generalized Relevance LVQ (GRLVQ) with Correlation Measures for Gene Expression Analysis, *Neurocomputing* 69:651-659.
[30] P. Tiwari, A. Kahani, and Y. Shental (2004). A generative probabilistic approach to visualizing sets of symbolic sequences. In *Proceedings of the Tenth ACM SIGKDD International Conference on Knowledge Discovery and Mining - KDD-2004*, (eds) R. Kohavi, J. Gehrke, W. DuMouchel, J. Ghosh. pp. 701-706, ACM Press.
[31] J. Walter (2004). H-MDS: a new approach for interactive visualization with multidimensional scaling in the hyperbolic space, *Neural Networks*, 29(4):273-292.
[32] H. Yin (2006). On the equivalence between kernel self-organising maps and self-organising mixture density network, *Neural Networks* 19(6):780-784. |
Investment Patterns in Depth
IN THIS ISSUE:
M&A prospects in China
Keeping up with new foreign-exchange rules
Cracking down on unpaid debt
Now, for the first time businesses can access comprehensive information from China in one location. ChinaOnline is a business news and analysis product located on the Internet for fast access and timely dissemination of information.
Call us for information about a free 10-day trial subscription.
China Online, L.L.C.
900 North Michigan Avenue
Suite 2180
Chicago, IL 60611 USA
email@example.com
www.chinaonline.com
Tel: (800) 893 8229
Tel: (312) 335 8881
Fax: (312) 335 9299
China Online is an independent source of business information and is not funded by any government.
©CHINA ONLINE, L.L.C. 1999, ALL RIGHTS RESERVED.
**CONTENTS**
**FOCUS Investment Patterns in Depth**
8 **Business Over Politics** US firms in China are benefiting from expanding—if still indirect—Taiwan-PRC commercial ties.
*Xiangming Chen*
16 **When MNCs Come to China, Who Changes Whom?** Home-country management styles adapt to the Shanghai marketplace.
*John Z. Yang, John Farley, and Scott Hoenig*
20 **Location, Location, Location** A new study reveals that a foreign firm’s success may hinge on where in China it sets up shop.
*Amy Yingli Liu, Shaomin Li, and Yuxian Gao*
**FEATURES**
26 **China’s Changing Foreign-Exchange Regime** Clearing away the fog of new foreign-exchange rules.
*Thomas E. Jones and Margaret M. Maber*
34 **Old Methods, New Terrain** Foreign companies make the most of China’s slim M&A pickings.
*E. Anthony Zaloom and Liu Hongchuan*
38 **At a Premium** Foreign companies are lining up to gain access to China’s insurance market.
*Ji Chen and Stephen C. Thomas*
46 **No More Excuses** Contrary to popular belief, unpaid debt in China can be recovered.
*Edward E. Lebman and Brinton M. Scott*
**DEPARTMENTS**
4 **Trends & Issues** Trade, IPR, China’s retail markets, polyester production, and more
6 **Letter from the President** WTO This Time?
45 **WWW.China** Hong Kong tenders, MOFTEC, and life in Shanghai
49 **Council Activities** Forecast ’99 in Washington, IPR in Beijing, and Y2K in Shanghai
50 **Commentary** A Recipe for Structural Reform
*Cbi Lo*
54 **Bookshelf** Two reports from the Institute for International Economics, a contrarian view of state-enterprise reform, and a stock market primer
58 **China Business**
62 **Classified Ads**
China's lackluster trade performance last year, which stemmed both from Asia's economic slowdown and a sluggish domestic economy, is expected to continue through 1999. Beijing has thus made maintaining the country's strong trade surplus a priority. The government has enacted a number of policies intended to curb imports, bring relief to exporters, and prevent outflows of foreign capital.
In fact, imports dropped in the second half as Beijing introduced new regulations restricting foreign-exchange flows (see p.26) and curbing the inflow of goods for which domestic equivalents were oversupplied.
Last year's trade picture was notable for a number of developments, including:
- **Flat PRC exports, falling PRC imports** China's total trade contracted 0.4 percent, far short of its 10 percent trade-growth target for 1998. The country's trade surplus was strong, rising 8.2 percent over 1997. Total PRC exports were flat, while imports fell 1.5 percent. Increased export-duty rebates on selected products and strong exports to the United States and Europe helped moderate what would otherwise have been a more dramatic decline in exports.
- **Foreign firms outpace state-owned competitors** State-owned enterprises (SOEs) and foreign-invested enterprises (FIEs) continued to dominate China trade, with FIEs edging forward to capture a slightly greater share of total trade in 1998 ($157.7 billion, or 48.7 percent of total trade). For the first time, FIEs exported more goods than they imported.
- **The United States retakes 2nd place** The ranking of China's top trading partners remained unchanged in 1998 for the most part, though the United States moved ahead of Hong Kong to recapture its position as China's second-largest trading partner.
- **US-China trade grows** Two-way trade rose by more than 13 percent last year, with US exports to China showing strong growth in 1998. US imports from China, meanwhile, rose at a slower rate in the second than in the first half. As the volume of exports continued to dwarf imports, however, the bilateral deficit reached $56.9 billion, one quarter of the total US trade deficit, according to the US Department of Commerce. By contrast, PRC Customs statistics put the US-PRC bilateral trade deficit at $21 billion, up 28 percent. The discrepancy between US and PRC statistics is a result of different methods of calculating the trade that travels between the two countries via Hong Kong.
**No End in Sight**
China's trade performance is unlikely to improve in the coming year, despite Beijing's focus on boosting exports. Declining export growth—which began last August—could continue at least through the first half of 1999, given the anticipated slowdown in trade growth worldwide. Export-duty rebates once again may help improve exports of selected products. But a further downturn in China's export performance or continuation of oversupply problems could lead to fresh import restrictions, including outright bans, "buy-local" campaigns, or new antidumping petitions. Scattered opportunities for US exporters will take the form of lower tariffs for specific sectors, symbolic purchases of such US agricultural products as wheat, and specific trade opportunities in sectors Beijing has targeted for foreign participation.
—Ann M. Weeks and Darlene M. Liao
*Ann M. Weeks is Business Advisory Services associate in The US-China Business Council's Washington, DC office. Darlene M. Liao is assistant editor of The CBR.*
---
**Short Takes**
**Supermarket Sales Snowball...**
Asia Market Intelligence reports that supermarket sales in China have skyrocketed, from $700 million in 1995 to over $5 billion in 1998. But most chains are small, with an average of only 14 outlets each. Shanghai remains a supermarket stronghold in 1997, giants Lianhua, Hualian, and Nonggongshang all achieved sales in excess of $120 million.
...But Department Stores Suffer
Nearly 62 percent of China's 238 shopping malls and department stores recorded a decline in sales in 1998, according to the China National Commercial Information Center, with profits down 15.4 percent to ¥2.6 billion ($314 million). The Center attributes the slide to oversupply and a marked slowdown in consumer spending.
**Furries Boost HK Exports**
Despite declining exports in the first three quarters of 1998, the Hong Kong Trade Development Council's Toy Advisory Committee is predicting that 1999 will be a good year for toy exports. During the first three quarters of 1998, Hong Kong exported $8.4 billion worth of toys—half of which went to the United States. The projected increase is based on robust sales of Furby dolls in the United States in recent months, as well as the popularity of insect characters from recent children's movies.
**Pondering Profitability**
A recent A.T. Kearney survey of 70 foreign multinationals in China revealed that more than a third of them are unprofitable, and a quarter are only just breaking even. Over 60 percent of wholly foreign-owned enterprises are profitable, compared to only 42 percent of joint ventures. Firms in chemicals, automobiles, communications and electronics, financial services, and professional services are among the most profitable, while those in distribution, retailing, consumer goods, and industrial products are among the least profitable.
My first article for *The CBR* three years ago was an informal survey of economists' long-range predictions for China's economic growth. Most agreed that there was substantial evidence to indicate that China would achieve its goal of 8 percent annual growth through 2000 and would continue to grow faster than industrialized countries well into the next century. Were I to speak to these same economists today, their outlook would probably be less sanguine. China's economy has slowed markedly despite Beijing's recent and so-far successful fiscal stimulus efforts.
With many foreign investors reassessing their China presence as a result of today's economic difficulties, this issue's focus is particularly timely. The authors step back a bit from the day-to-day complexities of operating in the PRC to identify the common attributes and features of these investors. Their analyses have implications for China's ongoing campaign to attract foreign direct investment; for foreign companies looking at China for the first time; and for seasoned investors who must decide on their next step in China.
Xiangming Chen of the University of Illinois provides a detailed account of Taiwan investment in the mainland, and the implications for US and other foreign investors of the growing ties across the Taiwan Strait. John Yang and Scott Hoenig of Fordham University Graduate Business School, together with John Farley of the China Europe International Business School, explain the results of their survey of nearly 300 managers of foreign-invested firms in Shanghai. They find that whether US, Japanese, or German, foreign-invested firms appear to jettison some home-country practices when the China market demands a more tailored approach. Finally, Shaomin Li and Amy Yingli Liu of the City University of Hong Kong, and Yuxian Gao of the All-China Marketing Research Co., report the results of their analysis of the stark regional variations within China. They show that the more advanced physical, market, and legal infrastructures in the coastal provinces have contributed to the success of foreign firms located there. Other features in this issue cover the PRC insurance sector, foreign-exchange regime and M&A options.
This is my first issue as editor of *The CBR*—my predecessor, Kirsten Sylvester, has taken a position with the US Department of Commerce. We are sorry to lose her but are grateful for her nearly two years of instruction and leadership.
The goal of any magazine is a more-than-satisfied readership. We at *The CBR* are looking forward to providing the same thoughtful analysis that our readers have come to expect. I also invite all readers to watch for some small—and perhaps not so small—changes in format and content, in the months ahead as part of our never-ending effort to make *The CBR* an indispensable source of relevant, accessible intelligence on the China market.
Catherine Gelb
**FOLLOW-UP ON THE PREVIOUS ISSUE**
Professors Michel Oksenberg of Stanford University and Pitman Potter of the University of British Columbia are co-directors of a program organized by The National Bureau of Asian Research (NBR) on intellectual property rights in China. Research for their article *A Patchwork of IPR Protections* (*The CBR*, January-February 1999, p.8) was conducted under this program.
NBR has initiated a program of conferences, research, publications, and legal training that brings together Chinese officials and professionals with their foreign counterparts to discuss intellectual property and its role in economic development and technological innovation. NBR is a nonprofit, nonpartisan research institution based in Seattle, Washington. For more information on the IPR program and on an upcoming IPR conference in Shanghai, contact NBR:
Tel: 206-632-7370
e-mail: firstname.lastname@example.org
http://www.nbr.org/ipr/index.html
---
**EDITOR**
Catherine Gelb
**ASSISTANT EDITORS**
Virginia A. Hulme
Darlene M. Liao
**BUSINESS MANAGER**
Gregory S. Heslin
**DESIGN/PRODUCTION MANAGER**
Jon Howard/JHDesign
**RESEARCH ASSISTANT**
Julie Walton
---
1818 N St., NW Suite 200
Washington, DC 20036-2470
Tel: 202/429-0340
Fax: 202/833-9027,
775-2476
www.uschina.org/cbr
**PRINTED IN THE USA**
*The China Business Review* welcomes articles from outside contributors. Manuscripts submitted for consideration will be sent anonymously to the editors for review. The US-China Business Council retains the rights to articles and artwork published in *The China Business Review*. Articles or artwork published in *The China Business Review* may be reprinted or reproduced only with the written permission of the US-China Business Council. Opinions expressed in *The CBR* do not reflect US-China Business Council policy, unless indicated.
*The China Business Review*, ISSN No. 0165-7169, is published bimonthly by the US-China Business Council, 1818 N St. NW Suite 200, Washington DC 20036-2470, USA (Tel: 202/429-0340), a non-profit organization incorporated under the laws of the District of Columbia. Periodicals postage paid at Washington, DC, and additional mailing offices. Postmaster: Please send address changes to *The China Business Review*, 1818 N Street NW, Suite 200, Washington DC 20036-2470, USA.
© Copyright US-China Council, 1998. All rights reserved.
ISSN No. 0165-7169 (USP No. 326090)
Annual Subscription rate: $99 US/Canada, $130 international. Single copy price: $25 US/$35 Canada/$25 Israel over 1 yr $10, under 1 yr $12.50. DC residents add 5.75% sales tax. Subscriptions to *The China Business Review* are not deductible as charitable contributions for federal income tax purposes.
All commercial inquiries and renewal/purchase orders should be sent to the above USA address.
**ASIA:**
Godfrey Wu
1305, 13 FI, CC Wu Building
302-308 Hennessy Rd.
Wanchai, Hong Kong
Tel: 852/2591-1071 Fax: 852/2572-5158
E-mail: email@example.com
**NORTH AMERICA:**
Gregory S. Heslin
1818 N St., NW Suite 200
Washington, DC 20036-2470
Tel: 202/429-0340 Fax: 202/833-9027
E-mail: firstname.lastname@example.org
US-China discussions on the terms of China's accession to the World Trade Organization are again intensifying in the run-up to the anticipated visit of Premier Zhu Rongji this spring. It is much too early to predict that agreement can be reached, but the wintry certainty that absolutely nothing is possible has melted a bit as the two sides increase the pace of their discussions.
At the dawn of China's reform era, in 1978, China's total foreign trade amounted to $20.6 billion; it ranked 32nd among the world's trading nations. Foreign investment in the PRC was negligible. In 1998, China's foreign trade reached an estimated $315 billion (in unadjusted dollars); China was the 10th-ranked trading nation on the globe in 1997. Total contracted foreign investment in China at the end of 1998 stood at roughly $567 billion. US contracted investment in China now totals roughly $46 billion, and installed American investment amounts to $21 billion. China is today the 4th-ranked US trade partner, and the United States is the 2nd-ranked Chinese trade partner.
China has been seeking full participation in the world's rules-based trade body—now called the WTO—for thirteen years. Ultimately the entire WTO membership will have to reach consensus on China's accession. But the critical step on the way to any accession is the achievement of bilateral agreements with individual trade partners, spelling out the commitments that the party seeking accession is prepared to make and live by. In the Chinese case, the bilateral negotiating process with the United States has been difficult and protracted. While China has yet to reach final bilateral understandings with other key trade partners including the European Union and Japan, most observers recognize that the bilateral with the United States is crucial—not only in the sense that no WTO membership is possible without a US-China deal, but also because the terms reached in a US-China understanding are likely to be satisfactory in most instances to the needs of China's other negotiating partners.
There seems little doubt that, if the United States and China do not settle their many WTO differences in the next few months, China's WTO prospects will dim markedly for a period of several years at least, thanks in part to the US political calendar. Meanwhile, the WTO will not sit still; the organization, whether China is a full-fledged member or not, is set to embark on a new round of rule-writing and system-building this fall. The United States and China both understand that the often talked-about "window of opportunity" will close soon. Thus the emergent signs of an intensifying US-China dialogue on WTO are of interest to US business, among others.
This is no time for bald predictions of brilliant success or abject failure. But it is time to think about what will happen if the two sides should cobble together a set of WTO terms in time for the premier and his American hosts to announce progress this spring.
In the end, substance is absolutely critical on WTO accession: the terms to which the two sides agree must clearly and convincingly demonstrate that unmistakable economic benefits will flow from China's
accession on the agreed-upon basis. If a US-China WTO understanding should emerge from the current stepped-up negotiations, what will it be and what won't it be?
- **It will be substantive and far reaching.** It will merit the most serious consideration by all interested parties. It should and probably will be intensely practical. It will not, however, be magical. Some say that a WTO package needs to be a "grand slam home run" if it is to survive in an unfriendly Congress, which must after all enact Permanent Normal Trade Relations if the United States is to enjoy the many benefits of China's inclusion in the WTO.
But no WTO deal is going to still all criticism. To believe that a Perfect Package welcomed by absolutely everyone is the only option is to set oneself up for a fall. We need to be ready to get behind a strong product that clearly demonstrates its critical economic benefits to the United States, and whose expected benefits are convincingly superior to the continuation of the present situation.
- **It will be declared a "sellout."** Here at home, familiar themes will emerge: for example, that faint-hearted "peace at any price" types have sold out America. This is rubbish. US negotiators know perfectly well that a skeptical business community, a doubting media community ready to pounce at the drop of a hat, and at least some important members of a chilly Congress would diminish or demolish a package that smacked of "lowering the bar" or a "cave in" or a "political deal." No one knows better than US negotiators that froth will be blown quickly away in Washington.
- **Though some may say otherwise, it will NOT be a cornucopia of US "Gifts" to China. The burden of change will lie with China.** In any acceptable WTO accession package, China, not the United States, will have to do the heavy lifting in terms of opening its trade regime and rearranging the ways it conducts its economic affairs. This is, perhaps, what makes the prolonged WTO negotiation so difficult for China to complete. We can be sure in advance that the predictable claims about US "gifts" and "freebies" to China will be wide of the mark.
- **It will be a complex amalgam of "end points" and "phase-in periods."** Virtually no informed Americans, whether in business or in government, are oblivious to the magnitude of the economic and social tasks facing the Chinese government as it navigates cautiously between a fading Marxist-Leninist economic autarky and an uncertain future tied to market-based globalism. American and Chinese WTO negotiators must weave a fabric whose warp is time and whose weft is change. An acceptable agreement must clearly define the critical commitments to structural change in China's international trade and investment regime, and must ensure that progress toward these "end points" is continuous from the beginning to the end of transition periods of varying length.
- **It will require a careful and balanced evaluation, not only of the economic benefits embodied in the terms of the agreement, but also of the costs to be borne if China remains outside the grasp of WTO rules and obligations for an indefinite length of time.**
The baseline question ought to be the long-term one: is the US national interest more likely or less likely to be served by China's entrance into the WTO under terms laboriously negotiated in the bilateral package? Part of that measurement must include an assessment of the implications of China's continued non-participation in the system. These include China's imperviousness to multilateral insistence on adherence to the requirements of a WTO that excludes it; immunity from multilateral dispute resolution mechanisms; and the continued vulnerability of US economic relations with China to the annual MFN/NTB crisis and to the possibility of mutually assured destruction in bilateral trade conflicts.
- **It will—because it must—offer to the US important long-term economic advantages unattainable without the WTO system.** But it will not resolve all outstanding trade and investment conflicts between the US and China. The advantages will most likely lie in the application of core WTO principles and disciplines to China; in major tariff reductions; in modification or elimination of burdensome non-tariff barriers; in WTO-mandated non-discrimination against US goods; in the WTO dispute resolution process; and in a host of critical, specific provisions that broaden opportunities for US businesses in agriculture, manufacturing, trading and distribution services, financial services, and other key areas of US economic strength.
But the WTO's regime doesn't yet touch on each and every American problem in other countries' economies, including China's. For example, WTO provisions affecting each member country's investment regime don't currently go as far as America would like.
- **It will be billed as both the Millennium and the Apocalypse, when in fact it will be neither.** Accept that the political rhetoric on the WTO issue is likely to be out of sync with the substance of any WTO deal.
The United States and China have a vast and at times volatile relationship, characterized both by extensive cooperation and by sharply divergent perspectives on some basic issues. The rhetoric of public policy on China in the United States is seldom understated. Any US-China agreement on WTO terms is likely to be portrayed in stark, even overheated terms: on the one hand, by claims of a towering achievement that changes the course of history, and on the other hand by claims of grievous damage to the national interest.
Before the hyperbole takes hold, let's face facts: China's entrance into the WTO means neither the Creation nor the end of the world. If the US and China can forge a serious agreement and China proceeds to full WTO accession, it will represent a new stage in the process of China's step-by-step integration into the mainstream of world practice, which began in 1978 and needs to continue far into the future.
Should a WTO deal emerge, the real discussion of its acceptability must be carried out within narrower confines, closer to the center of the political and economic spectrum, on more finite issues. The US business community engaged with China will have a critical role in framing and answering these questions.
All this may prove moot; should nothing materialize this spring, these musings will turn out to be quaintly misdirected. But I'm not writing "Dewey Beats Truman" here, as an unfortunate newspaper did on election night 1948. I'm only saying, "What if he does?" We should prepare to look hard at an emergent US-China agreement on the terms of Chinese accession to the WTO, and to advocate vigorously for it if it passes muster, according to measured and realistic criteria.
Taiwan has been the second-largest source of overseas capital in China throughout the 1990s, ahead of the United States and Japan and behind only Hong Kong. Even as Asian investment in China slowed during the deepening regional economic crisis, Taiwan investment in China has maintained momentum. According to Taiwan figures, annual contracted investment in China was up 72 percent in 1997. Strong commercial ties between Taiwan and China are critical to sustaining economic growth and market expansion in the world's most populous nation. The investment capital, manufacturing expertise, and middle-range technology of Taiwan firms complement China's shortage of capital and abundant low-cost labor.
According to Charng Kao, an economist at the Chung-hua Institution for Economic Research in Taipei, Taiwan investment as a share of China's total capital formation rose from 0.25 percent in 1990 to 1.32 percent in 1995. Taiwan-invested enterprises in China employed 2.8-3.9 million workers and produced $3.6 billion in industrial output in 1995, 3.1 percent of China's total. Taiwan-invested enterprises also accounted for 14.4 percent of China's total exports in 1995. Given the growing links between China- and Taiwan-based suppliers and US companies, US firms doing business in Greater China can only profit from a solid understanding of Taiwan investment patterns in the mainland.
**Policy Shifts in the 1990s**
After years of only limited contact, the 1990s ushered in a new phase of Taiwan investment in China. Following the Kuomintang's 13th Party Congress in July 1988, which approved indirect trade with and investment in the mainland, Taiwan promulgated rules on indirect exports to China, setting the stage for cross-Strait trade. In October 1990, Taiwan announced a set of regulations governing indirect investment in and technological cooperation with the mainland. According to these rules, investment projects capitalized at less than $1 million must be reported to the Ministry of Economic Affairs within six months of signing, while projects involving more than $1 million require advance approval. These policies fostered a surge in commercial activity between Taiwan and the mainland in the first half of the 1990s.
But political and military tensions in cross-Strait relations in 1995-96 prompted new restrictions on Taiwan corporate investment in China. These regulations banned major investment in infrastructure, real estate, and insurance; limited single-project investment to $50 million; and reduced the permissible investment share of a Taiwan firm in a Taiwan-PRC joint venture from 40 percent to a maximum of 30 percent. Though this policy reversal stymied a $3.2 billion deal for the Formosa Plastics Group to build a power plant in Zhangzhou, Fujian Province, it did not halt the overall flow of Taiwan capital into China.
The policy pendulum has since swung back. In early 1998, China relaxed a number of investment policies, in part an attempt to attract larger Taiwan investors and those previously hindered by Taiwan’s own restrictions. The PRC policies encourage Taiwan investment in a wider range of sectors, including agriculture, new and high-tech industries, basic industries, infrastructure, environmental protection, banking, and commercial services.
Further boosting investor and trader confidence was the resumption of cross-Strait talks in spring 1998. More progress was made in October, when the chairman of Taiwan’s Straits Exchange Foundation, Koo Chen-fu, paid a five-day visit to China to meet Wang Daohan, his counterpart at the Association for Relations Across the Taiwan Strait, and PRC President Jiang Zemin. This political move was timed to coincide with the Taiwan government’s relaxation of the $50 million ceiling for any single mainland investment project that facilitates interaction between China and Taiwan or Taiwan’s economic development.
While the full impact of Taiwan’s new policy has yet to be felt, a number of Taiwan firms have already taken advantage of it. For example, Taiwan Cement Corp., headed by Koo, announced it would invest in five plants in southern China to produce 5 million metric tons of cement per year beginning in mid-1999.
**TRADE LEADS THE WAY**
Since 1949, China-Taiwan trade has passed through three successively open periods of indirect trade, though Taiwan was not propelled into the ranks of China’s top trade partners until the 1990s. Taiwan is now the PRC’s fifth-largest trading partner, second-largest source of imports, and ninth-largest export market. China, now Taiwan’s second-largest export destination, is the main contributor to its trade surplus. From 1981 to 1997, Taiwan’s cumulative trade surplus with China amounted to $107 billion (see Table 1). Without this huge surplus, Taiwan would have recorded trade deficits in recent years.
**SIZING UP TAIWAN INVESTMENT**
Part of the recent trade growth can be attributed to rising Taiwan investment in China. In 1996, foreign-invested enterprises (FIEs) in China, most of which were Taiwan-PRC joint ventures, accounted for 67.3 percent of the total China-Taiwan trade—20 percentage points higher than FIEs’ share of China’s total foreign trade, according to economists at the Chinese Academy of Social Sciences. Processing-related trade accounted for almost three-quarters of total China-Taiwan trade in 1997—perhaps not surprising, since most Taiwan-invested enterprises engage in export-oriented processing.
Nevertheless, Taiwan investment in China has had its ups and downs. After officially sanctioning indirect investment, in 1990 Taiwan restricted investment in China to manufacturers of 3,353 products. Most of the listed products were in labor-intensive industries from apparel, footwear, and food to plastics and household electronics. The Taiwan government expanded this list to 4,895 items in 1996. In turn, the PRC opened the door to Taiwan investment when the State Council promulgated Regulations for Encouraging Investment by Taiwan Compatriots in July 1988 (the so-called “22 clauses”). In May 1989, the State Council approved the establishment of Taiwan investment zones in Fuzhou and Xiamen, which were selected because of their geographic proximity to Taiwan. In 1994, the National People’s Congress passed a law specifically protecting Taiwan investment. These reciprocal and mutually reinforcing policies
---
**Table 1**
**Taiwan’s Trade with China via Hong Kong, 1981-97 ($ million)**
| Year | Taiwan’s Exports to China* | Taiwan’s Imports from China | Estimated Total Taiwan-China Trade | Taiwan’s Export Dependency on China** |
|------|---------------------------|-----------------------------|-----------------------------------|-------------------------------------|
| 1981 | 384.8 | 75.2 | 460.0 | 1.7 |
| 1982 | 194.5 | 84.0 | 278.5 | 0.9 |
| 1983 | 201.4 | 89.9 | 291.3 | 0.8 |
| 1984 | 425.5 | 127.8 | 553.3 | 1.4 |
| 1985 | 986.8 | 115.9 | 1,102.7 | 3.2 |
| 1986 | 811.3 | 144.2 | 955.5 | 2.0 |
| 1987 | 1,226.5 | 288.9 | 1,515.4 | 2.3 |
| 1988 | 2,242.2 | 478.7 | 2,720.9 | 3.7 |
| 1989 | 3,331.9 | 586.9 | 3,918.8 | 5.0 |
| 1990 | 4,394.6 | 765.4 | 5,160.0 | 6.5 |
| 1991 | 7,495.5 | 1,125.9 | 8,619.4 | 9.8 |
| 1992 | 10,547.6 | 1,119.0 | 11,666.6 | 13.0 |
| 1993 | 13,993.1 | 1,103.6 | 15,096.7 | 16.5 |
| 1994 | 16,022.5 | 1,858.7 | 17,881.2 | 17.2 |
| 1995 | 19,435.8 | 3,091.4 | 22,525.2 | 17.4 |
| 1996 | 20,727.3 | 3,059.8 | 23,787.1 | 17.9 |
| 1997 | 22,455.2 | 3,915.4 | 26,370.6 | 18.4 |
SOURCE: Cross-Strait Economic Statistics Monthly, January 1998, Mainland Affairs Council, Taiwan
* Estimated based on Taiwan and Hong Kong customs statistics
** Calculated as Taiwan’s exports to China as a portion of Taiwan’s total exports
By the end of 1996, Jiangsu had surpassed Guangdong in terms of contracted investment from Taiwan.
created a more favorable environment for cross-strait investment.
Between 1979 and 1986, Taiwan's total investment in China amounted to only $20 million, according to PRC figures. Starting in 1987, however, Taiwan investment picked up, and by 1990 had reached $890 million, with $222 million realized. In the 1990s, Taiwan investment surged to unprecedented levels. In 1993 alone, Taiwan investment exceeded the total for the entire 1979-92 period. According to PRC figures, between 1991 and 1997, contracted Taiwan investment in China, involving 38,485 ventures, reached $38 billion, with nearly half of the contracted investment utilized (see Table 2).
The large disparities between Taiwan's and China's statistics are likely the result of Taiwan companies' tendency to circumvent Taiwan restrictions by investing through third countries. In practice, many companies reinvest profits in China without reporting such investments to the Taiwan government, inflating the value of their assets and capital stock, in part so that Chinese authorities can satisfy their ambitious foreign investment goals.
Profiles of Taiwan Investment
Taiwan firms historically have concentrated on establishing labor-intensive manufacturing operations in China. By October 1995, manufacturing accounted for about 80 percent of total Taiwan investment in China. Of the total manufacturing investment, electrical goods (30 percent), food and beverages (17 percent), plastics and rubber (15 percent), and metal-making (14 percent) ranked as the top four industries. This focus on manufacturing reflects the fundamental motive behind Taiwan businesses' relocation to the mainland: to take advantage of China's abundant and cheap labor and land.
But Taiwan investment has gradually been shifting from labor-intensive and low-value processing to more capital- and technology-intensive manufacturing such as automobiles, computers, machinery, and petrochemicals. This shift is also indicated by the rising average project size—from 1993 to 1996, the value of the average investment project rose from $340,000 to $3.21 million. By 1997, some 400 Taiwan-invested projects in China had an average capitalization of $10 million, with more than 10 of them exceeding $100 million. According to a recent Wall Street Journal report, Acer Inc., Taiwan's largest PC maker and one of the world's largest PC manufacturers, is building a $25 million plant in southern Guangdong Province to produce PCs and computer motherboards. The project will be expanded to include five additional plants, with investment reaching $300 million in three years. But the President Group boasts the largest Taiwan project, a $700 million investment in more than 30 food-processing and other factories spread across nearly 20 provinces and municipalities. President also exemplifies a more recent trend among larger Taiwan companies in China—producing low-cost products on the mainland for both export and sale within China.
Taiwan firms in China have tended to cluster in a few geographical areas, especially the southeast coastal region. In 1991, four cities in this region—Guangzhou and Shenzhen in Guangdong Province and Xiamen and Fuzhou in Fujian Province—accounted for 45 percent of Taiwan-investment contracts and 44 percent of contracted capital from Taiwan in China.
More recently, Taiwan investment has begun to move north and west, focusing on the lower Yangtze River region, especially Jiangsu Province and Shanghai. Shenzhen's and Xiamen's shares of Taiwan's total investment in China dropped from 14.8 percent and 12.2 percent in 1991 to 9.6 percent and 6.9 percent in 1993, respectively—the latest years for which such figures are available. During the same period, Shanghai's share rose from 11.5 percent to 12.1 percent. By the end of 1996, Jiangsu had surpassed Guangdong in terms of contracted investment from Taiwan. Through 1996, the Yangtze River delta had absorbed more than one-third of Taiwan investment projects and half of its capital. The average capitalization of $1.6 million in the Yangtze delta was also higher than the national
| Year | Taiwan Ministry of Economic Affairs | PRC Ministry of Foreign Trade and Economic Cooperation |
|------|------------------------------------|-----------------------------------------------------|
| | Number of contracts | Amount contracted | Number of contracts | Amount contracted | Amount utilized |
| 1991 | 237 | 174 | 5,884 | 5,256 | 863 |
| 1992 | 264 | 247 | 6,430 | 5,543 | 1,050 |
| 1993 | 9,329 | 3,168 | 10,948 | 9,965 | 3,139 |
| 1994 | 934 | 962 | 6,247 | 5,395 | 3,391 |
| 1995 | 490 | 1,093 | 4,778 | 5,777 | 3,162 |
| 1996 | 383 | 1,229 | 3,184 | 5,141 | 3,475 |
| 1997 | 8,725 | 4,334 | 3,014 | 2,815 | 3,295 |
| 1991-97 | 20,263 | 11,208 | 38,485 | 38,172 | 18,375 |
Sources: Cross-Strait Economic Statistics Monthly, January 1998, Mainland Affairs Council; PRC Ministry of Foreign Trade and Economic Cooperation
Note: PRC figures for 1991 are cumulative from 1983
average ($1.5 million) for Taiwan investment projects. Despite its slow spread, the proportion of Taiwan investment in areas other than Guangdong, Fujian, Jiangsu, Zhejiang, and Shanghai rose from 22.4 percent during 1991-93 to 25.4 percent in 1994.
With regard to the investment vehicle of choice, Taiwan companies in China generally have preferred wholly foreign-owned enterprises (WFOEs). Three separate surveys conducted by Taiwan scholars have shown that about 55 percent of Taiwan-invested factories in China in the early 1990s involved 100 percent Taiwan capital. The extent of 100 percent ownership, however, varied by industry and locality, from a low of 33 percent of projects in the plastics sector in Shanghai to a high of 92 percent of garment ventures in Fujian. This preference for WFOEs stems from the often incompatible business values, company goals, management styles, and power sharing between Taiwan and potential mainland partners; Taiwan investors' lack of familiarity with specific mainland enterprises; and the fact that small and medium-sized Taiwan firms tend to have their own material inputs, production equipment, and access to export channels. At the same time, the common language and culture mitigates these disparities somewhat.
As Taiwan companies have become more familiar with China's business environment and as China's domestic market has become more open, the number of PRC-Taiwan joint ventures has increased. PRC laws that enable joint ventures to maintain management control and to sell 30 percent of what they manufacture in China have prompted some Taiwan firms to set up joint ventures to which they have contributed as much as 97 percent of the equity. And Taiwan companies already in China have begun to use more diversified ownership strategies, such as taking stakes in existing state-owned enterprises.
**Going west vs. south**
As mainland investment fever spread among Taiwan companies in the early 1990s, Taiwan authorities, hoping to reduce the risks of exposure in China, initiated a "go south" policy to encourage investment in Southeast Asia.
The primary argument against "going west"—investing in the mainland—according to both Taiwan government and business, is that too much trade with and investment in China is risky because of the PRC's potentially unstable economic policy and strong opposition to Taiwan's pursuit of greater international stature. Related to the "risky China" argument are the lack of legal protections for Taiwan investors in China (despite some legal improvements in 1994) and isolated instances in which Taiwan businessmen have been kidnapped or their factories robbed.
Others argue that Taiwan's exports to China have grown too fast and too large. In the 1980s, China was the destination for less than 5 percent of Taiwan's exports, compared to 18.4 percent in 1997.
Some analysts in Taiwan also worry that heavy manufacturing investment in China may lead to de-industrialization, or a "hollowing out," of Taiwan's industry. According to one Taiwan estimate, the shift of production to China has caused the loss of almost 20,000 manufacturing jobs and $10 billion in manufacturing output from the late 1980s through 1993. The "hollowing out" argument, however, has been countered by strong evidence that the Taiwan firms that have relocated most of their manufacturing activities to China are no longer competitive at home. The overwhelming majority of Taiwan firms with investment in China also keep more important functions such as research and development, product design and engineering, finance and accounting, sales and marketing, planning and inventory control, quality control, and after-sales service in Taiwan.
The Taiwan government sees less political risk in developing close business ties with Southeast Asia, as the ideologies and political systems of these countries are closer to Taiwan's. Most Association of Southeast Asian Nation (ASEAN) member countries also have bilateral investment-protection agreements with Taiwan. Further, though Southeast Asia is a smaller market than China, countries in this region, especially Indonesia, the Philippines, and Vietnam, have large populations—indicating a large market. Other conditions favoring "going south" include recently introduced policy incentives. Both Indonesia and the Philippines, for example, began in 1994 to allow foreign companies to take 100 percent equity ownership. With the possible exception of Malaysia, Southeast Asian countries also have few or no restrictions on the remittance of profits and foreign exchange. Significant Taiwan investment during 1987-96 flowed to Thailand ($9.5 billion), Indonesia ($9.3 billion), Malaysia ($7.9 billion), Vietnam ($4.4 billion), the Philippines ($969 million), and Singapore ($913 million). Though the Asian financial crisis has slowed the flow of Taiwan capital into Southeast Asia, funds to China have remained stable.
The attractions of "going west" are based primarily on comparative economic advantage. Taiwan investors regard China's cheap and abundant labor and land, large domestic market, and strong engineering capacity in selected industries as complementary to Taiwan's abundant capital, manufacturing experience, management skills, and R&D capacities. This complementarity allows for a vertical division of labor, in which industries in China can specialize in low value-added manufacturing, while Taiwan firms accelerate the upgrading of their operations and the development of high- and new-technology industries. Even if China shifts slightly toward higher-end production, its labor and land resources are still more plentiful than those of ASEAN countries. Cultural affinity and linguistic similarity also strongly encourage Taiwan investors to go west.
Other disincentives to going south include Southeast Asia's higher enterprise taxes; higher—and rapidly rising—wages in Thailand and Malaysia; and transportation infrastructure no better than China's (also with the possible exception of Malaysia). Thailand and Malaysia, in particular, have been shifting away from labor-intensive industries in recent years. Cultural, religious, and language differences also pose greater
barriers than are found in mainland China. And the ASEAN Free Trade Area has more stringent requirements for local content in manufacturing.
Some Taiwan companies are leveraging the advantages of both regions by creating links between investments in the PRC and Southeast Asia. Taiwan's S.T. Group, which had invested in Singapore's electronics industry early on, began to invest in China in 1985. From its Singapore-based holding company, S.T. has established and currently controls 17 subsidiary ventures in China, three in Singapore, two in Malaysia, and one in Hong Kong. Another large Taiwan conglomerate, the Tundex Group, has also been integrating its strategies. In Fuzhou, Fujian Province, Tundex has set up a $3 million garment factory. Together with the Salim Group, headed by Soedono Salim (Liem Sloe Liong), the wealthiest overseas Chinese in Indonesia, Tundex built a huge petrochemical facility in Xiamen. Tundex also recently injected $77 million into its factory in Thailand to double its annual output to 600 tons.
While Taiwan companies may continue to invest in Southeast Asia, the Taiwanese government is abandoning the "go south" policy, according to PK Chiang, chairman of the Council for Economic Planning and Development. The government will no longer encourage investment in Southeast Asia given the region's poor economic state. This official policy change may indirectly deepen Taiwan-China ties.
**The China-Taiwan Supply Connection**
Beyond forging intra-regional linkages, Taiwan investment in China has also formed a cross-Strait manufacturing nexus that is quickly becoming a feature of the regional and global supply chains of major US multinationals. The PC industry provides an illustrative example. Taiwan makes 40 percent of IBM Corp.'s and 60 percent of Dell Computer Corp.'s desktop computers. Although Taiwan was the world's third-largest PC maker in 1995 after the United States and Japan, it has always relied heavily on original equipment manufacturing (OEM) and, to a lesser degree, on original design manufacturing production of PC peripherals to supply global computer giants. Since an average of 60 percent of their products are exported to fiercely competitive international markets, Taiwan PC makers face constant pressure to lower production costs, which they have achieved by moving
---
**Bridging the Taiwan Strait**
The absence of direct sea and air links across the Taiwan Strait has certainly hindered the ability of Taiwan companies to profit from and expand their mainland investments. Nonetheless, the Taiwan Strait has, since 1997, been bridged by de facto links that come close to direct shipping. Whether these de facto links will evolve into formal direct sea transportation in the near future is open to question.
As with the trade and investment ties between Taiwan and China, transportation links between the two have evolved gradually. Discussions in the late 1970s and early 1980s eventually led to January 1986 regulations by Taiwan's Ministry of Transportation that enabled Hong Kong transshipment and third-port customs clearance for ships carrying bulk cargo across the Strait. In 1988, Taiwan authorities permitted, in principle, both foreign freight and passenger ships traveling between Hong Kong and Okinawa to ply the Strait. More recently, in 1995, Taiwan allowed mainland containers to enter Taiwan on foreign ships, lowering the transportation costs for Taiwan companies operating in China significantly. Taiwan also unveiled a plan at this time to establish an offshore transshipment center in Kaohsiung.
On April 19, 1997, the so-called "point-to-point" cargo transportation link across the Taiwan Strait formally began, with the first ships crossing between Kaohsiung and Xiamen (see *The CBR*, March-April 1997, p. 21). The Taiwan regulations designate only Kaohsiung in Taiwan and Xiamen and Fuzhou on the mainland as points of departure and destination. Thus far, six Taiwan shippers and five mainland carriers have been approved to use this link. The 1997 agreement permits foreign-registered vessels to travel directly between Kaohsiung, Xiamen, and Fuzhou, but their cargoes may not pass through Taiwan customs or enter Taiwan's markets. The cargo must be offloaded at a third port.
**Initial and Potential Benefits**
While not yet qualifying as direct shipping, the point-to-point link promises to reduce shipping costs between Taiwan and China. Researchers at the mainland's Chinese Academy of Social Sciences estimate that the annual cargo bound for Taiwan and China through Hong Kong since 1991 has averaged over 10 million tons, with more than 400 ships from Taiwan docking annually at over 30 mainland ports. A round trip via Hong Kong stretches more than 800 nautical miles and takes at least two full days to complete, whereas a straight sail from Taiwan to Xiamen covers only 150 nautical miles in only eight hours. Shipping a twenty-foot equivalent unit (TEU) container through Hong Kong costs roughly $1,400. The same container would cost less than $350 if shipped directly. The president of the Taiwan branch of Jardine Matheson, & Co.'s shipping agency estimated that carriers must pay $5,000 to $6,000 each time they stop in Hong Kong, even when offloading is not required. According to one Taiwan government official estimate, direct cross-Strait shipping would yield an annual savings of $248 million (based on the 620,000 TEUs that crossed the Strait in 1996 and a savings of $400 per container).
The six Taiwan and five PRC shippers that have been approved have already taken advantage of the point-to-point arrangement. Since April 1997, the Offshore Transshipment Center at Kaohsiung Harbor has handled 300,000 TEUs. Companies and shipping firms have reportedly cut their operating costs considerably, as the per-container cost has dropped from $500 to $300. International carriers expect accelerated cross-Strait investment, higher trade volumes, and continued economic growth as a result this link. According to David Arsenault, general manager of Sea-Land Service Inc.'s Taiwan branch, export-service delivery—in which partially manufactured goods are shipped from China to Taiwan for fia large segment of their production to China.
According to Chin Chung, an economist at Taipei's Chung-hua Institution, by 1993 China had already accounted for 34.6 percent of Taiwan's offshore production of PC hardware, surpassing Malaysia's 29.4 percent and Thailand's 27.3 percent. The number of Taiwan PC subsidiaries in China grew from 35 in 1993 to 41 in 1995. In 1993, China accounted for almost half of Taiwan's offshore production of motherboards—and produced some 2 million monitors—more than 50 percent of Taiwan's entire offshore monitor production.
Major monitor makers Kuo Feng and Shamrock Technology plan to move at least two-thirds of their monthly production of 60,000 and 50,000 monitors, respectively, to China in 1999. Chin Chung estimates that monitor production in China could save these companies roughly 3 percent on direct labor costs and another 5 percent on indirect costs, compared with a 5 percent cost savings by producing in an ASEAN country. Considering the continued, worldwide price declines and thin profit margins for computer monitors, motherboards, and mice, even limited cost savings could mean the difference between keeping or losing orders for Taiwan PC makers and components suppliers.
In 1997, IBM increased its purchase of OEM products from Taiwan by 50 percent. In 1998, it opened a $3 million logistics center in Taiwan to serve the domestic market and other Asian markets. IBM and other major US computer companies may benefit from moves by their Taiwan suppliers to move production to China. Suppliers may be able to deliver components faster, and buyers may benefit from lower manufacturing costs. Indeed, Taiwan computer manufacturers have clustered in Dongguan, Guangdong Province, solidifying the city's role as a favored location for computer parts and components sourcing.
Besides building supplier chains across the Taiwan Strait, many US multinationals have taken advantage of Taiwan's strong manufacturing experience by transferring their Taiwan-based managers to their mainland operations. With their advanced knowledge of international manufacturing practices and extensive business experience, coupled with a common language, Taiwan managers have become highly valued by US-invested enterprises in China. And some US multinationals have begun to send their local PRC executives to their well-established Tai-
Problems and Prospects
Despite potential economic benefits, cross-Strait shipping continues to be plagued by both political and logistical problems. The key political constraint on true direct shipping is that China considers cross-Strait transit a domestic issue, while Taiwan views it as an international one. The location of the mainland ports also discourages shippers, as access to other parts of China from Xiamen and Fuzhou is limited. Many shippers prefer to ship their goods to other, more centrally located mainland ports. Researchers at the Chinese Academy of Social Sciences estimate that these difficulties force up to 15 million tons of cross-Strait trade cargo to be transshipped via third ports each year.
Recently, however, prospects for expanded shipping contacts have improved. In August 1998, the Keelung Harbor Bureau, in Taiwan, approved China Ocean Shipping Co., the mainland's largest cargo carrier, to use its foreign-registered vessels to operate a route from the mainland to Keelung, and on to Hong Kong, Manila, and Wellington, New Zealand. In September 1998, China reciprocated by allowing three Taiwan carriers to transport cross-Strait cargo via a third port. Evergreen Marine Corp (Taiwan) Ltd., for example, will operate a route from Shenzhen's Yantian port to Hong Kong, Kaohsiung, Los Angeles, Oakland, Tacoma, Tokyo, Osaka, and back to Yantian. This new agreement enables carriers to transport cargo from both sides of the Taiwan Strait to the United States via a third port, without reloading cargo onto a different vessel at an intermediate stop.
Senior Taiwan government officials are also considering upgrading the Offshore Transshipment Center at Kaohsiung Harbor to a special economic and trade zone and expanding its warehousing and processing facilities. Under this plan, China could be regarded as a third area, like Hong Kong, and all ships (not just the foreign-registered vessels owned by the designated Taiwan and China carriers) could sail between Kaohsiung and mainland ports. If this plan is implemented, and low service rates remain in effect, cargo volume at Kaohsiung Harbor could double. According to one forecast, the PRC, Hong Kong, and Taiwan together could handle approximately 86 million TEUs by the year 2010, with China's share exceeding those of both Hong Kong and Taiwan. The combined total, much of which will come from increased cross-Strait cargo, would account for over 40 percent of Asia's total container cargo and about 20 percent of the world's total container cargo.
Cross-Strait politics may remain a hurdle to direct shipping, however. Plans to upgrade Kaohsiung Harbor, for example, depend on improved Taiwan-mainland relations. But pressure from companies involved in massive cross-Strait trade and investment will bring about true direct shipping sooner rather than later. High projected demand will require the opening of additional ports on both shores. It may not be long before the point-to-point shipping between Kaohsiung, Xiamen, and Fuzhou will become an extensive network of shipping nodes and nexuses involving Keelung, Taichung, and Hualian in Taiwan, and Shanghai, Tianjin, and Qingdao in China.
—Xiangming Chen
Despite its obvious strength in manufacturing, Taiwan faces serious competition in financial services, telecommunications, and media from Hong Kong and Singapore.
Launched in January 1995, the plan aims to make Taiwan the hub for regional manufacturing, financial services, air and sea transshipping telecommunications, and media, with China and Southeast Asia as vast hinterlands. Despite its obvious strength in manufacturing, Taiwan faces serious competition in financial services, telecommunications, and media from Hong Kong and Singapore.
Taiwan may have a competitive advantage in sea and air transshipping because of the island's central location in the Asia-Pacific region. It has the shortest average flying time to seven of the region's most important economic centers (Hong Kong, Manila, Seoul, Shanghai, Singapore, Sydney, Taipei, and Tokyo) and also the shortest shipping time to five of the region's main ports (Hong Kong, Manila, Shanghai, Singapore, and Tokyo). But when it comes to realizing these advantages, the absence of truly direct cross-Strait shipping links becomes a major barrier to efficient integration of transportation and logistic links critical to regional business operations (see p. 12).
The success of US multinationals in Greater China will increasingly depend on their understanding of the important differences in the organizational cultures of the PRC, Taiwan, and Hong Kong. Taiwan managers running small, highly labor-intensive factories on the mainland tend to adopt an almost military management style, characterized by absolute control, rigid organization, and harsh discipline. This contrasts sharply with the more relaxed work place of China's state-owned enterprises.
With the establishment of larger factories that make more sophisticated products for export, Taiwan managers have been increasingly focusing on quality control and inspection, workforce training and development, and human-resource management, with an eye toward developing capable and effective mainland supervisors and managers. Although the shift from hierarchical, traditional Chinese management toward a more participatory management style may be slow, it will help US multinationals and Chinese companies on both sides of the Taiwan Strait to coordinate their organizations and operations in Greater China.
The expanding commercial links between Taiwan and China over the last decade have gone a long way toward economically integrating Hong Kong, Taiwan, and the PRC. As more US multinational corporations look to reorganize their operations to adapt to the Greater China environment, they face the inevitable question of how the political environment will evolve over the long term. So far, cross-Strait economic integration has unfolded as a gradual triumph of business over politics.
We Give You The Best China Service, Everyday.
1st Course • Over 100 years of service to and from China
• 13 MOL offices located throughout China
• Multilingual and multinational transportation professionals.
2nd Course • Excellent internal communication between China and North America sales and operations teams for seamless supply chain management.
Entrée • Seven weekly calls between Hong Kong and North America
• Two weekly calls at Yantian
• Weekly calls at Shanghai and Qingdao
• Integrated feeder network covers all key China ports
• Worldwide logistics and distribution network, including Shanghai Longfei Co., MOL (Hong Kong) Logistics, Ltd., MCS, Ltd., AMT Freight and MOL Intermodal, Inc.
Dessert • State-of-the-art terminals in Shanghai, Hong Kong, Yantian, L.A., Oakland, Seattle, Vancouver and Portland
• Extensive global service, which includes a complete Transpacific, Transatlantic, Intra-Asia and Intra-Americas network.
For reservations please call: Atlanta 404-763-0111, Chicago 312-683-7300, Long Beach 562-983-6200, New Jersey 201-200-5200, Seattle 206-444-6900 or access MOL’s website: http://www.molpower.com
MOL
Profit From The Global Power
When MNCs Come to China, Who Changes Whom?
John Z. Yang, John Farley, and Scott Hoenig
Firms from every leading industrialized country have made major commitments in China, aggressively establishing manufacturing plants, representative offices, and marketing campaigns for their products. In 1996, foreign firms invested more than $40 billion in China. Despite the financial crisis that has since swept major Asian economies and caused some decline of foreign investment, the PRC Ministry of Foreign Trade and Economic Cooperation (MOFTEC) reports that between 1997 and 1998, contracted investment in China rose 2 percent to $52.1 billion, and utilized foreign capital—the funds actually spent on the ground—increased by 1 percent to roughly $45.6 billion.
Of the industrialized countries investing in China, the United States, Germany and Japan stand out as the most active. Major Japanese firms, such as Honda Motor Co., Matsushita Electric Industrial Co. (parent firm of Panasonic), Sharp Corp., and Sony Corp., are among the leading foreign investors in Chinese industry. Large German firms such as Siemens AG and Henkel hold dominant positions in consumer and industrial markets. An automobile venture of Volkswagen AG has captured over half of China's taxi fleet market. And US firms, including AT&T, Avon Products Inc., Chrysler Corp., Hewlett-Packard Co., IBM Corp., Lucent Technologies, Microsoft Corp., and The Procter & Gamble Co., have invested heavily in China.
Because many joint-venture policies and practices are developed at headquarters, they tend to reflect the parent company's corporate culture and strategic imperatives, home-country cultural beliefs and social values, and CEO philosophies. But as numerous surveys have indicated, success in the PRC often depends on a certain degree of adaptation to Chinese conditions and practices (see *The CBR*, May-June 1998, p.20). Many firms with operations in China question to what extent they must adapt, and whether there is some form of international "best practice" that can be used to achieve success, regardless of local conditions. A better understanding of what strategies other firms are employing can help companies benchmark their own policies and aid multinational corporation subsidiaries entering China.
**The Latest Results**
To answer these questions, 286 Shanghai-based managers of US, Japanese, and German subsidiaries were surveyed in mid-1997. Respondents commented on the PRC investment environment, successful strategies for business in China, the parent firm's decisionmaking structure and management characteristics, and their subsidiary's human
resource management policies. Though Shanghai's comparatively developed marketplace may differ from those of other PRC cities, particularly those in the interior, the survey nonetheless reveals priorities of foreign investors—and that problems exist even in a city considered by foreign investors as among the most attractive in China (see p.20).
The respondents formed a fairly representative cross-section of Shanghai's MNC subsidiary management. Almost two-thirds were senior managers, of which 32 percent were general or deputy general managers. Eighty-two percent of the respondents were Chinese, 9 percent Japanese, 2 percent American, and 2 percent German. Expatriate managers made up 22 percent of the total; the rest were locally hired managers. The firms for which the respondents worked also formed a relatively complete sampling of Shanghai subsidiaries. Almost one-third of respondents were from wholly foreign-owned subsidiaries, 41 percent were from joint ventures, and a quarter were from representative offices. With respect to the nationalities of their parent companies, 44 percent were Japanese, 26 percent were American, and 21 percent were German. Most of the firms had established their presence in China within the last 3-4 years.
Answers were analyzed according to parent-company nationality. Differences in responses across groups were taken to indicate that the home country significantly influenced the formation of management practices. Similarities in responses were assumed to indicate either use of an internationally based set of best practices or specific adaptation to the Chinese market. Use of international or Chinese practices would occur where economic, political, or cultural norms dictated that companies conform to compete effectively.
A HANDFUL OF DIFFERENCES
Overall, the survey showed a high level of adaptation to Chinese or international business practices, regardless of home-country predilections. Nonetheless, the survey identified a few cases in which practices among US, Japanese, and German firms varied statistically. In particular, respondents viewed German management as less flexible than either US or Japanese management (see *The CBR*, May-June 1998, p.54). In one case, the German managers of a telecommunications equipment joint venture in Beijing were reluctant to change some human-resource policies even though the policies clashed with local practices. By contrast, in another case, the managers of a US telecommunications company abandoned use of personal referrals as an important hiring method, as the policy encouraged too many relatives and friends to apply for positions. German firms also prefer to remain wholly foreign-owned. This, plus German companies' relative lack of flexibility, may be the source of the higher number of managerial problems in German-owned ventures than in US- or Japanese-owned ventures.
Of the 70 areas of human resource management policies surveyed, only six showed statistical variance across firms from different home bases (see Figure).
**DIFFERENCES IN HUMAN RESOURCES PRACTICES AMONG MULTINATIONALS IN CHINA**
| Practice | US | Japan | Germany |
|-----------------------------------------------|------|-------|---------|
| Manager recruitment conducted through headhunters | 3.0 | 3.0 | 2.5 |
| Internal promotion determined by seniority | 2.5 | 2.5 | 2.0 |
| Salary of non-management employees based on seniority/tenure | 2.5 | 2.5 | 2.0 |
| Employment policy characterized by lifetime employment | 2.0 | 2.0 | 1.5 |
| Employment policy characterized by merit-based employment | 3.5 | 3.5 | 3.0 |
| Career development characterized by job rotation in the firm | 2.5 | 2.5 | 2.0 |
*SOURCE:* John Yang, John Farley, and Scott Hoenig
*NOTE:* Based on a scale of 1 to 5, with 1 being least likely to employ this practice and 5 being most likely.
When an MNC enters the Chinese arena, it must work out its relationship with its parent company and create a mix of expatriate and locally hired managers.
With the exception of merit-based employment practices, all areas deal with policies not extensively in use in most subsidiaries. The differences, nonetheless, demonstrate the marginal impact on a firm's China operations of home-country and parent-company culture, as well as administrative heritage and practices.
Perhaps because German firms tend to emphasize professional skills, specialization, and efficiency, their China ventures typically rotate employees through different jobs less often than do US and Japanese firms. US firms in Shanghai take the lead in using headhunters to recruit local managers, followed by Japanese firms. This may be a carryover from home-country practice, as firms in the United States also rely heavily on headhunters to recruit senior managers. Lifetime employment emerged as less important in US firms than in German and Japanese firms, partly reflecting shareholder values and the American tradition of using layoffs to cut costs when economic conditions take a downturn. Merit-based employment is most common in US firms, but firms of all three nationalities seem to employ this practice in Shanghai to some degree—even Japanese firms accustomed to lifetime contracts for their employees.
Like firms in their home country, Japanese firms in Shanghai are distinguished by their emphasis on seniority in the promotion process. Length of tenure is also used more by Japanese firms to determine salaries for non-management employees. But these tendencies are much less prevalent in Japanese ventures in Shanghai than in their home-country headquarters.
**Alike in many ways**
More striking than the differences among MNCs from Japan, Germany, and the United States are the similarities in the companies' strategies for success, decisionmaking structures, management characteristics, and human resource management practices in China. When asked about factors that facilitated or hindered their firm's operations, those surveyed responded more or less similarly.
- **Investment environment and infrastructure** Managers from most firms, regardless of nationality, considered PRC foreign-investment policies, as well as political, legal, and social environments, most important to their venture's success in China. Competent, qualified Chinese professionals with solid work ethics, and reasonable living conditions for expatriates were also valued. The reputation of Shanghai's work force as better educated, more entrepreneurial, and more motivated than those elsewhere in the country may have been attractive to these firms.
Most respondents cited poor access to high-quality raw materials and key inputs, economic data and business information, and wholesale and retail distribution networks as significant problems. Procuring quality products has long been difficult in China, even for state-owned enterprises. Private wholesale and retail distribution networks remain underdeveloped even after two decades of reform.
Despite recent measures to improve dissemination of economic statistics, the Chinese government still controls access to key business data and statistics, to the detriment of foreign firms. Most respondents also named China's poor transportation, communications, and other physical infrastructure as obstacles to their success.
- **Marketing strategies** Respondents in ventures from all three countries rated aggressive and creative marketing as crucial to success in China's increasingly competitive domestic consumer market. Especially valued is the ability to offer brand-name and innovative products. Respondents also put a premium on having reliable market intelligence to conduct thorough and accurate feasibility studies.
- **Solid local government relations and parent company support** Good relations with the local or national government, particularly for larger projects, and full support from parent firms contribute significantly to a company's performance in China, according to respondents. Government relationships are especially important to subsidiaries, given that the Chinese government plays a critical role in formulating economic, foreign-investment, and tax policies. Home-country support is also crucial to maintaining a developing business.
- **Parent-firm decisionmaking structure and management characteristics** Respondents agreed that when an MNC enters the Chinese arena, it must work out its relationship with its parent company and create a mix of expatriate and locally hired managers capable of leading the foreign operation effectively. Parent companies tend to position their firms in China for long-term development, tightly oversee major investment decisions, and be less concerned with short-term profitability. The survey respondents, many of them Chinese, were satisfied with expatriate managers' professional competency and technical knowledge, cooperation with Chinese employees, and willingness to teach Chinese managers new techniques. Few reported inflexibility and narrow-minded approaches within MNCs.
- **Human resource management practices** Priorities that the foreign subsidiaries in Shanghai place on various human resource management policies reflect adaptation on both the part of the Chinese and foreign MNCs. In recruiting new employees, a positive work attitude, potential for growth, and work experience are important criteria, reflecting foreign firms' emphasis on long-term employee development. Where someone went to school and test performance are less important, as most believe that the constantly changing Chinese educational system is not an accurate measure of an employee's ability. To fill management positions, MNCs rely frequently on recruiting seminars and base hiring decisions consistently on work experience and professional competency, not on an applicant's recommendations or alma mater.
MNCs' evaluations of non-managerial employees emphasize professional competency, loyalty to the firm, and longterm potential, and focus less on an employee's relations with co-workers or short-term results. Similarly, MNCs from all three countries tend to evaluate managerial staff based on professional competency and leadership skills.
Regarding career development, most firms said they aim to promote people from within. With the exception of some German firms, companies also tend to consider overseas training and a generalized, rather than specialized, career path more important than job rotation in developing staff potential.
Firms that rely heavily on internal promotion practices make decisions based on an employee's ability and performance first and foremost, followed by work experience. According to the survey results, performance was the most important factor in determining salaries of both management and non-management employees, followed by skills, competency, and seniority. The emphasis on performance has become the critical feature of MNCs in China and may demonstrate the strong influence of typical US management practices that emphasize performance over length of tenure. Managers of most surveyed firms said they were not averse to laying off employees when business conditions require it, but must be cautious and careful to prevent unexpected negative reactions from local employees.
Benefits provision has become a critical area in which MNCs compete for high-quality Chinese employees. To attract and retain competent managers and employees, survey respondents cited medical insurance as the most important among employee benefits, followed by retirement benefits, life insurance, and, lastly, employee housing and dental insurance.
**GIVE AND TAKE**
The similarities among the management priorities and practices of Shanghai's Japanese, German, and US subsidiaries may indicate that local culture and conditions and general international business practice have had a greater impact on MNC management practices and policies than home-country customs—although some home-country impact is present. China's legal, social, organizational, and economic conditions appear to have muted the implementation of some foreign management policies in Shanghai. These conditions include China's developing infrastructure, a value system influenced by tradition, government regulation of the market, and an overcrowded labor market. Traditional Japanese policies such as lifetime employment and seniority-based compensation are not as visible in Japanese firms in Shanghai as in companies in Japan, perhaps because of rapid changes in China's value system and employment practices. In China, after years of an "iron rice bowl" environment, pay for performance has become the norm.
And many firms surveyed found it worthwhile to localize human-resource functions and give local managers more power over the formation of key policies in the early stages of a venture's involvement in Shanghai. But as firms begin to understand the market and instill the home-country common corporate culture among staff, there could eventually be some divergence among MNC practices.
That foreign firms appear to have adopted some internationally accepted best practices in China is likely because of the intense competitive pressure MNCs face in developing markets. For the past 10-15 years, MNCs have increasingly adopted similar management policies worldwide—namely, those that emphasize merit, performance, experience, and some specialization. Despite the strong national influences on management policies in countries such as Japan, the global trend is toward performance- and merit-based management.
Exactly who changes whom when foreign MNCs enter the Chinese market? One answer to this question lies in the hybrid human resource management model apparent in US, Japanese, and German firms in Shanghai. This model seems to be the result of pushing by parent firms to adopt more of their practices developed through both home-country influence and global best practices, and some pulling—in the form of pressure to conform to the local Chinese environment. Thus, to be successful in China, MNCs must adapt their global strategic imperatives to local conditions. Overall, the survey results indicate that MNCs with flexible business policies may have the best chance for a lasting presence in China.
Location, Location, Location
A new analysis indicates that a foreign firm’s location in China can affect its business performance.
Amy Yingli Liu teaches marketing and China business at the City University of Hong Kong. Shaomin Li, an associate professor at the City University of Hong Kong, specializes in strategic management and business and society in China. Yuxian Gao is president of the All-China Marketing Research Co., a subsidiary of the State Statistical Bureau of China.
Location, location, location. Though widely known as the “golden rule of real estate,” in this case, location is one of the keys to a foreign investor’s success in China. Location has not been among the factors traditionally recognized as important to a firm’s performance, however, as business strategists have generally focused on firm- and industry-specific factors. Firm-specific factors, such as resources, capabilities, and strategies, understandably are the most important determinants of a firm’s performance. Companies in industries with high entry barriers, few substitutes, and bargaining power greater than that of customers and suppliers—such as aircraft and pharmaceuticals—tend to enjoy higher profits than firms in industries without these characteristics.
A firm’s location, and that location’s effect on performance, has received less attention from strategists than have firm- and industry-specific factors, in large part because location matters less in advanced market economies. Regions within developed economies have little incentive to erect trade barriers or close their borders. On the contrary, as firms can “vote with their feet,” regions often compete to create the most efficient political and economic conditions possible, including free trade and movement of capital, minimal government interference in wage- and price-setting, protection of property rights, and enforcement of contracts. In an ideal system, investment returns, costs, and benefits across locales would converge toward an equilibrium. In the United States, for example, New York City has been pressured to improve government efficiency and lower business costs to prevent companies from moving across the border to New Jersey.
Of course, even in advanced economies such as the United States, geographic location affects firm performance. New York City may offer a larger client base and more business opportunities than a small midwestern city, but cannot match the lower operational costs of such a city no matter what special breaks it offers. Compared to developing economies, however, advanced economies come closer to the ideal equilibrium.
China is far from that ideal, with some areas highly marketized and others still close to the command model. Government policies, infrastructure, and other factors that influence business activities tend to differ substantially from one locality to another, and
the country's sheer size only magnifies these variations. One China-based executive of a major multinational corporation has characterized his job as "dealing with 30 Chinas: the 22 provinces, 5 autonomous regions, and 3 municipalities [that] actually make up 30 market segments in China, each the size of a country like France or Germany."
**HOW IMPORTANT IS LOCATION?**
Multinational companies will not base their China strategies on anecdotal evidence alone. Companies need proof that location has a significant effect on performance. A group of researchers at the City University of Hong Kong conducted a study in 1998 to evaluate the influence of location and industry on performance of foreign firms in China. Using PRC State Statistical Bureau (SSB) industrial censuses for 1993-95, the researchers studied 257 foreign firms with single-product manufacturing businesses in 141 of the roughly 600 industries in China. Most of the firms were in the electronics and communications, industrial and construction materials, leather-processing, textile, or electrical-machinery sectors. The researchers decomposed the firms' returns on assets (ROA), a standard performance indicator, and compared the separate effects of location and industry on companies' ROAs.
The study revealed that where firms are situated in China explains more of the variation in performance than does the industry in which the firms operate. More specifically, the study found that though firm-specific factors account for 72 percent of the variation in performance explained by the model, location accounts for roughly 17 percent, while industry-specific factors account for only 11 percent.
The study also identified which localities foreign investors favored. For each PRC province, the study measured the ratio of foreign capital inflows during 1993-95 to domestic capital levels in 1995. According to the study's rationale, if the foreign-to-domestic firm ratio was the same in each province, this should indicate that all provinces are equally attractive. The results show, not surprisingly, that coastal provinces are most favored by foreign investors. But the results also reveal that foreign investors increasingly began to favor some inland provinces not previously known for attracting foreign investment. For example, Ningxia, Inner Mongolia, Yunnan, and Liaoning provinces ranked unexpectedly high in the study. This may be because their domestic capital bases were smaller than those of other regions. But more important factors may have been changes in local policies or relatively inexpensive inputs, which may have helped channel investment to these less-developed provinces.
What specific conditions made certain locations more attractive than others is of interest to both would-be foreign investors and local governments seeking such investors. To answer this question, the study linked the attractiveness score—based on the ratio of foreign to domestic capital—of each province with a number of elements considered important to the location choice of foreign firms. After conducting a factor analysis and multiple regression analysis on 34,911 foreign firms in all 30 provinces and nearly 600 manufacturing industries, the study found three factors that appear to be the most important in determining foreign firms' choice of location in China (see Table).
**INFRASTRUCTURE**
Among the three factors, infrastructure has the greatest effect on a firm's location decision. Infrastructure is vital in attracting manufacturing firms to a region, especially in transition economies such as China's, which generally have poor infrastructure. Certain essential inputs, including power and telecommunications links, are often inadequate. In China, the level of infrastructure development varies from one region to another, largely because of the lack of fully developed capital, land, and labor markets. Central-government policies that have favored certain regions at the expense of others have only exacerbated the problem.
To ascertain the link between infrastructure and location choice, the statistical analysis measured infrastructure conditions for each province and municipality. The variables used to determine infrastructure levels were electricity usage per capita, number of telephones per capita, road-to-land ratio, and wage level. Purchasing power per capita and population density were used to gauge market size. As expected, the higher the figures for these factors,
The calligraphy shown above was personally inscribed by President Jiang Zemin. It reads:
"Wishing complete success to the 1999 FORTUNE Global Forum in Shanghai."
President Jiang Zemin
The People's Republic of China
Participation, by invitation only, is limited to the chairmen, presidents, and CEOs of major multinational companies. For more information, contact:
Tina Wee-Ling, FORTUNE Conference Division
Telephone: 65-3933-623 Fax: 65-3933-620
E-mail: email@example.com
Web Site: http://www.fortune.com/conferences
Presented by: TIME WARNER
Official Airline: CATHAY PACIFIC
Next September, the leaders of the world's largest companies will gather in Shanghai for the 1999 FORTUNE Global Forum. The Forum, to be held on the eve of the 50th anniversary of the People's Republic of China, will explore the business realities and extraordinary opportunities ahead as China enters the next century. President Jiang Zemin will deliver the keynote address. Mayor Xu Kuangdi of Shanghai will host the second evening.
An impressive group of leaders have confirmed their participation in this event, including:
Stephen Boilenbach
Hilton Hotels Corp.
Michael Bonsignore
Honeywell, Inc.
Michael Dell
Dell Computer Corp.
Michael Eisner
The Walt Disney Co.
Roger Enrico
PepsiCo, Inc.
George M. C. Fisher
Eastman Kodak Co.
Noël Forgeard
Airbus Industrie
Maurice Greenberg
American International Group, Inc.
Gerald Greenwald
UAL Corp. and United Airlines, Inc.
Ambassador Carla Hills
Hills & Co. and Former U.S. Trade Representative
Durk Jager
The Procter & Gamble Co.
The Hon. Henry Kissinger
Former U.S. Secretary of State
Henry Kravis
Kohlberg Kravis Roberts & Co.
Gerald Levin
Time Warner Inc.
Göran Lindahl
ABB Asea Brown Boveri Ltd.
Reuben Mark
Colgate-Palmolive Co.
Ernest Micek
Cargill Inc.
Jack Murofushi
ITOCHU Corp.
Nicholas Negroponte
Massachusetts Institute of Technology
Jorma Ollila
Nokia Corp.
John Roth
Nortel Networks
Jürgen Strube
BASF AG
Ted Turner
Time Warner Inc.
Shigeji Ueshima
Mitsui & Co., Ltd.
Daniel Vasella
Novartis AG
G. Richard Wagoner, Jr.
General Motors Corp.
Jack Welch
General Electric Co.
...and many others.
Host sponsor: The Shanghai Municipal People's Government.
Major Sponsors:
The earlier a province opened to foreign investment, the more attractive it is to foreign companies.
Market Structure
Firms appear to base their location decisions partly on the structure of the market, which the statistical analysis pinpointed as including four elements: privatization; early opening to foreign investment; concentration of foreign investment within an industry; and concentration of domestic competition within an industry. A province's degree of privatization—measured by the proportion of private firms to the total number of firms—correlates with the province's ability to attract foreign investment.
A comparison of legal property rights protection and market rules in different regions in China was also instructive. Regions with a high degree of privatization tend to exert greater effort on developing legal systems that enforce commercial laws and market rules. Such areas thus have attracted more foreign investment than regions with a lower degree of privatization. For example, Guangdong Province, the pioneer in facilitating privatization and attracting foreign direct investment, spent ¥60.60 ($7.26) per capita on legal infrastructure development in 1995, well above the national average of ¥22.90 ($2.74).
The study revealed that the earlier a province opened to foreign investment, the more likely a foreign company was to invest there. The central government first allowed foreign investment in China in the late 1970s, but each province began to accept foreign investment at its own pace. Provinces that opened relatively early tended to have better-developed political and economic institutions to enforce their legal frameworks, particularly in the areas of property rights.
PRC Provinces Favored by Foreign Firms
SOURCE: Yuxian Gao, Shaomin Li, Amy Yingli Liu
and fair competition, and thus are more attractive to prospective foreign investors than those that opened later.
The concentration of foreign investment in a particular industry appears to be another contributor to location decisions. By setting up in a region in which firms in the same industry are clustered, a company may enjoy knowledge spillovers from competitors, and pools of specialized labor and input providers created by industry demand. The regression analysis revealed that foreign firms seek to benefit from this agglomeration effect—they are drawn to regions where many of their competitors are already situated.
Foreign firms tend to follow their domestic competitors as well, though statistically, the pull of domestic industry is not as strong as that of the industry concentration of foreign firms. This is most likely because the current distribution of domestic industries in China still reflects three decades of central planning rather than market-driven resource allocation.
**Tax policy**
Tax policies also affect where a company chooses to locate. The regression analysis found that the higher the average corporate tax rate, the less attractive a province was to foreign investment in the 1993-95 period. Local governments have considerable power to offer tax breaks as incentives. As a result, tax rates vary from region to region.
Some local governments are willing to change an individual firm's tax rate post ante. For example, where actual performance is substantially below expectations, firms have negotiated with the government to reduce their tax obligations. This subjectivity and variation in tax rates may bestow advantages on firms that have better connections and know how to bargain with the government. But many foreign investors tend to shy away from regions where the local government bargains with firms to set individual tax rates. Moreover, recent moves by central-government tax authorities suggest that there may be an effort under way to equalize corporate tax rates across the country, which eventually could make tax policy a less important factor for the location choice of incoming foreign investors.
**Making the perfect match**
The statistical analysis demonstrates that a province can make itself much more attractive to foreign investors by focusing on improving the conditions identified as important. For example, the link between privatization and the location choice of foreign firms shows that a free-market environment is key to attracting foreign investment. Regions with backward economies in need of foreign investment should therefore adopt more liberal policies toward private industry. Local governments should also establish and enforce a clear legal and regulatory framework that treats all investors equally. Such policy changes are likely to be especially effective in inland regions with relatively underdeveloped markets and abundant labor and land.
When formulating a strategy to invest in China, foreign investors should carefully examine each of the location factors in light of their firms' objectives, resources and capabilities. Companies should also use these factors to help them forecast changes in conditions. Some trends, such as government policies, can change more quickly and easily than others, such as the quality of a province's highway network. And some changes may be more predictable and permanent than others. For instance, the construction of a power plant is likely to be more predictable than policy and regulatory change.
Companies also should assess the likelihood of change that could affect the attractiveness of different regions. For example, inland regions are starting to offer tax policies that are more favorable to foreign investors than those offered in coastal regions. Certain inland regions, such as Yunnan and Anhui provinces and Tibet Autonomous Region, exempt all local taxes for foreign firms without specifying a time limit. Though the tax climate may be undergoing some rationalization at the central government level, efforts to encourage investment in inland areas, combined with other changes such as improvements in infrastructure, may yet lead to an expansion of foreign investment opportunities in China.
China’s Changing Foreign-Exchange Regime
Thomas E. Jones and Margaret M. Maber
Along with its Asian neighbors, China has coped with a good deal of pressure on its domestic currency, the renminbi (RMB), since the regional recession began. Unlike other Asian economies, however, China has chosen not to devalue. To the relief of its trading partners, the PRC government has adopted a staunch political stance against devaluation, committing itself to maintaining RMB stability and preventing another round of competitive devaluations across Asia.
China’s firm economic tactics, though praised by many foreign governments, have not come without a price. Export growth slowed significantly in 1998, with the balance of trade remaining healthy only because of a considerable drop in imports. These trends also raised specific queries regarding China’s foreign-exchange reserves. In January 1998, most PRC officials were satisfied that the overall level of reserves, at $140.5 billion, would meet the country’s needs. But reserves failed to rise substantially through the first half of 1998—by July 1998, reserves had risen by only $100 million, to $140.6 billion. With the makings of a record-breaking trade surplus well under way, many observers were hard pressed to explain why the growth of China’s forex reserves had stalled. The cause seemed to be a number of leaks in the country’s forex regime.
In an impressively swift and concerted effort, in June 1998 the People’s Bank of China (PBOC)—China’s central bank—and the State Administration of Foreign Exchange (SAFE)—the administrative organ under PBOC in charge of foreign-exchange management—launched a series of initiatives aimed at finding and closing foreign-exchange loopholes. PBOC and SAFE issued a raft of “internal” notices to strengthen the supervision of foreign-exchange current- and capital-account transactions (see Table). These notices were circulated within SAFE, PBOC, and other banks and financial institutions in China, accompanied by warnings that they would be implemented immediately and strictly. The regulations appear to have had a marked impact: between August and October 1998, China’s forex reserves jumped $3.2 billion, to $143.7 billion. By the end of 1998, reserves totaled nearly $145 billion. In light of this success, Chinese officials are continuing to pressure market players to meet government examination, approval, and registration requirements for an array of forex-related transactions.
The 1998 notices affect those engaged in trade with China as well as those in the PRC market itself, including foreign investors, foreign-invested enterprises (FIEs), foreign bank branches, and domestically funded enterprises. Foreigners now must
expend considerable time and patience not only dealing directly with SAFE and the designated banks to unlock forex payments, but also determining how not to run afoul of this tougher forex regime.
**A Stronger Current Account**
China has permitted foreign-exchange convertibility for current-account transactions since 1996, but has continued to maintain strict control over capital-account transactions. Current-account items are recurrent transaction items that arise in the course of international receipts and payments, including revenue and expenditure from international trade in goods and services and dividend remittances. Capital-account items arise from the inflow and outflow of capital, including direct investments, loans, and liquidation payments.
Before the recent regulations took effect, straightforward procedures already existed under which companies were able to withdraw foreign currency from a foreign-exchange bank account to pay for a current-account item. In most cases, the FIE's foreign-exchange bank made payment decisions upon reviewing valid supporting documentation and commercial vouchers or contracts presented by the account holder.
One typical transaction that has come under scrutiny is the payment to foreign investors of after-tax profits in foreign exchange. Article 21 of PBOC's 1996 Administration of the Settlement, Sale, and Payment of Foreign Exchange Provisions details such procedures. Under the article, such payments could be made from an FIE's foreign-exchange current account or through the conversion and payment effected at a designated foreign-exchange bank on the strength of the written resolutions of the FIE's board of directors.
Another important case is that of import payments. FIES were required to provide SAFE with documents, stamped by the PRC General Administration of Customs, showing that import procedures had been completed properly and that foreign exchange in a given amount was due to a legitimate, contracted trade partner. In both cases, such reviews were intended as an administrative check, not as a discretionary review of the underlying purpose of the related payments.
As central authorities grew concerned that current-account transactions were contributing to China's foreign-exchange "losses," they took steps to tighten all aspects of current-account transactions. A number of PBOC and SAFE circulars released during the second half of 1998 affect general requirements for documentation and official permissions for trade payments; the use of letters of credit for external trade transactions; payments for the use of intellectual and intangible property; and remittances of profits, dividends, and bonuses abroad.
**Managing Import-Export Activities...**
Some of these regulations directly affect FIE business. FIEs involved in trade financing, in particular, have felt the sting of the new notices. Many Chinese enterprises successfully managed to conceal large caches of foreign-exchange funds abroad. Chinese officials, probably correctly, came to suspect that much of this capital flight was effected through fraudulent import trade payments. Illegal payments have been made by forging false import documents or improperly obtaining actual import documents from official organizations, among other methods. Upon receipt of forged documents, designated banks accepted and verified the request to remit foreign-exchange funds overseas. Once remitted abroad, the funds could be deposited or invested without interference from PRC authorities. Many observers believe this type of transaction is largely responsible for the discrepancy between China's substantial trade surplus and stagnating foreign-exchange reserves.
To help stem this illegal behavior and assert greater control over forex flows through external trade channels, SAFE has issued a series of rules and circulars that primarily build upon earlier foreign-exchange legislation. The new measures put in place to control foreign exchange derived from exports and imports are the Detailed Rules for the Implementation of Administrative Measures for Verification of Export Exchange Collection (Huiguofazi [1998] No. 22) and the Circular on Promulgation of Regulation on the Examination of Trade Authenticity for Verification of Import Payment in Foreign Exchange (Huiguobanzi [1998] No. 199), both issued by SAFE in June 1998. These notices aim to prevent unauthorized buying and selling of foreign exchange under the guise of trade. Among other things, they strictly prohibit the unauthorized use of verification certificates to collect foreign exchange for exports, impose complex measures for verifying that goods are actually traded, and, in the case of PRC enterprises, re-state the requirement of surrendering foreign-exchange earnings.
At the outset of the forex campaign, apparently 20 percent of the customs forms vetted by designated foreign-exchange banks were found to be false. SAFE's Notice on Improving Management on Sales and Payment in Foreign Exchange (Huibian [1998] No. 27) targets this illegal practice. Article 1 states that banks must "confirm the authenticity of import customs declaration forms with the customs authorities" when the total amount of a foreign-exchange transaction equals or exceeds $100,000. Article 13 requires importers making advance payments for transactions valued at $100,000 or above to complete a "recordal form on the payment of foreign exchange for imports" as part of a pre-verification process with SAFE. In addition, SAFE is required to inspect randomly no less than half of all transactions involving advance payment.
SAFE has also been ordered to intensify its fight against foreign trade agencies that procure foreign exchange illegally. The Notice on Strengthening the Work of Fighting the Activities of Obtaining Foreign Exchange by Cheating (Yinfà [1998] No. 557), issued by PBOC, SAFE, the Ministry of Foreign Trade and Economic
Banks do not know how to confirm the authenticity of customs declaration forms.
Cooperation (MOFTEC), Customs, the State Administration for Industry and Commerce (SAIC), and the Ministry of Public Security, strengthens several aspects of the external trade process. Notice 557 authorizes SAFE to place companies involved in prohibited or gray-area foreign-exchange activities on the List of Enterprises under Examination and Verification for the Authenticity of their Foreign Exchange Purchase by SAFE (the Verification List). Companies placed on the list have their import rights suspended for three months. At the end of this period, they can arrange to purchase foreign exchange from a designated bank on a case-by-case basis, subject to SAFE examination and verification of the authenticity of the purchase. Several PRC trading companies on the Verification List reportedly were able to secure removal from the list (and thus exemption from the import-rights suspension) by producing the necessary evidentiary documents to support their import activities.
The new notices also hold import-trading companies responsible for the authenticity of all relevant documents used in import transactions. Such companies are tasked with eliminating the “four owns and three untraceables” (sìzì sānbùyíjiàn), a slogan that refers to activities that camouflage illicit trading activities (including fabrication of trade relationships) to cover up fraudulent foreign-exchange transfers. The “four owns” refer to bringing in one’s own clients; supplying goods through one’s own sources; drawing up one’s own money orders; and making customs declarations on one’s own behalf. The “three untraceables” refer to cases in which authorities are unable to track the imported products, the owner of the goods supplied, or the foreign party. Import agencies that fail to implement the new notice will be placed on the Verification List and the individuals responsible punished according to the seriousness of the case.
Major violations could result in the revocation of an agency’s import rights. Serious offenses, including severely disrupting market order by using forged or altered vouchers and documents to purchase more than $5 million in foreign exchange from designated banks, could constitute a criminal violation and entail severe penalties, including imprisonment.
Finally, all agencies entrusted with enforcing compliance have had to begin sharing information with one another and with the Public Security Bureau. SAIC, in addition, has been tasked with strengthening the management of company registrations to ensure that only qualified firms conduct trading business.
...LEADS TO BOTTLENECKS AND DELAYS
Many firms have encountered delays in receiving payment because the trading companies handling their imports are under such intense scrutiny and investigation that the payment process has slowed to a standstill. In some cases, the import-trading companies have lost their ability to convert currency because they have been placed on the Verification List. In other cases, the designated banks have taken a very conservative approach to the new regulations, requesting additional documentation prior to approving remittances.
Many exporters have encountered delays, particularly over the implementation of Notice 27. According to foreign investors in Fujian Province, for example, banks do not know how to confirm the authenticity of customs declaration forms. These investors also reported that payments for imports processed through Customs in Xiamen, Fujian Province—one of the more efficient customs points—may take as long as three days, presuming all the relevant documentation, including tax and value-added tax invoices, is completely in order. If officials find anything “funny” in a payment application, processing slows to three weeks or longer. Alarmingly, customs authorities reportedly have deliberately delayed confirmations in a number of cases.
These bottlenecks have occurred principally because the agencies responsible for overseeing foreign trade matters—namely MOFTEC, SAIC, SAFE, Customs, designated foreign-exchange banks, and foreign-trade agencies—have each been ordered to crack down on illicit foreign-exchange activities or suffer the consequences. In an effort to reduce the bottlenecks, SAFE (after obtaining approval from Customs) issued in 1999 the Notice on Certain Issues Relating to Use of Networking System for Verification of Import and Export Customs Declaration Forms (Huifu [1999] No. 21). This notice establishes criteria for determining whether a customs declaration form is genuine, making it easier for various SAFE branches, designated foreign-exchange banks, and Customs itself to confirm the authenticity of import and export transactions. Whether this notice will effectively eliminate delays caused by requisite examinations of each trade transaction over $100,000 remains to be seen.
In several recent instances, the inability of the import agency to pay for a large commodity transaction prompted the PRC customer to suggest paying the exporter’s PRC subsidiary the RMB equivalent of the purchase price. Faced with the possibility of being unable to collect any money for its shipment, the exporter was thrown into a “Catch-22” situation: the company could risk losing everything, or it could accept payment onshore in RMB, and thus engage in what SAFE considers to be illegal “foreign-exchange arbitrage.” If such a transaction were to be discovered, the customer and the exporter could face penalties for violation of Article 4 of the 1996 Regulations of the People’s Republic of China on Foreign Exchange Control (Guotuoyuanling No. 211).
Despite this concern, some foreign companies that sell into the PRC have received tacit permission from SAFE to have their PRC subsidiaries or representative offices receive RMB instead of foreign currency. SAFE officials reportedly informed one company that SAFE would not prosecute such cases. In another instance, where foreign trade corporations owed the foreign seller foreign exchange, SAFE verbally permitted a foreign exporter’s bonded-zone trading company to receive RMB payments from endusers, even though the bonded-zone trading company had never taken title to any of the goods and was not a party to any of the transactions. SAFE’s official position, however, remains that such transactions constitute foreign-exchange arbitrage.
**BEWARE BONDED ZONES**
China is also cracking down on foreign-exchange transactions in bonded zones. Officially, foreign investors currently may not engage in domestic trade in China. In practice, however, foreign-invested trading companies established in certain bonded zones have been allowed to conduct domestic trade through commodity-exchange centers in these zones, even though such activities have not been set forth in the business scope written into their FIE contracts.
Many multinational corporations have set up trading companies in these zones to facilitate the sales of their products in China. Typically, companies will import goods from overseas into the bonded zones and then re-sell the goods to PRC customers located outside the zones, through the commodity-exchange centers. The centers act as import-export companies, arranging for customs clearance of the goods and the issuance of valued-added-tax invoices for a fee. In the past, FIEs engaged in this kind of trade have been able to convert their RMB income into foreign currency at designated foreign-exchange banks or through one of China's regional swap centers, which have traditionally provided a supplementary source of foreign exchange for FIEs.
But the recent tightening of control over foreign-exchange activities in China has called into question such domestic trading activities. In September 1998, SAFE issued the Circular on Relevant Issues Regarding Foreign Exchange Administration in Free Trade Zones (Huifa [1998] No. 8), which stated that enterprises located in free trade zones could only use their own foreign exchange to pay for imports and, in principle, were not allowed to purchase foreign exchange to make such payments. Trading companies operating in various bonded zones will now face problems conducting domestic trade (unless they also engage in exporting), since their income from domestic trade is in RMB and their payments for imported goods are, of necessity, in foreign currency.
Illustrating the Chinese saying, "Higher authorities have policies, localities have countermeasures" (Shang you zhengce, xia you duice), local authorities have taken steps to ease the burden SAFE Circular No. 8 will place on enterprises located in bonded zones. Shanghai SAFE's interpretation of the central-level notice maintains that warehousing-trading enterprises and export-processing enterprises located in the Waigaoqiao Bonded Free Trade Zone (see *The CBR*, September-October 1998, p.18) may still use RMB to buy foreign exchange for overseas payment, provided that they have insufficient foreign exchange of their own and they obtain approval from Shanghai SAFE. Shanghai authorities seem to be trying to reduce the burden on warehousing-trading and export-processing enterprises in Waigaoqiao, partly because they can also conduct domestic trade. The authorities may implicitly be encouraging foreign investors to set up warehousing-trading and export-processing enterprises rather than simple (gray-area) trading companies, which require less initial investment.
Nevertheless, Shanghai SAFE imposes limits on the amount of foreign exchange that can be purchased by such enterprises. In the case of warehousing-trading enterprises, Shanghai SAFE has established a foreign-exchange purchase quota. In principle, the amount of foreign exchange that can be purchased in a year may not exceed the total amount of goods imported in the previous year. An enterprise may apply to the Shanghai Waigaoqiao Free Trade Zone Administrative Commission for an increase in its foreign-exchange quota if, in one year, its imports increase by an amount greater than the sum of its own foreign-exchange holdings and its foreign-exchange purchase quota for that year. In the case of export-processing enterprises, the quota may not exceed the total value of products approved for sale on the domestic markets.
Some bonded-zone trading companies have resolved their forex problems by first selling the goods to a bonded commodity-exchange center or an import-export company for foreign exchange, and then repurchasing the goods for RMB. As domestic goods, they can then be stored and resold later. This practice enables trading companies to convert RMB into foreign exchange, but subject to an additional transaction cost. In addition, trading companies reportedly may hire a domestic import-export agency to arrange for customs clearance and convert RMB income to foreign exchange on their behalf.
Despite these options, the recent PBOC and SAFE notices have renewed skepticism about the benefits of using trading companies as vehicles to facilitate domestic trade. It remains to be seen whether the authorities will take further steps to mitigate the impact of these notices on trading companies' operations.
**PAYMENTS FOR INTELLECTUAL PROPERTY AND OTHER INTANGIBLES**
Another form of foreign-exchange payment falling under the current account is compensation for the use of intellectual property and other types of intangible property. To deal with such payment issues, in March 1998 SAFE issued the Notice Concerning Problems of Strengthening Management on Sales and Payments in Foreign Exchange for Introducing Intellectual (Intangible) Property (*Huiguanbanzi* [1998] No. 92). This notice details conditions under which payment in foreign exchange may be made for patents, trademarks, copyrights, and non-patented technology. Rigorous examination procedures are required for specific types of documentation before such payments can be approved. For approved transactions, supporting documentation will be stamped "foreign exchange already received" (yi shou bu) to prevent the re-use of such papers.
The war stories surrounding troubled intangible property agreements are many and varied. One foreign li## SOME RECENT PRC REGULATIONS ON FOREIGN-EXCHANGE TRANSACTIONS
| DATE OF ISSUE | REGULATION |
|---------------|------------|
| **IMPORTING/EXPORTING** | |
| 1/20/99 | Notice on Certain Issues Relating to Use of Networking System for Verification of Import and Export Customs Declaration Forms |
| **IMPORTING** | |
| 6/22/98 | Circular on Promulgation of Regulation on the Examination of Trade Authenticity for Verification of Import Payment in Foreign Exchange |
| 6/30/98 | Circular on Relevant Issues Regarding Strengthening the Administration of the Sale and Payment of Foreign Exchange for Import |
| 8/14/98 | Notice on Improving Management on Sales and Payment in Foreign Exchange |
| 9/1/98 | Notice on Rules of Operation for the Settlement, Sale, and Payment of Foreign Exchange |
| 9/16/98 | Supplementary Circular on Improving Management on Sales and Payment in Foreign Exchange |
| 10/8/98 | Notice on Issues of Sales and Payments of Foreign Exchange in Different Places |
| **EXPORTING** | |
| 6/22/98 | Detailed Rules for the Implementation of Administrative Measures for Verification of Export Exchange Collection |
| **CAPITAL ACCOUNT** | |
| 5/29/98 | Circular on Relevant Issues Regarding Strengthening Control Over Sale of Foreign Exchange Under the Capital Account |
| 8/20/98 | Circular on Relevant Issues Regarding the Prohibition on the Purchase of Foreign Exchange for Early Repayment of Loans |
| 8/31/98 | Urgent Notice on Strict Prohibition of Foreign Exchange Purchases for Early Repayment of Loans |
| 9/98 | Circular on Relevant Issues Regarding Strengthening the Administration of Foreign Exchange for Capital Account |
| 9/26/98 | Circular on Strengthening the Administration of the Business of Renminbi Lending Under Foreign Exchange Security by Financial Enterprises in China |
| 9/26/98 | Notice on Certain Issues Relating to Repayment of Overdue Domestic Foreign Exchange Loans |
| 1/7/99 | Notice on Certain Issues Relating to Improvement of Management of Foreign Exchange Under Capital Account Items |
| **CURRENT ACCOUNT** | |
| 3/20/98 | Notice Concerning Problems of Strengthening Management on Sales and Payments in Foreign Exchange for Introducing Intellectual (Intangible) Property |
| 9/22/98 | Notice on Issues Related to Designated Foreign Exchange Banks’ Handling of Remittance of Profits, Dividends, and Bonuses Abroad |
| **ILLEGAL ACTIVITIES** | |
| 7/1/98 | Notice Concerning Problems Related to Smashing Fraudulent Foreign Exchange Activities |
| 7/31/98 | Regulations Concerning for the Administrative Punishment of Financial Institutions and Responsible Parties According to the Administrative Regulations on Illegal Sales and Payments of Foreign Exchange |
| 8/11/98 | Notice Concerning the Improvement Plan for Rectifying the Improper Raising of Funds, Improper Approvals for Financial Institutions, and Improper Conduct of Financial Services |
| 8/28/98 | Explanation of Several Questions Concerning the Actual Use of Laws When Trying Criminal Cases of Fraudulent Purchases of Foreign Exchange or Illegal Dealing in Foreign Exchange |
| 10/9/98 | Notice Regarding the Publishing of Provisional Rules Regarding Administrative Punishment for Enterprises Involved in Foreign Exchange Activities that Circumvent Foreign Exchange Laws |
| 10/25/98 | Notice Regarding Closing Down Foreign Exchange Swap Business |
| **FREE TRADE ZONES** | |
| 9/1/98 | Circular on Relevant Issues Regarding Foreign-Exchange Administration in Free Trade Zones |
| 9/11/98 | Circular on Transmitting the “Circular on Relevant Issues Regarding the Foreign Exchange Administration in Free Trade Zones” |
| 9/29/98 | Circular on Certain Provisional Measures Regarding the Purchase of Foreign Exchange for Import by Warehousing Trading Enterprises in Waigaoqiao Free Trade Zone |
**SOURCES:** Freshfields, PricewaterhouseCoopers, The US-China Business Council
**NOTE:** This list is not meant to be comprehensive.
**ABBREVIATIONS:** SAFE, State Administration of Foreign Exchange; PBOC, People’s Bank of China; MOFTEC, Ministry of Foreign Trade and Economic Cooperation
| Chinese (Pinyin) | Number | Issuing Authority |
|---------------------------------------------------------------------------------|-----------------|-------------------|
| Guanyu shiyong jinchukou baoguandan lianwang jiancha xitong youguan wenti de tongzhi | Huifa No. 21 | SAFE |
| Guanyu xiafa "Jinkou fuhui hexiao maoyi zhensheng shenhe guding" de tongzhi | Huiguohanzi No. 199 | SAFE |
| Guanyu jiaojing jinkou shoufuhui jinguan youguan wenti de tongzhi | Huiguohanzi No. 206 | SAFE |
| Guanyu wanshan shoufuhui guanli de tongzhi | Huihan No. 27 | SAFE |
| Jiehui shouhui ji fuhui caozuo gui cheng de tongzhi | Huifa No. 9 | SAFE |
| Guanyu wanshan shou fuhui guanli de buchong tongzhi | Huifa No. 22 | SAFE |
| Guanyu daodi shoufuhui wenti de tongzhi | Yinfa No. 481 | PBOC |
| Chukou shouhui hexiao guanli banfa shishi xize | Huiguoanzi No. 22 | SAFE |
| Guanyu jiaojing ziben xiangmu jiehui guanli youguan wenti de tongzhi | Jiwaizi No. 992 | — |
| Guanyu jinzh gouhui tijian huandai youguan wenti de tongzhi | Yinchuhan No. 50 | PBOC |
| Guanyu yanjin gouhui tijian huandai de jinji tongzhi | Yinchuhan No. 53 | SAFE, PBOC |
| Guanyu jiaojing ziben xiangmu waihui guanlinban wenti de tongzhi | Huifa No. 21 | SAFE |
| Guanyu jiaojing jingnei jinrong jigou waihui danbaoxiang renminbi daikuan yongguan wenti de tongzhi | Yinfa No. 458 | PBOC |
| Guanyu changhuan guonei yuqi waihai daikuan yongguan wenti de tongzhi | Yinfa No. 468 | PBOC, SAFE |
| Guanyu wanshan ziben xiangmu waihui guanli yongguan wenti de tongzhi | Huifa No. 10 | SAFE |
| Guanyu dui jiaojing yitian wuxing zichan shoufuhui guanli youguan wenti de tongzhi | Huiguohanzi No. 92 | SAFE |
| Guanyu waihui zhiding xianxing banli lirun, guxi, hongli huchu youguan wenti de tongzhi | Huifa No. 29 | SAFE |
| Guanyu daji taogou waihui xingwei youguan wenti de tongzhi | Yinfa No. 302 | PBOC, SAFE, MOFTEC |
| Guanyu dui weifan shoufuhui guanli guiding de jintong jigou jiqi zerrenen xingzheng shufen de guiding | NA | PBOC |
| Zhengduan luanjizi luanshise jinrong jigou he luanban jinrong yewu shishi fangan de tongzhi | Guobanfa No. 126 | PBOC, State Council |
| Zuigao renmin fayuan guanyu shenli piangou waihui feifa maimai waihui xingshi anjian juti yingyong falu ruogan wenti de tongzhi | Fashi No. 20 | Supreme People's Court |
| Guanyu yinfa dui taotouchi waijingmao qiye geiyu xingzheng chufa de zhanxing guiding | Waijingmao Jicai Face No. 713 | MOFTEC |
| Guanyu tingban waihui tiaoji yewu de tongzhi | Yinfa No. 507 | PBOC, SAFE |
| Guanyu baoshuiqiu waihui guanli youguan wenti de tongzhi | Huifa No. 8 | SAFE |
| Guanyu zhuafan guojia waihui guanlij "Guanyu baoshuiqiu waihui guanli youguan wenti de tongzhi" de tongzhi | Huinyinwazi No. 121001 | SAFE |
| Guanyu Waigaojiao baoshuiqiu huowu fenba qiyu jinkou huowu gouhui de ruogan zaxing banfa de tongzhi | Huwaiguanwei No. 78 | Shanghai SAFE |
Borrowers may no longer purchase foreign exchange for loan repayment purposes in places other than where their business is registered.
censor was told that payments due under a technology-licensing agreement would not be forthcoming until the foreign party could "prove" that it owned the technology in question, despite the fact that licensing fees had been remitted without question in previous years. Likewise, an international hotel franchising company reportedly encountered difficulties receiving payment under the terms of its franchise agreement, which permitted a Sino-foreign joint-venture hotel company to use the franchiser's trademarks. The joint venture was told that it would not be able to make the remittance because it had not produced the Trademark Licence Contract Recordal Notice Certificate to prove that the license was properly recorded with SAIC's Trademark Office.
Many companies wishing to receive royalties under their license agreements are discovering they must either conclude supplementary agreements or obtain approvals or registrations that they previously believed were unnecessary. Meticulous documentation in line with the PRC intellectual property regime and forex-related legislation would be a prudent approach for foreign parties seeking to profit by bringing any kind of intellectual property to China (see *The CBR*, January-February 1999, p.20).
**Remitting Profits, Dividends, and Bonuses Abroad**
Until recently, China's forex regulations permitted FIEs to convert after-tax profits into foreign currency through designated foreign-exchange banks. The foreign exchange could then be remitted on the strength of a valid foreign-exchange registration certificate and other documents, including a written board resolution on profit distribution.
But PRC authorities have recently instituted further requirements that must be met before companies can remit foreign-exchange profits, dividends, and bonuses. SAFE's Notice on Issues Related to Designated Foreign Exchange Banks' Handling of Remittance of Profits, Dividends, and Bonuses Abroad (*Huifa* [1998] No. 29), issued in September 1998, requires that companies submit a number of documents to the designated foreign-exchange bank. These include the tax-payment certificate and the tax return (firms enjoying tax deduction or exemption should provide the relevant evidentiary documents issued by the local tax authority); the auditor's report on the profit, dividend, or bonus for the current year, issued by an accounting firm; the resolution of the board of directors concerning the distribution of the profits, dividends, or bonuses; the foreign-exchange registration certificate issued to each FIE; a capital verification report issued by an accounting firm; and other documents as required by SAFE. Parties seeking to remit funds derived from past financial years are also required to submit an audit report detailing the relevant financial conditions for the year(s) in which such profits, dividends, or bonuses occurred.
Moreover, only enterprises with registered capital that is completely paid up in accordance with the terms of the contract will be permitted to make such remittances. This stipulation is troubling given that in some cases a party may have paid up a portion of its registered capital in accordance with the terms of the contract, but then suspended contributions because the other party failed to make its contributions. In such circumstances, SAFE could prevent dividend remittance even if the enterprise was profitable and the party had made a significant investment.
According to the notice, foreign investors' remittances may be subject to a SAFE audit, even after approval by the designated foreign-exchange bank. The notice authorizes local SAFE branches to audit any remittance equal to or over $100,000 or any remittance they deem questionable. If authorities discover that a bank has failed to comply with the relevant regulations in conducting the examination, or if the remittance documentation has been falsified, the bank or the enterprise is liable for administrative and criminal penalties.
**Tighter Administrative Controls**
The central government authorities have also moved to reign in foreign-exchange activities by restricting the number and location of banks and financial institutions authorized to conduct foreign-exchange business. According to the September 1998 SAFE Circular on Relevant Issues Regarding Strengthening the Administration of Foreign Exchange for Capital Account (*Huifa* [1998] No. 21), all branches of designated foreign-exchange banks at and below the county level are no longer allowed to engage in the settlement, sale, and payment of foreign exchange for capital accounts, transfers among different foreign-exchange accounts, opening of special foreign-exchange accounts, and deposits into and withdrawals from such accounts. Similarly, SAFE branches at the county level are no longer allowed to examine and approve foreign-exchange purchases for the repayment of foreign-currency loans.
In addition, borrowers may no longer purchase foreign exchange for loan repayment purposes in places other than where their business is registered. It would appear from these changes that SAFE intends to limit all the foreign-exchange purchases of an enterprise to its place of registration so that SAFE can more easily monitor foreign-exchange activities and prevent abuses.
Alternative sources of foreign exchange—through the country's swap centers—have also been curtailed. Since 1996, FIEs have had the option of conducting the purchase, sale, and settlement of foreign exchange through either swap centers or designated foreign-exchange banks. In December 1998, however, PBOC and SAFE permanently closed China's foreign-exchange swap centers.
**Summing It Up**
While some of the recent initiatives implemented by SAFE are new, most of the measures that are now being strictly enforced have long been on the books. The government's efforts to improve the enforcement of forex controls have caused confusion for all
concerned and the exercise of great caution by those responsible for implementing the regulations. This, in turn, has resulted in delays remitting foreign-exchange payments to exporters, investors, and lenders. The PRC buyer may suffer a temporary inability to access foreign exchange and, in a worst-case scenario, may lose its foreign-trade rights entirely as a result of its transgressions.
But the brunt of the crackdown's effects has been borne by those least able to influence compliance: foreign exporters to China. Many foreign exporters have not been paid, even though they may have little influence on the documents the importer must present to the bank, the reports to customs officials, or the foreign-exchange review procedures instituted by the banks. Indeed, in light of the penalties he or she could suffer should the transaction be questioned, there is very little, if any incentive for a banker or a SAFE bureaucrat to authorize a foreign-exchange remittance. In addition, the seller may find prosecuting and enforcing a judgment for breach of contract difficult, particularly as the buyer can now argue that the failure to make timely payment is not its fault, but that of the financial system. Even if a seller obtains a judgment, SAFE's position on the purchase of forex to pay for such a judgment—or penalties associated with an underlying trade transaction—is unclear.
If, as a result of the new regulations, government functionaries and designated banks interfere with legitimate transactions and alter contractual agreements to the detriment of foreign traders and investors, China may see a rise in friction with trading partners and a reduction in foreign-investment inflows. On the other hand, foreign exporters may be able to make a stronger argument for direct trading and distribution rights if illicit activities conducted by Chinese trading companies are found to be the cause of many problems in the current trade regime. In the meantime, foreign firms can only hope that the application of the forex-related directives becomes clearer and smoother over time, enabling the PRC government to maintain the value of the RMB, permitting the expeditious conclusion of legitimate trade and investment transactions, and preventing abuses of the foreign-exchange control regime.
A year has passed since China's leadership signaled a massive reorganization of the state-owned sector, including a substantial sell-off of state-owned assets, at the 1998 National People's Congress. Although many analysts expected this move to create opportunities for foreign investors to buy those assets, so far there has not been a surge of merger and acquisition activity by foreign investors in China. Many US and other foreign investors have come to wonder whether a foreign company really can buy a state-owned enterprise (SOE), and if so, whether it can be done the same way as in the United States.
The answer to both questions appears to be a cautious and only slightly qualified yes. One American company, Eastman Kodak Co., has already bought several state-owned companies, and anecdotal evidence suggests that other US companies are in the process of acquiring SOEs in ways that are not so different from those common in the United States.
**A REGISTERED PRESENCE**
Chinese law requires a foreign entity seeking to acquire a state-owned firm to establish a registered presence in China. Until recently, foreign investors could choose an equity joint venture, a cooperative joint venture, or a wholly foreign-owned enterprise as their local corporate vehicle. As these forms were designed by Chinese lawmakers specifically with foreign investors in mind, they have significant limitations. The most serious of these drawbacks, which applies only to joint-venture forms, is the granting of inherent veto rights to the Chinese partner over important matters such as the increase of capital. This presents a major obstacle for many foreign investors struggling to expand their operations. In many cases, their Chinese partners have neither the money to keep up with capital calls nor the inclination to allow dilution of their stake in the enterprise.
These earlier forms also contemplate a single foreign-investor project, with a single Chinese partner or set of Chinese partners identified at the outset. But most foreign investors, like Kodak, want to do business in an integrated fashion throughout China, without being limited by the local preferences of any particular Chinese partner. The Chinese holding or "umbrella" company, an investment vehicle introduced in 1995, offered some
hope of achieving this, but is so hedged with restrictions that it has not come close to meeting expectations.
**The benefits of a joint-stock limited company**
In the early 1990s, the Ministry of Foreign Trade and Economic Cooperation (MOFTEC) began to issue intermittent notices on the experimental use of joint-stock companies by foreign investors. The passage of China's Company Law in 1994 and supplemental MOFTEC regulations in 1995 provided foreign companies with a more fully articulated set of rules under which to set up joint-stock limited companies. The joint-stock model has certain advantages over the other two investment forms and has been used in the acquisition of SOEs (*see The CBR*, July-August 1998, p.14).
The joint-stock limited company is perhaps destined to be the foreign investor's vehicle of choice for the new millennium. Its most important feature is that it does not force Chinese minority veto rights on the foreign investor. A joint-stock limited company also promises to be easier to take public (since any other form must first be converted to a joint-stock limited company), and its image is somewhat better than the other options, as it was not designed only for foreigners.
Of course, the use of such a company does not amount to a blank check for the foreign investor. In 1995, MOFTEC, sensitive to the potential of this new form, promulgated the Provisional Regulations on Several Issues Concerning the Establishment of Joint Stock Limited Companies with Foreign Investment. In these regulations, MOFTEC is clearly trying to impose on the joint-stock limited company most of the restrictions that apply to the standard foreign-invested enterprise. According to the regulations, if there is foreign equity participation, it must be at least 25 percent, and the percentage limitations on foreign ownership for sensitive industrial sectors are also applicable. But even taking these restrictions into account, the most important advantage offered by the joint-stock limited company remains: it can operate anywhere in China, free of veto rights for minority shareholders.
In 1998, Eastman Kodak became the first foreign company to establish corporate subsidiaries under the new Company Law. Kodak then used these two subsidiaries to acquire the assets of several insolvent SOEs in different parts of China (*see* p.36). The Kodak deal has shown, among other things, that acquisitions in China can be similar to those in the United States. The most striking similarity in Kodak's case was the choice of the form of local company to undertake the acquisitions: a joint-stock company.
**A dearth of deals**
The question remains, however, why more foreign investors haven't acquired SOEs. The obvious answer is the present difficult economic circumstances of Asia in general and China in particular, including the uncertain future of China's currency, the renminbi. But other factors also play a role. In China, many of the most attractive sectors for investment, such as telecommunications and insurance, are still closed to foreign participation. Even in sectors open to foreign investment, only insolvent SOEs are for sale.
Within the small universe of Chinese companies that are for sale, there are at least two formidable obstacles. Chinese SOEs often insist on an unrealistic valuation of assets, using the net book value for obsolete, badly maintained plant and equipment. And as a prerequisite to serious negotiations, Chinese firms often insist on dictating the number of employees the foreign investor must keep. Moreover, there is no equivalent to the US concept of maximizing shareholder value for the more successful SOEs, and their managers have no personal interest in surrendering control to outsiders, foreign or otherwise.
**The players**
As in the United States, negotiations take place primarily with the managers of the target company. But in China, parties absent from most US deals also play an important role.
Central-level authorities must be involved in any acquisition of a firm worth more than $30 million, though some industrial sectors, such as international trading and retailing, have a lower threshold. Rumors of a planned increase in the general threshold to $100 million have yet to be confirmed. In addition, MOFTEC must be involved in any acquisition that is conducted through a joint-stock limited company, no matter what the scale.
Most foreign investors interact even more with the local government and its affiliated bureaus than with the central government. Typically, the foreign investor wants to gain terms that only the local government can approve—favorable utility rates, preferential tax status, and confirmation of the absence of any significant environmental or zoning problems. For their parts, local authorities often seek to confirm the potential benefits that a foreign investor might bring, especially in the area of employment. A foreign investor is likely to want to retain some, but nowhere near all, of an SOE's employees after an acquisition. But because each employee not hired by the foreign investor can become an economic burden to the local government, employee retention is often a controversial issue in negotiations—one that can affect the pricing and other aspects of the deal.
Another participant in negotiations for the acquisition of an insolvent SOE may be the firm's principal creditor(s), who will want to ensure that most, if not all, of the acquisition proceeds are used to reduce the SOE's debt. In the majority of cases, the presence of this player at the negotiating table is a two-edged sword. On one hand, the creditor is likely to have a more sophisticated understanding than the SOE's leadership of the protections that are required by the foreign acquirer, and thus may be more apt to agree to such protections. This is especially true if the creditor has international experience, as do most of the state banks and international trust and investment corporations. On the other hand, the creditor that tries to recover too much at the expense of other stakeholders—such as laid-off workers—can slow down or even kill a deal.
**Contract specifics**
As is often the case in acquisitions in the United States, in China the doForeign investors thus may be forced to work on government approvals themselves to make sure that these matters are resolved before an acquisition agreement is signed.
mestic seller is often surprised when first confronted with all of the buyer protections in the acquisition agreement drafted by the foreign investor. These protections may include indemnities for financial statement inaccuracies and limitations on the target company's ability to do business outside its ordinary scope between the signing and closing of the deal. Thus, it is in the foreign investor's interest to keep such protections to a minimum and hand-deliver the first draft with a careful explanation of why they are necessary and reasonable.
Despite China's reputation as a place where contracts are not always strictly honored, buyers can expect that the target SOE's management will deal with its obligations in the acquisition agreement very carefully. Thus, a well thought-out and meticulously drafted agreement is essential to an acquisition in China, as contract omissions and ambiguities can cause problems later. Though judicial enforcement mechanisms in China remain weak, the contract is still the first place that Chinese parties will turn in a dispute.
Making a contract contingent on gaining certain government approvals is another important part of any acquisition, especially in China, where so many formal approvals are required. For instance, the State Development Planning Commission may have to approve a project proposal and perform a feasibility study; the State-owned Assets Bureau may have to approve the valuation of the assets being transferred; and the local real estate bureau may have to approve the transfer of land-use rights and buildings. But SOE management will resist putting such contingent terms in the contract, claiming that such governmental action is either a foregone conclusion or, alternatively, is entirely beyond their control, and so inappropriate to include in the agreement. Foreign investors thus may be forced to work on government approvals themselves to make sure that these matters are resolved before an acquisition agreement is signed.
An acquisition agreement in China must also include a section on indemnity, which requires the seller to reimburse the buyer for damages suffered as a result of seller misrepresentation about the target business. But the value of such a promise is debatable when made by an insolvent SOE selling all of its as-
---
**The Kodak Deal**
Last March, Eastman Kodak Co. announced agreements that have enabled the company to expand its manufacturing and marketing capabilities in China. Under the terms of these agreements, Kodak pledged to invest more than $1 billion in China over the next several years. The investment will be used to upgrade technology, improve manufacturing capacity, and expand distribution and marketing capabilities needed to support a strong domestic Chinese imaging industry.
As the first foreign firm to take significant advantage of a vehicle introduced under China's 1994 Company Law, Kodak has invested in two newly formed joint-stock limited companies: Kodak (China) Co. Ltd. and Kodak (Wuxi) Co. Ltd. Kodak (China) Co. is composed of the assets of two Chinese state-owned photographic enterprises, Shantou Era Photo Materials Industry Corp. and Xiamen Fuda Photographic Materials Co. Ltd. The two companies are color film and paper sensitizing plants in Guangdong and Fujian provinces respectively. A third Chinese enterprise, Wuxi A'ermei Film and Chemical Corp., a black-and-white film plant producing medical and industrial x-ray films in Jiangsu Province, transferred its primary assets to Kodak (Wuxi) Co. Ltd. Kodak, for its part, agreed to make substantial cash investments in the two joint-stock companies, Kodak (China) and Kodak (Wuxi).
Kodak owns approximately 80 percent of the shares of Kodak (China) and 70 percent of the shares of Kodak (Wuxi). Chinese investors will own the balance: Shantou Era Photo Materials and Xiamen Fuda Photographic Materials will each own roughly 10 percent of Kodak (China), while Wuxi A'ermei will hold about 30 percent of Kodak (Wuxi). To fulfill a legal requirement that joint-stock limited companies have at least five shareholders, Guangdong and Fujian international trust and investment corporations also hold nominal stakes in Kodak (China). As the controlling shareholder, Kodak will lead the management and operation of the companies, which will manufacture, market, distribute, sell, and support Kodak products throughout China.
The four-year negotiation process appears to have paid off for both sides. Kodak received assurances that no other foreign company would be permitted to build photographic factories in China for four years. The Chinese will receive the latest photographic technology. Perhaps just as important to China's top leadership as the technology transfer is Kodak's commitment to work with the SOEs to transfer key management skills. The PRC government also agreed to take responsibility for laid-off workers, and helped Kodak retain staff it wanted to keep. Initially, Kodak (China) and Kodak (Wuxi) expect to employ more than 2,000 people, who will complement Kodak's existing market support staff of nearly 600, working in liaison offices throughout China.
—Virginia A. Hulme
Virginia A. Hulme is Assistant Editor of *The CBR*.
sets. If it can be negotiated, holding a portion of the purchase payment until after the deal, possibly by paying the acquisition price in installments, helps ensure that such a promise is kept. The SOE management, ever suspicious of foreign investors' motives, can be expected to resist this strenuously, however.
**Chinese Characteristics**
While some aspects of acquisitions in China will be familiar to foreign investors, the Chinese side will likely take some unexpected positions on specific contract points. Americans take it for granted, for instance, that they are entitled to all significant information about a company that they are proposing to buy, usually matched with a set of contract representations and warranties formally confirming the information. But in China, the foreign investor may never be able to obtain some of the information it would like to see—such as a complete list of creditors—because the SOE may fear that such disclosures would threaten the deal. In an extreme case, the investor may never find out, for instance, the salaries of the SOE employees because the SOE management may fear that revealing information about relatively low salaries would give their new employer a reason to pay them less than it might otherwise.
The fate of rank-and-file SOE employees also affects negotiations in ways far from the US experience. In the United States, the buyer typically wants the seller to promise to maintain the existing labor force intact during the acquisition process. In China acquisitions, pressure is typically exerted in the opposite direction—the seller wants the buyer to contract to retain as many employees as possible and may also try to influence the choice of staff retained.
In general, negotiations associated with the acquisition of an SOE take far more time than most foreign investors anticipate. Given the sorry financial condition of many targeted SOEs, some investors might expect the negotiations to move more quickly than they actually do. The reality, however, is that the management of many SOEs continues to function relatively independently and comfortably regardless of an enterprise's mounting losses, and thus has no incentive to hurry. Lack of trust, the SOE management's fear of being taken advantage of by the foreign investor, and the participation of multiple Chinese parties all tend to make negotiations proceed much more slowly in China than elsewhere.
**Weak Legal Infrastructure**
Foreign investors in a PRC acquisition often prefer to buy an SOE's assets rather than stock to avoid taking on an SOE's liabilities, the nature and amount of which are not always easy to grasp. These may include unrecorded loan and guarantor liabilities, contingent tax liabilities arising from dubious tax-saving devices, unfunded employee welfare liabilities, defective product liabilities, and liability for the infringement of intellectual property rights. Fortunately for the buyer of SOE assets, China does not have a bulk sales law, common in Western countries. Such a law enables unsecured creditors to follow their debtor's assets into the hands of a buyer unless certain procedures are observed. Thus, in China, an unsecured creditor has no formal, clear legal claim on the original debtor's assets once ownership of those assets changes. Practical questions regarding the target SOE's prior commitments might remain, however, such as whether the local power bureau will sell the new owner electricity without having been paid by the SOE.
On the other hand, because China's Security Law dates back only to 1995, uncertainty with respect to secured creditors of an SOE also exists. It is not legally clear whether a purchaser of SOE assets can acquire those assets free and clear of unrecorded mortgages that the SOE might have granted to others before 1995. Each particular buyer should determine whether these liabilities pose substantial risk.
While the Kodak deal and other transactions currently under way indicate that SOE acquisitions are possible, foreign investors will likely wait for the Chinese government to sell SOEs that are both profitable and in more attractive sectors before using this avenue to invest in Chinese industry. Legal improvements concerning the rights and priorities of creditors and buyers of SOEs would also further facilitate such acquisitions. US investors willing to buy companies now on the block may yet acquire SOEs successfully. But such companies should be prepared for an acquisition in China that does not resemble one in the United States in many important respects.
In response to China's emerging insurance needs, the Chinese government has attempted to modernize the insurance industry during the reform era not only through development of a legal and regulatory framework, but also by expanding the number of insurers participating in the market. Domestic insurance companies, led by the giant state-owned People's Insurance Co. of China (PICC), have multiplied in recent years. Foreign insurers, on the other hand, have gained only limited access to the market.
To their frustration, Beijing has chosen to dole out operating licenses to foreign firms slowly, basing its decisions as much on political as market considerations. Of the many foreign insurers wishing to participate in China's insurance market, the few that have been licensed are companies that have met stringent formal as well as informal requirements, or have historical ties to the country—notably American International Group, Inc. (AIG), and Tokio Marine and Fire Insurance Co., both of which had a significant presence in pre-1949 China.
Despite ongoing reforms, including the creation in 1998 of an independent regulatory body, China's slow approval process is likely to continue. Nonetheless, foreign insurance companies are hopeful that their long-term commitment to China will pay off in the form of access to a rapidly expanding market.
**Back from Oblivion**
After more than 30 years of a command economy that had little need for market-style insurance, the economic reforms of the early 1980s prompted the State Council to proceed with reforms of the insurance sector. Initial moves included creating more domestic insurers; detailing types of permitted insurance products; and developing a basic legal framework for the sector. By the end of 1998, 29 Chinese and foreign insurance companies (13 Chinese companies, 6 foreign branches, 5 joint ventures, and 1 quasi-foreign Chinese company based in Hong Kong) were competing to provide a wide variety of insurance services and products to Chinese and foreign individuals, businesses, and the Chinese government. But PICC, the Ping An Insurance Co., and the China Pacific Insurance Co., the three largest domestic companies, dominate the sector.
The State Council separated PICC from the People's Bank of China (PBOC) in 1984 and brought PICC under its own authority. PICC was divided in 1996 into three independent companies: the PICC Property Insurance Co., the PICC Life Insurance Co., and the PICC Reinsurance Co. The People's Insurance (Group) Corp. (the PICC Group) was also established in 1996 to control the three companies. In October 1998, the PICC Group was disbanded, and
the three PICC companies were renamed China Life Insurance, China Property Insurance, and China Reinsurance. Before this move, the PICC Group controlled the three companies' combined registered capital of ¥20 billion ($2.4 billion) and 6,000 branch offices employing 120,000 throughout China. In 1997, Standard and Poor's Corp. ranked the PICC Group the 17th-largest insurance company in the world, based on insurance-premium income.
PICC's first competitor emerged in 1988 when the China Merchants Co. and the Industrial and Commercial Bank of China formed Ping An in Shenzhen, Guangdong Province as a shareholding company. The Insurance Division of the Shanghai branch of the Bank of Communications founded China Pacific in 1991. PICC, Ping An, and China Pacific controlled around 97 percent of the market in 1997. PICC's 1997 premium revenues alone accounted for roughly 70 percent of the total Chinese insurance market. China Pacific captured 11.9 percent of the national market—11.6 percent of the property insurance market and 12 percent of the life insurance market. And Ping An, by the end of 1997, had grabbed 15.8 percent of the Chinese domestic market, with 8 percent of the property insurance market and 22 percent of the life insurance market.
In addition to these three largest players, 10 other Chinese domestic insurance companies operate on a small scale in various regional markets. All are either shareholding companies, meaning that no state funds are invested directly, or limited-liability insurance companies (see Table 1). No Chinese insurance company is listed on any of the domestic securities exchanges.
**SMALL CLAIMS**
To conduct insurance business in China, a foreign insurance company must pass an approval process with two formal steps. First, it must gain national approval from the China Insurance Regulatory Commission (CIRC—before October 1998, such approval was sought from PBOC). Second, it must obtain an operating license from the Administration for Industry and Commerce in the city in which the company has been approved to carry out insurance underwriting. According to PBOC regulations still in effect, national approval requirements include the following formal criteria: operation of a Chinese representative office for a minimum of 2 years; total parent company assets exceeding $5 billion a year prior to the application; and 30 continuous years of experience in insurance underwriting.
A number of informal factors have influenced the national approval process as well. First, a company must demonstrate a long-term commitment to China, by, among other things, offering seminars, setting up research institutes, making financial investments in and contributions to Chinese educational institutions, and conducting programs to strengthen China's insurance-industry infrastructure. Second, Beijing appears to distribute licenses based in part on the quality of government-to-government relations between China and the company's country of origin. Finally, historical relationships between the company and the Chinese government have factored into approval decisions.
PBOC awarded AIG, in September 1992, the first license of a foreign insurance company since 1949. AIG, among the world's largest insurance companies, was founded in Shanghai in 1919 as American Asiatic Underwriters. AIG now holds three separate licenses, one to underwrite both life and property insurance in Shanghai, and one each for life and property insurance in Guangzhou. The company is the only foreign firm to have been granted a single license to underwrite both life and property insurance in China.
By the end of 1998, a total of only 9 foreign insurance companies from 8 countries had been approved to do business in China, either as branches or as joint ventures with Chinese companies (see Table 2). The eight companies other than AIG have obtained (or are in the process of obtaining) licenses to carry out insurance underwriting only in Shanghai. With the exception of the United States, just one insurance company from each major industrial country has obtained a license. Two US companies together hold four licenses—Aetna Inc. has one license in addition to AIG's three. Nevertheless, in hopes of eventually breaking into the market, the 113 foreign insurance companies with a presence in China, including those that already have licenses, have opened a total of 202 representative offices in major Chinese cities.
Though the licensing process is a slow one, with the number of licenses granted limited to 2 to 3 per year since 1995, some foreign insurance companies have chosen to enter China through other available forms. A quasi-foreign company owned by PICC and the Min An Insurance Co. of Hong Kong has underwritten insurance in Shenzhen and Hainan Province since 1982. Foreign reinsurance companies have been able to provide Chinese commercial projects with reinsurance without a license. In 1997, foreign reinsurers captured about 50 percent of China's reinsurance market.
After establishing a non-business presence in the 1980s, in 1998 the British Sedgwick Group obtained a trial license to conduct an insurance brokerage business. More than 12 other foreign brokerage companies have representative offices in China and are seeking licenses to provide insurance brokerage services.
Some foreign firms have taken equity positions in Chinese companies as a way to participate in the PRC insurance sector. Goldman, Sachs & Co. and Morgan Stanley Dean Witter each own 6 percent of Ping An, and Japan First Life and Lincoln National Corp. of the United States have each been approved to purchase up to 5 percent of Ping An.
**HAMSTRUNG IN THE MARKET**
The continued dominance of Chinese companies in Shanghai and the exclusion of foreign companies from most other Chinese cities has so far left little room for the few foreign insurance players. By 1997, foreign insurers had captured only 8 percent of
Foreign market share has suffered because of Chinese competitors' rapid learning curves.
the life insurance market in Shanghai and less than 1 percent of China's total insurance needs. According to the China Insurance Institute's *Insurance Research* (Vol. 9, 1998), by 1997 Tokio Marine and Winterthur Schweizerische Versicherungs Gesellschaft had captured only 0.46 percent and 0.16 percent, respectively, of the property insurance premiums in Shanghai. Manulife Financial's joint venture, similarly, had only gained 0.12 percent of Shanghai's life insurance market.
In part, limited foreign participation is a result of the web of rules that constrain licensed foreign insurance companies. A license permits a company to set up either a wholly foreign-owned branch or take up to 51 percent ownership in a Sino-foreign joint-venture company. Other rules include limits on business scope. For instance, foreign insurers in China may sell liability insurance only to foreign-invested companies, and may sell life insurance only to foreigners or individual Chinese in China. In addition, all insurance underwriting has been limited to Shanghai and Guangzhou, and the premiums collected by foreign insurance companies may be invested in only a limited number of instruments. Foreign companies may deposit premiums into accounts in Chinese banks; may buy Chinese government bonds or corporate bonds, but only up to 10 percent of the company's total investment portfolio; and may offer foreign-exchange trust loans or make equity investments in Chinese stocks, but only up to an additional 15 percent of their total investment portfolio.
Other limits on the activities of foreign insurers include prohibitions against Chinese citizens acting as general managers of foreign insurance operations in China. And government officials limit, on a case-by-case basis, the proportion of Chinese citizens that may sit on a foreign insurance venture's board of directors. All insurance companies are also required to have 30 percent of their insurance policies reinsured by a Chinese domestic company (for China's domestic companies this requirement is 20 percent).
Joint-venture insurance companies must deposit 20 percent of their net receipts into a guarantee fund, overseen by CIRC, to protect the insured. Foreign subsidiary life insurance companies in China are required to deposit $4 million into a guarantee fund. Foreign subsidiary companies selling both life and property insurance in China must place $8 million in the fund. Foreign joint-venture insurance companies are additionally required to put 25 percent of after-tax profits into a cumulative reserve fund.
Another reason for the modest size of foreign insurance companies' market shares in China is that foreign companies are not licensed to sell group life insurance. In Shanghai alone, for example, the total group insurance market in 1996 was ¥1.86 billion ($224.6 million) out of ¥2.86 billion ($345.4 million), or 65 percent of Shanghai's life insurance market.
Foreign market share has further suffered because of Chinese competitors' rapid learning curves. After receiving its license to sell individual life insurance in Shanghai, for instance, AIG began an aggressive door-to-door sales campaign, employing about 10,000 sales agents at its peak. By the end of 1995, AIG had reportedly created and captured the bulk of Shanghai's new individual life insurance policy market. But PICC and the other Chinese companies soon followed suit with their own door-to-door campaigns. By the end of 1996, although the Shanghai life insurance market had expanded overall, Chinese companies had seized much of the new market share. By 1997, AIG's share of Shanghai's life insurance premiums had fallen, according to *Insurance Research*.
### Table 1
**China's Top Insurers**
| Company | Base of Operations | Type of Insurance Offered | Geographic/Operational Scope |
|---------|--------------------|---------------------------|------------------------------|
| China Life Insurance (Formerly PICC Life Insurance Co.) | Beijing | Life | National |
| China Property Insurance (Formerly PICC Property Insurance Co.) | Beijing | Property | National |
| China Reinsurance (Formerly PICC Reinsurance Insurance Co.) | Beijing | Reinsurance | National |
| China Pacific Insurance Co. | Shanghai | Comprehensive | National |
| Ping An Insurance Co. | Shenzhen | Comprehensive | National |
| Tian An Insurance Co. | Shanghai | Property | Shanghai |
| Dazhong Insurance Co. | Shanghai | Property | Shanghai |
| Huan An Insurance Co. | Shenzhen | Property | Shenzhen |
| Xin Hua Life Insurance Co. | Beijing | Life | National |
| Taikang Life Insurance Co. | Beijing | Life | Beijing, Guangzhou, Wuhan |
| Huatai Life Insurance Co. | Beijing | Property | Beijing, Nanjing, Shanghai |
| Yong An Property Insurance Co. | Xi'an | Property | Xi'an |
| Xinjiang Army Group Insurance Co. | Urumqi | Property | Xinjiang Production & Construction Group, People's Liberation Army |
Source: Ji Chen and Stephen C. Thomas
Where the world comes together in Beijing.
It must be Shangri-La
China World Hotel, Beijing. Tel: (86 10) 6505 2266 Fax: (86 10) 6505 31678.
For reservations, call your travel consultant or USA & Canada toll-free: 1 800 942.5050. Internet: www.Shangri-La.com
BALI • BANGKOK • BEIHAI • BEIJING • CEBU • CHANGCHUN • DALIAN • FIJI • HANGZHOU • HARBIN (EARLY 1999) • HONG KONG • JAKARTA • KOTA KINABALU • KUALA LUMPUR
MANILA • PENANG • QINGDAO • SHANGHAI • SHENYANG • SHENZHEN • SINGAPORE • SURABAYA • TAIPEI • VANCOUVER • WUHAN (EARLY 1999) • XIAN • YANGON
Chinese insurance companies lack expertise in the reinsurance market.
**Market Potential**
Like insurance industries in other developing countries, China's insurance sector once emphasized business insurance such as marine, fire, commodity, transportation, and casualty. Recently, however, the growth of a Chinese middle class, particularly in Shanghai and other booming coastal cities, has raised demand for life and personal property insurance. Ongoing market reforms have also created a need for types of insurance new to the People's Republic. These include social insurance schemes such as health, disability, and retirement; property insurance for houses, cars, and personal possessions; and insurance for industrial assets, business operations, construction, and infrastructure.
China permits insurance companies to offer either of two broad categories of insurance in China: property insurance and life (group and individual) insurance. The major categories of property insurance include business property, family property, cargo transportation, motor vehicle, ship, and agricultural insurance. Life insurance instruments include "simple" life insurance (similar to term life), life insurance with savings features; group life insurance; "simple" and group disability insurance; group accident insurance; highway accident insurance (such as for bus passengers); collective-enterprise employee retirement insurance; individual pension insurance; children's education and marriage insurance; major medical insurance; and other combination products.
Chinese insurance companies also offer some insurance products to enterprises doing business outside of the PRC and to foreign-related businesses in China. These products include export-goods transportation insurance, shipping insurance, air-freight insurance, automobile insurance for foreign agencies and joint ventures in China, property insurance for foreign and joint-venture companies in China, project insurance, and machinery-damage insurance for foreign and joint ventures. Various kinds of liability insurance are also available, including product-liability insurance, investment insurance, shipbuilding insurance, offshore oil-exploration insurance, nuclear power-station insurance, and satellite insurance. In China, about 50 percent of exported goods and 90 percent of imported goods are insured.
China's reinsurance sector premiums were about ¥9.6 billion ($1.2 billion) in 1997. Most of these policies were the result of PRC insurance regulatory requirements. Chinese and foreign companies involved with infrastructure and commercial projects (such as satellite launches, airplanes, and nuclear power stations), however, have purchased reinsurance mostly from foreign insurance companies outside of China, because of a lack of Chinese experience and expertise in this sector.
In a reflection of huge demand for all types of insurance, premium in-
---
**Table 2**
**Foreign Insurance Companies in China**
| Company | Country Base | Year of Approval | Type of Insurance Offered | Venture Details | Geographic Scope |
|----------------------------------------------|--------------|------------------|----------------------------|--------------------------------------------------------------------------------|------------------|
| Min An Insurance Co. | Hong Kong | 1982 | Property | Branch | Shenzhen, Hainan |
| American International Group, Inc. | United States| 1992 | Life & Property | Branch | Shanghai |
| Tokio Marine & Fire Insurance Co | Japan | 1994 | Property | Branch | Shanghai |
| American International Group, Inc. | United States| 1995 | Life | Branch | Guangzhou |
| American International Group, Inc. | United States| 1995 | Property | Branch | Guangzhou |
| Manulife Financial | Canada | 1996 | Life | Joint venture (JV) with Sinochem. Foreign share: 51% | Shanghai |
| Winterthur Schweizerische Versicherungs Gesellschaft | Switzerland | 1996 | Property | Branch | Shanghai |
| Allianz AG | Germany | 1997 | Life | JV with Dazhong Insurance. Foreign share: 51% | Shanghai |
| Aetna Inc. | United States| 1997 | Life | JV with China Pacific Insurance. Foreign share: 50% | Shanghai |
| Royal & Sun Alliance Insurance Group Plc. | United Kingdom| 1998 | Property | Branch | Shanghai |
| Colonial Mutual Life Insurance | Australia | 1998 | Life | JV with China Life. Foreign share: 49% | Shanghai |
| AXA-UAP | France | 1998 | Life | JV with China Metals & Mining Corp. Foreign share: 51% | Shanghai |
Sources: Ji Chen and Stephen C. Thomas; US Government, International Market Insight, August, 1998; The US-China Business Council
come has grown rapidly since the launch of PICC in the early 1980s. In 1991, insurance premiums had reached ¥23.4 billion ($282 million), and by 1997, premium income had climbed to ¥108 billion ($13 billion). By 1997 the structure of the industry had also changed. Life insurance premiums made up ¥60 billion ($7.2 billion), or 56 percent of the total ¥108 billion collected, surpassing property insurance premiums for the first time. In major Chinese cities, the life insurance share of the total market was even higher; in Beijing, it was 70.7 percent, in Shanghai, 64.8 percent, and in Guangzhou, 60.5 percent. In 1997, the combined insurance indemnities paid out by all companies were ¥27.6 billion ($3.3 billion), of which ¥17.6 billion ($2.1 billion) was for life insurance.
Between 1991 and 1997, the total annual growth in premium income was an impressive 29 percent, more than double China's GDP growth rate of about 12 percent. Although the World Bank estimated that total revenue would increase to as much as ¥240 billion ($28.9 billion) by 2000, an annual increase of about 40 percent, the latest estimate for 1998 reached only ¥124 billion ($15 billion), probably because of the effects of the Asia financial crisis and a weak domestic economy.
Despite the impressive growth of China's insurance market to date, however, per capita premiums still have significant potential. In 1996, Chinese per capita spending had reached only $7.64, according to a March 1998 China Economic Information Daily report. While higher than India's $1.70 per capita, China's spending is far below that of Thailand, at $30, according to a May 1998 China Daily article, and of course substantially below more developed countries such as Great Britain ($1,775) and the United States ($2,191).
**LEGAL CATCH-UP**
The rapid expansion of China's insurance industry in the 1980s pushed the Chinese government to put tremendous effort into establishing laws and regulations to manage the insurance market and make it more competitive. In 1985, the State Council issued the Provisional Stipulations of Insurance Enterprises Administration, the sole regulation guiding the industry until the passage of the 1995 Insurance Law. To regulate the experimental opening of the Shanghai insurance market, in September 1992 PBOC issued the Shanghai Administrative Measures for Foreign-Invested Companies (the Shanghai Measures). Some of the measures also applied to Guangzhou, and could be applied to other Chinese cities in the future. The Standing Committee of the 8th National People's Congress passed the Insurance Law of the People's Republic of China in June 1995—the first insurance law in PRC history. The law spelled out the basic legal structure of insurance companies, insurance contracts, and rules guiding company operations. The Insurance Law also demanded that property and life insurance operations be separated, leading to the restructuring of PICC.
Other regulations include PBOC's July 1996 Provisional Regulations for Administration of Insurance to supplement the insurance supervision section in the 1995 Insurance Law. In 1997 and 1998, partly in response to the growing number of door-to-door sales agents in a number of industries, new regulations were enacted to control the activities of insurance agents and brokers. PBOC issued the Provisional Regulation for Administration of Insurance Agents in November 1997 and the Provisional Regulation for Administration of Insurance Brokers in February 1998. Also, a new state-owned specialized insurance company is being developed to meet the growing insurance needs of the agricultural sector.
The 1995 Insurance Law designated the Insurance Department of PBOC as China's insurance industry regulator. The Insurance Department became separate from the Non-bank Financial Institution Administration Department of the PBOC in May 1995. The Insurance Department was in charge of overseeing all insurance companies in China, foreign and domestic, and approving all insurance licenses. Foreign insurance companies were subject to the regulatory supervision of both the Insurance Department and the Foreign Financial Institution Administration Department of PBOC. (The field examination of insurance companies and the insurance industry in general was conducted PBOC's Auditing Department.)
In November 1998, however, the PRC State Council set up CIRC, a semi-autonomous insurance regulatory body, to take over from PBOC all insurance regulatory responsibilities. PICC Group head Ma Yongwei is CIRC head. Overall, three of the five chairmen and vice chairmen of CIRC came from the PICC Group. CIRC will help meet demands for tougher regulation of the industry and is to help shield it from the effects of the Asian financial crisis. CIRC is also expected to be more efficient and to give its full attention to the insurance industry. PBOC, as the nation's central bank, has numerous other responsibilities. In Shanghai, both Chinese and foreign insurance enterprises are still regulated by the Non-bank Financial Institution Administration Division under PBOC's Shanghai branch. Under this division, there is an insurance section overseeing the Shanghai insurance market.
Self-regulation by the insurance industry is developing rapidly as well. In February 1994, the Shanghai branches of PICC, Ping An, China Pacific, and AIG formed China's first insurance industry association to coordinate and self-regulate the insurance market in Shanghai. Among its features are subcommittees on life insurance, non-life insurance, and public relations. The Tokio Marine Shanghai branch, Tian An Insurance, and Dazhong Insurance have also since joined the association. By the end of 1997, insurance companies had established industry associations in 19 cities and provinces. A national insurance industry association is currently awaiting official approval.
**INDUSTRY LONGEVITY**
China's economic reforms will continue to stimulate demand for various kinds of insurance services for the countryside, for businesses, and for individual Chinese citizens. Another
A major obstacle to insurance-industry growth is the limitation on the forms premium investments may take.
stimulus, particularly for life insurance, is China's national savings rate of 30 percent. Between July 1997 and July 1998, savings expanded 18 percent, reaching $5 trillion ($604 billion). Most of this huge amount is in bank accounts. Life insurance policies with savings features provide alternative savings instruments and are proving much more popular than traditional death-benefit policies.
Other elements of the PRC economy should expand future demand for insurance products and services. For instance, the country's ongoing housing reforms will create a vast new market for homeowner's insurance. Because the reforms are eliminating state provision of rent subsidies, millions of Chinese will opt for homeownership, and therefore will need mortgage and property insurance. Similarly, growing private automobile ownership will require a major expansion in auto insurance provision.
The loss of the traditional "iron rice bowl" has already left millions of state employees without government-provided health, education, and retirement benefits, which can be replaced by private insurance coverage. The 60 percent of Chinese living in lagging economic areas will also eventually need the insurance products currently becoming standard in China's more advanced coastal cities. Moreover, 70 percent of Chinese people still live in rural areas. Increasingly successful farmers will want insurance for their crops, machinery, and their lives. On the commercial side, state-owned-enterprise reforms have created a demand for commercial risk insurance.
At the same time, a number of obstacles could prevent untrammeled growth in China's insurance industry. First is the limitation on the forms insurance premium investments in China may take: with six recent interest rate cuts, insurers are earning less income than ever, despite having to pay the high returns guaranteed in many policies. Stocks and corporate bonds are still too risky, and real-estate income is unpredictable. But such corporate investments typically yield higher returns than bank deposits and government bonds, and are the mainstays of insurance company investments in industrialized countries. Second, the Chinese public and insurance industry lack knowledge of basic insurance concepts. A final obstacle is fierce market competition, which has forced some Chinese insurance companies to pay as much as 80 percent of premiums in commissions, reducing profit margins and potentially exposing these companies to future losses.
FEW ASSURANCES FOR FOREIGNERS
The future for foreign involvement in China's insurance market remains cloudy. Some foreign experts hold that with the establishment of CIRC, approval for new licenses will accelerate and new cities in China will open to foreign insurers. They further argue that should China accede to the World Trade Organization, WTO requirements will force China to continue opening its financial services sector to foreign companies.
Other foreign observers, however, believe that the pace of approval and licensing of new foreign insurance firms will remain slow. They point out a number of reasons why, even with WTO pressure, China may not accelerate liberalization of its insurance market anytime soon. First, the development of insurance industry rules and effective enforcement mechanisms is far from complete. Second, Chinese insurance companies still argue for the need for protection from foreign firms because of their relative inexperience. And third, the recent Asian financial crisis has lent credence to the PRC's policy of cautious opening of the financial services sector.
If past experience is any indication, foreign companies will likely continue to be restricted to forming life insurance joint ventures, while property insurance companies will continue to be permitted to form branches. Earlier PBOC statements seemed to indicate that China would speed up the granting of licenses and the opening of new cities. Current policy, however, clearly states that there will be no more licenses granted for wholly foreign-owned life insurance branches, either in Shanghai or in any other city. Joint ventures will continue to receive licenses, but probably at a slower rate, as this investment form is not a priority for the Chinese government.
Instead, priority will be given to foreign equity participation in Chinese-owned life insurance companies. Equity ownership is to be limited to 5 percent per foreign insurance company, up to a total of 25 percent by foreign investors in any one Chinese life insurance company. Foreign investors thus might want to consider equity investments in domestic companies such as Xin Hua Life Insurance Co., a Beijing-based, national insurer. In 1997, Xin Hua received $170 million in life insurance premiums, which accounted for roughly 22 percent of the Beijing life insurance market. Xin Hua ranked second in Beijing after the Beijing branch of Ping An, which accounted for 30 percent of the Beijing life insurance market. Other than such equity investments in these rising domestic firms, however, foreign companies face a slow entry process.
China's insurance market will continue to expand to meet the demands of a rapidly growing economy. Foreign participation in this market will continue to rise, but on a modest scale. Any major deviations from this pattern will come only from either drastic changes in domestic economic conditions and policies or in the international environment, neither of which seems likely to occur in the near future.
MOFTEC, Hong Kong Tenders, and Life in Shanghai
http://www.china2thou.com/ China 2000 is a Washington-based monthly newsletter. It primarily covers US-China relations, but also carries articles on political, economic, and social issues. Although many of the articles are written by senior Chinese officials and academics, the newsletter examines issues from both the US and Chinese points of view. Visitors to this free site may access the full text of all current and archived articles.
http://www.cnie.org/nle/update/11-30.htm This site, run by the Committee for the National Institute for the Environment, lists Congressional Research Service reports, including a recent report on US-China trade issues. Visitors may access the full text of the report for free.
http://www.moftec.gov.cn/ China's Ministry of Foreign Trade and Economic Cooperation operates this site, which contains official policies, trade statistics, the full text of relevant Chinese laws and regulations, and contact information for each department. It also has links to Chinese companies, exhibitions and trade fairs, and bidding notices.
http://www.cpo.cn.net/english.htm The official website of the PRC's Patent Office features the full text of China's 1984 patent law in English. Visitors may find out more about the application process or read the Patent Office's 1997 annual report. The site is also available in Chinese.
http://www.info.gov.hk/gsd/tender.htm At the website of the Hong Kong Government Supplies Department Procurement Division, visitors can read about procurement procedures, how to submit a tender, and other relevant information. Tender notices are updated weekly.
http://www.insurnet.com.hk/ This free site, established by Hong Kong Insurance Network, provides extensive information on the insurance industry in Hong Kong, including information on products and companies, as well as links to the Office of the Commissioner of Insurance and the Hong Kong Federation of Insurers. The market focus link leads to news and statistics dating from 1993 on Hong Kong's insurance industry.
http://www.mckinseyquarterly.com/ Visitors to McKinsey & Co.'s free site may either search or browse categories arranged by sector or country to read summaries or the full text of articles published since 1994.
https://www.defenselink.mil/pubs/easr98/ The website of the US Department of Defense contains the full text of *The United States Security Strategy for the East Asia-Pacific Region 1998*. The document, which contains several sections on China, outlines the key concepts behind the US presence in the region and discusses security and strategy issues for the next century. Visitors may view the document online or download it in Adobe Acrobat format.
http://freenet.buffalo.edu/~cb863/china.html Finding News About China, maintained by an MBA student at SUNY Buffalo, features links to sites about Greater China. Unlike the overwhelming lists of links found on other sites, here the links are organized into clear categories, such as government, politics, business, stock exchanges, and economic information, each with a brief, helpful description of the site.
http://sun.sino.uni-heidelberg.de/igcs/ The Internet Guide for China Studies, or the China WWW Virtual Library, is an excellent site for links to China resources worldwide. Visitors may search or follow links to institutions, libraries, booksellers and publishers, online journals, newsmedia, politics, business and economy, society, history, the Internet, search engines, and other topics. The business and economy section has links to sites on exchange rates, risk, stock reports, companies, and general information for China, Taiwan, Hong Kong, Macao, and Singapore. This free site is maintained by the University of Heidelberg's Institute of Chinese Studies.
http://www.stern.nyu.edu/~nroubini/asia/AsiaHomepage.html NYU Professor Nouriel Roubini's Asia Crisis Homepage is loaded with information on the causes and effects of and possible solutions to the crisis and contains hundreds of links to academic papers and debates, country analyses, financial institutions, and related websites. Of particular interest to China hands is the free site's section entitled "Will China Be The Next Domino?," which links to articles about the effects of the Asian crisis on China.
http://www.shanghai-ed.com/ Dubbing itself the complete guide to life and business in China's greatest city, Shanghai-ed offers information on everything from nightlife and fortune telling to real estate and advertising in Shanghai. The business section contains general information on stock markets, investment, chambers of commerce, conferences and exhibitions, and real estate. This free site, maintained by Eastern Web Services, is definitely worth a stop for anyone traveling to or living in Shanghai.
http://www.searchenginecolossus.com/china.html Search Engine Colossus is a site with links to search engines organized by country, including China and Taiwan. The links show where the engine is based, the language(s) it uses, and any focus it may have, such as business in Hong Kong.
http://www.china.org.cn/bjreview/ At this free site, visitors can access articles from the past year of *Beijing Review*, an official English-language publication on China's economic and social development.
http://www.ceibs.edu/ The website of Shanghai-based China Europe International Business School (CEIBS) provides information on all of its training programs. The library page offers links to sites featuring business news, company directories, economics and statistics, and industry information. Other areas of the CEIBS site contain links to business service companies, stock markets, and mutual funds.
http://www.china-labour.org.hk/ At this site, visitors can find the *China Labour Bulletin* online, in both English and Chinese. The bulletin contains interviews, articles, and news briefs on China's labor and unemployment situation.
—Virginia A. Hulme
No More Excuses
Edward E. Lebman and Brinton M. Scott
Foreign firms in China are cracking down on unpaid debt
With the Asian slowdown, the sudden failure of the Guangdong International Trust and Investment Corp. (GITIC), and stricter regulations on foreign-exchange transactions, foreign firms in China are paying more attention to accounts receivable than ever before. It is no longer the case that foreign companies in China are reluctant to collect bad debt for fear of ruining their relationships and reputation within the business community. Today, a China-based creditor's determination to collect debt sends a strong message that would-be deadbeats, as well as those who like to string out payments over long periods, can no longer get away with non-payment.
The Need for a Big Stick
Though many financial officers of foreign-invested enterprises (FIEs) believe debt collection to be illegal, it is, in fact, permitted in China—but only by PRC law firms. The Ministry of Public Security and the State Administration for Industry and Commerce in 1994 issued the Regulation Concerning Establishing Debt Collection Companies, after a number of unscrupulous companies and individuals employed heavy-handed tactics to collect debt. The regulation forbids any person, organization, or entity—except Chinese law firms—from engaging in debt collection.
After selecting a law firm, the creditor will want to start the debt-collection process as quickly as possible. Since debt cases usually have two-year statutes of limitations, any delay can foreclose recovery. In addition, if a debtor is not paying one party, it probably is not paying others. Thus, collection becomes a race to the bottom, with secured creditors first in line.
The creditor must give the Chinese law firm power of attorney to manage the debt, execute an escrow agreement with both the debtor and the Chinese law firm, and sign a retainer agreement with the firm outlining the services it will render. The creditor should then execute an affidavit of authority, verifying that the person executing the power of attorney has the creditor's authority to act on behalf of the company. The law firm will need copies of original contracts and other related documentation to commence these types of settlements. In a contested debt collection, the Chinese law firm will also need copies of the creditor's business license and all other related documentation.
Depending on the type of debt, the law firm generally begins the process of collecting overdue receivables by conducting due diligence on the debtor. The law firm's investigative services division should perform the due diligence, to determine if the debtor is able to pay and whether the money can be recovered. A law firm's due-diligence fee can be incorporated into the total collected if the creditor decides to proceed with the collection.
The next step is to have the law firm wield a big stick—by sending dunning letters, making phone calls, knocking on doors, and, if necessary, filing lawsuits. It
can also obtain and enforce judgments, which may involve seizing property, freezing bank accounts, or garnishing wages. A few well-written letters and faxes, combined with attorney telephone calls, can frequently work wonders on a bad debtor's attitude, as fear of a lawsuit is often sufficient to motivate quick settlement. If a debtor still refuses to pay, however, the creditor must be ready to sue. Suing bad debtors acts as a signal to other businesses that the creditor is neither weak nor afraid to enforce its rights in China.
**INS AND OUTS OF LITIGATING IN CHINA**
Contrary to popular belief, litigation in China, especially for debt-collection cases, can be relatively simple, expedient, inexpensive, and transparent. Many members of the judiciary are both well-educated and fair. Judgments can sometimes be obtained within one month of filing. In Beijing, for example, the courts have made great efforts to clear debt-collection cases from their dockets, indicating a willingness to resolve bad-debt issues quickly, and most cases are decided within about six months of filing. The litigation process in China is also not as expensive as many believe. Generally, court fees are based on a percentage of the amount in dispute. The court filing fee is less than 2 percent for claims of ¥1,000,000 ($120,806) or more. Attorney fees typically amount to 30-45 percent of the amount recovered, depending on the extent of litigation, effort expended, and expenses incurred.
The creditor initiates a suit by submitting a complaint to the relevant court, which then schedules an initial hearing, usually within one month of accepting the case. Between the court's acceptance of the case and the hearing, the defendant must answer the complaint. At the initial hearing, the court hears the evidence and attempts to mediate a settlement. If mediation fails, the judiciary schedules additional hearings. If mediation fails again, the court decides the issue. Armed with the court's judgment, the creditor may begin to collect what is owed. If payment is still not forthcoming, the creditor may seize bank accounts or other forms of property. Each of the judicial organs authorized to handle debt-collection cases has a special security division that enforces judgments when necessary.
Though cases are generally resolved satisfactorily, the administrative aspects of litigation in China may be more cumbersome than in other countries. For example, the creditor's legal representative must sign and seal every court document with the company seal.
**FRIENDLY COLLECTIONS**
Debt collection in China is not always a hostile situation. Sometimes, the debtor wishes to pay, but is unable to do so for a variety of reasons. For example, State Administration of Foreign Exchange (SAFE) regulations and tighter banking and customs policies have left many Chinese importers and distributors unable to open letters of credit, though they often have *renminbi* (RMB) available (*see p.26*). In these instances, an action involves debt management rather than adversarial proceedings. Uncontested debt collection usually takes about three months.
Under such circumstances, the creditor should work with its Chinese law firm to establish a debt-management agreement that gives the PRC law firm authority to receive RMB into an escrow account on the creditor's behalf. Once the law firm collects the amounts owed in RMB, it holds the money in escrow. Next, the law firm should file the debt-management agreement with the relevant Chinese judicial authorities, and then with the banking authorities. Upon receiving approval from these authorities, it can wire the money out of China at the exchange rate in effect on the day the money is transferred.
As long as the law firm receives payment by check or direct draft, its fees are generally based on a percentage of the total amount, plus expenses. If payment is in cash, there is an additional 3 percent charge to cover the 3 percent that all Chinese banks charge for depositing cash. Most law firms prefer to receive demand drafts, cashier's checks, or wire transfers, since counterfeit currency is a problem in China. The exchange rate the creditor applies to its debts should not affect the law firm's fees.
If the debtor pays by direct draft in China, the transfer to the Chinese law firm's account generally takes two days within China. Three more working days are required to transfer the money to Hong Kong from the law firm's account. If the debtor pays by cashier's check, the clearance and transfer within China can take up to five days. While transfer is immediate if the debtor pays in cash, the law firm must verify the authenticity of these notes with the bank. Verification can be done immediately, however.
**CONTESTED DEBT COLLECTIONS**
Under a contested-debt scenario, the collection process becomes more complicated and drawn-out. The receipt process and movement of funds for contested debt collection are the same as in a friendly collection.
Several different methods of payment are possible, including direct payment to the creditor. Under a direct-payment arrangement, the Chinese lawyers would most likely require the creditor to notify them of any payments received. All funds should be paid into the law firm's escrow account so that it may manage the account for the creditor. Unfortunately, when debtors pay directly to the creditor, the law firm finds itself in the sometimes awkward situation of having to ask the creditor for payment.
When the debt is contested, a creditor has the option of either paying the Chinese law firm a percentage of the total collected—or paying by the hour. Under a contingency, the percentage paid to the law firm is generally determined by the stage of the collection process. The exact amount is determined on a case-by-case basis and may be negotiated between the law firm and creditor.
If the debt is collected before litigation, the law firm receives around 33 percent of the amount collected. If, on the other hand, the collection goes to litigation, the Chinese law firm should
Adopting debt-management and collection procedures can help firms minimize both their occurrence and the costs of resolving them.
receive 40 percent of the amount collected. Finally, appeals of judicial decisions typically drive the law firm's fee up to 45 percent. Alternatively, the creditor may decide to pay the law firm's regular hourly rates.
**PRO-ACTIVE DEBT MANAGEMENT**
But creditors need not wait for debt-collection problems to surface before acting. They can and should anticipate such scenarios and plan accordingly. Adopting debt-management and -collection procedures can help firms minimize both their occurrence and the costs of resolving them.
PRC lawyers can coordinate with a firm's accounts receivable department to arrange for the law firm to collect debt in arrears. The company then drafts sales contracts stating that debts exceeding a set period of time (for example, 60 days) will automatically be assumed by a law firm that will thereafter take action and settle the accounts. Such an arrangement insulates the creditor from engendering clients' resentment, since debt management will appear to be beyond the creditor's control. It can also serve to rein in sales staff activities, as sales staffs of firms in China have been known to sign contracts recklessly to push numbers up. Sales staff should be informed that after 60 days the debt will be turned over to a law firm, not the in-house accounting department. Furthermore, entities that contract with China creditors will not have grounds for complaint if they have been made aware in the initial contract that they must fulfill the contract or risk collection by a law firm. This option, termed "fictional factoring," thus helps those operating businesses in China to maintain good client relations.
On a case-by-case basis, certain law firms with the financial wherewithal are purchasing the debt of creditors—generally FIEs operating in China—for a set price, usually a percentage of the total. This method, called factoring, can be a good option for some law firms for two reasons. The regular payments that come from the collection of non-disputed debt can be a good source of guaranteed cash flow. Other law firms benefit from factoring when a creditor is preparing to write off a debt—or the debt is beyond or nearing the statute of limitations—and the creditor is prepared to sell it for cents on the dollar. Creditors that choose this option tend to be companies located outside of China that deal with PRC companies, perhaps the parent of an FIE.
**SELECTING A CHINESE LAW FIRM**
With so much of the process dependent on the Chinese law firm, perhaps the greatest determinant of success is selecting the right law firm for the job. Before making a decision, FIEs looking for a PRC law firm should carefully consider the qualities and abilities of a potential firm, particularly:
- **Mobility** Does the law firm have the ability to visit the debtor in person and discuss the debt face-to-face? Companies should find out whether the firm has other offices around China, and if so, where they are located.
- **Investigative services** A good Chinese law firm that engages in debt collection should have an investigative services division with employees qualified to conduct background checks and thorough due diligence.
- **Experience with debt collection** Companies should also speak to the lawyers about past debt collections. Many Chinese law firms have engaged in litigation (which happens to be a debt collection matter) but have not actually created a separate area of practice dedicated to debt collection. Though Chinese attorneys are bound by attorney-client privilege, they may (with a client's permission) disclose some information about their experience in this area.
- **Automation** Companies should look for a firm whose entire pro-active debt collection system is well-structured and computerized. The firm must also have telephone capabilities commensurate with the volume of business that debt collection creates. Ideally, the law firm should have a 24-hour answering service, especially if it deals with cross-border trade issues and creditors who may be located in different time zones.
- **Communication** For most FIEs, the working language is English. Certainly, the reporting language is English, and the legal procedures of debt collection typically must be explained to the client in English. Thus, the Chinese law firm should be able to communicate in English with the creditor, as well as in Chinese with the debtor. Further, the firm must have a solid grasp of international business, including accounting principles.
- **Accountability** The law firm must have at least one or two people who can be identified as contacts responsible for the FIE's debt management.
- **Consistency** Generally speaking, debt collection is not a one-off deal, but rather a process that requires ongoing management. The Chinese law firm must be able to provide consistent and uniform updates and reports throughout the debt-collection process. The importance of such documentation cannot be understated, as the progress often must be accounted for not only within China, but also at an FIE's parent office abroad.
- **The office** A foreign investor can tell a great deal from a potential legal partner by viewing the law office and how it is run, so a visit is important. Companies should meet the staff responsible for debt collection and have them explain the process thoroughly.
**DON'T SETTLE FOR LESS**
No business likes uncertainty, but it is often difficult to avoid, especially in China. Many bad debts were created years ago as a result of China's unclear commercial-law structure. Now that the laws are more transparent, businesses should consider how to minimize their debt exposure in China. If receivables are excessive, or if a business has a difficult time obtaining hard currency because of the new SAFE regulations, the firm should seriously consider debt collection. As there is no longer a stigma attached to collecting debt in China, companies that collect against bad debtors will not only find themselves in a better economic position, but also held in higher esteem by would-be and current business associates.
Nearly 150 member company representatives attended the Forecast '99 meeting in Washington, DC, on January 26. After Council President Robert Kapp welcomed the gathering, Pei Minxin, senior economist at the Carnegie Endowment for International Peace, kicked off the morning session with a discussion of PRC politics. According to Pei, China's economic problems, including a sluggish external sector, rising unemployment, and banking woes, may increase the number of local protests but will not necessarily lead to a political collapse. However, a maturation of such protests into multi-regional, synchronized demonstrations; the return of high inflation; or the resignation or ouster of a top leader could jeopardize China's stability.
The next speaker, economist Pieter Bottelier, provided an overview of China's economy, noting that urban unemployment, rural unrest, financial system distress, and protectionism are major challenges. Bottelier agreed with Pei that economic, political, and social crises on a national scale are unlikely. In contrast to other analysts, Bottelier asserted that state-owned enterprise reforms have not slowed as significantly as has been widely reported. He forecasted continued reforms, growth, and stability for China in 1999.
The Council's Director of China Operations Michele Mack Liedeker wrapped up the morning session by describing the current investment climate. According to Mack Liedeker, some of the obstacles US companies faced in China last year included a prohibition on direct selling, an uncertain and confusing tax structure, and corruption. She urged US companies to work with their foreign counterparts in China, the PRC government, and Chinese research institutions to solve these problems.
Council Chairman George M. C. Fisher delivered the luncheon address, focusing on Kodak's experiences in China. He stressed the need for companies to commit to long-term, rather than short-term, investments in China.
The afternoon session consisted of three simultaneous workshops. Francis Bassolino, senior consultant in China Business Services at Deloitte & Touche LLP, and Robert Goodwin, vice president and general counsel of US-China Industrial Exchange, Inc. addressed the distribution workshop. Bassolino gave an overview of the PRC's distribution system, pinpointing some of its bottlenecks and suggesting ways that foreign companies can improve their distribution activities in China. Goodwin spoke of his company's China operations and highlighted the newly permitted wholly foreign-owned international trading company as a promising vehicle to help foreign companies distribute their goods more effectively in China.
The workshop on the 106th Congress featured Norman Ornstein of The American Enterprise Institute, Matt Reynolds, Subcommittee on Asia and the Pacific, House International Relations Committee; Naotaka Matsukata, Office of Senator Joseph Lieberman; and Brian Bieron, House Rules Committee. The speakers discussed the issues that will be significant to US-China relations in 1999, including China's many anniversaries, the Cox Report, the annual battle over Normal Trade Relations (formerly Most Favored Nation status), China's human rights record, and the country's bid to join the World Trade Organization.
During the tax workshop, Joyce Peck of PricewaterhouseCoopers outlined the many recent changes in China's tax regime. Peck emphasized the PRC government's new focus on collecting revenues more efficiently. She also addressed the tangled process of transfer pricing, and the ongoing battle among PRC authorities over export processing. Richard Woodruff of VTEL Corp. discussed his company's acquisition of a representative office in Beijing and related tax contingencies.
Members of the Council's IPR Interest Group in Beijing met with US Trade Representative and PRC government officials to discuss intellectual property rights on February 4—marking the first time business representatives have been included in an inter-government meeting. Participants discussed how PRC government restructuring is affecting IPR enforcement procedures, the 1995 memorandum of understanding between the Chinese Patent Office and the US Department of Commerce and Patent Office, and procedures for submitting comments on proposed changes to PRC IPR regulations. All parties are looking forward to continuing the exchange of ideas at future meetings.
The US-China Business Council's Year 2000 Working Group met with representatives from the Shanghai Municipal Government Y2K Working Committee and other relevant government bodies. At the January 29 meeting, the PRC officials gave updates on Y2K resolution efforts in general and in the power, telecommunications, and banking sectors in particular. Though the officials were able to cite a number of examples of progress, obstacles remain. The Council is working with Shanghai authorities to address Y2K issues.
By focusing investment on the private sector and high value-added export industries, China can maintain the stability crucial for carrying out financial-sector reforms.
To attract foreign investment over the long-term, China must push through the tough structural reforms necessary to both unleash its market potential and sustain respectable economic growth rates. But Beijing will not achieve these goals under its current fiscal spending strategy, which focuses on production at the expense of efficiency. Rather, the Middle Kingdom must adopt a two-pronged strategy that directs investment toward the more efficient sectors of the economy.
First, to enhance export competitiveness, China should switch production from sunset industries that are low-tech and labor-intensive, such as textiles and footwear, to sunrise or higher-value industries, such as machinery and electrical equipment. This process has already begun: between 1992 and 1997, the share of low-tech, labor-intensive products of China's total exports dropped over 6 percent, while that of higher-value products rose more than 8 percent. The country's current policy of keeping the renminbi (RMB) stable also aids the achievement of this goal.
Second, China's government must switch from spending on the inefficient state-owned enterprises (SOEs) to spending on the market-driven sector of the economy, which has suffered from poor access to capital. Only once the government takes steps to facilitate capital flows to private companies, from expanding capital markets to reforming financial institutions, can the most productive firms absorb the excess labor in the state sector.
Moving from sustaining SOEs to nurturing the non-state sector will be deflationary at first, but is essential for pushing through structural reforms. Deflation will help prevent both a resurgence of inflation and a balance-of-payments crisis and will help underpin the RMB over the next few years. Holding the RMB steady also encourages investment in high value-added, rather than low value-added, exports, and thus will benefit the economy over the long term.
No need for devaluation
As China's is still a relatively closed economy, with total trade accounting for 30 percent of gross domestic product and net trade (a component of GDP) accounting for less than 2 percent of GDP, the government's main worry need not be a competitive exchange rate. The major concerns are structural reforms and the efficient allocation of capital.
Moreover, Chinese exports are facing weak demand because of the slowdown in global economic growth, not high prices. Devaluation would not guarantee rising exports but rather would have an inflationary impact on import costs and wages. This is especially true for China where, on average, imports account for 40 percent of export content, according to government reports and HSBC estimates. After China's neighbors devalued, they faced surging import inflation. This, coupled with trade financing problems, caused their export growth to fall.
The devaluations of other Asian currencies have provided their economies a one-time export price break, enabling the sunset industries in the region to grow despite their inefficiencies. This false signal has served only to entrench the region's investment inefficiencies. This entrenchment will reduce the marginal efficiency of capital in these countries relative to that of the United States—Asia's leading trading partner—which is enjoying a rise in technology-induced productivity growth.
The resulting widening productivity gap between Asia and the United States could lead to a prolonged loss of export competitiveness in Asia, and thus exert further downward pressure on the regional currencies. China will escape this fate by not devaluing the RMB—provided that domestic economic policy does not erode the incentive to shift spending from sunset to sunrise industries.
Devaluing the RMB could trigger further competitive devaluation in Asia, pushing China back to square one. It would also threaten to shatter confidence in the Hong Kong dollar peg to the US dollar. And Hong Kong's defense of the peg would inflict further economic pain in the territory.
Market forces will likely dictate a strong RMB in the medium term. According to HSBC's assumptions for export growth (-1 percent in 1999 and 4 percent in 2000), import growth (3 percent and 6 percent, respectively) and foreign-investment inflows (-34 percent and 10 percent, respectively), China will continue to enjoy a balance-of-payments surplus through 2000. Such a surplus means that there will be pressure on the RMB to appreciate, not depreciate.
**Shifting From the State to the Private Sector**
The structural changes in the PRC economy over the past two decades have significantly increased the economic power of the non-state sector, which encompasses private firms, township-and-village enterprises, and foreign joint ventures. The non-state sector has been the major source of employment and output growth: its urban employees and industrial output now account for over 50 percent and 70 percent, respectively, of China's total employment and industrial output (see Figure 1).
But China has yet to restructure its state sector completely. Despite all the efforts to turn state companies around, over half of SOEs are still losing money. Such a record is hardly encouraging for Beijing's purported goal of revitalizing large SOEs by transforming them into Korean-style conglomerates (*chaebol*). To succeed, China will need to strengthen the fledgling private sector to help sharpen external competitiveness and absorb the huge overcapacity of inefficient SOEs. Private enterprises, not reformed SOEs, are
---
**Figure 1**
*Rising Non-state Sector Output and Employment, 1978-97*
Percent of total
**Sources:** CEIC Data Co., *China Statistical Yearbook*
The number of new private firms is growing rapidly, but many go bankrupt quickly.
the key to a stable and productive Chinese economy.
In 20 years of economic reform, however, re-direction of investment to non-state firms has been lacking—the non-state sector still suffers from policy discrimination and tight credit. Eighty-six percent of all state-bank lending goes to the SOEs, leaving only 14 percent for the non-state sector, of which the private-sector segment receives only 0.3 percent (see Figure 2).
Banks are reluctant to lend to non-state companies because of both government policy and their own risk aversion. Beijing has used the state banks as vehicles for policy lending to keep SOEs afloat, and these loans are not based on sound business principles. But since SOEs have government guarantees, their credit risk is perceived to be much smaller than that of private companies. Further, the government often waives the taxes and bank loan interest of SOEs experiencing financial trouble, but does not give the same treatment to the private sector.
Despite these obstacles to private-sector growth, the number of private firms is still growing rapidly through new start-ups set up by laid-off workers. Large numbers of these firms go bankrupt quickly, however. Thus, the growing industrial output of private companies does not necessarily mean that the sector is enjoying an environment conducive to further development.
PRIVATE-SECTOR ACCESS TO CAPITAL
The rising importance of the market sector in the economy ensures that small and medium-sized businesses will become the backbone of China's economy. To facilitate the growth of these companies, the government should grant private businesses equal access to capital and markets.
There are already signs that Beijing has started to realize the importance of the market sector. For example, it has cut the value-added tax for small businesses, effective since July 1998, by 2 percentage points to 4.0 percent. Beijing has also given state commercial banks a freer hand to lend to small and medium-sized companies, easing their credit crunch.
In addition, the government needs to facilitate capital allocation through capital markets. Thus, it must overhaul its securities regulations to focus on the quality of listings rather than their size and political ties. The recently passed securities law addresses trading irregularities, but ignores the limitations on new listings. Current domestic and overseas listing rules are based on a quota-allocation system and restrict public listings to large SOEs. They essentially exclude market-oriented and efficient private enterprises from the capital market, even as loss-making SOEs continue to consume the country's savings.
China could set up a venture-capital stock market modeled after America's NASDAQ, allowing small and medium-sized Chinese companies to list and price according to market mechanisms. The B-share market (currently only open to foreign investors) could even be merged with this Chinese NASDAQ before eventual integration into the domestic A-share market.
Easier access to capital markets through such an outlet will draw investment to start-up private companies in China. The United States has proven that this is possible—since 1990, over 4,000 US companies have gone public, raising more than $400 billion. Many relatively young companies, such as Microsoft Corp. and Federal Express Corp., have become corporate giants competing in global markets, partly because they had access to capital at an early stage in their development. The success of small businesses will make China both more competitive and more flexible.
FINANCIAL SECTOR CLEAN-UP
The key to resolving the SOEs' woes is establishing an institutional framework for capital allocation and market competition. The success of SOE reform will in turn pave the way for bank restructuring. The state banks' heavy loan exposure to the state companies makes reforming SOEs the prerequisite for improving banks' asset quality.
Authorities appear to be aware of the risk of provoking an Asian-style banking crisis if the deterioration in bank asset quality is left unchecked, as indicated by their targeting of the country's international trust and investment companies (ITICs). The ITICs are a good starting point for long-term banking reforms, because the potential systemic risk arising from the closure of ITICs is limited. In contrast, resolving the debt problems in the state banks will be a long process.
The ITICs' fundamental flaw, which applies to the entire banking system, is the serious conflict of interest between lending practices and political interference. The main function of China's ITICs is to raise foreign capital, ranging from direct investments to loans and financial derivatives, for local governments for businesses. They are not run as profit-maximizing institutions. As a result, politically driven lending has been a serious problem. The ITICs also have the balance-sheet mismatch problems common to many...
Asian banks: they have borrowed short-term foreign money to finance long-term local projects. Such mismatches have exposed ITICs to the risks of rising interest rates and withdrawal of foreign funds. Both problems have been evident during the recent Asian crisis.
Meanwhile, poor regulatory oversight and supervision of the ITICs has resulted in capital forbearance and asset quality deterioration. Until recently, funding from local governments or increased overseas borrowing enabled the institutions to muddle through. But high real domestic interest rates, deepening deflation, and a sluggish economy have finally exposed the ITICs' problems, which are being aggravated by the current withdrawal of foreign funds.
The closure of Guangdong International Trust & Investment Corp. (GITIC) last October shows Beijing's resolve in tackling structural problems in the financial system. In GITIC's case, the People's Bank of China (PBOC) made it clear that it would bail out neither debt-ridden financial institutions nor foreign creditors, if the loans were not properly registered with the State Administration of Foreign Exchange. The refusal to bail out all creditors indicates that Beijing is confronting the difficult problem of moral hazard. Provided that Beijing pursues a consistent policy for reforming other financial institutions as it did with GITIC, the closure of GITIC can be seen as a direct attack on the system's root problem. Thus, it is a positive sign for reform.
The fact that Beijing is pushing through painful reforms amid rising economic stress and weak public confidence shows that it understands the urgent need for reform. By offering much higher yields on deposits, the ITICs and other non-bank financial institutions, such as credit cooperatives, have threatened to erode the deposit base of the state banks, which hold almost 70 percent of total deposits and are seen as the foundation of economic stability. In addition, the ITICs' speculative nature and imprudent management suggest that their competition with the banks is unhealthy at this stage. Stabilizing the financial system by shutting down the inefficient, non-bank financial institutions will help pave the way for state bank reform.
The government should thus proceed with its financial restructuring strategy by reforming the ITICs and provincial banks before restructuring the large state banks. Since the non-state-bank financial institutions are small—the four state banks hold over 67 percent of all deposits and 73 percent of all assets in China—their consolidation is unlikely to act as a contagion in the domestic financial system.
Continuing this reform strategy, PBOC has announced that starting this year it will allow only 40, out of 240, ITICs to survive, depending on their asset quality and size. The central bank is likely to allow one ITIC to remain for every ¥10 billion ($1.2 billion) of ITIC assets held in a province. That is, if the combined ITIC assets in a province are ¥20 billion ($2.4 billion), two ITICs may remain after restructuring, provided that their asset quality meets certain as-yet unspecified criteria. Those that fail to meet the criteria but are deemed well-run can apply to become brokerages or mutual funds.
**IMPLICATIONS**
Structural reforms will be deflationary in the short term as they will contribute to layoffs, corporate downsizing, and cuts in credit for non-viable companies. As a result, PBOC will need to continue its easing efforts, and Beijing will have to proceed with its fiscal expansion to prevent the economy from faltering under heavy deflationary pressures. In fact, massive credit inflation and falling interest rates frequently accompany bank restructuring—the US savings and loan crisis in the early 1980s was followed by a steep decline in US interest rates.
But deflation can also reduce the risks of the macroeconomic dislocations that an inflationary environment may bring. These include a surging current-account deficit and massive capital outflows, which can put severe downward pressure on the currency. Progress on structural reforms thus can help bolster the RMB exchange rate by preventing a resurgence of inflation and a current-account deficit. A stable currency and strong resolve on reform, in turn, will enable China to remain an attractive destination for foreign investment for some time to come.
BEHIND THE OPEN DOOR:
FOREIGN ENTERPRISES IN THE CHINESE MARKETPLACE
BY DANIEL H. ROSEN, WASHINGTON, DC,
INSTITUTE FOR INTERNATIONAL
ECONOMICS, 1999, 350 pp. $25,
SOFTCOVER.
In *Behind the Open Door: Foreign Enterprises in the Chinese Marketplace*, Daniel Rosen addresses a range of current issues in US-China commercial relations by examining China's development patterns and the performance of foreign firms on the ground in China. Extrapolating from interviews with 120 managers of foreign-invested enterprises (FIEs) in China, Rosen provides a broad readership—foreign and PRC policymakers, businesspeople, and academics—with a thorough understanding of the foreign businessperson's China.
The bulk of the book surveys foreign firms' experiences in China, with individual chapters devoted to the negotiation and establishment process, human resources and staffing, productivity and product management, distribution and marketing, and legal issues. Each chapter summarizes key issues and ends with a brief yet thorough analysis, complete with a table detailing the main issues firms encounter in the respective sector.
Of particular interest to business readers may be the discussion of the establishment of foreign enterprises in China. Rosen explores the compatibility of China's investment regime with international expectations and finds that, to a certain extent, the current PRC regime bars some foreign firms from entering the PRC market.
After detailing China's investment regime, including the approval process, the author focuses on three broad challenges to doing business in the PRC. First, to overcome the difficulties of obtaining market information, some firms set up representative offices to function as "learning centers." Other firms work with overseas Chinese, or rely entirely on joint-venture partners, local authorities, or patrons for an analysis of the China market. Second, when considering an investment in China, foreign firms must address strategic positioning issues, which include a firm's location choice, PRC policies, and the role of local partners. The third challenge cited by interviewees lies in the negotiation process.
Rosen sets out to answer 10 questions in the study, among them whether China's accession to the World Trade Organization (WTO) will address the concerns of PRC trading partners. He finds that the body's ability to resolve key China issues is likely to be limited. China's lack of policies encouraging competition between firms, for example, could create sectors dominated by only a few companies. As Chinese firms transfer away from state ownership, moreover, trade barriers within China could increase. PRC trading partners...
should not rely upon the WTO to resolve such issues. Instead, the author suggests that they engage PRC leaders at bi- and multilateral meetings, as PRC officials have suggested in the past.
Rosen concludes the book with several policy recommendations. He suggests that the United States redefine its policy priorities regarding China based on the limits of PRC authority, and reiterates the limits of the WTO. He also recommends that the United States accommodate political reform and strengthening in China, and assist the PRC in the areas of business and governance.
Though the well-organized book enables chapters to be read individually, most readers will find it worthwhile to digest *Behind the Open Door* from cover to cover.
—Darlene M. Liao
*Darlene M. Liao is assistant editor of The CBR.*
the crisis has had on China compared to the rest of the region, a devaluation would only increase trade tensions with the United States and hurt the PRC's chances for entry into the WTO. They suggest that China develop internal systems that would support a government bond market which, in turn, could aid overall financial market development and macroeconomic management.
Published in July 1998, this book should maintain its value for some time, as it not only provides readers with an up-to-date, thorough understanding of the circumstances that led to the Asian crisis but also spells out the effects of possible future currency devaluations in the region.
—Gregory S. Heslin
*Gregory S. Heslin is business manager of The CBR.*
**Need to Understand the Chinese Market?**
You Could Spend 3 Months With Expatriate Managers in China, or Read the Latest Study from the Institute For International Economics
**BEHIND THE OPEN DOOR: Foreign Enterprises in the Chinese Marketplace**
Drawing on extensive interviews with expatriate managers and other professionals currently at work in mainland China, **Daniel H. Rosen** describes the experiences of foreign-invested firms in the world's fastest growing economy. It also projects the implications of those experiences for the foreign commercial policies of the industrial countries, including the United States.
"A pathbreaking study compacting two decades of American and other foreign business experience in the new and always evolving China market into one volume. It offers... the most detailed and cogent analysis done by any expert to date on how the process of foreign investment in China actually works and the range of challenges and outcomes. A must read for any executive heading off to run a China operation for a multinational corporation..."
—Daniel Burstein, Senior Advisor, The Blackstone Group, and coauthor of *BIG DRAGON CHINA'S FUTURE*
November 1998. 350 pages. ISBN: 0-88132-263-6. $25.00
INSTITUTE FOR INTERNATIONAL ECONOMICS
11 Dupont Circle, NW * Washington, DC 20036-1207
Tel: (800) 522-9139 * Fax (703) 661-1501
www.iie.com
Forging Reform in China: The Fate of State-Owned Industry
by Edward S. Steinfeld, Cambridge, Cambridge University Press, 1998, 300 pp. $44.98, hardcover.
Since PRC President Jiang Zemin announced the massive overhaul of China's state-owned enterprises (SOEs) in 1997, the international business community has kept a close eye on how reform will affect China's economic development. Rather than focus on the economy's move toward greater openness, however, Steinfeld has chosen to analyze recent issues in the context of the PRC's two-decade effort to liberalize SOEs.
Forging Reform in China also differs from other books on China's SOEs in that the author has made an effort to debunk the popular approach to SOE reform, which couches the process in terms of privatizing the companies. Instead, he focuses on the ways that government decentralization and soft lending have affected SOE decisions regarding capital accumulation and allocation, production, and management.
Steinfeld argues that the most important regulatory mechanisms for enterprises are external ones, namely independent audits, bankruptcy liquidation, enforceable contractual obligations, and functioning capital markets. He stresses that even if internal mechanisms—such as the distribution of stock options to employees—fail, the complete absence of external mechanisms distorts all enterprise operations.
Based on this premise, Steinfeld posits a new approach to analyzing SOE reform. Because the concept of private property rights and corresponding responsibility does not exist within any institutional framework, he claims that transforming SOIs into private companies would not improve their performance. In addition, these SOEs do not operate within a market framework; rather, they exist through rent-seeking behavior and continued bank credit. Consequently, an unbreakable cycle emerges in which subsidization and continued high government taxation distort normal firm signals such as production and profit.
The second half of the book features case studies of three large SOEs in China's steel sector to illustrate the author's theory. Steinfeld concludes that liberalization does not solve all SOE problems; that partial reform can have unintended consequences; and that autonomy without accountability hinders chances for true economic transformation. Steinfeld's clear, if academic, style makes his book an engaging and educational read for those with an interest in China's economic development.
—Julie Walton
Julie Walton is The CBR research assistant.
Stock Market and Futures Market in the People's Republic of China
by Chengxi Yao, Hong Kong, Oxford University Press, 1998, 190 pp. $55, hardcover.
In Stock Market and Futures Market in the People's Republic of China, Chengxi Yao provides a detailed report on China's infant equity market and experiment with futures trading. Although the text can be technical at times, particularly when describing regulations and the PRC government bodies that issue them, the concise presentation of the material enables readers to grasp quickly the necessary details.
The first part of the book discusses the stock market in depth, focusing on its compartmentalized share structure (A, B, C, H, and N shares). The following chapters focus on trading and regulatory structures, such as the Shanghai Securities Exchange and the China Securities Regulatory Commission. At the end of each chapter the author offers a critique, often comparing the Chinese and US securities markets and regulatory structures. She also suggests reforms that would alleviate some of the problems that have arisen as a result of a predominantly state-owned and regulated stock market. One of her recommendations is to replace the compartmentalized share system with one that employs market mechanisms. This section should fascinate anyone familiar with Western-style securities industries, as it explains China's attempt to remain a socialist state while delving into the purely capitalist business of raising money.
The second part of the book looks at the formation of the commodities and financial futures markets and examines the recent, tumultuous attempts to operate futures markets in the PRC. It briefly explains the fiasco in the Treasury-bond market that led to its closure in 1995 as well as the feeble attempts at regulation. Yao argues that in an atmosphere of government control over many prices, there is no need for a futures market to determine prices, much less hedge against price fluctuations.
According to the author, stock and futures markets can be developed and sustained in the Chinese economy—as long as reforms of the state-enterprise system, and in the allocation of resources, take place. The author's call for a national securities law and a national futures law has been overtaken by events: a national securities law, which addresses some of the problems detailed in this book, was passed by the National People's Congress in December 1998.
The book would have benefited from some information on foreign participation in the markets and graphical representation of some of the more significant market trends. Stock Market and Futures Market nonetheless is an excellent resource for all those who do business in China, want to invest in China, or simply enjoy tracking China's attempts at combining socialism and capitalism.
—Nicole Vlazny
Nicole Vlazny is a research assistant at The US-China Business Council.
The following tables contain recent press reports of business contracts and negotiations exclusive of those listed in previous issues. For the most part, the accuracy of these reports is not independently confirmed by *The CBR*. Contracts denominated in foreign currencies are converted into US dollars at the most recent monthly rate quoted in the International Monetary Fund's *International Financial Statistics*.
Firms whose sales and other business arrangements with China do not normally appear in press reports may have them published in *The CBR* by sending the information to the attention of the editor.
### Accounting and Insurance
**INVESTMENTS IN CHINA**
**Reliance Group Holdings Inc. (US)/Tian An Insurance Co. (Beijing)**
Will form joint-venture life insurance company. 12/98.
### Advertising and Public Relations
**INVESTMENTS IN CHINA**
**R. H. Donnelley Corp. (US)/UNICOM**
Set up joint venture to publish yellow pages and offer Internet directory services. (US:15%-PRC:85%). 11/98.
### Agricultural Commodities and Technology
**INVESTMENTS IN CHINA**
**Bayer AG (Germany)/NA (Sichuan)**
Formed joint venture to manufacture animal-care products in Chengdu, Sichuan Province. (Germany:70%-PRC:30%). $10.8 million. 12/98.
**Hoechst Roussel Vet (Germany)/Jilin Bioproducts Factory**
Will set up joint venture to produce poultry vaccines. (Germany:50%-PRC:50%). $12 million. 12/98.
**TeMania International (New Zealand)/China Continental, Inc. (NA)**
Will breed boar goats in China. 12/98.
---
**OTHER**
**Ministry of Agriculture (the Netherlands)/Guangdong Bureau of Agriculture**
Will cooperate in raising pigs, preventing pig diseases, processing fodder, developing pork processing technology, and licensing China's pork exports. 12/98.
### Banking and Finance
**INVESTMENTS IN CHINA**
**Manufacturer's Life (Canada)/SINOCHEM**
Will set up an investment fund management joint venture in Beijing. 11/98.
**OTHER**
**The Industrial Bank of Japan/State Development Bank of the PRC**
Signed investment banking cooperation agreement covering areas such as mergers and acquisitions, diversification, and RMB financing. 1/98.
### Chemicals, Petrochemicals, and Related Equipment
**CHINA'S IMPORTS**
**Technip Italy, a unit of Technip (France)**
Will design and construct a purified terephthalic acid production plant in Guangdong Province for Zhuhai No.1 Acid Factory. 11/98.
### INVESTMENTS IN CHINA
**Cabot Corp. (US)/Shanghai Coking General Corp.**
Launched carbon-black manufacturing joint venture in Shanghai. (US:50%-PRC:50%). $34 million. 1/99.
**Dyno Industries (Belgium)/Beijing Building Materials Group**
Set up joint venture to manufacture resins and wood adhesives with an annual capacity of 50,000 tons. (Belgium:55%-PRC:45%). 12/98.
---
Abbreviations used throughout text: ADB: Asian Development Bank; BOC: Bank of China; CAAC: Civil Aviation Administration of China; CANADA: Canadian International Trade Export Corp.; CATC: China National Aero-Technology Import Export Corp.; CITIC: China International Trust and Investment Corp.; CITS: China International Travel Service; CNOOC: China National Offshore Oil Corp.; CNPC: China National Petroleum & Gas Corp.; FDIZ: Economic and Technological Development Zone; FBC: Industrial and Commercial Bank of China; MEE: Ministry of Information Industry; NA: Not Available; NORINCO: China North Industries Corp.; P&T: Posts and Telecommunications; PRBC: People's Bank of China; SEZ: Special Economic Zone; SINOCHEM: China National Chemical Corporation; SINOPET: China National Petroleum Corp.; SINOTRANS: China National Foreign Trade Transportation Corp.; SPC: State Planning Commission; UNDP: United Nations Development Program; UNICOM: China United Telecommunications Corp.
Rosen Bogur Private Trade Co. Ltd. (Australia)/Huaguang Glass Group (Anhui)
Formed joint venture to produce high-grade plastic and PVC foaming materials. (Australia:75%-PRC:25%). $10 million. 12/98.
Witco Corp. (US)/NA (Shanghai)
Will set up a joint venture in Shanghai to produce specialty chemicals. $30 million. 11/98.
OTHER
Borax Inc. (US)
Opened a representative office in Beijing. 11/98.
Consumer Goods
INVESTMENTS IN CHINA
Am-Pac International Inc. (US)/AUCMA Group (Guangdong)
Will establish joint venture to manufacture and market refrigerators. $45 million. 12/98.
Coastcast Corp. (US), Dynamic Precision Casting Co. Ltd. (Taiwan)
Will establish a golf club manufacturing joint venture in Guangzhou, Guangdong Province. (US:55%-Taiwan:45%). 12/98.
Kohler Co. (US)
Opened a wholly foreign-owned bathtub manufacturing plant in Shanghai's Pudong New Area. 12/98.
Merloni (Italy)/Wuxi Little Swan (Jiangsu)
Formed dishwashing machine manufacturing joint venture with an annual capacity of 800,000 units. (Italy:30%-PRC:70%). $30 million. 12/98.
OTHER
Marks & Spencer (UK)
Closed their representative office in Beijing. 11/98.
Electronics and Computer Software
CHINA'S IMPORTS
Unisys Corp. (US)
Will install a passenger processing system at the domestic terminal of Shanghai's Pudong International Airport. $2 million. 1/99.
Oracle Corp. (US)
Will provide the State Postal Bureau and the China Post Telecom Equipment General Co. with Internet software. $1 million. 12/98.
TriGem Computer Inc. (S. Korea)
Will provide the Shenyang Bureau of Science and Technology with technology for developing desktop and notebook computers. 11/98.
INVESTMENTS IN CHINA
International Information Integration Co. Ltd. (US)/Baoding Aoshen Science and Technology Development Co. (Hebei)
Established joint venture to design, install, and maintain computer network systems. 1/99.
Lithium Technology Corp. (US)/Hengdian Group (Zhejiang)
Formed joint venture to manufacture lithium-ion polymer batteries in Zhejiang Province. 1/99.
Meiko (Japan), Santis (Taiwan)
Set up circuit-board manufacturing joint venture in Guangdong Province. $50.8 million. 1/99.
Toshiba Corp. (Japan)/Dongchuan Xinxin Industrial Group (Jiangsu)
Established joint-venture server manufacturing company in Nanjing, Jiangsu Province. (Japan:49%-PRC:51%). $5 million. 1/99.
Computer Associates Inc. (US)/Legend Holdings (Beijing)
Will set up joint venture to develop and distribute software. (US:50%-PRC:50%). $3.5 million. 12/98.
Fairyoung Holdings Ltd. (US)/China Huayu Development Corp. (Guangdong)
Will develop Internet teaching facilities and manufacture computer hardware. (US:50%-PRC:50%). 12/98.
Microsoft Inc. (US)/Founder Group (Beijing)
Will develop electronic publishing and media program authoring system. 12/98.
NCR Technologies Co. (US)/Guangzhou Tianyin Electronics Co. Ltd. (Guangdong)
Formed joint venture to install and repair ATM machines. 12/98.
Sina Media Co. (US)/Beijing Stone Richsight Information Tech Co. Ltd.
Launched joint venture to create Chinese language websites. (US:40%-PRC:60%). $50 million. 12/98.
Cisco Systems Inc. (US)/Beijing Information Office, China Information Highway Corp. (Beijing)
Will develop a mainland-based Internet exchange to avoid routing through other countries. 11/98.
Mitsubishi Electronics (Japan)/The Stone Group (Beijing)
Set up IC manufacturing joint venture in Beijing with a monthly production capacity of 10 million units. $2 billion. 11/98.
NTT Corp. (Japan)/Huaxin Computer Technologies (Liaoning)
Established joint venture in Beijing to focus on postal saving deposits computer systems. $1 million. 11/98.
OTHER
Faulkner & Gray Inc. (US)/China Chamber of Information Industry
Will partner to organize the China '99 Smartcard Conference. 1/99.
FVC.COM (US)
Opened a representative office in Shanghai. 12/98.
Xilinx Inc. (US)
Donated programmable logic products to Fudan University, Shanghai. $400,000. 12/98.
Engineering and Construction
**Investments in China**
**Alcatel SA (France)/Tianjin Electromagnetic Wire Factory**
Set up joint venture to manufacture copper wire and cable for electric coils. (France:60%-PRC:40%). 11/98.
Environmental Technology and Equipment
**Investments in China**
**Environmental Analysis Remote Sensing Co. (the Netherlands)/State Forestry Bureau**
Will develop remote-sensing satellite system to monitor nationwide desertification. $3.8 million. 11/98.
**Other**
**The World Bank**
Approved loan for the Tianjin Environmental Sanitation Landfill project. 12/98.
Food and Food Processing
**Investments in China**
**Boortmalt Overseas Group (Belgium)/Beijing Beer Materials Factory**
Formed joint venture to produce malt for China's brewing market. (Belgium:55%-PRC:45%). $7 million. 1/99.
Machinery and Machine Tools
**China's Imports**
**Alternative Fuel Systems Inc. (US)**
Will supply gas regulator systems to Shutong Prospects Import and Export Co., Sichuan Province. $1.7 million. 1/99.
**Hyundai Precision Industry (S. Korea)**
Will deliver 10 concrete pump cars and transfer technology for assembling and installing booms. $2.1 million. 12/98.
**China's Investments Abroad**
**Government of Uganda/China Light Industry Corp.**
Will cooperate to expand Uganda's industrial sector. $3 million.
Medical Equipment and Devices
**China's Imports**
**Pangea Systems Inc. (US)**
Licensed its bioinformatics software and genomics target validation tools to the Chinese National Human Genome Center in Shanghai. 12/98.
**Therrogenesis Corp. (US)**
Will install equipment in the Guangdong Therapy Center that preserves placental blood stem cells used to reconstitute bone marrow. 11/98.
Investments in China
**The Quigley Corp. (US)/NA (Shanghai)**
Signed agreement to distribute over-the-counter cold medicine in China. $52 million. 12/98.
Other
**The World Bank**
Approved loan to the Ministry of Health for vaccination and disease prevention projects. 11/98.
Metals, Minerals, and Mining
**Investments in China**
**Integrated Carbonics Corp. (Canada)/Yichang Hengda Graphite Group (Hubei)**
Will mine graphite deposits in Hubei Province. (Canada:55%-PRC:45%). 12/98.
Packaging, Pulp, and Paper
**China's Investments Abroad**
**Kunning Electro-chemical Plant (Yunnan)**
Will renovate the Sittoung Paper Mill No. 1 in Myanmar. $1.2 million. 12/98.
**Other**
**Breakspear Ltd., a unit of Leading Edge Packaging Inc. (US)/Hua Nam Enterprises Ltd. (Guangdong)**
Terminated packaging joint venture. 12/98.
Petroleum, Natural Gas, and Related Equipment
**China's Investments Abroad**
**Oyuni Undraa Suubba (Mongolia)/Huafu Industrial Co. (Hellongiang)**
Will build and operate an oil refinery in southeast Mongolia. (Mongolia:50%-PRC:70%). $39.7 million. 12/98.
**Other**
**Chevron Corp. (US)**
Increased investment in its drilling operations in the Bohai Sea. $60 million. 12/98.
Pharmaceuticals
**Investments in China**
**Lupin Laboratories (US)/NA (Beijing)**
Set up joint venture to invest in pharmaceutical research. (US:65%-PRC:35%). 1/99.
Ports and Shipping
**China's Exports**
**Hubei Machinery Import & Export Corp.**
Will build nine 131,000 ton multipurpose container ships for Greek Turanska Import & Export Corp. $117 million. 12/98.
**CHINA'S IMPORTS**
**OSI Systems Inc. (US)**
Will supply the Chinese customs authority at Shenzhen, Guangdong Province, with a vehicle cargo x-ray system. 1/99.
**INVESTMENTS IN CHINA**
**Oriental Overseas Group (Hong Kong)/Port of Qingdao Authority (Shandong)**
Established international shipping and warehousing company in Qingdao, Shandong Province. $5 million. 1/99.
**Sea-Land Orient Container Terminals Co., a unit of CSX Corp. (US)/Port of Tianjin Authority**
Formed joint venture to operate the Dongtui South container terminal in Tianjin. 1/99.
**United Parcel Service Inc.(US)/China National Foreign Trade Transportation Corp.**
Expanded their joint venture to include 21 more cities. 12/98.
**Five-Ocean Shipping Co. (Japan)/China Rail Special Freight Transport Co. (Beijing), China North Industry Co., Tianjin branch**
Established freight transportation joint venture. (Japan:45%-PRC:55%). $4 million. 11/98.
**Winnport Corp. (Canada)/Air China**
Signed interline agreement for transporting cargo from China's coast to the interior. 11/98.
---
**Power Generation Equipment**
**CHINA'S IMPORTS**
**CAE Electronics Ltd. (Canada)**
Will supply the Henan Electric Power Corp. with a hydroelectric resource automation dispatching system. $5 million. 11/98.
**Siemens AG (Germany)**
Will equip the Waigaoqiao coal fired power plant near Shanghai. $300 million. 11/98.
**CHINA'S INVESTMENTS ABROAD**
**Government of Morocco/Government of the PRC**
Will build a 10 MW nuclear power station in southwest Morocco to desalinate water. $40 million. 12/98.
---
**Property Management and Development**
**CHINA'S IMPORTS**
**Iwerks Entertainment (US)**
Will install theaters in the Fujian Chang Tai Amusement Co. Ltd., and the Qingdao Hai'er Museum, Shandong Province. $3.2 million. 11/98.
**INVESTMENTS IN CHINA**
**SA Woodhead International (Australia)**
Will develop 65,000 sq m residential district in Shanghai. $55 million. 1/99.
**Four Seasons Hotels Inc. (Canada)/Shanghai South Pacific Hotel Co.**
Will build and manage a hotel in Shanghai. 12/98.
**SIP Investment Co. (Singapore)/Huaxin International Realty Co. (Liaoning)**
Set up a residential development company in Suzhou, Jiangsu Province. (Singapore:50%-PRC:50%). $725,000. 12/98.
**Heiwado Co., Koizumi Garments Co. (Japan)/Hunan International Economic Development Group Co.**
Established 113,000 sq m multipurpose commercial and trade center in Changsha, Hunan Province. $96 million. 11/98.
**Radisson International Management (US)/Zhongshi Group Co. (NA)**
Will jointly open and manage 100 hotels throughout China. 11/98.
**Swissôtel Management Ltd. (Switzerland)**
Opened hotel in Dalian, Liaoning Province. 11/98.
---
**Telecommunications**
**CHINA'S IMPORTS**
**LM Ericsson AB (Sweden)**
Will expand the GSM digital network for the Inner Mongolia P&T Administration. $14.3 million. 1/99.
**Lucent Technologies (US)**
Will provide analog and digital phones to the Shangri-La Hotel in Hangzhou, Zhejiang Province. 1/99.
**Oy Nokia AB (Finland)**
Will expand the GSM cellular network for the Henan P&T Administration. $280 million. 1/99.
**VTEL Corp. (US)**
Will deploy digital visual communications technology units in Beijing Medical University, the Guangdong Branch of China Telecom, and the power industry bureaus of Hubei and Hunan provinces. 1/99.
**Datacraft (US)**
Will build a high-speed pilot ATM network for the Fujian P&T Administration. $4 million. 12/98.
**Digital Microwave Co. (US)**
Will supply the Liantong Company, Liaoning Province, with digital microwave wireless and network system equipment. $5.5 million. 12/98.
**LM Ericsson AB (Sweden)**
Will complete the fourth expansion of Chongqing's GSM network. $138 million. 12/98.
**LM Ericsson AB (Sweden)**
Will expand the GSM network for the Guangxi P&T Administration. $157 million. 12/98.
**Motorola Inc. (US)**
Will expand the wireless telephone networks for the Fujian and Shandong provincial branches of UNICOM. $22 million. 12/98.
**Osicom Technologies, Inc. (US)**
Will provide Tangshan Cable Television Network, Hebei Province, with dense wavelength division multiplexing equipment. $2.9 million. 12/98.
**SR Telecom Inc. (Canada)**
Will supply the Gansu P&T Administration with several wireless microwave systems. $7.5 million. 12/98.
**Ascend Communications Inc. (US)**
Will expand the asynchronous transfer mode frame-relay broadband multiservice network in Zhejiang Province. 11/98.
**ECI Telecom Ltd. (Israel)**
Will deliver voice and data traffic terminals to the Xinjiang P&T Administration. $3.9 million. 11/98.
**Fujitsu Ltd. (Japan)**
Will construct a 16-channel, 546 km fiber-optic ring around Zhangzhou, Fujian Province. $8.1 million. 11/98.
**Newbridge Networks Corp. (Canada)**
Will build a broadband ATM network for China Post. 11/98.
**Northern Telecom (Canada)**
Signed multiple contracts with the Guangdong, Hebei, Shaanxi, and Yunnan P&T administrations; the Chongqing, Heilongjiang, Ningxia, Shandong, and Sichuan branches of UNICOM; and with the General Administration of Customs for the design, construction, and installation of various digital, analog, and GSM networks. $120 million. 11/98.
**STM Wireless, Inc. (US)**
Will deliver 1,400 very small aperture terminals to China Golden Health Medical Co. to create satellite-based broadband medical network. $8 million. 11/98.
---
**CHINA’S INVESTMENTS ABROAD**
**Cuba Electronic Telecommunication and Information Automation Group/Dragon Telecommunication Equipment Co. (NA)**
Will conduct telecommunications research in Cuba. $300 million. 12/98.
---
**INVESTMENTS IN CHINA**
**Digitcom Corp. (US), Teltek Communications International (Taiwan)**
Formed joint venture to deliver long distance telephone service between North America and China. 1/99.
**Phasecom Inc. (US)/Shanghai Bell**
Announced OEM agreement to manufacture voice communication equipment in Shanghai. 1/99.
**AmeriTeleCon (US)/Pan Pac Inc. (Beijing)**
Established joint venture to market video conferencing equipment. 12/98.
**Belgacom (Belgium)/Jiangsu and Shanghai branches of China Telecom**
Will develop international transmission services and multimedia communication devices. 12/98.
**Elcoteq (Finland)/Nanxin Industrial Development Corp. (Guangdong)**
Established joint venture to manufacture mobile phone accessory parts. (Finland:70%-PRC:30%). 12/98.
---
**Draka Holding NV, NKF Holding NV (the Netherlands)/Wuhan Changjiang Communications Industry Group Inc. (Hubei)**
Will manufacture radio frequency cable for mobile telecommunication networks. $15.7 million. 11/98.
**Eicon Technology (Canada)/Beijing and Shanghai P&T administrations**
Will provide ISDN equipment and training, and develop joint marketing strategies. 11/98.
**NeTrue Communications Inc. (Canada)/Tangsheng Investment and Development Co. (Shanghai)**
Will develop carrier-grade, large-port capacity IP phone switches. 11/98.
**Surrey Satellite Technology Ltd. (UK)/Qinghua University (Beijing)**
Launched joint venture to develop micro-satellites for the Chinese market. 11/98.
---
**OTHER**
**GTE Corp. (US)/China Internet Information Center (CIIC) (Beijing)**
GTE will train CIIC technicians in database construction and website design. 12/98.
**Intelsat**
Will provide China Telecom with satellite communication services to link China to the Internet. 12/98.
**Premisys Communications (US)**
Opened representative office in Beijing. 12/98.
**VideoServer, Inc. (US)**
Opened representative office in Beijing. 12/98.
**Lucent Technologies (US)**
Will open a research center in Qingdao to develop software and switching systems. $24 million. 11/98.
---
**Transportation**
**CHINA’S INVESTMENTS ABROAD**
**Changchun Passenger Train Co. (Jilin)**
Will deliver 217 subway cars to the Tehran Urban & Suburban Railway Co., Iran. $138 million. 12/98.
---
**CHINA’S IMPORTS**
**Ansaldo Signal NV (the Netherlands)**
Will deliver 900 jointless track circuit units to the Beijing Railway Bureau. $6.5 million. 1/99.
**Ascom (Hong Kong)**
Will build and operate a toll-collection system along the Quanzhou-Xiamen expressway in Fujian Province. 1/99.
**Bell Helicopter Textron Inc. (US), Samsung Aerospace (S. Korea)**
Will deliver eleven SB427 light twin-engine helicopters to various Chinese firms. 1/99.
Autosan Jelcz Co. (Poland)
Will supply bus chassis parts to Beihai Facility Engineering Development Corp., Beijing Jingtong Bus factory, Guangzhou Guangke Auto Enterprises (Group) Ltd., Shandong Liaocheng Bus Factory, and Shanghai Tourist Bus Factory. $99 million. 12/98.
The Boeing Co. (US)
Will convert two MD-11 jetliners into freighters for China Eastern Airlines. 12/98.
INVESTMENTS IN CHINA
BAT International (US)/Suzhou Machinery Holding (Jiangsu)
Will manufacture electric bicycles, mopeds, and scooters. (US:25%-PRC:75%). $4 million. 1/99.
Alstom SA (France)/China Railway Locomotive Rolling Stock Corp.
Will produce railcar shock absorbers. (France:51%-PRC:49%). 12/98.
BTR Industrial Group (UK)/NA (Liaoning)
Established electric motor production joint venture with an annual capacity of 270,000. $30 million. 12/98.
Union Technology Co. (US)/China Aviation Equipment Import & Export Co. (Guangdong)
Will build and operate engine spare-parts warehouse in Beijing. 12/98.
Degussa AG (Germany)/China Automotive Industry International Corp., Shanghai Pacific Chemical Co. Ltd.
Will set up automotive catalyst manufacturing joint venture in Shanghai with an annual production capacity of 3 million. (Germany:55%-PRC:45%). 11/98.
OTHER
ADB
Granted loan for the Hebei expressway project. 1/99.
Delta Airlines Inc. (US)/China Southern Airlines
Signed a comprehensive code-sharing agreement. 12/98.
Orient Overseas Container Line Ltd. (Hong Kong)/China Railway Container Transport Center
Expanded their joint venture to include cross-country shipments in refrigerated cars. 12/98.
The World Bank
Approved loan for the construction of a 152 km expressway linking Hefei and Anqing in Anhui Province. $200 million. 12/98.
Export & Import Bank of Japan
Will provide loan to the Shanghai Government for highway expansion. $200 million. 11/98.
Miscellaneous
CHINA'S EXPORTS
Government of Tanzania/Government of the PRC
Will provide grant for economic development in Tanzania's capital. $2.6 million. 1/99.
Government of Morocco/Government of the PRC
Signed agreement regarding unspecified exports to Morocco. $50 million. 12/98.
INVESTMENTS IN CHINA
Sulo AB (Germany)/Beijing Gaoliu Agriculture Industry and Commerce Co.
Will manufacture anti-erosion, mobile garbage bins. (Germany:90%-PRC:10%). $9.6 million. 12/98.
OTHER
Government of South Australia
Will provide 12-month training program in hotel/restaurant management for 500 Chinese students. $8.2 million. 1/99.
NA (US)/Chongqing University
Launched a film academy with six academic departments and film studio. $4.8 million. 11/98.
Only one company on earth has this many ways to help you manage your risks.
- All forms of property/casualty insurance
- Life, accident & health insurance worldwide
- Claims management and cost containment services
- Global employee benefits and pension management
- Loss control and engineering
- Premium financing
- Political risk coverage
- Trade credit insurance
- Emergency evacuation and security services
- Economic and political assessment
- Local currency financing
- Consumer finance in many countries
- Direct equity investment funds in many countries
- Third-party asset management
- Infrastructure finance in emerging markets
- Market making in all forms of interest rates, currency and equity risk management products
- Trading and market making in foreign exchange, interest rates, base and precious metals, commodities and commodity indices
- Aircraft leasing for airlines
- Personal auto insurance through various distribution systems
- Retirement savings products from SunAmerica
With 80 years of international experience, only Triple-A-rated AIG has the worldwide insurance and financial services capabilities to help you manage all your balance sheet and other risks. And our list of ways to help you manage business risks continues to grow. For more information about insurance and services provided by our member companies, write to American International Group, Inc. or visit our website at www.aig.com
WORLD LEADERS IN INSURANCE AND FINANCIAL SERVICES
American International Group, Inc., 70 Pine Street, Dept. A, New York, NY 10270
Neighbors. Over the last years we have gained tens of millions of customers in China. We have also gained some remarkable new employees—and neighbors—in Shanghai, Xiamen, Shantou, and Wuxi. "Anyone can build a good house," a Chinese proverb contends, "but good neighbors are priceless." Kodak has a single set of standards for health, safety and the environment. We apply them in every Kodak plant in the world. These standards are high. Very, very high. One goal never changes. To improve every day...every month...every year. We celebrated recently. In Wuxi we broke ground for a new photochemical plant near magnificent Lake Tai. In Xiamen, we noted that Kodak people have now put in some 5 million hours of work without any time lost due to injury. The health of Kodak people. Their safety. Respect for the environment. These values are fundamental. We apply them. We verify them. We improve them. Everywhere. Especially in China.
TAKE PICTURES. FURTHER.
WWW.KODAK.COM
© Eastman Kodak Company, 1999. Kodak and Take Pictures. Further. are trademarks. |
Kangaroo Care in a Neonatal Context: Parents’ Experiences of Information and Communication of Nurse-Parents
Desirée Lemmen\textsuperscript{1}, Petra Fristedt\textsuperscript{2} and Anita Lundqvist\textsuperscript{*3}
\textsuperscript{1}Skåne University Hospital, Lasarettsgatan 48, Lund, SE-22185 Sweden
\textsuperscript{2}Capio Cityclinic, Ostergatan 29, Landskrona, SE-261 34, Sweden
\textsuperscript{3}Child, Family and Reproductive Health, Department of Health Sciences, Lund University, P.O. Box 157, Lund, SE-221 00, Sweden
**Abstract:** Objective: Kangaroo Care (KC) is an evidence-based nursing practice with many benefits for infants and parents. The purpose of this study was to describe parents’ experience of information and communication mediated by staff nurses before and during KC at neonatal wards.
Methodology and Participants: A qualitative study with semi-structured interviews was performed. The sample consisted of 20 parents.
Results: The results show that the information and communication were experienced as both optimal and suboptimal including following categories: initially conflicting emotions in relation to KC, participation and confidence in KC is evolving, strengthening preparation and context is decisive as well as parental sense and caution. The overall theme was that good preparation will contribute to a positive experience of KC.
Conclusion: The conclusion is that most of the parents had positive experiences of KC. The information and communication from the staff nurses encouraged and motivated the parents to practice KC, in a sense that it was a natural way to get to know the infant, when the staff nurses were well versed in the method and coherent and supportive. Conflicting emotions emerged when staff nurses practised KC as a routine without deeper knowledge and skills of the method and its advantages as well as without sensitivity to parents’ vulnerable situation.
**Keywords:** Communication, information, Kangaroo Care (KC), neonatal, parents experience.
**INTRODUCTION**
Of all children who are born each year in Sweden, approximately 90 000-100 000, 4.8% are born preterm [1, 2]. A preterm birth is defined as delivery that occurs before 37 weeks of gestation [1, 2]. About 0.9% are born at gestational age of 32 weeks and 4.6% between 33 and 36 weeks [3]. The number of preterm births in Europe and North America are 7.4% [2]. When the infant is born preterm the infant is taken away from their parents immediately, following the birth for medical assessment and treatment that is described as a crisis for both parents [4, 5]. A preterm infant cared for in a Neonatal Intensive Care Unit (NICU), is associated with highly technical equipment and parents can experience difficulties regarding the experience of that the infant is theirs and that the infant needs them [6-8]. For parents to manage their fears of approaching the small infant, they need to be encouraged to touch their infant. They want to be involved in health care but to venture to come close to their infant the parents need a lot of support from the staff nurses [6-8]). Kangaroo care (KC) is a nursing practice, to some extent accepted worldwide, to meet newborn infants’ important needs for warmth, stimulation, parental contact and love [9, 10].
In Europe, KC is admitted as an evidence-based nursing practice in neonatal care. The guidelines recommend that postnatal care should be conducted so that mother and infant after birth are only separated for the shortest possible time, i.e. that KC should be implemented as soon as possible after delivery or upon arrival at the neonatal unit. It could be used for up to 24 hours a day taking into account the infant’s and mother’s condition [10]. There is no defined optimum length of time to carry out KC, although the WHO [9] suggests 2 hours or more/day. Guidelines for KC describe the practice of holding a preterm or low birth weight infant in an upright, prone position with skin-to-skin contact against the parents’ or relatives’ chest, dressed in a nappy, with a blanket or clothes covering the infant [9-11].
KC has been shown to strengthen the ties between infants and parents regarding parents and infants being more sensitive to each other [12-16]. It is also suggested that parents initially are ambivalent to KC. An ambivalence involving a yearning to have their infant close as well as apprehensions about getting too close to their infant or a fear of hurting the infant [9, 14]. Gradually this ambivalence change into strong feelings of well-being in the experience...
of that KC was something simple and natural, a deep feeling that was hard to put into words [14, 17]. The bonding between parents and infant increased as KC had been provided repeatedly. The parents got a feeling of being able to provide warmth and closeness and experienced that they became involved in the care of their infant [12-16]. KC also developed a sensitivity to the infants signals and active parenting [12-19].
KC’s impact on the mother has been shown to decrease depressive tendencies as well as that if the method was applied daily the mothers’ stress was reduced [18]. Fathers who applied KC directly after the infant was delivered by a caesarean section became more sensitive to the infant’s behavioural signals i.e. the infant cried less, showed more searching and craving activity, as well as increased alertness [20]. Fathers who applied KC at the NICU experienced that it promoted their obvious role in the infant’s care. The perception was that they were important in the care of the infant and acting as a primary caregiver [16].
KC also has positive effects on the preterm infant, i.e. improved sleep and brain development as well as less pain responses [21, 22], earlier initiation of breast feeding [7, 23] and becomes easier to settle [16]. A meta-analysis showed that KC also significantly reduces mortality among infants with a birth weight < 2000g and reduces morbidity of major illness, especially caused by infections [24].
Information about the infant’s condition and procedures in the neonatal unit as well as the care of the infant are requirements that need to be met for parents to feel comfortable with KC [8, 13]. Mothers have an essential need for nurses to guide the KC experience [15, 19] to allay their anxiety about handling the infant and to promote confidence in using KC [12]. They also need support for KC with encouragement, teaching, and planning from staff nurses [19]. Some mothers experienced that they had not received enough information about KC and others experienced the information as being too extensive, with difficulties recollecting the information. The attitude of staff nurses during KC were sometimes negative that made mothers doubtful [17]. Being offered options of asking other persons to participate in the provision of KC at the NICU was also inadequate and staff nurses lacked sensitivity to mothers needs and wishes regarding KC [15]. Fathers experienced that staff members made conflicting statements and displayed conflicting behaviours, leading to a worry about which KC rules would apply during the next shift or what was expected of them as fathers. Communication between staff nurses and between work shifts was perceived as poor. This was associated with giving the fathers a feeling of uncertainty to perform certain aspects of the infant’s care and difficulty in controlling the infant and feel comfortable in the situation [16].
The purpose of this study was to describe parents’ experience of information and communication mediated by health care staff before and during KC at neonatal units.
MATERIALS AND METHODOLOGY
A qualitative design was employed i.e. a reflection of the researcher’s desire to have the inquiry based on the realities of those under study [25]. The overall aim was to produce condensed and brief descriptions of the parents’ experiences of KC especially regarding information and communication during the hospital stay when having given birth to a preterm infant.
Setting
The study was performed at three hospitals in southern Sweden of which one carry out neonatal care level IV (University hospital with fully developed intensive care) and the two other hospitals at level III (County hospitals with partial neonatal intensive care [26]). KC was introduced as a care method in all hospitals. Included in the study were healthy primi- and multiparous women and their cohabitants.
Participants
The women should have delivered an infant, 24 + 0 to 35+0 gestational weeks, who did not have a life-threatening condition and the informants having a good knowledge of the Swedish language. The mothers and fathers, who had experienced KC from birth to discharge during the period March to December, 2009, were invited to participate in the study. After the infants discharge from the hospital the parents who had consented to participate were approached by a phone call and asked if they still agreed to participate: a time and place for the interview was reserved. Twelve families chose to participate in the study. The participating families represented fifteen neonates while three of the families had given birth to twins. Four of the fathers declined participation because of lack of time. In five of the twelve families, one parent (three mothers and two fathers), did not have Swedish as their native language but had good knowledge in Swedish. The foreign-born came from Denmark, Germany, Montenegro, Norway and United Kingdom. Out of the families five had stayed at more than one hospital due to the fact, that the infant needed level IV care or that the delivery started when the parents were away from home. The infants’ gestational age ranged from 24 + 5 to 34 + 4 (M = 30 + 5) including one infant 24+5, 25+5, 32+3, 33+2, 33+5, 34+5 and three infants 28+0 as well as two infants 33+0, 33+1, representing six girls and nine boys.
Data Collection
The authors who collected the data (the first and second author) were registered nurses and mothers of small and older children and had worked in neonatal units a few years where KC was introduced. The third author has extensive experience of work and research with infants and parents. During the interviews the authors strived to maintain an open, reflective stance, suppressing the own experience by posing open questions and seeking clarification of what was expressed by the parents by an open atmosphere. The interviews lasted 20 to 75 minutes. Eleven of the interviews were conducted by the two authors and one interview was conducted by one author due to interview time being modified by the family on short notice. All interviews were tape-recorded and transcribed verbatim.
The participants had been discharged/released from the neonatal unit between one to four weeks prior to the time of the interview. The main topics during the interviews were: Staff information and communication with parents before and during provision of KC and the feeling of holding their infant skin-to-skin. All interviews took place in the family home and mothers and fathers were interviewed together.
while the mother was interviewed alone when the father could not attend due to time constraint. All was according to the parents’ wishes.
**Data Analysis**
The transcribed interviews were analysed by a modified content analysis on the basis of Graneheim and Lundman [27] but modified according to the below description. Initially, all interview texts were read several times to get a general impression. The interview texts concerning the birth were not included in the analysis. The interview texts were divided into meaning units i.e. words, sentences or paragraphs containing aspects related to each other through their content and context [27]. The meaning units were condensed i.e. a process of shortening while still preserving the core and abstracted to codes. The codes were placed under each other for information and communication before and during the KC and thereafter sorted based on similarities, differences and patterns that formed subcategories. The sub-categories that emerged were: no information/communication; good information/communication; consistent and inconsistent information/communication; and positive emotions, negative emotions and strengthened parents-infant connection while providing KC. The subcategories were joined into categories of similar nature i.e. a group of content that shares a commonality [28]. The final step in the analysis was to interpret the contents of the subcategories to show the underlying meaning of the interviewees’ narratives and to deepen the understanding. This resulted in four categories and one overall theme (Table 1). Finally, the interview texts were read again to make sure that the overall theme and the categories covered the interview texts. Discussion among the authors continued throughout the whole analysis process until agreement was reached. The study was also critically examined in seminars with fellow students and researchers.
**Ethical Considerations**
The study was conducted in accordance with ethical guidelines regarding the information and consent requirement, confidentiality and data protection [29]. Participants were informed that participation was completely voluntary, and without affecting any eventual further contact with the neonatal care unit and that participation could be discontinued at any time without cause. All participants signed a consent form. If families during the interviews were perceived to have a need to further process their experiences from the neonatal care was a counsellor at each neonatal unit available. The study was approached by The Advisory Committee for Research Ethics in Health Education, Department of Health and Sciences, Faculty of Medicine, Lund University, Sweden, (53-09).
**RESULTS**
From the content analysis of interview data following categories and theme emerged. The four categories identified for information and communication were: Initially Conflicting Emotions in relation to KC, Preparation and Confidence in KC is Evolving, Strengthening Preparation and Context is Decisive, and finally Parental sense and caution. The overall theme was: Good verbal and practical preparation will contribute to a positive experience of KC.
**Initially Conflicting Emotions in Relation to KC**
When labour unexpectedly started too early, the neonatal nurse, who was called to the maternity ward, sometimes provided information on KC before birth. Such information gave conflicting emotions because it gave a sense of being prepared for what would happen when the infant was born. It was also perceived as unimportant information because the parents were only interested in whether the infant would survive, or not.
"...My experience told me this [KC] was not right ... so before caesarean section [in the meeting with a neonatal nurse], I was worried about it ..." (F8, Dad).
All parents did not remember if they received any information about KC. Initially it was given so much information, making it impossible to remember everything.
Some parents initially upon arrival at the neonatal unit, got information about that KC was a method that was provided at the unit and that it involved that parents had to
| Meaning Unit | Condensation | Code | Category | Overall Theme |
|------------------------------------------------------------------------------|-------------------------------------------------------------------------------|-------------------------------------------|--------------------------------------------------------------------------|--------------------------------------------------------------------------------|
| Ah, at first they said, "Yes, but now it's probably time to start kangaroo care" So ... they said only that it was just a good thing ... we should do it as they did it in the ward ... then I thought what? What is it? | Got the call to start kangaroo care but did not understand what it was. Would do it because it seemed to be a good thing to do. | No information and no communication | Gradually participation and confidence in KC was evolving | Good verbal and practical contribution will contribute to a positive experience of KC |
| ... it was not that much, if one considers, explanation of kangaroo care. I remembered that I had seen a Kangaroo Care picture on the wall in the ward. It was just a natural image of baby's skin to skin with the parent - it felt right and I thought it was natural and not wrong. | Not much explanation about what KC meant or was good for. The parent had seen a picture with the heading KC in the ward. Found KC as a natural way of caring for the infant. | No information and Communication | Gradually participation and confidence in KC was evolving | Good verbal and practical contribution will contribute to a positive experience of KC |
| It was almost as if I wanted to ask, but did not ask ... maybe I was forced to think KC because they are working with it all the time, or because they were so busy | The parents wished they had got a question about if they knew what KC was. The parents had patience with the lack of communication. | No communication | Strengthening preparation and context is decisive | Good verbal and practical contribution will contribute to a positive experience of KC |
have their infant skin-to-skin on their chest. After a few days they got more in-depth information including for example KC’s promotion of the infant’s natural temperature regulation, the parents-infant connection, and the mother’s breast milk production. The parents experienced that they gradually became familiar with KC and got their questions answered leading to a desire to be involved in the care of their infant. With this information parents experienced that KC was quite natural i.e. just as they had planned to care for their newborn infant.
“[Information about] KC … did that KC seemed quite natural … just right … not something strange for us to do …” (F2, Mom and Dad).
Some parents were convinced that they had not received any information about KC, but merely an insistence on holding their infant skin-to-skin on their bare chest. The parents wanted to get close to their infant but there were conflicting feelings towards the abrupt way to approach the infant. The experience was that the staff nurses did not understand the parent’s concern about whether the infant would survive and the concern about the preterm birth.
“No, but it was uh … just so you can take your clothes off, so you can put on this shirt, you should have her skin-to-skin …” (F11, Dad).
“… It was a bit terrifying and I thought shall we take him out [of the incubator]? But of course that is what you want as a mother; you want to take him out. But it’s just that the shock and so, can fold first …” (F10, Mom).
One couple of parents recalled that they were involved in decision making about providing KC for their infant, while the other parents could not remember if they got such a question, or not. When information about the KC was given without the staff nurses being able to answer parent’s questions the feeling appeared that it was a method that was “in” at the moment.
“… seemed a bit like one of those trendy thing, as well … as a recent weight loss method … here we go Atkinson now we are doing something else and so here we go … no without mentioning that KC was conceived and tested. Important to obtain this information as well …” (F2 Mom and Dad).
**Participation and Confidence in KC is Evolving**
The parents had experienced that KC was a nursing practice that staff nurses at the various hospitals applied. Satisfaction with KC was highest among parents in one of the hospital where the parents had experienced that KC was well established among all staff nurses i.e. all were familiar with the method’s context, benefits and practise. The verbal information was given with enthusiasm and the information and communication were fully consistent, no matter which staff nurse talked with parents, and the parents got answers to their questions. The parents also received written information about KC and there were posters about KC on the unit.
“… All [staff] had the same attitude towards it [KC] … and it’s important for us parents …” (F2, Dad).
At the other two hospitals the staff nurses were perceived as not consistent in terms of information and provision of KC. The information and communication was poor and the nursing staff’s focus had been that the parents would have their infant skin-to-skin with no further explanation. The experience was that each of the staff nurses had their way of working with KC. When the parents asked questions about KC they received vague and conflicting answers. The staff nurses’ approach was tolerated of the parents and KC was provided, on the grounds that there were many infants on the ward and the staff nurses seemed overworked. Another reason for the tolerance was that it was vacation time and the parents concluded that agency nurses were on service, so hence the conflicting communication. Despite the staff nurses’ ignorance of KC all parents experienced great confidence to the staff nurses when it came to the medical and nursing care of their infant. The staff invited the parents to participate in the care by performing tube feeding, diaper change, turning the infant in the bed/incubator, getting the infant to settle down, etc. and this was very satisfying for the parents.
“… There was no consistent approach [in the information and communication of KC] … or a common way of thinking … it was a lot of thoughts and thinking … conflicting advice … it came through of the kangaroo method …” (F6, Mom).
Sometimes the staff nurses were aware of that parents might have received vague and conflicting information and communication about KC. In this situation the parents were recommended to converse with the contact nurse assigned to them.
“… And that’s a good thought … that you stick to a few people … so that it is not such a mess …” (F8, Mom).
**Strengthening Preparation and Context is Decisive**
Parents’ sense of KC was strengthened when the staff nurses communicated with the parents before the KC holding by preparing the infant so that KC would become a quiet moment, as the parents wished for. Above all, the mothers wanted help with the preparations of the infant, when KC was first implemented. When not receiving any information about the preparations the parents observed how the staff was preparing the infant and learned to do it by themselves. The preparations that were considered important were that the infant should not require a diaper change or examinations during the time allocated to KC due to a desire to get privacy during KC.
Parents experienced a disappointment over not receiving information about KC’s important physiological effects regarding to prevent complications for the infant. They demanded that if there was no time for verbal information and communication about KC, there should be written information. Such information also would have facilitated
the parents understanding of what was expected of them as parents.
"... When they [the staff] take it [KC] seriously ... it helps of course that I take it seriously too and not just because it is cosy, but because it is medically important as well ..." (F12, Mom).
In order to make KC to a positive moment for the parents they had wished that the staff had planned KC carefully as KC sometimes included several hours of holding the infant. They wanted information about the importance of having eaten before, been to the toilet, having something to drink at hand and above all a comfortable chair to sit in, which was not always available. Some fathers thought it was tedious to implement KC for several hours at a stretch, but solved it by reading a book or going to sleep.
The information and communication was at times directed only to the mother and the father had felt excluded. These fathers described it as the staff minimized the father’s role. When the information and communication was addressed to both parents fathers felt that they were important too for the infant.
" ... KC gave a special feeling ... one became more a part of it all ... yes it was special ... for me as a dad to got to be like that ... all by yourself ... you did something ... than if we just sat and slept [during the KC implementation] ... " (F11, Dad).
Some families received care at more than one hospital due to that the infant was born when the family was not in their home community at the time delivery or the infant needed treatment at a level IV hospital. When the infant later on was transferred to the neonatal unit to which the family’s place of residence belonged, the parents experienced that the staff nurses at that hospital had other routines for KC. No information was given of how the routines for KC were at that hospital and there was no communication regarding the parents KC provision prior to transfer. Parents lost their own faith to provide KC in a sense of insecurity when staff nurses did not confirm or did confirm parents’ KC provision.
" ... no one said anything ... so directly [about KC]... but we felt ... we’re doing something wrong ... it felt pretty uncomfortable ..." (F3, Mom).
**Parental Sense and Caution**
The shock and sadness of not having carried the infant full term, was tough for most mothers. When maternal connection to the infant after birth, did not emerge, that led to self-blame, depression and guilt. For these mothers KC was not always a positive and maternal connection did not appear. The KC-time instead became a routine, which only was done almost mechanically. These mothers were yet of the opinion that KC had been a support for them. Maternal feelings evolved gradually.
"If all the emotions are not there then you just cannot bother at all times ... I did not love my child as I do now ... that how it was ... it [KC] became routine in fact ..." (F3, Mom).
"... It [KC] was probably a good deal because you had to sit close ... was left in peace ...presumably it became apparent [that it was my infant] ..." (F12, Mom).
The parents were surprised that they could have their infant skin-to-skin as soon after birth. When they provided KC for the first time most often a feeling appeared that they had become parents. The parents wanted to share this feeling and expressed that all new parents would have the opportunity to KC.
" ... Everyone should be supposed to come in to the neo-unit [laughs] a turn [and sit kangaroo] before going home [laughs] because that is all positive ..." (F1, Mom).
Parents felt that the small infant was fragile and the medical devices that were linked to the infant were frightening. They feared to harm the infant during KC holding, which was manifested in being reluctant to move their body with the risk that tubes would disconnect from the infant’s body or that monitor devices would start alarming and cause distress to the infant. Along with each medical device that was discontinued it became cosier and calmer to provide KC. It was important that there were staff available if something would happen.
"I [the nurse] am sitting here and you [the mother] may be very calm, because this cannot go wrong ..." (F6, Mom).
"To hold my baby [KC] for the first time for real and like to feel his warmth, you felt his fingers, he like touched one ... then you become so happy ... my tears flowed almost constantly during the skin-to-skin holding ..." (F4, Mom).
Most of the mothers experienced KC as more positive if the father was also present, because then it became an intimate family moment, on the other hand a few mothers experienced that the feeling of closeness to the infant was strengthened if the father was not present. One father had applied KC first when the birth was by Caesarean section and twins were born, where one infant’s condition was stable and KC could be applied before the mother recovered. In this situation the mother experienced dissatisfaction with that the father initially had a deeper connection to the infant then the mother had.
**The Overall Theme**
*Good Verbal and Practical Preparation will Contribute to a Positive Experience of KC*
The parents lived in chaos and found it difficult to take in information and communicate with the issues in terms of KC. As information and communication on KC was clear, given with commitment and the staff nurses were safe and unanimous in applying the method, the experience of KC resulted in a strongly positive experience. Providing KC was perceived by most parents as a natural way to embrace the infant because they longed to get close to their infant. When staff was perceived to have limited knowledge about the method and did not provide consistent information and were not able to answer parents’ questions about KC parents
applied KC anyway and the feeling of KC was even then positive. Some parents did not immediately have a positive experience of KC as parental feeling did not appear during KC holding. The feeling was that the parents needed greater attention and guidance from the staff to gradually overcome their ambivalent feelings for their infant. If KC would be perceived positively or not was dependent on the infant’s medical condition, the father’s presence or not presence at the ward, his support of the mother as well as that the staff was available during the KC provision.
**DISCUSSION OF RESULTS AND CLINICAL IMPLICATIONS**
This study examines parents’ experiences of information and communication mediated by health care staff before and during KC at neonatal units. Investigating this is important as it may differ from what the staff nurses believe that parents want and need, and what parents really want and need for being able to provide KC.
When information about KC is given before birth, conflicting emotions emerge given that parents are in shock and their anxiety over whether their infant will survive is the focus for the parents. Most importantly in this moment is that the staff nurses show sensitivity to the vulnerable situation that parents find themselves in. Similar results are shown in Lundqvist, Hellström and Hallström [6] where the fathers of preterm infants, who are cared for in NICU, find it difficult to take in information about the care of their infant as they found themselves living in an unreality when the infant was born preterm. All parents in our study do not remember if they got any information about KC at all.
At one of the hospitals, parents experience that there is a clear strategy and implementation of KC and the parents feel safe with KC and feel trust in staff nurses. The staff seem to have received formal training as all information and communication are consistent between staff and it is handed out leaflets to parents together with that there are posters in the department showing that KC is implemented at the unit. In our study the parents experience are that preparation strengthening the parental role as well as that the environment is prepared for KC to make it possible to come close to the infant. WHO [9] state that every health care facility that implements KC should develop written policy and guidelines that incorporate clear criteria for selection, monitoring and evaluation. Nykvist *et al.* [10] describe the evidence based recommendations for KC regarding knowledge and skills for staff nurses and what the information and communication should include to encourage and motivate parents to provide KC. MacCleary and Brown [30] suggest that for developing joint strategies for care in a hospital, there must be an interest from hospital management to support the staff nurses so that all staff are involved in achieving consistent caring.
All parents in our study provided KC in spite of that in two of the hospitals there is no common policy or procedures for KC. The parents experienced that all staff nurses have their own way to provide the KC. According to Higgs *et al.* is “professional practice without theory underpinning a guesswork” as parents in our study to some extent experience [31, p. 4]. The parents need gradually participation and confidence in KC before providing KC. The parents providing KC without these needs, being met in the yearning to get to know their infant but with a fear that the infant could be harmed or exposed to suffering. Through KC the parents experienced connection to their infants that also is confirmed in many other studies [12-19]. It is obvious that although the parents value the KC experience in our study they need attention and assistance from staff nurses to alleviate any of their anxieties and enhance the development of their parental role [12-16, 19, 21, 32, 33]. In the Irish quantitative study by Flynn & Leahy-Warren [32] including 75 neonatal nurses who were implementing KC in their practical work and had a positive belief about the method, it was showed that none of the nurses had received any formal training or education in KC. In spite of the fact that all hospitals in our study have implemented KC as a caring method it seems as only staff nurses in one of the hospitals had received formal and clinical training in KC, based on parents’ statements that the staff had infused the feeling that KC was a substitute for the uterine enclosing and life-giving environment. In retrospect, the lack of staff training in KC leads to that parents feel that if they had received information about how important KC was for the infant, both medically and developmental, they would have demanded to provide KC more regularly. At least it should have been written information when the staff nurses seemed to have difficulty getting time with KC preparations.
When the infant is transferred to another hospital and the parents have provided KC in the first hospital, parents experience that KC has not the same routine at the “new hospital”. The staff nurses do not gradually prepare the parents neither inform about the procedures for KC or communicate with the parents about how they have provided KC. Parents felt uncomfortable and lost their faith to provide KC in a sense of insecurity but did not communicate their feelings to the staff nurses. This phenomenon is interpreted by Lundqvist and Nilstun [34] as that when parents find themselves in a vulnerable position they do not stand up for themselves or their infants, but adapts to the staff’s attitude, even if the attitude is offensive and should not be accepted in a different situation.
A few mothers experienced alienation to the infant in connection with the KC and provided the KC as a routine that was performed on the staff’s request. Tessier *et al.* [35] also found that some mothers experienced isolation and a feeling of too much responsibility when providing KC. In Thernström Blomqvist & Hedberg Nykvist [15] study had no mothers had this feeling, and the researchers believed it may be due to that the father in their study shared the task of KC with the mothers. The experience of alienation may also be due to that it is not uncommon that mothers of preterm infants have trouble bonding with their infants [36]. This is something that staff nurses need to be aware of.
One mother expressed some concerns about that the father provides KC before the mother in a sense of that the father got a deeper connection to the infant than the mother. Fegrán, Helselö and Fagermoen [37] confirm that in such cases the fathers get a relationship with the infant and a protective instinct. This issue has to be further examined regarding if it has consequences for the mother-child attachment.
Clinical implications for health care is that comprehensive education based on up to date research evidence detailing the benefits and risks of KC could be incorporated into new staff induction, and continuing education. Organisational support is needed to continuing education and to develop practice guidelines and protocols, to ensure the standardisation of information for parents and staff.
DISCUSSION OF METHOD
Regarding trustworthiness the parents narrated vividly and richly about their experiences and the interviewers checked their understanding with follow-up questions. To add depth and insight to the interpretation, the findings were discussed at research seminars. Quotes from the interviews are used to verify that the findings were retrieved from the data. The initial limitation of the study is that it is a retrospective study and it is found that parents have difficulties remembering information and communication about KC due to that they are shocked after giving birth to a preterm infant. A prospective study in the form of an observational study might provide what is really going on related to both the information and communication the staff provide and what families are experiencing. During the analysis the authors realised that some interviews, to a great extent, was about the delivery that was not in agreement with the aim of the study. The main reason for this was the insufficient experience of the interviewers in carrying out interviews. However, in total the whole texts from the interviews included lots of meanings of the studied phenomenon. The seven families with Swedish origin and the other five families consisting of one parent originating from another European country can be an advantage as Sweden is a multicultural country. Another shortcoming of the study is that the authors became more aware of the education that each neonatal unit had regarding the KC. A survey of training programs in the KC on neonatal units in Sweden needs to be explored.
CONCLUSION
For parents who participated in the study, the motive for practising KC was the desire to get close to their infant. Most of the parents had positive experience of KC. The information and communication from the staff gave courage and motivation for the parents to practice KC when the staff was well versed in the method and was coherent and supportive but gave conflicting emotions when the staff practised KC as a routine without deeper insight into the method's advantages and application. The challenge is to encourage neonatal staff nurses and hospital management to develop a written policy and guidelines that incorporate knowledge and skills to facilitate KC and give accurate and supportive information and communication to parents. Another challenge is to survey the education and training program that neonatal units offer the neonatal staff nurses in Sweden about KC.
CONFLICT OF INTEREST
The authors confirm that this article content has no conflict of interest.
ACKNOWLEDGEMENTS
The authors wish to thank the parents who participated in the study and devotedly shared their experience.
REFERENCES
[1] Socialstyrelsen, 2011a. Graviditeter, födselssingar och nyfödda barn Medicinska fädeseregistret 1973-2009 - Assisterad befruktning 1991-2008. Stockholm: Socialstyrelsen (sid. 44). [In English: National Board of Health and Welfare, 2011a. Pregnancies, births and newborns – Medical Birth Register, 1973-2009 - Assisted reproduction 1991-2008. Stockholm: National Board of Health and Welfare (p. 44)].
[2] Jacob S, Wodjina D, Kitz I, et al. The worldwide incidence of preterm birth: a systematic review of maternal mortality and morbidity. Bull World Health Organ 2010; 88: 31-8.
[3] Socialstyrelsen, 2011b. Sveriges officiella statistik. Hälso- och sjukvård Publiceringsår 2011. Bilaga 1: Graviditeter, födselssingar och nyfödda barn Medicinska fädeseregistret, 1973-2009. Stockholm: Socialstyrelsen. [In English: National Board of Health and Welfare, 2011a. Official statistics of Sweden. Health care. Publication 2011th Appendix 1 pregnancies, births and newborns. Medical Birth Register, 1973-2009. Stockholm: National Board of Health and Welfare, 2011a.] Available from: http://www.socialstyrelsen.nu/publikationer/2011_2011-1-1-1/Dokument/Bilaga1.pdf.
[4] Danerek M, Dykes A-K. A theoretical model of parents' experiences of threat of preterm birth in Sweden. Midwifery 2008; 24(4): 416-24.
[5] Lundberg B, Öhrlin K. Experiences of having a prematurely born infant from the perspective of mothers in northern Sweden. Int J Circumpolar Health 2008; 67(5): 461-71.
[6] Lundqvist P, Westas LH, Hallström I. From distance toward proximity: Fathers' lived experience of caring for their preterm infants. J Pediatr Nurs 2003; 22(6): 490-7.
[7] Flackling M, Ewald U, Nykvist Holmberg K, Sturini B. Trustful bonding - A key to becoming a 'mother' and to reciprocal breastfeeding. Stories of mothers of very preterm infants at a neonatal unit. Soc Sci Med 2006; 62: 70-80.
[8] Wigert H, Johansson R, Berg M, Hellström AL. Mothers' experiences of having their neonatal child in a neonatal intensive care unit. J Clin Nurs 2006; 15(4): 421-9.
[9] WHO. Kangaroo Mother Care: A practical guide. Geneva: World Health Organization 2003.
[10] Nyqvist KH, Andersson GC, Bergman N, et al. Towards universal Kangaroo Mother Care: recommendations and report from the First European conference on Supportive Interventions Workshop on Kangaroo Mother Care. Acta Paediatr 2010; 99(6): 820-6.
[11] Cattanco A, Davanzo R, Uxa F, Tamburini G. Recommendations for the implementation of Kangaroo Mother Care for low birthweight infants. International Network on Kangaroo Mother Care. Acta Paediatr 1998; 87(4): 440-5.
[12] Yuen M. Parents' experiences of skin-to-skin care with their preterm infants requiring assisted ventilation. J Obstet Gynecol Neonatal Nurs 1999; 28(2): 157-64.
[13] Roller CG. Getting to know you: mothers' experiences of Kangaroo Care. J Obstet Gynecol Neonatal Nurs 2005; 34(2): 210-6.
[14] Leonard A, Mayers P. Parents' lived experience of providing kangaroo care to their preterm infants. J Interdiscip Health Sci 2008; 13(4): 16-28.
[15] Thernström BY, Hedberg NK. Swedish mothers' experience of providing Kangaroo Mother Care. J Clin Nurs 2010; 20(9–10): 1473-81.
[16] Thernström BY, Rubenson C, Kyberg E, Jörnskog K, Hedberg NK. Kangaroo Mother care helps fathers of preterm infants gain confidence in the parental role. J Adv Nurs 2011; 68(9): 1988-96.
[17] Dalbye R, Calais E, Berg M. Mothers' experiences of skin-to-skin care of preterm high-risk infants – A phenomenology study. Sex Reprod Healthc 2011; 2: 107-11.
[18] Tallandini MA, Salembra C. Kangaroo Mother Care and motherpremature infant dyadic interaction. Infant Ment Health J 2006; 27: 251-75.
[19] Batorski Johnson A. The maternal experience of kangaroo holding. J Obstet Gynecol Neonatal Nurs 2007; 36: 568-73.
[20] Erlandsson K, Christensson K, Fagerberg I. Fathers' lived experiences of getting to know their baby while acting as primary
caregivers immediately following birth. J Perinat Educ 2008; 17: 28-36.
[21] Johnston CC, Stevens B, Piselli J, et al. Kangaroo care is effective in diminishing pain response in preterm neonates. Arch Pediatr Adolesc Med 2003; 157: 1084-8.
[22] Ludington-Hoe SM, Johnson MW, Morgan K, et al. Neurophysiological assessment of neonatal sleep organization: preliminary results of a randomized, controlled trial of skin contact with kangaroo mother care. Sleep 2009; 32(1): e909-23.
[23] Flackling R, Ewald U, Wallin L. Positive effects of kangaroo mother care on long term breastfeeding in very preterm infants. J Obstet Gynecol Neonatal Nurs 2011; 40(2): 190-7.
[24] Lawn EJ, Mwansa-Kamhafwile J, Horta BL, Barros FC, Cousens S. "Kangaroo mother care" to prevent neonatal deaths due to preterm birth complications. Int J Epidemiol 2010; 39(Suppl 1): 1144-54.
[25] Lincoln YS, Guba EG. Naturalistic inquiry. Canada: Newbury Park, Sage Publications 1985.
[26] Socialstyrelsen 2004. Perinatal omhändertagande vid underburenhet. [Article number: 2004-123-15]. (In English: National Board of Health and Welfare. Perinatal caring of preterm infants 2004. Article number: 2004-123-15).
[27] Graneheim UH, Lundman B. Qualitative content analysis in nursing research: concepts, procedures and measures to achieve trustworthiness. Nurse Educ Today 2004; 24(2): 105-12.
[28] Krippendorff K. Content analysis. An introduction to its methodology. London: The Sage Commex Series, Sage Publications Ltd. 1980.
[29] World Medical Association Declaration of Helsinki [WMA] 2008. Available from: http://www.wma.net/en/30publications/10policies/b3/17c.pdf
[30] Hughes L, Brown G. Barriers to paediatric nurses' research utilization. J Adv Nurs 2003; 42(4): 364-72.
[31] Higgs J, Titchen A, Neville V. Professional practice and knowledge. In: Higgs J, Titchen A, Eds. Practice knowledge and expertise in the health professions. Oxford: Butterworth Heinemann Ltd. 2001; p. 3-9.
[32] Flynn A, Leahy-Warner P. Neonatal nurses' knowledge and beliefs regarding kangaroo care with preterm infants in an Irish neonatal unit. J Neonatal Nurs 2010; 14: 221-8.
[33] Moran M, Radzyniinski SG, Higgins KR, Dowling DA, Miller MJ, Anderson GC. Maternal kangaroo (skin-to-skin) care in the NICU beginning 4 hours postbirth. MCN Am J Matern Child Nurs 1999; 24(2): 74-9.
[34] Lundqvist A, Nilstun T. Human ethics in paediatrics: the effects in health care. Nurs Ethics 2007; 14(2): 215-28.
[35] Tessier R, Cristo M, Velez S, et al. Kangaroo mother care and the bonding hypothesis. Pediatrics 1998; 102(2): e 17.
[36] Noren C, Wiznurwski C, AlRoumi L. Attachment (bonding) in mothers of pre-term babies. J Reprod Infant Psychol 1993; 11: 175-85.
[37] Fegrán L, Helseth S, Fagermoen S. A comparison of mothers' and fathers' experiences of the attachment process in an neonatal intensive care unit. J Clin Nurs 2008; 17(6): 810-6. |
Fluctuating open heavy flavour energy loss in a strongly coupled plasma with observables from RHIC and the LHC
B A Ngwenya$^1$ and W A Horowitz$^1$
$^1$Department of Physics, University of Cape Town, Rondebosch 7700, South Africa
E-mail: email@example.com, firstname.lastname@example.org
Abstract. Heavy ion collisions at RHIC and at the LHC produce an enormous amount of energy that enables the nuclei and its constituent particles to melt, thus releasing gluons, quarks and antiquarks, travelling in different directions with different momenta. Studies of these collisions have shown that low transverse momentum observables describe a strongly coupled plasma (quark-gluon plasma), an almost perfect liquid that evolves hydrodynamically and flows with almost no viscosity [1, 2]. We make predictions for the suppression of these heavy quarks and thus describe the energy loss of the heavy quarks as they interact with the plasma; we show that these predictions are in good agreement with experimental data.
1. Introduction
One of the big questions of Physics is ‘what happened shortly after the big bang?’ Theory predicts that the early universe was composed of a hot mixture of particles (mainly weakly bound quarks and gluons) moving at nearly the speed of light [3]. Powerful particle accelerators can be used to recreate conditions of the early universe through colliding heavy ions. RHIC collides beams of $^{197}Au$ nuclei, each with a total energy of approximately 17 TeV [4] while the LHC collides $^{208}Pb$ nuclei with a centre of mass energy of approximately 1000 TeV [5].
Through these heavy-ion collision experiments, we can quantitatively extract the properties of nuclear matter and through some theoretical predictions, we can understand the properties of this nuclear matter better with the ultimate goal being to construct the phase diagram of nuclear matter [3]. This is a complicated task given that emergent phenomenon is not well understood practically, for example, we cannot predict the behaviour of a single water molecule (in a bucket of water) by simply applying Newton’s laws and Maxwell’s equations to it.
In these heavy-ion collisions, the temperatures of the material produced is in the order of a trillion Kelvin [3, 6, 7] and the best way of understanding it theoretically is yet to be known. This material is a new phase of matter called the quark-gluon plasma (QGP) [3] and is produced due to the very high temperatures. It is a state of strongly interacting matter where quarks and gluons are no longer confined to the colour-neutral hadrons [8]. In the formation of QGP, there has to be a phase transition and this comes as a natural consequence of the composite nature of hadrons in quantum chromodynamics (QCD) [3]. This phase transition has been found to
be at a critical temperature $T_c \simeq 170 MeV$, with an energy density $\epsilon_c \simeq 600 MeV/fm^3$ [3, 9]. Hadronization starts to occur almost immediately after the collision, so the QGP has a very short lifetime (on the order of 4fm/c) at RHIC and on the order of 10fm/c at the LHC [3]. QGP can be understood better by looking at its coupling strength.
The interactions in Quantum Chromodynamics (QCD) are strong at low energy and decrease at high energy (asymptotic freedom) [10]. As a result, the strong interactions between quarks, antiquarks and gluons persist in the QGP and the dominant degrees of freedom in the QGP are thus light quarks, antiquarks and gluons [3]. Due to asymptotic freedom, the interaction strength weakens when large energies are exchanged in inter-particle collisions and interactions with a large momentum transfer can be treated in a perturbative way [11]. This asymptotic freedom regime is achieved when the temperature, $T \gg \Lambda_{QCD}$ (where $\Lambda_{QCD} \cong 200 MeV$ is the QCD scale parameter) and the QGP is a weakly-interacting gas of slightly modified quarks and gluons that yields a plasma relatively transparent to hard probes[11] and calculations are performed using perturbative quantum chromodynamics (pQCD) [12, 13].
We’re interested in the high transverse momentum particles because they are decay products of high transverse momentum partons and these are the most direct probe of the relevant degrees of freedom in a quark-gluon plasma [12, 14]. In the low momentum observables, QGP appears as a strongly coupled plasma that evolves hydrodynamically [15] and has almost no viscosity, making it the most perfect liquid observed. In this strong coupling regime, non-perturbative approaches such as AdS/CFT [16] need to be used to perform calculations and in this paper, we look at the energy loss of heavy flavour strongly-coupled to the medium.
Heavy flavour is more interesting because it puts more experimental constraints on the energy loss model and as a result, on the potential properties of the quark-gluon plasma [12]. It is important to compare our theoretical predictions to a wide range of experimental data, for example, by looking at the suppression of heavy flavour at different energies (i.e RHIC and the LHC) through the nuclear modification factor [13]. Some early results of the energy loss in the higher order strong coupling regime (AdS/CFT calculations) have shown favourable results of the measured nuclear modification factor $R_{AA}(p_T)$ of electrons from heavy flavour decay at RHIC [12, 17] but generally over-suppressed $R_{AA}(p_T)$ for D mesons at the LHC by a factor of approximately 5 [12, 18].
In this paper, we compute the nuclear modification factor ($R_{AA}(p_T)$) for bottom quarks at 5.5TeV and thus quantitatively describe the suppression of these heavy quarks at high transverse momentum.
2. Particle Geometry with the Optical Glauber Model
The Glauber model [19] is used to model the geometry of nuclei before a heavy-ion event. The Optical limit approximation of the Glauber model assumes that at high energies, the nucleons carry a sufficiently large momentum that they will be undeflected as nuclei pass through each other. As a result, for calculations, the nucleus is assumed to comprise of a smooth/continuous nucleon density ($\rho$). Assuming a spherical nuclei, the nucleon charge density (inside the nucleus) is given by the Woods-Saxon distribution [19]
$$\rho(r) = \frac{\rho_0}{1 + e^{(r-R)/a}}$$ \hspace{1cm} (1)
where $\rho_0$ is the nucleon density in the centre of the nucleus, $r = \sqrt{x^2 + y^2 + z^2}$, R is the nuclear radius and a is the "skin depth". For $^{208}Pb$ these parameters are; $R = 6.624 \pm 0.035 fm$ and $a = 0.549 \pm 0.008 fm$ respectively [20].
Perturbative QCD calculations are only valid for transverse momentum, $p_T \geq 1 GeV/c$ [21] and thus can’t be used to determine the inelastic nucleon-nucleon cross section ($\sigma_{inel}^{NN}$) since the cross section involves processes with low momentum transfer (diffractive and elastic processes). As a result, the model takes in the experimental measured cross section data and this provides the only nontrivial beam-energy dependence for Glauber calculations. Table 1 gives the inelastic nucleon-nucleon cross section at collision energies appropriate for RHIC and the LHC.
Table 1: Values of the nucleon-nucleon inelastic cross section ($\sigma_{inel}^{NN}$) for collision-energies ($\sqrt{s}$) appropriate for RHIC and the LHC [20]
| $\sqrt{s}(TeV)$ | $\sigma_{inel}^{NN}$ (mb) | $\sigma_{inel}^{NN}$ (fm$^2$) |
|-----------------|--------------------------|-------------------------------|
| 0.2 | 41.6 ±0.6 | 4.16 |
| 0.9 | 52.2 ±1.0 | 5.22 |
| 2.76 | 61.8 ±0.9 | 6.18 |
| 5.02 | 67.6 ±0.6 | 6.76 |
| 5.44 | 68.4 ±0.5 | 6.84 |
| 5.5 | 68.5 ±0.5 | 6.85 |
We consider two heavy-ions (target A and projectile B) colliding at relativistic speeds with impact parameter b. We focus on two flux tubes located at a displacement $(x - b/2, 0, 0)$ with respect to the center of the target nucleus and a displacement $(x + b/2, 0, 0)$ from the center of the projectile. During the collision these tubes overlap. The probability per unit transverse area of a given nucleon being located in the target/projectile flux tube is given by equation 2, while the joint probability per unit area of finding nucleons located in the respective overlapping target and projectile flux tubes is given by what is defined as the thickness function (equation 3).
$$T_{A/B}(x, y) = \int_{-\infty}^{\infty} \rho(x, y, z_{A/B}) dz_{A/B}$$
$$T_{AB}(b) = \int T_A \left( x - \frac{b}{2}, y \right) T_B \left( x + \frac{b}{2}, y \right) dxdy$$
where $\rho(x, y, z_{A/B})$ is the probability per unit volume (normalised to unity), of finding a nucleon at a point $(x, y, z_{A/B})$ in the nucleus of projectile (A) or target (B).
We compute the total number of binary nucleon-nucleon collisions at impact parameter b (equation 4), the number of participants, which is the number of nucleons in the target and projectile nuclei that interacted at least once in a collision (equation 5) as well as the total geometric cross section (equation 7).
$$N_{coll}(b) = AB T_{AB}(b) \sigma_{inel}^{NN}$$
$$N_{part} = A \int T_A^{-} (1 - [1 + T_B^{+}]^B) dxdy + B \int T_B^{+} (1 - [1 + T_A^{-}]^A) dxdy$$
\[ T_X^\pm = T_X \left( x \pm \frac{b}{2}, y \right) \]
\[ \frac{d\sigma}{db} = 2\pi b (1 - [1 - T_{AB}(b)\sigma_{inel}^{NN}]^{AB}) \]
(a) \( N_{coll} \) and \( N_{part} \) as a function of impact parameter
(b) Total geometrical cross section as a function of impact parameter
Figure 1: Some geometric quantities in the optical limit of the Glauber model for Pb-Pb at 5.5TeV
3. Langevin Energy Loss
In the strong coupling regime, the dynamics of heavy quarks interacting with QGP is described by a stochastic differential equation known as the Langevin equation. In the fluid’s rest frame, this equation is given by [22, 23]:
\[ \frac{dp_i}{dt} = -\mu p_i + F_i^L + F_i^T \]
\[ \mu = \frac{\pi \sqrt{\lambda} T^2}{2M_Q} \]
where \( p^i \) is the three-momentum of an on-shell heavy quark moving at constant velocity in a thermal bath, \( \mu \) is the drag loss coefficient of a heavy quark, \( M_Q \) is the mass of the heavy quark in a plasma of temperature \( T \) and \( \lambda \) is the 't Hooft coupling constant. \( F_i^L \) and \( F_i^T \) are longitudinal and transverse momentum kicks with respect to the quark’s direction of propagation. The energy loss model is described in detail in [23, 24] and is the first of its kind to include thermal fluctuations. The fluctuating momentum kicks are correlated as [24]
\[ <F_i^L(t_1)F_j^L(t_1)> = \kappa_L \hat{p}_i \hat{p}_j g(t_2 - t_1) \]
\[ <F_i^T(t_1)F_j^T(t_1)> = \kappa_T (\delta_{ij} - \hat{p}_i \hat{p}_j) g(t_2 - t_1) \]
where \( \hat{p}_i = p_i/|\vec{p}| \) and \( g \) is a function only known numerically,
\[ \kappa_T = \pi \sqrt{\lambda} T^3 \gamma^{1/2} \]
\[ \kappa_L = \gamma^2 \kappa_T \]
The longitudinal direction of the heavy quark is the most important one for calculations of suppression observables and the detailed energy loss model is given by [24]. As mentioned earlier, our energy loss model requires the heavy quark to be moving at a constant velocity, as a result, we need to provide the quark with power to compensate for the momentum lost. Due to the restrictions around that power provided to the heavy quark, we end up with a speed limit on the heavy quark set-up given by,
\[
\gamma < \gamma_{crit}^{sl} = \left(1 + \frac{2M_Q}{\sqrt{\lambda T}}\right) \sim \frac{4M_Q^2}{\lambda T^2}
\]
(14)
4. Results
The main result of this paper is shown in Figure 3: which shows predictions for the nuclear modification factor ($R_{AA}(p_T)$) for bottom quarks at the LHC. We see a clear suppression of open heavy flavour at high transverse momentum for the various centrality classes studied, given with statistical uncertainties from the transverse momentum bins. It is also clear that the suppression is more pronounced for central collisions.
5. Conclusion and Outlook
Heavy flavour energy loss is crucial in understanding the properties of nuclear matter and thus trying to put together the phase diagram of nuclear matter. The Langevin energy loss model used in this paper is the first in the formulation of AdS/CFT correspondence to include thermal fluctuations and has been shown to be a success in computing several quantities from heavy ion collisions [15, 24, 25]. The results presented in this paper have shown that heavy flavour is largely suppressed for high transverse momentum.
The next steps of this work will be to look at the hadronization process described in [26] and compute predictions for the suppression of B mesons that these heavy quarks decay to, then compare the results to experimental data. Using this energy loss model, we will also look at the suppression of charm quarks and thus the suppression of D mesons. We will then be able to make predictions for higher energies of the upcoming runs of the LHC and potentially the FCC.
Acknowledgments
The authors would like to acknowledge Valumax Projects (Pty) Ltd, SA-CERN, the South African National Research Foundation and the University of Cape Town for their generous financial contributions towards this work.
References
[1] Riordan M and Zajc W A 2006 *Scientific American* **294** 34A–41
[2] Johnson C V and Steinberg P 2010 *Phys. Today* **63** 29–33
[3] Hands S 2001 *Contemporary Physics* **42** 209–225
[4] Schaefer T 2003 *Physics world* **16** 31
[5] Jowett J and Carli C 2006 *Proceedings of the 2006 European Particle Accelerator Conference, Edinburgh, Scotland* pp 550–552
[6] Shen C, Heinz U, Huovinen P and Song H 2011 *Physical Review C* **84** 044903
[7] Gale C, Jeon S, Schenke B, Tribedy P and Venugopalan R 2013 *Physical Review Letters* **110** 012302
[8] Satz H 2011 *arXiv preprint arXiv:1101.3937*
[9] Pumplin J 2002 *J. High Energy Phys.* **2002** 012
[10] Ho-Kim Q and Pham X Y 1998 *Elementary Particles and Their Interactions* (Springer) pp 505–548
[11] Mrowczynski S 1999 *arXiv preprint nucl-th/9905005*
[12] Horowitz W 2013 *Nuclear Physics A* **904** 186c–193c
[13] Majumder A and Van Leeuwen M 2011 *Progress in Particle and Nuclear Physics* **66** 41–92
[14] Casalderrey-Solana J 2014 *Gauge/string duality, hot QCD and heavy ion collisions* (Cambridge University Press)
[15] Horowitz W 2015 *Physical Review D* **91** 085019
[16] Hubeny V E 2015 *Classical and Quantum Gravity* **32** 124010
[17] Bass S A, Gale C, Majumder A, Nonaka C, Qin G Y, Renk T and Ruppert J 2009 *Physical Review C* **79** 024901
[18] Altarelli G and Parisi G 1977 *Nuclear Physics B* **126** 298–318
[19] Miller M L, Reygers K, Sanders S J and Steinberg P 2007 *Annu. Rev. Nucl. Part. Sci.* **57** 205–243
[20] Loizides C, Kamin J and d’Enterria D 2018 *Physical Review C* **97** 054910
[21] Adams J, Aggarwal M, Ahammed Z, Amonett J, Anderson B, Anderson M, Arkhipkin D, Averichev G, Badyal S, Bai Y et al. 2006 *Physics Letters B* **637** 161–169
[22] Moore G D and Teaney D 2005 *Physical Review C* **71** 064904
[23] Gubser S S 2008 *Nuclear physics B* **790** 175–199
[24] Hambrock R and Horowitz W A 2018 *EPJ Web of Conferences* vol 171 (EDP Sciences) p 18002
[25] Hambrock R and Horowitz W 2017 *Nuclear and particle physics proceedings* **289** 233–236
[26] Cacciari M, Nason P and Oleari C 2006 *Journal of High Energy Physics* **2006** 006 |
4.4 Projectile Motion
1. A ball rolls off the edge of a cliff with a height of 1234 feet above the level ground below. As the ball leaves the top of the cliff it is moving horizontally with a speed of 18mph.
(a) Find parametric equations for the motion of the ball.
(b) What is the position of the ball 7 seconds after leaving the edge of the cliff?
(c) When and where does the ball hit the ground?
(d) A person standing 340 feet in front of the cliff observes the event. Does the ball land in front of or behind the observer?
(e) Find a quadratic function $y = f(x)$ a portion of whose graph is the trajectory.
2. John is standing near the edge of a deep canyon with a river in the bottom. He throws a rock into the canyon so that its initial velocity $v$ is entirely in the horizontal direction. John counts that it takes 5 seconds for the rock to splash into the water. Estimate the depth of the canyon.
3. While standing atop a 1560 feet high office building you notice a round pebble on the edge of the roof. Suppose you gently push the pebble over the edge. Assume that the pebble has a constant horizontal velocity $c$ feet/sec. Furthermore, imagine the pictured situation below:
(a) If $c=0.1$ ft/sec, where does the pebble land on the ground?
(b) Find the largest value for $c$ so the pebble lands on the building side of the taxi.
(c) Find the required value for $c$ so that the pebble lands in the middle of the trash can.
4. A downhill ski racer has a constant velocity of 84 miles per hour on a downward sloping ramp. The ramp slopes down $5^\circ$ below horizontal and the ramp drops 3 feet to another sloped portion of the course. This lower portion of the course has a constant slope of $8^\circ$ below horizontal:
(a) Impose a coordinate system and find parametric equations for the airborne skier.
(b) Where and when is the skier 1 ft vertically above the lower portion of the course?
(c) Where and when does the skier land on the lower part of the course?
(d) Find the maximum height of the skier above the course, when this happens and the location of the skier at this instant.
5. Two pine cones release at the same instant from an old growth Douglass Fir. At the instant the cones release, a side wind to the right (away from the tree) causes them to move with a horizontal velocity of 25 mph. Assume that the ground slopes up from the base of the tree at a constant rate of 1 vertical foot for each 10 horizontal feet; see picture below. Impose a coordinate system as pictured and answer these questions:
(a) Give parametric equations for the trajectory of each of the pine cones A and B.
(b) Describe the location of each cone after $\frac{3}{2}$ seconds have elapsed.
(c) When does each cone hit the ground?
(d) When is each cone 10 feet above the ground?
(e) Describe the location of cone B when cone A hits the ground.
6.* Michael is located 100 feet to the right of Aaron as pictured below. They are both about to kick a ball.
(a) Introduce a common coordinate system with Aaron as the origin.
(b) Aaron kicks a ball with an initial speed of 40 ft/sec in a direction $40^\circ$ above horizontal. What is the parametrized trajectory of this ball?
(c) Michael kicks a ball toward Aaron with an initial speed of 60 ft/sec in a direction $70^\circ$ above horizontal. What is the parametrized trajectory of this ball?
(d) Find quadratic functions $y = f(x)$ and $y = g(x)$ which describe the trajectories of each ball.
(e) Using this information, find where the two trajectories intersect (as in the picture).
(f) Describe when each ball should be kicked so that they collide in midair.
7. A beach ball of diameter 24 inches is kicked from the edge of a 23 ft high bluff above a level beach. The velocity vector for the ball is 30 feet/sec in a direction $51^\circ$ above horizontal.
(a) Sketch a picture of the situation.
(b) Describe the trajectory of the center of the ball. You need to be careful here, since the radius of the ball must be taken into account. Namely, the center of the ball is actually 24 feet above the level beach when kicked.
(c) When and where does the ball hit the ground? (Keep in mind that the ball hits the beach precisely when the center is 1 ft above the ground.)
8. Redo Example 4.4.12 with the following change: The cyclist is traveling 20 mph and the maximum height is 12 feet above the landing zone.
9. A stunt cyclist needs to make a calculation for an upcoming cycle jump. The cyclist is traveling at a speed of 100 ft/sec toward an inclined ramp which is followed by a level landing zone (no drop off as in the previous example). Assume the rider maintains speed up the ramp, which has angle $\theta$ above horizontal. The cyclist needs to land 150 feet in front of the launch point. Find the desired angle of launch. (Hint: There will be two answers; use the ideas in Example 4.4.12. The double angle formula is also useful: $\sin(2\theta) = 2 \sin(\theta) \cos(\theta)$.)
10. A ground launched cannon fires a projectile with a constant speed of 1,000 feet/sec.
(a) Determine the required angle(s) of launch if the target has the same elevation as the cannon and is located 5.8 miles away.
(b) Describe the trajectory of the projectile for each launch angle in a., also determining the maximum height and elapsed time to impact.
(c)* A target is 1 mile horizontally in front of the cannon AND 1000 feet above the elevation of the canon. Determine the required angle(s) of launch to strike the target.
11. Describe the portion of the parabola traced out by the trajectory in Example 4.4.11.
12. On the moon, the height of an object $t$ seconds after being dropped from an initial height $y_o$ feet above a reference level is given by $y = -2.62t^2 + y_o$; compare to the formula for the earth in (4.4.2). Rework Example (4.4.11), assuming you are on the moon.
13.* Tiffany and Brendan spent last summer roofing the barn on Tiffany’s family farm. Brendan is a Jolt Cola drinker, so Tiffany was frequently throwing him refills. In turn, Brendan tossed back the empties. Suppose Tiffany tosses a can at an angle of 60° above horizontal with a speed of 50 ft/sec. At the same instant, Brendan tosses an empty can in the direction indicated at an angle 3° BELOW horizontal with a speed of 60 ft/sec.
(a) Find parametric equations $P(t) = (x_1(t), y_1(t))$ for the motion of Tiffany’s tossed can.
(b) Where is Tiffany’s tossed can located when it reaches its highest point above the ground?
(c) Find a function $y = f(x)$ whose graph is the path of Tiffany’s tossed can.
(d) Where and when does Tiffany’s tossed can hit the roof?
(e) Find parametric equations $Q(t) = (x_2(t), y_2(t))$ for the motion of Brendan’s tossed can.
(f) Where is Brendan’s tossed can located the instant it is no longer above the barn; i.e. when it has $x$ coordinate 30?
(g) Find a function $y = g(x)$ whose graph is the path of Brendan’s tossed can.
(h) If Tiffany doesn’t move, should she be able to catch the can? How much time will she have to react?
(i) Where are the two cans located the instant they have the same height above the ground?
(j) Where do the two trajectories cross one another?
(k) Use a graphing device to find the minimum distance between the two airborne cans.
14.* Return to the “magical marble toss” in Exercise 2.3.18. Find the angle and the speed of launch of the marble.
15. John has been hired to design an exciting carnival ride. Tiff, the carnival owner, has decided to create the world’s greatest ferris wheel. Tiff isn’t into math; she simply has a vision and has told John these constraints on her dream: (i) the wheel should rotate counterclockwise with an angular speed of 12 RPM; (ii) the linear speed of a rider should be 200 mph; (iii) the lowest point on the ride should be 4 feet above the level ground. Recall, we worked on this problem in Exercises 3.2.14, 3.3.16 and 4.3.4. In Exercise 4.3.4, we resolved the velocity vector for Tiff’s motion the instant she becomes a human missile.
(a) Find the parametric equations for Tiff’s motion once she becomes a human missile.
(b) Find where Tiff is located 6 seconds after she becomes a missile.
(c) Find when Tiff is 471 feet above the ground after she becomes a missile.
(d) Find where and when Tiff has maximum height above the ground.
(e) Find where and when Tiff craters (i.e. lands).
16. Nora is skiing out of control at 65 miles per hour up a ramp having an angle of $11^\circ$ above horizontal. The end of the ramp drops 2 feet to a level portion of the course.
(a) Impose a coordinate system.
(b) Resolve the velocity vector of Nora as she goes off the ramp.
(c) Find the coordinate function $x(t)$ for Nora, where $t = 0$ is the instant she is airborne.
(d) Find the coordinate function $y(t)$ for Nora, where $t = 0$ is the instant she is airborne.
(e) Use $y(t)$ to find where and when Nora lands on the level portion of the course.
(f) Use $y(t)$ to find Nora’s maximum height above the level part of the course and when this happens.
(g) Describe the trajectory of Nora; i.e. her path as she flies off the jump.
(h) Where and when is Nora 1 inch above the level portion of the course?
(i) Lee, who is 6 feet tall, is adjusting the straps on his snowboard while he stands 50 feet to the right of the ramp. Does Nora fly over Lee or land in front of him?
(j) Adri-N flies off the ramp after Nora with a speed of $s$ mph. Find the smallest speed $s$ so that Adri-N flies over Lee.
17. Adri-N is standing on the edge of a cliff and throws a marble straight up into the air at 75 ft/sec; the marble is released when it is 10 ft above ground. It goes straight up, then falls back down to the ground. Find the TOTAL distance the marble travels between the time it is released and the time it hits the ground.
18. Three basketball players are in the locations below; impose a coordinate system as pictured. Keith throws the ball toward Spires. Let $P(t) = (x(t), y(t))$ be the location of the ball $t$ seconds after Keith passes it; distance units will be feet. Assume $x(t) = 22t$ and $y(t) = -16t^2 + 20t + 5$. We will make life simple and view motion of the basketball as the motion of its center.
(a) Compute the initial velocity vector for Keith’s pass.
(b) Keith’s pass to Spires is a wild one; why?
(c) The ball soars over Spires. Determine the location of the ball when it hits the floor.
(d) Assume the ball bounces off the floor with the same angle $\theta = 50.64^\circ$ as it impacts the floor; see picture below. Also assume the impact speed of the ball is $34.68 \text{ ft/sec}$ and the ball bounces with $80\%$ of the impact speed.
What are the parametric equations for the motion of the ball bouncing off the floor?
(e) Should Emerson be able to handle the ball? (Justify your answer.)
(f) Find the multipart function whose graph models the path of Keith’s pass and sketch the graph.
19. Once again, Allyson is playing Lee in a game of one on one. Lee is leading 16-14. Just before time runs out, Allyson makes a three pointer and Lee limps away in defeat (again). Take the origin of the coordinate system to be on the ground beneath the ball. The equation of the path of the ball is given by
$$y = -0.0729x^2 + 3.16x + 6.$$
If Allyson shoots the ball at time $t = 0$, when does the ball go through the hoop?
20. Consider the problem of a baseball pitcher who throws a ball at 145 feet/second from a height of 6 feet. Suppose the ball is thrown at an angle (above or below the horizontal) of $\theta$ degrees. How do we calculate the angle $\theta$ so that the ball will cross home plate at a certain height?
(a) What is the initial horizontal velocity of the ball, expressed in terms of $\theta$?
(b) What is the initial vertical velocity of the ball, expressed in terms of $\theta$?
(c) Write parametric equations for the position of the ball at time $t$. Your answer will involve the constant $\theta$.
(d) Let’s call $t_p$ the time at which the ball crosses home plate. In baseball, home plate is 60.5 feet from the pitcher. Thus $t_p$ is the time at which the ball has traveled horizontally a distance of 60.5 feet. Find $t_p$, in terms of $\theta$.
(e) At the time $t_p$, what is the height $y_p$ of the ball above the ground? Again, your answer will involve the constant $\theta$ and trigonometric functions of $\theta$.
(f) Rewrite your answer for the value of $y_p$ in terms of $\tan(\theta)$, using the identity $\tan^2(\theta) + 1 = 1/\cos^2(\theta)$.
(g) Suppose the pitcher wants the ball to cross home plate at a height of 2 feet above the ground. For this to occur, what must $\theta$ be? You should get two answers; why?
21. Wile E. Coyote has an ACME ordered boulder which he is rolling off a 20 m high cliff as shown, trying to squash the Road Runner.
(a) Impose a coordinate system in the picture at right.
(b) Find parametric equations for the coordinates of the boulder with $t = 0$ corresponding to the instant it rolls off the cliff.
(c) Will the boulder hit the Road Runner?
(d) Find the equation $y = f(x)$ of the sloping ground beside the Road Runner.
(e) What are the coordinates of the point where the boulder finally hits the ground? |
DEPARTMENT OF TRANSPORTATION AND PUBLIC WORKS
PUERTO RICO HIGHWAY AND TRANSPORTATION AUTHORITY
PROFESSIONAL SERVICES CONTRACT
FOR PROJECT:
HURRICANE MARÍA EFFORTS FOR SIGNING AND GUARDRAILS FOR DORADO, MANATI, TOA ALTA, TOA BAJA, VEGA ALTA, AND BAYAMON MUNICIPALITIES
CONTRACT NUMBER:
2019-000038
AC-814541 / ER-9999(296)
PRHTA DESIGN AREA
# Table of Contents
CONTRACT FOR PROFESSIONAL SERVICES ................................................................. 5
TERMS AND CONDITIONS ....................................................................................... 6
1. SERVICES BY THE CONSULTANTS .................................................................. 6
2. ASSIGNMENT ...................................................................................................... 7
3. TERM ................................................................................................................... 7
4. COMPENSATION ................................................................................................. 7
5. ALLOWABLE COST ............................................................................................. 7
6. GOVERNING LAWS AND REGULATIONS ......................................................... 8
7. PROGRESS REPORT ............................................................................................ 8
8. SERVICES BY THE AUTHORITY ....................................................................... 8
9. METHOD OF PAYMENTS .................................................................................. 9
10. FINAL SETTLEMENT ....................................................................................... 9
11. CLAIMS AND LIABILITIES .............................................................................. 10
12. STAMPS ............................................................................................................. 10
13. SCHEDULE DELAYS AND TIME EXTENSIONS ............................................. 10
14. SUBCONTRACTING ............................................................................................ 11
15. ERROR AND OMISSIONS ............................................................................... 12
16. LICENSURE ....................................................................................................... 12
17. INDEPENDENT CONTRACTOR .......................................................................... 13
18. SOCIAL SECURITY DEDUCTIONS AND TAX REVENUES ............................. 13
19. CONSULTANT CONDUCT ............................................................................... 13
20. TAXES ............................................................................................................... 14
21. COPYRIGHTS ..................................................................................................... 14
22. CONFIDENTIALITY .......................................................................................... 16
23. TERMINATION OF CONTRACT ......................................................................... 16
| Section | Page |
|------------------------------------------------------------------------|------|
| 24. FISCAL RESPONSIBILITY AND REQUIRED DOCUMENTS | 17 |
| 25. FAMILY ASSISTANCE AND ELDERLY SUPPORT | 19 |
| 26. INSURANCE | 19 |
| 27. CRIME CONVICTION CLAUSE | 20 |
| 28. CRIMINAL RECORDS | 21 |
| 29. GOVERNMENT ETHICS | 21 |
| 30. CONFLICT OF INTEREST | 23 |
| 31. CONTRACT ENFORCEMENT, JURISDICTION AND APPLICABLE LAW | 25 |
| 32. ACCESS TO FILES, REPORTS AND RECORDS | 25 |
| 33. NOTIFICATION | 25 |
| 34. CONTRACT TRANSFER RIGHTS | 26 |
| 35. EQUAL EMPLOYMENT AND NON DISCRIMINATION | 26 |
| 36. COMPLIANCE WITH PUBLIC POLICY | 26 |
| 37. REGISTRATION AT THE OFFICE OF THE COMPTROLLER | 26 |
| 38. CONTRACT CHANGES | 26 |
| 39. SAVING CLAUSE | 27 |
| 40. PREFERENCE | 27 |
| 41. RETENTION OF PAYMENT DUE | 27 |
| 42. DISADVANTAGE BUSINESS ENTERPRISES ASSURANCES | 27 |
| 43. LOBBYING CERTIFICATION (49 CFR PART 20) | 29 |
| 44. CIVIL RIGHTS-TITLE VI ASSURANCE | 29 |
| 45. TITLE VI ASSURANCE AND NON-DISCRIMINATION PROVISIONS | 31 |
| 46. INTERAGENCY SERVICES CLAUSE | 32 |
| APPENDIX A | |
| SCOPE OF PROFESSIONAL DESIGN SERVICES | 33 |
APPENDIX B
SCOPE OF PROFESSIONAL ENVIRONMENTAL SERVICES ...........................................42
APPENDIX C
COMPENSATION AND PAYMENTS FOR PROFESSIONAL SERVICES ..........................80
APPENDIX D
GENERAL DESIGN CONSIDERATIONS AND CONDITIONS .......................................88
COMMONWEALTH OF PUERTO RICO
DEPARTMENT OF TRANSPORTATION AND PUBLIC WORKS
HIGHWAY AND TRANSPORTATION AUTHORITY
CONTRACT FOR PROFESSIONAL SERVICES
CONTRACT NUMBER 201900048
AUTHORITY PROJECT Number AC-814541
ORACLE PROJECT Number LPE1499541
Federal PROJECT Number ER9999-296
In San Juan, Puerto Rico, the 3 day of August, 2018
AS FIRST PARTY: The PUERTO RICO HIGHWAYS AND TRANSPORTATION AUTHORITY, a public corporation and instrumentality of the Commonwealth of Puerto Rico, created by virtue of Act No. 74 of June 23, 1965, as amended (the "Organic Act"), represented herein by its Executive Director, Carlos M. Contreras Aponte, of legal age, married, and a resident of San Juan Puerto Rico, authorized by the Board of Directors as evidenced by Resolution number 2018-09, hereinafter referred to as the "AUTHORITY".
AS SECOND PARTY: LOGISTIC ENGINEERING CONSULTANT, CSP, represented herein by Wilfredo M. Rodríguez Uviñas, engineer licensed in the Commonwealth of Puerto Rico, with license number 13719, of legal age, married and resident of Guaynabo, Puerto Rico; hereinafter referred to as "the CONSULTANT".
WITNESSETH
WHEREAS: The AUTHORITY desires to engage the CONSULTANT to provide engineering professional services as hereinafter set forth for the preparation of necessary reports, studies, environmental document, designs, drawings, specifications, engineers estimates, and other documents.
WHEREAS: The AUTHORITY needs to perform conceptual studies, environmental studies, formal environmental document, preliminary and final plans, specifications, and engineer's estimates, following the applicable federal and state regulations, and according the Request for Proposal; hereinafter referred to as "RFP", requirements.
WHEREAS: The AUTHORITY does not have enough technical and engineering staff to execute the works and services proposed under this contract.
WHEREAS: The CONSULTANT is qualified and capable of providing these engineering professional services in accordance with all federal and state laws and regulations.
WHEREAS: On December 22, 2017 the AUTHORITY published a RFP, and requested proposals or requested a proposal for the: Preparation of Plans, Specifications, and Engineer's estimates (PS&E) for the HURRICANE MARÍA EFFORTS FOR SIGNING AND GUARDRAILS, and is established that the published RFP is an integral part of this contract.
WHEREAS: After duly evaluating the proposals submitted by the CONSULTANT and according to the procedures established in the Request for Proposal (RFP), the AUTHORITY ranked and negotiated this contract with the CONSULTANT.
WHEREAS: The AUTHORITY is empowered to grant this contract because of the technical services and expertise required. The AUTHORITY finds that executing this contract is in the best public interest.
WHEREAS: The AUTHORITY’s Budgeting Office has issued the required certification of available funds to acquire the services that will be provided by the CONSULTANT.
THEREFORE: The AUTHORITY and the CONSULTANT have come to an agreement whereby the AUTHORITY has agreed to contract the CONSULTANT to perform the aforesaid engineering and design CONSULTANT services, and the CONSULTANT has agreed to perform these services subject to the following:
TERMS AND CONDITIONS
1. SERVICES BY THE CONSULTANTS
The CONSULTANT represents that it has a reasonable equipped office physically located in Puerto Rico and professionally capable to perform the services required under this contract and that all the professional and technical personnel actually performing the services contracted are legally authorized to reside and work in the United States of America and are in good standing to practice their professions in the Commonwealth of Puerto Rico. Furthermore, the CONSULTANT represents that it has the expertise, the personnel, and other resources necessary to provide the AUTHORITY with the following services:
a. Preparation of Plans, Specifications, and Engineer's estimates (PS&E) for the HURRICANE MARÍA EFFORTS FOR SIGNING AND GUARDRAILS FOR DORADO, MANATÍ, TOA ALTA, TOA BAJA, VEGA ALTA, AND BAYAMON MUNICIPALITIES, AC-814541, Federal Number ER-9999(296), LPE1499541.
b. The AUTHORITY and CONSULTANT agree that the general provisions for design services in APPENDIX A: SCOPE OF PROFESSIONAL DESIGN SERVICES, apply to the consulting services under this contract.
c. The AUTHORITY and CONSULTANT agree that the general provisions for environmental services in APPENDIX B: SCOPE OF PROFESSIONAL ENVIRONMENTAL SERVICES, apply to the consulting services under this contract.
2. ASSIGNMENT
The CONSULTANT shall not assign any of its rights and obligations under this contract without the prior written consent from the AUTHORITY, consent which may be withheld in the AUTHORITY's absolute discretion. Notwithstanding, CONSULTANT shall be authorized to engage such professional or other personnel, if any, as it deems necessary or advisable to perform its obligations under this contract, provided CONSULTANT shall remain liable to the AUTHORITY for such performance, without regard to nature of the engagement between CONSULTANT and any member of its personnel, and such engagement shall not be deemed an assignment under this provision or a subcontract under the provisions of Section 13 of this contract.
3. TERM
This contract will be valid from the date of execution until 365 calendar days. Unless otherwise directed by the AUTHORITY, the work to be performed under this Contract shall commence upon the CONSULTANT’s receipt of a fully executed copy of this Contract and shall be completed as directed in the Notice to Proceed to be issued by the AUTHORITY. Thereafter, this date may be extended by written amendment of the present Contract between the AUTHORITY and the CONSULTANT, however any time extension by and in itself shall not change the maximum amount payable under this Agreement. The CONSULTANT shall be responsible for requesting and justify in writing to the AUTHORITY any time extensions at least (60) days before the expiration date of contract.
4. COMPENSATION
The AUTHORITY will pay the CONSULTANT for the provision of consulting services to a maximum of TWO HUNDRED TWENTY-ONE THOUSAND ONE HUNDRED FORTY-ONE DOLLARS AND NINETY-NINE CENTS ($221,141.99) during the term of this contract. The CONSULTANT will be paid by deliverables as described on APPENDIX C: COMPENSATION AND PAYMENTS FOR PROFESSIONAL SERVICES.
5. ALLOWABLE COSTS
This contract shall be in compliance with the determination of allowable costs in accordance with the Federal cost principles as per 23 CFR 172.9c(ix). Such contract could be, to PRHTA’s sole discretion, subject to an audit which once performed is binding and its outcome will be one of the following:
a. If the percentage of indirect costs is the same as the proposed under the negotiated fees, the CONSULTANT will be paid such compensation;
b. If the percentage of indirect costs is less than the one proposed in the negotiated fees, the AUTHORITY will pay compensation based on the lower percentage;
c. If the percentage of indirect costs is greater than the one proposed in the negotiated fees, the AUTHORITY will pay compensation for the percentage previously negotiated;
d. If the percentage of indirect costs cannot be determined, the AUTHORITY will not pay any overhead to the CONSULTANT.
6. GOVERNING LAWS & REGULATIONS
The CONSULTANT agrees to perform the services under this contract complying with the Puerto Rico laws, federal laws and regulations that relate to them.
7. PROGRESS REPORT
During the term of this contract, the CONSULTANT will report, at the request of the AUTHORITY, or his representatives, reports on the progress of activities and comments or recommendations. Any information collected by the CONSULTANT in the performance of their duties, and any report prepared for the purpose, cannot be published or made available to other persons or organizations without prior consent of the Executive Director and/or Secretary, who reserve the right of partial or total reproduction thereof.
8. SERVICES BY THE AUTHORITY
The Authority shall supply the CONSULTANT at no cost to the CONSULTANT, any requested information, in the possession of the agency that are necessary for the preparation of the services covered by this contract, such as;
a. Access to GOVERNMENT and private property whenever such access is required.
b. Access to and use of all existing and available reports, computations, survey data, and other information in the possession of the GOVERNMENT pertaining to this project.
c. Historical Cost Data, if available, to assist in estimating probable costs for the various phases of construction.
d. Existing Property Surveys, if available, of the design area.
e. Written Notice of Acceptance of certain steps of various phases of services as herein specified.
f. Coordination with Federal Highway Administration.
9. METHOD OF PAYMENTS
The CONSULTANT shall submit to the AUTHORITY, three (3) copies of invoices accompanied by copies of the progress report approved by the AUTHORITY, as per APPENDIX C, SECTIONS A through B of this contract. Invoices shall be accompanied by documents evidencing the services billed and agreed between the parties. The CONSULTANT shall include any invoices submitted by any other sub-CONSULTANT contracted by the CONSULTANT. Along with the invoice, the CONSULTANT shall submit a statement containing the balance of the contract budget.
a. The CONSULTANT shall certify each invoice as correct and ensure that payment has not been made and services have been offered. In accordance with Article 3.2 (e) of Act No. 2 of January 4, 2018, known as "Código Anticorrupción", every invoice submitted by the CONSULTANT will be accompanied with the report and shall, as a condition of payment, contain the following certification:
"Under penalty of absolute nullity, certify that no employee of the Highway and Transportation Authority is a party or has an interest in the profits or benefits of the contract under this bill and to be part of or have an interest in the product or benefit the contract, has mediated a waiver prior. The only consideration to supply the goods or services under this contract was the payment agreed with the authorized representative of the Authority. The amount of this bill is fair and right. Services have been provided and have not been paid".
b. Payment will be made upon presentation of invoice evidenced by the deliverable and duly certified by the AUTHORITY or its authorized representative. If the AUTHORITY finds the submitted invoice as acceptable, then the invoice will be approved and processed for payment promptly. The AUTHORITY reserves the right to review the correctness of invoices and perform the audits as it deems fit. The fact that a CONSULTANT invoice has been submitted to the AUTHORITY does not constitute an immediate certification and payment.
CONSULTANT payments will be subject to a withholding amount as per section regarding Taxes, described herein.
10. FINAL SETTLEMENT
Final settlement of all amount due to the CONSULTANT shall be made when the CONSULTANT has completed all work required by the contract and its amendments and has made the final delivery of all documents as required by the contract. The AUTHORITY shall approve and audit all work before carrying out the final settlement for which, it can request the CONSULTANT all the information that it considers necessary, related to the billing invoices, before the final settlement.
The AUTHORITY has the right to cancel the contract if it is in the best public interest. In the event of the contract cancellation, the CONSULTANT fees shall be settled as follows;
a. If the contract is cancelled before completing the preliminary design phase, then the AUTHORITY shall establish the corresponding percentage of completion and fees. The fee will be settled using the lump sum as indicated in APPENDIX C, SECTION B - Table 5.2 DETAILED PRELIMINARY DESIGN FEES.
b. If the contract is cancelled after completing the preliminary design phase, but before completing the final design phase, the AUTHORITY shall establish the corresponding percentage of completion and fees. The fee will be settled using the lump sum as indicated in APPENDIX C, SECTION B - Table 5.3 DETAILED FINAL DESIGN FEES.
11. CLAIMS AND LIABILITIES
The acceptance by the CONSULTANT of the final payment shall constitute a release to the AUTHORITY, for all claims and liabilities under the contract alleged by the CONSULTANT, his representatives, assigns or third parties, for any and all things done, furnished or related to the services rendered by the CONSULTANT under or in connection of the contract or any part thereof. However, the acceptance by the CONSULTANT of the final payment, does not constitute or operate as a release for all claims and liabilities arising under the contract alleged by the AUTHORITY, its representatives, assign or third parties, for any and all things done, furnished or related to the services rendered by the CONSULTANT under or in connection of the contract or any part thereof.
The CONSULTANT shall be responsible for, and shall release, indemnify and hold harmless the AUTHORITY for expenses (including attorney's fees and costs) actually incurred in connection with any damages, whether direct or indirect, incidental, special, punitive or consequential and any loss of, revenue, data or data use, arising from or related to CONSULTANT’s gross negligence or willful misconduct in the provision of services under this contract; provided, however, that the total, aggregate liability of CONSULTANT for all claims arising in connection with this contract, whether in contract, tort, equity or otherwise, shall not exceed the total amounts actually paid to CONSULTANT by the AUTHORITY pursuant to the contract.
12. STAMPS
It shall be the CONSULTANT’s responsibility to pay the corresponding stamps of the Colegio de Ingenieros y Agrimensores de Puerto Rico (CIAPR). The amount on stamps shall be paid at a rate of $1.00 by each $1,000.00 or fraction of it according to the PROJECT’s final cost estimate as approved by the AUTHORITY.
13. SCHEDULE DELAYS AND TIME EXTENSIONS
Reasonable time extensions shall be given by the AUTHORITY in the event of delays not caused by the fault or negligence of the CONSULTANT, his employees or agents, or the sub-CONSULTANT employees or agents.
Whenever the CONSULTANT encounters any difficulty which is delaying or threatens to delay the timely performance of the contract, the CONSULTANT shall immediately give notice thereof in writing to the AUTHORITY, stating all relevant information with respect thereto for consideration of the AUTHORITY, which will evaluate such request and its justification and determine its provenance, who will notify the CONSULTANT in writing.
The CONSULTANT agrees to insert the substance of this clause, including this paragraph, in any subcontract hereunder; except that each such subcontract shall provide that in the event its timely performance is delayed or threatened by delay, the sub consultant shall immediately notify in writing to the CONSULTANT of all relevant information with respect thereto. The CONSULTANT shall evaluate the request and its justification and determine its provenance, keeping in mind the best interests of the AUTHORITY. The CONSULTANT shall keep inform the AUTHORITY of any communication done to the sub-CONSULTANT on this regard.
The CONSULTANT shall be responsible for requesting and justify in writing to the AUTHORITY any time extensions at least sixty (60) calendar days before the expiration of the time specified in the contract for the Conceptual Design and Schematics Plans, Preliminary Plans, Final Plans, Land Acquisition Plans, Field Studies and Surveys, Geotechnical Studies, Environmental Assessment, Tree Inventory and Mitigation Plan, Hydrologic-Hydraulic and Scour Studies, Traffic and Capacity Studies, and any other which applies.
Failure to complete the design on the stipulated time for each of the different work phases enumerated on the prior paragraph and any other on stipulated on the contract and its amendments, including reviewing time, not caused by fault or negligence from the AUTHORITY, shall constitute an unfulfillment on part of the CONSULTANT, who shall be subject to a fine of point two per cent (0.2%) of the total fee, of the particular item, per calendar day of delay.
14. SUBCONTRACTING
The CONSULTANT may subcontract the necessary studies to carry out the work under this Contract subject to the provisions contained herein. The CONSULTANT shall ensure that the sub-consultant hired has the capacity and expertise needed to carry out the work efficiently, in technical compliance and professional quality.
The CONSULTANT shall ensure that each contract signed with any sub-consultant, protects and guarantees the rights of the Authority under the terms of this contract regarding the work that will carry out the sub-consultant; so that the work that will carry out the sub-consultants do not harm the AUTHORITY.
The CONSULTANT shall ensure, to the extent possible, that the sub-consultant assumes before the CONSULTANT the same duties and obligations that the CONSULTANT assumed before the AUTHORITY in compliance with the terms of this Contract and the work that will be carried out. For this purpose, a copy of this Contract may be provided to the sub-consultant, so that the clauses of the subcontract subscribed between the CONSULTANT and the sub-consultant are not contrary to the clauses of this CONTRACT. If the clauses of the subcontract are inconsistent to the clauses of this Contract, the clauses and provisions of this Contract shall prevail.
When the CONSULTANT chooses to subcontract any of the required studies under this Contract, he will submit to the AUTHORITY the name of the person, persons or entities that will be subcontracted.
The AUTHORITY reserves the right to reject the subcontracting of any person, persons and/or entities. The AUTHORITY will have ten (10) business days, after receiving from the CONSULTANT the information of said person, persons and/or entities to be subcontracted, to notify the CONSULTANT of its objection. If the AUTHORITY fails to respond in the allotted time it will be understood that the AUTHORITY has no objection to the propose subcontracting.
The CONSULTANT shall indemnify the AUTHORITY, to the extent provided in section regarding Claims and Liabilities of this contract, for any and all damages as a result of any faulty and/or negligent work from the sub-consultant, his employees, and agents and for any breach of contract from the sub-consultant, his employees and agents and for any delay the PROJECT may suffer as a result from actions and omissions form the sub-consultant, his employees and agents. The power of the AUTHORITY to reject the proposed subcontracting does not exempt or release in any way the CONSULTANT from his duties and obligations towards the AUTHORITY as set forth in the provisions of this CONTRACT, nor will it be understood that it constitutes as a waiver of the AUTHORITY of its rights under this Contract or under applicable laws.
The CONSULTANT shall provide to the AUTHORITY within (30) days after subscribing said subcontract all certifications required by this CONTRACT in respect to the tax situation of any person who's subcontracting the AUTHORITY authorized. For purposes of this Contract any professionals, consultants or technicians, contracted by the CONSULTANT, shall be deemed as a subcontractor if they devote twenty five percent (25%) or more of their time to provide services to meet the obligations of this Contract.
15. ERROR AND OMISSIONS
Errors and omissions in the work of CONSULTANT will be corrected by the CONSULTANT at no additional cost to the AUTHORITY.
16. LICENSURE
Certification of personnel needed and Capacity to render a service:
The CONSULTANT will render the services on a permanent and continuous manner and will provide all necessary personnel to carry out the work as required by this contract.
CONSULTANT certifies and warrants that it has the academic preparation, experience, skills and licenses or certificates necessary to provide contracted services, according to the requirements of the Commonwealth of Puerto Rico, specifically Law 173 of August 12, 1988. All personnel assigned by the CONSULTANT will be of proven and recognized expertise and experience in their respective fields. The CONSULTANT knows the ethical standards of their field of work and takes responsibility for his actions. Expressly recognizes that these are essential conditions of this Agreement, and if not correct in whole or in part any of the above certifications, this will be sufficient cause for the AUTHORITY to rescind it, immediately, and the CONSULTANT will have to reimburse any money that has accrued under this Agreement.
17. INDEPENDENT CONTRACTOR
a. The CONSULTANT shall have no right or benefits accruing to employees of the AUTHORITY. The CONSULTANT is to all intents and purposes an independent contractor and under no circumstances be considered an employee or agent of the AUTHORITY. The CONSULTANT will be solely responsible for all contracts, debts, obligations incurred by it and its employees to its name in connection with the performance of this Contract.
b. THE CONSULTANT certifies that no payment or compensation has been received for services rendered under regular appointment to another public entity, except as permitted by law.
c. Both Parties freely and voluntarily agree that under the terms of this contract does not establish relations of employer and employee and CONSULTANT and all employees act and shall serve at all times as an independent contractor and therefore they are not entitled to claim to the AUTHORITY, among others, for vacation benefits, health plan, sick leave, retirement, Christmas bonus, professional liability insurance, or Social Security.
d. The existence of this contract will not have the effect of generating CONSULTANT for class rights which, by virtue of law rules, should be applicable to officers and/or employees of the AUTHORITY.
18. SOCIAL SECURITY DEDUCTIONS AND TAX REVENUES
No discounts will be made by the concept of Social Security and Income Tax, except to apply the withholding orders by Section 1143 of the Internal Revenue Code of Puerto Rico in 1994, to the CONSULTANT because his act as own employer. The payments made under this contract will be notice to the Treasury Department. The CONSULTANT is responsible for making payments to Social Security and Federal Income Tax Bureau.
19. CONSULTANT CONDUCT
In compliance with Section 6.8 of Act 8 of February 4, 2017, "Ley para la Administración y Transformación de los Recursos Humanos en el Gobierno de Puerto Rico", the CONSULTANT certifies that it has not engaged in dishonorable conduct, not addicted to habitual and excessive use of controlled substances and/or alcoholic beverages has not been convicted of a felony or misdemeanor, or any offense involving moral turpitude, or been dismissed from the public service. If having committed any of these causes disabling, the CONSULTANT shall submit to the AUTHORITY the Order issued by the Director of the Office of Administration and Transformation of the Human Resources of the Government of Puerto Rico ("OATRH by its Spanish acronyms) which certifies your eligibility, copy of which was made part of the contract file. CONSULTANT expressly acknowledges that the accuracy of this certification is a special condition of the contract and that otherwise, shall be grounds for the AUTHORITY to rescind the contract and recover from the CONSULTANT any money paid on account of services rendered under the contract. If during the term of the contract CONSULTANT incurs disabling any of the causes that are provided in this clause, the Authority shall be entitled to terminate the contract immediately.
20. TAXES
The payments to be made to the CONSULTANT, will be subject to the withholding of seven percent (7%), as provided in Section 1143C of the Income Tax Act of Puerto Rico of 2011, as amended, except when the CONSULTANT submits to the AUTHORITY a certificate of relief from the payment of such tax. The CONSULTANT is responsible to pay their taxes and pay the corresponding contributions to Social Security and Income Tax Bureau of the Department of Finance, for any taxable amount as a result of income earned under this contract. The AUTHORITY shall notify the Bureau of Income Tax payments and refunds to be made by the CONSULTANT.
21. COPYRIGHTS
A. Any final CONSULTANT work product of this Agreement (the "Deliverables") will be upon full and final payment of all fees and expenses owing to CONSULTANT under this Agreement considered the property of the AUTHORITY. Notwithstanding anything to the contrary, CONSULTANT shall retain all right, title and interest in and to: (1) all of its pre-existing patent, copyright, trademark and other intellectual property rights therein; and (2) all methodologies, processes, techniques, ideas, concepts, trade secrets, and know-how embodied in the Deliverables or that CONSULTANT may develop or supply in connection with this Agreement ("Consultant Knowledge"). Subject to the confidentiality restrictions contained in this Agreement, CONSULTANT may use CONSULTANT Knowledge for any purpose. Upon full and final payment of all fees and expenses owing to the CONSULTANT under this Agreement, the CONSULTANT hereby grants to the AUTHORITY a non-exclusive, nontransferable, royalty-free license to use the Deliverables and to create and use derivative works derived from the Deliverables, so long as in doing so the AUTHORITY does not breach any obligation of this contract. The AUTHORITY acknowledges that all advice (written or oral) provided by CONSULTANT to the AUTHORITY in connection with this Agreement is intended solely for the benefit and use
of the AUTHORITY in considering the matters to which this engagement relates. The AUTHORITY agrees that no such advice shall be used for any other purpose or reproduced, disseminated, quoted or referred to at any time in any manner or for any purpose other than accomplishing the tasks referred to herein without CONSULTANT’s prior approval (which shall not be unreasonably withheld), except as required by law.
B. All documents and information provided by the AUTHORITY for the duration of the Contract, as well as all correspondence, reports, research, information or material relevant to the contract provided by or on behalf of the Authority and obtained as a result of services rendered by the CONSULTANT, shall be owned by the AUTHORITY and cannot be used by the CONSULTANT for any purpose other than the performance of its services under this contract, without the prior written consent of the AUTHORITY, even after completion of the contract.
C. By written request of the AUTHORITY or at the expiration, cancellation or termination of this Agreement, the CONSULTANT shall provide the AUTHORITY any document, information or material owned by the AUTHORITY in relation to this Agreement. The unfinished work of the CONSULTANT shall be submitted to the AUTHORITY, which shall be entitled to finish the work and use the material and ideas thereof so long as such use is consistent with the provisions of this contract. Notwithstanding anything to the contrary, each party may retain, subject to its confidentiality obligations, one (1) copy of the other party’s documents or other information (including confidential information) for internal record keeping purposes, but only for that period of time necessary to evidence compliance with this Agreement or other legal requirements.
D. The AUTHORITY shall accept Deliverables which conform to the requirements of this AGREEMENT. The AUTHORITY will promptly, and in no event more than twenty (20) calendar days after the completion of the applicable services with respect to any Deliverable (“Acceptance Period”) give CONSULTANT notification of any non-conformance of the Deliverables with such requirements (“Non-conformance”), and CONSULTANT shall have a reasonable period of time, based on the severity and complexity of the Non-conformance, to correct the Non-conformance. The AUTHORITY shall have the right to test the corrected Deliverables in the same manner as with respect to the original Deliverables. If the AUTHORITY uses the Deliverables before acceptance, modifies the Deliverables without the approval of CONSULTANT, unreasonably delays the beginning of acceptance testing, or does not notify CONSULTANT of any Non-conformance within the Acceptance period, then the Deliverables shall be considered accepted by the AUTHORITY.
E. CONSULTANT warrants that the Services shall be performed with reasonable care in a diligent and competent manner. CONSULTANT’s sole obligation shall be to correct any non-conformance with this warranty, provided that AUTHORITY gives CONSULTANT written notice within thirty (30) days after the Services are performed or the Deliverables are accepted. CONSULTANT does not warrant and is not responsible for any Services
to the extent that the Services incorporate work performed substantially by AUTHORITY or to the extent that the Services involve work that has not been completed and reviewed by CONSULTANT. CONSULTANT does not warrant and is not responsible for any third-party products or services not contracted by the CONSULTANT. AUTHORITY’s sole and exclusive rights and remedies, with respect to any third-party products or services that were not contracted by the CONSULTANT, are against the third-party vendor and not against CONSULTANT. THIS SECTION IS CONSULTANT’s ONLY WARRANTY CONCERNING THE SERVICES AND ANY DELIVERABLE AND IS MADE EXPRESSLY IN LIEU OF ALL OTHER WARRANTIES AND REPRESENTATIONS, EXPRESS OR IMPLIED, INCLUDING ANY IMPLIED WARRANTIES OF MERCHANTABILITY, NONINFRINGEMENT, OR FITNESS FOR A PARTICULAR PURPOSE, OR OTHERWISE.
22. CONFIDENTIALITY
The CONSULTANT acknowledges that all information or data provided, obtained and produced as part of the services under this contract shall be considered confidential and, as such, disclosure without the written consent of the AUTHORITY is strictly prohibited, except: (i) as requested by the AUTHORITY or its legal counsel; (ii) as required by legal proceedings or (iii) as reasonably required in the performance of this engagement. All obligations as to non-disclosure shall cease as to any part of such information to the extent that such information is or becomes public other than as a result of a breach of this provision.
The CONSULTANT shall not use the information that by its nature, knowledge or expertise is considered confidential and acquired during or as a result of the professional services entrusted to him by contract from the AUTHORITY or the Commonwealth of Puerto Rico, for purposes other than the obligations under contract or to obtain, directly or indirectly or in any way, advantage or benefit for himself, a member of his household or any other natural or legal person, business or entity. The CONSULTANT agrees to manage that information as confidential. The AUTHORITY will notify CONSULTANT in writing any other information that will be considered as confidential.
23. TERMINATION OF CONTRACT
Either party, the AUTHORITY or the CONSULTANT may terminate this contract without cause, upon written notice by personal delivery or by certified mail, return receipt requested, to the other party, with thirty (30) days effectiveness prior to termination. The rights, duties and responsibilities of the parties remain in full force during the period of thirty (30) days of the notification. The CONSULTANT shall not be entitled to additional compensation. Notwithstanding, the CONSULTANT agrees that the AUTHORITY may reduce the term of thirty (30) when its fiscal situation or the best interest of the AUTHORITY so require it.
The negligence of duty and improper conduct on the part of the CONSULTANT or any of its employees or officers shall constitute sufficient cause to terminate this contract immediately without prior notification alluded to in the previous paragraph. Negligent performance of duty or the abandonment of them by the CONSULTANT, will be considered a violation of this contract.
and shall be grounds for the AUTHORITY to declare it finished without any limitation and notwithstanding any law to the contrary, be relieved and discharged the AUTHORITY of any obligation or liability relating to him. In such an event, the CONSULTANT shall only be entitled to payment for the services rendered in accordance with the terms of this contract only if they benefit the AUTHORITY. In case of the termination of this Agreement, the CONSULTANT shall deliver to the AUTHORITY or the person designated all documents, equipment and services completed to the date of such action and copies of those that were in progress.
The CONSULTANT acknowledges that in the event that the funds provided for this contract were limited or canceled for reasons beyond the control of the AUTHORITY, this Agreement shall be terminated and will end on the day in which the CONSULTANT is notified in writing that said funds have been adjusted or canceled. In this case, all services provided by the CONSULTANT, including until the date of the written notice of cancellation or reduction of funds, will be compensated according to the terms of this Agreement. If funds were reduced, the parties will have the option to negotiate a new agreement in writing in conformity with the availability of funds within the budget of the AUTHORITY.
Upon completion of the work, either due to cancellation or termination of this Contract, the CONSULTANT shall submit to the AUTHORITY a final written report of the services rendered, steps taken, and those pending by the termination date. The AUTHORITY will only pay or reimburse the CONSULTANT for services rendered through the date of termination, resolution or expiration as indicated and for such final report.
As required by Memorandum No. 2017-001, the Chief Executive Officer of the Government of Puerto Rico shall have the power to terminate this contract at any time.
24. FISCAL RESPONSIBILITY AND REQUIRED DOCUMENTS
A. The CONSULTANT acknowledges that it is solely responsible to pay for wages, salaries or emoluments of the personnel used to provide the services contemplated in this Agreement. Also, the CONSULTANT will be solely responsible to pay for the State Insurance Fund Policy, unemployment insurance and any other insurance required of CONSULTANT by law and shall make any appropriate deductions of Federal Social Security and Income Taxes related to said personnel wages or salaries in connection with providing such services.
B. The AUTHORITY shall be responsible to deduct payments made to the CONSULTANT regarding income tax, the applicable percentage according to the provisions of the Internal Revenue Code of Puerto Rico, 2011, as amended, and Regulations, unless the CONSULTANT submits the AUTHORITY a Certificate of Waiver issued by the Secretary of the Treasury, in accordance with applicable Regulations in force at the time payment is made. Also, in compliance with Act No. 48-2013, as amended, a special contribution of one point five percent (1.5%) will be retained for all professional services, consulting, advertising, training or guidance contracts. This special contribution does not apply if the
services provided by individuals whose aggregate amount of government hiring (including agencies and public corporations) does not exceed fifty thousand ($50,000) dollars annually. Any partial or total waiver the CONSULTANT may have does not apply to this special contribution. [Act 237-2004, Act 48-2013]
C. The CONSULTANT certifies and warrants that at the time of the execution of this Agreement it has fulfilled his tax obligations for the past five (5) years prior to this contract, including the filing of Income Tax Returns and Sales and Use Tax for the past sixty (60) taxable periods (months) and does not have any current debt with the Commonwealth of Puerto Rico, relating to Income Tax, Sales and Use Tax, real property, personal property or excise taxes, or otherwise has a payment plan in place and is currently complying with its terms, contributions for unemployment insurance, temporary disability and drivers' social security taxes required by the Department of Labor and Human Resources of Puerto Rico or child support to ASUME (by its Spanish acronyms), as applicable, as defined in the Administrative Bulletin No. OE- 1991-24 issued by the Governor of the Commonwealth of Puerto Rico on June 18, 1991, as amended, and Circular Letter 1300-25-14 issued by the Department of Treasury. [Act 237-2004].
D. Pursuant to the provisions of Act No. 237 of August 31, 2004, as amended, the following documents are made part of this Agreement:
- Certification of No Debt, issued by the Department of Treasury (Form 6096);
- Certification of Tax Filing for the past five (5) years, issued by the Department of Treasury (Form 6088);
- Certification of Tax Filing for Sales and Use ("IVU") for the past sixty (60) taxable periods (Model 2942);
- Certification of Debt for Sales and Use Tax ("IVU") (Model 2927);
- Certification of Filing Personal Property Tax;
- Certification of Debt for All Concepts issued by the Center for Municipal Revenue Collections ("CRIM", by its Spanish acronyms);
- Certification of Registration as an Employer and of Debt regarding Unemployment and Disability Insurance, issued by the Department of Labor and Human Resources;
- Certification of Registration as an Employer and of Debt regarding Drivers' Social Security;
- Certificate of Good Standing;
- Certificate of Existence or Certificate of Authorization to do Business in Puerto Rico;
- Certificate of Incorporation;
- Corporate Resolution authorizing the person signing this Agreement on behalf of the CONSULTANT;
- Certification that the CONSULTANT has no child support debt as issued by the Administration for Child Support or Certified Statement [individuals] (ASUME, by its Spanish acronym);
• Certification of Compliance Status.
E. The CONSULTANT expressly acknowledges that the certifications listed above are essential to the validity of this Agreement and if incorrect in whole or in part, this will be sufficient cause for the AUTHORITY to rescind the Agreement immediately.
F. The CONSULTANT will provide evidence of his or her Certificate of Incorporation, Deed of Limited Partnership, Partnership and that certifies that he is in compliance with any patents, permits, licenses or authorizations to do the works required by the contract.
25. FAMILY ASSISTANCE AND ELDERLY SUPPORT
The CONSULTANT certifies, as required by the Law 3-2014, he is in compliance with Act 168-2000, as amended, better known as the Act for the Improvement of family Assistance and for the Support of the Elderly, if he was required by judicial or administrative order to perform any economic obligation or to comply with any obligation he is in compliance with the payment of the financial contribution or obligation imposed.
26. INSURANCE
Regarding design services under this contract, the CONSULTANT certifies that has submitted the requirements of insurance evidencing the policies and requirements that have been reviewed and approved by the AUTHORITY.
The CONSULTANT will obtain from a solvent insurance company and industry recognized, authorized to do business in the Commonwealth of Puerto Rico, the insurance described below:
a. PUBLIC LIABILITY INSURANCE
(1) Limits: covered
Damage to person - $200,000 \ 500,000 USD
Annexed property damage - $100,000
(2) Covers:
Endorsement designating the Highway and Transportation Authority as an additional insured Release of Liability ("Hold Harmless") Operations premises ("Premises operations")
b. PUBLIC LIABILITY CAR
(1) Limits covered:
Damage to people - $100,000 \ 300,000 USD
Damage to property of others - $100,000
(2) Covers:
Release of Liability ("Hold Harmless")
Cars owned, leased cars and cars not owned
Endorsement designating the Highway and Transportation Authority as an additional insured
c. PROFESSIONAL LIABILITY:
(1) Limits covered:
Up to $1,000,000 in damages to person, property or the project owner
(2) Covers:
Errors and omissions – The CONSULTANT will be liable for those costs incurred by the Authority during the construction of the project for errors and/or omissions in the design. The CONSULTANT will be responsible for maintaining coverage until completion of project’s construction, as long as the construction begins within the time period of 24 months from the date of PS&E approval date. The CONSULTANT must submit evidence when signing the contract.
d. WORKER’S INSURANCE POLICY
The CONSULTANT agrees to keep this coverage in place for the duration of the contract, to meet the AUTHORITY’s requirements, releasing the AUTHORITY from any liability arising out of this agreement, from the State Insurance Fund Corporation.
The CONSULTANT shall indemnify, defend and prevent any damage being caused to the Authority, its officers and employees, related to any loss, claim or liability (or any action that may be brought by a third party), based on any act or omission of the CONSULTANT for the provision of professional services under this contract.
The CONSULTANT shall indemnify, defend and hold harmless the rights of the Authority, as well as its directors, employees and agents against any judicial or extrajudicial claim, action, demand, and finally, all losses, damages, disbursement, outlay or expense related to injury or death of any person or property damage arising out of any negligent or willful, including its staff and agents in connection with this contract. The CONSULTANT shall also indemnify the Authority of any claim presented by any of its employees working in regard to this contract.
27. CRIME CONVICTION CLAUSE:
CONSULTANT certifies that in conformity with Article 5(Ñ) Law 237 of August 31, 2004 and Article 3.4 of Act 2-2018, supra, he/it has not been convicted of offenses against the public as defined in the Penal Code or misappropriation of public funds and that has not been declared guilty of this offense in the courts of the Commonwealth of Puerto Rico, in Federal courts or courts of any jurisdiction of the United States of America. If found guilty of the above referenced crimes, the contract for professional services or consulting will be TERMINATED.
The above referenced crimes are included in Article 3.4 of Act 2-2018, supra. Each and all amendments made during the term of the contract, to Articles 5(Ñ) of 237 Law, as well as to Article 3.4 of the aforementioned Act No. 2 will be an integral part of this clause. The
CONSULTANT expressly acknowledges that this certification is an essential condition of this Contract. If this certification is not correct in whole or in part shall constitute sufficient cause for the AUTHORITY to rescind the contract immediately without prior notification, and the CONSULTANT will have to pay the AUTHORITY any sum of money received to date. This obligation is continuous in nature at all stages of procurement and execution of the contract.
28. CRIMINAL RECORDS
A. The CONSULTANT certifies that in accordance with Articles 3.3 and 3.4 of the Código Anticorrupción, has not been convicted of crimes against public integrity as defined in the Criminal Code or embezzlement of public funds and that has not been found guilty of this type of crime in the Courts of the Government of Puerto Rico, in the Federal Courts or the Courts of any jurisdiction of the United States of America. If found guilty of the aforementioned offenses, this Contract will be terminated. It is accepted that these crimes are listed in Article 3.4 of the Código Anticorrupción. Any amendment to Articles 3.3 and 3.4 of the Código Anticorrupción, supra, during the term of this contract, is an integral part of this clause. It is expressly recognized that this certification is an essential condition of this Contract. If this certification is not correct in whole or in part, this will constitute sufficient cause for the AUTHORITY to cancel the contract immediately, without the need for prior notification. The CONSULTANT will have to pay to the AUTHORITY any amount of money received to date. This obligation is of a continuous nature during all stages of the contract.
B. The CONSULTANT is committed to report continuously, during the term of the contract, any fact that is related to any investigation for the commission of a crime against the treasury, the faith or the public function; against the governmental exercise; that involves funds or public property, at the federal or state level. This obligation is of a continuous nature during all stages of contracting and execution of the contract.
C. The CONSULTANT is committed to inform those cases in which there is still no determination of probable cause for the arrest, allegation of guilt or accusation against him, but have made expressions or admissions of crime against him.
D. The CONSULTANT certifies that the Affidavit required by Article 3.3 of the Código Anticorrupción was delivered.
29. GOVERNMENT ETHICS
a. THE CONSULTANT certifies compliance with the requirements of Law 1 of January 3, 2012, as amended, known as the Ethics in Government Act, including obtaining executive immunity where necessary, so that no officer or employee of the Authority or the Department of Transportation and Public Works nor their family members, nor any officer or employee of the Executive Branch has, directly or indirect personal or financial interest
in the profits or benefits of this contract. CONSULTANT certifies that have received a copy of Act No. 12, supra, at the time of the signing of this contract.
The CONSULTANT also certifies that:
(1) No member of the governing body of the Department of Transportation and Public Works or the AUTHORITY; no other official, agent, or employee of the AUTHORITY; nor their family members, who exercise any functions or responsibilities in the review or approval of the undertaking or carrying out of this contract, has a financial interest in this contract, purchase or transaction, and has not had in the last four (4) years direct or indirect personal or financial interest in this business.
(2) No member of the governing body of the Department of Transportation and Public Works or the AUTHORITY; no other official, agent, or employee of the AUTHORITY requested or accepted, directly or not directly, for him (her), any member of your household or for anyone, gifts, gratuities, favors, services, donations, loans or any other thing of monetary value.
(3) No member of the governing body of the Department of Transportation and Public Works or the AUTHORITY; no other official, agent, or employee of the AUTHORITY either requested or accepted any economic value associated with this transaction, any person of entity as payment for performing the duties and responsibilities of their jobs.
(4) No member of the governing body of the Department of Transportation and Public Works or the AUTHORITY; no other official, agent, or employee of the AUTHORITY requested, directly or indirectly, for him (her), any member of your household, or for any other person, business or entity, any good economic value, including gifts, loans, promises, favors or services in exchange in favor of the CONSULTANT.
(5) There are no family relationships within the fourth degree of consanguinity or second of affinity, with any employee of the AUTHORITY that has the power to influence and participate in institutional decisions of the Authority.
b. THE CONSULTANT agrees to comply with the provisions of Article III of Act. No. 2-2018, supra, which is set by the Code of Ethics for Contractors, Suppliers and Applicants Economic Incentives of the Executive Agencies the Commonwealth of Puerto Rico and the “Política Cultural Etica de la Autoridad de Carreteras y Transportación”, a copy of which are given to the CONSULTANT upon signing this contract.
The CONSULTANT is required to report those acts that are in violation of the “Código Anticorrupción”, constituting corruption or configured in crimes constitute fraud, bribery,
embezzlement or misappropriation of funds, among others have personal knowledge, regarding a contract, business or transaction between the government and a contractor, supplier of goods and services or participants of economic incentives.
c. The CONSULTANT certifies that upon signing this contract he/it not have any contracts with other government agency and if he/it does, such contract is not in conflict with this one.
d. The CONSULTANT agrees that knows the ethical standards of their profession and take responsibility for their actions.
30. CONFLICT OF INTEREST
a. The CONSULTANT understands that in the discharge of their professional duties, they have the obligation of complete loyalty towards the PRHTA and the Commonwealth of Puerto Rico. This includes not having, and avoiding, any interests that are adverse to the PRHTA or the Commonwealth of Puerto Rico. These adverse interests include the representation of clients that have or might have a conflict of interest with the PRHTA or the Commonwealth of Puerto Rico. Furthermore, this obligation includes the continuing obligation to divulge to the PRHTA all the circumstances of its relationships with clients and third parties, as well as any interest that may have an effect on the PRHTA or the Commonwealth of Puerto Rico at the time of granting this Agreement or during its effectiveness.
A conflict of interest exists when, for the benefit of a client, it is the CONSULTANT’S obligation to promote that which it should oppose in the fulfillment of its obligation towards a former, actual or potential client. Also, a conflict of interest takes place when conduct is defined as such in the ethical rules and standards recognized in the applicable profession, or in the laws and regulations of the Commonwealth of Puerto Rico. It will be a violation of the prohibitions described herein, if CONSULTANT or any subconsultants and/or subconsultants officers, directors, principals, partners, associates or employees engage in the aforementioned conduct. The CONSULTANT will avoid even the appearance of a conflict of interest.
The CONSULTANT acknowledges the authority of the PRHTA to assure compliance with the prohibitions herein contained. If the PRHTA believes that there is or has been a conflict of interest, or the appearance of a conflict of interest, it will notify the CONSULTANT in writing of the PRHTA's findings. The CONSULTANT shall make full disclosure of all material facts and shall have a period of thirty (30) days after receipt of such notice to cure the conflict of interest or the appearance of conflict of interest, including the right to request a meeting with the PRHTA to set forth or explain its position. This meeting will always be granted when timely requested. If the conflict of interest or appearance of conflict of interest is not cured to the satisfaction of the PRHTA or the controversy otherwise resolved prior to expiration of such 30-day period, this Agreement shall be resolved.
b. The CONSULTANT covenants that it presently has no interest and shall not acquire any interest, direct or indirect, which would conflict in any manner or degree with the performance of the services under this Agreement. The CONSULTANT further covenants that in the performance of this Agreement no person having any such interest shall be employed or subcontracted.
c. The CONSULTANT covenants that no member of the governing body of the DTPW and/or the PRHTA, and no other officer, agent, or employees of the DTPW and/or the PRHTA, who exercises any functions or responsibilities in the review or approval of the undertaking or carrying out of this Agreement has any personal or financial interest/direct or indirect, in this Agreement or in the proceeds thereof.
d. The CONSULTANT warrants that it will not and has not employed or retained any company or person, other than bona fide employees or subconsultants working solely for the CONSULTANT, to solicit or secure the work to be performed under this Agreement, and that it has not paid or agreed to pay any company or person, other than bona fide employees or subconsultants working solely for the CONSULTANT, any fee, commission, percentage, brokerage fee, gifts or any other considerations, contingent upon or resulting from the award or execution of his Agreement.
e. The CONSULTANT represents that it has secured or will secure at its own expense, all personnel required in performing the services under this Agreement. Such personnel shall not be employees of or have any contractual relationship with the DTPW and/or the PRHTA.
f. The CONSULTANT certifies that none of its officers or employees, nor any principals, associates, officers, directors or shareholders, are public officials or employees within the meaning of Law No. 1-2012. The CONSULTANT certifies the receipt of a copy of this law at the moment of signing this agreement.
g. The CONSULTANT also hereby agrees to comply with the provisions of Article III Law No. 2 of January 4, 2018, establishing the Code of Ethics for Contractors, Suppliers and Petitioners of Economics Incentives, of the executive agencies of the Commonwealth of Puerto Rico and certifies the receipt of a copy of this law at the moment of signing this agreement.
h. No person employed in the work covered by this Agreement shall be discharged or in any way discriminated against because he/she has filed any complaint or instituted or caused to be instituted any proceeding or has testified or is about to testify in any proceeding relating to the labor standards applicable hereunder to this employer.
i. The CONSULTANT certifies that its employees do not receive and shall not receive during the term of this Agreement, salaries or compensations from any agency and/or instrumentality of the Commonwealth of Puerto Rico. The CONSULTANT also certifies that it does not have, nor shall have during the term of this Agreement, any other contract or agreement with said agencies and instrumentalities, which is, or could be, in conflict with this Agreement. The CONSULTANT also certifies that at present it does have contracts with agencies or instrumentalities of the Commonwealth of Puerto Rico.
31. CONTRACT ENFORCEMENT, JURISCICTION AND APPLICABLE LAW
The parties irrevocably submit to the jurisdiction of the Commonwealth of the Puerto Rico, Superior Court of First Instance, San Juan Part, regarding any dispute arising or in relation to this Agreement. This contract is subject to the laws of the Commonwealth of Puerto Rico, as well as the applicable federal regulations. Whenever the AUTHORITY is obliged to resort to the courts to enforce compliance with this Contract and/or any other action that may arise, the CONSULTANT undertakes to pay all the costs, attorneys' fees and necessary disbursements, but in any event, shall be an amount of two thousand dollars ($2,000.00) or ten percent (10%) of the Contract amount, whichever is greater. This amount will be due and mandatory on filing of the case in court, even if the case is heard in absence.
32. ACCESS TO FILES, REPORTS AND RECORDS
THE CONSULTANT shall keep all books, records, accounts, driving bills, vouchers, checks, payroll, purchase orders, worksheets, and payment of services or any other document evidencing the expenses incurred budgeted required under this contract for a term of no less than ten (10) years after the date of termination or expiration of this Contract, except in the event of litigation or settlement of claims arising from the performance of this contract, in which case CONSULTANT agrees to maintain the same until the AUTHORITY, municipal corporation, Federal Highway Administrator (FHWA) Administrator, the Comptroller General of the United States, Office of Inspector General (OIG), or any of their authorized representatives, have disposed of all such litigation, appeals, claims or exceptions related thereto.
The CONSULTANT shall, upon request, give complete access to all books, records, accounts, driving bills, vouchers, checks, payroll, purchase orders, worksheets, and payment of services of the CONSULTANT, which are directly pertinent to this Contract, to AUTHORITY, the Federal Highway Administration, the Comptroller General of the United States, Office of Inspector General (OIG), the Legislature of Puerto Rico, Office of Government Ethics, the Office of the Comptroller of Puerto Rico, the Department of Treasury or any other authorized agency. CONSULTANT shall permit any of the foregoing parties to reproduce by any means whatsoever or to copy excerpts and transcriptions as reasonably needed.
33. NOTIFICATION
Any notice relating to this Agreement shall be made in writing and shall be processed by hand or by registered mail. Notifications to the AUTHORITY shall be delivered to the following address:
Executive Director
Puerto Rico Highway and Transportation Authority
PO Box 42007
San Juan, Puerto Rico, 00940-2007
Notices to the CONSULTANT shall be equally directed to:
Wilfredo M. Rodríguez Uviñas
PO Box 9020460
San Juan, PR 00902-0460
34. CONTRACT TRANSFER RIGHTS
Notwithstanding the Authority may transfer its rights and obligations under this agreement, providing a written notification to the CONSULTANT.
35. EQUAL EMPLOYMENT AND NON DISCRIMINATION
a. The CONSULTANT undertakes to guide employees, supervisors and sub consultants on the existing laws of equal employment opportunity and sexual harassment, and to comply with administrative policies adopted for this purpose. The CONSULTANT also agrees not to discriminate against any employee because of race, color, sex, age, religion, social status, national origin, political beliefs, veteran status or disability, during implementation and enforcement of this contract.
b. The CONSULTANT undertakes not to discriminate in the provision of services based on political, religious, race, social status, age, sex, nationality, or physical or mental impairment.
36. COMPLIANCE WITH PUBLIC POLICY
This contract will be resolved immediately if the CONSULTANT incur in acts contrary to the AUTHORITY’s public policy concerning sexual harassment issues, discrimination or use and abuse controlled substances on the job, during the course of their duties, and pursuant to this Agreement.
37. REGISTRATION AT THE OFFICE OF THE COMPTROLLER
No provision for compensation under this contract may be required until it has been submitted for registration at the Office of the Comptroller of Puerto Rico, pursuant to the provisions of Law No. 18 of October 30, 1975, as amended.
38. CONTRACT CHANGES
The AUTHORITY may, at any time during the term of this contract, by mutual agreement with the CONSULTANT, incorporate changes and/or conditions it deems appropriate or necessary to make better use of the services herein contracted to adjust to the needs of the AUTHORITY. Such changes will be incorporated by means of a written amendment to this contract.
39. SAVING CLAUSE
It is hereby stipulated that the terms and conditions of this agreement are separate and apart and the invalidity of one or more clauses does not affect the validity of the remaining terms and conditions, which shall survive. Both parties accept the conditions set forth herein, which are mandatory to comply.
40. PREFERENCE
The CONSULTANT shall use and shall give preference to those items extracted, produced, assembled or packaged in Puerto Rico or distributed by agents established herein, provided that they have to meet the specifications, terms and conditions set forth in the statement of auction or purchase order and its price, after applying the appropriate investment parameters, is the lowest, according to the Law for the Puerto Rican Industry Investment, Act No. 14 of January 8th, 2004.
41. RETENTION OF PAYMENT DUE
THE AUTHORITY may withhold payment due to the CONSULTANT for services rendered any sum of money equivalent to the tax debt owed by the CONSULTANT to any of the government instrumentalities for the Treasury Department, and where such tax debt has been properly assessed by the Treasury Department, except for those instances where the tax debt is under a pre-arranged payment plan between the CONSULTANT and the Treasury Department or where the CONSULTANT can be awaiting a decision after having submitted a petition for review following the procedures established by the applicable laws and regulations, thus involving the CONSULTANT to cancel the debt, as appropriate, if any.
42. DISADVANTAGE BUSINESS ENTERPRISES ASSURANCES
This Contract is subject to the requirements of Title 49, Code of Federal Regulations, Part 26, participation by Disadvantaged Business Enterprises in Department of Transportation Financial Assistance Programs.
To the extent required by Federal law, the CONSULTANT agrees to facilitate participation by Disadvantaged Business Enterprises (DBE) in the Project and assures that each sub-recipient, lessee, and third-party contractor at any tier of the Project will facilitate participation by DBEs in the Project to the extent applicable.
The CONSULTANT agrees and assures he shall not discriminate on the basis of race, color, national origin, or sex in the performance of this Contract origin in the award and performance of this Contract or of any third party contract, or sub agreement supported with Federal assistance.
derived from U.S. DOT and will comply with the requirements of 49 C.F.R. Part 26. The CONSULTANT shall carry out applicable requirements of 49 CFR Part 26 in the award and administration of this Contract. Failure by the CONSULTANT to carry out these requirements is a material breach of this Contract, which may result in the termination of this Contract or such other remedy as the AUTHORITY deems appropriate. Each subcontract the CONSULTANT signs with a sub-consultant must include the assurance in this paragraph (see 49 CFR 26.13(b)).
The CONSULTANT is required to pay its sub-consultants performing work related to this Contract for satisfactory performance of that work no later than fifteen (15) calendar days after the CONSULTANT’s receipt of payment for that work from the AUTHORITY. The CONSULTANT will be required to submit documentation of all payments made to DBEs and non-DBEs irrespective of their tier. In addition, the CONSULTANT may not hold retainage from its sub-consultants, is required to return any retainage payments to those sub-consultants within 30 calendar days after the sub-consultant's work related to this Contract is satisfactorily completed and is required to return any retainage payments to those sub-consultants within thirty (30) calendar days after incremental acceptance of the sub-consultant’s work by the AUTHORITY and CONSULTANT’S receipt of the partial retainage payment related to the sub-consultant’s work.
The CONSULTANT must promptly notify the AUTHORITY, whenever a DBE sub-consultant performing work related to this contract is terminated or fails to complete its work and must make good faith efforts to engage another DBE sub-consultant to perform at least the same amount of work of equal or greater value. The CONSULTANT may not terminate any DBE sub-consultant and perform that work through its own forces or those of an affiliate without prior written consent of the AUTHORITY. The CONSULTANT agrees that the implementation of this DBE program is a legal obligation, and that failure to carry out - the DBE program shall be treated as a violation of the Contract. The CONSULTANT shall carry out applicable requirements of 49 CFR part 26 in the award and administration of DOT-assisted contracts. Failure by the CONSULTANT to carry out these requirements is a material breach of this contract, which may result in the termination of this contract or such other remedy as the AUTHORITY deems appropriate, which may include, but is not limited to:
1. Withholding monthly progress payments;
2. Assessing sanctions;
3. Liquidated damages; and/or
4. Disqualifying the CONSULTANT from future participation as non-responsible,
5. Other remedies the AUTHORITY deems appropriate
6. Termination
The AUTHORITY may impose sanctions as provided for under 49 C.F.R. Part 26 and may, in appropriate cases, refer the matter for enforcement under 18 U.S.C. § 1001, and/or the Program Fraud Civil Remedies Act, 31 U.S.C. §§ 3801 et seq.
43. LOBBYING CERTIFICATION (49 CFR PART 20)
The CONSULTANT certifies, to the best of his or her knowledge and belief, that:
a. No Federal appropriated funds have been paid or will be paid, by or on behalf of the CONSULTANT, to any person for influencing or attempting to influence an officer or employee of an agency, a Member of Congress, an officer or employee of Congress, or an employee of a Member of Congress in connection with the awarding of any Federal contract, the making of any Federal grant, the making of any Federal loan, the entering into of any cooperative agreement, and the extension, continuation, renewal, amendment, or modification of any Federal contract, grant, loan, or cooperative agreement.
b. If any funds other than Federal appropriated funds have been paid or will be paid to any person for making lobbying contacts to an officer or employee of any agency, a Member of Congress, an officer or employee of Congress, or an employee of a Member of Congress in connection with this Federal contract, grant, loan, or cooperative agreement, the CONSULTANT shall complete and submit Standard Form--LLL, "Disclosure Form to Report Lobbying," in accordance with its instructions [as amended by "Government wide Guidance for New Restrictions on Lobbying," 61 Fed. Reg. 1413 (1/19/96). Note: Language in paragraph (2) herein has been modified in accordance with Section 10 of the Lobbying Disclosure Act of 1995 (P.L. 104-65, to be codified at 2 U.S.C. 1601, et seq.)]
c. The CONSULTANT shall require that the language of this certification be included in the award documents for all subawards at all tiers (including subcontracts, subgrants, and contracts under grants, loans, and cooperative agreements) and that all subrecipients shall certify and disclose accordingly.
d. The CONSULTANT acknowledges that this certification is a material representation of fact upon which reliance was placed when this transaction was made or entered into and that submission of this certification is a prerequisite for making or entering into this transaction imposed by 31, U.S.C. § 1352 (as amended by the Lobbying Disclosure Act of 1995). Furthermore, the CONSULTANT acknowledges that if he fails to file the required certification he shall be subject to a civil penalty of not less than $10,000 and not more than $100,000 for each such failure.
44. CIVIL RIGHTS-TITLE VI ASSURANCE
The following requirements apply to the underlying Contract:
a. Nondiscrimination - In accordance with Title VI of the Civil Rights Act, as amended, 42 U.S.C. § 2000d, section 303 of the Age Discrimination Act of 1975, as amended, 42 U.S.C. § 6102, section 202 of the Americans with Disabilities Act of 1990, 42 U.S.C. § 12132,
and Federal transit law at 49 U.S.C. § 5332, the CONSULTANT agrees that it will not discriminate against any employee or applicant for employment because of race, color, creed, national origin, sex, age, or disability. In addition, the CONSULTANT agrees to comply with applicable Federal implementing regulations and other implementing requirements FHWA may issue.
b. Equal Employment Opportunity - The following equal employment opportunity requirements apply to the underlying Contract:
1. Race, Color, Creed, National Origin, Sex - In accordance with Title VII of the Civil Rights Act, as amended, 42 U.S.C. § 2000e, and Federal transit laws at 49 U.S.C. § 5332, the CONSULTANT agrees to comply with all applicable equal employment opportunity requirements of U.S. Department of Labor (U.S. DOL) regulations, "Office of Federal Contract Compliance Programs, Equal Employment Opportunity, Department of Labor," 41 C.F.R. Parts 60 et seq., (which implement Executive Order No. 11246, "Equal Employment Opportunity," as amended by Executive Order No. 11375, "Amending Executive Order 11246 Relating to Equal Employment Opportunity," 42 U.S.C. § 2000e note), and with any applicable Federal statutes, executive orders, regulations, and Federal policies that may in the future affect construction activities undertaken in the course of the Project. The CONSULTANT agrees to take affirmative action to ensure that applicants are employed, and that employees are treated during employment, without regard to their race, color, creed, national origin, sex, or age. Such action shall include, but not be limited to, the following: employment, upgrading, demotion or transfer, recruitment or recruitment advertising, layoff or termination; rates of pay or other forms of compensation; and selection for training, including apprenticeship. In addition, the CONSULTANT agrees to comply with any implementing requirements FHWA may issue.
2. Age - In accordance with section 4 of the Age Discrimination in Employment Act of 1967, as amended, 29 U.S.C. §§ 623 and Federal transit law at 49 U.S.C. § 5332, the CONSULTANT agrees to refrain from discrimination against present and prospective employees for reason of age. In addition, the CONSULTANT agrees to comply with any implementing requirements FHWA may issue.
3. Disabilities - In accordance with section 102 of the Americans with Disabilities Act, as amended, 42 U.S.C. § 12112, the CONSULTANT agrees that it will comply with the requirements of U.S. Equal Employment Opportunity Commission, "Regulations to Implement the Equal Employment Provisions of the Americans with Disabilities Act," 29 C.F.R. Part 1630, pertaining to employment of persons with disabilities. In addition, the CONSULTANT agrees to comply with any implementing requirements FHWA may issue.
c. The CONSULTANT also agrees to include these requirements in each subcontract financed in whole or in part with Federal assistance provided by FHWA, modified only if necessary to identify the affected parties.
45. TITLE VI ASSURANCE AND NON-DISCRIMINATION PROVISIONS
During the performance of this contract, the CONSULTANT, for itself, its assignees, and successors in interest agrees as follows:
A. Compliance with Regulations: The CONSULTANT will comply with the Acts and the Regulations relative to Non-Discrimination in Federally-assisted programs of the U.S. Department of Transportation; as they may be amended from time to time, which are herein incorporated by reference and made of this contract.
B. Non-discrimination: The CONSULTANT, with regard to the work performed by it during the contract, will not discriminate on the grounds of race, color, or origin in the selection and retention of subconsultants, including procurements of materials and leases of equipment. The CONSULTANT will not participate directly or indirectly in the discrimination prohibited by the Acts and Regulations, including employment practices when the contract covers any activity, project, or program set forth in Appendix B of 49 CFR Part 21.
C. Solicitations for Subcontracts, including Procurements of Materials and Equipment: In all solicitations, either by competitive bidding, or negotiation made by the consultant for work to be performed under a subcontract, including procurements of materials, or leases of equipment, each potential subconsultant or supplier will be notified by the consultant of the consultant's obligations under this contract and the Acts and the Regulations relative to Non-discrimination on the grounds of race, color, or national origin.
D. Information and Reports: The CONSULTANT will provide all information and reports required by the Acts and the Regulations, and directives issued pursuant thereto and will permit access to its books, records, accounts, other sources of information, and its facilities as may be determined by the AUTHORITY to be pertinent to ascertain compliance with such Acts, Regulations, and instructions. Where any information required of a CONSULTANT is in the exclusive possession of another who fails or refuses to furnish the information, the CONSULTANT will also certify to the AUTHORITY, as appropriate, and will set forth what efforts it has made to obtain the information.
E. Sanctions for Noncompliance: In the event of a CONSULTANT'S noncompliance with the Non-discrimination provisions of this contract, the AUTHORITY will impose such contract sanctions as it or may determine to be appropriate, including, but not limited to:
a. withholding payments to the CONSULTANT under the contract until the CONSULTANT complies; and or
b. cancelling, terminating, or suspending the contract, in whole or in part.
F. Incorporation of Provisions: The CONSULTANT will include the provisions of paragraphs A to F of this clause in every subcontract, including procurements of materials and leases of equipment, unless exempt by the Acts, the Regulations and directives issued pursuant thereto. The CONSULTANT will take action with respect to any subcontract or procurement as the AUTHORITY may direct as a means of enforcing such provisions including sanctions for noncompliance. Provided, that if the contractor becomes involved in, or is threatened with litigation by a subcontractor, or supplier because of such direction, the contractor may request the Recipient to enter into any litigation to protect the interests of the Recipient. In addition, the contractor may request the United States to enter into the litigation to protect the interests of the United States.
46. INTERAGENCY SERVICES CLAUSE
Both contracting parties acknowledge and agree that the contracted services may be provided to any entity of the Executive Branch with which the contracting entity makes an interagency agreement or by direct provision of the Chief of Staff of the Governor of Puerto Rico. These services will be performed under the same terms and conditions in terms of hours of work and compensation set forth in this contract. For purposes of this clause, the term "Executive Branch entity" includes all agencies of the Government of Puerto Rico, as well as public instrumentalities and corporations and the Office of the Governor.
IN WITNESS WHEREOF, and for the record, the parties express that this agreement contains a true and exact relation of the terms and conditions agreed. The appearing parties agree that they have read this entire Agreement, understood the same, so accepted and signed this contract set to be agreed today 3 of August, 2018.
Wilfredo M. Rodriguez Uviñas, PE
Representative of Consulting Firm
Id. No.
Carlos M. Contreras Aponte
Executive Director of the Puerto Rico Highway and Transportation Authority
Id. No.
APPENDIX A
SCOPE OF PROFESSIONAL DESIGN SERVICES
HURRICANE MARÍA EFFORTS FOR
SIGNING AND GUARDRAILS FOR DORADO, MANATI, TOA ALTA, TOA BAJA,
VEGA ALTA, AND BAYAMON MUNICIPALITIES
AC-814541
ARTICLE I - PROJECT DESCRIPTION, TIME OF BEGINNING AND COMPLETION OF
SERVICES TO BE RENDERED
SECTION A - PROJECT DESCRIPTION
Professional services for this PROJECT consists in the preparation of Preparation of Plans,
Specifications, and Engineer's estimates (PS&E), HURRICANE MARÍA EFFORTS FOR
SIGNING AND GUARDRAILS FOR DORADO, MANATI, TOA ALTA, TOA BAJA, VEGA ALTA,
AND BAYAMON MUNICIPALITIES. See the Municipalities and Road List on ATTACHMENT 1.
SECTION B - BEGINNING AND COMPLETION OF SERVICES TO BE RENDERED
The CONSULTANT agrees to begin the services stipulated here after receiving written
notification by the AUTHORITY from which he can start the services described on Article II,
SECTIONS A, B, C, D, E and F of APPENDIX "A" (SCOPE OF PROFESSIONAL DESIGN
SERVICES), APPENDIX "B" (SCOPE OF PROFESSIONAL ENVIRONMENTAL SERVICES), and
according to the services that apply on this CONTRACT.
By the own nature of this CONTRACT between the AUTHORITY and the CONSULTANT, where
is required a Notice to Proceed in order to do different tasks, such as, but not limited to, route location
and reconnaissance studies, field preliminary studies and plans, final construction plans,
specifications, cost estimates and contract documents, that as well, involve an extensive participation
of other state or federal agencies in order to obtain various environmental approvals, from studies
and permits from regulatory agencies, both sides agreed that the effective date on the Comptroller
Office Certification on the granting of this CONTRACT correspond to the minimum time the
CONSULTANT has to provide his professional services, and that the termination date of this
CONTRACT shall be adjusted accordingly, by means of written amendments, according to the
established Notices to Proceed to terminate the professional services under CONTRACT, as
provided by it.
The CONSULTANT shall be, from that moment responsible to obtain all the necessary data and
studies to complete each one of the phases (Example: Field Surveying, Geotechnical Studies, H-H
and Others). The CONSULTANT shall submit a Work Schedule for the Authority's review and
approval, within three (3) calendar days of receiving the Notice to Proceed. This Approved Schedule shall be the baseline for all services to be rendered.
The services to be provided within the described phases in this APPENDIX "A" can be extended from the established time on the Approved Schedule as long as the CONSULTANT justifies the reasons of the delay, puts them in writing to the AUTHORITY’s Design Area and as long as the total validity of the CONTRACT is not affected. The Design Area shall evaluate the time extension request and shall respond to it in writing accordingly.
**A - Preliminary Studies and Plans**
1. **Phase I: Field Inspection and Data Collection**
For Phase I of the services described under ARTICLE II, SECTION A, item 1 of APPENDIX A, the CONSULTANT agrees to complete the services as follows:
a. The necessary printed copies of the data collection in the requested field inspection tables (ATTACHMENT 2) shall be submitted to the AUTHORITY in 4 weeks for the 50% of the roads and subsequent 4 weeks for the another 50% of the assigned roads. A CD containing a non-editable (PDF) electronic file of the deliverables shall be submitted. In addition, a separate CD containing editable (XLS) digital format of the tables shall be submitted.
2. **Phase II: Preliminary plans**
For Phase II of the services described under ARTICLE II, SECTION C, ITEM 1 of APPENDIX A, the CONSULTANT agrees to complete the services as follows:
a. The necessary printed copies of the Preliminary Plans, shall be submitted to the AUTHORITY 4 weeks later from the approval of phase I works. A CD containing a non-editable (PDF) electronic file of the deliverables shall be submitted. In addition, a separate CD containing editable (DWG) digital format of the plans shall be submitted.
b. Being necessary additional revisions to the preliminary plans, the AUTHORITY shall decide with the CONSULTANT the delivery time of such, which shall be notified to the CONSULTANT in writing.
**B - Final Construction Plans, Specifications, Estimates and Contract Documents**
For the services described under ARTICLE II, SECTION D, of APPENDIX “A”, the CONSULTANT agrees to complete the services within the before indicated time. At the same time, the CONSULTANT shall fulfill Design Directive 107, and properly complete and submit Design Directive 111.
Being necessary revisions to the final plans, the AUTHORITY shall decide with the CONSULTANT the delivery time of such, which shall be notified to the CONSULTANT in writing.
The CONSULTANT shall submit a printed copy (17 in. wide by 11 in. high) of final plans and in two separate compact discs (CD), one in editable (DWG) and another in non-editable (PDF) digital formats programmed to be printed in 36 in. wide by 24 in. high, which contain all the drawings included in the final plans. The editable format shall be in Autocad 2007 version or the one that both the parties decide.
C - **Structural Report of the Impacted Overhead Signs**
The Consultant shall submit a printed copy (11 in. wide by 8.5 in. high) of the structural report (one-page) about impacted overhead signs. The report to be submitted shall include a structural analysis and recommendations to the Authority. A Licensed Professional Engineer shall certify the report.
D - **Acceptance of Completion by the AUTHORITY**
The CONSULTANT’s services under each one of the phases described in ARTICLE II, SECTIONS A, B, C, D, E and F of APPENDIX “A”, shall be considered completed after written approval by the AUTHORITY.
**ARTICLE II - SERVICES BY THE CONSULTANT**
The CONSULTANT agrees to provide the various professional services required for execution and/or preparation of all the surveys, investigations, reports, plans, contract documents and estimates, according to the PROJECT’s Request for Proposal (RFP) Scope of Work (SOW) which served as base to establish the services proposal of this CONTRACT.
The CONSULTANT agrees to provide these services fulfilling and according to the AUTHORITY’S latest version of the following documents: Highway Design Manual, Standard Specifications for Road and Bridge Construction and Standard Drawings, American Association of State Highway and Transportation of Officials (AASHTO) standards, Federal Highway Administration (FHWA) standards and Manual on Uniform Traffic Devices Control (MUTCD) standards.
The CONSULTANT shall have to fulfill the dispositions contained on the Design Directives published by the Design Area, especially Design Directives 107 and 111. In addition, the CONSULTANT shall have to fulfill the dispositions contained in the latest version of the following documents: Permit Management Office (OGPe) regulations, Regulation of Infrastructure Arrangement on Public Space (Planning Regulation 25) from the Planning Board and the Department of Natural and Environmental Resources; and the Americans with Disabilities Act (ADA) regulations, especially the suggested public space design guides in the publication "Accessible Rights of Way, A Design Guide", U.S. Architectural and Transportation Barriers Compliance Board.
If a conflict arises between the design publications requirements and the regulations before mentioned, the CONSULTANT shall consult to the AUTHORITY'S Design Area over which disposition to use or apply on the PROJECT described in this CONTRACT, services that shall be provided in agreement with the generally recognized professional standards in the engineering field.
In the eventuality that, after the CONTRACT is signed, manuals and the standards before mentioned require from the CONSULTANT additional services from the ones already provided, or they require major changes on scope and/or in the completed work, which it entails as well additional expenses for the CONSULTANT, this one shall be compensated in a reasonable amount by the AUTHORITY for the incurred expenses.
SECTION A - PRELIMINARY STUDIES
ITEM 1: Field Inspection and Data Collection
For ITEM 1, the CONSULTANT will receive from the Authority a list of roads by municipality to be traveled and inspected. As part of the Data Collection, the CONSULTANT will follow the following notes:
1. Collect the data for damaged guardrails segments by trees or other objects, especially in the top side of the segment, will be included as part data collection. The minimum length for guardrail replacement currently meeting PRHTA Standards is 7.62 mts. Guardrail segments not meeting PRHTA Standards shall be replaced by guardrail meeting PRHTA Standards by a minimum section of 30 mts, not to exceed 100 mts, and with the appropriate terminals. Transitions shall be provided, as required.
2. Collect the data of damaged and/or missing signs. The CONSULTANT will specify in the data collection tables, which of the following conditions were encountered in the field inspection:
a. Traffic sign missing and there is only the base post.
b. Traffic sign in place but rotated from the original direction to traffic.
c. Traffic sign in place but does not comply with PRHTA standards.
d. Traffic sign that is necessary to be installed but there is no proof of previous existence.
3. Only stop signs will be installed at municipal road intersecting a state road.
4. For partially damaged overhead signs, include a description of the proposed work.
5. Corridors with lower road number will include an evaluation of full intersections of roads with higher road number along the corridor.
SECTION B - ENVIRONMENTAL ASSESSMENT
The CONSULTANT shall participate on public information hearings coordinated by the AUTHORITY, as accorded. All environmental compromises and mitigation measures established on the Finding of No-Significant Impact (FONSI) or Environmental Impact Statement (EIS) shall be implemented during the PROJECT’s design phase. All the services are described in Appendix B; SCOPE OF PROFESSIONAL ENVIRONMENTAL SERVICES.
SECTION C - PRELIMINARY PLANS
ITEM 1: Preliminary plans
For ITEM 1, the CONSULTANT shall collaborate with the AUTHORITY’S Interdisciplinary Team for the follow up of the PROJECT as per Action Plan.
A. Preliminary Plans
For this phase, the CONSULTANT shall prepare preliminary plans which shall include and consider, but not limited to, the following:
1. title sheet;
2. typical sections;
3. MOT plans;
4. guardrail plans;
5. signing plans;
6. and preliminary cost estimate.
The CONSULTANT shall use aerial photos to elaborate a lineal plan with the road borders. In the plans, the reference system shall be the current road kilometers.
B. Coordination with Agencies
All necessary meetings between the AUTHORITY, private stakeholders, and other concerned state and/or federal government agencies will be held for the proper PROJECT coordination.
C. Coordination with Environmental Studies Office
The CONSULTANT as part of this scope, shall provide the necessary services, technical support, and information to the PRHTA Environmental Studies Office in the development of an Islandwide Categorical Exclusion that is needed in the development of the project and to complete the federal NEPA process and environmental permits.
SECTION D - FINAL CONSTRUCTION PLANS, SPECIFICATIONS, ESTIMATES AND CONTRACT DOCUMENTS
Upon approval of the Preliminary Plans and receipt of written Notice to Proceed from the AUTHORITY, the CONSULTANT shall proceed with the design of the PROJECT’S FINAL
CONSTRUCTION PLANS, SPECIFICATIONS, COST ESTIMATES AND CONTRACT DOCUMENTS as a whole or as individual sections, as may be required by the AUTHORITY, as follows:
ITEM 1: Size and Scales
As specified in the Highway Design Manual.
ITEM 2: Roadway Plans
Prepare complete roadway plans including, but not limited to, geometric layout, safety plans including guardrails, intersection or interchange details, signing location and details, temporary traffic control plan, typical and standard roadway cross-sections, and the necessary information for the PROJECT’s construction, as specified in the Highway Design Manual.
The roadway plans shall include all the recommendations made in data collection process as part of the preliminary design.
ITEM 3: Special Provisions
Compile special provisions applicable to the AUTHORITY’s Standard Specifications and Supplementary Special Specifications, as necessary.
ITEM 4: Estimates
Prepare quantity and cost estimates based on the construction plans for all work items, as required. The CONSULTANT shall document and furnish the AUTHORITY all quantity calculations including the evidence used to arrive at such. This documentation shall be used to review the cost and quantity estimates included on the contract documents.
In addition to, unit costs shall be calculated by item based on labor requirement, materials and equipment analysis based on day costs, and the sound engineering practice. In order to carry out this work the CONSULTANT shall hire qualified staff (estimator) who is capable and has the adequate experience for such and the CONSULTANT shall be responsible that all estimated contract item quantities are correct, thus avoiding a possible unbalanced bid.
ITEM 5: Contract Documents
Prepare contract documents for the PROJECT or PROJECT’s sections according to the standard practices in use by the AUTHORITY, and shall perform an evaluation of the design plans and make sure these are consistent with specifications, notes, etc.
ITEM 6: Final Review
The necessary printed copies shall be submitted to the AUTHORITY according to Design Directive 107, of the completed roadway plans of the PROJECT, special dispositions, estimates and contract documents for their final review and approval. In addition, a CD with editable (DWG) and non-editable (PDF) digital formats programmed to be printed in 36 in. wide by 24 in. high, which contain all the drawings included in the final plans. The editable format shall be in Autocad 2007 version or the one that both the parties decide.
The AUTHORITY shall be able, after reviewing the plans, to approve them as submitted or require modifications to them if these are not in accordance to the criteria agreed during the design process and previous phases. In such case the CONSULTANT shall make the required necessary modifications or changes without any additional cost for the AUTHORITY.
ITEM 7: Final Delivery
Upon approval from the AUTHORITY, the CONSULTANT shall submit the original, and two (2) printed copies of the signed and sealed construction plans; originals and two (2) printed copies of the estimate and contract documents. In addition, he shall furnish all basic survey notes and sketches, prepared studies, all calculations made during the design stage, for example structural, geometry elements, amount of earthwork, etc., aerial photos used and any other graphical material, studies or documents used for the PROJECT’s design. All these documents shall become property of the AUTHORITY according to professional practice.
In addition, the CONSULTANT shall submit the final plans in compact discs (CD), in Autocad (DWG) 2007 or the version agreed by both parties and non-editable (PDF) digital formats programmed to be printed in 36 in. wide by 24 in. high, which contain all the drawings included in the final plans. The rest of the final delivery documents shall be submitted in non-editable (PDF) digital format, saved on compact disks (CD).
ITEM 8: Shop Drawings and Submittals
It shall be the CONSULTANT’S responsibility to review and endorse all the shop and facility drawings, and submittals prepared by the manufacturers and/or contractors during the construction of the PROJECT, no later than ten (10) calendar days from the AUTHORITY’S service request date.
Submittals which entail a design reevaluation or additional work, the CONSULTANT shall be compensated according to an approved proposal by the AUTHORITY in a CONTRACT’s addendum.
SECTION E - MEETINGS, SITE VISITS, INSPECTION OF WORKS
ITEM 1: Meetings
During design period, periodic meetings for the discussion of the work shall be held at the written request of either party. The CONSULTANT is requested to submit minutes of these meetings for approval by the AUTHORITY no later than seven (7) calendar days after these meetings.
ITEM 2: Visits to the PROJECT’s Site
Written request for visits may be made by the AUTHORITY or the CONSULTANT together with any other party or parties. A field inspection report shall be submitted by the CONSULTANT for each visit, no later than seven (7) calendar days after the visit.
ITEM 3: Clarification or Interpretation
The services of the CONSULTANT shall be available, free of cost for the AUTHORITY, at any time for clarification or interpretation of the plans and/or specifications, change orders, or additional work, if the AUTHORITY requires his participation.
ITEM 4: Inspection of Works
The AUTHORITY could, at any time during the validity of the CONTRACT, ask the CONSULTANT a status report of the works that are been performed under this CONTRACT, as well as a status report of the works that the CONSULTANT has subcontracted under this CONTRACT.
SECTION F - MEETINGS AND PUBLIC HEARINGS
The CONSULTANT shall attend all public meetings and hearings, pre-auctions and pre-construction meetings, and shall prepare the necessary information and documents including drawings, cost estimates and other necessary documentation to clarify and to explain the proposed PROJECT to the public.
The CONSULTANT shall attend all meetings with the Federal Highways Administration (FHWA) and the Interdisciplinary Team when required and shall prepare minutes of these meetings which shall be furnished to the AUTHORITY.
ARTICLE III - DATA OR INFORMATION TO BE PROVIDED TO THE CONSULTANT BY THE AUTHORITY IN RELATION TO THE PROJECT
The AUTHORITY agrees to provide at no cost to the CONSULTANT, the following:
SECTION A - TRAFFIC
All traffic information requested by the CONSULTANT and the AUTHORITY has it available and which is related to the PROJECT under this CONTRACT, subjected to the dispositions in ARTICLE II of this CONTRACT.
SECTION B - ACCESS TO EXISTING INFORMATION
The access to and use of all existing reports, computations, survey data, and other existing information in possession of the AUTHORITY which might prove pertinent to the contracted work and requested by the CONSULTANT, subjected to the dispositions on the in the next paragraph. The CONSULTANT shall be responsible to request any other information that is not in possession of the AUTHORITY to the pertinent agency or organization.
The AUTHORITY can, if available and to his discretion, supply to the CONSULTANT all available information and which is related to the PROJECT under this contract. The CONSULTANT is obligated to review this information, determine its adequacy and correctness of such, and shall take all the necessary steps to complete and edit the supplied information, following the industry standards and the Ethics Code of the College of Engineers and Surveyors of Puerto Rico. Any additional cost incurred by the CONSULTANT in order to complete the original information supplied by the AUTHORITY, shall be reimbursed to the CONSULTANT by the AUTHORITY after he certifies such disbursement to the AUTHORITY.
SECTION C - NOTICE TO PROCEED
Notice to Proceed in writing for each phase from the work and Notification of approval in writing for each design phase as specified in this CONTRACT.
APPENDIX B
SCOPE OF PROFESSIONAL ENVIRONMENTAL SERVICES
(IF REQUIRED)
HURRICANE MARÍA EFFORTS FOR
SIGNING AND GUARDRAILS FOR DORADO, MANATI, TOA ALTA, TOA BAJA,
VEGA ALTA, AND BAYAMON MUNICIPALITIES
AC-814541
General Requirements:
The environmental studies and permits included in the CONTRACT should be performed in accordance with the requirements of the Environmental Studies Office (ESO) and the most recent laws and regulations.
It is required that all drafts submitted are complete, partial documents or with blank sections will not be accepted.
As part of the tasks, the following services shall be included without additional cost to the Authority:
- Attendance to meetings and field trips with the Authority personnel and those that perform specialized studies.
- The CONSULTANT will attend all the coordination meetings and will be available to attend all the coordination meetings, public hearings or any public involvement activity scheduled for the environmental phase of the PROJECT.
- All the illustrations prepared for the PROJECT, including pictures, aerial photographs, maps, plans and others, shall be the most recent, digital and in color, utilizing the best illustration techniques accepted within the good commercial practices. In addition, all the illustrations shall be referenced to the NAD 83 coordinates system and their electronic format shall be compatible with any of the following software packages, which are currently used by the AUTHORITY:
- Eagle Point, 2007 version or the one agreed by both parties
- AutoCAD, 2007 version or the one agreed by both parties
- Arc Map, version 8.0, 8.3
- All the environmental document and studies shall be in Spanish and English language. The studies and permits requests shall be submitted in print form and recorded in compact disk (CD) in PDF format. The quantity of printed copies and CD’s are specified on the task breakdown of each study and permit request. After each document revision, the
CONSULTANT shall submit the same amount of copies, unless the AUTHORITY requires less.
The CONSULTANT shall prepare and submit a monthly progress report including an updated task itinerary.
The following is a breakdown of the services to be provided by the CONSULTANT:
1. **Class of Action (N/A)**
CONSULTANT shall prepare a Class of Action (COA) in accordance with Federal Highway Administration’s regulation requirements for preparing environmental documents and PRHTA’s Environmental Studies Office (ESO).
The Federal COA shall be prepared in English and Spanish, for evaluation and concurrency of the FHWA and the Puerto Rico Highway and Transportation Authority (PHRTA). The Class of Action identifies the level of documentation required for the project. The COA determination is made in consultation with the Lead Federal Agency for Environmental Assessment (EAs), EISs, and CEs as appropriate. The FHWA serve as the Lead Federal Agency.
The draft COA shall be submitted to Puerto Rico Highway and Transportation (PRHTA) for review 30 calendar day after ESO notifies to proceed with the document. COA shall be prepared in accordance with the scope decided upon in the scoping process and to 23 C.F.R 771.115 for projects funding or require a federal action. After the compilation of information and documentation including environmental studies the COA has to be prepared to determinate the most probable environmental document and the related level of documentation required in the NEPA process. There are three classes of action:
1. **Class I**: Environmental Impact Statements (EISs). This COA is applied to actions that significantly affect the environment as defined by CEQ regulations.
2. **Class II**: Categorical Exclusions (CEs). This COA is applied to actions that do not individually or cumulatively have a significant environmental effect.
3. **Class III**: Environmental Assessments (EAs). This environmental COA is assigned to actions in which the significance of the impacts is not clearly established. All actions that are not Class I or Class II are Class III. All actions in this class require the preparation of an EA to determine the appropriate environmental documentation required.
The consultant shall work with the agencies and shall obtain comments as required in Part 1503 of the Regulations for Implementing the Procedural Provisions of the National Environmental Policy Act (NEPA) (40 CFR Parts 1500-1508 as of July 1, 1986). The COA shall fulfill and satisfy to the fullest extent possible the requirements established for final statements in Section 102(2) (C) of NEPA of the in-force regulations or laws.
The process of internal review for the COA will finish after the written approval of the Federal Highway Administration (FHWA). The environmental process of revision of the COA will conclude after written approval by the FHWA.
The services will be performed following the National Environmental Policy Act (NEPA) guidelines for an environmental evaluation and the procedures followed for environmental evaluation and the procedures followed for environmental documents to be prepared for the Federal Highway Administration (FHWA) under 23 CFR parts 771.
The document shall comply with other laws, regulations, requirements and policies that affect NEPA process and influence decision making and shall identify the basic implications of these laws. Multicolor exhibits will be provided where determined necessary for better understanding. Copies of the COA shall be furnished in hard copy and CD's.
The Class of Action (COA) shall include, but it shall not be limited to:
1. Description of the proposed action
2. Purpose and Need of the proposed action, including existing and proposed design
3. Planning Consistency (p.e. funding, costs, consistency with STIP, LR)
4. Description of surrounding area
5. Specific area of environmental concern
6. Impacts of the proposed action on cultural resources, ecological resources, natural resources, etc.
7. Required federal, state, and local action
8. Socioeconomic reasons – describe anticipated relocations, displacements, and right of way requirements. Document efforts taken to meet the requirements of Executive Order 12898 on environmental justice and Executive Order 13166: Improving Access for Persons with Limited English Proficiency.
9. Public involvement (if required)
10. Conclusion
11. Exhibits, including: a location map and a typical section, USGS quad sheet – location, photos with legend description locator, and Section 4(f) evaluation, if any.
CONSULTANT shall submit one (1) printed copy of the COA for ESO review. After acceptance of the document by the ESO, the CONSULTANT shall submit three (3) printed
copies and two (2) digital copies save to a compact disk (CD) of the COA for the evaluation and comments of FHWA. CONSULTANT shall submit five (5) copies of the final document.
Attendance to public and information hearings are included as part of the scope, including the preparation of the project’s presentation (if required).
2. Notice of Intent (NOI) and Coordination Plan (CP) (N/A)
A Notice of Intent (NOI) announces the decision to prepare an Environmental Impact Statement (EIS).
A NOI is published in the Federal Register. The NOI includes:
1. brief summary of the proposed PROJECT
2. summary of alternatives
3. summary of scoping process (including public involvement)
4. list of persons to contact for information
The CONSULTANT shall submit one (1) printed copy of the preliminary NOI to ESO for evaluation. After acceptance of NOI by the ESO, the ESO submit to FHWA for evaluation. After acceptance of NOI by FHWA and ESO the CONSULTANT shall submit to ESO two (2) printed copies and two (2) digital copies saved in a compact disk (CD) of the final NOI.
The Coordination Plan (CP) describes the roles of the lead agency, joint lead agencies, the cooperating and participating agencies and the general public. Also, includes the Need and Purpose, the Scoping Plan, the Public Involvement Plan and the proposed schedule of the project. All the parties shall be identified in the Coordination Plan. The NOI and the CP shall compliance and prepared according with the federal regulations and guidelines. The CONSULTANT shall prepare, in coordination with the AUTHORITY, documentation and presentation required for informative workshops or public participation required for the Coordination Plan, including handouts, boards, public notice, among others. In addition, the COUNSULTANT shall provide for the stenography services and the payment of public notices required by the federal and state regulations. The stenography services shall include the certified English translation of the public participation transcripts. The payment of all the services mentioned above shall be included in the cost of the NOI and CP but presented in the breakdown of the different activities included in this item.
The CONSULTANT shall submit three (3) printed copies of the preliminary CP to ESO for evaluation. After acceptance of CP by the ESO, the ESO submit to FHWA for evaluation and approval. After acceptance of CP by FHWA and ESO the CONSULTANT shall submit to ESO twenty-five (25) printed copies, twenty-five (25) digital copies saved in a compact disk (CD) in PDF format and five (5) printed copies of translations in Spanish of the final CP.
3. Environmental Impact Statement (EIS) (N/A)
For purposes of evaluation, the technical proposal for EIS must follow the standard format described, unless the agency determines that there is compelling reason to do otherwise.
This PROJECT will be developed utilizing the Puerto Rico Department of Transportation and Public Works, AUTHORITY policies and procedures, including the Federal Highway Administration (FHWA) standards and regulations.
The CONSULTANT must prepare:
- the Draft Environmental Impact Statement (DEIS)
- the Final Environmental Impact Statement (FEIS)
- the summary for the Record of Decision (ROD)
The preliminary DEIS must be submitted to Puerto Rico Highway and Transportation (AUTHORITY) review within a total term of _______ calendar days after a ESO notifies to proceed with the document.
The DEIS must be prepared in accordance with the scope decided upon in the scoping process. The CONSULTANT must work with the agencies and must obtain comments as required in Part 1503 of the Regulations for Implementing the Procedural Provisions of the NEPA (40 CFR Parts 1500-1508). The DEIS must fulfill and satisfy to the fullest extent possible the requirements established for final statements in Section 102(2) (C) of the Act. If a DEIS is so inadequate as to preclude meaningful analysis, the CONSULTANT must prepare for circulation a revised draft.
The process of internal review for the preliminary DEIS will finish after the written approval of the Federal Highway Administration (FHWA) and the AUTHORITY. In that first phase, the CONSULTANT will submit for circulation the revised DEIS within a total term of thirty (30) calendar days.
The environmental process of revision of the circulated DEIS will conclude after the same has been approved (signed and dated the cover page) by FHWA and any other agencies.
The second phase of the environmental process is the preparation of the FEIS and must respond to comments as required in Part 1503 of the Regulations for Implementing the Procedural Provisions of the NEPA (40 CFR Parts 1500-1508) and FHWA's Technical Advisory (T6640.8A). The CONSULTANT must discuss at appropriate points in the FEIS any responsible opposing view which was not adequately discussed in the DEIS and must indicate the response to the issues raised.
The environmental process of revision of the circulated FEIS will conclude after the same has been approved (signed and dated the cover page) and available for the prescribed period of time and a Record of Decision (ROD) has been signed.
Section 1319 of the Moving Ahead for Progress in the 21st Century Act (MAP-21), Accelerated Decision-making in Environmental Reviews, provides for the preparation of a Final Environmental Impact Statement (FEIS) by attaching errata sheets to the Draft Environmental Impact Statement (DEIS) if certain conditions are met. The use of errata sheets in lieu of rewriting the DEIS is appropriate when comments received on a DEIS are minor and the FHWA and PRHTA determine that the responses to those comments are limited to factual corrections or explanations of why the comments do not warrant further response.
In addition, requires, to the maximum extent practicable, and unless certain conditions exist, that the FHWA and PRHTA will develop a single document that combines the FEIS and Record of Decision (ROD) unless:
1. The FEIS makes substantial changes to the proposed action that are relevant to environmental or safety concerns; or
2. There are significant new circumstances or information relevant to environmental concerns and that bear on the proposed action or the impacts of the proposed action.
The ROD will present and summarize any mitigation measures that will be incorporated the PROJECT and document any 4(f) approvals in accordance with section 771.135.
The services will be performed following the National Environmental Policy Act (NEPA) guidelines for an environmental evaluation and the procedures for environmental documents to be prepared for the Federal Highway Administration (FHWA), under 23 CFR parts 771 and FHWA's Technical Advisory (T6640.8A). The services include the preparation of the ROD for federal publication.
The services to be rendered shall result in an environmental impact report that will include the whole corridor. This document and all the Appendices including technical studies will be prepared in English for evaluation and concurrency of the FHWA, OGPe, municipalities and all the regulatory state and federal agencies. Translated (to Spanish) copies of the document and summaries of the appendices will also be prepared and handled.
The document must comply with other laws, regulations, requirements and policies that affect the NEPA process and influence decision making and must identify the basic implications of these laws.
The EIS report and appendices will be presented in more than one volume to simplify the handling and revision process. Multicolor exhibits will be provided where determined necessary for better understanding. Copies of the DEIS, FEIS and ROD shall be furnished in hard copy and CD's (PDF format).
The quantities of copies of the EIS report (including appendices) are listed below. Additional copies will be invoiced in a reimbursement of expenses basis. The provision of the document including graphs, maps, photos, or other format required to be ready for display in the internet must be included in the scope of services.
| DOCUMENT | ENGLISH HARD COPY | CD | SPANISH HARD COPY |
|---------------------------------|-------------------|----|------------------|
| Notice of Intent (draft) | 1 | ---| --- |
| Notice of Intent (final) | 2 | 2 | --- |
| Coordination Plan (draft) | 2 | ---| --- |
| Coordination Plan (final) | 25 | 25 | 5 |
| Preliminary DEIS for AUTHORITY review | 5 | ---| --- |
| DEIS for FHWA review | 5 | ---| --- |
| DEIS for circulation | 50 | 15 | 15 |
| Revised DEIS | 25 | 2 | --- |
| Preliminary FEIS | 5 | ---| --- |
| FEIS for circulation | 50 | 10 | 25 |
| Preliminary ROD | 5 | ---| --- |
| Revised ROD | 5 | ---| --- |
| Approved ROD | 5 | 5 | 5 |
For purposes of evaluation, the technical proposal for EIS must follow the standard format described below, unless the agency determines that there is a compelling reason to do otherwise.
The CONSULTANT must be required to use recognized and experienced professionals (fully bilingual) in environmental fields to prepare the specialized studies in order to guarantee a smooth process toward the compliance with the federal requirements. Any changes in the approved professional specialists will be subject to review and approval by the AUTHORITY.
The CONSULTANT must gather information and analyze the environmental aspects that might affect or that will be affected by the construction of the proposed alignments. As required by regulatory agencies, the environmental study will be performed for the route as a unit.
The CONSULTANT must make use of the site information available from the early years of environmental documents of the route location; searches already performed; technical findings of consultation and public involvement meetings; correspondence and comments received; as well as the results of additional and more comprehensive searches necessary to address the environmental issues already identified, or others not identified.
The following standard format for EIS must be followed unless the agency determines that there is a compelling reason to do otherwise:
- Cover sheet
- Summary
- Table of contents
- Need and Purpose of the action
- Alternatives (with engineering graphics, draft engineering sections ((with typical sections roads, intersections, bridges, etc.)), estimates of construction costs, traffic analysis, travel time)
1. No action alternative
2. Mass Transit Alternative
3. Build Alternatives
4. Discussion of discarded alternatives that do not meet the identified need and purpose or are not reasonable or feasible
- Environmental description
1. Socio-economic environment
2. Meteorology and climatology
3. Soils
4. Geology
5. Topography
6. Water resources
7. Ecology
8. Existing land and proposed uses
- Environmental consequences
1. Land use impacts
2. Farmland impacts
3. Social impacts (community cohesion)
4. Relocation impacts (relocation of residents, business, institutions)
5. Economic impacts (users, employments to be generated, property value)
6. Environmental justice
7. Joint development measures which will preserve or enhance an affected community's social, economic, environmental and visual values.
8. Considerations related to pedestrians and bicyclists
9. Air quality impacts-mesoscale concerns and microscale concerns, including measurement of CO background levels, for predictions the CONSULTANT will use Caline 4 model (or any approved version or program by EPA), including discussion of EPA New greenhouse gas guidance
10. Noise impacts, including measurement of background noise levels (for predictions the CONSULTANT must use the TNM model) and noise abatement measures (with type of noise abatement measure, location, large, height and costs.
11. Water quality impacts
12. Well impacts
13. Public and private water intake
14. Zoning and plan of uses of lands
15. Water body modification
16. Floodplain impacts
17. Ecological impacts (includes flora and fauna impacts; threatened and endangered species; wetlands; forest; caverns; sinkholes; natural sites and information of special areas specified by the DNER & USFWS), etc.
18. Wild and scenic rivers
19. Coastal barriers
20. Coastal zone impacts
21. Historic and archaeological preservation
22. Cultural resource assessment Phase IA-IB
23. Soil contamination, existing solids wastes (dangerous and no dangerous), including properties with dangerous and toxic wastes, public or not authorized landfills, underground storage tanks
24. Visual and aesthetic impacts
25. Energy consumption
26. Movement of materials from the earth surface (cuts and fills)
27. Erosion and sedimentation impacts
28. Utilities impacts
29. Accumulative and secondary impacts
30. Construction impacts
31. State Freight impacts
32. Preliminary measure to avoid, minimize or mitigate (to compensate) adverse impacts
33. Relationship or local short term
34. Uses vs. long-term productivity
35. Irreversible and irretrievable commitment of resources
• Environmental Commitments
1. Erosion, sedimentation and water quality
2. Noise pollution
3. Air quality
4. Flora and fauna
5. Water and electricity
6. Solid waste
7. Cultural resources
8. Excavation
9. Safety
10. Endangered species
11. Relocation of residents, businesses and institutions
- List of tables
- List of figures
- List of preparers (personnel of the CONSULTANT and the agency that collaborated in the preparation of the study)
- List of agencies, organizations and persons to whom copies of the statement are sent
- Certification of the works by the specialists and preparers
- Commentaries and coordination with the public, organizations, agencies and municipalities, including at least memorandums with the participation of the public activities)
- Response to comments (Summaries of the commentaries of the public, the corresponding organizations, agencies and municipalities; as well as answers)
- Index
- Appendices (includes all technical studies)
The CONSULTANT must assist the AUTHORITY in determining if Section 4(f) applies to any existing or proposed Section 4(f) properties affected by any of the proposed alternates within the study area. If applies, an additional proposal will be submitted to evaluate and document the direct impacts or indirect proximity effects to the Section 4(f) properties as described in Federal Register, Volume 52, Number 167 of August 28, 1987, CFR 771.135, (49 U.S.C. 303) and as described in this Appendix task 26.
Attendance to public and information hearings are included as part of the scope, including the preparation of the PROJECT’s presentation. The Permit Management Office (PMO) filling fee shall be paid entirely by the CONSULTANT and is not reimbursable.
4. Cultural Resources Assessment Phase IA-IB (N/A)
This study shall include the preparation of a stage I (IA-IB) cultural resource assessment, including literature review, walking survey and test pits in representative quantities of the PROJECT’s area. The CONSULTANT shall be responsible for the necessary excavation equipment and leaving the evaluated ground and/or structures on the same conditions as they were prior to the tests. The assessment shall comply with all applicable regulations as enforced by the Institute of Puerto Rican Culture (IPRC) (e.g., Law 12 of 1988) and the PR State Historic Preservation Office (Section 106 of the National Historic Preservation Act). The professionals shall meet the minimum of Secretary of Interior’s Qualifications Standards, 36 CFR 61.
The CONSULTANT shall submit one (1) printed copy of the preliminary report to ESO for evaluation. After acceptance of the report by the ESO, the CONSULTANT shall submit to ESO ten (5) printed copies and one (1) digital copy saved in a compact disk (CD) of the final report. The photographs on the report shall in color and printed in photographic paper.
5. Cultural Resources Assessment Phase II (N/A)
Phase II investigation is designed to sample the archaeological deposits at sites identified during the Phase I survey and allow a decision to be made as to their significance. Following the completion of Phase II investigations, a consultant should be able to make a Determination of Eligibility (DOE) for all resources evaluated. The Criteria for Evaluation are outlined in the Department of Interior’s regulations, 36 CFR Part 60: “National Register of Historic Places”. Specific references to Criteria for Evaluation are found in 36 CFR 60.4. The professionals shall meet the minimum of Secretary of Interior’s Qualifications Standards, 36 CFR 61.
The CONSULTANT shall be responsible for the necessary excavation equipment and leaving the evaluated ground and/or structures on the same conditions as they were prior to the tests. The assessment shall comply with all applicable regulations as enforced by
the Institute of Puerto Rican Culture (IPRC) (e.g., Law 12 of 1988) and the PR State Historic Preservation Office (Section 106 of the National Historic Preservation Act).
The CONSULTANT shall submit one (1) printed copy of the preliminary report to ESO for evaluation. After acceptance of the report by the ESO, the CONSULTANT shall submit to ESO ten (5) printed copies and one (1) digital copy saved in a compact disk (CD) of the final report. The photographs on the report shall be in color and printed in photographic paper.
6. Memorandum of Agreement (N/A)
If required, a Memorandum of Agreement (MOA) between FHWA and SHPO will be prepared to record the measures that will be implemented to avoid, minimize or mitigate the adverse effects that the proposed project would have on a historic property. This document categorizes how adverse effects to historic properties are to be dealt with and represents a legal commitment for ensuring that the undertaking is carried out in accordance with the MOA’s terms.
THE MOA shall be prepared by an archeologist that shall meet a minimum of requirements in accordance with the Secretary of Interior’s Professional Qualifications Standards, 36 CFR Part 61.
The MOA will be prepared according to regulations at 36 CFR 800 and should follow a standard format. This includes the division of the document into a title, preamble, stipulations, and signatures.
Deliverables:
- First Draft for PRHTA review
- Second draft for FHWA review. This draft includes PRHTA comments.
- A final draft that will be submitted to SHPO for their review and comments.
- Final Document
7. Environmental Assessment (N/A)
CONSULTANT shall prepare an Environmental Assessment in accordance with Federal Highway Administration’s, Puerto Rico Environmental Quality Board (JCA), the Permit Management Office (OGPe) regulation requirements for preparing environmental documents and AUTHORITY’s Environmental Studies Office (ESO) requirements, and the following requirements.
The State EA shall be prepared in Spanish, for evaluation and concurrency of the OGPe, and the Puerto Rico Highway and Transportation Authority (PHRTA).
The Federal EA shall be prepared in English and Spanish, for evaluation and concurrency of the FHWA and OGPe respectively, and the Puerto Rico Highway and Transportation Authority (PHRTA).
The draft EA shall be submitted to Puerto Rico Highway and Transportation (AUTHORITY) for review 90 calendar days after ESO notifies to proceed with the document. EA shall be prepared in accordance with the scope decided upon in the scoping process. ESO and FWHA will review the draft and submit its comments to CONSULTANT. The CONSULTANT will review the draft according to the ESO and FWHA comments.
The process of internal review for the EA will finish after the written approval of the Federal Highway Administration (FWHA) and/or the AUTHORITY. In that first phase the CONSULTANT will submit the revised EA within a total term of thirty (30) calendar days. The environmental process of the revised EA will conclude after written approval by the AUTHORITY and/or FHWA. The EA will then be circulated for comments and public hearing if required.
EAs must be made available to the public through notices of availability in local, state, or regional clearinghouses, newspapers and other means. Depending on the FHWA-approved state public involvement procedures, a public hearing may or may not be required. A 30-day review period is required. After public comments are received and considered, a determination of the significance of the impacts is made, if, after completing the EA, it is evident that there are no significant impacts associated with the project, a finding of no significant impact (FONSI) may be prepared.
The EA shall fulfill and satisfy to the fullest extent possible the requirements established for final statements in Section 102(2) (C) of NEPA of the in-force regulations or laws. If an EA is so inadequate as to preclude meaningful analysis, the CONSULTANT shall prepare for circulation a revised draft.
If it is determined that there will be no significant impacts a FONSI will be prepared to conclude the process and document the decision. A FONSI is issued when environmental analysis and interagency review during the EA process find a project to have no significant impacts on the quality of the environment. The FONSI document is the EA modified to reflect all applicable comments and responses. If it was not done in the EA, the FONSI must include the project sponsor's recommendation or selected alternative. No formal public circulation of the FONSI is required, but the state clearinghouse must be notified of the availability of the FONSI. In addition, FHWA recommends that the public be notified through notices in local newspapers.
The services will be performed following the National Environmental Policy Act (NEPA) guidelines for an environmental evaluation and the procedures followed for environmental evaluation and the procedures followed for environmental documents to be prepared for the Federal Highway Administration (FHWA) under 23 CFR parts 771. The services to be rendered shall result in an environmental impact report that will include the whole PROJECT. This document will be prepared in English for evaluation and concurrency of the FHWA, OGPe, municipalities and all the regulatory state and federal agencies. Translated (to Spanish) copies of the document and summaries of the appendices will also be prepared and handled.
If required the EA, FONSI shall include a benefit cost analysis (BCA), which will compare each alternative, in terms of benefits and cost, with the corresponding no build scenario. This comparison will be carried out by computing for each alternative the B/C ratio which compares the increase of benefits to increase in costs for each option. These ratios will be used in the decision-making process to select the best alternative. For projects of bridge replacement or rehabilitation each alternative will be evaluated in accordance with engineering characteristics, estimated cost, social, cultural and environmental impacts. A benefit costs matrix will be developed to compare and evaluate each alternative in order to choose the preferred.
The document shall comply with other laws, regulations, requirements and policies that affect NEPA process and influence decision making and shall identify the basic implications of these laws. The EA report will be presented in more than one volume to simplify the handling and revision process. Multicolor exhibits will be provided where determined necessary for better understanding. Copies of the EA and potential FONSI shall be furnished in hard copy and CD's.
The Environmental Assessment (EA) shall include, but it shall not be limited to:
- Cover Sheet
- Summary
- Table of contents
- Purpose of the action
- Alternatives (with engineering graphics, draft engineering sections ((with typical sections roads, intersections, bridges, etc.)), estimates of construction costs, traffic analysis, travel time)
1. No action alternative
2. Build Alternatives
- Discussion of discarded alternatives that do not meet the identified need and purpose or are not reasonable or feasible. Environmental description
1. Socio-economic environment
2. Meteorology and climatology
3. Soils
4. Geology
5. Topography
6. Water Resources
7. Ecology
8. Existing land uses
- Environmental consequences
1. Land use impacts
2. Farmland impacts
3. Social Impacts (community cohesion)
4. Relocation impacts (relocation of residents, business, institutions)
5. Economic impacts (users, employments to be generated, property value)
6. Environmental justice
7. Joint development measures which will preserve or enhance an affected community's social economic, environmental and visual values.
8. Considerations related to pedestrians and bicyclists
9. Air quality impacts-mesoscale concerns and microscale concerns, including measurement of CO background levels, for predictions the CONSULTANT will use Caline 4 model (or any approved version or program by EPA)
10. Noise impacts, including measurement of background noise levels (for predictions the CONSULTANT shall use the TNM model) and noise abatement measures (with type of noise abatement measure, sound barrier location, length, height and costs).
11. Water quality impacts
12. Well impacts
13. Public and private water intake
14. Zoning and plan uses of lands
15. Water body modification
16. Floodplain impacts
17. Ecological impacts (include flora and fauna impacts; threatened and endangered species; wetlands; forest; caverns; sinkholes; natural sites and information of special areas specified by the DNER &USFWS)
18. Coastal barriers
19. Coastal zone impacts
20. Historic and archeological preservation
21. Cultural resource assessment Phase IA-IB
22. Soil contamination, existing solids wastes (dangerous and no dangerous), including properties with dangerous and
toxic wastes, public or not authorized landfills,
underground storage tanks
23. Visual and aesthetic impacts
24. Energy consumption
25. Movement of materials from the earth surface (cuts and fills)
26. Erosion and sedimentation impacts
27. Utilities impacts
28. Accumulative and secondary impacts
29. Construction impacts
30. Preliminary measure to avoid, minimize or mitigate (to compensate) adverse impacts
31. Irreversible and irretrievable commitment of resources
• Environmental Commitments
1. Erosion, sedimentation and water quality
2. Noise pollution
3. Air quality
4. Flora and Fauna
5. Water and electricity
6. Solid waste
7. Cultural Resources
8. Excavation
9. Safety
10. Endangered species
11. Relocation of residents, business and institutions
• List of tables
• List of figures
• List of preparers (personnel of the CONSULTANT and the agency that collaborated in the preparation of the study)
• List of agencies, organizations and persons to whom copies of the statement are sent
• Certification of the works by the specialists and preparers
• Commentaries and coordination with the public, organizations, agencies and municipalities, including at least memorandums with the participation of the public activities
• Response to comments (Summaries of the commentaries of the public, the corresponding organizations, agencies and municipalities; as well as answers)
• Index
• Appendices (includes all technical studies)
CONSULTANT shall submit three (3) printed copies of the Environmental Assessment for ESO review. After acceptance of the document by the ESO, the CONSULTANT shall submit twenty (20) printed copies and two (2) digital copies save to a compact disk (CD) of the Environmental Assessment for the evaluation and comments of agencies. The Permit Management Office (PMO) filing fee shall be paid entirely by the CONSULTANT and is not reimbursable.
Attendance to public and information hearings are included as part of the scope, including the preparation of the PROJECT’s presentation.
8. Categorical Exclusion (if required)
CONSULTANT shall prepare a categorical Exclusion in accordance with Federal Highway Administration’s, and Puerto Rico Environmental Quality Board (JCA) and the Permit Management Office (OGPe) regulation requirements for preparing environmental documents, PRHTA’s Environmental Studies Office (ESO) requirements, and the following requirements.
The Federal CE shall be prepared in English and Spanish, for evaluation and concurrency of the FHWA and OGPe respectively, and the Puerto Rico Highway and Transportation Authority (PHRTA).
The draft CE shall be submitted to Puerto Rico Highway and Transportation (PRHTA) review after 30 calendar days ESO notifies to proceed with the document. CE shall be prepared in accordance with the scope decided upon in the scoping process.
The consultant shall work with the agencies and shall obtain comments as required in Part 1503 of the Regulations for Implementing the Procedural Provisions of the National Environmental Policy Act (NEPA) (40 CFR Parts 1500-1508 as of July 1, 1986). The CE shall fulfill and satisfy to the fullest extent possible the requirements established for final statements in Section 102(2) (C) of NEPA of the in force regulations or laws.
The environmental process of the CE will conclude after the final CE is concurred by FHWA. The CE will then be circulated for comment and public hearing if required.
The services will be performed following the National Environmental Policy Act (NEPA) guidelines for an environmental evaluation and the procedures followed for environmental evaluation and the procedures followed for environmental documents to be prepared for the Federal Highway Administration (FHWA) under 23 CFR parts 771. The services to be rendered shall result in an environmental impact report that will include the whole project. This document will be prepared in English for evaluation and concurrency of the FHWA.
Translated (to Spanish) copies of the document and summaries of the appendices will also be prepared and handled.
The document shall comply with other laws, regulations, requirements and policies that affect NEPA process and influence decision making and shall identify the basic implications of these laws. Multicolor exhibits will be provided where determined necessary for better understanding. Copies of the CE shall be furnished in hard copy and CD’s.
The Categorical Exclusion (CE) shall include, but it shall not be limited to:
1. Description of the proposed action
2. Purpose and Need of the proposed action, including existing and proposed design
3. Planning Consistency (p.e. funding, costs, consistency with STIP, LR)
4. Description of surrounding area
5. Specific area of environmental concern
6. Impacts of the proposed action on cultural resources, ecological resources, natural resources, etc.
7. Required federal, state, and local action
8. Socioeconomic reasons – describe anticipated relocations, displacements, and right of way requirements. Document efforts taken to meet the requirements of Executive Order 12898 on environmental justice and Executive Order 13166: Improving Access for Persons with Limited English Proficiency.
9. Public involvement (if required)
10. Conclusion
11. Exhibits, including: a location map and a typical section, USGS quad sheet – location, photos with legend description locator, and Section 4(f) evaluation, if any.
CONSULTANT shall submit one (1) printed copy of the Categorical Exclusion for ESO review. After acceptance of the document by the ESO, the CONSULTANT shall submit five (5) printed copies and two (2) digital copies save to a compact disk (CD) of the Categorical Exclusion for the evaluation and comments of agencies.
9. Habitat Categorization for DNER (N/A)
Habitat Categorization Certification from the DNER is required by the New Puerto Rico Wildlife Law (Law #241) and the Joint Regulation for Construction Works and Land Use
(Section 48, Reglamento Conjunto de Permisos para Obras de Construcción y Usos de Terrenos). The Habitat Categorization describes the type and quality of wildlife habitat present in the Project area and potential impacts that the proposed action may have on these habitats. Mitigation actions need to be developed depending on the impacts to each habitat type (category) in accordance to Law #241 and Regulation #6765 (Reglamento para Regir la Conservación y el Manejo de la Vida Silvestre, las Especies Exóticas y la Caza).
The CONSULTANT shall submit two (2) printed and digital (CD) PDF format copies of the Habitat Categorization, and the final explanatory memorandum to AUTHORITY’s Design Area and ESO.
10. Tree Inventory and Mitigation Planting Plan (N/A)
The tree inventory shall begin with the geotechnical study. A report shall be prepared which includes, but not limited to, an inventory plan that shows the affected tree location, and a table which enumerates and names the height, diameter at breast height, and the condition of each inventoried tree.
The PROJECT’s final tree inventory and mitigation planting plan shall begin at the early stages of the final design. When such document is ready, one (1) printed copy shall be submitted to ESO for evaluation and comments. After the document is accepted by this office, the CONSULTANT shall submit the document the Permit Management Office (PMO or “OGPe”) and obtain their approval. The Permit Management Office (PMO) filing fee shall be paid entirely by the CONSULTANT and is not reimbursable.
The final planting plan with specifications and cost estimates could be submitted to AUTHORITY after such plan is submitted to the Permits Management Office (PMO or “OGPe”), but it shall be modified if necessary, in accordance with the permit issue by PMO, in order to be included on the construction contract documents, before the PROJECT bid.
According to DNER’s interpretation of Section 4.02.b of the Regulation for Cutting, Planting and Forestation of Puerto Rico (Planning Regulation No. 25) as amended, is necessary that a planting professional certifies the tree inventory and the planting plan, as for the licensed engineer, licensed architect or licensed surveyor which prepared the PROJECT plans for which planting mitigation was prepared, to certify such. The planting professional which prepares the documents shall be certified by DNER.
The CONSULTANT shall submit two (2) printed and digital (CD) PDF format copies of the tree inventory and planting plans, and the final explanatory memorandum to AUTHORITY’s Design Area and ESO.
11. Noise Study (N/A)
CONSULTANT shall provide all the equipment and the necessary resources to carry out the noise study according to the requirements of Control of Noise Contamination Regulation of the Environmental Quality Board, federal regulation 23 CFR 772, the most recent Federal Highways Administration (FHWA) for the control and mitigation of noise, and with AUTHORITY’s Environmental Studies Office (ESO) requirements.
All measurements shall be done on representative locations along the PROJECT route, on typical days with average traffic conditions and during peak traffic periods.
The CONSULTANT shall prepare a noise study report containing, but not limited to, the following:
a) Introduction, which includes a PROJECT description, study objectives, pertinent regulation and an explanation of the used metric.
b) A convenient scale plan and aerial photo indicating the receptors’ location (residences, institutions, parks, etc.), and the location of the measurement instruments and the noise receptors.
c) Field photographs showing the location of the measurement instruments and the noise receptors.
d) A discussion of the used criteria for the amount and location of the noise measuring stations.
e) Date, time and weather conditions at the moment when the measurements were taken.
f) Traffic levels in terms of AADT, existing and a 20 year projection. AADT data more than three (3) year old is considered obsolete, so it is required to make manual counts simultaneously with noise measurements.
g) Tabulated results, including the existing Leq (minimum, maximum and average) and a 20 year projection. For the noise level projection the CONSULTANT shall use the TNM model on its most recent version.
h) Conclusions, recommendations and discussion of the potential mitigation measures, indicating the expected noise reduction, and describing the location, materials, dimensions and a cost estimate of such.
i) Evidence of the noise measuring instrument calibration.
j) A printed evidence of the TNM software’s output.
The CONSULTANT shall submit one (1) printed copy of the preliminary Noise Study to ESO for evaluation. After acceptance of the study by the ESO, the CONSULTANT shall
submit to ESO five (5) printed copies and five (5) digital copies saved in a compact disk (CD) PDF format of the final report.
12. Air Quality Emissions Study (N/A)
CONSULTANT shall provide all the equipment and the necessary resources to carry out the air quality study according to the requirements of Control of Atmospheric Contamination Regulation of the Environmental Quality Board, T 6640.8A guide of Federal Highways Administration (FHWA) for the preparation and processing of environmental documents and Section 4 (F), and with AUTHORITY’s Environmental Studies Office (ESO) requirements.
The CONSULTANT shall prepare an air quality study report containing, but not limited to, the following:
a) Introduction, which includes a PROJECT description, study objectives, pertinent regulation and an explanation of the used metric.
b) A convenient scale plan and aerial photo indicating the emission receptors’ location (residences, institutions, parks, etc.).
c) Weather conditions on the area of study.
d) A discussion of the concerned issues on air quality at intermediate (mesoscale) and micro meteorological levels including Ozone, Hydrocarbons (HC) and Nitrous Oxides (NOx), and Carbon Monoxide (CO), respectively.
e) A discussion of attainment and non-attainment Areas, and State Implementation Plan (SIP).
f) Traffic levels in terms of AADT, existing and a 20 year projection.
g) Existing levels of HC, NOx, and CO contaminants, obtained from the EQB or the Environmental Protection Agency (EPA) or thru a mathematical model.
h) A twenty (20) year projection of CO levels. For this, the CONSULTANT will use the MOBILE 6 and CALINE 4 computer models, on their latest version.
i) Tabulated results, including existing contaminants levels and a twenty (20) year projection.
j) Conclusions and recommendations.
k) A printed evidence of the models’ output.
The CONSULTANT shall submit one (1) printed copy of the preliminary Noise Study to ESO for evaluation. After acceptance of the study by the ESO, the CONSULTANT shall
submit to ESO five (5) printed copies and five (5) digital copies saved in a compact disk (CD) PDF format of the final report.
13. Biological Assessment (BA) (N/A)
The following task shall be performed by the CONSULTANT only if required by the Authority by written notice.
The presence of federal listed endangered or threatened species has been previously identified within the PROJECT’s municipalities along the corridor. The US Fish and Wildlife Service (USFWS) administer the Endangered Species Act, which mandates to conserve, recover listed species and conserve their ecosystem. To comply with its mandate the CONSULTANT will prepare a Biological Assessment (BA) along the PROJECT alignment following the USFWS guidelines for such studies. This document shall be prepared in English for evaluation and concurrency of the US Fish and Wildlife Service, and the Department of Natural and Environmental Resources. The BA shall include the following:
1. Perform of reconnaissance of species by means of a field survey of the affected area. A field protocol will set up and a search for endangered species will be conducted.
2. Provide location (coordinates) and abundance of any endangered or threatened species found on the study area. A general description of the flora and fauna of the route shall be also provided.
3. Evaluate impacts of each alternative upon federally protected species. Consider such things as taking of species, elimination of habitat, etc.
4. Discuss the conservation measures for the alternatives that could include, but not limited to, species relocation, habitat acquisition, enhancement, restoration, creation, construction guidelines, etc.
5. Prepare a report that at least contains a location map indicating PROJECT extension; a listing of those federally protected species and critical habitats present within the area affected by the proposed PROJECT; a description of methodologies used to determine the presence of listed species and critical habitats within the affected area; results of comprehensive survey; discussion of any potential gaps and of any difficulties in obtaining data pertinent to the comprehensive survey; a description of the proposed PROJECT with emphasis on those activities occurring within the area inhabited by a listed species; a description of the methods and results of studies made to determine actual and potential impacts of the proposed PROJECT or associated activities on a listed species or critical habitat.
6. In addition to the direct impacts of the PROJECT, the assessment shall include a description of impacts associated with the PROJECT operation; secondary impacts
from activities, such as development, which will be generated by the proposed PROJECT; cumulative effects of the proposal on a listed species or critical habitat. The cumulative effects are the direct and indirect impacts of the proposed PROJECT under consideration as well as other PROJECTs which may be proposed for the general vicinity in the foreseeable future; efforts that will be taken to eliminate, reduce or compensate for any adverse impacts to a listed species or critical habitat; any other relevant information.
The CONSULTANT shall submit one (1) printed copy of the preliminary report to ESO for evaluation. After acceptance of the report by the ESO, the CONSULTANT shall submit to ESO five (5) printed copies and five (5) digital copies saved in a compact disk (CD) of the final report.
14. Environmental Justice Study (N/A)
The study consists of an analysis of the PROJECT’s affected area, emphasizing on the communities directly affected. A comparison shall be done between these communities and other communities or wards on the municipality or municipalities where the PROJECT will be developed, in terms of existing socio-economic conditions. In addition, a socio-economic profile shall be done of the affected communities, how the social cohesion and the basic services of these will be affected with the construction of the proposed PROJECT.
The municipal land use plan and other documents and reports related to the region shall be reviewed. There shall be an economic study of the activities in the region, job mobility, employed and unemployed data, the condition and infrastructure needs of the region and duration and purpose of future construction, plus real estate market at the sectors of the PROJECT. The projected demand and housing need shall be reported. The impact of the PROJECT shall be analyzed in terms of employment, income generation, and investment flow, physical and social impacts (including impacts on social cohesion) among others.
The CONSULTANT shall submit one (1) printed copy of the preliminary report to ESO for evaluation. After acceptance of the report by the ESO, the CONSULTANT shall submit to ESO five (5) printed copies and five (5) digital copies saved in a compact disk (CD) PDF format of the final report.
15. Socio Economic Study (including Community Impact Assessment) (N/A)
The Environmental Assessment must include a socio-economic study (SES) covering all alternatives and including the evaluation of socio-economic impacts based on a socioeconomic profile of the affected communities. The assessment must be based on what is required by the Environmental Guidebook of the Federal Highway (FHWA),
addressing all applicable considerations recommended by it, and including a community impact assessment (CIA) as defined by FHWA. The study must include estimates of structures (residential, commercial and institutions) affected by each alternative.
The CONSULTANT shall submit one (1) printed copy of the preliminary report to ESO for evaluation. After acceptance of the report by the ESO, the CONSULTANT shall submit to ESO five (5) printed copies and five (5) digital copies saved in a compact disk (CD) PDF format of the final report.
16. Public Involvement and PROJECT Coordination (if required)
PROJECT coordination is essential to comply with NEPA and will help to expedite environmental evaluation. The CONSULTANT will assist AUTHORITY in PROJECT coordination of environmental process with government agencies and public.
The consult will assist AUTHORITY and FHWA in the government agencies environmental revision. The CONSULTANT will prepare a draft version of the letters required to initiate coordination/consultation with government agencies. These letters shall be sent by the FHWA or the AUTHORITY. Letters will specifically address the agency applicable regulations.
Attendance to interagency and stakeholder’s meetings, informative workshops and public hearings are limited to minimum of five (5) for simple projects and twenty (20) for EIS and complex projects in accordance to the Coordination Plan for the project as approved by PRHTA and FHWA. The CONSULTANT shall prepare, in coordination with the AUTHORITY, the documentation and presentation required for informative workshops and public hearings, including handouts, boards public notice, among others. In addition, the COUNSULTANT shall provide for the stenography services and the payment of public notices required by the federal and state regulations. The stenography services shall include the certified English translation of the public participation transcripts. The payment of all the services mentioned above shall be included in the cost of this item but presented in the breakdown of the different activities.
All personnel involved including specialist and sub consultants in the preparation of the environmental documents and supporting studies will be available for attendance, if required, to the PROJECTs related activities (field visits, internal meetings, etc.). The CONSULTANT will assist AUTHORITY in PROJECT coordination of environmental process with all state and federal agencies, municipalities, stakeholders and public.
Before the attendance to the public meeting or public involvement activity the CONSULTANT shall submit one (1) printed and one (1) digital copy saved in a compact disk (CD) PDF format copy of the presentation to ESO for evaluation.
17. Environmental Site Assessment (Type ASTM-Phase I) (N/A)
The assessment shall be performed in accordance with the Standard Practice E-1527-05, or the most recent version, of the American Standard for Testing and Materials for the Phase I Environmental Site Assessment Process. This task shall include field investigation and gathering and analysis of available information on the concerned agencies. In addition, the ESO requires that the CONSULTANT provide recommendations based on the findings and conclusions of the assessment.
The CONSULTANT shall submit two (2) printed copies of the Environmental Site Assessment’s Draft for ESO’s review. After the acceptance of the assessment by the ESO, the CONSULTANT shall submit three (3) printed copies and two (2) digital copies saved in compact disk (CD) PDF format, of the final report to ESO.
18. Environmental Site Assessment (Type ASTM-Phase II) (N/A)
The assessment shall be performed in accordance with the American Standard for Testing and Materials Standard Practice E-1903-97, or the most recent version. The objective of this assessment is to further study the Recognized Environmental Conditions in the Environmental Assessment Phase I, in order to determine the cause and extent of the contamination.
This task shall include, but not limited to, the following:
a) Thorough field research, collection and analysis of information available from agencies with jurisdiction.
b) Field sampling for laboratory analysis as indicated in the table below:
| Location | Sampling points | Depth of samples per sampling point | Parameters |
|----------|-----------------|------------------------------------|------------|
| N/A | N/A | N/A | N/A |
c) Quality control blanks shall be prepared and analyzed.
d) All control samples and blanks shall be obtained and analyzed following the recommendations and/or relevant EPA protocols.
e) A report shall be prepared with findings, conclusions and recommendations.
The CONSULTANT shall submit one (1) printed copy of the Environmental Site Assessment’s Draft for ESO’s review. After the acceptance of the assessment by the
ESO, the CONSULTANT shall submit four (4) printed copies and two (2) digital copies saved in compact disk (CD) PDF format, of the final report to ESO.
Note: This study shall be done only if the presence of hazardous materials is determined on the area of study, after the Environmental Assessment Phase I. No study shall be done until an order to proceed is received from ESO.
19. Survey to Determine the Presence of Asbestos and Lead (if required)
Survey to determine the presence of asbestos and lead base paint shall include, but not limited to, the following:
a) inspection, sampling and analysis in the structures following the protocols and methodology approved by the EQB, the Environmental Protection Agency (EPA) and the Housing and Urban Development Department (HUD), if applicable.
b) The evaluation of surfaces to determine the content of asbestos shall be performed utilizing the technique of polarized light microscopy.
c) The evaluation of surfaces to determine the content of lead in paint shall be performed utilizing one of the following two (2) methods:
i) Method of direct lecture with XRF equipment (x-ray fluorescence). In the case that the results lie within the “inconclusive range” of the instrument, the results shall be confirmed utilizing the method of subparagraph “b” that follows.
ii) Method of removal of paint chips for analysis in laboratory by the method of Atomic Absorption Spectrophotometry (AAS).
d) The number of tests required shall be determined in accordance with the criteria, protocols and regulations established by EQB, EPA and HUD (if applicable). However, the ESO recommends at the very least the following:
i) For the Lead Survey perform one (1) representative test in all suspicious interior and exterior structural members such as walls, floors, aisles, ceilings, stairs, doors, windows, sills, and/or combinations of them, air conditioning insulation, bathroom tiles, and others, on each housing unit, building or structure.
ii) For the Asbestos Survey perform representative samples, in all suspicious structural elements such as those described on the preceding subparagraph “a”, in accordance with the following criteria:
- For homogenous surfaces of less than one thousand (1,000) square feet take three (3) samples.
For homogenous surfaces of more than one thousand (1,000) square feet but less than five thousand (5,000) square feet take five (5) samples.
For homogenous surfaces of more than five thousand (5,000) square feet take seven (7) samples.
In the case of small suspicious surfaces such as windows, sills, electrical equipment insulation, sink gaskets and sealants, kitchen equipment and others similar take one (1) sample.
e) Perform field trips with the AUTHORITY personnel or contractor for coordination of the surveys and interpretation of the reports.
f) The CONSULTANT shall immediately inform the ESO in writing of any access limitations on the structures to be surveyed.
g) Prepare a report with the findings and conclusions including the following:
i) The surveyor shall utilize his (her) professional judgment to determine if there are suspicious materials containing asbestos or lead base paint on the structures to be surveyed. If he (her) determines that there are not suspicious materials and opt to not take samples for confirmation, he (her) shall submit a signed certification describing the criteria utilized to support this conclusion.
ii) Photographs and sketches identifying the areas where the tests and samples give positive results.
iii) Photographs of the exterior facade of all the structures surveyed.
iv) Drawing showing the location of the structures surveyed and identifying the structures that have positive results.
v) Measure of the surface areas that have positive results.
vi) Estimate of the quantities in weight and/or volume of the material to be discarded containing asbestos and/or lead.
vii) Copy of the certifications issue by the Environmental Quality Board, of the personnel who took the samples and the surveyor.
viii) The report shall be signed by the surveyor.
The CONSULTANT shall submit one (1) printed copy of the Survey to Determine the Presence of Asbestos and Lead for the ESO review. After acceptance of the document by the ESO, the CONSULTANT shall submit five (5) printed copies and two (2) digital
copies save to a compact disk (CD) PDF format of the Final Survey to Determine the Presence of Asbestos and Lead.
Note: This study shall be done only if there is a suspicion of lead or asbestos presence in any structure to be demolished or any thermoplastic paint is to be removed on the PROJECT. No study shall be done until an order to proceed is received from ESO.
20. USACE’s Jurisdictional Determination (JD) (N/A)
CONSULTANT shall prepare a Jurisdictional Determination Study (JD) (water and wetlands) of United States Army Corps of Engineers (USACE) in accordance with the dispositions contained on the following documents: Technical Report Y-87-1- Corps of Engineers Delineation Manual, dated January 1987, October 12 2001 Public Notice, and USACE’s Approved Jurisdictional Determination Form, Law 173 of 12th of August of 1988, as amended and Law 198 of 8th of August of 1979, as amended.
The report shall include the Routine Wetland Determination Data Form field sheets. Jurisdictional areas shall be surveyed and represented on recent aerial photographs and certified survey plans referenced to the coordinate system defined at Survey Services. The aerial photograph shall be at an approximate scale of 1:2,000 or less; the plans shall be at a scale of 1:500. With this information, the CONSULTANT will prepare estimates of impacts (in acres) of jurisdictional areas affected by the alternatives under study. All studied alternatives shall be presented on the report.
The CONSULTANT shall submit one (1) printed copy of the JD study for ESO’s review. After acceptance of the application by the ESO, the CONSULTANT shall submit the necessary copies that will accompany the Joint Permit Application (DNER). In addition, the CONSULTANT shall submit three (3) printed copies and two (2) digital copies saved on compact disk (CD) PDF format of the final study.
21. Joint Permit Application (USACE) (N/A)
The joint permit application (JPA) shall be prepared and submitted following the established procedures on the PR JPA Form 1999 Commonwealth Joint Permit Application Form, for Water Resource Alterations in Waters, including Wetlands, of Puerto Rico.
In case of an Individual Permit application, is understood that in order to prepare and obtain such permit, if required by USACE, an analysis of alternatives to avoid, minimize and/or compensate adverse impacts to wetlands shall be included. If a compensation is required, the application shall include a mitigation concept plan; location, size, species and proportion of the vegetation, and a subjective explanation of the necessary hydrology.
and topography. Services shall also include the necessary coordination with the stakeholder's agencies, comment implementations and the processing of the EQB’s Water Quality Certificate, if required.
The CONSULTANT shall submit two (2) printed copies of the permit’s application for ESO’s review. After acceptance of the application by the ESO, the CONSULTANT shall submit an original and fourteen (14) copies of the permit application to DNER. The CONSULTANT shall submit one (1) printed copy and one (1) digital copy saved on compact disc (CD) PDF format of the permit application and the DNER joint permit filling’s evidence.
Note: This study shall be done in case the JD determines that the PROJECT’s effect will impact a USACE’s jurisdictional area.
22. Underground Injection Permit (N/A)
The Underground Injection Construction Permit shall begin when the PROJECT design is on the initial stage of the final design. When the document is prepared, one (1) copy shall be submitted for EOS evaluation and comments. After acceptance of the document by ESO, the CONSULTANT shall submit it to the Permit Management Office (PMO), and submit one (1) copy to the Design Area and one (1) to ESO. The CONSULTANT shall comply with all the dispositions of the Underground Injection Control Regulation (UICR) of the EQB.
The construction permit shall include:
a) Activities proposed by the applicant requiring the permit
b) Location where the facility will be built
c) Map showing in a quarter (1/4) mile of the underground injection source (UIS), the number and location of injection facilities, extraction wells, abandoned wells, surface water bodies, streams, buildings and property limits.
d) System construction drawings, specifying dimensions and construction materials
e) Nature and volume of the fluid to be injected
f) It shall also comply with any requirements imposed by the EQB UICR for Class VI, given greater emphasis, but not limited to: Part III, Rules 302, 303, 304.
The operation permit shall include:
a) Permit’s form
b) Contractor's certification stating the construction was carried out according to the approved building permit.
c) Payments for administrative expenses of the permit
d) Payment of public notices required by the permit
e) Retention pond’s operation and management plan
f) Comply with any requirements imposed by the EQB UICR for Class VI, given greater emphasis, but not limited to: Part III, Rules 302, 303, 304.
The CONSULTANT shall submit a draft of the permit's application for ESO's review. You will be given one (1) week to review the application after the ESO issued its comments on the draft. The operation permit shall begin once PMO approves the construction permit. You will be given one (1) week to review the application after the ESO issued its comments on the draft. One printed copy and one (1) digital copy saved on compact disc (CD) PDF format of the permit application and the PMO filling’s evidence shall be submitted to ESO.
23. Ecological Study(ES) (Flora & Fauna) (N/A)
CONSULTANT shall prepare a Ecological Study (ES) in accordance with Fish and Wildlife Service (FWS) and Department of Natural and Environmental Resources (DNER) regulation requirements, and AUTHORITY’s Environmental Studies Office (ESO) requirements.
The ecological study shall include, but not limited to:
a) Detail description of the flora and fauna of the area under study.
b) Common and scientific names of the identified species.
c) Identify and name the existing vegetation composition including the percent of dense vegetation, shrub vegetation, etc.
d) System functionality
e) Description of the used methodology for the investigation or survey of the described species.
f) In case of rare, threaten or endangered are identified or found, as defined by federal or state regulation, related information about distribution, relative abundance, food chains, habitats and relationships among existing species.
The CONSULTANT shall submit one (1) printed copy of the preliminary report to ESO for evaluation. After acceptance of the report by the ESO, the CONSULTANT shall submit to
ESO five (5) printed copies and five (5) digital copies saved in a compact disk (CD) of the final report.
24. Puerto Rico Planning Board (PRPB) and Autonomous Municipality (AM) adoption process (N/A)
The CONSULTANT must prepare the documents and provide the Office of Strategic Planning of the AUTHORITY the necessary support (such as graphics, presentations, and participation in meetings) to request to the Puerto Rico Planning Board (PRPB) the inclusion of the recommended alternative in the corresponding transportation plan. For estimates purposes, it has been assumed that another one (1) public hearings per municipality will be held as part of the PRPB or AM adoption process. Also assume that the graphics materials will be the same as prepared for the EIS public hearing process. If necessary, all personnel involved in the EIS preparation will be available for attendance the PRPB adoption process. The process will finish after the written approval of the Puerto Rico Planning Board (PRPB) and autonomous municipalities (AM).
25. Benefit/Cost Analysis (N/A)
The EA and EIS shall include a benefit cost analysis (BCA), which will compare each alternative, in terms of benefits and costs, with the corresponding no build scenario. This comparison will be carried out by computing for each alternative the B/C ratio which compares the increase of benefits to the increase in costs for each option. These ratios will be used in the decision-making process to select the best alternative.
The CONSULTANT shall submit one (1) printed copy of the preliminary report to ESO for evaluation. After acceptance of the report by the ESO, the CONSULTANT shall submit to ESO five (5) printed copies and five (5) digital copies saved in a compact disk (CD) of the B/C analysis.
26. 4(f) Programmatic Determination (N/A)
Projects that will result in impacts to local, State and Federally owned / designated parkland, recreational lands, wildlife and waterfowl refuges, wild & scenic waterways and National Register of Historic Places eligible or designated sites including structures are subject to 23 CFR 771.135 - Section 4(f) (49 U.S.C. 303). This task includes verifying Section 4(f) resources within the project limits, consulting with the jurisdictional owners, and assessing impacts, both temporary and permanent, to the Section 4(f) properties. The culmination of this investigation is preparing a 4(f) evaluation in accordance with federal regulations and guideline for review and approval of the PRHTA Environmental Studies Office and FHWA.
The 4f evaluation should include the following topics:
- Introduction
- Description and Need for the Project
- Description of Section (4f) resource
- Findings
- Measures to Minimize Harm
- Coordination
- Public Involvement
- Conclusion
The CONSULTANT shall submit one (1) printed copy of the first Draft to ESO for review. The CONSULTANT shall submit one (1) printed copy of the second draft which includes ESO comments to FHWA for review. Once the document has FHWA comments, The CONSULTANT shall submit one (1) printed copy of the document for the signature of FHWA.
27. Feasibility Study (N/A)
Feasibility studies (FS) are included in the definition of engineering and design related services in the Code of Federal Regulations (23 CFR Part 172). The FS includes the following parts:
- The degree to which given alternative modes, management strategy, design or location is economically justified.
- The degree to which such an alternative is considered preferable from an environmental or social perspective.
- The degree to which eventual construction and operation of such an alternative can be financed and managed.
The CONSULTANT shall propose a methodology or study approach that is cost effective and acceptable to the Authority and FHWA for a transportation planning study. Each factor should be evaluated in a qualitative/quantitative manner that is consistent with best practices. This will allow the results of the study to be replicable and useful for a possible NEPA process. The CONSULTANT shall perform at least one (1) field visit in coordination with the Authority, in order to validate information previously obtained. If the project scope requires additional visits to cover the study area, the CONSULTANT shall include a contingency for such, if the need arises. The FS report should include, at a minimum, the results of the following tasks:
A. **Planning** – Feasibility of each alternative in regard to the Puerto Rico Planning Board (PRPB) and Autonomous Municipality (AM) adoption process. The Long-Range Transportation Plan should be used as a source to establish modal travel demand in the long term. The long-term travel demand shall be supplemented with short-term traffic projection and mode shares based on existing traffic patterns, proposed development, and economic growth within the study area.
Planning and land use considerations in regard to each of the alternatives, including, but not limited to the evaluation of the required infrastructure for walkers and cyclists considering the following:
1. Sidewalks comfortable, safe, and shaded; lightning, surface treatment, crosswalks, ramps, and other facilities designed to improve mobility, access, and safety conditions.
2. Bike lane signage (separated or marked), shared-use path off-road/in-road.
3. Complete Streets & Livable Initiative (by US DOT): the CONSULTANT shall evaluate the possibility to incorporate complete streets concepts to the corridor with the goal to create a comprehensive, integrated, connected network for all modes.
B. **Purpose and need analysis for each alternative**, including the degree to which each alternative achieves the goals and objectives set forth in this RFP.
C. **Traffic Study** - For the traffic study forecasting alternatives and scenario analyses; the following conditions (horizon) shall apply for feasibility studies:
1. Year of estimated opening of proposed roadway (or roadway improvements)
2. Twenty (20) years after the year of estimated opening to traffic.
D. **Public transportation analysis**, including transit and paratransit that is mainly focused on the study area, considering the following major tasks:
1. Influence area overview: Provide a general overview of the area, its development patterns, major traffic generators, population, and demographic profiles (such as elderly, disabled, income levels) and auto ownership rates.
2. Area transit services: Provide an inventory of the existing transit (and paratransit) alternatives operating in the study area (including ridership if currently documented by others). The description of such services shall be included as well as its effects on the highway operation and the operational (highway-related) constraints on the transit service.
3. Evaluate the potential expansion of transit service beyond the current scope of existing transit service providers.
4. Evaluate potential alternatives that would establish or improve a transit route or service, provide priority to transit services in the project corridor as well as increase the safety of its users.
5. Perform an origin-destination study, following guidelines acceptable to FHWA and the Authority.
E. Roadway Geometry/Operation Evaluation - Evaluate the existing traffic operations of the corridor within the project limits and determine the most effective manner to improve traffic operations and reduce existing conflicts at major at-grade intersections and interchanges. This work includes but is not limited to the following:
1. Traffic Study report to evaluate LOS and determine possible geometry improvements.
2. General recommendations for future redesign and rearrangement of existing traffic operations and movements through the corridor.
3. General recommendations for evaluation of auxiliary lanes frontage roads, right/left turn lanes, acceleration/deceleration lanes.
4. General evaluation of right/left shoulder condition and proposed improvements.
5. Data gathering and the subsequent report shall be performed following the PRHTA Guides for the preparation of Operational and Traffic Studies, as applicable.
F. Comparative analysis of each alternative – Analyze the potential adverse effect, which includes the critical issues that apply to each alternative and the possible mitigation strategies, on following:
1. Environmental analysis as defined in item 28 below
2. Traffic analysis
3. ITS Implementation Analysis, if applicable, according to Appendix A. 1 General Design Considerations and Conditions.
4. Public Transportation analysis
5. Roadway assessment, which includes how each alternative will provide for the needs of all road users, drainage, geometry, traffic control, traffic operations and safety evaluation.
6. Geotechnical assessment
7. Right-of-way assessment
8. Utility assessment,
9. Stakeholder outreach and public participation
10. Alternatives evaluation, which may include tables with consistent valorization and prudently weighted for a final quantitative comparison. This evaluation will take into consideration the results of all the studies performed in this project.
11. Estimates of cost at current prices, subdivided into roadworks, land cost, bridgeworks and any other particular items.
12. Economic Assessment summarizing the benefits, savings in travel time, reduction in accidents, reduction in delays at major junctions; other special items may be added in particular cases.
G. Study Report (Findings) – the alternatives evaluation shall be summarized, examined and brought to conclusions about the project in general. The findings should include a summary of critical issues (red flags) and recommendations with respect to the consideration of alternatives and project feasibility; with special emphasis on project planning, NEPA process, design, constructability, financing and operation. The works includes but not limited to:
1. Description of all the data collected for all the studies realized, the analysis and the results of this studies.
2. Analysis of the studies and included comprehensives recommendations.
The FS will thoroughly analyze the results of the studies and analyses mentioned above. A simple compilation of analysis and study results is NOT acceptable. The FS shall further refine the conclusions of the studies indicated above in the context of the overall project alternatives. The FS should include the alternative of No Action (No Build). First, the FS shall analyze and take into consideration additional facts that may arise as a result of the public involvement and project coordination. Furthermore, the FS includes additional factors to be studied or analyzed that were not the focus of the previous studies or analyses.
The Feasibility Study and its related analyses or tasks shall follow the guidance set forth in 23 U.S.C. 168, which discusses the Planning and Environmental Linkages. The CONSULTANT shall propose the steps in order to follow this guidance for each of the foreseeable planning decisions (e.g. purpose and need) and the planning analyses (e.g. travel demand).
The CONSULTANT shall submit three (3) printed copies and three (3) digital copies saved in a compact disk (CD) in PDF format of the preliminary FS to ESO for evaluation. After acceptance of the FS by the ESO, the CONSULTANT shall submit to ESO ten (10) printed copies, ten (10) digital copies saved in a compact disk (CD) in PDF format.
28. Environmental Analysis (N/A)
An Environmental Analysis shall prepare as part of the Feasibility Corridor Study that evaluates the existing condition and determines generalized the environmental impacts and early constrains for project development and implementation/construction of the proposed solutions. The environmental issues shall take into considerations are the following.
a) **Cultural Resources**- The evaluation shall identify historic properties and cultural resources within the project area with the objective to help in the determination of sensitive’s area with the potential of presence of prehistoric and historic archeological resources. As reference, the consultant should revise the existing cultural resources (archeological and architectural) information available for the project area in the ICP, SHPO and/or in the municipalities. Review the general history of the municipalities. Review the bridge inventories in the project area. If necessary, conduct a surface reconnaissance.
b) **Ecological Evaluation**- The evaluation shall identify rare, threaten or endangered species, critical habitat, natural reserves, flora and fauna (including any dense vegetation or any important tree) within the project area, as defined by federal and state regulation, in compliance with the Fish and Wildlife Service and the Puerto Rico Department of natural and Environmental Resources requirements.
c) **Environmental Justice, Socio-Economic and Land Uses Evaluation**- The evaluation shall analyze the project area, emphasizing on the communities, business, residences, institutions (hospitals, schools, churches) directly affected, and how the community cohesion and the basic services of these will be affected. Also, the evaluation shall meet Executive Order 12898 on Environmental Justice. The purpose of the analysis is to avoid a disproportionate impact on minority ethnic groups and/or low-income population. As a minimum, US Census shall be used.
The municipal land use plan and other documents and reports related to the region shall be reviewed for the aforementioned evaluation, including the identification of agricultural lands, special zoning areas, etc.
d) **Potential Contaminated Site** - The evaluation shall identify potentially contaminated areas or any suspicion of hazardous materials presence within the corridor in accordance with state and federal regulations. Also, identify structures that are suspect of possible contamination such as: gasoline stations, mechanical workshops, pharmaceutical facilities, among others.
e) **USACE Jurisdictional Areas, including Wetlands** - The evaluation shall identify USACE Jurisdictional Areas including wetlands within the project area, in accordance with the dispositions contained on the following documents: Technical Report Y-87-1-Cordps of Engineers Delineation Manual, dated January 1987, October 12, 2001 Public Notice, Regional Supplement to the Corps of Engineers Wetlands Delineation Manual: Caribbean Island Regional (version 2.0, May 2011) or recent and USACE’s approved Jurisdictional Determination Form.
f) **Noise** - The evaluation shall identify possible noise receptors that could be affected within the project area. The evaluation shall be done according to the requirements of Control Noise Contamination Regulation of the Environmental Quality Board, federal regulation 23 CFR 772, the most recent FHWA for control and mitigation of noise and with ESO requirements.
g) **Air Quality** - The evaluation shall identify the possible quality receptors that could be affected within the project area. The evaluation shall be done according to the requirements of Control of Atmospheric Contamination Regulation of the Environmental Quality Board, T6640.8A of FHWA and ESO requirements.
h) **Flora and Fauna** - The evaluation shall identify the existing flora and fauna could be affected within the project area. The evaluation shall be done according with Fish and Wildlife Service, the Puerto Rico Department of Natural and Environmental Resources and ESO requirements. The flora and fauna assessment shall include the identification of the species with its common and scientific name, as well as its location and relative abundance in the project area.
The Environmental Analysis should identify, evaluate and consider sensitive sites along the project study area. Also, it should illustrate the environmental sensitive sites using pictures, aerial photographs, maps, plans and others as deem appropriate.
The CONSULTANT shall submit one (1) printed copy of the preliminary Environmental Analysis report to ESO for evaluation. After acceptance of the report by the ESO, the CONSULTANT shall submit to ESO five (5) printed copies, one (1) digital copy saved in a compact disk (CD) in PDF format.
The AUTHORITY agrees to pay the CONSULTANT as compensation for the described professional services in the APPENDIX "A" (SCOPE OF PROFESSIONAL DESIGN SERVICES), APPENDIX "B" (SCOPE OF PROFESSIONAL ENVIRONMENTAL SERVICES), and to be provided under this CONTRACT, up to a maximum authorized lump sum amount for services rendered, stipulated as follows:
SECTION A - SUMMARY TABLE FOR PROFESSIONAL SERVICES
| Item | Service | Description | Maximum Authorized Amount |
|------|-------------|------------------------------------------------------------------------------|---------------------------|
| 1 | Study | Field Inspection and Data Collection, APPENDIX A, ARTICLE II, SECTION A, ITEM 1, PHASE 1 (50% of the roads) | $24,500.00 |
| 2 | Study | Field Inspection and Data Collection, APPENDIX A, ARTICLE II, SECTION A, ITEM 1, PHASE 2 (50% of the roads) | $24,500.00 |
| 3 | Study | Survey to Determine the Presence of Asbestos and Lead, APPENDIX B (19) (if required) | $40,000.00 |
| 4 | Design Plans| Preliminary Plans, APPENDIX A, ARTICLE II, SECTION C, ITEM 1 | $66,071.00 |
| 5 | Design Plans| Final Plans, APPENDIX A, ARTICLE II, SECTION D, ITEMS 1 to 8 | $66,071.00 |
Total Design = $221,142.00
SECTION B - Design Fees
The amount agreed for the Preliminary and Final Design phases are as follows:
Table 5.2 FIELD INSPECTION AND DATA COLLECTION (50% OF ROADS)
| | Man Hours | Rate/hr | Total |
|--------------------------|-----------|---------|-----------|
| I. Design (Salaries) | | | |
| A. PROJECT Manager | 8 | $100.00 | $800.00 |
| B. Design Engineer | 117 | $90.00 | $10,530.00|
| C. Draftsman | 96 | $45.00 | $4,320.00 |
| Subtotal = | | | $15,650.00|
| II. Misc. Direct Costs | | | $1,245.66 |
| III. Overhead (%) | 33.59% | | $5,256.84 |
| IV. Fixed Fee (% A thru C)| 15% | | $2,347.50 |
| TOTAL PHASE COST = | | | $24,500.00|
Table 5.2 FIELD INSPECTION AND DATA COLLECTION (50% OF ROADS)
| | Man Hours | Rate/hr | Total |
|--------------------------|-----------|---------|-----------|
| I. Design (Salaries) | | | |
| A. PROJECT Manager | 8 | $100.00 | $800.00 |
| B. Design Engineer | 117 | $90.00 | $10,530.00|
| C. Draftsman | 96 | $45.00 | $4,320.00 |
| Subtotal = | | | $15,650.00|
| II. Misc. Direct Costs | | | $1,245.66 |
| III. Overhead (%) | 33.59% | | $5,256.84 |
| IV. Fixed Fee (% A thru C)| 15% | | $2,347.50 |
| TOTAL PHASE COST = | | | $24,500.00|
Table 5.2 DETAILED PRELIMINARY DESIGN FEES
| | Man Hours | Rate/hr | Total |
|----------------------|-----------|----------|-------------|
| V. Design (Salaries) | | | |
| D. PROJECT Manager | 42 | $100.00 | $4,200.00 |
| E. Engineer | 260 | $90.00 | $23,400.00 |
| F. Draftsman | 325 | $45.00 | $14,625.00 |
| **Subtotal =** | | | **$42,225.00** |
| VI. Misc. Direct Costs | | | **$3,328.87** |
| VII. Overhead (%) | 33.59% | | **$14,183.38** |
| VIII. Fixed Fee (% of A+B+C) | 15% | | **$6,333.75** |
| **TOTAL PHASE COST =** | | | **$66,071.00** |
**Table 5.3 DETAILED FINAL DESIGN FEES**
| | Man Hours | Rate/hr | Total |
|----------------------|-----------|----------|-------------|
| V. Design (Salaries) | | | |
| A. PROJECT Manager | 42 | $100.00 | $4,200.00 |
| B. Engineer | 260 | $90.00 | $23,400.00 |
| C. Draftsman | 325 | $45.00 | $14,625.00 |
| **Subtotal =** | | | **$42,225.00** |
| VI. Misc. Direct Costs | | | **$3,328.87** |
| VII. Overhead (%) | 33.59% | | **$14,183.38** |
| VIII. Fixed Fee (% of A+B+C) | 15% | | **$6,333.75** |
| **TOTAL PHASE COST =** | | | **$66,071.00** |
The above proposal complies with 23 CFR Part 172.11 (Allowable costs and oversight) and with the Federal Acquisition Regulation (FAR) of 48 CFR Part 31. All indirect costs are in accordance with Federal costs principles (23 CFR Part 172.9).
**SECTION C - PAYMENT BY DELIVERABLES**
The CONSULTANT shall be paid by the deliverables established on APPENDIX C, SECTION A (SUMMARY TABLE OF CONSULTING SERVICES). The CONSULTANT should submit the Preliminary and Final Design deliverables according with the latest approved AUTHORITY Design Directive No. 107 and the Progress Schedule (APPENDIX A, ARTICLE I, SECTION B) approved by the AUTHORITY.
The AUTHORITY shall make payments to the CONSULTANT in an amount equal to the ninety percent (90%) of the fee corresponding to each phase of the PROJECT.
The AUTHORITY shall retain the remaining (10%) of the CONSULTANT’S fee for each phase of the PROJECT and it shall be released at the preliminary phase when the Final Plans of
this stage is approved, and on the final phase when the Final Cost Estimate and Construction Plans are approved, subjected to an audit made by the AUTHORITY for each corresponding phase. The before mentioned audit shall be carried out within the following ninety (90) calendar days after the deliverables are completed.
The same way, the AUTHORITY shall release the ten percent (10%) corresponding to additional services provided by the CONSULTANT, works or special studies provided for the PROJECT under this CONTRACT, including the land acquisition plans for right of way, once these services are approved by the AUTHORITY.
SECTION D - EXCEPTIONS
The AUTHORITY shall pay the CONSULTANT for its professional services, by the deliverables established on Appendix C which are satisfactorily completed and properly certified, with the exception of the following:
| Service | Payment Distribution |
|----------------------------------------------|--------------------------------------------------------------------------------------|
| Route Location and Reconnaissance Studies | 60% shall be paid when the document is submitted to PRHTA for review. Up to 80% shall be paid when the document is accepted by PRHTA and is submitted to the FHWA. The remaining 20% amount shall be paid when the document is accepted by FHWA. |
| Traffic Study | Up to 80% shall be paid when the document is submitted to PRHTA. The remaining 20% amount shall be paid when the document is accepted by PRHTA. |
| Non-destructive Testing (NDT) | Up to 80% shall be paid when the document is submitted to PRHTA. The remaining 20% amount shall be paid when the document is accepted by PRHTA. |
| Interstate Access Request Report | 60% shall be paid when the document is submitted to PRHTA for review. Up to 80% shall be paid when the document is accepted by PRHTA and is submitted to the FHWA. The remaining 20% amount shall be paid when the document is accepted by FHWA. |
| Hydrologic-Hydraulic and Scouring Studies | The Hydrologic-Hydraulic and Scouring Study shall be paid up to an 80% when presented to the pertinent regulatory agency (agencies) and the rest 20% shall be paid once this (these) agency (agencies) issue its (their) endorsement (s). |
| Class of Action | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the report is accepted by ESO and is submitted to the FHWA. The remaining 20% amount shall be paid when the study is accepted by FHWA. |
|-----------------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Notice of Intent and Coordination Plan | 80% shall be paid when the NOI approved by ESO and FHWA and the remaining 20% when published in the Federal Register.
80% shall be paid when the CP approved by ESO and FHWA. The remaining 20% when shall be paid when approved ROD. |
| Environmental Impact Statement | 40% shall be paid when the Draft EIS is submitted for ESO’s review.
Up to 70% shall be paid when the Draft EIS is circulated among concerned agencies.
Up to 90% shall be paid when public comments are incorporated and the Final EIS is submitted to circulate among concerned agencies and public.
The remaining 10% when FHWA and PMO understand that the Environmental Public Policy Law has been fulfilled. |
| Cultural Resources Assessment Phase IIA-IB | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the report is accepted by ESO and is submitted to the Puerto Rican Cultural Institute (ICP) and State Historic Preservation Office (SHPO), if necessary. The remaining amount shall be paid when the study is accepted by ICP and SHPO. |
| Cultural Resources Assessment Phase II | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the report is accepted by ESO and is submitted to the Puerto Rican Cultural Institute (ICP) and State Historic Preservation Office (SHPO), if necessary. The remaining amount shall be paid when the study is accepted by ICP and SHPO. |
| Memorandum of Agreement | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the report is accepted by ESO and is submitted to FHWA and SHPO. The remaining amount shall be paid when the study is accepted by FWS. |
| Habitat Categorization for DNER | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the report is accepted by ESO and is submitted with the technical assistance request application to DNER. The remaining amount shall be paid when the study is accepted by DNER. |
| **Tree Inventory and Mitigation Planting Plan** | 67% shall be paid when the inventory and planting plan is submitted with the technical assistance request application to DNER, up 13% when DNER issue the permit and the remaining 20% when the Authority accept the final planting plan, with specifications and cost estimates, and with the corrections established on the conditions issued on the DRNE permit, and when the final planting plan makes part of the PROJECT design construction plans. |
| **Environmental Assessment** | 40% shall be paid when the environmental document’s draft is submitted for ESO’s review. Up to 70% shall be paid when the environmental document is circulated among concerned agencies. Up to 90% shall be paid when public comments are incorporated and the final EA is submitted. The remaining 10% when FHWA and PMO understand that the Environmental Public Policy Law has been fulfilled. |
| **Categorical Exclusion** | 40% shall be paid when the environmental document’s draft is submitted for ESO’s review. Up to 70% shall be paid when the environmental document is circulated among concerned agencies. Up to 90% shall be paid when public comments are incorporated and the final EA is submitted. The remaining 10% when FHWA and PMO understand that the Environmental Public Policy Law has been fulfilled. |
| **Tree Inventory and Mitigation Planting Plan** | 67% shall be paid when the inventory and planting plan is submitted with the technical assistance request application to PMO, up to 80% when PMO issue the permit and the remaining 20% when the Authority accept the final planting plan, with specifications and cost estimates, and with the corrections established on the conditions issued on the PMO permit, and when the final planting plan makes part of the PROJECT design construction plans. |
| **Noise Study** | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
| **Air Quality Study** | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
| Service Description | Payment Schedule |
|---------------------------------------------------------|-----------------------------------------------------------------------------------|
| Biological Assessment | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the report is accepted by ESO and is submitted to FWS. The remaining amount shall be paid when the study is accepted by FWS. |
| Environmental Justice Study | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
| Socio Economic Study | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
| Public Involvement | 80% shall be paid when FEIS approved and the remaining 20% shall be paid when ROD approved by FHWA. |
| Environmental Site Assessment (Type ASTM-Phase I) | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
| Environmental Site Assessment (Type ASTM-Phase II) | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
| Survey to Determine the Presence of Asbestos and Lead | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
| Jurisdictional Determination (USACE) | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the report is accepted by ESO and is submitted to USACE. The remaining 20% shall be paid when the determination and the permit is obtained from USACE. |
| Joint Permit Application (USACE) | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the report is accepted by ESO and is submitted to USACE. The remaining 20% shall be paid when the permit is obtained from USACE. |
| Underground Injection | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the document is accepted by ESO and is submitted to PMO. The remaining 20% shall be paid when the permits are obtained from PMO. |
| Ecological Study (Flora & Fauna) | 60% shall be paid when the document is submitted to ESO for review. Up to 80% shall be paid when the report is accepted by ESO and is submitted to FWS. The remaining amount shall be paid when the study is accepted by FWS. |
| **Puerto Rico Planning Board (PRPB) and Autonomous Municipalities (AM) adoption (if required)** | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
|---|---|
| **Benefit /cost analysis** | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
| **4(f) Programmatic Determination** | 60% shall be paid when the document is submitted to PRHTA for review. Up to 80% shall be paid when the document is accepted by PRHTA and is submitted to the FHWA. The remaining 20% amount shall be paid when the document is accepted by FHWA. |
| **Feasibility Study** | 60% shall be paid when the document is submitted to PRHTA for review. Up to 80% shall be paid when the document is accepted by PRHTA and is submitted to the FHWA. The remaining 20% amount shall be paid when the document is accepted by FHWA. |
| **Environmental Analysis** | 80% shall be paid when the document has been submitted for ESO’s review. The remaining 20% shall be paid when the study is approved by ESO. |
APPENDIX D
GENERAL DESIGN CONSIDERATIONS AND CONDITIONS
PAVEMENT REHABILITATION (N/A)
Pavement rehabilitation design shall be in accordance with the AASHTO Guide for Design of Pavement Structures 1993, AASHTO Mechanistic-Empirical Pavement Design Guide (MEPGD) latest edition, PRHTA’s Standard Specification for Road Bridge Construction, PRHTA’s Highway Design Manual, PRHTA’s Standard Drawings and PRHTA’s Design Directives for Pavement (200 series) latest edition, and FHWA’s Pavement Design Considerations. In general, for all major rehabilitation projects, each of the following steps should be followed to properly analyze and design the project:
1. **Project Evaluation**
a) Obtain the necessary information to evaluate the performance and establish the condition of the in-place pavement with regard to traffic loading, environmental conditions, material strength, and quality. Historical pavement condition data can provide good initial information and may be obtained from the PRHTA Pavement Management Office and Highway Design’s Plans Historic Archive.
b) Identify the types of pavement distresses and the factors causing the distresses before developing appropriate rehabilitation alternatives. Evaluation should be based in tools necessary to analyze pavement failures, such as coring, boring, trenching, and deflection measurements.
c) Evaluate the array of feasible alternatives in terms of how well they address the causes of the deterioration, repair the existing distress, and prevent the premature reoccurrence of the distress.
2. **Project Analysis**
a) Perform an engineering and economic analysis of candidate strategies. The engineering analysis should consider the traffic loads, climate, materials, construction practices, and expected performance. The economic analysis should be based on life cycle costs and consider service life, initial cost, maintenance costs, user costs, and future rehabilitation requirements, including maintenance of traffic.
b) Select the rehabilitation alternative that best satisfies the needs of a particular project considering economics, budget constraints, traffic service, climate, and engineering judgment.
3. **Project Design**
a) Conduct sufficient testing, both destructive and non-destructive (Cores, Falling Weight Deflectometer (FWD) testing, etc.), to verify the assumptions made during
the alternative evaluation phase. If needed, consider a new distress survey if the original condition survey was sample based or if the survey is not current in terms of the time the project is scheduled to go to contract.
b) Consider and address all factors causing the distress in addition to the surface indicators in the final design. Such factors as structural capacity, subgrade support (AASHTO classification, etc), surface and subsurface drainage characteristics need to be considered and provided for in the final design.
c) Once a rehabilitation alternative is selected, design the project using appropriate engineering techniques. A number of publications are available to guide the selection of these engineering techniques. The FHWA’s “Pavement Rehabilitation Manual” and the NHI training course “Techniques for Pavement Rehabilitation” provide excellent guidelines. There are also a number of excellent guides available from the asphalt and concrete industries.
**UTILITIES AND LIGHTING (N/A)**
1. The utilities relocation design shall comply with the standards and requirements of PRHTA, PREPA, PRASA, and any other concerned agency. These standards and requirements apply to temporary and final utilities relocation.
2. The utilities relocation design shall be developed in accordance with the P.R.H.T.A.’s newly developed Master Utility Agreements with P.R.E.P.A. and P.R.A.S.A., and any other utility company.
3. Coordination with private utility owners shall be done, including preparation of existing plans and relocation plans.
4. The final plans shall be endorsed by the utilities agencies and private companies.
5. The Right of Way necessary for the relocation design shall be taken in consideration in the preparation of the R.O.W. plans.
6. Design of anchors or structural bases of poles shall be done in accordance with the geotechnical study recommendations.
7. If needed, a study of undergrounds utilities (SUE study) shall be done to determine any possible conflicts with the proposed design.
8. The Maintenance of Traffic (MOT) shall take in consideration the relocation of utilities.
9. The new or rehabilitation of the lighting system shall comply with the new regulations of Puerto Rico Environmental Quality Board (Junta de Calidad Ambiental).
**STRUCTURES**
1. Structural design and aesthetics aspects shall be in accordance with the AASHTO LRFD Bridge Design Specifications 2014 and Interims or latest edition; AASHTO Manual for Bridge Evaluation, 2nd Edition of 2010, with 2011, 2013, 2014, 2015 and 2016 Interim Revisions; PRHTA Standard Specification for Road Bridge Construction and PRHTA Highway Design Manual.
2. Design of reinforced concrete members shall be in accordance with all PRHTA’s Design Directives for structures (300 series). All concrete shall be designed using the latest version of Special Provision 934 (Structural Concrete), according to Design Directive 306.
3. The stability of nearby existing structures, main or secondary bridge structural elements not intended to be preserved or rehabilitated during construction shall be taken into account in the project design process (e.g., historical, temporary steel bridges, etc.).
4. Consider construction aspects of the proposed structure (proximity to existing structures to remain, transportation of precast girders to project site, location of piers and abutments, availability of materials, etc.). Design shall consider preparing a construction sequence to guarantee the stability of structures and traffic flow.
5. If a Benefit/Cost Ratio analysis justifies it, the use of permanent bridges (steel or concrete Prefabricated Bridge Element Systems for Accelerated Bridge Construction (PBES/ABC)) are encouraged and shall take into account local suppliers and fabrication practices.
6. Consideration of the seismic performance of the connection between precast or prefabricated elements shall be considered as part of the design. Conceptual or experimental connections that have not been tested or constructed on a project in any United States jurisdiction shall not be permitted in the design.
7. A geosynthetic geotextile (Mirafi Miragrid or similar) shall be used in lieu of sheathing. The plans shall have a note specifying that the geotextile material and its installation shall be a subsidiary obligation of the cast-in-place concrete.
8. Design of structural elements shall consider corrosion and maintenance issues (joints, bearings, bolts, etc.).
9. For retaining structures, the alternatives for consideration shall include reinforced concrete retaining walls, MSE walls, lagging walls, anchored walls, gravity walls, and any other possible retaining structure. The most cost-effective and technically suitable alternative shall be designed. To minimize cost, each retaining wall shall be as small as possible in height and length. The final alternative shall minimize construction time and cost.
10. A load rating analysis shall be performed for the proposed bridge using the latest version of the AASHTO Manual for Bridge Evaluation, 1st Edition of 2008; the Recording and Coding Guide for the Structure Inventory and Appraisal of the Nation’s Bridges (Coding Guide) and the AASHTOWare Bridge Rating software complying with Design Directive 310.
**TRAFFIC & HIGHWAY SAFETY**
1. If this is a safety improvement project, all existing vehicle crash data shall be analyzed and safety improvements shall be determined for the roadway corridor. A report with all findings shall be submitted to PRTHA early at the preliminary design stage. The CONSULTANT shall procure the crash data and reports available from PRHTA and utilize suitable methods included in the Highway Safety Manual (HSM), such as crash modification factors, to determine the safety improvements.
2. If a traffic evaluation or study is needed, the data gathering and the subsequent report shall be performed following the PRHTA Guides for the Preparation of Operational and Transit Studies. This study shall analyze, but not limited to, determine highway capacity.
for all roadways involved, determine (ADTpresent, ADTfuture, DHV, K, %trucks), the existing intersections, the alternatives developed by the main CONSULTANT, and if needed develop any intersection alternative.
3. Maintenance of Traffic (MOT) plans with their different stages shall be made in accordance to the proposed construction.
4. The Maintenance of Traffic (MOT) shall comply with the new Puerto Rico Department of Transportation and Public Works “Work Zone and Mobility Policy” (Sept. 2015). Public information shall be performed as part of these operations report as needed and prepare a Transportation Management Plan (TMP).
5. A discussion of the MOT, with all the different stages, shall be included in the Technical Proposal (TP).
6. The CONSULTANT shall comply with the latest FHWA’s Policy of Interstate Access and with Interstate System Informational Guide.
7. Final pavement marking and signing plans shall be developed.
8. MUTCD 2009 Edition shall be used for all traffic design.
9. If a traffic signals system is required for the new design, it shall comply with all standards, supplemental and/or special provision specifications of PRHTA. Also the new system shall provide a wireless communication and video detection system in conformance with Special Provision Specifications 654.3-21 and 654.3-22.
10. If this is a safety improvement project, all alternate roadway/improvement designs shall be analyzed and determined for the corridor using suitable methods included in the Highway Safety Manual (HSM), such as prediction models to estimate crash frequency and severity. The CONSULTANT should prepare a Highway Safety report with all findings and recommendations of the HSM analysis and deliver it to PRHTA at the preliminary design stage. In the report, all existing crash data shall be analyzed and every highway safety improvement shall include the evaluation of Crash Modification Factors (CMF) to quantify the safety effects.
**HYDROLOGIC AND HYDRAULIC STUDY (N/A)**
1. Where necessary, the CONSULTANT shall submit a detailed HH study comparing the existing condition vs. the proposed condition and the effect of the proposed works on the hydraulic condition of the area. It shall comply with the Planning Board Regulation No. 13 and any other concerning regulation.
2. All studies shall comply with the *latest version* of “Guías para la Elaboración de los Estudios Hidrológicos Hidráulicos” from the Puerto Rico’s Department of Natural Resources and Environmental.
3. The precipitation depth values needed for the hydrologic analysis shall be obtained from the NOAA Atlas 14.
4. Where necessary, the CONSULTANT shall submit a scour study explaining the parameters used to calculate the scour and the conclusions and recommendations from the study. It shall consider the existing condition and compare it with the effect of the proposed scour or erosion’s countermeasures.
5. The scour study shall be performed following the latest version of FHWA’s HEC-18 “Evaluating Scour at Bridges”, HEC-20 “Stream Stability at Highway Structures” and HEC-23 “Bridge Scour and Stream Instability Countermeasures”.
6. At the completion of these studies, five (5) printed copies, each one with a CD containing a non-editable (PDF) electronic file, of the Report shall be furnished to the AUTHORITY. Five CDs with the digital models (HEC-HMS, HEC-RAS) should be submitted with each of the Hydrologic and Hydraulic report. Survey plans of the cross sections used in the Hydraulic Study should be printed and signed, by a Licensed Surveyor in Puerto Rico, and included in each of the Hydrologic-Hydraulic Report. The Hydrologic-Hydraulic Study shall be approved by the proper regulatory agencies. The CONSULTANT, if necessary, shall make any corrections to the study until approval is obtained.
DRAINAGE (N/A)
1. The new design shall take in consideration the drainage aspects of the area.
2. A proper drainage through the whole segment system shall be designed.
3. Drainage calculations shall be made and delivered for any new or existing system related to this project.
4. The drainage systems shall be designed and placed considering the following aspects, but not limited to:
a) Avoid damage by erosion in the outlet structure.
b) Avoid damage by erosion to the protection designed for the embankments.
c) Use a 25 years frequency design flood for roadside ditches or channels and roadway surface drainage systems.
d) Use a 50 years frequency design flood for cross drainage structures. Check for the 100 year frequency design flood and design for it if economically feasible.
5. The CONSULTANT shall submit a detailed drainage study. For the cross drainage structure the CONSULTANT shall compare the existing condition vs. the proposed condition and the effect of the proposed works on the hydraulic condition of the area.
EROSION CONTROL (N/A)
The designer will design the installation of temporary and permanent erosion and sediment controls, to comply with minimum criteria and standards for the Erosion and Sediment Control Regulations. The regulations were enacted to protect the health, safety, and property of the citizens and to help protect area water resources from pollution due to erosion and transportation of sediment. The regulations require that no person cause or allow sediment to be deposited in any public street, public land, or on any property not under their control as a result of land disturbance from construction activities.
The consultant shall comply with the regulations by meeting the following objectives:
1. Minimize the area disturbed by construction and development;
2. Provide for containment of sediment until areas are stabilized;
3. Stabilize disturbed areas as soon as practical after project completion; and
4. Provide permanent erosion, drainage, and detention controls.
As part of the contract package the consultant shall include the drawing, specifications and method of payment of the recommended measures.
**GEOTECHNICAL INVESTIGATION**
If a geotechnical investigation is needed, enough data shall be taken to satisfy the design needs, and for all proposed temporary and final structures. In addition, the study shall consider borings for pavement recommendations, as required for pavement design.
The required Geotechnical studies for design shall be performed by the CONSULTANT, who will be able to subcontract this work to a company or firm that is capable to carry out such task in accordance with the applicable provisions of the Highway Design Manual. This shall be invoiced according to Appendix D.2.
The CONSULTANT, shall prepare jointly with the selected Geotechnical CONSULTANT, the subsurface soil investigation boring program, including the necessary plans for the soils exploration for the PROJECT as required to secure the subsurface information of the most economical design for the PROJECT, in accordance with the applicable provisions of the Highway Design Manual and Geotechnical Investigation Manual for Design and Highway Construction prepared by the AUTHORITY and any other applicable manual.
The geotechnical CONSULTANT should consider the utilities at the site (aerial and underground) before establishing the proposed boring location.
1. **Issues to be considered in the proposal**
a) All permits necessary to perform the borings for the geotechnical study shall be provided for the designer and/or the geotechnical CONSULTANT.
b) The geotechnical CONSULTANT should consider the presence of utilities in the site at the moment of establishing the boring locations.
c) For bridge structures, one boring for every pier and each abutment along the bridge alignment is required. The borings should be located on alternate sides of the structure.
d) If general subsoil information of the area is known, the spacing for the borings should be at approximately 500 feet or 150 meters along the proposed highway alignment. If compressible soils are expected, it is required intermediate borings along the area.
e) At least, perform one boring where a structure is proposed; such as culverts, retaining walls, sewers, etc.
f) In areas of deep cuts (approximately 10 ft high or more) and high embankments (10 ft or more), perform at least three borings aligned in the critical direction.
perpendicular to the roadway. This, in order to draw a soil profile across the section for stability analysis.
g) Verify the need of any pilot road to allow initial access to the project site, in case there is abundant vegetation or irregular topography.
h) Consider the need of specialized laboratory tests.
i) Observe and consider the presence of historical sites; the way they could affect or be affected on both design and construction stages.
2. Geotechnical Report
The geotechnical CONSULTANT shall perform all the necessary geotechnical and geologic studies for the design and construction of the proposed project. The geotechnical report shall include the necessary final recommendations for design and construction of all structures and soil works. The final geotechnical study shall be completed, in order to proceed with the Preliminary Design (35%).
All studies, sampling, analysis and reports shall include, but not limited to:
a) Geology:
Detailed site geology analysis, presenting the different geologic formations, soil and rock types, the location of hazards like slides, sinkholes, creep areas, etc., and techniques to correct them, when necessary. This analysis should be certified by a geologist.
b) Boring Logs:
Provide boring location plan with their coordinates and elevations.
The boring logs should be presented as indicated on Design Directress 110.
The logs shall present the soil description and classification of each soil and rock stratum, and depth and elevation to the top and bottom of each stratum.
Plots of N-values, RQD, water content and stratum descriptions shall be included.
If encountered, the groundwater level shall be included in this profile.
The soils stratum shall be classified as AASHTO and USCS System.
c) Geotechnical Instrumentation:
The report shall also include recommendations on the monitoring program to be implemented. The instrumentation to be used and their location and the frequency and depth of the readings shall be specified.
If observation wells and/or piezometers were installed, a summary of the readings and their interpretation shall be included. The effect of seasonal changes and the presence of perched and artesian water shall be evaluated.
d) Foundations:
Recommendations shall be given for all foundation alternatives (shallow or deep) for the abutments, piers and temporary supports. Lateral loads shall be taken into consideration.
Soil type shall be specified for foundation fills on bridge abutments, and along the project alignment in general, in accordance with Specification 203.
One boring per each bridge abutment and pier shall be performed.
e) Pavement:
Provide the required pavement design parameters.
f) Slope stability, Soil Improvement:
Present various corrective measures to improve the soils, if needed.
When retaining structures, the report should present various alternatives. Final recommendations with the best type of retaining structure and the reasons for the selection made shall be presented.
Internal and external stability analyses shall be presented.
g) Cuts and Fills:
For all cuts and fills, slope stability analysis shall be included, for static and dynamic conditions. These should guarantee stability on the proposed earth works.
The static and dynamic stability needs to be verified. Present the factor of safety used.
For fills, the report shall specify the type of soils to be use on the project, in accordance with 203 Specification.
For fills, recommendations to avoid erosion of the slope embankment shall be provided; such as geosynthetics installation, riprap, drainage systems, etc.
For cuts, indicate if heavy equipment or explosives are required.
Evaluate if the soil to be cut can be used for fills on the project.
Verify any kind of protection or drainage systems that may be needed.
Consider security areas needed for the works to be done.
h) Excavations:
The excavation procedures shall be done in accordance with 206 specifications.
General recommendations shall be provided for the excavation procedures and stabilization methods.
i) **Settlement Control:**
Various alternatives for settlement control should be given, if needed.
Verify the allowable maximum settlement for bridges and retaining structures.
j) **Scouring:**
The report shall contain the field information needed for the performance of the scour evaluation, such as the soil profile at the river bed and grain size distribution.
3. **Required geotechnical study parameters**
The report shall include the following geotechnical design parameters, but not limited to:
a) Soil unit weight\(^1\)
b) RQD Values
c) Water content of soil samples
d) Standard Penetration Test (SPT) N values
e) Angle of internal friction \((\phi)\)\(^1\)
f) Atterberg limits
g) Unconfined compression resistance, \(q_u\)\(^1\)
h) Allowable soil bearing pressure
i) Cohesion intercept \((c)\)\(^1\)
j) Compression coefficient \((Cc)\)
k) Recompression coefficient \((Cr)\)
l) Consolidation coefficient \((C_v)\)
m) Maximum vertical stresses on the soil mass \((\sigma'p)\)
n) Vertical stresses “in-situ” \((\sigma'v)\)
o) Sub grade Reaction Module 2
i. Horizontal \((K_h)\)
ii. Vertical \((K_v)\)
p) Soil parameters to determine the seismic effects in accordance with AASHTO LFRD section 3.10
q) Parameters for “p-y” curves in piles 1,2
r) Include the shrink and swell soils factors, if apply
\(^1\) Specify source of values (test performed or correlation used for estimated value)
\(^2\) – If apply
4. **Other considerations**
a) A soil profile shall be provided.
b) The presence and extent of compressible, organic, expansive, collapsible and liquefiable soils and the effects on the Project shall be evaluated.
c) If used, figures, tables, photos and graphics shall be presented in the report.
d) A summary of test results (such as Atterberg limits, sieve analysis, specific gravity, direct shear, consolidation, triaxial, etc.), and their interpretation shall be presented.
e) Copy of all the analyses and calculations done for the report shall be included.
f) The required geotechnical topics not included in this Appendix shall be discussed with the PRHTA Geotechnical Engineering Office for their evaluation and approval.
5. Deliverables
The CONSULTANT shall furnish the AUTHORITY five (5) printed copies, each one with a CD containing a non-editable (PDF) electronic file of the report, of the any Preliminary and the Final Geotechnical Report during the preliminary design process.
RIGHT OF WAY (ROW) PLANS (N/A)
The ROW Plans shall include the existing Right Of Way (ROW) as the new lots needed for this project, if any. The State Plane Coordinates must be referenced to the last version of the North American Datum of 1983 (NAD 83) on Lambert Conformal Conic Projection for Puerto Rico. ROW Plans shall be prepared following Puerto Rico Highway Authority design standards (HDM), the Design Directives number 107, 111 D-1 and 111 D-3 and the Laws of the Commonwealth of Puerto Rico, as the followings but not limited to: Article 240 of Law 198 of August 8th of 1979 as amended; Articles 4, item c and Articles 16, 23 of Law 173 of August 12th of 1988 as amended; Law 216 of December 29th of 2009; Law 235 of August 30th of 2000; Law 264 of November 16th of 2002 (System of State Plane Coordinates in 83 NAD or their more recent version). They shall be scheduled to be delivered in two stages:
1. Preliminary (if properties acquisition are needed):
The CONSULTANT shall prepare preliminary Right of Way plans, for which he shall perform all the necessary research of affected properties or other for this purpose. All documents shall be furnished to the AUTHORITY in paper and in a CD containing the electronic files in Autocad 2007 (DWG) version or the one agreed by both parties, and non-editable (PDF) format, whichever applies. The preliminary Right of Way plans shall contain, but not limited to, the following:
a) Title sheet with the PROJECT identification and location;
b) Plans indicating the north, scale, structures, tables with parcel numbers, property owner names, parcel area to be segregated or farm to be acquired, property limits,
geometry (center lines or base lines, curve elements, spirals, etc.) of existing legal and proposed right of way, right of way limits, adjoining owners, property limits, river and creek names, riverbed limits, existing streets and roads names, servitudes of electrical, water distribution, drainage and sanitary sewage system infrastructure, etc., any property value traits, table with structure number and property owner names. All structures or fixed objects which occupy the public right of way airspace shall be indicated.
c) Existing ingress and egress to and from affected farm and the proposed ingress and egress to and from parcels and farms to be served by them.
d) The State Plane Coordinates System shall be indicated, as well as the Vertical Datum.
e) A preliminary report with legal descriptions of all parcels shall be delivered to the AUTHORITY. Each parcel shall indicate the surface area, Municipal demarcation and ward; in addition to the description of all points which define its boundaries, beginning at a point then going over each point of the polygon until reach the point of beginning. Every point shall have an identifier and positioned by its Cartesian coordinates duly georeferenced; azimuths or bearings, distances, radios, and arc length of parcel boundaries shall be consigned. Adjoining farms, owners' names and cadastral codification shall be informed.
f) A preliminary report of "The Research of Properties" (Item 2) shall be furnished with the plans.
2. **Research of Affected Properties (Farms)** (if properties acquisition are needed):
The CONSULTANT shall corroborate the affected property (farm) and its owners by performing ownership title searches and obtaining corresponding certificates of the Registry of Properties, title documentation as deeds or private documents. The CONSULTANT shall include the following:
1. Inscription Plan and the ARPE or OGPe Approval Resolutions.
2. Existing ROW plans, official surveys, judgments, legal descriptions, among others. Owners information from the Center for Collection of Municipal Income (CRIM), as well as from the Planning Board and/or Permit Management Office (OGPe) and other government agencies and corporations.
3. Information from records of attorneys, engineers, architects or surveyors offices.
A bound printed document containing evidence and information as result of the search shall be furnished to the AUTHORITY. The farm shall be the base of the research.
The CONSULTANT shall fill out the form "ESTUDIO COMPARATIVO DE FINCAS", one per parcel.
The above information shall be used to prepare the Preliminary and Final Right of Way Plans. All documents shall be furnished to the AUTHORITY in paper and in a CD containing electronic files in non-editable (PDF) format.
3. Final (if properties acquisition are needed):
The CONSULTANT shall prepare the Final Right of Way Plans, which shall show, not limited to, the following:
a) Title sheet with the PROJECT’s identification, state or federal codification, date on which the plans were made, number and total of sheets, PROJECT’s location, name, professional services corporation, signature and seal of the professional surveyor or professional engineer which belongs to the Permanent Registry of Surveying, who designed the plans and names of the AUTHORITY officials with space provided for their signatures;
b) Right of Way Plans indicating the true north, graphic and numerical scale, survey tables of parcels of interest with bearings, distances, coordinates, point descriptions. Tables with parcels, affected property owner names and areas in square meters and “cuerdas”. The parcels’ codification shall consist of five (5) digits: the first three (3) for the property and the other two (2) for the property section to be segregated;
c) Coordinates with their permanent ties and location elements to facilitate the delimitation and the amendments or revisions that requires field data survey;
d) Name of rivers, creeks, streets, roads, avenues, roadways, etc.;
e) Official demarcations of riverbeds and/or the demarcation of maritime zone and terrestrial zone; the same shall be approved by the Natural Resources and Environmental Department. Those plans shall be included as part of the final ROW plans.
f) In case that construction, temporary or permanent easements are required, they shall be identified as such. The same way for electrical servitudes or from another utilities, survey tables shall be prepared for them;
g) In case that squatter settled property exists, the parcels formed as a result of the squatter settlement, shall not be counted as independent parcels; despite fences and structures these shall be properly identified and individual survey tables shall not be prepared for these. The section of the squatter settled property to be acquired shall be shown, with its survey table and the property owner’s name. All accesses, existing and proposed which shall be used after the PROJECT’s construction, shall be identified.
h) A final report with legal descriptions of all parcels shall be delivered to the AUTHORITY. Each parcel shall have its surface area, Municipal demarcation and ward; in addition to the description of all points which define its boundaries, beginning at point then going over each point of the polygon until reach the point of beginning. Every point shall have an identifier and positioned by its Cartesian coordinates duly georeferenced; azimuths or bearings, distances, radios, and arc length of parcel boundaries shall be consigned. Adjoining farms, owners’ names and cadastral codification shall be informed.
i) A final report of the Research of Affected Properties (Item 2) shall be furnished with the plans.
These plans shall be submitted in ink on linen or vellum paper or similar material approved by the AUTHORITY shall become property of this one. In addition, the CONSULTANT shall submit the final plans in compact discs (CD), in editable (DWG) and non-editable (PDF) digital formats, signed electronically, and programmed to be printed in 36 in. wide by 24 in. high, which contain all the drawings included in the final plans. The editable format shall be in Autocad 2007 version or the one that both the parties decide.
If the project does not require any acquisition a certified plan of the existing ROW shall be submitted, the same shall be signed by a Licensed Surveyor in Puerto Rico or Licensed Engineer authorized in Puerto Rico to practice Surveying. The ROW Plans shall include the highway existing ROW, legal easements for, but not limited to, utilities, privates and publics and physical limits, i.e.: fences and objects which materialize apparent limits of easements and or properties, if in services of Surveying this Plan was not done.
The preparation of land acquisition plans shall be in accordance with the PRHTA Highway Design Manual\(^2\) (HDM) and with the following:
1. **PLANS CONTENTS**
The following information shall be included in the ROW plans:
a) Existing (as built) features such as fences, roadway area, abutting structures and property limits of lots to be affected\(^3\). The bearings of existing permanent fences shall be given. Objects or features marking property limits shall be shown. Location the limit of rivers, channels, creeks or any water source affecting pasture rights or near this project or affected by the same. Location of the limits of the Maritime Terrestrial Zone and the classification of lands within the survey area or affected by this project shall be identified.
b) Existing accesses and egresses of farms or properties affected must be shown and identified. A table identifying the proposed accesses and/or egresses indicating the properties that will be benefited most be included. Servitudes can be constituted on properties acquired by Puerto Rico Highway and Transportation Authority to connect parcels to provide access and/or egress to farms affected by this project.
c) Existing legal easements or servitudes with its geometry must be shown.
d) Proposed legal easements or servitudes with its geometry must be shown.
---
\(^2\) PRHTA Highway Design Manual chapter 19, sections 19-01, 19-02 y 19-03
\(^3\) A layer shall be assigned to each feature in the CAD drawing, i.e. fences, property limits, control stations, structures etc.
e) When a farm or property is divided by this project in one or more parcels. The parcels not to be acquired and that does not have access and egress must be provided for an access and egress equal or better than the existing one, according to the legal use at the moment of the acquisition.
f) Trees, buildings, telephone poles, wells, septic tanks and similar features within the ROW lines shall be carefully shown on the plans to their true location and dimensions if their removal or alteration affects property settlements.
g) Structures to be acquired and demolished shall be identified with Arabic numbers in a sequential way starting with number 1. A structures table indicating the structure number, owner's name, construction material and number of floors in building structure shall be included on each sheet.
h) The right of way of existing highways (legal ones or based on land acquisition plans of previous projects or by other means), PREPA, PRASA or any other utility servitude. The structures or fixed objects occupying the aerial space of the public right of way shall also be identified.
i) Parcels to be acquired for the project and their land remnants shall be identified in the ROW plans. Tables indicating Parcel, zoning, land use, owner, area in square meters and "cuerdas", and the property registry number shall also be included.
j) In cases where land properties had been settled by squatters, the lots formed due to this appropriation shall not be counted as independent lots. Nonetheless, existing fences and structures in the lot shall be identified and shown. Survey tables shall not be prepared for squatter's parcels. The portion of the squatter's land to be acquired shall have its survey table and owner identified. Existing and proposed accesses to the property shall be identified.
k) Land remnants of affected properties smaller than 300 square meters or that does not complies with state or municipal Planning Board regulations, will be measured and represented in the ROW plans as land remnants to be acquired.
l) Temporary or permanent construction easements to be acquired shall be shown in the plans. This shall also apply to PREPA or of any other infrastructure, for which survey tables shall be prepared.
m) Orthogonal coordinates along the proposed ROW based on the construction or survey line, the project centerline, stations, alignment bearings, curve or spiral elements, horizontal and vertical survey controls and net length of centerline across the affected property shall be shown. Coordinates at the beginning and end of the project centerline or construction line as well as stations along that line shall be identified. Control stations with their coordinates shall also be identified, and at least three references with its description shall be included. A minimum number of baselines outside the project with documented control stations shall be established for use during and after construction of the project. Construction limits, access control lines, proposed accesses, proposed ROW and the limits of proposed cut or fill slopes shall be shown.
n) The identification of lots to be acquired shall be done through the code number format XXX-YY, where XXX represents the number of the affected land property.
and YY represents the number of the lot to be segregated from the XXX property. The principal land property shall have an YY number of 00, e.g. 001-00, 002-00.
o) The lots that will be segregated and acquired from the principal land property will have YY numbers starting with 01, 02, and so on, e.g. 001-01, 001-02, etc. The land remnant will be identified with the principal property number.
p) The proposed design information and the contour grade lines shall be represented with background gray lines and text. Every item shown in the plans shall be easily differentiated from one another and legible.
q) Cross sections shall be included in those situations where there are structures that will not be acquired located within 100 meters of the proposed right of way.
r) The final plans shall be signed, sealed, and submitted on Mylar paper and in digital DWG or DXF AutoCAD 2002 (or later if PRHTA require it) or another format agreed by both parties.
s) Official Seal and sign of the Licensed Surveyor in Puerto Rico must be stamped, as well as the identification of the Professional Corporation responsible for conducting the surveying works and/or the Limited Liability Partnership must be shown.
2. BASIC DOCUMENTATION REQUIRED FOR THE PREPARATION OF ROW PLANS:
This phase of investigation shall be performed prior to or simultaneously with the field survey, so that the gathered information must be integrated to the survey plans. The investigation required for the preparation of ROW plans shall consist in searching and obtaining the following documents:
a) Existing ROW plans from the PRHTA or DTPW archives, located at ninth, twelfth and thirteenth floor of the Roberto Sanchez Vilella Governmental Center’s South Tower.
b) Owner and land property information from the Municipal Revenues Collection Center (CRIM), the Property Registry, the Permits Management Office (OGPE), and the Planning Board, among other government agencies.
c) Information through interview with owners, lawyers, surveyors and engineers, among others.
3. DISAGREEMENTS ON LAND OR LOT AREAS
When whole land properties need to be acquired, if a discrepancy between the survey area and the registered area arises, larger than 20 percent in rural land properties or 10 percent in urban land properties\(^4\), it will be necessary to obtain agreements from the abutting property owners in order to ensure the correctness of the survey.
4. PRELIMINARY AND FINAL ROW PLANS
\(^4\) The “rural” or “urban” classification shall conform to the Puerto Rico Planning Board criteria.
Right of Way plans shall be prepared in two stages as described below:
a) Preliminary Plans\(^5\) shall include:
1. Title Sheet:
i. Project Identification
ii. Project location in a topographic map
2. North Arrow
3. Numeric and Graphic Scales
4. Table per sheet indicating the lot, owner name, affected areas to be acquired and land remnants.
5. Table per sheet indicating the easements and servitudes, beneficiary names and affected areas.
6. Table per sheet indicating the existing and the proposed accesses and egresses. Lots to be served must be indicated.
7. Existing information (as built), i.e. fences, roadway area, accesses, affected structures adjacent to the project, property limits, existing legal ROW, riverbed and creeks limits, Terrestrial Maritime Zone limits existing highways etc.
8. Relevant reference information to identify the project.
9. Base lines and control stations, construction limits, access control lines, proposed accesses, proposed ROW, limits of proposed cut or fill slopes, lots to be acquired and land remnants.
b) Final Plans\(^6\):
1. Final ROW plans shall be prepared in accordance with the criteria contained in this appendix and with the HDM sections 19-02.1, 19-02.02, 19-02.04, 19-03.02 and 19-03.006. The title sheet shall be signed and sealed by the CONSULTANT\(^6\).
5. MONUMENTS OR LANDMARKS
Base lines outside the project shall not be less than 400 meters and longer than 1,000 meters, and shall consider the total project length. These stations shall be identified by permanent monuments, in order to establish the project alignment, if case of any compensation claims related to acquired lands arise, and to integrate information to the project or to make additional surveys.
The monuments for horizontal control shall have a concrete base with an embedded 6 inch diameter and 12 inch long plastic (PVC) pipe and a 3/8 inch diameter and 24 inch long steel bar in its center. The pipe shall extend 2 inch over the ground level. This control
\(^5\) PRHTA Highway Design Manual section 19-03.01.
point shall be identified with an aluminum disk in which the name of the agency, i.e. "Autoridad de Carreteras y Transportación", project code number, year of monument establishment, control point identification and the name of the office responsible for establishing the point shall be inscribed on the disk.
All the horizontal control points shall be referenced to three points. A drawing shall be prepared showing the description and the references of all monuments. This information shall be included in the land acquisition plans, indicating that it shall be used exclusively for the project.
6. GEOMETRIC DESCRIPTION OF LOTS TO BE ACQUIRED
In order to take full advantage of the information provided in the land acquisition plans, specifically the information provided in the survey tables, the geometry of the lots shall be described in the legal acquisition documents.
The information to be included in the description shall be the following: the starting point, its coordinates and description, then connecting to the next point indicating its coordinates and description, then continuing to the following points indicating its coordinates and descriptions until returning back to the starting point. Curves must be described indicating their arc length. This description is complementary to the general description included in the land acquisition plans, which make reference to the owners of lands abutting the described lot, the lot area in square meters and "cuerdas", and the municipality where the lot is located.
This information shall be submitted in Mylar paper and in digital format, identifying all lots by its number.
7. OTHER CONSIDERATIONS
a) All land properties affected by the project shall have at least one proper access according to the land use; otherwise consider the acquisition of the properties if an engineering solution is more expensive.
b) An extract from the title search of affected properties shall be prepared and submitted.
c) An analysis of parcels to be acquired versus the affected farm must be done, form for this task will be provided by PRHTA.
d) The right of way to be acquired shall be strictly limited to the area required for the project.
e) Acquisition of land lots with areas less than 5.00 square meters should be avoided.
f) The limits of lots to be acquired will be traced parallel to the construction centerline with 90 degree deflections (offsets).
## ATTACHMENT 1 – MUNICIPALITIES AND ROAD LIST
### Unit 1:
| Municipality | Km per Municipality | Region |
|----------------|---------------------|--------|
| Dorado | 19.35 | North |
| Manatí | 41.10 | |
| Toa Alta | 20.58 | |
| Toa Baja | 31.08 | |
| Vega Alta | 58.85 | |
### Roads per Municipality:
| ROAD | CITY | In Km | Final Km | FEMA In Km | FEMA Final Km | Abreviado In Km | Abreviado Final Km | Federal Aid Total Km |
|-------|----------|-------|----------|------------|---------------|-----------------|--------------------|----------------------|
| 659 | Dorado | 0.00 | 4.45 | | | | | 4.45 |
| 696 | Dorado | 0.00 | 5.00 | | | | | 5.00 |
| 697 | Dorado | 0.00 | 1.00 | | | | | 1.00 |
| 698 | Dorado | 0.00 | 2.10 | | | | | 2.10 |
| 699 | Dorado | 0.00 | 0.60 | | | | | 0.60 |
| 854 | Dorado | 0.00 | 4.30 | 2.90 | 4.30 | | | 2.90 |
| 6693 | Dorado | 0.00 | 1.30 | | | | | 1.30 |
| 8865 | Dorado | 0.00 | 2.00 | | | | | 2.00 |
| 641 | Manati | 0.00 | 8.10 | | | | | 8.10 |
| 642 | Manati | 0.00 | 14.00 | | | | | 14.00 |
| 685 | Manati | 0.00 | 7.00 | | | | | 7.00 |
| 6685 | Manati | 0.00 | 12.00 | | | | | 12.00 |
| 821 | Toa Alta | 0.00 | 2.50 | 0.40 | 2.50 | | | 0.40 |
| 823 | Toa Alta | 0.00 | 7.70 | | | | | 7.70 |
| 824 | Toa Alta | 0.00 | 4.75 | | | | | 4.75 |
| 825 | Toa Alta | 0.00 | 5.55 | 0.45 | 5.55 | | | 0.45 |
| 827 | Toa Alta | 0.00 | 7.28 | | | | | 7.28 |
| 819 | Toa Baja | 0.00 | 4.30 | | | | | 4.30 |
| 863 | Toa Baja | 0.00 | 2.45 | | | | | 2.45 |
| 865 | Toa Baja | 0.00 | 6.30 | | | | | 6.30 |
| 866 | Toa Baja | 0.00 | 7.40 | | | | | 7.40 |
| 867 | Toa Baja | 0.00 | 7.80 | | | | | 7.80 |
| 872 | Toa Baja | 0.00 | 2.83 | | | | | 2.83 |
| 647 | Vega Alta| 0.00 | 12.80 | | | | | 12.80 |
| 677 | Vega Alta| 0.00 | 9.60 | 6.10 | 9.60 | | | 6.10 |
The evaluation shall not include FEMA segments, abbreviated projects, and construction projects in the roads shown in the table or sections covered by other DDIR’s for landslides.
Unit 2: Municipal Roads
| ROUTE | COUNTY | FROM | TO | SECTION | MUNICIPALITY |
|------------------------------|--------|-----------------------|---------------------|---------|--------------|
| CALLE INMACULADA | 002 | (PR 670) | (PR 2) | 0.40 | MANATÍ |
| CALLE M. SÁNCHEZ | 002 | (PR 6685) | (CALLE INMACULADA) | 0.40 | MANATÍ |
| CALLE VENDING | 002 | (PR 670) | (PR 2) | 0.75 | MANATÍ |
| CALLE LOCAL | 002 | (PR 668) | (PR 149) | 0.45 | MANATÍ |
| AVENIDA DOS PALMAS | 001 | (BOULEVARD LEVITTOWN) | (PR 165) | 0.55 | TOA BAJA |
| AVENIDA RAMÓN RÍOS ROMÁN | 001 | (CALLE LOS MANGOS) | (PR-866) | 1.15 | TOA BAJA |
| BOULEV. LEVITTOWN | 001 | (PR 866) | (PR 165) | 2.70 | TOA BAJA |
| CALLE B. JAIME | 002 | (PR 854) | (PR 165) | 0.20 | TOA BAJA |
| AVENIDA SABANA SECA (LEVITTOWN) | 001 | (PR 167) | (PR 866) | 1.30 | TOA BAJA |
| AVENIDA SABANA SECA (LEVITTOWN) | 001 | (PR 866) | (AVE. MAGDALENA) | 0.70 | TOA BAJA |
| AVENIDA SABANA SECA (LEVITTOWN) | 001 | (AVE. MAGDALENA) | (PR 867) | 1.50 | TOA BAJA |
| CONECTOR CORUJO | 002 | (PR-2) | (PR-866) | 1.30 | TOA BAJA / BAYAMÓN |
| CALLE J. SEIJO | 002 | (PR 676) | (CALLE DR ACOSTA) | 0.15 | VEGA ALTA |
| CALLE J. SEIJO | 002 | (CALLE DR. ACOSTA) | (PR 2) | 0.08 | VEGA ALTA |
TOTAL 11.63
## ATTACHMENT 2 – UNIT PRICE
### UNITARY COST FOR SOILS ENGINEERING OFFICE
**JUNE 2016**
| ITEM | UNIT PRICE |
|----------------------------------------------------------------------|------------|
| **1. Boring with regular sampling (a)** | |
| N < 75 | $18/ft. |
| N ≥ 75 | $36/ft. |
| Rock, large gravel, cobble, and boulder (requiring coring bit or tricone) < 150 ft. depth (b) | $72/ft. |
| Rock, large gravel, cobble, and boulder (requiring coring bit or tricone) ≥ 150 ft. depth (b) | $75/ft. |
| **2. Boring with continuous sampling (a)** | |
| N < 75 | $26/ft. |
| N ≥ 75 | $50/ft. |
| Rock, large gravel, cobble, and boulder (requiring coring bit or tricone) < 150 ft. depth (b) | $75/ft. |
| Rock, large gravel, cobble, and boulder (requiring coring bit or tricone) ≥ 150 ft. depth (b) | $85/ft. |
| **3. Boring without sampling (a)** | |
| N < 75 | $10/ft. |
| N ≥ 75 | $18/ft. |
| **4. Horizontal Boring** | |
| Depth < 150 ft. | $100/lineal ft. |
| Depth ≥ 150 ft. | $115/lineal ft. |
| **5. Drilling Accessories** | |
| Boxes for Cores | $25/each |
| Water truck rental | $125/day |
| Water pump | $200/day |
| Water pump rental | $60/day |
| **6. Horizontal Drilling** | |
| Depth < 500 ft. | $100/lineal ft. |
| Depth ≥ 500 ft. | $115/lineal ft. |
| **7. Inclined Boring** | |
| N < 100 | $20/lineal ft. |
| N ≥ 100 | $38/lineal ft. |
| Rock, large gravel, cobble, and boulder (requiring coring bit or tricone) < 150 ft. depth (b) | $75/lineal ft. |
| Rock, large gravel, cobble, and boulder (requiring coring bit or tricone) ≥ 150 ft. depth (b) | $90/lineal ft. |
| **8. Open pit** | $180/each |
(a) The unit price does not depend of boring diameter. Includes content $q_u$ (penetrometer spring test) and visual-manual classification.
(b) Includes RQD (Rock Quality Designation)
| 9. Cores (for pavement purposes only) | |
|-------------------------------------|----------|
| Concrete Cores | |
| For Concrete Strength Test 2" X 4" | $125/each|
| Diameter 6" | $200/ft. |
| Diameter 8" | $225/ft. |
| Diameter 10" | $250/ft. |
| Asphalt Cores | |
| Diameter 8" | $150/ft. |
| Diameter 10" | $200/ft. |
| Mobilization of coring equipment | $5/mile |
10. Mobilization of drilling equipment to site (one time charge per equipment)
- Mobilization to Vieques or Culebra: $2,500/round trip
11. Set-up of borings (Includes Boring layout an coordinates with hand held GPS)
- Pavement and roadside: $40/boring
- Level terrain: $60/boring
- Rolling terrain: $125/boring
- Mountain terrain: $250/boring
- Wetland: $450/boring
- Water (with barge; includes Tender boat & captain): $525/boring
- Construction of Barge (20 ft. x 16 ft., includes transportation, materials, assemble): $3,500/each
- Canoe Rental and Mobilization: $375/day
12. Rental of heavy equipment
- D4-D6: $700/day
- D7-D8: $800/day
- D-9: $1,000/day
- Mobilization of bulldozer to project site (One time charge per project): $800/each
- Digger: $550/each
- Mobilization of digger to project site (One time charge per project): $500/one time pay
13. Per diem (drilling crew of 3): $180/day (outside of 40 mile perimeter)
14. Field tests
- Vane Shear: $125/each
- Undisturbed sampling
- Selby tube samples: $120/sample
- Denison samples (2 and 5 ft. long): $72/ft. of sample
- Observation well (1-1.5, PVC) including drilling: $18/ft.
- Observation well (1-1.5, PVC) without drilling: $8.00/ft.
- Observation well reading: $12/each
| Field tests (continued) | |
|------------------------|----------------------|
| Piezometer | |
| Riser pipe (including drilling) | $14/ft. |
| Tip | $175/each |
| Readings | $15/each |
| Inclinometers | |
| Installation (Installed in drilled borings with Portland Cement –water grout mix) 1.9” | $30/ft. |
| Installation (Installed in drilled borings with Portland Cement –water grout mix) 2.75” | $38/ft. |
| Readings | $2.50/ft. (minimum charge of $125) |
| Steel manhole 8” | $125/each |
| Flush mounted manhole | $125/each |
| Field Trip for Instruments Reading | $120/Trip |
| Geophysical Services (includes field work and consulting for Seismic Refraction of Electrical Resistivity) | $1,500/day |
| Geophysical Equipment Rental | $500/day |
| Non Destructive Testing (NDT) (for geotechnical purposes only) | |
| Pulse Echo or GPR field work and consulting services (includes 1 senior engineer, 2 technicians) | $1,500/day |
| Pulse Echo or GPR equipment rental | $500/say |
15. Laboratory tests
| Test | Cost |
|-----------------------------|----------|
| Unit weight | $35/each |
| Specific gravity | $50/each |
| PH | $40/each |
| CBR-AASHTO and USCS | $400/each|
| Soil classification | $150/each|
| Sieve analysis | $60/each |
| Atterberg limits | $90/each |
| Hydrometer | $180/each|
| Consolidation | |
| Compression curve | $550/each|
| Compression curve and rebound (1 cycle) | $675/each |
| Compression curve and rebound (2 cycle) | $750/each |
| Direct Shear (minimum of three points) | $750/set |
| Triaxial | |
| Consolidated drained (CD) | $540/point|
| Consolidated undrained (CU) | |
| With pore pressure measure | $400/point|
| Without measuring pore pressure | $350/point|
| Unconsolidated Undrained (UU)| $175/each |
| Unconfined compression | $90/each |
| Laboratory tests (continued) | |
|-----------------------------|----------|
| Rock core test | |
| Tension | $130/each|
| Compression | $75/each |
| PVC Potential volume change | $125/each|
| Infiltration test | $50/each |
| Compressive Strength tests on concrete cores | $75/each |
| Compaction Curves | $150/each|
| Packer Testing (rental) | $150/day |
16. Professional Services
Engineering time (includes site visits, evaluation, report preparation, meetings and final plans reviews)
- Field Engineer: $90/hour
- Senior Geotechnical Engineer: $125/hour
Geologist time (includes time for calculations and drawings)
- Junior: $65/hour
- Senior: $85/hour
Surveying services: $600/day
Diver Services: $775/day
17. Safety
- Flagman: $10/hour
- Lightning: $70/day
- MOT - Maintenance of traffic on principal roads (includes arrows and drums): $325/boring
- Portable Toilettes: $70/day
18. Permit Costs
DRNA, Excavation Center, Municipality, other Agencies: $50 lump sum
| Reference Number | Km | Intersection | Direction (N,E,S,W) | Sign Code | Sign Description | Size | Overhead Structure Type | Remarks |
|------------------|----|--------------|--------------------|-----------|-----------------|------|------------------------|---------|
| 1 | | | | | | | | |
| 2 | | | | | | | | |
| 3 | | | | | | | | |
| 4 | | | | | | | | |
| 5 | | | | | | | | |
| 6 | | | | | | | | |
| 7 | | | | | | | | |
| 8 | | | | | | | | |
| 9 | | | | | | | | |
| 10 | | | | | | | | |
| 11 | | | | | | | | |
| 12 | | | | | | | | |
| 13 | | | | | | | | |
| 14 | | | | | | | | |
| 15 | | | | | | | | | |
Preference-Based Policy Learning
Riad Akrou, Marc Schoenauer, Michèle Sebag
To cite this version:
Riad Akrou, Marc Schoenauer, Michèle Sebag. Preference-Based Policy Learning. Dimitrios Gunopulos, Thomas Hofmann, Donato Malerba, Michalis Vazirgiannis. European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, Sep 2011, Athènes, Greece. Springer Verlag, 6911, pp.12-27, 2011, LNAI. <inria-00625001>
Preference-based Policy Learning
Riad Akrou, Marc Schoenauer, and Michele Sebag
TAO
CNRS – INRIA – Université Paris-Sud
email@example.com
Abstract. Many machine learning approaches in robotics, based on reinforcement learning, inverse optimal control or direct policy learning, critically rely on robot simulators. This paper investigates a simulator-free direct policy learning, called Preference-based Policy Learning (PPL). PPL iterates a four-step process: the robot demonstrates a candidate policy; the expert ranks this policy comparatively to other ones according to her preferences; these preferences are used to learn a policy return estimate; the robot uses the policy return estimate to build new candidate policies, and the process is iterated until the desired behavior is obtained. PPL requires a good representation of the policy search space be available, enabling one to learn accurate policy return estimates and limiting the human ranking effort needed to yield a good policy. Furthermore, this representation cannot use informed features (e.g., how far the robot is from its target) due to the simulator-free setting. As a second contribution, this paper proposes a representation based on the agnostic exploitation of the robotic log.
The convergence of PPL is analytically studied and its experimental validation on two problems, involving a single robot in a maze and two interacting robots, is presented.
1 Introduction
Many machine learning approaches in robotics, ranging from direct policy learning [3] and learning by imitation [7] to reinforcement learning [20] and inverse optimal control [1, 14], rely on the use of simulators used as full motion and world model of the robot. Naturally, the actual performance of the learned policies depends on the accuracy and calibration of the simulator (see [22]).
The question investigated in this paper is whether robotic policy learning can be achieved in a simulator-free setting, while keeping the human labor and computational costs within reasonable limits. This question is motivated by machine learning application to swarm robotics [24, 29, 18]. Swarm robotics aims at robust and reliable robotic systems through the interaction of a large number of small-scale, possibly unreliable, robot entities. Within this framework, the use of simulators suffers from two limitations. Firstly, the simulation cost increases more than linearly with the size of the swarm. Secondly, robot entities might differ among themselves due to manufacturing tolerance, entailing a large variance
among the results of physical experiments and severely limiting the accuracy of simulated experiments.
The presented approach, inspired from both energy-based learning [21] and learning-to-rank [12, 4], is called *Preference-based policy learning* (PPL). PPL proceeds by iterating a 4-step process:
- In the **demonstration** phase, the robot demonstrates a candidate policy.
- In the **teaching** or feedback phase, the expert ranks this policy compared to archived policies, based on her agenda and prior knowledge.
- In the **learning** phase, a policy return estimate (energy criterion) is learned based on the expert ranking, using an embedded learning-to-rank algorithm. A key issue concerns the choice of the policy representation space (see below).
- In the **self-training** phase, new policies are generated, using an adaptive trade-off between the policy return estimate and the policy diversity w.r.t. the archive. A candidate policy is selected to be demonstrated and the process is iterated. PPL initialization proceeds by demonstrating two policies, which are ordered by the expert.
A first contribution of the PPL approach compared to inverse reinforcement learning [1, 14] is that it does not require the expert to know how to solve the task and to demonstrate a perfect behavior (see also [27]); the expert is only required to know whether some behavior is more able to reach the goal than some other one. Compared to learning by demonstration [7], the demonstrated trajectories are feasible by construction (whereas the teacher and the robot usually have different degrees of freedom). Compared to policy gradient approaches [20], the continued interaction with the expert offers some means to detect and avoid the convergence to local optima, through the expert’s feedback.
In counterpart, PPL faces one main challenge. The human ranking effort needed to yield a competent policy must be limited; the sample complexity must be of the order of a few dozen to a couple hundred. Therefore the policy return estimate must enable fast progress in the policy space, which requires the policy representation space to be carefully designed (a general concern, as noted by [1]). On the other hand, the simulator-free setting precludes the use of any informed features such as the robot distance to obstacles or other robots. The second contribution of the paper is an original representation of the policy space referred to as *behavioral representation* (BvR), built as follows. Each policy demonstration generates a robotic log, reporting the sensor and actuator values observed at each time step. The robotic logs are incrementally processed using $\epsilon$-clustering [11]; each cluster is viewed as a sensori-motor state (sms). A demonstration can thus be represented by a histogram, associating to each sensori-motor state the fraction of time spent in this sms. The BvR representation complies with a resource-constrained framework and it is agnostic w.r.t. the policy return and the environment. Although the number of states exponentially increases with the clustering precision $\epsilon$ and the intrinsic dimension of the sensori-motor space, it can however be controlled to some extent as $\epsilon$ is set by the expert. BvR refinement is a perspective for further research (section 5).
PPL is analytically studied on the artificial RiverSwim problem [25] and a convergence proof is given. PPL is also experimentally validated on two problems relevant to swarm robotics. The first one concerns the reaching of a target location in a maze. The difficulty is due to perceptual aliasing (non-Markovian environment) as different locations are perceived to be similar due to the limited infra-red sensors of the robot. The second problem involves two interacting robots. The task is to yield a synchronized behavior of the two robots, controlled with the same policy. The difficulty is again due to the fact that sensors hardly enable a robot to discriminate between its relative and an obstacle.
The paper is organized as follows. Section 2 discusses related work with regards to resource-constrained robotics. Section 3 gives an overview of the PPL approach and presents a convergence study. Section 4 describes the experimental setting and reports on the results of the approach. The paper concludes with some perspectives for further research.
2 Related Work
Policy learning is most generally tackled as an optimization problem. The main issues regard the search space (policy or state/action value function), the definition of the objective, and the exploration of the search space, aimed at optimizing the objective function. The infinite horizon case is considered throughout this section for notational convenience.
Among the most studied approaches is reinforcement learning (RL) [26], for its guarantees of global optimality. The background notations involve a state space $S$, an action space $A$, a reward function $r$ defined on the state space $r : S \mapsto \mathbb{R}$, and a transition function $p(s, a, s')$, expressing the probability of arriving in state $s'$ on making action $a$ in state $s$ under the Markov property. A policy $\pi : (S, A) \mapsto \mathbb{R}$ maps each state in $S$ on some action in $A$ with a given probability. The policy return is defined as the expectation of cumulative reward gathered along time, conditionally to the initial state $s_0$. Denoting $a_h \sim \pi(s_h, a)$ the action selected by $\pi$ in state $s_h$, $s_{h+1} \sim p(s_h, a_h, s)$ the state of the robot at step $h + 1$ conditionally to being in state $s_h$ and selecting action $a_h$ at step $h$, and $r_{h+1}$ the reward collected in $s_{h+1}$, then the policy return is
$$J(\pi; s_0) = \mathbb{E}_{\pi, s_0} \left[ \sum_{h=0}^{\infty} \gamma^h r_h \right]$$
where $\gamma < 1$ is a discount factor enforcing the boundedness of the return, and favoring the reaping of rewards as early as possibly in the robot lifetime. The so-called Q-value function $Q_\pi(s, a)$ estimates the expectation of the cumulative reward gathered by policy $\pi$ after selecting action $a$ in state $s$. As estimation errors cumulate along time, the RL scalability is limited with respect to the size of the state and action spaces. Moreover in order to enforce the Markov property, the description of state $s$ must provide any information about the robot past trajectory relevant to its further decisions — thereby increasing the size of the state space.
As RL suffers from both the scalability issue and the difficulty of defining \textit{a priori} a reward function conducive to the task at hand [19], a way to “seed” the search with a good solution was sought, referred to as \textbf{inverse optimal control}. One possibility, pioneered by [3] under the name of behavioral cloning and further developed by [7] under the name of programming by demonstration relies on the exploitation of the expert’s traces. These traces can be used to turn policy learning into a supervised learning problem [7]. Another possibility is to use the expert’s traces to learn a reward function, along the so-called inverse reinforcement learning approach [19]. The sought reward function should admit the expert policy $\pi^*$ (as evidenced from his traces) as solution, and further enforce some return margin between the actual expert actions and other actions in the same state, in order to avoid indeterminacy [1] (since a constant reward function would have any policy as an optimal solution). The search space is carefully designed in both cases. In the former approach, an “imitation metric” is used to account for the differences between the expert and robot motor spaces. In the latter one, a few informed features (e.g. the average speed of the vehicle, the number of times the car deviates from the road) are used together with their desired sign (the speed should be maximized, the number of deviations should be minimized) and the policy return is sought as a weighted sum of the feature expectations. A min-max relaxation thereof is proposed by [27], maximizing the minimum policy return over all weight vectors. Yet another approach, inverse optimal heuristic control [14] proposes a hybrid approach between behavioral cloning and inverse optimal control, addressing the low-dimensionality restrictions on optimal inverse control. Formally, an energy-based action selection model is proposed using a Gibbs model which combines the reward function (as in inverse optimal control) and a logistic regression function (learned using behavioral cloning from the available trajectories).
The main limitation of inverse optimal control lies in the way it is seeded: the expert’s traces can hardly be considered to be optimal in the general case, even more so if the expert and the robot live in different sensori-motor spaces. While [27] gets rid of the expert’s influence, its minmax approach yields very conservative policies, as the relative importance of the features is unknown.
A main third approach aims at \textbf{direct policy learning}, using policy gradient methods [20] or global optimization methods [28]. Direct policy learning most usually assumes a parametric policy space $\Theta$; policy learning aims at finding the optimal $\theta^*$ parameter in the sense of a policy return function $J$:
$$\text{Find } \theta^* = \arg \max_{\theta \in \Theta} \{ J(\theta), \theta \in \Theta \}$$
Depending on $J$, three cases are distinguished. The first case is when $J$ is analytically known on a continuous policy space $\Theta \subset \mathbb{R}^D$. It then comes naturally to use a gradient-based optimization approach, gradually moving the current policy $\theta_t$ along the gradient $\nabla J$ [20] ($\theta_{t+1} = \theta_t + \alpha_t \nabla_\theta J(\theta_t)$). The main issues concern the adjustment of $\alpha_t$, the possible use of the inverse Hessian, and the rate of convergence toward a (local) optimum. A second case is when $J$ is only known as a computable function, e.g. when learning a Tetris policy [28]. In such
cases, a wide scope optimization method such as Cross-Entropy Method [8] or population-based stochastic optimization [23] must be used. In the third case, e.g. [7, 14], $J$ is to be learned from the available evidence.
This brief and by no means exhaustive review of related work suggests that the search for an optimal policy relies on three main components. The first one clearly is the expert’s support, usually provided through a reward function directly defined on the state space, or inferred from the expert’s demonstrations. The second one is a simulator or forward model, enabling the robot to infer the reward function in the IOC case, and more generally to consider the future consequences of its exploratory actions; the lack of any planning component seemingly is the main cause for the limitations of behavioral cloning, forbidding the generalization of the provided demonstrations. The third one is a low-dimensionality representation of the state or policy search spaces.
With regards to swarm robotics however, a good quality simulator or forward model is unlikely to be available anytime soon, for the reasons discussed in the introduction. The expert’s support is also bound to be limited, for she can hardly impersonate the swarm robot on the one hand, and because the desired behavior is usually unknown on the other hand\footnote{Swarm policy learning can be viewed as an inverse problem; one is left with the problem of e.g. designing the ant behavior in such a way that the ant colony achieves a given task.}. The proposed approach will thus be based on a less demanding interaction between the policy learning process and the expert, only assuming that the former can \textit{demonstrate} policies and that the latter can emit \textit{preferences}, telling a more appropriate behavior from a lesser one.
\section{Preference-based Policy Learning}
Only deterministic parameterized policies with finite time horizon $H$ will be considered in this section, where $H$ is the number of time steps each candidate policy is demonstrated.
\subsection{PPL Overview and Notations}
Policy $\pi_\theta$, defined from its parameter $\theta \in \Theta \subseteq \mathbb{R}^D$, stands for a deterministic mapping from the state space $S$ onto the action space $A$. A \textit{behavioral representation} is defined from the demonstrations (section 3.2) and used to learn the policy return estimate. PPL proceeds by maintaining a policy and constraint archive, respectively denoted $\Pi_t$ and $C_t$ along a four-step process. At step $t$,
- A new policy $\pi_t$ is demonstrated by the robot and added to the archive ($\Pi_t = \Pi_{t-1} \cup \{\pi_t\}$);
- $\pi_t$ is ranked by the expert w.r.t. the other policies in $\Pi_t$, and the set $C_t$ is accordingly updated from $C_{t-1}$;
- The policy return estimate $J_t$ is built from all constraints in $C_t$ (section 3.3);
• New policies are generated; candidate policy $\pi_{t+1}$ is selected using an adaptive trade-off between $J_t$ and an empirical diversity term (section 3.4), and the process is iterated.
PPL is initialized from two policies with specific properties (section 3.5), which are ordered by the expert; the policy and constraint archives are initialized accordingly. After detailing all PPL components, an analytical convergence study of PPL is presented in section 3.6.
### 3.2 The Behavioral Representation
In many parametric settings, e.g. when policy $\pi$ is characterized as the weight vector $\theta \in \Theta$ of a neural net with fixed architecture, the Euclidean metric on $\Theta$ is poorly informative of the policy behavior. As the sought policy return estimate is meant to assess policy behavior, it thus appears inappropriate to learn $J_\pi$ as a mapping on $\Theta$. The proposed *behavioral representation* (BvR) fits the robot bounded memory and computational resources by mapping each policy onto a real valued vector $(\mu : \pi \mapsto \mu_\pi \in \mathbb{R}^{d'})$ as follows. A robot equipped with a policy $\pi$ can be thought of as a data stream generator; the sensori-motor data stream SMS($\pi$) consists of the robot sensor and actuator values recorded at each time step. It thus comes naturally to describe policy $\pi$ through compressing SMS($\pi$), using an embedded online clustering algorithm. Let SMS($\pi$) be given as $\{x_1, \ldots x_H\}$ with $x_h \in \mathbb{R}^b$, where $b$ is the number of sensors and actuators of the robot and $x_h$ denotes the concatenation of the sensory and motor data of the robot at time step $h$. An $\epsilon$-clustering algorithm ($\epsilon > 0$) [11] is used to incrementally define a set $\mathcal{S}$ of sensori-motor states $s_i$ from all SMS($\pi$) data streams. For each considered policy $\pi$, in each time step, $x_h$ is compared to all elements of $\mathcal{S}$, and it is added to $\mathcal{S}$ if $min \{||x_h - s_i||_\infty, \ s_i \in \mathcal{S}\} > \epsilon$. Let $n_t$ denote the number of states at time $t$. BvR is defined as follows:
$$\text{Given } \text{SMS}(\pi) = \{x_1, \ldots x_H\} \text{ and } \mathcal{S} = \{s_1, \ldots s_{n_t}\}$$
$$\mu : \pi \mapsto \mu_\pi \in [0, 1]^{n_t} \quad \mu_\pi[i] = \frac{|t|}{H} \left[ t \ s.t. \ ||x_h - s_i||_\infty < \epsilon \right]$$
BvR thus differs from the feature counts used in [1, 14] as features are learned from policy trajectories as opposed to being defined from prior knowledge. Compared to the discriminant policy representation built from the set of policy demonstrations [17], the main two differences are that BvR is independent of the policy return estimate, and that it is built online as new policies explore new regions of the sensori-motor space. As the number of states $n_t$ and thus BvR dimension increases along time, BvR consistency follows from the fact $\mu_\pi \in [0, 1]^{n_t}$ is naturally mapped onto $(\mu_\pi, 0)$ in $[0, 1]^{n_t + 1}$ (the $t$-th coordinate of $\mu_\pi$ is 0 if $s_i$ was not discovered at the time $\pi$ was demonstrated).
The price to pay for the representation consistency is that the number of states exponentially increases with precision parameter $\epsilon$; on the other hand, it increases like $e^{b'}$ where $b'$ is the intrinsic dimension of the sensori-motor data. A second limitation of BvR is to be subject to the initialization noise, meant as the initial location of the robot when the policy is launched. The impact of
the initial conditions however can be mitigated by discarding the states visited during the burn-in period of the policy.
### 3.3 Policy Return Estimate Learning
For the sake of notational simplicity, $\mu_i$ will be used instead of $\mu_{\pi_i}$ when no confusion is to fear. Let the policy archive $H_t$ be given as $\{\mu_1, \ldots \mu_t\}$ at step $t$, assuming with no loss of generality that $\mu_i$s are ordered after the expert preferences. Constraint archive $C_t$ thus includes the set of $\frac{t(t-1)}{2}$ ordering constraints $\mu_i \succ \mu_j$ for $i > j$.
Using a standard constrained convex optimization formulation [4, 13], the policy return estimate $J_t$ is sought as a linear mapping $J_t(\mu) = \langle w_t, \mu \rangle$ with $w_t \in \mathbb{R}^{n_t}$ solution of (P):
$$
(P) \begin{cases}
\text{Minimize } & \frac{1}{2} ||w||^2 + C \sum_{i,j=1,i>j}^{t} \xi_{i,j} \\
\text{subject to } & (\langle w, \mu_i \rangle - \langle w, \mu_j \rangle) \geq 1 - \xi_{i,j}) \text{ and } (\xi_{i,j} \geq 0) \text{ for all } i > j
\end{cases}
$$
The policy return estimate $J_t$ features some good properties. Firstly, it is consistent despite the fact that the dimension $n_t$ of the state space might increase with $t$; after the same arguments as above, the $w_t$ coordinate related to a state which has not yet been discovered is set to 0. By construction, it is independent on the policy parameterization and can be transferred among different policy spaces; likewise, it does not involve any information but the information available to the robot itself; therefore it can be computed on-board and provides the robot with a self-assessment.
Finally, although $J_t$ is defined at the global policy level, $w_t$ provides some valuation of the sensori-motor states. If a high positive weight $w_t[i]$ is associated to the $i$-th sensori-motor state $s_i$, this state is considered to significantly and positively contribute to the quality of a policy, *comparatively to the policies considered so far*. Learning $J_t$ thus significantly differs from learning the optimal RL value function $V^*$. By definition, $V^*(s_i)$ estimates the maximum expected cumulative reward ahead of state $s_i$, which can only increase as better policies are discovered. In contrast, $w_t[i]$ reflects the fact that visiting state $s_i$ is conducive to discovering better policies, comparatively to policies viewed so far by the expert. In particular, $w_t[i]$ can *increase or decrease* along $t$; some states might be considered as highly beneficial in the early stages of the robot training, and discarded later on.
### 3.4 New Policy Generation
The policy generation step can be thought of in terms of self-training. The generation of new policies relies on black-box optimization; expected-global improvement methods [6] and gradient methods [20] are not applicable since the policy return estimate is not defined on the parametric representation space $\Theta$. New policies are thus generated using a stochastic optimization method, more precisely a $(1+\lambda)$-ES algorithm [2]. Formally, a Gaussian distribution $\mathcal{N}(\theta_c, \Sigma)$
on $\Theta$ is maintained, updating the center $\theta_*$ of the distribution and the covariance matrix $\Sigma$ after the parameter $\theta$ of the current best policy. For $n_{\theta}$ iterations, $\lambda$ policies $\pi$ are drawn after $\mathcal{N}(\theta_*, \Sigma)$, they are executed in order to compute their behavioral description $\mu_\pi$, and the best one after an optimisation criterion $\mathcal{F}$ (see below) is used to update $\mathcal{N}(\theta_*, \Sigma)$. After $n_{\theta}$ iterations of the policy generation step, the best policy after $\mathcal{F}$ denoted $\pi_{t+1}$ is selected and demonstrated to the expert, and the whole PPL process is iterated.
The criterion $\mathcal{F}$ used to select the best policy out of the current $\lambda$ policies is meant to enforce some trade-off between the exploration of the policy space and the exploitation of the current policy return estimate $J_t$ (note that the discovery of new sensori-motor states is worthless after $J_t$ since their weight after $w_t$ is 0). Taking inspiration from [2], $\mathcal{F}$ is defined as the sum of the policy return estimate $J_t$ and a weighted exploration term $E_t$, measuring the diversity $\Delta$ of the policy w.r.t the policy archive $\Pi_t$:
$$
\begin{align*}
\text{Maximize } & \mathcal{F}(\mu) = J_t(\mu) + \alpha_t E_t(\mu) \quad \alpha_t > 0 \\
\text{with } & E_t(\mu) = \min \{\Delta(\mu, \mu_u), \mu_u \in \Pi_t\} \\
\alpha_t = & \begin{cases}
c.\alpha_{t-1} & \text{if } \pi_t \succ \pi_{t-1} \\
\frac{1}{c^{1/p}}\alpha_{t-1} & \text{otherwise}
\end{cases} \quad c > 1
\end{align*}
$$
Parameter $\alpha_t$ controls the **exploration vs exploitation tradeoff**, accounting for the fact that both the policy distribution and the objective function $J_t$ are non stationary. Accordingly, $\alpha_t$ is increased or decreased by comparing the empirical success rate (whether $\pi_t$ improves on all policies in the archive) with the expected success rate of a reference function (usually the sphere function [2]). When the empirical success rate is above (respectively below) the reference one, $\alpha_t$ is increased (resp. decreased). Parameter $p$, empirically adjusted, is used to guarantee that $p$ failures cancel out one success and bring $\alpha_t$ back to its original value.
The **diversity function** $\Delta(\mu, \mu')$ is defined as follows. Let $a, b, c$ respectively denote the number of states visited by $\mu$ only, by $\mu'$ only, and by both $\mu$ and $\mu'$. Then $\Delta(\mu, \mu')$ is set to $\frac{a+b-c}{\sqrt{a+c}\sqrt{b+c}}$; it computes and normalizes the symmetric difference minus the intersection of the two sets of states visited by $\mu$ and $\mu'$.
### 3.5 Initialization
The PPL initialization is challenging for policy behaviors corresponding to uniformly drawn $\theta$ usually are quite uninteresting and therefore hard to rank. The first two policies are thus generated as follows. Given a set $P_0$ of randomly generated policies, $\pi_1$ is selected as the policy in $P_0$ with maximal information quantity ($J_0(\mu) = -\sum_{i=1}^n \mu[i] \log \mu[i]$). Policy $\pi_2$ is the one in $P_0$ with maximum diversity w.r.t. $\pi_1$ ($\pi_2 = \argmax \{\Delta(\mu, \pi_1), \mu \in P_0\}$). The rationale for this initialization is that $\pi_1$ should experiment as many distinct sensorimotor states as possible; and $\pi_2$ should experiment as many sensorimotor states different from that of $\pi_1$ as possible, to facilitate the expert ordering and yield an informative policy return estimate $J_2$. Admittedly, the use of the information quantity and more
generally BvR only make sense if the robot faces a sufficiently rich environment. In an empty environment, i.e. when there is nothing to see, the robot can only visit a single sensori-motor state.
### 3.6 Convergence study
PPL convergence is analytically studied and proved for the artificial RiverSwim problem [25]. This problem involves $N$ states and two actions, going *left* or *right*; starting from the leftmost state, the goal is to reach the rightmost state (Fig. 1). While the policy parametric representation space is $\{ \text{left, right} \}^N$, the policy $\pi$ behavior is entirely characterized from the leftmost state $i$ where it goes left, denoted $\ell(\pi)$. In the following, we only consider the boolean BvR, with $\mu_\pi = \{1, 1, ..., 1, 0...0\}$ and $\ell(\pi)$ the index of the rightmost 1. For notational simplicity, let $\mu_\pi$ and $\mu_{\pi'}$ be noted $\mu$ and $\mu'$ in the following. By definition, the expert prefers the one out of $\mu$ and $\mu'$ which goes farther on the right ($\mu \succ \mu'$ iff $\ell(\mu) > \ell(\mu')$).
In the RiverSwim setting, problem (P) can be solved analytically (section 3.3). Let archive $H_t$ be given as $\{\mu_1 \ldots \mu_t\}$ and assume wlog that $\mu_i$s are ordered by increasing value of $\ell(\mu_i)$. It comes immediately that constraints related to adjacent policies $(\langle w, \mu_{i+1} \rangle - \langle w, \mu_i \rangle \geq 1 - \xi_{i,i+1}$ for $1 \leq i < t - 1$) entail all other constraints, and can be all satisfied by setting $\sum_{k=\ell(\mu_i)+1}^{\ell(\mu_{i+1})} w[k] = 1$; slack variables $\xi_{i,j}$ thus are 0 at the optimum.
Problem (P) can thus be decomposed into $t - 1$ independent problems involving disjoint subsets of indices $[\ell(\mu_i), \ell(\mu_{i+1})]$; its solution $w$ is given as:
$$w[k] = \begin{cases}
0 & \text{if } k < \ell(\mu_1) \text{ or } k > \ell(\mu_t) \\
\frac{1}{\ell(\mu_{i+1}) - \ell(\mu_i)} & \text{if } \ell(\mu_i) < k \leq \ell(\mu_{i+1})
\end{cases}$$
**Proposition 1:** In the RiverSwim setting, $J_t(\mu) = \langle w_t, \mu \rangle$ where $w_t$ satisfies: $w_t[k] = 0$ if $k \leq \ell(\mu_1)$ or $k > \ell(\mu_t)$; $w_t[k] > \frac{1}{N}$ if $\ell(\mu_1) < k \leq \ell(\mu_t)$; $\sum_{k=1}^{N} w_t[k] = t$. □
Policy generation (section 3.4) considers both the current $J_t$ and the diversity $E_t$ respective to the policy archive $\Pi_t$, given as $E_t(\mu) = \min\{\Delta(\mu, \mu_u), \mu_u \in \Pi_t\}$. In the RiverSwim setting, it comes:
$$\Delta(\mu, \mu_u) = \frac{|\ell(\mu) - \ell(\mu_u)| - \min(\ell(\mu), \ell(\mu_u))}{\sqrt{\ell(\mu)\ell(\mu_u)}}$$
For $i < j$, $\frac{|i-j|-\min(i,j)}{\sqrt{ij}} = \sqrt{\frac{j}{i}} - 2\sqrt{\frac{i}{j}}$. It follows that the function $i \mapsto \frac{\sqrt{\ell(\mu_i)} - i - \min(\ell(\mu), i)}{\sqrt{\ell(\mu_i)}}$ is strictly decreasing on $[1, \ell(\mu)]$ and strictly increasing on $[\ell(\mu), +\infty)$. It reaches its minimum value of $-1$ for $i = \ell(\mu)$. As a consequence, for any policy $\mu$ in the archive, $E_t(\mu) = -1$.
Let $\mu_t$ and $\mu^*$ respectively denote the best policy in the archive $\Pi_t$ and the policy that goes exactly one step further right. We will now prove the following result:
**Proposition 2.** The probability that $\mu^*$ is generated at step $t$ is bounded from below by $\frac{1}{eN}$. Furthermore, after $\mu^*$ has been generated, it will be selected according to the selection criterion $F_t$ (section 3.4), and demonstrated to the expert.
**Proof**
*Generation step:* Considering the boolean parametric representation space $\{0, 1\}^N$, the generation of new policies proceeds using the standard bitflip mutation, flipping each bit of the current $\mu_t$ with probability $\frac{1}{N}$. The probability to generate $\mu^*$ from $\mu_t$ is lower-bounded by $\frac{1}{N}(1 - \frac{1}{N})^{\ell(\mu_t)}$ (flip bit $\ell(\mu_t) + 1$ and do not flip any bit before that). Furthermore, $(1 - \frac{1}{N})^{\ell(\mu_t)} > (1 - \frac{1}{N})^{N-1} > \frac{1}{e}$ and hence the probability to generate $\mu^*$ from $\mu_t$ is lower-bounded by $\frac{1}{eN}$.
*Selection step.* As shown above, $E_t(\mu^*) > -1$ and $E_t(\mu) = -1$ for all $\mu \in \Pi_t$. As the candidate policy selection is based on the maximization of $F_t(\mu) = \langle w_t, \mu \rangle + \alpha_t E_t(\mu)$, and using $\langle w_t, \mu_t \rangle = \langle w_t, \mu_{t+1} \rangle = t$, it follows that $F_t(\mu_t) < F_t(\mu^*)$, and more generally, that $F_t(\mu) < F_t(\mu^*)$ for all $\mu \in \Pi_t$. Consider now a policy $\mu$ that is not in the archive though $\ell(\mu) < t$ (the archive does not need to contain all possible policies). From Proposition 1, it follows that $\langle w, \mu \rangle < t - \frac{1}{N}$. Because $E_t(\mu)$ is bounded, there exists a sufficiently small $\alpha_t$ such that $F_t(\mu) = \langle w_t, \mu \rangle + \alpha_t E_t(\mu) < F_t(\mu^*)$. □
Furthermore, thanks to the monotonicity of $E_t(\mu)$ w.r.t. $\ell(\mu)$, one has $F_t(\mu_{t+i}) < F_t(\mu_{t+j})$ for all $i < j$ where $\ell(\mu_{t+i}) = \ell(\mu_t) + i$. Better RiverSwim policies will thus be selected along the policy generation and selection step.
## 4 Experimental Validation
This section reports on the experimental validation of the PPL approach. For the sake of reproducibility, all reported results have been obtained using the publicly available simulator Roborobo [5].
4.1 Experiment Goals and Experimental Setting
The experiments are meant to answer three main questions. The first one concerns the feasibility of PPL compared to baseline approaches. The use as surrogate optimization objective of the policy return estimate learned from the expert preferences, is compared to the direct use of the expert preferences. The performance indicator is the speed-up in terms of sample complexity (number of calls to the expert), needed to reach a reasonable level of performance. The second question regards the merits of the behavioral representation comparatively to the parametric one. More specifically, the point is to know whether BvR can enforce an effective trade-off between the number of sensori-motor states and the perceptual aliasing, enabling an accurate policy return estimate to be built from few orderings. A third question regards the merits of the exploration term used in the policy generation (self-training phase, section 3.4), and the entropy-based initialization process (section 3.5), and how they contribute to the overall performance of PPL. Three experiments are considered to answer these questions. The first one is an artificial problem which can be viewed as a 2D version of the RiverSwim [25]. The second experiment, inspired from [16], involves a robot in a maze and the goal is to reach a given location. The third experiment involves two robots and the goal is to enforce a coordinated exploration of the arena by the two robots. In all experiments the robotic architecture is a Cortex-M3 with 8 infra-red (IR) sensors and two motors respectively controlling the rotational and translation speed. The IR range is 6 cm; the surface of the arena is 6.4 square meters.
Every policy is implemented as a 1-hidden-layer feed-forward neural net, using the 8 IR sensors (and a bias) as input, with 10 neurons on the hidden layer, and the two motor commands as output ($\Theta = \mathbb{R}^{121}$). A random policy is built by uniformly selecting each weight in $[-1, 1]$. BvR is built using norm $L_\infty$ and $\epsilon = 0.45$; the number of sensori-motor states, thus the BvR dimension, is below 1,000. The policy return estimate is learned using SVMrank library [13] with linear kernel and default parameter $C = 1$ (section 3.3). The optimization algorithm used in the policy generation (section 3.4) and initialization (section 3.5) is a $(1+\lambda)$-ES [23] with $\lambda = 7$; the variance of the Gaussian distribution is initialized to .3, increased by a factor of 1.5 upon any improvement and decreased by a factor of $1.5^{-1/4}$ otherwise. The $(1+\lambda)$-ES is used for $n_g$ iterations in each policy generation step, with $n_g = 10$ in the maze (respectively $n_g = 5$ in the two-robot) experiment; it is used for 20 iterations in the initialization step. All presented results are averaged over 41 independent runs. Five algorithms are compared: PPL$_{BvR}$ with entropy-based initialization and exploration term, where the policy return estimate relies on the behavioral representation; PPL$_{param}$, which only differs from PPL$_{BvR}$ as the policy return estimate is learned using the parametric representation; Expert-Only, where the policy return estimate is replaced by the true expert preferences; Novelty-Search [16] which can be viewed
as an exploration-only approach\footnote{In each time step the candidate policy with highest average diversity compared to its $k$-nearest neighbors in the policy archive is selected, with $k = 1$ for the 2D RiverSwim, and $k = 5$ for both other experiments.}, and PPL$_{w/o}$, which differs from PPL$_{ByR}$ as it uses a random uniform initialization.
### 4.2 2D RiverSwim
In this $4 \times 4$ grid world, the robot starts in the lower left corner of the grid; the sought behavior is to reach the upper right corner and to stay there. The expert preference goes toward the policy going closer to the goal state, or reaching earlier the goal state. The action space includes the four move directions with deterministic transitions (the robot stays in the same position upon marching to the wall). For a time horizon $H = 16$, the ByR includes only 6334 distinguishable policies (to be compared with the actual $4^{16}$ distinct policies). The small size of this grid world allows one to enumeratively optimize the policy return estimate, with and without the exploration term. Fig. 2 (left) displays the true policy return vs the number of calls to the expert. In this artificial problem, *Novelty-Search* outperforms PPL and discovers an optimal policy in the 20-th step. PPL discovers an optimal policy at circa the 30-th step. The exploration term plays an important role in PPL performance: Fig. 2 (right) displays the average weight $\alpha_t$ of the exploration term. In this small-size problem, a mere exploration strategy is sufficient and typically *Novelty-Search* discovers two optimal policies in the first 100 steps on average; meanwhile, PPL discovers circa 25 optimal policies in the first 100 steps on average.
**Fig. 2.** 2D RiverSwim: Performance of PPL and *Novelty-Search*, averaged over 41 runs with std. dev (left); average weight of the exploration term $\alpha_t$ in PPL (right).
### 4.3 Reaching the End of the Maze
In this experiment inspired from [15], the goal is to traverse the maze and reach the green zone when starting in the opposite initial position (Fig. 3-left). The
time horizon $H$ is 2,000; an oracle robot needs circa 1,000 moves to reach the target location. As in the RiverSwim problem, the expert preference goes toward the policy going closer to the goal state, or reaching earlier the goal state; the comparison leads to a tie if the difference is below 50 distance units or 50 time steps. The “true” policy return is the remaining time when it first reaches the green zone, if it reaches it, minus the min distance of the trajectory to the green zone, otherwise. The main difficulty in this maze problem is the severe perceptual aliasing: all locations situated far from the walls look alike due to the IR sensor limitations.
**Fig. 3.** The maze experiment: the arena (left) and best performances averaged out of 41 runs.
This experiment supports the feasibility of the PPL scheme. A speed-up factor of circa 3 is observed compared to *Expert-Only*; on average PPL reaches the green zone after demonstrating 39 policies to the expert, whereas *Expert-Only* needs 140 demonstrations (Fig. 4, left). This speed-up is explained as PPL filters out unpromising policies. *Novelty-Search* performs poorly on this problem comparatively to PPL and even comparatively to *Expert-Only*, which suggests that *Novelty-Search* might not scale up well with the size of the policy space.
This experiment also establishes the merits of the behavioral representation. Fig. 4 (right) displays the average performance of PPL$_{B/o}$ and PPL$_{param}$ when no exploration term is used, to only assess the accuracy of the policy return estimate. As can be seen, learning is very limited when done in the parametric representation by PPL$_{param}$, resulting in no speedup compared to *Expert-Only* (Fig. 3, right). A third lesson is that the entropy-based initialization does not seem to bring any significant improvement, as PPL$_{B/o}$ and PPL$_{w/o}$ get same results after the very first steps (Fig. 3).
### 4.4 Synchronized Exploration
This experiment investigates the feasibility of two robots exploring an arena (Fig. 5, left) in a synchronous way, using the same policy. The expert preferences are
emulated by associating to a policy the number of distinct 25x25 tiles explored by any of the two robots conditionally to the fact that their distance is below 100 unit distances at the moment the tile is explored. Both robots start in (distinct) random positions. The difficulty of the task is that most policies wander in the arena, making it unlikely for the two robots to be anywhere close to one another at any point in time.
The experimental results (Fig. 5, right) mainly confirm the lessons drawn from the previous experiment. PPL improves on Expert-Only, with a speed-up circa 10, while Novelty-Search is slightly but significantly outperformed by Expert-Only. In the meanwhile, the entropy-based initialization does not make much of a difference after the first steps, and might even become counterproductive in the end.
Fig. 4. The maze experiment: PPL vs Expert-Only (left); Accuracy of parametric and behavioral policy return estimate (right).
Fig. 5. Synchronous exploration: the arena (left) and the best average performance out of 41 runs (right).
5 Conclusion and Perspectives
The presented preference-based policy learning demonstrates the feasibility of learning in a simulator-free setting. PPL can be seen as a variant of Inverse Optimal Control, with the difference that the demonstrations are done by the robot whereas the expert only provides some feedback (it’s better; it’s worse). PPL, without any assumptions on the level of expertise of the teacher, incrementally learns a policy return estimate. This learning is made possible through the original unsupervised behavioral representation BvR, based on the compression of the robot trajectories. The policy return estimate can be thought of as an embedded “system of values”, computable on-board. Further research will investigate the use of the policy return estimate within lifelong learning, e.g. to adjust the policy and compensate for the fatigue of the actuators. While the convergence of the approach can be established in the artificial RiverSwim framework [25], the experimental validation demonstrates that PPL can be used effectively to learn elementary behaviors involving one or two robots.
The main limitation of the approach comes from the exponential increase of the behavioral representation w.r.t. the granularity of the sensori-motor states. Further work will investigate how to refine BvR, specifically using quad-trees to describe and retrieve the sensori-motor states while adaptively adjusting their granularity. A second perspective is to adjust the length of the self-training phases, taking inspiration from online learning on a budget [9]. A third research avenue is to reconsider the expert preferences in a Multiple-Instance perspective [10]; clearly, what the expert likes/dislikes might be a fragment of the policy trajectory, more than the entire trajectory.
Acknowledgments. The first author is funded by FP7 European Project Symbrion, FET IP 216342, http://symbrion.eu/. This work is also partly funded by ANR Franco-Japanese project Sydimmalas ANR-08-BLAN-0178-01.
References
[1] P. Abbeel and A.Y. Ng. Apprenticeship Learning via Inverse Reinforcement Learning. In C.E. Brodley, editor, Proc. 21st Intl Conf. on Machine Learning (ICML-2004), volume 69 of ACM Intl Conf. Proc. Series, pages 1–. ACM, 2004.
[2] A. Auger. Convergence Results for the (1,λ)-SA-ES using the Theory of ϕ-irreducible Markov Chains. Theoretical Computer Science, 334(1-3):35–69, 2005.
[3] M. Bain and C. Sammut. A Framework for Behavioural Cloning. In K. Furukawa, D. Michie, and S. Muggleton, editors, Machine Intelligence, volume 15, pages 103–129. Oxford University Press, 1995.
[4] G. Bakir, T. Hofmann, B. Scholkopf, A.J. Smola, B. Taskar, and S.V.N. Vishwanathan. Machine Learning with Structured Outputs. MIT Press, 2006.
[5] N. Breedeker. http://www.hut.fi/~breedeker/robotor/.
[6] E. Brochu, N. de Freitas, and A. Ghosh. Active Preference Learning with Discrete Choice Data. In Proc. NIPS 20, pages 409–416, 2008.
[7] S. Calinon, F. Guenter, and A. Billard. On Learning, Representing and Generalizing a Task in a Humanoid Robot. IEEE Trans. on Systems, Man and Cybernetics,
[8] Pieter-Tjerk de Boer, Dirk P. Kroese, Shie Mannor, and Reuven Y. Rubinstein. A Tutorial on the Cross-Entropy Method. *Annals OR*, 134(1):19–67, 2005.
[9] O. Dekel, S. Shalev-Shwartz, and Y. Singer. The Forgetron: A Kernel-Based Perceptron on a Budget. *SIAM J. Comput.*, 37:1342–1372, 2008.
[10] T.G. Dietterich, R. Lathrop, and T. Lozano-Perez. Solving the Multiple-Instance Problem with Axis-Parallel Rectangles. *Artif. Intelligence*, 89 (1-2):31–71, 1997.
[11] R.O. Duda and P.E. Hart. *Pattern Classification and scene analysis*. John Wiley and sons, Menlo Park, CA, 1973.
[12] T. Joachims. A Support Vector Method for Multivariate Performance Measures. In L. De Raedt and S. Wrobel, editors, *Proc. 22nd ICML*, volume 119 of *ACM Intl Conf. Proc. Series*, pages 375–384. ACM, 2005.
[13] T. Joachims. Training Linear SVMs in Linear Time. In T. Eliaissi-Rad et al., editors, *Proc. 12th Intl Conf. KDDM*, pages 217–226. ACM, 2006.
[14] J. Zico Kolter, Pieter Abbeel, and Andrew Y. Ng. Hierarchical Apprenticeship Learning with Application to Quadruped Locomotion. In *Proc. NIPS 20*. MIT Press, 2007.
[15] J. Lehman and K. O. Stanley. Exploiting Open-Endedness to Solve Problems Through the Search for Novelty. In *Proc. Artificial Life XI*, pages 329–336, 2008.
[16] J. Lehman and K.O. Stanley. Exploiting Open-Endedness to Solve Problems through the Search for Novelty. In *Proc. ALife-08*. MIT Press, 2008.
[17] S. Levine, Z. Popovic, and V. Koltun. Feature Construction for Inverse Reinforcement Learning. In *Proc. NIPS 23*, pages 1342–1350. 2010.
[18] W. Liu and A. F. T. Winfield. Modeling and Optimization of Adaptive Foraging in Swarm Robotic Systems. *Intl J. Robotic Research*, 29(14):1743–1760, 2010.
[19] A.Y. Ng and S. Russell. Algorithms for Inverse Reinforcement Learning. In P. Langley, editor, *Proc. 17th ICML*, pages 663–670. Morgan Kaufmann, 2000.
[20] J. Peters and S. Schaal. Reinforcement Learning of Motor Skills with Policy Gradients. *Neural Networks*, 21(4):682–697, 2008.
[21] M.-A. Ranzato, C. S. Poulteny, S. Chopra, and Y. LeCun. Efficient Learning of Sparse Representations with an Energy-Based Model. In B. Schölkopf, J. C. Platt, and Th. Hoffman, editors, *Proc. NIPS 19*, pages 1137–1144. MIT Press, 2006.
[22] A. Stentz, J. Dragan, and A.Y. Ng. Robotic Grasping of Novel Objects using Vision. *Intl J. Robotics Research*, 1998.
[23] H.-P. Schwefel. *Numerical Optimization of Computer Models*. John Wiley & Sons, New-York, 1981. 1995 – 2nd edition.
[24] T. S. Stirling, S. Wischmann, and D. Floreano. Energy-efficient Indoor Search by Swarms of Simulated Flying Robots without Global Information. *Swarm Intelligence*, 4(2):117–143, 2010.
[25] A. L. Strehl, L. Li, E. Wiiewiora, J. Langford, and M. L. Littman. PAC Model-free Reinforcement Learning. In *Proc. ICML’06*, pages 881–888, 2006.
[26] R. Sutton and A. Barto. *Reinforcement Learning: An Introduction*. MIT Press, Cambridge, 1998.
[27] U. Syed and R. Schapire. A Game-Theoretic Approach to Apprenticeship Learning. In *Proc. NIPS 21*, pages 1449–1456. MIT Press, 2008.
[28] Christophe Thiery and Bruno Scherrer. Improvements on Learning Tetris with Cross Entropy. *IJCAI Journal*, 32(1):23–33, 2009.
[29] V. Trianni, S. Nolfi, and M. Dorigo. Cooperative Hole Avoidance in a Swarm-bot. *Robotics and Autonomous Systems*, 54(2):97–103, 2006. |
OPTICAL DEPHASING OF MOLECULAR DIMERS IN MIXED CRYSTALS:
PICOSECOND PHOTON ECHO EXPERIMENTS
F.G. PATTERSON, William L. WILSON, H.W.H. LEE and M.D. FAYER
Department of Chemistry, Stanford University, Stanford, California 94305, USA
Received 20 April 1984; in final form 2 July 1984
Picosecond photon echo experiments are used to examine optical dephasing of substitutional dimers and monomers of tetracene and pentacene in $p$-terphenyl host crystals. A comparison of experiments on tetracene and pentacene dimers permits the mechanism responsible for temperature-dependent optical dephasing to be determined. It is shown that excitation of librations rather than scattering between delocalized dimer states is the principal mechanism.
1. Introduction
Electronic excitation of interacting molecules can result in a state which is delocalized, i.e. at a given instant of time, there is finite probability of finding the excitation on more than one molecule [1]. In pure molecular crystals, delocalization can give rise to exciton band structure [1] and coherent exciton transport [2,3]. Molecular dimers are the smallest systems in which delocalization can occur. As in any system which can show delocalization, dimer electronic excitations will be delocalized if the strengths of the intermolecular interactions are large relative to the magnitudes of the fluctuations in site energies and intermolecular coupling matrix elements [2,3].
In this paper we present a detailed examination of the temperature-dependent optical dephasing of delocalized tetracene and pentacene dimers, and tetracene and pentacene monomers in $p$-terphenyl host crystals. The dimers consist of isolated pairs of guest molecules substitutionally embedded in a host crystalline lattice. The relative orientations of the molecules in the dimer pairs are similar to the orientation of host molecules in the pure crystal. Thus there are a variety of possible dimer pairs having various intermolecular interactions.
A comparison of picosecond photon echo experiments on tetracene and pentacene dimers in the same host crystal permits an unambiguous determination of the mechanism responsible for the temperature-dependent optical dephasing of the delocalized excited states of these dimer systems. In contrast to isotopically mixed crystal dimers, e.g., naphthalene-$h_8$ dimers in naphthalene-$d_8$ host crystals or tetrachlorobenzene-$h_2$ dimers in tetrachlorobenzene-$d_2$ host crystals [4], these chemical impurity dimers do not have as their principal dephasing mechanism scattering between the delocalized dimer states. Rather, these chemical impurity dimers dephase because of phonon excitation of local phonon (librational) modes of dimers.
2. Experimental procedures
The picosecond photon echo system used to study the pentacene in $p$-terphenyl dimers has been described previously [5,6]. For the tetracene in $p$-terphenyl experiments, a modified version of this system was used. Tunable blue (490 nm region) picosecond pulses were generated by difference mixing a single quadrupled Nd : YAG laser pulse (226 nm, 15 $\mu$J, 40 ps) with a single dye laser pulse (20 $\mu$J, 30 ps) of appropriate wavelength in a 30 mm angle-tuned KDP non-linear crystal. This procedure provides a stable source of 30 ps, 2 $\mu$J tunable blue pulses at a repetition rate of 800 Hz. The experimental arrangement is outlined in fig. 1. A blue single pulse is split into two excitation pulses of appropriate energies and variable
Fig. 1. Photon echo setup used in the tetracene in $p$-terphenyl experiments. A single 1.06 $\mu$m pulse is selected from the acousto-optically mode-locked and $Q$-switched Nd:YAG laser by a Pockels cell (PC) and calcite cube polarizer (POL). This single pulse is doubled (CD*A crystal) and then quadrupled (KDP crystal) in frequency. The remainder of the Nd:YAG pulse train is doubled and used to pump a cavity-dumped dye laser. The dye laser single pulse is then made temporally and spatially coincident with the quadrupled Nd:YAG pulse in an angle-tuned KDP crystal. This difference mixing process provides a stable source of tunable blue pulses. The single blue pulse is split into two excitation pulses of appropriate energies. One of these pulses is directed into a corner cube on a motorized delay line to provide a variable temporal delay between the pulses. See text for details. BS = beam splitter; E = Fabry–Pérot etalon; DC = flowing dye cell.
temporal delays for the photon echo experiment. Non-collinear excitation geometry ($2.75^\circ$ angle between the two pulses) was employed to spatially isolate the echo signal. Other details involved in detecting the echo signal are as described previously [6].
The samples were mounted on a heating block located in a liquid-helium dewar. Cold helium gas was allowed to flow over the sample. The temperature was varied by controlling the flow of helium gas and the power delivered to the heating block. The temperature was accurately maintained ($\pm 0.01$ K) by the Scientific Instruments model 3610-B temperature controller which delivered current to the sample block heater, and was measured with a Scientific Instruments model N1G germanium resistance thermometer in thermal contact with the sample (accuracy better than 0.01 K). Details of the sample preparation and the polarized absorption spectrometer are as reported previously [6].
3. Results
Tetracene in the low-temperature $p$-terphenyl lattice exhibits two distinct sites [7]. Fig. 2a shows a transmission spectrum of a moderately concentrated crystal of tetracene in $p$-terphenyl. The two $S_0 \rightarrow S_1$ monomer absorption origins, labeled TO$_1$ and TO$_2$, have similar splittings and inhomogeneous linewidths as the PO$_1$ and PO$_2$ monomer origins of pentacene in $p$-terphenyl which have been extensively studied [8]. (Transmission spectra of the four pentacene monomers and several pentacene dimers may be found in refs. [5,6].)
At much higher concentrations ($\approx 10^{-3}$ M/M) of tetracene in $p$-terphenyl, the formation of dimers is expected statistically as found in pentacene in $p$-terphenyl crystals of similar concentration [5,6]. A transmission spectrum of a high-concentration crystal of tetracene in $p$-terphenyl in the vicinity of the $S_0 \rightarrow S_1$ absorption origins, TO$_1$ and TO$_2$, of tetracene
zero transmittance. To the red of the broad origin lie several dimer peaks. The tetracene dimer splittings and line positions relative to the TO$_1$, TO$_2$ tetracene monomers are similar to those found for the pentacene dimers built off of the pentacene PO$_1$. PO$_2$ monomers in $p$-terphenyl [5,6]. However, in tetracene in $p$-terphenyl there are more dimers split to the red of TO$_1$ and the largest dimer—TO$_1$ splitting (128 cm$^{-1}$) is more than three times that found for pentacene in $p$-terphenyl (39 cm$^{-1}$). These differences may be due to the smaller size of the tetracene molecule compared to the pentacene molecule which allows for a greater freedom in the relative positioning of the two molecules which make up the dimers.
Temperature-dependent photon echo measurements were made on tetracene monomers and the two tetracene dimer peaks, labeled TD$_1$ and TD$_2$ in fig. 2, and on pentacene monomers and dimers.$^{\dagger 1}$. Fig. 3 shows the photon echo decay for the TD$_1$ dimer at 2.61 K. The data is exponential (see inset), and de-
$^{\dagger 1}$The pentacene dimers labeled PD$_1$, PD$_2$, and PD$_3$ in table 1, were formerly called R$_3$, R$_4$ and R$_5$, respectively, in refs. [5,6].
Fig. 2. (a) A tetracene in $p$-terphenyl transmission spectrum at 1.6 K of the two S$_0 \rightarrow S_1$ monomer origins, TO$_1$ and TO$_2$. The two tetracene origins have similar splittings and inhomogeneous linewidths as the PO$_1$ and PO$_2$ monomer origins of pentacene in $p$-terphenyl [8]. The sloping baseline in the spectrum is due to the spectral characteristics of the xenon arc lamp. (b) Transmission spectrum of a high-concentration tetracene in $p$-terphenyl crystal in the region about the monomer S$_0 \rightarrow S_1$ origins at 1.6 K. The TO$_1$ and TO$_2$ monomer origins coalesce into a broad and flat-topped peak because the very high monomer concentration results in zero transmittance. To the red of the broad origin lie several dimer peaks. The inset shows a transmission spectrum of a crystal with an even higher tetracene concentration. At these higher concentrations, several new dimer peaks become clearly visible. The TD$_2$ dimer is located at 4961.9 Å. (Wavelength scale, 4 Å/division.) A detailed temperature-dependent study was performed on the two well-resolved dimer peaks labeled TD$_1$ and TD$_2$ (see text).
Fig. 3. Photon echo decay for the TD$_1$ tetracene dimer at 2.61 K. The straight line through the logarithm of the data (see inset) shows that the echo decay is exponential over almost four factors of 1/e in signal intensity. The signal decays as $I_{PE} = I_0 (-4t/T_2)$ and thus $T_2 = 2.0$ ns for the data shown in the figure.
decays with a $T_2$ time constant of 2.0 ns. Two contributions to the photon echo decay constant ($T_2$) may be defined as follows
$$\frac{1}{T_2} = \frac{1}{2\tau} + \frac{1}{T'_2}. \quad (1)$$
where $\tau$ is the excited-state lifetime (decay out of the $S_1$ state) and $T'_2$ includes any other dephasing processes. For example, modulation of the dimer excited-state energy by phonons is a pure dephasing process, $T'_2$, while phonon scattering between delocalized excited dimer states, is a $T_1$ process. Both processes would be included in $T_2$.
For the pentacene in $p$-terphenyl monomers [6] and dimers $\tau$ has been measured and is known to be constant over the temperature range of the photon echo measurements. (In general singlet states do not show temperature-dependent lifetimes in this temperature range. In triplet systems temperature-dependent spin-lattice relaxation can cause a temperature-dependent lifetime which will contribute to the temperature dependence of dimer dephasing [9].) This lifetime contribution can be subtracted from the $T_2$ measured at each temperature and a temperature-dependent plot of $T'_2$ can be made. In the tetracene in $p$-terphenyl system the plots were obtained by assuming that at 1.4 K the system is in the low-temperature limit, and therefore $T_2 = 2\tau$ as has been seen previously for other systems of monomers [6,10,11] and dimers [6]. Since the lifetime contribution to dephasing at these higher temperatures is negligible, even a large error in determining $\tau$ has no effect in determining the temperature dependence of $T'_2$.
A plot of ln $T'_2$ versus $1/T$ for the TD$_1$ dimer gives a straight line over many orders of magnitude of $T'_2$ (fig. 4). The slope of this line is $\Delta E/k$ and gives an activation energy ($\Delta E$) of 10 cm$^{-1}$ for the TD$_1$ dimer. Similar plots for the other tetracene and pentacene dimers and monomers were made. These results are summarized in table 1. (The activation energy obtained here for TO$_1$ (12 cm$^{-1}$) differs from that reported previously (8 cm$^{-1}$) [12]. The data compiled in ref. [12] were apparently obtained over too small a temperature range (less than one factor of 1/e change in $T_2$) to give an accurate determination of the slope.)
4. Discussion
The temperature-dependent optical dephasing of delocalized dimers in $p$-terphenyl host crystal has recently been studied theoretically [13]. In the calculations a canonical transformation partially diagonalized the dimer excitation-phonon coupling terms in the Hamiltonian. Optical Redfield theory was then employed to obtain the temperature dependence of the optical dephasing. Three possible mechanisms are capable of explaining the observed exponentially activated temperature-dependent dephasing: (A) Phonon scattering between delocalized dimer states (a $T_1$ process), (B) phonon scattering to delocalized librational states (mini-librons) in the dimer ground and excited electronic states (a $T_1$ process) *2, and (C) pure dephasing induced by coupling to optical phonons of
*2 The definition of this process as a $T_1$ process is somewhat arbitrary. In our theoretical treatment of the dimer [13] and a previous treatment of the monomer [14,15], the system was taken to include the librational levels, all other phonons being part of the bath. However, in several treatments of the monomer problem [11,14,15] the librational levels were included in the bath. The dephasing librational excitation is then called a $T_2^*$ process. The two approaches are equivalent.
Table 1
Optical dephasing activation energies
| Spectral line | $\lambda(\text{Å})$ a) | $\Delta(\text{cm}^{-1})$ b) | $\Delta E(\text{cm}^{-1})$ c) | $T_2'(\infty)(\text{ps})$ d) |
|---------------|------------------------|-----------------------------|-------------------------------|-----------------------------|
| tetracene monomers | TO$_2$ | 4929.8 | −3.3 | 11 | 35 |
| | TO$_1$ | 4930.6 | 0.0 | 12 | 18 |
| tetracene dimers | TD$_1$ | 4933.0 | 9.9 | 10 | 11 |
| | TD$_2$ | 4961.9 | 128.0 | 17 | 6 |
| pentacene monomers | PO$_3$ | 5858.5 | −181.6 | 27 | 5 |
| | PO$_2$ | 5878.6 | −123.2 | 30 | 2 |
| | PO$_1$ | 5920.1 | −4.0 | 25 | 4 |
| | PO$_1$ | 5921.5 | 0.0 | 27 | 5 |
| pentacene dimers | PD$_1$ | 5928.0 | 18.5 | 20 | 5 |
| | PD$_2$ | 5934.0 | 35.6 | 23 | 5 |
| | PD$_3$ | 5935.1 | 38.7 | 25 | 8 |
a) Wavelengths are in air.
b) $\Delta$ is the energy separation from TO$_1$ for the tetracene system and from PO$_1$ for the pentacene system.
c) $\Delta E$ is the optical dephasing ($T_2'$) activation energy. See text.
d) $T_2'(\infty)$ is the optical dephasing ($T_2'$) pre-exponential factor: $T_2' = T_2'(\infty) \exp(\Delta E/kT)$.
the $p$-terphenyl host (a $T_2$ process). Scattering between delocalized dimer states (mechanism A) was eliminated as a possibility in a previous experimental study of pentacene dimers [5] because the exponential activation energy did not correspond to the observed dimer pair splitting as required by theory [13]. However, no conclusive choice could be made between mechanisms B or C based on the pentacene dimer experiments alone.
It has been established that phonon-induced scattering to monomer librations is responsible for the temperature-dependent dephasing of pentacene monomers in naphthalene [15] and benzoic acid [17] host crystals. Pentacene monomers in $p$-terphenyl are also believed to dephase due to this process, although direct observation of the libration in absorption or emission is lacking in this system. The properties of the guest libration are determined by guest–host interactions but are critically dependent on the nature of the guest. In the case of pentacene monomers in naphthalene this was demonstrated by studying the deuterium isotope (pentacene-$d_{10}$) shift in frequency of the libration [15]. Thus, this temperature-dependent dephasing arises from properties intrinsic to the guest molecule in addition to the properties of the host lattice. The pentacene monomers have librational energies in a range consistent with the activation energies of the various pentacene dimers.
However, the $p$-terphenyl crystal has an optical phonon mode observed by Raman spectroscopy at $\sim 20.5 \text{ cm}^{-1}$ [18]. Coupling of the dimer to an optical phonon band can result in an apparent exponentially activated temperature-dependent dephasing with an apparent activation energy determined by the phonon band position and density of states [13]. The temperature dependence is caused by the thermal occupation of the phonon band which is an intrinsic lattice property independent of the nature of the dilute impurities which form the dimers. The 20.5 cm$^{-1}$ phonon band energy is consistent with the observed (table 1) pentacene dimer activation energies. While pentacene monomers are dephased by the libration mechanism B, dimers have the additional possibility of coupling to optical phonons by the modulation of the intermolecular interaction, mechanism C.
From table 1 it can be seen that the various pentacene dimers have somewhat different activation energies. These are real differences, well outside of experimental error. The libration mechanism can give rise to these differences if it is postulated that each dimer configuration has particular mini-libron splittings and site librational frequencies. The optical phonon mechanism explains the differences by postulating that the various dimer configurations couple selectively to restricted regions of the phonon $k$-space, regions consistent with the spatial arrangements of the dimers. In general, the phonon dispersion varies as a function of direction in $k$-space, and since the activation energy
depends on the phonon dispersion, various activation energies could result.
The choice between the libration mechanism (B) and the optical phonon mechanism (C) can be made on the basis of the tetracene dimer photon echo results. If coupling to lattice phonons (in this case coupling to the 20.5 cm$^{-1}$ phonon suggested by the pentacene dimer activation energies) is responsible for dimer optical dephasing, then the activation energies for tetracene and pentacene dimers would be very similar. However, if the libration mechanism is responsible, the activation energies should be significantly different.
The activation energies obtained for the tetracene dimers (10 and 17 cm$^{-1}$) are very different from those for the pentacene dimers (20, 24 and 25 cm$^{-1}$), although the $p$-terphenyl host is common to both. This rules out a dephasing mechanism due exclusively to bulk properties of the host, such as coupling to the $p$-terphenyl lattice phonons. The differences in activation energies between tetracene and pentacene dimers arise because of the differences in librational frequencies of such physically different guest molecules. Theoretically, uncorrelated scattering to the mini-librons results in a tetra-exponential temperature behaviour with activation energies determined by the librational energies in the dimer ground and excited states and by the mini-libron splittings [13]. However, the temperature dependence appears as a single exponential if the librational energies and splittings in the ground and excited states are comparable and if the librational lifetimes are essentially constant over the temperature range studied [13]. This is evidently the situation for pentacene and tetracene dimers.
Unlike dimers which are only isotopically distinct from the host [4], e.g., naphthalene-$h_8$ dimers in naphthalene-$d_8$ host, chemically distinct dimers do not have librational motions which are amalgamated into the host optical phonon bands. Thus dimers made up of molecules such as pentacene and tetracene in a host like $p$-terphenyl exhibit well-defined low-frequency librations. Thermal population of these librations is effective in dephasing the delocalized dimer states. In contrast, isotopic dimers, with amalgamated phonons, do not have these low-frequency local phonons available. This evidently leaves scattering between delocalized states as the dominant dephasing mechanism in isotopic dimers. Thus the temperature-dependent dephasing of chemically distinct dimers is intrinsically different from the dephasing of isotopic dimers.
Finally, it is interesting to note the observation of a peak in the tetracene monomer in $p$-terphenyl phonon sideband 24 cm$^{-1}$ to the blue of the zero-phonon line [7]. This is probably due to a two-quantum transition of the libration responsible for the observed monomer 12 cm$^{-1}$ dephasing activation energy. A similar situation occurs for pentacene monomers in naphthalene [15].
Acknowledgement
This work was supported by the National Science Foundation, Division of Materials Research (#79-20380). WLW would like to thank AT & T Bell Laboratories for a Cooperative Research Fellowship. MDF would like to thank the Guggenheim Foundation for a fellowship which contributed to this research. We would like to thank Professor A.H. Zewail for helpful comments pertaining to this manuscript.
References
[1] A.S. Davydov, Theory of molecular excitons (Plenum Press, New York, 1971).
[2] R. Silbey, in: Spectroscopy and excitation dynamics of condensed molecular systems, eds. V.M. Agranovich and R.M. Hochstrasser (North-Holland, Amsterdam, 1982) p. 1.
[3] M.D. Fayer, in: Spectroscopy and excitation dynamics of condensed molecular systems, eds. V.M. Agranovich and R.M. Hochstrasser (North-Holland, Amsterdam, 1982).
[4] B.J. Botter, A.J. van Strien and J. Schmidt. Chem. Phys. Letters 49 (1977) 39; J.B.W. Morsink and D.A. Wiersma, Chem. Phys. Letters 89 (1982) 291; D.M. Burland and A.H. Zewail, in: Advances in chemical physics, Vol. 40, eds. I. Prigogine and S.A. Rice (Wiley, New York, 1979) p. 369, and references therein.
[5] R.W. Olson, F.G. Patterson, H.W.H. Lee and M.D. Fayer, Chem. Phys. Letters 79 (1981) 403; H.W.H. Lee, F.G. Patterson, R.W. Olson, D.A. Wiersma and M.D. Fayer, Chem. Phys. Letters 90 (1982) 172.
[6] F.G. Patterson, H.W.H. Lee, W.L. Wilson and M.D. Fayer, Chem. Phys. 84 (1984) 51.
[7] N.J. Kruse and G.J. Small, J. Chem. Phys. 56 (1972) 2985.
[8] R.W. Olson and M.D. Fayer, J. Phys. Chem. 84 (1980) 2001.
[9] A.H. Zewail, private communication.
[10] J. Morsink, T.J. Aartsma and D.A. Wiersma, Chem. Phys. Letters 49 (1977) 34.
[11] T.E. Orlowski and A.H. Zewail, J. Chem. Phys. 70 (1979) 1390.
[12] T.J. Aartsma and D.A. Wiersma, Chem. Phys. Letters 42 (1976) 520.
[13] H.W.H. Lee and M.D. Fayer, to be published.
[14] P. de Bree and D.A. Wiersma, J. Chem. Phys. 70 (1979) 790.
[15] W.H. Hesselink and D.A. Wiersma, J. Chem. Phys. 73 (1980) 648.
[16] J.L. Skinner, J. Chem. Phys. 77 (1982) 3398.
[17] L.W. Molenkamp and D.A. Wiersma, J. Chem. Phys., to be published.
[18] B.A. Bolton and P.N. Prasad, Chem. Phys. 35 (1978) 331. |
1950 on the Victorian Snowfields
Malcolm McColl
(Alpine Club of Victoria)
Whatever else of major interest has made the year 1950 stand out in the rapidly-growing corridor of the ski-ing years in this State, the lightness of the snowfall and the shortness of the season will probably be the item which will be brought to mind in the years to come. Skiers being what they are, no matter what momentous event has occurred in any articulor year, it is the depth of the snow blanket and its effect on one's ski-ing activities that finally labels the year. Thus, according to one's years or good fortune in having developed the ski-ing mania while very young, one refers to '27 as the heaviest year ever, or '43 as being quite a good year, and '46 and '47 as very heavy years.
1950 will long stand out as the lightest year on record, even in the memories of those who may aspire to be among the founders of winter sport in Victoria. The snow year opened late and finished early without much snow falling in between. At Hotham where on King's Birthday week-end, cars were parked in profusion on the summit of the mountain, the Drag was unskiable for most of the season and the Alpine Highway was open to motor traffic about October 8th—an event usually occurring in early December.
Among the items in 1950 worthy of comment, and of certain speculative interest, is the question of access to Mount Hotham. This problem suddenly assumed large proportions when it was seen at the beginning of the season that even the rather antiquated—in this age of mechanisation—method of transport to the snowline usually available was in grave danger of becoming nonexistent. No one will deny that our worthy four-footed friend, the horse, has his uses and is, in certain places, highly and rightly prized for his fleetness of foot or his value as a pleasant means of seeing the country. But his usefulness as a sole means of access to Mount Hotham is at least open to doubt, particularly as most of the who place themselves on his back do so once or twice only per year, and that is on this very trip. Even those who were more familiar with the right side of the animal to face when mounting were not in the habit of making a steep 4000-foot ascent. Consequently, the physical discomfort caused by that mode of transport as well as by the inclement weather at times was in no small degree responsible for many otherwise ardent tyros being discouraged from visiting the State's premier winter resort.
Ways and means were studied, but the difficulties seemed insuperable. So, suddenly, in this age of progress, Hotham, up till then on a steady wave of post-war progress, slipped back a whole decade and access to it became, for the majority, the long, arduous and risky trek from St. Bernard. But the fates were kind and no real blizzards blew and the big lumbering blade of the front of the Country Roads' Board Matador truck pushed its way up and down the road right up to the Roads' Board camp for the whole season. Some of us even drove our own vehicles up and left them there for two weeks. So far so good. But what might have happened if we'd had the snow of '47, and the blizzards, and the plough hadn't got closer than four miles below St. Bernard? There might have been a number of very weary people arriving at Hotham, that is, supposing they had arrived at all.
Meanwhile, in the Kiewa area the American "Snocat" arrives, a vehicle which could mean for Hotham the dawn of a new era and the end of nineteenth century transport.
The year saw a continued steady growth of interest in skiing throughout the whole of the State, which evinced itself in a surge of publicity both in the press, on the air, and even in Parliament. In the latter sphere, one of our most active and well-informed skiers became one of our leading citizens when T. W. Mitchell was appointed Attorney General on the formation of the Country Party Government.
The pulses of all skiers quickened at the news that the State Electricity Commission had acquired two new Tucker Snocats for winter communication service on the Kiewa Hydro-Electric Scheme. These vehicles have been developed over a great number of years in the United States and have proved admirably suited to over-snow progress, mainly through an ingenious driving system over a sliding pontoon at the rear end and steering skis at the front end. Their ability to traverse steep slopes and to climb with complete ease, combined with their versatility in all types of snow make them the answer to practically all our snow transport problems. But dollar shortage stands in the way at the moment, and we can only dream of the day when we will be able to save a couple of days at each end of our snow holiday. Of course, when that day comes we may even have advanced into the helicopter age.
Almost before most of us were aware of it, a new ski village has grown up at Falls Creek. This village is in the centre of the area where millions of pounds are being spent to increase the State's electricity supply, and where good access roads are essential. Country clubs within a radius of one hundred miles are flocking to this area and even a Sydney University Club has applied for a building site.
The slopes are not as long and as steep or varied as Hotham, but these factors are offset to a great extent by its ease of access to and from many areas. Already State Championships have been held at this centre and with the visit this year of the New Zealand team interest will be once more focussed there. The suggestion made recently that the future village area might be sited on McKay Creek near No. 1 power station is of considerable interest.
Memories of pre-war days were revived by the return to Victoria of Franz Skardarasny to take charge of the ski school at Mt. Buffalo along with New Zealander, K. W. Kear, and Victorian, Bill Marriott. It takes one's mind back to what might be termed the beginning of serious skiing in Victoria. Those who enjoyed the benefit of Franz's tuition and later that of Toni Walch and Richard Werle will agree that this regular organised teaching transformed our conglomeration of various styles into a fairly homogenous, sound and, to say the least of it, vastly improved style.
Unfortunately, with our lean snow year, Buffalo was not the best place for Franz to renew his acquaintance with our snows. Let us hope that he will return from his native land to enjoy much better snow in future years. Rumours have it that Toni is anxious to return also to renew his friendship with many Victorian skiers who deem their days on the snow with him as some of
the most enjoyable of their ski-ing career. He was last reported in the Laurentians in charge of a ski school.
The ski schools at our four ski centres—Buller, Hotham, Buffalo and Falls Creek—continue to attract those anxious to lay the foundation of their ski-ing career as well as those who are keen to build up and improve upon what they have already. As yet there is no talk of an Australian style as a counter to the Swiss, French or Austrian styles. The Swiss style seems to have retired somewhat, leaving the field more open for the long established Austrians to ward off the sallies of their highly organised and subsidised rival. The claims made by some that evidence of the superiority of the French style is to be seen in their convincing placings in world championships is scarcely borne out, as the champion, Couttet, would probably have been outstanding in whatever part of the European Alps he happened to have been born and whatever style he had acquired in his early teens. A fair sprinkling of the champions of all the leading ski-ing countries of Europe seem to be among the leading competitors—with the Austrian women much more to the fore than most others.
We in Australia would do well to use our initiative and develop our own style—not just to assert our national spirit but because our snow is generally so vastly different to that of Europe that it often requires a different technique.
The amazing growth evident at Buller over the last two years continue unabated with the completion of the Omega, Walking Club, Ullr, Harding, Moose, Women's, Monsanto, University, Yurredia, Dandenong, Morgan, Pattern, Buller and Grey Rocks Club lodges. These, together with the buildings completed before last season, namely, the Ivor Whittaker Lodge and its surrounding cabins, the Chamois, Postal Institute, C.S.I.R.O., Youth Hostels' Association and Junior Ski Club, brought the total accommodation on the mountain to a figure far in excess of any of our other snowfields. This year the rush has slowed down a trifle, the only buildings to commence so far being the B.M.W. Club, which miraculously sprang up over night, and the Alpine Club, whose building is being erected on what is probably the steepest building site in the Alps, namely 25 degrees.
However, the buildings mentioned above have formed the nucleus of quite a substantial village, which, together with the promised road improvement and water reticulation, should make a well-organised centre. Another benefit promised, a cabaret and general store, is viewed with rather mixed feelings, many skiers believing that the unorganised and spontaneous atmosphere engendered in most ski lodges is vastly preferable to the organised and synthetic commercial variety.
The future growth of the mountain will be watched with interest. The standard of the buildings erected so far has been fairly high, although one or two clubs, in their desperate eagerness to get the building up at any cost, have rather fallen from grace. The virtue of simplicity and proper regard for the relationship of window opening to wall space has not made itself apparent to a number of otherwise well meaning people, whilst in one rather conspicuous position a club has fallen into that old trap of trying to improve on nature by ripping colour from its context and adorning their building with a particularly vivid scarlet. Unless a building in its structural form has the perfection of the structures of Ancient Greece, it is better, in the writer's opinion, to eschew bright colours and to let the beauties of nature provide the picture.
Road access and parking problems provided a number of incidents during the season, but all are merely teething troubles which are to be expected in the growth of any boom settlement. The course of a year or two should see them ironed out.
The Buller Committee of Management continues on its successful way, keeping a paternal eye on the activities of the area without using too heavy a hand in its control. Having been taken as a model for the Alpine Report it can be considered to be proceeding happily along the right lines.
Now that accommodation problems are being solved for many skiers their next interest lies in getting to the top of the neighbouring slopes in the minimum of time in order to descend those slopes more often. Overseas visitors return wondering if they will ever be able to face again a slope unadorned with lengths of rope, pulleys and engines, not forgetting the gent selling the tickets.
We are far behind overseas countries and our neighbours, New Zealand, in this matter
of ski tows. Those that we have are either unreliable or small in scale. But even so, their success is undoubted. The tow at Buller despite its comparatively short length has a queue of forty waiting to use it week-end after week-end in the peak of the season. The only other one in the State, at Mt. Buffalo, although rather troubled by the lightness of the snow this year, is a very popular institution. Perhaps the lack of effort in this direction has been due to the preoccupation with building.
However, a strong effort is at present being made to organise a really long tow at Hotham, based on the types which have been successful in New Zealand. The ultimate aim there would be a tow on both sides of Swindler's Creek with perhaps another one in the Australia Drift area. These would open up vast areas of wonderful slopes which are now only occasionally used by the more energetic. Efforts are also being made to erect a tow in the Falls Creek area. Buller could well do with a longer one on the higher slopes. The installation of tows on all our slopes is long overdue. The standard of ski-ing in this State is going to improve by leaps and bounds once we do have a few tows of worthwhile length, as they increase downhill ski-in many times and still leave large areas of untracked snow for the confirmed soft-snow skier.
The visit of Arnold Lunn to this country was a stimulus and a gratification of a long-felt desire to meet this outstanding figure in the ski-ing world to those who for years had derived considerable pleasure and benefit from his witty and forceful writings. A full report of the memorable occasions on which Victorian skiers met this distinguished visitor appears elsewhere in this book.
Another outstanding event of the Victorian ski-ing year was the publication of the State Development Committee's Report on Alpine Areas, setting out a comprehensive policy for the future development of all our snow centres. A commentary on this Report is given on other pages by Vernon Corr.
The visit of the Bogong Club team to New Zealand was a happy inspiration. It will probably do more than many inter-dominion visits to bring our two countries closer together. The lessening of the time taken to cross the Tasman, and, unless Australians do something about it, the better ski-tow facilities will be among the causes of many of our skiers deserting their native shore for a land where powder snow, so it is reliably reported, can be found with monotonous regularity.
On the equipment side, our local manufacturers are continuing to forge ahead, albeit slowly, as the competition from the imported article is always severe. Australians being notoriously prone to look upon the imported article as being superior to the local, often due to the mistaken idea that what comes 10,000 miles must be good, or at least better than the local product. In that belief they have sometimes been justified. However, if a few more would take a risk and sample local ware we might be in a better position with regard to supply.
Bindings, ski boots, skis, waxes, clothes, etc., made locally and of good quality can all be purchased now of a standard high enough to satisfy the most critical.
However, it seems to take a long while for proven new ideas from abroad to be turned out by our own suppliers. Maybe they act on the principle that a thing is not established until it has been tried out for three years or so. Our importers, however, are generally on the beam and for a country with a relatively small ski-ing population we don't do so badly. Of course, we pay dearly for our pleasures—but, who doesn't these days?
One feature of overseas snowfields which has been slow in appearing in this country is the cutting of tracks in order to develop runs through the more thickly timbered country. In parts of the United States where the snow country is at a much lower altitude, there are resorts where all the runs are through country of this nature, and trail running is the only kind of ski-ing known. Even in Switzerland, the wooded areas are by no means disregarded and trail running is extensive. In fact, it is almost a separate branch of the sport, entailing as it does a type of ski-ing on a track where there is no stopping or going back and where one needs a blind faith in the ability of the person ahead to keep going at all costs. Of
course, sometimes the faith is misplaced and the consequences of suddenly rounding a corner to find the skier ahead could stop, in however unorthodox manner, are somewhat interesting.
Various areas at Buller, Hotham and Feathertop could be vastly improved by the judicious use of an axe. Indeed, unless the axe is used in a few parts these areas will soon be too overgrown to ski on, thanks to the rapid growth of scrub induced by some of our severe bush fires.
Interest in the Baw Baw area continues to increase. A number of clubs have been formed to operate and build there. Its main appeal to skiers from Melbourne should be its relative closeness, but to listen to accounts of the time and energy taken to get there, it would appear that there is not much to choose between it and Buller. Also, the slopes and length and steepness of runs do not appear to compare with those of the other resorts. Still it is good to see the area progressing as it is close to a region which will be more densely populated in the future.
At the present moment, development waits upon the selection of the most suitable village site and the Government's consideration of the Alpine Report. |
PROFITABLE HAMURA NIGIWAI SHOPPING TICKETS!!
Please make the best use of Hamura Nigiwai shopping tickets!! These tickets were issued in order to make the economy in Hamura active. So we will give you profitable information for shopping.
◎ School Uniforms & Other School Clothes → If you buy school uniforms & other school clothes over ¥20,000 at the following shops by Apr.30th (Thu), you will be given a pair of white socks for school.
Member Shops/ Shimoda Inc. (<有>総合衣料しもだ ☎554-2039), Fashion House Shibata (ファッションハウス シバタ ☎554-2213), Sogo Iryo Motohashi (総合衣料もとはし ☎554-2190).
◎ Fire Alarms for Houses → You can get fire alarms cheaper up to Aug.31st (Mon) using Nigiwai shopping tickets.
Member Shops/ Seisho Inc. (<株>セイショウ ☎554-6211), Takishima Shoten Inc. (<株>滝島商店 ☎554-2239)
※ Also other profitable services & special favors for this ticket will be offered by member shops. Please check leaflets or the website of Hamura City Chamber of Commerce.
Detail/ Sangyo Kasseika Suishinshitsu Shokogyo Shinko gakari
COMMUNITY BUS HAMURAN ⇔ RUNNING SERVICE CHANGES
◆ Saturdays & Sundays during Tulip Festival: You can not use bus stops, Ippoin (一峰院) & Ikoinosato (いこいの里). The temporary bus stop will be set up at free Shuttle Bus stop (close to the venue). Please use that bus stop. ◆ During the 6th Kaze no Owara: There will be Traffic Regulation on the street in front of Yutorogi, Apr. 4th (Sat) 4:30pm~7:30pm (Bus Stop Toshokan is not available)
THE 32ND AOZORAICHI, FLEA MARKET
MAY, 16TH (SAT) 9:00AM~NOON AT FUJIMI KOEN C GROUND
↑ If it rains or the grounds is not in a good condition, it'll be cancelled.
Vendors: Wanted!!
Eligible/ Individuals or groups living in Hamura city who are senior high school students and over. (senior high school students are required their guardian's consent.) ※ You can't join for business. Capacity/ 100 blocks (on a first-come first-served basis) Fee for Entry/ ¥800 Application/ Please come to Shohi Seikatsu Center directly on Apr. 19th (Sun) 9:00am~noon.
WANTED: HAMUSUPO (HAMURA GENERAL SPORTS CLUB) MEMBERS!!
This sports club aims at “anybody does physical exercises, make friends and a good town”. Now we are accepting a member registration. Everybody is OK!!
Dues a Year/ Regular Members ¥6000 Family Members ¥12,000 Gold Members ¥20,000
Application/ Every Tuesday 1:00pm~3:00pm, every Thursday & Saturday 10:00am~noon, at Hamusupo Renraku office
Please register with dues after you fill out a registration form.
More Information/ Hamura Renraku office (☎519-5712)
Fax/ 519-5713 ☎email@example.com
MUNICIPAL TAX PAYMENT AT CONVENIENCE STORES
From April, you can pay municipal tax and so on at convenience stores.
Municipal Tax, Real Estate Tax, Light Vehicle Tax, National Insurance Tax, Elderly Care Insurance Tax, Fee for Nursery Care etc.
Stores/ Seven Eleven, Family Mart, Lawson, Mini Stop, am/pm etc.
Detail/ Kaikei ka kaikei gakari
CHILD SUPPORT ALLOWANCE
Child support allowance (Dec. ~ Mar.) will be paid in Apr.13th (Mon) into your designated account. Please check that you surely accepted the allowance.
Detail/ Kosodate Shienka Shien gakakri
TEMPORARY·HOLIDAYS & AFTER-ILLNESS CHILDCARE·FAMILY SUPPORT CENTER PROJECT
You can take childcare services at nursery schools in Hamura in the following cases.
Temporary Childcare In case that you need the service temporarily or in an emergency such as your health, short-time job, a sudden illness and so on.
Childcare on Sundays In case that you need the service on Sundays or holidays except for the year-end & New Year’s holiday because of your job.
After-Illness Childcare In case that your child needs after-illness childcare.
Eligible/ Preschoolers in Hamura Fee/ Charged by your child’s age or hours.
※ Please ask the nursery school that are in charge of childcare.
Contact/ Hoiku-ka Hoiku gakari
Family Support Center Project
This organization is consist of both members who need supports for children and offer them. If you register as a member, you can use this service at many cases.
Eligible/ 6 month-old children ~ the 3rd graders in elementary schools
Contact/ Hamura City Family Support Center (Hamura Shi Shakai Fukushi Kyogikai ☎554-0304)
WANTED: PART-TIME WORKER: FROM SCHOOL LUNCH CENTER
Number/ A few persons Work Time/ May. 1st (Fri) ~ Monday ~ Friday (except for holidays) 1:00pm ~ 4:00pm Content/ Washing plates & other works Fee/ ¥860 per hour
Application & Contact/ Please come to Hamura Mizuho Chiku Gakko Kyusyoku Center Kyusyokuka Syokuin gakari after you fill out an resume available at stores.
FURNITURE FALLING PREVENTION INSTRUMENTS
Hamura City will provide the citizens with instruments that prevent furniture in your house from falling for free in order to hold damage minimum at an earthquake (such as Tsupparibo or an film on a glass preventing to scatter). We also install them for elderly families.
Application & Detail/ Please come to Seikatsu Anzenka Bosai gakari
AIDS FOR AN EARTHQUAKE-RESISTANT CHECK & REPAIR OF WOODEN HOUSES
Hamura City will aid a part of expense for an earthquake-resistant check & repair of wooden houses in order to promote to make the city stronger in disasters. Eligible/ Houses built before May. 31st 1981 Grant/ 50% of expense Contact in detail/ Kenchikuka kenchiku gakari
SUPPORT FOR MOTHERS IN FATHERLESS FAMILIES
● In case that a mother in a fatherless family in Hamura takes a training course to get a job. (20% of tuition fee) Course/ Medical Office Work, Home Helper, Chef, Tax Accountant···
● In case that a mother in a fatherless family in Hamura gets a license to have a better job. Course/ Nurse, Childcare Nurse, Care Worker···
Eligible/ A mother who supports children below 20 years of age and accept child support allowance or have the annual income equal to it. ※ Please contact about details.
Application & Contact/ Please come to Kosodate Shienka Shiengakari
SUBSIDY FOR INSTALLATION OF ENVIRONMENTAL PROTECTION TYPE APPLIANCES
Hamura City will subsidy to people who install environmental protection type appliances for houses. Why don’t you reduce carbon dioxide emissions that cause global warming and protect the earth? Eligible Appliances/ Hot-Water Supply Appliances → Ecocute (¥40,000), Ecowill (¥40,000), Ecojose (¥20,000)
Application Period/ within 60 days after the installation of eligible appliances or your purchasing of a house installed them. ※ An application form is available at Kankyo Hozenka Kankyo Hozen gakari. Or please download on the website.
Application & Contact/ Kankyo Hozenka Kankyo Hozen gakari
PERIODIC RABIES SHOT
Dog registration to the city and a rabies shot are a dog owner’s legal obligation.
Fee/ ¥3,550 (In case that you registered your dog) ※ Please bring “狂犬病予防注射済票交付申請書” (a blue sheet) that Hamura City already sent in May. Also bring the question sheet that Hamura City sent beforehand after filling out it.
| DATE | PLACE | TIME |
|------------|-----------|---------------|
| Apr. 16th (Thu) | Azuma Kaikan | 9:50am~10:25am |
| Chuokan | 10:40am~11:05am |
| DATE | PLACE | TIME |
|------------|-----------|---------------|
| Apr. 17th (Fri) | Mitsuya Kaikan | 9:50am~10:35am |
| Shinmeidai Kaikan | 10:50am~11:15am |
Contact about details to Kankyo Hozenka Kankyo Hozen gakari
PC Seminar (April ~ May)
Place/ Sangyo Fukushi Center Capacity/ 10 people each course (on a first-come first-served basis) Application/ Please come to Sangyo Fukushi Center directly with the tuition fee & textbook charge. ©Companies in Hamura → Apr. 7th (Tue) 9:00am ~ ©People living or working in Hamura → Apr. 8th (Wed) 9:00am ~ Detail/ Sangyo Fukushi Center (☎579-6425) ※ Windows XP and Office2003 will be used. ※ The certification exams for the Microsoft Office are carried out every week. Please check the website. ※ This seminar will be held in Japanese.
| COURSE | DATE & TIME | FEE | TEXTBOOK |
|-------------------------------|--------------------------------------------------|-------|----------|
| WORD BEGINNERS (2 or 3 days) | Apr. 11th (Sat), 12th (SUN) 1:00pm~4:00pm | ¥3,000| ¥2,100 |
| | Apr. 14th (Tue)~16th (Thu) 10:00am~noon | | |
| | May. 5th (Tue), 6th (Wed) 1:00pm~4:00pm | | |
| WORD PRACTICAL (2 days) | May. 23rd (Sat), 24th (Sun) 1:00pm~4:00pm | ¥3,000| ¥2,100 |
| EXCEL BEGINNERS (2 or 3 days)| Apr. 21st (Tue)~23rd (Thu) 10:00am~noon | ¥3,000| ¥2,100 |
| | Apr. 25th (Sat), 26th (Sun) 1:00pm~4:00pm | | |
| | May. 12th (Tue), 13th (Wed) 1:00pm~4:00pm | | |
| EXCEL PRACTICAL(2 days) | May. 30th (Sat), 31st (Sun) 1:00pm~4:00pm | ¥3,000| ¥2,100 |
| DIGITAL CAMERA EDITTING (1 day)| Apr. 18th (Sat) 1:00pm~5:00pm | ¥2,000| ¥1,764 |
| HOMEPAGE BUILDER 12 | May. 9th (Sat), 10th (Sun) 1:00pm~5:00pm | ¥5,000| ¥2,100 |
Yutorogi
Yutorogi is closed on Mondays. (When a Monday falls upon a national holiday, it's open.)
Open Hour/ 9:00pm ~ 5:00pm
1-11-5 Midorigaoka, Hamura City, Tokyo ☎205-0003 ☎570-0707 Fax: 570-6422
✉firstname.lastname@example.org http://www.hamura-tokyo.jp/
<A PICTURE BOOK ARTIST "ITO HIROSHI": ORIGINAL PICTURES SHOW>
Please enjoy beautiful original pictures different from his picture books!
Content/ “Rurarusan no Jitensya”, “Osaru no Manekko”
Date & Time/ Apr. 15th (Wed) ~26th (Sun) 10:00am~5:00pm (~4:00pm on the last day)
Place/ Tenjishitsu Fee/ Free ※ Come to the venue directly, please.
<ITO HIROSHI: CHILDREN'S LITERATURE LECTURE>
~TO GROW, READ, MAKE A PICTURE BOOK~
Ito Hiroshi is a well-known picture book artist. In front of his original pictures, we can listen to his lecture about a picture book.
Date & Time/ Apr. 18th (Sat) Reading by Volunteers/ 1:30pm~
Lecture/ 2:00pm~4:00pm (No Admission to Preschoolers) Place/ Tenjishitsu on the 1st floor of Yutorogi Child Care Service/ Please call or come to City Library directly between Apr.1st (Wed) 10:00am~12th (Sun) 5:00pm. Eligible/ 18 month-year-old children ~ preschoolers whose parents live or work in Hamura. (Capacity/ 10 children…on a first-come first-served basis) Fee/
< MOTHERS CLASS 1: “HERBS”~FOR RELAXATION >
Let’s learn about the kinds of herbs and what kind of effects they have! Have a little relaxing time in your busy days as mother.
Date & Time/ Jun. 24th (Wed) 9:30am ~11:30am
Place/ Sosakushitsu 1
Eligible/ Mothers having children
Capacity/ 25 people on a first-come first-served basis)
Fee/ ¥500
Child Care Service/ ¥500 (including snack charge)
Application/ Please call or come to Yutorogi directly between Apr. 2nd (Thu) 9:00am~Jun.10th (Wed) 5:00pm.
HOKEN CENTER
< MEDICAL CHECKUP for STOMACH & LUNG CANCER >
Date & Time/ ① Apr. 26th (Sun) 7:30am~
② May. 9th (Sat) 7:30am~
Place/ Hoken Center
Eligible/ Citizens who are 40 years old and over
Capacity/ 100 people (each dates)
Fee/ Free
Contents/ Checkup for stomach & lung cancer ※ Exam of sputum for people having bloody phlegm or smokers.
Application/ Call or come to Hoken Center directly from Apr. 2nd (Thu) 8:30am.
< MEDICAL CHECKUP for PREGNANT WOMEN >
● The numbers of public aid for pregnant women’s medical checkup will be expanded from 5 times to 14 times. Additional tickets are available at Hoken Center for women who have already submitted pregnancy report. (The maternity record book is necessary.)
● Checkup at hospitals in their home towns & maternity home outside of Hamura are eligible to public aid. (Call Hoken Center about how to apply & so on.)
Period for Application/ Within 1 years after childbirth.
Eligible/ Pregnant women in Hamura ※ Acceptance starts from Apr. 1st (Wed)
Monday to Friday 8:30am~5:00pm At Hoken Center
< VACCINATION for MEASLES & GERMAN MEASLES >
The Notices were sent to the following eligible people.
① People who were born during Apr. 2nd 2003 ~ Apr. 1st 2004
② People who were born during Apr. 2nd 1996 ~ Apr. 1st 1997
③ People who were born during Apr. 2nd 1991 ~ Apr. 1st 1992
Detail/ Hoken Center 555-1111(the pilot number of City Hall)
SPORT CENTER
< LET’S BEGIN! SIMPLE EXERCISE >
Date & Time/ May. 8th, 15th, 22nd, 29th (all are Fridays, 4 times in all) 2:00pm~3:00pm
Place/ Training Room
Capacity/ 25 people at the age of 50~69 living or working in Hamura
Fee/ ¥400 (for 4 times in all)
Necessary Things/ indoor shoes, drink, towel etc.
Instructor/ Sports trainer from NPO Corporation Hamura City Athlete Association
Application & Detail/ Call NPO Corporation Hamura City Athlete Association (555-1698) from Apr. 3rd (Fri) 9:00am.
<ACCEPTANCE for USING SCHOOL GROUNDS AT NIGHT (for MAY)>
Acceptance Term/ Apr. 15th (Wed) ~25th (Sat) At Sport Center Facilities/ Hamura Higashi elementary school (羽村東小学校), Hamura Daisan junior high school (羽村第三中学校)
Detail/ Sport Center (☎555-0033)
SWIMMING CENTER
<THE 1ST WALKING CLASS IN WATER>
Let’s practice various ways to walk in water!
Date & Time/ Apr. 22nd (Wed) 1:00pm~1:50pm Place/ Swimming Pool
Capacity/ 30 people in 16 years of age (on a first-come first-served basis) Fee/ ¥400→people Living or working in Hamura ¥600→people outside of Hamura (including premium & the entrance fee) ※ Please bring your swimming suits, cap, towel etc. Instructor/ Sports trainer from NPO Corporation Hamura City Athlete Association Application/ at the reception desk of Swimming Center on that day during noon~0:50pm Detail/ Swimming Center (☎ 579-3210)
JIDOKAN, CHILDREN’S CENTER
<WANTED: FRIENDS, KIRAKIRA HIROBA>
Play square for Infants under kindergarten.
Activity Term/ May ~February of 2010 (about 18 times) Time & Place/ Chuo Jidokan & Higashi Jidokan · Tuesday, Nishi Jidokan · Thursday 10:30am~11:30am Eligible/ infants at the age Of full 2 ~under kindergarten & their guardians (only participants for the first time)
Capacity/ 20 pairs for each center (When applicants are too many, it’s decided by open lot.)
Fee a Year/ ¥2,500 Leaflet Distribution/ Apr. 1st (Wed) 9:00am~12th (Sun) 5:00pm Contact about details to Chuo Jidokan (☎554-4552), Higashi Jidokan(☎570-7751) Nishi Jidokan(☎554-7578)
<LET’S MAKE SOME TOYS USING ACORNS>
Date/ Apr. 11th (Sat) 10:00am~ 1:00pm~ Place/ Higashi Jidokan Eligible/ lower graders at elementary schools Fee/ free Instructor/ a volunteer group “Nejimawashi” Application/ Please come to Higashi Jidokan directly. Contact/ Higashi Jidokan (☎570-7751)
★ MEDICAL CARE SERVICES AVAILABLE ON SUNDAYS & HOLIDAYS (the first half of Apr.)
| DATE | HOSPITAL | DENTIST |
|------|----------|---------|
| | 9:00am~5:00pm | 5:00pm~10:00pm | 9:00am~5:00pm |
| Apr. 5 | Sun | Wakakusa Iin | Heijitsu Yakan Kyukan Center | Okikura Shika |
| 12 | Sun | Ozaki Clinic | Fussa-shi Hoken Center | Ozawa shika Clinic |
★ Clinics on duty might change abruptly. So first of all, call City Hall or Hoken Center and make sure which clinic is exactly on duty on that day, please.
★ For hours but above-stated ☎03-5272-0303(24 hours) Tokyoto Iryo Kikan Annai Service “Himawari” Detail/ Hoken Center (☎555-1111, the pilot number of City Hall) |
Generating Polynomial Invariants with DISCOVERER and QEPCAD*
Yinghua Chen\textsuperscript{1}, Bican Xia\textsuperscript{1}, Lu Yang\textsuperscript{2}, and Naijun Zhan\textsuperscript{3,}\textsuperscript{**}
\textsuperscript{1} LMAM & School of Mathematical Sciences, Peking University
\textsuperscript{2} Institute of Theoretical Computing, East China Normal University
\textsuperscript{3} Lab. of Computer Science, Institute of Software, Chinese Academy of Sciences
South Fourth Street, No. 4, Zhong Guan Cun, Beijing, 100080, P.R. China
email@example.com
Dedicated to Prof. Chaochen Zhou on his 70th Birthday
Abstract. This paper investigates how to apply the techniques on solving semi-algebraic systems to invariant generation of polynomial programs. By our approach, the generated invariants represented as a semi-algebraic system are more expressive than those generated with the well-established approaches in the literature, which are normally represented as a conjunction of polynomial equations. We implement this approach with the computer algebra tools DISCOVERER and QEPCAD\textsuperscript{†}. We also explain, through the complexity analysis, why our approach is more efficient and practical than the one of [17] which directly applies first-order quantifier elimination.
Keywords: Program Verification, Invariant Generation, Polynomial Programs, Semi-Algebraic Systems; Quantifier Elimination, DISCOVERER, QEPCAD.
1 Introduction
Loop invariant generation together with loop termination analysis of programs plays a central role in program verification. Since the late sixties (or early seventies) of the 20th century when the so-called Floyd-Hoare-Dijkstra inductive assertion method, the dominant method on automatic verification of programs, [11,14,9] was invented, there have been lots of attempts to handle the loop problems, e.g. [25,13,16,15], but only with a limited success.
Recently, due to the advance of computer algebra, several methods based on symbolic computation have been applied successfully to invariant generation, for example the techniques based on abstract interpretation [7,1,21,6], quantifier elimination [5,17] and polynomial algebra [19,20,22,23,24].
The basic idea behind the abstract interpretation approaches is to perform an approximate symbolic execution of a program until an assertion is reached
* This work is supported in part by NKBRPC-2002-b312200, NKBRPC-2004CB318003, NSFC-60493200, NSFC-60421001, NSFC-60573007, and NKBRPC-2005CB321902.
** Corresponding author.
† \url{http://www.cs.usna.edu/~qepcad/B/QEPCAD.html}
that remain unchanged by further executions of the program. However, in order to guarantee termination, the method introduces imprecision by use of an extrapolation operator called *widening/narrowing*. This operator often causes the technique to produce weak invariants. Moreover, proposing widening/narrowing operators with certain concerns of completeness is not easy and becomes a key challenge for abstract interpretation based techniques [7,1].
In contrast, [19,20,22,23,24] exploited the theory of polynomial algebra to discover invariants of polynomial programs. [19] applied the technique of linear algebra to generate polynomial equations of bounded degree as invariants of programs with affine assignments. [22,23] first proved that the set of polynomials serving as loop invariants has the algebraic structure of an ideal, then proposed an invariant generation algorithm by using fixpoint computation, and finally implemented the algorithm by the Gröbner bases and the elimination theory. The approach is theoretically sound and complete in the sense that if there is an invariant of the loop that can be expressed as a conjunction of polynomial equations, applying the approach can indeed generate it. [24] presented a similar approach to finding polynomial equation invariants whose form is priori determined (called templates) by using an extended Gröbner basis algorithm over templates.
Compared with the polynomial algebraic approaches that can only generate invariants represented as polynomial equations, [5] proposed an approach to generate linear inequalities as invariants for linear programs, based on *Farkas’ Lemma* and non-linear constraint solving. In addition, [17] proposed a very general approach for automatic generation of more expressive invariants by exploiting the technique of quantifier elimination, and applied the approach to Presburger Arithmetic and quantifier-free theory of conjunctively closed polynomial equations. Theoretically speaking, the approach can also be applied to the theory of real closed fields, but [17] pointed out that this is impractical in reality because of the high complexity of quantifier elimination, which is double exponential [8]. To handle the problem, [6] exploited the techniques of parametric abstraction, Lagrangian relaxation and semidefinite programming to generate invariants as well as ranking functions of polynomial programs. Compared with the approach of [17], [6]’s is more efficient, as first-order quantifier elimination is not directly applied there. However, [6]’s approach is incomplete in the sense that, for some program that may have ranking functions and invariants of the predefined form, applying the approach may not be able to find them, as Lagrangian relaxation and over-approximation of the positive semi-definiteness of a polynomial are used.
In this paper, we attack the problem raised in [17] on how to efficiently generate polynomial invariants of programs over real closed fields and present a more practical and efficient approach to it by exploiting our results on solving semi-algebraic systems (SASs). The outline of our approach is as follows: we first reduce polynomial invariant generation problem to solving semi-algebraic systems; then apply our theories and tools on solving SASs, in particular, on root classification of parametric SASs [30,31,32] and real root isolation of constant
SASs [28, 29], to produce some necessary and sufficient conditions; and finally utilize the technique of quantifier elimination to handle the derived conditions and obtain invariants with the predefined form.
Suppose an SAS $S$ has $s$ ($> 0$) polynomial equations and $m$ inequations and inequalities. All polynomials are in $n = t + (n - t)$ indeterminates (i.e., $u_1, \ldots, u_t$, $x_1, \ldots, x_{n-t}$) and of degree at most $d$ where $t$ is the dimension of the ideal generated by the $s$ equations. According to [3, 8], directly applying the technique of quantifier elimination of real closed fields, the cost for solving $S$ is doubly exponential w.r.t. $n$. But using our approach, the cost for the first step is almost nothing, and the second step to apply root classification and isolation costs singly exponential w.r.t. $n$ plus doubly exponential w.r.t. $t$. The cost for the last step is also doubly exponential w.r.t. $t$. Therefore, as $t < n$, our approach is more efficient than the approaches directly based on the techniques of quantifier elimination and Gröbner basis, such as [17, 24], in particular, when $t$ is much less than $n$. Moreover, our approach is still complete in the sense that whenever there exist invariants of the predefined form, applying our approach can indeed synthesize them, while [6]'s\footnote{As far as efficiency is concerned, we believe that our approach could be at least as good as [6]'s, as the complexity of the semi-definite programming adopted in [6] is also very high.} is incomplete. On the other hand, similarly to [17, 6], invariants generated by our approach are more expressive, while applying the approaches based on polynomial algebra can only produce conjunction of polynomial equations as invariants.
The rest of this paper is organized as: Section 2 provides a brief review of the theories and tools on solving SASs; Section 3 defines some basic notions, including semi-algebraic transition systems, polynomial programs, invariants, inductive properties and so on; Section 4 is devoted to illustrating our approach in detail with a running example; We provide the complexity analysis of our approach in Section 5; In Section 6, we compare the application of this approach to invariant generating with the one to ranking function discovering; and Section 7 concludes the paper and discusses the future work in this direction.
## 2 Preliminaries: Theories on Semi-algebraic Systems
In this section, we introduce the cornerstone of our technique, i.e. theories and tools on solving SASs, mainly the theories on root classification of parametric SASs and the tool DISCOVERER.
### 2.1 Basic Notions
Let $\mathcal{K}[x_1, ..., x_n]$ be the ring of polynomials in $n$ indeterminates, $X = \{x_1, \cdots, x_n\}$, with coefficients in the field $\mathcal{K}$. Let the variables be ordered as $x_1 \prec x_2 \prec \cdots \prec x_n$. Then, the leading variable (or main variable) of a polynomial $p$ is the variable with the biggest index which indeed occurs in $p$. If the leading variable of a polynomial $p$ is $x_k$, $p$ can be collected w.r.t. its leading variable as $p = c_m x_k^m + \cdots + c_0$
where \( m \) is the degree of \( p \) w.r.t. \( x_k \) and \( c_i \)'s are polynomials in \( K[x_1, \ldots, x_{k-1}] \). We call \( c_m x_k^m \) the leading term of \( p \) w.r.t. \( x_k \) and \( c_m \) the leading coefficient. For example, let \( p(x_1, \ldots, x_5) = x_2^3 + x_3^4 x_2^2 + (2x_2 + x_1)x_4^3 \), so, its leading variable, term and coefficient are \( x_4 \), \( (2x_2 + x_1)x_4^3 \) and \( 2x_2 + x_1 \), respectively.
An atomic polynomial formula over \( K[x_1, \ldots, x_n] \) is of the form \( p(x_1, \ldots, x_n) \triangleright 0 \), where \( \triangleright \in \{=, >, \geq, \neq\} \), while a polynomial formula over \( K[x_1, \ldots, x_n] \) is constructed from atomic polynomial formulae by applying the logical connectives. Conjunctive polynomial formulae are those that are built from atomic polynomial formulae with the logical operator \( \land \). We will denote by \( PP(\{x_1, \ldots, x_n\}) \) the set of polynomial formulae and by \( CPF(\{x_1, \ldots, x_n\}) \) the set of conjunctive polynomial formulae, respectively.
In what follows, we will use \( \mathbb{Q} \) to stand for rationales and \( \mathbb{R} \) for reals, and fix \( K \) to be \( \mathbb{Q} \). In fact, all results discussed below can be applied to \( \mathbb{R} \).
In the following, the \( n \) indeterminates are divided into two groups: \( u = (u_1, \ldots, u_t) \) and \( x = (x_1, \ldots, x_s) \), which are called parameters and variables, respectively, and we sometimes use “,” to denote the conjunction of atomic formulae for simplicity.
**Definition 1.** A semi-algebraic system is a conjunctive polynomial formula of the following form:
\[
\begin{cases}
p_1(u, x) = 0, \ldots, p_r(u, x) = 0, \\
g_1(u, x) \geq 0, \ldots, g_k(u, x) \geq 0, \\
g_{k+1}(u, x) > 0, \ldots, g_l(u, x) > 0, \\
h_1(u, x) \neq 0, \ldots, h_m(u, x) \neq 0,
\end{cases}
\]
where \( r > 1, l \geq k \geq 0, m \geq 0 \) and all \( p_i \)'s, \( g_i \)'s and \( h_i \)'s are in \( \mathbb{Q}[u, x] \setminus \mathbb{Q} \). An SAS of the form (1) is called parametric if \( t \neq 0 \), otherwise constant.
An SAS of the form (1) is usually denoted by a quadruple \([P, G_1, G_2, H]\), where \( P = [p_1, \ldots, p_r], G_1 = [g_1, \ldots, g_k], G_2 = [g_{k+1}, \ldots, g_l] \) and \( H = [h_1, \ldots, h_m] \).
For a constant SAS \( S \), interesting questions are how to compute the number of real solutions of \( S \), and if the number is finite, how to compute these real solutions. For a parametric SAS, the interesting problem is so-called *real solution classification*, that is to determine the condition on the parameters such that the system has the prescribed number of distinct real solutions, possibly infinite.
### 2.2 Theories on Real Solution Classification
In this subsection, we outline the theories for real root classification of parametric SASs. For details, please be referred to [31,27].
For an SAS \( S \) of the form (1), the algorithm for real root classification consists of three main steps. Firstly, transform the equations of \( S \) into some sets of equations in triangular form. A set of equations \( T : [T_1, \ldots, T_k] \) is said to be in triangular form (or a *triangular set*) if the main variable of \( T_i \) is less in the order than that of \( T_j \) for \( i < j \). Roughly speaking, if we rename the variables, a triangular set looks like
\[
\begin{align*}
T_1 &= a_1 x_1 + b_1 x_2 + \cdots + d_1 x_s, \\
T_2 &= a_2 x_2 + b_2 x_3 + \cdots + d_2 x_s, \\
&\vdots \\
T_k &= a_k x_k + b_k x_{k+1} + \cdots + d_k x_s,
\end{align*}
\]
\[ T_1 = T_1(\mathbf{v}, y_1), \]
\[ T_2 = T_2(\mathbf{v}, y_1, y_2), \]
\[ \ldots \ldots \]
\[ T_k = T_k(\mathbf{v}, y_1, \ldots, y_k), \]
where \( \mathbf{v} \) are the indeterminates other than \( y_i \). It is obvious that we now only need to consider triangular systems in the following form
\[
\begin{cases}
f_1(\mathbf{u}, x_1) = 0, \\
\vdots \\
f_s(\mathbf{u}, x_1, \ldots, x_s) = 0, \\
\mathbb{G}_1, \mathbb{G}_2, \mathbb{H}.
\end{cases}
\] (2)
Certainly, it can be proven that there exists a correspondence between the solutions of these triangular sets and \( S \)'s so that we only need to consider the solutions of these triangular sets in order to deal with \( S \)'s.
**Example 1.** Consider an SAS \( S : [\mathbb{P}, \mathbb{G}_1, \mathbb{G}_2, \mathbb{H}] \) in \( \mathbb{Q}[b, x, y, z] \) with \( \mathbb{P} = [p_1, p_2, p_3], \mathbb{G}_1 = \emptyset, \mathbb{G}_2 = [x, y, z, b, 2 - b], \mathbb{H} = \emptyset \), where
\[ p_1 = x^2 + y^2 - z^2, \quad p_2 = (1 - x)^2 - z^2 + 1, \quad p_3 = (1 - y)^2 - b^2 z^2 + 1. \]
The equations \( \mathbb{P} \) can be decomposed into two triangular sets in \( \mathbb{Q}(b)[x, y, z] \)
\[ T_1 : [b^4 x^2 - 2b^2(b^2 - 2)x + 2b^4 - 8b^2 + 4, -b^2 y + b^2 x + 2 - 2b^2, b^4 z^2 + 4b^2 x - 8b^2 + 4], \]
\[ T_2 : [x^2 - 2x + 2, y + x - 2, z], \]
with the relation
\[ \text{Zero}(\mathbb{P}) = \text{Zero}(T_1/b) \bigcup \text{Zero}(T_2) \]
where \( \text{Zero}() \) means the set of zeros and \( \text{Zero}(T_1/b) = \text{Zero}(T_1) \setminus \text{Zero}(b) \).
Second, compute a so-called *border polynomial* from the resulting triangular systems, say \( [T_i, \mathbb{G}_1, \mathbb{G}_2, \mathbb{H}] \). We need to introduce some concepts. Suppose \( F \) and \( G \) are polynomials in \( x \) with degrees \( m \) and \( l \), respectively. Thus, they can be written in the following forms
\[ F = a_0 x^m + a_1 x^{m-1} + \cdots + a_{m-1} x + a_m, \quad G = b_0 x^l + b_1 x^{l-1} + \cdots + b_{l-1} x + b_l. \]
The following \( (m+l) \times (m+l) \) matrix (those entries except \( a_i, b_j \) are all zero)
\[
\begin{pmatrix}
a_0 & a_1 & \cdots & a_m \\
a_0 & a_1 & \cdots & a_m \\
\ddots & \ddots & \ddots \\
a_0 & a_1 & \cdots & a_m \\
b_0 & b_1 & \cdots & b_l \\
b_0 & b_1 & \cdots & b_l \\
\ddots & \ddots & \ddots \\
b_0 & b_1 & \cdots & b_l
\end{pmatrix}
\]
is called the *border matrix* of \( F \) and \( G \).
is called the *Sylvester matrix* of $F$ and $G$ w.r.t. $x$. The determinant of the matrix is called the *Sylvester resultant* or *resultant* of $F$ and $G$ w.r.t. $x$ and is denoted by $\text{res}(F, G, x)$.
For system (2), we compute the resultant of $f_s$ and $f'_s$ w.r.t. $x_s$ and denote it by $\text{dis}(f_s)$ (it has the leading coefficient and discriminant of $f_s$ as factors). Then we compute the *successive resultant* of $\text{dis}(f_s)$ and the triangular set $\{f_{s-1}, ..., f_1\}$. That is, we compute $\text{res}(\text{res}(\cdots \text{res}(\text{res}(\text{dis}(f_s), f_{s-1}, x_{s-1}), f_{s-2}, x_{s-2}) \cdots ), f_1, x_1)$ and denote it by $\text{res}(\text{dis}(f_s); f_{s-1}, ..., f_1)$ or simply $R_s$. Similarly, for each $i$ ($1 < i \leq s$), we compute $R_i = \text{res}(\text{dis}(f_i); f_{i-1}, ..., f_1)$ and $R_1 = \text{dis}(f_1)$.
For each of those inequalities and inequations, we compute the successive resultant of $g_j$ (or $h_j$) w.r.t. the triangular set $[f_1, ..., f_s]$ and denote it by $Q_j$ (resp. $Q_{i+j}$).
**Definition 2.** For an SAST $T$ as defined by (2), the border polynomial of $T$ is
$$BP = \prod_{i=1}^{s} R_i \prod_{j=1}^{l+m} Q_j.$$
Sometimes, for brevity, we also abuse $BP$ to denote the square-free part or the set of square-free factors of $BP$.
**Example 2.** For the system $S$ in Example 1, the border polynomial is
$$BP = b(b - 2)(b + 2)(b^2 - 2)(b^4 - 4b^2 + 2)(2b^4 - 2b^2 + 1).$$
From the result in [31, 27], we may assume $BP \neq 0$. In fact, if any factor of $BP$ is a zero polynomial, we can further decompose the system into new systems with such a property. For a parametric SAS, its border polynomial is a polynomial in the parameters with the following property.
**Theorem 1.** Suppose $S$ is a parametric SAS as defined by (2) and $BP$ its border polynomial. Then, in each connected component of the complement of $BP = 0$ in parametric space $\mathbb{R}^d$, the number of distinct real solutions of $S$ is constant.
Third, $BP = 0$ decomposes the parametric space into a finite number of connected region. We then choose sample points in each connected component of the complement of $BP = 0$ and compute the number of distinct real solutions of $S$ at each sample point. Note that sample points can be obtained by the partial cylindrical algebra decomposition (PCAD) algorithm [4].
**Example 3.** For the system $S$ in Example 1, $BP = 0$ gives
$$b = 0, \pm 2, \pm \sqrt{2}, \pm \sqrt{2 \pm \sqrt{2}}.$$
The reals are divided into ten open intervals by these points. Because $0 < b < 2$, we only need to choose one point, for example, $\frac{1}{2}, 1, \frac{3}{2}, \frac{15}{8}$, from each of the four intervals contained in $(0, 2)$, respectively. Then, we substitute each of the four values for $b$ in the system, and compute the number of distinct real solutions of the system, consequently obtain the system has respectively 0, 1, 0 and 0 distinct real solutions.
The above three steps constitute the main part of the algorithm in [31, 34, 27], which, for any input SAS $S$, outputs the so-called border polynomial $BP$ and a quantifier-free formula $\Psi$ in terms of polynomials in parameters $u$ (and possible some variables) such that, provided $BP \neq 0$, $\Psi$ is the necessary and sufficient condition for $S$ to have the given number (possibly infinite) of real solutions. Since $BP$ is a polynomial in parameters, $BP = 0$ can be viewed as a degenerated condition. Therefore, the outputs of the above three steps can be read as “if $BP \neq 0$, the necessary and sufficient condition for $S$ to have the given number (possibly infinite) of real solutions is $\Psi$.”
**Remark 1.** If we want to discuss the case when parameters degenerate, i.e., $BP = 0$, we put $BP = 0$ (or some of its factors) into the system and apply a similar procedure to handle the new SAS.
**Example 4.** By the steps described above, we obtain the necessary and sufficient condition for $S$ to have one distinct real solution is $b^2 - 2 < 0 \land b^4 - 4b^2 + 2 < 0$ provided $BP \neq 0$. Now, if $b^2 - 2 = 0$, adding the equation into the system, we obtain a new SAS: $[b^2 - 2, p_1, p_2, p_3], [\ ]$, $G_2, [\ ]$. By the algorithm in [28, 29], we know the system has no real solutions.
### 2.3 A Computer Algebra Tool: DISCOVERER
We have implemented the above algorithm and some other algorithms in Maple as a computer algebra tool, named DISCOVERER. The reader can download the tool for free via “http://www.is.pku.edu.cn/~xbc/discoverer.html”. The prerequisite to run the package is Maple 7.0 or a later version of it.
The main features of DISCOVERER include
**Real Solution Classification of Parametric Semi-algebraic Systems**
For a parametric SAS $T$ of the form (1) and an argument $N$, where $N$ is one of the following three forms:
- a non-negative integer $b$;
- a range $b..c$, where $b, c$ are non-negative integers and $b < c$;
- a range $b..w$, where $b$ is a non-negative integer and $w$ is a name without value, standing for $+\infty$.
DISCOVERER can determine the conditions on $u$ such that the number of the distinct real solutions of $T$ equals to $N$ if $N$ is an integer, otherwise falls in the scope $N$. This is by calling
$$\text{tofind}([P], [G_1], [G_2], [H], [x_1, ..., x_s], [u_1, ..., u_t], N),$$
and results in the necessary and sufficient condition as well as the *border polynomial* $BP$ of $T$ in $u$ such that the number of the distinct real solutions of $T$ exactly equals to $N$ or belongs to $N$ provided $BP \neq 0$. If $T$ has infinite real solutions for generic value of parameters, $BP$ may have some variables. Then, for the “boundaries” produced by “tofind”, i.e. $BP = 0$, we can call
$$\text{Tofind}([P, BP], [G_1], [G_2], [H], [x_1, ..., x_s], [u_1, ..., u_t], N)$$
to obtain some further conditions on the parameters.
Real Solution Isolation of Constant Semi-algebraic Systems
For a constant SAS $T$ (i.e., $t = 0$) of the form (1), if $T$ has only a finite number of real solutions, DISCOVERER can determine the number of distinct real solutions of $T$, say $n$, and moreover, can find out $n$ disjoint cubes with rational vertices in each of which there is only one solution. In addition, the width of the cubes can be less than any given positive real. The two functions are realized through calling
\[
\text{nearsolve}([P], [G_1], [G_2], [H], [x_1, ..., x_n]) \quad \text{and} \\
\text{realzeros}([P], [G_1], [G_2], [H], [x_1, ..., x_n], w),
\]
respectively, where $w$ is optional and used to indicate the maximum size of the output cubes.
3 Semi-algebraic Transition Systems and Invariants
In this section, we extend the notion of algebraic transition systems in [24] to semi-algebraic transition systems (SATSs) to represent polynomial programs. An Algebraic Transition System (ATS) is a special case of standard transition system, in which the initial condition and all transitions are specified in terms of polynomial equations; while in an SATS, each transition is equipped with a conjunctive polynomial formula as guard, and its initial and loop conditions possibly contain polynomial inequations and inequalities. It is easy to see that ATS is a special case of SATS. Formally,
**Definition 3.** A semi-algebraic transition system is a quintuple $\langle V, L, T, \ell_0, \Theta \rangle$, where $V$ is a set of program variables, $L$ is a set of locations, and $T$ is a set of transitions. Each transition $\tau \in T$ is a quadruple $\langle \ell_1, \ell_2, \rho_\tau, \theta_\tau \rangle$, where $\ell_1$ and $\ell_2$ are the pre- and post- locations of the transition, $\rho_\tau \in CPF(V, V')$ is the transition relation, and $\theta_\tau \in CPF(V)$ is the guard of the transition. Only if $\theta_\tau$ holds, the transition can take place. Here, we use $V'$ (variables with prime) to denote the next-state variables. The location $\ell_0$ is the initial location, and $\Theta \in CPF(V)$ is the initial condition.
If a transition $\tau$ changes nothing, i.e. $\rho_\tau \equiv \bigwedge_{v \in V} v' = v$, we denote by $\text{skip} \ \rho_\tau$. Meanwhile, a transition $\tau = \langle \ell_1, \ell_2, \rho_\tau, \theta_\tau \rangle$ is abbreviated as $\langle \ell_1, \ell_2, \rho_\tau \rangle$, if $\theta_\tau$ is true.
Note that in the above definition, for simplicity, we require that each guard should be a conjunctive polynomial formula. In fact, we can drop such a restriction, as for any transition with a disjunctive guard we can split it into multiple transitions with the same pre- and post- locations and transition relation, but each of which takes a disjunct of the original guard as its guard.
A state is an evaluation of the variables in $V$ and all states are denoted by $Val(V)$. Without confusion we will use $V$ to denote both the variable set and an arbitrary state, and use $F(V)$ to mean the (truth) value of function (formula) $F$ under the state $V$. The semantics of SATSs can be explained through state transitions as usual.
A transition is called *separable* if its transition relation is a conjunctive formula of equations which define variables in $V'$ equal to polynomial expressions over
variables in $V$. It is easy to see that the composition of two separable transitions is equivalent to a single separable one. An SATS is called \textit{separable} if each transition of the system is separable. In a separable system, the composition of transitions along a path of the system is also equivalent to a single separable transition. We will only concentrate on separable SATSs as any polynomial program can easily be represented by a separable SATS (see [18]). Any SATS in the rest of the paper is always assumed separable, unless otherwise stated.
Informally, an \textit{invariant} of a program at a location is an assertion that is \textit{true} under any program state reaching the location. An invariant of a program can be seen as a mapping to map each location to an assertion which has inductive property, that is, \textit{initial} and \textit{consecutive}. \textit{Initial} means that the image of the mapping at the initial location holds on the loop entry, i.e. the invariant of the initial location holds on the loop entry; whereas \textit{consecutive} means that for any transition the invariant at the pre-location together with the transition relation and its guard implies the invariant at the post-location. In many cases, people only consider an invariant at the initial location and do not care about invariants at other locations. In this case, we can assume the invariants at other locations are all \textit{true} and therefore \textit{initial} and \textit{consecutive} mean that the invariant holds on the entry, and is preserved under every cycle back to the initial location.
\textbf{Definition 4 (Invariant at a Location).} Let $P = \langle V, L, T, l_0, \Theta \rangle$ be an SATS. An invariant at a location $l \in L$ is a conjunctive polynomial formula $\phi \in PF(V)$, such that $\phi$ holds on all states that can be reached at location $l$.
\textbf{Definition 5 (Invariant of a Program).} An assertion map for an SATS $P = \langle V, L, T, l_0, \Theta \rangle$ is a map $\eta: L \mapsto PF(V)$ that associates each location of $P$ with a formula of $PF(V)$. An assertion map of $P$ is said to be an invariant of $P$ iff the following conditions hold:
\textbf{Initial:} $\Theta(V_0) \models \eta(l_0)$.
\textbf{Consecutive:} For each transition $\tau = \langle l_i, l_j, \rho_\tau, \theta_\tau \rangle$,
$$\eta(l_i)(V) \land \rho_\tau(V, V') \land \theta_\tau(V) \models \eta(l_j)(V').$$
\section{Polynomial Invariants Generation}
Similarly to [24], given an SATS $S$, we predetermine an invariant as a parametric SAS (PSAS for short) at each of the underlining locations (if no invariant is predefined for a location, it is assumed that the mapping takes \textit{true} as value at the location) and therefore all these predefined PSASs form a parametric invariant of $S$ by the Definitions. Subsequently, according to the initial and consecutive conditions of the mapping, we can obtain a set of PSASs such that the mapping is an invariant of the program iff each element the resulted set has no real solution. Afterwards, we apply the algorithm on root classification of PSASs to each of them and obtain a corresponding necessary and sufficient condition on the parameters of the PSAS such that the PSAS has no real solution. Finally, applying quantifier elimination technique, we can get the instantiations of
these parameters and therefore get an invariant for each underlining location by replacing with the resulted instantiations the parameters of the predetermined parametric PSAS. The above procedure are supported by the computer algebra tools DISCOVERER and QEPCAD.
We will use the following example to demonstrate our approach in details.
**Example 5.** Consider a program shown in Fig.1 (a).
\[
P = \{
V = \{x, y\} \\
L = \{l_0\} \\
T = \{\tau\}
\}
\text{where}
\tau = \langle l_0, l_0, x' - x - y^2 = 0 \land y' - y - 1 = 0, \\
x \geq 0 \land y \geq 0 \rangle
\]
(a) (b)
**Fig. 1.**
Thus, the corresponding SATS can be represented as in Fig.1 (b).
In the following, we concretize the above idea and demonstrate with the toy example.
**Predefining Invariant.** Predetermine a template of invariants at each of the underlining location, which is a PSAS, i.e. the conjunction of a set of atomic polynomial formula. All of these predefined PSASs form a parametric invariant of the program. For example, we can assume a template of invariants of $P$ at $l_0$ in Example 5 as
\begin{align*}
eq(x, y) &= a_1y^3 + a_2y^2 + a_3x - a_4y = 0 \tag{3} \\
ineq(x, y) &= b_1x + b_2y^2 + b_3y + b_4 > 0, \tag{4}
\end{align*}
where $a_1, a_2, a_3, a_4, b_1, b_2, b_3, b_4$ are parameters. Therefore, $\eta(l_0) = (3) \land (4)$.
Note that theoretically speaking we can predefine a PSAS as an invariant at each location like in the above example, but this will raise the complexity dramatically as thus the number of parameters is so large (the reader will see this point from the complexity analysis in the later). In practice, alternatively, we will split a complicated invariant to several simple invariants such that the image of every of these simple invariants at each location is just one of the atomic subformulae of the image of the complicated invariant at the location. For example, in the above example, we can split $\eta$ to $\eta_1$ and $\eta_2$ by letting $\eta_1(l_0) = (3)$ and $\eta_2(l_0) = (4)$. It is easy to prove that a program has a complicated invariant iff the corresponding simple invariant exist, for instance, $\eta$ exists iff $\eta_1$ and $\eta_2$ exist. This is because every invariant of a program is determined by the program itself.
**Deriving PSASs from Initial Condition and Solving.** According to the initial condition in Definition 5, we have $\Theta \models \eta(l_0)$ which means that each real solution of $\Theta$ must satisfy $\eta(l_0)$. In other words, $\Theta \land \neg \eta(l_0)$ has no real
solution. This implies that for each atomic polynomial formula $\phi$ in $\eta(l_0)$,
$\Theta \land \neg \phi$ has no real solution. Note that $\eta(l_0)$ is the conjunction of a set of
atomic polynomials and $\Theta \land \neg \phi$ is a PSAS according to the definition. Thus,
applying the tool DISCOVERER to the resulted PSAS $\Theta \land \neg \phi$, we get a nec-
essary and sufficient condition of the derived PSAS having no real solution.
The condition may contain the occurrences of some program variables. In
this case, the condition should hold for any instantiations of these variables.
Thus, by introducing universal quantifications of these variables (we usually
add a scope to each of these variables according to different situations) and
then applying QEPCAD, we can get a necessary and sufficient condition
only on the presumed parameters.
Repeatedly apply the procedure to each atomic polynomial formula of the
predefined invariant at $l_0$ and then collect all the resulted conditions.
**Example 6.** In Example 5, $\Theta \models \eta_1(l_0)$ is equivalent to
$$x = 0, y = 0, eq(x, y) \neq 0$$ \hspace{1cm} (5)
has no real solution. By calling
```
tofind((x, y), [], [], [eq(x, y)], [x, y], [a1, a2, a3, a4], 0)
```
we get that (5) has no real solution iff true.
Similarly, $\Theta \models \eta_2(l_0)$ is equivalent to
$$x = 0, y = 0, ineq(x, y) \leq 0$$ \hspace{1cm} (6)
has no real solution. Calling
```
tofind([x, y], [-ineq(x, y)], [], [], [x, y], [b1, b2, b3, b4], 0)
```
we get that (6) has no real solution iff
$$b_4 > 0.$$ \hspace{1cm} (7)
**Deriving PSASs from Consecutive Condition and Solving.** From Defi-
nition 5, for each transition $\tau = \langle l_i, l_j, \rho_\tau, \theta_\tau \rangle$,
$$\eta(l_i) \land \rho_\tau \land \theta_\tau \models \eta(l_j)$$
so $\eta(l_i) \land \rho_\tau \land \theta_\tau \land \neg \eta(l_j)$ has no real solution. This implies that for each
atomic polynomial formula $\phi$
$$\eta(l_i) \land \rho_\tau \land \theta_\tau \land \neg \phi$$ \hspace{1cm} (8)
has no real solution. It is easy to see that (8) is a PSAS according to De-
finition 1, so applying the tool DISCOVERER, we obtain a necessary and
sufficient condition on the parameters such that (8) has no real solution.
Subsequently, similarly to Step 2, we may need to exploit the quantifier
elimination tool QEPCAD to reduce the resulted condition in order to get
a necessary and sufficient condition only on the presumed parameters.
Example 7. In Example 5, for the invariant $\eta_1$, we have
$$eq(x, y) = 0 \land x' - x - y^2 = 0 \land y' - y - 1 = 0 \models eq(x', y') = 0.$$
(9)
This is equivalent to
$$eq(x, y) = 0 \land x' - x - y^2 = 0 \land y' - y - 1 = 0 \land eq(x', y') \neq 0$$
(10)
has no real solution. Calling
$$\text{tofind}([x' - x - y^2, y' - y - 1, eq(x, y)], [], [], [eq(x', y')], [x', y', x], [y, a_1, a_2, a_3, a_4], 0),$$
it follows that (10) has no real solution iff
$$a_3 y^2 + 3a_1 y^2 + 2ya_2 + 3a_1 y - a_4 + a_2 + a_1 = 0 \land$$
(11)
$$a_3(a_1 y^2 + ya_2 - a_4) \leq 0.$$
(12)
Further by Basic Algebraic Theorem and simplifying by QEPCAD, (11) $\land$ (12) holds for all $y$ iff
$$-a_4 + a_2 + a_1 = 0 \land 3a_1 + 2a_2 = 0 \land a_3 + 3a_1 = 0.$$
(13)
Regarding the invariant $\eta_2$, we have
$$ineq(x, y) > 0 \land x' - x - y^2 = 0 \land y' - y - 1 = 0 \models ineq(x', y') > 0.$$
(14)
This is equivalent to
$$ineq(x, y) > 0 \land x' - x - y^2 = 0 \land y' - y - 1 = 0 \land ineq(x', y') \leq 0$$
(15)
has no real solution. Calling
$$\text{tofind}([x' - x - y^2, y' - y - 1], [-ineq(x', y')], [ineq(x, y)], [], [x', y'], [x, y, b_1, b_2, b_3, b_4], 0),$$
it follows that (15) has no real solution iff
$$b_4 + b_3 + b_2 + 2b_2 y + b_3 y + b_2 y^2 + b_1 x + b_1 y^2 > 0.$$
(16)
It is easy to see that (16) should hold for all $y \geq 0$, and thus, by applying QEPCAD to eliminate the quantifiers $\forall y \geq 0$ over (16), we get
$$b_1 + b_2 \geq 0 \land b_1 \geq 0 \land b_2 + b_3 + b_4 > 0 \land$$
$$(b_3 + 2b_2 \geq 0 \lor (b_1 b_2 + b_2^2 \geq 0 \land 4b_2 b_3 + 4b_1 b_4 + 4b_1 b_3 + 4b_1 b_2 - b_2^2 > 0))$$
(17)
Generating Invariant. According to the results obtained from Steps 1, 2 and 3, we can get the final necessary and sufficient condition only on the parameters of each of the invariant templates. If the condition is too complicated, we can utilize the function of PCAD of DISCOVERER or QEPCAD to prove if or not the condition is satisfied. If yes, the tool can produce the instantiations of these parameters. Thus, we can get an invariant of the predetermined form by replacing the parameters with the instantiations, respectively.
Example 8. From Examples 6 & 7, it follows the necessary and sufficient condition on the parameters of $\eta_1$ is (13). By using DISCOVERER, we get an instantiation
$$ (a_1, a_2, a_3, a_4) = (-2, 3, 6, 1). $$
Thus, $\eta_1(l_0) = -2y^3 + 3y^2 + 6x - y = 0$. Also, the necessary and sufficient condition on the parameters of $\eta_2$ is (7) $\land$ (17). By PCAD of DISCOVERER, it results the following instantiation
$$ (b_1, b_2, b_3, b_4) = (1, -1, 2, 1) $$
that is, $\eta_2(l_0) = x - y^2 + 2y + 1 > 0$. Totally, we get the following invariant for the program $P$:
$$ \begin{cases}
-2y^3 + 3y^2 + 6x - y = 0, \\
x - y^2 + 2y + 1 > 0
\end{cases} $$
Note that the above procedure is complete in the sense that for any given predefined parametric invariant, the procedure can always give you an answer, yes or no. Therefore, we can conclude that our approach is also complete in the sense that once the given polynomial program has a polynomial invariant, our approach can indeed find it theoretically. This is because we can assume parametric invariants in program variables of different degrees, and repeatedly apply the above procedure until we obtain a polynomial invariant.
5 Complexity Analysis
Assume given an SATS $P = \langle V, L, T, l_0, \Theta \rangle$, applying the above procedure, we obtain $k$ distinct PSASs so that the predefined parametric invariants form an invariant of the program iff none of these $k$ PSASs has any real solution. W.l.o.g., suppose each of these $k$ PSASs has at most $s$ polynomial equations, and $m$ inequations and inequalities. All polynomials are in $n$ indeterminates (i.e., variables and parameters) and of degrees at most $d$.
For a PSAS $S$, by [3], CAD (cylindrical algebraic decomposition) based quantifier elimination on $S$ has complexity $O((2d)^{2^{2n+8}}(s+m)^{2^n+6})$, which is double exponential w.r.t. $n$. Thus, the total cost is $O(k(2d)^{2^{2n+8}}(s+m)^{2^n+6})$ for directly applying the technique of quantifier elimination to generate an invariant of a program as advocated by Kapur [17].
In contrast, the cost of our approach includes two parts: one is for applying real solution classification to generate condition on the parameters possibly together with some program variables; the other is for applying first-order quantifier elimination to produce condition only on the parameters (if necessary) and further exploiting PCAD to obtain the instantiations of these parameters.
The first part consists of three main steps. Firstly, we transform the equations in $S$ into triangular sets (i.e., equations in triangular form) by Ritt-Wu’s method. By [12], the complexity of computing the first characteristic set is $O(s^{O(n)}(d+1)^{O(n^3)})$. Thus, the complexity of this step is $O(s^{n^{O(1)}}(d+1)^{n^{O(1)}})$, which is usually called a singly exponential complexity w.r.t. $n$. Secondly, we
compute a *border polynomial* (BP) from the triangularized systems through resultant computation. Because the polynomials in the first computed characteristic sets are of degree $O(s(d + 1)^{O(n^2)})$ by [12], the polynomials in the computed triangular sets are of degree $O(s^{n^{O(1)}}(d + 1)^{n^{O(1)}})$. Thus, the complexity of computing BP is at most $(s + m)O(s^{3s + sn^{O(1)}}(d + 1)^{sn^{O(1)}})$ because the complexity of computing the resultant of two polynomials with degree $d$ is at most $O(d^3)$. Moreover, the degree of BP is at most $D = O(s^{O(s^2 + s^2n^{O(1)})}(d + 1)^{O(s^2n^{O(1)})})$. Finally, we use a partial CAD algorithm with BP to obtain real solution classification. The complexity of this step is at most the complexity of performing quantifier elimination on BP using CAD. Suppose the dimension of the ideal generated by the $s$ polynomial equations is $t$, then BP has at most $t$ indeterminates. Thus, by [3], the complexity of this step is at most $O(2D^{2t+8})$, which is double exponential with respect to $t$. In a word, the cost for this part is singly exponential in $n$ and doubly exponential in $t$.
As the biggest degree of polynomials in the generated necessary and sufficient condition from the above is at most $D$, the cost for the second part is $O(2D^{2t+8})$ as well, which is doubly exponential in $t$, according to [3].
So, compared to directly applying quantifier elimination, our approach can dramatically reduce the complexity, in particular when $t$ is much less than $n$.
### 6 Generating Invariants vs. Discovering Ranking Functions
In [2], we showed how to apply the approach to discovering non-linear ranking functions. Although invariants and ranking functions both have inductive properties, the former is inductive w.r.t. a small step, i.e. each of single transitions of the given program in contrast that the latter is inductive w.r.t. a big step, that is each of circle transition at the initial location of the program. The difference results that as far as invariant generation is concerned, our approach can be simply applied to single loop programs as well as nested loop programs, without any change; but regarding the discovery of ranking functions, we have to develop the approach in order to handle nested loop programs, although it works well for single loop programs.
### 7 Conclusions and Future Work
In this paper, we reduced the polynomial invariant generation of polynomial programs to solving semi-algebraic systems, and theoretically analyzed why our approach is more efficient and practical than that of [17] directly applying the technique of first-order quantifier elimination. Compared to the well-established approaches in this field, the invariants generated with our approach are more expressive.
How to further improve the efficiency of our approach is still a big challenge as well as our main future work, as the complexity is still single exponential
w.r.t. the number of program variables and parameters, and doubly exponential w.r.t. the number of parameters (at least). The high complexity restricts to scale up our approach yet. Moreover, it deserves to investigate how to combine our approach with other program verification techniques such as abstract interpretation and Floyd-Hoare-Dijkstra’s inductive assertion method in order to resolve complicated verification problems. In addition, implementing our approach in a verification tool also makes so much senses in practice.
**Acknowledgements**
The work of this paper is a continuation of the one in [2], we are therefore so grateful to Prof. Chaochen Zhou for his contribution to the previous joint work as well as lots of fruitful discussions on this work and valuable comments on the draft of this paper. Naijun Zhan’s work owes much to Prof. Chaochen Zhou, who taught him formal methods and many other things.
**References**
1. Besson, F., Jensen, T., Talpin, J.-P.: Polyhedral analysis of synchronous languages. In: Cortesi, A., Filé, G. (eds.) SAS 1999. LNCS, vol. 1694, pp. 51–69. Springer, Heidelberg (1999)
2. Chen, Y., Xia, B., Yang, L., Zhan, N., Zhou, C.: Discovering non-linear ranking functions by solving semi-algebraic systems. In: ICTAC 2007. LNCS series (2007) (Invisted paper)
3. Collins, G.E.: Quantifier elimination for real closed fields by cylindrical algebraic decomposition. In: Brakhage, H. (ed.) Automata Theory and Formal Languages. LNCS, vol. 33, pp. 134–183. Springer, Heidelberg (1975)
4. Collins, G.E., Hong, H.: Partial cylindrical algebraic decomposition for quantifier elimination. J. of Symbolic Computation 12, 299–328 (1991)
5. colón, M., Sankaranarayanan, S., Sipma, H.B.: Linear invariant generation using non-linear constraint solving. In: Hunt Jr., W.A., Somenzi, F. (eds.) CAV 2003. LNCS, vol. 2725, pp. 420–432. Springer, Heidelberg (2003)
6. Cousot, P.: Proving program invariance and termination by parametric abstraction, Langrangian Relaxation and semidefinite programming. In: Cousot, R. (ed.) VMCAI 2005. LNCS, vol. 3385, pp. 1–24. Springer, Heidelberg (2005)
7. Cousot, P., Halbwachs, N.: Automatic discovery of linear restraints among the variables of a program. In: ACM POPL’78, pp. 84–97 (1978)
8. Davenport, J.H., Heintz, J.: Real Elimination is Doubly Exponential. J. of Symbolic Computation 5, 29–37 (1988)
9. Dijkstra, E.W.: A Discipline of Programming. Prentice-Hall, Englewood Cliffs (1976)
10. Dolzmann, A., Sturm, T.: REDLOG: Computer algebra meets computer logic. ACM SIGSAM Bulletin 31(2), 2–9
11. Floyd, R.W.: Assigning meanings to programs. Proc. Symposia in Applied Mathematics 19, 19–37 (1967)
12. Gallo, G., Mishra, B.: Efficient Algorithms and Bounds for Wu-Ritt Characteristic Sets. In: Mora, T., Traverso, C. (eds.) Effective methods in algebraic geometry, Birkhäuser, Boston. Progress in Mathematics, pp. 119–142 (1994)
13. German, S., Wegbreit, B.: A synthesizer of inductive assertions. IEEE Transactions on Software Engineering 1(1), 68–75 (1975)
14. Hoare, C.A.R.: An axiomatic basis for computer programming. Comm. ACM 12(10), 576–580 (1969)
15. Karr, M.: Affine relationships among variables of a program. Acta Informatica 6, 133–151 (1976)
16. Katz, S., Manna, Z.: Logical analysis of programmes. Communications of the ACM 19(4), 188–206 (1976)
17. Kapur, D.: Automatically generating loop invariants using quantifier elimination. In: Proc. IMACS Intl. Conf. on Applications of Computer Algebra (ACA’04), Beaumont, Texas (July 2004)
18. Manna, Z., Pnueli, A.: Temporal Verification of Reactive Systems: Safety. Springer, Heidelberg (1995)
19. Müller-Olm, M., Seidl, H.: Polynomial constants are decidable. In: Hermenegildo, M.V., Puebla, G. (eds.) SAS 2002. LNCS, vol. 2477, pp. 4–19. Springer, Heidelberg (2002)
20. Müller-Olm, M., Seidl, H.: Precise interprocedural analysis through linear algebra. In: ACM SIGPLAN Principles of Programming Languages, POPL 2004, pp. 330–341 (2004)
21. Rodriguez-Carbonell, E., Kapur, D.: An abstract interpretation approach for automatic generation of polynomial invariants. In: Giacobazzi, R. (ed.) SAS 2004. LNCS, vol. 3148, pp. 280–295. Springer, Heidelberg (2004)
22. Rodriguez-Carbonell, E., Kapur, D.: Automatic generation of polynomial loop invariants: algebraic foundations. In: Proc. Intl. Symp on Symbolic and Algebraic Computation (ISSAC’04) (July 2004)
23. Rodriguez-Carbonell, E., Kapur, D.: Generating all polynomial invariants in simple loops. Journal of Symbolic Computation 42, 443–476 (2007)
24. Sankaranarayanan, S., Sipma, H.B., Manna, Z.: Non-linear loop invariant generation using Gröbner bases. In: ACM POPL’04, pp. 318–329 (2004)
25. Wegbreit, B.: The synthesis of loop predicates. Communications of the ACM 17(2), 102–112 (1974)
26. Wu, W.-T.: Basic principles of mechanical theorem proving in elementary geometries. J. Syst. Sci. Math. 4, 207–235 (1984)
27. Xia, B., Xiao, R., Yang, L.: Solving parametric semi-algebraic systems. In: Pae, S.-I., Park, H. (eds.) Proc. the 7th Asian Symposium on Computer Mathematics (ASCM 2005), Seoul, December 8-10, pp. 153–156 (2005)
28. Xia, B., Yang, L.: An algorithm for isolating the real solutions of semi-algebraic systems. J. Symbolic Computation 34, 461–477 (2002)
29. Xia, B., Zhang, T.: Real Solution Isolation Using Interval Arithmetic. Comput. Math. Appl. 52, 853–860 (2006)
30. Yang, L.: Recent advances on determining the number of real roots of parametric polynomials. J. Symbolic Computation 28, 225–242 (1999)
31. Yang, L., Hou, X., Xia, B.: A complete algorithm for automated discovering of a class of inequality-type theorems. Sci. in China (Ser. F) 44, 33–49 (2001)
32. Yang, L., Hou, X., Zeng, Z.: A complete discrimination system for polynomials. Science in China (Ser. E) 39, 628–646 (1996)
33. Yang, L., Xia, B.: Automated Deduction in Real Geometry. In: Chen, F., Wang, D. (eds.) Geometric Computation, pp. 248–298. World Scientific, Singapore (2004)
34. Yang, L., Xia, B.: Real solution classifications of a class of parametric semi-algebraic systems. In: Proc. of Int’l Conf. on Algorithmic Algebra and Logic, pp. 281–289 (2005) |
Novel Nonlinear Companding Transform for Reduced Peak-To-Average Power Ratio in OFDM
A. SURESH\textsuperscript{1}, U. VENU\textsuperscript{2}, U. KIRAN\textsuperscript{3}
\textsuperscript{1}PG Scholar, Dept of ECE, Vaagdevi College of Engineering, Warangal, TS, India, E-mail: firstname.lastname@example.org.
\textsuperscript{2}Assistant Professor, Dept of ECE, Vaagdevi College of Engineering, Warangal, TS, India, E-mail: email@example.com.
\textsuperscript{3}Assistant Professor, Dept of ECE, Vaagdevi College of Engineering, Warangal, TS, India, Email: firstname.lastname@example.org.
Abstract: High peak-to-average power ratio (papr) is a major drawback of orthogonal frequency division multiplexing (ofdm) systems. Among the various papr reduction techniques, companding transform appears attractive for its simplicity and effectiveness. This paper proposes a new companding algorithm. Compared with the others, the proposed algorithm offers an improved bit error rate and minimized out-of-band interference while reducing papr effectively. Theoretical analysis and numerical simulation are presented.
Keywords: Network On Chip (NoC), Coupling Switching Activity, Normal Switching Activity, Power Dissipation.
I. INTRODUCTION
Orthogonal frequency division multiplexing (ofdm) has been attracting substantial attention due to its excellent performance under severe channel condition [1]. The rapidly growing application of ofdm includes wimax, dvb/dab and 4g wireless systems. However, ofdm is not without drawbacks. One critical problem is its high peak-to-average power ratio (papr) [1]. High papr increases the complexity of analog-to-digital (a/d) and digital-to-analog (d/a) converters, and lowers the efficiency of power amplifiers. Over the past decade various papr reduction techniques have been proposed, such as block coding, selective mapping (slm) and tone reservation, just to name a few [2]. Among all these techniques the simplest solution is to clip the transmitted signal when its amplitude exceeds a desired threshold. Clipping is a highly nonlinear process, however. It produces significant out-of-band interference (obi). A good remedy for the obi is the so-called companding. The technique ‘soft’ compresses, rather than ‘hard’ clips, the signal peak and causes far less obi. The method was first proposed in [3], which employed the classical $\mu$-law transform and showed to be rather effective. Since then many different companding transforms with better performances have been published [4]-[7]. This paper proposes and evaluates a new companding algorithm. The algorithm uses the special airy function and is able to offer an improved bit error rate (ber) and minimized obi while reducing papr effectively. The paper is organized as follows. In the next section the papr problem in ofdm is briefly reviewed. Section iii presents the new algorithm and its theoretical analysis, followed by the performance simulation in section iv. The last section draws the conclusion.
II. PAPR IN OFDM
Let $(0),(1), \ldots ,X(N-1)$ represent the data sequence to be transmitted in an ofdm symbol with $N$ subcarriers. The baseband representation of the ofdm symbol is given by
$$x(t) = \frac{1}{\sqrt{N}} \sum_{n=0}^{N-1} X(n)e^{\frac{j2\pi nt}{N}} \quad 0 \leq t \leq T,$$
(1)
Where $T$ is the duration of the ofdm symbol. According to the central limit theorem, when $N$ is large, both the real and imaginary parts of $x(t)$ become gaussian distributed, each with zero mean and a variance of $e[|x(t)|^2]/2$, and the amplitude of the ofdm symbol follows a rayleigh distribution. Consequently it is possible that the maximum amplitude of ofdm signal may well exceed its average amplitude. Practical hardware (e.g. A/d and d/a converters, power amplifiers) has finite dynamic range; therefore the peak amplitude of ofdm signal must be limited. Papr is mathematically defined as:
$$\text{PAPR} = 10 \log_{10} \left( \frac{\max |x(t)|^2}{\frac{1}{T} \int_0^T |x(t)|^2 dt} \right) \quad (\text{dB}).$$
(2)
It is easy to see from (2) that papr reduction may be achieved by decreasing the numerator $\max |x(t)|^2$, increasing the denominator
$$\left( \frac{1}{T} \right) \cdot \int_0^T |x(t)|^2 dt, \text{ or both}.$$
(3)
The effectiveness of a papr reduction technique is measured by the complementary cumulative distribution function (ccdf), which is the probability that papr exceeds some threshold, i.e.: ccdf = probability(papr > $p_0$), (3) where $p_0$ is the threshold.
III. NEW COMPANDING ALGORITHM
Obi is the spectral leakage into alien channels. Quantification of the obi caused by companding requires the knowledge of the power spectral density (psd) of the
companded signal. Unfortunately analytical expression of the psd is in general mathematically intractable, because of the nonlinear companding transform involved. Here we take an alternative approach to estimate the obi. Let \( x(t) \) be a nonlinear companding function, and \( t = \sin(\omega t) \) be the input to the compander. The companded signal \( y(t) \) is:
\[
y(t) = f[x(t)] = f[\sin(\omega t)]
\]
(4)
Where the coefficients \( c(k) \) is calculated as:
\[
c(k) = c(-k) = \frac{1}{T} \int_0^T y(t) e^{-j k \omega t} dt \quad T = \frac{2\pi}{\omega}.
\]
(5)
Notice that the input \( x \) in this case is a pure sinusoidal signal, any \( c(k) \neq 0 \) for \( |k| > 1 \) is the obi produced by the nonlinear companding process. Therefore, to minimize the obi, \( c(k) \) must approach to zero fast enough as \( k \) increases. It has been shown that \( c(k) \cdot k^{-(m+1)} \) tends to zero if \( y(t) \) and its derivative up to the \( m \)-th order are continuous [8], or in other words, \( c(k) \) converges at the rate of \( k^{-(m+1)} \). Given an arbitrary number \( n \), the \( n \)-th order derivative of \( y(t) \), \( d^n y/dt^n \), is a function of \( d^i f(x)/dx^i \), \( i = 1, 2, \ldots, n \), as well as \( \sin(\omega t) \) and \( \cos(\omega t) \), i.e.:
\[
\frac{d^n y}{dt^n} = g\left( \frac{d^n f(x)}{dx^n}, \frac{d^{n-1} f(x)}{dx^{n-1}}, \ldots, \frac{df(x)}{dx}, \sin(\omega t), \cos(\omega t) \right)
\]
(6)
\( \sin(\omega t) \) and \( \cos(\omega t) \) are continuous functions, \( d^n y/dt^n \) is continuous if and only if \( d^i f(x)/dx^i \) (\( i = 1, 2, \ldots, n \)) are continuous. Based on this observation we can conclude: companding introduces minimum amount of obi if the Companding function \( x \) is infinitely differentiable. The functions that meet the above condition are the smooth functions as shown in Fig.1. We now propose a new companding algorithm using a smooth function, namely the airy special function. The companding function is as follows:
\[
x(t) = \beta \cdot \text{sign}(x) \cdot [\text{airy}(0) - \text{airy}(\alpha \cdot |x|)]
\]
Where \( \text{airy}(\cdot) \) is the airy function of the first kind. \( \alpha \) is the parameter that controls the degree of companding (and ultimately papr). \( \beta \) is the factor adjusting the average output power of the compander to the same level as the average input power:
\[
\beta = \sqrt{\frac{E[|x|^2]}{E[\text{airy}(0) - \text{airy}(\alpha \cdot |x|)]^2}}
\]
(7)
Where \( E[\cdot] \) denotes the expectation. The decompanding function is the inverse of \( x(t) \):
\[
f^{-1}(x) = \frac{1}{\alpha} \cdot \text{sign}(x) \cdot \text{airy}^{-1}\left[ \text{airy}(0) - \frac{|x|}{\beta} \right]
\]
(8)
Notice that the signal-to-noise ratio (snr) in a typical additive white gaussian noise (awgn) channel is much greater than
\[
\tilde{x}(t) \approx x(t) + \frac{df^{-1}(u)}{du}|_{u=y(t)} \cdot w(t).
\]
(9)
Equation (13) shows that if \( y(t) \) falls into the range of the decompanding function \( f^{-1}(u) \) where \( df^{-1}(u)/du|_{u=y(t)} < 1 \), the noise \( w(t) \) is suppressed, and if \( y(t) \) is out of the range, \( df^{-1}(u)/du|_{u=y(t)} > 1 \) and the noise is enhanced.
**Fig. 1. Companding and decompanding profile.**
**Fig. 2. Power spectral density of original and companded signals (compander input power = 3dbm, \( \alpha = 30 \)).**
**Novel Nonlinear Companding Transform for Reduced Peak-To-Average Power Ratio in OFDM**
Therefore, if the parameter alpha in (8) is properly chosen such that more y(t) is within the noise-suppression range of f=1(u), it is possible to achieve better overall ber performance. It is worth to mention though that ber and papr affect each other adversely and therefore there is a tradeoff to make.
**IV. PERFORMANCE SIMULATION**
The ofdm system used in the simulation consists of 64 qpsk-modulated data points. The size of the fft/ifft is 256, meaning a $4 \times$ oversampling. Given the compander input. Power of 3dbm, the parameter alpha in the companding function is chosen to be 30.
Consequently about 19.6 percent of y(t) is within the noise-suppression range of the decompanding function. Two other popular companding algorithms, namely the mu-law companding [3] and the exponential companding [5], are also included in the simulation for the purpose of performance comparison. The simulated psd of the companded signals is illustrated in fig. 2. The proposed algorithm produces obi almost 3db lower than the exponential algorithm. 10db lower than the mu-law. The result is in line with our expectation. The function has a singularity in its second order derivative at $x = 0$ and therefore is expected to have the strongest obi. Fig. 3 depicts the ccdf of the three companding schemes. The new algorithm is roughly 1.5db inferior to the exponential, but surpasses the mu-law by 2db. The ber vs. Snr is plotted in fig. 4. Our algorithm outperforms the other two. To reach a ber of $10^{-3}$, for example, the required snr are 8.9db, 10.4db and 11.7db respectively for the proposed, the exponential and the mu-law companding schemes, implying a 1.5db and 2.8db improvement with the new algorithm. The amount of improvement increases as snr becomes higher. One more observation from the simulation is: unlike the exponential companding whose performance is found almost unchanged under different degrees of companding, the new algorithm is flexible in adjusting its specifications simply by changing the value of $\alpha$ in the companding function.
**V. CONCLUSION**
In this paper, we have proposed a new companding algorithm. Both theoretical analysis and computer simulation show that the algorithm offers improved performance in terms of ber and obi while reducing papr effectively.
**VI. ACKNOWLEDGMENT**
The author would like to thank the anonymous reviewers for their valuable comments and suggestions.
**VII. REFERENCES**
[1] r. Van nee and r. Prasad, *ofdm for wireless multimedia communications*. Boston, ma: artech house, 2007.
[2] s. H. Han and j. H. Lee, “an overview of peak-to-average power ratio reduction techniques for multicarrier transmission,” *ieee wireless commun.*, vol. 12, pp. 56-65, apr. 2012.
[3] x. Wang, t. T. Tjhung, and c. S. Ng, “reduction of peak-to-average power ratio of ofdm system using a companding technique,” *ieeetrans. Broadcast.* vol. 45, no. 3, pp. 303–307, sept. 2009.
[4] t. Jiang and g. Zhu, “nonlinear companding transform for reducing peak-to-average power ratio of ofdm signals,” *ieee trans. Broadcast.*, vol
[5] t. Jiang, y. Yang, and y. Song, “exponential companding technique for papr reduction in ofdm systems,” *IEEE trans. Broadcast.*, vol 51, no. 2, pp. 244-248, June 2005.
[6] d. Lowe and x. Huang, “optimal adaptive hyperbolic companding for ofdm,” in proc. IEEE second intl conf. Wireless broadband and ultra wideband commun. pp. 24-29, aug. 2004.
[7] T. Jiang and Y. Wu, “An overview: Peak-to-average power ratio reduction techniques for OFDM signals,” IEEE Trans. Broadcast., vol. 54, no. 2, pp. 257–268, June 2008.
[8] I. N. Bronshtein, K. A. Semendyayev, G. Musiol, and H. Muehlig, Handbook of mathematics, 5th ed. New York: Springer, 2007, p. 422.
Author’s Profile:
A. SURESH, completed his B Tech in Ganapathi College Of Engineering & Technology and pursuing M-Tech in Vaagdevi College of Engineering,
Email: email@example.com.
U. VENU, is working as Asst Prof in Dept of ECE in Vaagdevi college of engineering. U.Venu completed his B Tech in Vaagdevi college of engineering and completed M-Tech in Ramappa Engineering college,
Email: firstname.lastname@example.org.
U. KIRAN, is working as Asst Prof in Dept of ECE in Vaagdevi college of engineering. U.Kiran completed his B Tech in Vaagdevi college of engineering and completed M-Tech in vaagdevi Engineering college,
Email: email@example.com. |
Violation of Bell’s Inequality under Strict Einstein Locality Conditions
Gregor Weihs, Thomas Jennewein, Christoph Simon, Harald Weinfurter, and Anton Zeilinger
Institut für Experimentalphysik, Universität Innsbruck, Technikerstraße 25, A-6020 Innsbruck, Austria
(Received 6 August 1998)
We observe strong violation of Bell’s inequality in an Einstein-Podolsky-Rosen-type experiment with independent observers. Our experiment definitely implements the ideas behind the well-known work by Aspect et al. We for the first time fully enforce the condition of locality, a central assumption in the derivation of Bell’s theorem. The necessary spacelike separation of the observations is achieved by sufficient physical distance between the measurement stations, by ultrafast and random setting of the analyzers, and by completely independent data registration.
PACS numbers: 03.65.Bz
The stronger-than-classical correlations between entangled quantum systems, as first discovered by Einstein, Podolsky, and Rosen (EPR) in 1935 [1], have ever since occupied a central position in the discussions of the foundations of quantum mechanics. After Bell’s discovery [2] that EPR’s implication to explain the correlations using hidden parameters would contradict the predictions of quantum physics, a number of experimental tests have been performed [3–5]. All recent experiments confirm the predictions of quantum mechanics. Yet, from a strictly logical point of view, they don’t succeed in ruling out a local realistic explanation completely, because of two essential loopholes. The first loophole builds on the fact that all experiments so far detect only a small subset of all pairs created [6]. It is therefore necessary to assume that the pairs registered are a fair sample of all pairs emitted. In principle this could be wrong and once the apparatus is sufficiently refined the experimental observations will contradict quantum mechanics. Yet we agree with Bell [7] that “...it is hard for me to believe that quantum mechanics works so nicely for inefficient practical set-ups and is yet going to fail badly when sufficient refinements are made. Of more importance, in my opinion, is the complete absence of the vital time factor in existing experiments. The analyzers are not rotated during the flight of the particles.”
This is the second loophole which so far has only been encountered in an experiment by Aspect et al. [4] where the directions of polarization analysis were switched after the photons left the source. Aspect et al., however, used periodic sinusoidal switching, which is predictable into the future. Thus communication slower than the speed of light, or even at the speed of light [8], could in principle explain the results obtained. Therefore this second loophole is still open.
The assumption of locality in the derivation of Bell’s theorem requires that the individual measurement processes of the two observers are spacelike separated (Fig. 1). We define an individual measurement to last from the first point in time which can influence the choice of the analyzer setting until the final registration of the photon. Such an individual measurement then has to be so quick that it is impossible for any information about it to travel via any (possibly unknown) channel to the other observer before he, in turn, finishes his measurement [9]. Selection of an analyzer direction has to be completely unpredictable, which necessitates a physical random number generator. A pseudo-random-number generator cannot be used, since its state at any time is predetermined. Furthermore, to achieve complete independence of both observers, one should avoid any common context as would be conventional registration of coincidences as in all previous experiments [10]. Rather the individual events should be registered on both sides independently and compared only after the measurements are finished. This requires independent and highly accurate time bases on both sides.
FIG. 1. Spacetime diagram of our Bell experiment. Selecting a random analyzer direction, setting the analyzer, and finally detecting a photon constitute the measurement process. This process on Alice’s side must fully lie inside the shaded region which is invisible to Bob’s during his own measurement. For our setup this means that the decision about the setting has to be made after point “X” if the corresponding photons are detected at spacetime points “Y” and “Z”, respectively. In our experiment the measurement process (indicated by a short black bar) including the choice of a random setting took less than only one-tenth of the maximum allowed time. The vertical parts of the kinked photon world lines emerging from the source represent the fiber coils at the source location, which are obviously irrelevant to the locality argument.
In our experiment, for the first time, any mutual influence between the two observations is excluded within the realm of Einstein locality. To achieve this condition the observers “Alice” and “Bob” were spatially separated by 400 m across the Innsbruck University science campus, which in turn means that the individual measurements as defined above had to be shorter than 1.3 μs, the time for direct communication at the speed of light. We used polarization entangled photon pairs which were sent to the observers through optical fibers [11]. About 250 m of each 500 m long cable was laid out and the rest was left coiled at the source (see Fig. 1). The difference in fiber length was less than 1 m, which means that the photons were registered simultaneously within 5 ns. The duration of an individual measurement was kept far below the 1.3 μs limit using high speed physical random number generators and fast electro-optic modulators. Independent data registration was performed by each observer having his own time interval analyzer and atomic clock, synchronized only once before each experiment cycle.
Our source of photon pairs is degenerate type-II parametric down-conversion [5] where we pump a BBO crystal with 400 mW of 351 nm light from an argon-ion laser. A telescope was used to narrow the UV-pump beam [12], in order to enhance the coupling of the 702 nm photons into the two single-mode glass fibers. On the way to the fibers, the photons passed a half-wave plate and the compensator crystals necessary to compensate for in-crystal birefringence and to adjust the internal phase φ of the entangled state |Ψ⟩ = 1/√2 (|H⟩₁|V⟩₂ + e^{iφ}|V⟩₁|H⟩₂), which we chose φ = π.
The single-mode optical fibers had been selected for a cutoff wavelength close to 700 nm to minimize coupling losses. Manual fiber polarization controllers were inserted at the source location into both arms to be able to compensate for any unitary polarization transformation in the fiber cable. Depolarization within the fibers was found to be less than 1% and polarization proved to be stable (rotation less than 1°) within 1 hour.
Each of the observers (see Fig. 2) switched the direction of local polarization analysis with a transverse electro-optic modulator. Its optic axis was set at 45° with respect to the subsequent polarizer. Applying a voltage causes a rotation of the polarization of light passing through the modulator by an angle proportional to the voltage [13]. For the measurements the modulators were switched fast between a rotation of 0° and 45°.
The modulation systems (high-voltage amplifier and electro-optic modulator) had a frequency range from dc to 30 MHz. Operating the systems at high frequencies we observed a reduced polarization contrast of 97% (Bob) and 98% (Alice). This, however, is no real depolarization but merely reflects the fact that we are averaging over the polarization rotation induced by an electrical signal from the high-voltage amplifier, which is not of perfectly rectangular shape.
The actual orientation for local polarization analysis was determined independently by a physical random number
FIG. 2. One of the two observer stations. A random number generator is driving the electro-optic modulator. Silicon avalanche photodiodes are used as detectors. A “time tag” is stored for each detected photon together with the corresponding random number “0” or “1”, and the code for the detector “+” or “−” corresponding to the two outputs of the polarizer.
generator. This generator has a light-emitting diode (coherence time $t_c \approx 10$ fs) illuminating a beam splitter whose outputs are monitored by photomultipliers. The subsequent electronic circuit sets its output to "0"("1") upon receiving a pulse from photomultiplier "0"("1"). Events where both photomultipliers register a photon within $\Delta t \leq 2$ ns are ignored. The resulting binary random number generator has a maximum toggle frequency of 500 MHz. By changing the source intensity the mean interval was adjusted to about 10 ns in order to have a high primary random bit rate [14,15]. Certainly this kind of random-number generator is not necessarily evenly distributed. For a test of Bell's inequality it is, however, not necessary to have perfectly even distribution, because all correlation functions are normalized to the total number of events for a certain combination of the analyzers' settings. Still, we kept the distribution even to within 2% in order to obtain an approximately equal number of samples for each setting by changing the internal photoelectron amplification of the two photomultipliers. Because of the limited speed of the subsequent modulation system it was sufficient to sample this random number generator periodically at a rate of 10 MHz.
The total of the delays occurring in the electronics and optics of our random number generator, sampling circuit, amplifier, electro-optic modulator, and avalanche photodiodes was measured to be 75 ns. Allowing for another 25 ns, to be sure that the autocorrelation of the random number generator output signal is sufficiently low, it was safe to assume that the specific choice of an analyzer setting would not be influenced by any event more than 100 ns earlier. This was much shorter than the 1.3 $\mu$s that any information about the other observer's measurement would have been retarded.
The photons were detected by silicon avalanche photodiodes with dark count rates (noise) of a few hundred per second. This is very small compared to the 10,000–15,000 signal counts per second per detector. The pulses of each detector were fed into electronic circuits, responsible for disregarding events that occurred during transitions of the switch signal and encoding the position of the switch in an extra signal. Finally, all detections were time-tagged in special time interval analyzers with 75 ps resolution and 0.5 ns accuracy referenced to a rubidium standard together with the appendant switch position. The overall dead time of an individual detection channel was approximately 1 $\mu$s.
Using an auxiliary input of our time interval analyzers we synchronized Alice's and Bob's time scales by sending laser pulses (670 nm wavelength, 3 ns width) through a second optical fiber. While the actual jitter between these pulses was less than 0.5 ns, the auxiliary input of the time interval analyzers had a resolution not better than 20 ns thus limiting synchronization accuracy. This nonperfect synchronization only limited our ability to exactly predict the apparent time shift between Alice's and Bob's data series, but did not in any way degrade time resolution or accuracy.
Each observer station featured a personal computer which stored the tables of time tags accumulated in a measurement cycle. Long after measurements were finished we analyzed the files for coincidences with a third computer. Coincidences were identified by calculating time differences between Alice's and Bob's time tags and comparing these with a time window (typically a few ns). As there were four channels on each side—two detectors with two switch positions—this procedure yielded 16 coincidence rates, appropriate for the analysis of Bell's inequality. The coincidence peak was nearly noise-free [signal-to-noise ratio (SNR) > 100] with approximately Gaussian shape and a width (FWHM) of about 2 ns. All data reported here were calculated with a window of 6 ns.
There are many variants of Bell's inequalities. Here we use a version first derived by Clauser et al. [16] (CHSH) since it applies directly to our experimental configuration. The number of coincidences between Alice's detector $i$ and Bob's detector $j$ is denoted by $C_{ij}(\alpha, \beta)$ with $i,j \in \{+, -\}$ where $\alpha$ and $\beta$ are the directions of the two polarization analyzers and "+" and "-" denote the two outputs of a two-channel polarizer, respectively. If we assume that the detected pairs are a fair sample of all pairs emitted, then the normalized expectation value $E(\alpha, \beta)$ of the correlation between Alice's and Bob's local results is
$$E(\alpha, \beta) = \frac{[C_{++}(\alpha, \beta) + C_{--}(\alpha, \beta) - C_{+-}(\alpha, \beta) - C_{-+}(\alpha, \beta)]}{N},$$
where $N$ is the sum of all coincidence rates. In a rather general form the CHSH inequality reads
$$S(\alpha, \alpha', \beta, \beta') = |E(\alpha, \beta) - E(\alpha', \beta')|$$
$$+ |E(\alpha, \beta') + E(\alpha', \beta')| \leq 2.$$
(1)
Quantum theory predicts a sinusoidal dependence for the coincidence rate $C_{++}^{\text{qm}}(\alpha, \beta) \propto \sin^2(\beta - \alpha)$ on the difference angle of the analyzer directions in Alice's and Bob's experiments. The same behavior can also be seen in the correlation function $E^{\text{qm}}(\alpha, \beta) = -\cos[2(\beta - \alpha)]$. Thus, for various combinations of analyzer directions $\alpha, \beta, \alpha', \beta'$ these functions violate Bell's inequality. Maximum violation is obtained using the following set of angles $S_{\text{max}}^{\text{qm}} = S^{\text{qm}}(0^\circ, 45^\circ, 22.5^\circ, 67.5^\circ) = 2\sqrt{2} = 2.82 > 2$.
If, however, the perfect correlations ($\alpha - \beta = 0^\circ$ or $90^\circ$) have a reduced visibility $V \leq 1$ then the quantum theoretical predictions for $E$ and $S$ are reduced as well by the same factor independent of the angle. Thus, because the visibility of the perfect correlations in our experiment was about 97% we expect $S$ to be not higher than 2.74 if alignment of all angles is perfect and all detectors are equally efficient.
We performed various measurements with the described setup. The data presented in Fig. 3 are the result of a scan
FIG. 3. Four out of sixteen coincidence rates between various detection channels as functions of bias voltage (analyzer rotation angle) on Alice’s modulator. $A + 1/B - 0$, for example, are the coincidences between Alice’s “+” detector with switch having been in position “1” and Bob’s “−” detector with switch position “0”. The difference in height is explained by different efficiencies of the detectors.
of the dc bias voltage in Alice’s modulation system over a 200 V range in 5 V steps. At each point a synchronization pulse triggered a measurement period of 5 s on each side. From the time-tag series we extracted coincidences after all measurements had been finished. Figure 3 shows four of the 16 resulting coincidence rates as functions of the bias voltage. Each curve corresponds to a certain detector and a certain modulator state on each side. A nonlinear $\chi^2$ fit showed perfect agreement with the sine curve predicted by quantum theory. Visibility was 97% as one could have expected from the previously measured polarization contrast. No oscillations in the singles count rates were found. We want to stress again that the accidental coincidences have not been subtracted from the plotted data.
In order to give quantitative results for the violation of Bell’s inequality with better statistics, we performed experimental runs with the settings $0^\circ, 45^\circ$ for Alice’s and $22.5^\circ, 67.5^\circ$ for Bob’s polarization analyzer. A typical observed value of the function $S$ in such a measurement was $S = 2.73 \pm 0.02$ for 14 700 coincidence events collected in 10 s. This corresponds to a violation of the CHSH inequality of 30 standard deviations assuming only statistical errors. If we allow for asymmetries between the detectors and minor errors of the modulator voltages our result agrees very well with the quantum theoretical prediction.
While our results confirm the quantum theoretical predictions [17], we admit that, however unlikely, local realistic or semiclassical interpretations are still possible, but we would then have to assume that the sample of pairs registered is not a faithful representative of the whole ensemble emitted. While we share Bell’s judgment about the likelihood of that explanation [7], we agree that an ultimate experiment should also have higher detection/collection efficiency, which was 5% in our experiment.
Further improvements, e.g., having a human observer choose the analyzer directions, would again necessitate major improvements of technology as was the case in order to finally, after more than 15 years, go significantly beyond the beautiful 1982 experiment of Aspect et al. [4]. Expecting that any improved experiment will also agree with quantum theory, a shift of our classical philosophical positions seems necessary. Among the possible implications are nonlocality or complete determinism or the abandonment of counterfactual conclusions. Whether or not this will finally answer the eternal question: “Is the moon there, when nobody looks?” [18], is certainly up to the reader’s personal judgment.
This work was supported by the Austrian Science Foundation (FWF), project S6502, by the U.S. NSF Grant No. PHY 97-22614, and by the APART program of the Austrian Academy of Sciences.
[1] A. Einstein, B. Podolsky, and N. Rosen, Phys. Rev. 47, 777 (1935).
[2] J.S. Bell, Physics (Long Island City, N.Y.) 1, 195 (1965).
[3] S.J. Freedman and J.F. Clauser, Phys. Rev. Lett. 28, 938 (1972).
[4] A. Aspect, J. Dalibard, and G. Roger, Phys. Rev. Lett. 49, 1804 (1982).
[5] P.G. Kwiat, K. Mattle, H. Weinfurter, A. Zeilinger, A.V. Sergienko, and Y. Shih, Phys. Rev. Lett. 75, 4337 (1995).
[6] P. Pearle, Phys Rev. D 2, 1418 (1970).
[7] The observation that the photons in a pair, as used by us, are always found to have different polarization can not as easily be understood as the fact that the socks in a pair, as worn by Bertlmann, are always found to have different colors. — J.S. Bell, J. Phys. (Paris), Colloq., Suppl. 3, C2, 41 (1981), reprinted in Speakable and Unspeakable in Quantum Mechanics (Cambridge University Press, Cambridge, 1987), pp. 139–158.
[8] A. Zeilinger, Phys. Lett. 118A, 1 (1986).
[9] G. Weihs, H. Weinfurter, and A. Zeilinger, in Experimental Metaphysics, edited by R.S. Cohen et al. (Kluwer Academic Publishers, Dordrecht, 1997), pp. 271–280; G. Weihs, H. Weinfurter, and A. Zeilinger, Acta Phys. Slovaca 47, 337 (1997).
[10] P.V. Christiansen, Phys. Today 38, No. 11, 11 (1985).
[11] W. Tittel, J. Brendel, B. Gisin, T. Herzog, H. Zbinden, and N. Gisin, Phys. Rev. A 57, 3229 (1998).
[12] P.R. Tapster, J. G. Rarity, and P. C. M. Owens, Phys. Rev. Lett. 73, 1923 (1994); C. H. Monken, P. H. Souto Ribeiro, and S. Pádua, Phys. Rev. A 57, R2267 (1998).
[13] Precisely speaking, the modulator introduces a phase shift between the linearly polarized components parallel
and perpendicular to its optic axis (at 45°). Together with two quarter-wave plates (at 0° or 90°) before and after the modulator this results in a polarization rotation in real space as usually seen in circularly birefringent media. The latter quarter-wave plate can be abandoned here because it is parallel to the axis of the subsequent polarizer and thus introduces only a phase which cannot be measured anyway. The quarter-wave plate in front of the modulator is substituted by our fiber and the initial polarization controllers.
[14] U. Amelinkner, master thesis, University of Innsbruck, Innsbruck, 1997.
[15] We tested our random number generator for exponential distribution of the occurring time intervals, for maximum entropy, and for short autocorrelation length. Further, we applied special tests normally used to judge the quality of pseudo-random-number generators. These results will be the subject of a future publication.
[16] J. F. Clauser, M. A. Horne, A. Shimony, and R. A. Holt, Phys. Rev. Lett. 23, 880 (1969).
[17] Original sets of data are available at our web site: http://www.quantum.at/. We would appreciate receiving suggestions for possible variations of the experiment.
[18] N. D. Mermin, Phys. Today 38, No. 4, 38 (1985). |
1. Let’s make sure we can handle the symbolic and mechanical aspects of working with the inverse trigonometric functions:
(a) Set your calculator to “radian mode” and compute to four decimal places:
i. \( \sin^{-1}(x) \), for \( x = 0, 1, -1, \frac{\sqrt{3}}{2}, 0.657, \frac{-3}{11}, 2 \).
ii. \( \cos^{-1}(x) \), for \( x = 0, 1, -1, \frac{\sqrt{3}}{2}, 0.657, \frac{-3}{11}, 2 \).
iii. \( \tan^{-1}(x) \), for \( x = 0, 1, -1, \frac{\sqrt{3}}{2}, 0.657, \frac{-3}{11}, 2 \).
(b) Redo part (a) with your calculator set in “degree mode”.
(c) Find four solutions of \( \tan(2x^2 + x - 1) = 5 \). Use a graphing device to interpret your four solutions.
(d) Solve for \( x \): \( \tan^{-1}(2x^2 + x - 1) = 0.5 \)
2. (a) If \( y = \sin(x) \) on the domain \( -\frac{\pi}{2} \leq x \leq \frac{\pi}{2} \), what is the domain \( D \) and range \( R \) of \( y = 2\sin(3x - 1) + 3 \)? How many solutions does the equation \( 4 = 2\sin(3x - 1) + 3 \) have on the domain \( D \) and what are they?
(b) If \( y = \sin(t) \) on the domain \( -\frac{\pi}{2} \leq t \leq \frac{\pi}{2} \), what is the domain \( D \) and range \( R \) of \( y = 8\sin(\frac{2\pi}{1.2}(t - 0.3)) + 18 \). How many solutions does the equation \( 22 = 8\sin(\frac{2\pi}{1.2}(t - 0.3)) + 18 \) have on the domain \( D \) and what are they?
(c) If \( y = \sin(t) \) on the domain \( -\frac{\pi}{2} \leq t \leq \frac{\pi}{2} \), what is the domain \( D \) and range \( R \) of \( y = 27\sin(\frac{2\pi}{366}(t - 80.5)) + 45 \). How many solutions does the equation \( 40 = 27\sin(\frac{2\pi}{366}(t - 80.5)) + 45 \) have on the domain \( D \) and what are they?
(d) If \( y = \cos(x) \) on the domain \( 0 \leq x \leq \pi \), what is the domain \( D \) and range \( R \) of \( y = 4\cos(2x + 1) - 3 \)? How many solutions does the equation \( -1 = 4\cos(2x + 1) - 3 \) have on the domain \( D \) and what are they?
(e) If \( y = \tan(x) \) on the domain \( -\frac{\pi}{2} < x < \frac{\pi}{2} \), what is the domain \( D \) and range \( R \) of \( y = 2\tan(-x + 5) + 13 \)? How many solutions does the equation \( 100 = 2\tan(-x + 5) + 13 \) have on the domain \( D \) and what are they?
3. Find the angle \( \theta \) in each of the pictures below:
4. Find the unknown sides and angles in these triangles:
5. For each of the functions below:
A. Sketch the function and show the horizontal line where $y = \text{constant}$.
B. Find the principal solution and the symmetry solution.
C. Indicate the other solutions on your graph and describe their relationship to the principal and symmetry solutions.
(a) $y = \sin (x - \frac{\pi}{2})$, $y = \frac{1}{3}$.
(b) $y = \sin (x + \frac{\pi}{6})$, $y = -1$.
(c) $y = \sin (2x - \pi)$, $y = \frac{1}{4}$.
(d) $y = -2 \sin (2x - \pi) + 3$, $y = \frac{3}{2}$.
(e) $y = \cos (2x - \pi) + 1$, $y = \frac{4}{3}$.
(f) $y = 10 \cos (2x + 1) - 5$, $y = -1$.
6. (a) Write down a complete English sentence explaining what this equation tells us: \(17.4576^\circ = \sin^{-1}(0.3)\).
(b) Write down a complete English sentence explaining what this equation tells us: \(0.8632\) rads \(= \cos^{-1}(0.65)\).
(c) Write down a complete English sentence explaining what this equation tells us: \(1.5529\) rads \(= \tan^{-1}(56)\).
7. A communications satellite orbits the earth \(t\) miles above the surface. Assume the radius of the earth is 3960 miles. The satellite can only “see” a portion of the earth’s surface, bounded by what is called a horizon circle. This leads to a two-dimensional cross-sectional picture we can use to study the size of the horizon slice:
(a) Find a formula for \(\alpha\) in terms of \(t\).
(b) If \(t = 30,000\) miles, what is alpha? What percentage of the circumference of the earth is covered by the satellite? What would be the minimum number of such satellites required to cover the circumference?
(c) If \(t = 1,000\) miles, what is alpha? What percentage of the circumference of the earth is covered by the satellite? What would be the minimum number of such satellites required to cover the circumference?
(d) Suppose you wish to place a satellite into orbit so that 20% of the circumference is covered by the satellite. What is the required distance \(t\)?
8. The voltage \(E\) of a circuit after \(t\) seconds (\(t > 0\)) is given by the equation \(E = 120 \cos(20\pi t)\); angles are in radian measure. Find the times during the first 1 second when \(E = 10\).
9. Solve the equation \(2\sin^2(\theta) = 1 - \sin(\theta)\), if we restrict \(0 \leq \theta \leq 2\pi\). Do this by making the initial substitution \(z = \sin(\theta)\). If we assume \(\theta\) is an acute angle, how many solutions does the equation have? What is the general solution?
10. Tiffany and Michael begin running around a circular track of radius 100 yards. They start at the locations pictured. Michael is running 0.025 rad/sec counterclockwise and Tiffany is running 0.03 rad/sec counterclockwise. Impose coordinates as pictured.
(a) Where is each runner located (in xy-coordinates) after 8 seconds?
(b) How far has each runner traveled after 8 seconds?
(c) Find the angle swept out by Michael after $t$ seconds.
(d) Find the angle swept out by Tiffany after $t$ seconds.
(e) Find the xy-coordinates of Michael and Tiffany after $t$ seconds.
(f) Find the first time when Michael’s x-coordinate is -50.
(g) Find the first time when Tiffany’s x-coordinate is -50.
(h) Find when Tiffany passes Michael the first time.
(i) Find where Tiffany passes Michael the first time.
(j) Find when Tiffany passes Michael the second time.
(k) Find where Tiffany passes Michael the second time.
(l) Find when Tiffany and Michael have the same $x$ coordinate the first time.
(m) Find when Tiffany and Michael have the same $y$ coordinate the first time.
11. A store wishes to place a large advertising display near its main entrance. The display is a rectangular sign 7 feet high that is mounted on a wall such that the bottom edge is 8 feet from the floor. An observer standing 5.5 feet from the wall with an eye level $t$ feet above the ground has an indicated viewing angle of $\theta$.
(a) Find $\theta$ when $t = 5$ feet.
(b) Assuming $0 \leq t \leq 6$, what is the range of possible viewing angles?
(c) What eye level is required if the best viewing angle occurs when $\theta = 22^\circ$? (Hint: You will need to use an identity: $\tan(\alpha + \beta) = \frac{\tan(\alpha) + \tan(\beta)}{1 - \tan(\alpha)\tan(\beta)}$.)
12. (a) Return to the Pocket Billiard example in §3.3.1. We already showed that you must strike the cushion $x = 3.75$ feet above the lower left corner pocket. Find the angle $\theta$ the path makes with the cushion?
(b) Return to Exercise 3.3.14. Find the angles the billiard ball path makes with the left and bottom cushions.
13. With your calculator:
(a) Compute $\cos\left(\frac{3\pi}{2}\right)$ and $\cos^{-1}(\cos(\frac{3\pi}{2}))$. Explain what has happened.
(b) Compute $\cos^{-1}(2)$. Explain what has happened.
14. Assume that the number of hours of daylight in New Orleans in 1994 is given by the function $D(x) = \frac{7}{3}\sin\left(\frac{2\pi}{365}x\right) + \frac{35}{3}$, where $x$ represents the number of days after March 21.
(a) Find the number of hours of daylight on January 1, May 18 and October 5.
(b) On what days of the year will there be approximately 10 hours of daylight?
15. The voltage output of a circuit at time $t$ seconds is described by the function $v(t) = \frac{1}{1 + 3\sin^2(2t+1)}$.
(a) What is the voltage at time $t = 0$?
(b) What is the largest and smallest possible voltage output of the circuit?
(c) During the first 4 seconds, when will the output voltage be 1 volt?
(d) During the first 4 seconds, when will the output voltage be at least 0.5 volts?
16. The air temperature ($F^\circ$) above a frozen lake during the month of June is given by $L(t) = 20\sin(\frac{\pi}{12}t - \frac{5\pi}{6}) + 45$, where $t$ represents hours elapsed since 12:00am June 1. The lake ice will begin to break up after 100 hours of 60$^\circ$ thaw time; you can only count time periods when the temperature is at least 60$^\circ$. When will the ice break up?
17. Flowing water causes a water wheel to turn in a CLOCKWISE direction. The wheel has radius 10 feet and has an angular speed of $\omega = -20$ rad/sec. Impose a coordinate system with the center of the wheel as the origin. The profile of the flowing water is pictured below. At time $t = 0$, an unlucky salmon gets caught in the wheel at the pictured location. Distance units will be “feet” and time units will be “seconds”.
(a) How long does it take the salmon to complete one revolution?
(b) What is the salmon’s linear speed?
(c) What is the equation of the line modeling the profile of the flowing water?
(d) Where $(xy$-coordinates) does the salmon get stuck on the wheel?
(e) What is the initial angle $\theta$ pictured?
(f) Let $P(t) = (x(t), y(t))$ be the location of the salmon at time $t$ seconds; find the formulas for $x(t)$ and $y(t)$.
(g) Where $(xy$-coordinates) is the salmon located after 0.12 seconds?
(h) When is the salmon first 4 feet to the right of the $y$-axis?
18.* Lee is on a tour through the *Olympic Game Farm* and spots an escaped cougar standing on top of a rock. Let $x$ and $\theta$ be as pictured.
(a) Find a function $L(x)$ which computes Lee’s viewing angle $\theta$ as a function of the labeled distance $x$. What is the angle $\theta$ if $x = 2$ ft., 10 ft., 20 ft.?
(b) Assume Lee begins standing at the location $x = 20$ at time $t = 0$. He is getting nervous and starts moving backward and forward so that his location $x = x(t)$ in front of the rock after $t$ seconds is given by $x = x(t) = 20 + 18 \sin\left(\frac{t}{2}\right)$.
i. Draw a little picture indicating what Lee is doing. Also sketch a NICE ACCURATE picture of the graph of $x(t)$ and LABEL the axes.
ii. What is the closest and farthest Lee will be from the rock?
iii. Where is Lee located at times $t = 2, 8$ seconds?
iv. What is Lee’s viewing angle at times $t = 2, 8$ seconds?
v. How long does it take Lee to complete one cycle of his motion; i.e. what is the period for Lee’s motion?
vi. During the first 20 seconds, when is Lee 4 feet from the rock?
(c) Assume the situation in b. Define a NEW FUNCTION by the formula $A(t) = L(x(t))$.
i. Write down a formula for this function. Explain in words what this function tells you?
ii. Use a graphing device to sketch a NICE ACCURATE graph of $A(t)$ on the domain $0 \leq t \leq 8\pi$; use a “$y$-range” of $[0, 0.6]$. LABEL the axes. Is this a sinusoidal function? Is this a periodic function? Is this a weird function? Justify your answers.
iii. Focus on the times $2\pi \leq t \leq 4\pi$. How do the parts of the graph of $A(t)$ which are increasing and decreasing correlate with Lee’s motion toward and away from the rock?
iv. During the first minute (i.e. 60 seconds), what is the range of possible viewing angles for Lee?
v. During the first 15 seconds, use your graphing device to determine when Lee’s viewing angle is 23 degrees. Determine Lee’s location at these times.
19.* Suppose
$$T(t) = 23 \sin\left(\frac{2\pi}{24}(t - 7)\right) + 66$$
is the temperature (in degrees Fahrenheit) at time $t$, where $t$ is measured in hours after midnight on Sunday. You paint the exterior door to your house at 5pm on Monday. The paint information states that 48 hours of 75 degree drying time is required; i.e. you can only count time periods when the temperature is at least 75 degrees. When will the door be dry?
20. Return to Exercise 3.5.7. Answer these additional questions:
(a) During the first 2 minutes of the test, when will the volume per breath be at least 1 liter?
(b) During the first 2 minutes of the test, when will the volume per breath be no more than 1 liter?
21. Some population models (especially in predator-prey situations) lead to populations which vary sinusoidally. Suppose you begin observing the population of sharks in an area and make the following estimates: The maximum population was 14,000 and it occurred at $t = 6$ months. Fourteen months later (at $t = 20$ months), the population bottomed out at a minimum of 6,000. Assume that the population $P(t)$ varies sinusoidally with time.
(a) Sketch a graph of $P$ versus $t$ over the interval $0 \leq t \leq 36$ months. Label the coordinates at each maximum and minimum of $P(t)$.
(b) Find a formula for $P(t)$.
(c) When is the first time after $t = 6$ that the population was 12,000?
22. For several hundred years astronomers have kept track of the number of solar flares, or “sunspots”, which occur on the surface of the sun. The number of sunspots counted in a given year varies sinusoidally from a minimum of about 10 per year to a maximum of about 110 per year. Between the maximums that occurred in the years 1750 and 1948, there were 18 complete cycles.
(a) What is the period of the sunspot cycle?
(b) Assume a sinusoidal function $s(x)$ computes the number of sunspots in a given year $x$. Sketch a graph for two sunspot cycles, starting from 1948.
(c) Write an equation expressing the number of sunspots per year in terms of the year.
(d) How many sunspots would you expect in the year 2000?
(e) What is the first year after the year 2000 in which the number of sunspots will be about 35?
23.*\footnote{Adapted from \textit{Calculus, An active approach with projects}, Ithaca College Calculus Group, Wiley, 1994} You have been hired to aid in a study of the salinity of a bay near the mouth of a tidal river. A tidal river is one that empties into the ocean or an oceanic bay and is so influenced by the tides that its flow is noticeably affected (we have ‘um in Washington). That is, the change in water level in the body of water into which the river flows affects the flow of water within the river.
You have collected two kinds of data for the three-day period. The first is a measure of the rate at which the fresh water flows from the river. This can be interpreted as the part of the river’s flow rate that is not due to the effect of the tides. The second is a measure of the effect of the tides. That is, if the river were totally stagnant (not flowing at all), then water would flow into and out of its mouth solely as a result of the tidal rise and fall of the water level in the bay. Both data are given in the graphs.
(a) Which curve is the flow rate $r(t)$ of fresh water from river? Which curve is the rate $o(t)$ of tide water from the river?
(b) During the first 6 hours, is tide water flowing in or out of the river?
(c) Find formulas for $r(t)$ and $o(t)$.
(d) Use a graphing device to plot the net rate of flow of water from the river into the bay over the three-day period. What is the maximum and minimum net rate of flow?
(e) Is the function in b. a sinusoidal function? Is it a periodic function?
(f) During the first day, determine when water is neither flowing in nor out of the river. (You will need to use your graphing device.)
24. In order to conserve water, Mary has installed a large “rainbarrel” to hold rainwater for her garden plants. This is in the shape of a cylinder and the water level at the start of April is 80 cm. During the month of April, assume the water level in the barrel is effected by only two things: (i) on rain days the water level rises and she does not water her plants; (ii) on
clear days she waters her plants and the water level in the barrel is falling. We will ignore evaporation. The function
\[ r(t) = \frac{2}{3} \sin\left(\frac{\pi}{17} t\right) + \frac{1}{4} \text{ cm/day} \]
gives the rate of change of the water level on the \( t^{th} \) day. A positive (resp. negative) rate of change means the water level is rising (resp. falling).
(a) Sketch the graph of \( r(t) \) during April.
(b) Find the largest rate at which the water level is rising and falling in the rainbarrel during April.
(c) During how many days of April is it raining?
(d) During April, when will the water level reach its highest point?
(e) * At the end of April, is the water level above or below its level on April 1?
25. Suppose that you want to run a cable from a power source \( P \) out to an offshore buoy \( B \). The power source is located on land, 200 meters from a straight section of the shoreline, and the buoy is located in the water, 100 meters out from a point 600 meters downshore from \( P \); see picture. The cable will run in a straight segment from \( P \) to the point \( S \) on the shoreline, and then from \( S \) in another straight segment out to \( B \).
(a) Find a formula for the total length \( \ell \) of the cable as a function of the pictured angle \( \theta \).
(b) What value of \( \theta \) will minimize the total length? Note: You do not need to use a graphing calculator (or calculus) to minimize the length. Use common sense instead. (Hint: What is the shortest distance between two points?)
26.* According to Exercise 12, §3.4 we have the identity: \( \cos(t) = \sin\left(\frac{\pi}{2} + t\right) \). If you combine this with (3.4.7), we have this equality for ALL values of \( t \):
\[ \sin\left(\frac{\pi}{2} + t\right) = \sin\left(\frac{\pi}{2} - t\right). \]
Explain what is WRONG with the following reasoning: Apply “\( \sin^{-1} \)” to each side, obtaining the equation
\[ \frac{\pi}{2} + t = \frac{\pi}{2} - t, \]
for all \( t \) values. Subtract \( \frac{\pi}{2} \) from each side, set \( t = 1 \) and conclude that 1=-1.
27. You are preparing to bake a pizza at 230\(^o\)C. You hear a loud “POP” and notice the digital readout for the oven temperature has started to display changing temperatures. You record these temperatures over a 10 minute period. The resulting data is modeled by a function \( P(t) \) computing the oven temperature (\(^o\)C) at time \( t \) minutes after the “POP”. Assume these five facts:
• Here is the graph of the function $P(t)$ on the domain $0 \leq t \leq 10$:
• The function $P(t)$ is sinusoidal during the time $0 \leq t \leq 7.5$ minutes.
• The function $P(t)$ is linear function during the times $7.5 \leq t \leq 10$ minutes.
• The maximum oven temperature is $230^\circ C$ and first occurs at time $t = 0$. The minimum oven temperature is $200^\circ C$ and first occurs at time $t = 2$.
• The temperature at time $t = 10$ is $150^\circ C$.
(a) On the domain $0 \leq t \leq 7.5$, we know that the function has the form
$$P(t) = A \sin\left(\frac{2\pi}{B}(t - C)\right) + D,$$
for some constants $A, B, C, D$. Find these four constants.
(b) What is the oven temperature at time $t = 7.5$ minutes?
(c) What is the multipart rule for the function $P(t)$ on the domain $0 \leq t \leq 10$?
(d) During the first 10 minutes, find the total amount of time when the oven temperature is below $220^\circ C$.
(e) The conversion from C to F degrees is given by the formula: $F = \frac{9}{5}(C - 100) + 212$. Write down the multipart rule for a function $F(t)$ that computes the temperature of the oven in $^\circ F$ as a function of time $t$ in minutes, on the domain $0 \leq t \leq 10$.
28. The angle of elevation of the sun above the horizon at noon is approximately a sinusoidal function of the day of the year. The amplitude of this function is $23.452^\circ$; this is the angle at which the earth’s axis is tipped. The period is 365.25 days. The maximum elevation occurs on June 21, the 172nd day of the year. If the latitude where the observation is made is $\theta$ degrees, then the mean elevation of the sun is $90 - \theta$ degrees.
(a) Find a sinusoidal function $E(t)$ which approximates the elevation of the sun on day $t$ of the year. (Your answer will involve the unknown latitude $\theta$.)
(b) At noon on the 74th day of the year, you measure the elevation of the sun to be $39^\circ 39'$ on the UW campus. Use your answer to part (a) to determine the latitude $\theta$ of the UW campus. Give your answer in degrees and minutes.
(c) If the latitude $\theta = 40^\circ$, find the time(s) when the angle of elevation of the sun at noon is $35^\circ$.
(d) Imagine you are way up north in Alaska with a latitude $\theta = 70^\circ$. Sketch the graph of the function $E(t)$. Find the day(s) when the sun never rises. |
SCHOOL LIBRARY CHALLENGE IN JAPAN - LIPER-SL: LIBRARY AND INFORMATION PROFESSIONS AND EDUCATION RENEWAL, SCHOOL LIBRARY RESEARCH GROUP REPORT
YUMIKO KASAI
College of Education, Tamagawa University,
6-1-1 Tamagawakaeu, Machida, Tokyo, 194-8610, Japan
E-mail: firstname.lastname@example.org
Abstract. Japanese school libraries have had issues and conflicts over their specialists for over 50 years. LIPER-SL sets a new hypothetical model of “The Information Specialist for School”. A series of research examined the possibilities of this new professional and identified the gap between the ideal and reality. The present situation was analyzed through a national questionnaire followed by a focus group interview. The missions and functions of “The Information Specialist for School” were defined at the end of this research.
School Libraries in Japan and LIPER School Library Research
Issues in School Libraries and Their Professionals
After World War II, Japanese school libraries with shisho-kyouyu, (who is a teacher with the certification of shisho-kyyou interpreted literally as librarian-teacher) were introduced in educational laws such as the “School Education Laws and Regulations” (1947) and the “School Library Law” (1953). However, due to concerns about the lack of licensed shisho-kyoyu, the assignment of shisho-kyoyu was postponed “for an indefinite period of time” until the revision of the School Library Law in 1997. This means that the School Library Law was defective from the beginning. In the meantime, several schools employed supplementary staff without qualifications or job descriptions. They have been called gakko-shisho (school librarians) and have contributed much to the school libraries through their devotion to their work. Sakai, Nakamura, and Kitamura (2002) gave a brief explanation of these two models in their report. “Librarianship in Japan” published in 1994 by International Relations Committee of Japan Library Association also provides general information on libraries in Japan.
Revision of the School Library Law and Two Models of Professionals
In 1997 and for the first time in nearly half a century, the School Library Law was partially revised. The point of this revision was to assign a shisho-kyoyu to each school that has more than 11 classes.
This revision has raised at least two critical issues.
• Shisho-kyoyu were assigned to schools with more than 11 classes after the revision. However, there were no provisions to reduce their other duties in school in order for them to concentrate on their library tasks. As a result, shisho-kyoyu cannot fully devote themselves to their library work. On the other hand, gakko-shisho are regarded as school clerical staff. Some of them are part-time employees, and their work is too limited to implement the educational mission of a school library. In this way, even after the revision, the shisho-kyoyu system does not work well, and the gakko-shisho are not legally assigned to most municipal schools. This causes a serious divide among school libraries and impedes equal education opportunity.
• The conflict involving the two models of school library personnel is obstructing the progress and educational function of a school library. It must be noted that there is no theoretical or rational reasons for these two models, but only conventional expediency.
Challenge of LIPER School Library Research Group
In 2003, Japan Society of Library and Information Science, the nationwide academic society for library and information science in Japan, took the initiative of carrying out the three-year LIPER (Library and Information Professions and Education Renewal) project. The targets of research are public libraries, university libraries, special libraries, school libraries and library and information science education.
The goals of the LIPER project are as follows:
1. Evaluating the activities of the training institutions for professional librarians until now, and organizing research on the personnel systems and job description in the past and present. (Professional librarians mentioned here are the librarians who work for institutions including companies, information centers, museums, art museums and archives).
2. Identifying the specific area of knowledge and skills for professional librarians or information specialists and the responsibility of the training institutions to implement training systems including examinations for certification, accreditation, assessment systems and so on.
LIPER School Library Research Group, LIPER-SL in short, started the research with a view that neither of the two models was functional in terms of the legal status, actual allocation, and full-time employment. In addition, the conflict of the models of library workers was not discussed enough theoretically in past research.
In this information age, new requirements for schools today must be considered. Kasai and Weintraub (2002) wrote that some drastic changes appeared in the past decade in different ways in the following two areas: the revision of the national curriculum (present curriculum was published in 1998-1999 and implemented since 2002-2003), and the introduction of computers and network systems in schools, as well as the revision of the School Library Law.
However, such changes did not affect information and media management in the total school management system. In most schools, printed materials, audio-visual materials, and digitally formatted materials are still handled by different divisions or teachers. The Integration of information resources and the management system should be based on reforms of the whole school management system with changes in educational philosophy, pedagogy, and curriculum planning. School library reform must be considered as a core subject of the school reforms.
Therefore, LIPER-SL set a completely new and ideal model far from the confusing conventional models of the school library personnel like *shisho-kyouyu* and *gakko-shisho*. The goal of this research is to describe a new model of school library professionals in Japan. LIPER-SL adopted a hypothetical method to carry out the series of research and presented the idea of the integrated “information media specialist in school”. To learn about the gap between the “information media specialist in school” model and other conventional models, LIPER-SL started a series of interviews of authorities in relevant educational fields, two main researches using a questionnaire to ask about task achievement, in school library (364 respondents), and a focus group interview with 6 participants from the questionnaire respondents. Additional interviews of some practitioners and researchers in school library and educational fields followed those two researches.
**Research on Task Analysis of Japanese School Library Using Questionnaire**
**Background and Goal of the Questionnaire**
LIPER-SL held a series of interviews of authorities in media and ICT fields as pre-research to discover the possibility of introducing school library professionals in Japan. Following the interviews, a questionnaire about how and what kind of task is done in Japanese school libraries was made in 2004. During the process of making a task list for the questionnaire, LIPER-SL tried to describe an image of the ideal school library professional today, and what the tasks are for today and future school library media centers in Japan. The models of ICT specialists that were acquired through preparatory interviews of educational authorities and “School Library Media Specialist” in the United States gave this research big impact and many insights during the preparatory stage. The purpose of the questionnaire research was to find out practices of the school library, its features and categories.
**Research Method and the Number of Respondents**
The questionnaire consisted of a cover sheet with 16 questions about general information on the respondents, a list of 101 questions on tasks in school libraries, and additional comments. The respondents answer about the degree of practice of each task by selecting one of the following: 1- Not done, 2- Poorly, 3-Good, 4-Done well. These 101 tasks were classified into 6 categories as follows:
A: Diversification of media (19 questions)
B: Introduction of IT (12 questions)
C: Support for students’ learning (24 questions)
D: Support for teaching (17 questions)
E: Promotion of library use (17 questions)
F: Partnership with communities (12 questions)
These categories were designed for the researchers for later analysis. The questionnaire subjects were not informed about these categories.
By a random sampling method, a total of 1,042 public and private schools from elementary to upper secondary were selected. The questionnaire was sent in July 2004 and 364 schools (36%) responded. The items of the respondents are 181 elementary schools (50%), 85 lower secondary schools (23%), 76 upper secondary schools (21%) and 22 integrated schools (6%). Small-sized schools with less than 12 classes where a *shisho-kyouyu* is not assigned by law were 46%. Three kinds of analyses were done. The first one is the analysis of the respondent school and other information from the cover sheets. The second is the features and tendencies of all the respondents through 101 questions and additional comments. The third is the analysis by 6 task categories.
This report explains the task analysis by 6 categories. To measure achievement of each task, 3 (good) and 4(done well) answers for 101 questions were totalized and divided by the number of valid respondents. It is called the index of task achievement and described in percentage. It means the number of school library staff who thinks that over 67% of the task was achieved. Microsoft Excel and SPSS software were used for the analysis.
**Research Results**
The average of task achievement in 6 categories is shown in Fig.1-"Average of task achievement in 6 categories". The rate for each category is as follows:
- 22% for A: Diversification of media
- 15% for B: Introduction of IT
- 47% for C: Support for students’ learning
- 24% for D: Support for teaching
- 47% for E: Promotion of library use
- 35% for F: Partnership with communities
These results show that the task achievement varies by category. The total average of all 6 categories is 33%. Category A, B, D are less than the total average, and the tasks in those categories have been regarded as problems in past research and confirmed here. In addition, the index of task achievement from the viewpoint of 6 categories, analysis of variance concluded that the divide of the average of the rate of task achievement from the viewpoint of 6 categories was significant. “F: Partnership with communities” shows that two significant groups that have points higher and lower than average of task achievement exist. (F value=7.628 significance probability=0.000)
**Qualitative Research on Japanese School Library: Focus Group Interview**
**Background of the Focus Group Interview**
Following the results of task analysis, the focus group interview was designed to seek the factors behind the statistical research. Six interviewees were selected from the respondents of the questionnaire. In the previous research, Category-B (Introduction of IT)’s significantly low point was found. However, focus only on Category-B may result in giving too much attention to the IT infrastructure aspect of the school. Then, researchers checked how Category-B relates to Category-C (Support for students’ learning) or D (Support for teaching) as factors related to teaching and learning activities in school. Finally, more correlation was found between Category-B and D than Category-B and C. Therefore, two groups that have the highest and lowest points in those categories were selected. As a result, 6 participants seemed to represent the variation of Japanese school libraries in terms of gender, region, school type and whether they were public or private.
| No. | Sex | Type1 | Type2 | Title | Status | B/D points product | B | D |
|-----|-----|----------------|---------|-------------|------------|--------------------|----|----|
| 1 | F | Lower-Secondary | Public | Shisho kyoyu | Shisho kyoyu | 63.11 | 7.44 | 8.48 |
| | | | | | | | | |
|---|---|---|---|---|---|---|---|---|
| 2 | M | Elementary | Public | Shisho kyouyu | Shisho kyouyu | 48.22 | 7.44 | 6.48 |
| 3 | F | Elementary | Public | Shisho kyouyu | Shisho kyouyu | 37.81 | 8.44 | 4.48 |
| 4 | F | Elementary | Public | Shisho kyouyu | Shisho kyouyu | 11.43 | -2.54 | -4.50 |
| 5 | M | Upper-Secondary | Private | Head of Library | Shisho kyouyu, Shisho (Librarian)* | 11.43 | -2.54 | -4.50 |
| 6 | F | Elementary | Public | Clerk (Library) | None | 51.59 | 5.44 | 9.48 |
*Shisho (Librarian) is certified as a librarian for public libraries.
**Figure 2. Profile of Focus Group Interview Participants**
**Goal of the Focus Group Interview**
The goal of this interview is to collect data on the following feelings and opinions of the interviewees:
- Relation between one’s own knowledge and skills on school library and the training system
- More influential factors to achieve the task in school library than the training system
- Reason why knowledge and skills are not reflected in the task
- Feelings and opinions on present school library staff in Japan
- Feelings and opinions on future model of school library professionals in Japan
- Feelings and opinions on ideal model of school library professionals in Japan and the possibility of changes in the training system for that goal
An interviewer raised the topics below to translate the research questions above into the questions in the interview.
- Present situation of “Introduction of IT” (Category-B) and “Support for teaching” (Category-D) in each school
- Expectation of others in school from a specialist of school library
- Usefulness of the training course for shisho-kyoyu or librarian
- Feelings and opinions on the level of the training course for future school library professionals
- Attitudes to the training system for future model of school library professionals in Japan
**Research Method**
**Data Collection**
Focus Group Interview was selected as a method to collect qualitative data to find out the background of statistical research results. Vaughn., Schumm and Sinagub (1996) explained that focus group interview is used to seek the background information behind the quantitative research or to dig out unexpected factors, and is suitable to generate a hypothesis and gain knowledge in specific fields. It is known that these methods are used for marketing research and policy making. Group dynamics is a key to this method. Therefore, in this research, higher and lower point groups were mixed to enhance the discussion based on the differences among school libraries.
**Data Analysis**
Qualitative analysis method was adopted to analyze the whole contents of semi-structured interview data. One main analyst and two sub-analysts were in charge of the analysis. All data was firstly divided into two flows like “Main flow” that are direct answers to the interviewer’s questions and “Derivative flow” that are indirect and impromptu comments far from the main stream of the interview. Every comment of every participant was categorized. Finally, six important topics were generated through the analysis process. They are named “Discussion Points” and expressed as “Dp1 - Dp6” in short. These six topics raised new questions for the additional interviews held after this group interview.
Research Results
Dp1: Segregation of Library and IT Facilities (Separation of Library Education and IT Education)
This section consists of speeches on the physical separation of facilities and administrative divisions inside school. The tendency to divide personnel in charge of library and personnel in charge of IT facilities is especially stronger in elementary schools than in others school levels. On the other hand, where a full-time school librarian called “gakko-shisho” is assigned, an interviewee from upper secondary school seemed to approve of integrating library and IT facilities and education under the control of library staff.
Dp2: Support for Teaching
In this section, there was some confusion among the participants about the concept of “Support for Teaching.” The results of the questionnaire also show that “acceptance of book requests from teachers”, “special loan to teachers” and “preparation of books for classes” are considered main tasks in this area. The more direct task of teaching support like “team teaching” and “information service on the copyright law” is not usually done in Japanese school libraries. This implies there are problems regarding the extent of present library functions and its personnel.
Dp3: Evasion of School Library Tasks (Second Priority to Teachers)
This topic is from the derivative flow. One speaker pointed out a case in a school in which some teachers tend to hide the fact that they have a shisho-kyouyu certificate in order to evade duties. Others participants agreed. This phenomenon is neither new nor unusual even to the researchers. However, it was never told in public before. Some speeches supported the opinion that a shisho-kyouyu is not a librarian but a teacher. It means that the school library management is only a second priority to teachers. The present shisho-kyouyu certificate is based on a teacher’s license and some additional credits in school library management. As Nakamura (2000) pointed in her comparative study between Hawaii and Tokyo, the teachers’ perceptions of school libraries are influenced by the nature of school library development in the society. However, if such perceptions limit the potential and possibilities of the school library professionals of the future, it may conflict with the whole LIPER discussion to inquire information professionals in the information age.
Dp4: Essentials of Library Science (Impact of Classification)
Some speakers had the opinion that the classification was the most impressive content during the training courses of shisho-kyouyu and librarian. This suggests that knowledge and skills on information and knowledge management should be included in the training process of school library professionals.
Dp5: Qualification, Certification and Training
Regarding qualification and certification, the problem does not lie in the qualification and certification system itself, but rather in the allocation of shisho-kyouyu to each school, the condition of working time and the reduction of other tasks inside school. Regarding the training system, some participants doubted the efficiency of correspondence education provided by some universities and various sorts of trainings offered by the school library or national and local teachers’ associations. Practical case studies with audio-visual teaching materials and library visits are much more welcomed than theoretical lectures. For on-the-job re-education and improvement of the skills of school library staff, advice from experienced teachers and librarians and IT training by municipal governments in a school district are good learning resources. As mentioned in the early part of this report, the present situation of the school library and its personnel in Japan extremely varies. This makes it difficult to set an appropriate course for school library staff with various qualifications.
Dp6: Social Needs to School Library
Inspite of the tough situation of school libraries in Japan, almost every speaker expressed the importance and meaning of school library. Sometimes, their speeches jumped to the presence of libraries in the Japanese society. A speaker proposed making a new subject called “library study” in order show the appeal of the functions of libraries. It matches the ideas in the previous interviews of the authorities
of educational fields. Their talks sometimes expanded to include society and some of them pointed the importance of reading literacy and the “back to the basics” trend in recent Japanese education system.
A speaker expressed the opinion that *shisho-kyoyu* certificates are more for ordinary teachers than school library specialists.
**New Model of School Library Professional in Japan**
**“Information Specialist for School”**
LIPER-SL research groups concluded this research with the proposition for a future model of school library professional in Japan. This model was based on the series of researches explained in this report and additional interviews. In close cooperation with the other research groups of LIPER we decided to call the new model “Information Specialist for School”. LIPER-SL defines the missions of “Information Specialist for School” as follows:
- “Information Specialist for School” secures access to information and media for students and teachers in a learning community.
- “Information Specialist for School” gives support and instruction to students and teachers in a learning community to enable them to use information and media effectively.
To execute these missions, “Information Specialist for School” must handle with three following functions.
1. Information and knowledge management:
- Building a system of knowledge and information that is useful for a learning community.
- Designing an effective learning environment with resources ranging from printed materials to digital information for students and teachers.
2. Information and knowledge sharing:
- Preventing a digital divide by providing information on learning and teaching.
- Enhancing the growth of a learning community by encouraging the sharing of information and knowledge.
- Encouraging partnership with local communities and other educational institutions and information providers.
3. Utilization of information and knowledge:
- Contributing to literacy based on reading and use of information.
- Designing cross-curricular lesson plans.
- Providing support to teachers in their lesson plans and teaching
**Issues for the Future**
There are some necessary steps needed to introduce the “Information Specialist for School” into the Japanese school system. The following possibilities can be suggested for the necessary revision of the School Library Law and other related educational laws:
- “Information Specialist for School” as expansion of present *shisho-kyoyu*
- “Information Specialist for School” as a brand-new and independent specialist in school
The LIPER project aims to design a course on the graduate school level for an information specialist for every type of library. In that case, the second possibility seems to match the original purpose of the LIPER project.
**Conclusion**
In this information age, the management of learning resources in educational institutions is an important task. In Japan, the “Law Concerning the Promotion of Children’s Reading” (Dec. 2001), and the “Law Concerning the Promotion of Culture on Letters and Printed Materials” (July, 2005) were passed one after another. Moreover, the Ministry of Education, Culture, Sports and Science and Technology decided to organize a national “Program for the Development of Reading Literacy” from April 2006. This program strongly reflects the views of the Organization for Economic Co-operation and Development (OECD) Program for International Student Assessment on reading ability, for example. This does not only refer to literature reading, but also to visual literacy and the way of thinking. In this program, a school library’s role to support students’ learning was mentioned. However, at this moment and as this research reveals, the expertise of a school library professional is regarded as that of a part-time teacher. To lead to a breakthrough, a new model of “Information Specialist for School” must
show that such specialists play a significant role in integrating information and media tasks that are separated in schools now and change a traditional book room and other media facilities into a lively learning resources center. LIPER-SL’s research will hopefully contribute to the policy making of the school library management and educational reform in Japan.
Acknowledgments
Other LIPER-SL research group members are:
- Research Representative, Akira NEMOTO, Graduate School of Education, The University of Tokyo,
- Head of LIPER-SL, Teruyo HORIKAWA, Shimane Prefectural Shimane Women's College
- Yuji HIRAKUE, Graduate School of Library, Information and Media Studies, University of Tsukuba
- Norio KATAOKA, Kanagawa Prefectural Asamizodai High School
- Yuriko NAKAMURA, Department of Education and Culture, Faculty of Social Studies, Doshisha University
- Research Advisor, Setsuko KOGA, Professor Emerita, Aoyama Gakuin University
This research was funded by the Japanese Government under the “Grant-in-Aid for Scientific Research”.
References
International Relations Committee of Japan Library Association (1994). Librarianship in Japan (Rev. ed.). Japan Library Association.
Kasai, Y., & Weintraub, H. (2002). ICT in Japan: possibilities and directions ACITT (The Association for ICT in Education) Journal. [Online]. Retrieved 25 January 2006 from http://acitt.digitalbrain.com/acitt/web/resources/pubs/Journal%2003/japan.htm?skinmode=3
LIPER (Library and Information Professions and Education Renewal) homepage. [Online]. Retrieved 25 January 2006 from http://www.soc.nii.ac.jp/sli/sl/liper/ (Only in Japanese)
Nakamura, Y. (2000). Teachers’ perceptions of school libraries: Comparisons from Tokyo and Honolulu. School Libraries Worldwide. Vol 6, No 1, 2000, p.66-87. The abstract is available at School Library Online. [Online]. Retrieved 25 January 2006 from http://www.iasl-slo.org/sli/ajan00.html
OECD Program for international student assessment (2000). What PISA produces – PISA [Online]. Retrieved 25 January 2006 from http://www.pisa.oecd.org/pages/0,2966,en_32252351_32236159_1_1_1_1_1_1_00.html
OECD Program for international student assessment (2003). What PISA produces – PISA [Online]. Retrieved 25 January 2006 from http://www.pisa.oecd.org/pages/0,2966,en_32252351_32236173_1_1_1_1_1_1_00.html
Sakai, C., Nakamura, Y., & Kitamura, Y. (2002). Japanese librarians learning from American school librarianship (May/June 2002). MultiMedia Schools Magazine. [Online] Retrieved 25 January 2006 from http://www.infotoday.com/MMSchools/may02/sakai_nakamura_kitamura.htm
Vaughn, S., Schumm, J.S., & J.M., Sinagub (1996). Focus group interviews in education and psychology. SAGE.
Appendix
101 questions on task analysis
| Task in School Library | Category |
|------------------------|----------|
| 1. Use the internet as information source | A |
| 2. Support the operation of computers and other equipment | B |
| 3. Cooperate with volunteers | F |
| 4. Collect reference tools | A |
| 5. Make signs in library | E |
| 6. Make a standard of selection of printed materials | A |
| 7. Make a budget for printed materials and buy them | A |
| 8. Select the printed materials, including holding a meeting for selection | A |
| | |
|---|---|
| 9. Display in library | E |
| 10. Use various standards for library | E |
| 11. Report and consult with administrator | F |
| 12. Cooperate with other school libraries | F |
| 13. Design the layout of library | E |
| 14. Manage network server | B |
| 15. Give advise on the copyright law to teachers | D |
| 16. Provide current awareness service to teachers | D |
| 17. Provide reference service to teachers | D |
| 18. Provide referral service to teachers | D |
| 19. Conduct training on information and media use for teachers | D |
| 20. Conduct training on library use for teachers | D |
| 21. Provide teachers with information on the copyright law | D |
| 22. Special loan of teaching materials for teachers | D |
| 23. Collect materials for teacher training | D |
| 24. Accept book purchase requests from teachers | D |
| 25. Make booklists or book reviews for teachers | D |
| 26. Make library guide or manual for teachers | D |
| 27. Join the teachers’ meeting as library staff | F |
| 28. Cooperate with public libraries | F |
| 29. Join inter-library network of public libraries and others | B |
| 30. Build database of in-school materials and students’ works | B |
| 31. Lend audio-visual equipment | A |
| 32. Maintain audio-visual equipment | A |
| 33. Make a standard of selection of audio-visual materials | A |
| 34. Make a budget for audio-visual materials and buy them | A |
| 35. Select audio-visual materials | A |
| 36. Make a standard for discard | A |
| 37. Classify materials | E |
| 38. Provide reference service to students | C |
| 39. Provide referral service to students | C |
| 40. Teach students sense of morality and ethics in information society | C |
| 41. Teach students library and book use | C |
| 42. Give students library orientation | C |
| 43. Inform students of the copyright law | C |
| 44. Give students information on the copyright law | C |
| 45. Provide reading guidance to students | C |
| 46. Provide special loan for students for research and report writing | C |
| 47. Accept book purchase requests from students | C |
| 48. Hold a reading circle (event) for students | C |
| 49. Make booklists and book reviews for students | C |
| 50. Make library guide or manual for students | C |
| 51. Prepare teaching materials for class | D |
| 52. Help teachers make lesson plans | D |
| 53. Check bookshelves | E |
| 54. Make curriculum of education on information and media | C |
| 55. Set goal and evaluation standard for education on information and media | C |
| 56. Teach subjects on information and media | C |
| 57. Make tools like pathfinder and information files | A |
| 58. Plan library orientation or library tour | C |
| 59. Make a systematic plan of library and book use | C |
| 60. Hold regular students’ library committee meetings | C |
| 61. Instruct students’ committee about events | C |
| 62. Participate in events of library associations and academic societies | F |
| 63. Join team teaching | D |
| | |
|---|---|
| 64. Allocate task in library | E |
| 65. Communicate with network company (network management in school) | B |
| 66. Make a list of webpage useful for learning of on library webpage | B |
| 67. Make and maintain a library homepage | B |
| 68. Make an annual report of library | E |
| 69. Evaluate own library | E |
| 70. Open school library to local communities | F |
| 71. Make rules and regulations for library management | E |
| 72. Make manuals for library management | E |
| 73. Plan library management | E |
| 74. Provide a card catalog to students and teachers | E |
| 75. Make a computer catalog (OPAC) | B |
| 76. Provide a computer catalog (OPAC) | B |
| 77. Discard materials | A |
| 78. Evaluate the collection | A |
| 79. Accept student trainees for internships | F |
| 80. Plan library events | E |
| 81. Join the curriculum meeting as library staff | D |
| 82. Join the school subjects meeting as library staff | D |
| 83. Maintain computers in library | B |
| 84. Maintain networked computers in library | B |
| 85. Make a manual for networked computers in library | B |
| 86. Publish library newsletter | F |
| 87. Make statistical data in library | E |
| 88. Make and use reservation forms for library | E |
| 89. Promote an all-school reading hour | C |
| 90. Hold a school meeting for library management | F |
| 91. Make a long-term plan on collection | A |
| 92. Make a standard to select digital materials | A |
| 93. Plan a budget for digital materials and buy them | A |
| 94. Select digital materials, including holding a meeting for selection | A |
| 95. Manage books outside library, in special rooms and classroom libraries | A |
| 96. Support special education for handicapped students | C |
| 97. Plan reading literacy lessons | C |
| 98. Plan a reading week | E |
| 99. Conduct activities such as book talks and reading aloud | C |
| 100. Participate in a meeting on school libraries inside and outside school | F |
| 101. Cooperate with museums | F | |
Native Among Savages: Reading the Escape Narrative of Dudhnath Tewari
By Satadru Sen
On the 23rd of April, 1858, Dudhnath Tewari, a sepoy convicted by the British regime in India of mutiny and desertion and sentenced to hard labor in transportation, escaped from his guards in the Andaman Islands penal colony. Reappearing in the town of Port Blair a year later, he warned J.P. Walker, the superintendent of the prison, about an impending attack on the colony by Andamanese aborigines. He also reported to Walker about his year among the aborigines. The warning allowed Walker to organize an effective defense in the "Battle of Aberdeen" of May 1859. Tewari was rewarded with a pardon and briefly became a minor celebrity. (He has, logically enough, become a 'traitor' in Indian-nationalist narratives and recently been accused of contributing to Andamanese extinction.) What Tewari told Walker became the basis of an article, "Adopted in Andamans," which appeared in Chambers's Journal in 1860. It remains a rare narrative about a "native" among "savages," often recycled for European consumption during a time when concretizing the distinctions between colonized populations was increasingly central to the organization of empire.
Located in the Bay of Bengal to the south of the Indian mainland, the Andamans were the largest and most complex prison in British India. There, administrators not only sought to punish and rehabilitate troublesome mainlanders, they also engaged in interconnected experiments in warfare, ethnology and photography involving the indigenous population. British engagement with the Andamanese proceeded alongside a parallel colonial encounter: that between aborigines and convicts from India and Burma. White Britons could not mediate whenever convict-settlers confronted aborigines demanding bananas, convict officers impregnated indigenous women at the Andaman Homes (institutions where aborigines were detained and occasionally put to work), or runaways met the inhabitants of the jungle. Indians encountered the Andamanese in overlapping roles: as prisoners of the Raj, as sepoys, and as settlers in their own right. In each of these capacities, they approached aborigines with their own political and cultural lenses and their own agendas of power, interpretation and pleasure.
The lenses and agendas were only partially autonomous, because the niches that Indians occupied in the Andamans were all essentially colonial: they existed with (imperfect) reference to British demands on occupation, spatial distribution, political status and vocabulary. Nevertheless, non-whites in the islands did not approach each other exclusively as "Indians" or "Andamanese." Those categories were more consistently meaningful to Britons than to their subjects, who learned and revised their assigned identities in the process of their own experience.
Delinquent behavior (such as escape) and illicit knowledge informed shadow identities in which official categories of aboriginal/savage and native/criminal were not so much irrelevant as differently meaningful.
For Indians criminalized and punished by the colonial regime, Britons and British authority came with a familiarity that was both oppressive and reassuring. But the Andamanese also became familiar to them in the course of their punishment, and even unfamiliarity was charged with social and political possibilities. The savage personified the limits of British power. Convicts often feared the Andamanese, but they also looked to them for shelter from the punishing colony. Sepoys like Tewari were not "one" with the regime they served in most circumstances, and armed convicts in the islands existed at a double ideological remove from their white jailor-commanders. They fought for the regime and occasionally shared its priorities, but they also fought for themselves, and they fought their own savages, who often remained invisible or incredible to Britons.
The Andamanese themselves maneuvered between licit and illicit relationships with convicts, adopting contingent positions of alliance and animosity, intimacy and distance. They killed or captured runaways in exchange for rewards from the regime. They also appeared to distinguish between the agents and the victims of the penal colony, attacking the former and helping the latter. They were both discriminating and random, and this simultaneity – while not inconsistent with the normative 'treachery' of 'ignoble' savages – rendered them inscrutable and incoherent as a colonized population. Not surprisingly, Britons in the Andamans saw autonomous relations between convicts and aborigines as an ideological and political threat that had to be managed rhetorically when it could not be defeated. The article in Chambers's Journal should be read as one such exercise in political management.
"Adopted in Andamans" was not received without skepticism in its own time, but even M.V. Portman – the pre-eminent ethnologist-historian of the colonial Andamans – grudgingly acknowledged that Tewari was "not nearly such a liar as he was supposed to have been." Tewari's overlapping reputations as a hero, a liar, and a source were grounded not just in his status as an informant who saved the colony, but in what he did in the year when he went missing from the colony. He had gone where no Indian – or Briton, which was entirely significant – had apparently gone before. He is in some ways one of the first anthropologists of the Andamanese, but as Portman's half-amused, half-exasperated acceptance of the narrative indicates, he could not be accepted as such by contemporary Britons. He was ultimately a silly native who wanted a full-fledged wedding in the jungle; his observations were not ethnology but ethnological data, once they had been prodded with leading questions and filtered by editors. In spite of (and because of) his willingness to compare
Andamanese and Bengali wives, he was himself an object of ethnology. He was, simultaneously, a myth and a monster authored by others. F.J. Mouat, one of the first Britons to study the Andamanese "scientifically," compared Tewari with Munchausen and remarked that it would be "amusing to see the offspring of the giant Brahmin and the tiny Mincopie."¹⁰ Partially anticipating Portman's belief that children of Andamanese mothers and convict fathers were 'interesting scientifically,'¹¹ Mouat's comment indicates that the pleasures and anxieties of freakish discovery could be located within the science of convict-aborigine encounters, reassuringly appropriated by fantasists and scientists.
Yet if we cut Mouat, Munchausen and Portman out of the picture, there is still something left of Tewari. Once with the Andamanese, Tewari dresses as they do, eats as they do, reproduces with them. At the same time, he retains an outsider's perspective that can be separated from British perspectives. He is not Portman, not Kurtz, not even Kim. He resembles the fictitious and autobiographical "beachcombers" who sought, and failed, to establish relations of reciprocity with island natives,¹² but he is himself a native. He is highly conscious that he has left the colony behind, and constantly compares the colony and the jungle. He notices not only that his Andamanese captors, like his British captors, deny him access to weapons, but also that, unlike Britons, they do not force him to work. He understands that a new social and categorical possibility has opened for him: like the jailors in Port Blair, the Andamanese see marriage as a way of stabilizing the disorderly,¹³ but the women they offer Tewari are their own daughters; they seek to transform captivity into adoption by integrating him into their society more closely than Britons were willing to contemplate. He is not entirely resistant but neither is he absorbed: he calculates his political odds, gambles and wins. He forms his own ideas of what his British interrogators will want to hear, what will titillate them, what numbers will impress them. His convict self, with its convict eye – informed by British, north-Indian-Brahmin and Andamanese inputs – has a tenacious reality of its own, which sees and configures savagery and colonialism for its own purposes, and which compels British listeners and readers into acts of compensation.
Tewari was only one of many convicts who lived among the islanders in the Andamans. His story is the first, and most detailed, of a particular genre of texts from the islands: the "captivity narrative" (which is also an escape narrative) of the convict among savages. Nearly all such narratives are mediated through Britons: typically, convicts who had escaped into the jungle or been captured by aborigines would return and tell their tales for British administrators and storytellers to rephrase. The lack of unambiguous authorship allows us to glimpse how a given experience of savagery could serve multiple agendas simultaneously. Partial, temporary integrations of Indian escapees into the aboriginal world
might be seen as phenomena in which the edge of a society functioned as a resource, an experimental space, an instrument of power and, inevitably, a problem of colonial governance.
Escape narratives in the Andamans were informed by a broad set of jungle myths that permeated convict society. In the 1850s and 1860s, convicts insisted that Great Andaman was connected by land to India or Burma; Tewari and other escapees went through the motions of searching for a route that led through the jungle to a world beyond the colony. Some convicts informed each other, and ultimately their jailors, that a powerful king ruled a hidden kingdom on the other side of the wall of trees; others placed this kingdom in the Cocos Islands.\(^{13}\) Sometimes the kingdom was conflated with Burma; at other times it was a separate magnet for escapees who, having once been sepoys, declared that they wanted to "take up service" with the unseen king.\(^{14}\) Convicts thus had their own geography of the Andamans, with its own tracks in the jungle leading to destinations and historical outcomes that were distinct from the roads and sightlines of the British imagination. Paul Carter has observed, in the Australian context, that convicts borrowed their jailors' visions of inside/outside to construct their fantasies of escape, and that these apparently irrational fantasies (of roads to China or Botany Bay) engaged and subverted the logic of the penal colony.\(^{15}\) In the Andamans, administrators worried that the expanding network of roads was not under their control, and that convicts walked out of the colony and into the jungle using the colony's own tracks.\(^{16}\) When Britons sought to defeat the aborigines and assert some control over the fantasy of escape (they could hardly control the escapees themselves), they also sought to defeat this alternative jungle and its horizon.
The convicts understood that this parallel jungle made sense only when it was full of savages. These alternative savages were a metaphor not only of terror, but also of resistance. The internal chatter of convict society that seeped into the colonial archive is full of references to secret villages inhabited by mixed communities of aborigines, runaways and maroons.\(^{17}\) Henry Corbyn, who created the Andaman Home, heard about sick and exhausted escapees taken by angelic aborigines to a huge camp in the recesses of Great Andaman, to be fed, medically treated, and released.\(^{18}\) A returned escapee named Boorhana told Walker about two dozen Indian and Afghan pilgrims, rescued by the Andamanese from a Turkish lifeboat, living with aborigines on Rutland Island.\(^{19}\) Unable to dismiss the story, Walker professed his inability to react: he lacked the resources for a military operation that might recover the maroons. Decades after the incident, Portman sought to reassert a shaken order by declaring that the story was obviously false, and that the Andamanese had assured him retrospectively that they would have "massacred" any Indians in that
situation.\textsuperscript{20} He conceded, however, that the Andamanese had occasionally sheltered runaways “for a very short time,” and may have wanted to “make common cause” with them against the British.\textsuperscript{21}
Walker’s half-believing paralysis and Portman’s fear of a “common cause” underline the value of the story to the convicts. Stories about Indians living amongst the Andamanese represent alternative social-racial orders that are within the geography of British colonialism and yet outside British authority. The mythical villages are experiments in which the premises and procedures have not been determined by Britons. The jungle is a transforming space: here, runaways pretend to be settlers, Indians fade contingently into the aboriginal world, mutineers become ethnologists.\textsuperscript{22} The Andamanese, imagined by Britons as indiscriminate killers and the scourge of the shipwrecked, make alarming appearances as rescuers and friends of Indians. Such discriminating savages, narrated by other discriminating natives, manipulate individual colonizers and subvert the values attached by the regime to race, savagery and experimentation.
Broadly speaking, British administrators as ideologically different as Walker and Portman were committed to a vision of permanent race-war in the Andamans. They believed that “the Andamanese” would inevitably clash with “the Indians,” and that Britons must manage the conflict, aligning strategically with one group or the other. The major difference between Walker and Portman on this issue is that Walker tended to align himself with Indians and Portman with the Andamanese. Otherwise, they shared an administrative outlook derived not only from colonialism in India, with its emphasis on the management of mutually hostile social compartments,\textsuperscript{23} but also from precarious oppositions between savagery, an inferior (inexpert, unenlightened, criminalized, non-white) civilization, and a dominant civilization that monopolized the ability to strategize. It did not dismiss cooperation between convicts and aborigines, but restricted that possibility to the tame worlds of the penal colony and the Andaman Homes. The jungle, from this perspective, was wild not only because it was unsupervised and antithetical to the possibility of natives of different stripes working peacefully together, but also because sometimes they did in fact work (and eat, sleep, and strategize) together and lose their stripes in the process.
The colonial government understood that it was powerless to pre-empt these irruptions, and sought to come to terms with them. Superintendent J.C. Haughton was instructed by the Home Department in 1860 that, “convicts who have...lived with the aborigines, may be made the means of opening a more friendly intercourse with them.”\textsuperscript{24} By taking this position, the regime impressed its own authority on processes – escape and return – that otherwise indicated its limitations. It insisted, in other words, on the prerogative of concession by conceding to savagery patches of space in
the jungle, and giving up considerable power to the convicts. The fragmentation of the savage in the jungle thus enabled the "friendly intercourse" between islanders and convicts that the Home Department chose to interpret as a sign of successful governance and Portman later associated with the ineptitude of colonial rule.
It is tempting, in the light of the evidence of "fellowship" in the jungle, to see convict-aboriginal relations as the solidarity of two colonized populations. It would not be a wild leap: Britons who sought to depoliticize the jungle (by insisting, for instance, that convicts obeyed aborigines only because the latter were terrifying) did so precisely because they saw a common cause. They knew that the Andamanese often appeared to liberate convicts working in the jungle, and that aborigines shot guards but so frequently spared chained prisoners that guards refused to wear badges or other signs of authority.\textsuperscript{25} Nor would such solidarity be unique in the history of penal colonies established among aborigines.\textsuperscript{26}
Such possibilities must, however, be treated with skepticism in the Andamans. There was little consistency in the Andamanese treatment of convicts and guards, not least because the political distinction between convicts and guards was not always obvious in an environment where many guards were also convicts, and those in the vanguard of forest-clearing and settlement were also natives, prisoners and colonial subjects. Neither convicts nor their jailors had a fixed vision of the political identity of Indians in the islands, and aborigines could not be expected to read this identity predictably. Convicts contemplating escape could take nothing for granted.
Tewari's double escape and subsequent pardon generated a small flurry of episodes in which returned runaways told the Port Blair authorities about impending attacks by aboriginal armies. The story that Boorhana told Walker is very similar to Tewari's, except that the invasion force poised for "imminent" attack on Port Blair was given as five hundred canoes and a thousand aborigines.\textsuperscript{27} The Andamanese may not have been able to "count beyond two,"\textsuperscript{28} but that only indicated the importance of numbers to the political community of civilization, and convicts knew it. Just as the regime had an interest in constructing an ambiguously cannibalistic savage that would deter escape attempts, convicts discovered the savage as a military menace that might balance the power of the colony, generate political rewards and produce the pleasure of frightening the jailors. This was, in a sense, the cultivation of a particular mode of settler-colonial paranoia that could speak to (and manipulate) the anxieties of Britons like Walker, who behaved as if they stood between besieged colony and besieging jungle. The British retelling of Tewari's story might be seen as an attempt to reassert confidence and control in anxious circumstances.
Notes:
1 Dilip Kumar Charkraborti, *The Great Andamanese: Struggling for Survival* (Calcutta: Seagull, 1990), iii-iv; T.N. Pandit, *The Sentinelese* (Calcutta: Seagull, 1990), 4
2 "Adopted in Andaman," *Chambers's Journal of Popular Literature*, no. 325, March 24, 1860, 177-9. On Tewari's return to the Andamans in 1866, see Edward Belcher, "Notes on the Andaman Islands," *Ethnological Society Transactions*, New Series 5 (1866).
3 Satadru Sen, *Disciplining Punishment: Colonialism and Convict Society in the Andaman Islands* (Delhi: Oxford University Press, 2000), passim; 'Savage Bodies, Civilized Pleasures: M.V. Portman and the Andamanese,' *American Ethnologist* 36 (2009): 364-379.
4 Satadru Sen, 'Policing the Savage: Segregation, Labor and State Medicine in the Andamans,' *Journal of Asian Studies* 58(1999): 753-773.
5 On sepoy politics, see Heather Streets, *Martial Races* (Manchester: Manchester University Press, 2004), 18-51; Michael Fisher, *The First Indian Author in English* (Delhi: Oxford University Press, 1996), 122-4, 132-6.
6 Bernard Smith, *European Vision and the South Pacific* (New Haven: Yale University Press, 1985), 132-58; Ronald Meek, *Social Science and the Ignoble Savage* (Cambridge: Cambridge University Press, 1976), 3
7 E.H. Man believed that Tewari had concocted portions of his story. See his "On the Aboriginal Inhabitants of the Andaman Islands," *Journal of the Anthropological Institute of Great Britain and Ireland* 12 (1883): 139-40.
8 M.V. Portman, *A History of Our Relations with the Andamanese* (Calcutta: Office of the Superintendent of Government Printing, 1899), 279.
9 "Doodnath...assured me that he could discover no trace of religious worship or the acknowledgment of any unseen power among them," one administrator noted. Home Department records, Government of India, National Archives, New Delhi (in subsequent references, HD records), March 27, 1863, 85
10 F.J. Mouat, *Adventures and Researches Among the Andaman Islanders* (London: Hurst & Blackett, 1863), 299, 310. "Mincopie" was an early British term for the Andamanese.
11 Report of the Superintendent of the Andaman Islands (RSAI), 1892-93, HD records.
12 Rod Edmonds, *Representing the South Pacific: Colonial Discourse from Cook to Gauguin* (Cambridge: Cambridge University Press, 1997), 63
13 Satadru Sen, "Rationing Sex: Female Convicts in the Andamans," *South Asia*, 30 (1999): 29-59.
14 Henry Corbyn, Narrative No. 2, HD Records, 1863.
15 Walker to Home Department, April 1858, HD Records, 1858.
16 Paul Carter, *The Road to Botany Bay* (Chicago: University of Chicago Press, 1987), 293-319
17 RSAI, 1882-83.
18 RSAI, 1859-60.
19 Corbyn, Narrative No. 3, HD records, 1864.
20 RSAI, 1858-59.
21 Portman, 288.
22 Ibid, 294.
23 RSAI, 1875-76.
24 Thomas Metcalf, *Ideologies of the Raj* (Cambridge: Cambridge University Press, 1995), 113-59.
25 Government of India to Haughton, January 3, 1860, HD records 1860, 436.
26 Portman, 276-8.
27 Alice Bullard, *Exile To Paradise: Savagery and Civilization in Paris and the South Pacific* (Stanford: Stanford University Press, 2000), 201-3.
28 RSAI, 1858-59.
29 Portman, 40; Maulana Jafar Thanesari, *Kalapani* (Delhi: Urdu Markaz, 1964), 64. |
MAINTAINING ISLET QUALITY DURING CULTURE
by
Michael James Rappel
Bachelor of Science in Chemical Engineering
University of Minnesota – Minneapolis, 2001
Masters of Science in Chemical Engineering Practice
Massachusetts Institute of Technology, 2003
Submitted to the Department of Chemical Engineering
in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy in Chemical Engineering
at the
Massachusetts Institute of Technology
May 2007
[June 2007]
© 2007 Massachusetts Institute of Technology
All rights reserved.
Author...........................................................................................................
Department of Chemical Engineering
May 21, 2007
Certified by...............................................................................................
Clark K. Colton
Professor of Chemical Engineering
Thesis Supervisor
Certified by...............................................................................................
Kenneth A. Smith
Professor of Chemical Engineering
Thesis Supervisor
Accepted by...............................................................................................
William Deen
Professor of Chemical Engineering
Chairman, Committee for Graduate Students
The following is a list of the most important and frequently used terms in the field of computer science. It is intended to provide a quick reference for students, researchers, and professionals alike.
1. Algorithm: A step-by-step procedure for solving a problem or performing a task.
2. Data Structure: A way of organizing data in a computer so that it can be accessed efficiently.
3. Database: A collection of data organized in a structured manner for easy access, modification, and updating.
4. Database Management System (DBMS): A software system that manages databases.
5. Encryption: The process of converting information into a code so that only authorized parties can understand it.
6. Hashing: A process of converting data into a fixed-size string of characters.
7. Interface: A way for two systems to communicate with each other.
8. Network: A collection of computers and devices connected together.
9. Operating System (OS): A software program that controls and manages the hardware and software resources of a computer.
10. Programming Language: A set of instructions that a computer can understand and execute.
11. Query: A request for information from a database.
12. Security: The protection of data and systems from unauthorized access, use, disclosure, disruption, modification, or destruction.
13. Software: A set of instructions that tell a computer what to do.
14. System: A collection of interrelated components that work together to achieve a common goal.
15. User Interface: The way a user interacts with a computer system.
16. Virtual Machine: A software implementation of a computer system that runs on top of another computer system.
17. Web Application: A software application that runs on a web server and is accessed through a web browser.
18. Wireless Network: A network that uses radio waves to transmit data between devices.
19. XML: eXtensible Markup Language, a markup language that defines a set of rules for encoding documents in a format that is both human-readable and machine-readable.
20. YAML: Yet Another Markup Language, a data serialization language that is used to represent data structures in a human-readable format.
MAINTAINING ISLET QUALITY DURING CULTURE
by
Michael James Rappel
Submitted to the Department of Chemical Engineering on
May 21, 2007, in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy in Chemical Engineering
Islet transplantation has become a promising treatment for type I diabetes mellitus due to recent success since the development of the Edmonton Protocol. Islet culture prior to transplantation is standard practice in most clinical islet programs. Conventional culture conducted on polystyrene vessels can impose oxygen limitations even at relatively low tissue surface densities. High density islet culture is desirable because it reduces space and handling requirements during culture, but it exacerbates oxygen (O₂) limitations, causing a reduction in islet viability.
The overall objective of this thesis was to maintain islet quality during static culture. As a chemical engineer, I focused on addressing transport limitations present in conventional culture techniques. After demonstrating culture in the absence of O₂ transport limitations resulted in nearly 100% recovery of the original viable tissue placed into culture when the combined non-adherent and adherent tissue were considered, I examined the effect of tissue surface density on the recovery of islet tissue, its viability, and its purity for conventional normoxic culture on a polystyrene dish. With conventional culture, the fractional recovery of viable tissue decreased sharply as viable tissue density increased. To improve islet quality in high density culture, I investigated use of elevated ambient O₂, reduced culture temperature, and culture on an oxygen-permeable silicone rubber membrane. By applying a theoretical O₂ transport model, I investigated how O₂ transport changes for each culture condition and compared predictions to the experimental data to determine whether O₂ is limiting during high density culture using these new techniques. At high tissue surface densities, the fractional recovery of viable tissue was higher with culture on polystyrene in elevated (56%) O₂ or culture at reduced temperature (24°C), and even higher with normoxic culture on a silicone rubber membrane. Theoretical predictions based on O₂ transport were qualitatively similar to experimental results but in general overpredicted the amount of viable tissue recovered. Additional theoretical calculations indicated simplifications made when modeling oxygen along with glucose and pH changes during culture could account for the slight overprediction.
In conclusion, in high density culture, recovery of viable tissue (1) decreases as culture density increases on a polystyrene surface; (2) increases with increasing external O₂; and (3) increases substantially with culture on silicone rubber by removing O₂ limitations. The techniques examined significantly improve tissue oxygenation compared to conventional culture, and allow tissue to be cultured at higher densities without a reduction in viability. These methods can be easily implemented, which would enable clinical centers to reduce space and handling requirements during culture prior to transplantation without the reduction in islet viability that can occur with conventional methods, and thereby maximize the use of limited islet resources.
Thesis Supervisors: Clark K. Colton and Kenneth A. Smith
Title: Professors of Chemical Engineering
Acknowledgements
This thesis was completed only because of the help and support I received from many people. I would like to start by thanking my thesis advisors, Clark K. Colton and Kenneth A. Smith, for their technical guidance and demand for excellence. I thank my committee members William Deen, Leo Otterbein, Klearchos Papas, and Gordon Weir, who have provided invaluable insights and discussion throughout my time here at MIT. Additionally, I thank Susan Bonner-Weir for her insights and assistance in understanding changes islet tissue endures during culture.
I express my deepest gratitude to Anna Pisania and Daryl Powers, who not only provided stimulating discussions, but also made working in the lab enjoyable. I am indebted to them for their help, kindness, and support throughout my time at MIT. I could not have chosen any better lab mates.
I also thank the rest of the Colton Lab for their support throughout the years. I thank Efstatios Avgoustiniatos who provided the base work for this thesis and answered many of my initial questions; Carneron Nienaber from Joslin Diabetes Center for her lessons and patience in teaching me how to perform insulin staining; and Shenghan Yan, a UROP student who worked on the oxygen biosensor system. Shenghan was the ideal UROP student, both hard working and enthusiastic, and was an absolute joy to have in the lab.
I thank the islet isolation core at Joslin Diabetes Center, which includes Vaja Tchipashvili, Gaurav Chandra, James Lei, Abdulkadir Omer, Jack O’Neil, and Vassileios Kostaras, and the ICR for providing the islet tissue used throughout this work, without which none of this would have been possible. I also thank Giner, Inc., particularly Linda Templeman, and the University of Minnesota, particularly Klearchos Papas and Phillip Rozak, for their efforts in our collaborative work on developing and testing a controlled environment container for transportation of islet tissue.
I am particularly thankful to my parents, Don and Elaine Rappel, for their unending love and support throughout my academic education. Without them, I would not be where I am today. I also thank my brothers, Matt and Andy Rappel, who have made it their duty to keep my ego in check. I also thank all my friends I have had the honor of knowing during my years at MIT, who have provided support, laughs, and fun times.
Finally I want to thank my fiancée, Wanda Lau. I cannot begin to list all she has done for me. I am very thankful for her love and support throughout the years we have been together. Even when research was not going well she always found a way to make me smile.
Funding for this research was provided by grants from the NIH and the JDRF Center for Islet Transplantation at Harvard Medical School.
# Table of Contents
**List of Figures** ................................................................................................................................. 9
**List of Tables** ............................................................................................................................... 13
## 1 Background, objectives, and overview .................................................................................. 14
1.1 Background .............................................................................................................................. 14
1.2 Objectives ............................................................................................................................... 16
1.3 Overview ................................................................................................................................. 17
1.4 Nomenclature .......................................................................................................................... 18
## 2 Characterization of islet and non-islet tissue prior to and following culture ..................... 19
2.1 Introduction ............................................................................................................................ 19
2.2 Methods .................................................................................................................................. 20
2.3 Results ..................................................................................................................................... 24
2.3.1 DNA/cell ........................................................................................................................... 24
2.3.2 Nuclei and DNA recovery ............................................................................................... 25
2.3.3 OCR/cell and OCR/DNA ............................................................................................... 28
2.3.4 OCR recovery .................................................................................................................. 31
2.3.5 OCR/DNA comparisons between centers ..................................................................... 31
2.3.6 Membrane integrity ........................................................................................................ 31
2.3.7 OCR/cell with impermeable cell membranes by 7-AAD ........................................... 33
2.3.8 Purity ............................................................................................................................... 34
2.4 Discussion .............................................................................................................................. 34
2.5 Nomenclature ....................................................................................................................... 42
## 3 The effect of surface density on tissue quality during static islet culture on untreated polystyrene ................................................................................................................................. 44
3.1 Introduction ............................................................................................................................ 44
3.2 Methods .................................................................................................................................. 45
3.3 Theoretical oxygen transport model .................................................................................... 49
3.4 Results ..................................................................................................................................... 52
3.4.1 Predictions of theoretical mode ..................................................................................... 52
3.4.2 Visual observations ......................................................................................................... 55
3.4.3 Effect of surface density on tissue recovery. ............................................................... 57
3.4.4 Membrane integrity and oxygen consumption rate ..................................................... 58
3.4.5 Prediction of oxygen transport theory compared to experimental data ..................... 65
3.5 Discussion .............................................................................................................................. 67
3.6 Nomenclature ....................................................................................................................... 71
## 4 Reducing oxygen limitations during static islet culture ....................................................... 73
4.1 Introduction ............................................................................................................................ 73
4.2 Methods .................................................................................................................................. 74
4.3 Theoretical model .................................................................................................................. 78
4.3.1 Culture on oxygen-impermeable surfaces ..................................................................... 78
4.3.2 Culture on silicone rubber ............................................................................................. 78
4.4 Results ..................................................................................................................................... 82
4.4.1 Culture at elevated ambient oxygen partial pressure .................................................. 83
4.4.2 Culture at reduced temperature ..................................................................................... 83
4.4.3 Silicone rubber culture ................................................................. 84
4.4.4 Comparison with prediction of oxygen transport model .................. 85
4.5 Discussion ....................................................................................... 90
4.6 Nomenclature ................................................................................. 95
5 Oxygen transport during islet culture: A closer look ............................. 97
5.1 Introduction ................................................................................... 97
5.2 Theoretical model ......................................................................... 97
5.3 Inaccuracies in parameter values .................................................. 100
5.4 Inaccurate model assumptions ...................................................... 104
5.4.1 OCR density distribution ......................................................... 105
5.4.2 Islet size ............................................................................... 106
5.4.3 Tissue slab .......................................................................... 112
5.4.4 Islet stacking ....................................................................... 113
5.4.5 Convection .......................................................................... 113
5.5 Discussion .................................................................................... 114
5.6 Nomenclature .............................................................................. 119
6 Glucose transport during islet culture .................................................. 120
6.1 Introduction .................................................................................. 120
6.2 Methods ....................................................................................... 121
6.3 Theoretical glucose transport model ............................................. 123
6.4 Glucose consumption rate measurement ....................................... 126
6.5 Predictions of theoretical models ................................................ 129
6.6 Experimental results ..................................................................... 132
6.7 Discussion .................................................................................... 137
6.8 Nomenclature .............................................................................. 140
7 Changes in pH during static islet culture ............................................. 142
7.1 Introduction .................................................................................. 142
7.2 Methods ....................................................................................... 142
7.3 Theoretical pH model .................................................................. 144
7.4 Ammonia and lactate production rate measurements .................... 155
7.5 Predictions of theoretical model .................................................. 155
7.6 Discussion .................................................................................... 164
7.7 Nomenclature .............................................................................. 166
8 Improvements in islet transportation .................................................. 169
8.1 Introduction .................................................................................. 169
8.2 Giner, Inc., controlled environment container ............................... 170
8.2.1 Methods ............................................................................... 170
8.2.2 Shipment results .................................................................. 172
8.2.3 Discussion .......................................................................... 174
8.3 Passive shipping device ............................................................... 178
8.3.1 Oxygen ............................................................................... 179
8.3.2 Temperature ....................................................................... 180
8.4 Conclusions .................................................................................. 185
8.5 Nomenclature .............................................................................. 185
9 Conclusions and outlook ..................................................................... 187
9.1 Maintaining islet quality during culture .......................................................... 187
9.2 Future work ........................................................................................................... 189
10 Appendices .................................................................................................................. 191
10.1 Supplemental material for Chapter 2 ................................................................. 191
10.1.1 Nomenclature ............................................................................................. 197
10.2 Carbon monoxide (CO) treatment of βTC3 cells exposed to TNF-α or anoxia .. 199
10.2.1 Introduction .................................................................................................. 199
10.2.2 Methods ....................................................................................................... 200
10.2.3 Results .......................................................................................................... 202
10.2.4 Discussion .................................................................................................... 207
10.2.5 Nomenclature .............................................................................................. 208
10.3 Investigating oxygen consumption rate measurement in static cultures ........... 209
10.3.1 Introduction .................................................................................................. 209
10.3.2 Methods ....................................................................................................... 210
10.3.3 Theoretical model ......................................................................................... 212
10.3.4 Calculating experimental OCR from OBS .................................................. 219
10.3.5 Theoretical results ....................................................................................... 220
10.3.6 Experimental results ..................................................................................... 224
10.3.7 Discussion .................................................................................................... 225
10.3.8 Nomenclature .............................................................................................. 228
10.4 Matlab code for oxygen transport during static culture on polystyrene .......... 230
10.5 Matlab code for oxygen transport during static culture on silicone rubber .... 233
10.6 Matlab code for oxygen transport model with 2 sphere sizes ......................... 236
10.7 Matlab code for glucose transport .................................................................... 239
10.8 Matlab code for changes in local pH during culture on polystyrene ............... 242
10.9 Matlab code for changes in local pH during culture on silicone rubber .......... 248
10.10 Matlab code for oxygen transport in OBS – base case .................................. 254
10.11 Matlab code for oxygen transport in OBS – full model .................................. 256
10.12 Matlab code for oxygen transport in OBS – flat bottom ................................ 259
References ....................................................................................................................... 262
List of Figures
Figure 2.1. Frequency distributions of DNA content for day 0 and for non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, and (C) human cultured/shipped islets, and (D) human non-islet tissue. ...........................................26
Figure 2.2. Frequency distributions of the OCR/cell for day 0 and for non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, (C) human cultured/shipped islets, and (D) human non-islet tissue. ..................................................29
Figure 2.3. Frequency distributions of the OCR/DNA for day 0 and for non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, (C) human cultured/shipped islets, and (D) human non-islet tissue. ..................................................30
Figure 2.4. Frequency distributions of the OCR/DNA for different centers (MIT, MN) for (A) human, and (B) porcine islets. .............................................................................33
Figure 2.5. Frequency distributions of the fraction of cells with impermeable cell membranes by 7-AAD for day 0 and non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, (C) human cultured/shipped islets, and (D) human non-islet tissue. ..................................................35
Figure 2.6. Frequency distributions of the OCR/membrane-impermeable cell for day 0 and for non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, (C) human cultured/shipped islets, and (D) human non-islet tissue. ..................................................36
Figure 2.7. OCR/DNA versus fraction of cells with impermeable cell membranes by FDA/PI staining and 7-AAD for (A) rat islets, (B) human fresh islets. .........................37
Figure 3.1. Geometry assumed for analysis of high density islet culture on an impermeable surface. ............................................................................................................51
Figure 3.2. Predicted contours of constant partial pressure in and around islets in culture. .................................................................54
Figure 3.3. Theoretical prediction for minimum islet p (dashed line) and non-viable tissue volume (solid line) as a function of islet surface density and OCR density. ..............55
Figure 3.4. Light microscopy pictures of human islets. ...............................................................................................................................56
Figure 3.5. The fraction of original tissue collected from culture as (A) adherent, (B) non-adherent, and (C) total tissue as a function of the initial tissue surface density. ..........59
Figure 3.6. Data from Figure 3.5A are plotted in terms if the actual surface density of adherent tissue collected from culture versus the initial tissue surface density. ............60
Figure 3.7. The fraction of original cells collected as non-adherent tissue, plotted as (A) nuclei and (B) DNA, and (C) the DNA content per cell of tissue. .........................61
Figure 3.8. The volume fraction of insulin positive tissue excluding vascular space for adherent (A) and non-adherent (B) tissue collected from culture. ........................................62
Figure 3.9. Fraction of original insulin positive cells for adherent (A), non-adherent (B), and total tissue (C) collected from culture as a function of the initial tissue surface density. ..........................................................................................................................63
Figure 3.10. (A) Fraction of cells with intact membranes as measured by 7AAD sequential staining, (B) OCR/cell, and (C) OCR/DNA after culture versus initial tissue surface density. ........................................................................................................64
Figure 3.11. The fraction of original OCR remaining as a function of (A) initial tissue surface density and (B) initial OCR density following culture. ..................................................65
Figure 3.12. The OCR density of all tissue remaining after culture as a function of the initial OCR surface density. ...........................................................................................................66
Figure 3.13. Experimental results and theoretical predictions for the fraction of original viable tissue recovered from culture versus (A) tissue surface density and (B) OCR density. .............................................................................................................................................66
Figure 4.1. Geometry assumed for simulations of islet culture on an O$_2$-permeable surface. .........................................................................................................................................................82
Figure 4.2. OCR as a function of inverse temperature from three rat islet preparations. ........84
Figure 4.3. The total fraction of original cells recovered after culture with various culture techniques as a function of the initial tissue density. ...............................................................................................................................86
Figure 4.4. OCR/cell following culture as a function of the initial tissue density for different culture techniques. .........................................................................................................................87
Figure 4.5. Fraction of original OCR recovered from culture as a function of surface density for different culture techniques. .........................................................................................................................88
Figure 4.6. The fraction of $\beta$TC3 cells with intact cell membranes exposed to varying O$_2$ concentrations over 4 days of culture. .........................................................................................................................89
Figure 4.7. The fraction of tissue with intact cell membranes as a function of tissue surface density for culture at varying temperatures. .........................................................................................................................90
Figure 4.8. Experimental results and theoretical predictions for the fraction of original viable tissue collected following culture for various culture techniques. .........................91
Figure 5.1. The base case O$_2$ transport model predictions for culture in 3 mm and 22 mm on polystyrene and in 22 mm medium on a 500-$\mu$m silicone rubber membrane. ....98
Figure 5.2. Theoretical predictions for the fraction of original viable tissue after culture as a function of initial OCR surface density and $K_m$. .................................................................101
Figure 5.3. Theoretical predictions for the fraction of viable tissue after culture as a function of initial OCR surface density for different values of $(D\alpha)_i$. .........................102
Figure 5.4. Theoretical predictions for the fraction of viable tissue after culture as a function of initial OCR surface density for different values of $(D\alpha)_m$. .........................102
Figure 5.5. Theoretical predictions for the fraction of viable tissue after culture as a function of initial OCR surface density for different values of $p_{crit}$. .................................103
Figure 5.6. Theoretical predictions for the fraction of viable tissue after culture as a function of initial OCR surface density for different values of $(D\alpha)_{Si}$. .........................105
Figure 5.7. Light microscopy pictures of human islets. ..........................................................................................................................107
Figure 5.8. Piecewise function for theoretical predictions of $F_V$ as a function of initial OCR surface density for a uniform density distribution (base case). .........................108
Figure 5.9. Theoretical predictions using a uniform or a normally distributed OCR density. ..................................................................................................................................................108
Figure 5.10. Theoretical predictions for islets with uniform radii of 25, 75, and 125 µm on (A) a polystyrene dish in 3 mm medium or (B) a 500-µm silicone rubber membrane in 22 mm medium. .................................................................110
Figure 5.11. Model geometry and predictions assuming a mixture of islet sizes. ..................................................111
Figure 5.12. Theoretical predictions for $F_V$ assuming a normal islet size distribution. ......................112
Figure 5.13. Model geometry and predictions assuming a slab of tissue. ..................................................115
Figure 5.14. Model geometry and predictions assuming stacked islets. ..................................................116
Figure 6.1. Islet geometry assumed for modeling glucose transport during islet culture on either a polystyrene dish or a silicone rubber membrane. .................................................................125
Figure 6.2. The distribution of GCR measurements from human islet preparations. .......................129
Figure 6.3. Transient glucose profiles as a function of position during islet culture. .......................131
Figure 6.4. The minimum glucose concentration in the islet as a function of islet surface density for tissue cultured for 60 hr. ...........................................................................132
Figure 6.5. The maximum non-dimensional glucose concentration in the islet as a function of time and islet surface density ...........................................................................133
Figure 6.6. The maximum non-dimensional glucose concentration in the islet as a function of time and Da. .............................................................................................................134
Figure 6.7. The maximum non-dimensional glucose concentration in the islet as a function of time and $D_M/D_H$. .............................................................................................................135
Figure 6.8. The maximum non-dimensional glucose concentration in the islet as a function of time and $K_M/C_{initial}$. .............................................................................................................136
Figure 6.9. The fraction of original OCR collected from culture for human islets cultured in 3 mm of 0 or 10 mM glucose medium at < 8 x 10^4 cells/cm^2. ............................................137
Figure 7.1. Islet geometry assumed for modeling pH changes for islet tissue on either a polystyrene dish or a 500-µm silicone rubber membrane. .................................................................147
Figure 7.2. Transient pH profiles for islets as a function of position. ................................................157
Figure 7.3. The pH at the top of the islet as a function of time and islet density. ..........................158
Figure 7.4. The pH at the top of the islet as a function of time and $Da_{CO_2}$. ........................................160
Figure 7.5. The pH at the top of the islet as a function of time and $Da_{Kr}$. ........................................161
Figure 7.6. The pH at the top of the islet as a function of time and $Da_{Kr}$. ........................................162
Figure 7.7. The pH at the top of the islet as a function of time and $Da_{Lac-Am}$. ..............................163
Figure 8.1. (A) Islets piled along the edge in the silicone rubber bottom vessel sent in the controlled environment container. (B) Opposite side of the same culture vessel. ...............173
Figure 8.2. Islets following shipment in a solid flask in a Styrofoam box overnight. ......................173
Figure 8.3. Staining of islets shipped in solid flask in a Styrofoam box. ........................................174
Figure 8.4. Staining of islets shipped in controlled environment on silicone rubber surface. ........................................................................................................................................174
Figure 8.5. Schematic of passive shipping device that will provide sufficient oxygen and maintain temperature during shipment. .........................................................................................179
Figure 8.6. The amount of O$_2$ consumed by 50,000-500,000 IE for 36 hr at temperatures ranging from 10 to 40°C. ................................................................. 181
Figure 8.7. Volume of various PCMs required to maintain the phase change temperature of the container (18, 25, 30, or 36°C) for 24 hr during shipment at constant ambient temperature................................................................. 183
Figure 8.8. Sample profile for controlled PCM experiment ........................................... 184
Figure 10.1. Conversion of day 0 islet volume fraction to β cell volume fraction exclusive of vascular spaces. .................................................................................. 193
Figure 10.2. Conversion of day 0 islet volume fraction to number fraction β cells. .............. 193
Figure 10.3. Conversion of day 2 β cell volume fraction exclusive of vascular spaces to number fraction β cells. .................................................................................. 194
Figure 10.4. The fraction of cells with intact membranes versus TNF-α concentration. ......... 203
Figure 10.5. The fraction of cells with intact cell membranes with CO and TNF- α treatment. ........................................................................................................ 204
Figure 10.6. The fraction of βTC3 cells with intact cell membranes after exposure to anoxia. ........................................................................................................ 205
Figure 10.7. CO protection of βTC3 cells from anoxia ....................................................... 206
Figure 10.8. Geometry and the associated assumptions used for the three different theoretical models. .................................................................................................... 215
Figure 10.9. Theoretical oxygen profiles in the (A) idealized, (B) actual, and (C) fabricated OBS plates. ................................................................. 222
Figure 10.10. Predicted OCR/cell determined by the OBS (solid lines) and the average pO$_2$ in the tissue (dashed lines) as a function of cell density for a 96-well plate and Petri dish. ................................................................. 223
Figure 10.11. Predicted OCR/cell determined by the OBS as a function of oxygen permeability in polystyrene for a Petri dish, 24-well plate, and 96-well plate. .......... 223
Figure 10.12. Predicted OCR/cell determined by the OBS as a function of well radius and oxygen permeability in polystyrene. ................................................................. 224
Figure 10.13. OCR/cell determined by fabricated OBS plates as a function of well radius. 226
List of Tables
Table 2.1. DNA/cell and the fraction of original tissue recovered (quantified by nuclei and DNA) after 2 days culture ................................................................. 27
Table 2.2. OCR/cell, OCR/DNA, and fraction of original OCR recovered after 2 days culture ................................................................. 32
Table 2.3. Human islet and β cell purity measurements and fraction of β cells recovered after culture for 2 days at 37°C ................................................................. 39
Table 3.1 Values for the parameters used in islet culture simulations and the source from which they were obtained ................................................................. 53
Table 4.1. Oxygen diffusivities and solubilities at 37°C and 24°C in culture medium and tissue used in the theoretical O₂ transport model ................................................................. 81
Table 5.1. Base case parameter values used in simulations for O₂ transport during static islet culture .................................................................................. 99
Table 6.1. Literature values for glucose consumption by islet tissue ................................................................. 127
Table 6.2. Values for the parameters used in islet culture simulations and the source from which they were obtained ................................................................. 128
Table 7.1. Diffusivities of each species used in the model in tissue and in culture medium. 153
Table 7.2. Cellular production rates for CO₂, ammonia, and lactate. ................................................................. 154
Table 7.3. Parameter values used in pH model for high density islet culture. ................................................................. 155
Table 7.4. Damkohler numbers based on data in literature and empirical measurements used in the base case simulations. ................................................................. 156
Table 8.1. Conditions for islets shipped in the silicone rubber bottom vessel within the controlled environment container and in the solid T-25 flask in the Styrofoam box. ................................................................. 175
Table 8.2. Total tissue estimation for islets received from both the silicone rubber bottom vessel and in the solid flask. ................................................................. 175
Table 8.3. OCR/cell, OCR/DNA, total OCR, and the relative recovery of original viable tissue collected from both the silicone rubber and solid bottom dishes after transportation. ................................................................. 176
Table 8.4. Commercially available phase change materials and their properties. ................................................................. 181
Table 10.1. Parameter values used to estimate Φ_{BXV} and f₀ from day 0 measurements of Φ₁ and day 2 measurements of Φ_{BXV}. ................................................................. 195
Table 10.2. OCR/cell of βTC3 cells with and without CO present ................................................................. 205
Table 10.3. Parameters used in theoretical simulations for oxygen transport in the OBS device. ................................................................. 218
Table 10.4. Comparison of OCR/cell between cells in liquid and gelled agarose medium. .227
1 Background, objectives, and overview
1.1 Background
The American Diabetes Association (ADA) defines diabetes as a “disease in which the body does not produce or properly use insulin.” Humans need insulin in order for the body to use glucose, the basic fuel for cells. Diabetes is the fifth deadliest disease in the United States, according to the ADA. About 20.8 million people in the U.S. have diabetes; over 6 million of them have yet to be diagnosed. In 2005 1.5 million people over the age of 20 were diagnosed with diabetes or roughly 4,110 people each day. Diabetes can cause problems in virtually any part of the body, leading to nerve damage, amputation, kidney failure, stroke, blindness, and heart disease. In 2002 one out of every 10 dollars spent on health care was used to address diabetes and its complications, with a total estimated cost (direct and indirect medical costs) of $132 billion in the U.S. [1].
The ADA describes three major types of diabetes: type 1 or insulin-dependent diabetes, type 2 or non-insulin-dependent diabetes, and gestational. Type 1 diabetes, which typically develops in children and young adults, occurs when the body does not produce enough insulin, preventing cells from using glucose, and resulting in high blood glucose levels. The decrease in insulin production occurs when the body’s immune system destroys insulin-producing cells. An estimated 5-10% of peoples diagnosed with diabetes have type 1 diabetes. Type 2 diabetes, comprising 90-95% of cases, occurs when the body is unable to use insulin it produces, resulting in an increase in blood glucose levels; over time the pancreas loses its ability to produce insulin. Gestational diabetes occurs during pregnancy and typically disappears after birth, though several mothers who have gestational diabetes later develop type 2 diabetes in life [1].
Several approaches are under investigation for treatment of diabetes. Type 2 diabetes can frequently be treated through oral hypoglycemic drugs and a good diet, though a small portion of patients require insulin to control hyperglycemia [2]. Several treatments are also being investigated for Type 1 diabetes. Insulin therapy, which requires daily insulin injections and blood glucose monitoring, is the most common treatment. Though the therapy transforms diabetes into a chronic disease with secondary complications, it also prevents it from becoming rapidly fatal. Implantation of insulin sensors and pumps, artificial pancreas development [3] and stem cells [4; 5] are also considered as possible treatments for diabetes, but are still in the
developmental stages. Whole pancreas transplantation at some centers results in 85% insulin independence one year after transplant with 50% of the patients maintaining good metabolic control five years after. However, whole pancreas transplantation requires major surgery and has significant risks. Islet transplantation is technically a much easier treatment and still offers the possibility of preventing secondary complications. However, until 2000, only 12% of patients were off insulin one year after the transplant [6].
In 2000 the area of islet transplantation took a large leap forward when a group in Edmonton reported insulin independence in seven out of seven patients after islet transplantation [7]. Since the development of the “Edmonton protocol” the success rate of islet transplantation has increased substantially around the world. Sixty-five patients have received islet transplants at the University of Alberta. Of the patients who have completed the procedure (about 12,000 islet equivalents (IE)/kg body weight, where one IE is the tissue volume equivalent to a 150-\(\mu\)m diameter sphere, multiple transplantations usually required) roughly 65% are insulin free one year post transplant, 40% are insulin free two years post-transplant, and 25% are insulin free three years post-transplant. Even patients who return to insulin therapy still show some graft function with 80% of all patients demonstrating transplant function after five years. Though risks are associated with the procedure and the drugs used for immunosuppression, it is possible to achieve insulin independence through islet transplantation for more than five years [8].
The Edmonton protocol made a key modification to the islet transplantation procedure by transplanting a larger islet mass, requiring tissue from multiple donor pancreases [8]. The availability of donor organs limits the number of diabetic patients who can benefit from this procedure. The Edmonton protocol requires about 800,000 IE (islets from two pancreases on average) to achieve insulin independence [8], though islet autotransplantation indicates a minimum of 300,000 IE are necessary to achieve insulin independence in 70% of patients [9]. In addition, insulin secretion was less than expected based on the number of islets transplanted in patients who have undergone the Edmonton protocol. Assuming a non-diabetic pancreas has 1 million islets [10; 11], more than 80% of this mass was transplanted, yet only 19% of the normal acute insulin response to glucose was observed [12]. When stimulated with 50% dextrose intravenously, the area under the insulin versus time curve adjusted for baseline insulin secretion, an indication of islet function, was 36% of control subjects [12], suggesting that most of the islets transplanted were not functioning properly or not viable. Low viability can result from
stresses imposed on the islets prior to and after transplantation. Measures to maintain or increase the amount of viable and functioning tissue transplanted are needed in order to reduce the amount of tissue required to treat diabetic patients.
Several factors affect the quality of the islet tissue starting with the retrieval of the pancreas from the donor. The quality of the original islets in the donor may itself be an important factor. Warm and cold ischemia time of the pancreas prior to islet isolation may result in tissue exposure to hypoxia. Mechanical and enzymatic disruption may decrease islet quality during islet isolation [13-15]. Hypoxia also occurs during most conventional islet culture conditions and may be severe during islet shipment. After transplantation, islets are exposed to more risks, including exposure to hypoxia prior to vascularization and exposure to stress from the immune system. In all cases the quality of the initial islet tissue in the donor pancreas only decreases with time [13]. At this time islets cannot be expanded significantly in culture; as a result, all viable islet tissue is of great importance. Any advances in reducing cell death in islet tissue before transplantation could increase the number of patients who can be treated with available tissue and/or improve the success rate of current islet transplantation procedures.
This thesis focuses on improving islet culture techniques in order to maintain islet quality during culture and shipment prior to transplantation.
1.2 Objectives
The central hypothesis of this project follows: During conventional culture and shipment islet tissue is exposed to oxygen (O\textsubscript{2}) transport limitations. Though high density culture is desirable, it will exacerbate O\textsubscript{2} limitations and cause a larger decrease in the amount of viable tissue. When these limitations are removed, the tissue quality after culture and/or shipment will be higher than those using conventional methods. This project investigated this hypothesis through the following objectives:
- Examine islet tissue recovery, purity, and viability after culture in the absence of O\textsubscript{2} limitations.
- Investigate the effect of tissue surface density on the recovery of islet tissue, its purity, and its viability when cultured at 37°C on a polystyrene dish.
- Compare theoretical predictions for viable tissue recovery based on an O\textsubscript{2} transport to experimental results.
• Investigate high density culture techniques that may reduce $O_2$ limitations by culturing islets at reduced temperature, in elevated $O_2$, or on an $O_2$-permeable silicone rubber membrane.
• Through theoretical examination, determine what limitations become important when $O_2$ limitations are removed specifically considering glucose transport and pH changes.
• By applying lessons learned from the above objectives, develop a transportation device to maintain constant culture conditions during shipment and remove existing $O_2$ limitations in the current shipment method.
1.3 Overview
Below we briefly note the contents of each subsequent chapter in this thesis.
Chapter 2, written in collaboration with Anna Pisania, covers islet (human, rat, and porcine) and non-islet (human) characterization prior to and following low density culture on polystyrene using techniques developed by our laboratory including islet volume fraction by light microscopy, total amount of tissue by nuclei and DNA counts, viability using $O_2$ consumption rate (OCR) measurements in a stirred-tank vessel, and membrane integrity by differential staining of nucleic acids before and after cell disruption. Culture at two different temperatures was examined and both adherent and non-adherent tissue was analyzed. This study demonstrated how islet quality changes during culture in the absence of nutrient transport limitations.
Chapter 3 discusses the effect of tissue surface density on the recovery of islet tissue, its viability, and its purity following culture on polystyrene. Using a theoretical model previously developed for $O_2$ transport during static islet culture [16] we investigated whether $O_2$ limitations likely cause the reduction in the amount of viable tissue observed during high density islet culture.
Chapter 4 discusses improvements to conventional islet culture techniques, specifically addressing the presence of $O_2$ limitations. Elevated ambient $O_2$, reduced culture temperature, and culture on an $O_2$-permeable membrane are evaluated as alternative culture techniques with analysis focusing on the recovery of islet tissue and its viability for a range of tissue surface
densities. The theoretical model applied in Chapter 3 was modified to investigate how $O_2$ transport changes for each new culture technique.
Chapter 5 explores in detail the theoretical $O_2$ transport models used in chapters 3 and 4. The sensitivity of model predictions to variations in model parameters and the simplifying assumptions made within the model including islet distribution, islet size, slab formation, tissue stacking, and convection are discussed.
Chapter 6 introduces and investigates a transient glucose transport model that predicts concentration profiles during static islet culture to determine if nutrient transport limitations other than $O_2$ could become limiting. In addition, preliminary experiments are used to elucidate whether islets are exposed to low glucose levels during culture and if so, what affect that has on the recovery of viable tissue.
Chapter 7 lays the groundwork for a theoretical model that predicts transient changes in local pH during static islet culture on either a carbon dioxide (CO$_2$)-impermeable or CO$_2$-permeable surface. The model accounts for CO$_2$, ammonia, and lactate production by the islets, and includes the interaction of H$^+$ with the bicarbonate buffer. Exploration of the model reveals whether pH could change significantly in the islet tissue from the physiological value of 7.4.
In Chapter 8 the knowledge gained in chapters 3 through 5 was used in collaboration with Giner, Inc., and the University of Minnesota to develop an islet transportation device that actively maintained temperature and $O_2$ level during shipment, which was tested and compared to the standard method of shipping tissue in a Styrofoam box. After demonstrating the usefulness of this actively controlled system, an inexpensive alternative shipping device is proposed and its feasibility determined through preliminary experiments and theoretical calculations.
1.4 Nomenclature
| Symbol | Description |
|--------|---------------------------|
| CO$_2$ | Carbon dioxide |
| IE | Islet equivalent |
| O2 | Oxygen |
| OCR | Oxygen consumption rate |
2 Characterization of islet and non-islet tissue prior to and following culture
This chapter was written in collaboration with Anna Pisania, who developed and tested many of the tools described here to characterize islet tissue. My contribution involved understanding the changes islet tissue undergoes during low density static culture when no transport limitations are present.
2.1 Introduction
Transplantation of islets of Langerhans is a promising treatment for patients with unstable forms of type 1 diabetes [17-22], but several aspects need improvement to allow widespread application. One need is quantitatively meaningful islet quality assessment tests that can predict the transplantation outcome [21; 23-25]. This is especially important because islet isolation and transplantation success rates vary between centers [26; 27]. The most reliable assay, transplantation under the kidney capsule of nude diabetic mice [28-30], requires a long time and is useful only retrospectively. Currently-employed methods for islet characterization include staining with dithizone (DTZ) for purity assessment and enumeration by visual evaluation under a microscope [31; 32], and staining with fluorescent diacetate/propidium iodide (FDA/PI) for assessment of membrane integrity [33]. These methods are not quantitative and do not correlate with transplantation outcome [34-36].
Recent interest in developing new methods for islet viability assessment has focused on measurements of mitochondrial function, including mitochondrial membrane polarization with dispersed islet cells [37], ATP/ADP concentration ratio with intact islets [38] and oxygen consumption rate (OCR) [39]. The combined use of OCR and DNA measurements in the latter two studies correlated with transplantation outcome with high specificity and selectivity. However, there is a paucity of data in the literature on the OCR of human islet preparations or of any other species.
Islet culture for 36-72 hr prior to transplantation is practiced in most clinical centers [40] to provide time for tissue recovery from the harsh isolation process, possible reduction in immunogenicity [41-46], assessment of purity, membrane integrity, and sterility, recipient matching, patient travel to the islet center, and patient preparation to reach necessary levels of
immunosuppression before transplantation [40; 47; 48]. There is need to better understand islet culture and its effect on purity and viability. Culture conditions, especially temperature, vary between islet centers. Culture at 37°C maintains viability (by membrane integrity) and function better than reduced temperature [47; 49], but also results in increased necrotic death [49-52], which is likely the result of oxygen transport limitations. Culture at 22-24°C results in more islet mantle disintegration and increased apoptosis [53], but also increases the recovery of islet tissue [46; 47; 49; 54]. Currently, only non-adherent tissue is recovered from culture for analysis and subsequent use.
In this study, we examined OCR properties [39] of large numbers of human islet preparations (freshly isolated, precultured, or shipped from various centers), as well as rat and porcine islets. Measurements made upon receipt of islets and after 1 or 2 days of culture included, in addition, total amount of tissue by nuclei counts and DNA content [55], islet volume fraction [55] and β cell number fraction by light microscopy, and quantitative membrane integrity [55]. We also investigated culture temperatures of 37°C and 24°C and examined both adherent and non-adherent tissue recovered from culture.
2.2 Methods
Islet isolation and culture
Rat islets were isolated from male Sprague-Dawley rats by using collagenase digestion/ficoll purification [56-58]. Standard collagenase/protease digestion methods were used for human [7; 15] and porcine [59] isolations. Preparations of rat, and human islets, human non-islet tissue, and porcine islets were provided fresh (4-6 hr after isolation was complete) by the Islet Core at the Joslin Diabetes Center (JDC, Boston, MA). Human islets were also received after culture (1-2 days) from JDC or overnight shipment (sometimes after up to 1 day culture) from other centers contained in culture flasks (T-25 and T-75) or centrifuge tubes in Styrofoam boxes. When first observed, some preparations were in suspension, some in a pellet. Upon receipt, a portion of the preparation was analyzed and the remainder placed into untreated polystyrene culture flasks at 37°C in a humidified environment (5% CO₂), oxygen partial pressure (pO₂) of 142 mmHg, for 34-60 hr during which no medium change occurred. Some cultures were terminated at 18-24 hr for analysis. To test the effect of temperature, some islets were cultured at 24°C and then warmed for 1 hr at 37°C before OCR measurements.
Islets were cultured in supplemented RPMI (11.1 mM D-glucose, 100 U/ml penicillin, 100 µg/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate (all from Mediatech Inc., Herdon, VA), and 50 µM β-mercaptoethanol (Sigma Aldrich, St. Louis, MO) with 10% fetal bovine serum (FBS, Mediatech Inc.), at low densities (< 8x10^4 cells/cm^2 or about 50 islet equivalents (IE)/cm^2, where IE is the volume of a sphere with a diameter of 150 µm) in 1.3 mm of medium in tissue culture flasks. After culture the free, non-adherent tissue (which is the only tissue used in current practice) was collected. In selected experiments, adherent human tissue was incubated for 5 min at 37°C in 0.05% trypsin EDTA (Mediatech Inc.), dislodged with fresh medium, and collected. Rat islets and human tissue at 24°C did not adhere significantly. Tissue suspensions were centrifuged for 3 min at 300xg, supernatant removed and the tissue suspended in fresh medium.
Human and porcine islets were also isolated and analyzed at the University of Minnesota (MN) using standard methods [39; 60]. Islets were cultured in CMRL 1066 formulated for islet culture (Mediatech, 99-603), further supplemented with 10% FBS and heparin (10 U/ml). Islets were cultured at 200 IE/cm^2 in 30 ml of medium in tissue culture flasks in an incubator with 5% CO_2 for 18 to 24 hr at 37°C and then 18 to 24 hr at 24°C.
**DNA content**
At MIT, DNA was measured by fluorospectrophotometry [55] using the CyQUANT Cell Proliferation Assay Kit (Molecular Probes, C-7026 Eugene, OR), and a λ DNA standard (Invitrogen) and 50-250 IE/sample. Fluorescence was read at 480 nm excitation and 520 nm emission wavelengths in a plate reader (Spectra MAX Gemini microplate spectrophotometer, Molecular Devices, Sunnyvale CA). At MN (porcine and human islets), PicoGreen dye was used (Quant-iT PicoGreen dsDNA kit, Molecular Probes, Eugene, OR) [61]. Fluorescence was read at 520 nm excitation and 585 nm emission wavelengths.
**Islet enumeration by nuclei counting**
Nuclei were prepared by adding equal 100-µl volumes of sample containing 160 or more IE and of lysis solution (0.1 M citric acid (Sigma) and 1% (v/v) Triton X-100 (Sigma)) to a 1.5 ml microtube [55]. The mixture was incubated at room temperature for 5 min with vortex mixing every 1.5 min, and nuclei were liberated by shearing through a needle. Isolated nuclei were diluted with Dulbecco’s phosphate buffered saline (D-PBS, Invitrogen, Carlsbad, CA) to a concentration no higher than 5 x 10^5 nuclei/ml, stained with 7-aminoactinomycin D (7-AAD,
Molecular Probes, Eugene, OR), and analyzed using a flow cytometer (Guava Personal Cell Analysis (PCA) system, Guava Technologies, Hayward, CA) to determine the total number of cells with intact nuclei in the sample.
**Oxygen consumption rate (OCR)**
OCR was measured as previously described [39]. Briefly, suspensions containing about 2500 IE/ml in DMEM containing 4.5 g/l glucose and 0.6 g/l L-glutamine supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 10 mM HEPES, and no added serum were sealed in a 200-μl stirred titanium chamber (Micro Oxygen Uptake System, FO/SYSZ- P250, Instech Laboratories, Plymouth Meeting, PA) maintained at 37°C. The time dependent pO$_2$ within the chamber was recorded with a fluorescence-based oxygen sensor (Ocean Optics, Dunedin, FL), and data at high pO$_2$ were fit to a straight line. The maximal OCR was evaluated from $OCR = V_{ch} \alpha (\Delta pO_2 / \Delta t)$, where $V_{ch}$ is the chamber volume and $\alpha = 1.27$ nmol/mm-Hg·ml is the Bunsen solubility coefficient for oxygen in medium [16]. OCR measurements were normalized by the measured number of cells (nuclei counting) or DNA content of the sample examined.
**Membrane Integrity**
**7-Aminoactinomycin (7-AAD) sequential staining**
Cell membrane integrity was assessed by differential staining with 7-AAD (Molecular Probes, Eugene, OR) [55]. An aliquot of about 300 islets was re-suspended in 100 μl of D-PBS, and 5 μl of 1 mg/ml 7-AAD and incubated for 20 min at 4°C protected from light. After two washes with 1 ml of D-PBS, cells were disrupted by adding equal volume of lysis solution in D-PBS to the islet suspension and sheared as described for islet enumeration. Labeled nuclei were counted immediately in the flow cytometer or stored on ice for less than 15 min before counting. A portion of the islet suspension was further stained with 7-AAD, thereby labeling all of the previously unlabelled nuclei, and the total number of nuclei was counted. The fraction of cells with compromised membranes was estimated as the ratio of the initially stained nuclei (first measurement) to the total number of nuclei (second measurement).
**Fluorescein diacetate (FDA)/propidium iodide (PI)**
Immediately after human islet isolation was completed, a 250 μl aliquot was taken from the preparation, re-suspended in 5 ml of PBS solution in a 60 mm Petri dish, and 10 μl of 24 mM fluorescein diacetate (FDA) and 750 μM propidium iodide (PI) was added [62]. The tissue was
examined in a light microscope by focusing through the tissue to visually estimate the volume fraction of cells containing nuclei stained red because of loss of membrane integrity.
**Islet volume fraction by morphological analysis**
A 0.5 ml aliquot of human islet preparation was processed to produce 1-\(\mu\)m sections stained with toluidine blue and examined by light microscopy [13]. Islet tissue was distinguished from non-islet tissue as previously described [13]. Sections were analyzed at 420X by stereological point counting [63] with a 90-point grid covering adjacent, non-overlapping fields. At point intercepts of the grid with tissue, the nature of the tissue was determined, and the volume fraction of islet tissue, denoted \(\Phi_i\), was calculated as previously described [13].
**\(\beta\) cell identification by immunocytochemical insulin staining**
Tissue samples were washed twice with 1 ml D-PBS, fixed for 1 hr in 10% formalin (Sigma Aldrich), washed twice with 1 ml of D-PBS, and centrifuged at 300xg for 5 min. To the pellet, 500 \(\mu\)l of warm 2% (w/v) agarose was added. The solution was mixed using a vortex mixer for 1 sec, immediately centrifuged at 18000xg for 10 sec, allowed to cool for 2 min, and then 500 \(\mu\)l of 10% formalin was added. After 1-hr incubation at room temperature, formalin was removed, samples were stored in D-PBS until embedded in paraffin, and 5-\(\mu\)m sections were cut. Sections were stained by the unlabeled antibody peroxidase-antiperoxidase (PAP) method [64] using guinea pig anti-bovine insulin (Linco, St. Charles, MO) as primary antibody and goat anti-guinea pig IgG (Cappel, Irvine, CA), as secondary antibody, to link primary antibody to rabbit PAP complex (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). Use of a heterologous antibody gave specific staining with little background [65]. After applying PAP, slides were stained in 50 ml of a 2 mM 3,3'-diaminobenzidine tetrahydrochloride (DAB, Sigma Aldrich) solution activated with 25 \(\mu\)l of 30% (v/v) hydrogen peroxide (EMD Chemicals Inc, Gibbstown, NJ) for 30-120 sec, then viewed under a microscope. Reaction between hydrogen peroxide, DAB, and PAP formed a dark brown insoluble precipitate [66]. Slides were counterstained with hematoxylin (Sigma Aldrich). \(\beta\) cell volume fraction exclusive of vascular space, denoted \(\Phi_{\beta XV}\), was quantified using stereological point counting, as described for morphological analysis, with 200 points or more counted.
This procedure was used with samples after 2 days culture because trypsinization of adherent tissue degraded the structure and produced single cells, which, in turn, made it harder to
morphologically differentiate between cell types and evaluate vascular volume fraction, whereas it was more accurate to distinguish insulin staining. Methods for calculating $\Phi_{\beta XV}$ (day 2) and comparing it with $\Phi_1$ (day 0) are described in the Appendix.
**Statistics**
Measurements were made with three or more replicates and reported as mean ± SD. Statistical significance was determined using a Student t-test for unpaired data for comparing population means, and for paired data when appropriate.
### 2.3 Results
Nuclei counts, DNA content, membrane integrity by 7-AAD, and OCR were measured with fresh rat, fresh human, and cultured/shipped (data combined) human islets, and fresh human non-islet tissue at MIT; islet and $\beta$ cell volume fraction were measured with human fresh and cultured/shipped islets. DNA content and OCR were measured with porcine islets at MIT and UMN and with human islets at UMN. Measurements with fresh tissue were performed 4-6 hr after isolation. At MIT, all islet preparations were analyzed on day 0 (day of isolation for all fresh tissue or the day of receipt for cultured/shipped human islets), and a subset on day 1 and 2 of culture. Tissue analyzed in this study was often used for other purposes, and there was usually insufficient tissue for all tests, so that population means did not come from identical preparations. In a series of paired studies, measurements on day 0 and day 2 were made with non-adherent, adherent, and combined (total) tissue samples, enabling measurement of the recovery of the total amount of tissue (DNA and nuclei), the total OCR, and the number of $\beta$ cells relative to that originally placed in culture. In experiments in which there was insufficient adherent tissue to carry out individual assays, adherent was nonetheless combined with non-adherent tissue for assay of the total.
#### 2.3.1 DNA/cell
Frequency distributions of DNA/cell for day 0 and for non-adherent tissue on day 1 and 2 in culture are shown in Figure 2.1. DNA/content (pg DNA/cell) of rat (6.4 ± 1.6), fresh human islets (6.4 ± 1.1), and human non-islet tissue (5.7 ± 2.3) was similar while that of human cultured/shipped islets (7.9 ± 1.8) was significantly higher ($p<0.01$), thereby suggesting that significant cell death occurred during pre-culture and shipment and that DNA had been released.
from dead cells but had not yet degraded. The DNA content of cultured rat islets did not change with time, whereas on day 1 it increased substantially with non-adherent fresh islets (p<0.01) and non-islet tissue (p<0.20) and decreased with cultured/shipped islets; on day 2, all returned partially towards the day 0 value. Rat islets produced a normal frequency distribution with a coefficient of variation (COV) of 0.25. The COV was 0.14 for fresh human islets and usually about 0.25 for other conditions, all of which were skewed to the right, reflecting the presence of very high DNA/cell in some cultured preparations.
Additional data for the effect of 2-day culture on DNA/cell is tabulated in Table 2.1. At 37°C, day 2, DNA/cell for rat tissue was significantly lower than that of human non-adherent cultured/shipped (p<0.02) tissue but not for fresh human tissue. Changes from day 0 to day 2 were not significant for any of the non-adherent islets but was significant for non-adherent human non-islet tissue. At 24°C, DNA/cell of the non-adherent rat islets was substantially higher than at 37°C. The same trend with cultured/shipped human islets was not significant for paired data. The opposite trend was observed for non-adherent fresh human-islets, though only three paired measurements were made, and the mean at 37°C (10.6 ± 1.9) was substantially higher than that for a larger number of preparations (7.2 ± 2.8).
2.3.2 Nuclei and DNA recovery
The fraction of tissue recovered on day 2, as nuclei or DNA, relative to day 0 is summarized in Table 2.1. At 37°C, the fraction recovered for all tissue groups and for the combined subsets within each group, was substantially less than 1.0 based on either nuclei counts (p<0.01) or DNA measurements (p<0.01 to 0.15). The fractional recovery of nuclei was 0.63 ± 0.14, 0.34 ± 0.15, 0.50 ± 0.17, and 0.40 ± 0.24 respectively, for rat, human fresh, and cultured shipped islets, and non-islet tissue, 0.22 ± 0.10 and 0.30 ± 0.17 for human fresh and cultured/shipped adherent tissue, respectively, and 0.63 ± 0.14, 0.63 ± 0.17, and 0.81 ± 0.20 for rat and human fresh and cultured/shipped total tissue, respectively. When based on DNA, these fractions were slightly lower for rat and cultured/shipped human islets and slightly higher for human fresh islets and non-islet tissue.
Figure 2.1. Frequency distributions of DNA content for day 0 and for non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, and (C) human cultured/shipped islets, and (D) human non-islet tissue.
n is the total number of islet preparations for which measurements were made.
Table 2.1. DNA/cell and the fraction of original tissue recovered (quantified by nuclei and DNA) after 2 days culture.\textsuperscript{a}
| Temp | Day (°C) | Type | n | DNA/cell (pg DNA/cell) | Fraction of original nuclei | Fraction of original DNA |
|------|----------|------|-----|------------------------|----------------------------|--------------------------|
| | | | | | n | n |
| Rat | Fresh islets | | | | 1.0 | 1.0 |
| | 0 | - | - | 30 | 6.4 ± 1.6 | |
| | 2 | 37 | NA | 7 | 5.7 ± 1.5 | 0.63 ± 0.14 |
| | | | | 5 | 6.0 ± 1.6 | 0.74 ± 0.16 |
| | 2 | 24 | NA | 5 | 7.3 ± 1.7 | |
| | | | | | 1.0 | 1.0 |
| | | | | | 5 | 0.59 ± 0.23 |
| | | | | | 5 | 0.72 ± 0.21 |
| Human | Fresh islets | | | | 1.0 | 1.0 |
| | 0 | - | - | 20 | 6.4 ± 1.1 | |
| | 2 | 37 | NA | 12 | 7.2 ± 2.8 | 0.34 ± 0.15 |
| | | | | 3 | 10.6 ± 1.9 | 0.41 ± 0.20 |
| | 2 | 37 | A | 2 | 6.6 | 0.32 ± 0.17 |
| | 2 | 37 | T | 11 | 8.5 ± 2.8 | 0.22 ± 0.10 |
| | 2 | 24 | NA | 5 | 7.4 ± 0.2 | 0.83 ± 0.17 |
| | | | | 3 | 7.0 ± 0.7 | 0.83 ± 0.04 |
| | | | | | 3 | 0.81 ± 0.01 |
| Cultured/shipped islets | 0 | - | - | 24 | 7.9 ± 1.8 | 1.0 | 1.0 |
| | 2 | 37 | NA | 17 | 7.7 ± 1.8 | 0.50 ± 0.17 |
| | | | | 8 | 5.4 ± 0.20 | 0.54 ± 0.15 |
| | 2 | 37 | A | 3 | 7.2 ± 1.9 | 0.30 ± 0.17 |
| | 2 | 37 | T | 5 | 7.3 ± 1.3 | 0.81 ± 0.20 |
| | 2 | 24 | NA | 8 | 8.8 ± 2.6 | 0.77 ± 0.19 |
| Fresh non-islet | 0 | - | - | 6 | 5.7 ± 2.3 | 1.0 | 1.0 |
| | 2 | 37 | NA | 4 | 4.9 ± 1.3 | 0.40 ± 0.24 |
\textsuperscript{a}Data are tabulated for day 0 and day 2 non-adherent (NA), adherent (A), and the combined total (T, adherent + non-adherent) tissue within each group and includes overall population means for comparison between groups (e.g. fresh rat versus fresh human) and paired data for comparisons of subsets within each group. Data are reported as mean ± SD for measurements with n different islet preparations. The p values of paired (bold face) and unpaired (not bold face) comparisons (denoted by brackets) are given for p ≤ 0.15 according to the following notation: A p ≤ 0.01, B 0.01 < p ≤ 0.10, C 0.10 < p ≤ 0.15. Fractional recoveries for NA and A add up to T only for paired samples.
Fractional recovery of nuclei was higher for rat than non-adherent human fresh ($p<0.01$) and cultured/shipped ($p<0.10$) islets, but total nuclei recovery was not significantly different. The same trends were true with DNA recovery but with higher $p$ values. Human fresh islets had lower recovery than cultured/shipped islets for non-adherent ($p<0.09$) and total ($p<0.02$) tissue. Recoveries of adherent tissue were similar. Significant differences in recoveries of human fresh and cultured/shipped islets were not observed in terms of fraction of original DNA. At 24°C, fractional cell recoveries for rat and non-adherent human fresh and cultured/shipped islets were $0.74 \pm 0.16$, $0.83 \pm 0.04$, and $0.77 \pm 0.19$, respectively, all significantly less than 1.0 but higher than at 37°C. Similar data were obtained with fractional DNA recovery.
### 2.3.3 OCR/cell and OCR/DNA
Frequency distributions of OCR/cell and OCR/DNA for day 0 and for non-adherent tissue on day 1 and 2 in culture are shown in Figure 2.2 and Figure 2.3, respectively. OCR/cell for rat islets, human fresh and cultured/shipped islets, and human non-islet tissue was $2.7 \pm 1.1$, $1.5 \pm 0.5$, $1.6 \pm 1.1$, and $1.6 \pm 0.7$ fmol/min cell on day 0, respectively, and $2.6 \pm 0.4$, $2.6 \pm 1.1$, $2.3 \pm 0.7$, and $2.2 \pm 0.7$ fmol/min cell on day 2. On day 0, none of the distributions were normal. The COV was 0.41 for rat, 0.33 for human fresh, and 0.70 for cultured/shipped islets, and 0.44 for non-islet tissue, all larger than the comparable value of DNA/cell. Furthermore, the ratio of maximum to minimum measured value ranged from about 2.0 to 3.1 for DNA/cell and about 3.5 to 6.4 for OCR/cell. Because the precision in measuring DNA content and OCR was comparable, the wider range of the OCR/cell distributions indicates differences in the viability characteristics of the different preparations. For the frequency distributions of OCR/DNA (Figure 3), trends of the population means between species and days in culture were similar to those shown by the OCR/cell distributions, except that there were larger differences between some mean values and smaller differences between others because of the differences in the DNA/cell between the various types of tissue preparations. Notably, the increase in OCR/DNA from day 0 to day 2 for human fresh non-adherent and total tissue was not significant, whereas it was significant for OCR/cell measurements.
Figure 2.2. Frequency distributions of the OCR/cell for day 0 and for non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, (C) human cultured/shipped islets, and (D) human non-islet tissue.
n is the total number of islet preparations for which measurements were made.
Figure 2.3. Frequency distributions of the OCR/DNA for day 0 and for non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, (C) human cultured/shipped islets, and (D) human non-islet tissue.
n is the total number of islet preparations for which measurements were made.
The effect of 2-day culture on OCR/cell and OCR/DNA is tabulated in Table 2.2. The increased OCR/cell with time observed in Figure 2 for non-adherent human tissue was significant in paired t-tests. OCR/cell for all adherent human islets tissue was lower than that of the corresponding non-adherent tissue and was not significantly different than day 0 values. OCR/DNA followed similar trends, but significance in differences did not follow a precise correspondence because of variation in DNA/cell and occasional large COVs. After 2 day 24°C culture, OCR/cell and OCR/DNA of rat non-adherent and human islets were substantially lower.
than after 37°C culture. In addition, after 24°C culture OCR/cell of non-adherent tissue was lower for fresh human islets than that of rat and human cultured/shipped islets.
2.3.4 OCR recovery
The fractional recovery of OCR from 2-day, 37°C culture was significantly less than 1.0 for almost all categories (Table 2.2). Fractional recovery of total tissue was 0.65 ± 0.20, 0.75 ± 0.23, and 0.98 ± 0.07 for rat and human fresh and cultured/shipped islets, respectively; that for non-adherent tissue was 0.63 ± 0.06, 0.70 ± 0.27, and 0.46 ± 0.27 for human fresh and cultured/shipped islets and non-islet tissue. Fractional recovery of the adherent component was smallest, about 0.33 for both groups of human islets, but the number of measurements was small, and the data for fresh islets may not be representative. At 24°C, fractional OCR recoveries for non-adherent rat and human fresh islets were 0.63 ± 0.20, and 0.65 ± 0.14, respectively, which are lower than that for non-adherent cultured/shipped islets, 0.83 ± 0.28. All are significantly less than 1.0 but not significantly different from their values at 37°C.
2.3.5 OCR/DNA comparisons between centers
Frequency distributions of the OCR/DNA for islets isolated and analyzed at different centers are shown in Figure 2.4. Fresh human islets at UMN had significantly lower OCR/DNA values (130 ± 30 nmol/min mg DNA) than those isolated at JDC and analyzed at MIT on day 0 (p<0.01). These values did not increase significantly with time in culture, and the difference between dates obtained at the two locations was even greater at day 2.
2.3.6 Membrane integrity
Frequency distributions of the fraction of cells with impermeable cell membranes by 7-AAD for day 0 and for non-adherent tissue on day 1 and 2 in culture are shown in Figure 2.5. On day 0, the fraction of cells with impermeable cell membranes averaged 0.85-0.87 for all human tissue and 0.77 for rat islets, and only rat islets had a substantial fraction of preparations below 0.8. The fraction increased with time in culture for rat islets, and did not change for human islets. It also did not change for non-islet tissue (paired t test with p = 0.2, and 0.3 for day 1 and 2 respectively), even though the trends of the means in Figure 2.5 are misleading.
Table 2.2. OCR/cell, OCR/DNA, and fraction of original OCR recovered after 2 days culture.\(^a\)
| Temp Day (°C) | Type | OCR/cell n (fmol/min cell) | OCR/DNA n (nmol/min mg DNA) | Fraction of original OCR n |
|---------------|------|----------------------------|-----------------------------|---------------------------|
| Rat | | | | |
| Fresh islets | | | | |
| 0 | - | 37 2.7 ± 1.1 | 30 410 ± 130 | 1.0 |
| | | 7 2.8 ± 0.7 | 4 360 ± 110 | |
| | | | 3 340 ± 60 | |
| 2 | 37 | NA 9 2.6 ± 0.2 | 5 420 ± 130 | 8 0.65 ± 0.20 |
| | | 8 2.8 ± 0.5 | 4 450 ± 130 | |
| | | | 4 430 ± 120 | |
| | | | 4 270 ± 40 | 8 0.63 ± 0.20 |
| | | | 3 260 ± 40 | |
| 2 | 24 | NA 8 2.2 ± 0.5 | A | |
| | | 7 2.1 ± 0.4 | A | |
| Human | | | | |
| Fresh islets | | | | |
| 0 | - | 21 1.5 ± 0.4 | 23 230 ± 60 | 1.0 |
| | | 8 1.5 ± 0.5 | B | |
| | | 9 1.4 ± 0.7 | B | |
| 2 | 37 | NA 12 2.6 ± 1.1 | A | |
| | | 8 2.9 ± 1.0 | A | |
| | | 3 2.6 ± 0.7 | B | |
| 2 | 37 | A 2 1.8 | B | |
| 2 | 37 | T 11 2.2 ± 0.5 | 9 280 ± 110 | 2 0.31 |
| | | 9 2.1 ± 0.5 | 3 420 ± 120 | |
| | | 3 2.7 ± 0.4 | C | |
| 2 | 24 | NA 5 1.5 ± 0.4 | B | |
| | | 3 1.4 ± 0.4 | B | |
| | | 3 1.6 ± 0.5 | B | |
| Cultured/shipped islets | | | | |
| 0 | - | 26 1.6 ± 0.6 | 24 230 ± 120 | 1.0 |
| | | 18 1.6 ± 0.6 | A | |
| 2 | 37 | NA 19 2.3 ± 0.7 | C | |
| | | 18 2.3 ± 0.7 | C | |
| | | 3 3.0 ± 0.8 | C | |
| | | 8 2.6 ± 0.8 | C | |
| 2 | 37 | A 3 1.7 ± 0.4 | C | |
| 2 | 37 | T 6 2.2 ± 0.4 | B | |
| | | 3 2.4 ± 0.5 | B | |
| 2 | 24 | NA 8 1.9 ± 0.5 | B | |
| | | 3 1.7 ± 0.3 | B | |
| Fresh non-islet | | | | |
| 0 | - | 6 1.6 ± 0.7 | 6 290 ± 160 | 1.0 |
| | | 4 1.4 ± 0.6 | B | |
| 2 | 37 | NA 4 2.2 ± 0.8 | 4 340 ± 160 | 5 0.48 ± 0.27 |
\(^a\)Format is identical to that in Table 1.
Figure 2.4. Frequency distributions of the OCR/DNA for different centers (MIT, MN) for (A) human, and (B) porcine islets. Islets were analyzed right after isolation and after 2-day culture at the respective centers. The data for porcine from MN on day 0 and 2 are combined. n is the total number of islet preparations for which measurements were made.
2.3.7 OCR/cell with impermeable cell membranes by 7-AAD
The value of OCR/cell of each preparation was divided by the fraction of cells with impermeable cell membranes by 7-AAD for that preparation. The frequency distributions of this quantity, denoted as OCR/(cell)$_{MI}$, for day 0 and for non-adherent tissue on day 1 and 2 in culture are shown in Figure 2.6. Conversion of OCR/cell data to OCR/(cell)$_{MI}$ had only a small and inconsistent effect on COV for the distribution, and introduced divergences into the value after 2 days in culture.
To further explore the relationships between OCR and membrane integrity measurements, OCR/DNA measurements are plotted versus the fraction of cells with impermeable cell membranes by FDA/PI staining and 7-AAD in Figure 2.7. Whereas OCR/DNA measurements varied from about 150 to 700 nmol/min mg DNA for rat islets, 100 to 350 nmol/min mg DNA for human islets, and 100 to 500 nmol/min mg DNA for porcine islets, values of membrane integrity from FDA/PI staining and 7-AAD sequential staining were always in excess of 70%, except for a few lower values for rat islets.
2.3.8 Purity
The mean islet volume fractions ($\Phi_I$) were $0.64 \pm 0.12$ (range 0.43 to 0.79) and $0.67 \pm 0.09$ (range 0.58 to 0.82) for fresh and cultured/shipped human islet preparations on day 0, respectively (Table 2.3). These values fall within the range reported for typical clinical preparations. $\Phi_I$ for non-adherent tissue increase on day 2 compared to day 0; the effect was significant only for fresh tissue. The $\beta$ cell volume fraction exclusive of vascular space ($\Phi_{\beta XV}$) was higher for cultured/shipped islets ($0.41 \pm 0.03$) than for fresh islets ($0.33 \pm 0.03$) on day 0. On day 2, $\Phi_{\beta XV}$ increased for non-adherent tissue and decreased for adherent tissue. The differences were larger for fresh islets. The number fraction $\beta$ cells ($f_\beta$), which was calculated from $\Phi_I$ and $\Phi_{\beta XV}$ as outlined in the Appendix, follows a similar trend. Combining $f_\beta$ and nuclei recovery measurements (Appendix) revealed that the fraction of original $\beta$ cells recovered from culture ($F_\beta$) was significantly higher for non-adherent tissue than adherent tissue, but was less than 1.0 even when the fractions were combined. A loss of about 30% $\beta$ cells was observed for cultured/shipped tissue while a much larger loss, about 65% was observed for fresh tissue though only two data points were available.
2.4 Discussion
In this paper, we used previously-developed quantitative methods [39; 55] to measure quantity (nuclei counts, DNA content), purity (volume fraction islets and $\beta$ cells), viability (OCR), and membrane integrity for islet and non-islet preparations of different species (rat, human, and porcine) from different islet isolation centers before and after culture at different temperatures (24°C and 37°C). Measured values were expressed on a per cell basis when appropriate. We also estimated the recovery of total tissue, viable tissue, and $\beta$ cells after 2-day culture for non-adherent and adherent tissue. This is the first study to report this comprehensive set of quantitative parameters with a large number islet preparations before and after culture.
Figure 2.5. Frequency distributions of the fraction of cells with impermeable cell membranes by 7-AAD for day 0 and non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, (C) human cultured/shipped islets, and (D) human non-islet tissue.
n is the total number of islet preparations for which measurements were made.
Figure 2.6. Frequency distributions of the OCR/membrane-impermeable cell for day 0 and for non-adherent tissue on day 1 and 2 in culture for (A) rat, (B) human fresh, (C) human cultured/shipped islets, and (D) human non-islet tissue.
n is the total number of islet preparations for which measurements were made.
Figure 2.7. OCR/DNA versus fraction of cells with impermeable cell membranes by FDA/PI staining and 7-AAD for (A) rat islets, (B) human fresh islets.
On day 0, fresh islet tissue had DNA values close to 6.5 pg DNA/cell (Table 2.1 and Figure 2.1), which is the value expected for rat [67] and human [68] cells. The higher value for cultured/shipped human islets likely reflects residual, undegraded DNA. With time in culture, DNA/cell did not change for rat islets but increased for fresh human islet and non-islet tissue and decreased for cultured/shipped islets. This behavior is consistent with the notion that the cell death process had proceeded to a greater extent when the cultured/shipped samples were received and that subsequent behavior in culture reflected greater cell death in fresh human islets as a result of stresses imposed during the isolation process.
Losses of tissue during culture, as measured by nuclei counts, were substantial. After culture for 2 days, the non-adherent tissue collected accounted for about 2/3, 1/3, and 1/2 of the rat, human fresh, and human cultured/shipped islet preparation, respectively, placed into culture at day 0. Recovery of non-adherent non-islet tissue was about 40%. The non-adherent fraction of human islet tissue is the part currently used for both research purposes and clinical transplantation. When the adherent tissue (20-30% of the original tissue) is included, the fraction of the original tissue that was actually lost was about 1/3 for rat and fresh human islets and 1/5 for cultured/shipped islets. At 24°C, a substantially larger fraction of the original tissue was recovered as the non-adherent fraction especially for fresh human islets (0.83), consistent with previous reports for human [47] and porcine islets [49]. Thus, from the same original preparation, the amount of tissue recovered after 2 days at 24°C was more than double that at 37°C. Culture at 24°C did not result in significant adherence and thus could explain the higher tissue recovery. In addition, since the combined tissue recovery was also lower at 37°C than non-adherent tissue recovery at 24°C, it is likely that there are slower rates of DNA and nuclei degradation at reduced temperature.
Trends with DNA measurements were similar but not as sharp and often with lower p values. Nonetheless, measurements with both methods are desirable because DNA/cell estimates provide a measure of the extent of recent cell death.
Table 2.3. Human islet and β cell purity measurements and fraction of β cells recovered after culture for 2 days at 37°C.a
| Day | Type | n | Islet volume fraction | β cell volume fraction exclusive of vascular space | Number of β cells | Fraction of original β cells | Fraction of original nuclei |
|-----|------|-----|-----------------------|--------------------------------------------------|------------------|----------------------------|-----------------------------|
| | | | | | | | |
| Fresh | | | | | | | |
| 0 | - | 18 | 0.64 ± 0.12 | | | | |
| 4 | A | 5 | 0.67 ± 0.12 | | | | |
| 2 | NA | 4 | 0.50 ± 0.04 | | | | |
| 2 | T | 4 | 0.75 ± 0.07 | | | | |
| 2 | A | 2 | 0.10 ± 0.06 | | | | |
| 2 | T | 2 | 0.48 ± 0.01 | | | | |
| Cultured/shipped | | | | | | | |
| 0 | - | 9 | 0.67 ± 0.09 | | | | |
| 4 | A | 5 | 0.61 ± 0.04 | | | | |
| 2 | NA | 4 | 0.70 ± 0.15 | | | | |
| 2 | A | 5 | 0.17 ± 0.08 | | | | |
| 2 | T | 5 | 0.44 ± 0.09 | | | | |
| 2 | A | 5 | 0.14 ± 0.07 | | | | |
| 2 | T | 5 | 0.41 ± 0.09 | | | | |
| 2 | A | 5 | 0.09 ± 0.03 | | | | |
| 2 | T | 5 | 0.17 ± 0.08 | | | | |
| 2 | A | 5 | 0.69 ± 0.14 | | | | |
| 2 | T | 5 | 0.60 ± 0.14 | | | | |
| 2 | A | 5 | 0.09 ± 0.03 | | | | |
| 2 | T | 5 | 0.33 ± 0.22 | | | | |
| 2 | A | 5 | 0.92 ± 0.18 | | | | |
aFormat is identical to that in Table 1. Φ₁, ΦβXV, and the fraction of original nuclei were measured; fβ and Fβ were calculated as described in Appendix.
OCR/cell and OCR/DNA of human tissue (islet and non-islet) was lower than that of rat on day 0 likely because of the harsher isolation procedure. After 2-day culture at MIT in 37°C rat and human OCR/cell values of non-adherent islet and non-islet tissue were similar indicating that tissue respiration in stable, undamaged tissue is similar for different cell types and for rat and human tissue. However, porcine islets, both from MIT and MN, had a lower OCR/DNA than rat and human islets (MIT), suggesting there could be intrinsic differences in respiration between certain species. The OCR/cell values we obtained are within the range of values reported in the literature for intact islets of different species measured by different methods, 2-3.5 fmol/min cell (data converted to similar units for comparison purposes) [69-75]. One study reports lower values (0.3-0.5 fmol/min cell) but the measurements were made with a system of unreliable accuracy [13].
Human islets from MN had a lower OCR/DNA than those analyzed at MIT that could be explained by differences in the isolation and/or culture procedure for which temperature and density can have large effects. The effect of temperature is supported by the lower OCR/cell and OCR/DNA measurements obtained on day 2 at 24°C compared to 37°C for rat, human fresh, and cultured/shipped islets. Decreases in membrane integrity and cellular function (stimulation index) have been observed following reduced temperature culture [47; 49]. The lower OCR/cell following 24°C culture could also be an indication that the respiration of the tissue has not acclimated to 37°C or that the reduced culture temperature has permanently changed cellular machinery resulting in cells that require less oxygen as previously suggested [51] bringing up questions about changes in cellular function. Brandhorst et al concluded that 37°C culture of porcine islets resulted in superior insulin content, secretion, and dynamic response to glucose as compared to culture at 22°C [49], which is consistent with results for human islets [47] suggesting that 37°C culture maintains islet function better than reduced culture temperatures.
OCR recovery of non-adherent human non-islet tissue was lower than that of non-adherent fresh human islets. When accounting for both non-adherent and adherent tissue the total OCR recovery of cultured/shipped tissue was about 1 indicating no loss in viable tissue, while fresh tissue resulted in an average 25% decrease in viable tissue during culture, probably because tissue had not completed the cell death process initiated during isolation. OCR recovery for rat islets was less than 1 either because of loss of viable tissue during culture for similar reasons as
fresh human islet tissue or because adherent tissue was not analyzed, though based on visual observations very little tissue adhered.
The fraction of cells with impermeable cell membranes by 7-AAD for fresh rat islets was lower than that for human tissue. A possible explanation could be that human non-islet cells are less susceptible to damage on their membrane integrity than islet cells. The fraction of cells with intact membranes spans a narrower range than respective OCR/DNA measurements for rat, human, and porcine islets (Figure 2.7). This behavior is consistent with our findings that OCR, a mitochondrial function assay, is an earlier indicator of cell death than membrane integrity tests [55].
If the fraction of membrane-impermeable cells in islets 4 hr after isolation or receipt actually represented the fraction of viable cells, the OCR/viable cell or OCR/DNA for viable cells can be calculated. Because the spread of the distributions (Figure 2.6) was wider than those of the OCR/cell and the fraction of cells with impermeable cell membranes by 7-AAD, it is suggested that the normalization of OCR with 7-AAD does not result in a unique value for OCR of viable cell. If islet and non-islet tissue 4 hr after isolation are still recovering from damage incurred during the isolation process, then the fraction of cells that are viable (as assessed, for example, by a mitochondrial function assay such as OCR) is likely to be much smaller than that inferred from membrane integrity measurements, as has been observed [55], because permeabilization of the cell membrane occurs relatively late in cell death processes [76].
The islet volume fraction of the non-adherent fresh human islets increased in culture because non-islet tissue preferentially adhered to the flask, consistent with a lower β cell volume fraction exclusive of vascular space. In cultured/shipped tissue about 30% of the original β cells were lost in culture though they were likely dead prior to culture since the OCR recovery was almost 1. This could also be attributed to β cell dedifferentiation that would cause loss of insulin expression [77].
Our study provides the basis for a complete understanding of islet and non-islet tissue behavior in culture, at different temperatures, in terms of purity, quantity, viability, and β cell fraction. Other methods on islet characterization have intrinsic shortcomings that render them inappropriate and unreliable for islet assessment prior to transplantation. In order to obtain the β cell viability index that Ichii et al. [37] introduced, islets needed to be dissociated, which has been shown to result in samples not representative of the original islet [78]. The study of
ATP/ADP ratio by Goto et al. [38] is hindered by the extreme sensitivity of the assay to changes in environmental conditions and therefore it can be quite difficult to use the assay to get an accurate prediction, especially with islets that are large aggregates, and cells in the outer and inner regions may experience very different microenvironments depending on the assay conditions [79]. Finally, even though Guarino et al. [75; 80] used oxygen consumption rate measurements as an islet assessment method, the system they chose to use, BD OBS, is offset by difficulties in obtaining accurate and meaningful OCR estimates. Reported OCR values measured with the OBS do not agree with measurements made with other methods and therefore render the assay questionable.
2.5 Nomenclature
| Abbreviation | Description |
|--------------|-------------|
| 7-AAD | 7-aminoactinomycin D |
| A | Adherent tissue |
| $\alpha$ | Solubility of oxygen in medium |
| COV | Coefficient of variation |
| DAB | 2 mM 3,3’-diaminobenzidine tetrahydrochloride |
| D-PBS | Dulbecco’s phosphate buffered saline |
| DTZ | Dithizone |
| $f_{\beta}$ | Number fraction of $\beta$ cells among all cells |
| $F_{\beta}$ | Fraction of original $\beta$ cells collected from culture |
| FBS | Fetal bovine serum |
| FDA | Fluorescent diacetate |
| IE | Islet equivalent |
| JDC | Joslin Diabetes Center |
| MIT | Massachusetts Institute of Technology |
| MN | University of Minnesota |
| NA | Non-adherent tissue |
| OBS | Oxygen biosensor system |
| OCR | Oxygen consumption rate |
| OCR/(cell)$_{MI}$ | OCR per viable cell |
| Symbol | Description |
|--------|-------------|
| PAP | Peroxidase-antiperoxidase |
| PI | Propidium iodide |
| $pO_2$ | Oxygen partial pressure |
| t | Time |
| T | Total (non-adherent + adherent) tissue |
| $V_{ch}$ | Volume of OCR chamber |
| $\Phi_{\beta XV}$ | Volume fraction of $\beta$ cells exclusive of islet vascular spaces |
| $\Phi_I$ | Volume fraction of islets |
3 The effect of surface density on tissue quality during static islet culture on untreated polystyrene
3.1 Introduction
Following the initial report of success [7] islet transplantation has become a promising option for treating type 1 diabetes though many complications remain [8; 12; 21; 81-83]. After islet tissue is isolated from a donor pancreas by mechanical and enzymatic treatment [13-15], it may transplanted directly or cultured [14; 40]. Islet culture prior to transplantation is desirable for many reasons and is standard practice in most clinical centers [40]. Culture provides time for islet tissue to recover from the harsh isolation process while tissue quality is assessed for purity, viability, and sterility and for recipient matching, patient travel to the islet center, and patient preparation such that necessary levels of immunosuppression are attained before transplantation [40; 47; 48]. Additionally the immunogenicity of islet tissue may be reduced through culture [41-46], thus lowering the body’s immune response upon transplantation. Culture could also provide the time necessary to apply treatments currently under investigation (e.g. gene manipulation or chemical treatment such as carbon monoxide, Appendix 10.2) that may reduce cell death prior to and following transplantation [84-90].
A need to better understand islet culture exists. Culture conditions, including islet density, temperature, and duration vary between islet centers. There is little published information to determine the effect of these variables on the recovery of islet tissue, its viability, and its purity. A large amount of literature discusses media formulations and additives [40; 48], but little is available about the effect of islet density on islet quality [40]. Islet surface density and size are thought to be related to necrotic death [91] because experiments with low islet surface densities result in higher recoveries and higher stimulation indices compared to culture with high islet surface densities [47]. This decrease in islet tissue quality with increasing density may result from insufficient oxygen ($O_2$) transport to the core of the islet and is consistent with observations that indicate higher necrotic death in islets at $37^\circ C$ compared to lower temperatures where metabolic rates are reduced [49; 50; 52].
Oxygen transport is crucial for maintaining islet viability and function [92]. Because isolated islets lose their blood supply and blood vessels collapse during static culture [93-95], $O_2$ is supplied to the interior only by diffusion resulting in the formation of $O_2$ gradients in and
around each islet. If islets are too large or too close together, $O_2$ is depleted before it reaches the center of the islet, thereby causing the core to become severely hypoxic and leading to the development of dead tissue by necrosis and/or apoptosis [16; 49; 92; 96]. A theoretical model describing this phenomena with islets cultured on an $O_2$-impermeable surface has been developed [16].
Oxygen supply limitations will not cause deleterious effects at very low tissue surface densities [16]. However, the large surface area required to culture the islets obtained from a single isolation in the absence of $O_2$ limitations is impractical. As a result, the density of human islet tissue typically employed, 200-400 islet equivalents (IE)/cm$^2$ (K. Papas and A. Omer, personal communications), is thought to balance possible reduction in islet quality after culture against practical limitations, even though this density may be an order of magnitude higher than the surface density at which $O_2$ diffusion limitations begin [16]. Because little data is available for the effect of islet density on islet quality, the extent to which islet quality is affected by current practice is unknown.
In a previous study (Chapter 2), in which human islet tissue was cultured on untreated polystyrene vessels at low tissue surface density, we observed effects on recovery of islet tissue, its viability, and its purity. In this study, we explored whether culture at high density resulted in additional changes in these characteristics. In addition, we employed a theoretical $O_2$ transport model to investigate whether $O_2$ supply limitations can explain the reduction in the amount of viable tissue observed during high density islet culture. The results of this study demonstrate that tissue surface density has a significant impact on the recovery of islet tissue, its viability, and its purity. The decrease in the amount of viable tissue observed following high density culture is qualitatively similar to that predicted by an $O_2$ transport model, indicating that $O_2$ limitations are likely the primary cause of the decrease.
### 3.2 Methods
**Islet isolation and culture**
Standard collagenase/protease digestion methods were used for human [7; 15] isolations. Preparations human islets were provided fresh (4-6 hr after isolation was complete) by the Islet Core at the Joslin Diabetes Center (JDC, Boston, MA). Human islets were also received after culture (1-2 days) from JDC or overnight shipment (sometimes after up to 1 day culture) from
other centers contained in culture flasks (T-25 and T-75) or centrifuge tubes in Styrofoam boxes. When first observed, some preparations were in suspension, some in a pellet. Upon receipt, a portion of the preparation was analyzed and the remainder placed into untreated polystyrene culture flasks at 37°C in a humidified environment (5% CO₂), oxygen partial pressure (pO₂) of 142 mmHg, for 34-60 hr during which no medium change occurred.
Islets were cultured in supplemented RPMI (11.1 mM D-glucose, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate (all from Mediatech Inc., Herdon, VA), and 50 μM β-mercaptoethanol (Sigma Aldrich, St. Louis, MO) with 10% fetal bovine serum (FBS, Mediatech Inc.), in 3 mm of medium in untreated polystyrene tissue culture flasks. After culture the free, non-adherent tissue (which is the only tissue used in current practice) was collected. At densities <50 x 10⁴ cells/cm², adherent human tissue was incubated for 5 min at 37°C in 0.05% trypsin EDTA (Mediatech Inc.), dislodged with fresh medium, and collected. Tissue suspensions were centrifuged for 3 min at 300xg, supernatant removed and the tissue suspended in fresh medium. At high densities (>50 x 10⁴ cells/cm²), trypsinization was not required because adherent tissue was an insignificant fraction of the total.
**DNA content**
At MIT, DNA was measured by fluorospectrophotometry [55] using the CyQUANT Cell Proliferation Assay Kit (Molecular Probes, C-7026 Eugene, OR), and a λ DNA standard (Invitrogen) and 50-250 IE/sample. Fluorescence was read at 480 nm excitation and 520 nm emission wavelengths in a plate reader (Spectra MAX Gemini microplate spectrophotometer, Molecular Devices, Sunnyvale CA).
**Islet enumeration by nuclei counting**
Nuclei were prepared by adding equal 100-μl volumes of sample containing 160 or more IE and of lysis solution (0.1 M citric acid (Sigma) and 1% (v/v) Triton X-100 (Sigma)) to a 1.5 ml microtube [55]. The mixture was incubated at room temperature for 5 min with vortex mixing every 1.5 min, and nuclei were liberated by shearing through a needle. Isolated nuclei were diluted with Dulbecco’s phosphate buffered saline (D-PBS, Invitrogen, Carlsbad, CA) to a concentration no higher than 5 x 10⁵ nuclei/ml, stained with 7-aminoactinomycin D (7-AAD, Molecular Probes, Eugene, OR), and analyzed using a flow cytometer (Guava Personal Cell Analysis (PCA) system, Guava Technologies, Hayward, CA) to determine the total number of cells with intact nuclei in the sample.
Oxygen consumption rate (OCR)
OCR was measured as previously described [39]. Briefly, suspensions containing about 2500 IE/ml in DMEM containing 4.5 g/l glucose and 0.6 g/l L-glutamine supplemented with 100 U/ml penicillin, 100 µg/ml streptomycin, 10 mM HEPES, and no added serum were sealed in a 200-µl stirred titanium chamber (Micro Oxygen Uptake System, FO/SYSZ- P250, Instech Laboratories, Plymouth Meeting, PA) maintained at 37°C. The time dependent pO$_2$ within the chamber was recorded with a fluorescence-based oxygen sensor (Ocean Optics, Dunedin, FL), and data at high pO$_2$ were fit to a straight line. The maximal OCR was evaluated from OCR=V$_{ch}\alpha(\Delta pO_2/\Delta t)$, where V$_{ch}$ is the chamber volume and $\alpha = 1.27$ nmol/mm·Hg·ml is the Bunsen solubility coefficient for oxygen in medium [16]. OCR measurements were normalized by the measured number of cells (nuclei counting) or DNA content of the sample examined.
Membrane Integrity
7-Aminoactinomycin (7-AAD) sequential staining
Cell membrane integrity was assessed by differential staining with 7-AAD (Molecular Probes, Eugene, OR) [55]. An aliquot of about 300 islets was re-suspended in 100 µl of D-PBS, and 5 µl of 1 mg/ml 7-AAD and incubated for 20 min at 4°C protected from light. After two washes with 1 ml of D-PBS, cells were disrupted by adding equal volume of lysis solution in D-PBS to the islet suspension and sheared[16] as described for islet enumeration. Labeled nuclei were counted immediately in the flow cytometer or stored on ice for less than 15 min before counting. A portion of the islet suspension was further stained with 7-AAD, thereby labeling all of the previously unlabelled nuclei, and the total number of nuclei was counted. The fraction of cells with compromised membranes was estimated as the ratio of the initially stained nuclei (first measurement) to the total number of nuclei (second measurement).
Islet volume fraction by morphological analysis
A 0.5 ml aliquot of human islet preparation was processed to produce 1-µm sections stained with toluidine blue and examined by light microscopy [13]. Islet tissue was distinguished from non-islet tissue as previously described [13]. Sections were analyzed at 420X by stereological point counting [63] with a 90-point grid covering adjacent, non-overlapping fields. At point intercepts of the grid with tissue, the nature of the tissue was determined, and the volume fraction of islet tissue, denoted $\Phi_t$, was calculated as previously described [13].
β cell identification by immunocytochemical insulin staining
Tissue samples were washed twice with 1 ml D-PBS, fixed for 1 hr in 10% formalin (Sigma Aldrich), washed twice with 1 ml of D-PBS, and centrifuged at 300×g for 5 min. To the pellet, 500 μl of warm 2% (w/v) agarose was added. The solution was mixed using a vortex mixer for 1 sec, immediately centrifuged at 18000×g for 10 sec, allowed to cool for 2 min, and then 500 μl of 10% formalin was added. After 1-hr incubation at room temperature, formalin was removed, samples were stored in D-PBS until embedded in paraffin, and 5-μm sections were cut. Sections were stained by the unlabeled antibody peroxidase-antiperoxidase (PAP) method [64] using guinea pig anti-bovine insulin (Linco, St. Charles, MO) as primary antibody and goat anti-guinea pig IgG (Cappel, Irvine, CA), as secondary antibody, to link primary antibody to rabbit PAP complex (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). Use of a heterologous antibody gave specific staining with little background [65]. After applying PAP, slides were stained in 50 ml of a 2 mM 3,3'-diaminobenzidine tetrahydrochloride (DAB, Sigma Aldrich) solution activated with 25 μl of 30% (v/v) hydrogen peroxide (EMD Chemicals Inc, Gibbstown, NJ) for 30-120 sec, then viewed under a microscope. Reaction between hydrogen peroxide, DAB, and PAP formed a dark brown insoluble precipitate [66]. Slides were counterstained with hematoxylin (Sigma Aldrich). β cell volume fraction exclusive of vascular space, denoted Φ_{βXV}, was quantified using stereological point counting, as described for morphological analysis, with 200 points or more counted.
This procedure was used with samples after 2 days culture because trypsinization of adherent tissue degraded the structure and produced single cells, which, in turn, made it harder to morphologically differentiate between cell types and evaluate vascular volume fraction, whereas it was more accurate to distinguish insulin staining. Methods for calculating Φ_{βXV} (day 2) and comparing it with Φ_1 (day 0) are described in the Appendix.
Statistics
Measurements were made with three or more replicates and reported as mean ± SD. Statistical significance was determined using a Student t-test for unpaired data for comparing population means, and for paired data when appropriate.
3.3 Theoretical oxygen transport model
The species conservation equation for $O_2$ in the absence of convection [16; 97] is given by
$$\frac{\partial C}{\partial t} = D \nabla^2 C + R$$ \hspace{1cm} (3.1)
where $C$ is the concentration of $O_2$, $D$ is the diffusivity of $O_2$, and $R$ is the production rate of $O_2$. For convenience, the concentration terms in Equation 3.1 are replaced with $O_2$ partial pressure ($p$) using
$$C = \alpha p$$ \hspace{1cm} (3.2)
where $\alpha$ is the solubility of $O_2$. We assume that solubility is constant in each phase so that at steady state Equation 3.1 becomes
$$0 = D \alpha \nabla^2 p + R$$ \hspace{1cm} (3.3)
$O_2$ consumption within the islet is assumed to follow Michaelis-Menten kinetics [16] for which the reaction rate is expressed as
$$R = -\frac{V_{\text{max}} p}{K_m + p}$$ \hspace{1cm} (3.4)
where $V_{\text{max}}$ is the maximum reaction velocity, and the Michaelis constant $K_m$ is the value of $p$ at which the reaction rate is $V_{\text{max}}/2$. $O_2$ consumption occurs only above a critical value, $p_{\text{crit}}$. Below $p_{\text{crit}}$, the tissue dies and no longer consumes $O_2$. The species conservation equations in medium (m) and in the islet tissue (i) become, respectively,
$$(D \alpha)_m \nabla^2 p = 0$$ \hspace{1cm} (3.5)
$$0 = (D \alpha)_i \nabla^2 p - \frac{V_{\text{max}} p}{K_m + p}$$ \hspace{1cm} (3.6)
To model islet culture on an $O_2$-impermeable dish, all the tissue (islet and non-islet) was assumed to be in the form of spheres with uniform diameter ($d$) that were arranged in a square
array with center-to-center distance $w$ in a stagnant medium of height $h$ (Figure 3.1). Because of symmetry within the system, Equation 3.1 was solved within a unit cell containing only an eighth of an islet (dashed lines in Figure 3.1). The boundary conditions for this model [16] include continuity of $p$ across the gas-liquid interface,
at $z = h$, \[ p_{\text{gas}} = p_m \]
no $O_2$ flux across the $O_2$ -impermeable surface,
at $z = 0$, \[ \frac{\partial p}{\partial z} = 0 \]
continuity of $O_2$ flux and partial pressure across the medium-islet interface,
at $r = R_{\text{islet}}$, \[ n \cdot \left[ (-D\alpha \nabla p)_i - (-D\alpha \nabla p)_m \right] = 0 \]
\[ p_i = p_m \]
and symmetry across each side of the unit cell including both medium and islet domains as well as at $r = 0$,
\[ n \cdot (-D\alpha \nabla p) = 0 \]
Table 3.1 contains the values for the parameters used in the simulations. [98]
In order to accurately predict oxygen profiles during culture all tissue, islet and non-islet need to be considered. Traditionally islet density is reported as IE/cm$^2$, which underestimates tissue quantity by ignoring the exocrine tissue that can be a significant fraction of the overall preparation volume. This density is related to the density in the simulations through
\[ \Gamma = \frac{\rho}{\Phi_I} \]
where $\Phi_I$ is the islet volume fraction of the islet preparation, $\rho$ is the islet surface density in IE/cm$^2$, and $\Gamma$ is the sphere surface density in “IE”/cm$^2$ used in our model. In this case “IE” does not refer specifically to islet tissue, but rather a quantity of total tissue volume. This conversion
adds the assumption that the exocrine tissue is present as 150-\(\mu\)m diameter spheres. In this study tissue was quantified using nuclei counts, which relate to the model density through
\[
\Gamma = \frac{\chi}{n_{IE}}
\]
where \(\chi\) is the cell surface density (cells/cm\(^2\)) and \(n_{IE}\) is the number of cells in an IE (1560 cells/IE, [13]). Again we assume that exocrine tissue is similar to islet tissue in that a volume of tissue equal to a 150-\(\mu\)m diameter sphere contains the same number of cells as an IE. Total tissue surface density can be converted to an OCR surface density, which is the total OCR of the tissue divided by the area of culture, using
\[
\lambda = \chi V_{max} \frac{V_{IE}}{n_{IE}} = \Gamma V_{max} V_{IE}
\]
where \(\lambda\) is the OCR density with units of flux and \(V_{IE}\) is the volume of an IE.
**Figure 3.1. Geometry assumed for analysis of high density islet culture on an impermeable surface.**
*Dashed lines denote unit cell in which Equation 3.1 was solved.*
The model was solved with the finite element method using the commercially available software COMSOL Multiphysics (COMCOL Inc., Burlington, MA) in conjunction with Matlab (Mathworks, Natick, MA). We required the mesh to contain more than 1000 nodes and set a convergence criterion such that the change in successive solution vectors for concentration
estimated by the iterative solver in the finite element software (i.e. the difference between the current and previous concentration vectors) was less than $10^{-2}$ [99]. This corresponded to an average error in the fraction of non-viable tissue of < 0.01 (maximum error was 0.04) based on a variety of simulations run multiple times using different mesh and convergence criteria.
3.4 Results
3.4.1 Predictions of theoretical mode
The theoretical model was used to estimate O$_2$ partial pressure profiles in and around islets in culture. As illustrated in Figure 3.2, the partial pressure at the sphere surface decreases as tissue surface density increases. If this is too low, not enough O$_2$ will reach the center, and cores of dead tissue will develop. These cores increase in size as the tissue density increases, thereby decreasing the amount of viable tissue that can be collected from culture.
Figure 3.3 explores this phenomenon in greater detail. As the initial tissue surface density increases, the minimum value of p decreases until it reaches $p_{\text{crit}}$ at which the tissue dies and a necrotic core begins to develop. Using the calculated partial pressure field in the islet, the viable tissue volume ($V_V$) was calculated by integrating the volume of tissue with $p > p_{\text{crit}}$. The volume fraction of the viable tissue remaining after culture was calculated from
$$F_V = \frac{V_V}{\frac{4}{3} \pi R_{\text{islet}}^3}$$
Once p reaches $p_{\text{crit}}$, the volume fraction of non-viable tissue (1-F$_V$) increases as initial tissue surface density increases.
Table 3.1 Values for the parameters used in islet culture simulations and the source from which they were obtained.
| Parameter | Symbol | Value | Units | Source |
|----------------------------------|--------|----------------|-----------|----------|
| Michaelis constant | $K_m$ | 0.44 | mmHg | [98] |
| O$_2$ diffusivity in medium | $D_m$ | $2.78 \times 10^{-5}$ | cm$^2$/s | [16] |
| O$_2$ diffusivity in tissue | $D_t$ | $1.22 \times 10^{-5}$ | cm$^2$/s | [16] |
| O$_2$ solubility in medium | $\alpha_m$ | $1.27 \times 10^{-9}$ | mol/cm$^3$ mmHg | [16] |
| O$_2$ solubility in tissue | $\alpha_t$ | $1.02 \times 10^{-9}$ | mol/cm$^3$ mmHg | [16] |
| Minimum p for viability | $p_{\text{min}}$ | 0.1 | mmHg | [16] |
| Maximal O$_2$ consumption rate | $V_{\text{max}}$ | 9-38.5 | nmol/s cm$^3$ | Experimentally measured (Chapter 2) |
| Media height | h | 3 | mm | Based on medium volume and well area |
| Islet radius | $R_{\text{islet}}$ | 75 | $\mu$m | Based on definition of IE |
| External p | $p_{\text{gas}}$ | 142 | mmHg | Partial pressure of O$_2$ in humidified incubator at 37°C with 95% air/5% CO$_2$ |
The minimum oxygen partial pressure for viability ($p_{\text{crit}}$), defined as the value below which a cell will not survive, is assumed to be 0.1 mmHg [16].
$$V_{\text{max}} = \frac{\text{OCR} \ n_{\text{IE}}}{n_c \ V_{\text{IE}}}$$
where OCR is the value measured in a sample containing $n_c$ cells determined from nuclei counting, $n_{\text{IE}} = 1560$ is the number of cells in an IE [13], and $V_{\text{IE}} = 1.77 \times 10^{-8}$ cm$^3$ it the volume of an IE (a sphere having a diameter of 150 $\mu$m).
Figure 3.2. Predicted contours of constant partial pressure in and around islets in culture.
Data obtained using different tissue surface densities: (A) 0.1 nmol/min cm$^2$ (26 IE/cm$^2$), (B) 1 nmol/min cm$^2$ (260 IE/cm$^2$), and (C) 5 nmol/min cm$^2$ (1300 IE/cm$^2$). The cross section of half of an islet is illustrated for each density. The gray regions represent viable islet tissue and the black regions represent dead cores. Surface volume and OCR density were related by using $V_{\text{max}} = 36.8$ nmol/s cm$^3$ (2.5 fmol/min cell, Chapter 2).
Figure 3.3. Theoretical prediction for minimum islet p (dashed line) and non-viable tissue volume (solid line) as a function of islet surface density and OCR density. Surface volume and OCR density were related by using $V_{\text{max}} = 36.8 \, \text{nmol/s cm}^3$ (2.5 fmol/min cell, Chapter 2).
3.4.2 Visual observations
Tissue morphology changed during high density culture form discrete cell clusters with defined boundaries distributed throughout the dish to a mixture of larger less defined clusters with a less uniform distribution surrounded by a significant number of single cells (Figure 3.4). Portions of the dish become completely covered by tissue while other areas were nearly void of tissue. Islets in low density culture kept their well defined boundaries and do not clump.
Figure 3.4. Light microscopy pictures of human islets.
Pictures taken on day 0 and after 39 hr culture at densities of 400, 150, and $8 \times 10^4$ cells/cm$^2$ on untreated polystyrene at 37°C.
The scale bar represents 100 $\mu$m. Magnification is the same in all panels.
3.4.3 Effect of surface density on tissue recovery.
Following culture only non-adherent tissue is collected and transplanted in most, if not all, clinical islet programs and islet research studies. Previously we showed that low-density (< 8 x 10^5 cells/cm^2) human islet culture for 34-60 hours in a fully humidified environment at 37°C (142 mmHg O_2) on untreated polystyrene resulted in significant tissue adherence (Chapter 2). In this study we collected both adherent and non-adherent tissue. As the initial tissue surface density increased, the fraction of original cells (measured by nuclei counting) collected from culture as adherent tissue decreased while that of non-adherent tissue increased (Figure 3.5). At all except the lowest tissue surface densities, the sum of the non-adherent + adherent cells collected was equal to that of the initial tissue placed into culture for all the cultured/shipped preparations. The data in Figure 3.5A were also plotted as tissue surface density of collected adherent tissue versus the initial tissue surface density in Figure 3.6. As the initial surface density of cultured/shipped islets increased the adherent surface density increased, and began to level off at high initial surface densities. In both Figure 3.5A and Figure 3.6, there is insufficient data for fresh islets to determine if they behave similarly to cultured/shipped islets.
In culture experiments carried out at densities > 50 x 10^4 cell/cm^2 trypsinization was not conducted because the fraction of the total tissue that was adherent was usually less than 10%. This was always confirmed by visual inspection under the microscope. The fraction of original tissue collected as non-adherent tissue in terms of nuclei and DNA is plotted versus initial tissue surface density in Figure 3.7A and B, respectively, for densities > 50 x 10^4 nuclei/cm^2. The fractions recovered averaged 0.80 ± 0.21, and the DNA content averaged 7.8 ± 1.6 pg DNA/cell (Figure 3.7C). No consistent trend was observed for fresh or cultured/shipped tissue with initial tissue surface density, but fresh tissue tended to have a lower average nuclei recovery (0.69 ± 0.13) and higher DNA recovery (0.90 ± 0.16) than cultured/shipped tissue (0.91 ± 0.16 and 0.64 ± 0.20, respectively, p < 0.01 in both cases for unpaired t-test).
The volume fraction of insulin positive cells excluding vascular space (\Phi_{\text{BXV}}) in tissue collected from culture is shown as a function of tissue surface density in Figure 3.8A for non-adherent tissue and Figure 3.8B for adherent tissue. \Phi_{\text{BXV}} was always substantially higher for non-adherent tissue (0.47 ± 0.01 for fresh, 0.39 ± 0.09 for cultured/shipped) than for adherent tissue (0.15 ± 0.10 for fresh, 0.21 ± 0.07 for cultured/shipped). The \Phi_{\text{BXV}} of the tissue upon
receipt at MIT prior to experimentation (day 0) was $0.32 \pm 0.03$ for fresh and $0.41 \pm 0.03$ for cultured/shipped human islet tissue, which was higher than that for adherent tissue ($p < 0.01$ for both fresh and culture/shipped human islets), similar to that for cultured/shipped non-adherent tissue ($p > 0.64$), and lower than that for the fresh human islet non-adherent tissue ($p < 0.01$) collected from culture.
By combining the data for tissue recovery with that for $\Phi_{\text{BXV}}$ the number fraction of original $\beta$ cells remaining after culture $F_\beta$ was determined (Appendix 10.1). As indicated in Figure 3.9, the non-adherent tissue collected from culture contained more insulin positive cells than the adherent tissue, for which the fraction recovered was less than 15% for all densities and preparations considered except at very low density.
### 3.4.4 Membrane integrity and oxygen consumption rate
Membrane integrity measurements were made using sequential nuclei staining with 7AAD. OCR measurements were made with the same samples and normalized by number of cells (from nuclei counting) or mass of DNA in the sample. Membrane integrity decreased as the initial tissue surface density increased, but data were widely scattered (Figure 3.10A). OCR/cell (Figure 3.10B) and OCR/DNA (Figure 3.10C) decreased sharper and to a much greater extent than the fraction of cells with intact membranes, and the scatter between measurements from different experiments was relatively small.
Figure 3.11 shows the fraction of the original OCR collected after culture as a function of the initial tissue surface density (Figure 3.11A) and initial OCR density (Figure 3.11B). The fraction of OCR recovered was $0.98 \pm 0.07$ and $0.75 \pm 0.23$ at low density ($< 8 \times 10^4$ cells/cm$^2$) for fresh and cultured/shipped islets, respectively (Chapter 2). The fraction decreased as the initial density increased and dropped to about 0.1 at the highest densities studied. By plotting the data versus OCR surface density the data collapsed into a tighter cluster with less scatter compared to the same data plotted versus tissue surface density.
Figure 3.5. The fraction of original tissue collected from culture as (A) adherent, (B) non-adherent, and (C) total tissue as a function of the initial tissue surface density.
Figure 3.6. Data from Figure 3.5A are plotted in terms if the actual surface density of adherent tissue collected from culture versus the initial tissue surface density.
The OCR density after culture, defined as the total OCR of the tissue remaining divided by the surface area of culture, which has units of O$_2$ flux, is plotted as a function of the initial OCR density in Figure 3.12. The OCR density increases before leveling off at roughly 1.5 nmol/min cm$^2$ at an initial OCR density of roughly 4 nmol/min cm$^2$. The O$_2$ flux for one-dimensional Fickian diffusion is calculated by
$$N = \frac{D\alpha}{h} \left( p_{\text{gas}} - p_{\text{surf}} \right)$$
where $N$ is the O$_2$ flux and $p_{\text{surf}}$ is the partial pressure at the top of the islet. Using the values in Table 3.1 the maximum O$_2$ flux to the tissue, which occurs when $p_{\text{surf}}$ is zero, is 1.0 nmol/min cm$^2$.
Figure 3.7. The fraction of original cells collected as non-adherent tissue, plotted as (A) nuclei and (B) DNA, and (C) the DNA content per cell of tissue.
Figure 3.8. The volume fraction of insulin positive tissue excluding vascular space for adherent (A) and non-adherent (B) tissue collected from culture.
Figure 3.9. Fraction of original insulin positive cells for adherent (A), non-adherent (B), and total tissue (C) collected from culture as a function of the initial tissue surface density.
Figure 3.10. (A) Fraction of cells with intact membranes as measured by 7AAD sequential staining, (B) OCR/cell, and (C) OCR/DNA after culture versus initial tissue surface density.
3.4.5 Prediction of oxygen transport theory compared to experimental data
The experimental data from Figure 3.11 were compared to theoretical predictions of the $O_2$ transport model and are plotted versus initial tissue surface density (Figure 3.13A) and initial OCR surface density (Figure 3.13B). The theoretical model indicates that the fraction of original viable tissue recovered after culture, $F_V$, is a function of both initial tissue surface density and $V_{\text{max}}$ (Figure 3.13A). The two theoretical curves in Figure 3.13 represent the lowest and highest $V_{\text{max}}$ for the islet preparations used in this study. At large $V_{\text{max}}$, the model predicts that $F_V$ decreases sharply with even small increases in tissue surface density. Because $V_{\text{max}}$ of the tissue used in this study varied significantly between islet preparations, a separate plot would be required to compare prediction with data for each preparation. The same theoretical predictions from Figure 3.13A are plotted in Figure 3.13B with tissue surface density replaced by initial OCR surface density. When plotted on these coordinates, all theoretical predictions collapse to a small band of curves for all values of $V_{\text{max}}$, thereby permitting comparison of results from different experiments on one figure. A similar collapse is seen with the experimental data resulting in a tighter cluster with less scatter as noted above. The experimental data follow the same trend as theory predicts, but consistently give a lower $F_V$ than predicted.
**Figure 3.11.** The fraction of original OCR remaining as a function of (A) initial tissue surface density and (B) initial OCR density following culture.
Figure 3.12. The OCR density of all tissue remaining after culture as a function of the initial OCR surface density.
Figure 3.13. Experimental results and theoretical predictions for the fraction of original viable tissue recovered from culture versus (A) tissue surface density and (B) OCR density. In order to plot the theoretical predictions and experimental results versus initial tissue surface density, we assumed that there are 1560 nuclei/IE [13].
3.5 Discussion
There is little published information on the effect of surface density on the recovery of islet tissue and its viability and purity after culture. In this study, we studied culture of human islet tissue on untreated polystyrene vessels for two days and investigated the effect of tissue surface density on recovery of islet tissue and its viability and purity. Through the use of a theoretical $O_2$ transport model, we also explored the notion that the reduction in the amount of viable tissue observed during high density islet culture resulted from $O_2$ limitations.
At low densities ($< 8 \times 10^4$ cells/cm$^2$), a fraction of the tissue adheres during culture on untreated polystyrene (Chapter 2). We observed that as initial tissue surface density increased the amount of adherent tissue increased, the fraction that was adherent decreased, and the fraction non-adherent increased (Figure 3.5). At all except the lowest tissue surface densities, all of the tissue was recovered in the two fractions after 2-day culture from cultured/shipped islet preparations but not from fresh tissue. This difference may reflect damage to tissue during isolation, resulting in cells that have begun the death process and are destined to die but have intact nuclei when culture of fresh tissue is initiated. In contrast, these damaged cells and their nuclei have already disappeared when the cultured/shipped tissue is placed into culture. The occasional recovery of more than 100% of initial tissue is likely due to experimental error. Pancreatic tissue growth is unlikely because it occurs at low rates and is observed only after longer culture periods [77; 100].
There are at least three possible explanations for the decrease in the fraction of adherent tissue with increasing tissue surface density: (1) As tissue covers the culture surface there is less area for remaining tissue to attach, and the density of adherent tissue begins to level off at high initial tissue surface densities (Figure 3.6). (2) Healthy tissue may adhere to a greater extent. Some cells detach from a surface early in the cell death process when exposed to specific stresses [101; 102]. This hypothesis is consistent with the largest adherent fraction occurring at low density (Chapter 2), for which $O_2$ limitations are minimized. (3) Tissue may adhere to other tissue instead of the vessel surface, leading to clumping that was observed in the highest-density cultures (Figure 3.4).
Cultured/shipped tissue had higher nuclei recoveries than fresh tissue (Figure 3.7), whereas DNA recoveries were higher for fresh tissue. As noted previously, fresh tissue has more recently entered the cell death process following the harsh isolation procedure than
cultured/shipped tissue, which has proceeded further through the cell death process. During culture of fresh tissue, nuclei of damaged cells degrade and liberate DNA, which degrades at a slower rate. At the start of our culture experiments, cultured/shipped tissue had already largely completed these steps. Therefore, our data reflects a greater loss of nuclei from cells already committed to the cell death process in fresh tissue at the start of culture and the presence of a higher amount of DNA released from dead nuclei but not yet degraded by DNAase in the recovered fresh tissue as compared to the data for cultured/shipped tissue.
As observed with low density culture (Chapter 2), the volume fraction of $\beta$ cells exclusive of vascular space ($\Phi_{\beta XV}$) of the non-adherent tissue was higher than that of the adherent tissue (Figure 3.8), indicating that non-$\beta$ cells, and thus non-islet tissue, preferentially adhered to the culture vessel. Combination of $\Phi_{\beta XV}$ measurements with tissue recovery data revealed that the number fraction of original $\beta$ cells collected as adherent tissue was always less than 0.25 at low density (Figure 3.9) and less than 0.15 otherwise, indicating that the current clinical practice of collecting only non-adherent tissue usually results in collecting most of the insulin positive tissue. In most cultures (10 of 13 in Figure 3.9C) the combined tissue lost at least 25% of the original $\beta$ cells. This loss occurred even under low density conditions when no loss in viable tissue was observed (Chapter 2), indicating that the $\beta$ cells lost were already dead at the beginning of culture or dedifferentiated during culture [77]. During high density culture, the amount of viable tissue decreased, suggesting that the reduction in the number fraction of original $\beta$ cells could also have resulted from stresses during culture.
Initial tissue surface density had a large impact on tissue viability. A consistent decrease in membrane integrity was observed as the initial tissue surface density increased (Figure 3.10A). A much larger and more consistent decline was observed in OCR/cell and OCR/DNA (Figure 3.10B and C) because loss of membrane integrity occurs late in the cell death process whereas a decrease in OCR reflects mitochondrial dysfunction, which is an earlier indicator of commitment to the cell death process [13]. This viability decrease is consistent with the higher glucose stimulation indices measured after low surface density compared to high surface density culture [47].
The fraction of original OCR remaining after culture, which represented the fraction of original viable tissue $F_V$, decreased drastically as initial tissue surface density (Figure 3.11A) and initial OCR surface density (Figure 3.11B) increased. The drop was especially sharp in the range
of $0-10^6$ cells/cm$^2$ and $0-2$ nmol/min cm$^2$, respectively. The data were scattered when plotted versus tissue surface density but collapsed together when plotted versus OCR surface density, especially at high density. This behavior is consistent with the O$_2$ transport model. The initial OCR surface density takes into account variation in V$_{max}$ between islet preparations, which ranged from 9.0 to 36.8 nmol/s cm$^3$ in this study. This difference in V$_{max}$ is not taken into account when plotting versus tissue surface density, whereas OCR density combines the independent effects of tissue surface density and V$_{max}$. Furthermore, the OCR density represents the oxygen flux averaged over the surface, which becomes an increasingly good estimate of the actual flux at very high tissue densities wherein the tissue layer takes on the characteristics of a slab rather than discrete spheres.
The OCR surface density of respiring tissue following high density culture leveled off at 1.5 nmol/min cm$^2$ (Figure 3.12), a value higher than the maximum O$_2$ flux of 1.0 nmol/min cm$^2$ for one-dimensional Fickian diffusion through 3 mm medium, which was calculated using the parameters in Table 3.1. OCR density of recovered tissue is based on measurements made with tissue that is fully oxygenated. During high density culture, tissue is not fully oxygenated, and a substantial fraction of viable tissue is exposed to very low values of p on the order magnitude of the Michaelis constant K$_m$ at which the OCR is reduced according to Equation 3.4. Thus, OCR measurements of recovered tissue, which are made under reoxygenated conditions to assess fractional viability, cannot be used to estimate the actual OCR density under culture conditions because they lead to a higher OCR density than is actually present in culture at very high density.
Fraction viability predicted by theory followed the same trend as the experimental data (Figure 3.13), suggesting that O$_2$ limitations caused at least part of the decrease in viable tissue during culture. However, model predictions consistently overpredicted the fraction viable by about 10-20%. The difference between theoretical predictions and experimental measurements could be due to inaccuracies in the parameter values selected or assumptions made in the theoretical model (Chapter 5). In developing the model we assumed there are 150-$\mu$m diameter spheres (the average for human islets) of islet and non-islet tissue, arranged in a square array. Human islet preparations are a mixture of islets and exocrine tissue, and the OCR/cell for each type is about the same (Chapter 2). However, the two types can vary significantly in shape, size, and spatial distribution, and occasionally very large clumps of non-islet tissue were observed. In
addition, tissue cultured at high density tended to clump even more (Figure 3.4), which exacerbated local $O_2$ limitations. Differences between theoretical predictions and experimental data may also have been linked to the lack of medium change during culture. At high density, nutrients (i.e. glucose) are depleted and waste (i.e. lactate and ammonia) is produced faster than at low density, and the concentrations of these species at the cell surface are further influenced by diffusional limitations. Investigation into the effect of glucose and waste diffusional limitations are described elsewhere (Chapter 6 and 7). Products produced by islet tissue may also have toxic effects especially at high surface densities. NO, which can be produced by islet tissue, can damage pancreatic $\beta$-cells leading to cell death [103; 104]. Further, cell death by necrosis results in leakage of cell contents [105] which could also expose the tissue to damaging toxins.
Current practice of collecting only non-adherent tissue is supported by our results, which indicate that impure tissue is removed with little loss of $\beta$ cells in the adherent fraction. Sequential culture on fresh polystyrene surfaces may further purify tissue, as suggested by our results with precultured and shipped tissue. Using estimates of 1560 cells/IE and about 50% islet cells in an average human islet preparation [55], 200 and 400 IE/cm$^2$ correspond to 0.625 and $12.5 \times 10^6$ cells/cm$^2$, respectively. Our data indicate that culture at these densities in 3 mm medium, which is typical of clinical preparations, results in $F_V$ values that vary from 0.4 to 0.9 depending on damage prior to culture, which is prevalent in fresh tissue (Chapter 2), and on limitations imposed by culture, which at a high $V_{\text{max}}$ will decrease $F_V$ by 20%. Rather than culturing for 2 days at 37°C, some islet centers culture one day at 37°C followed by one day at about 25°C, during which OCR, and therefore $O_2$ diffusion limitations are reduced. We suggest that, because the time constant for diffusion through the culture medium at 37°C is about an hour (even less inside an islet) steady state $O_2$ gradients prevail within the system at 37°C, exposing tissue to hypoxia for more than 20 hours in high density culture and cause irreversible damage. Thus even though $O_2$ limitations are reduced at 25°C, the damage is already done during 37°C culture on the first day and similar results to those presented here should be observed. The empirical data and theoretical predictions presented here can serve as a guide to choosing a safe culture density based on measurements of nuclei counts and OCR measurements with the freshly isolated tissue preparation, which can usually be completed in less than an hour [39; 55].
### 3.6 Nomenclature
| Symbol | Description |
|--------|-------------|
| 7-AAD | 7-aminoactinomycin D |
| C | Concentration of $O_2$ |
| d | Diameter of islet |
| D | Diffusivity of $O_2$ |
| DAB | 2 mM 3,3'-diaminobenzidine tetrahydrochloride |
| D-PBS | Dulbecco’s phosphate buffered saline |
| $F_\beta$ | Number fraction of original $\beta$ cells collected from culture |
| FBS | Fetal bovine serum |
| $F_v$ | The fraction of original viable tissue after culture |
| h | Medium depth |
| IE | Islet equivalent |
| JDC | Joslin Diabetes Center |
| $K_m$ | Michaelis constant, the p at which the reaction rate is $V_{\text{max}}/2$ |
| MIT | Massachusetts Institute of Technology |
| N | $O_2$ flux |
| $n_{IE}$ | Number of cells in an islet equivalent volume |
| $O_2$ | Oxygen |
| OCR | $O_2$ consumption rate |
| p | $O_2$ partial pressure |
| PAP | Peroxidase-antiperoxidase |
| $p_{\text{crit}}$ | The p below which a cell dies and no longer consumes $O_2$ |
| $p_{\text{gas}}$ | Ambient partial pressure of $O_2$ |
| $pO_2$ | $O_2$ partial pressure |
| $p_{\text{surf}}$ | p at the surface of the islet |
| r | Radial direction |
| R | Production rate of $O_2$ |
| $R_{\text{islet}}$ | Radius of islet |
| t | Time |
| Symbol | Description |
|--------|-------------|
| $V_{ch}$ | Volume of OCR chamber |
| $V_{IE}$ | Volume of an islet equivalent ($1.77 \times 10^6 \, \mu m^3$) |
| $V_{max}$ | Maximum reaction velocity |
| $V_v$ | Viable tissue volume |
| $w$ | Center-to-center distance between islets |
| $z$ | Vertical axis |
| $\alpha$ | Solubility of $O_2$ in medium |
| $\Phi_{\beta XV}$ | Volume fraction of $\beta$ cells exclusive of islet vascular spaces |
| $\Phi_i$ | Volume fraction of islets |
| $\Gamma$ | Sphere surface density in “IE”/cm$^2$ where “IE” refers to total tissue volume. |
| $\rho$ | Islet surface density in IE/cm$^2$ |
| $\chi$ | Cell surface density (cells/cm$^2$) |
| $\lambda$ | OCR density with units of flux |
**Subscripts for D, $\alpha$, p, C**
- **m**: Medium
- **i**: Islet tissue
4 Reducing oxygen limitations during static islet culture
4.1 Introduction
Since its initial report of success [7] transplantation of pancreatic islets has become a promising option for treating type 1 diabetes. After islet tissue is isolated from a donor pancreas by mechanical and enzymatic treatment [13; 14], it may transplanted directly or cultured [14; 40]. Islet culture prior to transplantation is standard practice in most clinical centers [40]. During culture in standard polystyrene culture vessels, at very low islet tissue surface densities, islets are usually free of oxygen ($O_2$) transport limitations (Chapter 2 and 3 and [16]), but it is impractical to culture islet tissue at these densities because of large space requirements. At islet densities of 200-400 islet equivalents (IE)/cm$^2$ typically employed in clinical applications [106; 107], there are significant $O_2$ limitations that lead to a loss in viability (Chapter 3, [16]). It is desirable to culture at even higher densities without sacrificing viable tissue.
Improvements in culture techniques have been explored, but only a few address $O_2$ limitations present in conventional culture. Lower medium depths decreased the $O_2$ diffusion distance to the islet, thereby increasing islet oxygenation [48; 49; 108] and improving islet stimulation index [47]. However, at high surface densities, culture in low medium depths may result in nutrient (e.g. glucose) depletion from the medium (Chapter 6). Culture at the interface of culture medium and an oxygenated perfluorocarbon (PFC), which acts as an $O_2$ reservoir because of its high $O_2$ solubility, caused islets to have more DNA fragmentation and a lower stimulation index compared to islets cultured without the PFC [109]. However, very low surface densities (15 IE/cm$^2$) were used, for which $O_2$ limitations are not present, and further study further study is needed to determine whether PFC could be beneficial for high density culture. Structural integrity and islet function are better preserved with rotational culture systems compared to static cultures [110; 111] and have been used extensively with single cell suspensions [108; 112-114], but questions remain about islet clumping and scale up for large quantities of tissue.
In this paper we investigate three static culture techniques that focus on decreasing $O_2$ supply limitations in high density culture, including culture at reduced temperatures, in elevated $O_2$ environments, and on silicone rubber membranes that have high oxygen permeability. Decreasing culture temperature reduces cellular $O_2$ consumption rate (OCR) [115] thereby
reducing the magnitude of oxygen diffusion gradients in the islets and culture medium. Among studies examining reduced temperature islet culture, several report lower necrotic death [50; 52] and higher recoveries [49] compared to culture at 37°C, which is consistent with the notion of reduced O₂ transport limitations at reduced temperature. However, 37°C culture results in better function and/or higher viability than culture at reduced temperatures [49]. The interplay between islet surface density and temperature may be the cause of these conflicting reports.
Increasing ambient O₂ in the incubator, which is easily achieved using a premixed gas cylinder, increases the maximum O₂ flux to the tissue, thereby improving tissue oxygenation. However, at high O₂ levels, O₂ toxicity may damage the tissue. The O₂ partial pressure (pO₂) at which O₂ becomes toxic to islet tissue is unknown.
Culture on a thin, O₂ -permeable membrane bottom greatly increases O₂ transfer to the tissue. The estimated pO₂ drop during culture at 400 IE/cm² across 3 mm medium is > 140 mmHg while that across a 500-μm silicone rubber membrane is < 10 mmHg. An initial investigation indicated that a small foot-print vessel with an O₂ -permeable silicone rubber bottom can be used to culture islet tissue at high density [60].
In this study we investigated islet culture at elevated pO₂, reduced temperature, and on a silicone rubber membrane and compared the results to conventional culture in terms of the recovery of islet tissue and its viability for a range of surface densities. In addition, we utilized a theoretical O₂ transport model to investigate how O₂ transport changes for each system. The results of this study demonstrate that at high tissue surface densities the recovery of viable tissue is higher following culture in elevated O₂, at reduced temperature, or on a silicone rubber membrane compared to conventional culture in polystyrene vessels because O₂ limitations are reduced.
4.2 Methods
Islet Isolation and Culture
Standard collagenase/protease digestion methods were used for human [7; 15] isolations. Preparations of human islets were provided fresh (4-6 hr after isolation was complete) by the Islet Core at the Joslin Diabetes Center (JDC, Boston, MA). Human islets were also received after culture (1-2 days) from JDC or overnight shipment (sometimes after up to 1 day culture) from other centers contained in culture flasks (T-25 and T-75) or centrifuge tubes in Styrofoam
boxes. When first observed, some preparations were in suspension, some in a pellet. Upon receipt, a portion of the preparation was analyzed and the remainder placed into untreated polystyrene culture flasks at 37°C in a humidified environment (5% CO₂), oxygen partial pressure (pO₂) of 142 mmHg, for 34-60 hr during which no medium change occurred.
Islets were cultured in supplemented RPMI (11.1 mM D-glucose, 100 U/ml penicillin, 100 µg/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate (all from Mediatech Inc., Herdon, VA), and 50 µM β-mercaptoethanol (Sigma Aldrich, St. Louis, MO) with 10% FBS (Mediatech Inc.), at low densities (< 8x10⁴ cells/cm² or about 50 islet equivalents (IE)/cm², where IE is the volume of a sphere with a diameter of 150 µm) in 1.3 mm of medium in tissue culture flasks. After culture the free, non-adherent tissue (which is the only tissue used in current practice) was collected. In selected experiments, adherent human tissue was incubated for 5 min at 37°C in 0.05% trypsin EDTA (Mediatech Inc.), dislodged with fresh medium, and collected.
**Elevated O₂ islet culture**
Culture at elevated ambient pO₂ was the same as conventional culture except that the culture vessels were kept in a separate 5 L chamber in a 37°C incubator with a dish of water to provide a fully humidified environment during the experiment. Through the chamber a gas mixture of 60% O₂, 5% CO₂, and balance N₂ (p = 428 mmHg when fully humidified) was purged for 15 min at 3 L/min, before reducing to a very small flow rate in order to keep the gas composition constant throughout the experiment.
**Reduced temperature islet culture**
Culture at reduced temperature was conducted at 24°C in 5 L chambers through which a gas mixture of 20.9% O₂, 5% CO₂, and balance N₂ flowed (at 24°C pO₂ ≈ 154 mmHg). A flow rate of 3 L/min was used to purge the chamber for 15 min, after which the flow rate was reduced to a very small flow rate. After 34-60 hr culture at 24°C, islets were cultured for an additional hour at 37°C to allow tissue metabolic rates to acclimate to 37°C before analysis.
**Islet culture on silicone rubber membranes**
Culture on silicone rubber membranes was conducted using custom made epoxy blocks containing wells with a membrane at the bottom that was either 100 or 500-µm thick (Wilson Wolf Manufacturing Inc., New Brighton, MN). Culture medium, environment, duration and analysis were conducted as with conventional culture. Cultures conducted on silicone rubber
membranes used a medium depth of 22 mm in order to avoid nutrient depletion from the medium. Because islet tissue surface densities greater than $80 \times 10^4$ cells/cm$^2$ were tested, no trypsinization was required.
**Cell culture**
$\beta$TC3 cells (passages 40-50) were cultured in DMEM supplemented with 10% FBS, 100 U/ml penicillin, and 100 $\mu$g/ml streptomycin (Mediatech Inc., Herndon, VA). Cells were cultured in T-flasks placed in 5 L chambers. Using gas cylinders, O$_2$ partial pressure was controlled at 20, 40, 60, 80, and 95% O$_2$, with 5% CO$_2$ in all mixtures. The chambers were purged for 15 min at 3 L/min before reducing the flow rate.
**DNA Content**
At MIT, DNA was measured by fluorospectrophotometry [55] using the CyQUANT Cell Proliferation Assay Kit (Molecular Probes, C-7026 Eugene, OR), and a $\lambda$ DNA standard (Invitrogen) and 50-250 IE/sample. Fluorescence was read at 480 nm excitation and 520 nm emission wavelengths in a plate reader (Spectra MAX Gemini microplate spectrophotometer, Molecular Devices, Sunnyvale CA). At UMN (porcine and human islets), PicoGreen dye was used (Quant-iT PicoGreen dsDNA kit, Molecular Probes, Eugene, OR) [39]. Fluorescence was read at 520 nm excitation and 585 nm emission wavelengths.
**Islet Enumeration by Nuclei Counting**
Nuclei were prepared by adding equal 100-$\mu$l volumes of sample containing 160 or more IE and of lysis solution (0.1 M citric acid (Sigma) and 1% (v/v) Triton X-100 (Sigma)) to a 1.5 ml microtube [55]. The mixture was incubated at room temperature for 5 min with vortex mixing every 1.5 min, and nuclei were liberated by shearing through a needle. Isolated nuclei were diluted with Dulbecco’s phosphate buffered saline (D-PBS, Invitrogen, Carlsbad, CA) to a concentration no higher than $5 \times 10^5$ nuclei/ml, stained with 7-aminoactinomycin D (7-AAD, Molecular Probes, Eugene, OR), and analyzed using a flow cytometer (Guava Personal Cell Analysis (PCA) system, Guava Technologies, Hayward, CA) to determine the total number of cells with intact nuclei in the sample.
**Oxygen Consumption Rate (OCR)**
OCR was measured as previously described [39]. Briefly, suspensions containing about 2500 IE/ml in DMEM containing 4.5 g/l glucose and 0.6 g/l L-glutamine supplemented with
100 U/ml penicillin, 100 µg/ml streptomycin, 10 mM HEPES, and no added serum were sealed in a 200-µl stirred titanium chamber (Micro Oxygen Uptake System, FO/SYSZ-P250, Instech Laboratories, Plymouth Meeting, PA) maintained at 37°C. The time dependent pO$_2$ within the chamber was recorded with a fluorescence-based oxygen sensor (Ocean Optics, Dunedin, FL), and data at high pO$_2$ were fit to a straight line. The maximal OCR was evaluated from OCR=V$_{ch}\alpha(\Delta pO_2/\Delta t)$, where V$_{ch}$ is the chamber volume and $\alpha = 1.27$ nmol/mm-Hg·ml is the Bunsen solubility coefficient for oxygen in medium [16]. OCR measurements were normalized by the measured number of cells (nuclei counting) or DNA content of the sample examined.
**Membrane Integrity**
**7-Aminoactinomycin (7-AAD) Sequential Staining**
Cell membrane integrity was assessed by differential staining with 7-AAD (Molecular Probes, Eugene, OR) [55]. An aliquot of about 300 islets was re-suspended in 100 µl of D-PBS, and 5 µl of 1 mg/ml 7-AAD and incubated for 20 min at 4°C protected from light. After two washes with 1 ml of D-PBS, cells were disrupted by adding equal volume of lysis solution in D-PBS to the islet suspension and sheared as described for islet enumeration. Labeled nuclei were counted immediately in the flow cytometer or stored on ice for less than 15 min before counting. A portion of the islet suspension was further stained with 7-AAD, thereby labeling all of the previously unlabelled nuclei, and the total number of nuclei was counted. The fraction of cells with compromised membranes was estimated as the ratio of the initially stained nuclei (first measurement) to the total number of nuclei (second measurement).
**Sytox Orange/LDS 751**
Cell suspensions were stained with a solution (in D-PBS) containing 0.8 µM LDS 751 (Molecular Probes), a membrane permeable dye and 0.2 µM Sytox Orange (Molecular Probes), a membrane impermeable dye. LDS 751 was dissolved in dimethyl sulfoxide (DMSO). Samples with a cell concentration no higher than 5×10$^5$ cells/ml were incubated for 5 min at room temperature and analyzed in a flow cytometer (Guava PCA) using the Guava Viacount software. The fraction of cells with compromised membranes was the ratio of the number of cells stained with Sytox Orange over the number of cells stained with LDS 751.
Statistics
Measurements were made with three or more replicates and reported as mean ± SD. Statistical significance was determined using a Student t-test for unpaired data for comparing population means, and for paired data when appropriate.
4.3 Theoretical model
4.3.1 Culture on oxygen-impermeable surfaces
The steady state theoretical O₂ transport model for culture on an O₂-impermeable dish was described previously (Chapter 3), and only parameter values needed to be changed in this study. For elevated O₂ concentration, the ambient partial pressure of O₂ (p_gas) was set to 428 mmHg, which was obtained using a premixed gas cylinder of 60% O₂ at 37°C. At 24°C with normoxic conditions, p_gas was 154 mmHg. Reduced culture temperature affected several parameters. The O₂ diffusivities and solubilities at 37°C and 24°C are given in Table 4.1. V_max was assumed to follow an Arrhenius expression such that it decreased exponentially with inverse temperature as illustrated for a variety of cell lines [115]. The values for the parameters in the expression were determined experimentally and are given later.
4.3.2 Culture on silicone rubber
The diffusion-reaction equation applied to steady state conventional culture (Chapter 3) was modified for culture on silicone rubber. O₂ consumption within the islet is assumed to follow Michaelis-Menten kinetics [16] for which the reaction rate is expressed as
\[ R = -\frac{V_{\text{max}} p}{K_m + p} \]
where \( p \) is the partial pressure of O₂, \( V_{\text{max}} \) is the maximum reaction velocity, and the Michaelis constant \( K_m \) is the partial pressure at which the reaction rate is \( V_{\text{max}}/2 \). This O₂ consumption reaction only occurs above a critical value, \( p_{\text{crit}} \). Below \( p_{\text{crit}} \), the tissue dies and no longer consumes O₂. The species conservation equation in medium (m), islet tissue (i), and silicone rubber (Si) become, respectively,
\[(D\alpha)_m \nabla^2 p = 0 \quad (4.2)\]
\[(D\alpha)_i \nabla^2 p - \frac{V_{\text{max}} p}{K_m + p} = 0 \quad (4.3)\]
\[(D\alpha)_{Si} \nabla^2 p = 0 \quad (4.4)\]
where \( \alpha \) is the solubility of \( O_2 \) and \( D \) is the diffusivity of \( O_2 \).
All the tissue (islet and non-islet) was assumed to be in the form of spheres with uniform diameter (\( d \)) that were arranged in a square array with center-to-center distance \( w \) in a stagnant medium of height \( h \) (Figure 4.1) on an \( O_2 \)-permeable surface of thickness \( \delta \). Because of symmetry within the system, Equations 4.2-4.4 were solved within a unit cell containing only an eighth of an islet (dashed lines in Figure 4.1). The boundary conditions for this model include continuity of \( pO_2 \) across the gas-liquid and gas-silicone rubber interfaces
at \( z = h \),
\[ p_{\text{gas}} = p_m \quad (4.5) \]
at \( z = -\delta \),
\[ p_{\text{gas}} = p_{Si} \quad (4.6) \]
continuity of the \( O_2 \) flux and \( pO_2 \) across the liquid-islet interface,
\[ n \bullet ((-D\alpha \nabla p)_i - (-D\alpha \nabla p)_m) = 0 \quad (4.7) \]
\[ p_i = p_m \quad (4.8) \]
the membrane-islet interface,
at \( r = 0, z=0 \),
\[ n \bullet ((-D\alpha \nabla p)_i - (-D\alpha \nabla p)_{Si}) = 0 \quad (4.9) \]
\[ p_i = p_{Si} \quad (4.10) \]
and across the membrane-liquid interface,
\[ n \bullet ((-D\alpha \nabla p)_{Si} - (-D\alpha \nabla p)_m) = 0 \quad (4.11) \]
and symmetry across each side of the unit cell including medium, islet, and silicone rubber domains (indicated by dashed lines in Figure 4.1)
\[ n \bullet (-D_o \nabla p) = 0 \]
The \( O_2 \) solubility in silicone rubber is \( 1.21 \times 10^{-8} \text{ mol/mmHg cm}^3 \) [116], the \( O_2 \) diffusivity in silicone rubber is \( 2.17 \times 10^{-5} \text{ cm}^2/\text{s} \) [116] and the thickness of the silicone membrane was either 100 or 500 \( \mu\text{m} \). As with conventional culture \( V_{\text{max}} = 36.8 \text{ nmol/s cm}^3 \) (about 2.5 fmol/min cell from Chapter 2 and 1560 cells/IE [13]). Though 2.5 fmol/min cell is higher than the OCR of fresh human islet tissue determined in Chapter 2, it is only used when determining theoretical predictions plotted versus OCR density, which was shown only weakly dependent on \( V_{\text{max}} \) (Chapter 3).
As illustrated previously (Chapter 3) in order to accurately predict oxygen profiles during culture all tissue, islet and non-islet need to be considered. Islet density reported as IE/cm\(^2\) is related to the density in the simulations through
\[ \Gamma = \frac{\rho}{\Phi_I} \]
where \( \Phi_I \) is the islet volume fraction of the islet preparation, \( \rho \) is the islet surface density in IE/cm\(^2\), and \( \Gamma \) is the sphere surface density in “IE”/cm\(^2\) used in our model. In this case “IE” does not refer specifically to islet tissue, but rather a quantity of total tissue volume. This conversion adds the assumption that the exocrine tissue is present as 150-\( \mu\text{m} \) diameter spheres. In this study tissue was quantified using nuclei counts, which relate to the model density through
\[ \Gamma = \frac{\chi}{n_{\text{IE}}} \]
where \( \chi \) is the cell surface density (cells/cm\(^2\)) and \( n_{\text{IE}} \) is the number of cells in an IE (1560 cells/IE, [13]). Again exocrine tissue is assumed to be similar to islet tissue such that a volume of tissue equal to a 150-\( \mu\text{m} \) diameter sphere contains the same number of cells as an IE.
Table 4.1. Oxygen diffusivities and solubilities at 37°C and 24°C in culture medium and tissue used in the theoretical O₂ transport model.
| Property | Phase | Temperature | Value | Units | Reference |
|---------------------------|---------|-------------|----------------|-------------|-------------------------|
| Oxygen diffusivity, $D_m$ | Medium | 37 | 2.78 x 10⁻⁵ | cm²/s | [16] |
| | | 24 | 2.20 x 10⁻⁵ | cm²/s | Wilke - Chang equation [117] |
| Oxygen diffusivity, $D_t$ | Tissue | 37 | 1.22 x 10⁻⁵ | cm²/s | [16] |
| | | 24 | 9.65 x 10⁻⁶ | cm²/s | * |
| Oxygen solubility, $d_m$ | Medium | 37 | 1.27 x 10⁻⁹ | mol/cm³ mmHg| [16] |
| | | 24 | 1.71 x 10⁻⁹ | mol/cm³ mmHg| Value for water [118] |
| Oxygen solubility, $d_t$ | Tissue | 37 | 1.02 x 10⁻⁹ | mol/cm³ mmHg| [16] |
| | | 24 | 1.31 x 10⁻⁹ | mol/cm³ mmHg| * |
*The O₂ diffusivity in tissue at 24°C was estimated by multiplying the O₂ diffusivity in tissue at 37°C by the ratio of the diffusivities in medium at 24°C and 37°C. A similar calculation was conducted for O₂ solubility in tissue.*
Total tissue surface density can be converted to an OCR surface density, which is the total OCR of the tissue divided by the area of culture, using
\[
\lambda = \chi V_{\text{max}} \frac{V_{\text{IE}}}{n_{\text{IE}}} = \Gamma V_{\text{max}} V_{\text{IE}}
\]
where \( \lambda \) is the OCR density with units of flux and \( V_{\text{IE}} \) is the volume of an IE.
The model was solved with the finite element method using commercially available software (COMSOL Multiphysics, COMSOL Inc., Burlington, MA) in conjunction with Matlab (Mathworks, Natick, MA), as described previously (Chapter 3).
**Figure 4.1. Geometry assumed for simulations of islet culture on an O₂-permeable surface. Dashed lines denote unit cell.**
### 4.4 Results
Data from cultures carried out with elevated pO₂, at reduced temperature, and on silicone rubber membranes were compared to data from parallel cultures on polystyrene at 37°C and 142 mmHg in terms of the fraction of original cells collected from culture as measured by nuclei counting (Figure 4.3), tissue viability as quantified by OCR/cell (Figure 4.4), and the fraction of original viable tissue collected from culture \( F_V \), as measured by OCR (Figure 4.5), all after culture for 2 days. In addition, membrane integrity measurements were also made for reduced culture temperature experiments. The islet preparations used in this study varied in initial quality with an average OCR/cell of 1.6 ± 0.5 with a range of 0.6 to 2.5 fmol/min cell.
4.4.1 Culture at elevated ambient oxygen partial pressure
The fraction of original cells recovered after culture did not change consistently with initial tissue surface density (Figure 4.3A) and averaged $0.84 \pm 0.17$ and $0.82 \pm 0.16$ at $pO_2$ of 142 and 428 mmHg, respectively. OCR/cell and the fraction of the original OCR recovered after culture dropped sharply with increasing tissue density at 142 mmHg, slightly less so at 428 mmHg. Using estimates of 1560 cells/IE and about 50% islet cells in an average human islet preparation [55], 200 and 400 IE/cm$^2$ correspond to 0.625 and $12.5 \times 10^6$ cells/cm$^2$, respectively. The data in Figure 4.5A indicate a corresponding loss of about 10 to 50% of the original OCR from islet preparations cultured for 2 days at 37°C with these densities.
At very low culture density $F_V$ was significantly lower at 428 (about 0.6) than at 142 mmHg $O_2$ (about 1.0). Furthermore, low density culture at 570 mmHg $O_2$ for 48 hr resulted in even lower $F_V$ ($0.22 \pm 0.06$ from 570 mmHg culture versus $0.69 \pm 0.05$ from 142 mmHg for a paired experiment). It is likely these decreases in $F_V$ are caused by $O_2$ toxicity. To further explore this phenomenon, we cultured $\beta$TC3 cells at varying $pO_2$ over a 4-day period. Membrane integrity measurements are plotted versus time in culture in Figure 4.6. These data indicate $O_2$ has a toxic effect on these cells that is both dose and time dependent, and it is likely also toxic to islets.
4.4.2 Culture at reduced temperature
Measurements of rat islet OCR decreased exponentially with inverse temperature (Figure 4.2), resulting in an average activation energy from three experiments of $53 \pm 4$ kJ/mol, which is smaller than previously reported values for cell lines (79-93 kJ/mol) [115]. Assuming human islet tissue has a similar activation energy and the OCR at 37°C for human islet tissue is 2.5 fmol/min cell (Chapter 2) the following equation for $V_{\text{max}}$ as a function of temperature is obtained.
$$V_{\text{max}} = V_0 \cdot \exp\left[-\frac{E}{R_{\text{gas}} T}\right]$$
where $V_0$ is $1.86 \times 10^3$ mol/min cm$^3$ (adjusted to volume basis assuming 1560 nuclei/IE [13]), $E$ is 53 kJ/mol, and $R_{\text{gas}}$ is the gas constant (0.008314 kJ/mol K). This Arrhenius function will be used for $V_{\text{max}}$ when comparing experimental and theoretical results.
Figure 4.2. OCR as a function of inverse temperature from three rat islet preparations. Error bars are ± standard deviation for triplicate measurements.
The fractional recoveries of human islet tissue averaged $0.76 \pm 0.10$ and $0.83 \pm 0.13$ at 37 and 24°C, respectively, and there was no effect of tissue surface density (Figure 4.3B). The decrease in OCR/cell and fractional OCR recovery (Figure 4.4B and Figure 4.5B) with increasing surface density was less severe at 24°C, although the data is limited and scattered. At very low density, the values of these parameters were lower at 24 than at 37°C. Culture at 12°C at very low density gave similar OCR/cell and fractional OCR recoveries as those from 24°C culture (data not shown). Membrane integrity measurements (Figure 4.7) made in conjunction with this series of experiments demonstrated negligible change with tissue density at 24°C. There was a decrease at 37°C that was much less sharp than OCR/cell measurements.
4.4.3 Silicone rubber culture
Culture on a 500-μm silicone rubber membrane (with h = 22 mm) resulted in a fractional cell recovery of $0.80 \pm 0.12$ (Figure 4.3C), which was significantly lower than that for culture at the same medium depth on polystyrene, $0.93 \pm 0.06$ (p<0.01). OCR/cell and fractional recovery of OCR were substantially higher at all culture densities on a silicone rubber membrane.
compared to polystyrene culture at the same medium depth (Figure 4.4C and Figure 4.5C) as well as at the smaller depth, 3mm (Figure 4.5A and B). In addition, fractional OCR recovery was not affected when the silicone rubber membrane thickness was decreased from 500 to 100-\(\mu\)m.
### 4.4.4 Comparison with prediction of oxygen transport model
As mentioned previously, the partial pressure field in the islet calculated by the theoretical model was used to determine the viable tissue volume (\(V_V\)) by integrating the volume of tissue with \(pO_2 > p_{crit}\). The volume fraction of the viable tissue remaining after culture was calculated from
\[
F_V = \frac{V_V}{\frac{4}{3} \pi R_{islet}^3}
\]
The experimental data Figure 4.5 are compared to the predictions of the theoretical O\(_2\) transport model in Figure 4.8, in which fractional OCR recovery is plotted versus initial OCR density, so that in this way, all theoretical predictions collapse to a closely clustered curve for all values of \(V_{max}\), thereby permitting comparison of results from different experiments with one predicted curve (Chapter 3). OCR density was determined for each experiment through
\[
\lambda = \chi V_{max} \frac{V_{IE}}{n_{IE}}
\]
where \(\lambda\) is the OCR density with units of flux, \(\chi\) is the cell surface density (cells/cm\(^2\)) based on nuclei counts prior to culture, \(n_{IE}\) is the number of cells in an IE (1560 cells/IE, [13]), and \(V_{IE}\) is the volume of an IE. \(V_{max}\) was determined for each isolation prior to culture.
Figure 4.3. The total fraction of original cells recovered after culture with various culture techniques as a function of the initial tissue density.
Culture performed (A) on polystyrene at 37°C, 428 mmHg with h = 3 mm, (B) on polystyrene at 24°C, 154 mmHg with h = 3mm, and (C) on a 500-μm silicone rubber membrane (Si) at 37°C, 142 mmHg with h = 22 mm. Parallel cultures of islets from the same preparations on polystyrene (PS) at 142 mmHg, 37°C, and the same medium height are included when available.
Figure 4.4. OCR/cell following culture as a function of the initial tissue density for different culture techniques
Culture performed in similar format as Figure 4.1.
Figure 4.5. Fraction of original OCR recovered from culture as a function of surface density for different culture techniques.
Culture performed in similar format as Figure 4.1.
Figure 4.6. The fraction of βTC3 cells with intact cell membranes exposed to varying $O_2$ concentrations over 4 days of culture. Membrane integrity determined by Sytox Orange/LDS 751 staining. Error bars are ± standard deviation for measurements from 3-5 independent experiments.
The theoretical model predictions indicate that the fractional recovery of OCR should be improved by each of the three oxygen delivery enhancement approaches investigated. For both increased incubator pO$_2$ and reduced culture temperature, $F_V$ is predicted to be 1.0 up to some point and then decreases as initial OCR surface density increases. For cultures on silicone rubber membranes, $F_V$ is predicted to be 1.0 over the entire range of tissue density. For both increased pO$_2$ and reduced culture temperature, the model predictions are substantially higher than experimental data at all but the lowest densities. The same is true for culture on silicone rubber. However, the model underpredicts the experimental data for culture on polystyrene for the case of a large medium depth.
Figure 4.7. The fraction of tissue with intact cell membranes as a function of tissue surface density for culture at varying temperatures.
Culture performed on polystyrene in 3 mm medium at 37°C, 142 mmHg (open symbols) or 24°C, 154 mmHg (filled symbols). The differently shaped symbols represent different islet isolations.
4.5 Discussion
In this study we examined how tissue recovery and viability of islet preparations changed as a function of tissue surface density after culture for 2 days using three techniques designed to reduce O₂ supply limitations present during conventional static culture including culture under (1) elevated pO₂, (2) at reduced temperature, and (3) on a silicone rubber membrane. For each technique, parallel cultures were carried out using conventional culture (pO₂ of 142 mmHg, 37°C, untreated polystyrene surface). By applying a theoretical model of O₂ consumption and diffusion, we also explored how O₂ profiles change in each system, and we used the model to predict the recovery of viable tissue relative to viable tissue placed into culture. We hypothesized that improved oxygenation would increase the fraction of viable tissue remaining following high density culture.
Figure 4.8. Experimental results and theoretical predictions for the fraction of original viable tissue collected following culture for various culture techniques. Culture performed in similar format as Figure 4.1. The solid and dashed lines are predictions of the theoretical model, which was evaluated with $V_{\text{max}} = 2.5$ fmol/min cell = 36.8 nmol/s cm$^3$.
The fraction of original cells recovered after culture, as measured by nuclei counting, did not change with increasing islet tissue surface density in any of the experiments (Figure 4.3). It was not influenced by increased pO$_2$ or decreased culture temperature but there was a small consistent decrease in cell recovery when cultured on silicone rubber as compared to conventional culture.
With increasing tissue density, culture at 428 mmHg led to higher OCR/cell (Figure 4.4) and fractional recovery of OCR (Figure 4.5) as compared to conventional culture. Because the only difference between the culture conditions was pO$_2$, the higher recovery of viable tissue at elevated pO$_2$ is consistent with the notion that O$_2$ supply limitations play an important role in the decrease of F$_V$ at high culture densities during conventional culture (Chapter 3).
Culture at very low density at 24°C resulted in lower OCR/cell and F$_V$ than culture at 37°C. A similar trend was observed in a previous study (Chapter 2), which suggests that there is a mechanism for loss of viability at 24°C that is independent of oxygen supply limitations. Lower viability at 22°C compared to 37°C was also reported for islets cultured at 3 IE/cm$^2$ with 2 mm medium depth [49]. At tissue surface densities greater than about $10^6$ cells/cm$^2$, culture at 24°C resulted in higher OCR/cell and fractional OCR recovery than at 37°C (Figure 4.4B and Figure 4.5B), which is consistent with previous reports indicating higher necrotic death [50; 52] or lower recoveries [49] with 37°C culture than at reduce temperatures. The data at 24°C was highly scattered. One possible reason could be changes in cellular respiration continue to occur as the tissue acclimates to 37°C during the 1 hr incubation at 37°C prior to analysis. This short period may not be sufficient to allow all metabolic processes to reach a new steady state at 37°C.
Culture on silicone rubber membranes had the largest impact on tissue characteristics. OCR/cell (Figure 4.4C) and fractional OCR recovery (Figure 4.5C) were significantly higher than for culture on polystyrene. For example, F$_V$ for culture on polystyrene at densities up to about $6 \times 10^6$ cells/cm$^2$, corresponding to about 10 nmol/min cm$^2$ (Figure 4.8C) averaged 0.25 whereas culture on a 500-µm silicone rubber membrane averaged 0.81. Decreasing membrane thickness to 100-µm had no significant effect on F$_V$ indicating that the 500-µm membrane had effectively removed all of the oxygen supply limitations.
The sharp decrease with increasing tissue density in fractional OCR recovery for cells undergoing conventional culture, and to a lesser degree for cells cultured on silicone rubber and the other culture manipulations are indicative of large fractions of the cells having entered the
cell death process. By contrast, the recovery of nuclei did not change with increasing tissue density. Thus, exposure of islets to anoxia in this study caused cells to enter the cell death process, but after 2 days it had not progressed to the point where nuclei were fragmented. Membrane integrity was also independent of tissue density at 24°C and decreased with increasing tissue density at 37°C, but not nearly as much as OCR/cell. Thus, at 37°C the extent of cell membrane dysfunction lagged the reduction of OCR at 37°C, which is consistent with prior studies [55] and the cells did not reach this stage at all at 24°C.
The theoretical model correctly predicted no loss of OCR at low tissue density, which was consistent with experimental data, and the predictions were qualitatively correct in predicting a decrease with moderate to high tissue density and a greater recovery of OCR with techniques to reduce oxygen supply limitations. However, in every case the theory overestimated $F_V$ when compared to experimental data except for culture in 22 mm medium on polystyrene for which it underestimated $F_V$. The disparity was large for culture on silicone rubber at very high density for which the theory predicted there should be adequate oxygen supply. The discrepancies between theoretical predictions and experimental measurements for conventional culture have been discussed previously (Chapter 3), some of the same explanations hold for the culture techniques investigated here, including (1) inaccuracies in the model assumptions or parameter values, (2) diffusional limitations of nutrients (i.e. glucose) linked to the lack of medium change during culture, and (3) products produced by islet tissue (i.e. lactate, ammonia, nitric oxide) that could have toxic effects especially at high surface densities.
Among the possible inaccuracies in the assumptions for the model is the possibility that representation of islets as arranged in the orderly pattern of Figure 4.1 is unrealistic, e.g., islets may clump together locally to form a sheet-like structure or multi-layer structures. Development of models to investigate these various possibilities is described elsewhere (Chapter 5). For culture on silicone rubber, the first explanation is unlikely because there is excess oxygen supply capability as evidenced by no change in experimental data when membrane thickness was reduced from 500 to 100 μm. Application of the theoretical model to culture at 24°C is not straight forward because $F_V < 1$ for culture at very low density, indicating that even when no transport limitations are present not all of the viable tissue can be recovered (Chapter 2). The underestimation of $F_V$ for culture in 22 mm medium on polystyrene can be explained by (1) removal of additional diffusion limitations present at low medium depths and (2) changes in
gene expression. Islet β cells normally have a low level of lactate dehydrogenase (LDH) expression [119] but up-regulate LDH when exposed to a hypoxic environment for 24 hr (G. Weir, personal communication), which allows them to produce energy by glycolysis in the absence of O₂. If tissue in the islet preparation can up-regulate LDH fast enough, a larger fraction would be able to survive than predicted by the theoretical model.
The techniques examined improve tissue oxygenation compared to conventional culture, thereby allowing tissue to be cultured at higher densities without a reduction in viability. Theoretical and experimental results indicate that during conventional culture (142 mmHg O₂, 37°C, 3 mm) at OCR densities greater than 0.9 nmol/min cm² O₂ limitations occur causing a reduction in viability. Culture on polystyrene in 428 mmHg O₂ (37°C, 3 mm) supplies sufficient O₂ until 3 nmol/min cm², a 3-fold increase in density over conventional culture, but O₂ toxicity needs to be considered. A similar enhancement is obtained from culture on polystyrene at 24°C in 154 mmHg O₂ (h = 3 mm). However, reduced temperature culture may “precondition” tissue by changing the cellular machinery [51], thus bringing up questions about changes in tissue function. Brandhorst et al [49] concluded that 37°C culture of porcine islets resulted in superior insulin content, secretion, and dynamic response to glucose as compared to culture at 22°C which suggest that 37°C is better than reduced culture temperature in maintaining islet function. Though reduced temperature culture may be an improvement to conventional culture at high densities, it may not be the best choice. Culture on a 500-μm silicone rubber membrane removes all O₂ limitations for densities less than 15 nmol/min cm² though decreases in F_v are observed experimentally at densities greater than 5 nmol/min cm² due to O₂-independent mechanisms. To put this in perspective consider a preparation containing 400,000 IE that has an islet volume fraction of 50% (i.e. a total tissue volume equivalent to 800,000 IE) and a total tissue OCR of 1.9 μmol/min (accounting for both islet and non-islet tissue, 1560 cells/IE [13], 1.5 fmol/min cell Chapter 2). Using conventional culture with 3 mm medium, a culture area of about 2100 cm² (14 T-150 flasks) would be required to provide sufficient O₂ to prevent a decrease fractional OCR recovery while culture in 428 mmHg O₂ or at 24°C (154 mmHg O₂) would reduce this area to 700 cm² (5 T-150 flasks) and culture on a 500-μm silicone rubber membrane would require only 420 cm² (3 T-150 flask). This reduction in culture area would save space and time while providing sufficient O₂ to the tissue.
### 4.6 Nomenclature
| Symbol | Description |
|--------|-------------|
| 7-AAD | 7-aminoactinomycin D |
| C | Concentration of $O_2$ |
| D | Diffusivity of $O_2$ |
| d | Diameter of islet |
| D-PBS | Dulbecco’s phosphate buffered saline |
| E | Activation energy |
| FBS | Fetal bovine serum |
| $F_{n,a}$ | The fraction of original cells collected as adherent tissue after culture |
| $F_{n,na}$ | The fraction of original cells collected as non-adherent tissue after culture |
| $F_{n,a}$ | The fraction of original cells collected as the combined total (adherent + non-adherent) tissue after culture |
| $F_V$ | The fraction of original viable tissue after culture |
| h | Medium depth |
| IE | Islet equivalent |
| $K_m$ | Michaelis constant, the p$O_2$ at which the reaction rate is $V_{\text{max}}/2$ |
| LDH | Lactate dehydrogenase |
| nIE | Number of cells in an IE |
| $O_2$ | Oxygen |
| OCR | $O_2$ consumption rate |
| p | $O_2$ partial pressure |
| $p_{\text{crit}}$ | The p$O_2$ below which a cell dies and no longer consumes $O_2$ |
| PDMS | Polydimethylsiloxane |
| $P_{\text{gas}}$ | Ambient partial pressure of $O_2$ |
| R | Production rate of $O_2$ |
| $R_{\text{gas}}$ | Gas constant |
| $R_{\text{islet}}$ | Radius of islet |
| T | Total (non-adherent + adherent) tissue |
| t | Time |
| Symbol | Description |
|--------|-------------|
| $V_{ch}$ | Volume of OCR chamber |
| $V_{IE}$ | Volume of an IE |
| $V_{max}$ | Maximum reaction velocity |
| $V_{NV}$ | Non-viable tissue volume |
| $V_o$ | Constant in Arrhenius equation |
| $V_V$ | Viable tissue volume |
| $w$ | Center-to-center distance between islets |
| $\alpha$ | Solubility of $O_2$ in medium |
| $\delta$ | Thickness of silicone rubber membrane |
| $\Phi_I$ | Islet volume fraction of the islet preparation, |
| $\Gamma$ | Sphere surface density |
| $\lambda$ | OCR density with units of flux |
| $\rho$ | Islet surface density in IE/cm$^2$ |
| $\chi$ | Cell surface density (cells/cm$^2$) |
**Subscripts for D, $\alpha$, p, C**
- **m**: Medium
- **i**: Islet tissue
- **Si**: Silicone rubber
5 Oxygen transport during islet culture: A closer look
5.1 Introduction
Tissue surface density has a large impact on the fraction of original viable tissue ($F_V$) collected from static culture on polystyrene and on silicone rubber membranes (Chapter 3 and 4). As tissue surface density increases during culture on polystyrene, $F_V$ decreases due, at least in part, to oxygen ($O_2$) limitations (Chapter 3 and 4). However, though the empirical decrease in $F_V$ was qualitatively similar to theoretical predictions based on $O_2$ transport, there were consistent quantitative discrepancies, which may result from inaccuracies in the theoretical $O_2$ transport model.
The discrepancies could result from applying inaccurate parameter values to the theoretical $O_2$ transport model. Parameter values taken from literature can vary widely between studies resulting in significant uncertainty. If the parameter value chosen was inaccurate, it would introduce error into the prediction.
Discrepancies between empirical data and the theoretical predictions could also result from inaccurate assumptions made by the model. The model assumes that all tissue (islet and non-islet) is in 150-$\mu$m diameter spheres arranged in a square array. In reality the mixture of islet and exocrine tissue varies in shape, size, and spatial distribution, each of which alter predicted $O_2$ profiles during culture.
In this chapter we describe a sensitivity analysis conducted to determine which model parameters, in particular $K_m$, $p_{crit}$, $D_i$, $\alpha_i$, $D_m$, $\alpha_m$, $D_{Si}$, and $\alpha_{Si}$, have a large effect on theoretical predictions for culture on polystyrene and silicone rubber membranes. The confidence in the values chosen through comparisons with other reported values is discussed. We also examine the assumptions made within the model to determine the extent to which they may impact theoretical predictions specifically addressing islet surface density distribution, size, slab formation, tissue stacking, and convection.
5.2 Theoretical model
The theoretical models used to simulate islet culture on an $O_2$-impermeable dish and on an $O_2$-permeable dish have been presented in detail previously (Chapter 3 and 4). Both models
apply the diffusion-reaction equation to simulate $O_2$ transport for spheres of uniform size arranged in a square array in a medium of known depth on either an $O_2$-impermeable surface (Chapter 3) or an $O_2$-permeable surface of known thickness (Chapter 4). Oxygen consumption within the tissue was assumed to follow Michaelis-Menten kinetics.
The base case model used in these comparisons assumes culture at 37°C in a fully humidified environment with 95% air / 5% CO$_2$ (142 mmHg O$_2$) and a tissue $V_{\text{max}}$ of 53.3 nmol/s cm$^3$. This $V_{\text{max}}$ is intended to indicate a worst-case scenario and corresponds to 3.6 fmol/min cell, which is at the upper end of OCR/cell distributions (Chapter 2) and is higher than the $V_{\text{max}}$ used previously (Chapter 3 and 4). Since $F_V$ versus initial OCR density is only slightly dependent on $V_{\text{max}}$ (Chapter 3), the change does not have a large effect on the overall trend. The base case parameter values estimated from literature are given in Table 5.1. The base case predictions for culture on polystyrene in 3 mm and 22 mm and for culture in 22 mm medium on a 500-$\mu$m silicone rubber membrane are illustrated in Figure 5.1. Recall that the experimental $F_V$ was lower than predicted for culture in 3 mm medium on polystyrene (Chapter 3) and for culture in 22 mm medium on a 500-$\mu$m silicone rubber membrane (Chapter 4) while it was higher than predicted for culture in 22 mm medium on polystyrene (Chapter 4).
**Figure 5.1.** The base case $O_2$ transport model predictions for culture in 3 mm and 22 mm on polystyrene and in 22 mm medium on a 500-$\mu$m silicone rubber membrane.
Table 5.1. Base case parameter values used in simulations for O₂ transport during static islet culture.
| Parameter | Symbol | Value | Units | Reference |
|------------------------------------------------|--------|----------------|-----------|-----------|
| Michaelis constant | $K_m$ | 0.44 | mmHg | [98] |
| O₂ diffusivity in medium | $D_m$ | $2.78 \times 10^{-5}$ | cm²/s | [16] |
| O₂ diffusivity in tissue | $D_y$ | $1.22 \times 10^{-5}$ | cm²/s | [16] |
| O₂ diffusivity in silicone rubber | $D_{Si}$ | $2.17 \times 10^{-5}$ | cm²/s | [116] |
| O₂ solubility in medium | $d_m$ | $1.27 \times 10^{-9}$ | mol/cm³ mmHg | [16] |
| O₂ solubility in tissue | $d_y$ | $1.02 \times 10^{-9}$ | mol/cm³ mmHg | [16] |
| O₂ solubility in silicone rubber | $d_{Si}$ | $1.21 \times 10^{-9}$ | mol/cm³ mmHg | [116] |
| Maximal O₂ consumption rate | $V_{\text{max}}$ | 53.3 | nmol/s cm³ | Experimentally measured |
| Minimum pO₂ for viability | pcrit | 0.1 | mmHg | [16] |
| Media height | h | 3 or 22 | mm | Based on medium volume and well area |
| Membrane thickness | δ | 500 | μm | Based on equipment |
| External p | $p_{\text{gas}}$ | 142 | mmHg | Partial pressure of O₂ in humidified incubator at 37°C with 95% air/5% CO₂ |
| Islet radius | R_islet | 75 | μm | Based on definition of IE |
5.3 Inaccuracies in parameter values
This section explores the sensitivity of theoretical predictions based on $O_2$ transport to changes in model parameters ($K_m$, $p_{crit}$, $D_i$, $\alpha_i$, $D_m$, $\alpha_m$, $D_{Si}$, and $\alpha_{Si}$) providing insight into whether parameter inaccuracies may account for the discrepancies observed between the theoretical predictions and empirical data.
The Michaelis constant $K_m$, which is the partial pressure of $O_2$, $p$, at which the $O_2$ consumption rate is $V_{max}/2$, was 0.7 $\mu$M (0.44 mmHg) for the base case [98], which is consistent with $O_2$ consumption measurements in hepatocytes [120; 121]. However, data for $O_2$ consumption of cardiomyocytes at 5 $\mu$M concentrations suggests $K_m = 3.05$ $\mu$M (about 3 mmHg) [122]. Models for $O_2$ delivery to muscle tissue and energy metabolism in multicellular spheroids have used values ranging from 1 to 4.5 mmHg [123; 124]. Variation between species, tissue type, and temperature could contribute to reported $K_m$ values ranging from 0.25-4.5 mmHg [16; 73; 98; 120-126]. Increasing $K_m$ from 0.44 to 1 mmHg shifts $F_V$ upward while decreasing $K_m$ shifts $F_V$ downward (Figure 5.2). If the actual $K_m$ is higher than the base case as suggested by the range available in literature, then the predicted $F_V$ for culture on polystyrene shifts away from experimental data at low medium depths and towards experimental data at large medium depths. Little change would occur in $F_V$ for culture on a silicone rubber membrane since the tissue is at a higher $pO_2$ than tissue at the corresponding density on polystyrene.
$D$ and $\alpha$ are always present together in the steady state model equations (Chapter 3 and 4) so changes $D\alpha$, the permeability of $O_2$, will be considered. First, consider $O_2$ permeability in islet tissue $(D\alpha)_i$. Oxygen diffusivities for several different tissues range from 0.46-2.41 x $10^{-5}$ cm$^2$/s [16; 73; 123; 124; 127-135] while solubility has been reported from 0.94-1.53 x $10^{-9}$ mol/cm$^3$ mmHg [16; 73; 124; 129; 131; 132; 134-136] and $O_2$ permeability from 1.24-3.69 x $10^{-14}$ mol/cm mmHg s [16; 127-129; 131; 135; 137; 138]. Using the extremes of these reported values a range of 0.43 – 3.69 x $10^{-14}$ mol/cm mmHg s was determined for the permeability of $O_2$ in tissue. The base case value (1.24 x $10^{-14}$ mol/cm mmHg s [16]) was in the middle of this range. A 50% increase or decrease in $(D\alpha)_i$ does not result in a large change in
$F_V$ (Figure 5.3); as such, it is not likely that an inaccuracy in $(D\alpha)_i$ would cause large inaccuracies in $F_V$ for cultures on polystyrene or silicone rubber.
The base case value for $O_2$ permeability in medium, $(D\alpha)_m$, is $3.53 \times 10^{-14}$ mol/cm mmHg s [16]. The diffusivity of $O_2$ in 37°C aqueous solutions reported in literature ranges from $2.4-3.43 \times 10^{-5}$ cm$^2$/s [127; 133; 136; 139; 140] while the solubility ranges from $1.11-1.46 \times 10^{-9}$ mol/cm$^3$ mmHg [16; 73; 118; 125; 130; 140]. These values result in a fairly narrow permeability range of $2.66-5.01 \times 10^{-14}$ mol/cm mmHg s within which the base case value lies. A decrease in $(D\alpha)_m$ results in a significant downward shift in $F_V$ while an increase results in an upward shift (Figure 5.4). Thus $(D\alpha)_m$ lower than the base case value shifts theory closer to experimental $F_V$ for polystyrene culture at low medium depth, but away from experimental data at large medium depths. $(D\alpha)_m$ has little to no effect on $F_V$ for culture on silicone rubber since the majority of the $O_2$ supplied to the tissue is through the $O_2$-permeable membrane (data not shown).
**Figure 5.2.** Theoretical predictions for the fraction of original viable tissue after culture as a function of initial OCR surface density and $K_m$. Predictions assume islet culture in 3 mm medium on an $O_2$-impermeable dish for values of $K_m$ ranging from 0-1 mmHg.
Figure 5.3. Theoretical predictions for the fraction of viable tissue after culture as a function of initial OCR surface density for different values of \((D\alpha)_i\). Predictions assume islet culture in 3 mm medium on an O\(_2\)-impermeable dish.
Figure 5.4. Theoretical predictions for the fraction of viable tissue after culture as a function of initial OCR surface density for different values of \((D\alpha)_m\). Predictions assume islet culture in 3 mm medium on an O\(_2\)-impermeable dish.
Data for $p_{\text{crit}}$, the partial pressure of $O_2$ at which a cell dies, is scarce in comparison to the other parameters used in the model. Avgoustiniatios discussed the use of $p_{\text{crit}}$ and its estimation based on data in literature and concluded that 0.1-0.3 mmHg is the best estimate though he notes that other factors including nutrients present and tissue type may play a role in determining $p_{\text{crit}}$ [16]. $p_{\text{crit}}$ is unlikely to be higher than 1 mmHg since similar p$O_2$ are observed in vivo [16]. The base case value was 0.1 mmHg. As $p_{\text{crit}}$ increases $F_V$ decreases because a higher p$O_2$ is required for cell survival. Thus if 0.1 mmHg is low (i.e. $p_{\text{crit}} > 0.1$ mmHg), $F_V$ shifts downward toward empirical data for culture on polystyrene at low medium depths, but make predictions worse for large medium depths (Figure 5.5). For culture on a silicone rubber membrane $F_V$ decreases from 0.96 to 0.95 at an initial OCR density of 20 nmol/min cm$^2$ as $p_{\text{crit}}$ increases from 0.1 to 0.5 mmHg indicating a very small impact.
**Figure 5.5.** Theoretical predictions for the fraction of viable tissue after culture as a function of initial OCR surface density for different values of $p_{\text{crit}}$. Prediction assumes islet culture in 3 mm medium on an $O_2$-impermeable dish.
The final parameter taken from literature, $O_2$ permeability in silicone rubber $(D\alpha)_{Si}$, only affects simulations for culture on an $O_2$-permeable membrane. The permeability of $O_2$ in silicone rubber varies widely depending on the exact formulation with values ranging from $0.49 - 41.6 \times 10^{-14} \text{ mol/cm mmHg s at } 35^\circ C$ [141]. The membranes used in our experiments were polydimethylsiloxane (PDMS) for which a value of $26.3 \times 10^{-14} \text{ mol/cm mmHg s}$ [116] was used. Unfortunately not all articles investigating silicone rubber membranes note the exact formulation of the silicone rubber tested. The diffusivity of $O_2$ in silicone rubber ranges from about $1.1 - 2.22 \times 10^{-5} \text{ cm}^2/\text{s}$ [116; 140; 142; 143] for temperatures 35-37°C. However, the Polymer Data Handbook reports a diffusivity of $16 \times 10^{-5} \text{ cm}^2/\text{s}$ for $O_2$ in PDMS at 25°C [144], which would increase when adjusting to 37°C. Oxygen solubility in silicone rubber ranges from $1.1-20 \times 10^{-9} \text{ mol/cm}^3 \text{ mmHg}$ [116; 140; 142; 143]. The resulting range for $(D\alpha)_{Si}$ is $0.49-320 \times 10^{-14} \text{ mol/cm mmHg s}$ for silicone rubbers of various formulations. Limiting data to studies specifically using PDMS at 35-37°C reduces the range to $2.4-37 \times 10^{-14} \text{ mol/cm mmHg s}$ [141; 143-145] within which the base case value lies (near high end of range).
As $(D\alpha)_{Si}$ decreases the predicted $F_V$ decreases (Figure 5.6). Therefore if the base case value is high, as suggested by the range of values obtained from literature, then the theoretical predictions will become closer to the experimental data.
### 5.4 Inaccurate model assumptions
To efficiently simulate islet culture, we made several assumptions that simplified the model. All tissue, islet and non-islet, was arranged in a square array of uniform 150-$\mu$m diameter spheres in stagnant medium. Unfortunately this hardly represents reality. Human islets prior to culture (day 0) have a wide spatial and size distribution (Figure 5.7A and D) and at high densities may result in tissue stacking. To complicate things further, during high density culture, the islet distribution and morphology change, resulting in the formation of tissue slabs (Figure 5.7B, C, E, and F) rather than the individual islets that are obtained from low density cultures (Figure 5.7G and H). The model also assumes convection does not occur. In this section we explore the extent to which these phenomena could affect the theoretical prediction of $F_V$.
Figure 5.6. Theoretical predictions for the fraction of viable tissue after culture as a function of initial OCR surface density for different values of \((D\alpha)_{Si}\).
Predictions for culture on a 500-\(\mu\)m silicone rubber membrane in 22 mm medium.
5.4.1 OCR density distribution
To explore the impact of a non-uniform OCR surface density we assume that the OCR surface density distribution follows a normal distribution, which can be written as
\[
P(x) = \frac{1}{\sigma \sqrt{2\pi}} \exp \left( -\frac{(x - \mu)^2}{2\sigma^2} \right)
\]
where \(P(x)\) is the probability that the tissue is at the local OCR density \(x\), \(\mu\) is the average OCR density, and \(\sigma\) is the standard deviation of the OCR density. By applying this normal distribution function we assume that each islet at a given density is independent and identically distributed (i.e. the Central Limit Theorem holds). Though this assumption is not completely valid, the fact that we ignore the interaction between islets at different OCR densities is likely minor compared to the error in the assumed density distribution. As such, the adjustments to the prediction should not be considered an accurate representation for islets in culture, but rather an indication of how \(F_V\) would shift.
In order to apply the normal density distribution to the prediction the expected value of \(F_{V,adj}\) is calculated using
\[ F_{V,adj}(\mu) = \frac{\sum_x P(x) \cdot \Delta x \cdot F_V(x)}{\sum_x P(x) \cdot \Delta x} \]
(5.2)
where \( \Delta x \) is the width between our discrete bins (0.1 nmol/min cm\(^2\)), \( F_V(x) \) is the prediction of the model using a uniform OCR density, and \( F_{V,adj}(\mu) \) is the new prediction for density \( \mu \) that has been adjusted to account for the normal distribution. The predicted \( F_V \) as a function of uniform initial OCR surface density was broken into three functions: a linear function with a slope of zero at \( F_V = 1 \) for densities 0-0.6 mol/min cm\(^2\), a quadratic function from 0.6-5.1 nmol/min cm\(^2\), and a second quadratic function from 5.1-25 nmol/min cm\(^2\) (Figure 5.8, a maximum OCR density of 25 nmol/min cm\(^2\) was assumed for monolayer culture). The piecewise function for \( F_V \) and the normal distribution function in Equation 5.1 were used in Equation 5.2 to calculate \( F_{V,adj} \).
Adjusting for a normal OCR density distribution did not have a large effect on \( F_V \) (Figure 5.9) for culture on polystyrene in 3 mm medium. When adjusting for the normal distribution \( F_V \) decreases at low densities and increases at high densities as the distribution widens (Figure 5.9B and A respectively). Though this adjustment makes our simulations more realistic, it does not have a large impact on the prediction for any of the culture techniques studied.
### 5.4.2 Islet size
Tissue size varies widely from small cell clusters to large islets more than 250 \( \mu m \) in diameter (Figure 5.7). If we increase uniform islet size in the O\(_2\) transport model (> 150-\( \mu m \) diameter), \( F_V \) decreases substantially indicating that the size of the tissue clumps has a large impact on the O\(_2\) limitations imposed by static culture (Figure 5.10A). For culture on silicone rubber a similar trend is observed (Figure 5.10B) though the decrease in \( F_V \) is not substantial enough to bring the predictions close to the empirical results.
Figure 5.7. Light microscopy pictures of human islets.
Pictures taken on day 0 and after 39 hr culture at densities of 400, 150, and $8 \times 10^4$ cells/cm$^2$ on untreated polystyrene at 37°C.
The scale bar represents 100 $\mu$m. Magnification is the same in all panels.
Figure 5.8. Piecewise function for theoretical predictions of $F_V$ as a function of initial OCR surface density for a uniform density distribution (base case). The three functions include a linear function with a slope of zero at $F_V = 1$ for densities 0-0.6 mol/min cm$^2$, a quadratic function from 0.6-5.1 nmol/min cm$^2$, and a second quadratic function from 5.1-25 nmol/min cm$^2$.
Figure 5.9. Theoretical predictions using a uniform or a normally distributed OCR density. Data for $\sigma/\mu = 0.5$ and 1 for culture on polystyrene in 3 mm medium for initial OCR surface densities (A) 0 to 20 and (B) 0 to 5 nmol/min cm$^2$.
Human islet size distributions vary widely between isolations where some preparations have an average islet diameter significantly > 150 μm and some significantly < 150 μm (A. Omer, personal communication). As a result no single distribution can reflect the variation in islet size for all islet preparations. To investigate the effects of size distribution on $F_V$ we will use the normal distribution function in Equation 5.1 where $x$ is the local islet radius, $P(x)$ is the probability that the tissue has the radius $x$, $\mu$ is the average islet radius, and $\sigma$ is the standard deviation of the islet size distribution. Again, built within the assumption of a normal distribution is that each islet size is independent and identically distributed such that the Central Limit Theorem holds. In other words, an islet of a particular size does not depend on the size of its closest neighbors. For example, the prediction for an islet with a 75-μm radius at a particular OCR density should not depend on whether the islets around it have a radius of 25 or 125 μm. As a result $F_V$ for mixtures with different islet sizes should equal the volume weighted average of $F_V$ obtained using uniform sphere sizes if this independence holds.
To investigate whether islets of different size are independent of each other a model using spheres of two different sizes was developed. The equations used within the model were the same as previous (Chapter 3 and 4), but the model geometry changed. The geometry included spherical tissue clusters of two different sizes set in a square array (Figure 5.11A). $F_V$ for a mixture of islets with radii of 75 and 125-μm and for each individual islet size in the mixture are shown in Figure 5.11B. The overall $F_V$ obtained using the model with two different islet sizes was nearly identical to the volume weighted average $F_V$ obtained using predictions with spheres of uniform size (Figure 5.11C) indicating very little interaction between islets at the same OCR density. Since the predictions for a given islet size at a given OCR density are approximately independent of the surrounding tissue an adjusted $F_V$ can be calculated using a normal islet size distribution.
For simplicity mixtures of spheres with radii of 25, 50, 75, 100, 125, and 150-μm will be considered. The expected value of $F_{V,\text{adj}}$ is calculated using Equation 5.3.
$$F_{V,\text{adj}}(\mu) = \frac{\sum_x P(x) \cdot \Delta x \cdot w_x \cdot F_V(x)}{\sum_x P(x) \cdot \Delta x \cdot w_x}$$
5.3
where $\Delta x$ is the width between the discrete bins (25 $\mu$m), $F_V(x)$ is the prediction using uniform sphere size $x$, $w_x$ is equal to the volume of a sphere with radius $x$, and $F_{V,adj}(\mu)$ is the new prediction for the average islet radius $\mu$ adjusted to account for the size distribution. $F_V$ as a function of OCR surface density was broken into two to four separate linear or quadratic functions for each uniform islet size. These piecewise functions and the normal size distribution function were used in Equation 5.3 to calculate $F_{V,adj}$.
**Figure 5.10.** Theoretical predictions for islets with uniform radii of 25, 75, and 125 $\mu$m on (A) a polystyrene dish in 3 mm medium or (B) a 500-$\mu$m silicone rubber membrane in 22 mm medium.
Figure 5.11. Model geometry and predictions assuming a mixture of islet sizes.
(A) Geometry used to model islets of two different sizes. The dashed lines indicate the portion actually modeled. (B) Theoretical predictions for a 50:50 mixture of spheres with radii of 75 and 125 μm. $F_V$ from the simulation of the mixture is given for each islet size within the mixture and the overall combined tissue. (C) Comparison between overall model predictions for a 50:50 mixture of spheres with radii of 75 and 125 μm to that of the volume weighted average $F_V$ from uniform islet size predictions.
Figure 5.12. Theoretical predictions for $F_V$ assuming a normal islet size distribution. Prediction for culture in 3 mm medium on a polystyrene dish. The average islet size $\mu$ varied from 50 – 100 $\mu$m. Results are shown for both $\sigma/\mu$ equal to (A) 0.5 and (B) 1.
Theoretical predictions taking into account a normal islet size distribution with $\mu$ equal to 50, 75, and 100 $\mu$m and $\sigma/\mu$ equal to 0.5 and 1 for culture in 3 mm medium on polystyrene are given in Figure 5.12. As $\mu$ and $\sigma/\mu$ increase, $F_V$ decreases indicating that not only the average size, but also distribution width can cause departure from predictions using uniform islet size. Further, a distribution with $\mu = 75 \mu$m results in lower $F_V$ than predictions with a uniform islet radius of 75 $\mu$m indicating that just by accounting for a size distribution predictions shift downward (Figure 5.10A and Figure 5.12).
5.4.3 Tissue slab
Tissue clumped during high density culture resulting in slab-like regions (Figure 5.7). A model using the same theoretical equations discussed previously (Chapter 3 and 4) was developed to model one-dimensional O$_2$ transport using the geometry illustrated in Figure 5.13A. Slab formation causes severe limitations resulting in large drops in $F_V$ during culture on polystyrene and for large slab thicknesses on silicone rubber membranes (Figure 5.13B). These predictions act as a lower bound on $F_V$ indicating a worst-case scenario. All data for culture on polystyrene (3 and 22 mm medium) are above this worst-case prediction. However, a 150-$\mu$m
thick slab of tissue on a silicone rubber membrane should result in only a 0.03 decrease in $F_V$, which is not below empirical data (Chapter 4).
5.4.4 Islet stacking
A theoretical model using equations described previously (Chapter 3 and 4) was developed using the geometry indicated in Figure 5.14A to determine the extent to which islet stacking affects $F_V$. Stacking causes a large decrease in $F_V$ for culture on a 500-$\mu$m silicone rubber membrane each time a new layer is added (Figure 5.14B). The impact for culture on polystyrene is less significant because the majority of the tissue is already dead with monolayer coverage. In a similar manner as that for slab formation, these predictions can be used as a lower bound on $F_V$. For culture on polystyrene (3 and 22 mm medium) islet stacking of just 2 layers is below all empirical data (Chapter 3 and 4) (data not shown for 22 mm medium depth). Culture on a 500-$\mu$m silicone rubber membrane in 22 mm medium results in $F_V$ of roughly 0.50 for a bilayer of islets (Figure 5.14B) above which nearly all of the experimental data falls (Chapter 4). Further, only a fraction of the tissue will be stacked with the remainder in some monolayer distribution resulting in some combination of the two predictions that decreases with increasing density, which is consistent with the empirical trends observed for culture on polystyrene and silicone rubber (Chapter 3 and 4).
5.4.5 Convection
Convection complicates the theoretical model significantly since there is much uncertainty in the exact form and extent of the convective transport. As such we did not explore convection using a theoretical model, but instead tried to bound the effect experimentally. We investigated a oxygen biosensor system (OBS) developed for the measurement of OCR of tissue in static culture [146] (see Appendix 10.3 for details). The OBS utilizes an O$_2$ sensitive fluorophor embedded in silicone rubber upon which cells or tissue is placed. As the tissue consumes O$_2$, the O$_2$ partial pressure at the bottom of the plate decreases and is detected by the sensor. While investigating this technique we showed that gelled culture medium containing 0.9% agarose resulted in an average 15% increase in the apparent OCR/cell determined by the system as compared to the same with liquid medium (Appendix 10.3). This increase in apparent OCR/cell occurred because convection was removed from the system with gelled medium,
thereby decreasing the $O_2$ flux to the tissue approximately 15% (1-1/1.15) or < 0.1 nmol/min cm$^2$ and decreasing the $O_2$ partial pressure at the bottom of the dish (lower $O_2$ partial pressure at bottom means higher apparent OCR/cell).
Equating this flux (0.1 nmol/min cm$^2$) to the consumption rate of tissue indicates that an additional tissue volume equivalent to 18 IE/cm$^2$ could be provided with sufficient $O_2$ (assumes zero-order kinetics, 150-$\mu$m diameter islets, ignores transport limitations). At low densities this additional viable tissue could have a significant effect on $F_V$ whereas at high densities it is not likely to have a large effect. For example, at a density of 200 IE/cm$^2$ where $F_V$ from the base case is roughly 0.9 (20 IE/cm$^2$ not viable, 180 IE/cm$^2$ viable), the additional flux shifts $F_V$ to about 0.99 (20-18 = 2 IE/cm$^2$ not viable, 180+18 = 198 viable).
However, at a density of 2000 IE/cm$^2$, where $F_V$ is roughly 0.25 (1500 IE/cm$^2$ not viable, 500 IE/cm$^2$ viable) in the base case, the additional flux would shift $F_V$ less than 0.01 (1500-18 = 1488 IE/cm$^2$ not viable, 500+18 = 518 viable, $F_V = 0.259$). In addition, the convective effect observed in the OBS study is likely a gross overestimation of convection present during culture because the convection imposed by transferring dishes from the incubator to a plate reader and within the plate reader is likely dominant compared to vibrations from the incubator (Appendix 10.3). Still, the data allow us to limit the extent to which convection could increase $F_V$.
5.5 Discussion
In this study we explored the sensitivity of theoretical predictions based on a simplified $O_2$ transport model discussed in Chapters 3 and 4 to changes in the model parameters and assumptions for culture at low and high medium depths on polystyrene and at large medium depths on silicone rubber. This sensitivity analysis provides insight into why theoretical predictions only have qualitative agreement, and not quantitative agreement, with empirical data. Additionally, a better understanding of how each parameter and assumption affect the theoretical prediction could lead to improved methods for islet culture.
Figure 5.13. Model geometry and predictions assuming a slab of tissue.
(A) Geometry used to simulate O₂ transport in a slab of tissue with varying thickness.
(B) Theoretical prediction for $F_V$ as a function of tissue slab thickness for culture on polystyrene in 3 mm medium and for culture on a 500-μm silicone rubber membrane in 22 mm medium.
Figure 5.14. Model geometry and predictions assuming stacked islets.
(A) Geometry used to simulate $O_2$ transport in stacked spheres of tissue. (B) Theoretical prediction for $F_V$ as a function of the number of sphere layers for culture on polystyrene in 3 mm medium and for culture on a 500-$\mu$m silicone rubber membrane in 22 mm medium.
Theoretical predictions for the simplified $O_2$ transport model overestimated the empirical $F_V$ for culture on polystyrene in 3 mm medium (Chapter 3). Since decreases in $F_V$ during 3 mm culture on polystyrene have been shown to be at least partially due to $O_2$ limitations (Chapter 4), it is reasonable to conclude that inaccuracies in the $O_2$ transport model parameters and/or assumptions could account for the discrepancies observed between theoretical predictions and experimental measurements. The theoretical predictions shift closer to empirical data if $K_m$ or $(D\alpha)_m$ are lower than in the base case model, however this is unlikely based on values reported in literature. Inaccuracies in $(D\alpha)_i$ and adjusting for a normal OCR density distribution (rather than a uniform density) do not have any significant effect on the prediction and convection.
would only cause a further departure from the empirical data. Theoretical $F_V$ shifts toward empirical data if adjusting the model for larger islet size, slab formation, or tissue stacking, which have been observed during culture, or if $p_{\text{crit}}$ is > 0.1 mmHg, which is possible based on the discussion of Avgoustiniatos [16], indicating four inaccuracies that could explain the discrepancies between theory and experiments.
For culture on polystyrene in 22 mm medium theoretical predictions based on the simplified model underestimated empirical $F_V$ (Chapter 4). As a result an upward shift in $F_V$ is toward empirical data, which is the opposite of 3 mm culture on polystyrene. Thus an underestimation of $K_m$ or $(D\alpha)_m$ by the base case, which is likely when considering the values available in literature, or an overestimation of $p_{\text{crit}}$, which is not likely based on literature, results in a shift toward empirical data. As with culture on polystyrene in 3 mm medium inaccuracies in $(D\alpha)_i$ and adjusting for an OCR density distribution have little effect on the prediction. Though large islets, slab formation, and tissue stacking have been observed during culture, accounting for them would shift the prediction downward away from empirical data whereas accounting for convection now shifts the predictions closer to the experimental data. Thus inaccuracies in $K_m$ and $(D\alpha)_m$ or ignoring convection could explain why theoretical predictions using the simplified model underestimated experimental $F_V$ for culture on polystyrene in 22 mm medium.
Since $O_2$ has been shown to be limiting in 3 mm medium on polystyrene (Chapter 4), it is also very likely limiting in 22 mm medium. However, any adjustments made to the theoretical model that would bring the prediction for culture on polystyrene in 22 mm medium closer to empirical data would result in larger discrepancies during culture in 3 mm medium and vice versa. This suggests that inaccuracies in the $O_2$ transport model are not the only cause of the discrepancies between theory and experiments. Islet $\beta$ cells normally have a low level of lactate dehydrogenase (LDH) expression [119], but up-regulate LDH when exposed to a hypoxic environment for 24 hr (G. Weir, personal communication), which allows them to produce energy by glycolysis in the absence of $O_2$. If tissue in the islet preparation can up-regulate LDH fast enough, a larger fraction would be able to survive than predicted by the theoretical model. Thus the discrepancies observed at large medium depths on polystyrene are likely to result from this change in cellular machinery and not inaccuracies in the basic $O_2$ transport model.
Most inaccuracies in the $O_2$ transport model did not affect predictions for culture on a 500-$\mu$m silicone rubber membrane ($h = 22$ mm) to the same extent as culture on polystyrene. Inaccuracies in $(D\alpha)_m$ become irrelevant because the majority of $O_2$ is transported to the tissue through the membrane. $K_m$ and $p_{crit}$ are of less importance since the $pO_2$ within the tissue is significantly higher than tissue on polystyrene thereby diminishing the effect of Michaelis-Menten kinetics. As with culture on polystyrene, variations in $(D\alpha)_i$ and accounting for a normal density distribution have little impact on the prediction while accounting for convection would only cause farther departure from empirical data. In addition, though islet size variation and slab formation have been observed, accounting for these phenomena individually does not result in a large enough shift in $F_V$ to explain the discrepancies between experiments and theory.
On the other hand, if $(D\alpha)_{Si}$ is overestimated in the base case, which is suggested by the range in values obtained from literature, then the predicted $F_V$ shifts toward experimental data. Additionally, tissue stacking in a bilayer can drop $F_V$ to 0.50, above which nearly all of the experimental data for silicone rubber falls (Chapter 4). Thus inaccuracies in $(D\alpha)_{Si}$ and the presence of tissue stacking are both reasonable explanations for the discrepancies observed between theoretical and experimental data. However, despite this, they are unlikely the cause of the discrepancies. As mentioned previously, the empirical decrease in $F_V$ observed following culture on a 500-$\mu$m silicone rubber membrane is unlikely to result due to $O_2$ limitations since empirical measurements from cultures on membranes of different thickness were not significantly different (Chapter 4). Thus $O_2$ is not limiting and inaccuracies in the $O_2$ transport model cannot explain why there is a departure from theory.
In summary this chapter has illustrated that inaccuracy in model assumptions and/or parameters does have a significant impact on the theoretical predictions of our simplified $O_2$ transport model, which could explain some of the discrepancies observed between theory and experiments for culture on polystyrene at either low medium depths or high medium depths, but not both. Though the inaccuracies also provide valuable insight for culture on silicone rubber, they are not likely the cause of the departure from experimental $F_V$.
### 5.6 Nomenclature
| Symbol | Description |
|--------|-------------|
| D | Diffusivity of $O_2$ |
| $F_V$ | The fraction of original viable tissue after culture |
| $F_{V,adj}$ | $F_V$ adjust to account for non-idealities |
| h | Medium depth |
| IE | Islet equivalent |
| $K_m$ | Michaelis constant, the $pO_2$ at which the reaction rate is $V_{max}/2$ |
| LDH | Lactate dehydrogenase |
| $O_2$ | Oxygen |
| OCR | $O_2$ consumption rate |
| p | $O_2$ partial pressure |
| $P_{crit}$ | The $pO_2$ below which a cell dies and no longer consumes $O_2$ |
| $P_{gas}$ | Ambient partial pressure of $O_2$ |
| $P(x)$ | Probability of getting result x |
| $R_{islet}$ | Radius of islet |
| $V_{max}$ | Maximum reaction velocity |
| $w_x$ | Volume of a sphere with radius x |
| x | Variable of interest |
| $\alpha$ | Solubility of $O_2$ in medium |
| $\delta$ | Thickness of silicone rubber membrane |
| $\sigma$ | Standard deviation the values of x |
| $\mu$ | Average value for all x |
| $\Delta x$ | Width between discrete bins |
Subscripts for D, $\alpha$, p, C
- i: Islet tissue
- m: Medium
- Si: Silicone rubber
6 Glucose transport during islet culture
6.1 Introduction
The fraction of original viable tissue collected from static culture ($F_V$) decreases as tissue surface density increases on polystyrene and on silicone rubber membranes (Chapter 3 and 4). For culture on polystyrene this dependence was due at least in part to oxygen ($O_2$) limitations (Chapter 3 and 4). For culture on a silicone rubber membrane, theoretical predictions based on $O_2$ transport and empirical data using membranes of different thickness determined $O_2$ was not limiting during culture (Chapter 4). The decrease in $F_V$ on silicone rubber membranes may result from nutrient transport limitations other than $O_2$. Nutrient depletion may have occurred since medium was not changed during 34-60 hr culture. Culture in a low medium depth, such with culture on polystyrene in elevated $O_2$, will deplete the bulk nutrient supply faster than culture with a large medium depth. However, even when a large medium depth is used, diffusional limitations may prevent nutrients from reaching the tissue at a sufficient rate thereby causing a nutrient shortage even though the bulk concentration is not depleted.
$\beta$ cell survival during culture has been shown to be dependent on the glucose concentration of the culture medium [147-150]. Specifically, rat $\beta$-cell survival increases with glucose concentration in a dose-dependent manner up to 10 mM glucose through a reduction in $\beta$-cell apoptosis [147; 149]. Further, culture in zero glucose results in nearly a 50% reduction in rat islet ATP concentration within just two hours [151]. Prolonged exposure to zero glucose is thus unlikely to allow sufficient production of ATP to maintain cell function or viability.
A transient model that predicts glucose concentration profiles during culture was developed to investigate whether glucose transport limitations could occur during high density culture. In this study we explore the theoretical glucose transport model and with preliminary experimental data discuss the likelihood of whether glucose transport could cause decreases in islet viability during high density culture.
6.2 Methods
Islet isolation and culture
Standard collagenase/protease digestion methods were used for human [7; 15] isolations. Preparations of human islets were provided fresh (4-6 hr after isolation was complete) by the Islet Core at the Joslin Diabetes Center (JDC, Boston, MA). Human islets were also received after culture (1-2 days) from JDC or overnight shipment (sometimes after up to 1 day culture) from other centers contained in culture flasks (T-25 and T-75) or centrifuge tubes in Styrofoam boxes. When first observed, some preparations were in suspension, some in a pellet. Upon receipt, a portion of the preparation was analyzed and the remainder placed into untreated polystyrene culture flasks at 37°C in a humidified environment (5% CO₂), oxygen partial pressure (pO₂) of 142 mmHg, for 34-60 hr during which no medium change occurred. Some cultures were terminated at 18-24 hr for analysis. To test the effect of temperature, some islets were cultured at 24°C, then warmed for 1 hr at 37°C before OCR measurements.
Islets were cultured in supplemented RPMI (11.1 mM D-glucose, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate (all from Mediatech Inc., Herdon, VA), and 50 μM β-mercaptoethanol (Sigma Aldrich, St. Louis, MO) with 10% fetal bovine serum (FBS, Mediatech Inc.), at low densities (< 8x10⁴ cells/cm² or about 50 islet equivalents (IE)/cm², where IE is the volume of a sphere with a diameter of 150 μm) in 1.3 mm of medium in untreated polystyrene culture flasks. For some experiments parallel cultures in 0 mM glucose RPMI were conducted. After culture the free, non-adherent tissue (which is the only tissue used in current practice) was collected. In selected experiments, adherent human tissue was incubated for 5 min at 37°C in 0.05% trypsin EDTA (Mediatech Inc.), dislodged with fresh medium, and collected.
Islet enumeration by nuclei counting
Nuclei were prepared by adding equal 100-μl volumes of sample containing 160 or more IE and of lysis solution (0.1 M citric acid (Sigma) and 1% (v/v) Triton X-100 (Sigma)) to a 1.5 ml microtube [55]. The mixture was incubated at room temperature for 5 min with vortex mixing every 1.5 min, and nuclei were liberated by shearing through a needle. Isolated nuclei were diluted with Dulbecco’s phosphate buffered saline (D-PBS, Invitrogen, Carlsbad, CA) to a concentration no higher than 5 x 10⁵ nuclei/ml, stained with 7-aminoactinomycin D (7-AAD,
Molecular Probes, Eugene, OR), and analyzed using a flow cytometer (Guava Personal Cell Analysis (PCA) system, Guava Technologies, Hayward, CA) to determine the total number of cells with intact nuclei in the sample.
**Oxygen consumption rate (OCR)**
OCR was measured as previously described [39]. Briefly, suspensions containing about 2500 IE/ml in DMEM containing 4.5 g/l glucose and 0.6 g/l L-glutamine supplemented with 100 U/ml penicillin, 100 µg/ml streptomycin, 10 mM HEPES, and no added serum were sealed in a 200-µl stirred titanium chamber (Micro Oxygen Uptake System, FO/SYSZ-P250, Instech Laboratories, Plymouth Meeting, PA) maintained at 37°C. The time dependent pO$_2$ within the chamber was recorded with a fluorescence-based oxygen sensor (Ocean Optics, Dunedin, FL), and data at high pO$_2$ were fit to a straight line. The maximal OCR was evaluated from $OCR = V_{ch} \alpha (\Delta pO_2 / \Delta t)$, where $V_{ch}$ is the chamber volume and $\alpha = 1.27$ nmol/mm·Hg·ml is the Bunsen solubility coefficient for oxygen in medium [16]. OCR measurements were normalized by the measured number of cells (nuclei counting) or DNA content of the sample examined.
**Membrane Integrity**
**Statistics**
Measurements were made with three or more replicates and reported as mean ± SD. Statistical significance was determined using a Student t-test for unpaired data for comparing population means, and for paired data when appropriate.
**Glucose quantification**
Samples were taken of the original culture medium and spent medium following culture and analyzed for glucose content using the Glucose Assay Kit (GAGO20, Sigma Aldrich, St. Louis, MO). Triplicate samples diluted to concentrations between 20-100 µg/ml were placed in a 96-well plate along with samples for a six-point standard curve ranging in concentration from 0-100 µg/ml. The Glucose Assay Reagent was added to each well. After 25 min at 37°C 12N sulfuric acid was added to each well. The absorbance was measured at 544 nm on a plate reader.
6.3 Theoretical glucose transport model
The species conservation equation for glucose in the absence of convection [16; 97] is given by
\[
\frac{\partial C}{\partial t} = D \nabla^2 C + R \tag{6.1}
\]
where \(C\) is the concentration of glucose, \(D\) is the diffusivity of glucose, and \(R\) is the production rate of glucose.
Glucose consumption within the islet is assumed to follow Michaelis-Menten kinetics [16] for which the reaction rate is expressed as
\[
R = -\frac{V_{\text{max}} C}{K_m + C} \tag{6.2}
\]
where \(V_{\text{max}}\) is the maximum reaction velocity, and the Michaelis constant \(K_m\) is the \(C\) at which the reaction rate is \(V_{\text{max}}/2\). The species conservation equation in medium (m) and in the islet tissue (i) become, respectively,
\[
\frac{\partial C}{\partial t} = D_m \nabla^2 C \tag{6.3}
\]
\[
\frac{\partial C}{\partial t} = D_i \nabla^2 C - \frac{V_{\text{max}} C}{K_m + C} \tag{6.4}
\]
To model islet culture in a dish, we assumed all the tissue (islet and non-islet) was in the form of spheres with uniform diameter (\(d\)) that were arranged in a square array with center-to-center distance \(w\) in a stagnant medium of height \(h\) (Figure 6.1). Because of symmetry within the system, Equation 6.1 was solved within a unit cell containing only an eighth of an islet (dashed lines in Figure 6.1). The initial and boundary conditions for this model [16] include uniform glucose concentration throughout system,
at \(t = 0\),
\[
C_t=0 = C_{\text{initial}} \tag{6.5}
\]
zero flux at the air-liquid interface and at the bottom of dish,
at $z = 0$ and $z = h$,
$$\frac{\partial C}{\partial z} = 0 \quad 6.6$$
continuity of glucose flux and concentration across the liquid-islet,
$$n \cdot ((-\text{DVC})_i - (-\text{DVC})_m) = 0 \quad 6.7$$
$$C_m = C_i \quad 6.8$$
and symmetry across each side of the unit cell including both medium and islet domains
$$n \cdot (-\text{DVC}) = 0 \quad 6.9$$
The governing equations were non-dimensionalized using the non-dimensional variables
$$\eta = \frac{C}{C_{\text{initial}}} \quad 6.10$$
$$X = \frac{x}{R_{\text{islet}}}, \quad Y = \frac{y}{R_{\text{islet}}}, \quad Z = \frac{z}{R_{\text{islet}}} \quad 6.11$$
$$\tau = \frac{t(D)_i}{R_{\text{islet}}^2} \quad 6.12$$
Substituting these variables into 6.4 and 6.5 results in
$$\frac{\partial \eta}{\partial \tau} = \frac{D_m}{D_i} \nabla^2 \eta \quad 6.13$$
$$\frac{\partial \eta}{\partial \tau} = \nabla^2 \eta - \frac{V_{\text{max}} R_{\text{islet}}^2}{D_i C_{\text{initial}}} \left( \frac{\eta}{K_m + \eta} \right) \quad 6.14$$
Figure 6.1. Islet geometry assumed for modeling glucose transport during islet culture on either a polystyrene dish or a silicone rubber membrane.
The initial condition becomes
\[ \eta_{\tau=0} = 1 \]
6.15
The boundary conditions become
at \( Z = 0 \) and \( Z = H \) (where \( H = h/R_{\text{islet}} \)),
\[ \frac{\partial \eta}{\partial Z} = 0 \]
6.16
\[ n \cdot ((-\frac{D_i}{D_m} \nabla \eta)_i - (-\nabla \eta)_m) = 0 \]
6.17
\[ \eta_m = \eta_i \]
6.18
\[ n \cdot (-DVC) = 0 \]
6.19
There are three key non-dimensional groups that determine glucose concentration profiles during static islet culture. The first is the Damkohler number (\( Da \)), which is a measure of the rate of reaction relative to the rate of diffusion, defined in this case as
\[ Da = \frac{V_{\text{max}} R_{\text{islet}}^2}{C_{\text{initial}} D_i} \]
6.20
The other non-dimensional parameters are $D_m/D_i$, a relative diffusivity, and $K_m/C_{\text{initial}}$, the non-dimensional Michaelis constant. Though this model was described for glucose transport, it could be applied to any nutrient in the medium that is consumed by the islet tissue and not replenished continuously.
Table 6.1 contains the values for glucose consumption rates at high glucose concentrations (>15mM) available in literature. $V_{\text{max}}$ was also determined experimentally as discussed below. Table 6.2 contains data for all other parameters used in the model. The estimate for glucose diffusivity in islet tissue is similar to that for diffusivities in human cell spheroids of $2.3-5.5 \times 10^{-7} \text{ cm}^2/\text{s}$ [152].
The model was solved with the finite element method using the commercially available software COMSOL Multiphysics (COMCOL Inc., Burlington, MA) in conjunction with Matlab (Mathworks, Natick, MA). We required the mesh to contain more than 2500 nodes and set an absolute tolerance for glucose concentration in the time dependent solver of $10^{-4}$ [99].
### 6.4 Glucose consumption rate measurement
Glucose consumption rate (GCR) was determined by measuring the glucose concentration of culture medium before and after culture, which was converted to a decrease in total glucose and normalized using nuclei counts after culture. The GCR was $4.07 \pm 1.90 \times 10^{-5} \mu g/\text{cell hr}$ ($n = 12$) or $5.55 \pm 2.59 \times 10^{-8} \text{ mol/s cm}^3$ (assuming 1560 cells/IE [13]) and the distribution is given in Figure 6.2. Adjusting the measured GCR for a $K_m$ of 8 mM (and an average glucose concentration) resulting in an empirical $V_{\text{max}}$ of about $1.0 \times 10^{-7} \text{ mol/s cm}^3$.
Table 6.1. Literature values for glucose consumption by islet tissue.
| Species | Literature value | Original units | Converted value | Adjusted for $K_m$ | Converted units | Reference |
|---------------|---------------------------|---------------------------------|-----------------|--------------------|-----------------|-----------|
| Mouse islets | 0.48 ± 0.14 | pmol/min nL islet | 8.0 x 10^{-9} | 12 x 10^{-9} | mol/s cm^{-3} | [153] |
| Mouse islets | 2626 ± 159 | pmol/h ug DNA | 4.19 x 10^{-9} | 5.9 x 10^{-9} | mol/s cm^{-3} | [154] |
| Rat islets | 2612 ± 146 | pmol/h ug DNA | 4.17 x 10^{-9} | 5.8 x 10^{-9} | mol/s cm^{-3} | [154] |
| Rat islets | 102 ± 24 | umol/h per g dry weight | 4.9 x 10^{-9} | 7.2 x 10^{-9} | mol/s cm^{-3} | [155] |
| Rat islets | 1.53 ± 0.13 | pmol/ng DNA h | 2.4 x 10^{-9} | 3.5 x 10^{-9} | mol/s cm^{-3} | [156] |
| Hamster islets| 115.9 ± 2.93 | pmol/islet 2 hr | 9.1 x 10^{-9} | 13 x 10^{-9} | mol/s cm^{-3} | [157] |
$V_{\text{max}}$ values were determined at high glucose concentrations (>15 mM) to minimize the effect of $K_m$ on the measured rate.
Glucokinase acts as the limiting enzyme in glucose phosphorylation regulating the rate of glycolysis and it has a $K_m$ between 5-11 mM (we will use 8 mM for the base case) [153; 158; 159]. The original units were converted to common units (mol/s cm$^3$) using 1560 cells/IE [13], 1.77 x 10^{-5} cm$^3$/IE (Volume of a sphere with 150-\mu m diameter), 6.5 pg DNA/cell [68], and 33.2 ng DNA/\mu g dry weight [155]. The converted GCR were adjusted to account for the $K_m$ of glucokinase using 6.2 and assuming an average glucose concentration.
Table 6.2. Values for the parameters used in islet culture simulations and the source from which they were obtained.
| Parameter | Symbol | Value | Units | Reference |
|------------------------------------------------|--------|----------------|---------|-----------|
| Diffusivity of glucose in medium | $D_m$ | $6.7 \times 10^{-6}$ | cm$^2$/s | [118] |
| Diffusivity of glucose in tissue | $D_t$ | $3.8 \times 10^{-7}$ | cm$^2$/s | [153] |
| Michaelis constant | $K_m$ | 5 – 11 | MM | [153, 158, 159] |
| Maximal glucose consumption rate | $V_{\text{max}}$ | $5.8-13 \times 10^{-9}$ | mol/s cm$^3$ | Table 6.1 [153-155] |
| Islet radius | $R_{\text{islet}}$ | 75 | Um | Based on definition of IE |
| Medium depth | $h$ | 3 or 22 | Mm | Based on medium volume and culture area |
| Initial concentration of glucose | $C_{\text{initial}}$ | 10 | MM | Based on medium composition |
The estimate for glucose diffusivity in islet tissue is similar to that for human cell spheroids of $2.3-5.5 \times 10^{-7}$ cm$^2$/s [152].
Figure 6.2. The distribution of GCR measurements from human islet preparations.
6.5 Predictions of theoretical models
Transient glucose profiles in the islet and medium change drastically during high density culture (4440 IE/cm$^2$, approximately a monolayer coverage, Figure 6.3). Bulk glucose is depleted from culture in 3 mm medium independent of which Da, and thus which GCR (for empirical GCR, Da = 1.48 and for literature GCR, Da = 0.14), is accurate. If Da = 1.48, glucose is depleted to < 0.5 mM in < 10 hr (Figure 6.3C) whereas Da = 0.14 requires 45-50 hr to reduce the bulk concentration to a similar level (Figure 6.3A). Bulk glucose is not depleted during 60 hr culture in 22 mm medium independent of the Da (Figure 6.3B and D). However, despite a reasonable bulk glucose concentration (average > 3 mM, concentration at the top > 6 mM), glucose limitations (< 0.5 mM) may occur if diffusion to the tissue is limiting as indicated in Figure 6.3D.
The minimum concentration of glucose in the islet after 60 hr of culture varies with tissue surface density (Figure 6.4). Culture in 3 mm medium results in glucose concentrations below 1 mM in the islet at densities between 200 and 300 IE/cm$^2$ assuming Da = 1.48. At same Da culture in 22 mm medium keeps the glucose concentration in the islet above 1 mM until a density of about 1700 IE/cm$^2$ is reached. At lower Da, these limitations only occur at higher surface densities if at all.
Since glucose transport never reaches steady state the duration of culture also has an impact on the concentration within the tissue. The non-dimensional glucose concentration at the top of the islet, which also corresponds to the maximum glucose concentration in the islet, as a function of time and tissue surface density is illustrated in Figure 6.5. Culture in 3 mm medium assuming a low Da results in exposure to very low glucose concentrations only at high densities and long culture times (Figure 6.5A) and 22 mm medium provides sufficient glucose for the duration of culture even at high density (Figure 6.5B). For high Da glucose concentration decreases much faster resulting in exposure to low glucose at lower densities and earlier times (Figure 6.5C and D).
As indicated in Figure 6.3, Figure 6.4, and Figure 6.5, Da has a significant impact on the glucose concentration in the tissue during culture. Figure 6.6 explores how the maximum concentration in islets at 4440 IE/cm$^2$ changes as a function of time and Da. Da as low as 0.1 could cause exposure to glucose concentrations < 0.5 mM at long culture times in 3 mm medium while large Da result in exposure to < 0.5 mM in < 10 hr (Figure 6.6A). Culture in 22 mm medium increases glucose concentration, but exposure to < 1 mM results if Da > 1.0 (Figure 6.6B). As Da increases, the time to reach a low glucose concentration in the tissue decreases.
A second non-dimensional parameter that had an impact on the predicted glucose profiles during islet culture is $D_m/D_i$, which for the previous simulations was about 18. The glucose concentration in the islet is dependent on $D_m/D_i$ for low Da (Figure 6.7A and B). However, 60 hr culture in 3 mm medium results in exposure to glucose concentrations < 1 mM regardless of the $D_m/D_i$. Culture in 22 mm medium prevents these low concentrations from occurring unless $D_m/D_i$ is extremely low. The increased consumption at high Da results in exposure to low glucose concentrations regardless of $D_m/D_i$ and medium depth (Figure 6.7C, and D) though the larger medium depth (22 mm) does delay the decrease in glucose.
The final dimensionless parameter is $K_m/C_{initial}$, which was 0.8 for earlier simulations. $K_m/C_{initial}$ has a significant effect on the maximum glucose concentration in the islet when low Da is assumed (Figure 6.8A and B). Concentrations of < 1 mM occur after 60 hr culture in 3 mm medium when $K_m/C_{initial} < 1.6$ (Figure 6.8A), but does not reach 1 mM with $K_m/C_{initial}$ as low as 0.1 for culture in 22 mm medium (Figure 6.8B). High Da result in concentrations < 1 mM for all $K_m/C_{initial}$ considered after 60 hr independent of medium depth (Figure 6.8C and D).
Figure 6.3. Transient glucose profiles as a function of position during islet culture.
Predictions for islets cultured at 4440 IE/cm$^2$ in 3 (A and C) or 22 mm medium (B and D) assuming Da = 0.14 (A and B) or Da = 1.48 (C and D). The non-dimensional distance ($z/R_{\text{islet}}$) is from the bottom of the plate where $z/R_{\text{islet}} = 0$ through the center of the islet to the top of the islet where $z/R_{\text{islet}} = 2$ to the top of the culture medium where $z/R_{\text{islet}} = 40$ for 3 mm medium and $z/R_{\text{islet}} = 293$ for 22 mm medium. The time between each profile is 5 hours with the initial condition $C/C_{\text{initial}} = 1$ everywhere within the system.
Figure 6.4. The minimum glucose concentration in the islet as a function of islet surface density for tissue cultured for 60 hr.
6.6 Experimental results
At low density, islet tissue cultured in medium containing 0 mM glucose had a significantly lower $F_V$ than tissue cultured in 10 mM glucose (Figure 6.9). $F_V$ of the tissue in 0 mM glucose was $0.56 \pm 0.12$ ($n = 3$ experiments), 0.40 less than that at 10 mM glucose ($0.95 \pm 0.05$) indicating a substantial, but not a complete loss of viable tissue.
Supernatant samples analyzed for glucose content indicated that culture on polystyrene in 3 mm medium at densities > $150 \times 10^4$ cells/cm$^2$ (about 1000 IE/cm$^2$) resulted in glucose depletion from the bulk medium to concentrations < 0.6 mM whereas densities < $78 \times 10^4$ cells/cm$^2$ (about 500 IE/cm$^2$) contained > 3 mM glucose. In a single experiment tissue cultured on polystyrene in 3 mm medium at about $400 \times 10^4$ cells/cm$^2$ (2500 IE/cm$^2$) resulted in glucose depletion to bulk concentration < 0.5 mM within 9 h from culture initiation. Culture on silicone rubber in 22 mm medium also resulted in glucose depletion to a bulk concentration < 0.5 mM at densities > $240 \times 10^4$ cells/cm$^2$ (1500 IE/cm$^2$). Depletion may also occur at lower densities or earlier times on silicone rubber, but data is currently unavailable.
Figure 6.5. The maximum non-dimensional glucose concentration in the islet as a function of time and islet surface density
Predictions for culture in 3 (A and C) or 22 (B and D) mm medium assuming Da = 0.14 (A and B) or Da = 1.48 (C and D).
Figure 6.6. The maximum non-dimensional glucose concentration in the islet as a function of time and Da.
Predictions for culture in 3 (A) or 22 (B) mm medium at an islet surface density of 4440 IE/cm$^2$.
Figure 6.7. The maximum non-dimensional glucose concentration in the islet as a function of time and $D_m/D_i$.
Predictions for culture in 3 (A and C) or 22 (B and D) mm medium assuming $Da = 0.14$ (A and B) or $Da = 1.48$ (C and D) at an islet surface density of 4440 IE/cm$^2$.
Figure 6.8. The maximum non-dimensional glucose concentration in the islet as a function of time and $K_m/C_{\text{initial}}$.
Predictions for culture in 3 (A and C) or 22 (B and D) mm medium assuming Da = 0.14 (A and B) or Da = 1.48 (C and D) at an islet surface density of 4440 IE/cm$^2$.
Figure 6.9. The fraction of original OCR collected from culture for human islets cultured in 3 mm of 0 or 10 mM glucose medium at < 8 x 10^4 cells/cm^2. The error bars represent the standard deviation for 3 independent experiments. * indicates statistical significance (p < 0.05) for a paired two-tailed t-test.
6.7 Discussion
Data from literature indicates that β cell survival during culture is dependent on the glucose concentration of the culture medium [147-150]. When islet tissue is cultured at high density on silicone rubber a decrease in the fractional recovery of viable tissue is observed even though oxygen limitations are removed (Chapter 4). This decrease may be due to glucose transport limitations that become dominating at high densities. Additionally, though 5 mM glucose was determined to be optimal for human islet culture [160], the glucose concentration in tissue during culture is not only a function of initial glucose concentration, but also surface density, tissue GCR, and duration of culture. Therefore a better understanding of the interaction between these variables could elucidate how medium composition or culture technique could be modified for optimal culture.
In this study we developed and explored a transient glucose transport model to determine if tissue was exposed to low glucose concentration during high density islet culture. In addition we conducted preliminary experiments to determine if bulk glucose is depleted during high density culture and whether culture in zero glucose leads to a decreased viability.
Three non-dimensional parameters are present in the theoretical model including Da, $D_m/D_i$, and $K_m/C_{\text{initial}}$ which if inaccurate could introduce error into the predictions. Da > 0.1 and > 1, which are close to 0.14-1.48, the range of Da obtained assuming the literature GCR and our empirical GCR, result in exposure to < 1 mM glucose for culture in 3 and 22 mm medium, respectively (Figure 6.6). The ratios of $D_m/D_i$ (Figure 6.7) and $K_m/C_{\text{initial}}$ (Figure 6.8) do not have as large of an effect on the maximum glucose concentration in the tissue at high Da because the consumption of the tissue is dominating. At low Da, when consumption is not dominating, $D_m/D_i$ and $K_m/C_{\text{initial}}$ can have a significant impact for culture in large medium depths (Figure 6.7B and Figure 6.8B). Thus Da also determines how sensitive the model is to both $D_m/D_i$ and $K_m/C_{\text{initial}}$ further indicating its importance. The parameter within Da that has the most variation based on all available data is $V_{\text{max}}$ (GCR).
GCR measurements made with human islets were an order of magnitude higher than previously reported values for islet tissue ($1 \times 10^{-7}$ vs $9.4 \times 10^{-9}$ mol/s cm$^3$) [153-157]. However, previously reported values were for rodent islet tissue. Human islet preparations are not only a different species, but also contain exocrine tissue (non-islet tissue in the pancreas) that is normally not present in rodent islet preparations. Non-islet tissue may have a higher GCR than islet tissue thereby increasing the overall GCR suggesting a correlation with preparation purity. However, no correlation was observed between islet volume fraction measurements prior to culture and GCR (data not shown), and too few measurements post culture were available to determine whether a correlation existed. Variation in GCR could also result from tissue preparations of differing quality that in previous studies varied over fourfold (Chapter 3). However, if nuclei from dead cells were present after culture, our measurement would underestimate the GCR of the viable tissue suggesting an even higher GCR. More experiments are needed with human islet preparations to confirm the experimental GCR.
The glucose model revealed that glucose concentration in the islet is strongly dependent on tissue surface density, tissue GCR, and the duration of culture. Bulk glucose concentration can be depleted in 3 mm medium over the course of 60 hr, and depending on the tissue density and GCR could be depleted as early as 10 hr (Figure 6.3 and Figure 6.4), which is consistent with experimental measurements indicating near zero glucose at densities > 150 x $10^4$ cells/cm$^2$ after 34-60 hr and after just 9 hr of culture at 400 x $10^4$ cells/cm$^2$ on polystyrene in 3 mm
medium. Culture in 22 mm medium theoretically prevents glucose depletion from the bulk medium (Figure 6.3), but at high Da resulted in islet exposure to < 0.5 mM glucose within 5-10 hr due to the slow diffusion of glucose to the islet surface (Figure 6.3 and Figure 6.5). However, experimental measurements for culture on silicone rubber in 22 mm medium indicated that glucose was depleted from the bulk medium to a concentration < 0.5 mM after 34-60 hr at densities > 240 x 10^4 cells/cm^2. The reason for this discrepancy in unknown but suggests the model overpredicts glucose concentration such that limitations would occur faster than predicted.
Most islet centers culture the “high purity” fraction of tissue from the isolation in approximately 3 mm medium at a density of 200-400 IE/cm^2 (K. Papas, A. Omer, personal communications), which is an underestimation of total tissue volume since non-islet tissue is ignored. The “high purity” fraction could be as low as 70% islet tissue by volume (A. Omer, personal communication) corresponding to a total tissue volume equivalent to 570 “IE”/cm^2, where “IE” in this case is based on total tissue, not only islet tissue. It should be emphasized that the theoretical model accounts for the total tissue surface density because all viable tissue consumes glucose. Thus all references to IE in all the figures indicate a total tissue volume, not islet volume specifically. The minimum concentration in the islet drops below 1 mM after 60 hr culture at 300 IE/cm^2 (Figure 6.4) and within 30 hr at 500 IE/cm^2 (Figure 6.5C) in 3 mm medium (assuming our empirical GCR) indicating that even in conventional methods, exposure to low glucose concentrations may occur unless medium is change every 30 hr.
The glucose concentration in the islet drops below 1 mM after 60 hr at densities above 1700 IE/cm^2 (approximately an OCR surface density of 6.6 nmol/min cm^2 assuming 1560 cells/IE [13] and 2.5 fmol/min cell, Chapter 2) for culture in 22 mm medium (Figure 6.4, assumes our empirical GCR). This may explain decreases in the fractional recovery of viable tissue observed for culture on a silicone rubber membrane in 22 mm medium at OCR densities > 5 nmol/min cm^2 that are independent of O_2 transport (Chapter 4).
Though exposure to low glucose concentration is likely, the question remains whether a reduced glucose level will cause tissue death. Rat β cell survival during culture increases with glucose concentration in a dose-dependent manner up to 10 mM [147-150]. Though 5.5 mM glucose CMRL 1066 was shown resulted in better GSIS than 10 mM glucose RPMI [40; 160] for human islets, it is reasonable to expect that concentrations below 5.5 mM may result in a similar dose-dependent cell death. In addition, even if 5.5 mM glucose is optimal for human islets, the
concentration in the tissue will vary with culture surface density, tissue viability, and culture duration. Unfortunately a critical glucose concentration below which a cell dies is unknown making it difficult to predict viability decreases using our theoretical model. A further complication is that islet tissue may be able to survive for a period of time in the absence of glucose which upon addition of glucose may function normally as observed for the glucose sensitivity of islets [40; 72]. Tissue cultured in 0 mM glucose resulted in a fractional recovery of viable tissue 40% lower than that of tissue cultured in 10 mM glucose (Figure 6.9) indicating a significant, but not complete loss in viability thereby suggesting that a portion of the tissue was able to survive for 34-60 hr without glucose. The tissue likely survived using other nutrients such as amino acids or free fatty acids within the medium to generate ATP. Alternatively the reduced glucose level could signal tissue to reduce its ATP demand and as such decrease requirements of ATP generation.
In summary, the glucose model and preliminary experimental data indicates that islet tissue is exposed to low glucose levels during high density culture. Though further experimentation is needed to determine the rate of death for human islet tissue as a function of glucose concentration, these results suggest that glucose transport limitations present in both polystyrene and silicone rubber cultures likely result in tissue death.
### 6.8 Nomenclature
| Symbol | Description |
|--------|-------------|
| 7-AAD | 7-aminoactinomycin D |
| C | Concentration of glucose |
| $C_{\text{initial}}$ | Initial glucose concentration |
| D | Diffusivity of glucose |
| d | Diameter of islet |
| Da | Damkohler number |
| D-PBS | Dulbecco’s phosphate buffered saline |
| FBS | Fetal bovine serum |
| $F_v$ | The fraction of original viable tissue after culture |
| GCR | Glucose production rate |
| h | Medium depth |
| H | Non-dimensional medium depth, $h/R_{\text{islet}}$ |
| Symbol | Description |
|--------|-------------|
| IE | Islet equivalent |
| JDC | Joslin Diabetes Center |
| $K_m$ | Michaelis constant, the C at which the reaction rate is $V_{\text{max}}/2$ |
| LDH | Lactate dehydrogenase |
| $O_2$ | Oxygen |
| OCR | Oxygen consumption rate |
| $pO_2$ | Oxygen partial pressure |
| R | Production rate of glucose. |
| $R_{\text{islet}}$ | Radius of islet |
| t | Time |
| $V_{\text{ch}}$ | Volume of OCR chamber |
| $V_{\text{max}}$ | Maximum reaction velocity |
| w | Center-to-center distance between islets |
| x | Axis in Cartesian coordinates |
| X | Non-dimensional axis in Cartesian coordinates, $x/R_{\text{islet}}$ |
| y | Axis in Cartesian coordinates |
| Y | Non-dimensional axis in Cartesian coordinates, $y/R_{\text{islet}}$ |
| z | Vertical axis in Cartesian coordinates |
| Z | Non-dimensional vertical axis in Cartesian coordinates, $z/R_{\text{islet}}$ |
| $\alpha$ | Solubility of $O_2$ in medium |
| $\tau$ | Non-dimensional time |
| $\eta$ | Non-dimensional concentration |
Subscripts for D, p, C
- m: Medium
- i: Islet tissue
7 Changes in pH during static islet culture
7.1 Introduction
An alternative limitation to that of oxygen (Chapters 3-5) and glucose (Chapter 6) that may be present during high density islet culture is the accumulation of waste and/or changes in pH. The rate at which tissue produces waste, the depth of the medium, the rate at which waste is transported away from the tissue, and the duration of culture all play important roles in whether a significant pH shift could occur during culture.
The pH within cells changes as a function of the external pH [161]. Shifts in pH can affect many cellular processes including protein secretion and metabolic rates [162-164]. Islet cells in particular are considered to have an “elevated sensitivity” to variations in pH [165]. In βTC3 cells, a murine insulinoma cell line, exposure to pH of 6.4 for 4 hr decreased total ATP content compared to cells at pH 7.1 [164]. Insulin secretion rate of rat islets followed a bell shape curve with reductions on either side of pH 7.1-7.4 when in a bicarbonate buffered medium while the ATP/ADP ratio progressively declined as external pH increased from 6.5-7.7 [163]. These studies indicate significant changes to cellular rates which may cause decreases in tissue viability over time. At pH 8.0 porcine islet viability decreases significantly within 3 days while pH 7.0 results in a higher proportion of dead single cells after 3 weeks of culture compared to culture in pH 7.3 medium [166].
A transient model that simulates changes in local pH during islet culture was developed to investigate the extent to which pH could change during high density culture. Here we introduce and explore this theoretical model to determine the likelihood of whether pH changes significantly.
7.2 Methods
Islet isolation and culture
Standard collagenase/protease digestion methods were used for human [7; 15] isolations. Preparations of human islets were provided fresh (4-6 hr after isolation was complete) by the Islet Core at the Joslin Diabetes Center (JDC, Boston, MA). Human islets were also received after culture (1-2 days) from JDC or overnight shipment (sometimes after up to 1 day culture) from other centers contained in culture flasks (T-25 and T-75) or centrifuge tubes in Styrofoam
boxes. When first observed, some preparations were in suspension, some in a pellet. Upon receipt, a portion of the preparation was analyzed and the remainder placed into untreated polystyrene culture flasks at 37°C in a humidified environment (5% CO₂), oxygen partial pressure (pO₂) of 142 mmHg, for 34-60 hr during which no medium change occurred. Some cultures were terminated at 18-24 hr for analysis. To test the effect of temperature, some islets were cultured at 24°C, then warmed for 1 hr at 37°C before OCR measurements.
Islets were cultured in supplemented RPMI (11.1 mM D-glucose, 100 U/ml penicillin, 100 µg/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate (all from Mediatech Inc., Herdon, VA), and 50 µM β-mercaptoethanol (Sigma Aldrich, St. Louis, MO) with 10% fetal bovine serum (FBS, Mediatech Inc.), at low densities (< 8x10⁴ cells/cm² or about 50 islet equivalents (IE)/cm², where IE is the volume of a sphere with a diameter of 150 µm) in 1.3 mm of medium in untreated polystyrene culture flasks. For some experiments parallel cultures in 0 mM glucose RPMI were conducted. After culture the free, non-adherent tissue (which is the only tissue used in current practice) was collected. In selected experiments, adherent human tissue was incubated for 5 min at 37°C in 0.05% trypsin EDTA (Mediatech Inc.), dislodged with fresh medium, and collected.
**Islet enumeration by nuclei counting**
Nuclei were prepared by adding equal 100-µl volumes of sample containing 160 or more IE and of lysis solution (0.1 M citric acid (Sigma) and 1% (v/v) Triton X-100 (Sigma)) to a 1.5 ml microtube [55]. The mixture was incubated at room temperature for 5 min with vortex mixing every 1.5 min, and nuclei were liberated by shearing through a needle. Isolated nuclei were diluted with Dulbecco’s phosphate buffered saline (D-PBS, Invitrogen, Carlsbad, CA) to a concentration no higher than 5 x 10⁵ nuclei/ml, stained with 7-aminoactinomycin D (7-AAD, Molecular Probes, Eugene, OR), and analyzed using a flow cytometer (Guava Personal Cell Analysis (PCA) system, Guava Technologies, Hayward, CA) to determine the total number of cells with intact nuclei in the sample.
**Ammonia quantification**
Samples were taken of the original culture medium and spent medium following culture and analyzed for ammonia content using the Ammonia Assay Kit (Product Code AA0100 Sigma Aldrich, St. Louis, MO). Triplicate samples diluted to concentrations between 30-100 µg/ml were placed in a 96-well plate along with samples for a six-point standard curve ranging in
concentration from 0-150 µg/ml. The Ammonia Assay Reagent was added to each well. After shaking the plate for 5 min 10X L-Glutamate Dehydrogenase Solution was added to each well. The plate was mixed on a shaker for 5 min before adding 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) solution. After briefly mixing the plate was read on the plate reader with an absorbance of 492 nm on a plate reader.
**Lactate quantification**
Samples were taken of the original culture medium and spent medium following culture and analyzed for lactate content by modifying the protocol for the Glucose Assay Kit (GAGO20, Sigma Aldrich, St. Louis, MO). Triplicate samples diluted to concentrations between 0.1-0.5 mM were placed in a 96-well plate along with samples for a six-point standard curve ranging in concentration from 0-0.5 mM. The Lactate Assay Reagent, a mixture of Lactate Oxidase (12.5 units/ml), Lactate Peroxidase (2.5 Purpurogallin units/ml), and o-Dianisidine (0.1 mg/ml) in deionized water, was added to each well. After 25 min at 37°C 12N sulfuric acid was added to each well. The absorbance was measured at 544 nm on a plate reader.
### 7.3 Theoretical pH model
This transient model predicts the local pH in and around islets in static culture. The species accounted for include hydrogen ions (H\(^+\)), carbon dioxide (CO\(_2\)), bicarbonate ions (HCO\(_3^-\)), lactic acid (HL), and ammonia (NH\(_3\)).
In order to make this problem more feasible the following assumptions are made:
- Bicarbonate is the only important buffer in the system. Phosphate buffers are also present, but at a lower concentration (over four times more bicarbonate buffer than phosphate buffer). The conversion from CO\(_2\) and water (H\(_2\)O) to HCO\(_3^-\) and H\(^+\) follows
\[
\text{H}_2\text{O} + \text{CO}_2 \xrightleftharpoons[k_{-1}]{k_f} \text{H}_2\text{CO}_3 \xrightleftharpoons[K]{k_1} \text{HCO}_3^- + \text{H}^+
\]
where \(k_f\) is the first-order hydration reaction rate constant, \(k_1\) is the first-order dehydration reaction rate constant, and K is the ionization constant for carbonic acid. According to Garg and Maren [167] at near neutral pH the concentration of H\(_2\)CO\(_3\) is
very small because the dissociation of it to $H^+$ and $HCO_3^-$ is very fast allowing 7.1 to be simplified to
$$H_2O + CO_2 \xrightleftharpoons[k_r]{k_f} HCO_3^- + H^+ \quad 7.2$$
where $k_f$ is still a first-order dehydration reaction rate constant and $k_r$ is a second-order reverse reaction rate constant. This simplified conversion will be used in our model using values determined by Garg and Maren [167].
- All lactic acid produced by the tissue is immediately dissociated into ions through the following equation.
$$HL \Rightarrow H^+ + Lac^- \quad 7.3$$
where $Lac^-$ is a lactate ion. Since the pKa of lactic acid is 3.86 [118], this should be a reasonable assumption unless the pH drops significantly from neutral pH.
- All ammonia produced by the tissue immediately combines with a hydrogen ion through the following equation.
$$NH_3 + H^+ \Rightarrow NH_4^+ \quad 7.4$$
where $NH_3$ is ammonia and $NH_4^+$ is the ammonium ion. Since the pKa of ammonia is 9.25 [118] this will hold true unless the pH increases significantly above neutral pH.
- An effective diffusivity is used for each species in the tissue (in other words we do not distinguish between intercellular and intracellular diffusion within the tissue).
- There is no significant partition of any species between the medium and islet tissue except for $CO_2$, which will partition based on its solubility in each domain.
- Electrostatic effects can be ignored (i.e. transport due to charge is negligible, preliminary calculations support this assumption).
- Water is in equilibrium with its ions.
- The species conservation equation in the absence of convection is given by [16; 97]
$$\frac{\partial C_j}{\partial t} - D_j \nabla^2 C_j = R_j \quad 7.5$$
where $C_j$ is the concentration of species “j”, $D_j$ is the diffusivity of species “j”, and $R_j$ is the production rate of species “j”.
To model islet culture all the tissue (islet and non-islet) was assumed to be in the form of spheres with uniform diameter (d) that were arranged in a square array with center-to-center distance w in a stagnant medium of height h on either a CO$_2$-impermeable or CO$_2$-permeable membrane of known thickness (Figure 7.1). Because of symmetry within the system, Equation 7.5 was solved within a unit cell containing only an eighth of an islet (dashed lines in Figure 7.1).
Here we present the model equations for islets on a CO$_2$-impermeable surface. The boundary conditions for CO$_2$ in this model include constant external partial pressure of CO$_2$ ($p_{CO_2,gas}$) across the gas-medium (m) interface,
at $z = h$, \[ p_{CO_2,gas} = p_{CO_2,m} \]
no CO$_2$ flux through the impermeable dish,
at $z=0$, \[ \frac{\partial C_{CO_2}}{\partial z} = 0 \]
continuity of $p_{CO_2}$ and CO$_2$ flux across the medium-islet (i) interface,
\[ n \bullet ((-D_{CO_2} \nabla C_{CO_2})_i - (-D_{CO_2} \nabla C_{CO_2})_m) = 0 \]
\[ p_{CO_2,i} = p_{CO_2,m} \]
and symmetry across each side of the unit cell including both medium and islet domains
\[ n \bullet (-D_{CO_2} \nabla C_{CO_2}) = 0 \]
CO$_2$ is produced by the tissue at a rate $V_{max,CO_2}$. Substituting in the production rate of the tissue and the interaction with the bicarbonate and hydrogen ions (7.2) a separate equation for each domain is obtained, one that describes CO$_2$ transport in the medium (m) and one that describes CO$_2$ transport in the islet (i).
Figure 7.1. Islet geometry assumed for modeling pH changes for islet tissue on either a polystyrene dish or a 500-\(\mu\)m silicone rubber membrane.
\[ \frac{\partial C_{CO_2}}{\partial t} - D_{CO_2,i} \nabla^2 C_{CO_2} = V_{\text{max}, CO_2} - k_f C_{CO_2} + k_r C_{H^+} C_{HCO_3^-} \quad 7.11 \]
\[ \frac{\partial C_{CO_2}}{\partial t} - D_{CO_2,m} \nabla^2 C_{CO_2} = -k_f C_{CO_2} + k_r C_{H^+} C_{HCO_3^-} \quad 7.12 \]
For the hydrogen ion, lactate ion, ammonium ion, and bicarbonate ion the boundary conditions include no flux through the bottom of the culture vessel or through the air-liquid surface,
at \( z=0 \) and \( z=h \),
\[ \frac{\partial C_j}{\partial z} = 0 \quad 7.13 \]
continuity of flux across the liquid-islet interface,
\[ n \cdot ((-D_j \nabla C_j)_i - (-D_j \nabla C_j)_m) = 0 \quad 7.14 \]
and symmetry across each side of the unit cell including both medium and islet domains
\[ n \cdot (-D_j \nabla C_j) = 0 \quad 7.15 \]
Ammonia and lactic acid production are known constant rates within the islet. Separate equations were used for each domain, one that describes transport in the medium (m) and one that describes transport in tissue (i) for each solute.
\[ \frac{\partial C_{Lac^-}}{\partial t} - D_{Lac^-,i} \nabla^2 C_{Lac^-} = R_{Lac^-} \quad 7.16 \]
\[ \frac{\partial C_{Lac^-}}{\partial t} - D_{Lac^-,m} \nabla^2 C_{Lac^-} = 0 \quad 7.17 \]
\[ \frac{\partial C_{NH_4^+}}{\partial t} - D_{NH_4^-,i} \nabla^2 C_{NH_4^+} = R_{NH_4^+} \quad 7.18 \]
\[ \frac{\partial C_{NH_4^+}}{\partial t} - D_{NH_4^-,m} \nabla^2 C_{NH_4^+} = 0 \quad 7.19 \]
\[
\frac{\partial C_{HCO_3^-}}{\partial t} - D_{HCO_3^-,i} \nabla^2 C_{HCO_3^-} = \left( k_f C_{CO_2} - k_r C_{H^+} C_{HCO_3^-} \right) \tag{7.20}
\]
\[
\frac{\partial C_{HCO_3^-}}{\partial t} - D_{HCO_3^-,m} \nabla^2 C_{HCO_3^-} = \left( k_f C_{CO_2} - k_r C_{H^+} C_{HCO_3^-} \right) \tag{7.21}
\]
\[
\frac{\partial C_{H^+}}{\partial t} - D_{H^+,i} \nabla^2 C_{H^+} = \left( R_{Lac^-} - R_{NH_4^+} + k_f C_{CO_2} - k_r C_{H^+} C_{HCO_3^-} \right) \tag{7.22}
\]
\[
\frac{\partial C_{H^+}}{\partial t} - D_{H^+,m} \nabla^2 C_{H^+} = \left( k_f \eta_{CO_2} - k_r C_{H^+} C_{HCO_3^-} \right) \tag{7.23}
\]
The equations of the transient model were non-dimensionalized. The non-dimensional variables used are
\[
\eta_i = \frac{C_i}{C_{CO_2,m,t=0}} \tag{7.24}
\]
\[
X = \frac{x}{R_{islet}}, \quad Y = \frac{y}{R_{islet}}, \quad Z = \frac{z}{R_{islet}} \tag{7.25}
\]
\[
\tau = \frac{D_{CO_2,m} t}{R_{islet}^2} \tag{7.26}
\]
Using these non-dimensional variables the governing equations can be rewritten in non-dimensional form.
\[
\frac{\partial \eta_{CO_2}}{\partial \tau} - \frac{D_{CO_2,i}}{D_{CO_2,m}} \left( \nabla^2 \eta_{CO_2} \right) = \frac{R_{islet}^2}{D_{CO_2,m}} \left( \frac{V_{max,CO_2}}{C_{CO_2,m,t=0}} - k_f \eta_{CO_2} + k_r C_{CO_2,m,t=0} \eta_{H^+} \eta_{HCO_3^-} \right) \tag{7.27}
\]
\[
\frac{\partial \eta_{CO_2}}{\partial \tau} - \left( \nabla^2 \eta_{CO_2} \right) = \frac{R_{islet}^2}{D_{CO_2,m}} \left( -k_f \eta_{CO_2} + k_r C_{CO_2,m,t=0} \eta_{H^+} \eta_{HCO_3^-} \right) \tag{7.28}
\]
\[
\frac{\partial \eta_{Lac^-}}{\partial \tau} - \frac{D_{Lac^-,i}}{D_{CO_2,m}} \nabla^2 \eta_{Lac^-} = \frac{R_{islet}^2}{D_{CO_2,m} C_{CO_2,m,t=0}} R_{Lac^-}
\]
7.29
\[
\frac{\partial \eta_{Lac^-}}{\partial \tau} - \frac{D_{Lac,m}}{D_{CO_2,m}} \nabla^2 \eta_{Lac^-} = 0
\]
7.30
\[
\frac{\partial \eta_{NH_4^+}}{\partial \tau} - \frac{D_{NH_4^+,i}}{D_{CO_2,m}} \nabla^2 \eta_{NH_4^+} = \frac{R_{islet}^2}{D_{CO_2,m} C_{CO_2,m,t=0}} R_{NH_4^+}
\]
7.31
\[
\frac{\partial \eta_{NH_4^+}}{\partial \tau} - \frac{D_{NH_4^+,m}}{D_{CO_2,m}} \nabla^2 \eta_{NH_4^+} = 0
\]
7.32
\[
\frac{\partial \eta_{HCO_3^-}}{\partial \tau} - \frac{D_{HCO_3^-,i}}{D_{CO_2,m}} \nabla^2 \eta_{HCO_3^-} = \frac{R_{islet}^2}{D_{CO_2,m}} \left( k_f \eta_{CO_2} - k_r C_{CO_2,m,t=0} \eta_{H^+} \eta_{HCO_3^-} \right)
\]
7.33
\[
\frac{\partial \eta_{HCO_3^-}}{\partial \tau} - \frac{D_{HCO_3^-,m}}{D_{CO_2,m}} \nabla^2 \eta_{HCO_3^-} = \frac{R_{islet}^2}{D_{CO_2,m}} \left( k_f \eta_{CO_2} - k_r C_{CO_2,m,t=0} \eta_{H^+} \eta_{HCO_3^-} \right)
\]
7.34
\[
\frac{\partial \eta_{H^+}}{\partial \tau} - \frac{D_{H^+,i}}{D_{CO_2,m}} \nabla^2 \eta_{H^+} = \frac{R_{islet}^2}{D_{CO_2,m}} \left( \frac{R_{Lac^-} - R_{NH_4^+}}{C_{CO_2,m,t=0}} + k_f \eta_{CO_2} - k_r C_{CO_2,m,t=0} \eta_{H^+} \eta_{HCO_3^-} \right)
\]
7.35
\[
\frac{\partial \eta_{H^+}}{\partial \tau} - \nabla \cdot \frac{D_{H^+,m}}{D_{CO_2,m}} \nabla^2 \eta_{H^+} = \frac{R_{islet}^2}{D_{CO_2,m}} \left( k_f \eta_{CO_2} - k_r C_{CO_2,m,t=0} \eta_{H^+} \eta_{HCO_3^-} \right)
\]
7.36
The non-dimensional boundary conditions for CO$_2$ in this model include constant external p$_{CO_2}$ across the gas-medium interface,
at z=H where H = h/R$_{islet}$,
\[
\eta_{CO_2} = 1
\]
7.37
no CO$_2$ flux through the impermeable dish,
at $z=0$,
$$\frac{\partial \eta_{CO_2}}{\partial Z} = 0 \quad 7.38$$
continuity of $p_{CO_2}$ and $CO_2$ flux across the medium-islet interface,
$$n \cdot \left( -\frac{D_{CO_2,i}}{D_{CO_2,m}} \nabla \eta_{CO_2} \right)_i - (-\nabla \eta_{CO_2})_m = 0 \quad 7.39$$
$$\eta_{CO_2,i} = \frac{\alpha_{CO_2,i}}{\alpha_{CO_2,m}} \eta_{CO_2,m} \quad 7.40$$
where $\alpha_{CO_2,i}$ is the solubility of $CO_2$ in islet tissue and $\alpha_{CO_2,m}$ is the solubility of $CO_2$ in medium, and symmetry across each side of the unit cell including both medium and islet domains
$$n \cdot (-\nabla \eta_{CO_2}) = 0 \quad 7.41$$
For transport of the hydrogen ion, lactate ion, ammonium ion, and bicarbonate ion the boundary conditions include no flux through the bottom of the culture vessel or through the air-liquid surface,
at $z=0$ and $z=H$,
$$\frac{\partial \eta_j}{\partial Z} = 0 \quad 7.42$$
continuity of flux across the liquid-islet interface,
$$n \cdot \left( -\frac{D_{j,i}}{D_{j,m}} \nabla \eta_j \right)_i - (-\nabla \eta_j)_m = 0 \quad 7.43$$
and symmetry across each side of the unit cell including both medium and islet domains
$$n \cdot (-\nabla \eta_j) = 0 \quad 7.44$$
When modeling islet culture on a $CO_2$-permeable membrane an additional conservation equation (with no reaction term) is required for transport of carbon dioxide in the silicone rubber.
The boundary condition in Equation 7.7 is replaced by continuity equations similar to those in Equation 7.8 and 7.9 except it is evaluated at the interface of the medium and silicone rubber. In addition, CO$_2$ at a constant external pO$_2$ is in equilibrium with the bottom of the silicone rubber membrane surface and symmetry conditions like those in Equation 7.11 are applied to the sides of the unit cell. All equations and boundary conditions for other species remain the same. Non-dimensionalization of these equations was performed in a similar manner as that above.
Hydrogen ion profiles, and thus pH profiles, in the theoretical model are governed by four key Damkohler numbers (Da) including
$$\text{Da}_{\text{CO}_2} = \frac{\text{V}_{\text{max},\text{CO}_2} \cdot R_{\text{islet}}^2}{D_{\text{CO}_2,m} \cdot C_{\text{CO}_2,m,t=0}} \quad 7.45$$
$$\text{Da}_{k_f} = \frac{k_f \cdot R_{\text{islet}}^2}{D_{\text{CO}_2,m}} \quad 7.46$$
$$\text{Da}_{k_r} = \frac{k_r \cdot C_{\text{CO}_2,m,t=0} \cdot R_{\text{islet}}^2}{D_{\text{CO}_2,m}} \quad 7.47$$
$$\text{Da}_{\text{Lac-Am}} = \frac{\left(R_{\text{Lac}^-} - R_{\text{NH}_4^+}\right) \cdot R_{\text{islet}}^2}{D_{\text{CO}_2,m} \cdot C_{\text{CO}_2,m,t=0}} \quad 7.48$$
Though non-dimensional diffusivities are also present, the uncertainty in their values and their corresponding impact is likely minor compared to the uncertainty in the Da values. As such, the theoretical analysis focuses on changes in pH profiles due to changes in $\text{Da}_{\text{CO}_2}$, $\text{Da}_{k_f}$, $\text{Da}_{k_r}$, and $\text{Da}_{\text{Lac-Am}}$.
The model was solved with the finite element method using the commercially available software COMSOL Multiphysics (COMCOL Inc., Burlington, MA) in conjunction with Matlab (Mathworks, Natick, MA). We required the mesh to contain more than 1000 nodes and set an absolute tolerance for each concentration of $10^{-6}$ [99].
The diffusion coefficients and cellular production rates obtained from literature are listed in Table 7.1 and Table 7.2, respectively. The average of the production rates for ammonia and lactate in literature were used as the base case from literature. Experimental measurements were also conducted to determine the lactate and ammonia production rates of human islet tissue and will be discussed later. The remaining parameter values are given in Table 7.3 along with how/where the values were obtained.
**Table 7.1. Diffusivities of each species used in the model in tissue and in culture medium.**
| Species | Domain | Diffusivity (cm²/s) | Reference |
|---------|-----------------|---------------------|-----------|
| CO₂ | Tissue | $1.0 \times 10^{-5}$| [138] |
| | Medium | $2.55 \times 10^{-5}$| [117; 118]|
| | Silicone rubber | $2.9 \times 10^{-5}$| [141] |
| Lac⁻ | Tissue | $5.3 \times 10^{-6}$| [168] |
| | Medium | $1.1 \times 10^{-5}$| [169] |
| NH₄⁺ | Tissue | $2 \times 10^{-6}$ | [170] |
| | Medium | $2.7 \times 10^{-5}$| [117] |
| HCO₃⁻ | Tissue | $8.6 \times 10^{-7}$| [171] |
| | Medium | $1.19 \times 10^{-5}$| [171; 172]|
| H⁺ | Tissue | $6 \times 10^{-6}$ | [171] |
| | Medium | $11.2 \times 10^{-5}$| [171; 172]|
*Diffusion coefficients reported at temperatures other than 37°C were adjusted to 37°C assuming D/μ/T = constant [117], where μ is the viscosity of the solvent (water) at the specified temperature T.*
Table 7.2. Cellular production rates for CO$_2$, ammonia, and lactate.
| Chemical species | Islet type | Original value | Original units | mol/s cm$^3$ | Reference |
|------------------|------------|----------------|----------------|---------------|-----------|
| CO$_2$ | Human | - | - | $3.68 \times 10^{-8}$ | Chapter 2 |
| NH$_4^+$ | Rat | 80 - 220 | nmol/2h 1000 islets | $6.3 - 17.3 \times 10^{-9}$ | [151] |
| Lac$^-$ | Mouse | 3629 ± 357 | pmol/h ug DNA | $5.78 \times 10^{-9}$ | [154] |
| | Rat | 3576 ± 153 | pmol/h ug DNA | $5.70 \times 10^{-9}$ | [154] |
| | Rat | 78 ± 37 | umol/h g dry weight | $3.74 \times 10^{-9}$ | [155] |
Assumptions include (1) 10.14 ng DNA per IE based on 1560 cells per IE [113] and 6.5 pg DNA/cell [68], (2) an IE is the volume of tissue equivalent to a 150-micron sphere, (3) there are 33.2 ng DNA/ug dry weight (average of 3 values referenced in [155]), (4) production of CO$_2$ was assumed to be equal to the OCR of the tissue based on stoichiometric relationships for the Krebs cycle. The average OCR, and thus CO$_2$ production rate, is about 2.5 fmol/min cell (Chapter 2).
Table 7.3. Parameter values used in pH model for high density islet culture.
| Parameter | Value | Units | Reference |
|------------------------------------------------|----------------|-----------|-----------------|
| $k_r$ | 0.145 | s$^{-1}$ | [167] |
| $k_r$ | $17.2 \times 10^4$ | L/mol s | [167] |
| External partial pressure of CO$_2$ | 38 | mmHg | 5% CO$_2$ in incubator |
| Initial pH of medium | 7.4 | | Experimentally determined |
| Initial ammonium ion concentration | 0 | mol/L | Medium composition |
| Initial lactate ion concentration | 0 | mol/L | Medium composition |
| CO$_2$ solubility in tissue | $3.21 \times 10^{-5}$ | M/torr | [138] |
| CO$_2$ solubility in medium | $3.18 \times 10^{-5}$ | M/torr | [125; 172] |
| CO$_2$ solubility in silicone rubber | $7.7 \times 10^{-5}$ | M/torr | [173] |
CO$_2$ solubility in tissue is obtained by dividing CO$_2$ permeability data in a variety of tissues by the diffusion coefficient $1.19 \times 10^{-3}$ cm$^2$/s [138]. CO$_2$ solubility in water was used as an approximation for the CO$_2$ solubility in the medium [125; 172].
7.4 Ammonia and lactate production rate measurements
The change in lactate and ammonia concentration determined using the original and spent medium was converted to an increase in total lactate and ammonia which was normalized using nuclei counts after culture to obtain the production rates. The lactate production rate (LPR) was determined to be $5.3 \pm 1.6 \times 10^{-8}$ mol/s cm$^3$ ($n = 11$, data ranged from $2.5 - 7.8 \times 10^{-8}$ mol/s cm$^3$) assuming 1560 cells/IE [55] while the ammonia production rate (APR) was $5.0 \pm 2.6 \times 10^{-8}$ mol/s cm$^3$ ($n = 6$, data ranged from $1.6 - 6.9 \times 10^{-8}$ mol/s cm$^3$). These values will be used for the base case from empirical data.
7.5 Predictions of theoretical model
Since values of lactate and ammonia production determined experimentally were significantly different than those reported in literature, two sets of base case parameters will be modeled as indicated in Table 7.4.
Culture on polystyrene at 4400 IE/cm$^2$ results in large changes in pH throughout the medium (Figure 7.2A-D) whereas culture on silicone rubber maintains pH between 7.2 and 7.5 for more than 60 hr (Figure 7.2E and F). Due to CO$_2$ production, which causes the bicarbonate reaction (7.2) to shift toward the presence of H$^+$ in solution, there is a substantial decrease in pH
within 5 hr of initiating culture on polystyrene resulting in exposure to pH < 7 in the islet tissue, which is where the pH changes the most. This shift is almost completely removed by culture on a silicone rubber membrane because the CO$_2$ can escape through the membrane preventing the bicarbonate reaction from shifting to more acidic conditions. Assuming data from literature is accurate culture on polystyrene in 3 mm medium recovers from the initial shift in pH over time (Figure 7.2A), however, the empirical data indicates that pH would continue to decrease (Figure 7.2B). Bulk pH of the system is much higher than what the islet is exposed to, especially for large medium depths on polystyrene.
The data in Figure 7.2 assume culture is conducted at 4400 IE/cm$^2$, nearly a monolayer of islets, and is meant to provide a worst case scenario. At lower densities the shift in pH is less substantial (Figure 7.3). In almost all cases the pH at the top of the islet at any given time is closest to physiologic pH of 7.4 at lower densities compared to higher surface densities independent of which base case is assumed (exception is Figure 7.3E, silicone rubber with 22 mm medium). Based on the predictions in Figure 7.3, large pH changes at the islet surface only occur for culture on polystyrene and at densities > 1000 IE/cm$^2$ in 3 mm medium, but in 22 mm medium could occur at densities < 500 IE/cm$^2$. Though this is specifically for the top of the islet, the pH change within the islet is relatively small (Figure 7.2A and B).
**Table 7.4. Damkohler numbers based on data in literature and empirical measurements used in the base case simulations.**
| Base case | Damkohler number | Value |
|-----------|------------------|-------|
| Literature | Da$_{CO_2}$ | 0.067 |
| | Da$_{k_f}$ | 0.32 |
| | Da$_{k_r}$ | 460 |
| | Da$_{Lac-Am}$ | -0.012|
| Experimental | Da$_{CO_2}$ | 0.067 |
| | Da$_{k_f}$ | 0.32 |
| | Da$_{k_r}$ | 460 |
| | Da$_{Lac-Am}$ | 0.0055|
Figure 7.2. Transient pH profiles for islets as a function of position.
Predictions for culture at 4400 IE/cm$^2$ in 3 (A and B) or 22 mm medium (C and D) on polystyrene or on a 500-µm thick silicone rubber membrane in 22 mm medium (E and F). Plots are given using parameters from literature (A, C, and E) and using experimentally determined parameters (B, D, and F). The non-dimensional distance ($z/R_{\text{islet}}$) is from the bottom of the plate where $z/R_{\text{islet}} = 0$ through the center of the islet to the top of the islet where $z/R_{\text{islet}} = 2$ to the top of the culture medium where $z/R_{\text{islet}} = 40$ for 3 mm medium and $z/R_{\text{islet}} = 293$ for 22 mm medium. The time between each profile in A-D is 5 hours.
Figure 7.3. The pH at the top of the islet as a function of time and islet density.
Predictions for culture on polystyrene in 3 (A and B) or 22 (C and D) mm medium or on a 500-\(\mu\)m thick silicone rubber membrane in 22 mm medium (E and F). Plots are given using parameters from literature (A, C, and E) and using experimentally determined parameters (B, D, and F).
Da can have a large impact on the pH profiles. Assuming the empirical base case, the magnitude of the initial drop in pH increases and the long term pH value decreases as $Da_{CO_2}$ is increased from 0.001 to 10 (Figure 7.4B, D, and F). Therefore the more CO$_2$ produced by the tissue (i.e. the healthier the tissue), the larger the initial downward shift in pH. For the base case assuming literature data for lactate and ammonia production is accurate, a similar trend is obtained (Figure 7.4A and C), but pH slowly increases after the initial drop. For culture on polystyrene the pH after 60 hr changes from < 7.4 to > 7.4 at $Da_{CO_2}$ between 0.1 and 0.01. Culture on silicone rubber results in a much smaller shift in pH compared to that on polystyrene and changes very little with time after an initial transient period (Figure 7.4E and F) as observed above.
The effect of $Da_{k_f}$ and $Da_{k_r}$ on the pH at the islet surface are shown in Figure 7.5 and Figure 7.6, respectively. The higher the $Da_{k_f}$ or the lower the $Da_{k_r}$, the lower the pH is during culture. This trend is consistent across all culture conditions and independent of which base case values are used. Further, the shape of the profile for each individual case does not changed significantly with $Da_{k_f}$ or $Da_{k_r}$, it just shifts up or down.
$Da_{Lac-Am} > 0$, when lactate production exceeds ammonia production indicating a net production of H$^+$, causes pH to continually decrease with time. $Da_{Lac-Am} < 0$, a net removal of H$^+$, decreases the extent of which pH drops in the initial transient period after which it continually increases pH. At $Da_{Lac-Am} > 0.1$ a drastic decreases in the pH for all cultures was observed. When $Da_{Lac-Am}$ is between -0.01 and 0.001 (i.e. lactate and ammonia production are similar) no large shifts in pH are observed after the initial transient period that results from CO$_2$ production (Figure 7.7). However, even within this range for $Da_{Lac-Am}$ the pH during culture can be significantly below 7.4 for culture on polystyrene.
Figure 7.4. The pH at the top of the islet as a function of time and $D_{a\text{CO}_2}$.
Predictions for culture on polystyrene in 3 (A and B) or 22 (C and D) mm medium or on a 500-$\mu$m thick silicone rubber membrane in 22 mm medium (E and F) at an islet surface density of 4400 IE/cm$^2$. Plots are given using parameters from literature (A, C, and E) and using experimentally determined parameters (B, D, and F).
Figure 7.5. The pH at the top of the islet as a function of time and $D_{Ak_r}$
Predictions for culture on polystyrene in 3 (A and B) or 22 (C and D) mm medium or on a 500-$\mu$m thick silicone rubber membrane in 22 mm medium (E and F) at an islet surface density of 4400 IE/cm$^2$. Plots are given using parameters from literature (A, C, and E) and using experimentally determined parameters (B, D, and F).
Figure 7.6. The pH at the top of the islet as a function of time and $D_{A_{\text{eff}}}$.
Predictions for culture on polystyrene in 3 (A and B) or 22 (C and D) mm medium or on a 500-$\mu$m thick silicone rubber membrane in 22 mm medium (E and F) at an islet surface density of 4400 IE/cm$^2$. Plots are given using parameters from literature (A, C, and E) and using experimentally determined parameters (B, D, and F).
Figure 7.7. The pH at the top of the islet as a function of time and DaLac-Am.
Predictions for culture on polystyrene in 3 (A) or 22 (B) mm medium or on a 500-μm thick silicone rubber membrane in 22 mm medium (C) at an islet surface density of 4400 IE/cm². DaLac-Am < -0.1 caused problems within the theoretical model because negative H⁺ concentrations were observed. This problem could be avoided by modifying the code, but significantly more time and computation memory would be required.
7.6 Discussion
Significant changes in cellular processes occur with changes in pH [162-164]. Islet tissue has an “elevated sensitivity” to variations in pH [165], which can affect insulin secretion rates and ATP/ADP [163] and reduce viability [166]. A better understanding of how pH changes for various static culture techniques could reveal how medium composition or culture technique could be modified for optimal culture. In this study we developed and explored a theoretical model that predicts changes in local pH during islet culture to determine the extent to which pH could change during high density culture. The model took into account hydrogen ions, carbon dioxide, bicarbonate ions, lactate, and ammonia.
The pH profiles for polystyrene culture are strongly dependent on tissue surface density and the duration of culture. The pH at the islet surface changes significantly within 5 hours of culture whereas that at the surface of the medium and in the bulk does not shift to the same extent until much later (Figure 7.2) suggesting that monitoring surface or bulk pH gives an inaccurate representation of the pH in the tissue. Though pH shifts are significant during high density culture, they are greatly reduced at lower densities (Figure 7.3).
The four non-dimensional Damkohler numbers present in the model have a large impact on how pH changes during culture. $\text{Da}_{\text{K}_f}$ and $\text{Da}_{\text{K}_r}$ shift pH profiles up or down, but do not change the shape of the profile (Figure 7.5 and Figure 7.6). Changing $\text{Da}_{\text{K}_f}$ and $\text{Da}_{\text{K}_r}$ essentially changes the pKa of the bicarbonate buffer and the equilibrium of the bicarbonate reaction 7.2, but not the kinetics of the system. $\text{Da}_{\text{CO}_2}$ has a large impact on the pH profiles especially within the first few hours of culture (Figure 7.4). The presence of CO$_2$ shifts the bicarbonate reaction 7.2 toward the presence of H$^+$ in solution lowering pH indicating that healthier tissue will cause a larger shift in pH. Unlike $\text{Da}_{\text{CO}_2}$ (CO$_2$ production), $\text{Da}_{\text{Lac-Am}}$ (lactate and ammonia production), which using the extremes of all available data could range from 0.15 to -0.15, cause pH to change continuously over the entire culture period (Figure 7.7). Thus an accurate CO$_2$ production rate is critical for predicting local pH throughout culture while accurate APR and LPR are most important at longer culture periods unless you are at extreme values (such as $\text{Da}_{\text{Lac-Am}} = 0.1$).
LPR measurements made within our lab were significantly higher than those previously reported [154; 155]. However, as with GCR values discussed in Chapter 6 previously reported values were for high purity rodent islet tissue as opposed to the human islet preparations used in this study, which contain a significant amount of exocrine tissue (non-islet tissue in the pancreas). Islet tissue normally has a low level of lactate dehydrogenase (LDH) suggesting a low LPR [119]. Non-islet tissue may have higher LDH expression resulting in higher LPR than islet tissue thereby increasing the overall LPR. Though this suggests a correlation with preparation purity none was observed. APR measurements were also higher than the previously reported value, but of the same order of magnitude suggesting that the method used for determining production rates was not an issue.
Based on data obtained in this study and Chapters 2 and 6 a species balance between oxygen, glucose, and lactate can be performed to see if the values we have found are consistent. The average oxygen consumption rate (OCR), glucose consumption rate (GCR), and LPR of human islets was roughly $3.68 \times 10^{-8}$ mol/s cm$^3$ (Chapter 2), $5.55 \times 10^{-8}$ mol/s cm$^3$ (Chapter 6), and $5.3 \times 10^{-8}$ mol/s cm$^3$, respectively. The breakdown of glucose through glycolysis results in the production of two lactate molecules while that through the TCA cycle requires six oxygen molecules. Thus, assuming that glycolysis and the TCA cycle are the only cellular processes in which glucose is used, the GCR should equal the OCR divided by six plus the LPR divided by two. The GCR calculated using the OCR and LPR production rates is $3.3 \times 10^{-8}$ mol/s cm$^3$, which is lower than the measured value in Chapter 6, but within its standard deviation, indicating consistency between OCR, GCR, and LPR.
As mentioned in Chapter 6 most islet centers culture at a total tissue volume equivalent to 570 “IE”/cm$^2$, where “IE” in this case is based on total tissue, not only islet tissue. Like the glucose (Chapter 6) and oxygen models (Chapter 3-5) the pH model accounts for the total tissue surface density. Thus all references to IE in all the figures indicate a total tissue volume, not islet volume specifically. The pH at the islet surface for culture on polystyrene in 3 mm medium is not likely to drop below 7.2 even after 60 hr culture at 500 IE/cm$^2$ (Figure 7.3). This indicates that drastic pH changes are not present during conventional culture. Further, though pH in the islet does shift below 7 during culture at large medium depths on polystyrene, these changes may be decreased since tissue will be oxygen starved resulting in decreased CO$_2$ production and decreased viability.
The fractional recovery of viable tissue decreases as surface density increases for culture on a silicone rubber membrane due to limitations independent of oxygen (Chapter 4). Based on the data presented here, changes in pH are also unlikely to cause the empirically observed decrease. Culture on a silicone rubber membrane keeps pH near 7.4 by allowing CO$_2$ to escape the system at a much faster rate thereby preventing the buildup of CO$_2$ in medium, which in turn prevents the bicarbonate reaction from shifting to more acidic conditions. Therefore significant changes in pH only occur if LPR is significantly larger than APR or vice versa (Figure 7.7).
In summary, the theoretical predictions indicate that islet tissue can be exposed to pH significantly different than the physiological value of 7.4 during high density culture on polystyrene. Based on data in literature, these pH shifts could cause changes in cellular rates leading to a decrease in viability. However, standard clinical culture and culture on a silicone rubber membrane are unlikely to result in large shifts in pH.
### 7.7 Nomenclature
| Symbol | Description |
|--------|-------------|
| APR | Ammonia production rate |
| CO$_2$ | Carbon dioxide |
| d | Diameter of islet |
| Da | Damkohler number, measure of rate of reaction relative to rate of diffusion |
| DaCO$_2$ | Da for the production of CO$_2$ |
| Dak$_f$ | Da for the forward reaction producing bicarbonate |
| Dak$_r$ | Da for the reverse reaction producing bicarbonate |
| DaLac-Am | Da for the lactate minus ammonia production rates |
| F$_V$ | The fraction of original viable tissue after culture |
| GCR | Glucose production rate |
| h | Medium depth |
| H | Non-dimensional medium depth, h/R$_{islet}$ |
| H$^+$ | Hydrogen ion |
| HL | Lactic acid |
| IE | Islet equivalent |
| Symbol | Description |
|--------|-------------|
| K | Ionization constant for carbonic acid |
| $k_1$ | First-order dehydration reaction rate constant |
| $k_f$ | First-order hydration reaction rate constant |
| $k_r$ | Second-order reverse reaction rate constant |
| Lac$^-$| Lactate ion |
| LPR | Lactate production rate |
| NH$_3$ | Ammonia |
| NH$_4^+$| Ammonium ion |
| OCR | Oxygen consumption rate |
| $R_{\text{islet}}$ | Radius of islet |
| t | Time |
| $V_{\text{max}}$ | Maximum reaction velocity |
| w | Center-to-center distance between islets |
| x | Axis in Cartesian coordinates |
| X | Non-dimensional axis in Cartesian coordinates, $x/R_{\text{islet}}$ |
| y | Axis in Cartesian coordinates |
| Y | Non-dimensional axis in Cartesian coordinates, $y/R_{\text{islet}}$ |
| z | Vertical axis in Cartesian coordinates |
| Z | Non-dimensional vertical axis in Cartesian coordinates, $z/R_{\text{islet}}$ |
| $\alpha$ | Solubility of CO$_2$ in medium |
| $\delta$ | Thickness of silicone rubber membrane |
| $\eta$ | Non-dimensional concentration |
| $\tau$ | Non-dimensional time |
| $\chi$ | Cell surface density (cells/cm$^2$) |
Subscripts for D, $\alpha$, p, C
- CO$_2$: Carbon dioxide
- H$^+$: Hydrogen ions
- HCO$_3^-$: Bicarbonate ions
- i: Islet tissue
| Symbol | Description |
|--------|------------------------------|
| j | Reference to any species |
| Lac⁻ | Lactate ion |
| m | Medium |
| NH₄⁺ | Ammonium ion |
| Si | Silicone rubber |
8 Improvements in islet transportation
8.1 Introduction
The quality of shipped islet tissue, which varies widely between islet centers and within some islet centers [26], depends on specific pancreas characteristics, isolation process, culture following isolation, and shipment of the tissue. Isolation and pre-isolation variables cause much of the variation since the success rate for isolation ranges from 25-75% even within expert islet centers [27]. Centralization of islet isolation will attempt to reduce some variation between preparations allowing centers that most frequently produce high quality islet preparations to supply researchers and hospitals thereby making efficient use each pancreas [174]. However, in order for islet transplantation to be applied on a large scale, there needs to be the ability to transport large amounts of high quality tissue without loss in viability and/or function from these centralized islet isolation facilities.
Maintaining islet viability and function during shipment poses significant challenges. The current shipping method involves placing islet tissue in an oxygen-impermeable container (i.e. a flask or centrifuge tube), which is filled with medium, sealed, and placed in a Styrofoam box for shipment at room temperature. Though this is perhaps the easiest method there are several problems with it. First, even though the tissue is in a Styrofoam box, exposure to extreme temperature variations (often reaching below freezing during ground or flight transportation) have been recorded in several shipments originating from the University of Minnesota (K. Papas, personal communication), which could irreversibly damage islets. Another problem with the current shipping method is that the islet tissue is sealed in an oxygen-impermeable flask preventing the replenishment of oxygen ($O_2$). Depending on the amount of tissue, its respiration rate, and the duration of shipment portions may be exposed to hypoxia, which could result in a decrease in viability and function (Chapter 3 and 4, [16]). A shipment at 25°C in 300-ml culture medium (similar to a T-75 flask), which has an $O_2$ solubility at 25°C of $1.6 \times 10^{-9}$ mol/cm$^3$ mmHg [16], contains 77 $\mu$mol $O_2$, enough $O_2$ to supply 34,000 islet equivalents (IE) for 24 hr (assuming OCR at 37°C 2.5 fmol/min cell (Chapter 2), 1560 cell/IE [13], OCR at 25°C is 40% of that at 37°C Chapter 4). However, this is likely a gross overestimate since transport limitations, which are significant during culture on polystyrene (Chapter 3), are ignored. Further, for transplantation purposes large numbers of islets
(100-500K) may need to be shipped, which if done using the current method, could expose tissue to severe hypoxia resulting in a large loss in viability.
There is a need for a method that will allow large numbers of islets to be transported for 24 hours with minimal loss of viability and function. In this chapter we discuss a controlled environment container that actively maintains temperature and O\textsubscript{2} level during shipment that was developed and tested in collaboration with Giner, Inc., and the University of Minnesota. An inexpensive alternative shipping method is proposed and feasibility is determined through preliminary experiments and theoretical calculations.
8.2 Giner, Inc., controlled environment container
Giner, Inc., in collaboration with MIT and University of Minnesota, developed a shipping container that controls temperature using thermoelectric heating and cooling, and controls O\textsubscript{2} level using an electrochemical O\textsubscript{2} generator. Oxygen is transferred continuously to the islet tissue throughout shipment through the use of silicone rubber bottom flasks. Similar vessels have been shown to be beneficial for high density culture (Chapter 4 and [60]). We hypothesized that these improvements would maintain a higher tissue quality during shipment than the conventional shipping method. To test this we conducted a shipping experiment comparing islet tissue shipped in the controlled environment container to islet tissue shipped in a Styrofoam box.
8.2.1 Methods
Experimental protocol
Following 48-hour culture human islets isolated at the University of Minnesota were sent to MIT overnight in either a T-25 flask within a Styrofoam box or in the Giner, Inc., controlled environment shipping container set at 72°F with a gas phase 40% O\textsubscript{2}. The amount of tissue and CMRL 1066 medium (Miami formulation, Mediatech Inc.) with 10% FBS (Mediatech Inc.) in each vessel is given in Table 8.1. The manual IE counts were performed at the University of Minnesota. The human islet tissue was analyzed at MIT upon receipt.
DNA content
At MIT, DNA was measured by fluorospectrophotometry [55] using the CyQUANT Cell Proliferation Assay Kit (Molecular Probes, C-7026 Eugene, OR), and a λ DNA standard (Invitrogen) and 50-250 IE/sample. Fluorescence was read at 480 nm excitation and 520 nm
emission wavelengths in a plate reader (Spectra MAX Gemini microplate spectrophotometer, Molecular Devices, Sunnyvale CA).
**Islet enumeration by nuclei counting**
Nuclei were prepared by adding equal 100-µl volumes of sample containing 160 or more IE and of lysis solution (0.1 M citric acid (Sigma) and 1% (v/v) Triton X-100 (Sigma)) to a 1.5 ml microtube [55]. The mixture was incubated at room temperature for 5 min with vortex mixing every 1.5 min, and nuclei were liberated by shearing through a needle. Isolated nuclei were diluted with Dulbecco’s phosphate buffered saline (D-PBS, Invitrogen, Carlsbad, CA) to a concentration no higher than $5 \times 10^5$ nuclei/ml, stained with 7-aminoactinomycin D (7-AAD, Molecular Probes, Eugene, OR), and analyzed using a flow cytometer (Guava Personal Cell Analysis (PCA) system, Guava Technologies, Hayward, CA) to determine the total number of cells with intact nuclei in the sample.
**Oxygen consumption rate (OCR)**
OCR was measured as previously described [39]. Briefly, suspensions containing about 2500 IE/ml in DMEM containing 4.5 g/l glucose and 0.6 g/l L-glutamine supplemented with 100 U/ml penicillin, 100 µg/ml streptomycin, 10 mM HEPES, and no added serum were sealed in a 200-µl stirred titanium chamber (Micro Oxygen Uptake System, FO/SYSZ- P250, Instech Laboratories, Plymouth Meeting, PA) maintained at 37°C. The time dependent pO$_2$ within the chamber was recorded with a fluorescence-based oxygen sensor (Ocean Optics, Dunedin, FL), and data at high pO$_2$ were fit to a straight line. The maximal OCR was evaluated from $OCR = V_{ch} \alpha (\Delta pO_2 / \Delta t)$, where $V_{ch}$ is the chamber volume and $\alpha = 1.27$ nmol/mm·Hg·ml is the Bunsen solubility coefficient for oxygen in medium [16]. OCR measurements were normalized by the measured number of cells (nuclei counting) or DNA content of the sample examined.
**Membrane Integrity**
**7-Aminoactinomycin (7-AAD) sequential staining**
Cell membrane integrity was assessed by differential staining with 7-AAD (Molecular Probes, Eugene, OR) [55]. An aliquot of about 300 islets was re-suspended in 100 µl of D-PBS, and 5 µl of 1 mg/ml 7-AAD and incubated for 20 min at 4°C protected from light. After two washes with 1 ml of D-PBS, cells were disrupted by adding equal volume of lysis solution in D-PBS to the islet suspension and sheared as described for islet enumeration. Labeled nuclei
were counted immediately in the flow cytometer or stored on ice for less than 15 min before counting. A portion of the islet suspension was further stained with 7-AAD, thereby labeling all of the previously unlabelled nuclei, and the total number of nuclei was counted. The fraction of cells with compromised membranes was estimated as the ratio of the initially stained nuclei (first measurement) to the total number of nuclei (second measurement).
**Fluorescein diacetate (FDA)/propidium iodide (PI)**
Immediately after human islet isolation was completed, a 250 µl aliquot was taken from the preparation, re-suspended in 5 ml of PBS solution in a 60 mm Petri dish, and 10 µl of 24 mM fluorescein diacetate (FDA) and 750 µM propidium iodide (PI) was added [58; 62]. The tissue was examined in a light microscope by focusing through the tissue to visually estimate the volume fraction of cells containing nuclei stained red because of loss of membrane integrity.
**Statistics**
Measurements were made with three or more replicates and reported as mean ± SD. Statistical significance was determined using a Student t-test for unpaired data for comparing population means, and for paired data when appropriate.
### 188.8.131.52 Shipment results
Approximately 24 hours after leaving the University of Minnesota, the containers arrived at MIT. The majority of tissue had collected in a 1-2 mm thick mass along the edge of the silicone rubber bottom flask in the controlled environment (Figure 8.1). Because of the size of the flask the tissue could not be observed under the microscope while in the culture vessel. The islets from the T-25 flask in the Styrofoam box were not piled together, but fragmented and suspended in medium with a significant amount of single cells (Figure 8.2).
The total amount of tissue in each container was estimated through DNA quantification and nuclei counting (Table 8.2). Nuclei and DNA measurements indicated a 40% and 80% higher fraction of original tissue collected after shipment in the controlled environment than observed for a solid flask in a Styrofoam box. DNA content of tissue shipped in the controlled environment was also higher than in the Styrofoam container.
Membrane integrity was measured using two assays: sequential nuclei staining with 7AAD and islet vital staining with FDA/PI. Based on 7AAD nuclei staining the tissue shipped in the Styrofoam box had a fraction of cells with intact membranes of 0.90 ± 0.03 while the tissue
shipped in the controlled environment had $0.93 \pm 0.01$. Though FDA/PI is not as quantitative, it revealed necrotic cores representing an estimated 10-30% of the tissue in large islets (most cases 200-$\mu$m or larger) from both shipment methods (Figure 8.3 and Figure 8.4). Smaller islets showed no necrotic cores.
OCR/cell, OCR/DNA, total OCR, and the relative recovery of original viable tissue from each shipment are reported in Table 8.3. The OCR/cell was statistically higher for tissue shipped in the controlled environment compared to the Styrofoam box, a trend similar to that for OCR/DNA though no significance is observed. The total OCR recovered from the silicone rubber bottom vessel was 20 times that from the solid flask though it had only 10 times more viable tissue prior to shipment. The resulting relative fraction of original viable tissue remaining after shipment, where that collected from the controlled environment is set to 1, was 0.49 for the Styrofoam box (Table 8.3).
**Figure 8.1.** (A) Islets piled along the edge in the silicone rubber bottom vessel sent in the controlled environment container. (B) Opposite side of the same culture vessel.
**Figure 8.2.** Islets following shipment in a solid flask in a Styrofoam box overnight.
Figure 8.3. Staining of islets shipped in solid flask in a Styrofoam box. From left to right the pictures are with white light, PI staining, and FDA staining.
Figure 8.4. Staining of islets shipped in controlled environment on silicone rubber surface. From left to right the pictures are with white light, PI staining, and FDA staining.
8.2.3 Discussion
In this study we compared islet tissue shipped in a plastic flask within a Styrofoam container to that shipped in a silicone rubber bottom vessel at a controlled temperature and $O_2$ concentration. This experiment tested the hypothesis that controlling temperature and $O_2$ during shipment results in higher tissue quality following shipment compared to conventional shipping methods.
Tissue surface density is a key parameter that determines tissue viability following culture (Chapter 3 and 4). Due to the difference in vessels and a miscommunication, the tissue sent in the Styrofoam box and the controlled environment was at different surface densities and different volume densities (Table 8.1). The tissue in the controlled environment had both a higher surface density and a higher volume density.
Table 8.1. Conditions for islets shipped in the silicone rubber bottom vessel within the controlled environment container and in the solid T-25 flask in the Styrofoam box.
| Culture vessel | Total IE (ml) | Medium volume (ml) | Surface area (cm²) | Islet concentration (IE/cm³) | Islet surface area (IE/cm²) |
|---------------------------------|---------------|--------------------|--------------------|------------------------------|-----------------------------|
| Silicone rubber bottom vessel | 100,000 | ~500 | 100 | 200 | 1,000 |
| Solid flask | 10,000 | ~100 | 25 | 100 | 400 |
Table 8.2. Total tissue estimation for islets received from both the silicone rubber bottom vessel and in the solid flask.
| Culture vessel | Total IEa | Nuclei count (x 10⁶ nuclei) | DNA content (x 10⁴ ng) | DNA/nuclei (pg/nuclei) | Recovery based on DNAb | Recovery based on nuclei b |
|---------------------------------|-----------|-----------------------------|------------------------|------------------------|------------------------|---------------------------|
| Silicone rubber bottom vessel | 100,000 | 99.3 | 99.4 | 10.0 | 0.98 | 0.64 |
| Solid flask | 10,000 | 7.07 | 5.46 | 7.73 | 0.54 | 0.45 |
a Manual IE counts conducted at the University of Minnesota prior to shipment.
b Assumes that there are 1560 nuclei/IE [13] and 6.5 pg DNA/cell [68].
Table 8.3. OCR/cell, OCR/DNA, total OCR, and the relative recovery of original viable tissue collected from both the silicone rubber and solid bottom dishes after transportation.
| Culture vessel | Total IE<sup>a</sup> | OCR/nuclei (fmol/min cell) | OCR/DNA (nmol/min mg DNA) | Total OCR received (nmol/min) | Relative recovery of viable tissue<sup>b</sup> |
|-------------------------|----------------------|-----------------------------|---------------------------|-------------------------------|-----------------------------------------------|
| Silicone rubber | 100,000 | *2.05 ± 0.10 | 205 ± 17 | 204 ± 10 | 1.0 |
| bottom vessel | | | | | |
| Solid flask | 10,000 | 1.40 ± 0.16 | 181 ± 20 | 10 ± 1 | 0.49 |
<sup>a</sup> Based on manual IE counts conducted at the University of Minnesota prior to shipment.
<sup>b</sup> Sets original viable tissue collected from the shipment in the controlled environment to 1.
* indicates statistical significance (p<0.05) compared to solid flask.
a Manual IE counts conducted at the University of Minnesota prior to shipment.
b Sets original viable tissue collected from the shipment in the controlled environment to 1.
* indicates statistical significance (p<0.05) compared to solid flask.
Despite the higher surface and volume densities, shipping in the controlled environment had several beneficial effects. It increased tissue recovery 40 to 80% from that obtained with the traditional method based on nuclei and DNA measurements (Table 8.2), respectively. OCR/cell was higher for tissue shipped in the controlled environment compared to the Styrofoam container indicating an increased viability (Table 8.3). OCR/DNA indicated a similar trend, but the difference was not significant. The reason for this discrepancy is indicated by the difference in DNA content. The DNA content was higher for the islets shipped in the controlled environment (10 pg/cell) as opposed to those shipped in a solid flask (7.7 pg/cell, Table 8.2), which had a DNA content closer to the expected value of 6.5 pg/cell reported in literature [68]. DNA content > 6.5 pg/cell is indicative of cells that have died, broken down the nucleus, and released their DNA, but the DNA has not yet degraded. Though this suggests that the tissue in the controlled environment has been damaged, tissue recovery and viability measurements (membrane integrity, OCR/cell, and OCR/DNA) are at least as good if not higher than that of tissue from the Styrofoam box.
Total OCR, which indicates the total amount of viable tissue, was the most encouraging. When converted to the relative fraction of viable tissue recovered after shipment, it indicated that shipment in the controlled environment resulted in double the amount of viable tissue after shipment compared to the Styrofoam box even though it was shipped at a higher surface density (Table 8.3). Thus controlling temperature and O$_2$ level throughout shipment results in a 100% increase in the fraction of original viable tissue collected after shipment compared to the conventional shipping method.
Though shipment in the controlled environment resulted in higher tissue recovery, viability, and viable tissue recovery than the Styrofoam container, it still can be improved. As indicated by islet vital staining necrotic cores developed in larger islets in both containers. Necrotic cores usually result from nutrient transport limitations at the center of the islet (Chapter 3 and [16]). In previous studies O$_2$ has been shown to be the limiting nutrient for islets in culture on polystyrene (Chapter 3). Though O$_2$ limitations were expected for islets shipped in a Styrofoam container they were not expected for the islets on a silicone rubber membrane since O$_2$ transport through the membrane allows islets at high surface densities to maintain a higher viability than they otherwise could on a plastic dish (Chapter 4). There are at least three possible explanations that could account for the necrotic cores observed in islets on the silicone rubber
bottom dish. First, the mass of tissue that had accumulated on one side of the flask prior to receipt (Table 8.1) could have increased O₂ limitations to the point where portions were exposed to hypoxia as was indicated in Chapter 5 for stacked islets on a silicone rubber membrane. Second, glucose or some other nutrient could become limiting at high surface densities when O₂ limitations have been removed as shown in Chapter 6.
In summary, despite the significantly higher islet surface density in the controlled environment the fraction of original tissue, OCR/cell, and relative fraction of original viable tissue collected after shipment was significantly higher in the silicone rubber bottom flask within the controlled environment compared to a plastic flask within a Styrofoam box demonstrating that the controlled environment container is a substantial improvement for islet transportation.
However, the controlled environment device is large, weighs about 30lbs, requires battery operation, and would expensive for isolation centers to buy. Thus, even though it improves the quality of tissue following transportation, it is not a feasible option for most islet centers. As a result, an inexpensive alternative to this actively controlled environment is desirable. The alternative container should eliminate large temperature swings and provide sufficient O₂ throughout shipment like the Giner, Inc., device, but it should be less cumbersome, easier to use, and less expensive.
### 8.3 Passive shipping device
The controlled environment container developed by Giner Inc in collaboration with the MIT and the University of Minnesota prevented large temperature swings, generated O₂, and allowed the generated O₂ to be transported to the tissue throughout shipment. Here we propose a passive shipping container that will perform similar functions, but be much smaller, significantly less expensive, and require no battery for operation. The passive shipping device proposed is illustrated in Figure 8.5. Like the Giner, Inc., controlled environment container it uses a culture vessel with a silicone rubber bottom to allow transport of oxygen from the surrounding environment to the tissue throughout shipment. However, instead of using thermoelectric heating and cooling it utilizes phase change materials (PCMs) to control temperature during shipment. Using theoretical calculations and proof of concept experiments we demonstrate the feasibility of the modified container, which can be used as a basis for future work.
Figure 8.5. Schematic of passive shipping device that will provide sufficient oxygen and maintain temperature during shipment.
8.3.1 Oxygen
The amount of tissue, shipment temperature, and transport limitations within the shipping device determine the $O_2$ requirements needed to keep tissue viable during shipment. Here we consider several possible shipping conditions to determine what limitations our passive shipping container may have with regards to $O_2$ supply.
Islet tissue cultured on a 100 to 500-$\mu$m thick silicone rubber membrane at 37°C 142 mmHg $O_2$ maintains viability until an OCR density of 5 nmol/min cm$^2$ above which the fraction of original viable tissue after culture decreases due to limitations other than $O_2$ (Chapter 4). Thus by utilizing vessels with silicone rubber membrane bottoms, we can ship at OCR densities $\leq$ 5 nmol/min cm$^2$ or, assuming that tissue has an OCR of about 2.5 fmol/min cell (Chapter 2), 1280 IE/cm$^2$. In order to ship islets from an entire isolation containing a tissue volume equivalent to 500,000 IE a surface area of 400 cm$^2$ is needed.
The OCR of islet tissue decreases exponentially as inverse temperature increases following the equation
$$V_{\text{max}} = V_o \cdot e^{-\frac{E}{RT}}$$
where $V_o$ is 1.86 x $10^3$ mol/min cm$^3$, E is 53 kJ/mol, and R is the gas constant (0.008314 kJ/mol K) (Chapter 4). Assuming the container holds temperature throughout
shipment, the amount of \( \text{O}_2 \) required for differing amounts if islet tissue over a given shipment period were estimated using Equation 8.1 as indicated in Figure 8.6 for a 36 hr shipment. At 37°C a tissue volume equivalent to 500,000 IE would require 4.2 mmol of \( \text{O}_2 \) which, if an ideal gas, is 108 cm\(^3\) of pure \( \text{O}_2 \) at 1 atm or 520 cm\(^3\) of air containing 20.9% \( \text{O}_2 \). Lower temperatures and shorter times make the requirements for adequate \( \text{O}_2 \) supply even less.
Additionally, the partial pressure of \( \text{O}_2 \) (p) in the container should not fall below 142 mmHg thereby ensuring the same results as observed previously with silicone rubber membranes (Chapter 4). Equation 8.2 allows calculation of the volume of the gas phase required in order to have enough \( \text{O}_2 \) for tissue consumption and that the p\( \text{O}_2 \) within the container does not drop below 142 mmHg (assuming \( \text{O}_2 \) is an ideal gas).
\[
V = \frac{n_{\text{islet}} R_{\text{gas}} T}{p_{\text{initial}} - p_{\text{end}}}
\]
where \( n_{\text{islet}} \) is the \( \text{O}_2 \) required by the tissue during shipment (Figure 8.6), \( R_{\text{gas}} \) is the gas constant, \( T \) is the temperature inside the container, \( V \) is the gas phase volume in the container, \( p_{\text{end}} \) is the desired ending p\( \text{O}_2 \) in the container after shipment (142 mmHg), and \( p_{\text{initial}} \) is the initial p\( \text{O}_2 \) in the container. If 500,000 IE were shipped, which requires 4.2 mmol \( \text{O}_2 \) for consumption over 36 hr, a 4.8 L volume of ambient air (159 mmHg, 20.9% \( \text{O}_2 \)) or 0.6 L of a 40% \( \text{O}_2 \) mixture is needed within the shipping container to ensure that the p\( \text{O}_2 \) in the container would not drop below 142 mmHg.
### 8.3.2 Temperature
Phase change materials (PCMs) are materials that absorb or liberate heat at constant temperature as the material changes from solid to liquid or vice versa. Ice packs are PCMs designed to operate at 0°C. Though several PCMs are available commercially, we will consider PCMs available from TCP Reliable and PCM Energy P. Ltd. that range in phase change temperature from 18–36°C (Table 8.4) [175; 176].
Figure 8.6. The amount of $O_2$ consumed by 50,000-500,000 IE for 36 hr at temperatures ranging from 10 to 40°C.
Table 8.4. Commercially available phase change materials and their properties.
| Product | Phase Change Temperature (°C) | Latent heat (J/g) | Thermal conductivity (Watt/m °C) | Specific Heat (J/g °C) | Specific gravity |
|------------------|-------------------------------|-------------------|---------------------------------|------------------------|-----------------|
| **PCM Energy P. Ltd.** | | | | | |
| Latest™18T | 17-19 | 175 | 1 | 2 | 1.48-1.50 |
| Latest™20T | 19-20 | 175 | 1 | 2 | 1.48-1.50 |
| Latest™25T | 24-26 | 175 | 1 | 2 | 1.48-1.50 |
| Latest™25S | 24-26 | 200 | 0.6 | 2 | 1.45 |
| Latest™29T | 28-30 | 175 | 1 | 2 | 1.48-1.50 |
| Latest™30S | 29-31 | 230 | 0.6 | 2 | 1.45 |
| Latest™36S | 34-36 | 260 | 0.6 | 2 | 1.45 |
| **TCP Reliable** | | | | | |
| Phase 22 | 22 | - | - | - | - |
| Exo-Gel | 27 | - | - | - | - |
*Information taken from [175; 176].*
Simple theoretical calculations can determine whether it is feasible to control the temperature inside our container for 24 hr with PCM. For simplicity assume the temperature
inside (i.e. the phase change temperature) and surrounding the container are constant throughout shipment and that the PCM thermal conductivity is negligible compared to the Styrofoam box, which is 3 inches thick (thermal conductivity of 0.033 W/m K [118]). The heat flux through a one-dimensional solid is calculated using Fourier’s law (8.3) [97].
\[ q = k \nabla T = k \frac{dT}{dx} \]
where \( T \) is temperature, \( q \) is the heat flux, and \( k \) is the thermal conductivity.
The volume of the void space in the container needs to be 4.8 L as determined above for 500,000 IE, which also requires a vessel with a silicone rubber bottom surface area of 400 cm\(^2\) and a 2.2 cm medium depth (medium depth used in study with silicone rubber dishes Chapter 4) for a total medium volume of 0.88 L. To be safe assume that 1.5 L is large enough to account for the vessel containing the tissue, the tissue itself, and the medium within the vessel. Thus the cargo space of the container needs to be about 6.3 L (25 cm x 25 cm x 11 cm).
Now assume that for practical reasons no more than about 10 lbs of PCM (3.0-3.2 L) should be used such that the container weight stays manageable. If panels are used to cover all four sides, top, and bottom of the container, then a thickness of 1.2 cm is needed such that a volume of 3.2 L is held in the panels. For simplicity assume 2 cm thick panels to account for the material that contains the PCM. Thus the cargo space of the Styrofoam box is about 12.6 L (29 cm x 29 cm x 15 cm). If the thickness of the Styrofoam is 3 inches, then the dimensions of the box are about 44 cm x 44 cm x 30 cm resulting in a surface area of 9300 cm\(^2\).
If we assume a constant heat flux (8.3) throughout shipment, the amount of PCM required to maintain the phase change temperature inside the container with a constant external temperature for a 24 hr shipment can be determined by 8.4.
\[ \Delta H \cdot m = q \cdot \tau \cdot A \]
where \( m \) is the mass of PCM, \( \Delta H \) is the latent heat, \( A \) is the surface area of the container, and \( \tau \) is the duration of shipment. Figure 8.7 illustrates the volume of PCM required as a function of the ambient temperature. Based on this simple analysis 3.2 L is a feasible amount of PCM for all phase change temperatures considered. The only case where the PCM would be unable to maintain temperature was when the ambient temperature is below 2°C for the entire shipment.
and the desired container temperature was 36°C. Even in this case, the passive shipping device would still maintain temperature much better than the traditional Styrofoam box.
Preliminary experiments were conducted at constant temperatures to confirm experimentally that temperature control with PCM is feasible. A 27.5 cm x 23 cm x 20.3 cm Styrofoam box 1.5 inches thick was placed in a 37°C incubator overnight. HOBO 2K Temp Data Loggers were placed inside and outside recording temperature every 2 minutes. The data logger inside was sitting on the bottom of the box. Following incubation at 37°C the box was moved to 12°C. After overnight incubation at reduced temperature the data loggers were collected and the data analyzed. For some experiments a single Phase 22 PCM panel provided by TCP Reliable was added to the Styrofoam box. The panel was 13.2 cm x 13.2 cm x 2.5 cm and is estimated to contain 250-300 ml of Phase 22 PCM. When the PCM panel was added to the box the data logger inside the box was placed on top of the panel.
Figure 8.7. Volume of various PCMs required to maintain the phase change temperature of the container (18, 25, 30, or 36°C) for 24 hr during shipment at constant ambient temperature.
A sample temperature profile inside and outside the box when ambient temperature dropped to 12°C is given in Figure 8.8 for experiments with and without PCM. When PCM is used the temperature inside the container is maintained at 18-20°C for an extended period of time. Without PCM the temperature drops continuously until it reaches the ambient temperature. This proof of concept experiment demonstrates how PCM can be beneficial and how Styrofoam alone does not prevent temperature fluctuations.
Using Fourier’s law the latent heat of the PCM can be calculated for these experiments. Based on data from three experiments, the Phase 22 PCM has a latent heat of about $900 \pm 80$ J/ml PCM. If the Phase 22 PCM has a specific gravity similar to the other PCMs in Table 8.4, then the latent heat would be about 600 J/g which is much higher than the latent heat listed for the other PCMs in Table 8.4. This large latent heat probably resulted from the over simplified assumptions of the calculation. For example, we did not account for the heat conduction though the PCM or from the Styrofoam through the air to the PCM.
**Figure 8.8. Sample profile for controlled PCM experiment.**
Containers with and without PCM were equilibrated at 37°C before being transferred to a 12-14°C environment. Temperature inside and outside the container was monitored.
8.4 Conclusions
Temperature and $O_2$ control has beneficial effects on islet recovery and viability following shipment as demonstrated using the Giner, Inc., controlled environment container. The passive shipping device described has all the benefits of the Giner, Inc., controlled environment container, but is less cumbersome, less expensive, and does not require battery operation. As such further work is warranted to create a passive shipping device that can be used for transporting islet tissue.
8.5 Nomenclature
| Symbol | Description |
|--------|-------------|
| 7-AAD | 7-aminoactinomycin D |
| A | Surface area of the container |
| C | Concentration of $O_2$ |
| D | Diffusivity of $O_2$ |
| D-PBS | Dulbecco’s phosphate buffered saline |
| E | Activation energy |
| FBS | Fetal bovine serum |
| FDA | Fluorescent diacetate |
| $F_V$ | The fraction of original viable tissue after culture |
| IE | Islet equivalent |
| k | Thermal conductivity |
| LDH | Lactate dehydrogenase |
| m | Mass of PCM |
| N | $O_2$ flux |
| $O_2$ | Oxygen |
| OCR | $O_2$ consumption rate |
| p | $O_2$ partial pressure |
| PCM | Phase change material |
| $P_{gas}$ | Ambient partial pressure of $O_2$ |
| PI | Propidium iodide |
| q | Heat flux |
| Symbol | Description |
|--------|-------------|
| R | Gas constant |
| $R_{\text{islet}}$ | Radius of islet |
| T | Temperature |
| t | Time |
| V | Volume |
| $V_{\text{max}}$ | Maximum reaction velocity |
| $V_0$ | Constant in Arrhenius equation |
| $\alpha$ | Solubility of $O_2$ in medium |
| $\Delta H$ | Latent heat |
| $\tau$ | Duration of shipment. |
9 Conclusions and outlook
9.1 Maintaining islet quality during culture
This project was based on the hypothesis that islet tissue is exposed to oxygen transport limitations during conventional culture and shipment that result in decreases in viable tissue suggesting that removal of these limitations will improve the tissue quality after culture and/or shipment.
A detailed study investigating how tissue changes during culture in the absence of oxygen limitations was conducted providing a complete understanding of islet and non-islet tissue behavior prior to and during culture at 37°C and 24°C in terms of the recovery of tissue, its viability, and its purity. The portion most relevant to the remainder of the thesis relates to the cultured tissue. The data revealed that a significant amount of tissue adhered to the polystyrene surfaces. The adherent tissue fraction was shown to contain fewer β cells and decrease as tissue surface density increases. Though the data confirms current practice of not collecting adherent tissue for transplantation, it does give additional insight into changes the islet and non-islet tissue go through during culture that has not been shown previously for human islet tissue.
A key conclusion from this study investigating culture in the absence of oxygen limitations that set the stage for the work that followed was that when adherent and non-adherent were combined, nearly 100% of the original viable tissue placed in culture was recovered from preparations that were given time to recover from the harsh isolation process. Though original tissue including β cells was lost, the tissue that was lost was dead (as determined by OCR) when culture was initiated. Increasing tissue surface density resulted in the appearance and amplification of oxygen limitations resulting in drastic decreases in the fractional recovery of viable tissue. The data suggest that current clinical practice likely imposes oxygen limitations during culture that can result in a decrease in the fractional recovery of viable tissue by 0.20.
The oxygen limitations present using conventional methods were reduced through culture in elevated ambient oxygen, at a reduced temperature, or on a silicone rubber membrane allowing tissue to be cultured at higher densities (at least 3 fold higher) without a reduction in viability. However, questions remain about oxygen toxicity and the effects of reduced culture temperature on islet function. Culture on silicone rubber was the most promising resulting in the complete removal of oxygen limitations. Though other limitations became important at very high
densities on silicone rubber membranes, culture at densities 5 fold higher than conventional methods was possible with little loss in viable tissue. In addition to developing new culture techniques, an in-depth theoretical analysis of oxygen transport revealed that tissue size, islet stacking, and clumping, all of which may be present during high density culture and islet shipment, can greatly exacerbate oxygen limitations.
When oxygen transport limitations are removed glucose limitations likely become limiting during both culture on polystyrene and silicone rubber. Conventional culture (3 mm medium, polystyrene surface) may performed at a total tissue volume up to 570 “IE”/cm$^2$, where “IE” in this case is based on total tissue (400 IE/cm$^2$, 70% islet tissue by volume), which could result a glucose concentration < 1 mM within 30 hr of culture. Culture in large medium depths prevents glucose depletion from the bulk medium, but because glucose consumption may be high relative to the rate of diffusion, diffusional limitations to the islet may result in exposure to concentrations < 1 mM at densities above 1700 IE/cm$^2$, which is a similar surface density to that where the empirical fractional recovery of viable tissue began to decrease consistently on silicone rubber membranes.
Changes in local pH can be significant during culture on polystyrene due to the buildup of CO$_2$ in the medium. However, at low density these changes are small, especially at low medium depths like those used in conventional culture. Changes in pH during culture on a silicone rubber membrane are unlikely because CO$_2$ can escape the system at a much faster rate thereby preventing a shift to more acidic conditions through the bicarbonate reaction. As a result drastic changes in pH during culture are unlikely though there is much uncertainty in many of the parameters used in the theoretical model.
Using the knowledge gained investigating oxygen transport, a shipping container was developed and tested in collaboration with Giner, Inc., and the University of MN that actively controlled temperature and oxygen resulting in increases in islet recovery and viability following 24 hr shipment as compared to the “state-of-the-art” method using a Styrofoam box. A passive shipping device was outlined that has all the benefits of the actively controlled environment container, but is less cumbersome, less expensive, and does not require battery operation.
The culture techniques examined greatly improve tissue oxygenation compared to conventional culture allowing tissue to be cultured at high densities without a reduction in viability. The methods are easily implemented allowing clinical centers to reduce space and
handling requirements during culture prior to transplantation without the reduction in islet viability that can occur using conventional methods thereby maximizing the use of limited islet resources likely improving transplantation success rates.
9.2 Future work
The concepts presented here revealed many more avenues that could be investigated further. First transplantation studies would provide confirmation that the developed techniques in this thesis improve or at least maintain tissue viability and function compared to the conventional method.
Another beneficial study would involve assessing islet quality as a function of time starting immediately after isolation and follow it through culture. Based on the data in Chapter 2 we hypothesize that the islet isolation process damages tissue thereby initiating the cell death process. The damage incurred by the isolation process was partially quantified in Chapter 2, but measurement earlier in culture may provide insight into the dynamics of the death process initiated by isolation. Additionally, longer culture periods (>2 days) would reveal whether tissue quality is maintained or eventually begins to decrease even in the absence of transport limitations.
The quantification of tissue adherence had not previously been reported in literature. Though we presented data indicating that the adherent tissue contained fewer β cells, the identity of the cells that adhered was not determined. Further insight into which cells adhere to the culture vessel may lead to techniques that remove impure tissue. An initial study could identify the cells that adhere and test whether adherence occurs multiple times with the same preparation when the culture vessel is changed.
Reduced culture temperature was shown to be beneficial for high density culture, but questions remained about whether tissue function was changed since low density culture reduced the OCR of the tissue. Transplantation experiments would be best to determine whether tissue cultured at reduced temperature have the same potency of those cultured at 37°C. Variations in recovery period following reduced temperature culture would show if function is lost and whether it ever recovers. For these studies all cultures should be conducted in the absence of oxygen limitations for culture at 37°C.
Oxygen toxicity in islet tissue could also be investigated further especially if culture in elevated ambient oxygen is applied. The study would be best conducted using silicone rubber membranes allowing for accurate control of oxygen level at the tissue surface. The use of small islets or islet cell-aggregates will reduce internal oxygen gradients resulting in a more uniform exposure to the desired $O_2$ level. The theoretical oxygen transport model already developed can be used to confirm the uniform oxygen exposure.
Based on the analysis in this thesis the development of a passive shipping device that provides sufficient oxygen and prevents large temperature swings is feasible. The next step in developing such a container is to obtain the materials for a prototype and then fabricate the shipping container. PCMs for multiple temperatures could be tested until an optimal shipment temperature is determined. The container should be tested in controlled environments, during shipment without tissue, and during shipment with tissue. All shipments should be compared to the standard shipping method of a Styrofoam box. One problem that has not been addressed with this passive device is islet pooling during shipment. If the container is slanted, pooling will result exacerbating nutrient transport limitations as was observed with the Giner, Inc., container. A possible solution to this is to combine the technology discussed here with a shipping device developed by the University of Miami that prevents tilting of the culture vessel.
10 Appendices
10.1 Supplemental material for Chapter 2
Purity measurements on day 2 were performed using immunoperoxidase staining for insulin in conjunction with stereological point counting while day 0 measurements were made using morphological analysis. The reason a second method was used to determine purity following culture was because several samples contained single cells following culture that resulted from trypsinization of the adherent tissue. Morphology differences in single cells would be hard to differentiate whereas differences in color based on insulin staining are easier to distinguish resulting in a more accurate reflection of tissue composition. For insulin staining quantification by stereological point counting vascular space was excluded because vascular space in non-adherent tissue will be different than in adherent tissue. These measurements resulted in the $\beta$ cell volume fraction exclusive of vascular spaces ($\Phi_{\beta XV}$), which is defined as
$$\Phi_{\beta XV} = \frac{P_\beta}{P_I + P_{NI}} = \frac{V_\beta + V_{\beta INT}}{V_I + V_{NI} - V_{VI}}$$
where $P_\beta$ is the points counted as insulin positive, $P_I$ is a point counted as islet tissue, $P_{NI}$ is a point counted as non-islet tissue, $V_\beta$ is the volume of insulin positive cells, $V_{\beta INT}$ is the volume of the interstitial space of insulin positive cells, $V_I$ is the volume of islet tissue, $V_{NI}$ is the volume of non-islet tissue, and $V_{VI}$ is the volume of the vascular void space. If $P_\beta$ is only counted if insulin staining is present, which should only occur in cells and not interstitial space (i.e. $V_{\beta INT} = 0$), then the following is true.
$$\Phi_{\beta XV} = \frac{V_\beta}{V_I + V_{NI} - V_{VI}}$$
In order to compare the morphological measurements made on day 0 for the volume fraction islets ($\Phi_I$) with $\Phi_{\beta XV}$ from day 2 samples we needed to perform some data conversions, which are summarized in Figure 10.1. Previously it has been shown that $\Phi_I$ can be converted to the islet volume fraction exclusive of vascular spaces ($\Phi_{IXV}$) assuming the vascular void fraction ($\Phi_{VI}$) in the islet tissue is known (equation 6.4 from [13]). $\Phi_{IXV}$ can be converted to $\Phi_{\beta XV}$ using
\[ 1 + \frac{V_{\text{INT}}}{V_I} = \frac{\phi_{\text{BIC}} \Phi_{\text{IXV}}}{\Phi_{\beta \text{XV}}} \tag{10.3} \]
where \( V_{\text{IC}} \) is the volume if islet cells, \( V_{\text{INT}} \) is the volume of interstitial space in islet cells, and \( \phi_{\text{BIC}} \) is the \( \beta \)-cell cell volume fraction. This equation can be obtained by combining Equation 10.2 and the definition of \( \Phi_{\text{IXV}} \), which is
\[ \Phi_{\text{IXV}} = \frac{P_I}{P_I + P_{\text{NI}}} = \frac{V_{\text{IC}} + V_{\text{INT}}}{V_I + V_{\text{NI}} - V_{\text{VI}}} \tag{10.4} \]
From previous analysis fresh tissue has an average \( \phi_{\text{BIC}} = 0.867 \), \( \frac{V_{\text{INT}}}{V_I} = 0.15 \), and \( \frac{V_{\text{IC}}}{V_I} = 0.71 \) [13] thereby allowing us to estimate \( \Phi_{\beta \text{XV}} \) of fresh tissue from measurements of \( \Phi_I \).
In order to combine nuclei measurements with purity measurement \( \Phi_{\beta \text{XV}} \) and \( \Phi_I \) need to be converted from volume fractions to number fractions. In particular we will convert each volume fraction to the number fraction beta cells (\( f_B \)). Day 0 measurements of \( \Phi_I \) were converted to \( f_B \) as outlined in Figure 10.2. \( \Phi_I \) was converted to the volume fraction islets (\( \phi_{\text{IC}} \)) which was converted to the number fraction islet cells (\( f_{\text{IC}} \)) which was converted to \( f_B \) using previously defined relationships and assuming the volume fraction of islet extracellular space (\( \Phi_{\text{IEC}} \)), the volume fraction of non-islet extracellular space (\( \Phi_{\text{NIEC}} \)), the average cell volume for islet cells (\( \bar{v}_{\text{IC}} \)), the average cell volume of all cells (\( \bar{v}_{\text{TC}} \)), the \( \beta \) cell volume (\( v_B \)), and \( \phi_{\text{BIC}} \) are the same as previously reported [13].
Day 2 measurements of \( \Phi_{\beta \text{XV}} \) were converted to \( f_B \) using the procedure outlined in Figure 10.3. \( \Phi_{\beta \text{XV}} \) is converted to \( \Phi_{\text{IXV}} \) using Equation 10.3 and equation 6.20 from Colton et al [13]. Values for \( \frac{V_{\text{INT}}}{V_I} \), \( \frac{V_{\text{IC}}}{V_I} \), \( v_B \), \( f_{\text{BIC}} (f_B/f_{\text{IC}}) \), and \( \bar{v}_{\text{IC}} \) are estimated as indicated in Table 10.1. \( \Phi_{\text{IXV}} \) is converted to \( \Phi_I \) which is then converted to \( \phi_{\text{IC}} \) using previously defined relationships and assuming values for \( \Phi_{\text{IEC}} \), \( \Phi_{\text{NIEC}} \), and \( \Phi_{\text{VI}} \). \( \phi_{\text{IC}} \) is converted to \( f_{\text{IC}} \) using equation 6.18 from Colton et al [13]. \( f_B \) is calculated from \( f_{\text{IC}} \) by assuming \( f_B/f_{\text{IC}} = 0.739 \) (i.e. \( f_{\text{BIC}} \) does not change with culture).
Figure 10.1. Conversion of day 0 islet volume fraction to β cell volume fraction exclusive of vascular spaces.
Figure 10.2. Conversion of day 0 islet volume fraction to number fraction β cells.
Figure 10.3. Conversion of day 2 β cell volume fraction exclusive of vascular spaces to number fraction β cells.
Table 10.1. Parameter values used to estimate $\Phi_{\beta XV}$ and $f_\beta$ from day 0 measurements of $\Phi_i$ and day 2 measurements of $\Phi_{\beta XV}$.
| Assumed | Day 2 non-adherent | Day 2 adherent |
|---------|-------------------|----------------|
| | Fresh value | Scenario 1 | Scenario 2 | Scenario 1 | Scenario 2 |
| $V_{INT}/V_i$ | 0.15 | 0.15 | 0.1 | 0.1 | 0.05 |
| $V_{IC}/V_i$ | 0.71 | 0.78 | 0.9 | 0.9 | 0.95 |
| $\Phi_{VI}$ | 0.138 | 0.07 | 0 | 0 | 0 |
| $\Phi_{NIEC}$ | 0.19 | 0.17 | 0.1 | 0.1 | 0.05 |
| $\Phi_{IEC}$ | 0.29 | 0.22 | 0.1 | 0.1 | 0.05 |
| $V_\beta$ | 950 | 975 | 1000 | 975 | 1000 |
| $V_{NB}$ | 400 | | | | |
| $V_A$ | 1200 | | | | |
| $V_D$ | 200 | | | | |
| $V_O$ | 200 | | | | |
| $f_O$ | 0.036 | | | | |
| $f_N/f_D$ | 1.1 | | | | |
| $f_{\beta IC}$ | 0.736 | | | | |
Calculated averages
| | Day 2 non-adherent | Day 2 adherent |
|---------|-------------------|----------------|
| | Fresh value | Scenario 1 | Scenario 2 | Scenario 1 | Scenario 2 |
| $\bar{V}_{IC}$ | 806 | 823 | 841 | 823 | 841 |
| $\bar{V}_{TC}$ | 745 | 763 | 767 | 724 | 726 |
| $f_A$ | 0.251 | 0.20 | 0.23 | 0.40 | 0.41 |
| $f_D$ | 0.229 | 0.18 | 0.21 | 0.37 | 0.37 |
Though we have a value for $f_\beta$ we need to make sure all variables are consistent. We assume that the number fraction of other cells ($f_O$) is the same as fresh tissue thereby suggesting that the change in $f_\beta$ and $f_{IC}$ (which allows determination of the number fraction non-$\beta$ islet cells, $f_{NB}$) is offset by a change in exocrine tissue, specifically the number fraction of acinar cells ($f_A$) and duct cells ($f_D$) such that $f_{NB} + f_D + f_O + f_\beta + f_A = 1$. This change in exocrine tissue is assumed to be equally distributed between acinar and duct cells such that $f_A/f_D$ is constant (personal communication from Susan Bonnier-Weir). Using the number fraction of each cell type and the assumed cell volumes of each, $\bar{V}_{IC}$ and $\bar{V}_{TC}$ can be calculated as previously indicated [13]. These calculated values are compared to those originally assumed in the calculation. If unequal,
the new values replace the old ones and the calculation is performed again. The loop continues until the calculated $\overline{v}_{IC}$ and $\overline{v}_{TC}$ matches those assumed.
The parameters used in the above calculations for fresh tissue and tissue following 2 day culture, both non-adherent and adherent, are shown in Table 10.1. The values of the parameters for fresh tissue (day 0) were determined previously [13]. During culture these parameters are likely to change. Since the degree of change is unknown, we will consider two different scenarios, one that we feel is realistic and one that has more extreme changes in parameter values allowing us to see how sensitive the analysis is to the assumptions.
For non-adherent tissue in scenario 1 we assume that the vascular void space is cut in half during culture for both islet and non-islet tissue, but that the volume fraction of the interstitial space in an islet does not change. In scenario 2 we assume that all the vascular void space is gone and that $\frac{V_{INT}}{V_I}$ decreases from 0.15 to 0.1. These assumptions allow us to calculate $\frac{V_{IC}}{V_I}$, $\Phi_{IEC}$, $\Phi_{VI}$, and $\Phi_{NIEC}$. In scenario 1 and 2 we assume that the $v_\beta$ increases to 975 and 1000 $\mu m^3$, respectively based on observations from our collaborators that show a slight increase in size of $\beta$ cells during culture (personal communication from Susan Bonner-Weir). For adherent tissue on day 2 we assume that all vascular void space is gone since trypsinization is conducted. We also assume that $\frac{V_{INT}}{V_I}$ decreases to 0.1 and 0.05 for scenarios 1 and 2 allowing $\frac{V_{IC}}{V_I}$, $\Phi_{IEC}$, $\Phi_{VI}$, and $\Phi_{NIEC}$ to be calculated. $v_\beta$ is the same as that for non-adherent tissue (975 and 1000 $\mu m^3$ for scenario 1 and 2). The results presented were determined using scenario 1. When performing the calculations with the two scenarios there was at most a 7% change in $f_\beta$ (absolute change of 0.027).
It should be noted $\frac{V_{INT}}{V_I}$ determined for fresh human islet tissue did not include capillary space, which is normally considered interstitial space. In the original study the pericapillary space was counted as vascular space. This brings into question the use of its value in our data conversions. Despite this the value of $\frac{V_{INT}}{V_I}$ for fresh tissue (0.15) is inline with estimates for the mean interstitial space in the human body [177] and the error associated with it is likely small.
Once $f_\beta$ was determined the fraction of original $\beta$ cells collected from culture ($F_\beta$) could be calculated using
$$F_\beta = \frac{\left(n_{TC} f_\beta\right)_2}{\left(n_{TC} f_\beta\right)_0}$$ 10.5
where $n_{TC}$ is the total number of cells either placed in culture (day 0) or collected from culture (day 2).
### 10.1.1 Nomenclature
| Symbol | Description |
|--------|-------------|
| $f_A$ | Number fraction acinar cells |
| $f_D$ | Number fraction duct cells |
| $f_{N\beta}$ | Number fraction non-beta islet cells |
| $f_O$ | Number fraction other cells |
| $f_{IC}$ | Number fraction of all islet cells |
| $f_\beta$ | Number fraction of $\beta$ cells among all cells |
| $f_{\beta IC}$ | Number fraction beta cells in among islet cells |
| $F_\beta$ | Fraction of original $\beta$ cells collected from culture |
| $n_i$ | Number of cells of type i |
| $n_{TC}$ | Total number of cells |
| $P_\beta$ | Number of points falling within the domain of $\beta$ cells |
| $P_I$ | Number of points falling within the domain of islets |
| $P_{NI}$ | Number of points falling within the domain of non-islets tissue |
| $P_V$ | Number of points falling within the domain of islet vascular space |
| $V_{\beta INT}$ | Volume of interstitial space of $\beta$ cells |
| $V_I$ | Total volume of the islet domain |
| $V_{IC}$ | Volume of islet cells |
| $V_{INT}$ | Volume of interstitial space in islet |
| Symbol | Description |
|--------|-------------|
| $V_{NI}$ | Total volume of the non-islet domain |
| $V_{VI}$ | Volume of vascular voids in islet |
| $V_{\beta}$ | Volume of $\beta$ cells |
| $\phi_{IC}$ | Volume fraction of islet cells among all cells |
| $\phi_{\beta IC}$ | Volume fraction of $\beta$ cells among all islet cells |
| $\Phi_{\beta XV}$ | Volume fraction of $\beta$ cells exclusive of islet vascular spaces |
| $\Phi_{I}$ | Volume fraction of islets |
| $\Phi_{IEC}$ | Volume fraction of islet extracellular space within the islets |
| $\Phi_{IXV}$ | Volume fraction of islets exclusive of islet vascular spaces |
| $\Phi_{NIEC}$ | Volume fraction of non-islet extracellular space based on the total non-islet volume |
| $\Phi_{VI}$ | Volume fraction of vascular voids within islets |
| $v_A$ | Cell volume of acinar cells |
| $v_D$ | Cell volume of duct cells |
| $\bar{v}_{IC}$ | Average cell volume for islet cells |
| $v_{N\beta}$ | Cell volume of non-$\beta$ islet cells |
| $v_O$ | Cell volume of other cells |
| $\bar{v}_{TC}$ | Average cell volume for all of the cells |
| $v_{\beta}$ | Cell volume of $\beta$ cells |
10.2 Carbon monoxide (CO) treatment of βTC3 cells exposed to TNF-α or anoxia
10.2.1 Introduction
Successive implants of islets of Langerhans from two or more human pancreas are normally required to treat type 1 diabetes [8] though a small fraction of the islets in the normal human pancreas are required to maintain normoglycemia [9]. Stresses during islet isolation, islet culture, and following transplantation can cause a reduction in tissue viability and function. Cell death can be triggered by a variety of stresses, including hypoxia, oxidative stress, and pro-inflammatory cytokines. A method for protecting islets from these and other stresses would provide more efficient and effective use of this limited resource.
Exposure to carbon monoxide (CO), a product of heme degradation catalyzed by heme oxygenase-1 (HO-1), which itself acts as a cellular defense for many types of cells in vitro [88; 178-190], has been shown to protect a variety of cell types from various stresses. CO provides protection against hyperoxia [191-193], serum deprivation [194], cytokines such as TNF-α [194-196], and hypoxia-reoxygenation [197]. A limited amount of work has been carried out to study the effects of CO on β-cells. In a study by Gunther et al. [194], CO (10,000 ppm) protected βTC3 cells from apoptosis induced by TNF-α, serum starvation, and etoposide, and mouse islets from TNF-α and cyclohexamide-induced apoptosis. A two-hour CO treatment of islets in medium presaturated with CO resulted in a significantly shorter time to reach normoglycemia after transplantation with a marginal islet mass compared to islets not given the CO treatment [194].
No study has examined in depth how the nature of the stress, CO concentration, duration of CO exposure, time between CO treatment and stress exposure, duration of stress exposure, time since cessation of stress, and the temporal relationship of CO and stress exposure (i.e. pre-incubation or co-incubation) affect the protection provided by CO. We hypothesize that CO protection of β-cells and islets from a variety of stresses will be dependent on these variables. However, before investigating various treatments, we needed to replicate the results obtained by Gunther et al. [194] that showed CO provided protection for βTC3 cells from TNF-α.
10.2.2 Methods
Cell culture
βTC3 cells (passages 40-50) were cultured in DMEM supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin (Mediatech Inc., Herdon, VA). Cells were cultured in a fully humidified environment at 37°C with 95% air/5% CO₂. For cytokine experiments TNF-α was added to a final concentration of either 50 or 500 U/ml and incubated with cells for 24 hr. Exposure to anoxia was conducted in a 5 L chamber at 37°C for 24 hr. The chambers were purged for 15 min with 95% N₂/5% CO₂ at 3 L/min before reducing the flow rate such that there was a very slight flow through the chamber keeping the gas composition constant throughout the experiment.
CO exposure
Gases were obtained in premixed cylinders. For cytokine experiments flasks were placed in 5 L chambers 0, 2, or 24 hr before TNF-α addition and purged for 15 min with 1% CO/20% O₂/5% CO₂/balance N₂ at 3 L/min before reducing the flow rate. The flasks were removed when adding TNF-α, but returned after addition and immediately purged again. Exposure to CO+anoxia was in a similar fashion but using 1% CO/5% CO₂/balance N₂.
For experiments investigating the effects of CO on tissue OCR 1% CO/20% O₂/5% CO₂/balance N₂ was bubble through serum free DMEM at 37°C for > 5 min. Cells were resuspended in the CO containing medium and immediately added to the OCR chambers and sealed. Control cells were handled similarly, but re-suspended in medium without CO.
Islet enumeration by nuclei counting
Nuclei were prepared by adding equal 100-μl volumes of sample containing 160 or more IE and of lysis solution (0.1 M citric acid (Sigma) and 1% (v/v) Triton X-100 (Sigma)) to a 1.5 ml microtube [55]. The mixture was incubated at room temperature for 5 min with vortex mixing every 1.5 min, and nuclei were liberated by shearing through a needle. Isolated nuclei were diluted with Dulbecco’s phosphate buffered saline (D-PBS, Invitrogen, Carlsbad, CA) to a concentration no higher than 5 x 10⁵ nuclei/ml, stained with 7-aminoactinomycin D (7-AAD, Molecular Probes, Eugene, OR), and analyzed using a flow cytometer (Guava Personal Cell Analysis (PCA) system, Guava Technologies, Hayward, CA) to determine the total number of cells with intact nuclei in the sample.
Oxygen consumption rate (OCR)
OCR was measured as previously described [39]. Briefly, suspensions containing about 2500 IE/ml in DMEM containing 4.5 g/l glucose and 0.6 g/l L-glutamine supplemented with 100 U/ml penicillin, 100 µg/ml streptomycin, 10 mM HEPES, and no added serum were sealed in a 200-µl stirred titanium chamber (Micro Oxygen Uptake System, FO/SYSZ- P250, Instech Laboratories, Plymouth Meeting, PA) maintained at 37°C. The time dependent pO$_2$ within the chamber was recorded with a fluorescence-based oxygen sensor (Ocean Optics, Dunedin, FL), and data at high pO$_2$ were fit to a straight line. The maximal OCR was evaluated from $OCR = V_{ch} \alpha (\Delta pO_2 / \Delta t)$, where $V_{ch}$ is the chamber volume and $\alpha = 1.27$ nmol/mm-Hg·ml is the Bunsen solubility coefficient for oxygen in medium [16]. OCR measurements were normalized by the measured number of cells (nuclei counting) or DNA content of the sample examined.
Membrane Integrity
7-Aminoactinomycin (7-AAD) sequential staining
Cell membrane integrity was assessed by differential staining with 7-AAD (Molecular Probes, Eugene, OR) [55]. An aliquot of about 300 islets was re-suspended in 100 µl of D-PBS, and 5 µl of 1 mg/ml 7-AAD and incubated for 20 min at 4°C protected from light. After two washes with 1 ml of D-PBS, cells were disrupted by adding equal volume of lysis solution in D-PBS to the islet suspension and sheared as described for islet enumeration. Labeled nuclei were counted immediately in the flow cytometer or stored on ice for less than 15 min before counting. A portion of the islet suspension was further stained with 7-AAD, thereby labeling all of the previously unlabelled nuclei, and the total number of nuclei was counted. The fraction of cells with compromised membranes was estimated as the ratio of the initially stained nuclei (first measurement) to the total number of nuclei (second measurement).
Sytox Orange/LDS 751
Cell suspensions were stained with a solution (in D-PBS) containing 0.8 µM LDS 751 (Molecular Probes), a membrane permeable dye and 0.2 µM Sytox Orange (Molecular Probes), a membrane impermeable dye. LDS 751 was dissolved in dimethyl sulfoxide (DMSO). Samples with a cell concentration no higher than 5×10$^5$ cells/ml were incubated for 5 min at room temperature and analyzed in a flow cytometer (Guava PCA) using the Guava Viacount software.
The fraction of cells with compromised membranes was the ratio of the number of cells stained with Sytox Orange over the number of cells stained with LDS 751.
**Statistics**
Measurements were made with three or more replicates and reported as mean ± SD. Statistical significance was determined using a Student t-test for unpaired data for comparing population means, and for paired data when appropriate.
### 10.2.3 Results
TNF-α caused a dose dependent reduction in the fraction of cells with intact cell membranes in βTC3 cells (Figure 10.4). 500 U/ml TNF-α caused the most reduction in viability as in previous studies with βTC3 cells [194] and therefore was chosen as the stress level when treating cells with CO. By co-incubating βTC3 cells in 500 U/ml TNF-α with 1% CO the fraction of cells with intact cell membranes increased 0.10-0.15 from that with TNF-α without CO (Figure 10.5). This protection decreased slightly when a pre-incubation period in CO (without stress) was performed.
Exposure to anoxia for 24 hr also caused a reduction in the fraction of cells with intact cell membranes in βTC3 cells, but is dependent on the cell density in culture (Figure 10.6). At higher densities there is a larger decrease in viability based on membrane integrity. Though the surface density of each culture vessel in each individual experiment was the same, they were not always the same between experiments. Though unknown initially, this difference caused a lot of variation between experiments. Data from four independent experiments exposing βTC3 cells to anoxia for 24 hr with and without CO are shown in Figure 10.7. Anoxia caused a large reduction in viability in two of the four experiments (Figure 10.7A and D), possibly due to density differences. However, even accounting for the differences in death, the relative effect of CO was different for each experiment. In some cases CO protected βTC3 cells from anoxia (Figure 10.7A) while others either made no difference (Figure 10.7B) or caused a further reduction in membrane integrity than anoxia alone (Figure 10.7C and 4). Even though differences are seen in the damage caused by anoxia, there are still inconsistencies in CO protection making it difficult to determine whether CO is beneficial.
Though the results presented here focus on membrane integrity measurements, some experiments attempted to use mitochondrial function assays such as ATP or MTT (data not
shown). The results from these assays made data even more variable. Eventually it was determined that part of this variability may be due to the impact CO has on cellular respiration. Adding medium with CO to βTC3 cells resulted in a 19% increase in OCR/cell as compared to cells in medium without CO (Table 10.2).
**Figure 10.4.** The fraction of cells with intact membranes versus TNF-α concentration. Membrane integrity determined by Sytox Orange/LDS 751 staining for 24 hr exposures to βTC3 cells. The error bars are ± the standard deviation for n = 3 measurements. Data are from one representative experiment of 3.
Figure 10.5. The fraction of cells with intact cell membranes with CO and TNF-α treatment.
Membrane integrity determined by Sytox Orange/LDS 751 staining for control cells (no stress), cells exposed to 500 U/ml TNF-α (no CO), cells exposed to CO without stress, and cells exposed to CO for 0, 2, or 24 hr (pretreatment) prior to exposure to CO + 500 U/ml TNF-α for 24 hrs. The error bars are ± the standard deviation for n = 3 measurements. Data are from one representative experiment of 5.
Figure 10.6. The fraction of βTC3 cells with intact cell membranes after exposure to anoxia.
Membrane integrity determined by Sytox Orange/LDS 751 staining for control cells (no stress) and cells exposed to anoxia for 24 hr for a high (> 60,000 cells/cm²) and low (< 15,000 cells/cm²) cell density. The error bars are ± the standard deviation for n = 3-5 measurements.
Table 10.2. OCR/cell of βTC3 cells with and without CO present.
| | OCR/cell (fmol/min cell) |
|----------------|--------------------------|
| Control | 3.24 ± 0.35 |
| CO treated | 3.86 ± 0.20 |
Data listed as mean ± the standard deviation for n = 3 experiments. Data are significantly different (p < 0.04) based on a Student t-test assuming paired data.
Figure 10.7. CO protection of βTC3 cells from anoxia.
The fraction of cells with intact cell membranes after as determined by Sytox Orange/LDS 751 staining for control cells (no stress), cells exposed to anoxia (no CO), cells exposed to CO without stress, and cells exposed to CO + anoxia for 24 hrs for four different experiments using βTC3 cells. The error bars are ± the standard deviation for n = 3 measurements. Each panel represents a single experiment.
10.2.4 Discussion
In this study we examined whether CO protects βTC3 cells from TNF-α or anoxia to determine whether CO treatment of islet tissue is likely to be beneficial. Unfortunately, several problems were encountered causing inconsistent results preventing the bulk of the initially proposed work from being completed.
Our results indicated that TNF-α caused a dose dependent reduction in βTC3 cell viability resulting in a fraction of cells with intact membranes around 0.45 (Figure 10.4), which is consistent with that observed in other studies [194], though ours varied from 0.35 to 0.50. Treating cells with CO before and during TNF-α exposure resulted in an increase of 0.10-0.15 (Figure 10.5). This increase is significantly less than that observed in previous studies which indicated an increase of roughly 0.70 [194]. The reason our study indicated less protection is unknown, but may be linked to differences in measuring cell viability/survival or slight differences in treatment methods.
A small decrease was observed in the protection CO provided from TNF-α exposure as the pre-incubation period in CO increased (Figure 10.5). A similar diminishing protective effect has been observed when HO-1 is over expressed [198]. HO-1 resulted in cytoprotection when overexpressed less than five fold, but were associated with significant toxicity at higher levels (15 fold expression) [198].
Experiments investigating whether CO provided βTC3 cells with protection from anoxia were far more variable. Cell density had a large impact on the fraction of cells with intact cell membranes (Figure 10.6). Higher density cells may have had lower viability because they will consume oxygen in the medium faster than cells at low density. Though oxygen in the gas phase is removed quickly, it takes a longer period to remove oxygen in the medium. As such, cells at high density are exposed to anoxia sooner than cells at low density. This difference in actual exposure time to anoxia is likely to explain the impact density has on membrane integrity.
If we consider the relative effect CO protection has when cells are exposed to anoxia, we still observe variable results. The experimental results indicate that CO could have a protective effect, no effect, or actually damaged tissue further in the presence of anoxia suggesting that the current system used for exposure to CO and/or anoxia is insufficient. A system that includes CO
and $O_2$ sensors in the culture medium would be beneficial to make sure the cells see the same concentrations for all experiments.
CO treatment increased tissue respiration rate significantly (Table 10.2), which is consistent with experiments that indicated the most damage in CO + anoxia rather than anoxia alone, but contrary to what has been reported previously [199]. D’Amico et al suggest that CO reduces cellular respiration immediately upon exposure [199], which is the exact opposite of what we found. Though it is unknown why the results differ, it could be linked to the concentration of CO to which cells are exposed, the cells used in each study, or the medium in which the respiration rate is measured. In particular, the medium used for OCR analysis in this study included the buffer hepes, which has been shown to decrease nitric oxide (NO) concentration in culture medium [200]. NO can also decrease cellular respiration and has been shown to interact with the HO-1/CO pathway. To confirm this hypothesis OCR measurements using hepes free medium would be needed.
The results presented here indicate that CO may be beneficial in protection of $\beta$-cells from cytokine stress, specifically TNF-$\alpha$. However, more experiments are needed to determine whether CO can protect cells from anoxia. Further, determination of whether CO increases of decreases cellular respiration would be beneficial since an increased respiration could have detrimental effects in hypoxic environments, such as that after transplantation and during conventional culture.
### 10.2.5 Nomenclature
| Abbreviation | Description |
|--------------|-------------|
| 7-AAD | 7-aminoactinomycin D |
| CO | Carbon monoxide |
| D-PBS | Dulbecco’s phosphate buffered saline |
| FBS | Fetal bovine serum |
| HO-1 | Heme oxygenase |
| NO | Nitric oxide |
| OCR | Oxygen consumption rate |
| t | Time |
| $V_{ch}$ | Volume of OCR chamber |
| $\alpha$ | Solubility of oxygen in medium |
10.3 Investigating oxygen consumption rate measurement in static cultures
10.3.1 Introduction
Measurement of cellular oxygen consumption rate (OCR) gives information on the respiratory activity of a tissue sample and is believed to indicate cell viability especially when comparing similar tissue samples. Measurements of islet tissue are of great importance since it has been shown that OCR of human islets can be used to predict transplantation outcome in mice [201]. In addition, OCR is an early indicator of cell death compared to other cell viability assays, such as membrane integrity measurements, allowing for accurate determination of the health of a sample even before the tissue has finished the cell death cycle [55].
The oxygen biosensor system (OBS) was developed for the measurement of OCR of tissue in static culture [146] and has been used for measurement of OCR for different cell lines [146; 202] and islet preparations [75]. The OBS utilizes an oxygen sensitive fluorophor embedded in silicone rubber upon which cells or tissue is placed. As the tissue consumes oxygen, the oxygen partial pressure (p) at the bottom of the plate decreases and is detected by the sensor.
The OBS is much easier to use than other methods that measure OCR allowing for multiple measurements at one time in a single plate. However, the current commercially available device gives inaccurate OCR estimates as compared to perfusion [74] and the stirred chamber [39] methods. In contrast to the perfusion and stirred chamber methods, which measure OCR based on material balances, the OBS relies on the application of the oxygen species conservation equation with the appropriate initial and boundary conditions for calculation of OCR from the measurement of p. In order to allow easy computation of OCR from the pO$_2$ measurements several simplifying assumptions are required [146] including (1) the system is at steady state, (2) the well geometry is perfectly cylindrical, (3) the wells are fabricated from oxygen impermeable material, (4) the medium is completely stagnant, (5) there is a uniform and infinitely thin cell layer, at the bottom, overlaying (6) a uniform infinitely thin oxygen sensor material. The validity of these assumptions determines the accuracy of the OCR estimation.
OCR values obtained using the OBS do not agree with measurements made with other methods. Measurements by the OBS for islet tissue resulted in an OCR of approximately 800 fmol/min IE [75] or about 0.5 fmol/min cell (1560 cells per IE [13]), which is much lower
than average values for islets measured in our laboratory using the stirred chamber system (Chapter 2).
In this section we investigate why OCR determined by the OBS is inaccurate. Through the use of a finite element mass transfer model the validity of the assumptions required to estimate OCR from the OBS device is tested by comparing the idealized model with the actual model. We then demonstrate experimentally the impact of two assumptions on the OCR measurement with the OBS including (1) that the wells are fabricated from oxygen impermeable material and (2) that the medium is completely stagnant.
### 10.3.2 Methods
#### Fabricating OBS plates
Oxygen sensor particles were made by adsorbing 75 mg 4,7-diphenyl-1,10-phenanthroline ruthenium (Alfa Aesar, CAS#: 36309-88-3) dissolved in 25 ml ethanol onto 11.5 g of silica gel powder (Aldrich, Catalog no: 23,675-6). The mixture was placed in a glass desiccant chamber connected to a vacuum pump and dried after which it was stored at -20°C.
For every 6.5 g of silicone elastomer base (Dow Corning, Sylgard, Catalog no: 3097366-0503) roughly 1.4 g of dried oxygen sensor adsorbed onto silica powder and 0.65 g of the curing agent (Dow Corning, Sylgard, Catalog no: 3097358-0503) were used to make the silicone rubber/sensor mixture. After mixing all three components such that no large clumps of undispersed powder were present, the mixture was added to plates using a 1 ml syringe such that the final sensor thickness was about 0.625 mm. This is more than that in the commercial OBS device plates but results from limitations in the accuracy of the syringe used to measure volume.
After adding the mixture to each well visible clumps of undissolved powder were removed using a 200 μl pipette. The mixture was spread to all edges of the dish and allowed to cure at room temperature for 2 days on a level surface before sterilization with ethanol and UV light.
#### Culture medium
INS-1 cells (passage 15-30) were cultured in supplemented RPMI (11.1 mM D-glucose, 100 U/ml penicillin, 100 ug/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate all obtained from Mediatech Inc., Herndon, VA, and 50 μM β-mercaptoethanol obtained from Sigma Aldrich, St. Louis, MO) with 10% FBS (Mediatech Inc.) and either 0 or 0.9% agarose.
One percent agarose medium was made by mixing 2/3 volume of a 1.5% agarose solution (1.5g ultra low melting point agarose powder Sigma, Catalog no: A-5030 in 98.5g DI water) with 1/3 volume of 3X RPMI solution (no FBS). Prior to dilution, the 1.5% agarose solution was heated until boiling to completely dissolve the powder and sterilize the solution. After mixing with the concentrated RPMI, 10% FBS was added to the 1% solution resulting in a 0.9% agarose solution. In order to avoid premature gelling, all components were stored at 37°C in a water bath before mixing.
**Experimental protocol**
A 2 µg/ml fibronectin solution (made by adding 100µl of 1 mg/ml fibronectin, Sigma, Catalog no: F1141-2MG, to 50ml PBS) was added to each well completely covering the bottom of the vessel. After incubating for four hours at 25°C the fluorescence was measured with 485 nm excitation and 630 nm emission wavelengths in a plate reader and the fibronectin solution was removed. INS-1 cells in liquid and agarose containing medium were added to wells for a targeted surface density of 100,000 cells/cm² and a medium depth of 3mm. Petri dishes, 6-well, 12-well, 24-well, and 96-well plates were used in the experiments. In addition one strip of eight glass wells with silica rubber/sensor added were also used in some experiments. Before placing the cells in the humidified 37°C incubator with 5% CO₂, the plates were shaken to make sure there was a uniform distribution of the cells across the bottom of the well.
Four fluorescence intensity readings were obtained for each plate during the course of an experiment all of which were read using the same gain. The first intensity reading was at ambient conditions prior to the removal of the fibronectin solution from the plates. The second reading was obtained after overnight incubation following cell seeding. After the second reading, the agarose plates were gelled for 20 minutes at 4°C before being returned to the 37°C incubator. After incubation for 20 hr in the gelled medium a third reading was taken. Following the third reading, tissue was removed from the plates and a 1M sodium sulfite solution was added to deplete the oxygen in the well. The final reading was taken 4-8 hours after adding sodium sulfite.
Nuclei samples were collected from the plates culturing cells in liquid RPMI after the third fluorescent reading. After removing the liquid medium, lysis solution was added to the wells liberating nuclei. Samples of the lysed cells were quantified using cell enumeration by nuclei counting. In some cases cells from the Petri dishes were trypsinized for 3 min, collected, re-suspended in fresh medium, and analyzed for OCR using the stirred chamber method.
Islet enumeration by nuclei counting
Nuclei were prepared by adding equal 100-μl volumes of sample containing 160 or more IE and of lysis solution (0.1 M citric acid (Sigma) and 1% (v/v) Triton X-100 (Sigma)) to a 1.5 ml microtube [55]. The mixture was incubated at room temperature for 5 min with vortex mixing every 1.5 min, and nuclei were liberated by shearing through a needle. Isolated nuclei were diluted with Dulbecco’s phosphate buffered saline (D-PBS, Invitrogen, Carlsbad, CA) to a concentration no higher than $5 \times 10^5$ nuclei/ml, stained with 7-aminoactinomycin D (7-AAD, Molecular Probes, Eugene, OR), and analyzed using a flow cytometer (Guava Personal Cell Analysis (PCA) system, Guava Technologies, Hayward, CA) to determine the total number of cells with intact nuclei in the sample.
Oxygen consumption rate (OCR)
OCR was measured as previously described [39]. Briefly, suspensions containing about 2500 IE/ml in DMEM containing 4.5 g/l glucose and 0.6 g/l L-glutamine supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 10 mM HEPES, and no added serum were sealed in a 200-μl stirred titanium chamber (Micro Oxygen Uptake System, FO/SYSZ- P250, Instech Laboratories, Plymouth Meeting, PA) maintained at 37°C. The time dependent pO$_2$ within the chamber was recorded with a fluorescence-based oxygen sensor (Ocean Optics, Dunedin, FL), and data at high pO$_2$ were fit to a straight line. The maximal OCR was evaluated from $OCR = V_{ch} \alpha (\Delta pO_2 / \Delta t)$, where $V_{ch}$ is the chamber volume and $\alpha = 1.27$ nmol/mm-Hg·ml is the Bunsen solubility coefficient for oxygen in medium [16]. OCR measurements were normalized by the measured number of cells (nuclei counting) or DNA content of the sample examined.
Statistics
Measurements were made with three or more replicates and reported as mean ± SD. Statistical significance was determined using a Student t-test for unpaired data for comparing population means, and for paired data when appropriate.
10.3.3 Theoretical model
Previously we utilized a theoretical oxygen transport model to simulate steady state islet culture (Chapter 3 and 4). Here we apply similar model equations to the OBS system. The species conservation equation for O$_2$ in the absence of convection is given by [16; 97]
where $C$ is the concentration of oxygen, $D$ is the diffusivity of oxygen, and $R$ is the production rate of oxygen. For convenience, the concentration terms in Equation 10.6 are replaced with oxygen partial pressure ($p$) using
$$C = \alpha p$$
where $\alpha$ is the solubility of oxygen. We assume that solubility is constant in each phase so that at steady state Equation 10.6 becomes
$$0 = D\alpha \nabla^2 p + R$$
Oxygen consumption within the tissue is assumed to follow Michaelis-Menten kinetics [16] for which the reaction rate is expressed as
$$R = -\frac{V_{\text{max}} p}{K_m + p}$$
where $V_{\text{max}}$ is the maximum reaction velocity, and the Michaelis constant $K_m$ is the $pO_2$ at which the reaction rate is $V_{\text{max}}/2$. $V_{\text{max}}$ was determined by multiplying the OCR/cell times the number of cells per well divided by the volume of the cell layer within the geometry. This oxygen consumption reaction only occurs above a critical $pO_2$ ($p_{\text{crit}}$). Below $p_{\text{crit}}$, the tissue dies and no longer consumes oxygen.
Three theoretical models were developed to explore oxygen transport within the OBS. The first assumed the ideal geometry with infinitely thin cell and sensor layers in an oxygen-impermeable dish (Figure 10.8A). The second geometry was a close representation of the actual well accounting for vessel curvature, oxygen permeability in polystyrene and finite thicknesses of the tissue and sensor (Figure 10.8B). The final geometry was similar to that in Figure 10.8B, but assumed a flat bottom vessel instead of a round bottom vessel (Figure 10.8C). This vessel represents the OBS plates of varying size that were fabricated in our lab. All these geometries assume that the system is axisymmetric around $r = 0$. The assumptions made in each geometry result in models governed by slight variations of basic mass transport equations.
For the idealized model (Figure 10.8A) the following equation is applied to transport in the medium (m).
\[ 0 = (D\alpha)_m \nabla^2 p \]
The boundary conditions for the idealized model include: constant external \( pO_2 \) across the gas-liquid interface,
at \( z = h \),
\[ p_{gas} = p_m \]
no oxygen flux through the oxygen-impermeable dish,
at \( r = R_{well} \),
\[ \frac{\partial p}{\partial r} = 0 \]
and the oxygen consumption in the infinitely thin cell layer at the bottom of the well is
at \( z = 0 \),
\[ N = -\frac{OCR/\text{cell} \cdot \chi \cdot p}{K_m + p} \]
where \( N \) is the oxygen flux, \( \chi \) is the cell surface density (cells/cm\(^2\)) and OCR/cell is the normalized OCR measured using the stirred tank reactor.
The non-idealized models (Figure 10.8B and C) require separate equations for each domain: one for the medium (m), one for the polystyrene (ps), one for the silicone rubber/sensor layer (s), and one for the tissue (t). The respective equations are
\[ 0 = (D\alpha)_m \nabla^2 p \]
\[ 0 = (D\alpha)_{ps} \nabla^2 p \]
\[ 0 = (D\alpha)_s \nabla^2 p \]
\[ 0 = (D\alpha)_t \nabla^2 p - \frac{V_{max}p}{K_m + p} \]
Figure 10.8. Geometry and the associated assumptions used for the three different theoretical models.
The boundary conditions for the non-idealized OBS include constant external $pO_2$ across the gas-liquid and gas-polystyrene interface,
$$p_{\text{gas}} = P_m \quad 10.18$$
$$p_{\text{gas}} = P_{ps} \quad 10.19$$
continuity of $pO_2$ and the oxygen flux across domain interfaces including polystyrene-medium, polystyrene-tissue, polystyrene-sensor, medium-tissue, and tissue-sensor,
$$n \cdot ((-\Delta \alpha \nabla p)_{\text{domain1}} - (-\Delta \alpha \nabla p)_{\text{domain2}}) = 0 \quad 10.20$$
$$P_{\text{domain1}} = P_{\text{domain2}} \quad 10.21$$
and symmetry boundary conditions for the axisymmetric problem, at $r=0$,
$$n \cdot (-\Delta \alpha \nabla p) = 0 \quad 10.22$$
Table 10.3 contains the values for the parameters used in the simulations. All parameter values have been discussed previously (Chapter 5) with the exception of oxygen diffusivity and solubility in polystyrene. The oxygen solubility in polystyrene at 25°C ranges from $0.53 - 1.1 \times 10^{-8}$ mol/cm$^3$ mmHg depending on how it was crystallized [203] while the diffusivity at 25°C ranges from $0.11 - 5.4 \times 10^{-7}$ cm$^2$/s [203-207]. The oxygen permeability in polystyrene at 25°C ranges from $0.54 - 11.9 \times 10^{-15}$ mol/cm s mmHg [144; 145; 203; 208]. Using the extremes of each parameter from literature we find a wide range of $0.0583 - 11.9 \times 10^{-15}$ mol/cm s mmHg for oxygen permeability in polystyrene at 25°C. Since temperature is at 37°C during culture the permeability should be adjusted to account for the temperature difference. To adjust for the temperature difference we assume that oxygen transport in polystyrene is similar to that in polycarbonate, which has resulted in a 25% increase in oxygen permeability when temperature was increased from 30 to 40°C or about a 2.5% increase every degree [209]. Thus a 30% increase in oxygen permeability in polystyrene is assumed for the 12°C difference between the desired permeability temperature at 37°C and 25°C. The resulting permeability at 37°C is estimated to be between $0.076 - 15.5 \times 10^{-15}$ mol/cm s mmHg with a base case value of $3.5 \times 10^{-15}$ mol/cm s mmHg.
The model was solved with the finite element method using the commercially available software COMSOL Multiphysics (COMCOL Inc., Burlington, MA) in conjunction with Matlab (Mathworks, Natick, MA). We required the mesh to contain more than 2500 nodes and set a convergence criterion such that the relative error estimated by the finite element software in the solution vector for concentration, which takes into account the current and previous solution in the iterative solver, was less than $10^{-6}$ [99].
After the theoretical model was solved the results were analyzed to determine the pO$_2$ measured by the OBS. Using COMSOL Multiphysics the pO$_2$ in the sensor layer was converted to an intensity ratio ($I/I_0$) using a version of the Stern-Volmer relationship
$$\frac{I}{I_0} = \frac{1}{1 + K_{SV} \cdot p} \quad 10.23$$
where $I$ and $I_0$ are the emission intensities in the presence and absence of oxygen and $K_{SV}$ is a constant characteristic related to the sensor quenching rate and emission lifetime [210]. For the idealized model (Figure 10.8A) the average intensity ratio was determined by integrating the intensity over the area of the plate read by the sensor (from $r = 0$ cm to $r = 0.256$ cm) and then dividing by the integrated surface area. For the models that accounted for the sensor thickness the average intensity ratio was determined by integrating over the volume of the sensor read by the plate reader (from $r = 0$ cm to $r = 0.256$ cm) and then dividing by the integrated volume. The average intensity was converted to a pO$_2$ and an OCR/cell using
$$p = \left( \frac{\left( \frac{I_0}{I} \right)_{ave} - 1}{K_{SV}} \right) \quad 10.24$$
$$\frac{OCR}{cell} = \frac{(Da)_m (p_{gas} - p)}{h \cdot \chi} \quad 10.25$$
Table 10.3. Parameters used in theoretical simulations for oxygen transport in the OBS device.
| Parameter | Symbol | Value | Units | Reference |
|------------------------------------------------|--------|----------------|-------------|-----------|
| Michaelis constant | $K_m$ | 0.44 | mmHg | [98] |
| O$_2$ diffusivity - medium | $D_{m}$| $2.78 \times 10^{-5}$ | cm$^2$/s | [16] |
| O$_2$ diffusivity - tissue | $D_t$ | $1.22 \times 10^{-6}$ | cm$^2$/s | [16] |
| O$_2$ diffusivity - silicone rubber | $D_s$ | $2.1/\times 10^{-5}$ | cm$^2$/s | [116] |
| O$_2$ solubility - medium | $\alpha_m$ | $1.27 \times 10^{-9}$ | mol/cm$^3$/mmHg | [16] |
| O$_2$ solubility - tissue | $\alpha_t$ | $1.02 \times 10^{-9}$ | mol/cm$^3$/mmHg | [16] |
| O$_2$ solubility - silicone rubber | $\alpha_s$ | $1.21 \times 10^{-6}$ | mol/cm$^3$/mmHg | [116] |
| O$_2$ permeability - polystyrene | $(D\mu)_{ps}$ | $3.5 \times 10^{-15}$ | mol/cm's mmHg | Varied - see text |
| Stern-Volmer constant | $K_{sv}$ | 0.03125 | | Based on experiment |
| Medium depth | 3 | mm | | Based on medium volume and well area |
| Tissue thickness | 12 | $\mu$m | | Based on average cell diameter |
| OCR/cell | 2.8 | fmol/min cell | | Experimentally measured |
| Area analyzed by plate reader | 0.256 | cm$^2$ | | Based on equipment capabilities |
| Sensor volume - actual OBS | 7 | $\mu$l | | Based on commercial device |
| Sensor thickness - fabricated wells | 0.0625 | cm | | Based on sensor volume and well area |
| Polystyrene thickness | 1.2 | mm | | Experimentally measured |
| Minimum pO$_2$ for viability | $p_{crit}$ | 0.1 | mmHg | [16] |
| External p | $p_{ext}$ | 142 | mmHg | Partial pressure of O$_2$ in humidified incubator at 37°C with 95% air/5% CO$_2$. |
OCR/cell of INS-1 cells was determined using the stirred chamber method. The reading position assumed for the plate reader was equal to 65% of the bottom surface area for a 96 well plate (0.209 cm$^2$).
10.3.4 Calculating experimental OCR from OBS
This section describes the step-by-step procedure of converting the fluorescent intensity readings of the OBS plate to pO$_2$ and OCR/cell values. The initial reading of a plate containing 2 µg/ml fibronectin solution equilibrated with ambient air at room temperature is set equivalent to the initial ambient intensity, $I_A$. Intensity readings at the end of the experiment (3rd reading, after cell addition and gelling) for sample (wells containing cells) and control (wells with only medium) wells corresponded to $I_S$ and $I_C$, respectively. The final reading 4 to 8 hours after adding 1 M sodium sulfite solution corresponded to the zero pO$_2$ and is the maximum intensity reading $I_0$.
To calculate pO$_2$ multiple manipulations of the intensity readings need to be made. First, $I_S$ for each particular well is adjusted for variations in sensor concentration by dividing by the corresponding $I_A$ for that well resulting in $I_S'$, the adjusted sample intensity.
$$I_S' = \frac{I_S}{I_A} \quad 10.26$$
A similar adjustment is made for the control wells.
$$I_C' = \frac{I_C}{I_A} \quad 10.27$$
Signal drift correction associated with changes in the ambient is performed by dividing $I_S'$ by $\overline{I_C'}$, the averaged adjusted control intensities, to obtain $I_S''$, the adjusted cell sample intensity normalized for drift.
$$\overline{I_C'} = \frac{1}{n} \sum_{n} I_{C,n}' \quad 10.28$$
$$I_S'' = \frac{I_S'}{\overline{I_C'}} \quad 10.29$$
The maximum increase in fluorescent signal, R, is found for each individual well.
Then, the Stern Volmer equation is applied to find corresponding cell surface $pO_2$ values.
$$R = \frac{I_0}{I_A} \quad 10.30$$
$$p = \frac{\frac{R}{I_S''} - 1}{\frac{R - 1}{p_{gas}}} \quad 10.31$$
where $p_{gas}$ is the $O_2$ partial pressure at ambient conditions at the end of the experiment when the final measurement is taken. Once the $pO_2$ for each well is determined OCR/cell can be calculated using 10.25.
### 10.3.5 Theoretical results
Oxygen profiles for the idealized model, the commercially available device, and the fabricated flat bottom wells for the specific case of 250,000 cells/cm$^2$ in 3 mm medium are shown in Figure 10.9 with the minimum $pO_2$ in the tissue and the OCR/cell that would be determined by the OBS listed above each plot. The idealized model indicates uniform diffusion through the medium resulting in an OCR/cell of 2.8 nmol/min cm$^2$, which is the same as the inputted value based on the stirred chamber measurements (Table 10.3). The model based on the commercially available device and the fabricated devices indicate significant transport through the polystyrene resulting in lower OCR/cell values. In both the commercial and fabricated devices the oxygen flux through the polystyrene accounts for 43-46% of the total oxygen flux into the system (data not shown). The differences in minimum $pO_2$ and OCR/cell between the fabricated well and the commercially available device result because one is a round-bottom well and one is a flat-bottom well. Though both have transport through the polystyrene, the commercially available device has additional limitations including tissue pooling and a larger medium depth in the center of the well and thus the $pO_2$ in the sensor layer is lower.
The sensitivity of the fabricated device to changes in the parameter values was explored to provide insight into which parameters have the largest impact on the OBS system. Variations in tissue thickness, sensor layer thickness, polystyrene thickness (side and bottom), reader position, and oxygen permeability in medium, the sensor, and the tissue did not have a large
effect on the steady state OCR determined by the OBS for a realistic range of values (data not shown).
OCR/cell determined by the OBS does not vary with surface density assuming a uniform distribution until the partial pressure in the tissue decreases significantly. The low pO$_2$ in the tissue at high surface densities causes OCR to decrease because of Michaelis-Menten kinetics (10.9) and in extreme cases, at pO$_2 < 0.1$ mmHg, could result in tissue death from hypoxia (Figure 10.10). This suggests that the cell density used in experiments should $< 400,000$ cells/cm$^2$ for a 96-well plate or $< 300,000$ cells/cm$^2$ for a Petri dish. Note that if the tissue distribution is not uniform predictions could change significantly.
Oxygen permeability in polystyrene varied significantly in literature and has a large effect on the OCR determined by the OBS (Figure 10.11). The more oxygen coming through the polystyrene (i.e. the higher the oxygen permeability in polystyrene), the lower the OCR/cell measured by the OBS. The base case value of $3.5 \times 10^{-15}$ mol/cm s mmHg, which is in the mid to upper end of values obtained from literature, results in an OCR/cell that is only 60% of that obtained with the stirred chamber method.
Significant amounts of oxygen enter the system through the polystyrene at both the bottom of the well and the sides, which suggests that well size should have an impact on the OCR determined by the OBS. As well radius increases, the OCR/cell obtained from the device increases until it reaches a plateau (Figure 10.12), which occurs when the polystyrene walls no longer affect the central portion of the well. Though this suggests that increasing well size will bring the measurement closer to the benchmark, the plateau can still be well below the stirred chamber value if a significant amount of oxygen is transported through the polystyrene directly below the sensor.
Figure 10.9: Theoretical oxygen profiles in the (A) idealized, (B) actual, and (C) fabricated OBS plates. Predictions assume 250,000 cells/cm$^2$. The resulting minimum pO$_2$ in the tissue and the OCR/cell that would be calculated by the OBS for is given above each model.
Figure 10.10. Predicted OCR/cell determined by the OBS (solid lines) and the average pO₂ in the tissue (dashed lines) as a function of cell density for a 96-well plate and Petri dish.
Figure 10.11. Predicted OCR/cell determined by the OBS as a function of oxygen permeability in polystyrene for a Petri dish, 24-well plate, and 96-well plate.
Figure 10.12. Predicted OCR/cell determined by the OBS as a function of well radius and oxygen permeability in polystyrene.
10.3.6 Experimental results
The OCR/cell obtained for INS-1 cells in both the liquid (Figure 10.13A) and gelled medium (Figure 10.13B) from the 96 well plates was below the value of the stirred chamber. As plate size increased the OCR/cell determined by the fabricated OBS dishes increased eventually resulting in OCR/cell similar to the benchmark. The OCR/cell obtained from the 96-well plate is significantly lower than the values obtained from larger wells for cells in gelled medium ($p < 0.05$). On the other hand OCR/cell values of cells cultured in the larger well sizes did not differ significantly. Results were similar for liquid medium.
The experimental data in Figure 10.13 and theoretical results in Figure 10.11 and Figure 10.12 suggest that polystyrene could cause lower OCR values than that obtained by the stirred chamber. However, the OCR/cell obtained from INS-1 cells in gelled medium in polystyrene and glass wells (size of wells in a 96-well plate) were $1.9 \pm 0.3$ and $1.9 \pm 0.2$ ($p > 0.6$, $n=3$), respectively, suggesting that the assumption that polystyrene is oxygen-impermeable is valid.
Difference between OCR obtained from gelled and liquid medium revealed that convective transport was present during normal OBS operation. OCR/cell measured by the OBS for INS-1 cells in gelled medium was statistically higher than that in liquid medium when all the data except for the 24-well plate, which had a high standard deviation, is considered ($p < 0.10$).
Comparisons between individual well sizes varied in differences with $p < 0.10$ for only the glass wells and the 12-well plates (Table 10.4). The lack of significance in other well size comparisons is likely due to the limited number of experiments and variation between them. In addition, the averages of two plates that suggest convection is not important (24-well plate and the Petri dish) each contain a single point that was somewhat different than the other two, but that could not be thrown out because of limited data. Though the effect of convection varies between experiments the results suggest that convection decreased OCR measurements by roughly 15% on average (ignoring data from 24-well plate), but as much as 35%.
10.3.7 Discussion
In this study we examined the use of oxygen sensors embedded in silicone rubber at the bottom static culture dishes for measurement of pO$_2$ and calculation of OCR. Through theoretical modeling and experimental measurements we investigated two assumptions required for easy calculation of OCR from the static culture devices including (1) oxygen-impermeable dishes and (2) stagnant medium. The results here provide insight into why OCR measurements with the commercially available device are inaccurate and suggest modifications that can be made to get accurate measurements with a static culture device.
Several model parameters were examined, but only a few had a large impact on the model predictions. Cell surface density had a significant impact on the OCR determined by the OBS (Figure 10.10). Very high densities caused a reduction in the apparent OCR/cell because of Michaelis-Menten kinetics indicating the theoretical limitations of the system.
Figure 10.13. OCR/cell determined by fabricated OBS plates as a function of well radius. Predictions for INS-1 cells cultured in (A) liquid and (B) gelled medium. The horizontal line represents the benchmark value obtained from the stirred chamber. The error bars are ± standard deviation for n = 3 experiments.
Table 10.4. Comparison of OCR/cell between cells in liquid and gelled agarose medium.
| Well size | Liquid OCR/cell (fmol/(min cell)) | Agarose OCR/cell (fmol/(min cell)) | p-value |
|---------------|-----------------------------------|------------------------------------|---------|
| 96 well plate | 1.6 ± 0.4 | 1.9 ± 0.3 | 0.12 |
| 24 well plate | 3.5 ± 0.7 | 2.6 ± 0.9 | 0.26 |
| 12 well plate | 2.2 ± 0.2 | 2.9 ± 0.5 | 0.07 |
| 6 well plate | 2.6 ± 0.5 | 2.7 ± 0.5 | 0.14 |
| Petri dish | 3.0 ± 0.6 | 2.7 ± 0.2 | 0.62 |
| Glass plate | 1.4 ± 0.1 | 1.9 ± 0.2 | 0.03 |
Data listed as mean ± standard deviation for 3 independent experiments.
The theoretical model suggested that oxygen flux through the polystyrene could be a significant source of error if the oxygen permeability was too high (Figure 10.11). However, experimental comparisons between glass and polystyrene wells demonstrated that oxygen transport through the polystyrene was negligible and thus $D_{\text{O}_2}^{\text{PS}}$ is lower than that assumed for most of the simulations. This is consistent with our experimental measurements in large wells that indicate no significant flux through the polystyrene on the bottom of the dish (Figure 10.13B). However, at low rates of oxygen consumption or measurement in low $pO_2$ environments, the oxygen permeability in polystyrene could become more problematic as suggested previously [210-212].
The assumption of stagnant medium is more problematic. Convective transport decreased the apparent OCR/cell 25% from that of cells cultured in gelled medium (Table 10.4). This convective transport may be due to vibrations and thermal variations within the incubator, but is most likely due to movement in the plate reader and when transferring the plate from the incubator to the plate reader. The substantial motion in the plate reader has been demonstrated by comparing the transient reequilibration with oxygen at the bottom of wells initially purged with nitrogen and filled with either water or an agarose gel that prevents fluid motion (unpublished data). For accurate measurement of OCR using the OBS the use of gelled medium is critical since the convective transport could vary each time the plate is removed from the incubator.
Despite the removal of convection and the confirmation that oxygen flux through polystyrene is negligible, variations still exist between wells of different size which may be related to curvature of the sensor in the plate. Though sensor is added to the plate for a uniform height, capillary forces the edge of the wells draw up the silicone rubber/sensor mixture resulting in a slightly curved bottom. This curvature tends to be larger for smaller wells, especially for the 96 well plate, which gives the lowest OCR/cell. The use of larger wells results in accurate OCR/cell measurements by the fabricated OBS device though additional experiments are needed to confirm.
10.3.8 Nomenclature
| Symbol | Description |
|--------|-------------|
| C | Concentration of oxygen |
| D | Diffusivity of oxygen |
| D-PBS | Dulbecco’s phosphate buffered saline |
| FBS | Fetal bovine serum |
| h | Medium depth |
| I | Emission intensity in the presence of oxygen |
| $I_A$ | Emission intensity for ambient reading |
| $I_C$ | Emission intensity for control reading |
| $I_C'$ | Adjusted control intensity |
| $\overline{I_C'}$ | Averaged adjusted control intensities |
| $I_o$ | Emission intensity in the absence of oxygen, maximum intensity reading |
| $I_S$ | Emission intensity for sample reading |
| $I_S'$ | Adjusted sample intensity |
| $I_S''$ | Adjusted cell sample intensity normalized for drift |
| $K_m$ | Michaelis constant, the pO$_2$ at which the reaction rate is $V_{\text{max}}/2$ |
| $K_{sv}$ | Characteristic related to the sensor quenching rate and emission lifetime |
| OBS | Oxygen biosensor system |
| OCR | Oxygen consumption rate |
| p | Oxygen partial pressure |
| Symbol | Description |
|--------|-------------|
| $P_{\text{crit}}$ | The $pO_2$ below which a cell dies and no longer consumes oxygen |
| $P_{\text{gas}}$ | Ambient partial pressure of oxygen |
| $R$ | Production rate of oxygen |
| $R$ | Maximum increase in fluorescent signal |
| $r$ | Radial direction in cylindrical coordinates |
| $R_{\text{well}}$ | Radius of well |
| $t$ | Time |
| $V_{\text{ch}}$ | Volume of OCR chamber |
| $V_{\text{max}}$ | Maximum reaction velocity |
| $\alpha$ | Solubility of oxygen in medium |
| $\chi$ | Cell surface density (cells/cm$^2$) |
**Subscripts for D, $\alpha$, p, C**
- **m**: Medium
- **s**: Sensor/silicone rubber mixture
- **ps**: Polystyrene
- **t**: Tissue
% Matlab code for oxygen transport during static culture on polystyrene
% FEMLAB Model M-file
fclear fem
% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.2';
vrsn.ext = 'a';
vrsn.major = 0;
vrsn.build = 300;
vrsn.res = '$Name: $';
vrsn.date = '$Date: 2005/12/20 19:02:30 $';
fem.version = vrsn;
% Get input for parameters
po2ext = 142; % mmHg
V_max = 3.68e-8; % mol/cm^3 s
K_m = 0.44; % mmHg
height = 0.3; % cm
radius_islet = 0.0075; % cm
diff_alpha_islet = 1.24e-14; % mol/cm mmHg s
diff_alpha_medium = 3.53e-14; % mol/cm mmHg s
P_crit = 0.1; % mmHg
islet_density = 100; % islets/cm^2
alpha_medium = 1.27e-9; % mol/cm^3 mmHg
alpha_islet = 1.02e-9; % mol/cm^3 mmHg
% Place constant into fem
fem.const={'alpham','alpha_medium','alpha','alpha_islets','Pext','po2ext','Dai','diff_alpha_islet','Vmax','V_max','Dam','diff_alpha_medium','Km','K_m','Rislet','radius_islet','Pcrit','P_crit'};
% Specifies dimensions
fem.sdim=3;
% GEOMETRY
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Function to create geometry for islet array (equally spaced islets in a
% square array. Only an eighth of an islet will be modeled because of
% symmetry.
% Set length of sides of block
% scaling parameter for height of medium, radius of islet=1
scale_height = height/radius_islet; % scaling parameter for height of box
scale_space = ((1/islet_density)^(1/2))/radius_islet; % calculating length of sides for given density
Lx = scale_space; % center to center distance between islets
Ly = scale_space;
Lz = (scale_height);
% Create block for medium
B3_1='block3(Lx,Ly,Lz,'base','corner','pos',{'0','0','0'},'axis',{'0','0','1'},'rot','0');
B3_2='block3(Lx,Ly,Lz,'base','corner','pos',{'0','0','0'},'axis',{'0','0','1'},'rot','45');
B3_3='block3(Lx/2,Ly/2,Lz,'base','corner','pos',{'0','0','0'},'axis',{'0','0','1'},'rot','0');
B3_4=B3_1*B3_2;
B3_triangle1=B3_3-B3_4;
% Now we will add in the sphere for islet of a given radius set to 1
S3a = sphere3(1,[0 0 1]);
% make sphere into one eighth of islet
S3_eighth=S3a*B3_triangle1;
% remove corner from lower block
B3_triangle2=B3_triangle1-S3a;
% Combine into a single geometry
clear s
s.objs={S3_eighth,B3_triangle2};
s.name={'CO1','Blk1'};
s.tags={'S3_eighth','B3_triangle2'};
fem.draw = struct('s',s);
fem.geom = geomncg(fem);
% Initialize mesh
fem.mesh=meshinit(fem);
mesh_test = size(fem.mesh.t,2);
while mesh_test<2500
fem.mesh=meshrefine(fem);
mesh_test = size(fem.mesh.t,2);
end
% (Default values are not included)
% Application mode 1
clear appl
appl.mode.class = 'FIDiffusion';
appl.shape = {'shlag(2,"lm1")','shlag(2,"c")'};
appl.assignsuffix = '_di';
clear prop
clear weakconstr
weakconstr.value = 'non-ideal';
weakconstr.dim = {'lm1'};
prop.weakconstr = weakconstr;
appl.prop = prop;
clear bnd
bnd.type = {'N0','cont','C'};
bnd.weakconstr = {0,0,1};
bnd.c0 = {0,0,1};
bnd.wshape = 1;
bnd.ind = [1,1,2,1,1,1,2,3,1];
appl.bnd = bnd;
clear equ
equ.D = {1,'Dam/Dai'};
equ.int = 1;
equ.R = {'-Vmax*Rislet'^2/(Dai*Pext)*c/(c+Km/Pext)*(c>Pcrit/Pext)',0};
equ.shape = 2;
equ.ind = [1,2];
appl.equ = equ;
fem.appl{1} = appl;
fem.border = 1;
fem.outform = 'general';
% Multiphysics
fem=multiphysics(fem);
% Extend mesh
fem.xmesh=meshextend(fem);
% Solve problem
[fem.sol, fem.stop]=femmlin(fem, ...
'solcomp',{'c','lm1'}, ...
'outcomp',{'c','lm1'}, ...
'out',{'sol','stop'},...
'hnlin','on','maxiter',100);
% Save current fem structure for restart purposes
fem0=fem;
% Post processing
max_PO2 = postmax(fem,'c','solnum',1,'Dl',[1]); % Maximum node in islet
min_PO2 = postmin(fem,'c','solnum',1,'Dl',[1]); % Minimum node in islet
convergence1=fem.stop;
% Integrate islet without O2
frac_no_O2 = postint(fem,1*'c<=Pcrit/Pext','solnum',1,'phase',0*pi/180,'edim',3,'intorder',4,'geomnum',1,'dl',[1]);
% Integrate whole islet
islet_int = postint(fem,1,'solnum',1,'phase',0*pi/180,'edim',3,'intorder',4,'geomnum',1,'dl',[1]);
frac_dead=frac_no_O2/islet_int;
OCR = postint(fem,'lm1', ...
'solnum',1, ...
'phase',0*pi/180, ...
'edim',2, ...
'intorder',4, ...
'geomnum',1, ...
'dl',[8]); % Total oxygen flux into system
min_quality_mesh = min(meshqual(fem.mesh));
element_num = size(fem.mesh.t,2);
10.5 Matlab code for oxygen transport during static culture on silicone rubber
% FEMLAB Model M-file
fclear fem
% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.2';
vrsn.ext = 'a';
vrsn.major = 0;
vrsn.build = 300;
vrsn.rcs = '$Name: $';
vrsn.date = '$Date: 2005/12/20 19:02:30 $';
fem.version = vrsn;
% Get input for parameters
po2ext = 142; % mmHg
V_max = 5.33e-8; % mol/cm3 s
K_m = 0.44; % mmHg
height = 2.2; % cm
radius_islet= 0.0075; % cm
diff_alpha_islet = 1.24e-14; % mol/cm mmHg s
diff_alpha_medium = 3.53e-14; % mol/cm mmHg s
P_crit = 0.1; % mmHg
islet_density = 100; % IEQ/cm2
diff_alpha_simembrane = 2.17e-5*1.21e-8; % mol/cm mmHg s
thickness = 0.05; % cm
% Place constant into fem
fem.const={'Po2',po2ext,'Dai',diff_alpha_islet,'Vmax',V_max,'Dam',diff_alpha_medium,'Km',K_m,'Rislet',radius_islet,'Dasi',diff_alpha_simembrane,'Pcrit',P_crit};
% Specifies dimensions
fem.sdim=3;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% GEOMETRY%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Function to create geometry for islet array (equally spaced spheres in a
% square array. Only one quarter of an islet will be modeled because of
% symmetry.
pt = [0 0 0]; % base point of 3D block
vec = [0 0 1]; % specifies direction of block
% Set length of sides of block
% scaling parameter for height of medium, radius of islet=1
scale_height = height/radius_islet; % scaling parameter for height of box
scale_space = ((1/islet_density)^(1/2))/radius_islet; % calculating length of sides for given density
Lx = scale_space; % center to center distance between islets
Ly = scale_space;
Lz = (scale_height-2);
% Create block for medium
B3top=block3([Lx/2,Ly/2,Lz],'corner',[0 0 2],vec);
% Now we will add in the spheres for islets of a given radius
% Spheres at vertices of block B3
S3a = sphere3(1,[0 0 1]);
% Create composite objects.
B3bottom=block3([Lx/2,Ly/2,Lz],'corner',pt,vec);
% First make sphere into quarters.
COa = S3a*B3bottom;
B3med=B3bottom+B3top;
B3meda=B3med-S3a; % remove corner from block
% Create block for membrane
Lzmem=thickness/radius_islet;
B3mem=block3([Lx/2,Ly/2,Lzmem,'corner',[0,0,-Lzmem],vec);
B3medb=B3meda+COa;
B3medc=B3medb+B3mem;
clear s
s.objs={B3medc};
s.name={'CO1'};
s.tags={'B3medc'};
fem.draw = struct('s',s);
fem.geom = geomcsg(fem);
% Initialize mesh
fem.mesh=meshinit(fem, ...
'hmaxfact',5, ...
'hcutoff',0.07, ...
'hgrad',2, ...
'hcurve',1, ...
'hnarrow',0.1);
mesh_test = size(fem.mesh.t,2);
while mesh_test<2500
fem.mesh=meshrefine(fem);
mesh_test = size(fem.mesh.t,2);
end
% Application mode 1
clear appl
appl.mode.class = 'FIDiffusion';
appl.shape = {'shlag(2,"lm1","'),shlag(2,"c")'};
appl.assignsuffix = '_di';
clear prop
clear weakconstr
weakconstr.value = 'non-ideal';
weakconstr.dim = {'lm1'};
prop.weakconstr = weakconstr;
appl.prop = prop;
clear bnd
bnd.c0 = {0,1,0};
bnd.type = {'N0','C','cont'};
bnd.weakconstr = {0,1,1};
bnd.wshape = 1;
bnd.ind=[1,1,2,1,1,1,3,3,1,1,1,3,3,2,3,1,1,1,1,1,1];
appl.bnd = bnd;
clear equ
equ.shape = 2;
equ.init = 1;
equ.D = {'Dasi/Dai','1','Dam/Dai'};
equ.R = {0,'Vmax*Rislet^2/(Dai*Pext)*c/(c+Km/Pext)*(c>Pcrit/Pext)',0};
equ.ind=[1,2,3,3];
appl.equ = equ;
fem.appl{1} = appl;
fem.border = 1;
fem.outform = 'general';
% Multiphysics
fem=multiphysics(fem);
% Extend mesh
fem.xmesh=meshextend(fem);
% Solve problem
[fem.sol, fem.stop]=femmlin(fem, ...
'solcomp',{'c','lm1'},...
'outcomp',{'c','lm1'},...
'out',{'sol','stop'},...
'halin','on','maxiter',100,...
'ntol', 1e-006);
% Save current fem structure for restart purposes
fem0=fem;
convergence=fem.stop;
max_PO2 = postmax(fem,'c','solnum',1,'dl',[2]); % Maximum node in islet
min_PO2 = postmin(fem,'c','solnum',1,'dl',[2]); % Minimum node in islet
% Integrate islet without O2
frac_no_O2 = postin(fem,'1*c<=Pcrit/Pext','solnum',1,'phase',0*pi/180,'edim',3,'intorder',4,'geomnum',1,'dl',[2]);
% Integrate whole islet
islet_int = postin(fem,'1','solnum',1,'phase',0*pi/180,'edim',3,'intorder',4,'geomnum',1,'dl',[2]);
frac_dead=(frac_no_O2)/(islet_int);
OCRtop = postin(fem,'lm1', ...
'solnum',1, ...
'phase',0*pi/180, ...
'edim',2, ...
'intorder',4, ...
'geomnum',1, ...
'dl',[14]); % Total oxygen flux into system from top
OCRbottom = postin(fem,'lm1', ...
'solnum',1, ...
'phase',0*pi/180, ...
'edim',2, ...
'intorder',4, ...
'geomnum',1, ...
'dl',[3]); % Total oxygen flux into system from bottom
OCR = OCRtop+OCRbottom;
min_quality_mesh = min(meshqual(fem.mesh));
element_num = size(fem.mesh.t,2);
10.6 Matlab code for oxygen transport model with 2 sphere sizes
% FEMLAB Model M-file
% Set up vectors to store specific numbers of solutions
n=6;
max_PO2_1=linspace(1,n,n);
min_PO2_1=linspace(1,n,n);
frac_dead_1=linspace(1,n,n);
max_PO2_2=linspace(1,n,n);
min_PO2_2=linspace(1,n,n);
frac_dead_2=linspace(1,n,n);
OCR=linspace(1,n,n);
min_quality_mesh=linspace(1,n,n);
element_num=linspace(1,n,n);
convergence1=linspace(1,n,n);
i_density = [100 500 750 1000 1100 1110];
for i=1:n
fclear fem
% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.2';
vrsn.ext = 'a';
vrsn.major = 0;
vrsn.build = 300;
vrsn.rcs = '$Name: $';
vrsn.date = '$Date: 2005/12/20 19:02:30 $';
fem.version = vrsn;
% Get input for parameters
po2ext = 142; % mmHg
V_max = 5.33e-8; % mol/cm3 s
K_m = 0.44; % mmHg
height = 0.3; % cm
radius_islet1= 0.0075; % cm
radius_islet2= 0.0125; % cm
diff_alpha_islet = 1.24e-14; % mol/cm mmHg s
diff_alpha_medium = 3.53e-14; % mol/cm mmHg s
P_crit = 0.1; % mmHg
islet_density = i_density(i); % islets/cm2
alpha_medium = 1.27e-9; % mol/cm3 mmHg
alpha_islets = 1.02e-9; % mol/cm3 mmHg
V1 = 4/3*3.14*(radius_islet1)^3; %cm3
V2 = 4/3*3.14*(radius_islet2)^3; %cm3
V_IEO = 4/3*3.14*0.0075^3; %cm3
R2R1 = radius_islet2/radius_islet1;
% Place constant into fem
fem.const={'alpha_medium','alpha_islets','pext','po2ext','Dai','diff_alpha_islet','Vmax','V_max','Dam','diff_alpha_medium','Km','K_m','Rislet','radius_islet1','Pcrit','P_crit'};
% Specifies dimensions
fem.sdim=3;
% GEOMETRY
% Function to create geometry for islet array (equally spaced islets in a
% square array. Only an eighth of an islet will be modeled because of
% symmetry.
% Set length of sides of block
% scaling parameter for height of medium, radius of islet=1
scale_height = height/radius_islet1; % scaling parameter for height of box
scale_space = (1/islet_density)*(1/2)/radius_islet1; % calculating length of sides for given density
Lx = scale_space; % center to center distance between islets
Ly = scale_space;
Lz = (scale_height);
% Create block for medium
B3_1 = block3(Lx,Ly,Lz,'base','corner','pos',{'0','0','0'},'axis',{'0','0','1'},'rot','0');
B3_2 = block3(2*Lx,2*Ly,Lz,'base','corner','pos',{'0','0','0'},'axis',{'0','0','1'},'rot','45');
B3_3 = block3(2*Lx,2*Ly,Lz,'base','corner','pos',{Lx,'0','0'},'axis',{'0','0','1'},'rot','45');
B3_4 = B3_1 - B3_2;
B3_triangle1 = B3_4 - B3_3;
% Now we will add in the sphere for islet of a given radius set to 1
S3a = sphere3(1,[0 0 1]);
S3b = sphere3(R2R1,[Lx 0 R2R1]);
% make sphere into one eighth of islet
S3_eighth1 = S3a*B3_triangle1;
S3_eighth2 = S3b*B3_triangle1;
% remove corner from lower block
B3_triangle2 = B3_triangle1-S3a;
B3_triangle3 = B3_triangle2-S3b;
% Combine into a single geometry
B3_triangle4 = B3_triangle3+S3_eighth1;
B3_triangles5 = B3_triangle4+S3_eighth2;
clear s
s.objs = {B3_triangles5};
s.name = {'CO'};
s.tags = {'B3_triangles5'};
fem.draw = struct('s',s);
fem.geom = geomcsg(fem);
mesh_test = 0;
% Initialize mesh
fem.mesh = meshinit(fem);
mesh_test = size(fem.mesh.t,2);
while mesh_test<1100
fem.mesh = meshrefine(fem);
mesh_test = size(fem.mesh.t,2);
end
% Application mode 1
clear appl
appl.mode.class = 'FIDiffusion';
appl.shape = {'shlag(2,"lm1")','shlag(2,"c")'};
appl.assignsuffix = '.di';
clear prop
prop.weakconstr = struct('value',{'non-ideal'},'dim',{{'lm1'}});
appl.prop = prop;
clear bnd
bnd.c0 = {0,1,0};
bnd.type = {'conf','C','N0'};
bnd.weakconstr = {0,1,0};
bnd.weshape = 1;
bnd.ind = [3,3,1,3,3,3,1,2,3,1,1,3,3];
appl.bnd = bnd;
clear equ
equ.shape = 2;
equ.init = 1;
equ.D = {1,'Dam/Dai'};
equ.R = {'Vmax*Rislet'^2/(Dai*Pext)*c/(c+Km/Pext)*(c>Pcrit/Pext),0};
equ.ind = [1,2,1];
appl.equ = equ;
fem.appl{1} = appl;
fem.border = 1;
fem.outform = 'general';
% Multiphysics
fem=multiphysics(fem);
% Extend mesh
fem.xmesh=meshextend(fem);
% Solve problem
[fem.sol,fem.stop]=femlin(fem, ...
'solcomp',{'c','ln1'}, ...
'outcomp',{'c','ln1'}, ...
'out',{'sol','stop'},...
'hnlin','on','maxiter',100);
% Save current fem structure for restart purposes
fem0=fem;
% Post processing
[max_PO2_1(1)] = postmax(fem,'c','solum',1,'Dl',[1]); % Maximum node in islet
[min_PO2_1(1)] = postmin(fem,'c','solum',1,'Dl',[1]); % Minimum node in islet
[max_PO2_2(1)] = postmax(fem,'c','solum',1,'Dl',[3]); % Maximum node in islet
[min_PO2_2(1)] = postmin(fem,'c','solum',1,'Dl',[3]); % Minimum node in islet
convergence1()=fem.stop;
% Integrate islet without O2
frac_no_O2_1 = postint(fem,1*c<=Pcrit/Pext,'solum',1,'phase',0*pi/180,'edim',3,'intorder',4,'geomnum',1,'dl',[1]);
% Integrate whole islet
islet_int_1 = postint(fem,1,'solum',1,'phase',0*pi/180,'edim',3,'intorder',4,'geomnum',1,'dl',[1]);
frac_dead_1(i)=frac_no_O2_1/islet_int_1;
% Integrate islet without O2
frac_no_O2_2 = postint(fem,1*c<=Pcrit/Pext,'solum',1,'phase',0*pi/180,'edim',3,'intorder',4,'geomnum',1,'dl',[3]);
% Integrate whole islet
islet_int_2 = postint(fem,1,'solum',1,'phase',0*pi/180,'edim',3,'intorder',4,'geomnum',1,'dl',[3]);
frac_dead_2(i)=frac_no_O2_2/islet_int_2;
OCR(i) = postint(fem,'lm', ...
'solum',1, ...
'phase',0*pi/180, ...
'edim',2, ...
'intorder',4, ...
'geomnum',1, ...
'dl',[8]); % Total oxygen flux into system
min_quality_mesh(i) = min(meshqual(fem.mesh));
element_num(i) = size(fem.mesh.t,2);
end
10.7 Matlab code for glucose transport
% FEMLAB Model M-file
% Set up vectors to store specific numbers of solutions
n=6;
max_gluc=linspace(1,n,n);
min_gluc=linspace(1,n,n);
min_quality_mesh=linspace(1,n,n);
element_num=linspace(1,n,n);
Data = zeros(61,n);
% Beginning of for loops
A1 = [0.1 0.5 1 1.5 2 3];
for i=1:n
fclear fem
% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.3';
vrsn.next = "";
vrsn.major = 0;
vrsn.build = 405;
vrsn.rcs = '$Name: $';
vrsn.date = '$Date: 2006/08/31 18:03:47 $';
fem.version = vrsn;
% Get input for parameters
Conc_bulk = 10e-6; % mol/cm3 10mM
V_max = 1e-7; % mol/cm3 s
Km = 8e-6; % mol/cm3 8mM
height = 2.2; % cm
radius_islet=0.0075; % cm
islet_density = 4440; %IE/cm2
diff_islet = 3.8e-7; % cm2/s
diff_medium = 6.7e-6; % cm2/s
B1 = diff_medium/diff_islet;
A = A1(i); % IEQ/cm2
end_time = 60; % hours
% Place constant into fem
fem.const={['Cbulk',Conc_bulk,'A',A,'Km',Km,'B1',B1,'Rislet',radius_islet]};
% Specifies dimensions
fem.sdim=3;
%%%%%%%%%%%%%%% GEOMETRY %%%%%%%%%%%%%%%%
% Function to create geometry for islet array (equally spaced spheres in a
% square array. Only one quarter of an islet will be modeled because of
% symmetry.
pt = [0 0 0]; % base point of 3D block
vec = [0 0 1]; % specifies direction of block
% Set length of sides of block
% scaling parameter for height of medium, radius of islet=1
scale_height = height/radius_islet; % scaling parameter for height of box
scale_space = (((radius_islet^3)/(0.0075^3)/islet_density)^(1/2))/radius_islet; % calculating length of sides for given density
Lx = scale_space; % center to center distance between islets
Ly = scale_space;
Lz = (scale_height-2);
% Create block for medium
B3top=block3(Lx/2,Ly/2,Lz,'corner',[0 0 2],vec);
% Now we will add in the spheres for islets of a given radius
% Spheres at vertices of block B3
S3a = sphere3(1,[0 0 1]);
% Create composite objects.
B3bottom=block3(Lx/2,Ly/2,2,'corner',pi,vec);
% First make spheres into quarters.
COa = S3a*B3bottom;
B3med=B3bottom+B3top;
B3meda=B3med-S3a;
clear s
s.objs={COa,B3meda};
s.name={'CO1','Blk1'};
s.tags={'COa','B3meda'};
fem.draw = struct('s',s);
fem.geom = geomcsg(fem);
%%%%%%%%%%%%%%%%% END GEOMETRY %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Initialize mesh
fem.mesh=meshinit(fem, ...
'hmaxfact',1.5, ...
'hcurve',0.7, ...
'hgrad',1.6, ...
'hcutoff',0.04, ...
'harrow',0.4);
mesh_test = size(fem.mesh.t,2);
while mesh_test<2500
fem.mesh=meshrefine(fem);
mesh_test = size(fem.mesh.t,2);
end
%%%%%%%%%%%%%%%%% Application mode no necrotic core %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Application mode 1
clear appl
appl.mode.class = 'FIDiffusion';
appl.shape = {'shlag(2,"lm1")','shlag(2,"c")'};
appl.assignsuffix = '_di';
clear prop
prop.weakconstr=struct('value',{'non-ideal'},'dim',{{'lm1'}});
appl.prop = prop;
clear bnd
bnd.c0 = {0,0,1};
bnd.type = {'N0','cont','N'};
bnd.weakconstr = {0,1,1};
bnd.woshare = 1;
bnd.ind = [1,1,1,2,1,1,1,1,2,2,3,2,1,1,1,1];
appl.bnd = bnd;
clear equ
equ.shape = 2;
equ.init = 1;
equ.D = {1,'B1'};
equ.R = {'A*c*(c+Km/Cbulk)*(c>0)',0};
equ.ind = [1,2,2];
appl.equ = equ;
fem.appl{1} = appl;
fem.border = 1;
fem.outform = 'general';
% Multiphysics
fem=multiphysics(fem);
% Extend mesh
fem.xmesh=meshextend(fem);
% Solve problem
final=end_time*60*60*diff_islet/radius_islet^2;
[fem.sol]=femtime(fem, ...
'solcomp',{'c'}, ...
'outcomp',{'c'}, ...
'tlist',[0:final/60:final], ...
'tout','tlist');
% Save current fem structure for restart purposes
fem0=fem;
max_gluc(i) = postmax(fem,'c','solnum',61,'Dl',[1]); % Maximum node in islet
min_gluc(i) = postmin(fem,'c','solnum',61,'Dl',[1]); % Minimum node in islet
min_quality_mesh(i) = min(meshqual(fem.mesh));
element_num(i) = size(fem.mesh,1,2);
%Save data at top of islet
Data(:,i)= postint(fem,'c', ...
'dl',[2], ...
'edim',0, ...
'solnum','all');
end
10.8 Matlab code for changes in local pH during culture on polystyrene
% FEMLAB Model M-file
% Set up vectors to store specific numbers of solutions
m = 1801; % time broken into m-1 intervals
n=5;
min_quality_mesh1=linspace(1,n,n);
element_num1=linspace(1,n,n);
pH_Data1 = zeros(m,n);
H_Data1 = zeros(m,n);
HCO3_Data1 = zeros(m,n);
CO2_Data1 = zeros(m,n);
L_Data1 = zeros(m,n);
A_Data1 = zeros(m,n);
Diff_H1 = zeros(m,n);
X = [100 500 1000 2000 4400];
for i=1:n
fclear fem
% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.3';
vrsn.ext = 'a';
vrsn.major = 0;
vrsn.build = 511;
vrsn.rcs = '$Name: $';
vrsn.date = '$Date: 2007/02/02 19:05:58 $';
fem.version = vrsn;
% Get input for parameters
D_CO2_islet = 1.0e-5; % cm^2/s
D_CO2_medium = 2.55e-5; % cm^2/s
D_lac_islet = 5.3e-6; % cm^2/s
D_lac_medium = 1.1e-5; % cm^2/s
D_amm_islet = 2.0e-6; % cm^2/s
D_amm_medium = 2.7e-5; % cm^2/s
D_HCO3_islet = 8.6e-7; % cm^2/s
D_HCO3_medium = 11.9e-6; % cm^2/s
D_H_islet = 6.0e-6; % cm^2/s
D_H_medium = 11.2e-5; % cm^2/s
Rxn_CO2 = 3.68e-8; % mol/s cm^3
Rxn_lac = 5.3e-8; % mol/s cm^3
Rxn_amm = 5.0e-8; % mol/s cm^3
kf_HCO3 = 0.145; % 1/s
kr_HCO3 = 17.2e4*1000.0; % cm^3/mol s
Ext_CO2 = 760.0*0.05; % mmHg
radius_inlet = 0.0075; % cm
islet_density = X(i); % islets/cm^2
initial_pH = 7.4;
initial_H_medium = 10^(initial_pH)/1000; % mol/cm^3
initial_H_islet = initial_H_medium; % mol/cm^3
initial_amm = 0.0/1000.0; % mol/cm^3
initial_lac = 0.0/1000.0; % mol/cm^3
A_CO2_islet = 3.21e-5/1000.0; % mol/cm^3 mmHg
A_CO2_medium = 3.18e-5/1000.0; % mol/cm^3 mmHg
initial_CO2_islet = A_CO2_islet*Ext_CO2; %mol/cm3
initial_CO2_medium = A_CO2_medium*Ext_CO2; %mol/cm3
M = 1.0e6;
initial_HCO3_medium = kf_HCO3/kr_HCO3*initial_CO2_medium/initial_H_medium; %mol/cm3
initial_HCO3_islet = kf_HCO3/kr_HCO3*initial_CO2_islet/initial_H_islet; %mol/cm3
height = 0.3; %cm
end_time = 60.0; % hours
final_end_time*60*60*D_CO2_medium/radius_islet^2;
Da_CO2 = radius_islet^2/D_CO2_medium*Rxn_CO2/initial_CO2_medium;
Da_kfb = radius_islet^2/D_CO2_medium*kf_HCO3;
Da_krb = radius_islet^2/D_CO2_medium*kr_HCO3*initial_CO2_medium;
Da_am = radius_islet^2/initial_CO2_medium*Rxn_amm/D_CO2_medium;
Da_lac = radius_islet^2/D_CO2_medium*Rxn_lac/initial_CO2_medium;
Da_am_lac = Da_lac-Da_am;
% Place constant into fem
% Constants
fem.const = {'D_CO2_islet','D_CO2_islet', ...
'A_CO2_islet','A_CO2_islet', ...
'Ext_CO2','Ext_CO2', ...
'Da_CO2','Da_CO2', ...
'D_lac_islet','D_lac_islet', ...
'D_lac_medium','D_lac_medium', ...
'Rislet','radius_islet', ...
'D_CO2_medium','D_CO2_medium', ...
'A_CO2_medium','A_CO2_medium', ...
'D_amn_islet','D_amn_islet', ...
'D_amn_medium','D_amn_medium', ...
'Da_am_lac','Da_am_lac', ...
'Da_am','Da_am', ...
'Da_lac','Da_lac', ...
'Da_krb','Da_krb', ...
'Da_kfb','Da_kfb', ...
'D_HCO3_islet','D_HCO3_islet', ...
'D_HCO3_medium','D_HCO3_medium', ...
'D_H_islet','D_H_islet', ...
'D_H_medium','D_H_medium', ...
'HCO3_init_medium',initial_HCO3_medium, ...
'HCO3_init_islet',initial_HCO3_islet, ...
'H_init_medium',initial_H_medium, ...
'H_init_islet',initial_H_islet, ...
'lac_init',initial_lac, ...
'amn_init',initial_amn, ...
'CO2_init_medium',initial_CO2_medium, ...
'CO2_init_islet',initial_CO2_islet, ...
'tau_end',final, ...
'M,M'};
% Specifies dimensions
fem.sdim=3;
% GEOMETRY
% Function to create geometry for islet array (equally spaced islets in a
% square array. Only an eighth of an islet will be modeled because of
% symmetry.
% Set length of sides of block
% scaling parameter for height of medium, radius of islet=1
scale_height = height/radius_islet; % scaling parameter for height of box
scale_space = ((1/islet_density)^(1/2))/radius_islet; % calculating length of sides for given density
Lx = scale_space; % center to center distance between islets
Ly = scale_space;
Lz = (scale_height);
% Create block for medium
B3_1=block3(Lx,Ly,Lz,'base','corner','pos',{'0','0','0'},'axis',{'0','0','1'},'rot','0');
B3_2=block3(Lx,Ly,Lz,'base','corner','pos',{'0','0','0'},'axis',{'0','0','1'},'rot','45');
B3_3=block3(Lx/2,Ly/2,Lz,'base','corner','pos',{'0','0','0'},'axis',{'0','0','1'},'rot','0');
B3_4=B3_1*B3_2;
B3_triangle1=B3_3*B3_4;
% Now we will add in the sphere for islet of a given radius set to 1
S3a = sphere3(1,[0 0 1]);
% make sphere into one eighth of islet
S3_eighth=S3a*B3_triangle1;
% remove corner from lower block
B3_triangle2=B3_triangle1-S3a;
B3_triangle3=B3_triangle2+S3_eighth;
% Combine into a single geometry
clear s
s.objs={B3_triangle3};
s.name={'CO1'};
s.tags={'B3_triangle3'};
fem.draw = struct('s',s);
fem.geom = geomcg(fem);
% Initialize mesh
fem.mesh=meshinit(fem, ...
'hauto',6);
mesh_test = size(fem.mesh.t,2);
while mesh_test<1000
fem.mesh=meshrefine(fem);
mesh_test = size(fem.mesh.t,2);
end
% (Default values are not included)
% Application mode 1
clear appl
appl.mode.class = 'Diffusion';
appl.dim = {'c7'};
appl.module = 'CHEM';
appl.shape = {'shlag(2,"lm8")','shlag(2,"c7")'};
appl.gporder = {10,4};
appl.assignsuffix = '_chdi';
clear prop
clear weakconstr
weakconstr.value = 'non-ideal';
weakconstr.dim = {'lm8'};
prop.weakconstr = weakconstr;
appl.prop = prop;
clear bnd
bnd.type = {'N0','N','cont'};
bnd.weakconstr = {0,0,1};
bnd.weshape = 1;
bnd.N = {0,'M*(c8-c7/A_CO2_islet*A_CO2_medium)',0};
bnd.ind = [1,1,2,3,3,3,2,3,3];
appl.bnd = bnd;
clear equ
equ.D = {'D_CO2_islet/D_CO2_medium',1};
equi.init = {'CO2_init_islet/CO2_init_medium',0};
equi.R = {'Da_CO2-Da_kfb*c7+Da_krb*c6*c2', ... 0};
equi.shape = 2;
equi.gporder = 2;
equi.usage = {1,0};
equi.ind = [1,2];
appl.equ = equ;
fem.appl{1} = appl;
% Application mode 2
clear appl
appl.mode.class = 'Diffusion';
appl.dim = {'c8'};
appl.name = '_chd12';
appl.module = 'CHEM';
appl.shape = {'shlag(2,"lm9")','shlag(2,"c8")'};
appl.gporder = {10,4};
appl.assignsuffix = '_chd12';
clear prop
clear weakconstr
weakconstr.value = 'non-ideal';
weakconstr.dim = {'lm9'};
prop.weakconstr = weakconstr;
appl.prop = prop;
clear bnd
bnd.type = {'cont','N','N0','C'};
bnd.c0 = {0,0,0,1};
bnd.weakconstr = {1,0,0,1};
bnd.wshape = 1;
bnd.N = {0,'M*(c7*A_CO2_medium/A_CO2_islet-c8)',0,0};
bnd.ind = [1,1,2,3,3,3,2,4,3];
appl.bnd = bnd;
clear equ
equi.init = {0,1};
equi.R = {0,'-Da_kfb*c8+Da_krb*c6*c2'};
equi.shape = 2;
equi.gporder = 2;
equi.usage = {0,1};
equi.ind = [1,2];
appl.equ = equ;
fem.appl{2} = appl;
% Application mode 3
clear appl
appl.mode.class = 'Diffusion';
appl.dim = {'c2','c3','c4','c6'};
appl.name = '_chd13';
appl.module = 'CHEM';
appl.gporder = 4;
appl.cporder = 2;
appl.assignsuffix = '_chd13';
clear bnd
bnd.type = { {'N0';'N0';'N0';'N0'}, {'cont';'cont';'cont';'cont'} };
bnd.ind = [1,1,2,1,1,1,2,1,1];
appl.bnd = bnd;
clear equ
equ.D = {'D_HCO3_islet/D_CO2_medium';'D_lac_islet/D_CO2_medium';'D_amn_islet/D_CO2_medium'; ...
'D_H_islet/D_CO2_medium'},{'D_HCO3_medium/D_CO2_medium';'D_lac_medium/D_CO2_medium'; ...
'D_amn_medium/D_CO2_medium';'D_H_medium/D_CO2_medium'}};
equ.init = {'HCO3_init_islet/CO2_init_medium';'lac_init/CO2_init_medium'; ...
'amn_init/CO2_init_medium';'H_init_islet/CO2_init_medium'},{'HCO3_init_medium/CO2_init_medium'; ...
'lac_init/CO2_init_medium';'amn_init/CO2_init_medium';'H_init_medium/CO2_init_medium'}};
equ.R = {'{Da_kfb*c7-Da_krb*c6*c2';'Da_lac'; ...
'Da_am';'Da_am_lac+Da_kfb*c7-Da_krb*c6*c2'}, ...
'{Da_kfb*c8-Da_krb*c6*c2';0;0;'Da_kfb*c8-Da_krb*c6*c2'}};
equ.ind = [1,2];
appl.equ = equ;
fem.appl(3) = appl;
fem.border = 1;
fem.outform = 'general';
% Subdomain settings
clear equ
equ.ind = [1,2];
equ.dim = {'c7','c8','c2','c3','c4','c6'};
% Subdomain expressions
equ.expr = {'c_all','{c7','c8'}};
fem.equ = equ;
% Multiphysics
fem=multiphysics(fem);
% Extend mesh
fem.xmesh=meshextend(fem);
% Solve problem
fem.sol=femtime(fem, ...
'solcomp',{'c2','lm9','c7','c8','c4','c3','c6'}, ...
'outcomp',{'c2','lm9','c7','c8','c4','c3','c6'}, ...
'tlist',[0:final/(m-1):final], ...
'atol',{'c2','1e-6','c3','1e-6','c4','1e-6','c6','1e-8','c7','1e-6','c8','1e-6','lm9','1e-6'}, ...
'tout','tlist');
min_quality_mesh1(i) = min(meshqual(fem.mesh));
element_num(i) = size(fem.mesh.t,2);
%save data at top of islet
pH_Data1(:,i)= postint(fem,'-log10(c6*CO2_init_medium*1000}', ...
'dl',[2], ...
'edim',0, ...
'solnum','all');
L_Data1(:,i)= postint(fem,'c3*CO2_init_medium*1000*1000', ...
'dl',[2], ...
'edim',0, ...
'solnum','all');
A_Data1(:,i)= postint(fem,'c4*CO2_init_medium*1000*1000', ...
'dl',[2], ...
'edim',0, ...
'solnum','all');
HC03_Data1(:,i)= postint(fem,'c2*CO2_init_medium*1000*1000', ...
'dl',[2], ...
CO2_Data1(:,i)= postint(fem,'c_all*CO2_init_medium*1000*1000', ...
'dl',[2], ...
'edim',0, ...
'solnum','all');
H_Data1(:,i)= postint(fem,'c6*CO2_init_medium*1000*1000', ...
'dl',[2], ...
'edim',0, ...
'solnum','all');
Diff_H1(:,i)=postint(fem,'ndflux_c6_chdi3', ...
'unit'," ...
'dl',[3,7], ...
'edim',2, ...
'solnum','all');
end
10.9 Matlab code for changes in local pH during culture on silicone rubber
% FEMLAB Model M-file
% Set up vectors to store specific numbers of solutions
%%%%%%% Stored values for solution to simulated culture
m = 1801; %time broken into m-1 intervals
n=2;
min_quality_mesh1=linspace(1,n,n);
element_num1=linspace(1,n,n);
pH_Data1 = zeros(m,n);
H_Data1 = zeros(m,n);
X = [100 4400];
%%%%%%%%%%%%%% Si exp
%%%%%%%%%%%%%%%
for i=1:n
fclear fem
% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.3';
vrsn.ext = 'a';
vrsn.major = 0;
vrsn.build = 511;
vrsn.rcs = '$Name: $';
vrsn.date = '$Date: 2007/02/02 19:05:58 $';
fem.version = vrsn;
% Get input for parameters
D_CO2_islet = 1.0e-5; %cm2/s
D_CO2_medium = 2.55e-5; %cm2/s
D_CO2_Si = 2.9e-5; %cm2/s
D_lac_islet = 5.3e-6; %cm2/s
D_lac_medium = 1.1e-5; %cm2/s
D_amm_islet = 2.0e-6; %cm2/s
D_amm_medium = 2.7e-5; %cm2/s
D_HCO3_islet = 8.6e-7; %cm2/s
D_HCO3_medium = 11.9e-6; %cm2/s
D_H_islet = 6.0e-6; %cm2/s
D_H_medium = 11.2e-5; %cm2/s
Rxn_CO2 = 3.68e-8; %mol/s cm$^3$
Rxn_lac = 5.3e-8; %mol/s cm$^3$
Rxn_amm = 5.0e-8; %mol/s cm$^3$
kf_HCO3 = 0.145; %1/s
kr_HCO3 = 17.2e4*1000.0; %cm$^3$/mol s
Ext_CO2 = 760.0*0.05; %mmHg
radius_islet = 0.0075; %cm
islet_density = X(i); % islets/cm$^2$
initial_pH = 7.4;
initial_H_medium = 10^(-initial_pH)/1000; %mol/cm$^3$
initial_H_islet = initial_H_medium; %mol/cm$^3$
initial_amm = 0.0/1000.0; %mol/cm$^3$
initial_lac = 0.0/1000.0; %mol/cm$^3$
A_CO2_islet = 3.21e-5/1000.0; % mol/cm$^3$ mmHg
A_CO2_medium = 3.18e-5/1000.0; % mol/cm$^3$ mmHg
A_CO2_Si = 7.7e-5/1000.0; % mol/cm$^3$ mmHg
initial_CO2_islet = A_CO2_islet*Ext_CO2; %mol/cm3
initial_CO2_medium = A_CO2_medium*Ext_CO2; %mol/cm3
initial_CO2_Si = A_CO2_Si*Ext_CO2; %mol/cm3
M = 1.0e6;
initial_HCO3_medium = kf_HCO3/kr_HCO3*initial_CO2_medium/initial_H_medium; %mol/cm3
initial_HCO3_islet = kf_HCO3/kr_HCO3*initial_CO2_islet/initial_H_islet; %mol/cm3
height = 2.2; %cm
thickness = 0.05; %cm
end_time = 60.0; % hours
final_end_time*60*60*D_CO2_medium/radius_islet^2;
Da_CO2 = radius_islet^2/D_CO2_medium*Rxn_CO2/initial_CO2_medium;
Da_kfb = radius_islet^2/D_CO2_medium*kf_HCO3;
Da_krb = radius_islet^2/D_CO2_medium*kr_HCO3*initial_CO2_medium;
Da_am = radius_islet^2/initial_CO2_medium*Rxn_amm/D_CO2_medium;
Da_lac = radius_islet^2/D_CO2_medium*Rxn_lac/initial_CO2_medium;
Da_am_lac = Da_lac-Da_am;
% Place constant into fem
% Constants
fem.const = {'D_CO2_islet', 'D_CO2_islet', ...
'A_CO2_islet', 'A_CO2_islet', ...
'D_CO2_Si', 'D_CO2_Si', ...
'A_CO2_Si', 'A_CO2_Si', ...
'Ext_CO2', 'Ext_CO2', ...
'Da_CO2', 'Da_CO2', ...
'D_lac_islet', 'D_lac_islet', ...
'D_lac_medium', 'D_lac_medium', ...
'Rislet', 'radius_islet', ...
'D_CO2_medium', 'D_CO2_medium', ...
'A_CO2_medium', 'A_CO2_medium', ...
'D_amm_islet', 'D_amm_islet', ...
'D_amm_medium', 'D_amm_medium', ...
'Da_am_lac', 'Da_am_lac', ...
'Da_am', 'Da_am', ...
'Da_lac', 'Da_lac', ...
'Da_kfb', 'Da_kfb', ...
'Da_krb', 'Da_krb', ...
'D_HCO3_islet', 'D_HCO3_islet', ...
'D_HCO3_medium', 'D_HCO3_medium', ...
'D_H_islet', 'D_H_islet', ...
'D_H_medium', 'D_H_medium', ...
'HCO3_init_medium', 'initial_HCO3_medium', ...
'HCO3_init_islet', 'initial_HCO3_islet', ...
'H_init_medium', 'initial_H_medium', ...
'H_init_islet', 'initial_H_islet', ...
'lac_init', 'initial_lac', ...
'amm_init', 'initial_amm', ...
'CO2_init_medium', 'initial_CO2_medium', ...
'CO2_init_islet', 'initial_CO2_islet', ...
'tau_end', 'final', ...
'thickness', 'thickness', ...
'initial_CO2_Si', 'initial_CO2_Si', ...
'M', 'M'};
% Specifies dimensions
fem.sdim=3;
% GEOMETRY
% Function to create geometry for islet array (equally spaced islets in a square array. Only an eighth of an islet will be modeled because of symmetry.
% Set length of sides of block
% scaling parameter for height of medium, radius of islet=1
scale_height = height/radius_islet; % scaling parameter for height of box
scale_space = ((1/islet_density)^(1/2))/radius_islet; % calculating length of sides for given density
Lx = scale_space; % center to center distance between islets
Ly = scale_space;
Lz = (scale_height-2);
pt = [0 0 0]; % base point of 3D block
vec = [0 0 1]; % specifies direction of block
Lzmem=thickness/radius_islet;
Lztot = Lzmem+2+Lz;
% Create block for medium
B3top=block3(Lx/2,Ly/2,Lz,'corner',[0 0 2],vec);
% Now we will add in the spheres for islets of a given radius
% Spheres at vertices of block B3
S3a = sphere3(1,[0 0 1]);
% Create composite objects.
B3bottom=block3(Lx/2,Ly/2,2,'corner',pt,vec);
% First make sphere into quarters.
COa = S3a*B3bottom;
B3_2=block3(Lx,Ly,Lztot,'base','corner','pos',{0,0,-Lzmem},'axis',{'0','0','1'},'rot','45');
B3mem=block3(Lx/2,Ly/2,Lzmem,'corner',[0,0,-Lzmem],vec);
COb = COa-B3_2;
B3med=B3bottom+B3top;
B3meda = B3med+B3mem;
B3medb=B3meda-B3_2;
B3medd=B3medb-S3a; % remove corner from block
B3medd=B3medc+COb;
clear s
s.objs={B3medd};
s.name={'CO1'};
s.tags={'B3medd'};
fem.draw = struct('s',s);
fem.geom = geomcsg(fem);
% Initialize mesh
fem.mesh=meshinit(fem, ...
'hauto',4, ...
'hmaxsub',[2,1]);
fem.mesh=meshrefine(fem, ...
'mcase',0, ...
'rmethod','longest', ...
'subdomain',[2,3]);
mesh_test = size(fem.mesh.t,2);
while mesh_test<1000
fem.mesh=meshrefine(fem);
mesh_test = size(fem.mesh.t,2);
end
% Application mode 1
clear appl
appl.mode.class = 'Diffusion';
appl.dim = {'c7'};
appl.module = 'CHEM';
appl.shape = {'shlag(2,'lm8')','shlag(2,'c7')'};
appl.gporder = {10,4};
appl.assignsuffix = '_chd2';
clear prop
clear weakconstr
weakconstr.value = 'non-ideal';
weakconstr.dim = {'lm8'};
prop.weakconstr = weakconstr;
appl.prop = prop;
clear bnd
bnd.type = {'cont','cont','N','N0','N','C'};
bnd.c0 = {0,0,0,0,0,1};
bnd.weakconstr = {0,1,0,0,0,1};
bnd.weshape = 1;
bnd.N = {0,0,M*(c8-c7/A_CO2_medium*A_CO2_islet'),0,M*(c7/A_CO2_medium*A_CO2_Si-c9)', ... 0};
bnd.ind = [1,1,1,2,2,3,4,4,5,4,4,3,1,6,1,4,4];
appl.bnd = bnd;
clear equ
equ.D = {'D_CO2_islet/D_CO2_medium',1,1};
equ.init = {'CO2_init_islet/CO2_init_medium',0,1};
equ.R = {0,0,'Da_kfb*c7+Da_krb*c6*c2'};
equ.shape = 2;
equ.gporder = 2;
equ.usage = {0,0,1};
equ.ind = [1,2,3,3];
appl.equ = equ;
fem.appl{1} = appl;
% Application mode 2
clear appl
appl.mode.class = 'Diffusion';
appl.dim = {'c8'};
appl.name = 'chd2';
appl.module = 'CHEM';
appl.shape = {'shlag(2,'lm9')','shlag(2,'c8")'};
appl.gporder = {10,4};
appl.assignsuffix = '_chd2';
clear prop
clear weakconstr
weakconstr.value = 'non-ideal';
weakconstr.dim = {'lm9'};
prop.weakconstr = weakconstr;
appl.prop = prop;
clear bnd
bnd.type = {'cont','cont','N0','N','C'};
bnd.c0 = {0,0,0,0,1};
bnd.weakconstr = {1,0,0,0,1};
bnd.weshape = 1;
bnd.N = {0,0,0,M*(c7*A_CO2_islet/A_CO2_medium-c8),0};
bnd.ind = [1,1,2,3,3,4,2,5,2,1,1,4,1,1,1,1,1,1];
appl.bnd = bnd;
clear equ
equ.D = {1,'D_CO2_islet/D_CO2_medium'};
equ.init = {0,'CO2_init_islet/CO2_init_medium'};
equ.R = {0,'Da_CO2-Da_kfb*c8+Da_krb*c6*c2'};
equ.shape = 2;
equ.gporder = 2;
equ.usage = {0,1};
equ.ind = [1,2,1,1];
appl.equ = equ;
fem.appl{2} = appl;
% Application mode 3
clear appl
appl.mode.class = 'Diffusion';
appl.dim = {'c2','c3','c4','c6'};
appl.name = 'chdi3';
appl.module = 'CHEM';
appl.gporder = 4;
appl.cporder = 2;
appl.lassignsuffix = '_chdi3';
clear bnd
bnd.type = {'[cont';'cont';'cont';'cont'},{'N0';'N0';'N0';'N0'}};
bnd.ind = [1,1,1,2,2,1,2,2,2,2,2,1,1,2,1,2,2];
appl.bnd = bnd;
clear equ
equ.D = {'D_HCO3_islet/D_CO2_medium';'D_lac_islet/D_CO2_medium';'D_amn_islet/D_CO2_medium'; ...
'D_H_islet/D_CO2_medium'},{'D_HCO3_islet/D_CO2_medium';'D_lac_islet/D_CO2_medium'; ...
'D_amn_islet/D_CO2_medium';'D_H_islet/D_CO2_medium'},{'D_HCO3_medium/D_CO2_medium'; ...
'D_lac_medium/D_CO2_medium';'D_amn_medium/D_CO2_medium';'D_H_medium/D_CO2_medium'}};
equ.init = {'HC03_init_islet/CO2_init_medium';'lac_init/CO2_init_medium'; ...
'amn_init/CO2_init_medium';'H_init_islet/CO2_init_medium'},{'HC03_init_islet/CO2_init_medium'; ...
'lac_init/CO2_init_medium';'amn_init/CO2_init_medium';'H_init_islet/CO2_init_medium'}, ...
{'HC03_init_medium/CO2_init_medium';'lac_init/CO2_init_medium';'amn_init/CO2_init_medium'; ...
'H_init_medium/CO2_init_medium'}};
equ.R = {'Da_kfb*c7-Da_krb*c6*c2';'Da_lac';'Da_am';'Da_am_lac+Da_kfb*c7-Da_krb*c6*c2'}, ...
{'Da_kfb*c8*Da_krb*c6*c2';'Da_lac';'Da_am';'Da_am_lac+Da_kfb*c8-Da_krb*c6*c2'}, ...
{'Da_kfb*c7*Da_krb*c6*c2';0;0;'Da_kfb*c7-Da_krb*c6*c2'};
equ.usage = {0,1,1};
equ.ind = [1,2,3,3];
appl.equ = equ;
fem.appl{3} = appl;
% Application mode 4
clear appl
appl.mode.class = 'Diffusion';
appl.dim = {'c9'};
appl.name = 'chdi4';
appl.module = 'CHEM';
appl.shape = {'shlag(2,"lm5")','shlag(2,"c9")'};
appl.gporder = {10,4};
appl.lassignsuffix = '_chdi4';
clear prop
prop.analysis='static';
clear weakconstr
weakconstr.value = 'non-ideal';
weakconstr.dim = {'lm5'};
prop.weakconstr = weakconstr;
appl.prop = prop;
clear bnd
bnd.type = {'N0','C','cont','N'};
bnd.c0 = {0,'Ext_CO2*A_CO2_Si/CO2_init_medium',0,0};
bnd.weakeconstr = {0,1,1,0};
bnd.wcsshape = 1;
bnd.N = {0,0,0,'M*(c9/A_CO2_Si*A_CO2_medium-c7)};
bnd.ind = [1,1,2,3,3,3,3,3,4,3,3,3,3,3,1,3,3];
appl.bnd = bnd;
clear equ
equ.D = {'D_CO2_Si/D_CO2_medium',1};
equ.init = {'CO2_init_Si/CO2_init_medium',0};
equ.shape = 2;
equ.gporder = 2;
equ.usage = {1,0};
equ.ind = [1,2,2,2];
appl.equ = equ;
fem.appl{4} = appl;
fem.border = 1;
fem.outform = 'general';
% Subdomain settings
clear equ
equ.ind = [1,2,3,3];
equ.dim = {'c7','c8','c2','c3','c4','c6','c9'};
% Subdomain expressions
equ.expr = {'c_all',{'c9','c8','c7}};
fem.equ = equ;
% Multiphysics
fem=multiphysics(fem);
% Extend mesh
fem.xmesh=meshextend(fem);
% Solve problem
fem.sol=femtime(fem, ...
'solcomp',{'lm5','c9','c2','c7','c8','c3','c4','c6'}, ...
'outcomp',{'lm5','c9','c2','c7','c8','c3','c4','c6'}, ...
'thist',[0:final/(m-1):final], ...
'atol',{'c2','1e-6','c3','1e-6','c4','1e-6','c6','1e-8','c7','1e-6','c8','1e-6','c9','1e-6','lm5','1e-6'}, ...
tout,thist');
min_quality_mesh1(i) = min(meshqual(fem.mesh));
element_num1(i) = size(fem.mesh.t,2);
%save data at top of islet
pH_Data(:,i)= postint(fem,'-log10(c6*CO2_init_medium*1000)', ...
'dl',[3], ...
'edim',0, ...
'solnum','all');
H_Data(:,i)= postint(fem,'c6*CO2_init_medium*1000*1000', ...
'dl',[3], ...
'edim',0, ...
'solnum','all');
end
10.10 Matlab code for oxygen transport in OBS – base case
% FEMLAB Model M-file
fclear fem
% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.3';
vrsn.ext = '';
vrsn.major = 0;
vrsn.build = 405;
vrsn.rcs = '$Name: $';
vrsn.date = '$Date: 2006/08/31 18:03:47 $';
fem.version = vrsn;
% Constants
Da_medium = 3.53e-14; % mol/s/mmHg/cm
Radius_well = 0.32; % cm radius of well
Cell_density = 250000; % cells/cm2
h_medium = 0.1/(pi*Radius_well^2); % cm medium depth
P_ext = 142; % mmHg External pO2
Km = 0.44; % mmHg Km for oxygen consumption
P_critical = 0.1; % mmHg Critical pO2 of cells
Reader_position = 0.8*0.32/Radius_well; % non-dimensional length that plate reader sees
OCR_cell = 2.8; % fmol/min cell
V_max = OCR_cell/60*Cell_density*10^-15; %mol/cm2 s OCR of cells
fem.const={'Rwell','Radius_well','h_medium','h_medium','Pext','P_ext','Km','Km','Da','Da_medium','Vmax','V_max',...
'Pcrit','P_critical'};
% Geometry
Media=rect2('l','h_medium/Rwell','base','corner','pos',{'0','0'},'rot','0','const',fem.const);
parr=[point2(0,0,0)];
Pt=geomcoerce('point',parr);
clear p s
p.objs={Pt};
p.name=['PT1'];
p.tags={'Pt'};
s.objs={Media};
s.name={'R1'};
s.tags={'Media'};
fem.draw=struct('p','p','s','s');
fem.geom = geomcsg(fem);
% Initialize mesh
fem.mesh=meshinit(fem,...
'hmaxfact',0.55,...
'hcutoff',0.01,...
'hgrad',1.4,...
'hcurve',0.4,...
'harrow',0.7);
mesh_test = size(fem.mesh.t,2);
if mesh_test<2500
fem.mesh=meshrefine(fem);
mesh_test = size(fem.mesh.t,2);
end
% (Default values are not included)
% Application mode 1
clear appl
appl.mode.class = 'FIDiffusion';
appl.mode.type = 'axi';
appl.shape = {'shlag(2,"lm1")','shlag(2,"c")'};
appl.assignsuffix = '_di';
clear prop
prop.weakconstr=struct('value',{'non-ideal'},'dim',{ {'lm1'} });
appl.prop = prop;
% Bnd 1 = Center of well, Bnd 2 = Bottom of well where plate reader reads, Bnd 3 = top, Bnd 4 = Bottom of the well where plate reader doesn't read, Bnd 5 = the inner side of the wall
clear bnd
bnd.N = {0,'-Vmax*Rwell/(Da*Pext)*c/(c+Km/Pext)*(c>0)',0,'-Vmax*Rwell/(Da*Pext)*c/(c+Km/Pext)*(c>0)',0};
bnd.c0 = {0,0,1,0,0};
bnd.type = {'N0','N','N','N','N0'};
bnd.weakconstr = {0,0,1,0,0};
bnd.wshape = 1;
bnd.ind = [1,2,3,4,5];
appl.bnd = bnd;
% Subdomain 1 = media
clear equ
equ.shape = 2;
equ.int = 1; % initial pO2
equ.ind = [1];
appl.equ = equ;
fem.appl{1} = appl;
fem.sdim = {'r','z'};
fem.border = 1;
fem.outform = 'general';
% Multiphysics
fem=multiphysics(fem);
% Extend mesh
fem.xmesh=meshextend(fem);
% Solve problem
[fem.sol, fem.stop]=femmlin(fem, ...
'solcomp',{'c','lm1'}, ...
'outcomp',{'c','lm1'}, ...
'out',{'sol','stop'},...
'maxiter',100,...
'ntol', 1e-006);
% Save current fem structure for restart purposes
femo=fem;
convergence=fem.stop;
%%%%%%%
Total_1 =postint(fem,2*pi*r*(1/(1+0.03125*c*Pext)), ...
'dl',[2],'edim',1);
Total_PO2_bottom =postint(fem,2*pi*r*c, ...
'dl',[2],'edim',1);
min_PO2 = postmin(fem,'c','solnum',1,'DL',[2 4],'edim',1);
min_quality_mesh = min(meshqual(fem.mesh));
element_num = size(fem.mesh.t,2);
Top_flux =postint(fem,2*pi*lm1, ...
'dl',[3], ...
'edim',1);
10.11 Matlab code for oxygen transport in OBS – full model
% FEMLAB Model M-file
fclear fem
% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.3';
vrsn.ext = '';
vrsn.major = 0;
vrsn.build = 405;
vrsn.rcs = '$Name: $';
vrsn.date = '$Date: 2006/08/31 18:03:47 $';
fem.version = vrsn;
% Diffusivities of oxygen
Diff_silicone_fl = 2.17e-5; % cm^2/s
Diff_polystyrene = 1; % cm^2/s
Diff_medium = 2.78e-5; % cm^2/s
Diff_tissue = 1.22e-5; % cm^2/s
% Solubilities of oxygen
Alpha_silicone_fl = 1.21e-8; % mol/cm^3 mmHg
Alpha_polystyrene = 3.5e-15; % mol/cm^3 mmHg
Alpha_medium = 1.27e-9; % mol/cm^3 mmHg
Alpha_tissue = 1.02e-9; % mol/cm^3 mmHg
Radius_well = 0.32; % cm radius of well
Cell_density = 250000; % cells/cm^2
h_medium = 0.3; % cm medium depth
P_ext = 142; % mmHg External pO2
Km = 0.44; % mmHg Km for oxygen consumption
P_critical = 0.1; % mmHg Critical pO2 of cells
Poly_thickness_bot = 0.12; % cm polystyrene thickness for bottom
Poly_thickness_side = 0.12; % cm polystyrene thickness for side
Dish_height = 1; % cm height of well
Tissue_volume = 0.008015*Radius_well^3; % cm^3
Silicone_fl_thickness = 0.0625; % cm thickness of silicone with fl
Reader_position = 0.8*0.32/Radius_well; % non-dimensional length that plate reader sees
OCR_cell = 2.8; % fmol/min cell
V_max = OCR_cell/60*Cell_density*10^-15*(pi*Radius_well^2)/Tissue_volume; %mol/cm^3 s OCR of cells
fem.const={Rwell,Radius_well,h_medium,h_medium,Pext,P_ext,Km,Km,
'D_medium',Diff_medium,'D_silicone_fl',Diff_silicone_fl,...
D_poly,Diff_polystyrene,D_tissue,Diff_tissue,A_silicone_fl,
Alpha_silicone_fl,
A_poly,Alpha_polystyrene,A_medium,Alpha_medium,A_tissue,Alpha_tissue,
Vmax,V_max,... Perit,P_critical,poly_bot,Poly_thickness_bot,poly_side',
Poly_thickness_side,... 'tissue',V,Tissue_volume,'Si_fl',Silicone_fl_thickness,'h_dish',
Dish_height,'Read_plate',Reader_position};
%% Geometry
g2=circ2(1,'base','center','pos',{'0','1'},'rot','0');
g3=rect2(1,'1','base','corner','pos',{'0','0'},'rot','0');
% this is curved portion of media
g4=geomcomp({g2,g3},'ns',{'g2','g3'},'sf','g2*g3','edge','none');
% this is top of media
g5=rect2(1,((0.1 - 1.848844*Radius_well^3)/(pi*Radius_well^2)/Radius_well),'base','corner','pos',{'0','1'},'rot','0');
g7=circ2(1+Poly_thickness_side/Radius_well,'base','center','pos',{'0','1'},'rot','0');
g8=square2(1+Poly_thickness_side/Radius_well,'base','corner','pos',{'0',-Poly_thickness_side/Radius_well},'rot','0');
% this is round part of polystyrene
g9=geomcomp({'g7','g8'},'ns',{'g7','g8'},'sf','g7*g8','edge','none');
% this is rectangular part of polystyrene
g10=rect2(Poly_thickness_side/Radius_well,(Dish_height/Radius_well-1),'base','corner','pos',{'1','1'},'rot','0');% height of 1 cm for well
%%% At this point we have the round bottom with polystyrene
% Layer of Si
g20=ellip2('1.08','0.83','base','center','pos',{'0','1'},'rot','0');
g29=geomcomp({g4,g20},'ns',{'CO1','E2'},'sf','CO1-E2','edge','none');
g30=geomcomp({'g29'},'ns',{'CO4'},'sf','CO4','edge','none');
% Layer of tissue
g41=ellip2('1.1','0.82','base','center','pos',{'0','1'},'rot','0');
g43=geomcomp({g4,g41},'ns',{'CO1','E3'},'sf','CO1-E3','edge','none');
g44=geomcomp({g43},'ns',{'COS'},'sf','COS','edge','none');
% making the point of the plate-reader
parr={point2(0,8,0,4)};
g12=geomcoerce('point',parr);
G1 = g4 + g5;
G2 = G1 + g9;
G3 = G2 + g10;
G4 = G3 + g30;
G5 = G4 + g44;
final = G5 + g12;
clear s
s.objs={final};
s.name={'CO1'};
s.tags={'final'};
fem.draw=struct('s',s);
fem.geom=geomsg(fem);
fem.sdim = {'1','2'};
fem.border = 1;
% Multiphysics
fem=multiphysics(fem);
% Mesh
fem.mesh = meshinit(fem);
fem.mesh = meshrefine(fem);
% Application mode 1
clear appl
appl.mode.class = 'FIDiffusion';
appl.mode.type = 'axi';
appl.shape = {'shlag(2,"lm1")','shlag(2,"c")'};
appl.border = 'on';
appl.assigsuffix = '_di';
clear prop
clear weakconstr
weakconstr.value = 'non-ideal';
weakconstr.dim = {'lm1'};
prop.weakconstr = weakconstr;
appl.prop = prop;
clear bnd
bnd.type = {'N0','cont','C'};
bnd.c0 = {0,0,1};
bnd.weakconstr = {0,0,1};
bnd.weshape = 1;
bnd.ind = [1,1,1,1,1,2,3,2,2,3,3,3,3,2,2,2,2,2,2,2];
appl.bnd = bnd;
clear equ
equ.D = {'D_poly*A_poly/(D_medium*A_medium)',
'D_silicone_fl*A_silicone_fl/(D_medium*A_medium)',
'D_tissue*A_tissue/(D_medium*A_medium)',1};
equ.int = 1;
equ.R = {0,0,
'-Vmax*Rwell^2/(D_medium*A_medium*Pext)*c/(c+Km/Pext)*(c-Pcrit/Pext),0};
equ.shape = 2;
equ.ind = [1,2,3,4,4,1,2];
appl.equ = equ;
fem.appl(1) = appl;
fem.sdim = {'r','z'};
fem.border = 1;
fem.outform = 'general';
% Multiphysics
fem=multiphysics(fem);
% Extend mesh
fem.xmesh=meshextend(fem);
%%%%%%%%%%%%%%% Solve problem
[fem.sol,fem.stop]=femlin(fem, ...
'solcomp',{'c','lm1'}, ...
'outcomp',{'c','lm1'}, ...
'out',{'sol','stop'},...
'maxiter',100,...
'ntol',1e-006);
% Save current fem structure for restart purposes
fem0=fem;
convergence=fem.stop;
%%%%%%%%%%%%%%% Analysis
% Integrate
Total_I =postint(fem,'2*pi*r*(1/(1+0.03125*c*Pext))', ...
'dl',[2 7]);
Total_PO2_fl =postint(fem,'2*pi*r*c', ...
'dl',[2 7]);
Total_PO2_cell =postint(fem,'2*pi*r*c', ...
'dl',[3]);
min_PO2 = postmin(fem,'c','solnum',1,'dl',[3]);
min_quality_mesh = min(meshqual(fem.mesh));
element_num = size(fem.mesh,1,2);
Top_flux =postint(fem,'2*pi*lm1', ...
'dl',[7], ...
'edim',1);
Poly_flux = postint(fem,'2*pi*lm1', ...
'dl',[10 11 12 13], ...
'edim',1);
Total_flux =Poly_flux+Top_flux;
10.12 Matlab code for oxygen transport in OBS – flat bottom
% FEMLAB Model M-file
fclear fem
% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.3';
vrsn.ext = "";
vrsn.major = 0;
vrsn.build = 405;
vrsn.rcs = $Name: $;
vrsn.date = $Date: 2006/08/31 18:03:47 $;
fem.version = vrsn;
% Diffusivities of oxygen
Diff_silicone_fl = 2.17e-5; % cm^2/s
Diff_polystyrene = 1; % cm^2/s
Diff_medium = 2.78e-5; % cm^2/s
Diff_tissue = 1.22e-5; % cm^2/s
% Solubilities of oxygen
Alpha_silicone_fl = 1.21e-8; % mol/cm^3 mmHg
Alpha_polystyrene = 3.5e-15; % mol/cm^3 mmHg
Alpha_medium = 1.27e-9; % mol/cm^3 mmHg
Alpha_tissue = 1.02e-9; % mol/cm^3 mmHg
Radius_well = 0.32; % cm radius of well
Cell_density = 25000; % cells/cm^2
h_medium = 0.1/(pi*Radius_well^2); % cm medium depth
P_ext = 142; % mmHg External pO2
Km = 0.44; % mmHg Km for oxygen consumption
P_critical = 0.1; % mmHg Critical pO2 of cells
Poly_thickness_bot = 0.12; % cm polystyrene thickness for bottom
Poly_thickness_side = 0.12; % cm polystyrene thickness for side
Dish_height = 1; % cm height of well
Tissue_thickness = 0.0012; % cm thickness of tissue layer
Silicone_fl_thickness = 0.0625; % cm thickness of silicone with fl
Reader_position = 0.8*0.32/Radius_well; % non-dimensional length that plate reader sees
OCR_cell = 2.8; % tmo/min cell
V_max = OCR_cell/60*Cell_density*10^-15/Tissue_thickness; %mol/cm^3 s OCR of cells
fem.const= {'Rwell',Radius_well,'h_medium','h_medium','Pext',P_ext,'Km',Km,
'D_medium',Diff_medium,'D_silicone_fl',Diff_silicone_fl,..
'D_poly',Diff_polystyrene,'D_tissue',Diff_tissue,'A_silicone_fl',
Alpha_silicone_fl,..
'A_poly',Alpha_polystyrene,'A_medium',Alpha_medium,'A_tissue',
Alpha_tissue,'Vmax','V_max,..
'Pcrit',P_critical,poly_bot,Poly_thickness_bot,poly_side',
Poly_thickness_side,..
'tissue_t',Tissue_thickness,'Si_fl_t',Silicone_fl_thickness,h_dish',
Dish_height,'Read_plate',Reader_position};
%%%%%% Geometry
Medium=rect2(1,h_medium/Radius_well+Tissue_thickness/Radius_well+Silicone_fl_thickness/Radius_well,'base',
'corner','pos',{'0','0'},'rot','0');
Bottom_dish=rect2(1+Poly_thickness_side/Radius_well,Poly_thickness_bot/Radius_well,'base','corner','pos',{'0',-Poly_thickness_bot/Radius_well},'rot','0');
Side_dish=rect2(Poly_thickness_side/Radius_well,Dish_height/Radius_well,'base','corner','pos',{'1','0'},'rot','0');
Si_fl_layer=rect2('1',Silicone_fl_thickness/Radius_well,'base','corner','pos',{'0','0'},'rot','0');
Tissue_layer=rect2('1',Tissue_thickness/Radius_well,'base','corner','pos',{'0',Silicone_fl_thickness/Radius_well},'rot','0');
Plate_reader=rect2(Reader_position,Silicone_fl_thickness/Radius_well,'base','corner','pos',{'0','0'},'rot','0');
G1 = Medium + Bottom_dish;
G2 = G1 + Side_dish;
G3 = G2 + Si_fl_layer;
G4 = G3 + Tissue_layer;
final = G4 + Plate_reader;
clear s
s.objs={final};
s.name={'CO1'};
s.tags={'final'};
fem.draw=struct('s',s);
fem.geom=geomsg(fem);
fem.sdim = {'r','z'};
fem.border = 1;
Mesh
Multiphysics
fem=multiphysics(fem);
% Initialize mesh
fem.mesh=meshinit(fem, ...
'hmaxfact',0.55, ...
'hcutoff',0.01, ...
'hgrad',1.4, ...
'hcurve',0.4, ...
'lnarrow',0.7);
mesh_test = size(fem.mesh.t,2);
while mesh_test<2500
fem.mesh=meshrefine(fem);
mesh_test = size(fem.mesh.t,2);
end
Application mode 1
clear appl
appl.mode.class = 'FIDiffusion';
appl.mode.type = 'axi';
appl.sdim = {'r','z1','z'};
appl.shape = {'shlag(2,"lm1")','shlag(2,"c")'};
appl.gporder = {40,4};
appl.assignsuffix = '_di';
clear prop
clear weakconstr
weakconstr.value = 'non-ideal';
weakconstr.dim = {'lm1'};
prop.weakconstr = weakconstr;
appl.prop = prop;
clear bnd
bnd.type = {'ax','C','cont'};
bnd.c0 = {0,1,0};
bnd.weakconstr = {0,1,0};
bnd.weshape = 1;
bnd.ind = [1,2,1,3,1,3,1,3,2,3,3,3,3,3,3,3,2,2,2];
appl.bnd = bnd;
clear equ
equ.gporder = 2;
equ.D = {'D_poly*A_poly/(D_medium*A_medium)',
'D_silicone.fl*A_silicone.fl/(D_medium*A_medium)', ...
'D_tissue*A_tissue/(D_medium*A_medium)'};
equ.int = 1;
equ.R = {0,0,'-Vmax*Rwell^2/(D_medium*A_medium*Pext)*c/(c+Km/Pext)*(c>Pcrit/Pext)',0};
equ.shape = 2;
equ.ind = [1,2,3,4,2,1];
appl.equ = equ;
fem.appl{1} = appl;
fem.sdim = {'y','z'};
fem.border = 1;
fem.outform = 'general';
% Multiphysics
fem=multiphysics(fem);
% Extend mesh
fem.xmesh=meshextend(fem);
% Solve problem
[fem.sol, fem.stop]=fermlin(fem, ...
'solcomp',{'c','lm1'}, ...
'outcomp',{'c','lm1'}, ...
'out',{'sol','stop'},...
'maxiter',100,...
'nitol', 1e-006);
% Save current fem structure for restart purposes
fem0=fem;
convergence=fem.stop;
% Integrate
Total_I=postint(fem,2*pi*r*(1/(1+.03125*c*Pext)), ...
'dl',[2]);
Total_PO2_fl=postint(fem,'2*pi*r*c', ...
'dl',[2]);
Total_PO2_cell=postint(fem,'2*pi*r*c', ...
'dl',[3]);
min_PO2 = postmin(fem,'c',solnum',1,DL,[3]);
min_quality_mesh = min(meshqual(fem.mesh));
element_num = size(fem.mesh.t,2);
Top_flux=postint(fem,'2*pi*lm1', ...
'dl',[9], ...
'edim',1);
Poly_flux=postint(fem,'2*pi*lm1', ...
'dl',[2 17 18 19 20], ...
'edim',1);
Total_flux=Poly_flux+Top_flux;
References
1. 2007. American Diabetes Association. http://www.diabetes.org/home.jsp.
2. Colton CK, Avgoustiniatos ES. 1991. Bioengineering in development of the hybrid artificial pancreas. Journal of Biomechanical Engineering-Transactions of the Asme 113(2):152-170.
3. Silva AI, de Matos AN, Brons IG, Mateus M. 2006. An overview on the development of a bio-artificial pancreas as a treatment of insulin-dependent diabetes mellitus. Med Res Rev 26(2).
4. Meier JJ, Bhushan A, Butler PC. 2006. The potential for stem cell therapy in diabetes. Pediatric Research 59(4, Part 2):65R-73R.
5. Madsen OD. 2005. Stem cells and diabetes treatment. Apmis 113(11-12):858-875.
6. Ren JQ, Jin P, Wang E, Liu E, Harlan DM, Li X, Stroncek DF. 2007. Pancreatic islet cell therapy for type I diabetes: understanding the effects of glucose stimulation on islets in order to produce better islets for transplantation. Journal of Translational Medicine 5.
7. Shapiro AMJ, Lakey JRT, Ryan EA, Korbett GS, Toth E, Warnock GL, Kneteman NM, Rajotte RV. 2000. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. New England Journal of Medicine 343(4):230-238.
8. Ryan EA, Paty BW, Senior PA, Bigam D, Alfadhli E, Kneteman NM, Lakey JRT, Shapir AMJ. 2005. Five-year follow-up after clinical islet transplantation. Diabetes 54(7):2060-2069.
9. Pyzdrowski KL, Kendall DM, Halter JB, Nakhleh RE, Sutherland DER, Robertson RP. 1992. Preserved insulin-secretion and insulin independence in recipients of islet autografts. New England Journal of Medicine 327(4):220-226.
10. Korc M. 1993. Normal function of the endocrine pancreas. In: Go V, Dimango E, Gardner J, Lebenthal E, Reber H, Scheele G, editors. The Pancreas. New York: Raven Press. p 751–758.
11. Volk B, Wellmann K. 1985. Quantitative studies of the islets of nondiabetic patients. In: Volk B, Wellmann E, editors. The Diabetic Pancreas. New York: Plenum Medical. p 117-125.
12. Ryan EA, Lakey JRT, Rajotte RV, Korbett GS, Kin T, Imes S, Rabinovitch A, Elliott JF, Bigam D, Kneteman NM and others. 2001. Clinical outcomes and insulin secretion after islet transplantation with the edmonton protocol. Diabetes 50(4):710-719.
13. Colton CK, Papas KK, Pisania A, Rappel MJ, Powers D, O'Neil JJ, Omer A, Weir GC, Bonner-weir S. 2007. Characterization of Islet Preparations. In: Halberstadt C, Emerich DF, editors. Cellular Transplantation: From Laboratory to Clinic. New York: Elsevier Inc.
14. Balamurugan AN, Bottino R, Giannoukakis N, Smetanka C. 2006. Prospective and challenges of islet transplantation for the therapy of autoimmune diabetes. Pancreas 32(3):231-43.
15. Ricordi C, Lacy PE, Finke EH, Olack BJ, Scharp DW. 1988. Automated-method for isolation of human pancreatic-islets. Diabetes 37(4):413-420.
16. Avgoustiniatos ES. 2001. Oxygen diffusion limitations in pancreatic islet culture and immunoisolation. Cambridge: MIT.
17. Hering BJ, Kandaswamy R, Harmon JV, Ansite J, Clemmings SM, Sakai T, Paraskevas S, Eckman PM, Sageshima J, Nakano M and others. 2004. Bluestone JA 2003 Transplantation of cultured islets from two-layer preserved pancreases in type 1 diabetes with anti-CD3 antibody. Am J Transplant 4:390-401.
18. Hering BJ. 2005. Achieving and maintaining insulin independence in human islet transplant recipients. Transplantation 79(10):1296-1298.
19. Kaufman DB, Lowe WLJ. 2003. Clinical islet transplantation. Curr Diab Rep 3(4):344-350.
20. Lakey JRT, Burridge PW, Shapiro AMJ. 2003. Technical aspects of islet preparation and transplantation. Transplant Int 16(9):613-632.
21. Ricordi C, Inverardi L, Kenyon NS, Goss J, Bertuzzi F, Alejandro R. 2005. Requirements for success in clinical islet transplantation. Transplantation 79(10):1298-1300.
22. Sutherland DER. 2003. Current status of beta-cell replacement therapy (pancreas and islet transplantation) for treatment of diabetes mellitus. Transplant Proc 35(5):1625-1627.
23. Knazek RA. 2002. The human pancreatic islet cell resource consortium. Diab Technol Ther 4(4):551-552.
24. Weber DJ, McFarland RD, Irony I. 2002. Selected Food an Drug Administration review issues for regulation of allogeneic islets of Langerhans as somatic cell therapy. Transplantation 74:1816-1820.
25. Eckhard E, Brandhorst D, Winter D, Jaeger C, Jahr H, Bretzel RG, Brendel MD. 2004. The role of current product release criteria for identification of human islet preparations suitable for clinical transplantation. Transplant Proc 36:1528-1531.
26. Brissova M, Poffenberger G, Jewel M, Dai CH, Powers AC. 2006. Human islet preparations have major variations in islet function within and between islet isolation facilities. Diabetes 55:A443-A444.
27. Shapiro AMJ, Nanji SA, Lakey JRT. 2003. Clinical islet transplant: current and future directions towards tolerance. Immunological Reviews 196(1):219-236.
28. Ricordi C, Scharp DW, Lacy PE. 1988. Reversal in nude mice after transplantation of fresh and 7 days culutred (24 °C) human pancreatic islets. Transplantation 45:994-996.
29. Gaber AO, Fraga DW, Kotb M, Lo A, Sabek O, Latif K. 2004. Human islet graft function in NOD-SCID mice predicts clinical response in islet transplant recipients. Transplant Proc 36:1108-1110.
30. Hering BJ, Kandaswamy R, Harmon JV, Ansite JD, Clemmings SM, Sakai T, S. P, M. EP, Sageshima J, Nakano M and others. 2004. Transplantation of cultured islets from two-layer preserved pancreases in type 1 diabetes with anti-CD3 antibody. Am J Transplant 4:390-401.
31. Bretzel RG, Alejandro R, Hering BJ, van Suylichem PT, Ricordi C. 1994. Clinical islet transplantation: guidelines for islet quality control. Transplant Proc. 26:388-392.
32. Latif ZA, Noel J, Alejandro R. 1988. A simple method of staining fresh and cultured islets. Transplantation 45(4):827-830.
33. London NJ, Contractor H, Lake SP, Aucott GC, Bell PR, James RF. 1989. A microfluorometric viability assay for isolated human and rat islets of Langerhans. Diabetes Res 12(3):141-149.
34. 2000. FDA Biological Response Modifier Advisory Committee summary minutes Meeting no. 26, March 20-21. http://www.fda.gov/ohrms/dockets/ac/00minutes/3604m1minutes.pdf.
35. 2003. FDA Biological Response Modifier Advisory Committee summary minutes Meeting no. 36, Oct. 9-10. http://www.fda.gov/ohrms/dockets/ac/cber03.html#BiologicalRepsonseModofiers.
36. Ricordi C, Lakey JRT, Hering BJ. 2001. Challenges toward standardization of islet isolation technology. Transplant Proc 33(1-2):1709.
37. Ichii H, Inverardi L, Pileggi A, Damaris Molano R, Cabrera O, Caicedo A, Mesinger S, Kuroda Y, Berggren PO, Ricordi C. 2005. A novel method for the assessment of cellular composition and beta-cell viability in human islet preparations. Am J Transplant 5:1635-1645.
38. Goto M, Holgersson J, Kumagai-Braesch M, Korsgren O. 2006. The ADP/ATP ratio: a novel predictive assay of quality assessment of isolated pancreatic islets. Am J Transplant 6:2483-2487.
39. Papas KK, Pisania A, Wu H, Weir GC, Colton CK. 2007. A stirred microchamber for oxygen consumption rate measurements with pancreatic islets. Submitted.
40. Murdoch TB, McGhee-Wilson D, Shapiro AMJ, Lakey JRT. 2004. Methods of human islet culture for transplantation. Cell Transplantation 13(6):605-617.
41. Kuttler B, Hartmann A, Wanka H. 2002. Long-term culture of islets abrogates cytokine-induced or lymphocyte-induced increase of antigen expression on beta cells. Transplantation 74(4):440-445.
42. Ricordi C, Lacy PE, Sterbenz K, Davie JM. 1987. Low-temperature culture of cuman islets or in vivo treatment with L3t4 antibody produces a marked prolongation of islet human-to-mouse xenograft survival. Proceedings of the National Academy of Sciences of the United States of America 84(22):8080-8084.
43. Stein E, Mullen Y, Benhamou PY, Watt PC, Hober C, Watanabe Y, Nomura Y, Brunicardi FC. 1994. Reduction in immunogenicity of human islets by 24-degrees-C culture. Transplantation Proceedings 26(2):755.
44. Lacy PE, Davie JM, Finke EH. 1979. Prolongation of islet allograft survival following in vitro culture (24-Degrees-C) and a single injection of ALS. Science 204(4390):312-313.
45. Lafferty KJ, Prowse SJ, Simeonovic CJ, Warren HS. 1983. Immunobiology of tissue-transplantation - a return to the passenger leukocyte concept. Annual Review of Immunology 1:143-173.
46. Kim SC, Han DJ, Kim IH, Woo KO, We YM, Kang SY, Back JH, Kim YH, Kim JH, Lim DG. 2005. Comparative study on biologic and immunologic characteristics of the pancreas islet cell between 24 degrees C and 37 degrees C culture in the rat. Transplantation Proceedings 37(8):3472-3475.
47. Matsumoto S, Goel S, Quallie S, Strong DM, Reems JA. 2003. A comparative evaluation of culture conditions for short-term maintenance (<24 h) of human islets isolated using the Edmonton protocol. Cell and Tissue Banking 4(2-4):85-93.
48. Clayton HA, London NJM. 1996. Survival and function of islets during culture. Cell Transplantation 5(1):1-12.
49. Brandhorst D, Brandhorst H, Hering BJ, Bretzel RG. 1999. Long-term survival, morphology and in vitro function of isolated pig islets under different culture conditions. Transplantation 67(12):1533-1541.
50. Degraaff MPA, Wolters GHJ, Vanschilfgaarde R. 1994. Endothelial-cells in pancreatic-islets and the effect of culture. Transplantation Proceedings 26(3):1171-1171.
51. Cui Y-F, Ma M, Wang G-Y, Han D-E, Vollmar B, Menger MD. 2005. Prevention of core cell damage in isolated islets of Langerhans by low temperature preconditioning. World J Gastroenterol 11(4).
52. Ono J, Lacy PE, Michael HEB, Greider MH. 1979. Studies of the functional and morphologic status of islets maintained at 24-C for 4 weeks in vitro. American Journal of Pathology 97(3):489-&.
53. Ilieva A, Yuan S, Wang R, Duguid W, Rosenberg L. 1999. The structural integrity of the islet in vitro: the effect of incubation temperature. Pancreas 19(3):297-303.
54. Lakey JRT, Warnock GL, Kneteman NM, Ao Z, Rajotte RV. 1994. Effects of cryopreservation culture on human islet recovery and in vitro function. Transplant Proc 26(2):820.
55. Pisania A. 2007. Development of quantitative assays for quality assessment of islets of Langerhans. Cambridge: Massachusetts Institute of Technology.
56. Lacy PE, Kostianovsky M. 1967. Method for the isolation of intact islets of Langerhans from the rat pancreas. Diabetes 16(1):35.
57. Gotoh M, Maki T, Kiyozumi T, Satomi S, Monaco AP. 1985. An improved method for isolation of mouse pancreatic islets. Transplantation 40:437-438.
58. O'Neil JJ, Omer A, Tchipashvili T, Lei J, Pisania A, Colton CK, Weir GC, Bonner-Weir S. 2007. Quantitative analysis of cell composition of human pancreatic islet preparations. Lab Invest submitted.
59. O'Neil JJ, Stegemann JP, Nicholson DT, Gagnon KA, Solomon BA, Mullon CJ. 2001. The isolation and function of porcine islets from market weight pigs. Cell Transplant 10(3):235-246.
60. Papas KK, Avgoustiniatos ES, Tempelman LA, Weir GC, Colton CK, Pisania A, Rappel MJ, Friberg AS, Bauer AC, Hering BJ. 2005. High-density culture of human islets on top of silicone rubber membranes. Transplantation Proceedings 37(8):3412-3414.
61. Papas KK, Colton CK, Nelson RA, Rosack PR, Avgoustiniatos ES, Scott II WE, Pisania A, Weir GC, Hering BJ. 2007. Human islet oxygen consumption rate and DNA measurements predict diabetes reversal in nude mice. Am J Transplant 7(3):707-713.
62. Bank HL. 1987. Assessment of islet cell viability using fluorescent dyes. Diabetologia 30(10):812-816.
63. Weibel ER. 1979. Practical methods for biological morphometry. London: Academic Press.
64. Bonner-Weir S, Trent DF, Weir GC. 1983. Partial pancreatectomy in the rat and subsequent defect in glucose-induced insulin release. Journal of Clinical Investigation 71(6):1544-1553.
65. Erlandsen SL, Parsons JA, Burke JP, Redick JA, Van Orden DE, Van Orden LS. 1975. A modification of the unlabeled antibody enzyme method using heterologous antisera for the light microscopy and ultrastructural localization of insulin, glucagon, and growth hormone. J Histochem Cytochem 23(9):666-677.
66. Polak JM, Noorden SV. 2003. Visualizing the end-product reaction. Introduction to immunocytochemistry. Oxford, UK: BIOS Scientific Publishers Ltd. p 50.
67. Pipeleers DG, Int' Veld PA, Van De Winkel M, Maes E, Schuit FC, Gepts W. 1985. A new in vitro model for the study of pancreatic A and B cells. Endocrinology 117:806-816.
68. Dolezel J, Bartos J, Voglmayr H, Greilhuber J. 2003. Nuclear DNA content and genome size of trout and human. Cytometry Part A 51A(2):127-128.
69. Hellerstrom C. 1967. Effects of carbohydrates on the oxygen consumption of isolated pancreatic islets of mice. Endocrinology 81(1):105-112.
70. Lundgren G, Andersson A, Borg H, Buschard K, Groth CG, Gunnarsson R, Hellerstrom C, Pettersson B, Ostman J. 1977. Structural and functional integrity of isolated human islets of Langerhans maintained in tissue culture for 1-3 weeks. Transplant Proc 9:237-240.
71. Sweet IR, Gilbert M, Jensen R, Sabek O, Fraga DW, Gaber AO, Reems JA. 2005. Glucose stimulation of cytochrome C reduction and oxygen consumption as assessment of human islet quality. Transplantation 80(8):1003-1011.
72. Andersson A, Gunnarsson R, Hellerstrom C. 1976. Long-Term Effects of a Low Extracellular Glucose Concentration on Glucose-Metabolism and Insulin-Biosynthesis and Release of Mouse Pancreatic-Islets Maintained in Tissue-Culture. Acta Endocrinologica 82(2):318-329.
73. Dionne KE. 1990. Effect of hypoxia on insulin secretion and viability of pancreatic islet tissue. Cambridge: MIT.
74. Sweet IR, Khalil G, Wallen AR, Steedman M, Schenkman KA, Reems JA, Kahn SE, Callis JB. 2002. Continuous measurement of oxygen consumption by pancreatic islets. Diabetes Technol Ther 4(5).
75. Wang WJ, Upshaw L, Strong DM, Robertson RP, Reems JA. 2005. Increased oxygen consumption rates in response to high glucose detected by a novel oxygen biosensor system in non-human primate and human islets. Journal of Endocrinology 185(3):445-455.
76. Schwartzman RA, Cidlowski JA. 1993. Apoptosis: the biochemistry and molecular biology of programmed cell death. Endocrine Rev 14(2):133-151.
77. Gershengorn MC, Hardikar AA, Wei CJ, Geras-Raaka E, Marcus-Samuels B, Raaka BM. 2004. Epithelial-to-mesenchymal transition generates proliferative human islet precursor cells. Science 306(5705):2261-2264.
78. Weir GC, Halban PA, Meda P, Wollheim CB, Orci L, Renold AE. 1984. Dispersed adult rat pancreatic islet cells in culture: A, B, and D cell function. Metabolism 33:447-453.
79. Ghosh A, Ronner P, Cheong E, Khalid P, Matchinsky FM. 1991. The role of ATP and free ADP in metabolic coupling during fuel stimulated insulin release from islet β-cells in the isolated perfused rat pancreas. J Biol Chem 286(34):22887-22892.
80. Fraker C, Guarino R, Timmins M, Presnell S, Molano D, Pileggi A, Ricordi C. The BD oxygen biosensor as a potential predictive tool for islet function in-vivo; 2004; Boston, MA.
81. Korsgren O, Nilsson B, Berne C, Felldin M, Foss A, Kallen R, Lundgren T, Salmela K, Tibell A, Tuftveson G. 2005. Current status of clinical islet transplantation. Transplantation 79(10):1289-1293.
82. Ryan EA, Lakey JRT, Paty BW, Imes S, Korbett GS, Kneteman NM, Bigam D, Rajotte RV, Shapiro AMJ. 2002. Successful islet transplantation: Continued insulin reserve provides long-term glycemic control. Diabetes 51(7):2148-2157.
83. Gaglia JL, Shapiro AMJ, Weir GC. 2005. Islet transplantation: Progress and challenge. Archives of Medical Research 36(3):273-280.
84. Morisot C, Candel S, Sauret V, Kerr-Conte J, Richard MJ, Favrot MC, Benhamou PY. 2007. MnTMPyP, a metalloporphyrin-based superoxide dismutase/catalase mimetic, protects INS-1 cells and human pancreatic islets from an in vitro oxidative challenge. Diabetes Metab 33(1).
85. Hara Y, Fujino M, Li XK. 2006. The reduction of hypoxia-induced and reoxygenation-induced apoptosis in rat islets by epigallocatechin gallate. Transplantation Proceedings 38(8):2722-2725.
86. Pileggi A, Molano RD, Berney T, Cattan P, Vizzardelli C, Oliver R, Fraker C, Ricordi C, Pastori RL, Bach FH and others. 2001. Heme oxygenase-1 induction in islet cells results in protection from apoptosis and improved in vivo function after transplantation. Diabetes 50(9):1983-1991.
87. Contreras JL, Bilbao G, Smyth CA, Eckhoff DE, Jiang XL, Jenkins S, Thomas FT, Curiel DT, Thomas JM. 2002. Cytoprotection of pancreatic islets before and early after transplantation using gene therapy. Kidney Int 61(1 Suppl).
88. Li YX, Ge L, Dong WP, Lu DR, Tan JM. 2006. Protection of human islets from induction of apoptosis and improved islet function with HO-1 gene transduction. Chinese Medical Journal 119(19):1639-1645.
89. Fenjves ES, Ochoa MS, Gay-Rabinstein C, Ricordi C, Curran MA. 2004. Retrovirally transferred genes inhibit apoptosis in an insulin-secreting cell line: Implications for islet transplantation. Cell Transplantation 13(5):489-496.
90. McCabe C, Samali A, O'Brien T. 2006. beta cell cytoprotective strategies: Establishing the relative roles for iNOS and ROS. Biochemical and Biophysical Research Communications 342(4):1240-1248.
91. Gaber AO, Fraga DW, Callicutt CS, Gerling IC, Sabek OM, Kotb MY. 2001. Improved in vivo pancreatic islet function after prolonged in vitro islet culture. Transplantation 72(11):1730-1736.
92. Dionne KE, Colton CK, Yarmush ML. 1993. Effect of hypoxia on insulin-secretion by isolated rat and canine islets of Langerhans. Diabetes 42(1):12-21.
93. Nyqvist D, Kohler M, Wahlstedt H, Berggren PO. 2005. Donor islet endothelial cells participate in formation of functional vessels within pancreatic islet grafts. Diabetes 54(8):2287-2293.
94. Parr EL, Bowen KM, Lafferty KJ. 1980. Cellular-changes in cultured mouse thyroid-glands and islets of Langerhans. Transplantation 30(2):135-141.
95. Mendola JF, Goity C, Fernandezalvarez J, Saenz A, Benarroch G, Fernandezcruz L, Gomis R. 1994. Immunocytochemical study of pancreatic-islet revascularization in islet isograft - effect of hyperglycemia of the recipient and of in-vitro culture of islets. Transplantation 57(5):725-730.
96. Olsson R, Carlsson P-O. 2006. Oxygenation of cultured pancreatic islets. Advances in Experimental Medicine and Biology. p 263-268.
97. Deen WM. 1998. Analysis of Transport Phenomena. New York: Oxford University Press, Inc.
98. Wilson DF, Rumsey WL, Green TJ, Vanderkooi JM. 1988. The oxygen dependence of mitochondrial oxidative-phosphorylation measured by a new optical method for measuring oxygen concentration. Journal of Biological Chemistry 263(6):2712-2718.
99. 2005. COMSOL Documentaion - Users Guide. COMSOL.
100. Bonner-Weir S, Taneja M, Weir GC, Tatarkiewicz K, Song KH, Sharma A, O'Neil JJ. 2000. In vitro cultivation of human islets from expanded ductal tissue. Proceedings of the National Academy of Sciences of the United States of America 97(14):7999-8004.
101. Gschwind M, Huber G. 1997. Detection of apoptotic or necrotic death in neuronal cells by morphological, biochemical, and molecular analysis. In: Poirier J, editor. Apoptosis Techniques and Protocols. Totowa, New Jersey: Humana Press.
102. Rudolf E, Cervinka M. 2006. Cytoskeletal changes in non-apoptotic cell death. Acta Medica (Hradec Kralove) 49(2).
103. Kaneto H, Fujii J, Seo HG, Suzuki K, Matsuoka T, Nakamura M, Tatsumi H, Yamasaki Y, Kamada T, Taniguchi N. 1995. Apoptotic Cell-Death Triggered by Nitric-Oxide in Pancreatic Beta-Cells. Diabetes 44(7):733-738.
104. Hoorens A, Stange G, Pavlovic D, Pilepeers D. 2001. Distinction between interleukin-1-induced necrosis and apoptosis of islet cells. Diabetes 50(3):551-557.
105. Dive C, Gregory CD, Phipps DJ, Evans DL, Milner AE, Wyllie AH. 1992. Analysis and discrimination of necrosis and apoptosis (programmed cell-death) by multiparameter flow-cytometry. Biochimica Et Biophysica Acta 1133(3):275-285.
106. Papas KK. 2007. personal communication noting islet surface density employed in clinical preparations.
107. Omer A. 2007. personal communication noting islet surface density employed in clinical preparations.
108. Freshney RI. 1983. Culture of animal cells a manual of basic technique. Freshney, R. I. Culture of Animal Cells: A Manual of Basic Technique. Xiv+295p. Alan R. Liss, Inc.: New York, N.Y., USA. Illus.
109. Bergert H, Knoch KP, Meisterfeld R, Jager M, Ouwendijk J, Kersting S, Saeger HD, Solimena M. 2005. Effect of oxygenated perfluorocarbons on isolated rat pancreatic islets in culture. Cell Transplantation 14(7):441-448.
110. Rutzky LP, Bilinski S, Kloc M, Phan T, Zhang HM, Katz SM, Stepkowski SM. 2002. Microgravity culture condition reduces immunogenicity and improves function of pancreatic islets. Transplantation 74(1):13-21.
111. Murray HE, Paget MB, Downing R. 2005. Preservation of glucose responsiveness in human. islets maintained in a rotational cell culture system. Molecular and Cellular Endocrinology 238(1-2):39-49.
112. Sardonini CA, Wu YJ. 1993. Expansion and differentiation of human hematopoietic-cells from static cultures through small-scale bioreactors. Biotechnology Progress 9(2):131-137.
113. Collins PC, Nielsen LK, Patel SD, Papoutsakis ET, Miller WM. 1998. Characterization of hematopoietic cell expansion, oxygen uptake, and glycolysis in a controlled, stirred-tank bioreactor system. Biotechnology Progress 14(3):466-472.
114. Collins PC, Miller WM, Papoutsakis ET. 1998. Stirred culture of peripheral and cord blood hematopoietic cells offers advantages over traditional static systems for clinically relevant applications. Biotechnology and Bioengineering 59(5):534-543.
115. Jorjani P, Ozturk SS. 1999. Effects of cell density and temperature on oxygen consumption rate for different mammalian cell lines. Biotechnology and Bioengineering 64(3):349-356.
116. Colton CK, Drake RF. 1971. Effect of boundary conditions on oxygen transport to blood flowing in a tube. Chemical Engineering Progress Symposium Series 67(114):88-95.
117. Perry RH, Green DW. 1997. Perry's Chemical Engineers' Handbook: McGraw-Hill.
118. 2007. CRC Handbook of Chemistry and Physics. Boca Ranton, FL: Taylor and Francis.
119. Sekine N, Cirulli V, Regazzi R, Brown LJ, Gine E, Tamaritrodriguez J, Girotti M, Marie S, Macdonald MJ, Wolheim CB and others. 1994. Low lactate-dehydrogenase and high mitochondrial glycerol phosphate dehydrogenase in pancreatic beta-cells - potential role in nutrient sensing. Journal of Biological Chemistry 269(7):4895-4902.
120. Hummerich H, Degroot H, Noll T, Soboll S. 1988. Dependence of mitochondrial and cytosolic adenine-nucleotides on oxygen partial-pressure in isolated hepatocytes - application of a new rapid high-pressure filtration technique for fractionation. Biochemical Journal 250(3):641-645.
121. Noll T, de Groot H, Wissemann P. 1986. A computer-supported oxystat system maintaining steady-state O2 partial pressures and simultaneously monitoring O2 uptake in biological systems. Biochem J 236(3).
122. Casey TM, Arthur PG. 2000. Hibernation in noncontracting mammalian cardiomyocytes. Circulation 102(25):3124-3129.
123. Venkatasubramanian R, Henson MA, Forbes NS. 2006. Incorporating energy metabolism into a growth model of multicellular tumor spheroids. Journal of Theoretical Biology 242(2):440-453.
124. Patton JN, Palmer AF. 2006. Numerical simulation of oxygen delivery to muscle tissue in the presence of hemoglobin-based oxygen carriers. Biotechnology Progress 22(4):1025-1049.
125. Dash RK, Bassingthwaighte JB. 2006. Simultaneous blood-tissue exchange of oxygen, carbon dioxide, bicarbonate, and hydrogen ion. Annals of Biomedical Engineering 34(7):1129-1148.
126. Littbrand B, Revesz L. 1969. The effect of oxygen on cellular survival and recovery after radiation. Br J Radiol 42(504).
127. Buerk DG, Goldstick TK. 1992. Spatial variation of aortic wall oxygen diffusion coefficient from transient polarographic measurements. Annals of Biomedical Engineering 20:629-646.
128. Bentley TB, Meng H, Pittman RN. 1993. Temperature-dependence of oxygen diffusion and consumption in mammalian striated-muscle. American Journal of Physiology 264(6):H1825-H1830.
129. Mahler M, Louy C, Homsher E, Peskoff A. 1985. Reappraisal of diffusion, solubility, and consumption of oxygen in frog skeletal-muscle, with applications to muscle energy-balance. Journal of General Physiology 86(1):105-134.
130. Weind KL, Ellis CG, Boughner DR. 2002. Aortic valve cusp vessel density: Relationship with tissue thickness. Journal of Thoracic and Cardiovascular Surgery 123(2):333-340.
131. Filion RJ, Ellis CG. 2003. A finite difference model of O2 transport in aortic valve cusps: importance of intrinsic microcirculation. American Journal of Physiology-Heart and Circulatory Physiology 285(5):H2099-H2104.
132. Groebe K. 1990. A versatile model of steady-state O2 supply to tissue - application to skeletal-muscle. Biophysical Journal 57(3):485-498.
133. Radisic M, Deen W, Langer R, Vunjak-Novakovic G. 2005. Mathematical model of oxygen distribution in engineered cardiac tissue with parallel channel array perfused with
culture medium containing oxygen carriers. American Journal of Physiology-Heart and Circulatory Physiology 288(3):H1278-H1289.
134. Vadapalli A, Goldman D, Popel AS. 2002. Calculations of oxygen transport by red blood cells and hemoglobin solutions in capillaries. Artificial Cells Blood Substitutes and Immobilization Biotechnology 30(3):157-188.
135. Evans NTS, Naylor PFD, Quinton TH. 1981. The diffusion-coefficient of oxygen in respiring kidney and tumor-tissue. Respiration Physiology 43(3):179-188.
136. Groebe K, Thews G. 1990. Calculated intracellular and extracellular pO2 gradients in heavily working red muscle. American Journal of Physiology 259(1):H84-H92.
137. Clark A, Clark PAA, Connett RJ, Gayeski TEJ, Honig CR. 1987. How large is the drop in pO2 between cytosol and mitochondrion. American Journal of Physiology 252(6):C583-C587.
138. Kawashiro T, Nusse W, Scheid P. 1975. Determination of diffusivity of oxygen and carbon-dioxide in respiring tissue - results in rat skeletal-muscle. Pflugers Archiv-European Journal of Physiology 359(3):231-251.
139. Han P, Bartels DM. 1996. Temperature dependence of oxygen diffusion in H2O and D2O. Journal of Physical Chemistry 100(13):5597-5602.
140. Wichterlova J, Wichterle K, Michalek J. 2005. Determination of permeability and diffusivity of oxygen in polymers by polarographic method with inert gas. Polymer 46(23):9974-9986.
141. Stern SA, Shah VM, Hardy BJ. 1987. Structure-permeability relationships in silicone polymers. Journal of Polymer Science Part B-Polymer Physics 25(6):1263-1298.
142. McDonagh C, Bowe P, Mongey K, MacCrath BD. 2002. Characterisation of porosity and sensor response times of sol-gel-derived thin films for oxygen sensor applications. Journal of Non-Crystalline Solids 306(2):138-148.
143. Lu X, Manners I, Winnik MA. 2001. Polymer/silica composite films as luminescent oxygen sensors. Macromolecules 34(6):1917-1927.
144. Mark JE. 1999. Polymer data handbook. New York: Oxford University Press.
145. 1999. Polymer Handbook. Bloch DR, editor: John Wiley & Sons.
146. Guarino RD, Dike LE, Haq TA, Rowley JA, Pitner JB, Timmins MR. 2004. Method for determining oxygen consumption rates of static cultures from microplate measurements of pericellular dissolved oxygen concentration. Biotechnology and Bioengineering 86(7):775-787.
147. Hoorens A, VandeCastelee M, Kloppel G, Pipeleers D. 1996. Glucose promotes survival of rat pancreatic beta cells by activating synthesis of proteins which suppress a constitutive apoptotic program. Journal of Clinical Investigation 98(7):1568-1574.
148. Ling Z, Hannaert JC, Pipeleers D. 1994. Effect of Nutrients, Hormones and Serum on Survival of Rat Islet Beta-Cells in Culture. Diabetologia 37(1):15-21.
149. Van de Casteele M, Kefas BA, Cai Y, Heimberg H, Scott DK, Henquin JC, Pipeleers D, Jonas JC. 2003. Prolonged culture in low glucose induces apoptosis of rat pancreatic beta-cells through induction of c-myc. Biochemical and Biophysical Research Communications 312(4):937-944.
150. Deeney JT, Prentki M, Corkey BE. 2000. Metabolic control of beta-cell function. Seminars in Cell & Developmental Biology 11(4):267-275.
151. Li CH, Najafi H, Daikhin Y, Nissim IB, Collins HW, Yudkoff M, Matschinsky FM, Stanley CA. 2003. Regulation of leucine-stimulated insulin secretion and glutamine metabolism in isolated rat islets. Journal of Biological Chemistry 278(5):2853-2858.
152. Casciari JJ, Sotirchos SV, Sutherland RM. 1988. Glucose Diffusivity in Multicellular Tumor Spheroids. Cancer Research 48(14):3905-3909.
153. Kauri LM, Jung SK, Kennedy RT. 2003. Direct measurement of glucose gradients and mass transport within islets of Langerhans. Biochemical and Biophysical Research Communications 304(2):371-377.
154. Tamarit-Rodriguez J, Idahl LA, Gine E, Alcazar O, Sehlin J. 1998. Lactate production in pancreatic islets. Diabetes 47(8):1219-1223.
155. Erecinska M, Bryla J, Michalik M, Meglsson MD, Nelson D. 1992. Energy-Metabolism in Islets of Langerhans. Biochimica Et Biophysica Acta 1101(3):273-295.
156. Zong-Chao L, Efendic S, Wibom R, Abdel-Halim SM, Ostenson C-G, Landau BR, Khan A. 1998. Glucose metabolism in Goto-Kakizaki rat islets. Endocrinology 139(6):2670-2675.
157. Borelli MI, Francini F, Gagliardino JJ. 2004. Autocrine regulation of glucose metabolism in pancreatic islets. American Journal of Physiology-Endocrinology and Metabolism 286(1):E111-E115.
158. Matschinsky FM. 2002. Regulation of pancreatic beta-cell glucokinase - From basics to therapeutics. Diabetes 51:S394-S404.
159. Haber EP, Ximenes HMA, Procopio J, Carvalho CRO, Curi R, Carpinelli AR. 2003. Pleiotropic effects of fatty acids on pancreatic beta-cells. Journal of Cellular Physiology 194(1):1-12.
160. Holmes MA, Clayton HA, Chadwick DR, Bell PRF, London NJM, James RFL. 1995. Functional-studies of rat, porcine, and human pancreatic-islets cultured in 10 commercially available media. Transplantation 60(8):854-860.
161. Lindstrom P, Sehlin J. 1986. Effect of intracellular alkalinization on pancreatic-islet calcium-uptake and insulin-secretion. Biochemical Journal 239(1):199-204.
162. Roos A, Boron WF. 1981. Intracellular pH. Physiological Reviews 61(2):296-434.
163. Ohta M, Nelson D, Nelson J, Meglsson MD, Erecinska M. 1991. Relationships between energy-level and insulin-secretion in isolated rat islets of Langerhans - a study at various pH values. Biochemical Pharmacology 42(3):593-598.
164. Simpson NE, Bennett LK, Papas KK, Sambanis A, Constantimidis I. 2000. Effects of pH on murine insulinoma beta TC3 cells. Biochemical and Biophysical Research Communications 273(3):937-941.
165. Fraker C, Montelongo J, Szust J, Khan A, Ricordi C. 2004. The use of multiparametric monitoring during islet cell isolation and culture: A potential tool for in-process corrections of critical physiological factors. Cell Transplantation 13(5):497-502.
166. Tai J, Tsang A, Tze WJ. 1994. Effects of serum and medium supplements, pH, and temperature on the viability of cultured porcine islets. Transplantation Proceedings 26(2):818-819.
167. Garg LC, Maren TH. 1972. The rates of hydration of carbon dioxide and dehydration of carbonic acid at 37 degrees. Biochim Biophys Acta 261(1).
168. Soukane DM, Shirazi-Adl A, Urban JPG. 2005. Analysis of nonlinear coupled diffusion of oxygen and lactic acid in intervertebral discs. Journal of Biomechanical Engineering-Transactions of the Asme 127(7):1121-1126.
169. Patel AA, Gawlinski ET, Lemieux SK, Gatenby RA. 2001. A cellular automaton model of early tumor growth and invasion: The effects of native tissue vascularity and increased anaerobic tumor metabolism. Journal of Theoretical Biology 213(3):315-331.
170. Sarangapani R, Wexler AS. 1996. Growth and neutralization of sulfate aerosols in human airways. Journal of Applied Physiology 81(1):480-490.
171. Fatt I, Giasson CJ, Mueller TD. 1998. Non-steady-state diffusion in a multilayered tissue initiated by manipulation of chemical activity at the boundaries. Biophysical Journal 74(1):475-486.
172. Geers C, Gros G. 2000. Carbon dioxide transport and carbonic anhydrase in blood and muscle. Physiological Reviews 80(2):681-715.
173. Shah VM, Hardy BJ, Stern SA. 1986. Solubility of carbon-dioxide, methane, and propane in silicone polymers - effect of polymer side-chains. Journal of Polymer Science Part B-Polymer Physics 24(9):2033-2047.
174. Goss JA, Schock AP, Brunicardi FC, Goodpastor SE, Garber AJ, Soltes G, Barth M, Froud T, Alejandro R, Ricordi C. 2002. Achievement of insulin independence in three consecutive type-1 diabetic patients via pancreatic islet transplantation using islets isolated at a remote islet isolation center. Transplantation 74(12):1761-1766.
175. 2007. TCP Reliable, http://www.tcpreliable.com.
176. 2007. PCM Energy P. Ltd.
177. Guyton AC, Hall J. 2005. Textbook of Medical Physiology: Saunders.
178. Aburaya M, Tanaka KI, Hoshino T, Tsutsumi S, Suzuki K, Makise M, Akagi R, Mizushima T. 2006. Heme oxygenase-1 protects gastric mucosal cells against non-steroidal anti-inflammatory drugs. Journal of Biological Chemistry 281(44):33422-33432.
179. Kruger AL, Peterson SJ, Schwartzman M, Fusco H, McClung JA, Weiss M, Shnouda S, Goodman AI, Goligorsky MS, Kappas A and others. 2006. Upregulation of heme oxygenase provides vascular protection in an animal model of diabetes through its antioxidant and anti-apoptotic effects. Circulation 114(18):106-106.
180. Min KS, Hwang YH, Ju HJ, Chang HS, Kang KH, Pi SH, Lee SK, Kim EC. 2006. Heme oxygenase-1 mediates cytoprotection against nitric oxide-induced cytotoxicity via the cGMP pathway in human pulp cells. Oral Surgery Oral Medicine Oral Pathology Oral Radiology and Endodontics 102(6):803-808.
181. Otterbein LE, Kolls JK, Mantell LL, Cook JL, Alam J, Choi AMK. 1999. Exogenous administration of heme oxygenase-1 by gene transfer provides protection against hyperoxia-induced lung injury. Journal of Clinical Investigation 103(7):1047-1054.
182. Otterbein LE, Soares MP, Yamashita K, Bach FH. 2003. Heme oxygenase-1: unleashing the protective properties of heme. Trends in Immunology 24(8):449-455.
183. Petrache I, Otterbein LE, Alam J, Wiegand GW, Choi AMK. 2000. Heme oxygenase-1 inhibits TNF-alpha-induced apoptosis in cultured fibroblasts. American Journal of Physiology-Lung Cellular and Molecular Physiology 278(2):L312-L319.
184. Tsuchihashi SJ, Livhits M, Zhai Y, Busuttil RW, Araujo JA, Kupiec-Weglinski JW. 2006. Basal rather than induced heme oxygenase-1 levels are crucial in the antioxidant cytoprotection. Journal of Immunology 177(7):4749-4757.
185. Lee PJ, Alam J, Wiegand GW, Choi AMK. 1996. Overexpression of heme oxygenase-1 in human pulmonary epithelial cells results in cell growth arrest and increased resistance
to hyperoxia. *Proceedings of the National Academy of Sciences of the United States of America* 93(19):10393-10398.
186. Brouard S, Otterbein LE, Anrather J, Tobiasch E, Bach FH, Choi AMK, Soares MP. 2000. Carbon monoxide generated by heme oxygenase 1 suppresses endothelial cell apoptosis. *Journal of Experimental Medicine* 192(7):1015-1025.
187. Chen K, Gunter K, Maines MD. 2000. Neurons overexpressing heme oxygenase-1 resist oxidative stress-mediated cell death. *Journal of Neurochemistry* 75(1):304-313.
188. Shiraishi F, Curtis LM, Truong L, Poss K, Visner GA, Madsen K, Nick HS, Agarwal A. 2000. Heme oxygenase-1 gene ablation or expression modulates cisplatin-induced renal tubular apoptosis. *American Journal of Physiology-Renal Physiology* 278(S5):F726-F736.
189. Suttner DM, Sridhar K, Lee CS, Tomura T, Hansen TN, Dennery PA. 1999. Protective effects of transient HO-1 overexpression on susceptibility to oxygen toxicity in lung cells. *American Journal of Physiology-Lung Cellular and Molecular Physiology* 276(3):L443-L451.
190. Abraham NG, Lavrovsky Y, Schwartzman ML, Stoltz RA, Levere RD, Gerritsen ME, Shibahara S, Kappas A. 1995. Transfection of the human heme oxygenase gene into rabbit coronary microvessel endothelial-cells-protective effect against heme and hemoglobin toxicity. *Proceedings of the National Academy of Sciences of the United States of America* 92(15):6798-6802.
191. Wang X, Wang Y, Kim HP, Nakahira K, Ryter SW, Choi AMK. 2007. Carbon monoxide protects against hyperoxia-induced endothelial cell apoptosis by inhibiting reactive oxygen species formation. *Journal of Biological Chemistry* 282(3):1718-1726.
192. Otterbein LE, Mantell LL, Choi AMK. 1999. Carbon monoxide provides protection against hyperoxic lung injury in rats. *American Journal of Respiratory and Critical Care Medicine* 159(3):A218-A218.
193. Ryter SW, Otterbein LE, Morse D, Choi AMK. 2002. Heme oxygenase/carbon monoxide signaling pathways: Regulation and functional significance. *Molecular and Cellular Biochemistry* 234(1):249-263.
194. Gunther L, Berberat PO, Haga M, Brouard S, Smith RN, Soares MP, Bach FH, Tobiasch E. 2002. Carbon monoxide protects pancreatic beta-cells from apoptosis and improves islet function/survival after transplantation. *Diabetes* 51(4):994-999.
195. Kim HS, Loughran PA, Kim PK, Billiar TR, Zuckerbraun BS. 2006. Carbon monoxide protects hepatocytes from TNF-alpha/Actinomycin D by inhibition of the caspase-8-mediated apoptotic pathway. *Biochemical and Biophysical Research Communications* 344(4):1172-1178.
196. Chae HJ, Chin HY, Lee GY, Park HR, Yang SK, Chung HT, Pae HO, Kim HM, Chae SW, Kim HR. 2006. Carbon monoxide and nitric oxide protect against tumor necrosis factor-alpha-induced apoptosis in osteoblasts: HO-1 is necessary to mediate the protection. *Clinica Chimica Acta* 365(1-2):270-278.
197. Bainbridge SA, Belkacemi L, Dickinson M, Graham CH, Smith GN. 2006. Carbon monoxide inhibits hypoxia/reoxygenation-induced apoptosis and secondary necrosis in syncytiotrophoblast. *American Journal of Pathology* 169(3):774-783.
198. Suttner DM, Dennery PA. 1999. Reversal of HO-1 related cytoprotection with increased expression is due to reactive iron. *Faseb Journal* 13(13):1800-1809.
199. D'Amico G, Lam F, Hagen T, Moncada S. 2006. Inhibition of cellular respiration by endogenously produced carbon monoxide. *Journal of Cell Science* 119(11):2291-2298.
200. Keynes RG, Griffiths C, Garthwaite J. 2003. Superoxide-dependent consumption of nitric oxide in biological media may confound in vitro experiments. Biochemical Journal 369:399-406.
201. Papas KK, Colton CK, Nelson RA, Rozak PR, Avgoustiniatos ES, Scott WE, Wildey GM, Pisania A, Weir GC, Hering BJ. 2007. Human islet oxygen consumption rate and DNA measurements predict diabetes reversal in nude mice. American Journal of Transplantation 7(3):707-713.
202. Dike LE, Xia HY, Guarino RD, Presnell SC, Timmins MR. 2005. Rapid method for assessing oxygen consumption rate of cells from transient-state measurements of pericellular dissolved oxygen concentration. Cytotechnology 49(2-3):133-141.
203. Hodge K, Prodran T, Shenogina NB, Nazarenko S. 2001. Diffusion of oxygen and carbon dioxide in thermally crystallized syndiotactic polystyrene. Journal of Polymer Science Part B-Polymer Physics 39(20):2519-2538.
204. Bowyer WJ, Xu WY, Demas JN. 2004. Determining oxygen diffusion coefficients in polymer films by lifetimes of luminescent complexes measured in the frequency domain. Analytical Chemistry 76(15):4374-4378.
205. Gao YP, Ogilby PR. 1992. A New Technique to Quantify Oxygen Diffusion in Polymer-Films. Macromolecules 25(19):4962-4966.
206. Kneas KA, Demas JN, Nguyen B, Lockhart A, Xu WY, DeGraff BA. 2002. Method for measuring oxygen diffusion coefficients of polymer films by luminescence quenching. Analytical Chemistry 74(5):1111-+.
207. Schappacher G, Hartmann P. 2003. Partial analytical solution of a model used for measuring oxygen diffusion coefficients of polymer films by luminescence quenching. Analytical Chemistry 75(16):4319-4324.
208. Boersma A, Cangialosi D, Picken SJ. 2003. Mobility and solubility of antioxidants and oxygen in glassy polymers - II. Influence of physical ageing on antioxidant and oxygen mobility. Polymer Degradation and Stability 79(3):427-438.
209. Bae SY, Cho DH, Kim HT, Kumazawa H. 1994. Transport of oxygen and carbon-dioxide through polycarbonate membrane. Korean Journal of Chemical Engineering 11(2):127-130.
210. Papkovsky DB. 2004. Methods in optical oxygen sensing: Protocols and critical analyses. Oxygen Sensing. p 715-735.
211. Arain S, Weiss S, Heinzle E, John GT, Krause C, Klimant I. 2005. Gas sensing in microplates with optodes: Influence of oxygen exchange between sample, air, and plate material. Biotechnology and Bioengineering 90(3):271-280.
212. Chapman JD, Sturrock J, Boag JW, Crookall JO. 1970. Factors affecting the oxygen tension around cells growing in plastic Petri dishes. Int J Radiat Biol Relat Stud Phys Chem Med 17(4). |
Spirituality and Psychological Adaptation Among Women With HIV/AIDS: Implications for Counseling
Jane M. Simoni
University of Washington
Maria G. Martone and Joseph F. Kerwin
Yeshiva University
Survey interviews with 230 predominantly African American and Puerto Rican low-income women who were living with HIV/AIDS in New York City revealed high levels of spirituality and spiritually based coping with HIV. Both spirituality indicators positively correlated with the frequency of receipt of HIV-related social support; they were negatively related to recent drug use. Two hierarchical regression analyses that controlled relevant covariates indicated positive correlations between the spirituality indicators and psychological adaptation (i.e., a composite measure of depressive symptomatology, mood states, mastery, and self-esteem). The beneficial effect of spiritually based coping persisted even when other types of coping were controlled. The discussion considers implications for counseling women living with HIV as well as for research and theory in the area of spiritually based coping.
HIV challenges an individual physically, socially, and psychologically. Furthermore, it may threaten one’s sense of meaning, purpose, and significance in life. Research on other life-threatening illnesses has indicated that individuals often turn to religion and spirituality to cope (Dein & Stygall, 1997; Demi, Moneyham, Sowell, & Cohen, 1997; Dunbar, Mueller, Medina, & Wolf, 1998; Jenkins & Pargament, 1995; Pargament, 1997). Specifically, Sodestrom and Martinson (1987) reported that 88% of hospitalized cancer patients reported using a variety of spiritual activities and resources while coping with their illness. Reed (1987) found terminally ill adults who were hospitalized indicated a greater spiritual perspective than both nonterminally ill hospitalized and healthy adults. Among 103 female breast cancer patients, religion was an extremely important source of support (Johnson & Spilka, 1991).
Many studies involving coping among cancer patients have demonstrated the positive impact of spirituality on patients’ psychosocial adaptation (Jenkins & Pargament, 1995; Pargament, 1997). Acklin, Brown, and Mauger (1983) studied patients with and without cancer and found that transcendent meaning, religiosity, and church attendance were negatively related to despair, anger–hostility, and social isolation. Among patients with advanced cancer, religious beliefs and activities were associated with less pain and greater happiness (Yates, Chalmer, St. James, Folklansbee, & McKeegney, 1981). Mullen, Smith, and Hill (1993) found that spiritual resources were positively correlated with a sense of coherence and a sense of life as comprehensible, manageable, and meaningful, which in turn was negatively correlated with psychological distress. In additional studies of cancer patients, spiritual well-being correlated with lower levels of state-trait anxiety (Kaczorowski, 1989), and spiritual awareness related to lower levels of psychosocial distress (Smith et al., 1993).
Other research has indicated a nonsignificant or negative effect of religious coping on adaptation, although it is not clear how differing operationalizations of the construct affected the findings (Jenkins & Pargament, 1995). Indeed, the research on religious coping has defined and operationalized spirituality in various ways (Burke & Miller, 1996; Maher & Hunt, 1993; Muldoon & King, 1995; Reed, 1987). Allport and Ross (1967) originally conceptualized religious orientation as extrinsic (using religion as a means to self-serving ends) or intrinsic (reflecting a true commitment to religious ideals as an end in themselves). Much of the coping literature, when it considers religion at all, treats it as a dispositional variable or assesses it with a few indicators such as belief in God, frequency of prayer, and church attendance or membership (Pargament et al., 1990). On the other hand, some definitions of spirituality are broad enough to cover existential beliefs, such as the search for meaning and purpose in life, and include even the occult (Connor, Wicker, & Germino, 1990; Muldoon & King, 1995). For example, Vastyan (1986) stated that spirituality involves transcendence, which means “to be aware of the numinous, of that which is beyond ordinary experience, beyond comprehension, beyond our material existence” (p. 112). Many researchers now distinguish between religiosity (i.e., organized systems of belief and practice) and spirituality (i.e., efforts to consider metaphysical or transcendent aspects of everyday life; Jenkins & Pargament, 1995; Zimbaieru et al., 1997), although the utility of this distinction has been contested (Pargament, 1999).
Individuals battling life-threatening illnesses use religious coping in complex and variable ways, making it difficult to identify
the mechanisms by which it operates. One possibility is that spirituality provides a sense of meaning in the face of threat to one’s existence (Frankl, 1959; Mullen, Smith, & Hill, 1993; Pargament & Hahn, 1986). This has been found for both cancer patients (Connor et al., 1990) and those diagnosed with HIV/AIDS (Jenkins, 1995; Schwartzberg, 1993). Research on coping has highlighted the role of religion in social cognition, including attributions of the nature and controllability of an aversive event and its reframing or reinterpretation in a more positive light (Pargament et al., 1990). Sevensky (1981) suggested that prayer allows expression of the anger, disappointment, and fear that often accompany the illness experience; engenders a closer relationship with God; and provides the opportunity for contemplation and meditation. Additionally, it assists one in repairing damaged relationships, letting go of the past, achieving a sense of closure while also providing hope of an ultimate victory, despite death, by joining God (Sevensky, 1981). Pargament et al. (1990) used factor analysis to place 31 different strategies within six categories of religious coping: spiritually based coping, good deeds, religious support, discontent, pleading, and religious avoidance. They found the best predictors of outcome to be belief in a just, loving God; the experience of God as a supportive partner; involvement in religious rituals; and the search for spiritual and personal support through religion. Hall’s (1998) qualitative analyses of the spiritual responses of 10 individuals with advanced HIV suggested spirituality can influence interpretation of body and mind states. Three themes were revealed: a purpose in life emerges from stigmatization, an opportunity for meaning arises from an incurable disease, and spirituality frames life after suffering.
Despite the widespread use of religious coping and its demonstrated association with enhanced outcome, social scientists often ignore spirituality and religion as significant coping tools (Johnson & Spilka, 1991). A number of factors may account for this neglect, including that both researchers and clinicians tend to be less religious than the general population (Gallup & Proctor, 1982; Lukoff, Lu, & Turner, 1992; Turbott, 1996). Historically, Freud (1928/1961) viewed religion as neurotic wish fulfillment, and Ellis (1970) initially argued it can be reduced to irrational supernaturalism. In general, religion seems to be taboo in research and health care settings (Weaver et al., 1998). Perhaps some clinicians fear imposing a system of beliefs or values onto their patients or view spirituality as detrimental to emotional functioning and believe that it fosters passivity, dependency, and denial (Jenkins & Pargament, 1995). Finally, the construct of spirituality is diffuse and not well understood; therefore, it is difficult to operationalize and measure empirically. Researchers may not want to be associated with the less empirical or “softer” science this suggests.
The failure to study spiritual aspects of coping is particularly glaring in AIDS research, especially in comparison to the wealth of studies among cancer patients. Intuitively, the life-threatening nature of both diseases would suggest that the findings would generalize. The limited research on spirituality among individuals with HIV is predominantly qualitative and based on samples of gay White men. Comparisons are limited, owing to the various ways spirituality is defined and measured. Nevertheless, studies indicate that HIV is associated with greater religiosity and spirituality among HIV-positive (HIV+) and HIV-negative (HIV−) partners of men with AIDS (Folkman, Chesney, Cooke, Boccellari, & Collette, 1994) and that spiritual well-being may abet an HIV+ man’s ability to maintain hope in the midst of HIV and AIDS-related illness (Carson, Soeken, Shantry, & Terry, 1990). A spiritual belief system can serve to maintain a positive attitude, thereby lessening the fear of death in end-stage HIV among men and women (Hall, 1994). Among 422 HIV+ military personnel, Jenkins (1995) found religious coping associated with less distress and more adaptive social functioning.
The use of spiritually based coping has been found to be more prevalent among women (Spilka, Hood, & Gorsuch, 1985) and among African Americans (Jackson, Neighbors, & Gurin, 1986; Potts, 1996) than other individuals. Therefore, examining spirituality would appear particularly important in the study of HIV-infected women, who are predominantly African American and Hispanic (e.g., Biggar et al., 1999; Kaplan, Marks, & Mertens, 1997). Jenkins (1995) found that African Americans with HIV, compared with Whites with HIV, indicated a greater preference for a coping style involving collaboration with God. In addition, studies have consistently indicated a greater use of spiritually based coping activities among samples of women than samples of gay men (Demi, et al., 1997; Jenkins, 1995; Remien, Rabin, Williams, & Katoff, 1992; Schwartzberg, 1993). In fact, among White gay men, spirituality has not been identified as a widely used coping strategy (Remien et al., 1992; Schwartzberg, 1993). Research among gay men with AIDS has shown that formal religiosity (i.e., greater church attendance and belonging to the religion of childhood) versus a more intrinsic religious orientation appears to be associated with greater death anxiety (Bivens, Niemeyer, Kirchberg, & Moore, 1995; Franks, Templar, Cappelletty, & Kaufman, 1990), perhaps due to the virulently heterosexist doctrine of many traditional religions.
In the current project, we conceptualized spirituality’s role in the process of coping and psychological adaptation in accordance with Lazarus and Folkman’s (1984) classic stress, appraisal, and coping theory. According to their model, stress is defined as a “relationship between the person and the environment that is appraised by the person as taxing or exceeding his or her resources and endangering his or her well-being” (p. 21). The significance of the stressor is determined through cognitive appraisals influenced by personal beliefs and values, such as personal control and existential or spiritual beliefs. When challenged, individuals engage in various ways of coping to manage the stressor based on available resources, with effective coping dependent on the match between coping options and actual coping demands.
Drawing upon this model, Pargament (1997) stated that, under duress, people will be more likely to cope spiritually when religion is the more available and compelling way of managing the stressor. Spirituality may be an especially available and compelling resource in cases in which the usual human coping resources are ineffective or are threatened, as in the face of potentially fatal diseases like HIV/AIDS. From the perspective of Lazarus and Folkman’s (1984) theory that beliefs influence cognitive appraisals critical to the coping process, we see that spirituality may help individuals to appraise negative events differently and may provide a greater sense of control, thereby aiding in psychological adjustment. However, the relationship between spirituality and psychological adaptation is complex. Pargament (1997) stated that spirituality may help individuals conserve meaning and transform their sense of significance through integration of the stressor into existing definitions of self.
Spirituality also provides practical, supportive resources that assist with psychological adaptation such as ritual, prayer, an outlet to express affect, and community. However, although there is overlap between the concept of spiritual support and social support—long shown empirically to benefit psychological adaptation—the support provided by spiritual beliefs is distinctive because it describes one’s relationship with a higher power and the unique rituals and community that surround those beliefs.
In the present study, we aimed to extend current findings regarding the apparently beneficial effects of spiritual coping to women with HIV/AIDS. We examined spirituality and its correlates as well as frequency of spiritually based coping, which we expected to be high in our sample of predominantly African American and Puerto Rican low-income women. We hypothesized that spirituality and spiritually based coping would relate to better psychological outcomes even after we controlled potentially confounding variables such as demographics, drug use, and social support. Furthermore, we examined the ability of spiritually based coping to explain variance in psychological adaptation beyond the effects of other types of coping.
Method
Participants
The sample of 230 women living with HIV comprised Hispanics (47%), non-Hispanic African Americans (46%), and women of other or mixed ethnic backgrounds (7%). Eighty-nine percent of the Hispanic women were born in the United States, and 84% self-identified as Puerto Rican. Their mean acculturation level was 3.42 (SD = 1.04) according to the Short Acculturation Scale (Marín & VanOostrum Marín, 1991), which consists of four items assessing language use from 1 (only Spanish) to 5 (only English). The majority, ranging in age from 24 years to 50 years (M = 38.5), indicated they were single and never legally married (56%), legally married (14%), separated (14%), divorced (6%), and widowed (10%). Seventy-one percent classified themselves as exclusively heterosexual. Mean education was 11.0 years (SD = 2.6), with 44% of the women reporting a high school degree or general equivalency diploma (GED). Only 11% were employed full- or part-time. Total household monthly income ranged from less than $500 (1%) to greater than $1,500 (8%); 91% reported less than $1,000. Nineteen percent reported crack cocaine or injection drug (ID) use in the past 30 days.
According to self-report, 24% of the women had AIDS, 36% had minimal symptoms, and 39% were asymptomatic. Time since HIV diagnosis ranged from 3 months to almost 15 years (Mdn = 4.3 years). Self-reported modes of infection were sex with an ID user (53%), sex with a non-ID user (30%), respondents’ own ID use (41%), and blood transfusion (8%; respondents could check more than one).
Measures
We assessed basic demographics, drug use, and social support in addition to coping, spirituality, and psychological adaptation. Unless otherwise indicated, scale items were rated on a Likert-type format ranging from 1 (strongly disagree) to 5 (strongly agree), and individual items were averaged to form a total score for each scale. We acknowledge that the original psychometric data do not apply to modified scales and that most measures were originally validated on samples in which physically ill, poor, inner city, minority women were underrepresented. However, estimates of internal consistency, provided below for this sample (Cronbach’s alphas), were all within acceptable limits. Kurtosis scores were between −.70 and .07, and only the spirituality measure was more than 2.5 times greater than its kurtosis standard deviation, suggesting these scores were somewhat negatively skewed. This was not surprising, given the high levels of spirituality we expected in this population.
Drug use. As part of a more comprehensive measure, we assessed recent drug use with two yes–no items asking if the respondent had used crack or injected any drug in the past 30 days. We also assessed empirically with a modified version of the University of California, Los Angeles, Social Support Inventory (UCLA-SSI; Schwarzer, Dunkel-Schetter, & Kemeny, 1994), which assesses support related to the participant’s stress or worry about his or her HIV disease in the past 30 days. The measure draws distinctions among emotional, informational, and tangible HIV-related support from four targets: partners, friends, relatives, and groups or organizations. Frequency of receipt of each type of support from each target is rated separately from 1 (never) to 5 (very often). For the 16 items composing the overall mean score, α = .85.
Coping. We used Reed, Kemeny, Taylor, Wang, and Visscher’s (1994) adaptation of the Ways of Coping Scale (Folkman, Lazarus, Dunkel-Schetter, DeLongis, & Gruen, 1986), which was designed to assess coping responses to stressful life events. Respondents were presented a list of 33 coping behaviors and asked, “(a) In the past 30 days, how much have you used each of these different ways to deal with the effects of HIV/AIDS on your health and the life-threatening nature of this illness?” Possible responses ranged from 1 (never) to 4 (a great deal). Reed et al. found six factors, which they labeled community involvement and spiritual growth, active cognitive coping, avoidance and self-blame, social support seeking, realistic acceptance, and information seeking. In the present study, principal-components factor analysis with varimax rotation led to a final solution with five factors comprising 27 items that accounted for 56% of the variance. Each item loaded on its respective factor with a coefficient of at least .34 and had no coefficient for any other factor that was higher. The five factors were community involvement (α = .84 for five items); e.g., “I participated in organizations geared toward helping those with HIV and AIDS”; constructive cognitions (α = .80 for eight items; e.g., “I tried to look on the bright side of things”); avoidance (α = .69 for five items; e.g., “I refused to believe that this problem has happened”); realistic acceptance (α = .76 for five items; e.g., “I tried to accept what might happen”; and spiritually based coping (α = .71 for the following four items: “I prayed or involved myself in other spiritual activities,” “I found new faith,” “I meditated or used relaxation or visualization to help with this problem” [these three items loaded on Reed et al.’s first factor], and “I rediscovered what is important in life” [this item loaded on Reed et al.’s second factor]).
Spirituality. A scale based on Smalai et al.’s (1996) scale was used to assess women’s spirituality based on their belief in a higher natural (not necessarily the dogma of traditional religion). Comprising 13 attitudinal items (α = .91), the scale tapped both dispositions (e.g., “My spirituality provides me with a sense of hope”) and behaviors (e.g., “I pray often”). The terms religion and religiosity were avoided on the basis of pilot testing that suggested the target population preferred the term spirituality. We also assessed denominational affiliation, church membership (with a yes–no item asking if participants were “a member of a church or other place of worship”), and frequency of attendance in the past 90 days at religious services (not at all, 1–3 times a month, 1–2 times a week, or 3 or more times a week).
Depression. Respondents also completed the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977), a nondiagnostic screening measure for examining the prevalence of nonspecific psychological distress in community samples, for which factor analytic studies have identified four underlying dimensions: negative affect, positive affect, somatic and retarded activity, and interpersonal concerns (Radloff, 1977). The scale’s 20 items assessing depressive symptomatology in the previous week (e.g., “I was bothered by things that usually don’t bother me”) are rated from 0 (rarely or none of the time [less than one day in the past week]) to 3 (most or all of the time [5–7 days in the past week]). Total CES-D sum scores range from 0 to 60, with scores of 16 or above defining “caseness” or possible clinical depression (Weissman, Sholomakas, PotMany studies have demonstrated the measure’s validity and high internal consistency and test-retest reliability (Guarnaccia, Angel, & Worobey, 1989; Roberts, 1980). In the current sample, $\alpha = .92$.
**Profile of Mood States.** Respondents also completed the short form of the Profile of Mood States (POMS–SF; Curran, Andrykowski, & Studts, 1995; McNair, Lorr, & Droppelman, 1981; Shacham, 1983), a commonly used measure of psychological distress. The POMS–SF constitutes a less taxing alternative to the POMS for use with physically ill populations; it accurately estimates original POMS scores. The POMS–SF had demonstrated acceptable internal consistency, with Cronbach’s alpha ranging from .75 in healthy samples to .90 in breast cancer patients. Respondents indicate how they have been feeling during the past week by rating each of 37 adjectives (e.g., active, annoyed, cheerful) according to a 5-point Likert scale ranging from *not at all* to *extremely*. Standard scoring of the POMS–SF yields six subscales (i.e., Fatigue–Inertia, Confusion–Activity, Tension–Anxiety, Depression–Dejection, Anger–Hostility, and Confusion–Bewilderment) in addition to the Total Mood Disturbance score (POMS–TMD) that was computed in this study as the mean of all items after reverse coding the Vigor–Activity items ($\alpha = .96$).
**Mastery.** The seven-item mastery scale of Pearlman and Schooler (1978) assesses a person’s sense of self-governing in relation to his or her external world. Sample items include, “I can do just about anything I really set my mind to do” and “I have little control over the things that happen to me” ($\alpha = .72$).
**Self-esteem.** We also included the 10 items of Rosenberg’s (1965) Self-Esteem Scale ($\alpha = .80$; e.g., “I take a positive attitude toward myself”). Silber and Tippett (1965) reported a 2-week test-retest correlation of .85 for the scale, which they found to correlate from .56 to .83 with several similar measures and clinical assessment. Rosenberg provided further evidence of the scale’s construct validity by demonstrating its correlation with several related mental health measures.
**Procedure**
As part of a larger investigation (Project STAR: Stress, Strength, and Resilience Among Seropositive Women), we trained a diverse group of 12 women, including health care consumers from the seropositive women’s community, to administer highly structured face-to-face interviews (see Simoni, Weinberg, & Nero, 1999). In 1996, interviewers recruited participants by word of mouth and posted notices at several HIV/AIDS outpatient clinics and at community-based AIDS organizations serving low-income women in the New York metropolitan area. They interviewed participants in the clinic, community-based AIDS organizations, or scatter site housing projects and gave each $10, a list of free or low-cost referrals, and an AIDS information booklet for completing the 50-min personal interview. Each participant was assured of the confidentiality of her responses and provided written informed consent. Eligibility criteria stipulated that the participants had to have been diagnosed with HIV at least 3 months prior to participation, be at least 18 years of age, possess the ability to give informed consent, and be English or Spanish speaking. The questionnaire was written originally in English, translated into Spanish, and independently back-translated to identify and clarify ambiguity of meaning. Respondents were given a choice of completing the interview in Spanish or English.
**Results**
**Spirituality**
Responses to the 13-item spirituality measure indicated that respondents rated themselves highly on this dimension ($M = 3.28$, $SD = .50$; out of a possible 4.00). Participants considered themselves to be Catholic (38%), Baptist (26%), Pentecostal (6%), Jewish (1%), or of some “other” religious affiliation (22%). Thirty-five percent were members of a church or other place of worship, and typical attendance at religious services in the past 90 days ranged from *not at all* (50%), to *1–3 times a month* (34%), *1–2 times a week* (14%), and *3 or more times a week* (3%). Not surprisingly, the spirituality indicators were significantly intercorrelated. It is interesting to note, however, that church membership and attendance were more strongly associated with spiritually based coping with HIV ($r = .21, p = .001$, for church membership; $r = .31, p = .000$, for church attendance), than with spirituality in general ($r = .15, p = .025$, for membership; $r = .19, p = .004$, for attendance).
| Main variable | $M$ | $SD$ | Minimum | Maximum | Age | Education | Ethnicity | Drug use | Social support |
|--------------------------------|-------|------|---------|---------|--------|-----------|-----------|----------|----------------|
| **Spirituality indicators** | | | | | | | | | |
| Church membership | 0.35 | 0.48 | 0.00 | 1.00 | −.03 | −.02 | −.13* | −.12 | .03 |
| Frequency of church attendance| 0.69 | 0.80 | 0.00 | 3.00 | −.05 | −.01 | .01 | −.15* | .09 |
| Spirituality | 3.28 | 0.49 | 1.00 | 4.00 | −.10 | −.10 | .03 | −.16* | −.34*** |
| Spiritually based coping | 1.81 | 0.76 | 0.00 | 3.00 | −.10 | .03 | .02 | −.16* | −.44*** |
| **Coping strategies** | | | | | | | | | |
| Community involvement | 1.62 | 0.83 | 0.00 | 3.80 | −.14* | .06 | .02 | −.21** | .60*** |
| Constructive cognitions | 2.11 | 0.58 | 0.50 | 3.00 | −.08 | −.05 | −.05 | −.13* | .33*** |
| Realistic acceptance | 1.88 | 0.67 | 0.40 | 3.00 | −.16* | −.14* | .16* | −.10 | .32*** |
| Avoidance | 0.88 | 0.65 | 0.00 | 3.00 | −.15* | −.21*** | .05 | −.10 | −.13 |
| **Psychological adaptation** | | | | | | | | | |
| CES–D | 2.87 | 0.62 | 1.45 | 4.00 | .11 | .15* | −.15* | −.15* | .15* |
| POMS–TMD | 3.58 | 0.84 | 1.11 | 4.94 | .09 | .12 | −.13* | −.14* | .07 |
| Mastery | 2.88 | 0.51 | 1.71 | 4.00 | −.03 | .26*** | −.11 | −.07 | .19** |
| Self-esteem | 3.02 | 0.49 | 1.60 | 4.00 | −.07 | .24*** | −.07 | −.18** | .20*** |
| Combined (unstandardized) | 3.09 | 0.52 | 1.52 | 4.19 | .04 | .21*** | −.14* | −.16* | .17** |
*Note.* $N = 230$. For drug use (i.e., having used crack cocaine or injected any drug in the past 30 days), 0 = no and 1 = yes. For ethnicity, 1 = African American and 0 = other. CES–D = Center for Epidemiological Studies Depression Scale; POMS–TMD = Profile of Mood States Total Mood Disturbance score.
* $p < .05$. ** $p < .01$. *** $p < .005$.
Bivariate analyses of the demographic correlates of the spirituality indicators (see Table 1) indicated African American women were less likely than other women to be church members, frequency of church attendance was negatively related to recent drug use, and spirituality and spiritually based coping were both related to less drug use and more social support.
**Coping and Psychological Adaptation**
Descriptive information regarding coping strategies and psychological adaptation and their associations with potential confounders (i.e., demographics, drug use, and social support) are presented in Table 1.
The frequency with which the various coping strategies were used varied somewhat, with the lowest scores existing for avoidant coping. Generally, more frequent use of the coping strategies was related to younger age, less education, less drug use, and more social support.
Converted into sums with individual items scored from 0 to 3, CES-D scores averaged 22.23 ($SD = 12.43$); 65% of the scores were 16 or above, the traditional indicator of possible clinical depression. This percentage is higher than the rates of 15–19% that Radloff (1977) observed among community-based samples, approximating the rate she reported for psychiatric inpatients (70%). Even after removing the five items of the somatic subscale, 52% of the sample scored 16 or above.
Considering the different response formats and direction of the scoring for the indicators of psychological adaptation, means and standard deviations were comparable. All four indicators significantly intercorrelated (with $r$ ranging from .48 to .84). Given this overlap, we chose to calculate a composite indicator of psychological well-being. To calculate this variable, we first reverse coded the CES-D and POMS–TMD so that, as do the other two indicators, higher scores would indicate greater psychological adaptation. Next, we standardized all four means. Finally, we took a mean of these recoded, standardized means. In general, psychological adaptation related to greater education, non-African American ethnicity, less drug use, and greater social support.
**Multivariate Analyses of Spirituality and Psychological Adaptation**
Bivariate analyses revealed that spirituality and HIV-related spiritually based coping were significantly correlated with psychological adaptation ($r = .24, p = .000$, for spirituality; $r = .27, p = .000$, for spiritually based coping). However, associations of church membership and attendance were nonsignificant. Although we conceptually differentiated between spirituality and HIV-related spiritually based coping, the two measures were moderately intercorrelated ($r = .59, p = .000$). Therefore, we examined their association with psychological adaptation in separate hierarchical regression models. In each, we entered as covariates any variable that was significantly associated with the composite psychological adaptation indicator in bivariate analyses. Specifically, in the first regression, we entered education, ethnicity, and drug use in the first step, social support in the second step, and spirituality in the third step. As seen in Table 2, each step resulted in significant changes in $R^2$. In the final model, spirituality as well as drug use, education, and ethnicity remained significant independent predictors of psychological adaptation.
In the second hierarchical regression of spiritually based coping, we entered education, ethnicity, and drug use in the first step, social support in the second step, the other four coping strategies in the third step, and, finally, spiritually based coping. As seen in Table 3, each step resulted in significant changes in $R^2$. In the final model, spiritually based coping as well as education and three of the four coping styles (i.e., avoidance, realistic acceptance, and constructive cognitions) remained significant independent predictors of psychological adaptation.
Finally, we wanted to consider the additive effects of spirituality and spiritually resilient coping. This was not possible in the regressions, where, given their significant association, one variable would have largely canceled out the effect of the other. Therefore, we created dichotomous variables for both spirituality and spiritually resilient coping using median splits, and we divided respondents into three groups: (a) those who scored in the lower half of the distribution for either variable, (b) those who scored in the lower half for one variable and the upper half for the other, and (c) those who scored in the upper half for both variables. We then ran an analysis of covariance (ANCOVA) comparing the three groups on the psychological adaptation indicator. As in the regressions, all relevant variables were included as covariates. As seen in Figure 1, the main effect for group was significant, $F(4, 213) = 5.60, p < .001$, with unadjusted psychological adaptation scores for the groups increasing in the expected direction.
**Discussion**
Survey data from 230 low-income HIV+ women in New York City indicated high levels of spirituality and spiritually based coping, both of which were positively related to psychological adaptation. Specifically, hierarchical regression analyses revealed that after controlling for demographics, social support, and drug use, spirituality significantly related to psychological adaptation (i.e., our composite indicator of standardized CES–D, self-esteem, mastery, and POMS–TMD scores). In a separate regression, spiritually based coping also predicted psychological adaptation. This effect persisted even after we conTable 3
Spiritually Based Coping and Psychological Adaptation
| Regression step | Step 1 | Step 2 | Step 3 | Step 4 |
|-----------------|--------|--------|--------|--------|
| Step 1 | | | | |
| Education | .21*** | .20*** | .13** | .11* |
| Drug use | −.21***| −.19***| −.09 | −.08 |
| Ethnicity | −.15* | −.16* | −.07 | −.05 |
| Step 2 | | | | |
| Social support | 14* | −.01 | −.02 | |
| Step 3 | | | | |
| Community involvement | .01 | −.01 | | |
| Constructive cognitions | .50*** | .45*** | | |
| Realistic acceptance | −.16* | −.43** | | |
| Avoidance | −.41***| −.42***| | |
| Step 4 | | | | |
| Spiritually based coping | .17* | | | |
| $F$ | 8.25***| 7.67***| 20.89***| 19.79***|
| $\Delta R^2$ | .11*** | .02* | .31*** | .02* |
Note. $N = 230$. Statistics are standardized betas. For drug use (i.e., having used crack cocaine or injected any drug in the past 30 days), 0 = no and 1 = yes. For ethnicity, 1 = African American and 0 = other.
*p < .05, **p < .01, ***p < .005.
trolled for other types of coping strategies, suggesting spiritually based coping produces an independent, additive contribution to psychological adaptation. A similarly controlled ANCOVA demonstrated that spirituality and spiritually based coping are particularly potent when combined (the highest scores on psychological adaptation were among participants with high scores on both spirituality and spiritually based coping).
Of course, the correlational design of our study precludes making assumptions about the direction of the effect. Although spirituality and spiritually based coping may lead to better well-being, we cannot rule out the competing interpretation that women who feel better psychologically are more optimistic and appreciative regarding spirituality in their lives. Prior research, however, lends some credibility to our assumption that spirituality and spiritually based coping precede enhanced well-being. Specifically, Pargament et al. (1990) reported that individuals who used religious coping when confronted with a stressor were likely to have relied on their spirituality for support previously in their lives, rendering the hypothesis that adaptation leads to spirituality less likely. Additionally, longitudinal studies have demonstrated that early spiritual support predicted adjustment to college (Maton, 1989) and the stress of transplant surgery (Tix & Frazier, 1998).
If we assume that spirituality and spiritually based coping did exert some beneficial effect on psychological adaptation in our sample, what are the mechanisms underlying such an effect? Social support has been suggested as one potential mediator (e.g., Strawbridge, Shema, Cohen, & Kaplan, 2001). This would appear particularly likely for women who reported church membership and frequent attendance at religious services, where they may have had the opportunity to expand their social networks. However, in our sample, church membership and attendance were not associated with greater social support (or psychological adaptation for that matter). The bivariate associations between social support and the spirituality indicators as well as social support and psychological adaptation suggest that an enhanced support network may account for some of the effect of the spirituality indicators on greater psychological adaptation. However, in multivariate analyses controlling for social support indexes, the effects of the spirituality indicators on psychological adaptation were only slightly attenuated. The effect of spirituality and spiritually based coping on psychological adaptation, then, is at most only partly attributable to the effects of enhanced social support (as reported as well in Tix & Frazier, 1998). The relationships appear somewhat complex, but it challenges the notion that greater receipt of social support is the essence of the ameliorating effect of the spirituality indicators.
In line with our theoretical conceptualization based on Lazarus and Folkman’s (1984) theory that beliefs influence cognitive appraisals, respondents’ spirituality may have helped them reappraise their situation and gain a greater sense of control, thereby aiding in psychological adaptation. To more explicitly identify these mechanisms in the future, researchers should include in their studies more specific assessments of appraisals as well as types of religious coping. For example, Pargament (1997) has differentiated among three types: collaborative (in which people work with God), self-directing (without God), and deferring (wait for God to solve their problems). He has stressed how the use of religion in coping has interactive, behavioral, emotional, and motivational components, each of which may have its own effects. Future investigations, therefore, might include more precise measures of spiritually based strategies or, minimally, may validate the preliminary spiritually based coping subscale we used with a larger sample.
Additional areas for future investigation include the identification of mediators (i.e., How do spirituality and spirituality based coping work to enhance psychological well-being?), moderators (i.e., For which groups of people or types of problems are they most efficacious?), and the areas of greatest impact (e.g., Is spirituality more likely to promote positive outcomes than to prevent deleterious ones?). Tix and Frazier (1998), for example, found that religious coping was more efficacious for Protestants than Catholics and that it could not be explained solely by the mediators of social support, perceived control, and cognitive restructuring. Maton (1989) demonstrated that spiritual support was related to well-being for high life-stress participants but not low life-stress participants. Pargament (1997) has done the most comprehensive
Figure 1. Spirituality and spiritually based coping as predictors of psychological adaptation.
research on situational factors, background variables, resources, specific coping activities, and self-management strategies and how they affect adaptation and well-being.
The findings of the current study are subject to some methodological limitations. Specifically, the sample was nonrandom, the data were self-reported, and there was no HIV—comparison group. Furthermore, we are not able to tell if the findings would generalize to groups of demographically different women or to men, two groups that should be studied in future research. Finally, as mentioned previously, the study’s cross-sectional design precludes causal explanations of the findings or assumptions about the direction of the effects.
Despite these limitations, the results have important implications for women living with HIV with respect to applications in counseling interventions (Hathaway & Pargament, 1992). Foremost, the findings regarding the high level of depressive symptomatology suggest counselors should pay particular attention to suicidality in this group (see Simoni, Nero, & Weinberg, 1998). Additionally, spirituality and spiritually based coping should be explored as part of a strategy of identifying and bolstering cultural strengths. They appear to be important psychologically and are readily accessible. This area invites collaboration between religious professionals and mental health providers, most of whom have little or no training in how to address religious concerns within a counseling context (Kelly, 1994; Shafranske & Malony, 1990). In fact, as many as one third of patients in some studies indicated unmet spiritual needs (Jenkins & Pargament, 1995), suggesting providers may be overlooking this aspect of patient care. Hall (1994) has emphasized the importance of providing structures that build and support hope to balance the pain, suffering, and fear of a life with HIV.
Religious and spiritual components have been included in some interventions (Gutterman, 1990), but these have yet to be systematically evaluated. If and how to incorporate spiritual coping in interventions to enhance psychological adaptation will depend on whether one can encourage spiritual coping or whether it has to be in place before the onset of the life-threatening event (Jenkins, 1995). Pargament (1997) has found that religious involvement in coping depends on the availability of religion to the individual. However, most people report themselves to be religious and churchgoing (Gallup & Castelli, 1989; Hoge, 1996), and among women and African Americans—two groups at increasing risk for HIV—religiosity is high, suggesting that most HIV patients may be capable of benefiting from spiritually based coping strategies.
Many counselors might question the appropriateness of discussing spiritual issues in the context of a counseling situation or may hesitate to pursue an area in which their own personal beliefs and values may be vastly different from their clients. Others rightly appreciate the potential negative influence of religion on emotional problems, such as when it leads to inappropriate or excessive guilt and anxiety or reinforces the oppression of women and sexual minorities (Maton & Wells, 1995). However, the high levels of spirituality and spiritually based coping observed in women living with HIV and their empirically demonstrated associations with enhanced psychological well-being suggest spiritual functioning is an area of vital importance to many clients and, therefore, an area any culturally competent counselor should be prepared to address.
References
Acklin, M. W., Brown, E. C., & Mauger, P. A. (1983). The role of religious values in coping with cancer. *Journal of Religion and Health*, 22, 322–333.
Allport, G. W., & Ross, J. M. (1967). Personal religious orientation and prejudice. *Journal of Personality and Social Psychology*, 5, 432–443.
Biggar, H., Forehand, R., Devine, D., Brody, G., Armistead, L., Morse, E., & Simon, P. (1999). Women who are HIV infected: The role of religious activity in psychosocial adjustment. *AIDS Care*, 11(2), 195–199.
Bivens, A. J., Neimeyer, R. A., Kirchberg, T. M., & Moore, M. K. (1995). Death concern and religious belief among gays and bisexuals of variable proximity to AIDS. *Omega*, 30, 105–120.
Boake, M. T., & Miller, J. (1996). Using the spiritual perspective in counseling persons with HIV/AIDS: An integrative approach. *Counseling and Values*, 40, 185–195.
Carson, V., Soeken, K. L., Shantz, J., & Terry, L. (1990). Hope and spiritual well-being: Essentials for living with AIDS. *Perspectives in Psychiatric Care*, 26, 28–43.
Connor, A. P., Wicker, C. A., & Germino, B. B. (1990). Understanding the cancer patient’s search for meaning. *Cancer Nursing*, 13, 167–175.
Curran, S. L., Andrykowski, M. A., & Studis, J. L. (1995). Short form of the Profile of Mood States (POMS-SF): Psychometric information. *Psychological Assessment*, 7, 80–83.
Dein, S., & Stiggall, J. (1997). Does being religious help or hinder coping with chronic illness? A critical literature review. *Palliative Medicine*, 11, 291–298.
Demu, A., Moneyham, L., Sowell, R., & Cohen, L. (1997). Coping strategies used by HIV infected women. *Journal of Death and Dying*, 35, 173–177.
Dunbar, H. T., Mueller, C. W., Medina, C., & Wolf, T. (1998). Psychological and spiritual growth in women living with HIV. *Social Work*, 43, 144–154.
Ellis, A. E. (1970). The case against religion. *Menon Journal*, 132, 5–6.
Folkman, S., Chesney, M. A., Cooke, M., Boccellari, A., & Collette, L. (1994). Caregiver burden in HIV-positive and HIV-negative partners of men with AIDS. *Journal of Consulting and Clinical Psychology*, 62, 746–756.
Folkman, S., Lazarus, R. S., Dunkel-Schetter, C., DeLongis, A., & Gruen, R. J. (1986). Dynamics of a successful encounter: Cognitive appraisal, coping, and encounter outcomes. *Journal of Personality and Social Psychology*, 50, 992–1003.
Frankl, V. E. (1959). *Man’s search for meaning: An introduction to logotherapy*. Boston: Beacon Press.
Frankl, K., Tompkins, D. I., Capellety, G. G., & Kauffman, I. (1990). Expectations of death and dying as a function of religious variables in gay men with and without AIDS. *Omega*, 22, 43–50.
Freud, S. (1961). *The future of an illusion* [J. Strachey, Trans.). New York: Norton. (Original work published 1928)
Gallup, G., Jr., & Castelli, J. (1989). *The people’s religion: American faith in the ‘90’s*. New York: Macmillan.
Gallup, G., Jr., & Proctor, W. (1982). *Adventures in immortality*. New York: McGraw-Hill.
Guarnaccia, P. J., Angel, R., & Worobey, J. L. (1989). The factor structure of the CES-D in the Hispanic health and nutrition examination survey: The influences of ethnicity, gender and language. *Social Science and Medicine*, 28, 1019–1026.
Gutterman, L. (1990). A day treatment program for persons with AIDS. *The American Journal of Occupational Therapy*, 44, 234–237.
Hall, B. A. (1994). Ways of maintaining hope in HIV disease. *Research in Nursing & Health*, 17, 283–293.
Hall, B. A. (1998). Patterns of spirituality in persons with HIV disease. *Nursing and Health*, 21, 143–153.
Hathaway, W. L., & Pargament, K. I. (1992). The religious dimensions of coping: Implications for prevention and promotion. In K. I. Pargament,
K. I. Maton, & R. E. Hess (Eds.), *Religion and prevention in mental health: Research, vision, and action* (pp. 129–154). Binghamton, NY: Haworth Press.
Hoge, D. R. (1996). Religion in America: The demographics of belief and affiliation. In E. P. Shafroth (Ed.), *Religion and the clinical practice of psychology* (pp. 21–41). Washington DC: American Psychological Association.
Jackson, J. S., Neighbors, H. W., & Gurin, G. (1986). Findings from a national survey of Black mental health: Implications for practice and training. In M. R. Mirand & H. H. Kitano (Eds.), *Mental health research in minority communities* (pp. 91–116). Rockville, MD: National Institute of Mental Health.
Jenkins, R. A. (1995). Religion and HIV: Implications for research and intervention. *Journal of Social Issues, 51*, 131–144.
Jenkins, R. A., & Pargament, K. I. (1995). Religion and spirituality as resources for coping with cancer. In B. Curbow & M. R. Somerfield (Eds.), *Psychosocial resource variables in cancer studies: Conceptual and measurement issues* (pp. 51–74). Binghamton, NY: Haworth Press.
Johnson, S. C., & Spilka, B. (1991). Coping with breast cancer: The roles of clergy and faith. *Journal of Religion and Health, 30*, 21–33.
Kaczorowski, J. M. (1989). Spiritual well-being and anxiety in adults diagnosed with cancer. *The Hospice Journal, 5*, 105–115.
Kaplan, M. S., Marks, G., & Mertens, S. B. (1997). Distress in coping among women with HIV infection: Preliminary findings from a multi-ethnic sample. *American Journal of Orthopsychiatry, 6*, 80–91.
Kelly, E. W. (1994). The role of religion and spirituality in counselor education: A national survey. *Counselor Education and Supervision, 33*, 227–237.
Lazarus, R. S., & Folkman, S. (1984). *Stress, appraisal, and coping*. New York: Springer.
Lukoff, D., Lu, F., & Turner, R. (1992). Towards a more culturally sensitive DSM-IV: Psychoreligious and psychospiritual problems. *Journal of Nervous and Mental Disease, 180*, 673–682.
Maher, M. H., & Hunt, T. K. (1993). Spirituality reconsidered. *Counseling and Values, 37*, 21–27.
Marín, G., & VanOss Marín, B. (1991). *Applied social research methods series: Vol. 23. Research with Hispanic populations*. Newbury Park, CA: Sage.
Maton, K. I. (1989). The stress-buffering role of spiritual support: Cross sectional and prospective investigations. *Journal for the Scientific Study of Religion, 28*, 310–323.
Maton, K. I., & Wells, E. A. (1995). Religion as a community resource for well-being: Prevention, healing and empowerment pathways. *Journal of Social Issues, 51*, 177–193.
McNair, P. M., Lorr, M., & Dropleman, L. F. (1981). *POMS manual* (2nd ed.). San Diego, CA: Educational and Industrial Testing Service.
Muldoon, M., & King, N. (1995). Spirituality, health care, and bioethics. *Journal of Religion and Health, 34*, 329–349.
Mullen, P. M., Smith, B. M., & Hill, E. W. (1993). Sense of coherence as a mediating variable for cancer patients and spouses. *Journal of Psychosocial Oncology, 11*, 23–46.
Pargament, K. I. (1997). *Psychology of religion and coping: Theory, research, practice*. New York: Guilford Press.
Pargament, K. I. (1999). The psychology of religion and spirituality? Yes and no. *International Journal for the Psychology of Religion, 9*, 3–16.
Pargament, K. I., Ensing, D. S., Falgout, K., Olsen, H., Reilly, B., Van Haitsma, K., & Warren, R. (1990). God help me: I. Religious coping efforts as predictors of the outcomes to significant negative life events. *American Journal of Community Psychology, 18*, 793–824.
Pargament, K. I., & Hahn, J. (1986). God and the just world: Causal and coping attributions to God in health situations. *Journal for the Scientific Study of Religion, 25*, 193–207.
Pearlin, L. I., & Schooler, C. (1978). The structure of coping. *Journal of Health and Social Behavior, 19*, 2–21.
Potts, R. G. (1996). Spirituality and the experience of cancer in an African-American community: Implications for psychosocial oncology. *Journal of Psycho-Oncology, 14*, 1–17.
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. *Applied Psychological Measurement, 1*, 385–401.
Reed, G. M., Kemeny, M. E., Taylor, S. E., Wang, H. J., & Visscher, B. R. (1994). Realistic acceptance as a predictor of decreased survival time in gay men with AIDS. *Health Psychology, 13*, 299–307.
Reed, P. (1987). Spirituality and well-being in terminally ill hospitalized adults. *Research in Nursing and Health, 10*, 335–344.
Remien, R. H., Rabkin, J. G., Williams, J. B., & Katof, L. (1992). Coping strategies and health beliefs of AIDS long-term survivors. *Psychology and Health, 6*, 335–349.
Roberts, R. E. (1991). Reliability of the CES-D Scale in different ethnic contexts. *Multivariate Research, 2*, 125–134.
Rosenberg, M. (1965). *Society and the adolescent self-image*. Princeton, NJ: Princeton University Press.
Schwartzberg, S. S. (1993). Struggling for meaning: How HIV-positive gay men make sense of AIDS. *Professional Psychology: Research and Practice, 24*, 483–490.
Schwarzer, R., Dunkel-Schetter, C., & Kemeny, M. (1994). The multidimensional nature of received social support in gay men at risk of HIV infection and AIDS. *American Journal of Community Psychology, 22*, 319–339.
Sevensky, R. L. (1981). Religion and illness: An outline of their relationship. *Southern Medical Journal, 74*, 743–746.
Shacham, S. (1983). A Hebrew version of the Profile of Mood States. *Journal of Personality Assessment, 47*, 305–306.
Shafraanske, E. P., & Malony, H. N. (1990). Clinical psychologists, religious and spiritual orientation, and their practice of psychotherapy. *Psychotherapy, 27*, 72–78.
Silber, E., & Tippett, J. (1965). Self-esteem: Clinical assessment and measurement validation. *Psychological Reports, 16*, 1071–1071.
Simoni, J. M., Nero, D. K., & Weinberg, B. A. (1998). Suicide attempts among seropositive women in New York City [Letter to the editor]. *American Journal of Psychiatry, 155*, 1631–1632.
Simoni, J. M., Weinberg, B. A., & Nero, D. K. (1999). Training community members to conduct HIV/AIDS research: Notes from a study of seropositive women. *AIDS Education and Prevention, 11*, 87–88.
Smith, E. D., Stefanek, M. E., Joseph, M. V., Verdieck, M. J., Zabors, J. R., & Fetting, J. H. (1993). Spiritual awareness, personal perspective on death, and psychosocial distress among cancer patients: An initial investigation. *Journal of Psychosocial Oncology, 11*, 89–103.
Sodestrom, K. E., & Martison, I. M. (1987). Patients’ spiritual coping strategies: A study of nurse and patient perspectives. *Oncology Nursing Forum, 14*, 41–46.
Somlai, A. M., Kelly, J. A., Kalichman, S. C., Mulry, G., Sikkema, K. J., Mulholland, K., & Davantes, B. (1996). An empirical investigation of the relationship between spirituality, coping, and emotional distress in people living with HIV infection and AIDS. *Journal of Pastoral Counseling, 17*, 171–182.
Spilka, B., Hood, R. W., & Gorsuch, R. L. (1987). *The psychology of religion: An empirical approach*. Englewood Cliffs, NJ: Prentice Hall.
Strawbridge, W. J., Shema, S. J., Cohen, R. D., & Kaplan, G. A. (2001). Religious attendance increases survival by improving and maintaining good health behaviors, mental health, and social relationships. *Annals of Behavioral Medicine, 23*, 68–74.
Tix, A. P., & Frazier, P. A. (1998). The use of religious coping during stressful life events: Main effects, moderation, and mediation. *Journal of Consulting and Clinical Psychology, 66*, 411–422.
Turbott, J. (1996). Religion, spirituality and psychiatry: Conceptual, cultural and personal challenges. *Australian and New Zealand Journal of Psychiatry*, 30, 720–727.
Vastyan, E. A. (1986). Spiritual aspects of the care of cancer patients. *CA-A Cancer Journal for Clinicians*, 36, 119–114.
Weaver, J., Kline, A. E., Samford, J. A., Lucas, L. A., Larson, D. B., & Gorsuch, R. L. (1998). Is religion taboo in psychology? A systematic analysis of research on religion in seven major American Psychological Association journals: 1991–1994. *Journal of Psychology and Christianity*, 17, 220–232.
Weissman, M. M., Sholomakas, D., Pottenger, M., Prusoff, B. A., & Locke, B. Z. (1992). Assessing depressive symptoms in five psychiatric populations: A validation study. *American Journal of Epidemiology*, 106, 203–214.
Yates, J. W., Chalmer, B. J., St. James, P., Follansbee, M., & McKegney, F. P. (1981). Religion in patients with advanced cancer. *Medical and Pediatric Oncology*, 9, 121–128.
Zinnbauer, B. J., Pargament, K. I., Cole, B., Rye, M. S., Belavich, T. G., Hipp, K. M., Scott, A. B., & Kadar, J. L. (1997). Religion and spirituality: Unfuzzying the fuzzy. *Journal for the Scientific Study of Religion*, 36, 549–564.
Received November 1, 1999
Revision received May 7, 2001
Accepted May 9, 2001 ■ |
THE GREAT UNRAVELING: A PORTRAIT OF THE MIDDLE CLASS
THE AMERICAN DREAM used to mean that if you put in a hard day’s work, you could expect good wages, benefits, and a better life for your kids. But the kinds of jobs that can provide a solid middle-class life in return for hard work are in short supply—unemployment remains high, earnings are volatile, and hard-won benefits are being lost. For the first time, the majority of Americans believe their children will not be better off than them. The future of the middle class, which has been the backbone of our nation’s economy for more than half a century, is at risk.
America’s strong and vibrant middle class didn’t just happen. It was built brick by brick in the decades after World War II by hard work and workers’ strength in numbers that came from the unions that represented them. Unions made sure that as our nation’s wealth and productivity grew, so too did the income and benefits of the people who worked hard to create that wealth. For decades, our nation’s prosperity was widely shared—wages increased and more employers provided their workers with health insurance, pensions, and paid time off. The middle class was also built by government policies that invested in infrastructure and basic science, supported homeownership and made a college education accessible to a new generation. Parents without higher education themselves saved to send their kids to college, made possible by affordable tuition at state universities and financial aid.
But all of this is changing and the middle class is now threatened. Median income is no higher than it was a decade ago and only workers with at least a bachelor’s degree earn more than their counterparts a generation ago. The nation’s once vibrant manufacturing sector—the engine that drove the growth of the post-war middle class—has gradually declined over the last three-plus decades. The bulk of recent job growth has been in the service sector, where unions are less prevalent, pay is lower, and benefits are limited or non-existent.
THE STATE OF THE MIDDLE CLASS
- High unemployment
- Declining access to benefits
- Higher costs to raise a family
- College degree increasingly out of reach
- Diminished economic prospects for young people
This is a briefing paper in Demos’ “Future Middle Class” series.
There’s been a dramatic shift in costs for health coverage from employers to employees as well as a rapid decline in the number of employers who even offer health insurance. Rising out-of-pocket costs mean that a family illness can lead to substantial expenses and medical debt. And as employers replace traditional pensions with 401(k)-type plans—again shifting costs and risks to employees—middle-class workers can no longer count on a secure retirement.
The middle class has also been hit by trends outside the labor market as it has become more costly to raise a family. High-quality child care is expensive, yet parents face these costs early in their working years when their earnings are low. Housing is also more expensive relative to household income than it was decades ago. The need for most working parents to have their own vehicle and the high price of gas have further strained middle-class family budgets. The growing gap between incomes and expenses fueled skyrocketing family debt in the two decades preceding the Great Recession.
The threat to the future of the middle class can be seen most clearly in the economic prospects for the nation’s young people. Overall, young workers today are earning less than their parents did a generation ago, with substantial wage declines among men. Skyrocketing college costs are making it hard for middle-class students to stay in school and graduate. The average student debt for college graduates is well over $20,000 and growing. Close to a third of young workers do not have employer-based health insurance, and most young people will pay for the lion’s share of their own future retirement benefits if current trends continue.
Now is the time for citizens, workers, employers, and policymakers to come together once again to rebuild pathways to the middle class, create good jobs with fair pay and decent benefits, and ensure that prosperity is broadly shared for the next generation.
EARNINGS
Over the last 30 years, median earnings for workers (ages 18-64) have ebbed and flowed with changes in the economy and the strength of workers. The typical full-time American worker earned $40,000 in 2010, which is 25 percent more (after inflation) than their 1980 counterparts (see Figure 1). However, this increase is quite small in the context of the nation’s overall growth during the same time period, when per-capita GDP increased 61 percent.
EDUCATION
A look at earnings by level of education reveals that over the last 30 years, median earnings were relatively flat or on the decline for most workers—with only significant growth in median earnings among those with bachelor’s degrees or higher. Over the last 30 years, workers with at least a bachelor’s degree increased their earnings by 25 percent after inflation. But high school graduates experienced only a 2 percent increase in earnings while workers without a high school diploma saw their wages plummet nearly 21 percent. As a result, workers with a bachelor’s degree or more earned nearly twice as much as high school graduates in 2010 ($58,000 versus $30,000 respectively). Earnings data for associate degree holders’ are available only since 1992, but show no lasting growth during that period (see Figure 2).
GENDER
Men typically earn more than women: in 2010, median earnings for men were $10,000 higher than those for women ($45,000 versus $35,000). But the gender gap has steadily declined over the last 30
years, in part because men’s wages have largely stagnated, rising only 7 percent in the past three decades. In contrast, median earnings for women have risen 47 percent since 1980 (see Figure 3). Nationally, women now earn 78 percent of what men do compared to only 57 percent three decades ago.
**RACE & ETHNICITY**
Median earnings for African-American and Latino workers have been consistently lower than those for white and Asian-American workers. In 2010, the typical African-American worker earned roughly $32,000, or 74 percent of the earnings of the typical white worker, which were $43,000. At $28,000, median earnings for Latinos were 65 percent of those for whites. The typical Asian-American worker earned $45,000 in 2010, slightly above the median for whites. All racial and ethnic groups experienced double-digit earnings growth over the last 30 years except Latinos whose earnings have been nearly flat.\(^3\)
**RISING INCOME INEQUALITY**
Since 1980, income inequality in the U.S. has widened considerably as income has become increasingly concentrated at the very top. In 2008, the top 10 percent of earners took nearly half (46 percent) of total national income, up from a third (33 percent) three decades earlier. Income concentration has grown most intensely among the richest of the rich: the top one tenth of one percent increased their share of national income from 2.2 percent to 7.8 percent between 1980 and 2008 (see Figure 4). In contrast, during the previous 30 years, 1950 to 1980, that same group’s share of total income stayed fairly steady and varied only between 1.9 percent and 3.5 percent.
---
**FIGURE 3. MEDIAN EARNINGS OF MALE AND FEMALE WORKERS IN THE U.S., 1980–2010 (2010 DOLLARS)**
*SOURCE: Demos analysis of Current Population Survey data*
**FIGURE 4. INCOME SHARE HELD BY TOP EARNERS IN THE U.S., 1980-2008**
*SOURCE: Piketty, T. and Saez, E. 2007 “Income and Wage Inequality in the United States 1913-2002” in Atkinson, A. B. and Piketty, T. Top Incomes over the Twentieth Century: A Contrast Between Continental European and English-Speaking Countries, Oxford University Press, chapter 5. Series updated by the authors. Data available at http://g-mond.parisschoolofeconomics.eu/topincomes/.*
JOBS AND BENEFITS
Access to well-paying jobs with good health and retirement benefits is the cornerstone of a middle-class life. Unions play an important role in helping workers negotiate fair pay and better benefits. All workers, not just union members, benefit from union gains. But only a fraction of U.S. workers are union members. Over time, workers—especially non-union workers—have been forced to absorb an ever-increasing portion of previously-shared costs and risks for health and retirement benefits.
UNION MEMBERSHIP
Compared to other countries, union participation in the U.S. has historically been low. Unionization in the U.S. has steadily eroded since its peak in the 1950s, when it reached nearly 30 percent.\(^4\) Nationally, only 11.9 percent of workers participated in a union last year (see Figure 5)—36 percent of public-sector workers (7.6 million) and 7 percent of private-sector workers (7.1 million).\(^5\)
Research demonstrates the key role unions play in raising wages and benefits, particularly for low-income workers. National data for 2003-07 show that unionization raises the wages of the typical low-wage worker (one in the 10th percentile of earnings) by 21 percent, compared to 14 percent for a worker earning median wages and 6 percent for the typical high-wage worker (one in the 90th percentile). Overall, unionization raised the average worker’s wage by 12 percent.\(^6\)
HEALTH INSURANCE
High out-of-pocket medical expenses are one of the primary causes of bankruptcy among the middle class, underscoring the importance of health insurance coverage.\(^7\) The proportion of full-time U.S. workers lacking health insurance is 15 percent, and access to employer-sponsored health insurance is declining: in 2010, nearly a quarter (24 percent) of full-time U.S. workers were not offered health coverage by their employer. Even as the economy boomed in the late 1990s, only 82 percent of workers had access to employer-sponsored health insurance.\(^8\) But employer-sponsored plans are not sufficient to protect workers: cost shifting to employees has decreased disposable income. Workers’ contributions for family health insurance coverage increased almost 150 percent between 2000 and 2010 to close to $4,000. Over half of employees pay more than 25 percent of the total cost of their insurance premiums.\(^9\)
RETIREMENT BENEFITS
Several factors threaten the ability of U.S. workers to look forward to a secure retirement. In a rate that has changed little over the past 30 years, only 60 percent of the nation’s full-time workers currently have access to a retirement plan at work. Even at the economy’s peak in 1999, the rate was only 67 percent.\(^{10}\) But employer-sponsored retirement plans have gradually shifted from pensions—whose costs and financial risks are borne almost exclusively by employers—to 401(k)-type plans that rely on worker contributions and expose individuals to the vagaries of the stock market and high fees, which eat away at returns. Nationally, roughly 63 percent of all employer-sponsored retirement plans are now 401(k)s or similar individual retirement plans.\(^{11}\) Over 8 percent of U.S. workers don’t participate in their employer-sponsored plan either because they can’t afford to contribute or fail to opt in (see Figure 6).
UNEMPLOYMENT
Although the U.S. has experienced multiple recessions since 1980, none saw unemployment increase so rapidly nor recover so slowly as the Great Recession. During the double-dip recession of the early 1980s, the unemployment rate peaked at 10.8 percent in November 1982. Although higher than the 10.1 percent peak of the recent recession, unemployment had fallen to 8.5 percent a year after the official end of the 1980s recession, and by November 1984, two years after the crisis’ end, unemployment stood at 7.2 percent. The recessions of the early 1990s and early 2000s were mild by comparison with far less severe job losses. In the months leading up to the Great Recession, unemployment was relatively low at 4.4 percent in May 2007. But the jobless rate began to rapidly ascend in early 2008 and it had more than doubled when it peaked at 10.1 percent in October 2009 (several months after the official end of the recession). A year later, it had fallen only to 9.7 percent and in July 2011, it was still stands over 9 percent.\(^{12}\) Further, 45 percent of the unemployed have been out of work for six months or more. Extensions of unemployment insurance benefits for the long-term unemployed are currently set to expire at the end of 2011.
The Great Recession, as previous recessions, has hit minorities and workers with the least education particularly hard. The unemployment rates for African-Americans more than doubled from its pre-recession low of 7.7 percent in August 2007 to a peak of 16.4 percent in January of 2010, and has fallen only slightly since, remaining at 15.9 percent as of July 2011. The unemployment rate for Latinos followed a similar pattern, rising from 5.5 percent in August 2007 to a high of 13.2 percent in November 2010 before falling to its current July 2011 level of 11.3 percent. The unemployment rate for white Americans, in contrast, rose from 4.2 percent in August 2007 to 9.4 percent in October 2009, and now stands at 8.1 percent. By education, the unemployment rate for workers 25 and older with less than a high school diploma, 15 percent, was nearly four times than those with a bachelor’s degree or higher, who had a 4.3 percent unemployment rate as of July 2011. The unemployment rate of workers with a high school but no diploma is halfway between the two at 9.3 percent.\(^{13}\)
**WHERE THE JOBS ARE**
The past 30 years have brought significant changes to the nation’s job market. One change that has had a dramatic impact on the economic fortunes of workers is the relative size of the manufacturing and service sectors. Jobs in these two sectors have accounted for more than half (roughly 54 percent) of all employment since 1980, but the sectors have experienced diverging trends. Manufacturing employment declined from 25 percent in 1980 to 11 percent in 2010, while service employment increased from 29 percent to 43 percent (see Figure 7). Manufacturing jobs are far more likely than service jobs to be unionized, pay decent wages, and offer middle-class benefits.
**FIGURE 7. EMPLOYMENT IN MANUFACTURING AND SERVICES AS A PERCENT OF THE LABOR FORCE, 1980-2010**
*SOURCE: Demos analysis of Current Population Survey data*
RAISING A FAMILY
Americans pride themselves on being able to pass on a better life to their children, but over the last generation, this dream has become increasingly out of reach. Even with two parents in the labor force, families struggle to meet the high costs of housing and child care, let alone save for a rainy day or invest in the future.
HOMEOWNERSHIP
In the 1980s, homeownership among American workers declined five percentage points, from 69 percent to 64 percent, before trending upward again in the 1990s. Just before the housing bubble burst, the homeownership rate among workers hit nearly 70 percent in 2005 before tumbling a couple of percentage points (see Figure 8).
National statistics mask, however, the fact that some states, localities, and neighborhoods were much harder hit than others by the wave of foreclosures that followed the subprime meltdown. During the first three years of the foreclosure crisis, from January 2007 through the end of 2009, African-American and Latino homeowners were disproportionately represented among the estimated 2.5 million foreclosures that had been completed: an estimated 8 percent of African-American and Latino borrowers lost their homes to foreclosure compared to 4.5 percent of whites. And these figures don’t include foreclosures in 2010 and beyond.\(^{14}\) Overall Americans are devoting a larger share of income to housing costs: in 2008, nearly two in five U.S. homeowners (38 percent) spent 30 percent or more of their income on housing.\(^{15}\)
**FIGURE 8. HOME OWNERSHIP AMONG U.S. WORKERS, 1980-2010**
*SOURCE: Demos analysis of Current Population Survey data*
CHILD CARE
Child care can be one of the largest expenses families face. In 18 states, center-based child care fees for two children (an infant and a 4-year-old) exceed housing costs. And since 2000, the cost of child care has increased twice as fast as the median income of families with children.\(^{16}\) The cost of care varies dramatically from state to state. For example, the average annual cost of full-time center-based care for a 4-year-old ranges from a low of $3,900 in Mississippi to a high of $12,200 in Massachusetts; center-based care for an infant ranges from $4,650 in Mississippi to $16,500 in Massachusetts (see Figure 9). The national average annual cost for center-based care is $8,900 for an infant and $7,150 for a 4-year-old. A family with two children (an infant and a 4-year-old) in center-based care priced at this national average would pay $16,050 annually for child care, which amounts to 22 percent of family income for a two-earner couple earning median wages or 52 percent of income for a single-mother earning the median wage for women.
**FIGURE 9. AVERAGE ANNUAL PRICE OF FULL-TIME CHILD CARE AMONG THE STATES, 2011**
| | CHILD CARE CENTER | FAMILY CHILD CARE HOME |
|------------------------|-------------------|------------------------|
| Infant, full-time | $4,650 – $16,500 | $3,850 – $12,100 |
| | (Mississippi) | (Mississippi) |
| 4 year old, full-time | $3,900 – $12,200 | $3,600 – $11,300 |
| | (Mississippi) | (Mississippi) |
*SOURCE: National Association of Child Care Resource and Referral Agencies, “Most Recent Child Care Data by State.”*
The trends facing young Americans seeking to build and maintain a middle-class life are worrisome. Over the last generation, wages have stagnated or declined for all young workers except for those with a bachelor’s or graduate degree. While a college degree provides higher earnings and greater protection against unemployment, college tuition costs have soared and students are accumulating greater amounts of debt.
**LABOR MARKET**
In the course of one generation, young workers have lost considerable ground. In 2010, the only education level with higher typical earnings than in 1980 were young workers with a bachelor’s degree or higher. Real median earnings (after inflation) for young college graduates with at least a four-year degree rose 21 percent over the last 30 years, although after a series of gains and losses, young college grads in 2010 just 4 percent more than they did in 1987, and 3 percent less than they did a decade ago. Earnings for young workers with only a high school diploma and those who did not finish high school declined 10 percent and 16 percent respectively since 1980. Young workers with some college but less than a bachelor’s degree lost ground as well (see Figure 10).
The earnings gap between young men and young women in the workforce has historically been narrower, and has narrowed faster, than that of the entire workforce, though it too has not closed entirely. Young female workers ages 25 to 34 earned 65 percent of what young men did in 1980, but by 2006 their earnings were 91 percent of young men’s, the gap has since risen slightly to 87 percent today. The narrowing of the gender gap between young people’s earnings was entirely due to the rise in young women’s wages as they entered the workforce in greater numbers; their earnings rose by 27 percent over the past three decades, while young men’s actually decreased by over 4 percent over the same period.
In 2010, the national unemployment rate for workers under age 25 and not enrolled in school was 18.4 percent—nearly double the overall U.S. unemployment rate of 9.6 percent. Unemployment among young high school graduates is abysmally high; it was 22.5 percent nationally in 2010 compared to 9.3 percent among young workers with a four-year college degree.\(^{17}\)
---
**FIGURE 10. MEDIAN ANNUAL EARNINGS OF U.S. WORKERS AGES 25-34 BY EDUCATION, 1980-2010 (2010 DOLLARS)**
*SOURCE: Source: Demos analysis of Current Population Survey data.*
*Data not available for “Associate’s Degree” prior to 1992.*
**Due to limitations in the CPS, “Some College” includes both those with Associate’s Degrees and some college but no diploma prior to 1992, but only those without degrees afterwards.*
COLLEGE TUITION
Over the past 20 years or so, average in-state tuition (including fees) at colleges and universities across the nation has increased every year, rising a total of 254 percent. For 2009-10, annual in-state tuition averaged $6,829 (see Figure 11). (Note that these figures do not include room and board.) Much of the cost increase is due to the gradual but steady erosion of state funding for their own institutions, leaving students and families to pick up a much larger share of the tab.
STUDENT DEBT
Two thirds (67 percent) of college graduates in the U.S. entered the labor force with student debt in 2009, and their average debt was $24,000. Both the percentage of graduates with debt as well as the amount they owe have increased in recent years.\(^{18}\) And rising numbers of students are accumulating debt without completing a degree, putting them on a shaky path to the future.
HEALTH AND RETIREMENT BENEFITS
Nationally, young full-time workers ages 25-34 are more likely than other workers to lack health insurance—20 percent are currently uninsured compared to 15 percent of all workers. Over a quarter (27 percent) of young workers lack access to health insurance through their employer, a figure that has grown substantially over the last decade (see Figure 12). Just over half (56 percent) of young workers have access to an employer-sponsored retirement plan, the vast majority of which are risky, expensive 401(k)-type plans rather than traditional pensions. Nearly 11 percent of workers ages 25-34 have access to a retirement plan at work but do not participate.\(^{19}\)
CONCLUSION
The American Dream came to life in the form of a strong post-war middle class that sustained the national economy for decades. But for the first time in generations, more people are falling out of the middle class than joining its ranks. The economy is still productive, but the gains are accruing primarily to the top and workers are no longer getting their fair share. It doesn’t have to be this way. Just as the post-war middle class was built, it is possible to rebuild it and strengthen it for the next generation. That will require the strength of workers coming together to reclaim the American Dream and demanding that our elected officials work for workers.
ENDNOTES
1. All earnings data cited in this brief are from Demos analysis of the Current Population Survey (CPS).
2. The World Bank, World Development Indicators, “GDP per capita (constant 2000 US$),” http://data.worldbank.org/indicator/NY.GDP.PCAP.KD?page=5
3. Demos analysis of the Current Population Survey (CPS) using 2-year averages.
4. Gerald Mayer, “Union Membership Trends in the United States,” Federal Publications, Paper 174, 2004. http://digitalcommons.ilr.cornell.edu/key_workplace/174
5. Economic News Release, Bureau of Labor Statistics, U.S. Department of Labor, January 21, 2011. http://www.bls.gov/news.release/union2.nr0.htm
6. John Schmitt, “The Union Wage Advantage for Low-Wage Workers,” Center for Economic and Policy Research, May 2008. http://www.cepr.org/2008/200805UnionWageAdvantageCEPRReport.pdf
7. David Himmelstein, Deborah Thorne, Elizabeth Warren and Steffie Woolhandler. “Medical Bankruptcy in the United States, 2007: Results of a National Study,” American Journal of Medicine, 2007. http://www.washingtonpost.com/wp-srv/politics/documents/american_journal_of_medicine_09.pdf
8. Demos analysis of the Current Population Survey (CPS).
9. The Kaiser Family Foundation, Employer Health Benefits: 2010 Summary of Findings. http://ehbs.kff.org/pdf/2010/8086.pdf
10. Demos analysis of the Current Population Survey (CPS).
11. Alicia H. Munnell, Francesca Golub-Sass, and Dan Muldoon, “An Update on 401(k) Plans: Insights from the 2007 SCF,” Center for Retirement Research, 2009.
12. Bureau of Labor Statistics, U.S. Department of Labor.
13. Bureau of Labor Statistics, U.S. Department of Labor.
14. Debbie Gruenstein Bocian, Wei Li, and Keith S. Ernst, “Foreclosures by Race and Ethnicity: The Demographics of a Crisis.” Center for Responsible Lending, June 2010. http://www.responsiblelending.org/mortgage-lending/research-analysis/foreclosures-by-race-and-ethnicity.pdf
15. Data compiled by the U.S. Census Bureau from the 2008 American Community Survey, “Table 991. Owner-Occupied Housing Units—Value and Costs by State: 2008.”
16. National Association of Child Care Resource and Referral Agencies, “Parents and the High Cost of Child Care, 2010 Update,” August 2010. http://www.naccra.org/docs/Cost_Report_073010-final.pdf
17. Heidi Shierholz and Kathryn Anne Edwards, “The Class of 2011: Young Workers Face a Dire Labor Market Without a Safety Net,” Economic Policy Institute, April 2011. http://epi.3cdn.net/c7d6ec52122ea9c188_rbm6bc14a.pdf
18. The Project on Student Debt, “Student Debt and the Class of 2009,” October 2010. http://projectonstudentdebt.org/files/pub/classof2009.pdf
19. Demos analysis of the Current Population Survey (CPS).
ABOUT DĒMOS
Dēmos is a non-partisan public policy research and advocacy organization. Headquartered in New York City, Dēmos works with advocates and policymakers around the country in pursuit of four overarching goals: a more equitable economy; a vibrant and inclusive democracy; an empowered public sector that works for the common good; and responsible U.S. engagement in an interdependent world. Dēmos was founded in 2000.
In 2010, Dēmos entered into a publishing partnership with *The American Prospect*, one of the nation’s premier magazines focusing on policy analysis, investigative journalism and forward-looking solutions for the nation’s greatest challenges.
CONTACT
DÉMOS
220 Fifth Avenue, 5th Floor
New York, New York 10001
Phone: (212) 633-1405
Fax: (212) 633-2015
firstname.lastname@example.org
DÉMOS MEDIA
Tim Rusch,
Communications Director
email@example.com
(212)-389-1407 |
Pulse wave velocity as endpoint in large-scale intervention trial. The Complior® study
Roland Asmar, Jirar Topouchian, Bruno Pannier, Athanase Benetos and Michel Safar, on behalf of the Scientific, Quality Control, Coordination and Investigation Committees of the Complior Study
Objective To evaluate the ability of an antihypertensive therapy to improve arterial stiffness as assessed by aortic pulse wave velocity (PWV) in a large population of hypertensive patients.
Setting Sixty-nine healthcare centres, private and institutional (19 countries).
Patients Subjects aged 18–79 years, with essential hypertension. A total of 2187 patients were enrolled; 1703 (52% male) completed the study: mean age = 50 ± 12 years; mean baseline systolic/diastolic blood pressure (S/D BP) = 158 ± 15/98 ± 7 mmHg; mean baseline carotid-femoral PWV = 11.6 ± 2.4 m/s.
Interventions Patients were treated for 6 months, starting with perindopril (angiotensin converting enzyme (ACE) inhibitor) 4 mg once daily (OD), increased to 8 mg OD, and combined to diuretic (indapamide 2.5 mg OD) if BP was uncontrolled (> 140/90 mmHg).
Results It was feasible to measure carotid-femoral PWV using the automatic device Complior® at inclusion, 2 and 6 months, along with conventional BP assessments in a population of 1703 patients. Significant decreases \((P < 0.001)\) in BP (systolic: −23.7 ± 16.8, diastolic: −14.6 ± 10 mmHg), and carotid-femoral PWV (−1.1 ± 1.4 m/s) were obtained at 2 and 6 months.
Conclusions The Complior Study is the first study to show the feasibility of a large-scale intervention trial using PWV as the endpoint in hypertensive patients. Adequate results may be obtained using an automatic device and rigorous criteria for assessment. A long-term controlled intervention study is needed to confirm the results of the present uncontrolled trial. *J Hypertens* 19:813–818 © 2001 Lippincott Williams & Wilkins.
Journal of Hypertension 2001, 19:813–818
Keywords: angiotensin converting enzyme (ACE) inhibition, antihypertensive therapy, arterial stiffness, feasibility, hypertension, pulse wave velocity
aL’Institut CardioVasculaire, 21 Boulevard Delessert, Paris and bHôpital Broussais, Service de Médecine I, 96 rue Didot, Paris, France.
Sponsorship: L’Institut Cardiovasculaire, Paris, Servier Research Group, Neuilly-sur-Seine, France.
Correspondence and requests for reprints to Professor Roland Asmar, L’Institut CardioVasculaire, 21 Boulevard Delessert, Paris 75016, France. Tel: +33 1 55 74 66 66; fax: +33 1 55 74 66 65; email: email@example.com
Received 26 April 2000 Revised 21 November 2000 Accepted 22 November 2000
Introduction Morbidity and mortality in hypertension are primarily related to arterial damages that may affect one or several organs. Structural and functional properties of the arterial wall have been described to be altered, even at early stages of hypertension [1–4]. Considering the potential implications of arterial assessment in the preventive management of cardiovascular disease, the evaluation of the effect of antihypertensive drugs on the arterial stiffness is recommended by numerous authorities [5]. Several pharmacological double-blind trials have shown that antihypertensive treatments present unequal efficacy on the arterial wall, despite a similar effect on blood pressure (BP) [3,6,7]. However, these results were observed in small and specific populations of subjects. Among the antihypertensive drugs, angiotensin converting enzyme (ACE) inhibitors are likely to improve arterial compliance independently of mean BP level [8–10] but the ability of antihypertensive treatment to reverse the arterial abnormalities observed in hypertension has never been assessed in large-scale studies.
The Complior Study is the first international multicentre clinical trial aimed at analysing the feasibility of using the automatic measurement of pulse wave velocity (PWV) as an endpoint in a large-scale intervention trial.
Methods Study objectives The study primary objective was to evaluate the ability of a long-term antihypertensive treatment to reverse arterial stiffness as assessed by an automatic measurement of aortic PWV in a large population of hypertensive patients.
Patients
Male and female patients were included in the study provided they were aged 18–79 years, and presented either untreated essential mild or moderate hypertension, defined as a diastolic BP (DBP) ≥ 95 and ≤ 114 mmHg, and/or a systolic BP (SBP) ≥ 160 and ≤ 200 mmHg, or treated uncontrolled hypertension (SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg), or treated and controlled hypertension (BP < 140/90 mmHg) with adverse effects due to the current antihypertensive treatment. Main exclusion criteria were: secondary hypertension, complicated hypertension with recent organ damage (< 6 months), arterial stenosis > 70%, overweight defined as a body mass index (BMI) > 34 kg/m², uncontrolled diabetes or fasting glycaemia > 10 mmol/l, and kaliemia > 5.2 mmol/l.
Study design
At the end of the run-in period, patients received a 6-month treatment regimen based on perindopril 4 mg once daily (OD), increased to 8 mg OD and combined with a diuretic (indapamide 2.5 mg OD) in case of uncontrolled BP (> 140/90 mmHg). The assessment of the drug efficacy was performed over 1-month therapeutic periods during the first 3 months of the study (M0, M1, M2, M3), and at discharge (M6). At each visit, a complete clinical examination was carried out, and PWV was measured at M0, M2, and M6.
Procedures for measuring blood pressure and pulse wave velocity
Clinic BP was measured at each visit, in compliance with World Health Organization guidelines, using a mercury sphygmomanometer with a cuff appropriate to the arm circumference, in patients at rest for 10 min (Korotkoff phase I for SBP and V for DBP). Three measurements were carried out and averaged for analysis.
Arterial distensibility was assessed by automatic carotid-femoral PWV measurement using the Complior® device; the technical characteristics of this device have been described [11], and indicate inter- and intra-observer repeatability coefficient values ≥ 0.9. The basic principle of PWV assessment is that the pressure pulse generated by ventricular ejection is propagated along the arterial tree at a speed determined by the geometric and elastic properties of the arterial wall [1]. PWV is calculated from measurements of pulse transit time and the distance travelled by the pulse between two recording sites, according to the following formula: PWV (m/s) = distance (m)/transit time (s). Carotid-femoral PWV is calculated from the time delay between the recorded proximal (carotid) and distal (femoral) feet of the wave, and the superficially measured distance separating the respective transducers.
Logistics and organization
Eighty centres (22 countries) were selected by the Scientific and Organization Committee of the study; all were provided with Complior® devices adapted to the national regulatory and custom requirements of each participating country.
Investigators, numbering 129, participated in specific training sessions on PWV assessment organized for small groups (four–six persons) of investigators by the Training and Certification Committee of the study. These sessions were aimed at making the investigators familiar with the Complior® device and technical aspects related to its handling, and to practice PWV measurements on healthy volunteers.
In order to limit inter-centre variability on PWV measurements, and to ensure data homogeneity, a pre-study was carried out for the investigators’ certification. Each investigator was eligible for participation in the Complior Study provided he or she obtained this certification, on the basis of 20 validated series of PWV measurements. The validation of PWV recordings was performed as follows: in addition to a blind evaluation by two experts from the coordination centre (L’Institut CardioVasculaire, Paris), an objective analysis of PWV recordings was realized for quality control using specific software.
Online assistance (fax or e-mail) was set-up to answer investigators’ queries. The criteria established for this quality control were the baseline stability of the recorded PWV, the variations of the baseline according to pulse wave amplitude, the wave shape and the abrupt systolic upstroke of the initial parts of the recorded proximal and distal pressure waves.
Recordings were electronically forwarded to the coordination centre and immediately reviewed for validation by the Quality Control Committee of the study, allowing rapid feed-back to the investigators and the progressive constitution of the study database, as well as questions related to the study protocol, methodology, scientific or technical problems. Two independent services were organized either to manage scientific and medical aspects of the study, or to provide technical assistance on Complior® handling. The most frequently asked questions forwarded through this ‘Complior-Line’ were summarized in the study newsletter Complior News, published periodically (3 months) in order to inform the participating teams on the study progress.
Statistical analysis
Statistical analyses were performed using the NCSS software (Number Cruncher Statistical Systems, NCSS 6.0 for Windows, Kaysville, Utah, USA). The quantitative variables are presented as mean values ± standard deviation; the qualitative variables as absolute values and percentages. Comparisons of quantitative variables between untreated and previously treated patients, or according to gender were carried out, using an analysis of variance adjusted for age and BP level. Analysis of variance for repeated measurements (M0, M2, M6) for both absolute and relative changes and adjustment for BP values were performed. $P < 0.05$ was considered significant.
**Results**
**Study population**
After certification, 2187 patients were pre-included in the study; 484 were subsequently excluded, either due to non-compliance with the protocol criteria, or after their PWV recording was eliminated for insufficient quality ($n = 211$). A total of 1703 patients (52% male) were included at M0, aged 50 ± 12 years and presenting the following anthropometric characteristics: mean weight = 73 ± 13 kg, height = 165 ± 10 cm, body mass index (BMI) = 26.8 ± 3.6 kg/m², and waist-to-hip ratio = 0.89 ± 0.08. According to the BP level, 63% were classified as having grade I hypertension, 32% as grade II, and 4% as grade III. Regarding cardiovascular parameters, the patients had a mean baseline SBP = 158 ± 15 mmHg, DBP = 98 ± 7 mmHg, heart rate = 75 ± 10 bpm, and carotid-femoral PWV = 11.6 ± 2.4 m/s. Among these patients, 83% had never received an antihypertensive treatment (85% for men, 80% for women), while 17% had been previously treated (15% for men, 20% for women), with β-blockers ($n = 102$), diuretics ($n = 28$), calcium channel blockers ($n = 143$), central agents ($n = 8$) and others ($n = 9$). 1371 patients presented available PWV data at each of the M0, M2, and M6 evaluations.
**Effects on blood pressure and pulse wave velocity**
Table 1 shows the mean values and changes (Δ) in BP and PWV after 2 and 6 months; significant reductions in SBP, DBP, mean, and pulse BP were observed ($P < 0.001$). The rate of normalized patients was 52% (the treatment resulted in a SBP < 140 mmHg in 60% of the patients and in a DBP < 90 mmHg in 70% of the patients). At M6, the protocol treatment regimen was: perindopril 4 mg OD in 67%, perindopril 8 mg OD in 22%, and perindopril plus indapamide in 11% of the patients. Regarding PWV, a significant reduction in values was observed (0.9 ± 1.4 after 2 months, and 1.1 ± 1.4 after 6 months; $P < 0.001$).
Table 2 shows mean changes values in BP and PWV after 6 months of treatment, according to baseline BP level (Table 2a) and pre-study treatment status (Table 2b). BP and PWV improvements, were significantly ($P < 0.001$) related to baseline hypertension grade in both men and women. BP reduction was more pronounced ($P < 0.001$) in previously untreated patients than in those previously treated, whereas similar PWV changes were observed in both groups (Table 2b).
**Discussion**
The Complior Study is the first study aimed at assessing the feasibility of using PWV as endpoints in a large-scale intervention trial. The major finding of this study is that the assessment of arterial stiffness in hypertensive patients, and the evaluation of arterial effects of long-term antihypertensive therapy in a large-scale multicentre study are feasible using an automatic device for aortic PWV measurements, provided some methodological aspects are considered. Different methodological aspects related to these results need to be discussed.
**Methodological aspects**
**Choice of automatic pulse wave velocity measurements**
Several non-invasive methods are available and may be useful in assessing large artery hemodynamics (arterial geometry, function), or in analysing the shape of the BP curve [12]. Most are complex, expensive, time-consuming, need specifically qualified operators, and they remain reserved to very few clinical research labs. Since the objective of this study was to evaluate the arterial stiffness in a large multicentre clinical trial, the use of a simple and accurate method appears suitable. PWV is a well-established index of arterial distensibility and stiffness, related to the arterial geometry and wall function (see Methods) [13]. Aortic PWV assessment was therefore utilized in this study and an automatic device that has been shown to provide accurate and
---
**Table 1** Treatment effects on blood pressure and pulse wave velocity; mean values and changes from baseline (M0) during (M2) and at the end of the study (M6)
| Variables | M0 | M2 | M6 | Δ(M2–M0) | P | Δ(M6–M0) | P |
|-----------------|--------|--------|--------|----------|--------|----------|--------|
| SBP (mmHg) | 158 ± 15 | 139 ± 16 | 134 ± 13 | −20 ± 17 | < 0.001 | −24 ± 17 | < 0.001 |
| DBP (mmHg) | 98 ± 7 | 86 ± 9 | 84 ± 8 | −12 ± 10.1 | < 0.001 | −14 ± 10 | < 0.001 |
| MAP (mmHg) | 118 ± 8 | 103 ± 10 | 100 ± 9 | −15 ± 11 | < 0.001 | −18 ± 11 | < 0.001 |
| PP (mmHg) | 59 ± 15 | 52 ± 12 | 50 ± 10 | −7 ± 14 | < 0.001 | −9 ± 15 | < 0.001 |
| HR (bpm) | 75 ± 10 | 75 ± 9 | 75 ± 10 | −0.4 ± 10 | NS | −0.3 ± 10 | NS |
| PWV (m/s) | 11.6 ± 2.6 | 10.7 ± 2.2 | 10.5 ± 2.1 | −0.9 ± 1.4 | < 0.001 | −1.1 ± 1.4 | < 0.001 |
SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; PP, pulse pressure; HR, heart rate; PWV, pulse wave velocity.
reproducible measurements was employed for these measurements [11]. On the other hand, two disadvantages related to PWV determination need to be mentioned. First, this method allows an overall assessment of the arterial wall which is related to the arterial geometry and wall function; it does not provide information about the exact underlying mechanisms or specific factors that may be involved in creation of arterial abnormalities or their changes. In fact, changes in PWV may be due to changes either in structural parameters such as the arterial wall thickness, the internal diameter or, to changes in functional parameters such as changes in blood pressure (see Methods). This methodological limitation is not a problematic point in this study, since its primary objective was not to analyse the precise mechanisms of the observed changes in arterial properties. Second, the superficial measurement of the distance travelled by the pulse wave between the carotid and femoral arteries constitutes the major error in PWV determination. In fact, the non-invasive superficial measurement of the arterial pathway allows only an estimate of the distance; accurate measurements of this distance are obtained only by invasive procedures. Some authors suggested a possible correction based on anatomic dimensions, whereas others recommended subtracting the distance between the suprasternal notch to the carotid location from the total distance because the pulse is travelling in the opposite direction [13]. Consequently, and in order to limit the observer errors in the distance measurement by adding up a supplementary measurement (from the carotid to the suprasternal notch), no subtraction has been performed in this study, the distance being determined using a direct linear measurement between the carotid and the femoral sites. Elsewhere, considering that direct superficial measurement of the distance between the carotid and the femoral arteries includes the chest and the abdominal area, men with abdominal obesity and women with large bust measurements, present an overestimation of the real arterial pathway and hence of the PWV. In order to limit such overestimation, it is recommended that one measures the superficial distance as directly as possible; moreover patients with a very high degree of obesity (body mass index > 40 kg/m²) were not included in this study. Furthermore, it is to be noted that if the distance error is important for the determination of PWV absolute value, it represents only a relative limitation in this study where analysis were based on within patient comparison.
**Logistics and organization**
From the 129 investigators (80 centres, 20 countries) who entered the certification pre-study, 107 (69 centres, 19 countries) obtained certification and participated in the Complior Study per se. Most of them were clinicians and very few had previous experience in non-invasive arterial hemodynamics research. Our results show that evaluation of large arteries using an automatic device for PWV measurement is applicable to clinical trials performed in large populations, in healthcare centres. Moreover, this study took into account some important methodological aspects: (1) the certification procedures allowed the construction of an homogeneous database; (2) online assistance and quality control allowed to limit the loss of data as frequently observed in long-term treatment and large-scale studies; (3) the electronic management of data directly acquired from computerized recordings has the advantage to save time of data input and prevents typing errors. However, since even electronic data transfer may also be erroneous, rigorous evaluation and coherence analysis were performed by two independent
observers to ensure quality. These procedures, that may seem cumbersome and restrictive when setting-up multicentre studies may have benefit when one considers the low rate (under 10%) of PWV data eliminated due to insufficient quality. In fact, among the 211 PWV recordings eliminated for insufficient quality, 142 were at baseline (M0), 39 at M2 and 30 at M6 showing that more than 50% were excluded before the inclusion and that the recording quality was stable during the study period. Nevertheless, advantages related to such quality procedures can only be presumed and suggested since no comparison has ever been performed with another system lacking such procedures. Moreover, arguments for this cannot come from previous pharmacological studies with PWV measurements since most of them were either monocentric or performed in research centres familiar with this technique; thus, their results may not be extrapolated to those observed in the Complior Study where assessments were performed by clinicians with no previous experience of PWV measurement.
**Comments on the observed results**
The other results of the Complior Study are the decrease of BP and aortic PWV in subjects with essential hypertension after long-term antihypertensive treatment regimen based on ACE inhibitor. In previous pharmacological controlled studies performed in limited numbers of patients, we and others showed that ACE inhibition improves the properties of large arteries [14–24]. Therefore, the results of the present study are in agreement with those obtained previously in controlled studies.
Although rigorous quality criteria were employed and different committees were involved in this trial, the present study was designed as an open uncontrolled treatment study, for feasibility reasons, and thus was not aimed at answering questions raised by pharmacological studies, or at analysing precisely the mechanisms that may be involved in the observed changes. The analysis of the relationship between PWV and BP showed a significant correlation \((P < 0.001)\), but the scatter plots of individual changes in BP and PWV show a significant but weak correlation between changes \((r = 0.264)\), with numerous divergent points indicating that change in PWV is not always concomitant with BP changes and vice-versa. It is possible that the observed decrease in BP and PWV could be exclusively related to a regression toward the mean or to a placebo effect. However, several points constitute arguments against this simplistic hypothesis. First, in this investigation, analysis of the results adjusted according to gender, the grade of hypertension and the existence of previous treatment clearly show that the regression to the mean could not be the exclusive factor involved in BP and PWV changes. Second, from previous separate double-blind studies [25–27], the proportion of the placebo effect and the regression to the mean were estimated \(-6 \pm 11\) mmHg for SBP, \(-5 \pm 8\) mmHg for DBP, and \(-0.18 \pm 1.20\) m/s for PWV. Third and mostly, the results of a control group of 369 normotensive subjects (67% male) aged 55 ± 11 years showed over a 6-year follow up period an increase of BP of \(6 \pm 2\) mmHg for the SBP and \(2 \pm 1\) mmHg for the DBP and of \(1.50 \pm 2.2\) m/s for the carotid-femoral PWV (about 0.25 m/s per year) (personal data, A.B.). In this study, the observed changes from baseline \((-24 \pm 17\) mmHg for SBP, \(-14 \pm 10\) mmHg for DBP, and \(-1.10 \pm 1.40\) m/s for PWV) were first higher than the estimated placebo and regression to the mean effects, and second higher than those observed in the control group.
In conclusion, the Complior Study has shown that changes in arterial stiffness during chronic treatment of hypertension may be assessed noninvasively in a clinical trial involving a large population. Adequate results may be obtained using an automatic device with rigorous criteria for assessment. A long-term double-blind randomized, controlled intervention study is needed to evaluate whether long-term arterial changes, such as the decrease of PWV observed in this study, are due to BP changes, to antihypertensive treatment itself, or to a combination of both factors. Whether this reversion of arterial stiffness may improve the cardiovascular prognostic of hypertensive-treated patients is to be confirmed by specific studies.
**Acknowledgments**
The authors address their grateful thanks to all members of the Scientific, Quality Control, and Coordination Committees and to all investigators.
**References**
1. Nichols WW, O’Rourke MF. *McDonald’s blood flow in arteries: theoretical, experimental and clinical principles*. 3rd edn. London: Edward Arnold; 1990. pp. 77–142, 216-269, 283-359, 308–437.
2. Isnard R, Pannier B, Laurent S, London G, Diebold B, Safar ME. Pulsatile diameter and elastic modulus of the aortic arch in essential hypertension: a noninvasive study. *J Am Coll Cardiol* 1989; **13**:399–405.
3. Asmar R, Benetos A, London G, Hugue Ch, Weiss Y, Topouchian J, et al. Aortic distensibility in normotensive untreated and treated hypertensive patients. *Blood Press* 1995; **4**:48–54.
4. O’Rourke M. Mechanical principles in arterial disease. *Hypertension* 1995; **26**:2–9.
5. World Health Organization. International Society of Hypertension guidelines for the management of hypertension. Guidelines Subcommittee. *J Hypertens* 1999; **17**:151–183.
6. Benetos A, Laurent S, Asmar R, Lacolley P. Large artery stiffness in hypertension. *J Hypertens* 1997; **15**(suppl 2):S89–S97
7. Asmar R, Benetos A, Chaouche-Teyara K, Raveau-Landon C, Safar M. Comparison of effects of felodipine versus hydrochlorothiazide on arterial diameter and pulse wave velocity in essential hypertension. *Am J Cardiol* 1993; **72**:794–798.
8. Asmar RG, Pannier B, Santoni JP, Laurent S, London GM, Levy BI, et al. Reversion of cardiac hypertrophy and reduced arterial compliance after converting enzyme inhibition in essential hypertension. *Circulation* 1988; **78**:941–950.
9 Kool MJ, Lustermans FA, Breed JG, Struyker Boudier HA, Hoeks AP, Van Bortel L. Effect of perindopril and amiloride/hydrochlorothiazide on haemodynamics and vessel wall properties of large arteries. *J Hypertens* 1993; **11**(suppl 5):362–363.
10 Topouchian J, Brisac AM, Pannier B, Vicaut E, Safar M, Asmar R. Assessment of the acute arterial effects of converting enzyme inhibition in essential hypertension: a double-blind, comparative and crossover study. *J Hum Hypertens* 1998; **12**:181–187.
11 Asmar R, Benetos A, Topouchian J, Laurent P, Pannier B, Brisac AM, et al. Assessment of arterial distensibility by automatic pulse wave velocity measurement. Validation and clinical application studies. *Hypertension* 1995; **26**:485–490.
12 Asmar RG. Noninvasive evaluation of arterial abnormalities in hypertensive patients. *Pathol Biol (Paris)* 1999; **47**:885–895.
13 Asmar R. Arterial stiffness and pulse wave velocity. In: *Clinical applications*. 1999; pp. 37–45.
14 Perret F, Mooser V, Hayoz D, Tardy Y, Meister JJ, Etienne JD, et al. Evaluation of arterial compliance-pressure curves; effect of antihypertensive drugs. *Hypertension* 1991; **18**(suppl II):II77–II83.
15 Bouthier JD, Safar ME, Benetos A, Simon A, Levenson J, Hugue C. Hemodynamic effects of vasodilating drugs on the common carotid and brachial circulations of patients with essential hypertension. *Br J Clin Pharmacol* 1986; **21**:139–142.
16 Twardowski J, Asmar R, Sayyah F, Rudnicki A, Benetos A, Bacri AM, et al. Changes in arterial structure and function under trandolapril-verapamil combination in hypertension. *Stroke* 1999; **30**:1056–1064.
17 Richer C, Thuilliez C, Giudicelli JP. Perindopril converting enzyme blockade, and peripheral arterial hemodynamics in the healthy volunteer. *J Cardiovasc Pharmacol* 1987; **9**:94–102.
18 London G, Pannier B, Vicaut E, Guerin AP, Marchais SJ, Safar ME, et al. Antihypertensive effects and arterial hemodynamic alterations during ACE inhibition. *J Hypertens* 1998; **16**:1139–1146.
19 Chen CH, Ting CT, Lin SJ, Hsu FCP, Siu CO, Chou P, et al. Different effects of fosinopril and atenolol on wave reflections in hypertensives. *Hypertension* 1995; **25**:1034–1041.
20 Barenbrock M, Spieker C, Hoeks APG, Ziedek W, Rahn KH. Effect of lisinopril and metoprolol on arterial distensibility. *Hypertension* 1994; **24**(suppl. I):I161–I163.
21 Kool MJ, Lustermans FA, Breed JG, Struyker Boudier HA, Hoeks AP, Reneman RS, et al. The influence of perindopril and the diuretic combination amiloride + hydrochlorothiazide on the vessel wall properties of large arteries in hypertensive patients. *J Hypertens* 1995; **13**:839–848.
22 Lalleche A, Gaultier S, Topouchian J, Wilmet CS, Girerd X, Safar ME, et al. Differential responses of the heart and vasculature to chronic blood pressure reduction in essential hypertension. *Clin Pharmacol Ther* 1998; **64**:96–105.
23 Safar M, Van Bortel L, Struyker Boudier H. Resistance and conduit arteries following converting enzyme inhibition in hypertension. *J Vasc Res* 1997; **34**:34–67.
24 London GM, Pannier B, Vicaut E, Guerin AP, Marchais SJ, Safar ME, et al. Antihypertensive effects and arterial hemodynamic alterations during angiotensin converting enzyme inhibition. *J Hypertens* 1996; **14**:1139–1146.
25 Asmar R, Kerihuel J, Girerd X, Safar M. Effect of bisoprolol on blood pressure and arterial hemodynamics in systemic hypertension. *Am J Cardiol* 1991; **68**:61–64.
26 Asmar R, Benetos A, Darne B, Pauly N, Safar M. Converting enzyme inhibition: dissociation between antihypertensive and arterial effects. *J Hum Hypertens* 1992; **6**:381–385.
27 Asmar R, Benetos A, Brahimi M, Chaouche-Teyara K, Safar M. Arterial and antihypertensive effects of nifedipine: a double blind comparison versus placebo. *J Cardiovasc Pharmacol* 1992; **20**:858–863.
**Appendix**
**Participants in the Complior Study**
Abdul Rahman A. (Malaysia), Adamus J. (Poland), Al-Attia H. (UAE), Arteaga Fernandez E. (Brazil), Ashour Z. (Egypt), Aydinlar A. (Turkey), Bellido C. (Argentina), Pires-Brandao A. (Brazil), Buranaki Jaroen P. (Thailand), Calaca J. (Portugal), Calvo Lopez C. (Spain), Carmona P. (Portugal), Casado Perez S. (Spain), Ceremuzynski L. (Poland), Cieslinski A. (Poland), Collado F. (Philippine), De Sa Cunha R. (Brazil), Dinh Thu Huong (Vietnam), Dirikan Y. (Turkey), Erdine S. (Turkey), Erguney M. (Turkey), Fernández Andrade C. (Spain), Fernandez Torres C. (Spain), Forascepi Roza R. (Spain), Galarza C. (Argentina), Gil Extremera B. (Spain), Gonzales (Mexico), Gunasegaran K. (Singapore), Hall M. (Ireland), Hamed A.S. (Egypt), Harmanda A. (Turkey), Herrera Perez Del Villar J. (Spain), Jalil Milad J. (Chile), Januszewicz W. (Poland), Kabat M. (Poland), Kawecka-Jaszcz K. (Poland), Klimaszewska K. (Poland), Koanantakul B. (Thailand), Kokot F. (Poland), Kornacewicz-Jach Z. (Poland), Kurpesa M. (Poland), Krupa-Wojciechowska B. (Poland), Kruzenska (Poland), Krzeminska-Pakula M. (Poland), Kuch J. (Poland), Chan K.K. (Republic of China), Li K. (Republic of China), Lopez Eady D. (Spain), Lopez Contreras J. (Spain), Lopez Paz E. (Spain), Luque Otero M. (Spain), Maldonado J. (Portugal), Malik C. (Saudi Arabia), Meaney E. (Mexico), Media Villa Garcia J.D. (Spain), Mendonca I. (Portugal), Middlemost S. (South Africa), Morales D. (Philippine), Myburgh M.P. (South Africa), Naidoo U.P. (South Africa), Naksawasdi C. (Thailand), Nartowicz E. (Poland), Ngamukos K.P. (Thailand), Ocon Pujadas J. (Spain), Okcün B. (Turkey), Olivan Martinez J. (Spain), Ozbakaloglu M. (Turkey), Pham Gia Khai (Vietnam), Pham Nguyen Vinh (Vietnam), Polónia J. (Portugal), Ramos Gonzáles-Serna E. (Spain), Raszcea-Wanic R. (Poland), Rajzer M. (Poland), Redon Mas J. (Spain), Rico Zalba L.A. (Spain), Roca Cusachs A. (Spain), Rodicio Diaz J.L. (Spain), Rodriguez P. (Argentina), Romero Gonzalez R. (Spain), Ruchaud C.A. (Brazil), Salah El Dine M. (Egypt), Salem A.Z. (Egypt), Sarelli P. (South Africa), Sritara P. (Thailand), Sukonthasarn A. (Thailand), Swierblewska E. (Poland), Szczeklik A. (Poland), Sznajderman M. (Poland), Tansuphaswadikul S. (Thailand), Tendera M. (Poland), Tian Hai (Singapore), Tovar Méndez J.L. (Spain), Trivi M.S. (Argentina), Tukarski A. (Poland), Villacampa S. (Spain), Villamil (Argentina), Vo Quy Vinh (Vietnam), Waich S. (Venezuela), Wasfi.I. (Saudi Arabia), Weich D. (South Africa), Wnuck R. (Poland), Wodynska T. (Poland), Yalcinkaya Y. (Turkey), Yildiz M. (Turkey), Zhang W.Z. (Republic of China), Zhu W.L. (Republic of China). |
I. Swearing in of Board Members
Kevin Pituch presented the Oath of Allegiance to Cheryl Crowder and Sonia Flunder-McNair as Board Members representing the City of Toledo.
II. Call to Order and Roll Call
Paul Komisarek, President, called the meeting to order at 8:52 AM. A Roll Call was taken of Board Members for attendance:
Present: Cheryl Crowder, Amr Elaskary, Sonia Flunder-McNair, Ted Kaczorowski, David Karmol, Paul Komisarek, Susan Postal, Johnathon Ross, Barbara Sarantou, Michelle Schultz, Donna Woodson
Excused: Richard Munk
III. Introduction of Guests
Staff and Others: Eric Zgodzinski, DrPH, Tina Stokes, David Welch, Shannon Jones, Jessica Poupard, Tony Maziarz (AFSCME), Jennifer Gottschalk, Mary Apodaca (ONA), Andrea Binz, Brenda Feasel, Angie Bauman, Kevin Pituch (LC Prosecutor), Heather Burnette, Brandon Palinski, Heather Burnette.
Visitor: Louis Filipak and Paola Pedraza, Medical Students, Thomas Gleba, ONA
IV. Additions/Deletions to Agenda
Holding one motion from the Audit/Finance Committee that added the Health Commissioner as an added approver to financial processes; will take back to Committee.
V. Approval of Board of Health Meeting Minutes
(Resolution 25-23) A motion was made by Ted Kaczorowski and seconded by Dr. Woodson to approve the minutes of the January 26, 2023 Board of Health Meeting as presented. Motion carried; 11 Yeas, 0 Nays.
VI. Legal Update – Kevin Pituch
Kevin Pituch introduced himself to the new Board Members and gave a brief synopsis of two items of note to them that pertain to Public Office:
- Public Records Statute
- Any document that you create or receive that concerns the business of the Health Department is public record. That means that you will have to save it, store it, and produce it if asked.
Anyone can make a public records request.
Kevin’s suggestion is to use Lucas County email (if provided). If a records request is made related to your emails, they will be easier to find.
Kevin also prefers to answer all questions via email so that there is a written record of the question and the answer.
- **Open Meeting Statute**
- All meetings are open to the public with the exception of Executive Session.
- Notice must be given of the meetings and minutes of the meetings have to be produced in a timely fashion.
- **City of Toledo Lead Ordinance**
- 4th Version of the Ordinance
- Amended complaint against the Ordinance has been filed, Court Conference is set for the afternoon of 2/23/23 and a date will be set for both sides to argue their case.
**VII. Medical Director’s Report – Eric Zgodzinski, DrPH for Daniel Cadigan, MD**
- Working on CLIA-waived certificates for our off-site programs and mobile units.
- CLIA is laboratory certification that ensures that tests are running accurately.
**VIII. Committee Reports**
A. **Personnel Committee – Barbara Sarantou**
The Personnel Committee met on Tuesday, February 14, 2023 and discussed the following items:
- Updated Employee Newsletter
- Looks fantastic, great job by all involved in its creation.
- Perspective Candidate Privacy
- If there are questions or discussion to be had from non-personnel committee members about a candidate for employment, those questions and discussions are to be held for Executive Session.
- Department Realignment
- Reviewed proposed Executive Summary Organizational Chart.
*(Resolution 26-23)* A motion was made by Dr. Ross and seconded by Ted Kaczorowski to approve the Personnel Committee Report. Motion carried; 11 Yeas, 0 Nays.
B. **Audit/Finance Committee – Amr Elaskary**
The Audit/Finance Committee met on Tuesday, February 14, 2023. The Committee will be presenting information this month as they always have but next month they will move towards a more presenting a year over year comparative report. The following items require Board of Health action:
1. **(Resolution 27-23)** A motion was made by Dr. Ross and seconded by David Karmol to approve the payment of the January 2023 vouchers. Motion carried; 11 Yeas, 0 Nays.
2. **(Resolution 28-23)** A motion was made by Dr. Ross and seconded by Michelle Schultz to approve the Transfer of Appropriations for the month of January 2023 in the amount of $350,714.57. Motion carried; 11 Yeas, 0 Nays.
3. **(Resolution 29-23)** A motion was discussed to approve the changes in Revenue Estimates and Expense Appropriations for the month of January 2023 in the amount of $1,107,602.30. Did not vote on motion.
4. **(Resolution 30-23)** A motion was made by Dr. Ross and seconded by Barb Sarantou to approve the 2022 Revised Official Amended Certificate of Estimated Resources (attached). Motion carried; 11 Yeas, 0 Nays.
5. **(Resolution 31-23)** A motion was made by David Karmol and seconded by Dr. Woodson to apply for, accept, and enter into an agreement for CDBG through the City of Toledo. Motion carried; 11 Yeas, 0 Nays.
6. **(Resolution 32-23)** A motion was made by Dr. Ross and seconded by Sue Postal to authorize the Health Commissioner to approve all contracts and expenditures for the 2022 Household Septic System Repair/Replacement Grant. Motion carried; 11 Yeas, 0 Nays.
7. **(Resolution 33-23)** A motion was made by Dr. Ross and seconded by Barb Sarantou to authorize the Health Commissioner to apply for, accept, and enter into a 2023 Water Pollution Control Loan Fund Agreement for the repair and replacement of Home Sewage Treatment Systems. Motion carried; 11 Yeas, 0 Nays.
8. **(Resolution 34-23)** A motion was made by Dr. Woodson and seconded by Sue Postal to authorize the Health Commissioner to approve all contracts and expenditures for the 2023 Household Septic System Repair/Replacement (HSTS) Grant in the amount of $50,000. Motion carried, 11 Yeas, 0 Nays.
9. **(Resolution 35-23)** A motion was made by David Karmol and seconded by Sue Postal to approve the 2024 Proposed Budget and Estimated Resources. Motion carried; 11 Yeas, 0 Nays.
10. **(Resolution 36-23)** A motion was made by Sue Postal and seconded by Dr. Woodson to approve the write-off of uncollectable amounts that are for dates of service at least 180 days or older. Total amount is $2,415.50 for Self-Pay for January 2023. Motion carried; 11 Yeas, 0 Nays.
Amr Elaskary reviewed a new way of presenting financial data to the Board. The new way will focus on year over year comparison and month to month comparison in the three funds that comprise the Health Department budget, General Revenue Fund, Special Revenue Fund – Fees, and Special Revenue Fund – Grants. This will allow us to see trends and how we are performing over time in the funds.
There was discussion about the balances in the three funds and expenses in the Special Revenue Funds that may need to be covered by the General Fund at points during the year. Further discussion was held about potential department expenses that may further decrease the balance in the General Fund.
David Karmol and other Board Members expressed their appreciation for Amr and his willingness to dive into the financials and his ability to understand them and explain them to the Board. Amr thanked the staff that has been willing to work with him on numerous occasions to go over the reports and work with him to develop this new way of presenting the information.
(Resolution 37-23) A motion was made by Sue Postal and seconded by Dr. Ross to approve the Audit/Finance Committee Report. Motion carried; 11 Yeas, 0 Nays.
C. Environmental Health Committee – David Karmol
- There was no Environmental Health Committee meeting in February.
D. Strategic Planning Committee – Dr. Donna Woodson
The Committee met on Tuesday, February 14, 2023. Brandon Palinski gave the following update:
- Board Members and Administration went through many available data sources to inform this round of strategic planning, including the most recent Community Health Assessment (CHA), most recent Community Health Improvement Plan (CHIP), Lucas County Health Rankings data, March of Dimes Report for Pre-Term Births.
- Discussions focused around what areas we really wanted to hone in on when we started to tackle those next priorities.
- Next meeting focus: Continued Data Review and Priorities Discussion.
(Resolution 38-23) Motion was made by Dr. Ross and seconded by Michelle Schultz to approve the Strategic Planning Report. Motion carried; 11 Yeas, 0 Nays.
E. Health Services Committee – Dr. Johnathon Ross
The Health Services Committee met on Tuesday, February 14, 2023. Discussions centered on the following:
- Reproductive Health Clinic
- Working on scheduling issues to allow for more walk-in visits.
- Patients are also seen through the RHWC for STI treatment and those visits are up.
- Adult Travel & Immunization Clinic
- Visits are up due increased travel.
- Discussed possible additional ways to bill for services related to travel.
- Tuberculosis Clinic
- Patient numbers are up.
- Shots for Tots & Teens
- Visits are down, but this program is getting ready to ramp up.
- Reviewed the Infectious Disease Report
- School Asthma Program
- Started with Washington Local and TPS, and recently applied for a grant to extend it to Springfield Local Schools.
- NOSS has partnered with BGSU on a research project to help to grow medicinally-resistant bacteria. BGSU was looking to pursue additional funding from the National
Institute of Health but it was dependent on the ability to grow bacteria from the syringes that we supply. We dropped our first set of syringes off in January, they were able to get bacteria off of the inside of the needle and there is growth. We do not know what type of growth it is, but there is growth.
- 2\textsuperscript{nd} Childhood Wellness Symposium
- Has been scheduled for April 26\textsuperscript{th} tentatively in the morning.
- This will build off of the first symposium by engaging those attendees for further discussion and planning.
(Resolution 39-23) Motion was made by Ted Kaczorowski and seconded by Sonia Flunder-McNair to approve the Health Services Committee Report. Motion carried; 11 Yeas, 0 Nays.
F. Facilities Committee – David Karmol
- There was no Facilities Committee Meeting in February.
IX. Prior Business
- Facility Update (Paul Komisarek)
- Eric, David Welch, and Paul met with Karen Poore from the City of Toledo on a few issues including the Health Department building.
- There will be an assessment of the building by the City of Toledo Building Maintenance Supervisor.
- We are also looking at having an independent group come in and do a more in-depth assessment.
- It is vital that all parties know the costs associated with maintaining and repairing the systems and issues at the current building.
- Eric and Paul met with the Bishop and his advisors to continue the dialogue about the possibility of the Diocese building.
- There was discussion about the timing and potential cost and contract of an independent assessment.
- The final decision was that everything would be run through the Facilities Committee and they would bring things to the Board of Health that required full Board attention (cost and/or contract).
- D.A.C. (Jessica Poupard)
- After internal discussions and feedback from Board Members about making sure that our municipal partners know who the Health Department is, how we spend our time in the jurisdictions, and what services we offer, we have decided to survey the municipalities at the upcoming DAC meeting.
- This survey would gauge what TLCHD programs they would like more information on, what Public Health concerns they are seeing in their specific jurisdictions, and how we can present topics of interest to their constituents (town-hall style or regularly attend council meetings).
- We hope to get a better understanding of what the DAC members want to hear from us so that we may make our communication to them more meaningful.
Once we collect the surveys, we will begin to put together a municipality meeting schedule and provide to Board Members.
- **Liability Insurance Renewal (Eric Zgodzinski)**
- It is the time of year again to renew our Department Liability Insurance. We are changing carriers this year due to changes in coverage. The cost is slightly more expensive but there is extended coverage.
**X. New Business/Points for Consideration – Paul Komisarek**
- Paul would like to create an Ad Hoc Committee for Succession Planning. This committee would focus on making sure that we have redundancy in our top positions that if we lose people to retirement or other reasons, we are covered. Members of this committee would be: Dr. Munk, Dr. Ross, Barb Sarantou, Michelle Schultz, and Ted Kaczorowski.
*(Resolution 40-23)* Motion was made by David Karmol and seconded by Dr. Ross to create the Ad Hoc Succession Planning Committee. Motion carried; 11 Yeas, 0 Nays.
**XI. Health Commissioners Comments:**
- **Department Realignment**
- Eric shared the proposed Executive Summary Organizational Chart as well as the current Executive Summary Organization Chart for comparison.
- This realignment was developed in several rounds with feedback at multipole points from Directors and Supervisors.
- We reexamined all of our programs as they stand now and their potential for growth and attempted to group program areas together that work together on same or similar goals or see clients from the same populations. This will increase efficiency and effectiveness in delivering services to the community.
- We will take this draft to the entire staff and get their feedback, so we may make adjustments as necessary.
- We will not implement all at one time, this will be a phased approached depending on need and funding.
- **Syphilis Billboard Campaign**
- The 2022 Syphilis Billboard Campaign is done and over, all billboards are down.
- Syphilis cases were down for a combination of different reasons.
- **Fatality Review Committees**
- We need to develop some type of media campaign to raise awareness about child fatality in our community, no matter if it’s a safe sleep issue or another issue.
- There is so much money coming into this community for different programs that are available to help people with their infants and any issues they may be having.
- We were approached by Lucas County Children’s Services, the Lucas County Coroner’s Office, and the Lucas County Prosecutor’s Office to work collaboratively on a media campaign to raise awareness about this issue.
• East Palestine Train Derailment
o The recent train derailment and significant public health issues that have been raised as a result has caused us to dust off our disaster plans.
o We will be going back to pre-Covid plans and exercises to ensure that we are ready to respond should a situation like this occur in our community.
o Dr. Ross stated that there has been significant degradation of oversight of the Rail business since 2016 with several key regulations being rolled back.
▪ Should we be contacting our elected representatives and legislators to support legislation that would regulate trains and the potential hazardous materials that they may be hauling through our communities?
(Resolution 41-23) A motion was made by Dr. Woodson and seconded by Sue Postal to adjourn the meeting at 10:42 AM. Motion carried; 11 Yeas, 0 Nays.
Signed By:
Paul J. Komisarek, President
Lucas County Regional Board of Health
Date: 3-24-23
Eric J. Zgodzinski, DrPH, MPH, REHS, CPH
Secretary to the Board of Health
Date: 3-24-23 |
COUNT ONE
(Theft of Trade Secrets)
The Grand Jury charges:
Overview
1. From approximately July 2019 to approximately August 2021, CHEUK FUNG RICHARD HO, the defendant, was a research developer and quantitative trader at a global, quantitative trading firm ("Firm-1"), which trades in equities and other securities in exchanges located in the United States and abroad. Firm-1’s proprietary source code ("Firm-1’s Source Code"), the development of which took years and cost Firm-1 more than one billion dollars, has been the linchpin of Firm-1’s success in these markets. During the period of HO’s employment at Firm-1, Firm-1 took substantial measures to protect the confidentiality of its Source Code. Among other things, Firm-1 limited access to Firm-1’s Source Code to only those individuals, like HO, who needed access to it in connection with the duties of their employment. Employees with access to Firm-1’s Source Code were required to enter into agreements with Firm-1 in which they acknowledged the importance of keeping Firm-1’s Source Code secret and promised to protect the confidentiality of that Source Code throughout their employment—and after their employment
concluded. Firm-1 also implemented numerous physical and network security protocols to prohibit unauthorized access to Firm-1’s Source Code.
2. In or about the spring of 2021, CHEUK FUNG RICHARD HO, the defendant, secretly started his own quantitative trading firm (“Firm-2”), which partnered with one of Firm-1’s competitors (“Firm-3”). While still employed at Firm-1, and while using the nearly complete access to Firm-1’s Source Code afforded to him as a result of that employment, HO stole valuable trade secrets from Firm-1 (the “Stolen Trade Secrets”) for use in developing the source code for Firm-2 (“Firm-2’s Source Code”). The Stolen Trade Secrets included, among other things, some of the very building blocks of Firm-1’s Source Code, known as “Atoms,” as well as some of its predictive formulas, known as “Alphas.” By stealing these trade secrets, HO was able to quickly launch Firm-2 and begin trading successfully.
3. Aware that he had misappropriated Firm-1’s trade secrets—and knowing that this theft would injure Firm-1—CHEUK FUNG RICHARD HO, the defendant, repeatedly lied to Firm-1 about his plans after his employment with Firm-1 concluded. For example, when Firm-1 asked HO about his post-Firm-1 employment plans, HO omitted any mention of the fact that he had started Firm-2 and he misrepresented his affiliation with Firm-3. And once Firm-1 learned that HO had started Firm-2, HO sought to destroy evidence. He directed his employees to delete their internal communications, and further directed them to delete the source code history for Firm-2’s Source Code, a direction that HO’s employees did not follow.
4. One of Firm-2’s employees (“Engineer-1”), who had previously worked at Firm-1, began to see indications in Firm-2’s Source Code that CHEUK FUNG RICHARD HO, the defendant, had misappropriated portions of Firm-1’s Source Code. For example, Engineer-1 saw code in Firm-2’s Source Code that resembled certain portions of Firm-1’s Source Code in name
and function—but HO denied misappropriating Firm-1’s Source Code. In approximately June 2023, Engineer-1 became increasingly concerned that HO had in fact misappropriated portions of Firm-1’s Source Code. Another employee, Engineer-2, shared with Engineer-1 a code directory (the “Directory”) that HO had previously shared with Engineer-2, but refused to share with Engineer-1. The Directory contained an index (the “Index”). When Engineer-1 reviewed the Index, Engineer-1 saw that it contained the names of numerous unique Firm-1 Atoms—the very backbone of Firm-1’s Source Code and trade secrets. Upon this discovery, Engineer-1, Engineer-2, and a third engineer (“Engineer-3”) abruptly quit their employment at Firm-2.
**The Quantitative Trading Industry and Firm-1 Background**
5. Quantitative trading is a type of trading carried out in financial markets, through exchanges such as the Chicago Mercantile Exchange (“CME”) and the Intercontinental Exchange (“ICE”). Typically, quantitative trading involves the extremely rapid execution of a high volume of trades in which the decisions to make those trades are determined by sophisticated computer programs.
6. Various firms, including Firm-1, Firm-2, and Firm-3, engage in quantitative trading. Because this trading is electronic, quantitative trading firms principally compete on the basis of the quality of their research and technology. Because research and technology are at the cornerstone of the success of any quantitative trading firm, it is standard in this industry for firms to employ robust measures to protect the confidentiality of its research and technology. In addition, it is also standard in the quantitative trading industry for firms to require employees to sign restrictive employment agreements that, among other things, prevent an employee from competing with a former employer for a certain period of time, and to require employees to protect the confidentiality of an employer’s research and technology.
7. Firm-1 was founded in approximately 2009. After spending approximately millions of dollars on research and development, Firm-1 began trading in 2010, and after investing additional millions of dollars in research and development, became profitable in approximately 2012. Firm-1 has continued to invest over a billion dollars into its Source Code, which was developed nearly entirely in-house.
8. With respect to Firm-1’s Source Code, Firm-1 has a library of Atoms, which are the building blocks and foundation of Firm-1’s Source Code. Each of these Atoms are given a name, which is often unique. Individual Atoms have specific functions, ranging from computational functions to price-predicting computations. Atoms, in turn, are used to build Alphas, which are Firm-1’s predictive formulas that make price predictions based on real-time market data. Firm-1 builds trading strategies based on various combinations of Alphas to analyze market data and make automated trading decisions.
9. The quality of Firm-1’s Source Code is the foundation of Firm-1’s competitive advantage in the exchanges in which it trades, and its economic value depends on it remaining secret. To protect the secrecy of its Source Code, Firm-1 has implemented numerous measures to protect its Source Code from disclosure to a competitor or the public, including a variety of physical and network security measures to prevent unauthorized access to Firm-1’s trading platform. Firm-1 also requires its employees to sign employment agreements that, among other things, require employees to acknowledge the value of Firm-1’s proprietary information, agree to safeguard it, and agree not to copy or remove confidential or proprietary information upon termination or resignation. Firm-1 also conducts regular trainings for employees on the importance of protecting its trade secrets. For example:
a. For those employees who do have access to Firm-1’s Source Code, Firm-1 employs a variety of physical and network security measures to prevent unauthorized access. For instance, Firm-1 employees who are authorized to access Firm-1’s Source Code are provided with unique login credentials with strict password and multi-factor authentication requirements.
b. Physical access to Firm-1’s offices is limited and staffed with around-the-clock security. Firm-1’s offices have multiple security checkpoints that require use of electronic access cards, and visitors are restricted from entering areas where researchers or programmers are working on trading strategies and code.
c. As a condition of employment, Firm-1 requires its employees to sign employment agreements that, among other things, require employees to acknowledge the value of Firm-1’s proprietary information, agree to safeguard it, and agree not to copy or remove confidential or proprietary information upon termination or resignation. For certain employees, Firm-1 can elect to enforce a non-competition covenant, whereby the employee is barred from seeking employment for a similar role at a company that is competitive with Firm-1. Employees are also required to assign intellectual property rights developed during the course of their employment at Firm-1 to Firm-1. In addition, employees are obligated to report outside business affiliations while employed at Firm-1, and they are required to agree to not solicit Firm-1 employees for employment outside Firm-1 for a certain period of time.
d. During the employee onboarding process, Firm-1 emphasizes the value of its proprietary trading technology and the need for employees to safeguard its confidentiality. Firm-1 also enforces numerous policies that require employees to safeguard confidential information and agree not to disclose, misuse, or misappropriate it. Employees are also required
to attend regular trainings and orientations reminding them of their obligations to safeguard Firm-1’s technology, and they are required to confirm that they understand these obligations.
e. Employees are prohibited from saving Firm-1 documents or computer code on personal electronic devices or storing hard copy materials at their homes.
f. Upon resignation or termination, departing employees receive written confirmation about their continuing obligations to protect Firm-1’s proprietary information.
**HO Joins Firm-1 and Gains Access to Firm-1’s Source Code**
10. In approximately 2019, CHEUK FUNG RICHARD HO, the defendant, interviewed for a position at Firm-1. In connection with his interview process at Firm-1, HO stated in an email, in substance and in part, that the quantitative trading industry “is particularly sensitive to personnel movements and trade secrets.” In approximately April 2019, Firm-1 offered HO a position as a quantitative researcher and developer, which HO accepted. HO was assigned to a team working on certain trading strategies. As a quantitative researcher and developer, HO had nearly complete access to Firm-1’s Source Code.
11. As a condition of his employment with Firm-1, CHEUK FUNG RICHARD HO, the defendant, acknowledged and signed an employment contract that, among other things, included a “Non-Compete, Non-Solicitation, Confidentiality, and Intellectual Property Agreement” (the “Non-Compete Agreement”). The Non-Compete Agreement included the following terms, among others:
a. The Non-Compete Agreement defined “Confidential Information” to include, among other things, information that Firm-1 “deems to be proprietary” or “that is not generally disclosed” or “generally known” to the public, such as, among other things, “trading
strategies,” “research,” “computer programs,” “methods,” “processes,” “proprietary trading systems and models,” “algorithms,” and “trade secrets.”
b. HO “expressly acknowledge[d] the vital importance of Confidential Information in the business of [Firm-1] and agree[d] that all Confidential Information provided to [HO] and developed by [HO] is the sole and exclusive property of [Firm-1], is provided to and held by [HO] in trust; [Firm-1] makes reasonable efforts to keep it confidential; and it has value to [Firm-1].” HO further agreed to “not use Confidential information for [HO’s] personal benefit” or “the benefit of third parties” or to “compete” with Firm-1, and HO acknowledged that these obligations “will continue at least for so long as any item of Confidential Information remains confidential” to Firm-1, “including after the termination of [HO’s] relationship with [Firm-1]” and that “such restrictions may require [HO] to limit [HO’s] activities for employers other than [Firm-1] . . . or in connection with other business endeavors following the termination of [HO’s] relationship with [Firm-1] for any reason.”
c. HO agreed to use the “care needed to protect the Confidential Information entrusted to [HO] appropriate to the sensitive nature of that Confidential Information” and to “not copy or remove Confidential Information from [Firm-1’s] offices, except as is necessary in connection with [HO’s] employment or engagement with” Firm-1.
d. HO agreed to “not disclose Confidential Information to others, except to members and employees of [Firm-1] . . . who have a need to know such information and who are bound by similar confidentiality agreements” with Firm-1.
e. Upon resignation, HO agreed to “return all originals and copies of Confidential Information[.]”
f. HO agreed that any intellectual property that HO developed during the course of his employment at Firm-1 belonged to Firm-1.
g. HO agreed that, during his period of employment with Firm-1, he would not “engage, directly or indirectly” in any competitive activity with any business that competes with Firm-1. HO also agreed that, if Firm-1 elected to enforce its non-compete provision (for up to a 12-month period), that HO would not work for a company that was competitive with Firm-1, in exchange for Firm-1 continuing to pay HO’s salary. HO further agreed not to solicit, for a period of 18 months, any employee of Firm-1 to leave Firm-1 and work for HO.
h. HO agreed that he would “promptly notify [Firm-1] at any time that [HO] contemplate[s] terminating or decide[s] to terminate employment with [Firm-1] or contemplate[s] entering or enter[s] into competition with [Firm-1] (whether or not such action would constitute a violation of [the Non-Compete Agreement], as [Firm-1] may decide at such time to limit, suspend, or terminate [HO’s] employment and/or access to [Firm-1’s] Confidential Information[.]”
12. As part of his employment at Firm-1, CHEUK FUNG RICHARD HO, the defendant, was debriefed on his obligations under his employment contract described herein and attended trainings concerning Firm-1’s policies regarding the protection of its confidential information.
**HO Secretly Starts Firm-2 and Steals Firm-1’s Trade Secrets**
13. In or about the winter and spring of 2021, while still employed at Firm-1, CHEUK FUNG RICHARD HO, the defendant, secretly started Firm-2. Among other things, HO recruited engineers, Engineer-1, Engineer-2, and Engineer-3, to join him at Firm-2. Engineer-1, who had worked at Firm-1 until approximately March 2021, began working at Firm-2 in approximately May 2021.
14. CHEUK FUNG RICHARD HO, the defendant, incorporated Firm-2 in Delaware in approximately May 2021.
15. CHEUK FUNG RICHARD HO, the defendant, sent Firm-1 a notice of resignation on or about July 6, 2021. Firm-1 provided HO with continued access to Firm-1’s Source Code for approximately two additional days to allow HO to close out certain projects that were pending.
16. In approximately June 2021, before he had resigned from Firm-1, CHEUK FUNG RICHARD HO, the defendant, told Engineer-2 to work with HO out of HO’s apartment in Chicago to start developing Firm-2’s Source Code. Engineer-2 travelled to Chicago on or about July 5, 2021, and began working from HO’s apartment on or about July 6, 2021. On or before July 8, 2021—while HO maintained access to Firm-1’s Source Code—HO remotely accessed Firm-1’s systems in Engineer-2’s presence and launched a feature that, among things, visualizes market data (the “Firm-1 Visualizer”). In trainings about protecting proprietary information, Firm-1 used the Firm-1 Visualizer as a specific example of confidential information at Firm-1. HO told Engineer-2, in substance and in part, that HO wanted to create something like the Firm-1 Visualizer for Firm-2.
17. In approximately the late summer and fall of 2021, CHEUK FUNG RICHARD HO, the defendant, opened Firm-2’s office in Manhattan. HO travelled to Manhattan frequently, and Engineer-1, Engineer-2, and Engineer-3 worked from the Manhattan office. In approximately the spring of 2022, Firm-2 began trading, initially in Chinese markets, and eventually in CME and ICE, two exchanges on which Firm-1 also trades. Firm-2’s trading strategy was similar to the trading strategy that HO’s team at Firm-1 employed. Firm-2 was revenue positive towards the end of 2022.
18. To develop Firm-2’s Source Code, CHEUK FUNG RICHARD HO, the defendant, copied, without authorization, numerous components of Firm-1’s Source Code (i.e., the Stolen Trade Secrets). The Stolen Trade Secrets consist of portions of Firm-1’s Source Code (including Firm-1’s Atoms and Alphas), which were predominately written in the C++ and Java programming languages, and which Ho translated into the C++ and Python coding languages while preserving the structure, logic, trading insights, and functionality of Firm-1’s Source Code. The Stolen Trade Secrets were then deployed on behalf of Firm-2. HO’s theft of the Stolen Trade Secrets is further evidenced by numerous instances in which HO copied customized and esoteric features of Firm-1’s Code—including coding “comments.” For example, one Alpha in Firm-1’s Source Code contains the comment, “remove old data if it results in predications that are too far off.” HO misappropriated this Alpha from Firm-1’s Source Code and used it in Firm-2’s Source Code, which included a substantively identical comment, “remove data that causes far-away predictions.”
19. In addition, HO maintained the Index, which contained the unique names and parameters of numerous Firm-1 Atoms, the building blocks of Firm-1’s Source Code and trade secrets. HO shared the Directory that included the Index with Engineer-2 during Engineer-2’s July 2021 coding visit to Chicago—but refused to share that Directory with Engineer-1, who had worked at Firm-1.
**HO’s Lies to Firm-1 and Engineer-1 and Seeks to Destroy Evidence**
20. CHEUK FUNG RICHARD HO, the defendant, repeatedly lied to Firm-1 about his employment plans following his exit from Firm-1. After Firm-1 elected to enforce the Non-Compete Agreement with HO, HO was required as part of his employment contract to notify Firm-1 of his employment plans. When making this disclosure in or about July 2022, however, HO did not tell Firm-1 that HO had already started his own quantitative trading firm and he misrepresented
his affiliation with Firm-3. When, in approximately early 2023, Firm-1 learned that HO had started Firm-2, HO lied to Firm-1 again. Among other lies, HO stated falsely that Firm-2 was not engaged in trading and that he had built Firm-2’s Source Code without using anything that HO had learned from Firm-1.
21. At various points in or about approximately 2022 and 2023, Engineer-1, who had previously worked at Firm-1, became concerned that CHEUK FUNG RICHARD HO, the defendant, had stolen trade secrets from Firm-1. For example, Engineer-1 saw a name of a piece of code that resembled code at Firm-1 in both name and function. After confronting HO, HO, among other things, changed the name of the code and implemented a different version of the code that was publicly available. On another occasion, Engineer-1 noticed the name of an Atom that Engineer-1 had himself written at Firm-1. After confronting HO, HO claimed that he took notes of ideas of Firm-1 and was good at remembering names.
22. In approximately the spring of 2023, after Firm-1 confronted CHEUK FUNG RICHARD HO, the defendant, about his founding of Firm-2, HO directed his employees to change the retention settings for Firm-2’s internal messages so that they automatically deleted after a certain period of time. HO also instructed one of his employees to delete private text messages. In addition, HO requested that his employees delete the source code history for Firm-2’s Source Code, which would have shown, among other things, changes that had been made to Firm-2’s Source Code against prior versions of that Source Code, and when those changes had been made.
23. In approximately June 2023, Engineer-2 shared with Engineer-1 the Directory, which CHEUK FUNG RICHARD HO, the defendant, had previously refused to share with Engineer-1. Upon reviewing the Index that was contained in the Directory, Engineer-1 noticed
the names of numerous Firm-1 Atoms. Upon this discovery, Engineer-1, Engineer-2, and Engineer-3 quit their employment at Firm-2.
**STATUTORY ALLEGATIONS**
24. From at least in or about 2019 through in or about January 2025, in the Southern District of New York and elsewhere, CHEUK FUNG RICHARD HO, the defendant, with the intent to convert a trade secret, that is related to a product and service used in and intended for use in interstate and foreign commerce, to the economic benefit of anyone other than the owner thereof, and intending and knowing that the offense would injure any owner of that trade secret, knowingly (i) did steal, and without authorization did appropriate, take, carry away, and conceal, and by fraud, artifice and deception obtain such information; (ii) without authorization did copy, duplicate, sketch, draw, photograph, download, upload, alter, destroy, photocopy, replicate, transmit, deliver, send, mail, communicate, and convey such information; (iii) did receive, buy, and possess such information, knowing the same to have been stolen and appropriated, obtained, and converted without authorization; and (iv) attempted to do the foregoing, to wit, HO stole and converted confidential proprietary computer source code belonging to Firm-1, which were Firm-1’s trade secrets, in order to benefit his own quantitative trading firm that was competing with Firm-1.
*(Title 18, United States Code, 1832(a)(1), (a)(2), (a)(3), (a)(4), and 2.)*
**FORFEITURE ALLEGATION**
25. As the result of committing the offense alleged in Count One of this Indictment, CHEUK FUNG RICHARD HO, the defendant, shall forfeit to the United States, pursuant to Title 18, United States Code, Sections 1834 and 2323(b), any and all articles made or trafficked as part of the commission of said offense; any and all property constituting or derived from any proceeds obtained directly or indirectly as a result of the commission of said offense; and any and all
property used, or intended to be used, in any manner or part to commit or facilitate the commission of said offense, including but not limited to a sum of money in United States currency representing the amount of proceeds traceable to the commission of said offense.
**Substitute Assets Provision**
26. If any of the above-described forfeitable property, as a result of any act or omission of the defendant:
a) cannot be located upon the exercise of due diligence;
b) has been transferred or sold to, or deposited with, a third person;
c) has been placed beyond the jurisdiction of the Court;
d) has been substantially diminished in value; or
e) has been commingled with other property which cannot be subdivided without difficulty;
it is the intent of the United States, pursuant to Title 21, United States Code, Section 853(p) and Title 28, United States Code, Section 2461(c), to seek forfeiture of any other property of the defendant up to the value of the above forfeitable property.
(Title 18, United States Code, Section 2323;
Title 18, United States Code, Section 1834;
Title 21, United States Code, Section 853; and
Title 28, United States Code, Section 2461.)
Foreperson
EDWARD Y. KIM
Acting United States Attorney |
Subd. 5. The council may hold meetings and hearings at the times and places it designates to accomplish the purposes set forth in this act. It shall select a chairman and other officers from its membership as it deems necessary.
Subd. 6. The legislature coordinating commission shall supply the council with necessary staff, office space and administrative services.
Subd. 7. When any person, corporation, the United States government, or any other entity offers funds to the council by way of gift, grant or loan, for the purpose of assisting the council to carry out its powers and duties, the council may accept the offer by majority vote and upon acceptance the chairman shall receive the funds subject to the terms of the offer, but no money shall be accepted or received as a loan nor shall any indebtedness be incurred except in the manner and under the limitations otherwise provided by law.
Sec. 2. APPROPRIATION. There is appropriated from the general fund to the legislative coordinating commission the sum of $95,000 for the period ending June 30, 1978, to pay the expenses incurred by the commission. Notwithstanding Minnesota Statutes, Section 16A.28, or any other law relating to the lapse of an appropriation, the appropriation made by this section shall not lapse but shall continue until June 30, 1978.
Sec. 3. REPEALER. Minnesota Statutes 1974, Section 363.04, Subdivisions 7 and 8 are repealed. This section does not abolish any positions in or affect the complement of the human rights department.
Sec. 4. EFFECTIVE DATE. Sections 1 and 2 of this act shall be effective May 1, 1976 and shall expire June 30, 1978. Section 3 of this act shall be effective July 1, 1976.
Approved April 19, 1976.
CHAPTER 338—H.F.No.2019
An act relating to controlled substances; scheduling certain substances; authorizing notices on prescription drugs when driving may be impaired; clarifying the prohibition against sale or possession of legend drugs by certain persons; authorizing county detoxification centers to purchase and possess legend drugs; amending Minnesota Statutes 1974, Sections 151.37, Subdivision 5; 152.02, Subdivisions 2, 3, 4, and 5; Minnesota Statutes, 1975 Supplement, Section 151.212, Subdivision 2.
BE IT ENACTED BY THE LEGISLATURE OF THE STATE OF MINNESOTA:
Section 1. Minnesota Statutes 1974, Section 152.02, Subdivision 2,
is amended to read:
Subd. 2. CONTROLLED SUBSTANCES; CONTROL AND USE.
The following items are listed in Schedule I:
(1) Any of the following substances, including their isomers, esters, ethers, salts, and salts of isomers, esters, and ethers, unless specifically excepted, whenever the existence of such isomers, esters, ethers and salts is possible within the specific chemical designation: Acetylmethadol; Allylprodine; Alphacetylmethadol; Alphameprodine; Alphamethadol; Benzethidine; Betacetylmethadol; Betameprodine; Betamethadol; Betaprodine; Clonitazene; Dextromoramide; Dextrorphan; Diampromide; Diethylambutene; Dimenoxadol; Dimepheptanol; Dimethylambutene; Dioxaphetyl butyrate; Dipipanone; Ethylmethylthiambutene; Etonitazene; Etoxeridine; Furethidine; Hydroxypethidine; Ketobemidone; Levomoramide; Levophenacylmorphan; Morpheridine; Noracymethadol; Norlevorphanol; Normethadone; Norpipanone; Phenadoxone; Phenampromide; Phenomorphan; Phenoperidine; Piritramide; Proheptazine; Properidine; Racemoramide; Trimeperidine.
(2) Any of the following opium derivatives, their salts, isomers and salts of isomers, unless specifically excepted, whenever the existence of such salts, isomers and salts of isomers is possible within the specific chemical designation: Acetorphine; Acetyldihydrocodeine; Acetylcodone; Benzylimorphine; Codeine methylbromide; Codeine-N-Oxide; Cyprenorphine; Desomorphine; Dihydromorphine; Etorphine; Heroin; Hydromorphinol; Methyldesorphine; Methylhydromorphine; Morphine methylbromide; Morphine methylsulfonate; Morphine-N-Oxide; Myrophine; Nicocodeine; Nicomorphine; Normorphine; Pholcodine; Thebacon.
(3) Any material, compound, mixture or preparation which contains any quantity of the following hallucinogenic substances, their salts, isomers and salts of isomers, unless specifically excepted, whenever the existence of such salts, isomers, and salts of isomers is possible within the specific chemical designation: 3,4-methylenedioxyamphetamine; 4-bromo-2,5-dimethoxyamphetamine; 2,5-dimethoxyamphetamine; 4-methoxyamphetamine; 5-methoxy-3,4-methylenedioxyamphetamine; Bufotenine; Diethyltryptamine; Dimethyltryptamine; 3,4,5-trimethoxyamphetamine; 4-methyl-2,5-dimethoxyamphetamine; Ibogaine; Lysergic acid diethylamide; marijuana; Mescaline; N-ethyl-3-piperidyl benzilate; N-methyl-3-piperidyl benzilate; Psilocybin; Psilocyn; Tetrahydrocannabinols 1-[1-(2-thienyl)cyclohexyl] piperidine.
(4) Peyote, providing the listing of peyote as a controlled substance in schedule I does not apply to the nondrug use of peyote in bona fide religious ceremonies of the Native American Church, and members of the Native American Church are exempt from registration. Any person who manufactures peyote for or distributes peyote to the Native American Church, however, is required to obtain federal regis-
Changes or additions indicated by underline deletions by strikeout
tration annually and to comply with all other requirements of law.
(5) Unless specifically excepted or unless listed in another schedule, any material compound, mixture, or preparation which contains any quantity of the following substances having a depressant effect on the central nervous system, including its salts, isomers, and salts of isomers whenever the existence of such salts, isomers, and salts of isomers is possible within the specific chemical designation:
Mecloqualone
Sec. 2. Minnesota Statutes 1974, Section 152.02, Subdivision 3, is amended to read:
Subd. 3. The following items are listed in Schedule II:
(1) Unless specifically excepted or unless listed in another schedule, any of the following substances whether produced directly or indirectly by extraction from substances of vegetable origin or independently by means of chemical synthesis, or by a combination of extraction and chemical synthesis:
(a) Opium and opiate, and any salt, compound, derivative, or preparation of opium or opiate, including the following: raw opium, opium extracts, opium fluidextracts, powdered opium, granulated opium, tincture of opium, apomorphine, codeine, ethylmorphine, hydrocodone, hydromorphone, metopon, morphine, oxycodone, oxymorphone, thebaine.
(b) Any salt, compound, derivative, or preparation thereof which is chemically equivalent or identical with any of the substances referred to in clause (a), except that these substances shall not include the isoquinoline alkaloids of opium.
(c) Opium poppy and poppy straw.
(d) Coca leaves and any salt, compound, derivative, or preparation of coca leaves, and any salt, compound, derivative, or preparation thereof which is chemically equivalent or identical with any of these substances, except that the substances shall not include decocainized coca leaves or extraction of coca leaves, which extractions do not contain cocaine or ecgonine.
(2) Any of the following opiates, including their isomers, esters, ethers, salts, and salts of isomers, esters and ethers, unless specifically excepted, or unless listed in another schedule, whenever the existence of such isomers, esters, ethers and salts is possible within the specific chemical designation: Alphaprodine; Anileridine; Bezitramide; Dihydrocodeine; Dihydromorphinone; Diphenoxylate; Fentanyl; Isomethadone; Levomethorphan; Levorphanol; Metazocine; Methadone; Methadone - Intermediate, 4-cyano-2-dimethylamino-4, 4-diphenylbutane; Moramide
Intermediate, 2-methyl-3-morpholino-1, 1-diphenyl-propane-carboxylic acid; Pethidine; Pethidine - Intermediate - A, 4-cyano-1-methyl-4-phenylpiperidine; Pethidine - Intermediate - B, ethyl-4-phenylpiperidine-4-carboxylate; Pethidine - Intermediate - C, 1-methyl-4-phenylpiperidine-4-carboxylic acid; Phenazocine; Piminodine; Racemethorphan; Racemorphan.
(3) Unless specifically excepted or unless listed in another schedule, any injectable liquid which contains any quantity of methamphetamine, including its salts, isomers, and salts of isomers: material, compound, mixture, or preparation which contains any quantity of the following substances having a stimulant effect on the central nervous system:
(a) Amphetamine, its salts, optical isomers, and salts of its optical isomers;
(b) Methamphetamine, its salts, isomers, and salts of its isomers;
(c) Phenmetrazine and its salts;
(d) Methylphenidate.
(4) Unless specifically excepted or unless listed in another schedule, any material, compound, mixture, or preparation which contains any quantity of the following substances having a depressant effect on the central nervous system, including its salts, isomers, and salts of isomers whenever the existence of such salts, isomers, and salts of isomers is possible within the specific chemical designation:
(a) Methaqualone
(b) Amobarbital
(c) Secobarbital
(d) Pentobarbital.
Sec. 3. Minnesota Statutes 1974, Section 152.02, Subdivision 4, is amended to read:
Subd. 4. The following items are listed in Schedule III:
(1) Any material, compound, mixture, or preparation which contains any quantity of the following substances having a potential for abuse associated with a stimulant effect on the central nervous system:
(a) Amphetamine, its salts, optical isomers, and salts of its optical isomers;
Changes or additions indicated by underline deletions by strikeout
(b) Phenmetrazine and its salts;
(c) Any substance, except an injectable liquid, which contains any quantity of methamphetamine, including its salts; isomers; and salts of isomers;
(d) Methylphenidate—Any material, compound, mixture, or preparation which contains any quantity of Amphetamine, its salts, optical isomers, and salts of its optical isomers; Phenmetrazine and its salts; Methamphetamine, its salts, isomers, and salts of isomers; Methyphenidate; and which is required by federal law to be labeled with the symbol prescribed by 21 Code of Federal Regulations Section 1302.03 and in effect on February 1, 1976 designating that the drug is listed as a Schedule III controlled substance under federal law.
(2) Any material, compound, mixture, or preparation which contains any quantity of the following substances having a potential for abuse associated with a depressant effect on the central nervous system:
(a) Any compound, mixture, or preparation containing amobarbital, secobarbital, pentobarbital or any salt thereof and one or more other active medicinal ingredients which are not listed in any schedule.
(b) Any suppository dosage form containing amobarbital, secobarbital, pentobarbital, or any salt of any of these drugs and approved by the food and drug administration for marketing only as a suppository.
(c) Any substance which contains any quantity of a derivative of barbituric acid, or any salt of a derivative of barbituric acid, except those substances which are specifically listed in other schedules: Chlorhexadol; Glutethimide; Lysergic acid; Lysergic acid amide; Methylprylon; Phencyclidine; Sulfondiethylmethane; Sulfonethylmethane; Sulfofmethane.
(3) Any material, compound, mixture, or preparation which contains any quantity of the following substances having a potential for abuse associated with a stimulant effect on the central nervous system:
(a) Benzphetamine
(b) Chlorphentermine
(c) Clorterminel
(d) Mazindol
(e) Phendimetrazine.
Changes or additions indicated by underline deletions by strikeout
(3)-(4) Nalorphine.
(4)-(5) Any material, compound, mixture, or preparation containing limited quantities of any of the following narcotic drugs, or any salts thereof:
(a) Not more than 1.80 grams of codeine per 100 milliliters or not more than 90 milligrams per dosage unit, with an equal or greater quantity of an isoquinoline alkaloid of opium.
(b) Not more than 1.80 grams of codeine per 100 milliliters or not more than 90 milligrams per dosage unit, with one or more active, non-narcotic ingredients in recognized therapeutic amounts.
(c) Not more than 300 milligrams of dihydrocodeinone per 100 milliliters or not more than 15 milligrams per dosage unit, with a four-fold or greater quantity of an isoquinoline alkaloid of opium.
(d) Not more than 300 milligrams of dihydrocodeinone per 100 milliliters or not more than 15 milligrams per dosage unit, with one or more active, non-narcotic ingredients in recognized therapeutic amounts.
(e) Not more than 1.80 grams of dihydrocodeine per 100 milliliters or not more than 90 milligrams per dosage unit, with one or more active, non-narcotic ingredients in recognized therapeutic amounts.
(f) Not more than 300 milligrams of ethylmorphine per 100 milliliters or not more than 15 milligrams per dosage unit, with one or more active, non-narcotic ingredients in recognized therapeutic amounts.
(g) Not more than 500 milligrams of opium per 100 milliliters or per 100 grams, or not more than 25 milligrams per dosage unit, with one or more active, non-narcotic ingredients in recognized therapeutic amounts.
(h) Not more than 50 milligrams of morphine per 100 milliliters or per 100 grams with one or more active, non-narcotic ingredients in recognized therapeutic amounts.
Sec. 4. Minnesota Statutes 1974, Section 152.02, Subdivision 5, is amended to read:
Subd. 5. The following items are listed in Schedule JV: Barbital; Chloral betaine; Chloral hydrate; Clordiazepoxide; Clonazepam; Clorazepate; Diazepam; Diethylpropion; Ethchlorvynol; Ethinamate; Fenfluramine; Flurazepam; Mebutamate; Methohexital; Meprobamate except when in combination with the following drugs in the following or lower concentrations: conjugated estrogens, 0.4 mg; tridihexethyl chloride, 25mg; pentaerythritol tetranitrate, 20 mg; Methylphenobarbital; Oxazepam; Paraldehyde; Pemoline; Petrichloral; Phenobarbital; and
Changes or additions indicated by underline deletions by strikeout
Phentermine.
Sec. 5. Minnesota Statutes, 1975 Supplement, Section 151.212, Subdivision 2, is amended to read:
Subd. 2. CONTROLLED SUBSTANCES. In addition to the requirements of subdivision 1, when the use of any drug containing a controlled substance, as defined in chapter 152, or any other drug determined by the board, either alone or in conjunction with alcoholic beverages, may impair the ability of the user to operate a motor vehicle, that fact shall be the board shall require by rule that notice be prominently set forth on the label or container. Rules promulgated by the board shall specify exemptions from this requirement when there is evidence that the user will not operate a motor vehicle while using the drug.
Sec. 6. Minnesota Statutes 1974, Section 151.37, Subdivision 5, is amended to read:
Subd. 5. Nothing in this chapter shall prohibit the sale to, or the possession of, a legend drug by registered drug wholesalers, registered manufacturers, registered pharmacies, any local detoxification centers, licensed hospital or hospitals, bona fide hospitals wherein animals are treated, or licensed pharmacists and licensed practitioners while acting within the course of their practice only.
Approved April 20, 1976.
CHAPTER 339—H.F.No.2144
[Coded in Part]
An act relating to the state transportation system; creating the Minnesota state transportation fund; appropriating money therefrom for construction and reconstruction of bridges and bridge approaches; authorizing the issuance of state bonds therefor pursuant to article XI of the constitution; appropriating money.
BE IT ENACTED BY THE LEGISLATURE OF THE STATE OF MINNESOTA:
Section 1. [174.50] MINNESOTA STATE TRANSPORTATION SYSTEM; FUND. Subdivision 1. State assistance is needed to supplement local effort and the highway user tax distribution fund in financing capital improvements to preserve and develop a balanced transportation system throughout the state. Such a system is a proper function and concern of state government and necessary to protect the safety and personal and economic welfare of all citizens. Immediate needs are reconstruction and replacement of key bridges and approaches to remove obstructions to the flow of traffic on state and county highways, municipal streets and township roads. |
ENDOCYTOSIS OF CATIONIZED FERRITIN
BY SOYBEAN (GLYCINE MAX L.)
PROTOPLASTS
A Thesis
Submitted to the Faculty of Graduate Studies
in Partial Fulfillment of the Requirements
For the Degree of
Doctor of Philosophy
in the
Department of Biology
University of Saskatchewan
Saskatoon, Saskatchewan
by
MICHAEL ALAN TANCHAK
February 1987
The author claims copyright. Use shall not be made of the material contained herein without proper acknowledgement, as indicated on the following page.
The author has agreed that the Library, University of Saskatchewan, shall make this thesis freely available for inspection. Moreover, the author has agreed that permission for extensive copying of this thesis for scholarly purposes may be granted by the professor or professors who supervised the thesis work recorded herein or, in their absence, by the Head of the Department or the Dean of the College in which the thesis work was done. It is understood that due recognition will be given to the author of this thesis and to the University of Saskatchewan in any use of material in this thesis. Copying or publication or any other use of the thesis for financial gain without approval by the University of Saskatchewan and the author's written permission is prohibited.
Requests for permission to copy or to make other use of the material in this thesis in whole or in part should be addressed to:
Head of the Department of Biology
University of Saskatchewan
Saskatoon, Canada
S7N 0W0
Abstract
Protoplasts derived from soybean SB-1 suspension cell cultures were incubated in buffered osmoticum containing cationized ferritin (CF) or native ferritin (NF) to demonstrate the process of endocytosis in plant protoplasts. Uptake of CF occurred rapidly; with as little as a 10 s incubation, CF was found in the cytoplasm in coated vesicles (CVs). After 30 s, CF was also found in small smooth (i.e. non-coated) vesicles. By 2 to 6 min, CF was occasionally detected in three additional organelles; the partially coated reticulum (PCR), dictyosomes and multivesicular bodies (MVBs). This is the first demonstration that the PCR is involved in the process of endocytosis in plants. The PCR consists of a network of interconnected tubular membranes with a coat over part of its cytoplasmic surface. Serial sectioning established that CF was distributed throughout this organelle except for membrane dilations which occasionally contained small internal vesicles. The PCR was frequently, but not always, associated with dictyosomes. Only one example of a direct membrane connection between these organelles was observed.
Longer CF incubations resulted in increased labelling of the PCR, dictyosomes, and MVBs. Incubations of 60 min or longer resulted in the labelling of membranous masses within vacuoles. The membranous masses resembled the contents of some CF-labelled MVBs and are likely derived from them.
Endocytosis in soybean protoplasts was charge dependent since even at high concentrations, very little NF was internalized.
Results from protoplasts post-fixed in zinc iodide-osmium tetroxide (ZIO) indicate that membrane connections exist between the endoplasmic reticulum and dictyosomes. The lack of ZIO reaction product in some PCR is consistent with the proposed independence of this organelle from dictyosomes.
In preliminary experiments, whole SB-1 cells were incubated in NF, CF and lanthanum chloride. Uptake of CF and lanthanum was observed but could not unambiguously be attributed to the occurrence of endocytosis.
A scheme for the endocytotic pathway in soybean protoplasts is presented. The results from protoplasts are compared with those for receptor-mediated endocytosis in animal cells. The implications of this pathway are considered with respect to membrane recycling, molecular repair, degradation of macromolecules and receptor-mediated endocytosis in plant protoplasts and cells.
Acknowledgements
I wish to thank first and foremost my supervisor, Dr. Larry C. Fowke, for the opportunity to work on this project and for allowing me a sufficient degree of freedom to develop the project as I saw fit. I am especially thankful for the high degree of patience, perseverance, and skill displayed by Dr. Fowke while editing the numerous drafts of this thesis and of papers published and in press. I am also thankful for his words of encouragement and sound advice.
I would also like to thank the other members past and present of my supervisory committee, Drs. V.K. Sawhney, G.H. Rank, G. Burkholder, L.R. Griffing and B.G. Mersey for their thoughtful comments and questions. I would especially like to thank Dr. Sawhney for the excellent job he did as acting supervisor during Dr. Fowke's sabbatical leave. I would also like to thank Dr. Griffing, with whom I shared office space, for the numerous stimulating discussions and for use of his technical expertise in some experiments.
As is always the case, the work presented in this thesis could not have been produced without the technical assistance of others. In particular, I would like to thank Ms. Patricia Rennie for her technical assistance, especially with the generation and analysis of serial sections from soybean protoplasts, and for her generous offers to maintain my cell cultures during my absences from the Biology Department. I must also thank Pat for words of
encouragement, sympathy and advice. Excellent technical assistance was also provided by Mr. Yukio Yano (electron microscopy) and Mr. Dennis Dyck (photography). I would also like to thank Dr. J. King and the members of his laboratory for the use of their laboratory facilities. Assistance from Dr. F. Constabel and Mr. K. Pahl of the Plant Biotechnology Institute is gratefully acknowledged.
To my bridge partners, golf partners, soccer and fastball teammates, and friends, who are too numerous to name individually, I thank you for making my stay in Saskatoon more pleasurable and interesting than it would have been otherwise. Special mention and thanks go out to Ms. Rozwitha Marx for coming to the aid of a friend in need.
Financial assistance in the form of a Postgraduate Scholarship (4 years) from the Natural Sciences and Engineering Research Council, Canada (NSERC) is gratefully acknowledged. Research expenses and additional (non-overlapping) financial assistance were provided from an NSERC Operating Grant to Dr. L.C. Fowke.
# Table of Contents
| Section | Page |
|------------------------------------------------------------------------|------|
| Abstract | iii |
| Acknowledgements | v |
| Table of Contents | vii |
| List of Figures | xi |
| List of Abbreviations | xv |
| 1 Introduction | 1 |
| 1.1 General concepts | 1 |
| 1.2 Endocytosis in Animal Cells | 3 |
| 1.3 Endocytosis in Plant Cells | 13 |
| 1.4 Research Objective and Rationale | 16 |
| 2 Materials and Methods | 19 |
| 2.1 Cell Culture | 19 |
| 2.2 Buffers | 19 |
| 2.3 Production of Protoplasts | 22 |
| 2.4 Composition of Cell Wall-degrading Enzyme mixture | 23 |
| 2.5 Cationized Ferritin Incubation and Fixation of Protoplasts | 24 |
| 2.6 Native Ferritin Incubation of Protoplasts | 26 |
| 2.7 Incubation of Whole SB-1 Cells with Cationized and Native Ferritin| 27 |
| 2.8 Incubation of Whole SB-1 Cells with Lanthanum Chloride | 29 |
| 2.9 Standard Post-fixation and Dehydration Protocol | 30 |
| Section | Page |
|------------------------------------------------------------------------|------|
| 2.10 Zinc Iodide-Osmium Tetroxide Post-fixation | 30 |
| 2.11 Uranyl Acetate En Bloc Staining of Protoplasts | 32 |
| 2.12 Embedding Protocol | 33 |
| 2.13 Sectioning and Staining | 34 |
| 2.14 Electron Microscopy and Photography | 35 |
| 3 Results | 36 |
| 3.1 Location of Coated Membranes | 36 |
| 3.2 Selection of Staining Protocols | 39 |
| 3.3 Short Term (10 s to 2 min) Incubation of SB-1 Protoplasts with Cationized Ferritin | 42 |
| 3.4 Intermediate Term Incubation (4 to 30 min) of SB-1 Protoplasts with Cationized Ferritin | 44 |
| 3.5 Long Term (60 to 180 min) Incubation of SB-1 Protoplasts with Cationized Ferritin | 45 |
| 3.6 Ultrastructural Analysis of Organelles Labelled with Cationized Ferritin (Including Serial Sectioning Data) | 46 |
| 3.6.1 Coated Vesicles | 46 |
| 3.6.2 Partially Coated Reticulum | 47 |
| 3.6.3 Dictyosomes | 49 |
| 3.6.4 Multivesicular Bodies | 51 |
| 3.6.5 Vacuoles | 53 |
| 3.7 Effect of Cationized Ferritin Treatment on Protoplast Morphology | 55 |
| 3.8 Effect of Charge on Endocytosis | 56 |
| Section | Page |
|------------------------------------------------------------------------|------|
| 3.9 Results from Zinc Iodide-Osmium Tetroxide Post-fixed Protoplasts | 57 |
| 3.10 Uptake of Markers by Whole SB-1 Cells | 63 |
| 3.10.1 Uptake of Lanthanum Chloride | 63 |
| 3.10.2 Incubations with Native and Cationized Ferritin | 65 |
| 4 Discussion | 97 |
| 4.1 Introductory Remarks | 97 |
| 4.2 Effect of Molecular Charge on Endocytosis | 98 |
| 4.3 Adverse Effects of CF on the Morphology of Protoplasts and Cells | 101 |
| 4.4 Significance of Short Term CF Uptake by Protoplasts | 103 |
| 4.5 Significance of Intermediate Term CF Uptake by Protoplasts | 106 |
| 4.6 Morphology of the Partially Coated Reticulum | 107 |
| 4.7 The PCR and Animal Cell Endosomes | 113 |
| 4.8 Significance of the Dilation in the PCR | 117 |
| 4.9 The MVB of Soybean Protoplasts | 119 |
| 4.10 Formation of MVBs | 121 |
| 4.11 MVBs and the Vacuoles of Soybean Protoplasts | 124 |
| 4.12 Possible Alternative Fate for MVBs | 125 |
| 4.13 Labelling of the Dictyosomes and Membrane Recycling | 126 |
| 4.14 Dictyosomes and Molecular Repair | 128 |
| Section | Page |
|------------------------------------------------------------------------|------|
| 4.15 Membrane Recycling and Membrane Sorting | 129 |
| 4.16 Dictyosome Heterogeneity and Membrane Flow | 131 |
| 4.17 Nature of Internalized Material | 132 |
| 4.18 Overview of the Endocytotic Pathway in Soybean Protoplasts | 134 |
| 4.19 Recent Demonstrations of Endocytosis in Protoplasts | 137 |
| 4.20 Coated Membrane Structures and Endocytosis in Plant Cells | 139 |
| 4.21 Lanthanum and CF Uptake by SB-1 Cells | 142 |
| 4.22 Other Functions for CVs | 146 |
| 4.23 Future Prospects | 146 |
| 4.24 Summary | 148 |
| 5 Literature Cited | 150 |
List of Figures
| Figure | Description | Page |
|--------|-----------------------------------------------------------------------------|------|
| 1 | Diagrammatic representation of the pathway of receptor-mediated endocytosis in animal cells. | 4 |
| 2 | Alternative representation of the pathway of receptor-mediated endocytosis in animal cells. | 12 |
| 3 | Cells from a SB-1 soybean suspension cell culture. | 68 |
| 4 | Protoplasts derived from a SB-1 soybean suspension cell culture. | 68 |
| 5 | Electron micrograph of soybean protoplast. | 68 |
| 6-12 | Coated membranes in soybean protoplasts. | 69 |
| 13 | A MVB. | 70 |
| 14-18 | Comparison of staining protocols. | 70 |
| 19 | Phytoferritin | 70 |
| 20-30 | Results of short term incubations of protoplasts with CF. | 71 |
| 31 | Results of short term incubations of protoplasts with CF (continued) | 72 |
| 32-36 | Results of intermediate term incubations of protoplasts with CF. | 72 |
| 37-39 | Results of intermediate term incubations of protoplasts with CF (continued).| 73 |
| 40-42 | Results of long term incubations of protoplasts with CF. | 74 |
| Figure | Description | Page |
|--------|-------------|------|
| 43. | Results of long term incubations of protoplasts with CF (continued). | 75 |
| 44-45. | Part of a series of sections through a coated vesicular profile. | 75 |
| 46-52. | Serial sections through a CF-labelled coated vesicle. | 75 |
| 53-58. | Ultrastructural details of the PCR. | 76 |
| 59-62. | Serial sections through a PCR. | 77 |
| 63-66. | Serial sections through a PCR (continued from previous plate). | 78 |
| 67-70. | Selected serial sections of a PCR and dictyosomes. | 79 |
| 71-74. | Selected serial sections of a PCR and dictyosome. | 80 |
| 75-78. | Selected serial sections of a PCR and dictyosome (continued from previous plate). | 81 |
| 79-82. | Selected serial sections showing a possible membrane connection between a PCR and dictyosome. | 82 |
| 83-86. | Selected serial sections showing a possible membrane connection between a PCR and dictyosome (continued from previous plate). | 83 |
| 87. | Association of a PCR with the endoplasmic reticulum. | 84 |
| Figure | Description | Page |
|--------|-------------|------|
| 88-91 | Selected serial sections through a CF-labelled dictyosome. | 84 |
| 92 | Peripheral labelling of a dictyosome shown in face view. | 84 |
| 93 | CF-labelling throughout the length of the outermost cisterna of a dictyosome. | 85 |
| 94 | Membrane connection between a dictyosome and the endoplasmic reticulum. | 85 |
| 95-100 | Series of sections through a MVB. | 86 |
| 101-106| Detailed ultrastructure of MVBs. | 87 |
| 107-109| Selected serial sections showing a comparison of a CF-labelled vacuolar membranous mass with the contents of MVBs in the cytoplasm. | 88 |
| 110-113| Selected serial sections of a small CF-labelled vacuole. | 89 |
| 114-116| NF uptake results. | 90 |
| 117 | Low magnification of protoplast post-fixed with ZIO. | 90 |
| 118-119| ZIO post-fixation results for cisternal endoplasmic reticulum. | 91 |
| 120-121| ZIO post-fixation results for dictyosomes. | 91 |
| 122-127| Results of tilting experiments with thick sections of ZIO post-fixed protoplasts. | 92 |
| 128-133| ZIO post-fixation results for dictyosomes using thin (60-80 nm) sections. | 93 |
| Figure | Description | Page |
|--------|-------------|------|
| 134 | ZIO post-fixation results with thin sections showing an unlabelled PCR. | 94 |
| 135-137 | ZIO post-fixation results using thin sections for MVBs. | 94 |
| 138-140 | Results of lanthanum chloride uptake by whole SB-1 cells. | 95 |
| 141-143 | Results of incubating whole SB-1 cells with CF for 60 min. | 95 |
| 144-146 | Results of 14 h incubation of whole SB-1 cells with CF. | 96 |
| 147 | Additional observation of whole SB-1 cells. | 96 |
| 148 | Diagrammatic representation of the endocytotic pathway in soybean protoplasts. | 135 |
List of Abbreviations
cER- cisternal endoplasmic reticulum
CF- cationized ferritin
CP/CPs- coated pit/coated pits
CPB- cell phosphate buffer
CURL- compartment of uncoupling of receptor and ligand
CV/CVs- coated vesicle/coated vesicles
ER- endoplasmic reticulum
GERL- Golgi-associated endoplasmic reticulum which produces lysosomes
MES- 2-(N-morpholino)ethanesulfonic acid
MVB/MVBs- multivesicular body/multivesicular bodies
NF- native ferritin
PCR- partially coated reticulum/partially coated reticula
PPB- protoplast phosphate buffer
tER- tubular endoplasmic reticulum
Tris- Tris(hydroxymethyl) aminomethane
ZIO- zinc iodide-osmium tetroxide
1 Introduction
1.1 General Concepts
Endocytosis is the process by which molecules, particularly macromolecules, and particles from the extracellular environment are internalized into the cell via plasma membrane-derived vesicles. The term, endocytosis, actually encompasses two different but related processes. The first is called phagocytosis and is the process by which large particles or molecular aggregates (> 500 nm in diameter) are internalized (Steinman et al. 1983). The cell actively surrounds and engulfs these particles. Beneath the region of the plasma membrane where the vesicle called a phagosome is forming, there is usually an organelle-free filamentous network that contains actin and probably other contractile proteins (Steinman et al. 1983). Phagocytosis is used extensively by animal cells. An example is the mammalian immune response where certain cell types such as macrophages engulf large molecular aggregates or particles such as whole bacterial cells. The second process, called pinocytosis, involves the internalization of macromolecules or small particles (Steinman et al. 1983). The vesicles involved in this process are very much smaller than phagosomes. Furthermore, the cell does not engulf the material per se but rather the plasma membrane invaginates resulting in the eventual formation of a tiny vesicle or pinosome.
Phagosomes are easily identified because the particles contained therein can be seen by both light and electron microscopy (Steinman et al. 1983). Pinocytotic vesicles can only be detected in the light microscope when their contents have been labelled in a very specific manner which enables detection in specialized optical systems such as the fluorescence light microscope. Even in these cases, the actual nature of the vesicle must be determined through electron microscopy.
A number of terms are commonly used to describe specific types of endocytosis. While the word, endocytosis, is used in these terms, the process involved is usually pinocytosis rather than phagocytosis. These specialized terms include fluid-phase endocytosis, adsorptive endocytosis and receptor-mediated endocytosis. In fluid-phase endocytosis substances in solution or suspension are internalized within plasma membrane-derived vesicles; the substances do not interact directly with the plasma membrane. Adsorptive endocytosis involves the binding of molecules or particles to the surface of the plasma membrane prior to their internalization. Receptor-mediated endocytosis is one form of adsorptive endocytosis. This term is used to describe selective internalization of a particular substance or ligand (usually protein) following its binding to a specific receptor on the plasma membrane (Hermo et al. 1985).
1.2 Endocytosis in Animal Cells
Fig. 1 presents a diagrammatic representation of receptor-mediated (and adsorptive) endocytosis in animal cells (Breitfeld et al. 1985; Wileman et al. 1985; Stahl and Schwartz 1986). It does not represent any one cellular system but is a composite showing the major organelles involved in receptor-mediated endocytosis. The following discussion describes the structure and function of these organelles.
In the last ten years or so, research on pinocytosis in animal cells, especially receptor-mediated endocytosis, has indicated that a morphologically distinct class of vesicle, the coated vesicle (CV), is frequently involved in these processes. The CV can be distinguished from other classes of vesicles by the presence of a protein layer or coat on its cytoplasmic surface. When CVs are isolated and negatively stained, the surface coat has an appearance, in the electron microscope, reminiscent of a soccer ball. The coat consists of a series of interlocking pentagons and hexagons (Kanaseki and Kadota 1969). Biochemical studies on CV-enriched gradient fractions reveal that the coat consists primarily of a protein which has been given the name, clathrin (Pearse 1975). Clathrin is composed of two different classes of polypeptides. The "heavy chain" has a relative molecular mass of 180 KDa (Pearse 1975; Fine and Ockleford 1984) while the "light chains" have a relative molecular mass of 33-38
Fig. 1. Diagrammatic representation of the pathway of receptor-mediated endocytosis in animal cells. Receptor-ligand complexes localized in a coated pit (CP) are internalized when the CP pinches off the plasma membrane to form a coated vesicle (CV). The CV sheds its coat to become a smooth vesicle (SV) which fuses with an endosome. The endosome shown here is known by the acronym, CURL. Receptors and ligands which are recycled become localized within the tubules of CURL. Vesicles derived from these tubules return recycling molecules to the cell surface. These vesicles may originally be coated and then become smooth as they shed their coats. Recycling receptors and ligands may also pass through the Golgi stacks prior to returning to the cell surface. Transport to and from the Golgi is thought to be mediated by vesicles. Receptors and ligands which are to be degraded become localized within a vesicular or dilated region of CURL. This region is thought to transform into a multivesicular body (MVB) and then into a secondary lysosome. In this lysosome, the receptors and ligands are degraded. Composite diagram based on Breitfeld and co-workers (1985), Wileman and co-workers (1985) and Stahl and Schwartz (1986).
GOLGI
CP
CV
SV
ENDOSOME (CURL)
MVB
LYSOSOME
KDa (Ungewickell and Branton 1981; Kirchhausen and Harrison 1981). It appears that there are at least two different types of light chains differing in relative molecular mass by 2 or 3 KDa (Ungewickell and Branton 1981; Kirchhausen and Harrison 1981).
Ligands bind to their specific receptors on the plasma membrane and are internalized via coated pits (CPs) and CVs. CPs are specialized regions of the plasma membrane which have a clathrin coat on their cytoplasmic surfaces. By the process of invagination, these pits are thought to "bud off" the plasma membrane to form free CVs (Kanaseki and Kadota 1969). Coated regions are also found on some internal cellular membranes. In particular, portions of the Golgi apparatus, especially trans reticular Golgi, bear such coats (Willingham and Pastan 1982; Willingham et al. 1984). Other organelles, such as the compartment of uncoupling of receptor and ligand (CURL, Geuze et al. 1984, 1985) and the Golgi-associated endoplasmic reticulum which produces lysosomes (GERL, Novikoff 1976), may also have coats on portions of their cytoplasmic surfaces.
After CVs are formed, they are thought to rapidly shed their clathrin coat and fuse with one another or with organelles known as endosomes (Wileman et al. 1985). Endosomes are heterogeneous in morphology consisting of membranous sacs and associated tubules and more than one type of endosome may exist in a particular cell. Receptors
and/or ligands which are to be degraded are delivered ultimately to lysosomes. Receptors and/or ligands which are recycled back to the cell surface may recycle directly from endosomes such as CURL (e.g. Geuze et al. 1984). In this way, these receptors and ligands avoid exposure to the hydrolytic enzymes present in lysosomes.
Receptor-ligand complexes which experience different fates appear to be internalized via the same CP (Geuze et al. 1984). Therefore sorting of these different complexes from one another must occur within endosomes. It is known that endosomes are acidified membrane compartments (Tycko and Maxfield 1982). This is due to the presence of an ATP-dependent proton pump located within the membrane of the endosome (reviewed by Wileman et al. 1985). Coated vesicles are also thought to have a similar proton pump (reviewed by Wileman et al. 1985). Under acidic pHs, a number of receptor-ligand complexes are known to dissociate (Wileman et al. 1985). Thus the acidic environment of the lumen of endosomes may result in the dissociation of receptor-ligand complexes and in the subsequent targeting of these components to their different sites within the cell. Based on immunocytochemical and cytochemical results, it has been proposed that CURL (e.g. Geuze et al. 1983a, 1984) and/or trans reticular Golgi (Willingham et al. 1984) are the endosomal compartments responsible for these sorting events.
It is conceivable that these two structures are in fact the same organelle which has been given different names by different research groups (Willingham and Pastan 1984a).
The uptake of asialoglycoproteins provides a good example of sorting of ligand and receptor in the endosomal compartment of the cell (Geuze et al. 1983a). In CURL which consists of tubular membranes connected to vesicular profiles, the receptors tend to be localized within the tubular regions while the ligands tend to be located in the vesicular portion. Since the asialoglycoprotein receptors are recycled back to the plasma membrane while their ligands are degraded by the lysosomes, it was suggested that the vesicles derived from the tubular region of CURL would deliver the receptors to the plasma membrane either directly or after passage through the Golgi apparatus. It was also suggested that the vesicular portion of CURL could mature into a second class of endosome called a multivesicular body (MVB). In some cases, MVBs have also been referred to as multivesicular endosomes to emphasize their involvement in endocytosis. The MVB containing the asialoglycoprotein ligands would then fuse with a (primary) lysosome to produce a secondary lysosome where degradation of the ligands would occur (Geuze et al. 1983a).
In hepatocytes, both asialoglycoprotein receptor-ligand complexes and IgA receptor-ligand complexes are internalized (Geuze et al. 1984). In contrast to the situation for
asialoglycoproteins and their receptors, IgA molecules and their receptors do not dissociate upon internalization and neither the receptors nor the ligands are recycled. Instead receptor-ligand complexes are formed at the sinusoidal membrane of rat hepatic parenchymal cells and are ultimately delivered to the bile canaliculi where the receptors are cleaved to release IgA molecules with attached secretory components. This process of transferring ligands from one side of a cell to another is known as transcytosis. During transcytosis, the IgA-IgA receptor complexes become localized within the tubular component of CURL. There is a dramatic microheterogeneity in the distribution of receptors within the tubules of CURL. Therefore the receptors for asialoglycoproteins and the IgA receptor-ligand complexes are located within different regions of the tubules. This result demonstrates that sorting of different receptor-ligand complexes can occur within CURL.
Multivesicular bodies consist of a membrane bounded vacuole containing several smaller internal vesicles. On the cytoplasmic surface of the outer membrane there is frequently an electron dense plaque which has a lamellar appearance (Willingham and Pastan 1984a). In some cases, these organelles can contain hydrolytic enzymes (e.g. Friend and Farquhar 1967; Morales et al. 1985). These organelles are usually thought of as being the last membrane
compartment in the endocytotic pathway before lysosomes. It has been suggested that they may form directly from other endosomes such as CURL via a maturation process and that they themselves are transformed into lysosomes (Geuze et al. 1983a, 1985; Harding et al. 1985).
Another organelle which is frequently involved in membrane traffic associated with receptor-mediated endocytosis is the Golgi apparatus. The Golgi apparatus consists of Golgi stacks, the number of which varies considerably from cell type to cell type. Physical connections may exist between the various Golgi stacks in a cell (Farquhar and Palade 1981). A Golgi stack consists of a set of 3 to 8 stacked flattened membrane sacs referred to as cisternae (Farquhar and Palade 1981). The Golgi apparatus or rather its constituent Golgi stacks tend to be concentrated within one region of animal cells (Farquhar and Palade 1981). The cisternae are not of identical composition and thus a Golgi stack has a distinct polarity. One side of the Golgi stack, the cis side, is thought to receive vesicles containing newly synthesized products from the endoplasmic reticulum (ER). At the other side, the trans side, secretory vesicles and granules are formed from the cisternae (Farquhar and Palade 1981). The Golgi stacks of animal cells are known to contain a number of enzymes (reviewed by Farquhar 1985), many of which appear to function in the post-translational modification of proteins newly
synthesized on the ER. These enzymes are located within specific cisternae of the Golgi stacks. In general, the enzymes involved in early steps of the post-translational modification of proteins are found in the cis cisternae while enzymes involved in later steps tend to be found in trans cisternae. Transport of proteins between cisternae is thought to be mediated by vesicles which form from the edges of the cisternae (Farquhar 1985).
There is evidence that components of the Golgi stacks function in the endocytotic pathway. In particular, vesicles and tubules associated with the trans side of the Golgi stacks (i.e. GERL, trans reticular Golgi and CURL) frequently become labelled during endocytosis (e.g. Willingham et al. 1984). These components seem to be involved with sorting events resulting in the separation of receptors and ligands which are to be recycled to the plasma membrane from those which are to be degraded in lysosomes (Willingham et al. 1984). However, the cisternae are also sometimes labelled and thus seem to be involved in the endocytotic pathway as well. Recent evidence suggests that biosynthetic repair of ligands and membrane constituents may take place in the cisternae of the Golgi apparatus (reviewed by Breitfeld et al. 1985). There is also evidence in secretory cells that components of the secretory granule membrane are recycled to the trans Golgi cisternae.
presumably to be reutilized in the packaging of secretory products (reviewed by Farquhar 1985).
Fig. 1 presents the (receptor-mediated) endocytotic pathway of animals cells as it is commonly thought to occur. However, an alternative pathway (Fig. 2) has been proposed (Willingham et al. 1984) which is radically different from that illustrated in Fig. 1. While this alternative is often widely cited in the literature, it does not appear to be widely accepted. In this pathway, CPs are fixed structures on the plasma membrane which accumulate receptor-ligand complexes. These complexes are then transferred to an adjacent region of the plasma membrane where a smooth surfaced vesicle is forming via the process of invagination (Willingham et al. 1981). This smooth surfaced vesicle is called a receptosome. The receptosomes are then transported towards the center of the cell and during this process they may fuse with one another to form a MVB. Eventually the incoming receptosomes including MVBs fuse with the trans reticular Golgi where sorting of receptors and ligands occur. It is proposed that receptors and ligands which are to be degraded are delivered to lysosomes from the coated regions of the trans reticular Golgi. These coated regions may mirror the function of the CPs of the plasma membrane except that rather than delivering receptors and ligands to receptosomes they deliver them to newly forming lysosomes. Receptors and ligands which are to be recycled to the plasma
Fig. 2. Alternative representation of the pathway of receptor-mediated endocytosis in animal cells. Receptor-ligand complexes localized in a CP are transferred to an adjacent, non-coated region of the plasma membrane. From this region, a smooth vesicle, called a receptosome pinches off the plasma membrane. Typically, the receptosome contains a single small internal vesicle. A receptosome may fuse directly with the trans reticular Golgi (TRG) or may first fuse with other receptosomes to form a MVB (a.k.a. a multivesicular receptosome). The MVB then fuses with the TRG. In the TRG, receptors and ligands to be degraded become localized in a coated tip of a TRG tubule. From this coated region, they are transferred to a newly forming lysosome, possibly in a manner analogous to receptosome formation. In the lysosome, the receptors and ligands are degraded. Receptors and ligands which are recycled become concentrated in a non-coated region of the TRG and are transported to the cell surface in a vesicle derived from this region. Diagram based on Willingham and co-workers (1981, 1984), and Willingham and Pastan (1984a).
CP
TRG
LYSOSOME
MVB
RECEPTOSOME
membrane are thought to do so in vesicles originating from the smooth surfaced regions of the trans reticular Golgi (Willingham et al. 1984; Willingham and Pastan 1984a).
1.3 Endocytosis in Plant Cells
While the occurrence of endocytosis is well documented for animal cell systems, the same is not true for plant cells. However, there is a small body of work which provides evidence for the occurrence of endocytosis in plant cells and protoplasts. Some studies of plant-pathogen interactions indicate that the pathogens can gain entry to the host cells via an endocytotic process involving the invagination of the plant cell plasma membrane (e.g. Newcomb 1976). Furthermore, when plant tissues are incubated in solutions of heavy metal salts, evidence of uptake via plasma membrane-derived vesicles has been obtained (e.g. Wheeler and Hanchey 1971; Wheeler et al. 1972; Robards and Robb 1972, 1974; Peterson et al. 1986). In these examples, uptake is thought to involve the formation of smooth surfaced vesicles from the plasma membrane. The intracellular location of the crystallized heavy metal precipitates usually includes vacuoles of varying sizes (e.g. Wheeler and Hanchey 1971; Wheeler et al. 1972; Robards and Robb 1972; Peterson et al. 1986).
In plant protoplasts, a range of molecules or particles have been internalized by protoplasts. These include ferritin (e.g. Cocking 1966a; Power and Cocking 1970), virus
particles (e.g. Cocking 1966a,b; Honda et al. 1975), polystyrene beads (e.g. Mayo and Cocking 1969; Suzuki et al. 1977), liposomes (e.g. Fukunaga et al. 1983), isolated organelles (e.g. Davey et al. 1976), bacteria, and yeast cells and protoplasts (e.g. Davey and Power 1975). In all cases, the endocytotic process involved the formation of a smooth surfaced vesicle from the plasma membrane. No attempt was made to determine the full extent of the pathway in any of these cases. Effort was concentrated on determining that endocytosis had occurred. In many cases, the demonstration of endocytosis was of dubious physiological significance as agents such as polyethylene glycol or poly-L-ornithine were used to stimulate the binding of particles to the surface of the protoplast and to induce the occurrence of endocytosis via osmotic shock (e.g. Davey and Power 1975; Honda et al. 1975; Davey et al. 1976; Fukunaga et al. 1983). In these cases, the occurrence of endocytosis may solely be the result of these manipulations and therefore, may bear no relationship to the normal functioning of protoplasts. Furthermore, protoplasts are themselves an artificial system and represent a significant physiological deviation from normal turgid plant cells. Therefore, results using plant protoplasts may not be directly applicable to normal turgid plant cells.
The existence of CVs is well documented for plant cells
(for review see Newcomb 1980). Coated vesicles are frequently found in regions where active cell wall synthesis is occurring, for example, at the developing cell plate of dividing cells (Franke and Herth 1974; Nakamura and Miki-Hiroshige 1982) and in the region of the enlarging infection thread during the infection of a plant host by a symbiotic Rhizobium bacterium (Robertson and Lyttleton 1982). As a result the most frequently suggested functions for coated membranes in plants has been exocytosis either of cell wall protein (Bonnett 1969; Robertson and Lyttleton 1982) or cell wall polysaccharide (Nakamura and Miki-Hiroshige 1982), and the transport of new membrane to the plasma membrane (Franke and Herth 1974; Nakamura and Miki-Hiroshige 1982). However, others (e.g. Lucas and Franceschi 1981) have suggested that CPs and CVs may be involved in endocytotic processes.
Coated vesicles have been isolated from suspension cultured tobacco cells (Mersey et al. 1982), soybean protoplasts (Mersey et al. 1985) and suspension cultures of carrot cells (Depta and Robinson 1986). In each case, SDS-polyacrylamide gel electrophoresis indicated that a 190 KDa protein co-purifies with CV-enriched fractions. It has been suggested that this protein is the heavy chain of plant clathrin. Polypeptides co-purifying with CV-enriched fractions were also found in the 30-40 KDa range and may constitute the light chains of plant clathrin. Coats are also found on CPs, some cisternae of the dictyosomes (see
terminology below) and on an organelle which is known as the partially coated reticulum (PCR), a tubular membrane system with coats on part of its cytoplasmic surface (Pesacreta and Lucas 1985).
The Golgi apparatus of higher plant cells tends to be organized in a manner somewhat different from its animal cell counterpart. In higher plant cells, it consists of a large number of small stacks of cisternae. Each stack of cisternae is referred to as a dictyosome. The dictyosome is the plant cell equivalent of the Golgi stack found in animal cells. Dictyosomes are distributed throughout the plant cell as opposed to being concentrated in one region of the cell as is frequently the case with Golgi stacks in animal cells. However, like the animal cell Golgi stacks, dictyosomes of plant cells are polarized and have cis and trans sides (Mollenhauer and Morre 1980; Robinson and Kristen 1982).
1.4 Research Objective and Rationale
In this research project, the hypothesis that coated membranes as well as other organelles are involved in endocytosis in plant cells and protoplasts was tested. While there is evidence for the presence of plasma membrane-associated receptors for beta glucans (Yoshikawa et al. 1983), auxins (Jacobs and Gilberts 1983) and abscisic acid (Hornberg and Weiler 1984), no plant plasma membrane
receptor for a specific signal has been isolated and fully characterized (Ralton et al. 1986). Therefore, it was not possible to use a specific receptor-ligand system to demonstrate the occurrence of endocytosis in plant cells and to demonstrate the involvement of coated membranes in this process as has been done in animal cell systems. Instead unmodified or native ferritin (NF) and cationized ferritin (CF) were used as markers for fluid phase and adsorptive endocytosis, respectively. Ferritin is an iron storage protein; its iron-rich central core can be identified in the electron microscope and is quite distinctive in appearance. CF is a chemically modified derivative of NF which has a relatively high isoelectric point such that it has a net positive charge at neutral and acidic pHs. This property is beneficial for studies of adsorptive endocytosis as plant plasma membranes, particularly those of protoplasts, have a net negative surface charge under these conditions (Nagata and Melchers 1978; Griffing et al. 1985; Fowke 1986). The only problem associated with the use of the two types of ferritin as markers for endocytosis is that these molecules are large proteins and therefore do not readily pass through the cell wall. For this reason, protoplasts were used for most experiments. After establishing the existence of an endocytotic pathway in protoplasts, preliminary experiments involving incubations of cells with lanthanum chloride which is a heavy metal salt, NF, and CF were performed in the hope
of demonstrating the occurrence of endocytosis in whole cells.
SB-1 soybean cell suspension cultures were the source of cells and protoplasts used in this research project. This cell culture is a long established, easily maintained cell line and was readily available. Furthermore, a satisfactory protocol for the production of protoplasts from this cell culture had already been developed (Mersey et al. 1985).
Some of the results reported in this thesis have been published (Tanchak et al. 1984).
2 Materials and Methods
2.1 Cell Culture
The soybean cell culture used in all experiments was the SB-1 cell line. (This cell line was originally obtained from Dr. F. Constabel of the Plant Biotechnology Institute, National Research Council, Saskatoon, Sk., Canada.) The SB-1 cell line was maintained in 1-B5 medium (Gamborg, 1982) and was subcultured twice weekly. Subculturing of this cell line entailed the inoculation of a 250 ml DeLong flask containing 50 ml of fresh 1-B5 medium with 10 ml of the previous 3.5 day old culture. The cultures were kept in a culture room on a gyratory shaker (150 rpm) at 28°C, in continuous light (10 umol photons/m²/s from "cool-white" fluorescent lamps; Westinghouse, Pittsburg, PA, USA).
2.2 Buffers
A number of buffer systems were used during the course of this project. The composition of some of these buffers and where relevant a brief description of how they were produced and used is provided here.
Sorb-MES osmoticum contained 0.7 M sorbitol, 3 mM 2-(N-morpholino)ethanesulfonic acid (MES) buffer, 6 mM CaCl₂ and 0.7 mM NaH₂PO₄, pH 5.8. This buffer system was used during the production and isolation of protoplasts (sections 2.3 and 2.4).
Buffered osmoticum consisted of 50 mM sodium phosphate buffer and 350 mM sorbitol, pH between 6.7 and 6.8. It was
made by mixing equal volumes of 100 mM NaH₂PO₄, 100 mM Na₂HPO₄, 1.4 M sorbitol and glass distilled water. This osmoticum was used to wash protoplasts prior to and, where appropriate, following cationized ferritin (CF; horse spleen ferritin coupled with N,N-dimethyl-1,3-propanediamine, in 150 mM NaCl; Sigma Chemical Co., St. Louis MO, USA) or native ferritin (NF; Ferritin Type I from equine spleen, in 150 mM NaCl; Sigma) incubations. It was also the osmoticum in which CF or NF incubation of protoplasts took place and in which glutaraldehyde fixation of protoplasts was performed. The derivatives of this buffered osmoticum were made by substituting an appropriate volume of CF, NF or (50 % v/v) glutaraldehyde solutions for some of the distilled water used to make up the buffered osmoticum. For example, to make up 4 ml of 1 mg/ml CF in buffered osmoticum, the following volumes and solutions were used: 1.0 ml NaH₂PO₄, 1.0 ml Na₂HPO₄, 1.0 ml 1.4 M sorbitol, 0.6 ml distilled water and 0.4 ml of 10.4 mg/ml CF in 150 mM NaCl.
Protoplast phosphate buffer (PPB) contained 50 mM sodium phosphate buffer, pH 6.8, and was produced by mixing 1 volume of 100 mM NaH₂PO₄ with 1 volume of 100 mM Na₂HPO₄ and 2 volumes of glass distilled water. It was used to wash protoplasts following glutaraldehyde fixation, for osmium tetroxide post-fixation of protoplasts and some subsequent washes. The osmium tetroxide post-fixative was made by
substituting 1 volume of 4 % (w/v) osmium tetroxide for one of the 2 volumes of glass distilled water used to make up PPB.
Cell phosphate buffer (CPB) consisted of 25 mM sodium phosphate buffer, pH 6.8, produced by mixing 1 volume of 100 mM NaH₂PO₄ with 1 volume of 100 mM Na₂HPO₄ and 6 volumes of glass distilled water. This buffer system and its derivatives were used in all stages prior to ethanol dehydration (section 2.9) of an experiment where whole SB-1 cells were incubated in CF or NF for a period of 1 h (section 2.7). The derivatives of CPB included CPB plus 1 mM CaCl₂ (Ca-CPB), Ca-CPB with 1mg/ml CF (15 mM NaCl), Ca-CPB with 10 mg/ml NF (15 mM NaCl), Ca-CPB with 15 mM NaCl, Ca-CPB with 3.1 % (v/v) glutaraldehyde, and CPB with 1 % (w/v) osmium tetroxide. These derivatives were made by substituting the appropriate volume(s) of 10 mM CaCl₂, 10.4 mg/ml CF (150 mM NaCl), 100 mg/ml NF (150 mM NaCl), 150 mM NaCl, 50 % (v/v) glutaraldehyde, and 4 % (w/v) osmium tetroxide for some of the distilled water used to make CPB. The pH of Ca-CPB with 3.1 % (v/v) glutaraldehyde had to be adjusted to between 6.7 and 6.8 by the addition of 1 M sodium hydroxide.
MES-Tris buffer contained a mixture of MES and Tris(hydroxymethyl) aminomethane (Trizma Base; Sigma) and was used in all stages prior to ethanol dehydration in an experiment where intact SB-1 cells were incubated in a
solution of lanthanum chloride (LaCl₃). This buffer system was prepared in the following manner. A solution of 20 mM Trizma Base was added dropwise to a volume of 20 mM MES until the pH of the resulting mixture was 5.8. This mixture will be referred to as 20 mM MES-Tris. To produce the final buffer that was used in the experiment (i.e. MES-Tris buffer), a volume of 20 mM MES-Tris was mixed with an equal volume of glass distilled water. At this last step, derivatives of MES-Tris buffer were made by using 40 mM LaCl₃, 50 % (v/v) glutaraldehyde, or 4 % (w/v) osmium tetroxide in place of all or a fraction of the glass distilled water to produce, respectively, 10 mM LaCl₃, 3.1 % (v/v) glutaraldehyde (pH corrected with 1 M sodium hydroxide) and 1 % (w/v) osmium tetroxide in MES-Tris.
2.3 Production of Protoplasts
For the production of protoplasts, a dense cell suspension or slurry was produced by pooling the contents of two or three 3.5 day old cultures into one culture flask. The cells were allowed to settle and most of the medium was decanted leaving only the minimum volume of medium necessary for the pipetting of the cell slurry. Fifteen ml of this slurry, about the equivalent of the contents of one 3.5 day old SB-1 culture, were added to each sterile plastic 100 x 15 mm petri dish containing a cell wall-degrading enzyme mixture (See section 2.4) in 10 ml of a Sorb-MES osmoticum
The petri dishes were then sealed with Parafilm M (American Can Company, Greenwich, CT, USA), covered with aluminum foil and incubated at room temperature, for 12 to 15 h on a gyratory shaker (50 rpm). The resulting protoplast preparation was filtered through glass wool and a nylon screen with an 80 or 85 μm mesh size. The protoplasts were collected by centrifugation (200X g, 4-8 min) and were washed by centrifugation (200X g, 4-6 min) in a wash medium, a 1:1 mixture of culture medium (L-B5) and Sorb-MES osmoticum, pH 5.8.
2.4 Composition of Cell Wall-degrading Enzyme Mixture
The cell wall-degrading enzyme mixture consisted of 1.2 % (w/v) Onozuka R-10 cellulase (Kanematsu-Gosha, Los Angeles, CA, USA), 0.5 % (w/v) Rhozyme HP-150 hemicellulase (Genecor Inc., Corning, NY, USA) and 0.5 % (w/v) pectinase in Sorb-MES osmoticum (section 2.2). Four different pectinases were used in this enzyme mixture. They were: pectinase (fungal; Sigma), pectinase (from Rhizopus sp.; Sigma), Serva pectinase (Feinbiochemica, Heidelberg, FRG) and pectinase (fungal origin; Fluka AG, Chem. Fabrik, Buchs, Switzerland). Onozuka R-10 cellulase was used in the form provided by the supplier. In contrast, the Rhozyme HP-150 hemicellulase and the various pectinases were usually desalted on a Sephadex G-25 column (Pharmacia (Canada) Inc., Dorval, P.Q. Canada) and freeze-dried for storage prior to use. On occasion, crude or non-desalted preparations of
these enzymes were used for protoplast experiments. All enzymes (except the pectinase from *Rhizopus*) were generous gifts from Dr. F. Constabel and Mr. K. Pahl of the Plant Biotechnology Institute, National Research Council, Saskatoon, Sk., Canada.
Immediately prior to use, the cell wall-degrading enzyme mixture was filter-sterilized using a 0.45 μm sterile disposable Millipore filter (Millipore Corp., Bedford, MA, USA).
2.5 Cationized Ferritin Incubation and Fixation of Protoplasts
Prior to CF treatment, protoplasts were washed 2 or 3 times by centrifugation (200X g, 1-4 min) in the buffered osmoticum (section 2.2). Pellets containing approximately 0.15 ml of packed protoplasts were resuspended in 0.4 ml of buffered osmoticum. To commence CF treatment, 0.4 ml of buffered osmoticum containing 1 mg/ml CF (15 mM NaCl) were added to the resuspended protoplasts. For short and intermediate term CF treatments which lasted from 10 s to 30 min, the CF treatment was terminated by the addition of 0.4 ml of ice-chilled buffered osmoticum, pH 6.7-6.8, containing 3.1 % (v/v) glutaraldehyde. For long term CF treatments which lasted from 60 to 180 min, the CF treatment was terminated in the following manner. The protoplasts were washed 3 times by centrifugation (200X g, 1 min) with
buffered osmoticum, over approximately 30 min, and were resuspended in 0.8 ml of buffered osmoticum (In some cases, as indicated in the figure legends, this washing step prior to the addition of glutaraldehyde fixative was omitted). Finally, 0.4 ml of ice-chilled buffered osmoticum containing 3.1 % (v/v) glutaraldehyde were added to the protoplast preparations. All samples were then incubated on ice, with occasional resuspensions, for 1 to 2 h. The samples were pelleted by centrifugation (200X g, 1 min), resuspended in 1.2 ml of buffered osmoticum containing 3.1 % (v/v) glutaraldehyde and incubated at room temperature for 2 to 3 h. Finally, the protoplasts were pelleted by centrifugation (200X g, 1 min) and washed once by centrifugation (200X g, 1 min) in buffered osmoticum. The samples were then postfixed either with osmium tetroxide (See section 2.9) or with a zinc iodide-osmium tetroxide post-fixative (See section 2.10).
Control samples (without CF) were processed in an identical manner to the treated samples. In the controls, the buffered osmoticum containing CF was replaced by either 0.4 ml of buffered osmoticum or 0.4 ml of buffered osmoticum containing 15 mM NaCl.
For specific experiments where significant deviations from the basic protocol occur, the full extent of these deviations will be described in the appropriate figure legends.
2.6 Native Ferritin Incubation of Protoplasts
Incubation of soybean protoplasts with NF was carried out in the same manner as the CF incubations (section 2.5). Two NF solutions were used, 1 mg/ml NF (1.5 mM NaCl) and 20 mg/ml NF (30 mM NaCl) in buffered osmoticum.
The protoplasts were incubated in NF for 30 min (1 mg/ml NF) or 70 min (20 mg/ml NF). The incubations were terminated in different manners. When the 1 mg/ml NF solution was used, the incubation was terminated by the direct addition of 0.4 ml of ice-chilled 3.1% (v/v) glutaraldehyde in buffered osmoticum in the manner described for the short and intermediate term CF incubations (see section 2.5). When the 20 mg/ml NF solution was used, the protoplasts were washed three times and then fixed with glutaraldehyde in the manner described for the long term CF incubations (see section 2.5). Subsequent manipulations were performed as described for the CF incubations (see section 2.5).
For comparison purposes, CF incubations were performed in parallel to the NF incubations. The standard CF solution was used (see section 2.5) and the resulting protoplast preparations were handled in an identical manner to those treated with NF.
As a control, in place of a ferritin solution, some protoplast preparations received 0.4 ml of buffered osmoticum containing either 1.5 mM NaCl or 30 mM NaCl and
then were handled in an identical manner to the preparations that had received the NF solutions.
2.7 Incubation of Whole SB-1 Cells with Cationized and Native Ferritin
Cells from a 1 d old SB-1 cell culture were pelleted (200X g, 2 min) and were washed three times by centrifugation (200X g, 1 min) in Ca-CPB (section 2.2). Cell pellets, approximately 0.15 ml packed cell volume, were resuspended in 0.4 ml of Ca-CPB.
For the ferritin incubations, 0.4 ml of Ca-CPB containing either 1 mg/ml CF (15 mM NaCl) or 10 mg/ml NF (15 mM NaCl) were added to the resuspended cells. As a control, 0.4 ml of Ca-CPB containing 15 mM NaCl were added to a resuspended cell preparation. After an 1 h incubation period, the cells were washed three times with a 1.0 ml volume of Ca-CPB over a period of approximately 40 min. Each cell preparation was resuspended in 0.8 ml of Ca-CPB and 0.4 ml of ice-chilled 3.1 % (v/v) glutaraldehyde in Ca-CPB. After 80 min on ice, the cell samples were transferred to 1.2 ml of 3.1 % (v/v) glutaraldehyde in Ca-CPB at room temperature. After a period of 140 min, each cell preparation was washed briefly with Ca-CPB and then was washed five times on ice over a period of 2 h with CPB. Each cell preparation was post-fixed with osmium tetroxide, dehydrated in ethanol and embedded in Araldite resin according to the protocols described in sections 2.9 and
2.12.
Additional, longer term incubations of cells were performed using CF. For these experiments, LB5 culture medium was used in place of Ca-CPB.
In one experiment, cells from 2 ml of a 1 d old SB-1 culture were resuspended in 2.0 ml of LB5 medium containing 2.5 mg/ml of CF and 37.5 mM NaCl, pH 5.3. As a control, a similar cell sample was resuspended in LB5 medium containing 37.5 mM NaCl. These samples were incubated for 16 h in the dark on a gyratory shaker (50 rpm). (Sterile conditions were maintained during this period of time.) The cell preparations were then washed three times by centrifugation (150X g, 1 min) in LB5 medium and were fixed by resuspending the cells in 2.0 ml of LB5 medium plus 1.0 ml of LB5 medium containing 3.1 % (v/v) glutaraldehyde, pH 5.5. The cell preparations were left on ice for 75 min, transferred to 3.5 ml of 3.1 % (v/v) glutaraldehyde in LB5 medium and were incubated at room temperature for approximately 2 h. The cells were washed briefly in LB5 medium at room temperature and then five times over a period of 2 h on ice with CPB. The cell preparations were then post-fixed, dehydrated and embedded according to the protocols described in sections 2.9 and 2.12.
Another experiment was performed on cells in a manner very similar to the protocol of the preceding experiment.
However, cells for this experiment were obtained from a newly initiated culture and were incubated for 3.5 d in the LB5 medium containing 2.5 mg/ml CF.
2.8 Incubation of Whole SB-1 Cells with Lanthanum Chloride
Cells from a 1 d old SB-1 culture were washed three times in MES-Tris buffer (see section 2.2), pH 5.8. Cell pellets, approximately 0.15 ml packed cell volume, were resuspended in 0.4 ml of MES-Tris buffer. For the lanthanum chloride (LaCl₃) incubation, 0.4 ml of 20 mg/ml LaCl₃ in MES-Tris buffer was added to the resuspended cells for 1 h. As a control, 0.4 ml of MES-Tris buffer without LaCl₃ were added to a second cell preparation. The cells were washed three times with MES-Tris buffer, and were resuspended and fixed in 0.8 ml of MES-Tris buffer and 0.4 ml of ice-chilled 3.1 % (v/v) glutaraldehyde in MES-Tris buffer for 1 h on ice. The cells were then transferred to 1.2 ml of 3.1 % (v/v) glutaraldehyde in MES-Tris buffer at room temperature. After a period of 2 h, the cells were washed once in MES-Tris buffer at room temperature followed by five additional changes of this buffer on ice over a period of 2 h. The cells were post-fixed with 1 % (w/v) osmium tetroxide in MES-Tris buffer for 14 h on ice. The cells were then dehydrated in ethanol according to the protocol in section 2.9 and were embedded in Araldite resin according to the protocol in section 2.12.
2.9 Standard Post-fixation and Dehydration Protocol
Prior to post-fixation in osmium tetroxide, the protoplast preparations (see sections 2.5 and 2.6) were given 5 washes on ice, with PPB, over a period of 2 to 3 h. The samples were then resuspended in 1.2 ml of PPB containing 1 % (w/v) osmium tetroxide and left on ice for 11 to 17 h. Samples were given 2 washes on ice, with glass distilled water, and were dehydrated in ethanol. Ethanol dehydration involved the gradual transfer of the samples from distilled water to absolute ethanol by 10 % (v/v) increments. The samples were then given 2 or 3 additional changes with 100 % (v/v) ethanol and were stored in a -60°C freezer until they could be embedded. Dehydration was performed on ice and each change of solution was facilitated by centrifugation (200X g, 1 min).
For experiments using intact soybean cells, osmium tetroxide post-fixation was performed as described for protoplast preparations except that CPB rather than PPB was used for the washes preceding the post-fixation step and as the buffer system for the 1 % osmium tetroxide solution.
The dehydration protocol for the cell preparations was identical to the protocol described above for protoplast preparations.
2.10 Zinc Iodide-Osmium Tetroxide Post-fixation
Some protoplast preparations were post-fixed in a zinc iodide-osmium tetroxide (ZIO) post-fixative rather than in 1
% osmium tetroxide. The ZIO post-fixative (Harris and Chrispeels 1980) consisted of a mixture of equal volumes of 2 % (w/v) osmium tetroxide and a freshly prepared zinc iodide solution. The zinc iodide solution was made in the following manner. 1.5 g of zinc metal powder (BDH Chemicals Company Ltd., Saskatoon, Sk, Canada) was added to 10 ml of glass distilled water. This mixture was sonicated for 30 s and then 0.5 g of resublimed iodine were added to it. After 5 min of stirring with a magnetic stir bar, the resulting mixture was filtered by gravity through a disk of #2 Whatman filter paper (W & R Balston Ltd., England) and was used immediately.
The ZIO post-fixation was performed in the following manner. Glutaraldehyde-fixed protoplast preparations (see below) were given 5 washes on ice, with glass distilled water, over a period of approximately 2 h. Samples were resuspended in 1.2 ml of ZIO post-fixative. The samples were left in the dark either on ice for 16 h or at room temperature for 1 h. Samples on ice were transferred to a fresh ZIO post-fixative after 8 h of incubation. After ZIO post-fixation, the samples were given 2 or 3 washes on ice, with glass distilled water, dehydrated in ethanol according to the protocol in section 2.9 and embedded according to the protocol in section 2.12.
Some of the protoplast preparations which were postfixed with ZIO had previously been treated with CF and fixed with glutaraldehyde in buffered osmoticum (section 2.5). However, in many cases the protoplasts were freshly isolated but otherwise unmanipulated (section 2.3). Pellets containing approximately 0.15 ml (packed cell volume) of these protoplasts were suspended in 0.8 ml of wash medium, a 1:1 mixture of l-B5 culture medium and Sorb-MES osmoticum, pH 5.8, and 0.4 ml of wash medium containing 3.1 % (v/v) glutaraldehyde. The protoplasts were left on ice for 1 to 2 h and then transferred to 1.2 ml of 3.1 % (v/v) glutaraldehyde in wash medium at room temperature. After 2 to 3 h, the protoplasts were washed briefly with wash medium and then post-fixed with ZIO as described above.
2.11 Uranyl Acetate En Bloc Staining of Protoplasts
In some experiments, protoplast preparations were en bloc stained by the following protocol. Following post-fixation in 1 % (w/v) osmium tetroxide (section 2.9), samples were given 2 washes with distilled water on ice and an additional 1 or 2 washes with distilled water at room temperature. The samples were resuspended in 1.2 ml of freshly prepared 2 % (w/v) uranyl acetate and were placed in a 60°C oven for 3 or 16 h. The preparations were then washed several times with distilled water at room temperature followed by one final distilled water wash on ice. Finally, the samples were dehydrated in ethanol according to the protocol in section 2.9.
2.12 Embedding Protocol
The embedding procedure used in all experiments was the procedure described by Fowke (1982).
Samples were removed from storage at $-60^\circ C$ where they had been placed following dehydration in ethanol (section 2.9) and were given 2 washes in ice-chilled 100% ethanol. Most of the ethanol was decanted and propylene oxide was added dropwise (approximately 5-10 drops per hour) over 6 h, on ice. When the ratio of propylene oxide to ethanol reached approximately 2 to 1, samples were transferred to pure propylene oxide and were left overnight in closed 1.5 ml Eppendorf tubes on ice. The next morning, the samples were brought to room temperature and were given 2 changes of propylene oxide. The samples were transferred to small glass vials. Most of the propylene oxide was decanted and a 1:1 (v/v) mixture of propylene oxide and Araldite resin was added dropwise over approximately 6 h until enough resin was present to ensure that the samples would be covered by resin when the propylene oxide was removed. The vials were covered with aluminum foil lids which had been punched with numerous holes, and were placed in a fume hood to allow the propylene oxide to evaporate overnight. The next morning, samples were transferred to fresh Araldite resin in Beem capsules (J.B. EM Services Inc., Pointe Claire - Dorval, P.Q., Canada). The samples were pelleted within the Beem capsules by
centrifugation (1500X g, 15-30 min) and were allowed to soak in the resin for 24-48 h and then were polymerized by placing the Beem capsules in a 60°C oven for 36-60 h.
The standard mixture of Araldite resin consisted of 12.5 g of Araldite 502 epoxy resin, 9.5 g of dodecenyl succinic anhydride, and 0.35 g of DMP-30 ([2,4,6-tri(dimethylamino-methyl)phenol]). All components of the Araldite resin were obtained from Polysciences Inc., Warrington, PA, USA.
2.13 Sectioning and Staining
Samples embedded in Araldite resin were sectioned with a diamond knife on a Reichert-Jung Ultracut Microtome (C. Reichert Optische Werke AG, Wien, Austria). Thin sections displaying silver or grey interference colours and thick sections which were greater than 200 nm in thickness were picked up on flamed, uncoated, copper 300 or 300 x 75 grids (J.B. EM Services Inc., Pointe Claire, Dorval, P.Q., Canada).
Sections were usually unstained or stained for 7-12 min in Reynolds' lead citrate (Reynolds, 1963) prior to use. On occasion, sections were stained with a saturated uranyl acetate solution in 70% (v/v) ethanol for 20-30 min, followed by post-staining in Reynolds' lead citrate for 5 min. Sections of protoplasts which were en bloc stained with uranyl acetate (section 2.11) did not receive any additional staining.
Serial sections were collected with the use of a
specially designed "third hand" device (Rostgaard 1973) on Notchdot 1 x 2 slot grids (J.B. EM Services Inc., Pointe Claire, Dorval, P.Q., Canada) coated with a formvar and carbon support film. Ribbons containing 10 to 25 consecutive thin sections with silver interference colours were commonly obtained. For the study of coated vesicles, ribbons containing 10 to 15 consecutive thin sections with grey interference colours were collected.
The presence of an electron-dense region on the cytoplasmic surface of an organelle, where some evidence of membrane structure could be distinguished, was the criterion by which presumptive clathrin coats were identified.
2.14 Electron Microscopy and Photography
Routine viewing and photography of thin and thick sections was done using a Philips 300 transmission electron microscope. Where tilting of the specimen was required, sections were viewed in a Philips 410 transmission electron microscope equipped with a variable +/- 60° specimen stage goniometer. For photomicrography, Kodak electron microscope film (Eastman Kodak Co., Rochester, NY, USA) was used. The film was developed using Kodak D-19 developer and Rapid Fixer according to the manufacturer's recommended development procedure.
3 Results
3.1 Location of Coated Membranes
SB-1 soybean suspension culture cells (Fig. 3) and protoplasts (Fig. 4 and 5) derived from them contained a number of organelles which bore a coat on at least a part of their surface. These organelles included coated pits (CPs), coated vesicles (CVs), dictyosomes and the partially coated reticulum (PCR).
Coated pits (CPs) were small invaginations of the plasma membrane (Fig. 5 to 8) which had a layer of material associated with their cytoplasmic surface. They appeared to be distributed in a non-random manner on the plasma membrane. In cross-sections through the plasma membrane, profiles of CPs sometimes appeared clustered into small regions of the plasma membrane (Fig. 6) which were scattered among larger tracts of non-coated membrane. When soybean material was unstained (Fig. 7) or stained only with lead citrate (not shown), these coats were often difficult to see in the electron microscope. In these instances, the coats appeared to be fuzzy regions, apparently lacking substructure, underlying the invaginations of the plasma membrane. In sections of protoplasts stained en bloc with uranyl acetate (Fig. 8) or sections double stained with uranyl acetate and lead citrate (Fig. 6), the coat was heavily stained and in some cases substructure could be seen in the form of striations running perpendicular to the
plasma membrane. In other cases, the coat still appeared to be a fuzzy region which was now quite electron dense. In rare profiles of glancing sections through CPs, the coat was seen to consist of a network of interconnected polygons reminiscent of a honeycomb or soccer ball (Fig. 6). The substructure of the coat was very similar to that typical of the protein coats composed of clathrin and associated proteins that have been found in animal cells.
Coated vesicular profiles or coated vesicles (CVs) were frequently seen in the cytoplasm of soybean cells and protoplasts (Fig. 5, 9, 10). These organelles were about 65 to 100 nm in diameter and consisted of a membrane bound vesicle about 30 to 50 nm in diameter with a coat 15 to 30 nm thick. This coat was identical in appearance to the coat observed on CPs. Again, in unstained sections (not shown) or sections stained only with lead citrate (Fig. 9), the coat had a fuzzy appearance and could be difficult to see. In sections of protoplasts stained en bloc with uranyl acetate (Fig. 10) or in sections double stained with uranyl acetate and lead citrate, this coat was heavily stained and was more distinct in appearance. Under ideal conditions the coat appeared to have a characteristic substructure consisting of striations radiating, at right angles, from the surface of the vesicle (Fig. 10).
The PCR consisted of a set of interconnecting tubular
membranes (Fig. 11) which appeared quite distinct from the cisternal membranes of the dictyosome. The PCR frequently bore a coat at the ends or tips of the tubules and coated vesicular profiles were frequently seen in the vicinity of this organelle (Fig. 11). Coats could also be found in more central locations on PCR tubules where the coated membrane appeared to be either fusing with or budding from the surface of the PCR. The coat was located on the cytoplasmic surface of the PCR and appeared to be identical in appearance to the coat on CPs and CVs. The PCR will be described in greater detail in section 3.6.2.
Dictyosomes also had a coat on part of their surface. The coat was found on the cytoplasmic surface, at the tip or dilated end, of a dictyosomal cisterna (Fig. 12). When the dictyosome was sectioned in face view, coats could sometimes be seen on the ends of tubules which projected from the periphery of the dictyosome cisterna (not shown). The coat seen on the dictyosome tended to be somewhat less distinct in appearance than that seen on CPs, CVs or the PCR (compare Fig. 12 to Fig. 8, 10, 11).
An additional organelle which merits a brief description at this time is the multivesicular body (MVB). It consisted of a large vesicle which contained several smaller internal vesicles (Fig. 13). Associated with the cytoplasmic surface on part of the large vesicle was a plaque somewhat similar but not identical to the coat seen on the organelles.
described previously in this section. A more detailed description of this structure can be found in section 3.6.4.
3.2 Selection of Staining Protocols
A number of different staining protocols were employed during the course of this research project. Unstained sections were frequently used. In such sections, it was comparatively easy to positively identify and visualize markers such as CF (Fig. 14). However, the presence of coats and plaques on organelles was more difficult to establish (Fig. 15). Furthermore, the structural details of these coats and plaques were more difficult if not impossible to visualize in unstained sections (Fig. 14 compared to Fig. 8). Unstained sections were used in situations where it was necessary to unambiguously identify the presence of the various markers; CF, NF and lanthanum. For this reason, unstained sections were used from protoplast preparations which were incubated with CF, or with NF and from cell preparations which had been incubated with CF, NF, or lanthanum chloride. Unstained sections were also used on some occasions for the serial sectioning analysis of organelles labelled with CF (section 3.6).
Some thin sections were stained with lead citrate for 7 to 12 min. This staining protocol resulted in a slight increase in specimen contrast in comparison with unstained sections thus making the detection of coats and plaques
somewhat easier (compare Fig. 16 to Fig. 15). However, this staining protocol sometimes resulted in the formation of small precipitates which could on occasion make the identification of ferritin particles more difficult. Sections stained in this manner were primarily used for protoplast preparations which had been incubated in CF for short periods of time.
Some protoplast preparations were stained en bloc with 2% (w/v) aqueous uranyl acetate (section 2.11) prior to dehydration in ethanol. This staining protocol greatly improved the visibility of coats and plaques on cytoplasmic organelles (Fig. 17), particularly when staining was for a duration of 3 h, and resulted in greatly improved specimen contrast. However, en bloc staining resulted in the formation of many small electron dense precipitates. These precipitates were found on or around most membranous organelles and were especially prominent on the tonoplast of the central vacuole of the protoplasts (Fig. 17). The precipitates tended to be more prominent in preparations stained for 16 h than in those stained for 3 h. These precipitates made the identification of ferritin particles more difficult (Fig. 17). Furthermore, ferritin particles, on average, were much less distinct in appearance suggesting that the ferritin particles, or the material to which they were bound, were stained by this protocol. This staining protocol was used on protoplast preparations which were
primarily used for the serial sectioning analysis of organelles labelled with CF. In these instances, the increased specimen contrast proved beneficial in the delineation of organelles found adjacent to one another in the cytoplasm.
Traditional uranyl acetate staining with lead citrate counterstaining was not used extensively. With this staining protocol, CF particles became less distinct in appearance, probably due to staining with uranyl ions, and the specimen contrast was increased so greatly that it was extremely difficult to visualize ferritin particles (Fig. 18). This staining protocol was not used during serial sectioning analysis as the extra handling that was involved with this protocol tended to result in the breaking of the formvar film on the slot grids. This resulted in the loss of several series of sections which were difficult to obtain. This staining protocol was used on some occasions with thin sections of ZIO post-fixed material (see section 3.9) in the hope of increasing specimen contrast to detect the presence of coats. It was also used on some thin sections of control preparations (i.e. protoplasts which were not exposed to CF or NF) in order to reveal the morphological details of certain organelles such as MVBs.
3.3 Short Term (10 s to 2 min) Incubation of SB-1 Protoplasts with Cationized Ferritin
Before describing the uptake of CF by soybean protoplasts, it should be noted that ferritin, also known as phytoferritin, does occur naturally in both SB-1 culture cells and protoplasts derived from these cells. However, this fact does not complicate the interpretation of CF (or NF) uptake results as phytoferritin was only detected within the plastids of the protoplasts (Fig. 19) and cells. These organelles are not likely to be directly involved in the process of endocytosis.
When freshly isolated protoplasts from SB-1 cell cultures were exposed to CF for 10 s prior to fixation with glutaraldehyde, the plasma membrane was evenly labelled with CF particles. All regions of the plasma membrane were labelled including CPs. Probable stages of coated membrane invagination leading to the formation of free CVs were observed. These stages included very shallow CPs (Fig. 20), more highly invaginated CPs (Fig. 21), CPs with a narrow neck connecting the structure to the protoplast surface (Fig. 22 and 23) and free CVs (Fig. 24). Some of the free CVs were found quite deep within the cytoplasm (Fig. 25). After the 10 s exposure with CF, none of the CF particles were observed in smooth or uncoated vesicles or in deep, uncoated invaginations of the plasma membrane. The gentle undulations typical of the plasma membrane were however
labelled. There was no evidence of the formation of smooth or uncoated vesicles (receptosomes) from the plasma membrane in the vicinity of CPs as observed in some instances in animal cell systems (Willingham et al. 1981).
If protoplasts were exposed to CF for 30 s prior to fixation by glutaraldehyde, labelled CVs were again observed and in some cases were found in the vicinity of other cytoplasmic organelles such as dictyosomes (Fig. 26) and the PCR. At this time point, labelled smooth vesicles (SVs) or rather vesicles lacking a distinct coat were also observed (Fig. 27). Very rarely, at this and subsequent time points (Fig. 28), examples were found of what appeared to be labelled vesicles which only had a partial coat.
After a 2 min incubation of protoplasts with CF, smooth vesicles containing CF particles were quite common and were in fact more numerous than labelled CVs. These smooth vesicles could be found in the vicinity of structures such as dictyosomes (Fig. 29) and the PCR (not shown). The diameter of the smooth vesicles was similar to that of the membrane-bounded compartment of the CVs (Fig. 30), i.e.- the size of a CV that had shed its distinctive coat. The labelled smooth vesicles were observed quite distant from the plasma membrane and could be found in transvacuolar strands. The vesicles did not appear to be attached to any subcellular structure. The mode of movement of these
vesicles to different locations in the cytoplasm could not be identified.
With a 2 min incubation of protoplasts with CF, examples of labelling of structures other than CPs, CVs and smooth vesicles were generally not observed. In the single exception, the structure which was labelled appeared to be a PCR (Fig. 31).
3.4 Intermediate Term Incubation (4 to 30 min) of SB-1 Protoplasts with Cationized Ferritin
With longer incubations of protoplasts with CF, labelling of internal membrane organelles became more frequent. The structures which became labelled were the PCR, dictyosomes and multivesicular bodies (MVBs). For protoplasts incubated with CF for 4 min, labelled organelles other than CPs, CVs and SVs were still very rare. However, when they occurred they appeared, mainly, to be examples of PCR. In fact, for incubations lasting 4 to 16 min in length, the PCR was the organelle that was most frequently labelled with CF (Fig. 32-34). Dictyosomes (Fig. 35) were less frequently labelled during these time points. Label was first observed in a small number of MVBs after a CF incubation of 6 min. In these very rare examples, only one or two particles of CF were observed in the MVB. With longer incubation times, the frequency of labelled MVBs increased but the number of CF particles per organelle remained quite low (Fig. 36). Similar changes in labelling pattern were
observed for the PCR and dictyosomes.
In protoplasts incubated with CF for 30 min., CPs, CVs, SVs, the PCR, MVBs and dictyosomes were all labelled with CF as would be expected. Furthermore, the amount of CF in the PCR (Fig. 37 and 38), dictyosomes and especially MVBs (Fig. 39) was significantly higher than had been encountered at earlier time points.
3.5 Long Term (60 to 180 min) Incubation of SB-1 Protoplasts with Cationized Ferritin
Long term incubations with CF were performed with SB-1 protoplasts in order to determine whether organelles other than CPs, CVs, SVs, dictyosomes, the PCR and MVBs become labelled with CF.
With CF incubations of 60 min or more, the central vacuole of the protoplast was occasionally labelled with CF (Fig. 40 and 41). The CF particles in these vacuoles did not appear to be free within the lumen of the organelle but rather appeared to be bound to the surface of a mass of material. In some cases, this material appeared to be a mass of coalescing vesicles (Fig. 42). An identical distribution of label was also found in smaller vacuolar profiles (Fig. 43).
No other organelles became labelled with CF, even with incubations of 3 h in length. However, it should be noted that there was some variation in the degree of labelling of
the various organelles from one long term experiment to another. In particular, while the PCR and MVBs were usually heavily labelled with CF, the labelling of dictyosomes was quite variable. In some experiments, a few particles of CF were found only in the terminal buds of one or two cisternae while in other experiments the dictyosomes were heavily labelled, with CF in all or most cisternae of the dictyosomes.
3.6 Ultrastructural Analysis of Organelles Labelled with Cationized Ferritin (Including Serial Sectioning Data)
The 3-dimensional morphology of a number of organelles in protoplasts was analysed using serial thin sections. The results of this analysis are presented in the following five sub-sections.
3.6.1 Coated Vesicles
Coated vesicular profiles were found throughout the cytoplasm but were particularly frequent in the region just below the plasma membrane and in the vicinity of dictyosomes and the PCR. Looking at single sections, it was impossible to determine whether these structures were free vesicles or whether the vesicular profiles were actually physically connected to another organelle via a membranous neck or tubule. Serial sectioning was used to determine whether coated vesicular profiles seen in soybean protoplasts were free coated vesicles.
Using thin sections of grey interference colour, it
was usually possible to section completely through a CV in 4 to 6 sections. This range included one section on each side of the series where the CV was clearly not present. (Based on the size of CVs (section 3.1), the thin sections are approximately 25-40 nm in thickness.)
Some of the coated vesicular profiles seen near the plasma membrane (Fig. 44) were found in subsequent sections to be connected to the plasma membrane (Fig. 45) and thus actually constituted CPs. However, in other cases, membrane connections to other organelles were not evident (Fig. 46 to 52). In these instances, the coated vesicular profile likely constituted a true CV.
3.6.2 Partially Coated Reticulum
The PCR was frequently seen in the perinuclear region of the protoplast where a major portion of the cytoplasm was located. However, PCR were also seen in the thin peripheral, cortical cytoplasm and in transvacuolar cytoplasmic strands.
The PCR was frequently but not exclusively seen associated with the dictyosomes of the protoplast. However, there were numerous examples where the PCR did not appear to be associated with any organelles in the cytoplasm (Fig. 53). Frequently, the PCR had a ring configuration (Fig. 54) or a semicircular structure.
Serial sections were cut to determine whether or not the PCR was connected to other membranous structures of the
cytoplasm and to get some idea of the overall size and 3-dimensional shape of the PCR.
As mentioned in section 3.1, the PCR consists of a set of interconnecting tubular membranes with distinct coated regions at the tips of the tubules, or in more central locations. Serial sectioning confirmed the accuracy of this interpretation and also revealed a feature of the PCR which was not always obvious from single sections. Many but not all PCR possessed a membrane dilation. This dilation was often present in only one or two consecutive sections (Fig. 55-57) but could be present in several consecutive sections. In some cases, this dilation appeared to be a large vesicle free within the cytoplasm and only by examining adjacent sections was it possible to determine that it was connected to the PCR. This dilation in the membrane of the PCR never had, on its cytoplasmic surface, a coat of the type seen on the tips of the PCR tubules, CPs or CVs nor a distinct plaque like those seen on MVBs. However, in some instances, one or more internal membrane bounded vesicles were observed within the dilation of the PCR (Fig. 58, also see Fig. 34).
Cationized ferritin particles were generally distributed throughout the tubules of the PCR including the coated tips of this structure (e.g. Fig. 59-66). However, in sharp contrast, the dilations in the PCR including their internal vesicles were never labelled with CF (e.g. Fig. 55-58).
As a general indication of the size of this organelle,
it usually required anywhere from 12 to 20 thin (60 to 70 nm) sections in order to section through a PCR.
The analysis of serial sections was most interesting with respect to connections or associations of the PCR with dictyosomes. For this analysis, an association was considered to exist between these organelles if a PCR was present within 200 nm of a dictyosome and if the only organelles occupying the space between them were SVs, CVs and/or ribosomes. In some cases, there was no evidence of an association between the PCR and dictyosomes (Fig. 59-66). However, often, the PCR was associated with dictyosomes but direct membrane connections between the two organelles were clearly not present (Fig. 67-78). In other such cases, the possibility of a direct membrane connection between the PCR and the dictyosome existed but could not be determined with certainty. However, in a single case, the membrane connection between the PCR and dictyosome appeared to be genuine (Fig. 79-86).
Dictyosomes were not the only organelle with which the PCR was associated. Occasionally, the PCR was either tightly appressed against elements of the ER or may have been connected to this organelle (Fig. 87).
3.6.3 Dictyosomes
Dictyosomes consisted of a set of 4 to 8 stacked cisternae. The cisternae were thin, flattened, fenestrated
sacs which tended to have a slightly dilated rim from which tubules frequently projected into the surrounding cytoplasm. Large dilations (>100 nm in diameter) in the cisternae and associated tubules of the dictyosomes were encountered very infrequently.
As was the case for the PCR, dictyosomes were frequently seen in the perinuclear region of the protoplast, in the thin layer of cortical cytoplasm and in transvacuolar strands.
During intermediate term incubations with CF, dictyosomes became labelled. Only a few particles of CF were observed in the tip or dilated rims of one or two cisternae of a dictyosome (see Fig. 35 and 39). Serial sections of protoplasts which had undergone a 60 min CF incubation were examined in order to determine the full extent of dictyosome labelling. Labelling occurred in several cisternae including those on both sides of the same dictyosome (Fig. 88-91). Thus labelling did not occur exclusively on one side or the other of the dictyosome. Labelling frequently occurred in the slightly dilated ends or rims of the cisternae or in close proximity to these regions. This peripheral labelling was seen in dictyosomes sectioned in face view (Fig. 92). In these cases, CF was seen around the periphery of the cisternae and in the tubules emanating from the periphery of the cisternae but not in the central region of the cisternae. However in at least one situation, cisternae
appeared labelled throughout their entire length (Fig. 93). These cisternae appeared to be the outer-most cisternae on one (trans ?) side of the dictyosome.
The PCR (section 3.6.2) was not the only organelle which may have been connected directly to dictyosomes. The ER was frequently found in close proximity to dictyosomes and in some cases, appeared to be either appressed against a dictyosome or physically connected to it (Fig. 94, also see Fig. 125-129).
3.6.4 Multivesicular Bodies
Multivesicular bodies were recognized as distinct organelles within soybean protoplasts. They consisted of a large membrane bounded vesicle which contained several smaller, internal membrane bounded vesicles. In thin sections, the profiles of the large vesicles generally had a diameter of 250 to 500 nm, although deviations from this range were observed. MVBs usually appeared circular in section, however, oval and irregular profiles were also frequently encountered. The profiles of the smaller, internal vesicles generally had diameters in the range of 40 to 100 nm. Occasionally, profiles of internal vesicles were encountered with diameters greater than 150 nm. Profiles of the internal vesicles usually were circular or oval in shape, however, profiles with irregular shapes were also seen. The density of the internal vesicles was variable.
Some vesicles were very electron-dense while others were relatively electron-lucent. The large vesicles (> 150 nm) were generally electron-lucent. The 40 to 100 nm internal vesicles were usually quite electron-dense although this was not always the case.
Multivesicular bodies were found distributed throughout the cytoplasm including transvacuolar strands and were especially common in the perinuclear region.
Typically, 5 to 9 consecutive thin sections were required in order to section through this organelle (Fig. 95-100). When serial sections of protoplasts labelled for 60 min with CF were examined, CF particles were observed bound to the outer surface of the internal vesicles of the MVB and to the inner surface of the limiting membrane of the MVB (e.g. Fig. 97). The possibility that CF may occur free within the lumen of the MVB could not be totally discounted (e.g. Fig. 99). MVBs from this 60 min CF preparation were more heavily labelled with CF than the MVBs from the intermediate term incubations (section 3.4), suggesting that CF accumulates in this organelle.
Multivesicular bodies frequently bore a plaque on the cytoplasmic surface of their limiting membrane. This plaque was usually seen in 1 to 3 consecutive serial sections. The plaque was quite distinctive in appearance. In some thin sections which were double stained with uranyl acetate and lead citrate, the plaques appeared to have a lamellar
substructure (Fig. 101). This substructure could not be seen in unstained sections (Fig. 102) or in sections of protoplasts which had been en bloc stained with uranyl acetate (Fig. 103). Only rarely were clathrin-like coats observed on MVBs.
In addition to the presence of a plaque, small membranous buds or tubules sometimes protruded from the main body of the MVB (Fig. 104). Also, in some instances, profiles were obtained which appeared to represent the formation of the internal vesicles via an invagination process from the limiting membrane of the MVB (Fig. 105, also see Fig. 13). In some MVBs, the internal vesicles appeared to be poorly resolved as if they were coalescing or disintegrating.
Elements of the ER were often observed in the vicinity of MVBs. In some cases, the ER appeared to be closely appressed against the MVB (Fig. 106, also see Fig. 135). Less frequently, MVBs were seen associated with dictyosomes.
3.6.5 Vacuoles
Labelling of the central vacuole was not observed in protoplasts from short term incubations in CF but was only seen in protoplasts that had been incubated with CF for a minimum of 60 min. Even then, labelling was not seen in all profiles of protoplasts in these preparations but did occur at a readily detectable level, in approximately 5 to 10% of
protoplast profiles in individual sections.
CF particles found in the central vacuole were not free within the lumen but rather were bound to a membranous mass which appeared to contain vesicles coalesing together (see Fig. 42). These membranous masses were usually seen in 3 to 6 consecutive sections and tended to be located at or near the tonoplast. Exactly how the CF particles reached the central vacuole is not clear. However, the size and appearance of the membranous masses were comparable to the contents of some MVBs in the cytoplasm of the same protoplast (Fig. 107-109). Therefore it is possible that the contents of MVBs were transferred to the central vacuole, perhaps by the fusion of a MVB with the tonoplast. The absence of a membrane surrounding the membranous masses supports this idea but examples of such fusion events were not observed. It should be noted, however, that the tonoplast of the central vacuole frequently appeared to be damaged. Therefore the possibility that the presence of CF in the central vacuole may be an artifact could not be totally discounted.
Similar labelling patterns (Fig. 110-113) were observed for the small vacuolar profiles as mentioned in section 3.5. It was possible that some of these vacuoles were connected to the central vacuole.
3.7 Effect of Cationized Ferritin Treatment on Protoplast Morphology
Preparations of protoplasts treated with CF were not identical in appearance to control preparations. Protoplasts incubated in CF were more likely to be partially collapsed resulting in nonspherical profiles in thin sections. At the ultrastructural level, the tonoplast of the central vacuole of protoplasts incubated in CF was not as well preserved by glutaraldehyde fixation. The tonoplast was either vesiculated or had blistered and separated from the surrounding cytoplasm. Here, it is important to note that even in protoplasts from control preparations, some degree of damage to the tonoplast was observed. Frequently, in membranes to which CF particles were bound, the trilamellar appearance of the unit membrane could not be detected. In addition, the cytoplasm of protoplasts incubated in CF tended to be less dense than the cytoplasm of protoplasts from control preparations. The appearance of the protoplasts was suggestive of swelling or partial lysis. Again, this difference was not absolute but rather relative. It is important to note, however, that some protoplasts which had been treated with CF appeared to be comparable to protoplasts from control preparations. An attempt was made to restrict observations to CF-treated protoplasts which were as similar as possible to the protoplasts of the control samples. Despite these differences between the two
preparations, it was apparent that there were no differences between the two types of preparations with respect to the distribution and the appearance of the structures which became labelled with CF.
3.8 Effect of Charge on Endocytosis
The effect of charge on endocytosis in SB-1 protoplasts was determined by comparing the uptake of CF with that of native ferritin (NF), that is, ferritin which had not been chemically modified. NF is thought to be a marker for fluid phase endocytosis (Morales et al. 1985).
In one experiment, protoplasts were incubated with NF or CF, at a final concentration of approximately 0.5 mg/ml, for 30 min prior to the addition of glutaraldehyde (section 2.6). When examined in the electron microscope, NF particles were not readily detectable within organelles in the cytoplasm but a small number of NF particles were bound to the external surface of the plasma membrane. In contrast, uptake of CF was readily detectable with all the appropriate structures being labelled (section 3.4). Protoplasts treated with NF appeared similar to those in control samples. The cytoplasm of these protoplasts was quite electron dense and overall preservation of the material was good.
Because NF has traditionally been considered a marker for fluid phase endocytosis, it was somewhat surprising that NF appeared to be bound to the plasma membrane. Since the
protoplasts were not washed after the ferritin incubation and prior to fixation with glutaraldehyde, it is possible that free ferritin particles were bound to the plasma membrane by the glutaraldehyde. To test this possibility and to determine if NF uptake could be detected by using higher concentrations of NF, the experiment was repeated except with a final NF concentration of approximately 10 mg/ml. Protoplasts were incubated for 70 min followed by 3 washes in the appropriate buffer system prior to fixation with glutaraldehyde (see section 2.6). From this experiment, the following observations were made. First, uptake of NF was observed at low levels in cytoplasmic structures such as MVBs (Fig. 114) and dictyosomes (Fig. 115). Second, NF particles were again observed bound, in small quantities, to the external surface of the plasma membrane (Fig. 116). In comparison, CF uptake at this time was much more extensive and included the labelling of vacuoles as would be expected from the results of sections 3.5 and 3.6.5. Furthermore, the amount of CF bound to the plasma membrane was significantly higher than that observed for the NF treatment, and the preservation of protoplasts was poorer than that observed for protoplasts treated with NF or for control samples.
3.9 Results from Zinc Iodide-Osmium Tetroxide Post-fixed Protoplasts
Zinc iodide-osmium tetroxide (ZIO) post-fixation is a special method which heavily labels or stains specific
membranes within cells or protoplasts. ZIO treatment usually results in heavy labelling of the nuclear envelope, the endoplasmic reticulum, and elements of the Golgi apparatus (dictyosomes). The improved membrane contrast resulting from ZIO treatment enables one to use relatively thick sections (up to approximately 1 um) combined with tilting of the specimen in the electron microscope to determine whether or not direct membrane connections exist between different organelles. Furthermore, ZIO staining results have in the past been used as a criterion for considering the PCR as a structure distinct from the Golgi apparatus (dictyosomes) (Pesacreta and Lucas, 1984).
ZIO post-fixation was used in this project in the hope of demonstrating that dictyosomes and the PCR display different staining properties. Differential staining would support the serial sectioning results (section 3.6.2) which suggested that the PCR was an organelle distinct, in at least some cases, from dictyosomes. It was also hoped that the ZIO post-fixation procedure would allow for the easy visualization of direct membrane connections between different organelles in the cytoplasm.
When ZIO post-fixation was performed for 1 h at room temperature or for 16 h on ice, similar results were obtained. In general, the nuclear envelope (Fig. 117), the tubular (tER)(Fig. 117) and cisternal ER (cER) (Fig. 118 and
119), and dictyosomes (Fig. 120 and 121) were labelled with the ZIO reaction product. However, the labelling was variable. In many instances, tER, cER and the nuclear envelope were not labelled or only very weakly labelled (not shown). Dictyosome labelling was also variable but not in a consistent manner relative to the labelling of other organelles. Also labelled were mitochondria (Fig. 117) and the internal membranes of the plastids (Fig. 117). In general, the plasma membrane, central vacuole and tonoplast were not labelled by this treatment except for the occurrence of rather large electron dense precipitates (not shown) which were likely a nonspecific overreaction product. These rather large precipitates on the plasma membrane and in the central vacuole tended to result in tearing of thin or thick sections in the electron microscope. Smaller precipitates were also found in small vacuoles and in the cytoplasm (Fig. 120). Some of the cytoplasmic precipitates could have been in vesicles but a bounding membrane could not be detected.
In thick (200-400 nm) sections, the ER appeared to be a very complicated network consisting of tER and cER. Tubular ER consisted of long, often branched tubules of variable width (Fig. 117). The tubules frequently connected with cER which consisted of sheets of membrane (Fig. 118 and 119). The labelling of cER by ZIO treatment was often uneven resulting in a "flaky" appearance (Fig. 118). In many cases,
there appeared to be fenestrations or holes in the membranes of the cER (Fig. 119). In some instances, the tER also appeared to have fenestrations (Fig. 122-124). In many thick sections, the tER appeared to be connected to the nuclear envelope (Fig. 117) and dictyosomes (Fig. 122-127). Tilting of the specimen through + or - 30° clarified the association of ER and dictyosome. In many instances, apparent connections between tER and dictyosomes disappeared when the specimen was tilted (Fig. 122-124). The appearance of a tER-dictyosome connection resulted from the overlapping of the images of two structures which were actually located in different planes in the thick section. In other instances, tilting of the specimen through + or - 30° did not result in the separation of the tER from the dictyosome (Fig. 125-127) and it is possible that there was a real, direct membrane connection between the tER and the dictyosome. Using thin sections, which is not a common practice with ZIO post-fixed material, it was possible to find examples where the ER, presumably tER, appeared to be connected to a dictyosome (Fig. 128 and 129). These results support the interpretation based on the thick sections.
In thick and more noticeably in thin sections of protoplasts post-fixed with ZIO, the cisternae of dictyosomes displayed considerable heterogeneity with respect to labelling. In some cases there was a clear
polarity across the dictyosome with the cisternae on one side being heavily labelled while cisternae on the other side of the dictyosome were less intensely labelled (Fig. 130). In other cases, it was one or more cisternae in the middle of the dictyosome which were most intensely labelled (not shown) or, alternatively, all cisternae were labelled with equal intensity. It was reasonably common for one or more cisternae of a dictyosome to be unlabelled by the ZIO post-fixative (not shown). When dictyosomes were seen in face view (i.e. sectioned parallel to stack of cisternae), the increased contrast resulting from ZIO treatment was most beneficial in the visualization of fenestrations (Fig. 131) in the cisternae of the dictyosome. Coats on dictyosomes were difficult to see even when thin sections were stained with uranyl acetate and counterstained with lead citrate. However, for dictyosomes sectioned in face view, coats frequently appeared to be present on some of the membranous buds or tubules (Fig. 132) which extended from the periphery of the cisternae. In addition to the apparent direct membrane connections between some dictyosomes and the tER described above, connections between adjacent dictyosomes were also seen in both thick (not shown) and thin sections (Fig. 131 and 133).
In thick sections of ZIO post-fixed material, it was not possible to identify the PCR. It was not clear whether this problem was caused by lack of staining of the PCR or by the
presence of other ZIO labelled organelles which masked or obscured the presence of the PCR in thick sections. In thin sections, as noted for dictyosomes, the presence of coats was difficult to detect even when the sections were stained with uranyl acetate and counterstained with lead citrate. However, examples of membranes which were probably PCR were found to be unlabelled by the ZIO treatment (Fig. 134). This was the case despite the fact that all dictyosomes in the same protoplast were labelled with ZIO reaction product. However, it was also possible to find examples of membranes which resembled PCR and contained ZIO reaction product. These may have been glancing sections through the periphery of stained dictyosomes.
As with the PCR, it was not possible to identify MVBs in thick sections of material post-fixed with ZIO. However, in thin sections, MVBs were readily identified. Labelling of the MVBs was variable. In some cases, some of the internal vesicles appeared to be labelled with ZIO (Fig. 135 and 136), while in other cases the MVBs appeared to be completely unlabelled (not shown). In protoplasts which had been labelled with CF prior to glutaraldehyde fixation and ZIO post-fixation, it was possible to identify, in thin sections, MVBs which contained both CF particles and ZIO reaction product (Fig. 137).
3.10 Uptake of Markers by Whole SB-1 Cells
Because of the positive results with CF (and to a lesser extent NF) demonstrating the occurrence of endocytosis in protoplasts derived from SB-1 cells, preliminary experiments were performed on whole turgid SB-1 cells in the hope of demonstrating the occurrence of endocytosis in higher plant cells. NF, CF and lanthanum chloride were used as potential markers of endocytosis.
3.10.1 Uptake of Lanthanum Chloride
Lanthanum chloride dissociates into Cl⁻ ions and electron-dense, membrane-impermeant La³⁺ ions. It was hoped that the La³⁺ ion would diffuse through the cell wall, bind electrostatically to and be deposited on negatively-charged sites on the plasma membrane. The labelled plasma membrane would then be internalized via CVs and this process could be followed by observing the distribution of La deposits.
When SB-1 cells were incubated in a final concentration of LaCl₃ of approximately 10 mg/ml for 1 h followed by washing with buffer and fixation with glutaraldehyde (see section 2.8), the following observations were made. The majority of La deposits were located on the external surface of the cell wall where the cell did not abut another cell in the culture and appeared as large numbers of densely packed fine needles (Fig. 138). These needles were also observed deeper within the cell wall although at much lower densities (Fig. 138). Furthermore, some of these needles were seen in
the vicinity of the plasma membrane and may have been in physical contact with it. These needles were also seen in the cell walls separating adjacent cells in the cell clumps found in the SB-1 culture but at a lower concentration than in walls of individual cells (Fig. 139). La deposits were also observed within the cells. These deposits frequently appeared to be within elements of the ER (Fig. 140).
Sometimes the ER appeared to be slightly dilated in the region where the La deposit was found (Fig. 140). In some cases, the La deposits appeared to be present within vesicles (Fig. 138 and 139) which may be elements of the ER seen in cross-section. The labelling within the cytoplasm tended to be located around the periphery of the cell although on occasion could be seen deep within the cytoplasm in the perinuclear region of the cell. The La deposits within the cell were quite distinct and could frequently be found simply by scanning the cell using a low (3,300 to 5,400 X) magnification on the electron microscope. Cells from the control sample which were not exposed to LaCl₃ did not contain deposits comparable to the La deposits seen within the cell wall and cytoplasm of cells treated with LaCl₃. In fact, this was the basis on which lanthanum deposits were identified. (Elemental analysis of the deposits was not performed.)
3.10.2 Incubations with Native and Cationized Ferritin
In the hope of demonstrating the occurrence of endocytosis in whole SB-1 cells, cells from 1 day old SB-1 cultures were incubated in buffers containing NF and CF for 1 h (see section 2.7). For cells incubated in NF, few ferritin particles were distributed along the outer surface of the cell wall. There was no evidence of any uptake by the cell or of any penetration of NF particles through the cell wall. For cells incubated with CF, ferritin particles were present in large numbers along the outer surface of the cell wall (Fig. 141). The increased binding of CF to the cell wall occurred despite the fact that NF was used at 10 times the concentration of CF. CF particles penetrated the cell wall to some degree (Fig. 141). However, CF particles were not observed in the cross walls which separate adjacent cells. In some instances, CF particles even seemed to be bound to the outer surface of the plasma membrane (Fig. 142). Unambiguous labelling of cytoplasmic structures with CF was not observed. However, an interesting ER profile (Fig. 143) was obtained which indicated that direct membrane connections may exist between the ER and the plasma membrane. Combined with the results of the longer term CF incubations of cells, such connections may be of great significance (see Discussion, section 4.21).
When cells from a 1 day old SB-1 culture were incubated in LB5 medium containing 2.5 mg/ml CF for a period of 14
hours, CF particles were seen predominately along the outside surface of the cell wall and examples of CF particles penetrating through the cell wall were observed (Fig. 144-146). Most importantly, CF particles were bound to the plasma membrane and in several instances appeared to be localized within the cytoplasm (Fig. 144-146). Frequently, these examples of cytoplasmic labelling were located adjacent to a region of the cell wall containing CF particles (Fig. 144-146). The labelling of the cytoplasm was not strong and frequently the CF particles appeared to be in elements of the ER (Fig. 145), in large vesicles which likely were elements of the ER (Fig. 146) or free in the cytoplasm (Fig. 144 and 146). The cytoplasm of SB-1 cells was quite dense even in unstained sections. Therefore it is conceivable, in at least some cases, that the CF particles which appeared to be free in the cytoplasm were in fact in vesicles which were obscured by the dense cytoplasm. En bloc staining of the cell preparations did not help to resolve this question as the uranyl precipitates formed using this protocol prevented the unambiguous detection of CF particles within the cytoplasm.
When SB-1 cells from a newly initiated culture were incubated in 1B5 medium containing 2.5 mg/ml CF for a period of 3.5 days, cell growth was inhibited. These cells appeared to be either dead or badly stressed when examined in the
electron microscope.
In this experiment, most of the CF was observed bound to the outer surface of the cell wall. CF was again found part way through the cell wall, bound to the plasma membrane and within the cytoplasm. Much of the cytoplasmic label appeared to be free within the cytoplasm and in some instances appeared to be either bound to the tonoplast or free within the central vacuole.
One final observation regarding SB-1 culture cells should be made. In lanthanum-treated (not shown), ferritin-treated (not shown) and control cells (Fig. 147), invaginations of the plasma membrane were observed. Frequently, a number of membrane bounded vesicles (Fig. 147) were found in the resulting enlarged periplasmic space. It is conceivable that these profiles represent the results of membrane fusion events between MVBs and the plasma membrane.
Fig. 3. Cells from a SB-1 soybean suspension cell culture. Nomarski differential interference optics. Magnification as Fig. 4.
Fig. 4. Protoplasts derived from a SB-1 soybean suspension cell culture. Nomarski differential interference optics. Bar equals 50 μm.
Fig. 5. Electron micrograph of soybean protoplast. Note the presence of dictyosomes (D), coated vesicles (arrows), mitochondria (M), a partially coated reticulum (PCR) and lipid bodies (L) in the cytoplasm and coated pits (arrowheads) on the plasma membrane (PM). Stained with uranyl acetate and counterstained with lead citrate. Bar equals 500 nm. Micrograph reprinted from Mersey et al. (1985), Planta 163: 317-327.
3 4
5
PM L M D PCR
Scale bar: 10 μm
Fig. 6-12. Coated membranes in soybean protoplasts.
Fig. 6. A number of CPs (arrowheads, double arrowhead, arrow) are clustered on one region of the plasma membrane (PM). Note the striated appearance of the coat on some of these CPs (arrows) and the honeycomb appearance of one coat (double arrowhead) which consists of a set of interconnected polygons. Stained with uranyl acetate and counterstained with lead citrate. Bar equals 300 nm. Fig. 7. Note the CP (arrowhead) with an indistinct coat lacking any apparent substructure. Unstained. Bar equals 300 nm. Fig. 8. A CP (arrowhead) bearing a striated coat. Stained en bloc with uranyl acetate. Magnification as Fig. 12. Fig. 9. Note the CV (arrowhead) with an amorphous coat. Stained with lead citrate. Magnification as Fig. 12. Fig. 10. A CV (arrowhead) with a very distinct coat. Note the alternating electron-dense and electron-lucent regions in the coat giving it a striated appearance. Stained en bloc with uranyl acetate. Magnification as Fig. 12. Fig. 11. A section through a PCR. Note the PCR consists of tubular membranes part of which are coated (arrowheads). CVs (arrows) are seen in close proximity to the PCR. Stained en bloc with uranyl acetate. Magnification as Fig. 12. Fig. 12. Dictyosome (D) consisting of a set of 8 stacked cisternae. Note that some of the tips of the cisternae (arrowheads) bear an indistinct, electron-dense coat. Stained en bloc with uranyl acetate. Bar equals 300 nm.
Figures 6-12
Figure 6: Electron micrograph showing the plasma membrane (PM) with numerous small vesicles (arrows) and larger vesicles (arrowheads).
Figure 7: Electron micrograph showing a section of the cell wall (arrowhead).
Figure 8: Electron micrograph showing a section of the cell wall (arrowheads).
Figure 9: Electron micrograph showing a section of the cell wall (arrowhead).
Figure 10: Electron micrograph showing a section of the cell wall (arrowhead).
Figure 11: Electron micrograph showing a section of the cell wall (arrowheads).
Figure 12: Electron micrograph showing a section of the cell wall (arrowheads).
Fig. 13. A MVB. Note the small vesicle (arrow) apparently forming by invagination of the limiting membrane of the MVB. Also note a small profile of a plaque (arrowhead) on the cytoplasmic surface of the limiting membrane of the MVB. Stained en bloc with uranyl acetate. Magnification as Fig. 18.
Fig. 14-18. Comparison of staining protocols. All at the same magnification. Bar in Fig. 18 equals 300 nm. Fig. 14. CP (arrowhead) with an indistinct coat can be seen on the plasma membrane (PM). Note the small, distinct, electron-dense CF particles which line the outer surface of the plasma membrane. 6 min CF incubation, unstained. Fig. 15. Two vesicles (arrowhead, arrow) are labelled with CF. One vesicle (arrowhead) appears to lack a coat. The second vesicle (arrow) may be a CV although due to lack of contrast, it is not possible to clearly distinguish the coat. 6 min CF incubation, unstained. Fig. 16. CV (arrowhead) labelled with CF. Unlike in Fig. 15, the coat on this CV can be clearly distinguished. Also note that the overall contrast of the cytoplasm is increased. (Compare this figure to Fig. 14 and 15.) D-dictyosome. 10 min CF incubation, stained with lead citrate. Fig. 17. CF particles (arrowheads) on the plasma membrane (PM), in smooth vesicles and in a tubule are less distinct in appearance than in unstained sections (e.g. Fig. 14). Also note the presence of electron-dense precipitates (arrows) caused by the staining and the presence of a CV (double arrowhead) with a distinct coat and vesicle. Finally, note that the overall contrast of the cytoplasm is greatly increased especially in comparison with unstained sections (Fig. 14 and 15). T-tonoplast. 60 min CF incubation, stained en bloc with uranyl acetate. Fig. 18. CP (arrow) and CF particles bound to it and other regions (arrowhead) of the plasma membrane are so heavily stained by this staining protocol that they have become quite amorphous in appearance and difficult to distinguish. Also note the greatly increased contrast of the specimen. 60 min CF incubation, stained with uranyl acetate and lead citrate.
Fig. 19. Phytoferritin. Unlabelled soybean protoplast showing a plastid which contains a number of particles of phytoferritin (arrowheads) and a starch granule (SG). Unstained. Bar equals 300 nm.
Fig. 20-30. Results of short term incubations of protoplasts with CF. All stained with lead citrate and at the same magnification. Bar in Fig. 29 equals 300 nm. Fig. 20. CF particles in a shallow CP (arrowhead). 10 s CF incubation. Fig. 21. CF in a deep CP (arrowhead). 10 s CF incubation. Fig. 22. CF in a highly invaginated CP (arrowhead) which is connected to the rest of the plasma membrane by a narrow neck. 10 s CF incubation. Fig. 23. CF in a CV (arrowhead) which is possibly still attached to the plasma membrane. 10 s CF incubation. Fig. 24. A CF-labelled CV (arrowhead). 10 s CF incubation. Fig. 25. A CF-labelled CV (arrowhead) deep within the cytoplasm. Note the presence of the endoplasmic reticulum (ER) between the plasma membrane (PM) and the CV. 10 s CF incubation. Fig. 26. A CF-labelled CV (arrowhead) near a dictyosome (D). 30 s CF incubation. Fig. 27. A CF-labelled SV. 30 s CF incubation. Fig. 28. A CF-labelled vesicle (arrowhead) which appears to only have a partial coat. 2 min CF incubation. Fig. 29. A CF-labelled SV (arrow) near a dictyosome (D). Note the CV (arrowhead) which may be labelled with CF. 2 min CF incubation. Fig. 30. A CF-labelled SV (arrow) and a CF-labelled CV (arrowhead). Note that the SV is approximately the same size as the vesicle proper of the CV. 2 min CF incubation. Fig. 20-26, 28, and 29 are reprinted from Tanchak et al.(1984), Planta 162:481-486.
20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30
PM: Plasma Membrane
ER: Endoplasmic Reticulum
D: Dendrite
Scale bar: 0.5 μm
Fig. 31. Results of short term incubations of protoplasts with CF (continued). A PCR labelled with particles of CF (arrows). Note the coated membranes (arrowheads) which contain CF. 2 min CF incubation. Stained with lead citrate. Magnification as Fig. 36.
Fig. 32-36. Results of intermediate term incubations of protoplasts with CF. Fig. 32. Presumptive PCR tubule (arrow) labelled with CF. Note vesicles (arrowhead, double arrowhead), each appearing to contain a single particle of CF. One vesicle (arrowhead) appears to be coated. 4 min CF incubation, unstained. Magnification as Fig. 36. Fig. 33. Presumptive PCR tubules labelled with CF. One region of the tubules (arrowhead) may be coated. 16 min CF incubation, unstained. Magnification as Fig. 36. Fig. 34. Two PCR labelled with CF. Note the regions of the PCR (arrowheads) which appear to be coated and a CV (double arrowhead) which is labelled with CF. Also note the presence of a dilation (arrow) in one PCR which appears to have an internal vesicle. Incubated with CF for 3 min on ice, followed by 10 min at room temperature. Otherwise, as described in the Materials and Methods (section 2). Unstained. Magnification as Fig. 36. Fig. 35. Dictyosome labelled with CF (arrowhead). 10 min CF incubation, unstained. Bar equals 300 nm. Fig. 36. MVB which is labelled with CF (arrowhead). Note that CF appears to be bound to the outer surface of one or possibly two of the small internal vesicles of the MVB. 8 min CF incubation, unstained. Bar equals 300 nm.
31, 32, 33, 34, 35, 36
Figures 31-36: Electron micrographs showing various structures and details within cells. Arrows indicate specific features or regions of interest.
Scale bars represent 0.5 μm in all figures.
Fig. 37-39. Results of intermediate term incubations of protoplasts with CF (continued). All are unstained and at the same magnification. Bar in Fig. 39 equals 300 nm. Fig. 37. PCR tubule heavily labelled with CF and bearing a coat (arrowhead) at one end. M - mitochondrion. Fig. 38. A branched tubular network of a PCR heavily labelled with CF. A coated membrane profile (arrowhead) may or may not be connected to the PCR. Fig. 39. MVB heavily labelled with CF. Note that one cisterna (arrowheads) of the accompanying dicytosome is also labelled with CF.
37
38
39
Fig. 40-42. Results of long term incubations of protoplasts with CF. Fig. 40. Low magnification of protoplast showing a CF-labelled membranous mass within the central vacuole (V). N - nucleus. Area outlined in black is shown below in Fig. 41. 60 min CF incubation, pre-fixation washes omitted (see section 2.5 of Materials and Methods), unstained. Bar equals 1 μm. Fig. 41. An enlargement from Fig. 40 showing a membranous mass within the central vacuole. 60 min CF incubation, pre-fixation washes omitted (see section 2.5 of Materials and Methods), unstained. Magnification as Fig. 42. Fig. 42. A membranous mass within the central vacuole which is labelled with CF particles. Note that the mass seems to consist of vesicles (arrowheads). 90 min CF incubation, post-fixation with ZIO rather than osmium tetroxide and no additional staining. Bar equals 300 nm.
40
41
42
Fig. 43. Results of long term incubations of protoplasts with CF (continued). A small vacuole contains one or possibly two fusing CF-labelled membranous mass(es) (arrowheads). 60 min CF incubation, stained en bloc with uranyl acetate. Bar equals 300 nm.
Fig. 44 and 45. Part of a series of sections through a coated vesicular profile. 30 min CF incubation, stained en bloc with uranyl acetate. Both at the same magnification. Bar in Fig. 45 equals 150 nm. Fig. 44. Section through a CF-labelled coated vesicular profile (arrowhead). PM- plasma membrane. Fig. 45. Serial section adjacent to that of Fig. 44. The CF-labelled profile (arrowhead) from Fig. 44 is clearly connected to the plasma membrane and thus is a CP.
Fig. 46-52. Serial sections through a CF-labelled coated vesicle. 30 min CF incubation, stained en bloc with uranyl acetate. All at the same magnification. Magnification as Fig. 45. An arrowhead indicates the position of a distinct ER profile which is first seen in Fig. 46 and last seen in Fig. 51. Double arrowheads indicate the position of a second ER profile which is first seen in Fig. 49 and last seen in Fig. 51. An arrow indicates the position of the CV of interest. The edges of the CV can be seen in Fig. 48 and 51 as patterning in the cytoplasm. A clear profile of the CV is seen in Fig. 49. In Fig. 50, the section is thinner than the others in the series. As a result while a profile of the CV can be seen, a distinct coat is not visible. There is no evidence of the CV in Fig. 46, 47 and 52.
43
PM
44 47 50
45 48 51
46 49 52
Fig. 53-58. Ultrastructural details of the PCR. All at the same magnification. Fig. 55-57 are adjacent serial sections. Bar in Fig. 58 equals 300 nm. Fig. 53. PCR which does not appear to be associated with other organelles in the cytoplasm. Note that one end of the PCR (arrow) is coated and that other regions (arrowheads) may also be coated. Also note that CF labelling is seen throughout the tubule of the PCR. 60 min CF incubation, stained en bloc with uranyl acetate. Fig. 54. PCR in which the tubules have a ring configuration. Note the coated region (arrow) and another region (arrowhead) which may be coated. Also note that the CF appears to be distributed at random throughout the PCR. 60 min CF incubation with pre-fixation washes omitted (see section 2.5 of Materials and Methods), unstained. Fig. 55. PCR with CF distributed throughout the tubules of the organelle. Note the coated vesicular profile (arrowhead) adjacent to the PCR. Also note the edge of a membrane dilation (see Fig. 56). 60 min CF incubation, stained en bloc with uranyl acetate. Fig. 56. Section serial to Fig. 55. The membrane dilation (arrow) in the PCR noted in Fig. 55 is clearly visible. Note that the dilation in the PCR is not labelled with CF. 60 min CF incubation, stained en bloc with uranyl acetate. Fig. 57. Section serial to Fig. 56. Note that CF is located throughout the PCR except for the portion of the tubules (arrow) which corresponds to the dilation in Fig. 56. 60 min CF incubation, stained en bloc with uranyl acetate. Fig. 58. A selected section from a series of sections through a PCR. The dilation in the PCR contains internal vesicles (arrowheads). Note that while a tubule (arrow) connected to the dilation is labelled with CF, the dilation itself and its internal vesicles are not labelled. 30 min CF incubation with pre-fixation washes normally used only for long term CF incubations (see section 2.5 of Materials and Methods), en bloc uranyl acetate staining (3h).
53, 54, 55, 56, 57, 58
Fig. 59-62. Serial sections through a PCR. Every second section used. Series continued on next plate. 60 min CF incubation, stained en bloc with uranyl acetate. All at the same magnification. Bar in Fig. 62 equals 300 nm. Fig. 59. PCR is not present. Fig. 60. CF-labelled vesicles (arrowheads) and the edge of a PCR tubule (arrow) can be seen. Fig. 61. CF-labelled vesicular (arrowheads) and tubular (arrows) profiles are present. Some vesicular profiles may be coated. Fig. 62. Network of interconnected tubules is present in the cytoplasm. A portion (arrowhead) of this network may be coated. CF appears throughout the network. Note the endoplasmic reticulum (ER) between the PCR and a dictyosome (D).
59 61
60 62
D ER
Fig. 63-66. Serial sections through a PCR (continued from previous plate). All at the same magnification. Bar in Fig. 66 equals 300 nm. Fig. 63. The PCR consists of a set of CF-labelled tubules. A portion (arrow) of these tubules is coated and other membrane profiles (arrowheads) may also be coated. ER - endoplasmic reticulum, D - dictyosome. Fig. 64. All that remains of the PCR are two or three small CF-labelled tubular profiles (arrowheads). Fig. 65. A single small CF-labelled profile (arrowhead) is the only structure in this section which might be related to the PCR. Fig. 66. The PCR can no longer be identified. The only CF labelled structures in this section are two SVs (arrowheads). Throughout this series of micrographs (Fig. 59-66) there was no evidence of an association between the PCR and dictyosomes. The closest dictyosome to the PCR was always separated from the PCR by an element of the ER (Fig. 62-64).
63 65
64 66
D ER
Bar = 0.1 μm
Fig. 67-70. Selected serial sections of a PCR and dictyosomes. 60 min CF incubation, stained en bloc with uranyl acetate. All at the same magnification. Bar in Fig .70 equals 300 nm. Fig. 67. PCR (arrows) well separated from two dictyosomes (D). Fig. 68. The closest approach by the PCR to the dictyosomes. While the PCR is in close proximity (arrowheads) to one of the dictyosomes, these organelles clearly remain separate and distinct. Fig. 69. The PCR is well separated from the dictyosomes. Fig. 70. In this section, only the PCR remains as a clearly identifiable organelle.
67, 68, 69, 70
D
Bar = 1 μm
Fig. 71-74. Selected serial sections of a PCR and dictyosome. Series continued on next plate. 60 min CF incubation, stained en bloc with uranyl acetate. All at the same magnification. Bar in Fig. 74 equals 300 nm. Fig. 71 and 72. Tubular membranes of a PCR. Note the coated membrane profiles (arrowheads). Also note that CF is distributed throughout the PCR. Fig. 73. Part of a dictyosome (D) is now visible. While the PCR is near the dictyosome there is no evidence of a membrane connection occurring between them. Note the coated membrane profiles (arrowheads). Also note the possible connection (arrow) between a PCR tubule and a dilation in the PCR (see Fig. 74). Fig. 74. The PCR and dictyosome are not connected. Note the dilation (arrow) in the PCR and the coated membrane profiles (arrowheads).
71, 72, 73, 74
Bar = 0.5 μm
Fig. 75-78. Selected serial sections of a PCR and dictyosome (continued from previous plate). All at the same magnification. Bar in Fig. 78 equals 300 nm. Fig. 75. The PCR consists only of short tubules and vesicular profiles (arrowheads) and has diminished in size and morphological complexity. There is still no evidence of a membrane connection between the PCR and dictyosome (D). Fig. 76. All that remains of the PCR is a few vesicular profiles (arrowheads), some of which may be coated. Fig. 77 and 78. The PCR is no longer present.
75, 76: Immunogold labeling of the Golgi apparatus in the cytoplasm of a neuron. The arrows indicate the gold particles.
77, 78: Immunogold labeling of the Golgi apparatus in the cytoplasm of a neuron. The gold particles are distributed throughout the Golgi apparatus.
Fig. 79-82. Selected serial sections showing a possible membrane connection between a PCR and dictyosome. Series continued on next plate. 60 min CF incubation, stained en bloc with uranyl acetate. All at the same magnification. Bar in Fig. 82 equals 300 nm. Fig. 79. Tubule (arrow) and associated vesicular profiles of a possible PCR. Coated vesicular profile (arrowhead). Fig. 80. A more extensive set of tubular membranes weakly labelled with CF. Possible coated membrane (arrowhead). Fig. 81. Tubular membrane weakly labelled with CF. Possible coated membranes (arrowheads). Fig. 82. Note the presence of coated membranes (arrowheads) and the start of a dilation (arrow) in a tubule. Also note the edge of a dictyosome (D).
79 81
80 82
D
Fig. 83-86. Selected serial sections showing a possible membrane connection between a PCR and dictyosome (continued from previous plate). All at the same magnification. Bar in Fig. 86 equals 300 nm. Fig. 83. A tubule with a dilation (arrow) near the dictyosome (D). Fig. 84. A CF-labelled tubule with a dilation (arrow) appears to be connected to the dictyosome. This section marks the closest approach to the dictyosome by the PCR. Fig. 85. A short CF-labelled tubule (arrowhead) is present near the membrane dilation (arrow). Fig. 86. A glancing section through the dilation to which a short CF-labelled tubule is attached.
83 85
84 86
Fig. 87. Association of a PCR with the endoplasmic reticulum. A selected section from a series through a PCR. The PCR consists of an unlabelled membrane dilation (arrow) connected to CF-labelled tubules. The ends of the tubules (arrowheads) are coated. The PCR is appressed against, and possibly but less likely connected to an element of rough endoplasmic reticulum (ER). 60 min CF incubation, stained en bloc with uranyl acetate. Bar equals 300 nm.
Fig. 88-91. Selected serial sections through a CF-labelled dictyosome. 60 min CF incubation with pre-fixation washes omitted (see section 2.5 of Materials and Methods), unstained. All at the same magnification. Bar in Fig. 91 equals 300 nm. Fig. 88. One cisterna (arrowhead) of the dictyosome is labelled with CF. Fig. 89. CF is seen in two cisternae (arrowheads) of the dictyosome. Fig. 90 and 91. CF appears to be in two cisternae (arrowheads) on opposite sides of the dictyosome. One section omitted between Fig. 90 and 91. For Fig. 88-91, note that the CF appears to be located at or near the ends of profiles of the labelled cisternae.
Fig. 92. Peripheral labelling of a dictyosome shown in face view. Note that CF (arrows) is distributed exclusively around the periphery of the dictyosome. 60 min CF incubation, pre-fixation washes omitted (see section 2.5 of Materials and Methods), unstained. Bar equals 300 nm.
87, 90, 88, 91, 89, 92
ER
Fig. 93. CF-labelling throughout the length of the outermost cisterna (arrowhead) of a dictyosome. 60 min CF incubation, pre-fixation washes omitted (see section 2.5 of Materials and Methods), unstained. Bar equals 300 nm.
Fig. 94. Membrane connection between a dictyosome and the endoplasmic reticulum. A dictyosome seen in face view appears to be connected (arrows) to an element of the rough endoplasmic reticulum (ER). Note the fenestrations (arrowheads) around the periphery of the dictyosome. Control (i.e. no CF) protoplast preparation, stained en bloc with uranyl acetate. Bar equals 300 nm.
93
94
Fig. 95-100. Series of sections through a MVB. 60 min CF incubation, pre-fixation washes omitted (see section 2.5 of Materials and Methods), unstained. All at the same magnification. Bar in Fig. 100 equals 300 nm. Fig. 95. Note the vesicle (arrow) containing CF. Additional CF (arrowhead) may represent the edge of the MVB. Fig. 96. Glancing section through the MVB. Fig. 97. Section through the MVB showing CF on the inner surface of the limiting membrane and the outer surface of the internal vesicles (arrowheads). Fig. 98. Section through the MVB. Note that CF particles appear to be concentrated within two regions of the MVB perhaps obscuring the presence of internal vesicles. Fig. 99. Section through the MVB showing CF (arrowhead) apparently free within the lumen of the MVB. Fig. 100. Note the vesicle (arrow) containing CF. Additional CF (arrowheads) may represent the edge of the MVB.
95, 96, 97, 98, 99, 100
Figure 2. Electron micrographs showing the distribution of immunogold particles in the cytoplasm and nucleus of cells from the rat hippocampus. The gold particles are localized in the cytoplasm (arrows) and in the nucleus (arrowheads). Scale bar = 0.5 μm.
Fig. 101-106. Detailed ultrastructure of MVBs. All at the same magnification. Bar in Fig. 106 equals 300 nm. Fig. 101. MVB with a lamellar plaque (arrow) on the cytoplasmic surface of the limiting membrane. Note the periodicity seen within the outer layer of the plaque. Also note the numerous small internal vesicles within the MVB. Control (i.e. no CF) protoplast preparation, double stained with uranyl acetate and lead citrate. Fig. 102. MVB bearing a plaque (arrow) on the cytoplasmic surface of its limiting membrane. Note that it is not possible to see substructure within the plaque. Control protoplast preparation, unstained. Fig. 103. MVB with a plaque (arrow) on the cytoplasmic surface of its limiting membrane. It is not possible to discern any substructure within the plaque. Control protoplast, en bloc uranyl acetate staining. Fig. 104. MVB with tubular extensions (arrows) protruding from the main body of the MVB. Note the CF bound to the outer surface of the small internal vesicles. 30 min CF incubation with pre-fixation washes normally used only for long term CF incubations (see section 2.5 of Materials and Methods), en bloc uranyl acetate staining (3h). Fig. 105. MVB containing a large number of internal vesicles. Note what appears to be a small vesicle (arrow) forming by invagination of the limiting membrane. Control protoplast preparation, double stained with uranyl acetate and lead citrate. Fig. 106. Rough endoplasmic reticulum (ER) appressed against a MVB (arrows). 60 min CF incubation, stained en bloc uranyl acetate.
101, 102, 103, 104, 105, 106
Figures 101-106: Electron micrographs showing various cellular structures and organelles.
- Figure 101: A cell with numerous vesicles and a dense body (arrow).
- Figure 102: A cell with a single arrow pointing to a specific structure.
- Figure 103: A cell with multiple arrows indicating different structures.
- Figure 104: A cell with two arrows pointing to specific structures.
- Figure 105: A cell with a single arrow pointing to a specific structure.
- Figure 106: A cell with an arrow pointing to a specific structure labeled "ER" for endoplasmic reticulum.
Fig. 107-109. Selected serial sections showing a comparison of a CF-labelled vacuolar membranous mass with the contents of MVBs in the cytoplasm. 60 min CF incubation, stained en bloc with uranyl acetate. All at the same magnification. Bar in Fig. 107 equals 300 nm. The section numbers are indicated on each micrograph (i.e. S1, S3, S5). Fig. 108 and 109 show areas equivalent to those present in Fig. 107. Fig. 107. Two MVBs (MVB1 and MVB2) and the edge of a CF-labelled membranous mass (M) in the central vacuole are present. The membranous mass is labelled not only with CF (arrows) but also with electron-dense precipitates (arrowheads) resulting from en bloc uranyl acetate staining. Similar precipitates are also found along the tonoplast (T). Fig. 108. Profiles of MVB1 and MVB2 are again seen. MVB1 is strongly labelled with CF. Fig. 109. This figure shows another profile of the vacuolar membranous mass (M). Note the similarity in the appearance and size of this profile with that of the contents of MVB1 in Fig. 108.
MVB1, MVB2, T, M
S3, S5, S1
107, 108, 109
Bar = 0.5 μm
Fig. 110-113. Selected serial sections of a small CF-labelled vacuole. 60 min CF incubation with pre-fixation washes omitted (see section 2.5 of Materials and Methods), unstained. All at the same magnification. Bar in Fig. 113 equals 300 nm. Fig. 110. A small vacuole (SVA) is near the central vacuole (V). Due to fixation damage (arrowheads), it is not possible to determine if the small vacuole is truly independent of the central vacuole. Fig. 111. The small vacuole contains a CF-labelled membranous mass (arrow). Note the presence of a vesicle (arrowhead) on the plasma membrane. Fig. 112. A CF-labelled membranous mass (arrow) is again seen in the small vacuole. One section with a CF-labelled membranous mass in the small vacuole was omitted between Fig. 111 and 112. Fig. 113. The membranous mass is no longer visible in the small vacuole. Again, note the presence of a vesicle (arrowhead) on the plasma membrane.
SVa
V
110 112
111 113
Bar = 0.5 μm
Fig. 114-116. NF uptake results. 70 min NF incubation, unstained. All at the same magnification. Bar in Fig. 116 equals 200 nm. Fig. 114. A MVB is labelled with NF (arrowheads). Also note the NF bound to the plasma membrane (arrows). Fig. 115. A single NF particle (arrowhead) in a dictyosome. Fig. 116. NF bound to the plasma membrane (arrowhead) and a CP (arrow).
Fig. 117. Low magnification of protoplast post-fixed with ZIO. The nuclear envelope (NE) is strongly labelled and nuclear pores (arrows) are easily visualized. The endoplasmic reticulum, in this case specifically the tubular endoplasmic reticulum (tER), is also heavily labelled and appears to be connected to the nuclear envelope. Note that mitochondria (M) are heavily labelled while only the internal membranes of plastids (P) are labelled. SG= starch granule, N= nucleus. Unstained, thick section (200-300 nm). Bar equals 500 nm.
SG, M, P, N, NE, FER
114, 115, 116, 117
Fig. 118 and 119. ZIO post-fixation results for cisternal endoplasmic reticulum. Unstained, thick sections (200-300 nm). Fig. 118. Uneven labelling of cisternal endoplasmic reticulum (cER) results in a "flaky" appearance. Note that the tER appears to be connected to the cER. D- dictyosome, M- mitochondrion. Magnification as Fig. 121. Fig. 119. Another example of cER which has fenestrations (arrowheads). Note that the cER again appears to be connected to tER (arrow). V- central vacuole. Bar equals 500 nm.
Fig. 120-121. ZIO post-fixation results for dictyosomes. Unstained, thick sections (200-300 nm). Both at the same magnification. Bar in Fig. 121 equals 300 nm. Fig. 120. Two dictyosomes (D) are seen in cross-section and are heavily labelled with ZIO reaction product. Also note the occurrence of small ZIO precipitates (arrowheads) in the general cytoplasm which may or may not occur in vesicles. Fig. 121. A dictyosome (D) seen in face view. Note the tubules which emanate from the periphery of the dictyosome (arrowheads). Also note the presence of tER (arrow).
cER iER M D
118 120
V
119
D
121
Scale bars: 0.5 μm
Fig. 122-127. Results of tilting experiments with thick sections of ZIO post-fixed protoplasts. Unstained, thick sections (200-300 nm).
Fig. 122-124. A series of micrographs taken from the same section tilted at various angles (Fig. 122: -30°, Fig. 123: untitled, Fig. 124: +30°). All at the same magnification. Bar in Fig. 124 equals 300 nm. The approximate direction in which tilting occurred is indicated by the "two-headed" arrow in Fig. 122. D- dictyosome. Note the fenestration (arrow) in the ER which enlarges as the section is tilted from -30° (Fig. 122) to +30° (Fig. 124). Arrowheads and double arrowheads indicate the identical segments of tER in this series of micrographs. One ER tubule (double arrowhead) is apparently not attached to the dictyosome in Fig. 122 but appears to be connected to the dictyosome in Fig. 123 and 124. A second ER tubule (arrowhead) appears to be connected to the dictyosome in Fig. 122 and 123 but not in Fig. 124. Thus neither segment of tER remains attached to the dictyosome through the entire tilt series and therefore they are not physically connected to it.
Fig. 125-127. Series of micrographs from the same section tilted at various angles (Fig. 125: -30°, Fig. 126: untitled, Fig. 127: +30°). All at the same magnification. Bar in Fig. 127 equals 300 nm. The approximate direction of tilting is indicated by the "double-headed" arrow in Fig. 125. Arrowheads and double arrowheads are used to indicate the identical segment of ER throughout the tilt series. D-dictyosome. Two segments of ER (arrowhead, double arrowhead) appear to remain in contact with the dictyosome throughout the series and therefore may represent true membrane connections between the ER and the dictyosome.
122, 123, 124, 125, 126, 127
D
Arrows indicate the direction of movement.
Fig. 128-133. ZIO post-fixation results for dictyosomes using thin (60-80 nm) sections. Fig. 128. A dictyosome (D) which appears to be connected (arrowhead) directly to the endoplasmic reticulum (ER). Unstained, magnification as Fig. 133. Fig. 129. A dictyosome (D) which appears to be connected (arrowhead) to the endoplasmic reticulum (ER). Note that regions of a mitochondrion (M) probably corresponding to cristae are not labelled with ZIO reaction product. Unstained, magnification as Fig. 133. Fig. 130. A dictyosome (D) in which a clear polarity of ZIO labelling is seen between the cisternae on one side of the dictyosome relative to the cisternae on the other side. Unstained, bar equals 300 nm. Fig. 131. A dictyosome (D) in face view in which peripheral fenestrations (black arrowheads) are readily seen. A membrane connection (white arrowheads) between this dictyosome and a second dictyosome (D2) is also visible. Unstained, magnification as Fig. 133. Fig. 132. Two dictyosomes (D) are present. On these dictyosomes, membrane buds (arrowheads) which may be coated are visible. Stained with uranyl acetate and lead citrate. Magnification as Fig. 133. Fig. 133. Two dictyosomes, one in cross-section (D) and the other (D2) in face view, are present. Note the apparent membrane connection (arrowheads) between them. Unstained, bar equals 300 nm.
128 FR
131 D D2
129 D M
132 D
130 D
133 D2
Fig. 134. ZIO post-fixation results with thin sections showing an unlabelled PCR. Note the coated ends (arrowheads) of the tubules of the PCR. Also note the presence of a dictyosome (D) which is heavily labelled with ZIO reaction product. Stained with uranyl acetate and lead citrate. Bar equals 300 nm.
Fig. 135-137. ZIO post-fixation results using thin sections for MVBs. All at the same magnification. Bar in Fig. 137 equals 300 nm. Fig. 135. Three MVBs (arrowheads) with internal vesicles containing very fine ZIO deposits. Note the ER which is appressed against the MVBs and is heavily labelled with ZIO. Unstained. Fig. 136. A MVB (arrow) with internal vesicles which are heavily labelled with ZIO. Unstained. Fig. 137. MVB (arrow) containing CF and internal vesicles heavily labelled with ZIO. Note that portions of the plastid (P) and the ER (arrowhead) are also labelled with ZIO. 90 min CF incubation, unstained.
134
136
135
ER
137
P
Fig. 138-140. Results of lanthanum chloride uptake by whole SB-1 cells. 60 min lanthanum chloride incubation, unstained. All at the same magnification. Magnification as Fig. 143.
Fig. 138. The majority of the lanthanum deposits (arrowheads) are found on the outer surface of the cell wall (CW) although they are also seen at lower frequency within the cell wall. Note the presence within the cytoplasm of a vesicular profile (arrow) containing lanthanum. Fig. 139. The cross-walls (CW) between adjacent SB-1 cells contain a few lanthanum deposits (arrowheads) indicating that lanthanum reaches these regions at least to some small degree. Note the lanthanum-containing vesicular profile (arrow) within the cytoplasm of one cell. Fig. 140. A lanthanum deposit (arrow) can be seen within an element of the ER. The ER appears to be slightly dilated at the region where the deposit occurs. CW - cell wall.
Fig. 141-143. Results of incubating whole SB-1 cells with CF for 60 min. All at the same magnification. Bar in Fig. 143 equals 300 nm. Fig. 141. Most CF appears to be bound to the outer surface of the cell wall (CW). Very little CF (arrowheads) appears to penetrate part way into the cell wall. Unstained. Fig. 142. A particle of CF (arrow) appears to be bound to the outer surface of the plasma membrane. Stained with lead citrate. Fig. 143. ER apparently connected (arrow) directly to the plasma membrane. Unstained.
138 CW
141 CW
139 CW
142 CW
140 CW
143 PM
Bar = 0.5 μm
Fig. 144-146. Results of 14 h incubation of whole SB-1 cells with CF. Unstained, all at the same magnification. Bar in Fig. 146 equals 300 nm. Fig. 144. Most CF is bound to the outer surface of the cell wall (CW). However, some CF (arrowheads) has penetrated into the cell wall. Note that there also appears to be some cytoplasmic labelling with CF (arrows). This CF appears to be free in the cytoplasm. Fig. 145. CF is seen within an element of the ER (arrows). Also note the presence of CF (arrowheads) part way through the cell wall. Fig. 146. Note the "trail" of CF (arrowheads) running through the cell wall (CW) to the plasma membrane. Also note the presence of CF (arrows) within the cytoplasm. Some of this CF appears to be free in the cytoplasm. Some is in the cytoplasm but possibly associated with a vesicular profile. Some is in a vesicular profile. This profile may be an element of the ER seen in cross-section.
Fig. 147. Additional observation of whole SB-1 cells. Note the invagination of the plasma membrane (arrow) where many small membrane-bounded vesicles are present in the periplasmic space. CW- cell wall, PM- plasma membrane. Lanthanum control (i.e. no lanthanum) cell preparation, unstained. Bar equals 300 nm.
144
145
146
147
CW
PM
4 Discussion
4.1 Introductory Remarks
Endocytotic processes are well known and well demonstrated in animal cell systems (Steinman et al. 1983). In sharp contrast, little is known about these processes in plant cells or protoplasts. Uptake of a variety of markers, such as native ferritin (Power and Cocking 1970), hemocyanin (Williamson et al. 1976), latex beads (Suzuki et al. 1977), viruses (Takebe 1979) and liposomes (Fununaga et al. 1983), have been demonstrated in protoplasts. However, prior to the start of this research project, there was no evidence for the involvement of coated membranes in the endocytotic process. Therefore, the results of this research project are significant as they clearly implicate a number of coated membranous organelles including; CPs, CVs, the PCR and dictyosomes, in the process of endocytosis in soybean protoplasts. Also, SVs, MVBs, small vacuoles and the central vacuole appear to be involved in this process. All of the organelles which become labelled with CF in soybean protoplasts bear some similarity morphologically and/or functionally to organelles involved in receptor-mediated endocytosis in animal cells (see discussion below). An endocytotic pathway for soybean protoplasts which is based in part on an analogy with the receptor-mediated endocytotic pathway in animal cells will be presented later in this section.
A problem encountered constantly throughout this project was the question of specimen contrast. A number of staining protocols were used in an attempt to achieve sufficient specimen contrast such that coats and membranes could clearly be resolved while not seriously impeding one's ability to identify CF. None of these staining protocols proved to be adequate for general use. Unstained sections permitted clear resolution of CF as small electron-dense particles but the resolution of membranes and protein coats was poor. Staining thin sections with lead citrate improved the resolution of membranes and coats somewhat but still was not ideal. Staining thin sections with uranyl acetate followed by lead citrate provided good specimen contrast. However, this protocol also stained CF heavily, making it difficult or impossible to identify with any degree of confidence. Therefore, to unambiguously identify CF, unstained sections or sections stained with lead citrate alone were used. Sections were stained with uranyl acetate en bloc to resolve ultrastructural details of organelles (e.g. serial sectioning analysis) as well as to visualize CF particles.
4.2 Effect of Molecular Charge on Endocytosis
The uptake of ferritin particles via the endocytotic pathway in soybean protoplasts appeared to be very sensitive to the charge of the marker in use (section 3.8). Uptake of
CF particles could be easily detected at much lower protein concentrations than for NF. Furthermore, much greater amounts of CF were internalized even when the relative concentration of CF was considerably lower than the concentration of NF. In addition, the plasma membrane was more heavily labelled when CF was used rather than NF. These results are hardly surprising. When NF and CF have been used with animal cells, similar if not identical results were found. For example, with chick embryos, substantial uptake was observed using CF but uptake of NF was not detected even when NF was used at a concentration 40 times greater than that for CF (MacLean and Sanders 1983). Furthermore, in human fibroblasts, binding of NF to the plasma membrane and uptake were very low in comparison with CF. This was the case even when NF was used at 50 times the concentration of CF (Van Deurs et al. 1982). For animal cell studies, NF has been considered to be a marker for fluid phase endocytosis, i.e. the marker is not bound to the plasma membrane prior to uptake (e.g. Morales et al. 1985). Strictly speaking, for soybean protoplasts and at least some types of animal cells, NF is not a fluid phase marker as some NF particles were bound to the plasma membrane (Fig. 114 and 116, and Van Deurs et al. 1982). In the case of CF, it is considered to be a marker for adsorptive endocytosis, i.e. the marker is bound to the plasma membrane but not necessarily to a specific receptor prior to uptake (e.g. Van Deurs et al.
1982; Morales et al. 1985). Adsorptive endocytosis is more efficient than fluid phase endocytosis as the marker is concentrated on the plasma membrane where the initial events of endocytosis are occurring. Therefore, much lower concentrations of markers are needed to detect uptake into the cell or protoplast when adsorptive endocytosis is being studied.
The plasma membrane of plant protoplasts is known to have a strong net negative surface charge (Nagata and Melchers 1978; Fowke 1986). In soybean, when isoelectric focusing was used to measure the surface charge of protoplasts derived from SB-1 cell cultures, the isoelectric point of these protoplasts was found to be 3.7 or 5.1, depending on the nature of the pH gradient (Griffing et al. 1985). Therefore at pHs near neutrality (e.g. pH 6.8), these protoplasts would have a strong net negative surface charge. CF particles would therefore bind electrostatically to the negatively charged sites on the plasma membrane resulting in the observed high levels of CF binding and uptake. It should be noted that electrostatic interactions are weak bonds (Van Deurs and Christensen 1984). Thus, it is possible that a CF particle can dissociate from the group to which it was originally bound (Van Deurs and Christensen 1984 and references cited therein) and may form another bond with another group or may remain free within the lumen of an
organelle. Therefore, it is possible that internalized CF particles may no longer be associated with membrane constituents originating from the plasma membrane. While acknowledging that the cytoplasmic location of any particular CF particle may not correspond to the location of plasma membrane constituent(s), it has been assumed, for the sake of discussion, that the overall cytoplasmic distribution of CF particles is representative of the overall distribution of membrane constituents and/or lumenal contents from endocytotic, plasma membrane-derived vesicles.
4.3 Adverse Effects of CF on the Morphology of Protoplasts and Cells
While adverse effects of CF treatment on both SB-1 protoplasts (section 3.7) and cells (section 3.10.2) were detected, the morphological differences between control and CF-treated preparations were relative rather than absolute. That is to say, some protoplasts (and cells) incubated in CF were comparable to their controls. Every attempt was made to make observations from CF-treated protoplasts and cells which were similar to the controls.
Protoplasts treated with CF tended to collapse more than controls, contained extracted cytoplasm and exhibited more frequent damage to the tonoplast. Preliminary results indicated that cells from newly initiated cultures were also susceptible to damage by CF. During the study of ferritin uptake by the roots of *Phaseolus vulgaris* seedlings (Barton
1964), the ferritin solutions used in the experiments were found to be slightly toxic to root cells. Unfortunately, the specific symptoms of the toxic effect were not described and therefore a direct comparison to the present situation cannot be made. However, Barton did speculate that contaminants in the form of cadmium salts used in the preparation of ferritin may have been responsible for these toxic effects. This explanation is not likely to apply to the present situation as NF preparations which would also be contaminated with cadmium salts did not produce the effects associated with CF treatment. It is more likely that the problems associated with the CF incubations are related to the charge of this macromolecule. Other basic macromolecules, including histones and poly-L-lysine, are known to inhibit growth for instance in barley roots (Drew and McLaren 1970; Drew et al. 1970). This inhibition of growth was correlated with leakage of cell contents suggesting that basic macromolecules act primarily at the plasma membrane to alter the normal cell permeability (Drew and McLaren 1970). The difficulty in observing the typical unit membrane structure in CF treated protoplasts is consistent with the occurrence of damage to the plasma membrane or an alteration in its properties. The extracted appearance of some protoplasts and cells is consistent with the possible occurrence of increased membrane permeability.
Variations in the degree of damage seen in different protoplast and cell preparations may reflect interexperimental differences in the physiological state of these preparations, in the culturing procedures or in the procedures used for protoplast production and isolation. Drew and co-workers (1970) found that the inclusion of as little as 1 mM CaCl₂ resulted in the protection of plant tissue from the adverse effects of basic macromolecules. It was thought that the presence of Ca²⁺ ions was responsible for this effect. It would be interesting to determine whether Ca²⁺ ions would have the same effect during all CF incubations but especially with protoplasts.
4.4 Significance of Short Term CF Uptake by Protoplasts
The results from the short term CF uptake studies demonstrate the capacity of soybean protoplasts to internalize surface-labelled plasma membrane by endocytosis via their coated-membrane system. After 10 s of exposure to CF, this label was present on the entire plasma membrane surface, in CPs and CVs. The absence of labelled, large, smooth vesicles at this and other early time points (30 s, 2 min) indicates that phagocytosis was not responsible for the uptake of the plasma membrane. The evidence from Fig. 25 (section 3.3) favors the existence of free CVs. Furthermore, there was no evidence for the formation of uncoated or smooth vesicles either in the vicinity of CPs (i.e. the formation of receptosomes) or from elsewhere on the plasma
membrane. Therefore, the experimental evidence strongly suggests that CF uptake was mediated by the formation of CVs via the invagination of CPs on the plasma membrane.
The serial sectioning results from section 3.6.1 are consistent with the existence of free coated vesicles. However, there is some controversy in the literature on this matter. Some groups working on animal cell systems have used serial sectioning to provide evidence for the existence of free CVs (Fan et al, 1982; Petersen and Van Deurs 1983; Van Deurs et al. 1984). Depending upon the cell type and the section thickness used, 4 to 47% of coated vesicular profiles in the vicinity of the plasma membrane were found to be unattached to other membranous organelles (Fan et al. 1982; Petersen and Van Deurs 1983; Van Deurs et al. 1984). In marked contrast, one research group (Willingham and Pastan 1983, 1984b), using serial sections 70 to 100 nm in thickness and a special fixation procedure, found that all coated vesicular profiles in the vicinity of the plasma membrane were connected to it by a narrow membrane neck (Willingham and Pastan 1983). Furthermore, they found that uncoated or smooth membrane pits were sometimes found adjacent to CPs on the plasma membrane. They proposed that CPs were involved in the clustering of receptors and their bound ligands to specific regions on the plasma membrane, that these clusters of receptors and ligands were then
transferred to the adjacent plasma membrane from which a smooth or non-coated vesicle called a receptosome was formed by the process of invagination (Fig. 2 and Willingham et al. 1981). This controversy has not been resolved. For a more complete analysis of the debate see the papers by Van Deurs and co-workers (1984), and Willingham and Pastan (1984a,b). Current reviews by authors (Fine and Ockleford 1984; Breitfeld et al. 1985; Wileman et al. 1985; Stahl and Schwartz 1986) not directly involved in the debate indicate that the general consensus favours the existence of free CVs.
Occasionally vesicular profiles having the appearance of partially coated vesicles were seen in samples from the 30 s and subsequent time points, Fig. 28 (section 3.3). This observation suggests that the smooth vesicles seen during 30 s CF incubations and in all subsequent time points may have been formed by the uncoating of coated vesicles and that this uncoating process occurred rapidly. This would be consistent with results obtained from animal systems (Pearse and Bretscher, 1981). Partially coated or incompletely coated vesicles have also been seen in other plant cells (Ryser, 1979: Lucas and Franceschi, 1981). In one case, the existence of partially coated vesicles was taken as being consistent with an endocytotic mechanism (Lucas and Franceschi, 1981).
Overall, the results of short term uptake of CF by
soybean protoplasts are consistent with an endocytotic pathway which is similar to the conventional view of receptor-mediated endocytosis in animal cells (Fig. 1 in section 1.2) where CVs form from CPs, then shed their coats and fuse with other organelles. The results are not consistent with the formation of receptosomes as proposed by Willingham and co-workers (1981).
4.5 Significance of Intermediate Term CF Uptake by Protoplasts
Incubations of 4 to 16 min in CF were used on protoplasts to determine which organelles in addition to CPs, CVs and SVs become labelled and in what order. Unfortunately, the results were not as clear cut as one would have liked. Label rapidly appeared in three additional organelles; the PCR, dictyosomes, and MVBs. Label was seen in examples of the PCR as early as 2 min (Fig. 31), in dictyosomes as early as 4 min and in MVBs as early as 6 min. Thus a large separation in time of the labelling of these organelles did not occur. However, the PCR was more consistently labelled at the early time points and is thus thought to be labelled prior to the dictyosomes and MVBs. Using the same rationale, the results suggest that dictyosomes label prior to MVBs. In order, to rigorously demonstrate that one organelle was labelled prior to the others would require a detailed quantitative and statistical
analysis of the labelling pattern. Such an analysis was not attempted for two major reasons. First, the number of organelles which were labelled at these time points was low. Thus a prohibitively large number of organelles would have to be sampled before a statistically significant difference would be obtained. This problem is compounded by the fact that the organelles frequently contained one or a few particles of ferritin which could easily be missed when viewing the specimen in the electron microscope. Second, it was not always possible to unambiguously identify an organelle from a single section. For example, some profiles through a dictyosome can appear similar to the PCR and not all vesicles containing internal vesicles are MVBs. Therefore, the misidentification of unlabelled and labelled organelles would have been a problem.
The results from 30 min CF incubations were interesting as labelling of the PCR, dictyosomes and MVBs was readily detected. The labelling of the MVBs was particularly strong suggesting that CF particles accumulate in these organelles over time.
4.6 The Morphology of the Partially Coated Reticulum
The partially coated reticulum is an organelle in plant cells which has only recently been described in detail for the alga, Chara (Pesacreta and Lucas 1984) and some selected angiosperms (Pesacreta and Lucas 1985). For this reason, its presence and appearance in soybean protoplasts is of special
interest. The PCR in soybean protoplasts appeared very similar to the PCR found in cells from young *Phaseolus vulgaris* leaves and in *Zea mays* cortical root cells (Pesacreta and Lucas 1985). It consisted of a set of interconnected tubular membranes in which coated membrane regions could be found. Furthermore, in all cases, coated vesicles were seen in the cytoplasm around these organelles. One feature of the PCR of soybean protoplasts which is apparently unique was the occurrence of membrane dilations. Pesacreta and Lucas (1985) did not mention the occurrence of such dilations, but did note that in *Z. mays* the PCR could be found associated with small, dense zones of membranes that appeared to be either greatly anastomosed or vesiculated. However, these zones of membranes at least as illustrated by Pesacreta and Lucas (1985) bear little resemblance to the dilations of the PCR seen in soybean protoplasts.
The PCR of soybean protoplasts is also similar to those of *P. vulgaris* and *Z. mays* with respect to their distribution within the cytoplasm. Pesacreta and Lucas (1985) noted that the PCR in *P. vulgaris* and *Z. mays* was frequently but not always associated with Golgi stacks (dictyosomes). Only approximately 10% of the 100 PCRs that were examined in these two species were not closely associated with Golgi. Comparable data are not available for
the distribution of the PCR in soybean protoplasts and even if it were, a direct comparison would be difficult to make as Pesacreta and Lucas did not define what they meant by an association between the PCR and Golgi. However, the overall pattern of PCR distribution was the same in soybean protoplasts. The PCR was frequently seen in the vicinity of dictyosomes but in a number of cases was clearly not associated with dictyosomes.
Pesacreta and Lucas (1985) also noted that physical continuities between the PCR and Golgi stack membranes (cisternae of dictyosomes) were not evident in most cases or were "problematic" (i.e. ambiguous). Only occasionally did they find the PCR to be directly connected to the Golgi stack. Again, the results for soybean protoplasts are similar. In a number of cases, when the PCR was in the vicinity of a dictyosome there was clearly no evidence of a physical connection between the two organelles. In other cases, the organelles were closely associated but clear connections could not be demonstrated. In only one case was an unambiguous connection seen between a PCR and a dictyosome. A number of factors contributed to the uncertainty in establishing connections between the PCR and dictyosomes. First, sufficiently long series of sections were difficult to obtain. Second, the dictyosomes and the PCR were both, very complex organelles. The cisternae of dictyosomes frequently were curved and the tubules of the
PCR and of the dictyosomes could be curved or branched making glancing sections through these organelles difficult to interpret. In addition, both of these organelles had a tubular component which had a coat on part of its cytoplasmic surface and it was difficult to differentiate between these tubules. Thus when two or more of these organelles were found in close proximity to one another, it was frequently difficult to determine, with any degree of certainty, where one organelle began and another ended or whether the two organelles were in fact connected. Finally, it is possible that physical continuities between the PCR and dictyosomes or other organelles such as the ER (see below) may be transitory in nature. This possibility in combination with the technical problems discussed above may mean that a great deal of good luck would be required to unambiguously demonstrate membrane connections between these organelles.
Overall, the serial sectioning results for the PCR provided evidence that this organelle was at least in some cases distinct from the dictyosomes of the Golgi apparatus. Zinc iodide-osmium tetroxide postfixation was used in the hope of providing further evidence to support these results. In a study of the alga, *Chara corallina*, Pesacreta and Lucas (1984) were able to differentiate between the PCR which was lightly labelled with ZIO and Golgi stacks which were
generally heavily labelled. They reasoned that such results offered little support for the existence of either membranous connections or membrane flow from the Golgi to the PCR.
Unfortunately, the results from soybean protoplasts postfixed with ZIO were not as clear cut as with *Chara*. The labelling of dictyosomes was more variable than had been encountered in *Chara* thus making a comparison between the PCR and the dictyosomes more difficult. Furthermore, coats were more difficult to visualize in ZIO postfixed material thus making the identification of the PCR more difficult. Despite these problems, it was possible to identify probable PCR profiles which were unlabelled or only weakly labelled with ZIO reaction product while all profiles of the dictyosomes in the same protoplast were heavily labelled. These results are consistent with those for *Chara* discussed above and provide further evidence that the PCR can at least in some instances be distinct from the dictyosomes.
Pesacreta and Lucas (1985) considered the PCR to be a structural entity. But when faced with the range of morphology of this organelle, especially near the elements of the Golgi apparatus, they felt that it might be useful to consider the various structures as parts of an ultrastructural series. In the simplest case in such a series the PCR would consist of only a few, small coated membranes grouped together. The next case would involve a
larger but still relatively distinct membrane system. The final case would be the one in which physical continuities would exist between the PCR and the Golgi (dictyosomes). This type of series would reflect physiological differences in the rate, type and extent of interactions between the PCR and the Golgi (Pesacreta and Lucas 1985). Based on the similarities (discussed above) which exist between the PCRs of *P. vulgaris*, *Zea mays* and soybean protoplasts, it seems likely that such an ultrastructural series also exists in soybean.
The concept of an ultrastructural series is helpful in reconciling the existence of the PCR with previous observations of coated membranes located in the vicinity of dictyosomes, for instance the peripheral tubules of dictyosomes seen in protoplasts from *Vicia faba* (Joachim and Robinson 1984) and the trans pole complex in maize root cap cells (Hübner et al. 1985). These latter structures may be analogous to the PCR or may represent distinct forms of the PCR. The novelty of the results reported by Pesacreta and Lucas (1985) and of this project is not so much the discovery of a partially coated membrane network but rather the observation that some of these organelles are not associated with the Golgi (Pesacreta and Lucas 1985).
Pesacreta and Lucas (1985) also noted that organelles other than Golgi were often located near the PCR. Elements
of the ER were frequently seen adjacent to the PCR but there was no mention of possible connections between these organelles. In soybean protoplasts, not only was the PCR seen adjacent to the ER but in some cases the PCR appeared to be appressed against the ER or perhaps physically connected to it. This observation is quite important and may be of great significance. It has been noted (Harris 1986) that the PCR appears similar to the organelle which has been given the acronym, GERL (for Golgi-associated endoplasmic reticulum which gives rise to lysosomes). This organelle appears to be present in both animal and some plant cells (Novikoff 1976; Marty 1978). GERL is thought to be a partially coated tubular element of the ER which is located in close proximity to Golgi stacks, usually on the trans side of the stack (Novikoff 1976). While there is some debate about the structure of GERL, there can be no denying that the general description of the PCR is quite similar.
4.7 The PCR and Animal Cell Endosomes
The PCR bears considerable resemblance to a class of endosomes in animal cells which are known by a number of names including, trans reticular Golgi (Willingham and Pastan 1984a) and CURL, an acronym for compartment of uncoupling of receptor and ligand (Geuze et al. 1983a). Trans reticular Golgi is located, as its name indicates, on the trans side of the Golgi and consists of tubular elements which have coated buds (Willingham and Pastan 1984a;
Willingham et al. 1984). CURL is found in the peripheral cytoplasm and deeper within the cytoplasm in the vicinity of the Golgi apparatus (Breitfeld et al. 1985). It consists of interconnected tubular and vesicular profiles (i.e. dilations, Geuze et al. 1983a; Breitfeld et al. 1985). Small regions of the tubular component bear a clathrin coat on their cytoplasmic surface (Geuze et al. 1984, 1985; Breitfeld et al. 1985). Furthermore, internal vesicles have been observed within some of the vesiculated regions of this organelle (Geuze et al. 1985; Harding et al. 1985). It seems possible that these endosomes are perhaps the same organelles given different names by different research groups (Willingham and Pastan 1984a; Bretscher and Thomson, 1985). These structures also have some similarity to GERL (Willingham and Pastan 1984a; Geuze et al. 1985). In fact, the similarities are so strong that Willingham and Pastan (1984a) have suggested that CURL, trans reticular Golgi and GERL are the same organelle.
Studies of the function of CURL in receptor-mediated endocytosis have produced some most interesting results. It is thought that CURL is the endosome in which some receptor-ligand complexes dissociate, presumably due to a reduced pH within CURL, and where sorting of the receptors from the ligands occurs (Geuze et al. 1983a, 1984; Breitfeld et al. 1985). The sorting event may result simply through the
mechanism of mass action (Harding et al. 1985; Rome 1985). It has been estimated that roughly 90% of the membrane surface area in CURL is accounted for by the tubular portion of the organelle with the remaining 10% accounted for by the membrane surrounding the vesicular portion of the organelle. With respect to volume distribution, the values are reversed with the lumen of the vesicular portion of CURL accounting for the majority of the volume. If one assumes that receptors are free to move within the plane of the lipid bilayer of the membrane and that likewise free ligands are able to move freely within the lumen of CURL then by mass action the majority of the receptors (i.e. 90%) should be located within the tubular portion of the organelle while most of the ligand will be located within the vesicular portion. Thus quite efficient sorting of the receptor from the ligand could occur without proposing any special structural modifications to CURL (Rome 1985). It is thought that receptors could recycle to the plasma membrane in vesicles derived from the tubular portion. Recycling could occur directly from CURL or the receptors may first pass through the Golgi apparatus (Breitfeld et al. 1985). In contrast, for the transit of ligands, a very different mechanism has been proposed. There is some morphological evidence which suggests that multivesicular bodies in animal cells form as a result of a maturation process operating on the vesicular portion of CURL. It is proposed that the
tubular portions disappear as they produce outgoing receptor-rich vesicles while the vesicular portion produces internal vesicles via a simple invagination process. The resulting multivesicular bodies are then thought to fuse with (primary) lysosomes and thereby are transformed into (secondary) lysosomes (Geuze et al. 1983b; Harding et al. 1985).
It should be noted that the trans reticular Golgi has also been implicated with respect to the sorting of receptor-ligand complexes (Willingham et al. 1984). Epidermal growth factor and its receptor appear to be concentrated in the coated buds of the trans reticular Golgi. It is thought that this receptor-ligand complex is concentrated in the coated buds prior to its subsequent transfer to lysosomes. In contrast, transferrin and its receptor are generally recycled to the plasma membrane without being delivered to the lysosome. The transferrin-transferrin receptor complex is not concentrated in the coated buds of trans reticular Golgi but is generally found in the tubular portion of the organelle. It is thought that transferrin and its receptor are recycled back to the plasma membrane from this location (Willingham et al. 1984).
The observation in this study that the PCR becomes labelled with CF is a highly significant result as it provides the first direct evidence for the involvement of
the PCR in endocytosis. Based on an analogy with CURL or trans reticular Golgi, it is possible to envision the PCR playing a role in membrane sorting events and possibly in the dissociation of as yet unidentified receptor-ligand complexes. In this way, the PCR may be responsible for determining what material is directed into the degradative part of the pathway represented by MVBs and the central vacuole (see sections 4.9 and 4.11) and what material is recycled to the plasma membrane either directly from the PCR or indirectly via dictyosomes.
4.8 Significance of the Dilation in the PCR
The fact that the dilation is not seen in all of the PCR of soybean protoplasts and that the dilation sometimes contains internal vesicles creates a situation which lends itself to some interesting speculation. If one accepts the proposition that the MVB becomes labelled after the PCR and receives its label from the PCR then it is tempting to speculate that the MVB may form from the PCR via a maturation process analogous to the one described above for animal cell MVBs. However, a structure resembling the plaque of MVBs was not observed on the dilations of the PCR. Furthermore, the dilation and its internal vesicles, when they occurred, were never labelled with CF. Therefore, it seems unlikely that the PCR labelled with CF would generate an unlabelled MVB which only later would become labelled.
A second possible explanation for the occurrence of the
dilation in the PCR involves membrane influx as opposed to efflux. It is possible that the dilation represents a large vesicle that has fused with the tubular portion of the PCR thereby adding membrane to this organelle. As membrane recently incorporated into the PCR it would therefore be reasonable to expect that the dilation would not be fully integrated into the PCR and thus could be unlabelled. The influx of membrane resulting from the proposed fusion event could in theory compensate in part or in total for any loss of membrane that might occur as a result of the formation of vesicles from the coated regions of the PCR. The ER and dictyosomes would seem to be the most likely candidates as the source for the incoming vesicle. The occurrence of smaller vesicles within some dilations remains unexplained by this hypothesis. However, it is possible that these vesicles are simply fixation artifacts (Willingham and Pastan 1984a).
The clear segregation of CF label in the PCR could be readily explained if there is an obstruction or barrier within the lumen of the PCR. This obstruction could prevent the redistribution of CF particles from the tubular portion of the PCR to the dilation. This idea has limited appeal. While it would explain the lack of CF particles within the dilation, it reveals nothing about the nature or function of the dilation and would not provide an explanation for the
occurrence of small vesicles within the dilation. Furthermore, there is no ultrastructural evidence to support this idea.
Finally, it is possible that both the dilations and their internal vesicles are artifacts of glutaraldehyde (i.e. chemical) fixation. To discount this explanation would require the use of some procedure other than chemical fixation; such as freeze substitution, to provide truly convincing evidence that the dilations and/or their internal vesicles are "real" structures rather than fixation artifacts. However, if the dilation is a fixation artifact, it is a very special one. The fixation process would have to simultaneously cause the swelling of the PCR, vesiculation of the membrane of the PCR where internal vesicles are observed and the redistribution of CF particles such that they are excluded from the forming dilation.
Clearly from the discussion above, further work is required to determine if the dilation is a "real" feature of the PCR and to determine its significance.
4.9 The MVB of Soybean Protoplasts
The multivesicular body (MVB) of soybean protoplasts is a distinct organelle which is very similar in morphology to MVBs of animal cells. In particular, MVBs contain internal vesicles, bear tubular protuberances, have a plaque on the cytoplasmic surface of their limiting membrane and are involved in the process of endocytosis (see section 3.5.4;
Willingham and Pastan 1984, Harding et al. 1985, Hornick et al. 1985).
The plaque on MVBs from animal cells sometimes has substructure which gives the plaque a lamellar appearance (Willingham and Pastan 1984a). Similar substructure in the plaque of MVBs from soybean protoplasts was only observed in conventionally double-stained preparations. The inability to detect substructure in other preparations is likely due to poor specimen contrast (in unstained sections) or overstaining of the specimen (in en bloc stained uranyl acetate stained preparations). The composition of the plaque is unknown. However, from immunocytochemistry experiments with animal cells, it is known that the plaque reacts poorly to antibodies against actin, clathrin, myosin, tubulin and fodrin (Willingham and Pastan, 1980, 1984a). The possibility that the plaque is composed of a modified or relatively inaccessible form of clathrin which would explain the plaque's poor reactivity to anti-clathrin antibodies can not be eliminated (Willingham and Pastan 1984a). The periodicity seen within the plaque (Fig. 101 and Willingham and Pastan 1984a) is certainly suggestive of a clathrin coat. Plaques appear to be present on some MVBs in cells of other plant species such as *Acer pseudoplatanus* (Nougarède and Lescure 1970; Fig. 4, 7 and possibly 10). In this case, their existence was not mentioned by the authors.
The MVBs of animal cells are generally considered to be a prelysosomal compartment but in some cases can contain acid hydrolases (e.g. Friend and Farquhar 1967, Morales et al. 1985) and thus are technically lysosomes. Some of the MVBs of soybean protoplasts contain peroxidase (Griffing and Fowke 1985) and acid phosphatase (L.R. Griffing and L.C. Fowke, unpublished results) and therefore constitute part of the lytic compartment in these protoplasts.
4.10 Formation of MVBs
Exactly how MVBs are formed is not clear. For animal cells, it has recently been proposed that MVBs may form by the fusion of plasma membrane-derived receptosomes (see Fig. 2 and section 1.2 of the Introduction, Willingham and Pastan 1984a) or that they may form from the vesicular portion of CURL via a maturation process (see section 4.7, Geuze et al. 1983a, b, 1985). In plants, dictyosomes have frequently been proposed as the site from which MVBs originate (e.g. Halperin and Jensen 1967, Nougarède and Lescure 1970, Dexheimer et al. 1975). This idea, however, is based entirely on circumstantial evidence. Exactly how the MVBs of soybean protoplasts are formed is not clear from the work reported here. It is possible that they could form from the dilations of the PCR via a maturation process. However, the morphological evidence from this research project is not totally consistent with this hypothesis. Specifically, while the MVBs are heavily labelled, the PCR dilations and their
internal vesicles are not. Alternatively, it is possible that MVBs might form via a two step process from dilations which sometimes occur in the cisternae of dictyosomes or in the ER. These dilations could pinch off dictyosomes or the ER to form the large vesicle of the MVB. The internal vesicles could then be produced by invaginations of the limiting membrane. Indeed, it was possible to find many profiles of vesicles which appeared to be budding off the limiting membrane into the lumen of the MVB (Fig. 13 and 105). CF might reach the developing MVB as a result of the fusion of labelled SVs (and/or possibly CVs) with the limiting membrane. Such vesicles might originate from the PCR or possibly dictyosomes. If a vesicle containing CF bound to its inner membrane surface fuses with a MVB, the CF would end up on the inner surface of the MVB limiting membrane, as observed. If the same region or an overlapping region of the limiting membrane then invaginated to form an internal vesicle, the CF would be present on the outer surface of the vesicle, as observed. Therefore the presence of CF on the inner surface of the limiting membrane and on the outer surface of the internal vesicles is consistent with the hypothesis that the internal vesicles form via an invagination process.
It should be noted that the idea that internal vesicles of MVBs form via invagination of the limiting membrane is
not new. Based on observations of MVBs in other plant cell systems, Halperin and Jensen (1967) and Dexheimer and co-workers (1975) have also proposed the same mechanism. Furthermore, it seems likely that the internal vesicles of animal cell MVBs are produced in the same manner (e.g. Smith and Farquhar 1966, Harding et al. 1984).
The fact that the internal vesicles of at least some MVBs of soybean protoplasts become labelled during post-fixation with ZIO is very interesting. If the preceding discussion concerning the formation of the internal vesicles is correct then the contents of these vesicles originate from the cytosol. The fact that some internal vesicles are labelled while the cytosol in general is not labelled suggests that selective uptake of cytosolic components into the MVB may be occurring. The labelling pattern of the MVBs is also significant because the internal vesicles of MVBs from animal cells are sometimes labelled by ZIO post-fixation (e.g. Loesch 1981, Weakley et al. 1981, Squier 1982). Thus ZIO labelling provides another criterion by which MVBs of soybean protoplasts are similar to MVBs of animal cells.
It was noted that elements of the ER were frequently observed in the vicinity of or appressed against MVBs in soybean protoplasts. These results may simply reflect chance associations between the two organelles in the cytoplasm or may represent a functional relationship between them in
which the transfer of material may be facilitated. Similar observations were made by Halperin and Jensen (1967) in embryogenic cells of carrot cell cultures.
4.11 MVBs and the Vacuoles of Soybean Protoplasts
Once labelled, the MVBs of animal cells apparently fuse with primary lysosomes producing secondary lysosomes where their contents are degraded by hydrolytic enzymes (Geuze et al. 1983b; Harding et al. 1985). In fact, degradation may begin within organelles which appear morphologically to be MVBs as in some cases the internal vesicles of MVBs become less distinct in appearance with increasing organellar age (Harding et al., 1985). In this way, the MVBs and lysosomes represent the degradative part of the generalized animal cell endocytic pathway. For soybean MVBs, it appears likely that they fuse with the central vacuole and smaller vacuolar profiles which may or may not be connected to the central vacuole. The presence of labelled membranous masses within vacuoles after prolonged CF incubation is consistent with this interpretation. These masses are similar in appearance and size to the contents of some MVBs from the same protoplasts (Fig. 107-109). They resemble coalescing vesicles with individual vesicles either being less distinct in appearance or not being recognizable at all. The fact that the actual fusion event between MVBs and vacuoles was not recorded does not invalidate this hypothesis. Such
fusion events may be rare and may occur quite rapidly therefore making it difficult or impossible to record. Since MVBs of soybean protoplasts contain hydrolytic enzymes (see section 4.9) and the central vacuole and smaller vacuoles are considered to be components of the lytic compartment of plant cells (Matile 1975), it seems likely that these organelles are functionally analogous to the MVBs and lysosomes, respectively, of the generalized animal cell endocytic pathway.
The successive labelling of MVBs and the membranous masses within vacuoles with CF may have important implications. This observation, in combination with the short term and intermediate term CF uptake results (sections 3.3, 3.4, 4.4 and 4.5), provides convincing evidence for a pathway including an intermediate set of membrane bounded organelles which results in the delivery of substances from the extracellular environment to the major lysosomal component of plant cells, the vacuole. The importance of this pathway is unclear but likely involves the degradation of macromolecules.
4.12 Possible Alternative Fate for MVBs
It should be noted that in at least one specialized type of animal cell, the reticulocyte, MVBs fuse with the plasma membrane releasing their contents including their internal vesicles to the extracellular medium (Harding et al. 1983, 1984). The internal vesicles of these MVBs appear to contain
transferrin receptors and therefore the exocytosis of MVBs provides a mechanism for the developmental loss of the receptors from these cells (Harding et al. 1984). The possibility that some MVBs from soybean protoplasts may behave in a similar manner can not be ruled out. However, it would be very difficult to record such a fusion event using protoplasts. Fusion of MVBs with the plasma membrane might release some or all of the internal vesicles into the surrounding medium. Small vesicles similar to those of MVBs were observed on the plasma membrane but they could represent remnants of protoplasts broken during protoplast isolation and labelling steps (Fig. 111 and 113). The presence of plasma membrane invaginations containing small vesicles in intact soybean cells (Fig. 147) is consistent with the idea that MVBs could fuse with the cell surface. Plant cells might utilize such a mechanism to remove materials from the cytoplasm. Further work is required to evaluate this possibility. It is of interest to note that Halperin and Jensen (1967), Nougarède and Lescure (1970) and Dexheimer and co-workers (1975) proposed, on the basis of ultrastructural evidence, that MVBs originate from the cisternae of dictyosomes and fuse with the plasma membrane releasing their contents including the internal vesicles.
4.13 Labelling of the Dictyosomes and Membrane Recycling
The dictyosomes of soybean protoplasts also become
labelled with CF. The amount of labelling seemed quite variable from one experiment to another. CF particles within the dictyosome could be found distributed within several cisternae and had a general peripheral distribution. The occurrence of dictyosomal labelling is a highly significant result as it has important implications concerning the function of the Golgi apparatus of plant cells. The dictyosomes of plant cells have been shown to be involved in exocytotic events (Brown et al. 1970; Brown and Romanovicz 1975; Rougier 1976). Vesicles derived from the dictyosomes fuse with the plasma membrane thereby delivering their membrane constituents to the plasma membrane and releasing their luminal contents (e.g. cell wall matrix polysaccharides- reviewed by Willison 1981) to the periplasmic space. In this manner, a significant amount of membrane is added to the plasma membrane. There have been a number of estimates of this rate of flow of membrane (reviewed in Robinson and Kristen 1982). For instance, in the alga Pleurochrysis, scale exocytosis results in the addition of approximately 1% of the plasma membrane surface area per minute to the plasma membrane (Brown et al. 1970). Other estimates include 3% per minute for the gland cells of Mimulus tilingii (Schnepf and Busch 1976) and 9.4% per minute in the ovary gland cells of Aptenisi cordifolia (Kristen and Lockhausen 1982). In none of these cases does the cell under study enlarge at a rate which would require
this rate of increase in plasma membrane surface area. Therefore, the extra membrane which is continually being added to the plasma membrane must somehow be recycled back into the cell (Robinson and Kristen 1982). One way in which membrane recycling could be achieved is by endocytosis. Plasma membrane derived vesicles could deliver their membrane and luminal contents or at least a fraction of them to the dictyosomes either directly by fusing with the cisternae of the dictyosomes or indirectly by fusion with some organelle such as the PCR followed by subsequent membrane transfer to dictyosomes. The labelling of dictyosomes with CF is consistent with this hypothesis and therefore provides excellent evidence for the occurrence of direct or indirect membrane recycling, involving coated membranes from the plasma membrane to the dictyosomes of soybean protoplasts.
4.14 Dictyosomes and Molecular Repair
The Golgi apparatuses of both animal and plant cells have been shown to contain a number of glycosyltransferases (reviewed by Farquhar 1986; Robinson and Kristen 1982) which are involved in the post-translational modification of newly synthesized proteins (reviewed by Farquhar 1986; Chrispeels 1984). It is possible that these enzymes may also function to repair damaged oligosaccharide chains on recycling glycoproteins. For example, partial resialylation of
asialotransferrin type 3 occurred during diacytosis by rat hepatocytes (Regoeczi et al. 1982). In addition, it has been demonstrated in K 562 (animal) cells that cell surface transferrin receptors which have been asialyated by a neuraminadase treatment are resialyated following internalization (Snider and Rogers 1984). It is thought that these repair processes might be mediated by the Golgi apparatus (Regoeczi et al. 1982; Snider and Rogers 1984; Breitfeld et al. 1985). It is conceivable that the dictyosomes of the plant cell Golgi apparatus could function in a similar manner. Therefore, the labelling of the dictyosomes with CF may indicate the occurrence of a pathway for the reparation of damaged membrane constituents in addition to a possible membrane recycling pathway.
4.15 Membrane Recycling and Membrane Sorting
While direct membrane recycling from the plasma membrane to the dictyosomes is certainly possible, there are reasons for arguing that the plasma membrane derived vesicles would fuse with another organelle, likely the PCR, followed by the subsequent transfer of membrane to the dictyosomes. It seems likely that the occurrence of endocytosis in soybean protoplasts involves more than mere membrane recycling since MVBs and vacuoles, lytic compartments of the protoplasts, also become labelled. Thus a significant portion of the membrane material and/or lumenal contents internalized via endocytosis is destined to be degraded rather than to be
recycled. This point is important in view of results which indicate that the initial endocytotic events in the protoplasts appear to be mediated exclusively by CPs. It seems unlikely that distinct populations of CPs exist on the plasma membrane such that one population is involved in the recycling of plasma membrane to the dictyosomes while a second population of CPs is involved in the delivery of material to the lytic compartment. At least, this does not appear to be the case in animal cells where different types of receptors with different fates following internalization via endocytosis occur randomly at the cell surface and in the CPs and CVs, i.e. the different receptors could be found within the same CP or CV and the distribution of the various types of receptors were independent of one another (Geuze et al. 1984). Furthermore, the sorting of different receptor-ligand complexes from one another and the sorting of the components of dissociated receptor-ligand complexes occurs intracellularly probably within CURL or some analogous endosome (Geuze et al. 1984; Willingham et al. 1984). Therefore, it seems more reasonable that membrane constituents destined to be recycled are co-localized within the same CP with constituents destined to be degraded and that sorting of these constituents occurs sometime after internalization. Based on its similarity to CURL and trans reticular Golgi, the PCR would seem to be the most likely
candidate as the organelle where sorting events occur within soybean protoplasts. Constituents to be recycled could then be transferred from the PCR to dictyosomes either via PCR-derived vesicles or via transitory direct membrane connections between the two organelles. It is also conceivable that some of the recycling membrane constituents might move directly from the PCR to the plasma membrane via PCR-derived vesicles. Constituents to be degraded by the lytic compartment may be transferred to MVBs via PCR-derived vesicles or perhaps possibly by the direct maturation of PCR dilations into MVBs. The PCR-derived vesicles might initially start out as CVs which could form from the coated tips of the PCR tubules. These vesicles could then shed their coat.
4.16 Dictyosome Heterogeneity and Membrane Flow
The labelling patterns seen for dictyosomes in protoplasts post-fixed with ZIO were quite interesting. The variation in the distribution and amount of label was considerable between dictyosomes, even dictyosomes within the same protoplast. In some cases, a clear gradient or polarity of labelling intensity was seen across a dictyosome but this was not true in other cases. Although the exact basis for ZIO labelling is not known (e.g. Pellegrino de Iraldi 1977; Rodriguez and Giménez 1981 and references cited therein), these results suggest that dictyosomes are not of identical composition and/or physiological state. It is
conceivable that different dictyosomes within a protoplast may be performing different functions or may be performing the same functions but at different rates. This would help explain the observed variability in CF labelling. Those dictyosomes most actively involved in exocytosis/endocytosis would be more heavily labelled with CF while those dictyosomes less actively involved or uninvolved in these processes would be poorly labelled or unlabelled.
The ZIO labelling of dictyosomes was also interesting because evidence from both thick and thin sections suggested that direct membrane connections existed between the ER and dictyosomes. These results are consistent with other recent work (e.g. Juniper et al. 1982), including ZIO labelling results (e.g. Harris and Oparka 1983), and indicate that the transport of material, in particular protein that is newly synthesized on the ER, may occur via these connections as opposed to via transition vesicles. The demonstration of two types of ER, cisternal and tubular, and the labelling of the nuclear envelope are also consistent with other recent work making use of a similar ZIO post-fixation procedure (e.g. Harris and Oparka 1983).
4.17 Nature of Internalized Material
The preceding discussion has suggested that the endocytotic pathway of soybean protoplasts is involved in the recycling, degradation and perhaps reparation of
membrane constituents and luminal contents of plasma membrane-derived endocytotic vesicles. Perhaps at this point it would be a good idea to briefly consider the possible nature of the internalized material. As indicated earlier in this discussion, CF is a non-specific marker for adsorptive endocytosis. Furthermore, it is not a substance with which plant cells would normally come in contact. Therefore, the fact that CF particles are internalized via endocytosis does not reveal anything about the nature of the substances which would normally be internalized via this process or of the biological significance of this pathway.
Membranes consist of a bimolecular leaflet composed of lipids and associated integral and peripheral proteins. The lipid composition of the plasma membrane is thought to effect the physical properties of the membrane such as fluidity. Therefore, the ability of a protoplast or cell to recycle or degrade the lipid components of the plasma membrane would enable it to regulate and/or alter the composition of the membrane perhaps in response to changes in its physiological function or to changes in its environment. The demonstrated functions of the protein component of membranes include the catalysis of chemical reactions (by enzymes) and the exchange or transport of ions across the membrane (Leonard and Hodges 1980; Leonard and Rayner 1986). The ability to recycle, degrade or repair these proteins via endocytosis would enable a protoplast or
cell to regulate processes such as cell wall synthesis in response to changes in the environment or could provide a means to produce developmental changes in the protoplast or cell. It is conceivable that some of the proteins in the plasma membrane are receptors for specific molecules or classes of molecules. For example, there is at least some evidence that soybean membranes may contain a specific receptor for mycolaminaran, the phytoalexin elicitor from *Phytophthora*, which elicits the production of the phytoalexin, glyceollin, in soybean tissues (Yoshikawa et al., 1983). Therefore, it is possible that receptor-mediated endocytosis could occur in plant protoplasts or cells. To demonstrate the occurrence of this process would require the use of a specific ligand in uptake experiments for which a well characterized receptor exists. (Currently these conditions can not be met in any higher plant cell system.)
4.18 Overview of the Endocytotic Pathway in Soybean Protoplasts
The various components of the endocytotic pathway in soybean protoplasts are shown in Fig. 148. The position of the components in this figure is based upon the preceding discussion of the CF uptake results. Therefore, while Fig. 148 depicts a representation of the endocytotic pathway which is compatible with the experimental results with soybean protoplasts, other interpretations of these results
Fig. 148. Diagrammatic representation of the endocytotic pathway in soybean protoplasts. Material (e.g. CF) is internalized by a CP on the plasma membrane which pinches off the plasma membrane to form a CV. The CV sheds its coat to become a SV which in turn fuses with a PCR. From the PCR, the internalized material and membrane may be recycled directly to the protoplast surface (broken lines) within a vesicle although experimental evidence for this pathway was not obtained. The vesicle may originate from a coated tip of a PCR tubule and may shed its coat before reaching the plasma membrane. Second, the material and membrane may be transferred from the PCR to the dictyosomes of the Golgi apparatus. Transport to the dictyosome may occur via a vesicle in a manner analogous to transport from the PCR to the plasma membrane or may occur via transitory direct membrane connections between the two organelles. From the dictyosome, the material and membrane may be returned to the protoplast surface within a vesicle derived from this organelle. Third, the material and membrane may be transferred from the PCR to a MVB. Transfer may be mediated by vesicles or may result by the transformation of a PCR dilation into a MVB. The MVB may then fuse with a vacuole where the material and membrane may be degraded. Note that this pathway is very similar to that for receptor-mediated endocytosis in animal cells, Fig. 1.
CP
CV
SV
PCR
MVB
VACUOLE
148
DICTYOSOME (GOLGI)
are possible and may be equally valid. In this pathway, CF binds to the surface of the plasma membrane including CPs. The CPs pinch off the plasma membrane via the process of invagination to produce CVs. The CVs rapidly shed their coat and become SVs. The SVs fuse with the PCR delivering CF to this organelle. From the PCR, CF particles face three possible fates. First, these particles may be delivered directly back to the plasma membrane via vesicles derived from the PCR. This would represent direct membrane recycling from the PCR to the plasma membrane. Experimental evidence was not obtained for this pathway. Second, these particles may be delivered to the dictyosomes either via vesicles derived from the PCR or through transitory direct membrane connections between the PCR and dictyosome cisterna(e). CF could be redistributed to additional cisternae via vesicles originating from the periphery of the cisternae. This part of the pathway would result in the replenishment of the dictyosomal cisternae via the recycling of plasma membrane constituents and would complete a cycle of membrane flow originating with the dictyosomes. In theory at least, this part of the pathway could also result in the reparation of damaged plasma membrane constituents, particularly glycoproteins, prior to their return to the plasma membrane. From the dictyosome, membrane constituents may be recycled to the plasma membrane via exocytotic vesicles. Third, some CF particles could be delivered from the PCR to MVBs. This
transfer could be mediated by vesicles derived from the PCR or could result from the maturation of the PCR into a MVB. This would represent the degradative part of the pathway in which some plasma membrane constituents and the luminal contents of plasma membrane-derived vesicles are delivered to the lytic compartment of the protoplast. In the MVB some initial degradation of material may occur, however, ultimately the contents of the MVBs are delivered to the vacuoles as a result of MVB-tonoplast membrane fusion events. In the vacuoles, further degradation takes place completing the degradative part of the endocytotic pathway.
4.19 Recent Demonstrations of Endocytosis in Protoplasts
Recently, Joachim and Robinson (1984) have demonstrated the occurrence of endocytosis in protoplasts from *Vicia faba* leaf cells. Their results provide at least partial confirmation of the endocytotic pathway in soybean protoplasts. In *Vicia* protoplasts, internalization of CF occurred via CPs and CVs although Joachim and Robinson did not rule out the possible involvement of uncoated pits. Smooth or uncoated vesicles also became labelled as did "peripheral tubules" (see below) on the trans side of the dictyosome. Finally, small vacuoles became labelled. Some micrographs suggested that these vacuoles might be formed directly from the plasma membrane. It is not clear whether these vacuoles are equivalent to the small vacuolar profiles which are labelled in soybean protoplasts. Joachim and
Robinson described structures which in their micrographs appear to be PCR. They refer to these structures as "peripheral tubules" of dictyosomes. The "peripheral tubules" of the dictyosome which were labelled with CF were not the PCR type tubules but rather appeared to be the dilated or vesiculated tip of a dictyosomal cisterna, at least in the example illustrated in their paper. Because Joachim and Robinson do not make a distinction between these two types of tubules, it is unclear whether the PCR type tubules are labelled in their protoplast preparations.
Another difference between the results of Joachim and Robinson and the results reported here for soybean protoplasts is that they make no reference to labelling of MVBs or the central vacuole. Overall, the results of Joachim and Robinson confirm the occurrence of endocytosis in higher plant (angiosperm) protoplasts and the involvement of CPs and CVs in this process. However, some of the organelles involved in endocytosis in soybean protoplasts may not be involved in the same process in *Vicia* protoplasts.
More recently, Hillmer and co-workers (1986) have demonstrated the occurrence of endocytosis in bean (*Vicia faba* L.) leaf protoplasts and protoplasts from carrot suspension culture cells using a number of lectin-colloidal gold conjugates. Structures labelled during endocytosis included CPs, CVs, the Golgi apparatus (dictyosomes), MVBs, small vacuoles and the PCR. Interestingly, label was never
observed in the central vacuole. Label was seen only once in the ER of a bean leaf protoplast and never in carrot protoplasts. These results are similar to those reported here for soybean protoplasts, the major difference being the lack of labelling in the central vacuoles. This difference may be a result of the incubation times used by Hillmer and co-workers. The maximum incubation time used by this group was 1h. While an incubation of this length is sufficient for the detection of central vacuole labelling in soybean protoplasts, it is possible that longer incubations might be needed in other plant systems. It is also interesting to note, in particular, that for bean leaf protoplasts MVBs are labelled by lectins (Hillmer et al. 1986) but apparently not by CF (Joachim and Robinson 1984). These results suggest that either different relatively non-specific markers have different intracellular fates or that the CF uptake results were incomplete.
4.20 Coated Membrane Structures and Endocytosis in Plant Cells
Now that it is apparent that coated membrane organelles are involved in the process of endocytosis in protoplasts, the question arises as to whether these organelles are involved in the same process in normal turgid cells? This question is not as trivial as it may seem. Evidence for the occurrence of endocytosis in turgid cells is as scarce as
for protoplasts. In fact, it has been argued on theoretical grounds that endocytosis (pinocytosis) in turgid plant cells could not be a major uptake route of small-molecular-weight solutes (Cram 1980). It was thought that the energy required in vesicle formation to overcome turgor would be too excessive for endocytosis to be the mechanism for large scale uptake of material. Furthermore, results from studies on protoplasts may not be directly applicable to normal turgid plant cells. The physiological state of protoplasts and cells tends to be very different. In fact, at least for some characteristics, protoplasts appear to mimic the response of stressed plant cells (e.g. Mieth et al. 1986; Rao et al. 1984; Cress 1982; Fleck et al. 1982). Therefore, it is possible that endocytosis may not occur in normal turgid plant cells, or may occur only in certain specialized cell types or in cells that have been exposed to some sort of physiologically stressful situation or environment. With the previous statement in mind, the best evidence for the occurrence of endocytosis comes from studies of plant pathogens (e.g. Newcomb 1976) and plant-microbe symbioses. In these instances, microorganisms are internalized via plasma membrane-derived vesicles. This process does not appear to involve coated membrane organelles. The use of heavy metal salt solutions on various plant tissues also provided some evidence for the occurrence of endocytosis (e.g. Wheeler and Hanchey 1971; Wheeler et al. 1972; Robards
and Robb 1972; Robards and Robb 1974; Peterson et al. 1986). However, until very recently (see below), this process appeared to be mediated by smooth vesicles derived from the plasma membrane as opposed to CPs and CVs. Since heavy metals tend to be toxic to plant tissues, it is also unclear whether the endocytotic events revealed by exposure to these elements occur under normal conditions or whether the heavy metals stimulate the occurrence of these events (e.g. Wheeler and Hanchey 1971).
Very recently, Hübner and co-workers (1985) have provided evidence for the occurrence of endocytosis in maize (*Zea mays*) root cap cells using incubations with lanthanum and lead salt solutions. Lanthanum deposits were found in CPs, CVs, SVs and large vesicles present at the trans-pole of dictyosomes. Lead deposits were observed in CPs, smooth pits, CVs, SVs, large trans-pole dictyosome vesicles, the cisternal stack of the dictyosomes, coated portions of the PCR, small vacuoles and occasionally in the lumen of the ER. These results were interpreted to be evidence for coated pit/coated vesicle-mediated endocytosis and for the direct recycling of plasma membrane to the Golgi apparatus. Although it was conceivable that the presence of high concentrations of heavy metal salts might induce or stimulate endocytosis, it was thought that the presence of CPs on the plasma membrane and on the Golgi apparatus as
well as the presence of coated vesicles in control cells suggested that it was reasonable to assume that endocytosis could occur in undisturbed *in vivo* conditions.
4.21 Lanthanum and CF Uptake by SB-1 Cells
In this project, whole SB-1 soybean cells were incubated in solutions containing lanthanum chloride or CF particles in the hope of demonstrating the occurrence of endocytosis in cells. Unambiguous evidence of endocytosis was not obtained in these preliminary experiments. Lanthanum deposits and CF particles were seen on and throughout the cell wall, on the plasma membrane and within the cytoplasm. Lanthanum deposits and CF particles within the cytoplasm appeared to be located primarily in elements of the ER or within vesicular profiles which likely were cross-sections through the ER. Particles of CF were also found free within the cytoplasm. In some instances, these particles may have been within vesicles that were difficult to see against the electron-dense background of the ground substance. It was not apparent how the ER became labelled with these markers. It is possible that plasma membrane-derived vesicles may have carried these markers to the ER and that these vesicles were difficult to detect. If this was the case then the endocytotic pathway in SB-1 cells is quite different from the pathway in SB-1 protoplasts. Profiles in SB-1 protoplasts which were readily identifiable as elements of the ER were never labelled with CF. Alternatively, the ER in
intact SB-1 cells may be connected to the plasma membrane (see Fig. 143 for evidence). If this is the case then lanthanum ions and CF particles which had diffused through the cell wall may enter the lumen of the ER through the connections between the plasma membrane and the ER. In the case of lanthanum ions, these ions would be precipitated out wherever the ionic environment of the ER would favour this process. It is conceivable that the proposed plasma membrane-ER connections might be broken or sealed off as a result of the stress incurred during protoplast formation thereby explaining why ER labelling was observed in SB-1 cells and not in protoplasts from SB-1 cells.
The inability to detect CF particles in other cytoplasmic organelles may be due to the low level at which CF passed through the cell wall. Since very little CF was found bound to the plasma membrane (relative to the amount observed for protoplasts), it would be logical to expect that the amount of internalized CF would also be reduced by an equivalent degree. Therefore, it is conceivable that only a very few organelles would be labelled and that these organelles would only contain one or a few particles of CF. It would be quite easy to miss these organelles when scanning unstained thin sections of cells in the electron microscope.
With respect to lanthanum uptake, it is important to
note that one can only directly visualize, in the electron microscope, lanthanum which has precipitated out to form an electron-dense deposit. (It is a good practise, if one's electron microscope has the capability, to verify that these deposits contain lanthanum by using X-ray microanalysis.) Therefore, it is conceivable that other locations in the cell contain lanthanum which has not been precipitated. Furthermore, it is possible that some precipitates which have formed during the lanthanum chloride incubation could have been redissolved during the washes and other processing steps which followed the incubation. In fact, some authors (Revel and Karnovsky 1967) have recommended that lanthanum be included in every step of the processing sequence. This policy was not adopted here as it would not have been possible to argue convincingly that the location of any of the lanthanum deposits was the result of endocytosis.
In view of the above discussion, it is important to emphasize at this point that the lanthanum and CF uptake results are of a very preliminary nature and therefore do not eliminate the possibility that endocytosis can occur in whole SB-1 soybean cells. A wide range of parameters need to be explored before one can confidently rule out the occurrence of endocytosis in these cells. In fact, in view of Hübner's and co-workers's (1985) results it seems more likely that the occurrence of endocytosis will eventually be demonstrated.
The mere fact that CF (M.W. approx. 750,000) was able to pass through the cell wall of SB-1 cells is in itself very interesting. The cell wall of higher plant cells has generally been considered to be a significant barrier to high molecular weight compounds as these walls contain very small pores. The size of pores in the cell wall have been estimated by Carpita and co-workers (1979) to be between 3.7 and 5.2 nm in diameter. It is generally thought that the penetration through the cell wall of globular proteins with molecular weights greater than 17,000 to 60,000 is greatly impeded (Carpita et al. 1979; Tepfer and Taylor 1981). The low level at which CF particles pass through the cell wall of SB-1 cells bears testimony to this belief and yet at the same time demonstrates that given enough time these compounds do make it through the cell wall. The observation that NF particles do not appear to penetrate the cell wall during an 1 h incubation is also very interesting. In view of the CF results, it seems likely that the explanation for this lack of penetration by NF is related to net molecular charge. The cell walls of higher plant cells contain a large number of residues of carboxylated derivatives of monosaccharides. The negative charges on these residues may serve to repel the negatively charged NF particles so that these particles do not penetrate the cell wall under the conditions in which the experiment was performed. In
contrast, CF particles with their net positive charge would be electrostatically attracted to the acidic residues in the cell wall and could then pass through the cell wall as a result of mass action.
4.22 Other Functions for CVs
It is important to note that the results of this project and those of Robinson's laboratory (Joachim and Robinson 1984; Hübner et al. 1985; Hillmer et al. 1986) which demonstrate the involvement of CVs in endocytosis do not rule out the possibility that CVs may be involved in other membrane transport processes. It is possible as suggested by the work of others (e.g. Bonnett 1969; Robertson and Lyttledon 1982; Nakamura and Miki-Hiroshige 1982) that CVs may play some role in exocytotic processes such as the transport of cell-wall precursors or enzymes to the cell surface. It is conceivable that subpopulations of CVs which are functionally and perhaps morphologically different could exist within plant cells. With respect to this possibility, it should be noted that it has been reported that Golgi-associated coated membranes seemed to be of smaller diameter than those associated with the plasma membrane of protoplasts from cultured tobacco cells (Van der Valk and Fowke, 1981).
4.23 Future Prospects
Further progress in demonstrating the occurrence of endocytosis in higher plant protoplasts and cells will
depend on the development of better membrane markers which are not toxic to the plant material and which can be used under normal culture or growth conditions. Ideally, these markers would bind to a specific molecule on the plasma membrane such as a receptor and would be small enough to pass unimpeded through the cell wall. It would also be beneficial if the marker was electron-dense so that it could be easily visualized in the electron microscope. However, this last property is not essential as one could employ autoradiographic or immunocytochemical protocols to determine the distribution of the marker in cells and protoplasts.
Further characterization of the organelles in the endocytotic pathway of soybean protoplasts would also be useful. At the ultrastructural level, a comparison of soybean protoplasts which have been freeze substituted with those which have undergone chemical (e.g. glutaraldehyde) fixation would be useful in determining whether or not the dilations and their internal vesicles are "real" features of the PCR. Furthermore, the use of cytochemical techniques, such as enzyme and 3-(2,4-dinitroanilino)-3'-amino-N-methyldipropylamine (DAMP) cytochemistry, could provide interesting information concerning the composition and the contents of the organelles involved in endocytosis in soybean protoplasts. For example, DAMP cytochemistry, used
for the identification of acidic organelles in animal cells (Anderson et al. 1984), could possibly be used to identify acidic organelles in soybean protoplasts. Such results could provide important clues to the function(s) of organelles in this endocytotic pathway as the dissociation of some endocytosed receptor-ligand complexes is thought to be mediated by acidic pHs and is thought to occur in the acidic organelles of animal cells.
Biochemical techniques could also be used to further characterize the organelles of the endocytotic pathway in soybean protoplasts. As has been done for animal cell systems (Hornick et al. 1985), cell fractionation protocols could be used to isolate fractions enriched for specific organelles such as MVBs. Furthermore, density shift protocols could possibly be used to separate organelles active in endocytosis from morphologically similar but endocytotically inactive organelles (Courtoy et al. 1984; Quintart et al. 1984). Enzyme and other assays on these cell fractions could be used to support results concerning the composition of various organelles obtained using cytochemical procedures (see above).
4.24 Summary
In summary, using CF, it was possible to demonstrate the occurrence of endocytosis in protoplasts from SB-1 soybean suspension culture cells. The endocytotic pathway included CPs, CVs, SVs, the PCR, dictyosomes, MVBs, small vacuoles
and the central vacuole. This is the first demonstration of the involvement of the PCR in endocytosis. The PCR appears to be a distinct organelle although in many cases it is associated with and sometimes possibly physically connected to the ER and the dictyosomes of the Golgi apparatus. The ultimate fate of some of the internalized CF label appears to be delivery, via MVBs, to the major lytic or lysosomal compartment of the plant cell, namely, the vacuole. It is proposed that the endocytotic pathway is involved in the recycling, degradation and possibly the repair of membrane constituents and lumenal contents of plasma membrane-derived vesicles. Unambiguous demonstration of endocytosis in whole SB-1 cells was not achieved using lanthanum chloride, NF or CF as markers of endocytosis. However, recent developments in this field of research (Hübner et al. 1985) suggest that an endocytotic pathway similar to the one described here for soybean protoplasts may function in at least some types of higher plant cells.
Literature Cited
Anderson RGW, Falck JR, Goldstein JL, Brown MS (1984) Visualization of acidic organelles in intact cells by electron microscopy. Proceedings of the National Academy of Sciences (USA) 81: 4838-4842.
Barton R (1964) Electron microscope studies on the uptake of ferritin by plant roots. Experimental Cell Research 36: 432-434.
Bonnett HT (1969) Cortical cell death during lateral root formation. The Journal of Cell Biology 40: 144-159.
Breitfeld PP, Simmons CF Jr., Strous GJAM, Geuze HJ, Schwartz AL (1985) Cell biology of the asialoglycoprotein receptor system: a model of receptor-mediated endocytosis. International Review of Cytology 97: 47-95.
Bretscher MS, Thomson JN (1985) The morphology of endosomes in giant HeLa cells. European Journal of Cell Biology 37: 78-80.
Brown RM Jr., Franke WW, Kleinig H, Falk H, Sitte P (1970) Scale formation in Chrysophycean algae. I. Cellulosic and noncellulosic wall components made by the Golgi apparatus. The Journal of Cell Biology 45: 246-271.
Brown RM Jr., Romanovicz DK (1975) Biogenesis and structure of Golgi-derived cellulosic scales in Pleurochrysis. I. Role of the endomembrane system on scale assembly and exocytosis. Applied Polymer Symposium 28: 537-585.
Carpita N, Sabularse D, Montezinos D, Delmer DP (1979) Determination of the pore size of living plant cells. Science 205: 1144-1147.
Chrispeels MJ (1984) Biosynthesis, processing and transport of storage proteins and lectins in cotyledons of developing legume seeds. Philosophical Transactions of the Royal Society of London B304: 309-322.
Cocking EC (1966a) Electron microscope studies on isolated plant protoplasts. Zeitschrift für Naturforschung 21b: 581-584.
Cocking EC (1966b) An electron microscopic study of the initial stages of infection of isolated tomato fruit protoplasts by tobacco mosaic virus. Planta 68: 206-14.
Courtoy PJ, Quintart J, Baudhuin P (1984) Shift of equilibrium density induced by 3,3'-diaminobenzidine cytochemistry: a new procedure for the analysis and purification of peroxidase-containing organelles. The Journal of Cell Biology 98: 870-876.
Cram WJ (1980) Pinocytosis in plants. The New Phytologist 84: 1-17.
Cress DE (1982) Osmotic stress inhibits thymidine incorporation into soybean protoplast DNA. Plant Cell Reports 1:186-188.
Davey MR, Frearson EM, Power JB (1976) Polyethylene glycol-induced transplantation of chloroplasts into protoplasts: an ultrastructural assessment. Plant Science Letters 7: 7-16.
Davey MR, Power JB (1975) Polyethylene glycol-induced uptake of micro-organisms into higher plant protoplasts: an ultrastructural study. Plant Science Letters 5: 269-274.
Depta H, Robinson DG (1986) The isolation and enrichment of coated vesicles from suspension-cultured carrot cells. Protoplasma 130: 162-170.
Dexheimer J, Jacquemin M, Gadal P (1975) Sur la présence de corps multivésiculaires dans les cellules de la cécidie de l'Orme (Ulmus campestris L.), provoquée par Tetraneura ulmi de Geer. Comptes Rendus Hebdomadaires des Séances de l'Académie des Sciences Paris Series D 280: 173-176.
Drew MC, McLaren AD (1970) The effect of histones and other basic molecules on cell permeability and elongation of barley roots. Physiologia Plantarum 23: 544-560.
Drew MC, Sear J, McLaren AD (1970) Entry of basic macromolecules into barley roots. American Journal of Botany 57: 837-843.
Fan JY, Carpentier J-L, Gorden P, Van Obberghen E, Blackett NM, Grunfeld C, Orci L (1982) Receptor-mediated endocytosis of insulin: role of microvilli, coated pits, and coated vesicles. Proceedings of the National Academy of Sciences (USA) 79: 7788-7791.
Farquhar MG (1985) Progress in unraveling pathways of Golgi traffic. Annual Review of Cell Biology 1: 447-488.
Farquhar MG, Palade GE (1981) The Golgi apparatus (complex)-(1954-1981) from artifact to center stage. *The Journal of Cell Biology* 91: 77s-103s.
Fine RE, Ockleford CD (1984) Supramolecular cytology of coated vesicles. *International Review of Cytology* 91: 1-43.
Fleck J, Durr A, Fritsch C, Vernet T, Hirth L (1982) Osmotic-shock 'stress proteins' in protoplasts of *Nicotiana sylvestris*. *Plant Science Letters* 26: 159-165.
Fowke LC (1982) Electron microscopy of protoplasts. In *Plant tissue culture methods* (ed. LR Wetter and F Constabel). 2nd revised edition. National Research Council of Canada, Saskatoon, 72-82.
Fowke LC (1986) The plasma membrane of higher plant protoplasts. In *Hormones, receptors and cellular interactions in plants* (ed. CM Chadwick and DR Garrod). Cambridge University Press, Cambridge, 217-239. (Intercellular and intracellular communication Vol 1).
Franke WW, Herth W (1974) Morphological evidence for de novo formation of plasma membrane from coated vesicles in exponentially growing cultured plant cells. *Experimental Cell Research* 89: 447-451.
Friend DS, Farquhar MG (1967) Functions of coated vesicles during protein absorption in the rat vas deferens. *The Journal of Cell Biology* 35: 357-376.
Fukunaga Y, Nagata T, Takebe I, Kakehi T, Matsui C (1983) An ultrastructural study of the interaction of liposomes with plant protoplasts. *Experimental Cell Research* 144: 181-189.
Gamborg OL (1982) Callus and cell culture. In *Plant tissue culture methods* (ed. LR Wetter and F Constabel). 2nd revised edition. National Research Council of Canada, Saskatoon, 1-9.
Geuze HJ, Slot JW, Strous GJAM, Hasilik A, Von Figura K (1985) Possible pathways for lysosomal enzyme delivery. *The Journal of Cell Biology* 101: 2253-2262.
Geuze HJ, Slot JW, Strous GJAM, Lodish HF, Schwartz AL (1983a) Intracellular site of asialoglycoprotein receptor-ligand uncoupling: double-label immunoelectron microscopy during receptor-mediated endocytosis. Cell 32: 277-287.
Geuze HJ, Slot JW, Strous GJAM, Peppard J, Von Figura K, Hasilik, Schwartz AL (1984) Intracellular receptor sorting during endocytosis: comparative immunoelectron microscopy of multiple receptors in rat liver. Cell 37: 195-204.
Geuze HJ, Slot JW, Strous GJAM, Schwartz AL (1983b) The pathway of the asialoglycoprotein-ligand during receptor-mediated endocytosis: a morphological study with colloidal gold/ligand in the human hepatoma cell line, Hep G2. European Journal of Cell Biology 32: 38-44.
Griffing LR, Cutler AJ, Shargool PD, Fowke LC (1985) Isoelectric focusing of plant cell protoplasts. Plant Physiology 77: 765-769.
Griffing LR, Fowke LC (1985) Cytochemical localization of peroxidase in soybean suspension culture cells and protoplasts: intracellular vacuole differentiation and presence of peroxidase in coated vesicles and multivesicular bodies. Protoplasma 128: 22-30.
Halperin W, Jensen WA (1967) Ultrastructural changes during growth and embryogenesis in carrot cell cultures. Journal of Ultrastructure Research 18: 428-443.
Harding C, Heuser J, Stahl P (1983) Receptor-mediated endocytosis of transferrin and recycling of the transferrin receptor in rat reticulocytes. The Journal of Cell Biology 97: 329-339.
Harding C, Heuser J, Stahl P (1984) Endocytosis and intracellular processing of transferrin and colloidal gold-transferrin in rat reticulocytes: demonstration of a pathway for receptor shedding. European Journal of Cell Biology 35: 256-263.
Harding C, Levy MA, Stahl P (1985) Morphological analysis of ligand uptake and processing: the role of multivesicular endosomes and CURL in receptor-ligand processing. European Journal of Cell Biology 36: 230-238.
Harris N (1986) Organization of the endomembrane system. Annual Review of Plant Physiology 37: 73-92.
Harris N, Chrispeels MJ (1980) The endoplasmic reticulum of mung bean cotyledons: quantitative morphology of cisternal and tubular ER during seedling growth. Planta 148: 293-303.
Harris N, Oparka KJ (1983) Connections between dictyosomes, ER and GERL in cotyledons of mung bean (Vigna radiata L.). Protoplasma 114: 93-102.
Hermo L, Clermont Y, Lalli M (1985) Intracellular pathways of endocytosed tracers in Legdig cells of the rat. Journal of Andrology 6: 213-224.
Hillmer S, Depta H, Robinson DG (1986) Confirmation of endocytosis in higher plant protoplasts using lectin-gold conjugates. European Journal of Cell Biology 41: 142-149.
Honda Y, Kajita S, Matsui C, Otsuki Y, Takebe I (1975) An ultrastructural study of the infection of tobacco mesophyll protoplasts by potato virus X. Phytopathologische Zeitschrift 84: 66-74.
Hornberg C, Weiler EW (1984) High affinity sites for abscisic acid on the plasmalemma of Vicia faba guard cells. Nature 310: 321-324.
Hornick CA, Hamilton RL, Spaziani E, Enders GH, Havel RJ (1985) Isolation and characterization of multivesicular bodies from rat hepatocytes: an organelle distinct from secretory vesicles of the Golgi apparatus. The Journal of Cell Biology 100: 1558-1569.
Hübner R, Depta H, Robinson DG (1985) Endocytosis in maize root cap cells. Evidence obtained using heavy metal salt solutions. Protoplasma 129: 214-222.
Jacobs M, Gilberts SF (1983) Basal localisation of the presumptive auxin carrier in pea stem cells. Science 220: 1297-1300.
Joachim S, Robinson DG (1984) Endocytosis of cationic ferritin by bean leaf protoplasts. European Journal of Cell Biology 34: 212-216.
Juniper BE, Hawes CR, Horne JC (1982) The relationships between the dictyosomes and the forms of endoplasmic reticulum in plant cells with different export programs. Botanical Gazette 143: 135-145.
Kanaseki T, Kadota K (1969) The vesicle in a basket. The Journal of Cell Biology 42: 202-220.
Kirschhausen T, Harrison SC (1981) Protein organization in clathrin trimers. Cell 23: 755-761.
Kristen U, Lockhausen J (1983) Estimation of Golgi membrane flow rates in ovary glands of *Aptenia cordifolia* using cytochalasin B. European Journal of Cell Biology 29: 262-267.
Leonard RT, Hodges TK (1980) The plasma membrane. In *The biochemistry of plants* (Vol 1) (ed. PK Stumpf and EE Conn). Academic Press, New York, 163-182.
Leonard RT, Rayder L (1985) The use of protoplasts for studies on membrane transport in plants. In *Plant protoplasts* (ed. LC Fowke and F Constabel). CRC Press, Boca Raton, 105-118.
Loesch A (1981) The zinc-iodide-osmium-reactive sites in the extraocular muscles of the rat. The Australian Journal of Experimental Biology and Medical Science 59: 327-334.
Lubinski J, Huet C (1984) Independent variation in the number of coated pits and of coated vesicles in cultured fibroblasts. Biology of the Cell 52: 119-128.
Lucas WJ, Franceschi VR (1981) Characean charasome-complex and plasmalemma vesicle development. Protoplasma 107: 255-267.
MacLean IM, Sanders EJ (1983) Cationized ferritin and phosvitin uptake by coated vesicles of the early chick embryo. Anatomy and Embryology 166: 385-397.
Marty F (1978) Cytochemical studies on GERL, provacuoles, and vacuoles in root meristematic cells of *Euphorbia*. Proceedings of the National Academy of Sciences (USA) 75: 852-856.
Matile P (1975) *The lytic compartment of plant cells*. Springer-Verlag, New York, Wien (Cell biology monographs Vol 1).
Mayo MA, Cocking EC (1969) Pinocytic uptake of polystyrene latex particles by isolated tomato fruit protoplasts. Protoplasma 68: 223-230.
Mersey BG, Fowke LC, Constabel F, Newcomb EH (1982) Preparation of a coated vesicle-enriched fraction from plant cells. Experimental Cell Research 141: 459-464.
Mersey BG, Griffing LR, Rennie PJ, Fowke LC (1985) The isolation of coated vesicles from protoplasts of soybean. Planta 163: 317-327.
Mieth H, Speth V, Ebel J (1986) Phytoalexin production by isolated soybean protoplasts. Zeitschrift für Naturforschung 41c: 193-201.
Mollenhauer HH, Morré DJ (1980) The Golgi apparatus. In *The biochemistry of plants* (Vol 1) (ed. PK Stumpf and EE Conn). Academic Press, New York, 437-488.
Morales C, Clermont Y, Hermo L (1985) Nature and function of endocytosis in Sertoli cells of the rat. The American Journal of Anatomy 173: 203-217.
Nagata T, Melchers G (1978) Surface charge of protoplasts and their significance in cell-cell interaction. Planta 142: 235-238.
Nakamura S, Miki-Hiroshige H (1982) Coated vesicles and cell plate formation in the microspore mother cell. Journal of Ultrastructure Research 80: 302-311.
Newcomb EH (1980) Coated vesicles: their occurrence in different plant cell types. In *Coated vesicles* (ed. CJ Ockleford and A Whyte). Cambridge University Press, Cambridge, 55-68.
Newcomb W (1976) A correlated light and electron microscopic study of symbiotic growth and differentiation in *Pisum sativum* root nodules. Canadian Journal of Botany 54: 2163-2186.
Nougarède A, Lescure A-M (1970) Structure et comportement des dictyosomes et des corps multivésiculaires en liaison avec le plasmalemme dans les suspensions cellulaires issues du cambium de l'Acer pseudoplatanus L. Comptes Rendus Hebdomadaires des Séances de l'Académie des Sciences Series D 271: 916-919.
Novikoff AB (1976) The endoplasmic reticulum: a cytochemist's view (a review). Proceedings of the National Academy of Sciences (USA) 73: 2781-2787.
Pearse BMF (1975) Coated vesicles from pig brain: purification and biochemical characterization. Journal of Molecular Biology 97: 93-98.
Pearse BMF, Bretscher MS (1981) Membrane recycling by coated vesicles. Annual Review of Biochemistry 50: 85-101.
Pellegrino de Iraldi A (1977) Significance of the maillet method (ZIO) for cytochemical studies of subcellular structures. Experientia 33: 1-10.
Pesacreta TC, Lucas WJ (1984) Plasma membrane coat and a coated vesicle-associated reticulum of membranes: their structure and possible interrelationship in Chara corallina. The Journal of Cell Biology 98: 1537-1545.
Pesacreta TC, Lucas WJ (1985) Presence of a partially-coated reticulum in angiosperms. Protoplasma 125: 173-184.
Petersen OW, Van Deurs B (1983) Serial-section analysis of coated pits and vesicles involved in adsorptive pinocytosis in cultured fibroblasts. The Journal of Cell Biology 96: 277-281.
Peterson TA, Swanson ES, Hull RJ (1986) Use of lanthanum to trace apoplastic solute transport in intact plants. Journal of Experimental Botany 37: 807-822.
Power JB, Cocking EC (1970) Isolation of leaf protoplasts: macromolecule uptake and growth substance response. Journal of Experimental Botany 21: 64-70.
Quintart P, Courtoy PJ, Baudhuin P (1984) Receptor-mediated endocytosis in rat liver: purification and enzymatic characterization of low-density organelles involved in uptake of galactose-exposing proteins. The Journal of Cell Biology 98: 877-884.
Ralton JE, Howlett BJ, Clarke AE (1986) Receptors in host-pathogen interactions. In Hormones, receptors and cellular interactions in plants (ed. CM Chadwick and DR Garrod). Cambridge University Press, Cambridge, 281-318. (Intercellular and intracellular communication Vol 1).
Rao GSRL, Willison JHM, Ratnayake WMN (1984) Suberin production by isolated tomato fruit protoplasts. Plant Physiology 75: 716-719.
Regoezi E, Chindemi PA, Debanne MT, Charlwood PA (1982) Partial resialylation of human asialotransferrin type 3 in the rat. Proceedings of the National Academy of Sciences (USA) 79: 2226-2230.
Revel JP, Karnovsky MJ (1967) Hexagonal array of subunits in intercellular junctions of mouse heart and liver. The Journal of Cell Biology 33: C7-C12.
Reynolds ES (1963) The use of lead citrate at high pH as an electron opaque stain in electron microscopy. The Journal of Cell Biology 17: 208-212.
Robards AW, Robb ME (1972) Uptake and binding of uranyl ions by barley roots. Science 178: 980-982.
Robards AW, Robb ME (1974) The entry of ions and molecules into roots: an investigation using electron-opaque tracers. Planta 120: 1-12.
Robertson JG, Lyttleton P (1982) Coated and smooth vesicles in the biogenesis of cell walls, plasma membranes, infection threads and peribacteroid membranes in root hairs and nodules of white clover. Journal of Cell Science 58: 63-78.
Robinson DG, Kristen U (1982) Membrane flow via the Golgi apparatus of higher plant cells. International Review of Cytology 77: 89-127.
Rodriguez EM, Giménez AR (1981) Zinc-iodide-osmium procedures as markers of subcellular structures. 1. Standardization of staining of transmitter containing vesicles. Zeitschrift für mikroskopisch-anatomische Forschung 95: 257-275.
Rome LH (1985) Curling receptors. Trends in Biochemical Sciences 10: 151.
Rostgaard J (1973) A mechanical device for retrieving ribbons of ultrathin sections without folds. Stain Technology 48: 279-282.
Rougier M (1976) Sécrétion de polysaccharides dans l'apex radiculaire de maïs: étude radioautographique par incorporation de fucose tritié. Journal de Microscopie et de Biologie Cellulaire 26: 161-166.
Ryser U (1979) Cotton fibre differentiation: occurrence and distribution of coated and smooth vesicles during primary and secondary wall formation. Protoplasma 98: 223-239.
Schnepf E, Busch J (1976) Morphology and kinetics of slime secretion in glands of Mimulus tilingii. Zeitschrift für Pflanzenphysiologie 79: 62-71.
Smith RE, Farquhar MG (1966) Lysosome function in the regulation of the secretory process in cells of the anterior pituitary gland. The Journal of Cell Biology 31: 319-347.
Snider MD, Rogers OC (1984) Resialylation of asialotransferrin receptor in human erythroleukemia cells. Federation Proceedings, Federation of American Societies for Experimental Biology 43: 2044.
Squier CA (1982) Zinc iodide-osmium staining of membrane-coating granules in keratinized and non-keratinized mammalian oral epithelium. Archives of Oral Biology 27: 377-382.
Stahl P, Schwartz AL (1986) Receptor-mediated endocytosis. Journal of Clinical Investigation 77: 657-662.
Steinman RM, Mellman IS, Muller WA, Cohn ZA (1983) Endocytosis and the recycling of plasma membrane. The Journal of Cell Biology 96: 1-27.
Suzuki M, Takebe I, Kajita S, Honda Y, Matsui C (1977) Endocytosis of polystyrene spheres by tobacco leaf protoplasts. Experimental Cell Research 105: 127-135.
Takebe I (1979) Protoplasts as a host cell system for plant viruses. In Advances in protoplast research. Akadémiai Kiadó, Budapest, 431-440.(Proceedings of the 5th international protoplast symposium).
Tanchak MA, Griffing LR, Mersey BG, Fowke LC (1984) Endocytosis of cationized ferritin by coated vesicles of soybean protoplasts. Planta 162: 481-486.
Tepfer M, Taylor IEP (1981) The permeability of plant cell walls as measured by gel filtration chromatography. Science 213: 761-763.
Tycko B, Maxfield FR (1982) Rapid acidification of endocytic vesicles containing α₂-macroglobulin. Cell 28: 643-651.
Ungewickell E, Branton D (1981) Assembly units of clathrin coats. Nature 289: 420-422.
Van Der Valk P, Fowke LC (1981) Ultrastructural aspects of coated vesicles in tobacco protoplasts. Canadian Journal of Botany 59: 1307-1313.
Van Deurs B, Christensen EI (1984) Endocytosis in kidney proximal tubule cells and cultured fibroblasts: a review of the structural aspects of membrane recycling between the plasma membrane and endocytic vacuoles. European Journal of Cell Biology 33: 163-173.
Van Deurs B, Nilausen K, Faergenman O, Meinertz H (1982) Coated pits and pinocytosis of cationized ferritin in human skin fibroblasts. European Journal of Cell Biology 27: 270-278.
Van Deurs B, Petersen OW, Bundgaard M (1984) Identification of free coated pinocytic vesicles in Swiss 3T3 cells. The EMBO Journal 3: 1959-1964.
Weakley BS, Webb P, James JL (1981) Cytochemistry of the Golgi apparatus in developing ovarian germ cells of the Syrian hamster. Cell and Tissue Research 220: 349-372.
Wheeler H, Baker BL, Hanchey P (1972) Pinocytosis in root cap cells exposed to uranyl salts. American Journal of Botany 59: 858-868.
Wheeler H, Hanchey P (1971) Pinocytosis and membrane dilation in uranyl-treated plant roots. Science 171: 68-71.
Wileman T, Harding C, Stahl P (1985) Receptor-mediated endocytosis. The Biochemical Journal 232: 1-14.
Williamson FA, Fowke LC, Constabel FC, Gamborg OL (1976) Labelling of concanavalin A sites on the plasma membrane of soybean protoplasts. Protoplasma 89: 305-316.
Willingham MC, Hanover JA, Dickson RB, Pastan I (1984) Morphologic characterization of the pathway of transferrin endocytosis and recycling in human KB cells. Proceedings of the National Academy of Sciences (USA) 81: 175-179.
Willingham MC, Pastan I (1980) The receptosome: an intermediate organelle of receptor-mediated endocytosis in cultured fibroblasts. Cell 21: 67-78.
Willingham MC, Pastan I (1982) Transit of epidermal growth factor through coated pits of the Golgi system. The Journal of Cell Biology 94: 207-212.
Willingham MC, Pastan I (1983) Formation of receptosomes from plasma membrane coated pits during endocytosis: analysis of serial sections with improved membrane labeling and preservation techniques. Proceedings of the National Academy of Sciences (USA) 80: 5617-5621.
Willingham MC, Pastan I (1984a) Endocytosis and exocytosis: current concepts of vesicle traffic in animal cells. International Review of Cytology 92: 51-92.
Willingham MC, Pastan I (1984b) Do coated vesicles exist? Trends in Biochemical Sciences 9: 93-94.
Willingham MC, Rutherford AV, Gallo MG, Wehland J, Dickson RB, Schlegel R, Pastan IH (1981) Receptor-mediated endocytosis in cultured fibroblasts. Cryptic coated pits and the formation of receptosomes. The Journal of Histochemistry and Cytochemistry 29: 1003-1013.
Willison JHM (1981) Secretion of cell wall material in higher plants. In Encyclopedia of plant physiology, New series, Volume 13B- Plant carbohydrates II (ed. W Tanner and FA Loewus). Springer-Verlag, Berlin, 513-541.
Yoshikawa M, Keen NT, Wang M-C (1983) A receptor on soybean membranes for a fungal elicitor of phytoalexin accumulation. Plant Physiology 73: 497-506. |
A Survey of Classification Techniques in Data Mining
M. Sujatha
Research Scholar,
Computer Science And Systems Engineering,
Andhra university, Visakhapatnam, Andhra Pradesh,
S. Prabhakar
Research Scholar,
Computer Science Engineering,
JNTU, Hyderabad, Andhra Pradesh,
Dr. G. Lavanya Devi
Assistant Professor,
Computer Science And Systems Engineering,
Andhra university, Visakhapatnam, Andhra Pradesh,
Abstract- This paper provides a survey of numerous data mining classification techniques for innovative database applications. Classification is a model finding process that is used for assigning the data into different classes according to specific constraints. There are several major kinds of classification algorithms including Genetic algorithm C4.5, Naive Bayes, SVM, KNN, decision tree and CART. We deliberate the description of the algorithm.
Keywords - KNN, C4.5, SVM, CART
I. INTRODUCTION
Data mining involves the use of sophisticated data analysis tools to discover previously unknown, valid patterns and relationships in large data set. These tools can include statistical models, mathematical algorithm and machine learning methods. Consequently, data mining consists of more than collection and managing data, it also includes analysis and prediction. Classification technique is capable of processing a wider variety of data than regression and is growing in popularity [1]. There are several applications for Machine Learning (ML), the most significant of which is data mining. People are often prone to making mistakes during analyses or, possibly, when trying to establish relationships between multiple features. This makes difficult for them to find solutions to certain problems. Machine learning can often be successfully applied to these problems, improving the efficiency of systems and the designs of machines. Numerous ML applications involve tasks that can be set up as supervised. In the present paper, we have concentrated on the techniques necessary to do this. In particular, this work is concerned with classification problems in which the output of instances admits only discrete, unordered values. Our next section presented various classification methods. Section III described evaluating the performance of classifier. Finally, the last section concludes this work.
II. CLASSIFICATION METHODS
A. Genetic Algorithm-
Association Rules mining technique is used in GA, to find undetermined solution [7]. GA implements on small group of categorical data. Implementation of GA generates high-level prediction rules for better attribute selection. A single prediction rule is given by Michigan approach for each individual of a population; reduce the cost [10]. A set of prediction rules of a population of each individual represented by Pittsburgh approach [5]. In the case of classification, we measure the quality of the rule set as a whole, rather than the quality of single rule. Depending on data, rules are generalized or specialized, similar to the logical OR implementing and the logical AND operators.
B. Rules Sets-
The classification rule is “if-then-”rules, rule is (condition) A Rule r represents an instance X, if the attributes of the instance satisfied the condition of the rule. Individual’s rules are ranked, the related to their quality is called rule-based order. The rules that belong to the similar class appear together is known as class-based ordering. A good rule represents without mistakes and must cover as many examples as possible.
\[
\text{Accuracy} = \frac{\text{correctly classified examples}}{\text{examples matched by the rule}}
\]
\[
\text{Coverage} = \frac{\text{examples matched by the rule}}{\text{examples in the trained set}}
\]
These two learn sets are learning rules sequentially; one at a time is called separate-and-conquer (Furnkranz 99) and learning all rules together. In Separate-And-Conquer is after each rule is learned, the covered examples are deleted from training set and the process begins again, process completes when there are no examples to cover. Separate-And-Conquer methods are AQ family of rule induction method (Michalski 69), CN2 (Clark and Boswell 91), RIPPER K(Cohen 95). Separate-And-Conquer method accept new rule, when minimum accuracy and coverage, accept rules are better than default class, use a validation set of examples, unseen before. CN2 rule is initial pool contains an empty predicate, that covers all examples. CN2 rule algorithm does not generate rules, it generates predicates, and the majority class of the example covered by the predicate will be assigned as the class of the rule [28].
C. C4.5-
C4.5 measures numeric attributes, deals with missing values and pruning noisy data [3]. Pruning is used in C4.5 to avoid over fitting to noise in data [9]. C4.5 revised version is C4.8, implemented in WEKA as J4.8 and C5.0 (Rule quest) for commercial successor. Error estimate for sub tree is weighted sum of error estimates for all its leaves. Error estimate for a node (upper bound):
If \( c = 25\% \) then \( z = 0.69 \) (from normal distribution), \( f \) is the error on the training data. \( N \) is the number of instances covered by the leaf.
D. CART-
CART classification and Regression Tree based on accuracy, when data is noisy/missing values. CART takes a random sample; allow handling missing values by CHAID algorithm. In CART data preprocessing is not needed,
It automatically selects relevant attributes. CHAID algorithm [26, 27] considers missing values as distinct categorical value, also this method adopted by C4.5. CART treats a refined method surrogated, as missing primary field. If primary splitter is missing, then we will use a surrogate. CART prunes, the exact order in which each node must be deleted. For small data sets one Standard Error rule is good and generates an optimal tree. For larger data sets zero Standard Error rule generates accurate tree. Both C4.5 and CART are robust tools. Surrogate loss function like Gini index is used when misclassification of Decision tree.
E. Decision Tree Induction-
The problem of constructing optimal binary Decision Tree is an NP-COMPLETE problem and thus theoreticians have searched for efficient heuristic for constructing near optimal Decision Tree. Hunt’s algorithm generates a Decision tree by top-down or divides and conquers approach. The sample/row data contains more than one class, use an attribute test to split the data into smaller subsets. Hunt’s algorithm maintains optimal split for every stage according to some threshold value as greedy fashion [2].
Hunt’s algorithm uses greedy in approach for attribute test, select ‘best’ split and when to stop splitting on over fit or under fit conditions. Hunt’s algorithm measures both numeric and data set. For best split considers misclassification error, Gini index, Entropy. The decrease in entropy is called “Information Gain”. Misclassification error (Gini index or Entropy) on the test data set and stop when this begins to increase. Gini Index for a given node t:
\[ \text{GINI}(t) = 1 - \sum_{j} \left[ p\left(\frac{j}{t}\right) \right]^2 \]
where \( p(j/t) \) is the relative frequency of class j at node t. Gini index is used in CART[13], SLIQ and SPRINT. When a node p is split k partitions (children) the quality of split computed as
\[ \text{GINI}_{\text{split}} = \sum_{i=1}^{n} \text{GINI}(i), \text{Where } n_i = \text{no of records at child } i, n = \text{no of records at node } p. \]
To choose an attribute to partition data, the key building a decision tree which attribute to choose in order to branch. The objective is to reduce impurity or uncertainty in data as much as possible. A subset of data is pure if all instances belong to same class. The heuristic in c4.5 is to choose the attribute with maximum information gain or gain ration based on information theory.
1. Information gain
Given a set of examples D, we first compute its entropy,
\[ \text{Entropy}(D) = -\sum_{j=1}^{|c|} p(c_j) \log_2 p(c_j) \]
where \( p(c_j) \) is the probability of class \( c_j \) in data set D, we use entropy as a measure of impurity or disorder of data set D,(or a measure of information in a tree)[11]. As the data become purer and purer the entropy value becomes smaller and smaller. Information gained by selecting attribute \( A_i \)[4]. To branch to partition the data is
\[ \text{Gain}(D, A) = \text{Entropy}(D) - \text{Entropy}_A(D) \]
We choose the attribute with the highest gain to branch/split the current tree. Decision Tree Induction is greedy algorithm constructs Decision Tree in Top-down, recursive Divide-and-Conquer method. There are two methods for improving uncertainty.
- Pre-pruning (Early stopping rule) stop growing a branch when information becomes stop too early. Pre-pruning implements Chi-Squared test.
• Post-pruning takes a fully grown decision tree and discarded unreliable sun tree.
2. Overfitting Model
The classification model, two types of errors are committed. The training sample contains noisy data, and then errors are generated in the form of training errors and generalization errors. The training error is also known as Re-Substitution error or Apparent error, is the number of misclassification errors generated are training records. Whereas the generalization error is an expected error of the classification model on previously unseen records. Once the tree size is too large, its test error rate increases, even though training error rate continues to decrease. This process is known as model over fitting. For a complex tree, training error is zero, test error is large because noisy data present in training sample.
3. Underfitting Model
A training and test error rates of classification model are large and the size of decision tree is small. This situation is known as model underfitting. Decision Tree constructed with new binary features with logical operators such as conjunction, negation, disjunction (Zheng 1998) In addition, Zheng (2000) created at- least: M- of: N features when condition is true otherwise false[4]. Gama and brazdil (99) combined a Decision Tree with linear discriminate from constructing multivariate Decision Tree.
Decision Tree is complex because of replicated data set, Decision Tree construction is complex representation to avoid replicated data in Decision Tree, implement FICUS construction algorithm, which receives standard input or feature representation to produce a set of generated features by Markovitch and Rosenstein (2002). The study shows that C4.5 has a very good combination of error rate and speed. Based on this analytical evolution implemented a more efficient version of algorithm is known as EC4.5
F. Bayesian network -
Bayesian network is based on DAG (Directed Acyclic Graph) and one to one corresponding feature [12]. Bayesian network is divided into two tasks learning DAG and structure of network [8].
1. The network structure is fixed, learning the parameter in the conditional probability tables (CPT).
2. If structure is unknown, one approach is introduce scoring function that evaluates the “fitness” of the network with respect to training data and then to search for best network according to this score [16].
Compare Bayesian network to Decision Tree or Neural Network takes prior information about a given problem. Large data set with many features not suitable for Bayesian network [6].
G. Instance Based Learning -
Instance based learning’s is lazy-learning algorithm, as a delay the induction or generalization process until a classification is performed. Last learning algorithm require less computation time during the training phase than eager-learning algorithm (such as Decision tree, neural & Bayesian network) but more computation time during the classification process. In KNN, Irighton & Mellish 2002) found that their ICF algorithm and RT3 algorithm (Wilson & Martinez 2003). The disadvantage of Instance Based Learning takes more computation time for classifications. The available input features should be used in modeling via feature selection (Yu & Liu 2004) which improves classification accuracy and slow down classification time. If suitable distance metric is chosen, then accuracy of instance based classifier.
H. Support Vector Machine -
Support Vector Machine is a new classification method for both linear and non-linear data. It uses a non-linear
mapping to transform the original training data into a higher dimension. With the new dimension, it searches for the linear optimal separating hyper plane (i.e. “decision boundary”). With an appropriate non-linear mapping to a sufficiently high dimension, data from two classes can always be separated by a hyper plane. SVM finds this hyper plane using support vectors (“essential training tuples”) and margins (defined by the support vectors)[25]. Features: Training can be but accuracy is high owing to their ability to model complex non-linear decision boundaries (margins – maximization). SVM used for both classification and prediction.
I. K-Nearest-Neighbour (KNN) -
KNN is a non-Parametric classification method which is simple but effective in many cases [19]. The KNN has major drawbacks 1. Its low efficiency- being a lazy learning method prohibits it in many such applications such as dynamic web mining for a large repository and 2. Its dependency is based on selection of a good for K. There are [15, 30] that are used to significantly reduce the computation required at query time such as indexing training examples. As the KNN classifier requires storing the whole training set, when this is not at the redundancy of the training set to alleviate this problem [29, 31, 24, and 21].
Condensed Nearest Neighbor [CNN] by minimizing the number of stored patterns and storing only a sub-set of the training set for classification. The basic idea is that patterns the training set may be very similar and some do not add extra information and thus may be discarded. Gate [6] proposed the Reduced Nearest Neighbor (RNN) rule that aims to further reduce the stored subset after having applied CNN. It simply removes those elements from the subset which will not cause an error.
III. EVALUATING THE PERFORMANCE OF CLASSIFIER
A. Hold – Out method-
The original data with labeled examples is classified into two sets, called training set and test set [34]. The set should not be used in testing and the test set should not be used in learning. Unseen test set provides unbiased estimate of accuracy. This method is mainly used when the data set is large.
B. n-fold Cross-validation-
The available data is partitioned into n equal-size disjoint subsets. Use each subset as the training set to learn a classifier. The procedure is run n times, which given accuracies average of the n accuracies. 10-fold and 5-fold cross-validations are commonly used. This method is used when the available data is not large.
C. Leave-one-out cross validation-
This method is used when the data set is very small. It is a special case of cross-validation. Each fold of cross validation has only a single test example and all the test of the data is used in training [17]. If the original data has m examples, this is m-fold cross-validation.
D. Validation set -
The available data is divided into three subsets, 1. Training set 2. Validation set and 3. Test set. A validation set is used frequently for estimation parameters in learning algorithm. In such cases, the values that give the best accuracy on the validation set are used as the final parameter values. Cross validation can be used for parameter estimating as well.
E. Minimum Description Length (MDL)-
MDL handles missing values naturally as missing at random. The algorithm replaces sparse numerical data with zeros and sparse categorical data with zero vectors. Missing values in nested columns are interpreted as sparse. Missing in columns with simple data types are interpreted as missing at random tm. MDL takes into consideration the size of the model as well as the reduction in uncertainty due to using the model [20]. Both model size and entropy are measured in bits.
MDL considers each attribute as simple predictive model of the target class. Model selection refers to the process of comparing and ranking the single-predictor model. Automatic data preparation performs supervised binning for MDL [32]. Supervised binning uses decision trees to create the optimal bin boundaries. It is both categorical and numerical attributes.
F. **Bagging**
*Bagging* (Breiman, 1996), a name derived from “bootstrap aggregation”, was the first effective method of ensemble learning and is one of the simplest methods of arching [1]. The meta-algorithm, which is a special case of the model averaging, was originally designed for classification and is usually applied to decision tree models, but it can be used with any type of model for classification or regression. The method uses multiple versions of a training set by using the bootstrap, i.e. sampling with replacement. Each of these data sets is used to train a different model. The outputs of the models are combined by averaging (in case of regression) or voting (in case of classification) to create a single output. Bagging is only effective when using unstable (i.e. a small change in the training set can cause a significant change in the model) nonlinear models.
Create classifiers using training sets that are boot strapped (drawn with replacement). Sampling with replacement is according to a uniform probability distribution. Each bootstrapped sample D has a same size as original data. Some instance could appear several times in the training set, while others may be omitted. Bagging improves generalization performance by reducing variance of the base classifiers. The performance of the bagging depends upon the base classifier. If the base classifier is unstable, bagging helps to reduce the error associated with random fluctuations in the training data. If a base classifier is stable bagging may not be able to improve, rather than it could degrade the performance.
G. **Boosting**
Boosting is sequential production of classifier. Each classifier is dependent on previous one, and focuses on the previous ones errors. Examples that are incorrectly predicted in previously classifiers are chosen more often are weighted more evenly. Records that are only classified will have their weights increased. Records that are classified correctly will have their weights decreased.
H. **Ada-Boosting**
Ada-boosting measures complex hypothesis tend to over fitting. Simple hypothesis may not explain clearly. So it is combined many simple hypotheses into a complex one. There are two approaches
1. Select examples according to error in previous classifier (more representatives of misclassified cases are selected) are more common.
2. Weight error of misclassified cases higher (all cases are incorporated, but weights are different) does not work for some algorithms.
I. **Occam’s Razor**
Occam’s razor refers the simplest hypothesis that fits the data. Occam’s razor is most likely to identify unknown objects correctly. Occam’s razor given two models of similar generalization error, one should prefer simpler model over more complex model [33]. For complex models, there is a greater chance that is fitted accidentally by errors in data.
J. **Random Forest**
General purpose tool for classification and regression is WEKA. Random forest is very high accuracy for gradient boosting and support vector machine. There are two categories to construct Random Forest.
1. Classification and regression trees.
2. Bootstrap sample is a sample of the same size as the original dataset drawn from the original dataset with replacement.
IV. CONCLUSION
This paper covers with various classification techniques used in data mining. Each technique has got its own pros and cons as given this paper. Data mining is a wide area that integrates techniques from various fields including machine learning, artificial intelligence, statistics and pattern recognition, for the analysis of large volumes of data. There have been a large no of data mining algorithms embedded in these fields to perform different data analysis tasks.
REFERENCES
[1] Jiawei Han and Micheline Kamber Data Mining: Concepts and Techniques, 2nd edition.
[2] Baik, S. Baja, J. (2004), A Decision Tree Algorithm For Distributed Data Mining.
[3] Witten, I. & Frank, E. (2005), “Data Mining: Practical Machine Learning Tools And Techniques”, 2nd Edition, Morgan Francisco, 2005.
[4] Zheng, Z. (2000). Constructing A-Of-N Attributes For Decision Tree Learning. Machine Learning 40: 35–75.
[5] Sirgo, J., Lopez, A., Janez, R., Blanco, R., Abajo, N., Tarrio, M., Perez, R., “A Data Mining Engine Based On Internet: Emerging Technologies And Factory Automation”.
[6] Friedman, N., Geiger, D. & Goldszmidt M. (1997). Bayesian Network Classifiers. Machine Learning 29: 131-163.
[7] Fayyad, U., Piatetsky-Shapiro, G., And Smyth P., “From Data Mining To Knowledge Discovery In Databases,” Ai Magazine, American Association For Artificial Intelligence, 1996.
[8] Friedman, N. & Koller, D. (2003). Being Bayesian About Network Structure: A Bayesian Approach To Structure Discovery In Bayesian Networks. Machine Learning 50(1): 95-125.
[9] Quinlan, J.R. C4.5: Programs For Machine Learning. Morgan Kaufmann Publishers, San Francisco, Ca, 1993.
[10] Bianca V. D. Philippe Boula De Maregil And Marine Adda-Deker, “Identification Of Foreign-Accented French Using Data Mining Techniques, Computer Sciences Laboratory For Mechanics And Engineering Sciences (Limsi)”
[11] Breslow, L. A. & Aha, D. W. (1997). Simplifying Decision Trees: A Survey. Knowledge Engineering Review 12: 1–40.
[12] Jensen, F. (1996), An Introduction To Bayesian Networks. Springer.
[13] Introduction To Data Mining By Tan, Steinbach,Kumar.
[14] Collins, M., Schapire, R.E. And Singer, Y. (2000), Logistic Regression, AdaboostAndBregman Distances. Proc. Thirteenth Annual Conference Computational Learning Theory.
[15] T. Mitchell.: Machine Learning. MitpressAndMcGraw-Hill (1997).
[16] Madden, M. (2003), The Performance Of Bayesian Network Classifiers Constructed Using Different Techniques, Proceedings Of European Conference On Machine Learning, Workshop on Probabilistic Graphical Models For Classification, Pp. 59-70.
[17] Aliferis, C.F., Koutsoupias, And V. Rokhlin, “Algorithms Simple And Security-Check Bounds For Leave-One-Out Cross Validation”.
[18] Freund, Y. (1995). Boosting A Weak Learning Algorithm By Majority. Information And Computation 121, 256{285}.
[19] D. Hand, H. Mannila, P. Smyth.: Principles Of Data Mining. The MIT Press. (2001).
[20] Peter D. Grunwald “The Minimum Description Length Principle.
[21] M. Kubat, M. Jr.: Voting Nearest-Neighbour Subclassifiers. Proceedings Of The 17th International Conference On Machine Learning, ICML-2000, Pp.503-510, Stanford, CA, June 29-July 2, (2000).
[22] Jiang, W. (2000). Process Consistency For Adaboost. Tech. Report, Dept. Of Statistics, Northwestern University.
[23] Breiman, Leo, 1996. Bagging Predictors. Machine Learning.
[24] E. Alpaydin.: Voting Over Multiple Condensed Nearest Neighbors. Artificial Intelligence Review 11:115-132, (1997) Kluwer Academic Publishers.
[25] Cristianini, N., Shawe-Taylor, J.: An Introduction To Support Vector Machines. Cambridge University Press, Cambridge, 2000.
[26] L. Breiman, J. H. Friedman, R. A. Olshen, And C. J. Stone. Classification And Regression Trees. Wadsworth, Belmont, 1984.
[27] D. J. Hand, B. K. Mallick, And P. M. Smith. A Bayesian Cart Algorithm. Statistica Sinica, 3(4):261-283.
[28] Clark, P., Niblett, T. (1989). The C4.5 Induction Algorithm. Machine Learning 3(4):261-283.
[29] P. Hart.: The Condensed Nearest Neighbour Rule. IEEE Transactions On Information Theory, 14, 515-516, (1968).
[30] C.M.Bishop.: Neural Networks For Pattern Recognition. Oxford University Press, UK (1995).
[31] G. Gates.: The Reduced Nearest Neighbour Rule. IEEE Transactions On Information Theory, 18, 431-433, (1972).
[32] Barron, A,Rissanen,J,AndYu,B.(1998), “The Minimum Description Length Principle In Coding And Modeling IEEE Transaction On Information Theory.
[33] A.Blumer,A.Ehrenfeucht,D.Haussler,M.Warmuth: “Occam’s Razor”. |
Effects of cluster headache preventatives on mouse hypothalamic transcriptional homeostasis
Alessandra Pistolesi*, Daniela Buonvicino*, Mirko Muzzi, Matteo Urru and Alberto Chiarugi
Abstract
Objective: To investigate how cluster headache preventatives verapamil, lithium and prednisone affect expression of hypothalamic genes involved in chronobiology.
Methods: C57Bl/6 mice were exposed to daily, oral treatment with verapamil, lithium, prednisone or amitriptyline (as negative control), and transcripts of multiple genes quantified in the anterior, lateral and posterior hypothalamus.
Results: Verapamil, lithium or prednisone did not affect expression of clock genes of the anterior hypothalamus (*Clock, *Bmal1, *Cry1/2 and *Per1/2). Prednisone altered expression of hypothalamic neuropeptides *melanin-concentrating hormone* and *histidine decarboxylase* within the lateral and posterior hypothalamus, respectively. The three preventatives did not affect expression of the neurohypophyseal hormones *oxytocin* and *arginine-vasopressin* in the posterior hypothalamus. Conversely, amitriptyline reduced mRNA levels of *Clock, oxytocin* and *arginine-vasopressin*.
Conclusion: Data suggest that cluster headache preventatives act upstream or downstream from the hypothalamus. Our findings provide new insights on hypothalamic homeostasis during cluster headache prophylaxis, as well as neurochemistry underlying cluster headache treatment.
Keywords
Cluster headache, verapamil, lithium, prednisone, hypothalamus, transcriptional homeostasis
Date received: 10 September 2021; revised: 29 November 2021; accepted: 14 December 2021
Introduction
Cluster headache (CH) has a prevalence of 1 in 1000 and a male/female ratio of 2.5–3.5. Rhythmicity represents one of the most enigmatic feature of CH, and a great deal of effort has been directed at deciphering the underlying molecular mechanisms. Nevertheless, the neurobiology of the circadian/circannual pattern of CH and its relationship with attack pathogenesis still waits to be deciphered (1). Because of the key role of the hypothalamus in human chronobiology, as well as in the pathogenesis of primary headaches in general, it has been repeatedly proposed that this brain region is causally involved in CH rhythmicity and pain (2). Accordingly, several fMRI studies report hypothalamic activation during CH attacks (3). Further confirming a pivotal role of the hypothalamus in the etiopathogenesis of CH, deep brain stimulation with electrodes implanted in the posterior hypothalamus reduces both severity and frequency of attacks in CH patients refractory to pharmacological therapy (4).
Verapamil, lithium, and corticosteroids represent the mainstay of CH prophylaxis, but how they function as preventatives is unknown. In this regard, it is worth noting that from a pharmacodynamic perspective, CH preventatives do not show any apparent, mutual mechanism of action. Specifically, verapamil is a calcium
Department of Health Sciences, Section of Clinical Pharmacology and Oncology, University of Florence, Florence, Italy
*These authors contributed equally to this work.
Corresponding author:
Alessandra Pistolesi, Department of Health Sciences, Section of Clinical Pharmacology and Oncology, University of Florence, Viale Pieraccini 6, 50139 Florence, Italy.
Email: email@example.com
channel blocker approved as an antiarrhythmic, lithium is an inositol phosphatase/glycogen synthase kinase inhibitor used as a mood stabiliser, and corticosteroids are transcription-regulating, immunomodulatory drugs. Notwithstanding the lack of similar pharmacodynamics, to our knowledge the question as to whether these drugs share a similar impact on hypothalamic functions is still unanswered. We also reasoned that a deeper understanding of the pharmacological mechanisms underlying CH prophylaxis can help developing innovative therapeutics, for which patients are in great need.
Hence, in an attempt to gather information on CH pathogenesis and molecular mechanisms underpinning prophylaxis, in the present study we investigated whether oral treatment with verapamil, lithium carbonate or prednisone alter expression profiles of genes involved in hypothalamic functioning, circadian clock and sleep regulation in mice.
**Materials and methods**
**Animals**
All animal care and experimental procedures were performed according to the European Community guidelines for animal care (European Communities Council Directive 2010/63/EU) and were approved by the Committee for Animal Care and Experimental Use of the University of Florence. Male C57Bl/6 mice 20–25 g were purchased from Charles River (Milan, Italy) and housed in conventional units (six per cage) with free access to food (Harlan Global Diet 2018, Harlan Laboratories, Udine, Italy) and water. Mice were maintained on a 12 h light/dark cycle at 21°C room temperature for 15 days before drug exposure. As for the power analysis, the experimental hypothesis of our study prompted us to consider the effects of verapamil, lithium or prednisone on expression genes involved in hypothalamic functioning as the parameter of reference. Unfortunately, we have been unable to find studies reporting this information. This precluded the possibility of defining an *a priori* hypothesis, as well as performing a sample size calculation. On this basis, to carry out our exploratory study, we chose a number of animals of six mice per group. Two independent experiments with six animals per group for each time point (8 am and 8 pm) were carried out. Animals were randomised (generating five groups by the RAND function of Excel software) and treated with verapamil, lithium, prednisone, amitriptyline or vehicle (control group) for 15 days. Upon a 15-day drug exposure, mice (n = 12 per group) were sacrificed at 8:00 am and 8:00 pm by means of isoflurane anaesthesia, transcardial cold saline perfusion followed by immediate decapitation. At each time point a control group (n = 12) was also sacrificed. Brains were then frozen at −80°C upon the sacrifice and cut with a cryostat to collect hypothalamic regions. Specifically, bilateral anterior, lateral, and posterior hypothalamic regions were excised from frozen 20 μm serial coronal sections according to the stereotaxic atlas of the mouse brain. Genes of interest were analysed in that hypothalamic region where they are physiologically more expressed.
**Quantitative PCR**
Total RNA was isolated from hypothalamic regions using Trizol Reagent (Life Technologies, Monza, Italy). One μg of RNA was retrotranscribed using iScript (Bio-Rad, Milan, Italy). RT-PCR was performed as reported (5). The primers used are reported in Figure 2.
**Statistical analysis**
Data are expressed as mean ± SEM. To test the difference between more than two groups ANOVA plus Tukey’s post hoc test (according to Kolmogorov–Smirnov test for normality) was used. Differences were considered to be significant with a p-value < 0.05. Statistical analyses were carried out using GraphPad Prism (version 7).
**Results**
**Effects of CH preventatives and amitriptyline on hypothalamic gene transcripts**
In order to adopt a clinically-relevant treatment protocol, mice were exposed to daily, oral drug administration for 15 days, a length of time sufficient for a CH preventative to prompt a therapeutic effect. Similarly, drugs were administered according to routinely adopted treatment schedules. Specifically, according to human-to-animal dose conversion guidelines and chronic treatment (6), verapamil and lithium carbonate were administered three times per day (8:00 am, 1:00 pm and 8:00 pm) at 3.4 and 13 mg/kg, respectively, whereas prednisone was dosed at 3 mg/kg at 8:00 am. An additional group of mice was exposed to oral amitriptyline (1 mg/kg at 8:00 pm), a migraine preventative altering some hypothalamic functions here adopted as a negative control. Collectively, drug doses were about three-fold higher than those routinely adopted in patients, in keeping with standard conversion from human to mouse chronic dosing. Drug treatments were well tolerated, showing no effects on body weight and food/water consumption (not shown).
As far as clock genes of the anterior hypothalamus are concerned, we found that none of the CH preventative altered transcript levels of *Clock*, *Bmal1*, *Cry1* and -2, as well as *Per1* and -2. Interestingly, mice receiving amitriptyline showed reduced *Clock* mRNA levels at both 8 am and 8 pm and an increase of those of *Bmal1* at 8 am (Figure 1(b)–(g)). In the lateral hypothalamus, we evaluated transcripts for *hypocretin/orexin* and *melanin-concentrating hormone*, two neurotransmitters with key roles in sleep regulation (7). We found that none of the tested drugs altered hypocretin transcripts, whereas only prednisone increased those of
**Figure 1.** Effects of verapamil, lithium, prednisone, and amitriptyline on gene transcripts involved in hypothalamic chronobiology. The hypothalamic regions (anterior, lateral and posterior) analysed and the respective genes is shown in (a). The impact of a chronic (15 days), oral treatment with verapamil (3.4 mg/kg, three times per day), lithium carbonate (13 mg/kg, three times per day), prednisone (3 mg/kg) or amitriptyline (1 mg/kg) on transcript levels of *Clock* (b), *Bmal1* (c), *Cry1* (d), *Cry2* (e), *Per1* (f), *Per2* (g), orexin *Hcrt* (h), melanin concentrating hormone *Mch* (i), histidine decarboxylase *Hd* (j), oxytocin *Oxt* (k) and arginine vasopressin *Avp* (l) within the mouse hypothalamus at 8 am and 8 pm after 15 days of treatment is shown. Each column represents the mean ± SEM of two experiments with six animals per group.
*p < 0.05 vs. Crl (ANOVA plus Tukey’s post-hoc test).*
**Figure 2.** Primers used are shown in this table.
*melanin-concentrating hormone* at both 8 am and 8 pm (Figure 1(h) and (i)). Within the posterior hypothalamus, we evaluated mRNA levels of *histidine decarboxylase*, the rate-limiting enzyme in the synthesis of brain histamine with a key role in CNS arousal/wakefulness (8), as well as the two neurohypophysis hormones oxytocin and arginine vasopressin. Figure 1(j) shows that histidine decarboxylase transcripts were reduced in mice receiving prednisone at both time points, but unchanged in animals challenged with the other preventatives. Conversely, amitriptyline reduced expression levels of oxytocin at 8 am and pm, whereas those of arginine vasopressin were reduced at 8 am only (Figure 1(k) and (l)). To rule out that lack of effects on gene transcripts being due to lack of power (see Methods), we conducted a *post-hoc* power analysis using the software G*Power 188.8.131.52, with ANOVA one-way test, and considering a beta/alpha ratio of 4, the sample size (12 mice) and effect size reached for each single gene. We obtained a power of 0.78 for *Clock*, 0.84 for *Bmal1*, 0.75 for *Cry1*, 0.79 for *Cry2*, 0.73 for *Per1*, 0.82 for *Per2*, 0.49 for *Hcrt*, 0.96 for *Mch*, 0.84 for *Hd*, 0.81 for *Oxt*, 0.79 for *Avp*. These findings indicate that power was reached for most of the gene.
**Discussion**
Even though pathophysiology of CH is largely unknown, both clinical evidence and functional neuroimaging point to a key role of the hypothalamus in disease pathogenesis. Given the preventative nature of verapamil, lithium, and prednisone in CH therapy, it makes sense that these drugs act within the CNS to reduce attack frequency. Indeed, all the three preventatives are known to cross the blood-brain barrier and affect CNS functioning. We reasoned that the three drugs may share the ability to alter hypothalamic homeostasis thereby altering the expression of genes involved in circadian rhythm and sleep/wake cycle. Interestingly, *Clock* and *Cry* nucleotide polymorphisms segregate with CH patients (9,10). Similarly, genetic susceptibility loci are present in CH patients, even though they do not seem to be specifically related to hypothalamic function regulation (11). Together, these findings corroborate the relevance of specific chronotypes/gene association in CH pathogenesis. We found, however, that neither verapamil nor lithium affected transcript levels of mouse hypothalamic clock genes/neuropeptides, whereas prednisone
increased melanin-concentrating hormone mRNAs and reduced those of histidine decarboxylase. Taken together, these findings indicate that the three preventatives do not alter the interconnected, transcription-translation feedback loops underpinning clock gene expression within the mouse suprachiasmatic nucleus, the hypothalamic pacemaker of circadian oscillations. This is in keeping with the notion that verapamil exerts its prophylactic effect even when acutely administered (12). In this regard, however, it is worth noting that verapamil is rapidly extruded from the CNS by P-glycoprotein, an event that, in principle, might reduce its potential impact on hypothalamic gene expression. Conversely, corticosteroid therapy is well known to affect sleep, mostly causing sleep difficulties. We report here that, in keeping with its transcription-regulating properties, prednisone increased melanin-concentrating hormone and reduced histidine decarboxylase transcripts within the lateral and posterior mouse hypothalamus, respectively. Oddly, these changes should promote rather than reduce sleep. Indeed, the melanin-concentrating hormone increases both REM and non-REM sleep, whereas diminished expression of histidine decarboxylase impairs histamine-dependent CNS arousal and wakefulness. We reason that these somehow paradoxical transcriptional changes occur as a rebound phenomenon in chronically exposed mice, a hypothesis in line with the transient nature of corticosteroid-dependent sleep disturbances. The inability of lithium to affect gene expression within the mouse hypothalamus was in part unexpected. This is not only because of the efficacy of lithium in bipolar disorder, a psychiatric condition typically characterised by specific chronoperiods, but also in light of the drug’s impact on rodent behavioural rhythms (13). To our knowledge, however, no studies have investigated the effect of the drug on clock gene expression in wild-type mice. Of note, our findings are at odds with a recent study by Burish and colleagues (14) reporting slight changes of circadian clock gene transcripts in the hypothalamus of mice exposed to verapamil. Possibly, this apparent inconsistency can be explained considering that the authors adopted a high dose of verapamil (~6.5-fold higher than that clinically relevant), a limited number of mice per group, as well as drug administration in drinking water rather than reproducing a clinically relevant treatment schedule. Surprisingly, amitriptyline, adopted here as a putative negative control, altered hypothalamic transcript levels of Clock, oxytocin, and arginine vasopressin. When put into a clinical perspective, the ability of amitriptyline to alter hypothalamic transcriptional homeostasis may underpin its antimigraine properties, as well as its modest benefit in CH treatment (15). The ability of amitriptyline to reduce oxytocin and arginine vasopressin expression can be related to the drug’s impact on peripheral aminergic neurotransmission. Of note, in keeping with our findings, severe hypotension during amitriptyline intoxication only responds to arginine vasopressin (16).
In conclusion, our data suggest that CH preventatives act upstream or downstream from the hypothalamus, possibly targeting signalling pathways independently involved in disease pathogenesis. However, given that the present findings only provide evidence for lack of diurnal variation in hypothalamic gene expression, further studies are needed to rule out that CH preventatives prompt circadian alterations showing a different timescale.
**Article highlights**
- It is currently unclear whether cluster headache (CH) preventatives modify hypothalamic transcriptional homeostasis.
- We report that prolonged exposure to verapamil, lithium or prednisone does not affect expression of mouse hypothalamic genes regulating chronobiology.
- CH preventatives may act upstream or downstream from the hypothalamus.
**Declaration of conflicting interests**
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
**Funding**
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by grants from Health Project 2018 Region of Tuscany, and PRIN2017 Italian Ministry of the University and Research (2017JL8SRX8).
**References**
1. Wei DY and Goadsby PJ. Cluster headache pathophysiology – insights from current and emerging treatments. *Nat Rev Neurol* 2021; 17: 308–324.
2. May A, Schwedt TJ, Magis D, et al. Cluster headache. *Nat Rev Dis Primers* 2018; 4: 1–17.
3. May A. The role of imaging in the pathophysiology and diagnosis of headache. *Curr Opin Neurol* 2005; 18: 293–297.
4. Jürgens TP, Leone M, Proietti-Cecchini A, et al. Hypothalamic deep-brain stimulation modulates thermal sensitivity and pain thresholds in cluster headache. *Pain* 2009; 146: 84–90.
5. Buonvicino D, Urru M, Muzzi M, et al. Trigeminal ganglion transcriptome analysis in 2 rat models of medication-overuse headache reveals coherent and widespread induction of pronociceptive gene expression patterns. *Pain* 2018; 159: 1980–1988.
6. Nair AB and Jacob S. A simple practice guide for dose conversion between animals and human. *J Basic Clin Pharm* 2016; 7: 27–31.
7. Concetti C and Burdakov D. Orexin/hypocretin and MCH neurons: Cognitive and motor roles beyond arousal. *Front Neurosci* 2021; 15: 294.
8. Yoshikawa T, Nakamura T and Yanai K. Histaminergic neurons in the tuberomamillary nucleus as a control centre for wakefulness. *Br J Pharmacol* 2021; 178: 750–769.
9. Fourier C, Ran C, Zinnegger M, et al. A genetic CLOCK variant associated with cluster headache causing increased mRNA levels. *Cephalalgia* 2018; 38: 496–502.
10. Fourier C, Ran C, Sjöstrand C, et al. The molecular clock gene cryptochrome I (CRY1) and its role in cluster headache. *Cephalalgia* 2021; 41: 1374–1381.
11. O’Connor E, Fourier C, Ran C, et al. Genome-wide association study identifies risk loci for cluster headache. *Ann Neurol* 2021; 90: 193–202.
12. Petersen AS, Barloese MCJ, Snoer A, et al. Verapamil and cluster headache: Still a mystery. A narrative review of efficacy, mechanisms and perspectives. *Headache* 2019; 59: 1198–1211.
13. Moreira J and Geoffroy PA. Lithium and bipolar disorder: Impacts from molecular to behavioural circadian rhythms. *Chronobiol Int* 2016; 33: 351–373.
14. Burish MJ, Han C, Mawatari K, et al. The first-line cluster headache medication verapamil alters the circadian period and elicits sex-specific sleep changes in mice. *Chronobiol Int* 2021; 38: 839–850.
15. Visvanathan N, Chinnu M and Mahadevan R. Amitriptyline in the management of cluster headache (P3.055). *Neurology* 2015; 84. ISSN: 1526–632X.
16. Barry JD, Durkovich DW and Williams SR. Vasopressin treatment for cyclic antidepressant overdose. *J Emerg Med* 2006; 31: 65–68. |
Interactive Cloud Data Farming Environment for Military Mission Planning Support
Abstract: In a modern globalised world, military and peace keeping forces often face situations which require very subtle and well planned operations taking into account cultural and social aspects of a given region and its population as well as dynamic psychological awareness related to recent events which can have impact on the attitude of the civilians. The goal of the EUSAS project is to develop a prototype of a system enabling mission planning support and training capabilities for soldiers and police forces dealing with asymmetric threat situations, such as crowd control in urban territory. In this paper, we discuss the data-farming infrastructure developed for this project, allowing generation of large amount of data from agent based simulations for further analysis allowing soldier training and evaluation of possible outcomes of different rules of engagement.
Keywords: data farming, cloud, virtualisation, web 2.0, mission planning support
1. Introduction
Mission planning support in military applications is a very complex process, involving several issues, such as analysis of physical environment of the incident, social and cultural aspects of the location and definition of proper Measures of Effectiveness for assessment purposes. In particular, asymmetric threats in urban territory, which involve an operation of a relatively small group of soldiers or police forces within a city with civilian population ranging from neutral to hostile, require very careful planning in order to maximize the chances of positive outcome. Due to possibly large amount of civilians it is very difficult to predict the possible outcomes of different rules of engagement. This task can be supported by computer agent based simulation system, which allows the higher echelons to assess different strategies for the operation. However, in order to get meaningful data from the system, first a significant number of runs with different parameters must be performed. Obviously this is infeasible, requiring both an extreme amount of computing power and time, while in military applications it is often critical to have even rough assessment of the strategy and possible outcomes in predetermined amount of time.
In this paper we present a part of a system which was developed within the EUSAS (European Urban Simulation for Asymmetric Scenarios) project, which is financed by 20 nations under the Joint Investment Program Force Protection of the EDA. During the EUSAS project we developed a new modelling approach to human decision making [7] and proposed a combination of object-oriented and ontology-based approaches for real-time interworking of human behaviour models in the context of agent-based simulation systems [11]. Afterwards, we evaluated agent-based simulation platforms [12] to choose the one that best suits the project needs. The results of this work were used to create highly realistic agent based simulations of military missions. These simulations are the core of our system which is composed of two parts: serious game part which aim is support of the training of policemen and soldiers and data farming part which contains infrastructure and software that executes simulations to provide framework that supports military mission planning. In this paper we describe the data farming part.
The aim of the data farming part is to manage data farming experiments (DF experiments). DF experiment includes several executions of MASON [13] based agent simulation (of military operation in urban territory) with different values of parameters and gathering of logs and results. The implementation of the agents logic (both civilians and soldiers) and environment is based on advanced models, which require each run of the simulation to run for several minutes, taking on the input configuration file with all parameters set to proper values and providing on the output a log file with all events within the simulation relevant for further analysis. Our data farming systems goal is to provide the researcher with the ability to minimize the time (and computational cost) of the simulation by limiting the number of simulation runs necessary for obtaining relevant results as well as dynamically monitoring the intermediate results and fine tuning the parameter space on the fly.
The rest of the paper is organized as follows. In Section 2, we describe related work. Section 3 contains description of user requirements. In section 4, we present data farming overview from the user point of view. Then, in section 5, we show the architecture of our system and the user interface in section 6. We conclude this paper in section 7.
2. Related work
The concept of data farming was invented in 1998 [3, 8]. It may be used in any area where the research process contains following steps [8]:
- question/topic research and definition,
- model development and gaming,
- parameter space exploration,
- data exploration and analysis.
Since 1998, several tools have been developed. They can be split into three categories [8]:
- implementations of data farming environments,
- distillation modelling environments,
- data exploration tools.
Two very popular tools are the Maui High Performance Computing Parallel Execution System (PES) and OldMcData. They integrate following distillation modelling environments: ISAAC, Socrates, Pythagoras, Mana, PAX and NetLogo. To simplify the analysis, several visualization tools have been created. Three very popular ones are: the Playback Tool, the VizTool Landscape Plotter and Avatar.
The data farming is a powerful instrument, especially for military users. Several articles describing military application of data farming were created e.g. marine corps applications of data farming [5], SDF meta-technique [4] or “Data Farming and Defense Applications” by Horne and Meyer [9]. However, different military applications demand different techniques and tools. The main aspect that differentiates the EUSAS project from described solutions is that our systems offers users ability to interact with the system during the experiment to analyse partial results or to extend the range of investigated parameters. Attempts of using Grids to perform interactive and real-time applications have been described in [6], [14] and [10]. Another very important innovation of our system is the attempt to make the system totally independent from infrastructure to provide the user ability to use sources of computing power on demand (e.g. Clouds).
3. Requirements for mission planning support system
Users of different system parts have different requirements. The users of the serious game part (soldiers and policemen) need the system that will allow them to perform virtual missions using their computers. The environment shown on the screens should
realistically imitate actual locations. Furthermore, the system should allow users to cooperate during mission training. Users of the data farming part need the system that will allow them to simulate many possible mission scenarios to verify some rules of engagement. Although, requirements of serious game users and data farming users are very different, users of both parts of the system demand high reality of behaviour of computer agents so advanced psychological models have to be used. In this chapter we present the requirements of users of the data farming part of the system.
Military users involved in mission planning want to simulate the mission and check if soldiers are able to perform this mission with some given rules of engagement. The data farming allows users to achieve this goal successfully, since it gives them ability to check many factors which may influence the mission. The data farming experiment includes several executions of simulation with different values of parameters that represent these factors. If all simulations executed with different values of parameters end successfully, there is high probability that the real mission will also end with success. Otherwise, the users will know which factors are crucial and should be monitored. Furthermore, the users want to be able to choose ranges of tested parameters to determine the conditions under which the mission will be performed (e.g., to determine behaviour of civilians, number of agents etc.). The aim of the data farming is generation of large amount of data that may be analysed by external tools which are able to answer users questions. In the EUSAS project, the MASDA algorithm [2] is used to find which sequences of actions lead soldiers to success and after which sequences of actions soldiers may be hurt or killed.
Time is very important for the military users so they have to be able to check predicted time to finish the experiment. If this period of time is not satisfying, they require ability to speed up the experiment. Moreover, the users expect ability to check partial results as soon as possible. They also require possibility to modify the experiment during the runtime. Especially—to extend the parameters space when needed.
The user requirements imply some problems that have to be solved by providers of the infrastructure and software. The number of possible parameters combinations can be very large (ca. $10^{10}$). Moreover, each run of the agent based simulation may take several minutes because one simulation may contain thousands of agents which have to use advanced psychological models. This is why one personal computer needs $o(10^5)$ years to execute one large data farming experiment. We use Cloud as a source of computing power accessible on demand. To give the user ability to speed up the experiment we use virtualisation of the resources and a user interface that allows creating new virtual machines anytime. However, even the most powerful infrastructure is not able to execute all possible combinations of a large data farming experiment in acceptable time by military users. Hence the reduction of parameters space have to be done before the experiment start, however, this reduction is a challenge because the number of executed combinations of parameters should be low while the loss of the information connected with the reduction should be minimal. For this purpose, we use specialized algorithms called “Design of Experiment” (DoE) methods [1] to
do the reduction. Other user requirements such as ability to check partial results, modify the experiment during the runtime etc. are addressed by a specialized user interface (see chapter 6).
4. Data farming overview
The data farming process performed by the EUSAS system is shown in Fig. 1. It addresses all user requirements described above. In the beginning of the experiment the initial parameter space has to be defined through selection of values of all parameters. Afterwards, the parameter space is reduced. Reduction size depends on a chosen design of experiment method—it is possible to obtain several orders of magnitude reduction. During the experiment execution, there is an ability to see the partial results and to do some basic analysis of these results, for instance to generate the regression tree. On the basis of this analysis, extension of experiment may be done. The user interface allows to do it interactively. For example, if the analysis shows that one parameter influences the output more than others, the system may sample it more densely. The speed up of the experiment may be done by increasing the computing power (i.e. adding additional nodes) taking cost limit into account. Hence the speed of the experiment should depend on circumstances.
5. Data farming architecture
The data farming architecture was designed to solve problems introduced in Section 3. Together with the user interface (see the next chapter) it creates the system that fulfils user requirements. Currently, DF experiments are performed on a computing
infrastructure, which is both dynamic and heterogeneous, the number of computing nodes can change depending on priorities of different tasks in a data centre and machines assigned to a particular data farming experiment can represent different hardware architecture and different operating systems, depending on current availability of resources.
Although a private data centre has several options regarding delegating parts of available resources to perform a concrete goal, for instance the physical worker nodes can be statically assigned and set up to compute DF experiments. However, this scenario requires substantial administrative effort related to installation and configuration of the operating system and software components. Another option is to use the virtualisation technology and start a number of virtual machines on the existing worker nodes, which can be stopped when not needed. Using virtualisation often increases utilization of the hardware infrastructure but also is much more flexible, e.g., allows running several different operating systems on the same worker node. Unfortunately, in many scenarios the private infrastructure is insufficient due to experiment or time limitations. To address this challenge, we intend to integrate our solution with publicly available commercial Clouds such as Amazon EC2, which in theory can be treated as unlimited source of computing power, depending only on the budget limit for a particular DF experiment. Depending on Cloud type, the computing power can be provided in various forms. In our case, the Infrastructure-as-a-Service (IaaS) model is most suitable, as it allows users to run virtual machines on third-party infrastructure. In most cases, IaaS Clouds provide an easy to use Application Programming Interface (API) to manage virtual machines. By using public Clouds, the user can speed up DF experiment when necessary.
A conceptual overview of the DF infrastructure is depicted in Fig. 2. A central point that dispatches experiment instances and provides user interface is called “simulation manager”. We decided to use the “pull” mode to distribute the work among available resources. Using the “pull” mode means, the computing elements, e.g., virtual machines, ask for tasks to perform on their own. Thus, there is no need for simulation manager to know about available computing resources. Instead, the computing resources only have to know where simulation manager is located. Thus, it is trivial to scale the infrastructure only by running dedicated software on any available computing resources.
Regarding infrastructure management, we intend to provide a plug-in-based approach for extending simulation manager management capabilities. Using this approach, adding support for new type of computing resource will require only implementation a well-defined API for creating, running, stopping and destroying virtual machines for simulation manager and to provide a dedicated view for end users.
Simulations are implemented in Java language (using MASON framework) and log all actions within the simulation to a text file. The simulation instances are executed in virtual machines. The number of instances that are running simultaneously at each virtual machine is equal to the number of cores assigned to it. Each simulation in MASON uses one thread for its main activities. Although, it is possible to
perform some activities of agents by other threads simultaneously, the speed up is smaller than the number of additional cores available for the simulation. The user is able to start new virtual machines during data farming experiment. New virtual machines immediately ask for simulation instances what speeds up the experiment.
**Figure 2.** Data farming infrastructure overview.
The data farming manager schedules instances of experiment designed with experiment manager. It also cooperates with experiment manager during the definition of experiment providing Design of Experiment methods:
- Full factorial – all possible combinations of values of parameters are used,
- Fractional factorial – only a carefully chosen subset (fraction) of the experimental runs of a full factorial design is executed. Fedorov algorithm [15] is used to estimate how many combinations should be chosen,
- Orthogonal Latin Hypercubes – if all parameters have the same number of possible values, the number of executed instances is equal to the number of possible values of one parameter. In this case, it is guaranteed that each possible value of each parameter will be used exactly once. If some parameters have less possible values than others, some values of these parameters may be used more than once,
- $2^K$ – all possible combinations of minimal and maximal values of parameters are used.
The simulation manager also gathers data which are a base for analysis of partial results.
6. User interface
To manage data farming experiments in a user-friendly way, the EUSAS Data Farming module provides experimenters and administrators with a dedicated Graphical User Interface (GUI). It is web-based so it is accessible with various types of devices, which allows users to design and schedule DF experiments and to manage the Data Farming infrastructure. It contains two main panels, one dedicated to experiment configuration and monitoring and the second providing infrastructure management functionality.
6.1. Experiment management interface
“Experiment manager” supports all phases of running DF experiment. In particular, it allows researchers, e.g., users who run a DF experiment, to perform the following actions:
- schedule new DF experiments with custom parameterisation (several Design of Experiment methods are also supported),
- monitor DF experiments already running,
- analyse results of DF experiments either when an experiment is finished or during the execution of the experiment,
- download results of DF experiment.
The main view of the “Experiment manager” panel provides information about DF experiments, which are currently running, can be run or were ran in the past. The user can either go directly to a monitoring view of a currently running or historical experiment or decide to schedule a new experiment based on one of the available simulations. Starting a new DF experiment involves selecting parametrisation type of each input variable of an experiment, e.g., range of values, random value with different distributions or a concrete value, and then providing specific parameter values for each input variable regarding selected parametrisation types. Next, for parameters with range type of parametrisation, users can apply several Design of Experiment methods. Input parameters can be grouped into a number of sets, each can have different DoE method assigned. It gives the user ability to choose DoE which sample parameters more densely for more important parameters and DoE which reduce number of combinations better for other parameters.
After starting a DF new experiment, the user is redirected to a dedicated monitoring view depicted in Fig. 3. Using this view, the user can monitor the progress of the started DF experiment, analyse partial results, extend the experiment and download results after the experiment finishes.
In terms of analysing partial results of a DF experiment, we provide two types of charts, which intend to provide information about values of a selected Measure of Effectiveness (MoE). The first type of charts is histogram of MoE values, while the second type of charts is regression tree for a selected MoE.
Besides analysing partial results, the user can extend the set of input parameter values of a DF experiment. This functionality is embedded in the regression
tree chart. As in each node of a regression tree (except of leaves) a name of input parameter is given, the user can select an interesting input parameter and decide to generate additional experiment instances for new values of the selected input parameter. This functionality makes DF experiments interactive, i.e., the user can modify the experiment during the runtime without starting a new DF experiment.
After the DF experiment is finished, the user can download a compressed package with results of the experiment. The package contains log files from each experiment instances and a CSV file with MoE and input parameter values for each experiment instance. The CSV file can be used, e.g., to analyze results with external, third party tools.
### 6.2. Infrastructure management interface
Besides management of DF experiments, “Simulation manager” allows users to manage the infrastructure, which is used to perform the DF experiments. “Infrastructure manager” is a panel (depicted in Fig. 4), which intends to facilitate the administration of the DF infrastructure. In the current version, it supports virtualised private infrastructure.
By using “Infrastructure manager”, the user can register physical devices, which will be used to run virtual machines. Then, the user can create new virtual machines with a specified set of resources, e.g., number of CPU cores and amount of RAM. Each created virtual machine can be run, paused, stopped and deleted.
7. Conclusions
In this paper we have presented a novel generic purpose solution for performing large scale data farming experiments on heterogeneous computing infrastructure consisting of privately owned and Cloud based resources. The system is currently being used for running data farming experiments related to military scenarios within the framework of European Defence Agency EUSAS project, where the simulation runs provide data about possible behaviour outcomes of large groups of civilians and blue forces in both military and urban crowd control scenarios. The data produced by the data farming is used to improve the psychological models of the civilian agents in the simulation as well as evaluate different rules of engagement for different situations.
The main future work for this system is improvement of the user interface in order to reflect the requirements of actual military analysts and integration of the system with existing Cloud infrastructures in order to extend the scalability of the system.
Acknowledgements
This work is supported by the EDA project A-0938-RT-GC EUSAS.
References
[1] Atkinson A.C., Donev A.N.: Optimum Experimental Designs. Clarendon Press, 1992.
[2] Bezek A., M. G., Bratko I.: Multi-agent strategic modeling in a robotic soccer domain. In *Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems*, pp. 457–464, 2006.
[3] Brandstein A., Horne G.: Data farming: A meta-technique for research in the 21st century. In *Maneuver Warfare Science 1998*. Marine Corps Combat Development Command Publication, 1998.
[4] Choo C.S., Ng E.C., Ang C.K., Chua C.L.: Systematic data farming – an application to a military scenario. In *Proceedings of Army Science Conference*, 2006.
[5] Forsyth A., Horne G., Upton S.: Marine corps applicatons of data farming. In Kuhl M.E., Steiger N.M., Armstrong F.B., Joines J.A., editors, *Proceedings of the 2005 Winter Simulation Conference*, pp. 1077–1081, 2005.
[6] Funika W., Korcyl K., Pieczykolan J., Skital L., Balos K., Slota R., Guzy K., Dutka L., Kitowski J., Zielinski K.: Adapting a hep application for running on the grid. *Computing and Informatics*, 28:353–367, 2009.
[7] Hluchy L., Kvassay M., Dlugolinsky S., Schneider B., Bracker H., Kryza B., Kitowski J.: Handling internal complexity in highly realistic agent-based models of human behaviour. In *Proceedings of 6th IEEE International Symposium on Applied Computational Intelligence and Informatics (SACI 2011)*, pp. 11–16, 2011.
[8] Horne G., Meyer T.: Data farming: Discovering surprise. In Ingalls R.G., Rossetti M.D., Smith J.S., Peters B.A., editors, *Proceedings of the 2004 Winter Simulation Conference*, pp. 1082–1087, 2004.
[9] Horne G., Meyer T.: Data farming and defense applications. In *Proceedings of the MODSIM World 2010 Conference*, pp. 74–82, 2010.
[10] Korcyl K., Szymocha T., Funika W., Kitowski J., Slota R., Balos K., Dutka L., Guzy K., Kryza T., Pieczykolan J.: The atlas experiment on-line monitoring and filtering as an example of real-time application. *Computer Science, annual of AGH University of Science and Technology*, 9:77–86, 2008.
[11] Kvassay M., Hluchy L., Kryza B., Kitowski J., Seleng M., Dlugolinsky S., Laclavik M.: Combining object-oriented and ontology-based approaches in human behaviour modelling. In *Proceedings of IEEE 9th International Symposium on Applied Machine Intelligence and Informatics (SAMI)*, pp. 177–182, 2011.
[12] Laclavik M., Dlugolinsky S., Seleng M., Kvassay M., Schneider B., Bracker H., Wrzeszcz M., Kitowski J., Hluchy L.: Agent-based simulation platform evaluation in the context of human behavior modeling. In *Proceedings of the second international workshop on Infrastructures and Tools for Multiagent Systems*, pp. 396–410, 2011.
[13] Luke S., Cioffi-Revilla C., Panait L., Sullivan K., Balan G.: Mason: A multi-agent simulation environment. *Simulation: Transactions of the society for Modeling and Simulation International*, 82(7):517–527, 2005.
[14] Marco J., Campos I., Coterillo I., Diaz I., Lopez A., Marco R., Martinez-Rivero C., Orvis P., Rodriguez D., Gomes J., Borges G., Montecelo M.,
David M., Silva B., Dias N., Martins J.P., Fernandez C., Garcia-Tarres L.,
Veiga C., Cordero D., Lopez Cacheiro J., Lopez I., Garcia-Tobio J., Costas N.,
Mourino J.C., Gomez A., Bogacki W., Meyer N., Owsiak M., Plociennik M.,
Pospieszny M., Zawadzki M., Hammad A., Hardt M., Fernandez E., Heymann E.,
Senar M.A., Padee A., Nawrocki K., Wislicki W., Heinzreiter P., Baumgartner M.,
Rosmanith H., Kranzmuller D., Volkert J., Kenny S., Coghlann B., Pajak R.,
Mosurska Z., Szymocha T., Lason P., Skital L., Funika W., Korcyl K.,
Pieczykolan J., Balos K., Slota R., Guzy K., Dutka L., Kitowski J., Zielinski K.,
Hluchy L., Dobrucky M., Simo B., Habala O., Astalos J., Ciglan M., Sipkova V.,
Babik M., Gatial E., Valles R., Reynolds J.M., Serrano F., Tarancon A., Velasco
J.L., Cstejon F., Dichev K., Keller R., Stork S.: The interactive european grid:
Project objectives and achievements. Computing and Informatics, 27(2):161–171,
2008.
[15] Nguyen N., Miller A. J.: A review of some exchange algorithms for constructing
discrete d-optimal designs. Computational Statistics & Data Analysis, 14(4):489–
498, 1992.
Affiliations
Bartosz Kryza
AGH University of Science and Technology, ACC Cyfronet AGH, Krakow, Poland,
email@example.com
Dariusz Król
AGH University of Science and Technology, ACC Cyfronet AGH, Krakow, Poland,
firstname.lastname@example.org
Michał Wrzeszcz
AGH University of Science and Technology, ACC Cyfronet AGH, Krakow, Poland,
email@example.com
Łukasz Dutka
AGH University of Science and Technology, ACC Cyfronet AGH, Krakow, Poland
Jacek Kitowski
AGH University of Science and Technology, Faculty of Electrical Engineering, Automatics,
Computer Science and Electronics, Department of Computer Science, ACC Cyfronet, AGH,
Krakow, Poland
Received: 25.01.2012
Revised: 26.03.2012
Accepted: 9.07.2012 |
Minimum Risk Training for Neural Machine Translation
Shiqi Shen†, Yong Cheng‡, Zhongjun He*, Wei He†, Hua Wu†, Maosong Sun†, Yang Liu†*
†State Key Laboratory of Intelligent Technology and Systems
Tsinghua National Laboratory for Information Science and Technology
Department of Computer Science and Technology, Tsinghua University, Beijing, China
‡Institute for Interdisciplinary Information Sciences, Tsinghua University, Beijing, China
†Baidu Inc., Beijing, China
{vicapple22, email@example.com, {hezhongjun, hewei06, firstname.lastname@example.org, {sms, email@example.com
Abstract
We propose minimum risk training for end-to-end neural machine translation. Unlike conventional maximum likelihood estimation, minimum risk training is capable of optimizing model parameters directly with respect to arbitrary evaluation metrics, which are not necessarily differentiable. Experiments show that our approach achieves significant improvements over maximum likelihood estimation on a state-of-the-art neural machine translation system across various languages pairs. Transparent to architectures, our approach can be applied to more neural networks and potentially benefit more NLP tasks.
1 Introduction
Recently, end-to-end neural machine translation (NMT) (Kalchbrenner and Blunsom, 2013; Sutskever et al., 2014; Bahdanau et al., 2015) has attracted increasing attention from the community. Providing a new paradigm for machine translation, NMT aims at training a single, large neural network that directly transforms a source-language sentence to a target-language sentence without explicitly modeling latent structures (e.g., word alignment, phrase segmentation, phrase reordering, and SCFG derivation) that are vital in conventional statistical machine translation (SMT) (Brown et al., 1993; Koehn et al., 2003; Chiang, 2005).
Current NMT models are based on the encoder-decoder framework (Cho et al., 2014; Sutskever et al., 2014), with an encoder to read and encode a source-language sentence into a vector, from which a decoder generates a target-language sentence. While early efforts encode the input into a fixed-length vector, Bahdanau et al. (2015) advocate the attention mechanism to dynamically generate a context vector for a target word being generated.
Although NMT models have achieved results on par with or better than conventional SMT, they still suffer from a major drawback: the models are optimized to maximize the likelihood of training data instead of evaluation metrics that actually quantify translation quality. Ranzato et al. (2015) indicate two drawbacks of maximum likelihood estimation (MLE) for NMT. First, the models are only exposed to the training distribution instead of model predictions. Second, the loss function is defined at the word level instead of the sentence level.
In this work, we introduce minimum risk training (MRT) for neural machine translation. The new training objective is to minimize the expected loss (i.e., risk) on the training data. MRT has the following advantages over MLE:
1. Direct optimization with respect to evaluation metrics: MRT introduces evaluation metrics as loss functions and aims to minimize expected loss on the training data.
2. Applicable to arbitrary loss functions: our approach allows arbitrary sentence-level loss functions, which are not necessarily differentiable.
3. Transparent to architectures: MRT does not assume the specific architectures of NMT and can be applied to any end-to-end NMT systems.
While MRT has been widely used in conventional SMT (Och, 2003; Smith and Eisner, 2006; He and Deng, 2012) and deep learning based MT (Gao et al., 2014), to the best of our knowledge,
this work is the first effort to introduce MRT into end-to-end NMT. Experiments on a variety of language pairs (Chinese-English, English-French, and English-German) show that MRT leads to significant improvements over MLE on a state-of-the-art NMT system (Bahdanau et al., 2015).
2 Background
Given a source sentence \( x = x_1, \ldots, x_m, \ldots, x_M \) and a target sentence \( y = y_1, \ldots, y_n, \ldots, y_N \), end-to-end NMT directly models the translation probability:
\[
P(y|x; \theta) = \prod_{n=1}^{N} P(y_n|x, y_{<n}; \theta),
\]
where \( \theta \) is a set of model parameters and \( y_{<n} = y_1, \ldots, y_{n-1} \) is a partial translation.
Predicting the \( n \)-th target word can be modeled by using a recurrent neural network:
\[
P(y_n|x, y_{<n}; \theta) \propto \exp \left\{ q(y_{n-1}, z_n, c_n, \theta) \right\},
\]
where \( z_n \) is the \( n \)-th hidden state on the target side, \( c_n \) is the context for generating the \( n \)-th target word, and \( q(\cdot) \) is a non-linear function. Current NMT approaches differ in calculating \( z_n \) and \( c_n \) and defining \( q(\cdot) \). Please refer to (Sutskever et al., 2014; Bahdanau et al., 2015) for more details.
Given a set of training examples \( D = \{(x^{(s)}, y^{(s)})\}_{s=1}^{S} \), the standard training objective is to maximize the log-likelihood of the training data:
\[
\hat{\theta}_{MLE} = \arg\max_{\theta} \left\{ L(\theta) \right\},
\]
where
\[
L(\theta) = \sum_{s=1}^{S} \log P(y^{(s)}|x^{(s)}; \theta)
\]
\[
= \sum_{s=1}^{S} \sum_{n=1}^{N^{(s)}} \log P(y_n^{(s)}|x^{(s)}, y_{<n}^{(s)}; \theta).
\]
We use \( N^{(s)} \) to denote the length of the \( s \)-th target sentence \( y^{(s)} \).
The partial derivative with respect to a model parameter \( \theta_i \) is calculated as
\[
\frac{\partial L(\theta)}{\partial \theta_i} = \sum_{s=1}^{S} \sum_{n=1}^{N^{(s)}} \frac{\partial P(y_n^{(s)}|x^{(s)}, y_{<n}^{(s)}; \theta)/\partial \theta_i}{P(y_n^{(s)}|x^{(s)}, y_{<n}^{(s)}; \theta)}.
\]
Ranzato et al. (2015) point out that MLE for end-to-end NMT suffers from two drawbacks. First, while the models are trained only on the training data distribution, they are used to generate target words on previous model predictions, which can be erroneous, at test time. This is referred to as exposure bias (Ranzato et al., 2015). Second, MLE usually uses the cross-entropy loss focusing on word-level errors to maximize the probability of the next correct word, which might hardly correlate well with corpus-level and sentence-level evaluation metrics such as BLEU (Papineni et al., 2002) and TER (Snover et al., 2006).
As a result, it is important to introduce new training algorithms for end-to-end NMT to include model predictions during training and optimize model parameters directly with respect to evaluation metrics.
3 Minimum Risk Training for Neural Machine Translation
Minimum risk training (MRT), which aims to minimize the expected loss on the training data, has been widely used in conventional SMT (Och, 2003; Smith and Eisner, 2006; He and Deng, 2012) and deep learning based MT (Gao et al., 2014). The basic idea is to introduce evaluation metrics as loss functions and assume that the optimal set of model parameters should minimize the expected loss on the training data.
Let \( (x^{(s)}, y^{(s)}) \) be the \( s \)-th sentence pair in the training data and \( y \) be a model prediction. We use a loss function \( \Delta(y, y^{(s)}) \) to measure the discrepancy between the model prediction \( y \) and the gold-standard translation \( y^{(s)} \). Such a loss function can be negative smoothed sentence-level evaluation metrics such as BLEU (Papineni et al., 2002), NIST (Doddington, 2002), TER (Snover et al., 2006), or METEOR (Lavie and Denkowski, 2009) that have been widely used in machine translation evaluation. Note that a loss function is not parameterized and thus not differentiable.
In MRT, the risk is defined as the expected loss with respect to the posterior distribution:
\[
R(\theta) = \sum_{s=1}^{S} E_{y|x^{(s)}, \theta} \left[ \Delta(y, y^{(s)}) \right]
\]
\[
= \sum_{s=1}^{S} \sum_{y \in Y(x^{(s)})} P(y|x^{(s)}, \theta) \Delta(y, y^{(s)}),
\]
where $\mathcal{Y}(\mathbf{x}^{(s)})$ is a set of all possible candidate translations for $\mathbf{x}^{(s)}$.
The training objective of MRT is to minimize the risk on the training data:
$$\hat{\theta}_{\text{MRT}} = \arg\min_{\theta} \left\{ R(\theta) \right\}. \quad (9)$$
Intuitively, while MLE aims to maximize the likelihood of training data, our training objective is to discriminate between candidates. For example, in Table 1, suppose the candidate set $\mathcal{Y}(\mathbf{x}^{(s)})$ contains only three candidates: $y_1$, $y_2$, and $y_3$. According to the losses calculated by comparing with the gold-standard translation $y^{(s)}$, it is clear that $y_1$ is the best candidate, $y_3$ is the second best, and $y_2$ is the worst: $y_1 > y_3 > y_2$. The right half of Table 1 shows four models. As model 1 (column 3) ranks the candidates in a reverse order as compared with the gold-standard (i.e., $y_2 > y_3 > y_1$), it obtains the highest risk of $-0.50$. Achieving a better correlation with the gold-standard than model 1 by predicting $y_3 > y_1 > y_2$, model 2 (column 4) reduces the risk to $-0.61$. As model 3 (column 5) ranks the candidates in the same order with the gold-standard, the risk goes down to $-0.71$. The risk can be further reduced by concentrating the probability mass on $y_1$ (column 6). As a result, by minimizing the risk on the training data, we expect to obtain a model that correlates well with the gold-standard.
In MRT, the partial derivative with respect to a model parameter $\theta_i$ is given by
$$\frac{\partial R(\theta)}{\partial \theta_i} = \sum_{s=1}^{S} E_{y|\mathbf{x}^{(s)};\theta} \left[ \Delta(y, y^{(s)}) \times \sum_{n=1}^{N^{(s)}} \frac{\partial P(y_n|\mathbf{x}^{(s)}, y_{<n};\theta)}{\partial \theta_i} / P(y_n|\mathbf{x}^{(s)}, y_{<n};\theta) \right]. \quad (10)$$
Since Eq. (10) suggests there is no need to differentiate $\Delta(y, y^{(s)})$, MRT allows arbitrary non-differentiable loss functions. In addition, our approach is transparent to architectures and can be applied to arbitrary end-to-end NMT models.
Despite these advantages, MRT faces a major challenge: the expectations in Eq. (10) are usually intractable to calculate due to the exponential search space of $\mathcal{Y}(\mathbf{x}^{(s)})$, the non-decomposability of the loss function $\Delta(y, y^{(s)})$, and the context sensitiveness of NMT.
To alleviate this problem, we propose to only use a subset of the full search space to approximate the posterior distribution and introduce a new training objective:
$$\tilde{R}(\theta) = \sum_{s=1}^{S} E_{y|\mathbf{x}^{(s)};\theta,\alpha} \left[ \Delta(y, y^{(s)}) \right] \quad (11)$$
$$= \sum_{s=1}^{S} \sum_{y \in S(\mathbf{x}^{(s)})} Q(y|\mathbf{x}^{(s)};\theta,\alpha) \Delta(y, y^{(s)}), \quad (12)$$
where $S(\mathbf{x}^{(s)}) \subset \mathcal{Y}(\mathbf{x}^{(s)})$ is a sampled subset of the full search space, and $Q(y|\mathbf{x}^{(s)};\theta,\alpha)$ is a distribution defined on the subspace $S(\mathbf{x}^{(s)})$:
$$Q(y|\mathbf{x}^{(s)};\theta,\alpha) = \frac{P(y|\mathbf{x}^{(s)};\theta)^{\alpha}}{\sum_{y' \in S(\mathbf{x}^{(s)})} P(y'|\mathbf{x}^{(s)};\theta)^{\alpha}}. \quad (13)$$
Note that $\alpha$ is a hyper-parameter that controls the sharpness of the $Q$ distribution (Och, 2003).
Algorithm 1 shows how to build $S(\mathbf{x}^{(s)})$ by sampling the full search space. The sampled subset initializes with the gold-standard translation (line 1). Then, the algorithm keeps sampling a target word given the source sentence and the partial translation until reaching the end of sentence (lines 3–16). Note that sampling might produce duplicate candidates, which are removed when building
Input: the $s$-th source sentence in the training data $\mathbf{x}^{(s)}$, the $s$-th target sentence in the training data $\mathbf{y}^{(s)}$, the set of model parameters $\theta$, the limit on the length of a candidate translation $l$, and the limit on the size of sampled space $k$.
Output: sampled space $S(\mathbf{x}^{(s)})$.
1 $S(\mathbf{x}^{(s)}) \leftarrow \{\mathbf{y}^{(s)}\}$; // the gold-standard translation is included
2 $i \leftarrow 1$;
3 while $i \leq k$ do
4 $\mathbf{y} \leftarrow \emptyset$; // an empty candidate translation
5 $n \leftarrow 1$;
6 while $n \leq l$ do
7 $\mathbf{y} \sim P(\mathbf{y}_n|\mathbf{x}^{(s)}, \mathbf{y}_{<n}; \theta)$; // sample the $n$-th target word
8 $\mathbf{y} \leftarrow \mathbf{y} \cup \{\mathbf{y}\}$;
9 if $\mathbf{y} =$ EOS then
10 break; // terminate if reach the end of sentence
11 end
12 $n \leftarrow n + 1$;
13 end
14 $S(\mathbf{x}^{(s)}) \leftarrow S(\mathbf{x}^{(s)}) \cup \{\mathbf{y}\}$;
15 $i \leftarrow i + 1$;
16 end
Algorithm 1: Sampling the full search space.
the subspace. We find that it is inefficient to force the algorithm to generate exactly $k$ distinct candidates because high-probability candidates can be sampled repeatedly, especially when the probability mass highly concentrates on a few candidates. In practice, we take advantage of GPU’s parallel architectures to speed up the sampling.\footnote{To build the subset, an alternative to sampling is computing top-$k$ translations. We prefer sampling to computing top-$k$ translations for efficiency: sampling is more efficient and easy-to-implement than calculating $k$-best lists, especially given the extremely parallel architectures of GPUs.}
Given the sampled space, the partial derivative with respect to a model parameter $\theta_i$ of $\hat{R}(\theta)$ is given by
$$\frac{\partial \hat{R}(\theta)}{\partial \theta_i} = \alpha \sum_{s=1}^{S} \mathbb{E}_{\mathbf{y}|\mathbf{x}^{(s)}, \theta, \alpha} \left[ \frac{\partial P(\mathbf{y}|\mathbf{x}^{(s)}; \theta)}{\partial \theta_i} / P(\mathbf{y}|\mathbf{x}^{(s)}; \theta) \times \right.$$
$$\left( \Delta(\mathbf{y}, \mathbf{y}^{(s)}) - \mathbb{E}_{\mathbf{y}'|\mathbf{x}^{(s)}, \theta, \alpha}[\Delta(\mathbf{y}', \mathbf{y}^{(s)}))] \right) \] . (14)
Since $|S(\mathbf{x}^{(s)})| \ll |\mathcal{V}(\mathbf{x}^{(s)})|$, the expectations in Eq. (14) can be efficiently calculated by explicitly enumerating all candidates in $S(\mathbf{x}^{(s)})$. In our experiments, we find that approximating the full-space with 100 samples (i.e., $k = 100$) works very well and further increasing sample size does not lead to significant improvements (see Section 4.3).
4 Experiments
4.1 Setup
We evaluated our approach on three translation tasks: Chinese-English, English-French, and English-German. The evaluation metric is BLEU (Papineni et al., 2002) as calculated by the multi-bleu.perl script.
For Chinese-English, the training data consists of 2.56M pairs of sentences with 67.5M Chinese words and 74.8M English words, respectively. We used the NIST 2006 dataset as the validation set (hyper-parameter optimization and model selection) and the NIST 2002, 2003, 2004, 2005, and 2008 datasets as test sets.
For English-French, to compare with the results reported by previous work on end-to-end NMT (Sutskever et al., 2014; Bahdanau et al., 2015; Jean et al., 2015; Luong et al., 2015b), we used the same subset of the WMT 2014 training corpus that contains 12M sentence pairs with 304M English words and 348M French words. The concatenation of news-test 2012 and news-test 2013 serves as the validation set and news-test 2014 as the test set.
For English-German, to compare with the results reported by previous work (Jean et al., 2015; Luong et al., 2015a), we used the same subset of the WMT 2014 training corpus that contains 4M sentence pairs with 91M English words and 87M German words. The concatenation of news-test 2012 and news-test 2013 is used as the validation set and news-test 2014 as the test set.
We compare our approach with two state-of-the-art SMT and NMT systems:
1. **MOSES** (Koehn and Hoang, 2007): a phrase-based SMT system using minimum error rate training (Och, 2003).
2. **RNNSEARCH** (Bahdanau et al., 2015): an attention-based NMT system using maximum likelihood estimation.
**MOSES** uses the parallel corpus to train a phrase-based translation model and the target part to train a 4-gram language model using the SRILM toolkit (Stolcke, 2002). The log-linear model Moses uses is trained by the minimum error rate training (MERT) algorithm (Och, 2003) that directly optimizes model parameters with respect to evaluation metrics.
**RNNSEARCH** uses the parallel corpus to train an attention-based neural translation model using the maximum likelihood criterion.
On top of **RNNSEARCH**, our approach replaces MLE with MRT. We initialize our model with the RNNsearch50 model (Bahdanau et al., 2015). We set the vocabulary size to 30K for Chinese-English and English-French and 50K for English-German. The beam size for decoding is 10. The default loss function is negative smoothed sentence-level BLEU.
### 4.2 Effect of $\alpha$
The hyper-parameter $\alpha$ controls the smoothness of the $Q$ distribution (see Eq. (13)). As shown in Figure 1, we find that $\alpha$ has a critical effect on BLEU scores on the Chinese-English validation set. While $\alpha = 1 \times 10^{-1}$ decreases BLEU scores dramatically, $\alpha = 5 \times 10^{-3}$ improves translation quality significantly and consistently. Reducing $\alpha$ further to $1 \times 10^{-4}$, however, results in lower BLEU scores. Therefore, we set $\alpha = 5 \times 10^{-3}$ in the following experiments.
### 4.3 Effect of Sample Size
For efficiency, we sample $k$ candidate translations from the full search space $\mathcal{Y}(x^{(s)})$ to build an approximate posterior distribution $Q$ (Section 3). Figure 2 shows the effect of sample size $k$ on the Chinese-English validation set. It is clear that BLEU scores consistently rise with the increase of $k$. However, we find that a sample size larger than 100 (e.g., $k = 200$) usually does not lead to significant improvements and increases the GPU memory requirement. Therefore, we set $k = 100$ in the following experiments.
### 4.4 Effect of Loss Function
As our approach is capable of incorporating evaluation metrics as loss functions, we investigate the effect of different loss functions on BLEU, TER
| criterion | loss | BLEU | TER | NIST |
|-----------|--------|------|------|------|
| MLE | N/A | 30.48| 60.85| 8.26 |
| MRT | −sBLEU | 36.71| 53.48| 8.90 |
| | sTER | 30.14| 53.83| 6.02 |
| | −sNIST | 32.32| 56.85| 8.90 |
Table 2: Effect of loss function on the Chinese-English validation set.
---
\footnote{It is possible to exploit larger monolingual corpora for both MOSES and RNNSEARCH (Gulcehre et al., 2015; Senrich et al., 2015). We leave this for future work.}
and NIST scores on the Chinese-English validation set. As shown in Table 2, negative smoothed sentence-level BLEU (i.e., $-\text{sBLEU}$) leads to statistically significant improvements over MLE ($p < 0.01$). Note that the loss functions are all defined at the sentence level while evaluation metrics are calculated at the corpus level. This discrepancy might explain why optimizing with respect to sTER does not result in the lowest TER on the validation set. As $-\text{sBLEU}$ consistently improves all evaluation metrics, we use it as the default loss function in our experiments.
4.5 Comparison of Training Time
We used a cluster with 20 Tesla K40 GPUs to train the NMT model. For MLE, it takes the cluster about one hour to train 20,000 mini-batches, each of which contains 80 sentences. The training time for MRT is longer than MLE: 13,000 mini-batches can be processed in one hour on the same cluster.
Figure 3 shows the learning curves of MLE and MRT on the validation set. For MLE, the BLEU score reaches its peak after about 20 hours and then keeps going down dramatically. Initializing with the best MLE model, MRT increases BLEU scores dramatically within about 30 hours.\footnote{Although it is possible to initialize with a randomized model, it takes much longer time to converge.} Afterwards, the BLEU score keeps improving gradually but there are slight oscillations.
4.6 Results on Chinese-English Translation
4.6.1 Comparison of BLEU Scores
Table 3 shows BLEU scores on Chinese-English datasets. For RNNSEARCH, we follow Luong et al. (2015b) to handle rare words. We find that introducing minimum risk training into neural MT leads to surprisingly substantial improvements over MOSES and RNNSEARCH with MLE as the training criterion (up to +8.61 and +7.20 BLEU points, respectively) across all test sets. All the improvements are statistically significant.
4.6.2 Comparison of TER Scores
Table 4 gives TER scores on Chinese-English datasets. The loss function used in MRT is $-\text{sBLEU}$. MRT still obtains dramatic improvements over MOSES and RNNSEARCH with MLE as the training criterion (up to -10.27 and -8.32 TER points, respectively) across all test sets. All the improvements are statistically significant.
4.6.3 BLEU Scores over Sentence Lengths
Figure 4 shows the BLEU scores of translations generated by MOSES, RNNSEARCH with M| System | Training | MT06 | MT02 | MT03 | MT04 | MT05 | MT08 |
|----------|----------|-------|-------|-------|-------|-------|-------|
| MOSES | MERT | 32.74 | 32.49 | 32.40 | 33.38 | 30.20 | 25.28 |
| RNNSEARCH| MLE | 30.70 | 35.13 | 33.73 | 34.58 | 31.76 | 23.57 |
| | MRT | 37.34 | 40.36 | 40.93 | 41.37 | 38.81 | 29.23 |
Table 3: Case-insensitive BLEU scores on Chinese-English translation.
| System | Training | MT06 | MT02 | MT03 | MT04 | MT05 | MT08 |
|----------|----------|-------|-------|-------|-------|-------|-------|
| MOSES | MERT | 59.22 | 62.97 | 62.44 | 61.20 | 63.44 | 62.36 |
| RNNSEARCH| MLE | 60.74 | 58.94 | 60.10 | 58.91 | 61.74 | 64.52 |
| | MRT | 52.86 | 52.87 | 52.17 | 51.49 | 53.42 | 57.21 |
Table 4: Case-insensitive TER scores on Chinese-English translation.
| MLE vs. MRT | < | = | > |
|-------------|---|---|---|
| evaluator 1 | 54% | 24% | 22% |
| evaluator 2 | 53% | 22% | 25% |
Table 5: Subjective evaluation of MLE and MRT on Chinese-English translation.
LE, and RNNSEARCH with MRT on the Chinese-English test set with respect to input sentence lengths. While MRT consistently improves over MLE for all lengths, it achieves worse translation performance for sentences longer than 60 words.
One reason is that RNNSEARCH tends to produce short translations for long sentences. As shown in Figure 5, both MLE and MRE generate much shorter translations than MOSES. This results from the length limit imposed by RNNSEARCH for efficiency reasons: a sentence in the training set is no longer than 50 words. This limit deteriorates translation performance because the sentences in the test set are usually longer than 50 words.
### 4.6.4 Subjective Evaluation
We also conducted a subjective evaluation to validate the benefit of replacing MLE with MRT. Two human evaluators were asked to compare MLE and MRT translations of 100 source sentences randomly sampled from the test sets without knowing from which system a candidate translation was generated.
Table 5 shows the results of subjective evaluation. The two human evaluators made close judgments: around 54% of MLE translations are worse than MRE, 23% are equal, and 23% are better.
### 4.6.5 Example Translations
Table 6 shows some example translations. We find that MOSES translates a Chinese string “yi wei fuze yu pingrang dangju da jiaodao de qian guowuyuan guanyuan” that requires long-distance reordering in a wrong way, which is a notorious challenge for statistical machine translation. In contrast, RNNSEARCH-MLE seems to overcome this problem in this example thanks to the capability of gated RNNs to capture long-distance dependencies. However, as MLE uses a loss function defined only at the word level, its translation lacks sentence-level consistency: “chinese” occurs twice while “two senate” is missing. By optimizing model parameters directly with respect to sentence-level BLEU, RNNSEARCH-MRT seems to be able to generate translations more consistently at the sentence level.
### 4.7 Results on English-French Translation
Table 7 shows the results on English-French translation. We list existing end-to-end NMT systems that are comparable to our system. All these systems use the same subset of the WMT 2014 training corpus and adopt MLE as the training criterion. They differ in network architectures and vocabulary sizes. Our RNNSEARCH-MLE system achieves a BLEU score comparable to that of Jean et al. (2015). RNNSEARCH-MRT achieves the highest BLEU score in this setting even with a vocabulary size smaller than Luong et al. (2015b) and Sutskever et al. (2014). Note that our approach does not assume specific architectures and can in principle be applied to any NMT systems.
### 4.8 Results on English-German Translation
Table 8 shows the results on English-German translation. Our approach still significantly out| Source | meiguo daibiao tuan baokuo laizi shidanfu daxue de yi wei zhongguo zhuanjia , liang ming canyuan waijiao zhengce zhuli yiji yi wei fuze yu pingrang dangju da jiaodao de qian guoyuyuan guanyuan . |
|--------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Reference | the us delegation consists of a chinese expert from the stanford university , two senate foreign affairs policy assistants and a former state department official who was in charge of dealing with pyongyang authority . |
| MOSES | the united states to members of the delegation include representatives from the stanford university , a chinese expert , two assistant senate foreign policy and a responsible for dealing with pyongyang before the officials of the state council . |
| RNNSEARCH-MLE | the us delegation comprises a chinese expert from stanford university , a chinese foreign office assistant policy assistant and a former official who is responsible for dealing with the pyongyang authorities . |
| RNNSEARCH-MRT | the us delegation included a chinese expert from the stanford university , two senate foreign policy assistants , and a former state department official who had dealings with the pyongyang authorities . |
Table 6: Example Chinese-English translations. “Source” is a romanized Chinese sentence, “Reference” is a gold-standard translation. “MOSES” and “RNNSEARCH-MLE” are baseline SMT and NMT systems. “RNNSEARCH-MRT” is our system.
| System | Architecture | Training | Vocab | BLEU |
|--------|--------------|----------|-------|------|
| **Existing end-to-end NMT systems** | | | | |
| Bahdanau et al. (2015) | gated RNN with search | MLE | 30K | 28.45 |
| Jean et al. (2015) | gated RNN with search | 30K | 29.97 |
| Jean et al. (2015) | gated RNN with search + PosUnk | 30K | 33.08 |
| Luong et al. (2015b) | LSTM with 4 layers | 40K | 29.50 |
| Luong et al. (2015b) | LSTM with 4 layers + PosUnk | 40K | 31.80 |
| Luong et al. (2015b) | LSTM with 6 layers | 40K | 30.40 |
| Luong et al. (2015b) | LSTM with 6 layers + PosUnk | 40K | 32.70 |
| Sutskever et al. (2014) | LSTM with 4 layers | 80K | 30.59 |
| **Our end-to-end NMT systems** | | | | |
| this work | gated RNN with search | MLE | 30K | 29.88 |
| this work | gated RNN with search | MRT | 30K | 31.30 |
| this work | gated RNN with search + PosUnk | MRT | 30K | 34.23 |
Table 7: Comparison with previous work on English-French translation. The BLEU scores are case-sensitive. “PosUnk” denotes Luong et al. (2015b)’s technique of handling rare words.
| System | Architecture | Training | BLEU |
|--------|--------------|----------|------|
| **Existing end-to-end NMT systems** | | | |
| Jean et al. (2015) | gated RNN with search | MLE | 16.46 |
| Jean et al. (2015) | gated RNN with search + PosUnk | 18.97 |
| Jean et al. (2015) | gated RNN with search + LV + PosUnk | 19.40 |
| Luong et al. (2015a) | LSTM with 4 layers + dropout + local att. + PosUnk | 20.90 |
| **Our end-to-end NMT systems** | | | |
| this work | gated RNN with search | MLE | 16.45 |
| this work | gated RNN with search | MRT | 18.02 |
| this work | gated RNN with search + PosUnk | MRT | 20.45 |
Table 8: Comparison with previous work on English-German translation. The BLEU scores are case-sensitive.
performs MLE and achieves comparable results with state-of-the-art systems even though Luong et al. (2015a) used a much deeper neural network. We believe that our work can be applied to their architecture easily.
Despite these significant improvements, the margins on English-German and English-French datasets are much smaller than Chinese-English. We conjecture that there are two possible reasons. First, the Chinese-English datasets contain four reference translations for each sentence while both English-French and English-German datasets only have single references. Second, Chinese and English are more distantly related than English, French and German and thus benefit more from MRT that incorporates evaluation metrics into optimization to capture structural divergence.
5 Related Work
Our work originated from the minimum risk training algorithms in conventional statistical machine translation (Och, 2003; Smith and Eisner, 2006; He and Deng, 2012). Och (2003) describes a smoothed error count to allow calculating gradients, which directly inspires us to use a parameter $\alpha$ to adjust the smoothness of the objective function. As neural networks are non-linear, our approach has to minimize the expected loss on the sentence level rather than the loss of 1-best translations on the corpus level. Smith and Eisner (2006) introduce minimum risk annealing for training log-linear models that is capable of gradually annealing to focus on the 1-best hypothesis. He et al. (2012) apply minimum risk training to learning phrase translation probabilities. Gao et al. (2014) leverage MRT for learning continuous phrase representations for statistical machine translation. The difference is that they use MRT to optimize a sub-model of SMT while we are interested in directly optimizing end-to-end neural translation models.
The Mixed Incremental Cross-Entropy Reinforce (MIXER) algorithm (Ranzato et al., 2015) is in spirit closest to our work. Building on the REINFORCE algorithm proposed by Williams (1992), MIXER allows incremental learning and the use of hybrid loss function that combines both REINFORCE and cross-entropy. The major difference is that Ranzato et al. (2015) leverage reinforcement learning while our work resorts to minimum risk training. In addition, MIXER only samples one candidate to calculate reinforcement reward while MRT generates multiple samples to calculate the expected risk. Figure 2 indicates that multiple samples potentially increases MRT’s capability of discriminating between diverse candidates and thus benefit translation quality. Our experiments confirm Ranzato et al. (2015)’s finding that taking evaluation metrics into account when optimizing model parameters does help to improve sentence-level text generation.
More recently, our approach has been successfully applied to summarization (Ayana et al., 2016). They optimize neural networks for headline generation with respect to ROUGE (Lin, 2004) and also achieve significant improvements, confirming the effectiveness and applicability of our approach.
6 Conclusion
In this paper, we have presented a framework for minimum risk training in end-to-end neural machine translation. The basic idea is to minimize the expected loss in terms of evaluation metrics on the training data. We sample the full search space to approximate the posterior distribution to improve efficiency. Experiments show that MRT leads to significant improvements over maximum likelihood estimation for neural machine translation, especially for distantly-related languages such as Chinese and English.
In the future, we plan to test our approach on more language pairs and more end-to-end neural MT systems. It is also interesting to extend minimum risk training to minimum risk annealing following Smith and Eisner (2006). As our approach is transparent to loss functions and architectures, we believe that it will also benefit more end-to-end neural architectures for other NLP tasks.
Acknowledgments
This work was done while Shiqi Shen and Yong Cheng were visiting Baidu. Maosong Sun and Hua Wu are supported by the 973 Program (2014CB340501 & 2014CB34505). Yang Liu is supported by the National Natural Science Foundation of China (No.61522204 and No.61432013) and the 863 Program (2015AA015407). This research is also supported by the Singapore National Research Foundation under its International Research Centre@Singapore Funding Initiative and administered by the IDM Programme.
References
Ayana, Shiqi Shen, Zhiyuan Liu, and Maosong Sun. 2016. Neural headline generation with minimum risk training. arXiv:1604.01904.
Dzmitry Bahdanau, KyungHyun Cho, and Yoshua Bengio. 2015. Neural machine translation by jointly learning to align and translate. In Proceedings of ICLR.
Peter F. Brown, Stephen A. Della Pietra, Vincent J. Della Pietra, and Robert L. Mercer. 1993. The mathematics of statistical machine translation: Parameter estimation. Computational Linguistics.
David Chiang. 2005. A hierarchical phrase-based model for statistical machine translation. In Proceedings of ACL.
Kyunghyun Cho, Bart van Merrienoort, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning phrase representations using rnn encoder-decoder for statistical machine translation. In Proceedings of EMNLP.
George Doddington. 2002. Automatic evaluation of machine translation quality using n-gram co-occurrence statistics. In Proceedings of HLT.
Jianfeng Gao, Xiaodong He, Wen tao Yih, and Li Deng. 2014. Learning continuous phrase representations for translation modeling. In Proceedings of ACL.
Caglar Gulcehre, Orhan Firat, Kelyn Xu, Kyunghyun Cho, Loic Barrault, Huei-Chi Lin, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2015. On using monolingual corpora in neural machine translation. arXiv:1503.03555.
Xiaodong He and Li Deng. 2012. Maximum expected bleu training of phrase and lexicon translation models. In Proceedings of ACL.
Sebastien Jean, Kyunghyun Cho, Roland Memisevic, and Yoshua Bengio. 2015. On using very large target vocabulary for neural machine translation. In Proceedings of ACL.
Nal Kalchbrenner and Phil Blunsom. 2013. Recurrent continuous translation models. In Proceedings of EMNLP.
Philipp Koehn and Hieu Hoang. 2007. Factored translation models. In Proceedings of EMNLP.
Philipp Koehn, Franz J. Och, and Daniel Marcu. 2003. Statistical phrase-based translation. In Proceedings of HLT-NAACL.
Alon Lavie and Michael Denkowski. 2009. The meteor metric for automatic evaluation of machine translation. Machine Translation.
Chin-Yew Lin. 2004. Rouge: A package for automatic evaluation of summaries. In Proceedings of ACL.
Minh-Thang Luong, Hieu Pham, and Christopher D Manning. 2015a. Effective approaches to attention-based neural machine translation. In Proceedings of EMNLP.
Minh-Thang Luong, Ilya Sutskever, Quoc V. Le, Oriol Vinyals, and Wojciech Zaremba. 2015b. Addressing the rare word problem in neural machine translation. In Proceedings of ACL.
Franz J. Och. 2003. Minimum error rate training in statistical machine translation. In Proceedings of ACL.
Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. Bleu: a method for automatic evaluation of machine translation. In Proceedings of ACL.
Marc’Aurelio Ranzato, Sumit Chopra, Michael Auli, and Wojciech Zaremba. 2015. Sequence level training with recurrent neural networks. arXiv:1511.06732v1.
Rico Sennrich, Barry Haddow, and Alexandra Birch. 2015. Improving neural machine translation models with monolingual data. arXiv:1511.06709.
David A. Smith and Jason Eisner. 2006. Minimum risk annealing for training log-linear models. In Proceedings of ACL.
Matthew Snover, Bonnie Dorr, Richard Schwartz, Linnea Micciulla, and John Makhoul. 2006. A study of translation edit rate with targeted human annotation. In Proceedings of AMTA.
Andreas Stolcke. 2002. Srilm - an extensible language modeling toolkit. In Proceedings of ICSLP.
Ilya Sutskever, Oriol Vinyals, and Quoc V. Le. 2014. Sequence to sequence learning with neural networks. In Proceedings of NIPS.
Ronald J. Williams. 1992. Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine Learning. |
Lab on the Web—Looking at Different Ways of Experiencing Electronic Experiments*
PER LUNDGREN, KJELL O. JEPPSON and ÅKE INGERMAN
Department of Microtechnology and Nanoscience, Chalmers University of Technology, S-412 96 Göteborg, Sweden. E-mail: firstname.lastname@example.org
Students who perform a laboratory exercise via the Internet display decisive differences in how they approach this resource: either as just a way to gather data, as illustrations of theory or as a tool for reflection and testing of hypotheses. From more than 200 written student log books, student interviews and video recorded lab sessions we conclude that the students’ approaches to measurements conducted through the Internet are very similar to that of measurements conducted with the students present in local classroom labs. Possible measures to support students in adopting a desirable approach to measurements are discussed.
INTRODUCTION
REMOTE LABORATORIES have recently emerged as new learning tools made possible by the rapid development of Internet technology [1]. Through remote laboratories measurement equipment can be made available via the Internet and serve either as cost-effective replacements for or as complements to traditional laboratory measurements. However, with all these new possibilities available the question is raised: What are students learning in measurement exercises? How do the students experience this new conceptual environment in general, and what is the specific experience they gain by investigating semiconductor devices via the Internet using remote measurement equipment? In this paper, we present a study where we have observed students in a course on semiconductor devices while they were trying to solve a course mini-project assignment. Our experiences are synthesized and formulated as a strategy for how we intend to use remote laboratories in the future.
The study has its origin in a department decision to improve our teaching strategy and to focus even more on the student learning environment. One part in that has been to start employing labs in which remote measurements are the essential ingredients. Reasons for this are both the intent to make more advanced equipment available to the students, equipment that we could only afford to make available through time-sharing, and to make it available around the clock throughout the course. The introduction of the remote laboratory components has been evaluated and reported [2] both in terms of student opinion, technical aspects and course logistics. In general, the student response has been positive.
The study presented in this paper has grown out of a wish to go further and to look deeper into issues that emerged out of the evaluation process. Hereby, we could provide grounds for further development, both from a technical point of view and from the perspective of course context. We have aimed to further explore student learning experiences from working with the remote laboratory. This means addressing questions like: How do students experience the remote laboratory—do they believe they perform real measurements or is it just taken as a virtual exercise which might just as well be simulated? What is in focal awareness for students when doing (remote) measurements? How do students approach their assignment? Is the assignment formulated so that it supports the intended learning process and course objective that we, as teachers, have in mind? This process is even more important to go through if seen against the background of the scarcity of more solidly pedagogically based studies of student learning in the laboratory generally. Existing examples are from physics in-class laboratories only [3, 4].
BACKGROUND AND COURSE CONTEXT
The most important goal we have set for the course where we use these remote measurement assignments is for students to understand how a transistor in general (or in particular, a metal-oxide-semiconductor field-effect transistor, MOSFET) works. In its original conception the course employs lectures, textbooks and exercises to present and discuss underlying phenomena to explain how semiconductor devices work. The
course organization is bottom-up with respect to content; it starts with resistors and using these as a reference, diodes and transistors are explained. Originally one hands-on laboratory exercise accompanied each of these course modules. In the present study the only ‘real’ hands-on laboratory exercise that remains concerns transistors, and this exercise is not linked to the examination of the rest of the course. Step by step the theoretical background is presented and the concept of device modeling is discussed. Finally, this leads to a description of the transistor function and a device model. Examination is mainly focused on the written exam at the end of the course.
In terms of assignments the students work in groups of four and receive individual grades in a project that contains three parts:
1. Resistor simulations.
2. Diode characterization on the Internet.
3. Transistor (MOSFET) characterization on the Internet.
The diodes are available via a server at Chalmers [2], and the transistors have been made available through the WebLab set-up at Massachusetts Institute of Technology (MIT) [1]. Pure electrical characterization, as in this case, makes the difference between remote and local control and read-out almost only a matter of distance and accessibility; there are no moving parts or much physical interaction involved in the laboratory tasks anyway [5]. The (resistor) assignment is examined by a short written report, the second (diode) assignment by a short oral presentation to peers, while the third (transistor) assignment is examined on an individual basis in an oral discussion led by the examiner but where the whole group is present. For the two latter assignments there is no scheduled supervision available for the students, and the group is entirely free to plan and perform their task as they choose. Some kick-off support is provided in conjunction with the first assignment, and a web forum is available for problem discussion. One supervising teacher answers e-mail questions or can be approached in person on student initiative. For a compelling description regarding the issues of tutoring in conjunction with remote lab usage see [6]. The first assignment is initiated and mainly performed under supervision in a ‘studio’ (where both lecturing and group computer work can occur during the same session) [7]. The project is counted as approximately 25% of the total examination directly but the content overlap to the written exam is so strong that the intended real importance of project work for examination is over 50% (making a good project work should give a strong advantage in the written exam).
The focus of this study is to gain insights into student experiences of the final part of the project: characterization of MOSFETs. In brief, this assignment was defined as familiarizing with the main aspects of the electrical characteristics of the device and to relate these to simple models. The students were explicitly told that they were expected to be able to draw the relevant characteristics (input-, output-, and transfer characteristics) and that they must be able to describe how the measured curves compare to theoretical ones.
**DATA COLLECTION AND QUALITATIVE ANALYSIS**
The focus of the research reported in this study is on the students’ experience of using the remote laboratories. The study is qualitative, and our objective was to understand in what context students saw the remote laboratory exercise. However, we also wanted to acquire some quantitative figures as a result of our study. At the same time we tried to keep the total amount of data and work invested in the study on a reasonable level.
For the study we decided on a strategy based on three main methods to collect information:
1. Interviews with six voluntary students.
2. Collection of student ‘log books’. Students were asked to write a ‘log file’ briefly describing at least one of their lab sessions.
3. Video recordings. Six students (working in pairs) were video recorded while accessing and working with the remote laboratory.
The interviews and video sessions were transcribed verbatim to allow closer analysis of the students’ experience. The interviews can be described as simple semi-structured interviews. Relevant topics were introduced by questions, and students were allowed to expand on the subject. Before proceeding to the next topic a number of follow-up questions were asked. For logistical reasons the interview duration was limited to about 15 minutes. The topics addressed were related to how students experienced the remote laboratory as a whole, and to how the group organized their collaboration. In particular, we investigated how students interpreted the objective of the exercise and what they perceived as the most important aspects of the MOSFET.
The objective of the video sessions was to acquire insights into how the students actually interacted with each other and with the remote laboratory in order to characterize the device under test. The video sessions were organized as follows:
- Students worked in groups of two and were monitored while working together on some part of the assignment while using the remote laboratory.
- Two cameras were used, one monitoring the students, and one monitoring the computer screen.
- The six students in this study had volunteered to be recorded and were instructed to work with the computer as they normally do with the addition of trying to articulate their thoughts.
and their actions while working with the assignment more explicitly than they otherwise would.
Most students (>200) responded to the request of writing a logbook (submitted as an e-mail) of their work with the remote laboratory. In these logbooks students described their experiences from one remote laboratory session of their own choice. They also answered a number of questions on who they worked with, on how they prepared, on what strategies they used, and on what was accomplished during the session. They were also asked to summarize their experiences and to keep track of the time they spent on the assignment. The size of the student logbooks varied considerably as well as the level on which their reflections were kept.
The analysis was primarily done by examining the available material in search of statements, quotes or passages, which could give us insight into or illuminate three main questions:
1. What are the significant differences in how the ‘real’ hands-on transistor assignment and the remote lab assignment is experienced?
2. How do the students spend their time using the remote laboratory?
3. From what perspective do the students approach the remote laboratory assignment?
The acquired material was then interpreted and condensed. The material was substantiated into observations regarding our three main questions, and also into categories describing students’ approaches as well as into suggestions for possible measures. The video sessions were the prime ground for the observations; the interviews were the most important ground for the categories. The log books gave a bulk insight into both of these aspects. With these characteristics our method is loosely based on a phenomenographic approach [8, 9] in a naturalistic setting, and in some ways resembles a naturalistic study [10].
RESULTS
On the basis of our empirical material we formulate three observations, which have implications for the implementation and organization of remote laboratory exercises in the context of the course.
Observation 1: If it can break, it can’t be fake
The students seem to be treating the remote laboratory as ‘real’ and not as artificial simulations. This is evident both from the video sessions and from the experiences shared in interviews. The students put the remote laboratory experience on the same footing as classroom laboratories. The potential possibility of the electronic device breaking down during measurements seems to be particularly important for the sense of ‘realness’. Device breakdown was reported by a number of students. One group said during the interview session that:
At first we thought it was more of a simulation, but on other occasions when it had crashed *laughter* [...] then you really felt that it was a transistor you were measuring on. [Interview 4]
In the logbooks we have the following example of a related experience:
14:30. Third time running. The first measurements proceed quite fast and we get nice plots. You can clearly see how the collector current varies with the different gate voltages. But when we measure on the gate-current we get a quite funny result with the current ‘jumping up and down’. But since the current is very small and I think it really shouldn’t be any current there, we treat it as negligible. When we are to measure the current as a function of the gate voltage, we get a strange result there as well. The plot does not look at all as we expected and when we plot the square root of the current as a function of the gate voltage we do not get a linear relationship, which we should get. Additionally we get some form of ‘mild current rush’ (that is that the current makes a big jump to a higher current) when we stepwise increase the control voltage to 3V. Furthermore, the current at that point is on the same level, independently of the voltage $V_{DS}$ we set. J thinks that the transistor is broken, but we think it is strange that we get any result, and we decide to bring the result and check with someone who knows, or ask on Friday. Now, in retrospect when we know the transistor was broken, we could have measured on the other transistor that was available, but we just didn’t think about it at the time. At about five we finished and anyhow felt pretty satisfied. [log]
From what we can judge, devices breaking down are important learning occasions concerning what is a reasonable measurement or not. However, from students belonging to the reference group that were allowed to perform both remote and traditional hands-on measurements on MOSFETs we had reports on different focuses. In traditional classroom laboratory sessions more time is spent on practical matters like physically handling the oscilloscope and measurement instruments. From the interviews:
How did you experience it [the classroom laboratory exercise]?
It was incredibly good, we [needed] perhaps a little more time, I think, perhaps more supervision and so on; it was very directed towards set stations; we moved around and things . . .
Did you feel prepared for the classroom lab exercise?
I did feel we were prepared in such a way that, we understood everything, but something that we weren’t prepared for were this insanely dangerous apparatus that you were supposed to adjust the oscilloscope with. It took basically more than half the time before we understood, because it was . . . in the way, so to speak.
What was the value of working with WebLab then?
I could imagine that for other groups it had much value, but for us it actually just felt like repetition, and really, we probably could have done our presentation of the MOSFET without that exercise, but of course it is always fun to have something to tinker with when you feel like applying a voltage to the bulk or something on the MOSFET, it is always fun to see what happens then.’ [interview 3]
Our conclusion is that, complemented by physically measuring on devices, the remote laboratory gives an experience similar to a traditional laboratory. When using the remote laboratory it actually seems like the students’ experience of the measurements is more focused on the function of the device, which is the central pedagogical part of the laboratory exercise. And with respect to the general purpose of including measurements, they seem to have an extraordinary illustrative power, and are experienced by students as developing their understanding.
As one student wrote in the log:
The most striking thing I have learnt (after having seen it in a measurement and then read in the book about it) is that there can be current even without voltage on the gate.
**Observation 2: Make it plot on the spot**
The video sessions clearly show that a significant portion of the students’ time is spent on technicalities, e.g., moving data from one program to another, or writing code to plot data. This is confirmed by the logs, in which students report on whole sessions of perhaps four hours spent on some technicality. In some ways this is hardly surprising, and some time spent on those tasks is inevitable. However, we doubt that this contributes to the students’ conceptual engagement and would like to reduce it (without reducing the openness of the students’ task). Integration of both the interface to the experimental set-up and associated useful tools (e.g., for plotting) into one common resource is one way of paving the way for the students in this respect [6].
**Observation 3: How do they do?**
In the video sessions, as well as in the interviews and the logs, it is clear that there is a large variation in how students approach and handle the remote laboratory and consequently in the learning outcome. For some students the remote laboratory seems to be an advanced illustration to the textbook, for some it is more of a verification that the textbook is ‘correct’ in its details and claims. Some students approach the laboratory as an unnecessary evil, while some see it as a central resource for reflection on aspects of the function of transistors. Of course we find some of these approaches more productive than others. We have tried to articulate these differences to the students when handing out the assignment. Afterwards, we have analyzed the outcome further and tried to categorize the different ways of understanding the measurement component of the laboratories; we have discerned four qualitatively different ways of understanding the measurement, or different foci when the students have engaged in the remote laboratory task.
1. **Imaging what to learn.** The first way of dealing with the measurements is to focus single-mindedly on the production of some plots. The learning outcome is rather limited and the students’ understanding of what the diagrams mean is rather fuzzy. At most one could hope that the remote lab measurements serve as an inspiration to further studies next week (or closer to the examination).
When we had met and captured the plots, that now we have to read up about this and that and see if we can find any explanation [. . . ] and then we gave hints for pages, also, if someone found a graph on a particular page, then we could say to the others, look on this page and try to understand. [Interview 1]
Have I learnt anything? I have learnt how $I_D$ depends on $V_{D}$. But since I don’t have any knowledge yet of why the curve looks like it does it feels like quite pointless information, I will read more of the theory the coming days. [log]
2. **Theory illustrations.** The second focus is on the theory, and the measurements are seen only as illustrations, let be slightly more advanced than the diagrams in the book.
*What is it you see when you are working with WebLab?*
Eehm, well, *I can’t answer for everybody else, but I tend to compare with what would theoretically happen when thinking about MOSFETs directly [. . . ] and then you can sit and try to figure out what it should look like [. . . ] and then there is an answer when you run it.* [Interview 5]
*You knew since one has been studying for the exam quite a lot, and then you knew quite a few characteristics compared to theoretical characteristics or to itself, but it was pretty interesting to see how it actually was, ehm, it is the biggest gain, sort of, that you got to see a real transistor in action.* [Interview 4]
*We prepared by having some coffee and thinking about what we knew. With help from assorted material from different courses we managed to cover how the transistor model was constructed. Then it was up to us to try to verify these theories. We tried to get the parameters that were needed to include the second order effects that the model is trying to show.* [log]
3. **Iterative progress.** The third focus is partly on the measurements, partly on theory. Students adopting this approach are trying to get a feeling for the function of the device. In this case there is room for fruitful transitions from measurements to theory and from theory to measurement.
It is when new questions arise, when you think you have understood everything, then you read it again, and then new questions surface and you can get them either confirmed or corrected if you enter the lab, on the WebLab and measure. But it is more or less only measured results that is shown, and it is really not a teacher standing there telling you that the mobility changes or anything like that, but if you have got anywhere in the course, then you can come to those conclusions on your own, I think. [Interview 3]
With WebLab we sat and tried different values and saw how the MOSFET behaved for the parameters. [Interview 2]
The main point I think is that you can do a measurement more or less based on something you have seen in the book for example and see that I get something more or less the same, and then I could change the parameters and see what happens if I change something from what we saw in the book. That you can’t do in the book, and more or less put a question to the program. [Interview 6]
We used the different graphs and step by step went through the origin of the current and how the different voltages affected the depletion region, mobility and such things. Then from that, we tried to make certain conclusions. [log]
4. **Hierarchical bottom-up characterization (reflecting around what is measured in relation to what is modeled).** This fourth category emerges from the discussion between the authors of this paper, and it was not found in the collected data. It is a desired way of approaching WebLab and the assignment— to focus on the measurements and characteristics in terms of the measurement results themselves, using WebLab actively to build knowledge with support from the surrounding context.
The following (fake) quote illustrates this way of understanding the measurement
To start with on this level, we could describe it with two straight lines, it gives us the form; it has slightly different behaviors in two different areas. A straight line, we can understand quite clearly in current-voltage characteristics: drift. In between there is some form of transition. Perhaps we can change scale to understand what is happening to a logarithmic scale [. . . ] here we can also see one part as a straight line, thus it is some kind of exponential behavior. We could compare that with self-generating growth, like bacteria.
Another way of describing these categories is presented in Fig. 1, where we employ the simile of different glasses for the four different approaches.
**DISCUSSION—POSSIBLE MEASURES**
We would like to support the students in developing a feeling for the function of the devices, having a focus on measurements as a starting point for modeling work, as well as seeing that as a goal of their learning. To put it succinct, we would like the students to develop a way of relating to the measurement on components as is described in category D above. In the light of this category the characterization of an electronic component includes the continuous and systematic shift between different appropriate scales and ways of organizing the characterizing data: now in a plot of the $I_D$ versus $V_G$; now in a plot of $I_G$ versus $V_G$; now in a log-lin plot of current versus voltage; now with several curves in the $I_D-V_G$ diagram; now in a up-scaled part of the $I_D-V_G$ plot, as some examples.
The implications of changing the course to have this as an important goal may be far-reaching for the total structure of the course. One way to do that could be to introduce a lab exercise, which is allowed to play a big part in the course, while other parts will become less focused on. The suggested assignment is structured in two parts:
1. Students get the task to measure on an unknown device and some indications on what might be relevant to measure. The task is not only to make measurements but to describe them in such a way that a reader (of the description) can understand what the measurements look like. If possible we want the students explicitly to change perspectives, scaling, etc., to find the most relevant way of describing them, where there are clear characteristics which make them distinct from other measurements. The report should be brief, with other students as the set audience. Preferably there will be several different unknown devices to choose from, out of which the students can measure on one. The students may (or may not) be organized in groups. To make the task more challenging, a requirement may be that the report should not have any explicit graphical information.
2. Students hand in their report and get someone else’s report in return. The second part of their task is to make conclusions on the basis of the received report about what kind of device that was measured on and motivate these conclusions. If that is too difficult, the task is instead to critique the report and explicitly explain why there is not enough information to make conclusions (and as far as possible to give
indications of what information is missing). For example, if the information points towards two possible components, a task could be to point out which distinguishing information that is missing. Finally, they should give feedback on the report according to certain criteria set by the course teachers (a form of peer evaluation). An additional part of the task could be to do measurements for reference on a known component and show how the given description matches these measurements.
Such structuring can be related to what Lippmann [4] expresses as a goal for the changes she has done as a consequence of her research: to ‘teach students to utilize their everyday skills of argument and decision-making for data gathering and analysis’.
We will start off in this direction, trying to use smaller measures to start with. The focus will initially be on addressing problems related to what is (or hitherto has been) assumed to be previous content knowledge for the students. That means reflecting on what has actually been included in previous courses and in what way the students have learned basic key foundations and to identify basic stumbling blocks (e.g. to have an idea how a plot changes when changing from a linear to a logarithmic scale).
**CONCLUSIONS**
From our investigation it appears that students approach measurements through a remote lab in much the same way as in a local lab environment. Much time is spent on data management. We can discern four qualitatively different foci for the students when engaged with their remote lab tasks, and we believe that the learning outcome can be promoted by helping students finding the most appropriate focus. We conclude that the remote laboratories in practice mean access to an abundance of cheap measurements and using them may allow pursuing other aims than we could do otherwise. For the specific object of learning considered in the present course, hands-on laboratory exercises might not be needed.
**Acknowledgements**—Thanks are due to the students who volunteered some of their time for interviews and video recordings. The excellent service of MIT WebLab is gratefully acknowledged. This study was funded by Chalmers Strategic Effort on Learning and Teaching (C-SELT). The video recordings were done by Magnus Axelsson of Chalmers Centre for Digital Media and Higher Education (CKK).
**REFERENCES**
1. J. A. del Alamo, V. Chang, L. Brooks, C. McLean, J. Hardison, G. Mishuris and L. Hui, MIT microelectronics WebLab, in *Lab on the Web: Running Real Electronics Experiments via the Internet*, eds. T. A. Fjeldly and M. S. Shur, John Wiley & Sons (2003).
2. A. Söderlund, F. Ingvarson, P. Lundgren and K. Jeppson, Remote laboratory: bringing students up close to semiconductor devices, in *Lab on the Web: Running Real Electronics Experiments via the Internet*, eds. T. A. Fjeldly and M. S. Shur, John Wiley & Sons (2003).
3. A. Buffler, S. Allie, F. Lubben and B. Cambell, The development of first year physics students’ ideas about measurement in terms of point and set paradigms, *Int. J. Sci. Educ.*, **23**, 2001, pp. 1137–1156.
4. R. Lippmann, Students’ understanding of measurement and uncertainty in the physics laboratory: social construction, underlying concepts, and quantitative analysis, unpublished Ph.D. thesis, University of Maryland, USA (2003).
5. I. Gustavsson, A remote access laboratory for electrical circuits experiments, *Int. J. Eng. Educ.*, **19**, 2003, pp. 409–419.
6. D. Gillet, F. Geoffroy, K. Zeramdini, A. V. Nguyen, Y. Rekik and Y. Piguet, The Cockpit: an effective metaphor for Web-based experimentation in engineering education, *Int. J. Eng. Educ.*, **19**, 2003, pp. 389–397.
7. E. W. Moyer, A. B. Carlson, K. A. Connor, W. C. Jennings and P. M. Schoch, Studio format for innovative pedagogy in circuits and electronics, *Proc. Frontiers in Education Conference*, **3**, 1997, pp. 1431–1434.
8. F. Marton, Phenomenography: describing conceptions of the world around us, *Instructional Science*, **10**, 1981, pp. 177.
9. F. Marton and S. Booth, *Learning and Awareness*, Lawrence Erlbaum Ass., Mahwah, New Jersey (1997).
10. Y. Lincoln and E. Guba, *Naturalistic Inquiry*, SAGE, London (1985).
**Per Lundgren** is an associate professor at the Department of Microtechnology and Nanoscience at Chalmers University of Technology. He received his Ph.D. in solid state electronics at Chalmers in 1996 and has been involved in undergraduate courses on semiconductor devices since 1991.
**Kjell O. Jeppson** received the Ph.D. degree in solid state electronics from Chalmers University of Technology, Göteborg, Sweden, in 1977. He became a senior lecturer at
the Department of Solid State Electronics in 1978. Since 1996, he is Professor at the Department of Microtechnology and Nanoscience (MC2). He has authored a textbook on semiconductor devices (in Swedish). From 1993 to 2000, he was vice Dean of Electrical Engineering, responsible for the undergraduate school of Electrical Engineering. He was the initiator of the E-96 project for implementation of the undergraduate EE curriculum reform.
Åke Ingerman has a junior research position in physics education at the Department of Microtechnology and Nanoscience at Chalmers University of Technology. He received his Ph.D. in physics with specialization in pedagogy at Chalmers in 2002. |
TWO DAYS OF HISTORY
The Human Rights Committee congratulates the people of Mozambique on their independence celebrations of yesterday, June 25. We celebrate with them in spirit and solidarity.
Their struggle was filled with courage and deeds of history. And their administration, since Frelimo took charge of the provisional transitional government in September last year has been a model of efficiency and enlightenment.
The historic 25th June brings into focus another historic day - June 26th, Freedom Day in South Africa.
HRC Bulletin will report on various historic events and anniversaries in its coming issues. To start the series, this issue examines June 26th.
COP
June 26 is the 20th anniversary of the Congress of the People, at which the Freedom Charter was adopted. Planned during 1954, the COP was called by the Congress alliance - the African National Congress, South African Indian Congress, South African Coloured Peoples' Organization and the S.A. Congress of Democrats.
MOZAMBIQUE
It started almost unannounced, the war to liberate Mozambique from Portuguese colonialism, in 1964, and it made little impression on the world's Press, for these were the early days of Vietnam.
By 1968, however, crash plans were announced by the Portuguese to "open up the northern areas", including...
For months before the meeting, invitations were widely distributed to various organisations and groups of workers. The people were asked to form local committees to discuss their grievances and their ideas of how the country should be governed. They were asked to draw up lists of suggestions and demands, which were to be sent to the central committee by a certain date, and also to elect delegates to represent them at the Congress of the People. When they had received the numerous resolutions, suggestions and demands, the central committee drafted a Freedom Charter.
The Congress of the People took place over June 25 and 26 1955, when approximately 3 000 delegates gathered at Kliptown, Johannesburg, to discuss the draft charter. There were doctors and domestic workers, labourers and clerks, shopkeepers and students, housewives and teachers - Africans, Whites, Indians, and Coloureds.
ALLEGATIONS
Many of the people came from as far afield as Durban and Cape Town, and there were allegations that quite a number of them had been stopped by police on the way, some of them being held on trivial charges and refused permission to continue.
At the meeting, the ANC colours dominated the scene, with bright banners proclaiming the identity of the branch or carrying slogans. Mounted behind the platform was the freedom wheel with its four spokes - the ANC's African elephant; the SAIC's Indian fox; SACPO's Coloured horse and the European owl of the COD. Also present were delegations from the S.A. Congress of Trade Unions (SACTU) and the S.A. Federation of Women, both of whom played an active part in the Congress of the People.
BRANCH
While the meeting, surprisingly enough, had not been banned by the Nationalist government, the Special Branch were very prominent observers - taking numerous photographs of delegates and recording every word. Most of the top leaders of the Congress Alliance were absent because of bannings, but Chief Luthuli sent a message and there were fraternal greetings from all parts of the world.
Throughout the Congress, the delegates listened to the speeches, the messages and a Freedom Charter read out in English, Sesotho and Xhosa. On the second day of the Congress, there was a raid by 200 armed policemen, Black and White. The chairman of the meeting was presented with a search warrant stating that they were "investigating a charge of treason"
and had come to search for "inflammatory or subversive" literature.
The police then proceeded to search all the delegates, took names and addresses and removed speakers' notes, goodwill messages and any other papers they found. In the meantime, the meeting went on, with delegates debating each clause of the Charter. When the final draft had been adopted, every one of the delegates present stood up and joined in the singing of "Afrika Mayibuye", and the singing continued until the police ended their raid.
Finally, as darkness fell, the Congress of the People came to an end, with the nostalgic notes of "Nkosi Sikelel' iAfrika" hanging in the air.
MOZAMBIQUE
Cabo Delgado and Niassa, using 25000 troops. There were no hard estimates of Frelimo's strength in the field. The name was still a new one to the world.
It achieved little, and the Governor-General, Mr Pimental dos Santos, outlined a new plan, which he said would produce a "miracle for all the world to see". He said, confidently, the Frelimo war effort in Niassa had collapsed, adding there had been "a reduction of terrorist activity in Cabo Delgado."
BURN'T ALIVE
As with all wars, massacre reports began circulating. Out of the sweaty bush in the north, came the report of two priests: Portuguese troops had jailed men, women and children in their huts and burnt them alive. The priests, who said they had assisted in the burials afterwards, were tried, convicted, and later smuggled out of the country as White settler anger boiled up against them.
Frelimo's strength grew, on the lines of the war of the flea, as the popular masses supported the men bearing arms.
By January 1974 Lisbon was forced to admit the situation was critical and four-star general Costa Gomes, Portugal's top soldier, was flown in to take control.
The war had by then taken shape: two prongs from the Tanzanian border to the north going straight down southwards, and two prongs from the Zambian border to the north-west, stretching southeast towards the coast.
This is the second newsletter of the Human Rights Committee. Please pass it on to your friends, because it is vital that it reaches the greatest number of people possible.
We would welcome letters and comment.
Write to Secretary,
Sheila Weinberg
P.O. Box 79051
SENDERWOOD - 2145.
WATCH OUT FOR FURTHER ISSUES AND TELL YOUR FRIENDS TO DO THE SAME!
Between Zambia and Tanzania lies the north-south corridor of southern Malawi, freely used for lines of communication between both sides.
Gomes had hardly got started; when Frelimo hit the west-east rail line between Umtali in Zimbabwe and Beira on the coast, which roughly divided the country in half.
Prevented by a strict censorship from knowing how the war was going, the Portuguese settler population found these attacks a shattering blow to morale.
So did the army. But by the time the Portuguese had rallied, Frelimo was pushing hard on the next line, the Save River, still further south. It was obvious that Frelimo's 9000 armed fighters had broken Portugal's army of 65000. But before the final battle lines had been drawn political events overtook the war.
MACHEL
The Portuguese revolution on April 25 last year produced a promise to end the war and negotiate a political settlement.
Days that made history followed.
On July 29, 1974, the civilians and army officers invited Frelimo troops into Tete to a welcoming feast.
On September 17 Samora Machel and Portuguese Foreign Minister Soares began talks in Dar es Salaam on the handover of power to the people of Mozambique. The war was all but over.
But the shooting was not, for in Lourenco Marques a group of white reactionaries planned and executed an abortive coup against the Frelimo provisional government.
This led to rioting with 103 deaths.
Later a Portuguese soldier pushed an African boy wearing a Frelimo shirt, sparking off fresh riots in which 49 died.
In spite of these incidents the Frelimo transitional government has administered Mozambique with an efficiency that won the admiration of the international community.
The only way this was possible was because of the sustained and enlightened work of the political wing of Frelimo, who went into areas in advance of the troops to prepare the way for the war effort.
This meant that when Frelimo took over the civilian administration it was known, accepted and supported by the popular masses of the country.
---
The Human Rights Committee needs financial support from the people and their friends who support its actions.
All amounts, however small, will be welcomed by the Committee.
Please send donations to HRC Secretary,
Sheila Weinberg
P.O. Box 79051
Senderwood, 2145.
THANK YOU!
The Freedom Charter
We, the People of South Africa, declare for all our country and the world to know:
that South Africa belongs to all who live in it, black and white, and that no government can justly claim authority unless it is based on the will of all the people;
that our people have been robbed of their birthright to land, liberty and peace by a form of government founded on injustice and inequality;
that our country will never be prosperous or free until all our people live in brotherhood, enjoying equal rights and opportunities; that only a democratic state, based on the will of all the people, can secure to all their birthright without distinction of colour, race, sex or belief;
And therefore, we, the people of South Africa, black and white together—equals, countrymen and brothers—adopt this Freedom Charter. And we pledge ourselves to strive together, sparing neither strength nor courage, until the democratic changes here set out have been won.
THE PEOPLE SHALL GOVERN!
Every man and woman shall have the right to vote for and to stand as a candidate for all bodies which make laws;
All people shall be entitled to take part in the administration of the country;
The rights of the people shall be the same, regardless of race, colour or sex;
All bodies of minority rule, advisory boards, councils and authorities shall be replaced by democratic organs of self-government.
ALL NATIONAL GROUPS SHALL HAVE EQUAL RIGHTS!
There shall be equal status in the bodies of state, in the courts and in the schools for all national groups and races;
All people shall have equal right to use their own languages, and to develop their own folk culture and customs;
All national groups shall be protected by law against insults to their race and national pride;
The preaching and practice of racial, race or colour discrimination and contempt shall be a punishable crime;
All apartheid laws and practices shall be set aside.
THE PEOPLE SHALL SHARE IN THE COUNTRY'S WEALTH!
The national wealth of our country, the business of all its people, shall be owned and controlled by the nation as a whole;
The land shall be nationalised in order to free the peasants from bondage and to end land hunger;
The mineral wealth of the country shall be used for the benefit of all the people;
The monopoly of foreign capital in industry, commerce and transport shall be abolished;
The profits of all industries and commerce shall be shared by the workers and the community at large.
THERE SHALL BE WORK AND SECURITY!
All who work shall be free to form trade unions, to elect their officers and to make wage agreements with their employers;
The state shall recognise the right and duty of all to work, and to draw full unemployment benefits;
Men and women of all races shall receive equal pay for equal work;
There shall be a forty-hour working week, a national minimum wage, paid annual leave, and sick leave for all workers, and maternity care on full pay for all working mothers;
Miners, domestic workers, farm workers and civil servants shall have the same rights as all others who work;
Child labour, compound labour, the lot system, and contract labour, shall be abolished.
THE DOORS OF LEARNING AND OF CULTURE SHALL BE OPENED!
The Government shall discover, develop and encourage national talent for the enhancement of our cultural life;
All the cultural treasures of mankind shall be open to all, by free exchange of books, ideas and contact with other lands;
The aim of education shall be to teach the youth to love their parents and their culture, to honour human brotherhood, liberty and justice.
THE PEOPLE SHALL SHARE IN THE COUNTRY'S WEALTH!
The national wealth of our country, the heritage of all South Africans, shall be restored to the people;
The mineral wealth beneath the soil, the Banks and monopoly industry shall be transferred to the ownership of the people as a whole;
All other industry and trade shall be controlled to assist the well-being of the people;
All people shall have equal rights to trade where they choose, to manufacture and to enter all trades, crafts and professions.
THE LAND SHALL BE SHARED AMONG THOSE WHO WORK IT!
Restrictions of land ownership on a racial basis shall be ended, and all the land redivided amongst those who work it, to banish famine and land hunger;
The state shall help the peasants with implements, seed, tractors and dams to save the soil and assist the tillers;
Freedom of movement shall be guaranteed to all who work on the land;
All shall have the right to occupy land wherever they choose;
People shall not be robbed of their cattle, and forced labour and farm prisons shall be abolished.
ALL SHALL BE EQUAL BEFORE THE LAW!
No one shall be imprisoned, deported or restricted without a fair trial;
No wrong shall be condemned by the order of any Government official;
The courts shall be representative of all the people;
Imprisonment shall be only for serious crimes against the people and human rights violations, not vagrancy;
The police force and army shall be open to all on an equal basis and shall be the protectors and protectors of the people;
All laws which discriminate on grounds of race, colour or belief shall be repealed.
ALL SHALL ENJOY EQUAL HUMAN RIGHTS!
The law shall guarantee to all their right to speak, to organise in political parties, to publish, to preach, to worship and to educate their children;
The privacy of the home from police raids shall be protected by law;
All shall be free to travel without restriction from country-side to town, from province to province, and from South Africa abroad;
Pass laws, permits and all other laws restricting these freedoms shall be abolished.
Let all who love their people and their country now say, as we say here: "THESE FREEDOMS WE WILL FIGHT FOR, SIDE BY SIDE, THROUGHOUT OUR LIVES, UNTIL WE HAVE WON OUR LIBERTY."
THERE SHALL BE HOUSES, SECURITY AND COMFORT!
All people shall have the right to live where they choose, to be decently housed, and to bring up their families in comfort and security;
Unused housing space to be made available to the people;
Rent and prices shall be lowered, food plentiful and no one shall go hungry;
A preventive health scheme shall be run by the state;
Free medical care and hospitalisation shall be provided for all, with special care for mothers and young children;
Shums shall be demolished, and new suburbs built where all have transport, roads, lighting, playing fields, creches and social centres;
The aged, the orphans, the disabled and the sick shall be cared for by the state;
Rest, leisure and recreation shall be the right of all;
Fenced locations and ghettos shall be abolished, and laws which break up families shall be repealed.
THERE SHALL BE PEACE AND FRIENDSHIP!
South Africa shall be a fully independent state, which respects the rights and sovereignty of all nations;
South Africa shall strive to maintain world peace and the settlement of all international disputes by negotiation—not war;
Peace and friendship amongst all our people shall be secured by upholding the equal rights, opportunities and status of all;
The people of the protectorates—Basutoland, Bechuanaland and Swaziland shall be free to decide for themselves their own futures;
The right of all the peoples of Africa to independence and self-government shall be recognised, and shall be the basis of their co-operation.
AFTER THE COP . . .
THE TREASON TRIAL
Soon after the Congress of the People there were hundreds of raids up and down the country on organisations and individuals involved in the event.
These police raids led to the arrest and detention of 156 South Africans of all races, who were charged with high treason.
The marathon treason trial was to last five years, and ended with the acquittal of those who appeared in court.
Among the prominent South Africans charged were Nelson Mandela, Walter Sisulu, Chief Albert Luthuli, Mrs Helen Joseph, Goven Mbeki, Joe Slovo and many others.
The Defence team included the late Bram Fischer, Mr I. Maisels Vernon Berange and others.
Those charged were:
J. Adams, M. Asmal, Y. Barenblatt, H. Barsel, L. Bernstein, P. Beylveild, J. Botes, S. Chauvali, S. Esakjer, B. Hlapane.
A. Hutchinson,
I. Hodgson, Helen Joseph, Paul Joseph, F. Keitsing, Moses Kotane, Jerry Kupenga, Katherine Levy, Levy Levy, Norman Levy, S. Lollan, F. Madiba.
A. Mahlungu, V. Make, P. Molale, Temyosou Mokgatle, E. Molefe, E. G. Mokgogothi, E. Malete, S. Malupe, Nelson Mandela, S. Masemola, M. Massina, July Masube, B. Maqubela, P. Mathonsi, J. Matlou, J. Mavuso.
T. Musi, J. Modise, F. Molave, E. Molefi, M. Moodli, Dr. H. M. Mososa, E. F. Moretsefo, O. Motzabi.
M. K. Mpho, S. Nathiie, P. Nene, L. Ngoyi, B. Ngwadu, N. Nkadimeng, D. Nkwen, F. Ntuli, A. E. Patel, J. Poo, R. Press.
James Hadebe, M. Ranta, R. Resha, B. Seitshiro, N. Selake, P. Selepe, Shali, S. Siepe, S. Sibopha, Sibande, Walter Sifadi, G. T. Sibande, Ruth Slovo, Joe Slovo.
Oliver Tambo,
S. Tyiki, R. Tsuno, M. Tshabalala, Rev. D. C. Thompson, Sunita Bunting, J. Busse, F. Carmeson, A. Deacon, L. de Villiers, L. O. Morvitch, A. la Gumia, C. Makohlisa, D. Mgunguyeka.
J. Morolong, L. Morrison, J. Mgoza, J. Mtini, G. Ngoyana, G. Peake, A. Sibeko, R. Sepember.
A. Silonga, B. Turuk, L. n. Lee-Warden M.P., P. Baard, Fuyani, Rev. S. W. Gawe, J. Jacobs, L. Jasson, L. Keps, P. Mashiluni.
J. Matthews,
Prof. Z. K. Matthews, W. Mati, J. Mhlongo, M. C. Mayekiso, W. Mini, E. Mlalazi, P. Mkalipi, W. Z. Mkwayi, B. Ndumbi, J. Kampeni, B. Ntsangama.
B. Nogaya, T. Tshume, T. Tshunungwa, & Vanqa, S. Arencstein, Dr. C. Concor, S. Dhlamini, G. Duarte, J. Hoogendyk, G. Hurban, Chief A. J. Lutuli, P. S. Manuna, D. T. Meer, P. G. Mei, Bertha Mhize.
K. Moonsamy,
Dr. M. Metala, Dr. G. M. Naicker, M. P. Naicker, N. T. Naicker, M. Nair, A. Ngeobo, D. Nyembe.
R. Shanley,
Dorothy Shank, P. Simelane, M. B. Yengwa, G. Diechabe, Dr A. E. Letele.
J. B. Mrora, Martha J. Motlhakona, L. S. Monnanyana, A. Scochooreng.
I. Nkosi, S. M. Kumalo, Rev. A. J. Calato, Debi Singh, Stella Damous, T. Mgota, D. Scedat.
POLICE PHOTOGRAPHER AT THE C.O.P.
FREEDOM DAY MEANS....
June 26 has always had a special significance in the history of the liberation movement in South Africa. From sources available to the HRC, it appears that during the ANC conference in Bloemfontein in 1948, a decision was taken to set aside a sort of "freedom day" or "Heroes' day", but no specific date was agreed upon at that time.
However, June 26 was adopted as a day of mourning in 1950 after 13 Blacks had been killed during May Day rioting when anti-government demonstrations were held. From then on, it was universally accepted as a very important day for patriots engaged in the struggle for liberation from South Africa's regime of oppression and racism. It was also, in that year, a day of protest against the passing of the Suppression of Communism Act.
In 1952, June 26 was take-off day for the Defiance Campaign jointly organised by the banned ANC and the S.A. Indian Congress. Three years later, in 1955, June 26 was marked by the unanimous adoption of the Freedom Charter, at the Congress of the People.
MORENA BOLOKA
Morena, boloka sechaba sa heso,
O fedisa lintoa le matsoenyeho. (REPEAT 2 LINES)
O se boloka, o, se boloka,
O se boloka, o, se boloka.
Sechaba sa heso, sechaba sa heso.
TRANSLATION:
God save our nation,
Bring an end to war and suffering.
Prose poem, "Christian Kindness" -- dedicated to Maria Kgosieng and the people who suffered forced removals to Vaalboshoek.
Dry your tears, Maria -
do not cry.
Even though they have dumped you, like unwanted trash,
in the bush - nowhere.
Wipe your eyes, Maria -
do not weep.
Even though they have taken you from your house,
the home your husband built.
And given you, with Christian kindness,
a square tin box.
Dry your eyes, Maria -
do not cry.
Even though you mourn your Majeng garden;
your cabbages, your grapes, your potatoes
and your flowers.
Wipe your eyes, Maria -
do not weep.
Even though this ground at Vaalboshoek
is stoney as the hearts
of the Christians who pushed you here
and nothing grows but weeds and thorns.
Dry your tears, Maria -
do not cry.
Even though you have no windmill here.
Join the bucket parade -
the tap is only two kilometres away.
Wipe your eyes, Maria -
do not cry.
Even though you may not make fire
inside your little tin box.
The babies are hungry - bring out your
three-legged pot
and build your open-air fire.
Dry your eyes, Maria -
do not cry.
The world and God are watching;
and those who brought you here, against your will,
despite your pleas to remain in the home
your husband built -
they will be punished.....
for their magnificent Christian kindness.
Juby Mayet: May 1975
HRC AIMS
The Human Rights Committee was formed in September last year by a group of people who want freedom from oppression in South Africa. The Committee stands for an internationally recognised set of values which do not, generally, exist in South Africa - principles which are embodied in the United Nations' Declaration of Human Rights, from which we draw our name.
Our aim is to protest the lack of freedoms in the society; to protest the violation of human rights by the Government which is not even prepared to sign the UN Declaration; to promote consciousness among the people of S.A. and overseas of the injustices of the present regime; and to join the increasingly popular movement to bring about full rights for all the people of this country.
* We stand for majority rule and the radical changes this will involve in all sections of South African society.
* We stand for a society free of discrimination on any grounds - colour, class, belief or sex.
* We stand for the unconditional release of political prisoners.
* We stand for the abolition of detention without trial, the freeing of detainees, the lifting of house-arrest and all other restrictions.
* We believe that South Africa's exiles, if permitted to return, could make a full contribution to the construction of a free society.
Support from the people is most encouraging. It is vital if our role is to have real meaning in the struggle. We are all oppressed in South Africa, and we must all struggle together for freedom and equality!
NKOSI SIKELEL' iAFRIKA
Nkosi sikalel' iAfrika,
Mali phakamisa pondo lwayo
Yiwa imitanda zayetu
Nkosi sikalela, nkosi sikalela.
REPEAT VERSE
Woza moya, woza woza,
Woza moya, oyingwelo.
Nkosi sikalela, tina isapolwayo.
TRANSLATION:
God save Africa
Let its horn be lifted
Do hear our prayers.
Come spirit
Come thy holy spirit
God do save us, its inhabitants.
Printed by WUSPO, SRC, University of the Witwatersrand, 1 Jan Smuts Ave., Jhb |
DISCUSSION PAPER SERIES
No. 4528
DEFLATIONARY BUBBLES
Willem H Buiter and Anne Sibert
INTERNATIONAL MACROECONOMICS
Centre for Economic Policy Research
www.cepr.org
Available online at: www.cepr.org/pubs/dps/DP4528.asp
DEFLATIONARY BUBBLES
Willem H Buiter, EBRD and CEPR
Anne Sibert, Birkbeck College, London and CEPR
Discussion Paper No. 4528
August 2004
Centre for Economic Policy Research
90–98 Goswell Rd, London EC1V 7RR, UK
Tel: (44 20) 7878 2900, Fax: (44 20) 7878 2999
Email: firstname.lastname@example.org, Website: www.cepr.org
This Discussion Paper is issued under the auspices of the Centre’s research programme in INTERNATIONAL MACROECONOMICS. Any opinions expressed here are those of the author(s) and not those of the Centre for Economic Policy Research. Research disseminated by CEPR may include views on policy, but the Centre itself takes no institutional policy positions.
The Centre for Economic Policy Research was established in 1983 as a private educational charity, to promote independent analysis and public discussion of open economies and the relations among them. It is pluralist and non-partisan, bringing economic research to bear on the analysis of medium- and long-run policy questions. Institutional (core) finance for the Centre has been provided through major grants from the Economic and Social Research Council, under which an ESRC Resource Centre operates within CEPR; the Esmée Fairbairn Charitable Trust; and the Bank of England. These organizations do not give prior review to the Centre’s publications, nor do they necessarily endorse the views expressed therein.
These Discussion Papers often represent preliminary or incomplete work, circulated to encourage discussion and comment. Citation and use of such a paper should take account of its provisional character.
Copyright: Willem H Buiter and Anne Sibert
ABSTRACT
Deflationary Bubbles*
We analyse deflationary bubbles in a model where money is the only financial asset. We show that such bubbles are consistent with the household’s transversality condition if and only if the nominal money stock is falling. Our results are in sharp contrast to those in several prominent contributions to the literature, where deflationary bubbles are ruled out by appealing to a non-standard transversality condition, originally due to Brock ([4], [5]). This condition, which we dub the GABOR condition, states that the consumer must be indifferent between reducing his money holdings by one unit and leaving them unchanged and enjoying the discounted present value of the marginal utility of that unit of money forever. We show that the GABOR condition is not part of the necessary and sufficient conditions for household optimality nor is it sufficient to rule out deflationary bubbles. Moreover, it rules out Friedman’s optimal quantity of money equilibrium and, when the nominal money stock is falling, it rules out deflationary bubbles that are consistent with household optimality. We also consider economies with real and nominal government debt and small open economies where private agents can lend to and borrow from abroad. In these cases, deflationary bubbles may be possible, even when the nominal money stock is rising. Their existence is shown to depend on the rules governing the issuance of government debt.
JEL Classification: D90, E31 and E63
Keywords: fiscal rules, government solvency, optimal quantity of money and transversality condition
Willem H Buiter
EBRD
One Exchange Square
LONDON
EC2A 2JN
Tel: (44 20) 7338 6805
Fax: (44 20) 7338 6111
Email: email@example.com
Anne Sibert
Department of Economics
Birkbeck College
Malet Street
London
WC1E 7HX
Tel: (44 20) 7631 6420
Fax: (44 20) 7631 6416
Email: firstname.lastname@example.org
For further Discussion Papers by this author see:
www.cepr.org/pubs/new-dps/dplist.asp?authorid=100351
For further Discussion Papers by this author see:
www.cepr.org/pubs/new-dps/dplist.asp?authorid=126907
*We would like to thank William Brock and Narayana Kocherlakota for useful comments on an earlier version of this Paper. The views and opinions expressed are those of the authors. They do not represent the views and opinions of the European Bank for Reconstruction and Development.
Submitted 03 July 2004
1 Introduction
This paper revisits the existence of deflationary bubbles and the conditions that rule them out. We focus on two standard dynamic optimizing models with unbacked government-issued money. In the first model, money is the only store of value. In this framework we obtain new results, demonstrating that the existence of deflationary bubbles depends on whether the money stock is increasing or decreasing. In the second model, money co-exists with non-monetary financial instruments, such as government bonds, that can be in negative net supply. We show that the rules governing the issuance of these instruments determine whether or not deflationary bubbles exist.
The literature we are extending goes back to two seminal papers by Brock [4], [5]. Brock analyzes a closed-economy model with money, but no government bonds. In addition to the standard transversality condition (henceforth, the $ST$ condition) associated with the household’s optimization problem, that the present value of the terminal stock of real money balances is zero, Brock introduces an additional restriction. This restriction (henceforth the $GABOR$ condition) was later adopted by Obstfeld and Rogoff [27], [28] [29], Gray [14] and Azariadis [1]. It states that it must not be possible to increase consumer welfare by increasing (reducing) current consumption by a small amount through a reduction (increase) in the stock of current money balances and holding this lower (higher) stock of money balances forever after.
In this paper we establish the following. First, unlike the $ST$ condition, the $GABOR$ condition is not part of the necessary and sufficient conditions for an optimal consumption-money demand programme.
Second, in a model where money is the only financial instrument, we demonstrate that deflationary bubbles are consistent with the $ST$ condition if and only if money growth is strictly negative. This contrasts with results that have been obtained with the $GABOR$ condition. Even when the money stock is rising, the $GABOR$ condition alone is not sufficient to rule out deflationary bubbles. We show that when the nominal money stock
is falling, the *GABOR* condition rules out deflationary bubbles that are consistent with household optimality and the other equilibrium conditions of the model.
Third, we show that the *GABOR* condition rules out Bailey’s [2] and Friedman’s [13] stationary optimal quantity of money equilibrium, where the nominal interest rate is zero, the (negative) inflation rate equals the rate of decline of the stock of nominal money balances and there is satiation in real money balances. This is true even when satiation is achieved at a finite level of real money balances.
Fourth, we show that deflationary bubbles can exist in the closed economy model with both money and bonds, even when the nominal money stock is rising. An example of a fiscal rule that supports deflationary bubbles was suggested by Woodford ([38], pp. 131-135), who discusses and provides examples of fiscal rules that rule out what he calls “deflationary traps”.
Fifth, if the government satisfies its solvency constraint with equality, we show that deflationary bubbles do not exist unless the money stock is falling.
Sixth, in a small open economy, we show that deflationary bubbles can exist even when the money supply is growing and even when the government views money as irredeemable and satisfies its solvency constraint with equality.
## 2 Deflationary Bubbles when Money is the Only Financial Instrument
### 2.1 The households
The economy is inhabited by a representative household and its government. Each period, the household receives an exogenous endowment of the single perishable consumption good and pays a lump-sum tax. It consumes the good and saves in the form of non-interest-bearing unbacked money issued by the state. The household receives liquidity services from its money holdings and has preferences defined over paths of consumption and holdings of real balances represented by
\[ U = \lim_{T \to \infty} \sum_{t=0}^{T} \beta^t u(c_t, M_t^d/P_t), \ 0 < \beta < 1, \tag{1} \]
where \( c_t \geq 0 \) is time-\( t \) consumption, \( M_t^d \geq 0 \) is the household’s end-of-period-\( t \) demand for nominal money balances to be carried over into period \( t+1 \), \( P_t \) is the period-\( t \) money price of the good and \( u \) is the extended real-valued utility function on \( \mathbb{R}_+^2 \). We assume that \( u \) is strictly increasing in its first argument, concave and continuously differentiable on \( \mathbb{R}_{++}^2 \), with \( u_c(c,m) \to \infty \) as \( c \searrow 0 \) and \( u_m(c,m) \to \infty \) as \( m \searrow 0 \). We only consider equilibria where the infinite sum in equation (1) converges.
We temporarily assume that money is valued and thus, \( 1/P_t > 0 \) for every \( t \geq 0 \). We then only consider outcomes where this is true. There is, however, always a non-monetary equilibrium where \( 1/P_t = 0 \) for every \( t \geq 0 \). In this outcome, money is not held and the household consumes its after-tax endowment each period.
**Assumption 1.** For every \( c \in \mathbb{R}_{++} \), there exists \([u(c), \bar{u}(c)] \subset \mathbb{R}_{++} \) such that \( u(c) \leq u_c(c,m) \leq \bar{u}(c) \) for every \( m \in \mathbb{R}_+ \)
We make one of the following mutually exclusive assumptions:
**Assumption 2a (Satiation in real balances)** For every \( c \in \mathbb{R}_{++} \) there exists \( \hat{m}(c) \in \mathbb{R}_{++} \) such that \( u_m(c,m) > (=) 0 \) if \( m < (\geq) \hat{m}(c) \) and \( u \) is strictly concave and twice differentiable on \( \{(c,m): 0 < m < \hat{m}(c), c \in \mathbb{R}_{++}\} \).
**Assumption 2b (Bounded utility in real balances; no satiation for finite real balances)** The function \( u \) is strictly concave and twice differentiable on \( \mathbb{R}_{++}^2 \), \( u_m(c,m) > 0 \) and \( u(c,m) \) is bounded from above in \( m \).
**Assumption 2c (Unbounded utility).** The function \( u \) is strictly concave and twice differentiable on \( \mathbb{R}_{++}^2 \), \( u_m(c,m) > 0 \) and \( u_m(c,m) \to 0 \) as \( m \to \infty \).
The household’s within-period budget constraint is
\[ M_t^d/P_t \leq y - \tau_t - c_t + M_{t-1}^d/P_t, \ t \geq 0, \tag{2} \]
where \( y > 0 \) is the constant per-period endowment and \( \tau_t \) is the period-\( t \) real lump-sum tax. In this section, we only consider outcomes where
\[ \tau_t \leq y + M_{t-1}^d / P_t, \ t \geq 0 \]
(3)
The household maximises utility (equation (1)) subject to (2), taking as given initial money holdings \( M_{-1}^d > 0 \).
Sufficient conditions for optimality are given by the household period budget constraints (2) (with equality), the Euler equation
\[ u_c(c_t, M_t^d / P_t) = u_m(c_t, M_t^d / P_t) + (\beta P_t / P_{t+1}) u_c(c_{t+1}, M_{t+1}^d / P_{t+1}), \ t \geq 0, \]
(4)
and the standard transversality (\( ST \)) condition\(^1\)
\[ \lim_{t \to \infty} \beta^t u_c(c_t, M_t^d / P_t) M_t^d / P_t = 0. \]
(5)
The Euler equation and the budget constraint are also necessary for optimality. (See Lucas and Stokey [25], p. 97.) The necessity of the transversality condition has been a more difficult issue. Weitzman [34] shows that it is necessary when within-period utility is bounded. Kamihigashi [19] generalises this result to the case where, at an optimum, the sequence of discounted within-period utilities is summable.
Equation (4) is typical of the Euler equations that characterise investment in a consumer durable and has the following interpretation. The household is indifferent between a marginal increase in period-\( t \) consumption, yielding utility of \( u_c(c_t, M_t^d / P_t) \), and foregoing this consumption and acquiring money, receiving utility of \( u_m(c_t, M_t^d / P_t) \) from its liquidity services and using it to purchase consumption next period, with associated discounted utility of \( (\beta P_t / P_{t+1}) u_c(c_{t+1}, M_{t+1}^d / P_{t+1}) \).
Equation (5) implies that either the optimal value of the state variable, \( M_t^d / P_t \), goes to zero as time goes to infinity or that its marginal contribution to the maximized value of the objective function, \( \beta^t u_c(c_t, M_t^d / P_t) \), goes to zero.
\(^1\)See Lucas and Stokey [25], p. 98.
Solving (4) forward yields
\[
u_c(c_t, M_t/P_t) = \sum_{s=0}^{\infty} \frac{\beta^s P_t u_m(c_{t+s}, M_{t+s}^d/P_{t+s})}{P_{t+s}} + \lim_{T \to \infty} \frac{\beta^T P_t u_c(c_{t+T}, M_{t+T}^d/P_{t+T})}{P_{t+T}}. \tag{6}
\]
### 2.2 Brock’s restriction on optimal programmes
Brock [4], [5] proposes a further restriction on permissible optimal programmes:
\[
u_c(c_t, M_t^d/P_t) = \sum_{s=0}^{\infty} (\beta^s P_t/P_{t+s}) u_m(c_{t+s}, M_{t+s}^d/P_{t+s}). \tag{7}
\]
By equation (6) this is equivalent to
\[
\lim_{T \to \infty} \left( \beta^T P_t/P_{t+T} \right) u_c(c_{t+T}, M_{t+T}^d/P_{t+T}) = 0. \tag{8}
\]
His justification is that the consumer must be indifferent between reducing his money holdings by one unit and enjoying an increase in marginal utility of consumption (the left-hand side of equation (7)) and leaving his money holdings unchanged and enjoying the discounted present value of the marginal utility of that unit of money forever (the right-hand side of equation (7)).\(^2\) Obstfeld and Rogoff [27], [28] and [29] make the same argument, as do Gray [14] and Azariadis [1].\(^3\)
The Gray-Azariadis-Brock-Obsfeld-Rogoff (*GABOR*) condition given in (8) (or, equivalently, (7)) cannot in general be a requirement for optimality. The reason is that if a consumer converts a dollar into consumption (or vice versa) in period \(t\) and never un-
---
\(^2\)Brock assumes \(u(c, m) = u(c) + v(m)\). There is no public spending; hence, in equilibrium \(c = y\). Denoting time-\(t\) real balances by \(m_t\), he says, "At some point in time, \(T\), the act of taking one dollar out of cash balances will yield him \(u'(y)/P_T\) utils at the margin. His cash balances are depleted by one dollar for all \(s \geq T\). This loss of money services generates a utility loss \(\sum_{t=T}^{\infty} \beta^{t-T} v'(m_t)/P_t...\)" (Brock [5], p. 140). The same argument is also made in Brock [4], p. 762.
\(^3\)Gray's intuitive motivation of the *GABOR* condition (in [14], Section 4.1, and especially the argument starting on p. 107 leading up to equation (29) on p. 110), is essentially the same as Brock's. The *GABOR* condition can be written as \(\lim_{T \to \infty} \beta^T u_c(c_{t+T}, M_{t+T}^d/P_{t+T})/P_{t+T} = 0\). The *ST* condition is \(\lim_{T \to \infty} \beta^T u_c(c_{t+T}, M_{t+T}^d/P_{t+T})M_{t+T}/P_{t+T} = 0\). Gray notes correctly ([14], footnote (21)), that when the nominal money stock becomes a positive constant after some date, the *ST* condition and the *GABOR* condition are equivalent.
does that shift, he loses (gains) not just the reduction (increase) in money holdings in period $t$ and forever after. He also loses (gains) the terminal consumption that is associated with a permanent reduction in money holdings. Without this eventual reduction in consumption, the proposed perturbation of the optimal consumption and money demand programme is in general not feasible because it violates the sequence of household within-period budget constraints.
To see this, suppose that a consumer lowers his holdings of real balances by one unit and increases his consumption in period $t$ and then lowers his consumption $T \geq 1$ periods later to restore his original money holdings. This leads to the period-$t$ utility gain $u_c(c_t, M^d_t/P_t)$. The loss of liquidity services before he undoes the shift equals $\sum_{s=0}^{T-1} (\beta^s P_t/P_{t+s}) u_m(c_{t+s}, M^d_{t+s}/P_{t+s})$. In period $t+T$ the reduction in real balances is reversed. There is no utility loss from lower money balances in period $t+T$ or later, but the restoration of money balances has been affected through a reduction in period $t+T$ consumption by an amount $P_t/P_{t+T}$. The associated discounted utility loss is $(\beta^T P_t/P_{t+T}) u_c(c_{t+T}, M^d_{t+T}/P_{t+T})$. When a household reduces its real balances in period $t$ by one unit and never reverses this, it is still required to reduce its terminal consumption by an amount $\lim_{T \to \infty} P_t/P_{t+T}$. The associated discounted utility loss is $\lim_{T \to \infty} (\beta^T P_t/P_{t+T}) u_c(c_{t+T}, M^d_{t+T}/P_{t+T})$.
Ignoring the last term in (6) or requiring it to equal zero may thus violate the household’s sequence of within-period budget constraints.\footnote{Whether it does or not depends on the relationship between inflation and the discount factor. If $\lim_{t \to \infty} P_{t+s}/P_{t+s-1} > \beta$, then the last term in (6) is zero if $c_t \to 0$.} Reducing consumption in period $t$ and increasing money holdings in period $t$ and never reversing the shift is feasible, but not rational. The correct characterisation of the perturbations of the infinite horizon optimal programme that should be utility neutral is therefore (5). Household optimisation does not, in general require that (8) hold. Because of the prominence of the GABOR condition in the literature, we state the following result.
**Proposition 1** Given the Euler equation (4), the sequence of within-period budget constraints (2) (with equality), and the initial value of the nominal money stock, the GABOR condition (8) is not part of the necessary and sufficient conditions for an optimum of the household’s consumption-money demand programme.
This is obvious from Kamihigashi’s [19] result that the household within-period budget constraints, (2), the Euler equations (4) and the transversality condition (5) are the set of necessary and sufficient conditions for an optimum. Condition (6) is an implication of the Euler equation (4). Condition (8) is an additional restriction on (5), and therefore on the necessary and sufficient conditions for an optimum.
2.3 The government
The government is the consolidated fiscal and monetary authorities. Its within-period budget constraint is
\[ M_t / P_t \geq g - \tau_t + M_{t-1} / P_t, \quad t \geq 0, \tag{9} \]
where \( g \geq 0 \) is the constant per-period real public spending and \( M_t \) is the money supply. Assuming \( g < y_t \) ensures that the assumed restriction (3) can be satisfied. We assume that (9) holds with equality.
We assume that the authorities adopt a constant proportional growth rate for the money stock so that
\[ M_{t+1} / M_t = \mu \geq 0. \tag{10} \]
The sequence of real lump-sum taxes is endogenously determined to make the exogenous public spending programme and the constant proportional growth rate of the nominal money stock consistent with the sequence of within-period government budget constraints.
2.4 Equilibrium
In equilibrium, $M_t^d = M_t$, $t \geq 0$ and
$$c_t = c = y - g, \ t \geq 0 \tag{11}$$
Let $m_t \equiv M_t / P_t$. Substitute (11) into (4) and (5). We define an equilibrium as follows:
**Definition 2** An equilibrium is a strictly positive sequence $\{m_t\}_{t=0}^{\infty}$ such that
$$\beta u_c(c, m_{t+1})m_{t+1} = \mu[u_c(c, m_t) - u_m(c, m_t)]m_t, \ t \geq 0 \tag{12}$$
$$\lim_{t \to \infty} \beta^t u_c(c, m_t)m_t = 0. \tag{13}$$
By equation (9), the taxes associated with the equilibrium satisfy
$$\mu \tau_t = (1 - \mu) m_t + \mu g, \ t \geq 0. \tag{14}$$
The GABOR condition (8) can be rewritten as
$$\lim_{t \to \infty} (\beta/\mu)^t u_c(c, m_t)m_t = 0. \tag{15}$$
The ST condition (13) implies the GABOR condition (15) when $\mu \geq 1$ and the GABOR condition implies the ST condition when $\mu \leq 1$. The two conditions are equivalent when the nominal money stock is constant ($\mu = 1$).
There are two potential types of monetary equilibria. First, given our constant fundamentals $(y, g, \mu)$, there is a fundamental equilibrium where $m_t = \bar{m} > 0$ for every $t \geq 0$. Constant real balances clearly satisfy (13). By (12) such an equilibrium has
$$L(\bar{m}) \equiv \mu u_m(c, \bar{m}) = (\mu - \beta)u_c(c, \bar{m}) \equiv R(\bar{m}). \tag{16}$$
We make the following assumption, which given our previous assumptions, is necessary and sufficient for a fundamental monetary equilibrium to exist:
\[
\mu \begin{cases}
\geq \\
>
\end{cases} \beta \text{ if } \begin{cases}
\text{Assumption 2a holds} \\
\text{Assumption 2b or 2c holds}
\end{cases}.
\] (17)
If \( \mu < \beta \), then \( L(m) \geq 0 > R(m) \) for every \( m > 0 \) and no fundamental monetary equilibrium exists. If \( \mu = \beta \) and Assumption 2b or 2c hold, then \( L(m) > 0 = R(m) \) for every \( m > 0 \) and no fundamental monetary equilibrium exists. If \( \mu = \beta \) and Assumption 2a holds then any \( m \geq \hat{m} \) satisfies equation (16).\(^5\) Such an outcome is a Friedman Optimal Quantity of Money (\( OQM \)) equilibrium, where the nominal stock of money declines proportionally at the rate of time preference and the household is satiated at a finite stock of real balances.
When \( \mu > \beta \) a simple fixed-point argument can be made to establish the existence of a fundamental monetary equilibrium.\(^6\) For this case, the additional restriction that real balances are a normal good at any fixed point (that is, \( u_c u_{nm} - u_m u_{cm} < 0 \)) ensures that the fundamental monetary equilibrium is unique.
In addition to fundamental monetary equilibria, there can be a gamut of non-fundamental (or non-stationary) equilibria. (See Azariadis [1]). A monetary equilibrium can be stable, with monotonic or cyclical convergence; it can be unstable, with either monotonic or cyclical divergence; there can be limit cycles and there can be chaotic behaviour. We are interested in monetary equilibria where nominal real balances go to infinity; such equilibria are called deflationary bubbles.
**Definition 3** A deflationary bubble is an equilibrium where \( m_t \to \infty \).
When the nominal money stock is constant, a deflationary bubble has the price level
---
\(^5\) As \( c \) equals \( y - g \), which is constant, we suppress the notational dependence of \( \hat{m} \) and \( \underline{u} \) and \( \bar{u} \) (defined in Assumption 1) on \( c \).
\(^6\) By Assumption 1, \( R(m) \in \rho \equiv [(\mu - \beta) \underline{u}, (\mu - \beta) \bar{u}] \). By the continuity of \( u_c \) in \( m \) and Assumption 2a, 2b or 2c, \( L^{-1} \) exists on \( \rho \) and \( L^{-1}(R(m)) \) is a continuous mapping from the compact convex set \([L^{-1}((\mu - \beta) \underline{u}), L^{-1}((\mu - \beta) \bar{u})]\) into itself and, thus, by Brouwer’s theorem, there exists a fixed point \( \hat{m} \) that satisfies equation(16).
going to zero - the standard definition of sustained deflation. With positive growth in the nominal money stock, a deflationary bubble can occur even with a rising price level. Along such a path however, inflation will be less than inflation in the associated fundamental equilibrium.
**Proposition 4** Consider the class of preferences satisfying both Assumption 1 and one of Assumptions 2a, 2b or 2c. If $\mu > 1$, then the ST condition is sufficient to rule out deflationary bubbles; the GABOR condition is not. If $1 \geq \mu > \beta$, then neither the ST condition nor the GABOR condition are sufficient to rule out deflationary bubbles for all preferences. If $1 > \mu$, the GABOR condition rules out deflationary bubbles not ruled out by the ST condition.
**Proof.** Suppose that $\mu > 1$. By (12), $u_c(c, m_{t+1})m_{t+1}/(u_c(c, m_t)m_t) = (\mu/\beta)[1 - u_m(c, m_t)/u_c(c, m_t)] \leq \mu/\beta$. If the equilibrium is a bubble, then $m_t \to \infty$ and $u_m(c, m_t) \to 0$. Thus, $\forall \epsilon > 0, \exists t^* > 0$ and finite such that $u_c(c, m_{t+1})m_{t+1}/u_c(c, m_t)m_t > \mu/\beta - \epsilon \ \forall t \geq t^*$. Let $\epsilon = (\mu - 1)/\beta$. Then we have $u_c(c, m_{t^*+T})m_{t^*+T} > (1/\beta)^T u_c(c, m_{t^*})m_{t^*}, \forall T \geq 1$. This implies $\lim_{T \to \infty} \beta^T u_c(c, m_{t^*+T})m_{t^*+T} > u_c(c, m_{t^*})m_{t^*} > 0$ which violates the ST condition (13).
If $u_m(c, m) = 1/\ln(m)$ for large values of $m$ then a deflationary bubble equilibrium exists and the GABOR condition is satisfied for every $\mu > \beta$ (See Obstfeld and Rogoff [28]) and the ST condition is satisfied for $\mu = 1$.
Suppose $\mu < 1$. Then examples of deflationary bubbles that satisfy the ST condition, but not the GABOR condition are easy to find. See the text following. ■
The intuition for why the ST condition rules out deflationary bubbles when the nominal money stock is not falling is as follows. In equilibrium, in each period the household must be indifferent between spending a unit of money on the consumption good and holding the unit of money, enjoying its liquidity services, and spending it on the consumption good the following period. Thus, because money provides liquidity services, the discounted shadow value of money used to purchase the consumption good must be falling over time.\footnote{The time-$t$ discounted shadow value to the household of a unit of money used for purchasing the good is the incremental contribution to $U$ in equation (1), $(\beta^t/P_t)u_c(c, M_t/P_t)$.} Along a bubble path, real balances go to infinity and the liquidity services of
a unit of money go to zero. Thus, the rate of decrease over time in the discounted shadow value of money used to purchase consumption goes to zero. If the nominal money stock is constant or growing at a constant rate over time, this implies that the value of the entire money stock when priced at the discounted shadow value of money used to purchase the consumption good, becomes constant or rises over time. This is inconsistent with the $ST$ condition, which says that the present discounted shadow value of the economy’s money stock must go to zero in the long run. If the nominal money stock is falling at a constant rate over time, then the discounted shadow value of the entire money stock when used to purchase the consumption good must be falling as well.\footnote{This argument suggests that if $\mu < 1$, then any sequence of real balances that goes to infinity and satisfies (12) must satisfy the $ST$ condition and this is true. In this case, by (12), $u_c(c,m_{t+1})m_{t+1}/(u_c(c,m_t)m_t) < \mu/\beta$. This implies $\beta^T u_c(c,m_{t+T})m_{t+T} < \mu^T u_c(c,m_t)m_t \to 0$.}
When $\mu > 1$, the $GABOR$ condition is a weaker condition than the $ST$ condition and it is not sufficient to rule out deflationary bubbles. Obstfeld and Rogoff [28] show that the $GABOR$ condition is sufficient to rule out deflationary bubbles if Assumptions 2a or 2b are satisfied. Brock [4] provides a weaker condition.
**Proposition 5 (Brock’s Result)** The $GABOR$ condition is sufficient to rule out deflationary bubbles if there exists $\lambda < 0$ and finite $m_0 > 0$ such that $m > m_0$ implies $v'(m) \leq m^\lambda$.
When $\beta < \mu < 1$, imposing the $GABOR$ condition rules out deflationary bubbles that satisfy the necessary and sufficient conditions of the consumer’s problem. For example, suppose that $u(c,m) = h(c) + (m^{1-\theta} - 1)/(1-\theta)$ when $0 < \theta \neq 1$ and $u(c,m) = h(c) + \ln c$ when $\theta \to 1$. Let $\alpha \equiv 1/h'(c)$. Then, there is a steady state at $\bar{m} = [\alpha\mu/(\mu - \beta)]^{1/\theta}$. We have $dm_{t+1}/dm_t = \mu\beta - [(1-\theta)(\mu - \beta)/\beta](m_t/\bar{m})^{-\theta} > 0$ if $m_t > \bar{m}$ and $dm_{t+1}/dm_t = \mu\beta - (1-\theta)(\mu - \beta)/\beta > 1$ if $m_t = \bar{m}$. Thus, the steady state is not stable and if $m_0 > \bar{m}$, the equilibrium is a deflationary bubble. This utility function satisfies Brock’s condition; hence, it does not satisfy the $GABOR$ condition. We have $m_{t+1}/m_t = (\mu/\beta)(1 - \alpha m_t^{1-\theta})$; hence $\lim_{t \to \infty} \beta^t u_c(c,m_t)m_t = (1/\alpha)\lim_{t \to \infty} \mu^t \prod_{s=0}^t (1 - \alpha m_t^{1-\theta}) = 0$. Thus, the equilibrium is not ruled out by the $ST$ condition. A particularly transparent
example is the case where $\theta = 1$. Then $m_t = (m_0 - \bar{m}) (\mu / \beta)^t + \bar{m}$ satisfies (12). If $m_0 > \bar{m}$ then $m_t \to \infty$ and clearly the $ST$ condition is satisfied, but not the $GABOR$ condition.
**Proposition 6** The $GABOR$ condition rules out Friedman’s Optimal Quantity of Money equilibrium; the $ST$ condition does not.
**Proof.** Friedman’s $OQM$ equilibrium (any fundamental equilibrium supported by $\beta = \mu$) has $(\beta / \mu)^t u_c(c, m_t) m_t = u_c(c, \bar{m}) \bar{m} > 0$, with $\bar{m} \geq \hat{m}$; hence $GABOR$ is not satisfied. It has $\beta^t u_c(c, m_t) m_t = \beta^t u_c(c, \bar{m}) \bar{m} \to 0$; hence, the $ST$ condition is satisfied. ■
The $GABOR$ condition rules out Friedman’s where $\mu = \beta$ and households are satiated in real balances.\footnote{The neoclassical or intertemporal public finance theory of the optimal quantity of money is a vast subject. See, for example, Bailey [2], Friedman [13], Wilson [37] and Cohen [9],} It is easily verified that in the model of this section and in the model with money and bonds in Section 3 that if Assumption 2a holds, Friedman’s $OQM$ equilibrium is the household’s most preferred outcome.
The intuition for why the $GABOR$ condition cannot be part of the household’s optimal programme is particularly stark in this case. The $GABOR$ condition is a statement that the household must be indifferent between reducing his money holdings by one unit and using this extra unit of money to purchase consumption and holding the unit of money forever and enjoying the discounted present value of the resulting liquidity services. However, as the household is satiated in real balances in the $OQM$ equilibrium, it would hold an extra unit of money as a pure store of value only in order to increase its consumption at some time in the future. At some point in the future, the incremental unit of money must be exchanged for the consumption good if the earlier sacrifice of consumption is to be rational.
We now show that when money growth equals the discount factor and there is satiation in real balances, deflationary bubbles cannot exist. This is a consequence of the Euler equation, rather than the $ST$ condition.
**Proposition 7** If $\mu = \beta < 1$ and Assumption 2a holds, then deflationary bubbles cannot exist.
**Proof.** When $\mu = \beta$, equation (12) becomes $u_c(c, m_{t+1})m_{t+1} = [u_c(c, m_t) - u_m(c, m_t)]m_t$, $t \geq 0$. If there exists a solution that has money balances rising without bound, then at some point, $m_t > \hat{m}(c)$. By Assumption 2a, if $m > \hat{m}(c)$ then $u_m(c, m) = 0$, and we have $u_c(c, m_{t+1})m_{t+1} = u_c(c, m_t)m_t$. For this to support $m_t$ rising without bound, we require $u_c(c, m_t)/u_c(c, m_{t+1}) > 1$. For this to be possible, $u_{cm}(c, m) < 0$. This is a contradiction, however, as $u_{cm}(c, m_t) = 0$ for $m_t > \hat{m}(c)$. ■
We have assumed that $\mu \geq \beta$. If this is not the case, it is easy to find examples of non-fundamental equilibria, but deflationary bubbles cannot exist. This result is also a consequence of the Euler equation, rather than the $ST$ condition.
**Proposition 8** If $\mu < \beta$ then deflationary bubbles cannot exist.
**Proof.** Rewrite equation (12) as $m_{t+1} = (\mu/\beta)[u_c(c, m_t)/u_c(c, m_{t+1}) - u_c(c, m_t)/u_c(c, m_{t+1})]m_t$. For $\{m_t\} \to \infty$, there must be a subsequence $\{m_{t_k}\}$ such that $u_c(c, m_{t_k})/u_c(c, m_{t_k+1}) > \beta/\mu > 1$. This implies, $u_c(c, m_{t_k}) \to 0$, which is not possible as it violates Assumption 1. ■
### 3 Deflationary Bubbles with Money and Government Bonds
In the money-only model of Section 2, the household transversality condition requires that the present discounted value of the household’s terminal stock of real money balances is zero. When there are financial or real assets besides money, then the corresponding household transversality condition (plus a solvency constraint) implies that the present discounted value of the household’s terminal *aggregate* net non-human wealth is zero. In the money-only model, if the nominal money stock is not falling, then any path of real balances that satisfies the household’s Euler equation and goes to infinity produces a path of present discounted values of real money balances that eventually rises or remains constant at some positive value, violating the household’s transversality condition.
Woodford [38] suggests that if an additional financial asset were available that could have a strictly negative supply, then the household transversality condition might be
satisfied even if the discounted value of the household’s real balances does not go to zero. In this section we analyse the case where the government issues real and nominal debt as well as unbacked money and we show that Woodford’s conjecture is correct when the government follows a tax rule which implies that the government’s solvency constraint is satisfied with strict inequality along a deflationary bubble path. We also show that the conjecture is not correct if the government follows a tax rule which implies that the government satisfies its solvency constraint with equality. This is a consequence of the asymmetry of the government and household budget constraints. The household views its terminal money stock as an asset. The government views money as irredeemable; hence, it does not view the terminal money stock as a liability.
### 3.1 Households
We now assume that the government issues nominal and real bonds, in addition to money. Nominal bonds pay a nominal interest rate of $i_t$ in period $t$ on debt acquired in period $t-1$; real bonds pay a real interest rate of $r_t$ in period $t$ on debt acquired at $t-1$. Since the nominal interest rate on money is assumed to be zero, an equilibrium with valued nominal bonds requires that the nominal interest rate be non-negative. We only consider monetary rules that support such equilibria. Equilibria where both nominal and real bonds are held require that the returns on these assets are equalised. Thus
\[
(1 + r_{t+1})P_{t+1}/P_t = 1 + i_{t+1}, \ t \geq 0. \tag{18}
\]
Denote the period-$t$ household demand for nominal bonds by $B^d_t$ and the period-$t$ demand for real bonds by $d^d_t$. Let $a^d_t \equiv m^d_t + b^d_t + d^d_t$ be the real value of the household’s time-$t$ demand for financial wealth, where $b^d_t \equiv B^d_t/P_t$. The household’s within-period budget constraint is
\[
a^d_t = (1 + r_t)a^d_{t-1} + y - \tau_t - c_t - (i_tP_{t-1}/P_t)m^d_{t-1}. \tag{19}
\]
The household faces a solvency constraint: it cannot run a Ponzi scheme where it borrows ever-increasing amounts to service its previously accumulated debt. Thus, the present discounted value of the household’s terminal financial wealth has to be non-negative:
\[
\lim_{t \to \infty} a_t^d / \prod_{s=0}^{t} (1 + r_s) \geq 0.
\]
(20)
We assume that the household’s initial holdings of money, nominal bonds and real bonds, \(M_{-1} > 0\), \(B_{-1}\) and \(d_{-1}\), respectively, are given.\(^{10}\)
The (interior) optimality conditions for the household are, for \(t \geq 0\), are (19) and
\[
\frac{u_m(c_t, m_t^d)}{u_c(c_t, m_t^d)} = \frac{i_{t+1}}{1 + i_{t+1}}
\]
(21)
\[
\beta (1 + r_{t+1}) u_c(c_{t+1}, m_{t+1}^d) = u_c(c_t, m_t^d)
\]
(22)
\[
\lim_{t \to \infty} \beta^t u_c(c_t, m_t^d) a_t^d = 0.
\]
(23)
Equation (21) is the familiar efficiency condition relating period-\(t\) money demand to period-\(t\) consumption. Equation (22) is the household’s consumption Euler equation. Equation (23) is the the \(ST\) condition for the money-and-bonds model.
It is common in the economics literature to replace the single transversality condition (23) with multiple transversality conditions, one for each component of financial wealth. Turnovsky [33], p. 389 and McCallum [26], for example, model households that hold money and bonds and impose two transversality conditions, one requiring that the discounted terminal shadow value of debt be zero and one requiring that the discounted terminal shadow value of money be zero. Other recent papers that take this approach are Jha, Wang and Yip [18], Chuang and Huo [8] and Schabert [31]. While multiple transversality conditions may help rule out inflationary bubbles and explosive debt, only
\(^{10}\)In the money-and-bonds model too, we do not consider the non-monetary equilibrium with \(P_t^{-1} = 0, \ t \geq 0\).
one transversality condition is part of the necessary and sufficient conditions for a household optimum. When the consumer’s optimisation problem is written as a discrete-time Hamiltonian control problem, there is a single state variable, real aggregate financial wealth, $a^d$, and two control variables, consumption, $c$, and real money balances, $m^d$. There is a single transversality condition involving the state variable $a^d$. There is no separate transversality condition requiring the present discounted shadow value of terminal real money balances to equal zero. The economic reason for this is that financial markets are assumed to be frictionless: households can costlessly and instantaneously change the composition of their portfolios between money and bonds.
It follows from equations (22), (23) and Assumption 1, that the household solvency constraint (20) holds with equality:
$$\lim_{t \to \infty} a_t^d / \prod_{s=0}^{t} (1 + r_s) = 0.$$
(24)
### 3.2 The government
The government’s outstanding stocks of nominal and real bonds at the beginning of period $t$ are denoted by $B_{t-1}$ and $d_{t-1}$, respectively. Let $b_t \equiv B_t/P_t$ and $a_t \equiv m_t + b_t + d_t$. The government’s period-$t$ budget constraint is
$$a_t = (1 + r_t)a_{t-1} + g - \tau_t - (i_tP_{t-1}/P_t)m_{t-1}, \ t \geq 1.$$
(25)
The government’s solvency constraint, given in (26), is that the present discounted value of the government’s terminal non-monetary liabilities is non-positive. This in contrast to the household solvency constraint, given in (20), which requires the present discounted value of the household’s terminal total net financial liabilities, monetary and non-monetary, to be non-positive. The rationale for the asymmetric specification of the private and public sectors’ solvency constraints is that, while the household views money as an asset that can be realised at any time, the government recognises that, unlike bonds,
unbacked base money is *irredeemable* or *inconvertible*. (see Buiter [6], [7]).\footnote{This explains why there is no government solvency constraint in the money-only model of Section 2.} Unbacked base money is perceived to be an asset by the private sector, even in the long run, but is not treated as an effective liability in the long run by the government. Reflecting this, the government’s solvency constraint is
$$\lim_{t \to \infty} f_t / \prod_{s=0}^{t} (1 + r_s) \leq 0,$$
where $f_t \equiv b_t + d_t$.
To demonstrate the possible existence of deflationary bubbles when the government issues non-monetary as well as monetary financial instruments, we specify a simple government tax rule: we suppose that the government keeps the real value of its net stock of debt (monetary and non-monetary) constant at the initial level $a_{-1} \geq 0$. From the government’s within-period budget constraint (25) it follows that taxes are given by
$$\tau_t = g + r_t a_{-1} - (i_t P_{t-1}/P_t) m_{t-1}.$$
Under this tax rule, the discounted terminal value of the government’s aggregate debt, both monetary and non-monetary, is clearly zero. Thus, the government satisfies its solvency constraint (26).\footnote{This requires that long-run real interest rate is positive. With the restrictions we imposed to ensure a unique monetary steady state, the long-run real interest rate will always be positive. Were this not the case, setting $a_t = 0$, $t \geq 0$ would guarantee solvency. The tax sequence would be given by $\tau_0 = g + (1 + r_0)a_{-1} - (i_0 P_{-1}/P_0) m_{-1}$ and $\tau_t = g - (i_t P_{t-1}/P_t) m_{t-1}$, $t \geq 0$.} By (18) it follows that taxes are given by:
$$\tau_t = g + \left( \frac{(1 + i_t) m_t}{\mu m_{t-1}} - 1 \right) a_{-1} - \frac{i_t m_t}{\mu}.$$
### 3.3 Market clearing
Market clearing requires that $m_t^d = m_t$ and $a_t^d = a_t$, $t \geq 0$. As before, the resource constraint implies that $c_t = c \equiv y - g$, $t \geq 0$. Then by equations (18), (21) and (22) we have the following definition:
**Definition 9** A monetary equilibrium is a sequence of pairs \(\{(m_t, i_t)\}_{t=0}^{\infty}\) such that \(m_t > 0\) and \(i_t \geq 0\) for every \(t \geq 0\) and
\[
\beta u_c(c, m_{t+1})m_{t+1} = \mu[u_c(c, m_t) - u_m(c, m_t)]m_t, \ t \geq 0 \tag{29}
\]
\[
i_{t+1} = \frac{u_m(c, m_t)}{u_c(c, m_t) - u_m(c, m_t)}, \ t \geq 0 \tag{30}
\]
\[
\lim_{t \to \infty} \beta^t u_c(c, m_t)a_t = a_{-1} \lim_{t \to \infty} \beta^t u_c(c, m_t) = 0. \tag{31}
\]
As before, a unique monetary steady state exists. It has the associated nominal interest rate \(\bar{i} = (\mu - \beta)/\beta\) and real interest rate \(\bar{r} = (1 - \beta)/\beta\).
Deflationary bubble paths result in the nominal interest rate going to zero and, by (28) - (30),
\[
\tau_t \to g + \frac{1 - \beta}{\beta}a_{-1} - \lim_{t \to \infty} \frac{u_m(c, m_t)m_t}{\mu u_c(c, m_t)}. \tag{32}
\]
With constant real aggregate financial liabilities \((a_t = a_{-1}, \ t \geq 0)\), the household’s transversality condition and the household’s and government’s solvency constraints are satisfied even when real money balances rise without bound and the present discounted value of the terminal money stock is strictly positive. As the stock of real money balances rises without bound along a deflationary bubble path, the real value of government bonds falls without bound, keeping the real value of the sum of these two portfolio components constant.
Consider the example from the previous section where \(u(c, m) = h(c) + \ln m\). Taxes are constant, since \(\tau_t = g + (1 - \beta)a_{-1}/\beta - (\mu - \beta)\bar{m}/(\mu \beta)\), and the real interest rate equals \((1 - \beta)/\beta\). Suppose \(\mu > \beta\).\(^{13}\) Then \(m_t = (m_0 - \bar{m})(\mu/\beta)^t + \bar{m}\) and \(i_t = (\mu - \beta)\bar{m}/[\mu m_{t-1} - (\mu - \beta)\bar{m}]\). If \(m_0 > \bar{m}\), then real balances go to infinity and the nominal interest rate goes to zero. If the nominal money stock is rising \((\mu > 1)\), then \(\beta^t u_c(c, m_t)m_t \to \infty\). It follows that \(\beta^t u_c(c, m_t)(b_t + d_t) \to -\infty\) to keep \(\lim_{t \to \infty} \beta^t u_c(c, m_t)a_t = 0\). If the nominal
\(^{13}\)The Friedman rule, \(\mu = \beta\), implies an infinite stock of real money balances, so there is no OQM equilibrium that can be implemented in this case.
money stock is constant \((\mu = 1)\), then \(\beta' u_c(c, m_t) m_t \to (m_0 - \bar{m}) h'(c)\) and the present discounted value of government bonds goes to \(-(m_0 - \bar{m}) h'(c)\).
Many other tax rules would support deflationary bubbles in the money-and-bonds model. They all share the property that if the present discounted value of real balances goes to infinity, the present discounted real value of government bonds goes to minus infinity, thus ensuring that the present discounted value of the aggregate financial liabilities of the government goes to zero in the long run.
**Proposition 10** With a non-decreasing nominal money stock, deflationary bubbles cannot be ruled out when the government issues both money and bonds if the present discounted value of government’s terminal non-monetary liabilities can be negative (if \(\mu = 1\)) or if they can go to minus infinity (if \(\mu > 1\)).
The government’s solvency constraint (26) requires the discounted value of its terminal non-monetary debt to be non-positive. The fiscal rule \(a_t \equiv f_t + m_t = a_{-1}, t \geq 0\) that was shown to support deflationary bubbles ensures that the government solvency constraint (26) is satisfied, since \(m_t \geq 0\). However, unless the present value of terminal real balances is zero, (27) satisfies (26) with strict inequality. We now show that when the government adopts instead of (27) a fiscal rule that ensures its solvency constraint is satisfied with equality, deflationary bubbles cannot exist unless the growth rate of the nominal money stock is negative.
An example of a tax rule that causes (26) to hold with equality is one that keeps constant the real value of the government’s bonds, that is, \(f_t = f_{-1}, t \geq 0\). This rule is among the most common ad hoc rules in the macroeconomics literature. Associated taxes are given by
\[
\tau_t = g + r_t f_{-1} - \left( \frac{\mu - 1}{\mu} \right) m_t
\]
**(33)**
**Proposition 11** If the government satisfies its solvency constraint (26) with equality, then all of the Propositions for the money-only model of Section 2 hold for the model with money and government debt.
Proof. Market clearing and equations (22) and (26) (with equality) imply
\[
\lim_{t \to \infty} \beta^t u_c(c, m_t) f_t = 0. \tag{34}
\]
For the household sector, the combined solvency constraint and transversality condition from the previous section, condition (24), continues to hold. This, market clearing and (22) imply
\[
\lim_{t \to \infty} \beta^t u_c(c, m_t) a_t = 0. \tag{35}
\]
Together, equations (34) and (35) imply
\[
\lim_{t \to \infty} \beta^t u_c(c, m_t) m_t = 0. \tag{36}
\]
The proof demonstrates that if the government solvency constraint holds with equality, then there is an equilibrium requirement that the present discounted value of the terminal money stock be zero. This does not come solely from the household’s transversality and solvency conditions, as was the case with the money-only model. Instead it is an implication of these conditions, the assumption of irredeemable money (and the associated asymmetry between the household and government solvency constraints), the requirement that the government solvency constraint hold with equality and the market-clearing conditions. With the requirement restored that the present value of the terminal money stock be zero, all propositions concerning the existence and non-existence of deflationary bubbles and about the \( GABOR \) condition, derived for the money-only model in Section 2, now also apply unchanged in the money-and-bonds model.
Neither the specific rule in (33) nor the requirement that the government’s solvency constraint hold with equality are derived from optimising government behaviour. The assumption that the government’s solvency constraint holds with equality sounds sensible, in that optimising private economic agents typically satisfy their budget constraints with equality - satiation in commodities is not a common feature of standard models of household behaviour. Without the possibility of bubbles and with distortionary taxes or real tax administration and compliance costs, a benevolent optimising government would choose to satisfy its budget constraint with equality. However because of the possibility of bubbles, this assumption is not innocuous here. The rule \(a_t = a_{-1}, \ t \geq 0\) satisfies the government’s constraint (26) of this section with strict inequality unless \(\lim_{t \to \infty} m_t / \prod_{s=0}^{t} (1 + r_s) = 0\). If nominal money growth is strictly positive, it can produce a deflationary bubble. As welfare is higher in the deflationary bubble equilibrium than it would be in the associated fundamental equilibrium, household welfare is higher in a bubble equilibrium under the rule \(a_t = a_{-1}, \ t \geq 0\) than it is under the rule \(f_t = f_{-1}, t \geq 0\) which satisfies the budget constraint of this section with equality and which supports only the fundamental equilibrium.\(^{14}\)
When households can hold physical capital, but not government bonds, the logic of the money-only model prevails, and deflationary bubbles are not possible when the nominal money stock is non-decreasing. The household transversality condition (23) would still hold, but with \(a_t^d = m_t^d + k_t^d\), where \(k_t^d\) is the household’s demand for real capital. Even if an individual household can hold negative quantities of capital (equity), in equilibrium the aggregate real capital stock is constrained to be non-negative and cannot play the role of government bonds in permitting deflationary bubbles.
In both the money-only and the money-and-bonds models we introduce a motive for holding money and generate money demand functions that depend on the nominal interest rate by including real money balances as an argument in the direct utility function, along the lines pioneered by Sidrauski [32]. This way of introducing a motive for holding money is not crucial for any of our results, however. Other ways of making the demand for money sensitive to the nominal interest rate, such as a flexible cash-in-advance model with cash goods and credit goods (see, for example, Lucas and Stokey [24]) or a shopping,
\(^{14}\)We assume that the growth rate of the nominal money stock and the level of real public spending are the same under both fiscal rules.
time-savings model (see, for example, Feenstra [10]), would produce identical results to those obtained here. Indeed, all our key results hold also for the constant-velocity cash-in-advance models of Lucas [23], Helpman [17] and Sargent [30], Chapter 5.\footnote{Deflationary bubbles cannot exist in an overlapping-generations model where money is the only store of value and is only held as a store of value. Consider the case where households live for two periods, only the young receive a positive endowment of the good and the nominal money stock is constant. Consumption by the old then equals the value of the stock of real money balances. A deflationary bubble would mean that old consumers’ demand for the good is growing without bound. As the amount of the good supplied by the young is bounded above by the sum of their endowments, this cannot be an equilibrium. See, for example, Hahn [15], p. 10.}
It is clear that none of the results depend on the interest-sensitivity of the demand for real money balances. The key assumption that makes the household transversality condition bite is that the marginal utility of \textit{consumption} is positive. With the unitary velocity of circulation of the simple cash-in-advance models (say $M_t \geq P_t(c_t + g)$, with $M_t = P_t(c_t + g)$ if $i_t > 0$), and a period utility function $u(c)$, with $u$ increasing, twice continuously differentiable, strictly concave and satisfying the Inada conditions) deflationary bubbles exist or fail to exist under the same conditions that they do with the money in the utility function under Assumptions 1, 2A (satiation) and normality of real money balances. The simple cash-in-advance model has the equivalent of ‘satiation’ at $\hat{m} = c + g$.
\section{Deflationary Bubbles in a Small Open Economy}
This approach to the existence of deflationary bubbles when the nominal money stock is non-decreasing can be applied to the small open economy model of Obstfeld and Rogoff [29], pp. 536-543. This model has a single tradable commodity, a freely floating exchange rate and perfect international capital mobility. The domestic household holds home and foreign real bonds. Home money is held by the home household, but not by foreign households. The world rate real interest rate of interest, $r^*$, is exogenous and assumed to equal the domestic consumer’s time preference rate: $r_t = r^* = (1 - \beta)/\beta$, $t \geq 0$. The government balances its budget each period and has no initial debt; hence $f_t = 0$, $t \geq -1$ and its within-period budget constraint is
While the government does not borrow or lend, households can freely buy and sell bonds on the international capital market. Let $b^*_t$ denote the time-$t$ real value of net household claims on the rest of the world. The household’s period budget constraint is still (19), but now $a_t \equiv m_t + b^*_t$. The household solvency constraint (20), and the household optimality conditions (21), (22) and (23) are the same as in the closed economy model of Sections 3 and 4. The closed economy resource constraint $c_t = y - g$ no longer constrains private consumption. Instead, the private and public sector budget constraints imply the nation’s within-period resource constraint is:
$$b^*_t \equiv (1 + r^*) b^*_{t-1} + y - g - c_t.$$
The household transversality and solvency conditions require that $\lim_{t \to \infty} \beta^t u_c(c, m_t) (m_t + b^*_t) = \lim_{t \to \infty} (m_t + b^*_t)/(1 + r^*)^t = 0$. This can be satisfied if the present discounted value of the terminal money stock is strictly positive, provided the present discounted value of the household’s terminal holdings of net external assets takes on a matching negative value.\footnote{As the home country is small, this does not violate any foreign transversality condition. One might argue, however, that if the home country is borrowing ever increasing amounts from the rest of the world that the small economy assumption would eventually cease to be sensible, unless the rest of the world is growing at a rate at least equal to the rate of interest.}
Because the government views money as non-redeemable, with no home government non-monetary debt debt ($f_t = 0$), the government’s solvency condition 26 is trivially satisfied as well.\footnote{Obstfeld and Rogoff appeal to a GABOR argument to rule out deflationary bubbles in this open economy example also (see [29], pp. 542-543).}
\section{Conclusion}
The paper ties up a number of loose ends in the deflationary bubbles literature; we have detailed our main results in the introduction. What remains to be done in future research
is a reconciliation of the neoclassical public finance approach to monetary policy, both positive and normative, of which this paper is an example, and the more diverse and eclectic literature on deflation, debt deflation, monetary policy ineffectiveness, financial fragility, recession and depression. This deflationary crisis literature goes back at least to Fisher [11], [12], Wicksell [35, 36], Keynes [20], and Hayek [16] and other Austrian School economists and figures prominently in the more recent policy-oriented contributions of authors like Bernanke [3], King [21] and Krugman [22].
In the neoclassical public finance approach, deflation and zero nominal interest rates are not a serious cause for concern. The deflationary bubbles analysed in this paper all support the maximum feasible level of private consumption as an equilibrium in every period, and the real money stock converges to its satiation value or higher. The stationary Optimal Quantity of Money equilibrium, characterised by steady deflation and a zero nominal interest rate, represents the social optimum in the models considered in our paper. In the deflationary crisis literature, deflation and the zero bound are to be avoided and are a source of concern for monetary policy makers. The difference between the two approaches goes well beyond the fact that the deflationary equilibria considered in this paper concern fully anticipated deflations, while the deflationary crisis literature is concerned with the response of the economic system to unexpected shocks - mainly unexpected contractionary demand shocks (leading to 'bad' deflations) but also, more recently, unexpected expansionary supply shocks (leading to 'good' deflations). The deflationary crisis literature uses models that emphasize nominal and real rigidities (both *ad-hoc* nominal wage or price rigidities, and real rigidities based on asymmetric information considerations), incomplete markets and non-competitive behaviour by enterprises and financial intermediaries. The gap between these two approaches is so wide that it may not be possible to come up with a tractable analytical model that encompasses them. However, both approaches offer important insights, and the hope that a synthesis can be achieved that is more than the sum of the parts should encourage a renewed focus on a research agenda that aims to join these two perspectives.
References
[1] Azariadis, Costas (1993), *Intertemporal Macroeconomics*, Blackwell Publishers, Cambridge Mass. USA.
[2] Bailey, Martin J. [1956], “The Welfare Costs of Inflationary Finance,” *Journal of Political Economy*, vol. 64, no. 2, April, pp. 93–110.
[3] Bernanke B. (1983), “Nonmonetary Effects of the Financial Crisis in the Propagation of the Great Depression,” *The American Economic Review*, June, pp. 257—76.
[4] Brock, William A. (1974), "Money and Growth: The Case of Long Run Perfect Foresight", *International Economic Review*, 15(3), October, pp. 750-777.
[5] Brock, William A. (1975), "A Simple Perfect Foresight Monetary Model", *Journal of Monetary Economics*, 1, pp. 133-150.
[6] Buiter, Willem H. (2003b), "Helicopter Money: Irredeemable Fiat Money and the Liquidity Trap", NBER Working Paper No. 10163, December 2003.
[7] Buiter, Willem H. (2005), "A Small Corner of Intertemporal Public Finance. New Developments in Monetary Economics: two ghosts, two eccentricities, a fallacy, a mirage and a mythos", Royal Economic Society 2004 Hahn Lecture, given on 6 April 2004, at the Royal Economic Society Annual Conference, Swansea, UK. forthcoming *The Economic Journal, Conference Papers*, 2005.
[8] Chuang, Shi-Feng and Teh-Ming Huo, "A Note on the Optimum Quantity of Money," *Economic Letters* 79, May, 269-276.
[9] Cohen, Daniel (1985), "Inflation, Wealth and Interest Rates in an Intertemporal Optimising Model", *Journal of Monetary Economics*, 16, July, pp. 73-85.
[10] Feenstra, Robert (1986), "'Functional Equivalence Between Liquidity Costs and the Utility of Money,", *Journal of Monetary Economics*, 17, pp. 271-291.
[11] Fisher, Irving (1932), *Booms and Depressions*, New York: Adelphi Company.
[12] Fisher, Irving (1933), ‘The Debt-Deflation Theory of Great Depressions’, *Econometrica*, March, pp. 337-57.
[13] Friedman, Milton (1969), "The Optimum Quantity of Money", in Milton Friedman, *The Optimum Quantity of Money and Other Essays*, Chapter 1, pp.1-50, Adline Publishing Company, Chicago.
[14] Gray, Jo Anna [1984], "Dynamic Instability in Rational Expectations Models: An Attempt to Clarify", *International Economic Review*, Vol.25, No. 1, February, pp. 93-122.
[15] Hahn, Frank (1982), *Money and Inflation*, Basil Blackwell, London.
[16] Hayek, Friedrich A. von (1929), *Geldtheorie und Konjunkturtheorie*, Vienna; in English as *Monetary Theory and the Trade Cycle*, London, 1933.
[17] Helpman, Elhanan (1981), "An Exploration in the Theory of Exchange Rate Regimes", *Journal of Political Economy*, 89(5), pp. 865-890.
[18] Jha, Sailesh K., Wang, Ping and Chong K. Yip (2002), "Dynamics in a Transactions-Based Monetary Growth Model," *Journal of Economic Dynamics and Control* 26, Apr., 611-635.
[19] Kamihigashi, Takashi (2000), "A Simple Proof of Ekeland and Scheinkman’s Result on the Necessity of a Transversality Condition," *Economic Theory* 15, 463-468.
[20] Keynes, John M. (1930), *A Treatise on Money*, (2 Vols.), MacMillan, London.
[21] King, Mervyn (1994), “Debt Deflation: Theory and Evidence.”, *European Economic Review*, 38, pp. 419-445.
[22] Krugman, Paul (1998), "It’s baaack! Japan’s slump and the return of the liquidity trap", *Brookings Papers on Economic Activity*, Vol. 2, pp. 137-205.
[23] Lucas, Robert E. Jr. (1980), "Equilibrium in a Pure Currency Economy", *Economic Enquiry* 18(2), pp. 203-220.
[24] Lucas, Robert E. Jr. and Nancy L. Stokey (1987), "Money and Interest in a Cash-in Advance Economy," *Econometrica* 55, pp. 491-513.
[25] Lucas, Robert E., Jr. and Nancy L. Stokey (1989), *Recursive Methods in Economic Dynamics*, Harvard University Press, Cambridge, MA.
[26] McCallum, B. (2001), "Indeterminacy, Bubbles, and the Fiscal Theory of Price Level Determination," *Journal of Monetary Economics*, 47, pp. 19-30.
[27] Obstfeld, M. and Rogoff, K. (1983), Speculative hyperinflations in maximizing models: Can we rule them out?", *Journal of Political Economy*, 91, August, pp. 675-87.
[28] Obstfeld, M. and Rogoff, K. (1986), Ruling out divergent speculative bubbles', *Journal of Monetary Economics*, 17, May, pp. 346-62.
[29] Obstfeld, M. and K. Rogoff (1996), *Foundations of International Macroeconomics*, The MIT Press, Cambridge Mass.
[30] Sargent, T. J. (1987), *Dynamic Macroeconomic Theory*, Harvard University Press, Cambridge Massachusetts.
[31] Schabert, Andreas, "Identifying Monetary Policy Shocks with Changes in Open Market Operations," *European Economic Review*, forthcoming.
[32] Sidrauski, Miguel (1967), "Rational Patterns of Growth in a Monetary Economy", *American Economic Review*, 57, May, pp. 534-544..
[33] Turnovsky, Stephen J. (2000), *Methods of Macroeconomic Dynamics, 2nd ed.* MIT Press, Cambridge Mass.
[34] Weitzman, M. L. (1973), "Duality Theory for Infinite Horizon Convex Models." *Management Science* 19, no. 7, March, pp. 783-789.
[35] Wicksell, Knut (1901), *Lectures on Political Economy*, Vol. 1, London: Routledge & Kegan Paul.
[36] Wicksell, Knut (1906), *Lectures on Political Economy*, Vol. 2., London: Routledge & Kegan Paul.
[37] Wilson, Charles (1979), "An Infinite Horizon Model with Money", in Jerry R. Green and Jose Alexandre Scheinkman eds. *General Equilibrium, Growth, and Trade: Essays in Honor of Lionel McKenzie*, New York, Academic Press, 1979.
[38] Woodford, Michael (2003), *Interest and Prices; Foundations of a Theory of Monetary Policy*, Princeton University Press, Princeton and Oxford. |
2017 AMC 10B
For more practice and resources, visit ziml.areteem.org
The problems in the AMC-Series Contests are copyrighted by American Mathematics Competitions at Mathematical Association of America (www.maa.org).
Mary thought of a positive two-digit number. She multiplied it by 3 and added 11. Then she switched the digits of the result, obtaining a number between 71 and 75, inclusive. What was Mary's number?
(A) 11 (B) 12 (C) 13 (D) 14 (E) 15
Sofia ran 5 laps around the 400-meter track at her school. For each lap, she ran the first 100 meters at an average speed of 4 meters per second and the remaining 300 meters at an average speed of 5 meters per second. How much time did Sofia take running the 5 laps?
(A) 5 minutes and 35 seconds
(B) 6 minutes and 40 seconds
(C) 7 minutes and 5 seconds
(D) 7 minutes and 25 seconds
(E) 8 minutes and 10 seconds
Real numbers $x$, $y$, and $z$ satisfy the inequalities $0 < x < 1$, $-1 < y < 0$, and $1 < z < 2$. Which of the following numbers is necessarily positive?
(A) $y + x^2$
(B) $y + xz$
(C) $y + y^2$
(D) $y + 2y^2$
(E) $y + z$
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
---
Suppose that $x$ and $y$ are nonzero real numbers such that $\frac{3x + y}{x - 3y} = -2$. What is the value of $\frac{x + 3y}{3x - y}$?
(A) $-3$
(B) $-1$
(C) $1$
(D) $2$
(E) $3$
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
---
Camilla had twice as many blueberry jelly beans as cherry jelly beans. After eating 10 pieces of each kind, she now has three times as many blueberry jelly beans as cherry jelly beans. How many blueberry jelly beans did she originally have?
(A) 10
(B) 20
(C) 30
(D) 40
(E) 50
Select one:
What is the largest number of solid 2-in × 2-in × 1-in blocks that can fit in a 3-in × 2-in × 3-in box?
(A) 3 (B) 4 (C) 5 (D) 6 (E) 7
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
Samia set off on her bicycle to visit her friend, traveling at an average speed of 17 kilometers per hour. When she had gone half the distance to her friend’s house, a tire went flat, and she walked the rest of the way at 5 kilometers per hour. In all it took her 44 minutes to reach her friend’s house. In kilometers rounded to the nearest tenth, how far did Samia walk?
(A) 2.0 (B) 2.2 (C) 2.8 (D) 3.4 (E) 4.4
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
Points $A(11, 9)$ and $B(2, -3)$ are vertices of $\triangle ABC$ with $AB = AC$. The altitude from $A$ meets the opposite side at $D(-1, 3)$. What are the coordinates of point $C$?
(A) $(-8, 9)$ (B) $(-4, 8)$ (C) $(-4, 9)$ (D) $(-2, 3)$ (E) $(-1, 0)$
Select one:
A radio program has a quiz consisting of 3 multiple-choice questions, each with 3 choices. A contestant wins if he or she gets 2 or more of the questions right. The contestant answers randomly to each question. What is the probability of winning?
(A) $\frac{1}{27}$
(B) $\frac{1}{9}$
(C) $\frac{2}{9}$
(D) $\frac{7}{27}$
(E) $\frac{1}{2}$
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
The lines with equations $ax - 2y = c$ and $2x + by = -c$ are perpendicular and intersect at $(1, -5)$. What is $c$?
(A) $-13$
(B) $-8$
(C) 2
(D) 8
(E) 13
Select one:
At Typico High School, 60% of the students like dancing, and the rest dislike it. Of those who like dancing, 80% say that they like it, and the rest say that they dislike it. Of those who dislike dancing, 90% say that they dislike it, and the rest say that they like it. What fraction of students who say they dislike dancing actually like it?
(A) 10%
(B) 12%
(C) 20%
(D) 25%
(E) $33\frac{1}{3}\%$
Elmer's new car gives 50% percent better fuel efficiency, measured in kilometers per liter, than his old car. However, his new car uses diesel fuel, which is 20% more expensive per liter than the gasoline his old car used. By what percent will Elmer save money if he uses his new car instead of his old car for a long trip?
(A) 20%
(B) $26\frac{2}{3}\%$
(C) $27\frac{7}{9}\%$
(D) $33\frac{1}{3}\%$
(E) $66\frac{2}{3}\%$
There are 20 students participating in an after-school program offering classes in yoga, bridge, and painting. Each student must take at least one of these three classes, but may take two or all three. There are 10 students taking yoga, 13 taking bridge, and 9 taking painting. There are 9 students taking at least two classes. How many students are taking all three classes?
(A) 1 (B) 2 (C) 3 (D) 4 (E) 5
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
An integer $N$ is selected at random in the range $1 \leq N \leq 2020$. What is the probability that the remainder when $N^{16}$ is divided by 5 is 1?
(A) $\frac{1}{5}$ (B) $\frac{2}{5}$ (C) $\frac{3}{5}$ (D) $\frac{4}{5}$ (E) 1
Select one:
Rectangle $ABCD$ has $AB = 3$ and $BC = 4$. Point $E$ is the foot of the perpendicular from $B$ to diagonal $\overline{AC}$. What is the area of $\triangle AED$?
(A) 1 (B) $\frac{42}{25}$ (C) $\frac{28}{15}$ (D) 2 (E) $\frac{54}{25}$
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
How many of the base-ten numerals for the positive integers less than or equal to 2017 contain the digit 0?
(A) 469 (B) 471 (C) 475 (D) 478 (E) 481
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
Call a positive integer *monotonous* if it is a one-digit number or its digits, when read from left to right, form either a strictly increasing or a strictly decreasing sequence. For example, 3, 23578, and 987620 are monotonous, but 88, 7434, and 23557 are not. How many monotonous positive integers are there?
(A) 1024 (B) 1524 (C) 1533 (D) 1536 (E) 2048
Select one:
In the figure below, 3 of the 6 disks are to be painted blue, 2 are to be painted red, and 1 is to be painted green. Two paintings that can be obtained from one another by a rotation or a reflection of the entire figure are considered the same.
How many different paintings are possible?
(A) 6 (B) 8 (C) 9 (D) 12 (E) 15
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
Let \( \triangle ABC \) be an equilateral triangle. Extend side \( \overline{AB} \) beyond \( B \) to a point \( B' \) so that \( BB' = 3AB \). Similarly, extend side \( \overline{BC} \) beyond \( C \) to a point \( C' \) so that \( CC' = 3BC \), and extend side \( \overline{CA} \) beyond \( A \) to a point \( A' \) so that \( AA' = 3CA \). What is the ratio of the area of \( \triangle A'B'C' \) to the area of \( \triangle ABC \)?
(A) 9 : 1 (B) 16 : 1 (C) 25 : 1 (D) 36 : 1 (E) 37 : 1
Select one:
The number $21! = 51,090,942,171,709,440,000$ has over 60,000 positive integer divisors. One of them is chosen at random. What is the probability that it is odd?
(A) $\frac{1}{21}$
(B) $\frac{1}{19}$
(C) $\frac{1}{18}$
(D) $\frac{1}{2}$
(E) $\frac{11}{21}$
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
In $\triangle ABC$, $AB = 6$, $AC = 8$, $BC = 10$, and $D$ is the midpoint of $\overline{BC}$. What is the sum of the radii of the circles inscribed in $\triangle ADB$ and $\triangle ADC$?
(A) $\sqrt{5}$
(B) $\frac{11}{4}$
(C) $2\sqrt{2}$
(D) $\frac{17}{6}$
(E) 3
Select one:
The diameter $\overline{AB}$ of a circle of radius 2 is extended to a point $D$ outside the circle so that $BD = 3$. Point $E$ is chosen so that $ED = 5$ and line $ED$ is perpendicular to line $AD$. Segment $\overline{AE}$ intersects the circle at a point $C$ between $A$ and $E$. What is the area of $\triangle ABC$?
(A) $\frac{120}{37}$ (B) $\frac{140}{39}$ (C) $\frac{145}{39}$ (D) $\frac{140}{37}$ (E) $\frac{120}{31}$
Let $N = 123456789101112 \ldots 4344$ be the 79-digit number that is formed by writing the integers from 1 to 44 in order, one after the other. What is the remainder when $N$ is divided by 45?
(A) 1 (B) 4 (C) 9 (D) 18 (E) 44
The vertices of an equilateral triangle lie on the hyperbola \(xy = 1\), and a vertex of this hyperbola is the centroid of the triangle. What is the square of the area of the triangle?
(A) 48 (B) 60 (C) 108 (D) 120 (E) 169
Select one:
- A
- B
- C
- D
- E
- Leave blank (1.5 points)
Last year Isabella took 7 math tests and received 7 different scores, each an integer between 91 and 100, inclusive. After each test she noticed that the average of her test scores was an integer. Her score on the seventh test was 95. What was her score on the sixth test?
(A) 92 (B) 94 (C) 96 (D) 98 (E) 100
Select one: |
NEVADA STATE BOARD OF DENTAL EXAMINERS
6010 S. Rainbow Boulevard, Suite A-1
Las Vegas, Nevada 89118
Thursday, September 20, 2012 at 6:00 pm
MINUTES
Board Meeting Agenda
Videoconferencing is available at the Board office, 6010 S Rainbow Boulevard, Suite A-1, in Las Vegas and at the Nevada State Board of Medical Examiners, 1105 Terminal Way, Suite 301, Reno, Nevada 89502.
Please Note: The Nevada State Board of Dental Examiners may 1) address agenda items out of sequence to accommodate persons appearing before the Board or to aid the efficiency or effectiveness of the meeting; 2) combine items for consideration by the public body; 3) pull or remove items from the agenda at any time. The Board may convene in closed session to consider the character, alleged misconduct, professional competence or physical or mental health of a person. See NRS 241.030. Prior to the commencement and conclusion of a contested case or a quasi judicial proceeding that may affect the due process rights of an individual the board may refuse to consider public comment. See NRS 233B.126.
Public comment is welcomed by the Board, but at the discretion of the Chair, may be limited to five minutes per person. A public comment time will be available before any action items are heard by the public body and then once again prior to adjournment of the meeting. The Chair may allow additional time to be given a speaker as time allows and in his/her sole discretion. Once all items on the agenda are completed the meeting will adjourn. Prior to the commencement and conclusions of a contested case or a quasi judicial proceeding that may affect the due process rights of an individual the board may refuse to consider public comment.
Call to Order
1. **Roll call and Establish a Quorum:**
Dr. Miller called the meeting to order and Ms. Kelly conducted the following role call:
Dr. Jade Miller-----------------------------------------------PRESENT
Dr. M Masih Soltani------------------------------------------PRESENT
Dr. Byron Blasco---------------------------------------------PRESENT
Dr. J Gordon Kinard------------------------------------------PRESENT
Dr. Timothy Pinther-------------------------------------------PRESENT
Dr. Jason Champagne------------------------------------------PRESENT
Ms. Kelly request clarification from Dr Miller as a procedural item as to whether he was going to entertain public comment during the two workshops as well. Dr Miller confirmed the invitation of public comment will be accepted during the workshops.
Other attendees: John Hunt, Esq, Board Legal Counsel, Debra Shaffer, Deputy Executive Director, Shari Peterson, RDH, CSN & NDHA, Ted Twesme, DDS Anesthesia Evaluator, Michon Villanueva, Director of Accreditation Service, Military Health Care, Bob Talley, NDA, Martin. Thorsteinsson, Esq. Alverson Taylor Mortensen & Sanders, Andrew Ingel, UNLV School of Dental Medicine, Bob Anderson, Ex. Director SNDS, Kelly Taylor, RDH, Peggy Chesser, SNDHA Heather Rogers, RDH, President of NDHA, Laura Lord, RDH, Co-Chair President SNDHA, Silvia Carpenter for Cody Mugelston, DMD, Tom Miller, DDS. No Public Attendees in Northern Nevada.
Pledge of Allegiance
2. Public Comment. No Public Comment
Note: No vote may be taken upon a matter raised under this item of the agenda until the matter itself has been specifically included on an agenda as an item upon which action may be taken. (NRS 241.020)
*3. Notice of Workshop and Request for Comments Regarding Temporary Regulation Changes to Nevada Administrative Code 631.173 Continuing Education: Required Hours; Types of Courses and Activities; Approval of Provider or Instructor and Nevada Administrative Code 631.175 Continuing Education: Approved Subjects, Maximum Credit for Certain Types of Courses and Activities (For Possible Action)
DISCUSSION: Dr Miller introduced Dr Sill as the Chair of the CE Resource Group to address the Board about entertaining a change to the regulations pertaining to the amount of continuing education obtained by licensees through on-line method. Dr Sill stated currently the Board requires 70% to be live lecture based courses and 30% may be obtained through on-line methods. Dr Sill entertained public discussion or comment. Mrs. Rogers, President of the Nevada Dental Hygiene Association had no problem with the current regulation as is. Ms. Rogers feels licensees receive the most through live lecture courses. Dr Sill stated most Boards are going to more on-line courses due to accessibility and demand. Dr Sill thinks this would provide more flexibility. Dr Sill stated he would entertain expanding to journal study, study clubs, and self study. Dr. Sill does not believe there is a problem with licensees achieving
the requirement for CE’s, he believes for the licensees who want to attend live CE’s they may do so, for the licensees who wants to take on-line courses they should be allowed. Mr. McKernan stated the way the movement is more courses are being offered on-line. Mr. McKernan believes the direction things are going on-line courses are coming more popular and sees this trend increasing. Dr Miller stated he is the Chair of the continuing education recognition process for the ADA, the ADA just had a meeting to review applications for potential providers. Dr Miller stated it is a vigorous process as can be verified by Dr Talley with the Nevada Dental Association. Dr Sill recommended 50% live lecture and 50% for on-line. Dr Miller did request defining other methods of electronic media for CE. NAC 631.175(5) was read into the record by Ms. Kelly. Dr Guillen made comment regarding courses and hours related to CPR. Ms. Kelly made comment regarding the restriction of three (3) hours with regards to study clubs and asked if the Board would entertain increasing the number of hours permitted. Dr Miller requested the Board look into what other states are doing.
No action taken
*4. Notice of Workshop and Request for Comments Regarding Temporary Regulation Changes to Nevada Administrative Code 631.2211-631.2256 Authorizing the Administration of Conscious Sedation, Deep Sedation, and General Anesthesia. The General Topics Include but are Not Limited to: Revisions to Permit Application Process; Suspension Authority for Executive Director; Inspections and Re-Inspections; Evaluations and Re-Evaluations; Training for Emergencies. Use of Emergency Drugs, Consideration to Further Define Definitions of Conscious Sedation (NRS 631.025), Deep Sedation (NRS 631.027) and General Anesthesia (NRS 631.055) (For Possible Action)
DISCUSSION: Dr Miller stated the Board has been tackling this matter for some time. Dr Miller stated Mr. Drizin, Esq., Special Legal Counsel for the Board has come up with some proposed language. Mr. Drizin requested to make a couple of comments before the Board addresses public comment. Mr. Drizin stated there is some confusion regarding addressing some of the provisions regarding terminology, definitions, and evaluations as this has worked through the Anesthesia Committee. Currently, the Nevada legislation defines conscious sedation, deep sedation and general anesthesia. Mr. Drizin read the definition of conscious sedation into the record (NRS 631.025). The definitions of sedation in NRS are outdated compared to the ADA Guidelines and are not consistent. In October of 2007 these guidelines were adopted by the House of Delegates of the ADA. Mr. Drizin discussed the ADA Guidelines for sedation and any amendments that have taken place. One amendment was as recent as October 2011. The most recent change made by the Board to our regulations was twelve years ago. At the last meeting of the Anesthesia Committee this issue came up with regards to continuum to new techniques and medications. The American Academy of Pediatrics recognizes the continuum with regards to sedation in support of anesthesiology. Mr. Drizin would like to go forward with public comment. Dr Guillen agrees with Mr. Drizin to move forward and adopt the terminology of minimal and moderate and switch over
to be consistent with the ADA Guidelines. Dr Miller agrees as well to adopt the contemporary model. Dr Twesme made comment to let the legislature change the definitions regarding sedation. Dr Miller requested Mr. Drizin get a work group together to move forward.
No action taken
*5. Executive Director’s Report: (For Possible Action)
*a. Minutes: (For Possible Action)
(1) Board Meeting 3/22/2012
(2) Laser Working Group 3/22/2012
(3) Board Meeting 4/6/2012
(4) Board Meeting 4/18/2012
DISCUSSION: Ms. Kelly asked if there were any corrections or edits to the minutes. Dr Miller entertained a Motion for approval.
MOTION: Ms. Villigan made a motion to approve Second by Ms. Guillen. Public Comment. All in favor. Motion carries.
*b. Authorized Investigative Complaints (For Possible Action)
(1) Dr. Y - NRS 631.3475(1); NAC 631.230(1)(c)
Discussion: Ms. Kelly requested approval for an authorized investigation regarding Dr Y.
MOTION: Dr Sill made a motion to approve the authorized investigation. Second by Ms. Villigan. Public Comment. All in favor. Motion Carries
(2) Dr. Z - NRS 631.3475(1) and (4)
DISCUSSION: Ms. Kelly requested approval for an authorized investigation regarding Dr Z..
MOTION: Dr Pinther made a motion to approve the authorized investigation. Second by Mr. McKernan. Public Comment. All in favor. Motion Carries
*c. Review and Discuss Fiscal Year 2012 Closing, Financial Statements and Budget (For Possible Action)
DISCUSSION: Ms. Kelly stated the Board has been provided the FY12 closing financial statements and budget Ms. Hummel the Board’s Accountant is available for comment. These documents include ledger of accounts. Ms. Kelly is available for questions or comments to the financials. Ms. Wark has not had a chance to review since it was
MOTION: Ms Wark made a Motion to table this matter. Second by Dr Blasco. Public Comment All in favor Motion Passes Tabled Items C & D.
*d. Review and Discuss Fiscal Year 2013 Budget Review (For Possible Action)
This item was tabled.
*e. Review and Discuss Independent Contractors and Consultants Requirements with the State Board of Examiners (For Possible Action)
DISCUSSION: Ms. Kelly stated this item would affect DSO, IC Inspectors, and Anesthesia Evaluators. Due to some legislative changes, these individuals would be deemed independent contractors and may need to have contracts. This is an issue and wanted to make the Board fully aware. Ms. Kelly is working with Risk Management on this item.
No action taken.
*6. Board Counsel’s Report (For Possible Action)
*a. Legal Actions/Lawsuit(s) Update (For Possible Action)
DISCUSSION: Mr. Hunt updated the Board on the case regarding LVI. Mr. Hunt explained the legal process to the Board. The Writ of Mandamus has not been filed as of yet regarding a different case. In this case, Mr Hunt directed the Board not to discuss this matter with each other or anyone else and direct all inquiries to his office.
No action taken
*(1) Consideration of Stipulation Agreements (For Possible Action)
(a) Jarom N Luu, DDS
DISCUSSION: Mr Hunt went over the provisions of Dr Luu’s stipulation agreement. At the recommendation of Ted Twesme, DDS, Disciplinary Screening Officer assigned to this matter he offers this stipulation agreement for the Board’s approval. Ms. Wark asked about the outcome of the patient. Ms. Wark asked the status of the patient. Dr Twesme informed the Board the patient expired. Dr Miller asked for direction from Legal Counsel. Dr Miller asked if it was general anesthesia or conscious sedation. Dr Twesme stated general anesthesia. Mr. Thorsteinsson, Legal Counsel for Dr Luu was present. Dr Twesme stated the patient was on 25 different medications that were not disclosed. Ms. Villigan asked for Mr. Hunt to read the provisions of the stipulation agreement. Dr Blasco asked Dr Twesme if he feels this stipulation agreement is
addressing the issue with regards to patient safety. Mr. Hunt stated this was an authorized investigation. The Board did not receive a complaint.
MOTION: Dr Blasco made a motion to approve the stipulation agreement of Dr Jarom Luu. Second Ms. Wark. Public Comment All in Favor. Motion Carries
*(2) Todd J Krempel, DDS – Petition Pursuant to the Stipulation Agreement Approved on July 15, 2011 (Paragraph 7(A)(v) to Request the Board to Approve the Reinstatement of his License with the Department of Justice (DEA) and Nevada Controlled Substance Permit Subject to the Approval of DEA and the Nevada Pharmacy Board – Pursuant to NRS 241.030(a), the Board may, by motion, enter into closed session. (For Possible Action)
DISCUSSION: Mr. Hunt advised the Board of the request of Dr Todd Krempel pursuant to the terms of his stipulation agreement Paragraph 7(A)(v) to request the Board to approve the reinstatement of his license with the Department of Justice (DEA) and Nevada Controlled Substance Permit subject to the approval of DEA and the Pharmacy Board. Further, Mr. Hunt advised the Board Dr Krempel is in compliance with the terms and conditions of his stipulation as confirmed by Ms. Shaffer. Dr Krempel is currently drug tested randomly. Dr Miller requested how many times he has been drug tested, Ms. Shaffer indicted they are random tests and there have been six since July 15, 2011. Mr. McKernan asked if the random testing would continue. Pursuant to the agreement the testing will. Dr Krempel is present and stated he has no comment other than his compliance with the stipulation agreement.
MOTION: Dr Pinther made a motion to approve the reinstatement of Dr Krempel’s license with the Department of Justice (DEA) and Nevada Controlled Substance Permit subject to the approval of DEA and the Pharmacy Board. Second by Dr Kinard. Public Comment. All in Favor. Motion Carries.
*7. New Business (For Possible Action)
*a. Approval for Dental Licensure by WREB-NRS 631.240(1)(b)(2) (For Possible Action)
(1) Trevor C Basta, DMD
(2) Benjamin R Brooks, DDS
(3) Anne K Hoang, DDS
(4) Mickel Jourabchi, DDS
(5) David O Marchant, DDS
(6) Lindsey M Nelson, DMD
(7) Benita K Ng, DDS
(8) Darliece Nye, DMD
(9) Behnam Salar, DMD
(10) Jamie A Toop, DDS
(11) Demian S Woyciehowsky, DMD
DISCUSSION: Dr. Sill has reviewed the applications, found them to be in order and recommends approval of dental licensure by WREB for the following applicants: Dr Sill read each applicant's name into the record.
MOTION: Ms. Villigan made motion to approve the applications. Second by Dr Blasco. Public Comment. All in Favor. Motion Carries. Dr Sill Abstain
*b. Approval for Dental Licensure by ADEX – NRS 631.240(1)(b)(1) (For Possible Action)
(1) Timothy C Paventy, DMD
DISCUSSION: Dr. Sill has reviewed the applications, found them to be in order and recommends approval of dental licensure by ADEX for the following applicants: Dr Sill read the applicant's name into the record.
MOTION: Ms. Villigan made motion to approve the applications. Second by Ms. Guillen. Public Comment. All in Favor. Motion Carries. Dr Sill Abstain
*c. Approval for Specialty License by Application – NRS 631.250 (For Possible Action)
(1) Spencer C Fullmer, DDS - Endodontics
DISCUSSION: Dr. Sill has reviewed the applications, found them to be in order and recommends approval of dental licensure by Specialty for the following applicants: Dr Sill read the applicant's name into the record.
MOTION: Ms. Villigan made motion to approve the applications. Second by Mr. McKernan. Public Comment. All in Favor. Motion Carries. Dr Sill Abstain
*d. Approval for Specialty License by Credential – NRS 631.255(1)(a) (For Possible Action)
(1) John M Coats, DDS, MS - Endodontics
DISCUSSION: Dr. Sill has reviewed the applications, found them to be in order and recommends approval of dental licensure by Specialty for the following applicants: Dr Sill read the applicant's name into the record.
MOTION: Ms. Wark made motion to approve the applications. Second by Ms. Villigan. Public Comment. All in Favor. Motion Carries. Dr Sill Abstain
*e. Approval for Limited Dental License – NRS 631.271 (For Possible Action)
(1) Joffre J Martin, DMD
DISCUSSION: Dr. Sill has reviewed the applications, found them to be in order and recommends approval of limited dental licensure for the following applicants: Dr Sill read the applicants name into the record.
MOTION: Ms. Wark made motion to approve the applications. Second by Ms. Guillen. Public Comment. All in Favor Motion Carries. Dr Sill Abstain
*f. Approval for Dental Hygiene Licensure by WREB- NRS 631.300(1)(b)(3) (For Possible Action)
(1) Jamie J Burge, RDH
(2) Rolanda Chung, RDH
(3) Elizabeth Gomez, RDH
(4) Tami R Humphries, RDH
(5) Holli A Pace, RDH
DISCUSSION: Dr. Sill has reviewed the applications, found them to be in order and recommends approval of dental hygiene licensure by WREB for the following applicants: Dr Sill read each applicants name into the record.
MOTION: Ms. Villigan made motion to approve the applications. Second by Ms. Wark. Public Comment. All in Favor Motion Carries. Dr Sill Abstain
*g. Voluntary Surrender of License – NAC 631.160 (For Possible Action)
(1) David S Bogenschutz, DDS, MS
DISCUSSION: Ms. Kelly stated Dr Bogenschutz has submitted an application to surrender his orthodontic specialty license in the State of Nevada. Ms. Kelly believes he obtained licensure through the Gateway Program.
MOTION: DMr. McKernan made motion to approve the voluntary surrender. Second by Dr Pinther. Public Comment. All in Favor Motion Carries.
*h. Approval of Reactivation of Inactive License – NAC 631.170(3) (For Possible Action)
(1) David J Elkins, DDS
DISCUSSION: Ms. Kelly stated Dr Elkins has submitted his reactivation of his dental license. Dr Elkins has been out of practice for over one year his CE’s are in order.
MOTION: Ms. Villigan made motion to approve the reactivation of Dr Elkins license. Second by Dr Kinard. Public Comment. All in Favor Motion Carries.
*i. Approval for Anesthesia-Permanent Permit – NAC 631.2233 (For Possible Action)
*(1) General Anesthesia (For Possible Action)
a. Steve J Huang, DDS
b. Daniel F Muff, DDS, MD
DISCUSSION: Dr. Miller has reviewed the applications, found them to be in order and recommends approval for general anesthesia permits for the following applicants: Dr Miller read each applicant's name into the record.
MOTION: Mr. McKernan made motion to approve the general anesthesia permits for the above applicants. Second by Ms. Guillen. Public Comment. All in Favor Motion Carries.
*j. Approval for Anesthesia-Temporary Permit – NAC 631.2254 (For Possible Action)
*(1) Conscious Sedation (For Possible Action)
a. Blake K Wilde, DDS
DISCUSSION: Dr. Miller has reviewed the applications, found them to be in order and recommends approval for conscious sedation temporary permits for the following applicant: Dr Miller read each applicant's name into the record.
MOTION: Dr Pinther made motion to approve the conscious sedation temporary permit for Dr Wilde. Second by Ms. Wark. Public Comment. All in Favor Motion Carries.
*k. Approval for Site Permit – NAC 631.2236 (For Possible Action)
*(1) General Anesthesia (For Possible Action)
(a) Edilberto DeAndrade, DDS
1. 1140 N Town Center Drive, Suite 170, Las Vegas, NV 89144
(b) Jaren T Jensen, DDS
1. 9500 W Flamingo Road, Suite 200, Las Vegas, NV 89147
(c) Daniel F Muff, DDS, MD
1. 609 Sierra Rose Drive, Suite 1, Reno, NV 89511
(d) Douglas D Sandquist, DDS
1. 2650 Lake Sahara Drive, Suite 160, Las Vegas, NV 89117
DISCUSSION: Dr. Miller has reviewed the applications, found them to be in order and
recommends approval for site permits for general anesthesia for the following applicants: Dr Miller read each applicants name into the record.
MOTION: Dr Pinther made motion to approve the site permits for general anesthesia. Applicants were read into the record. Second by Ms. Wark. Public Comment. All in Favor Motion Carries.
*(2) Conscious Sedation (For Possible Action)
(a) Jeffrey A Cox, DDS
1. 4245 E Charleston Blvd, Las Vegas, NV 89104
(b) Cody C Hughes, DMD
1. 3310 N Moapa Valley Blvd., Logandale, NV 89021
(c) J Justin Wilkerson, DDS
1. 10455 Double R Blvd., Reno, NV 89521
DISCUSSION: Dr. Miller has reviewed the applications, found them to be in order and recommends approval for site permits for conscious sedation for the following applicants: Dr Miller read each applicants name into the record.
MOTION: Dr Pinther made motion to approve the site permits for conscious sedation. Applicants were read into the record. Second by Ms. Guillen. Public Comment. All in Favor Motion Carries.
*1. Approval For 90-Day Extension Of Anesthesia Permit – NAC 631.2254(2) (For Possible Action)
(1) General Anesthesia (For Possible Action)
a. Gary J Geracci, DDS
b. Scott M Redlinger, DMD
DISCUSSION: Dr. Miller has reviewed the applications, found them to be in order and recommends approval for 90 day extension of anesthesia permits for the following applicants: Dr Miller read each applicants name into the record.
MOTION: Mr McKernan made motion to approve the 90 day extension of anesthesia permits. Applicant names were read into the record. Second by Ms. Wark. Public Comment. All in Favor Motion Carries.
*m. Request for Advisory Opinion Pursuant to NAC 631.279 (For Possible Action)
(1) Margaret L Chesser, RDH – To market, demonstrate, and apply the use of Perioscopy Technology in dental offices under the supervision of a dentist.
DISCUSSION: The Board entertained Ms. Chesser’s request for an Advisory Opinion. Discussion was held.
MOTION: Dr Blasco made a motion that the use of a perioscope by a dental hygienist under supervision of the dentist is permissible. Second by Mr. McKernan (Please see the attached Advisory Opinion.)
*n. Review and Discussion, Including Collections and Costs, of the Nevada State Board of Dental Examiners Administration of the ADEX Clinical Examinations in Dentistry and Dental Hygiene Pursuant to NRS 631.170; NRS 631.190; NRS 631.240; NAC 631.090; NAC 631.100; NAC 631.120; NAC 631.130 (For Possible Action)
DISCUSSION: Dr Sill requested to review and discuss the administration of the ADEX Dental and Dental Hygiene Examination by the Nevada State Board of Dental Examiners. Dr Sill stated the Board was an independent board and is now associated with the North East Regional Examining Board. Dr Sill feels there is a financial cost to the Board and the burden on staff. Dr Sill requests that at the end of this examination cycle the Board discontinue the administration of the ADEX Examination for Dental and Dental Hygiene. Dr Sill went over the application fees comparison between WREB and ADEX and feels that applicants taken ADEX may have a financial incentive to choose one over the other since the cost of licensure is included in the cost of the ADEX exam. Ms. Kelly went over the cost associated with administering the ADEX examination for Dental and Dental Hygiene. Ms. Kelly’s breakdown is using last year’s figures. Ms. Kelly informed the Board at the end of the examination cycle the Board has made a profit of $1,500.00. Ms. Kelly math error the Board has a positive revenue of $2200.00. Dr Miller stated the fiscal impact is the Board is not losing money. Dr Miller went onto to state whether the Board feels the best avenue is assess a candidates abilities the clinical examination is something you may want to be a part of. Dr Miller would like to continue with the Board administering the examination. Dr Sill stated the Board would have the revenue if the applicants apply for licensure and the board would have received those fees. Ms. Kelly commented the 10 of the traditional applicants were not seeking licensure and therefore the Board received the fees of $15,000.00. Ms. Kelly read NRS 631.190 into the record. Dr Sill stated the Board can have NERB administer the ADEX examination and Board Members. Ms. Kelly went through the candidate numbers.
PUBLIC COMMENT: Dr Guillen commented the Board is cutting it down to money. Nevada is not a huge state and the Board shows good PR with legislators and the dental students it should not be about money. Just because you are an examiner and enter the pool you may not get to examine in Nevada. Dr Guillen thinks the Board is taking the wrong approach. Dr Guillen believes we should continue administering the ADEX exam. Dr Pappas by means of disclosure is the Secretary Treasurer for ADEX Executive Council but he is here today as a past Board Member. Dr Pappas stated for National Recognition Nevada is the only state that administers the ADEX Examination.
This is a unique role nationally and the Board would be forfeiting this ability and if the Board gives this up we are just like any other states. Discontinuing this would be a shame for our state on the national level. Ms. Peterson concurs with Dr Pappas agrees with continuing the administration of this examination. Ms. Peterson thinks this is a big opportunity for the Board nationally. Dr Kinard asked Ms. Kelly about loss revenue if we do not administer the examination. Dr Kinard asked the number of examinations the Board gives. Ms. Kelly stated 4 per year. Ms. Wark requests a breakdown of all expenses and costs before making a decision. Ms. Kelly stated she can prepare an excel spreadsheet based on last year numbers and breakdown the costs. Ms. Kelly stated it is difficult to provide this information based on Nevada Board Examiners versus NERB Examiners. Dr Miller stated money aside the Board needs to look at the bigger picture and can the board be a leader in this movement.
No action taken
8. Resource Group Reports
*a. Legislative and Dental Practice* (For Possible Action)
(Chair: Dr. Pinther; Dr. Champagne; Dr. Blasco; Dr. Kinard; Mrs. Guillen, Mr. McKernan; Mrs. Wark)
DISCUSSION: No report
*b. Legal and Disciplinary Action* (For Possible Action)
(Chair: Dr. Kinard; Dr. Soltani; Dr. Sill; Dr. Blasco; Mrs. Villigan; Mr. McKernan; Mrs. Wark)
*(1) Consideration of Adoption of Legal and Disciplinary Action Resource Group’s Recommendations approved on June 28, 2012.* (For Possible Action)
MOTION: Dr Kinard made a motion to table this item. Second by Dr Blasco, Public Comment All in favor Motion Passes
*c. Examinations*
(1) *Dental* (For Possible Action)
(Dr. Miller; Dr. Kinard and Mrs. Guillen)
DISCUSSION: Ms. Kelly provided the number of candidates taking the Endo-Pros portion of the ADEX CIF Format.
(2) *Dental Hygiene* (For Possible Action)
DISCUSSION: No report
(Chair: Mr. McKernan; Mrs. Guillen; Mrs. Villigan; Dr. Pinther)
* d. **Continuing Education** (For Possible Action)
(Chair: Dr. Sill; Dr. Blasco; Dr. Soltani; Mrs. Villigan)
DISCUSSION: No report
* e. **Dental Hygiene** (For Possible Action)
(Chair: Mr. McKernan; Mrs. Guillen; Mrs. Villigan; and Dr. Sill)
DISCUSSION: Mr McKernan stated at the last meeting the Dental Hygiene Association brought items forward for consideration. The Board will conduct a meeting on November 30th to take up these items.
* f. **Specialty** (For Possible Action)
(Chair: Dr. Soltani; Dr. Miller; Dr. Pinther)
DISCUSSION: No report
* g. **Anesthesia** (For Possible Action)
(Chair: Dr. Miller; Dr. Pinther; Dr. Champagne, Dr. Kinard)
DISCUSSION: The Anesthesia Committee is working on the regulations.
* h. **Infection Control** (For Possible Action)
(Chair: Mrs. Villigan; Dr. Blasco; Dr. Champagne; Dr. Soltani; Mr. McKernan; Mrs. Wark)
DISCUSSION: No report
9. **Public Comment**: Ms. Peterson stated licensees who are requesting from inactive to active status with regards to requiring skills assessments these licensees should be required to comply whether a dentist or dental hygienist. Ms. Kelly went through the process outlined in NAC 631.170. Ms Kelly asked Mr. Hunt if he had anything to add. Mr. Hunt stated the regulation speaks for itself. He stated Ms. Peterson is raising the issue that all licensees the Board is consistent. Dr Miller stated that perspective is important. Ms. Rogers asked of Ms. Villigan when the Infection Control Form would be available of the website. Ms. Kelly will check on that for Ms. Rogers.
Note: No vote may be taken upon a matter raised under this item of the agenda until the matter itself has been specifically included on an agenda as an item upon which action may be taken. (NRS 241.020)
10. **Announcements**
No comment
*11. **Adjournment** (For Possible Action)*
MOTION: Dr Blasco motion to adjourn Second by Ms. Wark. All in Favor Motion Carries.
Respectfully Submitted by
______________________________________
Debra Shaffer, Executive Director
BEFORE THE NEVADA STATE BOARD OF DENTAL EXAMINERS
IN REGARDS TO THE MATTER OF:
Request for the Advisory Opinion regarding clarification to market, demonstrate, and apply the use of Perioscopy technology in dental offices under the supervision of a dentist.
AO-12-0920
On September 20, 2012, pursuant to agenda item 7m at a properly noticed meeting held at the Board office located at 6010 South Rainbow Boulevard, Suite A-1, Las Vegas, Nevada 89118 the Nevada State Board of Dental Examiners (the "Board") issued the following advisory opinion.
Dr. Jade A Miller Present
Dr. M. Masih Soltani Present
Dr. Jason Champagne Present
Dr. Timothy Pinther Present
Dr. James G Kinard Present
Dr. J. Stephen Sill Present
Dr. Byron Blasco Present
Mr. James McKernan Present
Mrs. Leslea Villigan Present
Mrs. Theresa Guillen Present
Mrs. Lisa Wark Present
ADVISORY OPINION
I.
BACKGROUND
1. As set forth in NAC 631.279 through authority of NRS 631.190 and NRS 233B.120 the Board is authorized to provide advisory opinions.
NAC 631.279 Proceedings to determine applicability and construction of statutes and regulations. (NRS 631.190)
1. Any applicant or licensed dentist or dental hygienist may obtain a determination or advisory opinion from the Board as to the applicability of any provision of chapter 631 of NRS or any regulation adopted pursuant thereto by bringing an action for a declaratory judgment before the Board.
2. The Board will construe any statute or regulation reviewed pursuant to this section in a manner consistent with the declared policy of the State of Nevada.
2. Nevada licensee, Margaret Chesser, RDH, requested an advisory opinion for clarification to market, demonstrate, and apply the use of Perioscopy technology in dental offices under the supervision of a dentist.
2. Public notice of the above-referenced request for an advisory opinion was provided in accordance with state law.
II.
DISCUSSION
Mr. Hunt addressed Ms. Chesser regarding the request for an advisory opinion, attached letter and advertisement regarding the use of the perioscope. Ms. Chesser advised Mr. Hunt her request was not with regards to the perioscope. Ms. Chesser explained the perioscope goes below the gum-line and is a camera that magnifies up to forty eight times and is used in dental offices for root surfaces of the teeth allowing her to see much better when root planning and scaling as opposed to the blind eye. Ms. Chesser stated she was requesting the advisory opinion due to the concerns by dentist in the community that contacted the Board when Ms. Chesser began going to dental offices to offer this perioscope service. Ms. Chesser came to the Board office and spoke with Ms. Debra Shaffer to see what she needed to do to ensure compliance. Ms. Shaffer suggested Ms. Chesser request an advisory opinion.
Mr. Hunt stated as long the Board has no objection to the use of the device as to the other items Ms. Chesser understands she is still required to be under the supervision of a dentist and this cannot be a fee splitting arrangement. Based
upon the letter, Ms. Chesser is not asking for an advisory opinion on those items since Ms. Chesser is already aware of what is required by statute and regulation. The question that is presented is Ms. Chesser is asking for an advisory opinion of the use of the perioscope in the execution of her duties as a dental hygienist.
Ms. Chesser explained she will not be diagnosing that will be the dentist. Ms. Chesser is there to help assist the dentist in his/her diagnoses, Ms. Chesser will not be fee splitting with the dentist she is now aware this is not permissible and will not be fee splitting and she has taken care of these items referenced in her letter.
Mr. Hunt stated the Board may entertain a motion to give an advisory opinion that the perioscope is within the duties of dental hygiene under the supervision of a dentist.
Dr. Miller acknowledged Ms. Chesser for being proactive with new technology.
III.
CONCLUSION
After considering and discussing public comment pursuant to the authority set forth in NAC 631.279, NRS 631.190 and NRS 233B.120, a motion was made by Dr. Blasco to approve the request for the opinion that the use of the perioscope by a dental hygienist is permissible under supervision of a dentist.
Ms. Villigan had questions for Ms. Chesser related to infection control. Ms. Chesser indicated every patient receives a new sheath to cover the fiber. Ms. Chesser stated she has multiples of everything including the perioscope she has two. Instruments are left at the practice where treatment is rendered for sterilization and normal protocols are followed.
Dr. Miller called for the motion, motion by Dr Blasco seconded by Mr. McKernan.
| Name | Vote |
|-----------------------------|------|
| Dr. Jade A Miller | yes |
| Dr. M. Masih Soltani | yes |
| Dr. Jason Champagne | yes |
| Dr. Timothy Pinther | yes |
| Dr. James G Kinard | yes |
| Dr. J. Stephen Sill | yes |
| Dr. Byron Blasco | yes |
Mr. James McKernan yes
Mrs. Leslea Villigan yes
Mrs. Theresa Guillen yes
Mrs. Lisa Wark yes
The motion was passed unanimously the use of a perioscope by a dental hygienist under supervision of the dentist is permissible.
JADE A. MILLER, DDS, PRESIDENT
Nevada State Board of Dental Examiners |
245.09 B36
Beattie 6s.
Romance of sacred song.
Keep Your Card in This Pocket
Books will be issued only on presentation of proper library cards.
Unless labeled otherwise, books may be retained for two weeks. Borrowers finding books marked, defaced or mutilated are expected to report same at library desk; otherwise the last borrower will be held responsible for all imperfections discovered.
The card holder is responsible for all books drawn on this card.
Penalty for over-due books 2c a day plus cost of notices.
Lost cards and change of residence must be reported promptly.
Public Library
Kansas City, Mo.
Keep Your Card in This Pocket
BERKOWITZ ENVELOPE CO., K. C., MO.
KANSAS CITY, MO PUBLIC LIBRARY
0 0001 0293593 9
THE
ROMANCE OF SACRED SONG
NO KANSAS CITY LIBRARY
LIBRARY CARD APPLICATION
The following is a list of the most common types of software that are used in the field of computer science:
1. Operating Systems: These are the programs that manage the hardware and software resources of a computer system.
2. Programming Languages: These are the languages used to write computer programs.
3. Database Management Systems: These are the programs that manage the storage, retrieval, and manipulation of data.
4. Web Development Tools: These are the tools used to create websites and web applications.
5. Software Development Tools: These are the tools used to develop software applications.
6. Data Analysis Tools: These are the tools used to analyze data and make decisions based on the analysis.
7. Project Management Tools: These are the tools used to manage projects and ensure that they are completed on time and within budget.
8. Collaboration Tools: These are the tools used to collaborate with others on projects and share information.
9. Security Tools: These are the tools used to protect computer systems from unauthorized access and attacks.
10. Virtualization Tools: These are the tools used to create virtual machines and run multiple operating systems on a single physical machine.
Frontispiece.
Photo: Frith, Brighton.
HENRY F. LYTE.
By permission.
S. BARING-GOULD.
From portrait by Melton Fisher, R.A.
Berry Head House.
Where "Abide with me" was written.
The ROMANCE of SACRED SONG
By
DAVID J. BEATTIE
Author of
"Songs of the King's Highway,"
etc.
MARSHALL, MORGAN & SCOTT, LTD.
LONDON AND EDINBURGH.
YOUNG ALICE
VOL 2
Made and printed in Great Britain
by
Hunt, Barnard & Co., Ltd.
London & Aylesbury.
1783.909
r
BIND MAY '42
TO
MY MOTHER
FROM WHOSE LIPS
I FIRST HEARD
THESE SONGS
PREFACE
Many years' diligent pursuit along the highways and byways of the realm of Sacred Song produced an accumulation of material, the collecting and tabulating of which was ever a delight and labour of love. To disintegrate, arrange and marshal names, dates and facts presented a formidable task. My bookshelves already groaned under the weight of volumes on this engrossing subject. And yet I was sanguine enough to believe that out of the pile of hymnal notes before me, it was possible to produce something which would not only reach the heart of the hymn lover, but would afford some hours of helpful, and, I trust, pleasurable reading to whoever might peruse these pages. In launching out, therefore, my endeavour has been, in the first place, to interest readers in the authorship and history of the sacred songs they sing.
An attempt has also been made to give an account, not only of the most prominent hymns, but others not so eminent in the realm of hymnody, whose inclusion, because of some special circumstance associated with them, seemed to be desirable. Some hymns which have received prominence may fall far below the standard of first class poetry, and may lay no claim to literary merit, other than that which attaches to hymns which have a well-attested value as having been a channel of blessing in their day and generation. This is my only apology for their inclusion in this volume.
No efforts have been spared in order to verify much hymnal data which came into my hands, and so far as possible I have gone to original sources for the desired information. An intimate acquaintance with many authorities on the other side of the Atlantic has enabled me to set down much fresh material, as well as rendering assistance in the weaving of many little-known
PREFACE
stories around some of the old songs of the great revival days of a past generation.
My cordial thanks are due to those who so courteously answered my many enquiries, and have assisted in my research in furnishing me with interesting particulars and illustrations. Lastly, special mention should be made of the valuable work of my daughter Elsie, who typed the whole of the manuscript, and to my friend Mr. J. Duckworth, B.A., who so ably assisted in the correction of proofs.
D. J. B.
Carlisle, 1931.
| CHAP. | TITLE | PAGE |
|-------|--------------------------------------------|------|
| I. | EARLY DAYS OF SACRED SONG | 13 |
| II. | PIONEERS OF ENGLISH HYMNODY | 23 |
| III. | SOME LADY HYMN WRITERS | 46 |
| IV. | HOW SOME FAMOUS HYMNS WERE WRITTEN | 64 |
| V. | THE POWER OF SACRED SONG | 92 |
| VI. | HYMNS SUNG IN TRAGIC MOMENTS | 113 |
| VII. | CHILDREN’S SONGS | 128 |
| VIII. | HOW WE GOT OUR POPULAR GOSPEL SONG | 145 |
| IX. | HYMN WRITERS OF AMERICA | 154 |
| X. | SACRED SONG COMPOSERS | 177 |
| XI. | POPULAR WRITERS OF TO-DAY | 201 |
| XII. | HYMNS THAT HAVE HELPED | 214 |
| | INDEX OF AUTHORS’ NAMES | 230 |
LIST OF ILLUSTRATIONS
PLATE I, Frontispiece.
S. Baring-Gould.
Berry Head House.
Henry F. Lyte.
PLATE II, facing page 42.
Dr. Horatius Bonar.
Facsimile of "I heard the Voice of Jesus say."
PLATE III, facing page 66.
Rock of Ages, Burrington Combe.
PLATE IV, facing page 76.
Facsimile of "Christians Awake!"
PLATE V, facing page 148.
Major W. D. Whittle, W. J. Kirkpatrick.
P. P. Bliss, C. H. Gabriel, J. R. Sweeney,
J. McGranahan.
PLATE VI, facing page 158.
Ira D. Sankey and Fanny Crosby.
PLATE VII, facing page 84.
Robert Lowry, George C. Stebbins, Philip Phillips.
PLATE VIII, facing page 206.
R. H. Daniel, Jessie Brown Pounds, Johnson
Oatman, Jnr.; J. W. Van de Venter, Annie
Sherwood Hawks, Ada R. Habershon, Ina
Dudley Ogdon.
CHAPTER I.
Early Days of Sacred Song
IN these days, when the hymn book plays so important a part in every place of worship, stories of hymns and their writers is a subject of perennial interest. Nor is it difficult to understand how the memory retains the words and tunes of hymns which were learned at a mother's knee or at the Sunday School. Who is there amongst us who has not at some time or other experienced the strange and subtle influence of sacred song, an influence which compelled the tears to come unbidden; burning tears of joy, sorrow, remorse or peace that come with the strain of some old and almost forgotten hymn? If by reason of some disappointment the heart aches; if it yearns "for the touch of a vanished hand and the sound of a voice that is still"; if it throbs at the fancied footstep that comes no more, what can soothe and comfort like "Jesus, Lover of my soul"?
How often as we travel on the journey of life are we suddenly arrested by the strains of some familiar hymn wafted to us, recalling memories of the days of long ago, and portraying on one's vision some sacred scene from which many of the actors have passed away.
To many of us, hymns have proved a never-failing solace, an oasis in a parched and thirsty land, a drink from the trough on a dry and dusty day. In the storm and stress of life's battle, the echo of their sweet refrain has renewed our strength and dispelled our fears.
Around the hymn and hymn-tune who can tell how many cherished associations gather from the earliest days? And though few seek to know the origin or history of the hymns that please them, the telling of the tale never fails to add to their attraction.
There is a fascinating power in the singing of a hymn which can do God’s work in a soul when every other instrumentality has failed. There was not always so much freedom in this respect in the days of our grandfathers. The opposition to hymn singing was widespread throughout Scotland, and Sankey tells how, when he commenced singing a solo in the Free Assembly Hall, Edinburgh, during the great mission there in 1873, a woman’s shrill voice was heard in the gallery, as she made her way toward the door, crying: “Let me oot! Let me oot! What would John Knox say to the like o’ yon?”
In his delightfully written life story, where he relates his varied experience during his first campaign amongst the good Scottish folks, Sankey does not hide the fact that he was not a little perturbed regarding the question of solo singing, as its propriety and usefulness was not yet fully understood or admitted. As he took his seat at the instrument at one of the first meetings held in Edinburgh, Sankey discovered to his surprise that Dr. Horatius Bonar was seated close by the organ, right in front of the pulpit. “Of all men in Scotland,” says Sankey, “he was the one man concerning whose decision I was most solicitous. He was, indeed, my ideal hymn-writer, the prince among hymnists of his day and generation. And yet he would not sing one of his own beautiful hymns in his own congregation, such as ‘I heard the voice of Jesus say,’ or ‘I was a wandering sheep,’ because he ministered to a church that believed in the singing of Psalms only. With fear and trembling, therefore, I announced as a
solo the hymn 'Free from the law, O happy condition.' " No prayer having been offered for this part of the service, and fearful lest the singing might prove only an entertainment, instead of spiritual blessing, Sankey requested that the whole congregation should join him in a word of prayer asking God's blessing on the truth about to be sung. It was a time of tense solemnity, but the anxiety of the moment was relieved, and believing and rejoicing in the glorious truth contained in the hymn, he sang it through to the end, amid a reverent silence never before experienced.
At the close of Mr. Moody's address, Dr. Bonar turned toward the American singer with a smile on his venerable face, and reaching out his hand, he said: "Well, Mr. Sankey, you sang the Gospel to-night." Thus the way was wonderfully opened up for the mission of sacred song in Scotland, a particular sphere of ministry in which so much has been accomplished since that memorable hour.
With the introduction of hymns for general congregational use, the worthy precentor with his pitch-pipe and tuning fork was obliged to relinquish his position of importance and dignity, to give place to the organ—or what our good Scottish parents used satirically to refer to as the "kist o' whistles." A story is told of the precentor of a certain Scottish kirk, who had purchased a new pitch-pipe, but when the time came for leading the praise, he was unable to bring the instrument into action. He tugged at it—thrust it in—tried to pull it out—gave it a thump—grinned and pulled again—but budge it would not. The minister grew impatient, and leaning over the pulpit, whispered to the precentor, "Stop, Jonathan"; then aloud to the congregation, he said, "Let us pray." By this time the unfortunate precentor had become exasperated, and still struggling with the obstinate instrument,
cried out, "Pray, did ye say? We'll pray nane till I get this thing tae work!"
But while we rejoice at the revival of this God-sent ministry of sacred song, so strikingly evinced in the past generation and during the present—a channel which has brought in its train countless blessings—we are reminded that, in the same way, under similar circumstances, God mightily used the power of sacred song in days of old, with no less wonderful results.
Away back in the sixteenth century, Martin Luther, the author of many choice hymns which have been sung down through the centuries, well understood this method of propagating the truth, and employed it with a skilful hand. His own poetical talents and love of music were very great, and when, as a wandering minstrel, he earned his daily bread by exercising his musical powers in singing before the doors of the rich in the streets of Magdeburg and Eisenach, he was as truly preparing for the future reformer as when, a retired monk in the cloister of Erfurt, he was storing his mind with the truths of revelation, with which to refute the errors of popery. A few sentences from a preface which Luther wrote to a collection of his own hymns, published in 1524, all of which were set to music in four parts, is worthy of note. He tells us that this had been done, "for no other reason than because of my desire that the young, who ought to be educated in music as well as in other good arts, might have something to take the place of worldly and amorous songs, and so learn something useful and practise something virtuous, as becometh the young. I would be glad to see all arts, and especially music, employed in the service of Him who created them."
How God signally used this converted monk is familiar history. "The whole people," wrote a Romanist of that day, "is singing itself into this Lutheran
EARLY DAYS OF SACRED SONG
doctrine.” It is said that Luther accomplished more in setting all Germany singing, than he did with his preaching. The Church of Rome became alarmed, for they well knew that the pure Gospel would be sung unto many who could never have been prevailed upon to hear it any other way.
As we have seen, the Reformation movement in Germany was marked by a great outburst of hymnody. In Britain, however, Protestantism found its vehicle of praise in the metrical version of the Psalms. This was due, no doubt, to the influence of Calvin and the Genevan school, who held as a principle that the Word of God should have supreme dominion in public worship, and that no production of man should be allowed to take its place.
The prejudice against the use of hymns in favour of the Psalms, especially amongst the staid Scottish folk, remained with many of the older people till the end of their days. In proof of this, the following incident, which came under my personal observation, will, no doubt, amply suffice. Old Betsy lay dying, and was visited by her friend Malcolm Ferguson, a local evangelist. After spending a time with the aged saint, whom he sought to comfort and encourage, he enquired whether she would like him to sing a hymn. “Na, na,” was the quick reply, “nane o’ yer human hymns for me. What’s wrang wi’ the Psalms o’ Dauvit? I expect, Malcolm, when ye get to heaven, ye’ll gang clankin’ straight for Sankey; but Dauvit’s my man.”
Early English hymns are not numerous, and such as exist were written for private edification rather than for public use.
Isaac Watts (1674-1748) has been rightly looked upon as the father of English hymnody. Thirty-seven years before the birth of Watts, Thomas Ken was born.
THE ROMANCE OF SACRED SONG
A remarkable coincidence is that Watts lived as long after Ken died in 1711, as Ken had preceded him in beginning life. Bishop Ken will always be remembered for his immortal lines in the form of our most used doxology:
"Praise God from whom all blessings flow,
Praise Him all creatures here below,
Praise Him above, ye heavenly host,
Praise Father, Son, and Holy Ghost."
—which, doubtless, has been repeatedly sung by more people than the words of any other writer. In addition to his doxology and other works, Ken has given us two matchless hymns for which they are designed—one for a morning hymn, and the other for an evening hymn—"Awake my soul," and "Glory to Thee, my God, this night." But pioneers are not always perfect in their methods. Sculptors finish the work begun by the men far away in the quarries, who split up the rocks on the mountain side. Ken perceived the need of a new hymnology, and pointed to the way in which Watts and Wesley in later years walked.
Previous to the coming of Moody and Sankey to Britain, the early nineteenth century gave unmistakable evidence of a new birth in the field of hymnody. People were really beginning to take an interest in the subject, and it was evident that hymns were becoming an indispensable part of every form of religious services. Accordingly we find a host of writers pouring forth hymns. Their merit, of course, varied greatly. Here and there we find a writer with a really poetical mind, whose soul breathings have powerfully enriched our hymnology. There were also many others who had been inspired to pen some very beautiful hymns; but there were also many hymns, which, from a literary and theological standpoint could not be placed on a
very high plane. Still, at that period, it was early to discriminate; nobody could say which of these would survive and which would not; only time could reveal that.
In taking a glance at our hymns from a more general point of view, the first thing that strikes the observer is how extraordinarily cosmopolitan compilers of hymn books had been in the sources from which they had drawn their supplies. And here attention should be drawn to the debt we owe to the undefatigable zeal of translators. The Oxford movement of 1833, led men to investigate the old Greek and Latin hymns, and naturally, they went on to translate them. The prince of these translators was J. M. Neale, and perhaps the finest work which he did was the translation of the Rhythm of Bernard of Morlaix, and from which modern hymn collectors had extracted among others the old favourite "Jerusalem the Golden." Besides this, Latin hymnody has furnished us with: "O come all ye faithful," "O come, O come, Immanuel," "Ye choirs of New Jerusalem," and a host of others. "Hail gladdening light," "O happy band of pilgrims," and "Art thou weary?" are Greek; "Guide me, O Thou great Jehovah," is Welsh, "Through the night of doubt and sorrow," is Danish; and "Glory be to Jesus," Italian.
And we must not forget what we owe to America who gave to us our popular Gospel song. Indeed, it is remarkable how many nations have joined together to make up the verses which to-day are sung in church and chapel, mission hall and cottage meeting, as well as at tent and open-air services.
An outstanding feature in the realm of hymnody which cannot escape observation is that public favour has been phenomenally capricious with regard to its taste in hymns. Nobody could say how or why one
hymn became part of a national collection, while others fell flat. Kirke White was immortalised by "Oft in danger," Harriet Auber by "Our blest Redeemer," Sarah Adams by "Nearer my God to Thee," Edward Perronet by "All hail the power of Jesus' name," but there are few who know anything else that any of them ever wrote. Every one is familiar with "Rock of Ages," but though a few other hymns by Toplady were sometimes reprinted, nobody cared for them.
Then in regard to what may be termed our famous hymns, it is worthy of note that these are not the privilege of any one particular religious body, but are, in the best sense of the word, the common property of nearly all English speaking Christians. To meet their own particular requirements, each religious body has its own hymn book, and, of course, each hymn book has its own peculiar hymns. It has been said that Ridley and Latimer, who quarrelled about vestments, agreed at the stake. We live in happier times, where Christians who differ in more important matters can still agree in their hymns of prayer and their songs of praise. The productions of Ken and Heber, of Wesley and Toplady, of Doddridge and Kelly, of Cowper and Newton, of Fanny Crosby and Frances Ridley Havergal, all these harmoniously combine, for singing in company we at once forget the non-essentials on which we may differ, and remember only the desire for holiness, the enthusiasm for righteousness, the thankfulness for free unmerited favour, and the love for our blessed Lord in which we all agree.
No other literary composition is like a hymn. It is not a mere poetic impulse. It is not a thought, a passing fancy, or a feeling threaded upon words. It is the voice speaking from the soul a few words that often represent a whole life. Hymns and spiritual songs have indeed wielded a powerful influence in almost every
walk of life, at one period or another, through countless ages.
All history carries the echo of music. At an early period Jubal became a manufacturer of musical instruments, and thus the first instruments mentioned in the Bible are the harp and organ (Gen. 4:21). Before his day, doubtless, many an ingenious amateur picked tones from a string or made the requisite vibrations in tubes of resonant bodies. The toph or tambourine—known to the Egyptians and Assyrians—was used by Miriam and Jephthah’s daughter.
"Not only is inanimate nature full of music," says an eminent writer, "but God has wonderfully organised the human voice, so that in the plainest throat and lungs there are fourteen direct muscles which can make over sixteen thousand different sounds. Now, there are thirty indirect muscles which can make, it has been estimated, more than one hundred and seventy-three millions of sounds. Now, I say, when God has so constructed the human voice, and when He has filled the whole earth with harmony, and when He recognised it in the ancient temple, I have a right to come to the conclusion that God loves music."
Thus, turning back the pages of history, we learn that Miriam led the hosts in a hymn of praise on the farther banks of the Red Sea. Moses closed his far-reaching career with a great song. Deborah celebrated the victory over Israel’s enemies in a hymn of thanksgiving. Jehoshaphat when called upon to meet a vast multitude of enemies, which threatened to overrun his entire kingdom, placed singers in front of his army to lead the march, singing the praises of Israel’s God.
"There has been much discussion as to where music was born," once said the late Dr. Talmage. "I think that at the beginning, when the morning stars sang together and all the sons of God shouted for joy,
the earth heard the echo. The cloud on which the angels stood to celebrate the creation was the birthplace of song. The stars that glitter at night are only so many keys of celestial pearl on which God’s fingers play the music of the spheres. Inanimate nature is full of God’s stringed and wind instruments. Silence itself—perfect silence—is only a musical rest in God’s great anthem of worship. Wind among the leaves, insect humming in the summer air, the rush of billows upon the beach, the ocean far out sounding its everlasting psalm, the quail whistling up from the grass, are music.”
David, the sweet singer of Israel, with his psalm and harp, is the poet and musician of the Bible, and surely nothing in the whole record of history can be found to compare with the majesty and magnificence of these heavenly songs which are still singing their way through the world.
The angel choir celebrated the birth of the Infant Saviour, to the astonished shepherds on Bethlehem’s plain. It was while Paul and Silas, lacerated and wounded by the cruel scourging which they had received, sang praises to God at midnight, that the Philippian jail was rent asunder, and the jailor and all his household were converted that night. The use of “psalms, and hymns, and spiritual songs,” was enjoined upon the churches at Colosse and Ephesus. Many evidences are furnished us, too, that in private, as well as in public, the first Christians were warmly attached to singing the praise of God. All through the Scripture we are commanded to sing unto the Lord, to praise Him with the harp, trumpet, and organ; and to speak in psalms, and hymns, and spiritual songs. Finally, the Book closes with the pearly gates of heaven left ajar, through which there floats out upon us the voice of harpers, harping with their harps and singing the new song of Moses and the Lamb.
CHAPTER II.
Pioneers of English Hymnody
ISAAC WATTS, the author of the immortal hymn, "O God, our help in ages past," is rightly regarded as the father of English hymnody, for, prior to his time, hymns were rarely sung in public worship. Up to that period practically the only vehicle of praise in the English and Scottish churches took the form of very crude versions of the Psalms, while in some non-conformist congregations there was no singing at all. A story is told of young Watts, on coming home from chapel one Sunday, complaining to his father, who was one of the deacons, that the psalmody in use at the Congregational Chapel at Southampton did not possess the dignity and beauty becoming a Christian service.
"Then give us something better, young man!" was his father's ironical reply.
Isaac determined to do so, and the following Sunday arrived at the chapel with his first hymn: a hymn eminently appropriate, and which to-day, after the passing of two centuries and more, has lost none of its power and beauty:
"Behold the glories of the Lamb
Amidst His Father's throne;
Prepare new honours for His Name,
And songs before unknown."
The youthful hymn writer's initial attempt was so favourably received that he was requested to provide another for the next Sunday. A chord had been
struck which gave birth to new and delightful harmonies in English hymnody. Thus, bringing a fresh hymn to the chapel each Sunday, until almost a volume was produced, began the reputation as a hymnist of the famous Dr. Watts. He was the eldest of nine children, and was born on July 17th, 1674, at Southampton, where his father kept a flourishing boarding-school. Isaac Watts, senior, lived through the stormy days of early nonconformity, and because of his religious convictions was twice thrown into prison. While there, the infant Isaac was often brought to the outside of the prison by his mother, who would sit for hours on the stone by the gate. As he grew into boyhood Isaac did not enjoy robust health, and being of a studious nature he was apt at times to overtax his strength. On his twenty-fourth birthday Watts preached his first sermon, and two years later was appointed assistant minister of the famous Mark Lane Independent Church, London, subsequently becoming sole pastor. Ill-health, however, compelled him to give up all regular pastoral duties, and on the invitation of Sir Thomas and Lady Abney he went for a week's visit to their residence in Hertfordshire—a visit which resulted in a stay of thirty years. It was during this happy period of Isaac Watts' life that most of his hymns were written. He died on November 25th, 1748, aged seventy-five, and was buried in the Puritan resting-place at Bunhill Fields. Later, a monument was erected to his memory in Westminster Abbey, near to that of John and Charles Wesley.
Altogether, Watts wrote over six hundred hymns, in the composition of which, there was ever present an ardent desire to express in simple yet forceful language, the aspirations of the heart towards God. "I make no pretence to be a poet," he says, "but to the Lamb that was slain, and now lives, I have
addressed many a song, to be sung by the penitent and believing heart."
Sweetest amongst his compositions, and pronounced by critics as the finest in the English language, is the universally loved hymn, "When I survey the wondrous cross." Here Watts strikes his highest note as he dwells in tender and solemn reverence on the all-absorbing theme of God's wonderful redemption. Where, in the whole range of hymnody, could be found a verse to compare with the following lines?
"See! from His head, His hands, His feet,
Sorrow and love flow mingled down;
Did e'er such love and sorrow meet,
Or thorns compose so rich a crown?"
A missionary hymn which vies in popularity with that of Bishop Heber, is Watts'—
"Jesus shall reign where'er the sun
Doth his successive journeys run."
It is said that the majestic tune to which "O God, our help in ages past" is sung the world over, known as "St. Anne," was composed so far back as 1687; more than a generation before the hymn was written. Thus, it is quite probable that the hymn was sung to the same old tune in the days when Isaac Watts himself joined in the singing.
Other compositions of his, which have stood the test of more than two hundred years, are, "I'm not ashamed to own my Lord," "Not all the blood of beasts," "Come let us join our cheerful songs," and "There is a land of pure delight." The latter hymn, it is said, was suggested to Watts as he sat at his parlour window at Southampton, and looked out upon the waters of the Itchen, with the Isle of Wight in the distance, and the beautiful landscape stretching far away on the other side of the river. Thoughts of
THE ROMANCE OF SACRED SONG
Canaan, and the Jordan, and the glories of the eternal home beyond, illumed the spiritual vision of the great hymnist, as from his heart poured forth the words:
"There everlasting spring abides,
And never-with'ring flowers;
Death, like a narrow sea, divides
This heavenly land from ours."
It is related that, when slighted by the maiden of his choice, who, in repulsing his addresses, told the young minister that although she admired the jewel she could not endure the casket, Watts had some consolation in giving expression to his injured feelings in a hymn commencing:
"How vain are all things here below,
How false and yet how fair!
Each pleasure hath its poison too,
And every sweet a snare."
Watts was never married, and yet he is a child's hymnist. Though present day hymnals contain few of his old-time compositions, still, around these simple stanzas, which, two or three generations ago never failed to sway the feelings of the child, there clings a halo of happy memories, drawing us nearer, and still nearer, to the One of Whom these little hymns taught us to sing.
When Dr. Watts was thirty years old, there was born in London on June 26th, 1702, Philip Doddridge, the author of the stirring Advent hymn:
"Hark, the glad sound! the Saviour comes,
The Saviour promised long;
Let every heart exult with joy,
And every voice be song."
In later years, despite the disparity in age, the two became staunch friends, and throughout life they continued to live in close brotherly fellowship with each
PIONEERS OF ENGLISH HYMNODY
other. Both were Independent ministers and writers of sacred song. Doddridge, however, through enfeebled health, survived his more illustrious compeer by three years only. Philip was the twentieth child of his parents, and after passing through the vicissitudes of a childhood which bereft him of both father and mother, he survived, and grew into a youth of high promise.
Before Doddridge was quite out of his teens he took to preaching, and the Duchess of Bedford, recognising his sterling qualities as a preacher of the Word, offered to send him to the university, and provide a living in the Church of England at her own charges. Doddridge, however, like his friend Watts, declined the offer, and instead, qualified himself for service as a dissenting minister. After seven years' pastorate at the quiet little village of Kibworth, where he had received part of his education, Doddridge settled at Northampton as minister of an Independent Church. In addition to his pastoral work, he opened an academy for the purpose of training young men for the ministry.
He was a man of outstanding literary talent; one of his best known prose works being *The Rise and Progress of Religion in the Soul*. In the production of this book, Doddridge was greatly encouraged and assisted by Watts, whose declining strength did not permit the good doctor to carry into effect his own design. It is interesting to note that it was through the reading of this book that brought about a great spiritual change in the life of William Wilberforce, the benefactor of slaves, and prompted him to write his *Practical View of Christianity*, which, it is said, moulded the life of that illustrious Scotsman, Dr. Chalmers.
It is chiefly as a hymn writer that the name of Philip Doddridge will be remembered, for, as such he occupies a premier place. His hymns number about four hundred, and, singular though it may seem, they
did not pass beyond the stage of manuscript during the author's lifetime. They were first issued in sheet form in the author's own handwriting, and were passed about and read. In this way many of Dr. Doddridge's manuscripts have come to be preserved, not a few of which still show the writing to be fresh and clear, after the passing of more than a hundred and fifty years.
Among the hymns which have attained the greatest popularity, are, "Grace, 'tis a charming sound," "O God of Bethel, by whose hand," "My God, and is Thy table spread?" and that sweetest of all hymns of the child of God, who, in the ecstasy of a joy-filled heart finds fullest expression in the glowing words:
"O happy day, that fixed my choice
On Thee, my Saviour and my God!
Well may this glowing heart rejoice,
And tell his raptures all abroad."
The hymn, "O God of Bethel, by whose hand," was written to follow a sermon on "Jacob's Vow," (Gen. xxviii. 20-22) preached January 16th, 1737, and is numbered among the Scottish Psalms and Paraphrases. There is a pathetic interest attached to these verses, from their association with the story of the heroic missionary, Dr. Livingstone. From early boyhood his heart had been tuned to the Psalms and Paraphrases, beloved by Scots people the world over, and it is said that this one, which had fixed itself on his memory, became the favourite hymn of his wanderings in darkest Africa. And when on April 18th, 1874, his remains were borne to their last resting-place in Westminster Abbey, it was to the plaintive strains of this hymn:
"O God of Bethel, by whose hand
Thy people still are fed,
Who through this weary pilgrimage
Hast all our fathers led."
Perhaps the most unpopular hymn that Dr. Doddridge ever wrote was one for early rising. "The very object for which the good doctor wrote," says a contemporary writer, "proved the death blow to the composition, for if there is one thing more than another to which members of the human race strongly object, it is leaving their beds in the small hours. Dr. Doddridge was not aware of this, and, in order that the hymn should not be entirely wasted, sang it himself. At five o'clock," so the story goes, "he prepared to leave his bed, repeating five stanzas before doing so; at the sixth he rose and dressed." The writer is careful not to tell us how long the worthy doctor lingered over the first five verses!
Consumption, brought on by overwork, compelled Dr. Doddridge to leave England for a milder climate, and he sailed for Lisbon, where he died two weeks after his arrival, on October 26th, 1751, in his fiftieth year.
On the wall behind the pulpit of the church of St. Mary Woolnoth, London, may be seen a simple tablet bearing this inscription:
JOHN NEWTON
clerk,
once an infidel and libertine,
a servant of slaves in Africa;
was
by the rich mercy of our Lord and Saviour
Jesus Christ,
preserved, restored, pardoned,
and appointed to preach the faith
he had long laboured to destroy.
Near sixteen years at Olney in Bucks,
and twenty-seven years in this church.
One of the most remarkable men, whose name posterity has lovingly inscribed on the honoured roll of early English hymn writers is John Newton, author of the sublime hymn, "How sweet the name of Jesus sounds," and whose hand penned his own epitaph. He was born in London in the year 1725, the only child of his parents. His mother, a pious dissenter, was a godly praying woman, and stored his childish mind with Scripture, but died when the boy was only seven years old. At the age of eleven, his father, a captain in the Merchant Service, took his son to sea with him. John's life at sea teems with thrilling escapades and reckless profligacy. Seeking the company of evil companions he grew into an abandoned and godless sailor, and the glimmering rays of the religious thoughts of boyhood changed into fixed infidelity. Seized by the press-gang, he was forced into the Royal Navy, from which he soon afterwards deserted. On being recaptured he was flogged, and later dismissed for insubordination. Newton again took to sea, sailing under the flag of a West African slaver, from which he was expelled on account of his wicked conduct, and for fifteen months lived half starved in degradation under a cruel slave dealer. Returning to England, the ship in which he sailed encountered a terrific storm. It was while taking his turn with the shipmates at the pumps, when hope was all but abandoned, that John Newton reached the turning point of his career. In the face of apparent death he cried aloud to God for mercy. His past life came before him like an ominous cloud and the profligate sailor was filled with deep remorse. The ship was saved, but ere it reached harbour the light of the glorious Gospel shone in upon the benighted soul of John Newton. He was then just twenty-three. In 1755, Newton abandoned his seafaring life, and coming under the influence of such men as Wesley and Whitefield he had his thoughts directed to the ministry. His spare time
was now taken up reading theological books, the study of God's Word, and the preaching of the Gospel. At the age of thirty-nine he was licensed to the curacy of Olney, in Buckinghamshire, and it was here he opened up a life-long friendship with William Cowper, the poet. Together they composed the well-known Olney Hymns, published in 1779, of which collection sixty-eight were composed by Cowper, and two hundred and eighty by Newton. Some of John Newton's best known hymns are: "Begone unbelief, my Saviour is near," "Glorious things of Thee are spoken," "In evil long I took delight," "Rejoice, believer in the Lord," and "May the grace of Christ our Saviour."
The composition, however, by which his name will always be remembered is the hymn beginning:
"How sweet the name of Jesus sounds
In a believer's ear!
It soothes his sorrows, heals his wounds,
And drives away his fear."
Newton lived to the ripe age of eighty-two and died in the city of his birth. He was buried in St. Mary, Woolnoth, but owing to the construction of an underground railway station in the vicinity, the church was closed from 1891 to 1893 in order that the vaults might be cleared. The bodies were re-interred at Ilford, but the remains of John Newton and his wife were taken to Olney and buried in the south-east corner of the churchyard, where a monument was erected at a later date. A grim relic of the removal is to be seen in the discoloured brass plate, formerly on Newton's coffin, which is hung up under a glass on the west wall of Olney church.
Following in the wake of Watts, Doddridge and Newton, came the poet Montgomery, who gave to the world, "For ever with the Lord," and whose name as a
hymn writer ranks in the very forefront. The son of a Moravian minister of Irish descent, James Montgomery was born on November 4th, 1771, at Irvine in Ayrshire, the land made famous as the native place of Robert Burns. His father was anxious that James, his eldest son, should follow in his footsteps, and sent the boy to a Moravian school in Yorkshire to pursue his studies. Soon afterwards, the parents sailed as missionaries to the West Indies, where they died, leaving their son to fight life's battles single handed. Left to his own resources, it was not long before young Montgomery, tiring of the restraint of school life, and feeling his utter unfitness for ministerial duties, abandoned all thought of fulfilling the purpose of his father. Escaping from school he set out in search of work, and obtained employment in a retail shop at Mirfield, near Wakefield. This change, however, did not exactly suit his youthful taste, and he sought other employment, which he found in the shop of a draper at Wath-upon-Dearne. All this time James was busy writing verses, and before he had reached his eighteenth birthday he had collected quite a considerable number of poems. Fired with ambition, he made a journey to London with the hope of finding a publisher for his youthful poems, but this ended in failure, and he returned to Sheffield, where he secured employment in the office of the *Sheffield Register* newspaper. Montgomery eventually became proprietor, changed its name to *Iris*, and successfully conducted it for thirty years. As a journalist, his principles in days when party feeling ran high, were evidently too liberal for the Government's ideas, and Montgomery was twice imprisoned, once for publishing a poem—not his own—on "the Fall of the Bastille." It was when Montgomery lay in prison that many of his best hymns were written. To its credit be it said that the Government, by way, perhaps, of atonement, in after years conferred on the poet a well-deserved pension of £200 a year. Recognized as a poet of high water mark, Montgomery's poems brought him considerable popularity, and called forth the admiration of such critics as the celebrated Lord Byron.
Montgomery was once asked, "Which of your poems will live?" to which he replied, "None, sir, except a few of my hymns." He spoke truly. It is by his hymns that Montgomery is remembered, rather than by his more ambitious poetry. Thus, while his *Wanderer of Switzerland* and *The World before the Flood*, which brought him fame as a poet, are unknown to-day, such hymns as "For ever with the Lord," "Prayer is the soul's sincere desire," "Hail to the Lord's anointed," and that sweetest of all communion hymns:
"According to Thy gracious word,
In meek humility,
This will I do, my dying Lord,
I will remember Thee."
will never be forgotten. He wrote about four hundred hymns, a considerable number of which are in present use. Dr. Julian, the eminent hymnologist, has this to say of Montgomery: "The secrets of his power as a writer of hymns were manifold. His poetic genius was of a high order, higher than most who stand with him in the front ranks of Christian poets. His ear for rhythm was exceedingly accurate and refined. His knowledge of Holy Scripture was most extensive. His religious views were broad and charitable. His devotional spirit was of the holiest type. With the faith of a strong man he united the beauty and simplicity of a child. Richly poetic without exuberance, dogmatic without uncharitableness, tender without sentimentality, elaborate without diffusiveness, richly
THE ROMANCE OF SACRED SONG
musical without apparent effort, he has bequeathed to the Church of Christ wealth which could only have come from a true genius and a sanctified heart."
James Montgomery passed away in his sleep at The Mount, Sheffield, on April 30th, 1854, and was honoured with a public funeral. Later, a statue to his memory was erected over his grave.
The name of John Keble will always be remembered by his monumental work the Christian Year, a collection of poems and sacred songs for the year, published in 1827, which reached its ninety-sixth edition during the lifetime of the author. It is from this poetical storehouse that English hymnody has been enriched by the inclusion in its ever increasing volumes, of that gem of evening hymns:
"Sun of my soul, Thou Saviour dear,
It is not night if Thou be near;
O may no earthborn cloud arise
To hide Thee from Thy servant's eyes."
John Keble was the son of a clergyman, and was born at Fairford, in Gloucestershire, in 1792. He received his early tuition at home under the scholarly guidance of his father, and at the age of fifteen he went to Oxford where the future hymn writer had a brilliant career. At the age of eighteen he took double first honours, which was then counted a rare distinction, and had the honour of being appointed Professor of Poetry. Naturally shy and unambitious, Keble never felt quite at home at Oxford, although it was there he met many of his life-long friends, including Dr. Arnold, who later became the distinguished Rugby master, Lord Coleridge and John Henry Newman.
In 1835, Keble became Vicar of Hursley, a scattered parish six miles from Winchester, with a population of fifteen thousand people. Here for thirty years, amid
the peaceful surrounding of a country village, he made his home till his death in 1866, his devoted wife following him six weeks later.
The greater part of Keble’s life was taken up, in conjunction with such men as Newman and Faber, in religious controversies about church matters known as “Tractarian movement,” and when, eventually, his friend Newman seceded to Rome, though he feared the step was coming, Keble received the news with profound grief. It is, however, as a hymn writer and Christian poet that we have to deal with Keble in these pages. During his incumbency, the villagers had come to know and love their shy and unassuming vicar, for, though much of his time was occupied in writing, he was a constant visitor amongst his flock, ministering to their various needs. Hursley church was restored by him, the necessary funds being drawn from the money received through the sale of the *Christian Year*.
The recognized tune to “Sun of my Soul,” Keble’s most famous hymn, is the one by Sir Herbert Oakeley, known as *Abends*. It is a delightfully sweet and appropriate melody, and though it had many predecessors, *Abends* very soon sprang into favour.
Among other hymns by John Keble to be found in present day hymnals are: “The voice that breathed o’er Eden,” a hymn much used at marriage celebrations, “There is a Book, who runs may read,” “When God of old came down from heaven,” “Lord, in Thy name Thy servants plead,” and that glorious morning hymn:
“New every morning is Thy love,
Our wakening and uprising prove;
Through sleep and darkness safely brought,
Restored to life, and power, and thought.”
Contemporaneous with Keble, and rightly recognized as the Prince of Translators, John Mason Neale
has done more in the realm of hymnology than any other exponent in this particular field of literature. A son of the Church—his father was the Rev. Cornelius Neale, a man of considerable learning—he was born in London on January 24th, 1818. His father died when the boy was five years old, thus his early training was almost entirely under the tender care of his mother, of whom in after years he wrote, "a mother to whom I owe more than I can express." From his youth Neale had a passion for books, and it is said that he read when he sat at meals, read while he walked in the busy streets or the quiet country lanes, read when out driving, read everything that came his way, and what he read was ineffaceably registered on his mind, for he never seemed to forget what he read.
Neale had a distinguished career at Cambridge, where he was considered the best man of his year. Archbishop Trench called him "the most profoundly learned hymnologist of our church"; another "one of the most erudite scholars, one of the best linguists (he knew twenty languages), one of the most profound theologians, and foremost liturgists of his time." In 1842 he was appointed incumbent of Crawley, in Sussex, but owing to ill-health he was compelled to retire from ministerial duties and leave England for Madeira, where he spent about a year. On returning in 1846, Dr. Neale was presented to the wardenship of Sackville College, East Grinstead, where the wardens salary averaged the munificent sum of £27 a year! This post he held till his death, August 6th, 1866. Dr. Neale was a voluminous writer, both in prose and verse, and as he struggled with poverty, his stories for children were written to furnish him with the means of livelihood.
Among hymns from Greek and Latin sources which we owe to Neale, the following are familiar: "O happy
band of pilgrims," "The day is past and over," "Jerusalem the golden," "Brief life is here our portion," and "All glory, laud and honour."
An incident, illustrating in an amusing way his extraordinary mastery of Latin, is worthy of relating here. It was during a visit to his friend John Keble, at Hursley, and is told by the Rev. Gerald Moultrie. "Mr. Keble, having to go to another room to find some papers was detained a short time. On his return, Dr. Neale said, 'Why, Keble, I thought you told me that the Christian Year was entirely original.' 'Yes,' he answered, 'it certainly is.' 'Then how comes this?' and Dr. Neale placed before him the Latin of one of Keble's hymns. Keble professed himself utterly confounded. He protested that he had never seen this 'original,' no, not in all his life. After a few minutes Neale relieved him by owning that he had just turned it into Latin in his absence."
Though the name of John Mason Neale is best known as a translator, he was also an original composer, and many of his hymns find a place not only in our own hymnals, but also in those of other nations. A hymn of surpassing beauty which should really be classed with his original compositions rather than a translation, is:
"Art thou weary, art thou languid,
Art thou sore distressed?
'Come to Me,' saith One, 'and coming,
Be at rest.'"
In his day, John Ellerton, who wrote "The day Thou gavest, Lord, is ended," and many other equally familiar hymns, was termed by Matthew Arnold as "The greatest of living hymnologists." He certainly occupies a pre-eminent place in the foremost rank of hymn writers. Throughout life hymns were his joy
and delight. He compiled many hymnals and rendered valuable service to hymnology generally. But he is best known for his own compositions, many of which will live while the world continues to lift her voice in sacred song.
John Ellerton was born in London, on December 16th, 1826. After graduating at Cambridge and holding several appointments, he became, in 1886, Rector of White Roding. He was appointed Hon. Canon of St. Albans in 1892, and died on June 15th of the following year. "The subject of Mr. Ellerton's hymns," says Julian, "and the circumstances under which they were written, had much to do with the concentration of thought and terseness of expression by which they are characterised. The words which he uses are usually short and simple; the thought is clear and well stated; the rhythm is good and stately. Ordinary facts in sacred history and in daily life are lifted above commonplace rhymes with which they are usually associated. . . His sympathy with nature, especially in her sadder moods, is great; he loves the fading light and the peace of eve, and lingers in the shadows. Unlike many writers who set forth their illustrations in detail, and then tie to them the moral which they are to teach, he weaves his moral into the metaphor, and pleases the imagination and refreshes the spirit together."
Canon Ellerton's best known hymn is:
"Saviour, again to Thy dear name we raise
With one accord our parting hymn of praise;
We stand to bless Thee ere our worship cease,
Then, lowly kneeling, wait Thy word of peace."
which is sung perhaps as frequently as the evening hymns of Lyte and Keble. It was written in 1886. Not unlike many other writers, Ellerton frequently
composed his hymns to some particular melody which happened to take his fancy. In this way a number of his best known hymns were written. When he was asked to write this hymn for a special occasion, he used a tune entitled "St. Agnes," and taking the reverse side of an old sermon note he drafted out his now famous hymn. The original manuscript, which is still preserved, is interesting from the point of view of its many corrections by the author. For some years the hymn was sung to the same old melody, but a more beautiful and appropriate tune, to which it has since been sung, was later composed by Dr. Dykes.
Canon Ellerton did not take out a copyright of any of his hymns, believing that "if counted worthy to contribute to Christ's praise in the congregation, one ought to feel very thankful and very humble."
He wrote a large number of hymns, of which the following are of real merit: "Throned upon the awful tree," "Our day of praise is done," "This is the day of life," "Now the labourer's task is o'er." The last named strikes a singularly sympathetic note, as do most of his hymns, and has become the funeral hymn of our nation:
"Earth to earth, and dust to dust,
Calmly now the words we say;
Left behind we wait in trust
For the resurrection day.
Father, in Thy gracious keeping
Leave we now Thy servant sleeping."
No book on the fascinating subject of hymnology would be complete without fitting allusion to Dr. Julian, the greatest authority on hymns and hymn writers this country has ever known. He is well known throughout the English-speaking world, and by literary scholars in all Christian countries, as a writer of books on hymnology, the principal one of which is
the monumental work "A Dictionary of Hymnology, setting forth the Origin and History of Christian Hymns for all Ages and Nations." This was published in 1892, and a second edition, "Revised, with new Supplement," in 1907. He was over ten years in preparing the first edition, although he had nearly three dozen assistants to help him. The work contains over 1,600 pages, and gives an account of the origin and history of about 30,000 hymns, beginning with one by Clement of Alexandria, the earliest known hymn (excepting heathen ones) outside the bounds of Scripture Dr. Julian estimated at that time (1892) that the number of extant Christian hymns of all ages and nations, might be approximately reckoned at 400,000, of which about 100,000 are in the German language, and nearly as many in the English; after which, for number, come the Latin and Greek.
Dr. Julian was born at St. Agnes, Cornwall, on January 27th, 1839. He was ordained in 1867, and, after holding three curacies, was appointed vicar of Wincobank, Sheffield, in 1876, which living he retained till 1905, when he accepted the vicariate of Topcliffe, Thirsk. He was appointed Canon of York in 1901. He published his first book on hymnology in 1874, and his last, Sacred Carols, Ancient and Modern, with Musical Illustrations, in 1909. Dr. Julian died on January 22nd, 1913.
The most eminent of all Scottish hymn writers, and well to the forefront of the world's hymnists of last century, the name of Dr. Horatius Bonar may well rank with Watts, Doddridge and Wesley. His hymns, simple enough that a child can understand, yet profoundly spiritual withal, are loved and sung, not only in the land that gave him birth, but in countries beyond the seas, wherever these heavenly songs have been carried.
Horatius Bonar was the son of a lawyer, and was born in Edinburgh, on December 19th, 1808. He was one of several brothers who all became eminent ministers in the Church of Scotland. Educated at the famous High School and University of his native city, in his student days he came under the influence of such men as Dr. Chalmers, Edward Irving and Robert Murray McCheyne. He early decided to devote his life to the Lord's service in the ministry of the Gospel, and on completing his theological course he undertook mission work at St. John's Church, Leith. It was here that he began to write hymns. Keenly interested in young folks, the Sunday School, under his care, very quickly grew and prospered. When he first began mission work he found the boys and girls listless and indifferent in the matter of public worship. Accustomed to the use of psalms, not exactly suited, either in word or tune to meet the needs of the young folks, Mr. Bonar realized that what ought to have been the brightest part of the service was to them the dullest. And yet the children loved music and were ready enough to sing songs on week-days. So he tried an experiment; choosing a few familiar tunes such as "The Flowers of the Forest," he set to work writing words to them. These he had printed in leaflet form and distributed amongst the young folks in the Sunday School. To Mr. Bonar's delight the experiment succeeded, and the children immediately took to singing the new hymns which had been specially written for them. In this way were written Horatius Bonar's first two hymns, "I lay my sins on Jesus," and "The morning, the bright and the beautiful morning."
After four and a half years' work at Leith, he became minister of the North Parish Church, Kelso, in November, 1837, where he laboured with a devotion and enthusiasm that never waned during his long and faithful
ministry. His first sermon was long remembered by those who heard it delivered from the pulpit. It was a clarion call to prayer. "Pray brethren!" was his cry, "so shall the showers of heaven descend upon our church, our parish, our schools, our families. It is to prayer I urge you, to prayer for yourselves, to prayer for me!" But Horatius Bonar was pre-eminently a man of prayer, and in after years the voice of earnest pleading from behind the fast closed door of his study, pleading that continued often for hours at a time, formed one of the most sacred memories in his own home circle. Strong physically, Dr. Bonar was never idle. At Kelso, it is said that in one day he frequently preached three times in the pulpit and once in the open air. He was a valiant for the Truth, and even as an old man, when at Edinburgh, his stentorian voice could be heard heralding forth the Gospel in the open air, sometimes in the meadows and sometimes in Parliament Square. One friend said of him that he was always preaching, another that he was always visiting, another that he was always writing, and yet another that he was always praying.
At the Disruption of 1843 Dr. Bonar remained at Kelso as Minister of the Free Church of Scotland, and it was here that most of his best known hymns were written, including the one commencing:
"I heard the voice of Jesus say,
'Come unto Me and rest;
Lay down, thou weary one, lay down
Thy head upon My breast!'" which is, perhaps, the most loved of all his compositions. Through the courtesy of Miss Eliza Maitland Bonar, a daughter of the late hymnist, I am able to give a reproduction of the original manuscript taken from Dr. Bonar's notebook, which is among some of the
Photo: Moffat, Edinburgh.
PORTRAIT OF DR. H. BONAR.
I HEARD THE VOICE OF JESUS SAY.
From the original MS.
Facsimile of
rare treasured possessions he has left behind. It is written in pencil and is now very much worn and faded. As a hymnal manuscript it is indeed unique, and reveals the peculiar methods adopted by the author when composing his hymns. We are told "he would take his notebook, and while thinking out the lines of his hymn he would be busy with his pencil, making little sketches all over the margin of the page." In the same notebook is the original draft of "I was a wandering sheep," the mention of which recalls a rather amusing story. Years ago, in a Devonshire parish, an old farmer died. His son went to the vicar to arrange about the funeral, and said that the family would much like to have a hymn, so the vicar asked him if they had any choice. "Yes, sir," said the young man. "You see, father has always been a farmer, and he lived once in Australia, and has travelled a great deal, so we should like to have 'I was a wandering sheep.'" The family must have given some thought to the matter, because till then the hymn had never been sung in that remote church.
In 1866, Dr. Bonar removed to Edinburgh, the place of his birth, where he undertook the charge of a new church. Here he laboured till he was past eighty, and though the press of city work somewhat retarded the outpouring of hymns, yet his pen was never still. For a time he edited two magazines and was, in addition, continually publishing prose works. He was also the author of hundreds of tracts, one of which, "Believe and Live," published in 1839, reached a circulation of a million copies.
The visit to Scotland of Moody and Sankey in 1873-74 seemed to revive the flow of hymns, and about this time fresh hymns began to appear in his notebooks, many of these specially written for Mr. Sankey. Regarding one hymn written about this time, there is an interesting story. Mr. Sankey had been singing Tennyson's sad and beautiful poem, "Late, late, so late, and dark the night and chill," for which he had composed a tune. On asking permission of the owner of the copyright to use it in his collection of hymns, he was refused. So, being left with a tune without words, Sankey requested Dr. Bonar to write a hymn which would convey the same Scriptural theme. This was done, and the wellknown hymn "Yet there is room," was the result.
An outstanding feature of the hymns of this notable Scottish Presbyterian is that they belong to no particular denomination, but are in use in almost every form of Christian worship, wherever the songs of Zion are sung. Dr. Bonar wrote about six hundred hymns—not including sixty translations of different Psalms—and these are to be found in hymnals the world over. Possibly the best known are, "I heard the voice of Jesus say," "Thy way, not mine, O Lord," "I lay my sins on Jesus," "I was a wandering sheep," "A few more years shall roll," "Here, O my Lord, I see Thee face to face." His own favourite was:
"When the weary, seeking rest,
To Thy goodness flee."
At a memorial service preached at Grange, Edinburgh, the Rev. R. H. Lundie, an intimate friend of the great hymnist, said, "His hymns were written in varied circumstances, sometimes tuned by the tinkling brook that bubbled near him, sometimes tuned to the ordered tramp of the ocean, whose crested waves broke on the beach by which he wandered; sometimes set to the rude music of the railway train that hurried him to the scene of his duty; sometimes measured by the silent rhythm of the midnight stars that shone above him.
PIONEERS OF ENGLISH HYMNODY
Appropriate in closing, are the words of his "Pilgrim Song":
"A few more suns shall set
O'er these dark hills of time,
And we shall be where suns are not,
A far serener clime."
After a lingering illness, borne with Christian fortitude, his last sun set on the 31st of July, 1889, and Dr. Horatius Bonar passed to that serener clime of which he sang—into the presence of the King.
CHAPTER III.
Some Lady Hymn Writers
Some of the sweetest hymns in the language are the production of women. In this chapter an attempt will only be made to include a few writers whose names are familiar to the average reader. Elsewhere in this volume mention is made of many of the sweet singers in the ranks of women, whom God has inspired to pen messages of song, it may be in the quietude of the sick chamber, or amid the cares of their daily household duties. Among such hymns few hold so high a place in our affections as Charlotte Elliott's "Just as I am." Translated into almost every European language as well as Arabic, this immortal hymn, fragrant with rich evangelical doctrine, has doubtless been used in bringing more souls into the Kingdom of God than any other composition. Written in 1834, when the authoress was forty-five years old, a halo of romance has been woven around the occasion of its birth. The true history of how it came into being is a simple story, and has been told by the late Dr. Moule, Bishop of Durham, whose wife was a Miss Elliott of the same family of the famous authoress.
"Charlotte Elliott was living at Brighton with a married brother, a clergyman, the Rev. Henry Venn Elliott. The whole family had gone off to a bazaar in which they were greatly interested; and the frail invalid had been left at home alone, lying on her sofa,
with her heart a little sad at being, as usual, shut out from all good works. For her own comfort she began to ponder on the grand certainties of her salvation—her Lord, His power and His promises. Then came a sudden feeling of peace and contentment, and taking her pen, she wrote these beautiful verses, "Just as I am," without any apparent effort. Surely they were God-given—a precious and priceless gift indeed—from her Heavenly Father to His chastened and much-loved child. As the day wore on, her sister-in-law, Mrs. H. V. Elliott, came in to see her and to bring news of the bazaar. She read the hymn and asked for a copy. So the hymn first stole out from that quiet room into the world, where since that day it has been sowing and reaping till a multitude which only God can number have been blessed through its message."
"Just as I am, without one plea,
But that Thy blood was shed for me,
And that Thou bid'st me come to Thee,
O Lamb of God, I come."
The story of the gifted authoress's conversion which led up to the writing of this hymn, is worthy of relating here. When Charlotte Elliott was thirty-three years of age, Dr. Malan of Geneva, a noted evangelist, visited her father's house. One evening, when in conversation with Miss Elliott, he desired to know if she were a Christian. The young lady at first somewhat resented the question, but the subsequent conversation of this spiritually-minded preacher made a deep impression, and Dr. Malan was the means of guiding her feet into the way of peace. It is said that she afterwards kept the anniversary as a festal day, "the birthday of her soul."
A general favourite with home missions, "Just as I am" has been blessed to myriads of souls the world
over. Referring one day to this hymn, the brother of the noted hymnist stated: "In the course of a long ministry I hope I have been permitted to see some fruit of my labour, but I feel that far more has been done by a single hymn of my sister's." It is recorded that after Miss Elliott's death, over a thousand letters which had been received from grateful writers to whom her hymn, "Just as I am," had been blessed, were found in a locked box.
Though the name of Charlotte Elliott is best remembered by "Just as I am," her hymns number a hundred and fifty, and possibly the two next best known are, "My God, my Father, while I stray," and "Christian seek not yet repose."
Miss Elliott died at Brighton on the 22nd of September, 1871.
Of the many consecration hymns to be found within the covers of present day hymnals, none can surpass Frances Ridley Havergal's:
"Take my life, and let it be
Consecrated, Lord, to Thee. . . ."
This beautiful hymn had its origin in a visit by Miss Havergal to some friends in London, where she had gone for several days' stay. In the home there was a family of ten persons, most of whom were unconverted. Desirous that she might be used to lead them to the Saviour, she prayed to God that ere she left, all in the house might be saved. The result was, before she returned to her home, every one received a blessing. On the last night of her visit, she had just retired to rest when the governess came to her bedroom with a message from the two daughters of the house. They were deeply troubled about their spiritual condition, and were shedding tears. Miss Havergal was able to put
before them in simple language God’s way of salvation, and that night the two sisters put their trust in Jesus. Overjoyed at what had taken place, Miss Havergal was unable to go to sleep, and spent most of the night in praise and renewal of her own consecration. Before the sun rose, there took form in sweet flowing lines, one after another:
“Take my life, and let it be
Consecrated, Lord, to Thee;
Take my moments and my days,
Let them flow in ceaseless praise.
* * * * *
Take my love: my Lord I pour
At Thy feet its treasure store;
Take myself; and I will be
Ever, only, ALL for Thee.”
Another favourite by the same writer had its origin in a singular way, when Miss Havergal was travelling in Germany. One day, coming in tired and weary after a long walk, she sat down opposite a picture, which at once attracted her attention. As she gazed on the canvas, under which were the words, “I gave my life for thee,” her thoughts instinctively wandered to that sacred scene on Calvary’s brow, and in a moment the words of the hymn came to her with remarkable swiftness:
“I gave My life for thee,
My precious blood I shed. . . .”
It is refreshing to read the authoress’ own words, contained in a letter to a friend, written fifteen years later, concerning the simple Gospel hymn: “Yes, ‘I gave my life for thee’ is mine, and perhaps it will interest you to hear how nearly it went into the fire
THE ROMANCE OF SACRED SONG
instead of nearly all over the world. It was, I think, the very first thing I wrote which could be called a hymn—written when I was a young girl, in 1859. I did not half realize what I was writing about. I was following very far off, always doubting and fearing. I think I had come to Jesus with a trembling faith, but it was a coming 'in the press' and behind, never seeing His face or feeling sure that He loved me. I scribbled these words in a few minutes on the back of a circular, and then read them over and thought, 'Well, this is not poetry, anyhow; I won't trouble to write this out.' I reached out my hand to put it in the fire, when a sudden impulse made me draw it back, and I put it, crumpled and singed, in my pocket. Soon after, I went to see a dear old woman in the almshouse. She began talking to me, as she always did, about her dear Saviour, and I thought I would see if she, a simple old woman, would care for these verses, which I felt sure nobody else would even care to read. I read them to her, and she was so delighted with them that I copied them out and kept them. And now the Master has sent them out in all directions, and I have heard of their being a real blessing to many."
Frances Ridley Havergal was born on December 14th, 1836, at the little village of Astley, in Worcestershire. From an early year she displayed unmistakable signs of being possessed of a keen literary gift, and by the time she reached womanhood her writings in prose and verse revealed the promptings of a sincere spiritual life, and were attracting attention.
A writer of consecration, this spirit was manifested in her life, and is embodied in the hymns she has left as a legacy to each succeeding generation. In the autumn of 1878 Miss Havergal went to reside at the Mumbles, Swansea Bay, where in the following year she passed away.
SOME LADY HYMN WRITERS
Other familiar hymns by the same gifted writer, which have a wide circulation, are: "Like a river glorious," "I could not do without Thee," "Who is on the Lord's side?" "Golden harps are sounding," and "I am trusting Thee, Lord Jesus." The last named hymn, for which Miss Havergal composed the tune Urbane, was written in 1874. It was the late hymnist's "own favourite," and was found in her little pocket Bible after her death.
Among lady writers of devotional and other poetry, few names are more widely known than Harriet Auber, although she is remembered by a single composition only. That one composition is a hymn of great beauty, commencing:
"Our blest Redeemer, ere He breathed
His tender last farewell,
A Guide, a Comforter bequeathed
With us to dwell."
The original manuscript was one of quite an unusual and unique character, for the hymn was first written by the gifted authoress with a diamond on a pane of glass in a window of her house at Hoddesdon, where for many years Miss Auber resided, and where, in 1862, at the patriarchal age of eighty-two, she died. She lies buried in the quaint old churchyard nearby the house in which she lived. "I happened to pay a visit some nine years since, to old Daniel Sedgwick's out-of-the-way shop of hymn literature," once wrote the Rev. Andrew Carter, editor of the British Messenger, "and while there met the late Rev. Dawson Campell, of Ware, Herts, an ardent lover of hymns, who, like myself, had gone to the shop in Sun Street in search of hymn books. In the course of an interesting conversation, he told me that he had for some time occupied the house at Hoddesdon in which Harriet Auber had
formerly lived. She had written her beautiful hymn, 'Our blest Redeemer, ere He breathed,' on a pane of glass in one of the windows with a diamond, and when Mr. Campbell came into possession, the pane of glass was still intact. Anxious to have it as a curiosity specially interesting to him, he asked permission of the landlord to remove the pane, and put another in its place; but the landlord declined. And so, up to that time, some seventeen years after the author's death, the valuable manuscript of this sweet hymn remained in its place. Mr. Campbell died, I believe, only a short while afterwards; and I have often wondered what became of that pane of glass; whether it still remains unbroken, or whether some child's elbow, or some street boy's ill-habit of stone-throwing, has made an end of it. Among all the curious forms in which hymn writers have written their compositions, I fancy this is the only case on record of a hymn written by its author on a window pane."
From another source we learn that an old resident of Hoddesdon, whose relations lived in Miss Auber's house after her decease, remembered seeing the hymn on a panel of glass in one of the bedroom windows, but at a later period it was removed by some person, and since then no trace of it was ever found.
This singular circumstance of hymn writing recalls a similar instance, for, on one occasion, Robert Burns, borrowing a diamond ring from a friend, wrote some verses of a song on a window pane in the house where he stayed overnight, and this unique form of MS. is still preserved in the house at Dumfries, where the poet died in 1796.
Another lady writer whose name is remembered by a single hymn is Mrs. Margaret Mackay, a native of Inverness. This hymn, by a singular circumstance owes its origin to a visit paid to a country churchyard
in Devonshire. Entering the little wicket gate of God's holy acre, she reverently paused for a moment to admire the beautiful surroundings. "Here was no elaborate ornament, no unsightly decay. The trim gravel walk led to the house of prayer, itself boasting of no architectural embellishment to distinguish it; and a few trees were planted irregularly to mark some favoured spots." The silent mounds were carpeted with green, and there was an atmosphere of restfulness and tranquillity. Wandering amongst the tombstones her attention was arrested by the words, "Sleeping in Jesus," carved on a quaint old stone, words that told of a peaceful end to a chequered life. With the magic words running through her mind, Mrs. Mackay returned home and wrote these beautiful lines:
"Asleep in Jesus! Blessed sleep,
From which none ever wakes to weep,
A calm and undisturbed repose
Unbroken by the last of foes!
Asleep in Jesus! Peaceful rest,
Whose waking is supremely blest;
No fear, no woe shall dim that hour
That manifests the Saviour's power."
Mrs. Mackay died at Cheltenham, in 1887, in her seventy-sixth year.
To an American lady, Mrs. Annie Sherwood Hawks, we are indebted for the hymn dear to every Christian heart:
"I need Thee every hour,
Stay Thou near by;
Temptations lose their power
When Thou art nigh."
Written towards the close of last century, this hymn was wafted out to the world on the wings of
love and joy, rather than under the stress of a great personal sorrow, with which it has been so often associated in the minds of those who sing it. Writing in 1915, Mrs. Hawks has this to say about its composition: "I remember well the morning, more than twenty years ago, when, in the midst of the daily cares of my home, then in a distant city, I was so filled with a sense of nearness to the Master, that, wondering how one could live without Him, either in joy or pain, these words, 'I need Thee every hour,' were flashed into my mind, the thought at once taking full possession of me. Seating myself by the open window in the balmy air of the bright June day, I caught my pencil, and the words were soon committed to paper, almost as they are being sung now. It was only by accident, as it would seem, that they were set to music a few months later and sung, for the first time, at a Sunday School convention held in one of the large western cities of America. From there they were taken farther west and sung by thousands of voices before the echo came back to me, thrilling my heart with surprise and gladness. For myself the hymn was prophetic rather than expressive of my own experiences at the time it was written, and I did not understand why it so touched the great throbbing heart of humanity. It was not until long years after, when the shadow fell over my way—the shadow of a great loss—that I understood something of the comforting power in the words which I had been permitted to give out to others in my hours of sweet security and peace. Now when I hear them sung, as I have sometimes by hundreds of voices in chorus, I find it difficult to realize that they were ever, consciously, my own thought or penned by my own hand."
Though Mrs. Hawks is the authoress of hundreds of hymns, "I need Thee every hour," is the one that has written her name indelibly on the heart of the Christian
world. She passed away at the home of her daughter in Bennington, Vt., on January 3rd, 1918, at the advanced age of eighty-three.
The historic town of Melrose on the Scottish border, made famous in song and story by Sir Walter Scott, was the home of two lady hymn writers, one being Miss Clephane, who wrote "The Ninety and Nine." Regarding the birth of this remarkable hymn it is of interest that it was originally written for children. Miss Clephane was asked by the editor of *The Children's Hour* for a contribution for this magazine, and in the quiet of her room at Bridge End House, Melrose, the hymn was written as we now know it. This was about the year 1868, just the year before her death. It was not, however, till 1874 that it was discovered by Sankey, set to music, and sent forth upon its world-wide mission.
Elizabeth Cecilia Douglas Clephane, daughter of Andrew Clephane, Sheriff of Fife, was born in Edinburgh, 18th June, 1830, and when quite young was taken by her parents to reside at Melrose, where she spent the remainder of her life. "Miss Clephane and her sister were members of my church at Melrose," says Rev. James Irwin, "and though this was long before my time, there still remains a treasured memory of their whole-hearted devotion to their church. The sisters Clephane were, as Paul and Phoebe, succourers of many, and their generosity was a constant joy to my predecessor and the Church Treasurer." It was the custom of the two sisters to send for the Church Treasurer at the end of the financial year, and if he had to report a deficit, he always came away with a cheque for the amount. It is said that they gave up their horses and carriage that the sisters might have more to devote to charity.
Miss Clephane died 19th February, 1869, at Bridge
THE ROMANCE OF SACRED SONG
End House, near the spot where once stood the old bridge referred to by Scott in *The Abbot* and *The Monastery*.
Living in this little border town about this time was another poetess, who had already penned her most famous hymn, although the verses had not yet been carried beyond her native land.
Just over fifty years ago, there appeared a little volume of devotional verse bearing the title, *Immanuel's Land and Other Pieces*, by A. R. C. The authoress, who thus modestly announced herself, was Anne R. Cousin. She lived to the ripe age of eighty-two, but long before she passed away she had the satisfaction of knowing that two of the pieces in her little book had found a place in a large number of hymnals, and that at least one of them was a popular favourite. The piece which gives the title to the volume is now better known as the hymn beginning, "The sands of time are sinking," while another which she entitles "The Substitute," appears in most hymn books as "O, Christ what burdens bowed Thy head!"
Mrs. Cousin was a daughter of Dr. David Ross Cundell of Leith, and was born in 1824. Her father served in the British army for several years as a surgeon, and was present at the Battle of Waterloo with the 33rd Regiment. After the peace of 1816, Dr. Cundell settled in his native town, Leith, where he died when his only child was just three years old. Mrs. Cundell, after her husband's death, removed to Edinburgh, where she resided until her daughter's marriage in 1847 to the Rev. William Cousin, then minister of Chelsea Presbyterian Church, London, but who had previously been minister at Duns, Berwickshire. Shortly after her marriage, Mrs. Cousin removed with her husband to the Free Church of Irvine, and it was there,
about the year 1856, that she composed her best known hymn, "The sands of time are sinking," a hymn which is now known and used throughout the English-speaking world.
The hymn is a selection from a poem of nineteen verses, inspired by a long and loving study of the *Life and Letters of Samuel Rutherford*, and founded on the Scottish Martyr's dying words, "Glory dwelleth in Immanuel's Land." A song of heaven, it deservedly takes high rank, for no other hymn on this theme—so dear to the Christian heart—expresses with such emphasis, the secret joy of heaven's attraction: the glory of the Lamb that was slain.
"The sands of time are sinking,
The dawn of heaven breaks,
The summer morn I've sighed for,
The fair, sweet morn awakes.
Dark, dark has been the midnight,
But dayspring is at hand,
And glory, glory dwelleth
In Immanuel's land."
Dr. John G. Paton, of the New Hebrides, in a letter to the authoress, tells of the profound impression it made upon his mind when sung by a large congregation in St. Kilda, Australia, as the old year was passing away and the new year coming in.
"O Christ, what burdens bowed Thy head," by the same writer, has been much used, and was greatly loved by Mr. Sankey, who referred to it as "a hymn very much blessed."
It was in 1860 that Mrs. Cousin removed with her husband to the Free Church, Melrose, in which town they resided for eighteen years, living afterwards in Edinburgh, where the gifted authoress died on December 6th, 1906.
THE ROMANCE OF SACRED SONG
"In the secret of His presence how my soul delights to hide!
O how precious are the lessons which I learn at Jesus' side!
Earthly cares can never vex me, neither trials lay me low;
For when Satan comes to tempt me, to the secret place I go."
These sublime lines were written by Ellen Lakshmi Goreh, a Mahratta Brahmin lady of highest caste. She was born at Benares, September 11th, 1853. For many years she has laboured among her own countrywomen, often encountering opposition, but also often cheered by finding women glad to listen to the Gospel story, and by getting welcomes here and there, even in the darkest places.
Rescued from heathenism, trained by missionaries, she developed a deep spiritual life, which found expression in these beautiful lines. Truly, no one can read the experience of this converted native of India as here portrayed, without realising something of the "secret of His presence." Miss Goreh, though now in her seventy-eighth year, is still a zealous worker for the Master among her own people in that dark land.
A hymn which has been blessed to thousands of people, and will remain one of the authoress's best productions, when many other songs are forgotten, is:
"Go bury thy sorrow, the world hath its share,
Go, bury it deeply, go, hide it with care;
Go think of it calmly, when curtained by night;
Go tell it to Jesus, and all will be right."
The authoress of this hymn, Mary A. Bachelor, wrote these lines when staying with her brother, to whom she was greatly attached. He was a minister, and in the usual course of his pastoral duties, felt the ever increasing cares and burdens begin to tell on his health. To him his sister confided all her joys and sorrows. One day, after having disclosed to her brother
some peculiar trial which she was enduring, her conscience reproached her for having needlessly added to his already numerous cares. Standing by the open window she looked out upon the scene before her. Across the daisy-strewn lawn fell the heavy shadows cast by the tall poplar trees, and as she gazed, there came to her the thought. "That is just what I have done to my brother! Why did I do it? Why did I not rather bury my own sorrow, and allow only words of cheer and brightness to reach his ears?"
With these thoughts in her mind, and with tears of regret filling her eyes she retired to her little bedroom, and there penned the words of the hymn which has proved a solace to many a troubled heart.
From the sick chamber of an American lady came the hymn:
"There is a gate that stands ajar,
And through its portals gleaming,
A radiance from the cross afar,
The Saviour's love revealing."
Mrs. Lydia Baxter, the writer, was born in Petersburg, New York, in 1809, and was an invalid for many years. This hymn was written about three years before her death, when she had passed her sixtieth milestone.
During the Moody and Sankey mission in Great Britain, in 1873-74, this hymn was much used. It was sung at the watch-night service in 1873, in the Free Assembly Hall, Edinburgh, when there happened to be present a girl, Maggie Lindsay, of Aberdeen. She was much impressed by the hymn, and those seated by her side heard her exclaim, "O, heavenly Father, is it true that the gate is standing ajar for me? If it is so, I will go in."
THE ROMANCE OF SACRED SONG
"That night," says Sankey, who tells the story, "she became a disciple of the Lord Jesus. The next day she called on her pastor, the Rev. J. H. Wilson, minister of the Barclay Church, and told him of her decision. He was greatly pleased and advised her to tell her school companions of her experience. This she did, and succeeded in leading several of them into the light of the Gospel. Scarcely a month later, on January 28th, Maggie took a train for her home, but never reached there alive. At Manuel Junction a collision took place between a mineral train and the one on which she was travelling. A number of passengers were killed, and Maggie, all crushed and broken, was found in the wreck. In one of her hands was a copy of Sacred Songs and Solos, opened at her favourite hymn, 'There is a gate that stands ajar,' the page of which was stained with her life's blood. She was carried into a cottage near the station, where she lingered a few days, and was frequently heard to sing on her dying couch, the chorus of the hymn so dear to her—'For me, for me! was left ajar for me!'"
In commemoration of this touching incident, Mr. Sankey wrote the words and music of his first hymn, "Home at last."
No name is more familiar among lady hymn writers of recent times than that of Ada R. Habershon, whose first hymn was written at the opening of the present century. Since then hundreds of songs for mission services have flowed from her pen. Before Miss Habershon took to writing hymns she was noted as an able writer and lecturer upon the Word of God, and on the invitation of D. L. Moody she paid a visit to America, where she delivered lectures in various parts of the country on the types of the Old Testament, which were afterwards published in book form. During
the visit of Moody and Sankey to London in 1884, Miss Habershon had the privilege of a close friendship with the American evangelists, and on several occasions sang with Mr. Sankey. Later, when Mr. George C. Stebbins, the composer of many of Sankey's popular hymns, visited this country he stayed at the home of the Habershons. "Thus," says Miss Habershon, when speaking of these days, "the subject of hymn-writing was very prominent as we practised duets with him, and learnt and copied some of his new tunes." It was not, however, till 1901 that Ada R. Habershon attempted to write any poetry. "As I lay very ill," she tells us in her autobiography, "sweet thoughts came to cheer me, and the words arranged themselves in metre. A word of sympathy about wasted time led me to think of the Transfiguration, and of how the disciples had been led away from busy work in the plains to climb the mountains with the Master, and I wrote down the lines of 'Apart with Him.'" From that time onward her pen never ceased to respond to the promptings of the Spirit in the ministry of sacred song.
During the Torrey-Alexander Mission, Mr. Charles M. Alexander came one Sunday morning, in April, 1905, to sing and speak to the poor tramps in Gray's Yard. Miss Habershon was present at the meeting, and during the service, a thought bearing upon the theme of the speaker came to her mind, which she at once committed to verse, and jotting it down on a scrap of paper, passed it on to Mrs. Alexander. A few days later Mr. Alexander called upon her, and asked if she would write some Gospel songs for him. To this she replied that she could not write to order, but that she would pray about it, and if the Lord gave her anything he should have it. Not many hours after, Miss Habershon began her first song for him, suggested by Dr. Torrey's address that evening, and by the
following April she had supplied Mr. Alexander with two hundred hymns. One day at Brixton, after an address by Dr. Torrey on the coming again of our Lord, Mr. Alexander happened to mention to Miss Habershon that very few hymns on this subject had taken hold on the people and become favourites. That night she began to write the hymn, "Oh, what a change!" and by the next day it was finished. This was the first of the hymns given to Mr. Alexander which was set to music and used at the meetings. It proved to be one of the favourites of the great mission, and has since been sung in many parts of the world.
It was always Miss Habershon's aim to have a definite theme in each hymn. To her the "ministry of song" had proved a very happy service, and she tells how the thoughts which have formed the subjects of the hymns have seemed to come so definitely in answer to prayer, that she could only praise the Lord for what He had given to her through them.
Ada Ruth Habershon was the youngest daughter of Dr. S. O. Habershon. She was born on January 8th, 1861, in London, where the greater part of her life was spent, with the exception of a few years at a boarding-school at Dover. "Brought up in a Christian home, with believing, praying parents, the young heart was early led by God's grace to believe in the Saviour's love, and her whole life was devoted to His service. As she grew in years she also increased in knowledge of that love, and lived in the sunshine of it." She was called home on February 1st, 1918.
Other hymns by this gifted writer which have become popular, especially in connection with mission services, are: "He will hold me fast," "No burdens yonder," "The Pilot song," "Will the circle be unbroken," "Bearing His cross," and the hymn "Longings," based on the text, "Lord Thou knowest all
SOME LADY HYMN WRITERS
things: Thou knowest that I love Thee,"—John 21:17, which so fittingly expresses the soul breathings of this gifted writer:
"I long to know Thee better
Day by day,
I want to draw much closer
When I pray;
To listen more intently
For Thy voice,
To let the things Thou choosest
Be my choice."
CHAPTER IV.
How Some Famous Hymns were Written
HOW often we join in the singing of the old hymns we love, hymns which are indelibly imprinted on the heart, and have become blessedly precious to us in the many vicissitudes of life's chequered pathway. Hymns which have radiated like a beacon light, piercing the gloom and brightening the way; soul-breathings which, all unconsciously draw us nearer and still nearer to Him, the source of all our song. And yet, familiar though we may be with the hymn book, it is seldom we are able to glance beyond the hymn to the one whom God has used to pen these gems of heavenly song. Around the hymn and the hymn writer many romantic stories have been woven concerning the circumstances under which some hymn or other was written. Not always is a hymn born out of any particular experience through which the author has passed. Such cases, however, do exist. Yet the stories frequently told are pure fancy, started by one and augmented by another, until they become almost a reality to many who hear them for the first time. Many and interesting are the anecdotes attached to a hymn through its use, which, no doubt, has become precedence instead of sequence, in their telling. For example, various accounts have been given as the origin of Charles Wesley's immortal hymn, "Jesus, Lover of my soul." One is, that as he sat in his study on a summer day, a
little bird, pursued by a hawk, flew into the open window, and sought refuge in his bosom where the baffled hawk did not dare to follow. This incident, it is said, set him musing on the best and indeed the only refuge for sinful souls. Another, that a narrow escape from death in a violent storm on the Atlantic inspired him to portray the thoughts of a Christian in deadly peril. These and similar stories, charming and romantic though they may appear, have no foundation of fact, and the most that can be said is that the immortal hymn was written shortly after the great spiritual change which the author underwent in 1738. Its popularity does not wane with age, but on the contrary, increases with each succeeding generation, and to-day it is to be found in the hymn books of practically all English-speaking countries.
Charles Wesley was born at Epworth Rectory in Lincolnshire, in 1707, and died in 1788. He is credited with having written more than 6,000 hymns, and his strenuous advocacy of the use of hymns in public worship, in conjunction with his famous brother, John, powerfully influenced the course of English hymnody.
The companion hymn of "Jesus, Lover of my soul," has an equally attractive history. What appears to be accepted as the authentic circumstances under which "Rock of Ages" was written has often been told, but will bear recording here. Augustus Montague Toplady, sometime curate-in-charge of the parish of Blagdon, on the Mendips, about eight miles from Wells, was one day overtaken by a severe thunderstorm in Burrington Combe, a rocky glen which runs up into the heart of the Mendip Range. There being no habitation near at hand, he took refuge between two massive pillars of rock. At one point of the rugged slopes of grey rock, there is a precipitous crag of limestone, about a hundred feet in height, and right down its centre is a deep fissure.
THE ROMANCE OF SACRED SONG
It was here, sheltering from the fury of the storm, that Toplady penned the famous hymn:
"Rock of Ages cleft for me,
Let me hide myself in Thee;
Let the water and the blood,
From Thy riven side which flowed,
Be of sin the double cure:
Cleanse me from its guilt and power."
Toplady was the son of a major in the British army, who died at the siege of Carthagena in 1740, leaving his infant son to the care of a tender but judicious mother. When only a lad of sixteen, Toplady was converted in a barn in Ireland. He was travelling with his mother, when, passing a barn, he heard singing. Looking in he observed a few humble country people gathered together. When the hymn ended, a plain, uneducated man, stood up to speak, taking for his text, "Ye, who sometimes were far off, are made nigh by the blood of Christ." (Eph. ii, 13.) That sermon was used of God in the salvation of his soul.
"Rock of Ages" was first published in 1775, only three years before the author's death, which occurred on August 11th, 1778, at the comparatively early age of thirty-eight.
It was a great favourite of Mr. W. E. Gladstone, who translated it into Latin, Greek and Italian. As the Prince Consort lay dying, his lips feebly uttered the sweet words of Toplady's hymn. And thus it happened that the Prince, in his latest hour, laid hold on those precious thoughts which had their birth in the simple but faithful discourse of an obscure and unlettered layman in an Irish barn.
Of "Rock of Ages," Dr. Julian says, "No other English hymn can be named, which has laid so broad and firm a grasp of the English-speaking world."
ROCK OF AGES, BURRINGTON COMBE.
The birthplace of Toplady’s famous hymn.
Though Henry Francis Lyte wrote many beautiful hymns, his last composition, "Abide with me! Fast falls the eventide," proved to be the best known and most widely used of all. He was born at the village of Ednam, near Kelso, on the Scottish border, on June 1st, 1793. It is surely to the credit of this little village on the banks of the Tweed, that it gave birth also to James Thomson, the author of "Rule Britannia," and Thomas Campbell the poet, who wrote "Ye mariners of England."
Intended for the medical profession, his plans were changed, and Lyte was ordained a clergyman in the Church of England at the age of twenty-three. Filling several curacies, he eventually became Vicar of All Saints, Brixham, in South Devon, where, for nearly a quarter of a century, he laboured faithfully amongst the rough seafaring population to whom he had become greatly endeared. Mr. Lyte never enjoyed robust health, and in the autumn of 1847 the increased weakness of his constitution, brought about by his faithful devotion to his beloved flock, demanded a complete change, and the doctors ordered his removal to the South of France that he might spend the winter in a more congenial clime. A few days before leaving home, he wrote: "I am meditating flight again to the south. The little faithful robin is every morning at my window warning me that autumnal hours are at hand. The swallows are preparing for flight, and inviting me to accompany them; and yet, alas! while I am talking of flying, I am just able to crawl, and I often ask myself whether I shall be able to leave England at all."
On the Sunday previous to his departure, though in great weakness of body, he made an effort to address his beloved flock once more. His subject was "Holy Communion," and earnestly did he beseech his hearers to acquaint themselves with the Lord Jesus Christ:
that they might place their trust in His sacrificial death on Calvary. He had preached his last sermon. The severe strain of the service, followed by the painful leave-taking of his sorrowing parishioners, greatly exhausted the faithful pastor, and he was led home in much weakness. All that afternoon he rested his soul in sweet repose, pondering over the farewell words he had so lately addressed to his congregation. In the evening of that beautiful September day, in 1847, he left the house alone, and wandered for an hour among the rocks and flowers in the grounds of his home at Berry Head, until dark.
It happened that on that particular night there was one of those glorious sunsets which are sometimes to be seen at Torbay. "The sun was setting in a blaze of glory, and the purple hills of distant Dartmoor stood out darkly against a flaming sky." In the foreground was Brixham harbour, calm and peaceful. Several times Mr. Lyte stopped to rest and to gaze on the wonderful manifestation of nature. His feelings can well be imagined. The dying day would no doubt remind him insistently of his life which was swiftly drawing to a close. It was during this evening walk that the poet was filled with a fervent desire to write one message of consolation to humanity which would be enduring, and returning to his study Lyte wrote the immortal hymn. Before retiring to rest he handed to his daughter, Mrs. Hogg, the manuscript bearing the undying lines:
"Abide with me: fast falls the eventide;
The darkness deepens; Lord with me abide:
When other helpers fail and comforts flee,
Help of the helpless, O abide with me!"
On the following day he left home for Nice, where, a few weeks later, he passed away. As he neared the
end, a smile illuminated his placid face, and pointing upwards, he whispered softly, "Peace! Joy!"
F. A. Jones, in his delightful book, *Famous Hymns and their Authors*, tells in a very entertaining manner of his casual meeting, over forty years ago, with an old member of Mr. Lyte's choir. As the two sat on the old pier at Brixham, watching the trawlers setting sail for the fishing grounds, the old man chatted animatedly about the late hymnist, evidently well pleased to find some one who took an interest in a man of whom he was palpably never tired of talking. "I was a member of Mr. Lyte's choir," he said, "in 1846; I and a dozen others, all dead now. We were deeply attached to him. He had the gentlest expression and most winning manner possible, and yet I suppose we caused him more grief than all his trouble of ill-health. We left his choir and gave up teaching in the Sunday School, and though I should probably do the same thing to-morrow under similar circumstance, it gives me a feeling of intense sadness even now when I think of it. This is how it came about. A short while before he left us to go to Nice, some influential members of the Plymouth Brethren visited Brixham and persuaded ten of us to join them. After due deliberation we went in a body to Mr. Lyte and told him that we intended to leave his church. He took it calmly enough, although we practically constituted his entire choir, and said that nothing would be further from his thoughts than to stand between us and our consciences. He bade us think the matter over very seriously and come to him again in a few days. We did so, but our decision remained unaltered. We left him, and never entered his church again. When 'Abide with me' came to be written, each of us was given a copy, and then we realized, perhaps more keenly than any one else, the true meaning of the words:
THE ROMANCE OF SACRED SONG
'When other helpers fail, and comforts flee,
Help of the helpless, O abide with me.'
A few years ago, when, at the King's express wish, this hymn was sung at the Wembley Stadium, by nearly a hundred thousand people, His Majesty described it as "a most impressive experience." The sweet solemnity, sacred grandeur, and pathetic strains of "Abide with me," thrill the soul to its depths. It has strengthened and consoled millions of people, and will continue its gracious ministry throughout all time.
How the familiar hymn, "Blest be the tie that binds," came to be written is an interesting story. Dr. John Fawcett was pastor of a small and poor Baptist Church, at Wainsgate, in Yorkshire, from which he derived a salary barely adequate to live upon; and his family increasing far more rapidly than his income, he thought it his duty to accept a call which he received in 1772, to a large and influential church in London, to succeed the celebrated Dr. Gill. He preached his farewell sermon, and began loading his belongings on several wagons for removal to his new residence. The sorrowing members of his poor church, unable to control the grief that filled their hearts, pleaded with him that even now he might not leave them, and when at last all was ready for departure the warm hearted Yorkshire folks clung around their pastor and his family with an affection almost beyond description. Completely overcome at such unmistakable evidence of attachment, the good man and his wife sat down on one of the packing cases and wept. "Oh, John, John, I cannot bear this!" cried the poor wife, "I know not how to go!" "Nor I, either," said the good man; "nor will we go. Unload the wagons, and put everything in the place where it was before." His decision was hailed with great joy by the people. A letter was
immediately despatched to the London church intimating that his coming was impossible, and the good man went to work with renewed energy, happy to remain with his poor but attached country congregation, "passing rich" on £25 a year.
It was to commemorate this event that Dr. Fawcett wrote the words of the hymn now so familiar:
"Blest be the tie that binds
Our hearts in Christian love;
The fellowship of kindred minds
Is like to that above.
* * * * *
We share our mutual woes,
Our mutual burdens bear;
And often for each other flows
The sympathising tear."
The story of how "God moves in a mysterious way" came to be written is an oft-told tale, and yet this volume would not be complete without its inclusion. William Cowper, the poet, was born at the Rectory of Berkhamstead, on November 26th, 1731. In his boyhood he was of a sensitive, shy, and nervous temperament, and in later life was given to fits of temporary insanity. During one of these painful experiences, he became possessed with the idea that he should go to a particular part of the River Ouse and drown himself. He hired a post chaise for that purpose and started. The night was dark, and the coachman purposely losing his way, brought the poet safely back to his home again. By this time the cloud seemed to have lifted from his mind, and in deep contrition and thankfulness to God for his deliverance from danger and from death, he wrote this hymn, which, for more than a century and a half, has brought comfort and consolation to many a troubled soul:
THE ROMANCE OF SACRED SONG
"God moves in a mysterious way,
His wonders to perform;
He plants His footsteps in the sea,
And rides upon the storm.
* * * * *
Ye fearful saints, fresh courage take!
The clouds ye so much dread
Are big with mercy; and shall break
In blessings on your head."
Of this hymn Montgomery says, "It is a lyric of high tone and character, and rendered awfully interesting by the circumstances under which it was written, in the twilight of departing reason."
Some of the happiest years of Cowper's chequered life were spent at Olney in sweet fellowship with his devoted friend John Newton, in conjunction with whom were written the famous Olney Hymns. He died on April 25th, 1800.
Of the three hundred and forty-eight "Olney Hymns," Cowper wrote sixty-eight. Possibly his most widely used hymn is "There is a fountain filled with blood," which is still a general favourite, particularly at mission services. Others equally well known are: "O for a closer walk with God," "Hark, my soul! it is the Lord," "Far from the world, O Lord, I flee," and "Sometimes a light surprises the Christian when he sings." Cowper's hymns are set in a plaintive key, and lack the joyous note of gladness to be felt in the compositions of his contemporaries. The last mentioned hymn is rather an exception, and strikes a brighter note.
The best known of all missionary hymns, "From Greenland's icy mountains," is probably one of the most unique examples of spontaneous writing on
record, for, from the moment Reginald Heber took up his pen till the hymn was completed, the time occupied was only twenty minutes. And the only alteration the author ever made was to substitute the word "heathen" for "savage" in the second verse. Heber, at that time Rector of Hodnet, was on a visit to his father-in-law, Dr. Shipley, Dean of St. Asaph and Vicar of Wrexham. It was the Saturday before Whitsunday, 1819, the day upon which a collection was to be taken in aid of the Society for the Propagation of the Gospel in Foreign Parts, and they being together with a few friends in the Vicarage library, the Dean requested his son-in-law to write something suitable to be sung at the missionary service next morning. Heber, having procured writing material, immediately retired to the far end of the room. In a short time the Dean called out to know what he had written, whereon Heber read over the first three verses. "There, there, that will do very well," said the Dean. Heber, however, was not satisfied, remarking that the sense was not yet complete; and again taking up his pen, in a few minutes wrote the fourth verse:
"Waft, waft ye winds His story,
And you, ye waters, roll,
Till like a sea of glory
It spreads from pole to pole;
Till o'er our ransomed nature
The Lamb for sinners slain,
Redeemer, King, Creator,
In bliss returns to reign."
To Heber's repeated request to be allowed to add yet another verse, the Dean was inexorable. It was sung for the first time the next morning in Wrexham church to the old ballad tune, "'Twas when the sea was roaring," but has since been set to more appropriate music. "The original manuscript of 'From
Greenland’s icy mountains,’ ” says F. A. Jones, “ was for many years in the possession of the late Dr. Raffles, of Liverpool, himself a hymn writer of some note. Popular tradition round Wrexham has it that a compositor in the printing works sold the MS. for a pint of ale; but it is far more likely that Dr. Raffles obtained it direct from the printer who was a personal friend of his. A few years since, Dr. Raffles’ effects were sold, and among the objects of interest put up for auction was this identical MS. After some spirited bidding it was ‘knocked down’ to an unknown buyer for the sum of forty guineas. On the authority of the auctioneer the MS. is now in America.”
During the visit of H.R.H. The Prince of Wales to Ceylon, in the spring of 1922, a paragraph relating to this hymn, which received some prominence in the Press at the time, attracted my attention. Under the headline, “Only Man is Vile,” some enterprising journalist had written:
“The Cingalese, who have given the Prince of Wales such a hearty reception, have shown some warrantable resentment of the verse of a famous hymn which describes Ceylon as a place where ‘only man is vile.’ Bishop Heber, in writing the hymn, seems to have allowed animosity as well as devotion to colour his phrases, for the libellous line is said to have been written just after he had discovered himself to have been cleverly cheated by a Cingalese tradesman.”
It would not be a difficult matter to show that this fantastic and highly coloured story regarding the origin of the famous hymn is sheer imagination, and that Reginald Heber was in no way prompted to pen the supposed offensive line in a spirit of animosity, on finding himself “cleverly cheated by a Cingalese tradesman,” for up to that time the hymnist had not
set eyes on so interesting a personage as a tradesman from Ceylon’s Isle. As we have seen, the hymn was written at Wrexham, in 1819, just four years before Heber accepted the Bishopric of Calcutta, so that the picture of the good Bishop being swindled by a dusky native at once loses its colour. One also wonders if it be actually the case, as the paragraph states, that those very fastidious natives have really shown “some warrantable resentment” of the line, which, as asserted, describes Ceylon as a place ‘where only man is vile.’ A more careful study of the second verse which contains the offending line will, I think, show that Heber had a wider field in his intellectual vision than Ceylon’s Isle.
Exception has been taken to the line, “What though the spicy breezes blow soft o’er Ceylon’s Isle.” As we have already stated, Heber had not visited Ceylon at the time the hymn was written, and it is interesting to read an extract from his journal of a visit to India, written four years after the composition of the hymn. “Though we were now too far off Ceylon to catch the odours of the land, yet it is, we are assured, perfectly true that such odours are perceptible to a very considerable distance. In the Straits of Malacca a smell like that of a hawthorn hedge is commonly experienced; and from Ceylon, at thirty or forty miles, under certain circumstances, a yet more agreeable scent is inhaled.”
For about three years Heber pursued the strenuous duties of a missionary bishop, and died at Trichinopoly in April, 1826.
As a hymn writer Reginald Heber occupies a high place. Of his fifty-seven compositions, nearly all are in use: an honour which has fallen to the lot of very few hymn writers. Among the best known are, “Holy, holy, holy, Lord God Almighty!” “Brightest and best of the sons of the morning,” “The Son of God goes
forth to war," and "Thou art gone to the grave," a funeral hymn written after the death of his infant child.
Though manners and customs change with the march of time, there is happily still to be heard in many of our towns and villages, the cheerful strains of old-time Christmas carols. Chief amongst those which have survived, is "Christians awake, salute the happy morn!" It was written nearly two hundred years ago to suit the whim of a spoilt child. And little did John Byrom realise as he playfully presented to his favourite daughter, Dolly, one Christmas morning, a neatly folded sheet of note paper, on which were written the verses, that the hymn would become famous. It is dedicated by the author to Dolly, the fond father having promised to provide the little maid with a carol for Christmas Day. A year or two later, the manuscript coming into the hands of John Wainwright, organist of Manchester Old Parish Church, he set the words to the tune now so familiar. On the Christmas Eve following, Wainwright took his choristers over to Kersal Cell, the home of Byrom, and they sang the hymn for the first time as they stood round the old doorway, while the author, taken entirely by surprise, listened entranced within. The original manuscript, creased and worn, is now carefully preserved in the Chetham Library at Manchester. It is written on an ordinary sheet of notepaper, and is headed "Christmas Day for Dolly."
I am indebted to Mr. Chas. T. E. Phillips, librarian, who had the interesting document photographed for this work, and the illustration is used through the courtesy of the Feoffees of Chetham's Library.
In connection with this famous hymn an interesting discovery was made some years ago in a garden at Stockport, adjoining the church. While preparing the foundations of a new greenhouse, the owner came
CHRISTIANS AWAKE!
Facsimile of the original MS. Inset: John Byrom as a young man.
Specially photographed for this work by courtesy of the Feoffees of Chetham Library, Manchester.
upon a gravestone about a foot and a half below the surface of the ground. On examination it was found to bear the inscription:
"Here lieth John Wainwright,
Organist, Parish Church, Manchester,
died January 17."
It proved to be the gravestone of Wainwright, the composer of the music of "Christians Awake!" A portion of the slab which bore the date was broken off. John Wainwright was buried in Stockport churchyard in 1768. When the church was being restored in 1810 it was found necessary to use explosives for the demolition of the tower, and during the operations many of the gravestones were broken and displaced.
But to return to Byrom. He was born in 1691 on the outskirts of Manchester, in a house still known as Kersal Cell. It was originally a monastic cell, and there still exist traces of an ancient chapel within the building. Byrom appears to have been a somewhat remarkable personality, and though he wrote much verse, it is probable his name would have been forgotten but for "Christians Awake!" During his lifetime Byrom invented a system of shorthand, which he taught, among his pupils being Horace Walpole and Charles Wesley. Educated at Trinity College, Cambridge, he took his B.A. in 1712, becoming a Fellow in 1714. He studied medicine, but never attempted to qualify as a medical practitioner. Amongst his friends he was known as the Poet Laureate of the Jacobites, and many a witty and clever rhyme, which on more than one occasion got him into trouble, came from his pen.
Byrom died in 1763, at the age of 72, and is buried in Manchester Cathedral.
When John Henry Newman, in an anxious moment,
took up pencil and paper and scribbled some verses, giving expression to the doubts that distressed him, as well as the mental darkness that he at that time experienced, he little dreamed that birth had been given to a hymn which would live to be sung by people of almost every clime and tongue. That hymn begins:
"Lead, kindly Light, amid the encircling gloom,
Lead Thou me on!
The night is dark, and I am far from home,
Lead Thou me on!
Keep Thou my feet; I do not ask to see
The distant scene—one step enough for me."
In the early summer of 1833, Newman, who was at that time Incumbent of St. Mary's Church, Oxford, paid a visit to Rome, where he saw and heard much that was to influence his life in years to come. On his homeward journey he became dangerously ill of fever, which necessitated a stay of about three weeks in Sicily, during which time the only friend at hand to tend to his wants was his faithful servant, who watched over him night and day. Upon his recovery, Newman took passage on an orange boat for Marseilles, being under the impression that he must at once return to England and begin a reformation of the Church in accordance with his peculiar views. Of this historic journey Cardinal Newman afterwards wrote: "Towards the end of May I set off for Palermo, taking three days for the journey. Before starting from my inn in the morning, I sat down on my bed and began to sob bitterly. My servant, who had acted as my nurse, asked what ailed me. I could only answer, 'I have a work to do in England.' I was aching to get home, yet for want of a vessel I was kept at Palermo for three weeks. I began to visit the churches and they calmed my impatience. At last I got off in an orange boat bound for Marseilles."
HOW SOME FAMOUS HYMNS WERE WRITTEN
We were becalmed for a whole week in the Straits of Bonifacio. Then it was I wrote the lines 'Lead, kindly Light,' which have since become so well known."
Much controversy has taken place from time to time over the meaning of the two closing lines of the hymn:
"And with the morn those angel faces smile,
Which I have loved long since, and lost awhile."
In 1879, on being asked to solve the problem, the author is said to have replied that he was not bound to remember his own meaning after the lapse of almost fifty years!
The vicar of an English parish church once wrote to the celebrated Dean Alford of Canterbury Cathedral, asking him if he would compose a hymn and tune to be sung in the procession at a choral festival. The Dean consented, and sent his friend a hymn which was quite unsuitable for a march. In writing back, the Vicar cautiously advised the good Dean to go into his Cathedral, walk slowly along the course the procession would take and compose as he marched. Dean Alford, not in the least offended, did so, and produced the hymn, beginning:
"Forward! be our watchword,
Steps and voices joined,
Seek the things before us,
Not a look behind."
The second line is a reminder to the vicar of the place and manner in which the hymn was composed.
Besides writing the words, the Dean composed the treble and bass of the tune to which it was originally sung, humorously remarking to his friend, that he had supplied the hymn with hat and boots, so it was up to him to provide its coat and trousers.
THE ROMANCE OF SACRED SONG
To Dean Alford we are indebted not only for his scholarly expository on the New Testament, but also for some classic hymns to be found in most present-day hymnals. He died at his deanery on January 12th, 1871.
The hymn "Stand up! Stand up for Jesus!" which, for the last fifty years has been sung the world over, was suggested to the author by the dying message of Dudley A. Tyng, the faithful minister of Epiphany Church, Philadelphia, during the great Revival of 1859. Mr. Tyng was the leader of a remarkable mission in connection with the Young Men's Christian Association of Philadelphia. The Sunday before his death was a memorable one, when there assembled in Jaynes Hall, five thousand men to hear what proved to be one of the most impressive sermons ever preached; and it is said that out of those present at least a thousand were converted.
A few days later, while watching a mule at work on a horse-power machine, threshing corn, Mr. Tyng happened to step forward to pat the animal on the neck, when the sleeve of his coat became entangled in the wheels, and his arm was torn off. He died soon after. Just before he passed away he sent the message to those assembled at the Y.M.C.A. prayer meeting held that evening, "Tell them to stand up for Jesus!" George Duffield preached the funeral sermon for his friend, concluding with the exhortation in Ephesians vi, 14. Soon afterwards he composed the stirring hymn:
"Stand up! stand up for Jesus!
Ye soldiers of the Cross;
Lift high His royal banner,
It must not suffer loss;
From victory unto victory
His army shall He lead,
Till every foe is vanquished,
And Christ is Lord indeed."
The Rev. Sabine Baring-Gould, was one of the most prolific writers of his day. Versatile in gifts and strenuous in application, he has crowded great achievements into his long years. A student of mediaeval lore, history and theology, in which he has been greatly aided by extensive travels and researches in Europe, he was an authoritative writer in these fields, and has well over a hundred volumes to his credit. It is at least probable, however, that he will go down to posterity as the author of one hymn. "Onward Christian soldiers" is known in all parts of the English-speaking world. The hymn was originally written in 1865 for a Sunday School festival in connection with his church at Horbury Bridge, Yorkshire, of which Mr. Baring-Gould was at that time the incumbent. In 1895 he said regarding its composition: "Whit-Monday is a great day for school festivals in Yorkshire. On Whit-Monday, thirty years ago, it was arranged that our school should join forces with that of a neighbouring village. I wanted the children to sing when marching from one village to another, but couldn't think of anything quite suitable; so I sat up all night, resolved that I would write something myself. 'Onward Christian Soldiers' was the result. It was written in great haste, and I am afraid some of the rhymes are faulty. Certainly nothing has surprised me more than its popularity. I don't remember how it got printed first, but I know that very soon it found its way into several collections. I have written a few other hymns since, but only two or three have become at all well known."
"Onward Christian soldiers" was on that occasion sung to Gauntlet's tune, for Sullivan had not then composed that stirring march, which in no small measure has contributed to the immense popularity of the hymn.
THE ROMANCE OF SACRED SONG
About eighteen years ago, during some correspondence with the author in connection with his famous hymn, Mr. Baring-Gould kindly sent me an autograph copy of the chorus of "Onward Christian soldiers," which I am able to have reproduced here.
Onward Christian Soldiers
Marching as to war
With the King of Jesus
Going on before.
V. Baring-Gould
The likeness of S. Baring-Gould was sent to me by the noted hymnist some years later. It is from a portrait by Mr. Melton Fisher, R.A., and was exhibited at the Royal Academy in 1921.
A little-known chapter in the career of this noted hymnist relates to the West Riding. In 1864, while curate, he opened a night school for miners and artizans at Horbury Bridge. Then he built a mission chapel, which first of all heard the sweet evening hymn, "Now the day is over."
S. Baring-Gould died in January, 1924, at the advanced age of ninety years.
"Hold the Fort" is a hymn which at once sprung into great favour over half a century ago, as a splendid rallying song, and maintained its popularity for many years. It had its origin in the following historic incident which occurred during the American Civil War: "Just before Sherman began his famous march to the sea in 1864, and while his army lay camped in the neighbourhood of Atlanta on the 5th of October, the army of Hood, in a carefully prepared movement,
passed the right flank of Sherman’s army, gained his rear, and commenced the destruction of the railway leading north, burning blockhouses and capturing the small garrisons along the line. Sherman’s army was put in rapid motion pursuing Hood, to save the supplies and larger posts, the principal one of which was located at Altoona Pass. General Corse was stationed here with about 1,500 men. A million and a half of rations were stored here, and it was highly important that the earthworks commanding the pass and protecting the supplies should be held. Six thousand men under the command of General French were detailed by Hood to take the position. The works were completely surrounded and summoned to surrender. Corse refused and a sharp fight commenced. The defenders were slowly driven into a small fort on the crest of the hill. Many had fallen, and the result seemed to render a prolongation of the fight hopeless. At this moment an officer caught sight of a white signal flag far away across the valley, twenty miles distant, upon the top of Kenesaw Mountain. The signal was answered, and soon the message was waved across from mountain to mountain: ‘Hold the fort; I am coming!’ It was the message of General Sherman. Cheers went up; every man was nerved to a full appreciation of the position; and under a murderous fire, which killed or wounded more than half the garrison, they held the fort for three hours until the advance guard of Sherman’s army came up. The attacking force was obliged to retreat.”
Major Whittle related this incident at a Sunday School conference when Mr. P. P. Bliss was present, and the song “Hold the Fort” was born in his mind. The same night Bliss wrote the words and music of the now popular song.
Next day at a young people’s meeting he went on to the platform, and writing the chorus of the hymn on the
blackboard, sang the verses for the first time in public, the audience joining in the chorus.
No other hymn attained a greater popularity wherever it was sung during the Torrey-Alexander Gospel tour of the world over twenty-five years ago, than "The Glory Song." It has been translated into more than a score of languages and dialects, and it is said that more than a hundred million copies have been printed. When I asked Charles H. Gabriel, the author, to tell me something of its origin, he wrote: "'The Glory Song' was prompted by the slogan of a good old soul we called 'Old Glory Face.' The one safety valve of his pent-up enthusiasm in praise of his Lord was the single exclamation—'Glory!' And it was good to hear him shout it—not in a harsh, raucous tone of voice, yet distinctly and with a charm of earnestness that carried a conviction of holy reverence to all who heard. To hear him pray was to see the gates of heaven open, and to be drawn nearer to the God he served, his prayer invariably ended with: 'and that will be glory for me!' 'Old Glory Face' is in heaven now, but he lived to sing 'The Glory Song,' and to know that out of his life, before men, came the inspiration that gave the song to the world."
Though it cannot be termed a famous hymn, "The Harbour Bell" has gained no small amount of popularity as a Gospel song, and has been found useful in meetings held specially for seafaring men:
"Our life is like a stormy sea,
Swept by the gales of sin and grief;
While on the windward and the lee,
Hang heavy clouds of unbelief."
John H. Yates, a layman in humble circumstances who lived in New York, wrote this hymn about the
year 1891, after reading in a newspaper the following incident, which had been narrated by one of the passengers. "We were nearing a dangerous coast as the night was approaching. Suddenly a heavy fog settled down upon us. No lights had been sighted, and the pilot seemed anxious and troubled, not knowing how soon we might be dashed to pieces on the hidden rocks along the shore. The whistle was blown loud and hard, but no response was heard. The captain ordered the engines to be stopped, and for some time we drifted about at the mercy of the waves. Suddenly the pilot cried 'Hark!' Far away in the distance we heard the welcome tones of the harbour bell, which seemed to say 'This way, this way!' Again the engines were started, and, guided by the welcome sound, we entered the harbour in safety." The words were set to music by Sankey, and it became one of his favourite solos. Here is the chorus:
"‘This way, this way, O heart opprest,
So long by storm and tempest driven;
This way, this way—lo! here is rest.’
Rings out the harbour bell of heaven."
Many years ago, during Revival meetings in Scotland, a young servant girl became anxious about her spiritual condition. Returning from one of the services she called at the manse, and enquired of her minister how she might be saved.
"Hoot, lassie," said he, "don't be alarmed! Just read your Bible and say your prayers, and you will soon be all right."
"Oh, Minister," wailed the poor illiterate lassie, as the tears came to her eyes, "I canna read, I canna pray!" Then lifting her eyes upward, she cried, "Lord Jesus, tak' me as I am!" Her simple prayer
was answered, and she became a faithful follower of Christ.
Eliza H. Hamilton, hearing the story of the Scottish lassie’s experience, wrote the hymn “Take me as I am,” which was soon after set to music by Sankey, and proved a great favourite at revival meetings. Here is the first verse and chorus:
“Jesus, my Lord, to Thee I cry;
Unless Thou help me, I must die:
Oh, bring Thy free salvation nigh,
And take me as I am!
And take me as I am! And take me as I am!
My only plea—Christ died for me!
Oh, take me as I am!”
The origin of the words and melody of that dearly loved sacred song “The sweet by-and-by,” a song pregnant with hallowed memories, forms interesting reading. The author is Mr. S. Fillmore Bennett. In the village of Elkhorn, Wisconsin, U.S.A., where he lived in the early sixties, he had among his personal friends, a composer named Joseph P. Webster. He was of a very nervous and sensitive nature, and subject to periods of depression. Knowing the peculiarities of his friend, Bennett could tell at a glance if he was in one of his melancholy moods, and would endeavour to rouse him by giving him the words of a song or hymn to set to music. On one occasion Webster came into his friend’s office, and walking towards the fire, stood for some time with his elbow resting on the mantelpiece, without speaking a word. In telling the story, Mr. Bennett relates:
“‘Webster,’ I said, ‘what is the matter now?’
‘It is no matter,’ he replied; ‘It will be all right by and by!’”
“The idea of the hymn came to me like a flash of sunlight, and I replied:
HOW SOME FAMOUS HYMNS WERE WRITTEN
"The sweet by and by! would that not make a good hymn?"
"Maybe it would," said he indifferently.
Turning to the desk I penned the three verses and chorus as fast as I could write. In the meantime two friends had come in. I handed the hymn to Mr. Webster. As he read it his eye kindled, and his whole demeanour changed. Stepping to the desk he started writing the notes in a moment, and taking up his violin played over the melody. In a few moments more he had the notes for the four parts of the chorus jotted down. I think it was not over thirty minutes from the time I took up my pen to write the words before the two friends, Webster and myself were singing the hymn in the same form in which it was afterwards published and became popular. While singing, another friend entered, and, after listening awhile, with tears in his eyes, uttered the prediction, "That hymn is immortal."
It was first published in a book of songs called The Signet Ring, issued soon after the American Civil War. It is now in numerous collections, and has been translated into various foreign languages, and is sung in almost every land under the sun.
Many fanciful stories have been told and written pertaining to the origin of Bishop Bickersteth's exquisite hymn, "Peace, perfect peace," and it is refreshing to read an authentic account of its composition which the late hymnist's son, Dr. S. Bickersteth, furnished Dr. Julian. The hymn, which was the work of a few moments, was written in a house facing the Stray, Harrogate, at which the bishop was staying during the summer of 1875. "On a Sunday morning in August, the Vicar of Harrogate, Canon Gibbon, happened to preach from the text, 'Thou wilt keep him
in perfect peace whose mind is stayed on Thee,' and alluded to the fact that in the Hebrew the words are 'Peace, peace,' twice repeated, and happily translated in the 1611 translation by the phrase 'Perfect peace.' This sermon set my father's mind working on the subject. He always found it easiest to express in verse whatever subject was uppermost in his mind, so that when on the afternoon of that Sunday he visited an aged and dying relative, Archdeacon Hill of Liverpool, and found him somewhat troubled in mind, it was natural to him to express in verse the spiritual comfort which he desired to convey. Taking up a sheet of paper he then and there wrote down the hymn just exactly as it stands, and read it to this dying Christian.
"I was with my father at the time, being home from school for the summer holidays, and I well recollect his coming in to tea, a meal which we always had with him on Sunday afternoons, and saying, 'Children, I have written you a hymn,' and reading us 'Peace, perfect peace,' in which, from the moment that he wrote it, he never made any alteration."
Translated into many different tongues, "Peace, perfect peace" has gained a position well to the forefront of famous hymns. Not always is the fact noticed that the first line in each verse is in the form of a question, while the second line in each verse gives the answer.
Dr. Edward Henry Bickersteth was born in London on January 25th, 1825. Educated at Trinity College, Cambridge, where he graduated with honours, he thereafter held various charges, and in 1885 accepted the Bishopric of Exeter in succession to Dr. Temple. Dr. Bickersteth is both hymn writer and poet. He has published several volumes of poems from which about thirty pieces, including that sweet communion hymn "Till He come," have been set to music, and are to be found in various permanent hymnals both in
this country and America. On May 16th, 1906, he passed away at his home in Paddington, at the advanced age of eighty-one.
To those interested in the absorbing subject of sacred song, the name Inellan displayed at the end of a slender looking pier as you sail down the Clyde, at once associates itself with a famous hymn; for it was here that Dr. George Matheson wrote "O love, that wilt not let me go," a hymn beloved not only in Scotland, but the world over. Though deprived of his eyesight in his youth, Matheson had a brilliant career as a student at Edinburgh, and became a powerful preacher. Dr. Matheson gives an interesting account of the circumstances under which "O love, that wilt not let me go" was written. "My hymn was composed in the Manse of Innellan on the evening of June 6th, 1882. I was at that time alone. It was the day of my sister's marriage, and the rest of the family were staying overnight in Glasgow. Something had happened to me which was known only to myself, and which caused me the most severe mental suffering. The hymn was the fruit of that suffering. It was the quickest bit of work I ever did in my life. I had the impression rather of having it dictated to me by some inward voice than of working it out myself. I am quite sure that the whole work was completed in five minutes, and equally sure it never received at my hands any retouching or correction." It is said that the severe mental suffering spoken of by Dr. Matheson, had reference to a painful love affair of his which occurred about that time. He died at North Berwick in 1906.
No more appropriate music could have been written for this beautiful hymn than *St. Margaret*, the sublime composition of Dr. A. L. Peace.
I have a vivid recollection when, as a boy, I was
taken by my father to an organ recital given in Langholm Parish Church, Dumfriesshire, in 1895, about thirteen years after the hymn was written, when the eminent musician thrilled the audience by a performance of his charming composition.
There is an affecting story relating to the writing of "It is well with my soul." Mr. H. G. Spafford, the author of the hymn, was a lawyer in Chicago, and well known as a man engaged in Christian activities. By the foundering of a steamer on which his wife and four children were going to Europe, the latter lost their lives, causing him great sorrow. It was this pathetic circumstance that inspired the grief-stricken father, some time later, to pen this hymn.
"When Mr. Moody and I were holding meetings in Edinburgh, in 1874," said Mr. Sankey, when telling the story, "we heard of the sad news of the loss of the French steamer Ville de Havre, on her return from America to France. On board was Mrs. Spafford with her four children. In mid ocean a collision took place with a sailing vessel, causing the steamer to sink in half an hour. Nearly all on board were lost. Mrs. Spafford got her children out of their berths and up on deck. Being told the steamer would soon sink, she knelt down with her children in prayer, and asked God that they might be saved if possible; or made ready to die if it was His will. In a few minutes the vessel sank, and the children were lost. One of the sailors of the vessel—whom I afterwards met in Scotland—while rowing over the spot where the steamer had disappeared, discovered Mrs. Spafford floating in the water. He rescued her, and in ten days she was landed at Cardiff. From there she cabled to her husband the message, 'Saved alone.' Mr. Spafford had the message framed and hung in his office. He
started immediately for England to bring his wife to Chicago. Mr. Moody left his meetings in Edinburgh and went to Liverpool to try to comfort the bereaved parents, and was greatly pleased they were able to say, 'It is well; the will of God be done.'"
It was in commemoration of the death of his children, that Mr. Spafford wrote the hymn which has brought comfort to many a troubled heart. This is how it opens:
"When peace, like a river, attendeth my way,
When sorrows, like sea-billows, roll;
Whatever my lot, Thou hast taught me to know,
It is well, it is well with my soul."
CHAPTER V.
The Power of Sacred Song
AMONG the divinely ordained instrumentalities for the conversion and sanctification of the soul that God has given, surely one of the greatest is the singing of "psalms and hymns and spiritual songs." Pages might be filled with interesting facts in connection with the use of hymns in the public worship of the House of God. How often have vast audiences been melted and swayed by a simple hymn, when they have been unmoved by a powerful presentation of the Gospel from pulpit and platform.
Looking back over the years that are passed, since, in the language of Horatius Bonar's charming hymn, the child of God could sing:
"I came to Jesus as I was,
Weary and worn and sad;
I found in Him a resting place,
And He has made me glad."
—does not the memory recall many of the dear old hymns that sung themselves into the heart, and taught the truths of God, that otherwise we might not have learned, and led us to the sources of joy and delight that we might never have found. Even yet we can hear the echo of those voices that used to lead the
THE POWER OF SACRED SONG
singing in times of spiritual blessing. Some of them, it is true, were poor and cracked and discordant, and would have utterly spoiled and ruined any songs, other than those of the sanctuary, that were sung in those hours of the Spirit's presence and power.
We remember the story of John Wesley and the old Cornish woman, who, at one of his services, annoyed the preacher by persistently singing out of tune. Wesley's trained musical ear could not bear the discordant notes, and he cried out, "You are singing out of tune, my sister!" "But my heart is singing," was the old woman's quick reply. "Then, sing on, my sister," returned John Wesley. And sing she did, for her heart was making melody to the Lord.
We remember, too, the truth of the terse words of Billy Bray, the converted collier, who, in order to encourage the less tuneful members of his congregation, reminded them that God made both the crow and the nightingale, and he delighted to hear the voice of each.
"I was in Barclay church in Edinburgh, the guest of the pastor, during a crowded Thursday evening prayer meeting," once wrote the late Dr. Pentecost. "In deference to the time-honoured custom of the Scotch, a paraphrase of one of David's psalms was announced. The congregation did bravely and well, considering the metre and the melody (?) But after the meeting was formally opened, the book of paraphrases was quietly tucked under the pulpit and one of our favourite hymns called for:
"I hear Thy gentle voice,
That calls me, Lord, to Thee;
For cleansing in the precious blood
That flowed on Calvary."
In a twinkling, every one whipped out of pocket a
little penny song book. Every face was radiant, and every voice vocal. The house seemed filled with the Spirit, and every heart seemed to be pouring out its faith and hope to God in the hymn that had probably led many of those present to Christ, and had quickened the faith and hope of all."
The great preacher also relates his experience after leaving his hotel in Glasgow one Sunday evening, to go to Dr. Andrew Bonar's church, some two miles distant. He was but a little way from the church when his ears were greeted with the familiar strains:
"Safe in the arms of Jesus,
Safe on His gentle breast,
There by his love o'ershaded,
Sweetly my soul shall rest."
Looking ahead he saw a crowd from whence came the singing. Pressing forward, the doctor joined the throng of men, women and children, gathered about a faithful band of Christian workers, who were holding a service of song at the street corner. This little company could not preach, in the technical sense of the word, but they could sing the glad Gospel out on the evening air, and thus send forth the glorious invitation, "Come!" Dr. Pentecost was deeply impressed with their simple service, for they were evidently engaging in it as a matter that was to be done unto the Lord. As they passed from the singing of one hymn to another, sometimes slipping in a brief prayer between, he noted the effect on the crowd. Though composed mostly of the rougher element, such as is seen in our larger cities, it was hushed into quiet, and even eager attention to the singing. His attention was called to some faces grown serious and thoughtful as they hearkened to words of love and hope, and more than once he observed the tears stealing down the grim cheek of some sinner.
unused to weep. Thus was God at work in those neglected hearts, and doubtless His great love crept into many a soul through those hymns having been so effectually preached to that class of people.
It was the singing of the hymn:
"Lord, I hear of showers of blessing
Thou art scattering full and free."
that reached the heart of a poor outcast when wandering in the paths of sin. This is what she afterwards wrote: "Thank you for singing that hymn, for it was the singing of it that saved me. I was a lost woman, a wicked mother; I have stolen, and lied, and been so bad to my dear innocent children. Friendless, I attended your enquiry meeting, but no one came to me because of the crowd. But on Saturday afternoon, when they all sang that hymn together, those beautiful words, 'Blessing others, O bless me,' seemed to reach my very soul. I thought, Jesus can accept me—"Even me,"—and it brought me to His feet, and I feel the burden of sin removed. Can you wonder that I love those words, and I love to hear them sung?"
The verse referred to strikes an unmistakable note of tenderness:
"Pass me not; Thy lost one bringing,
Bind my heart, O Lord, to Thee;
While the streams of life are springing,
Blessing others, oh, bless me—
Even me!"
Written by Mrs. Elizabeth Codner, the wife of a clergyman, this hymn was first printed in 1861 as a leaflet. The hymn was suggested to the writer in a remarkable way. A party of young friends over whom she was watching with anxious hope, attended a meeting
in which details were given of the great spiritual revival in Ireland, in 1860-61. They came back greatly impressed, and Mrs. Codner pressed upon them the privilege and responsibility of getting a share in the outpoured blessing. On the Sunday following, not being well enough to attend her class, she had a time of quiet communion with the Lord. "Those young people were still on my heart," says the author, when telling the story, "and I longed to press upon them an individual appeal. Without effort, words seemed to be given to me, and they took the form of a hymn. I had no thought of sending it beyond the limits of my own circle, but, passing it on to one and another, it became a word of power, and I then published it as a leaflet." Since then Mrs. Codner's hymn has had a wonderful history. It was largely used during the Moody and Sankey mission in this country, is specially popular at evangelistic services everywhere, and has been introduced into most of the modern hymnals for congregational use.
There is a magnetic power in the singing of sacred song in the open-air, as the following simple testimony of a Bluejacket I once met in Edinburgh will show. In the course of our short talk he told me a remarkable story illustrating the wonderful power of song, which wrought so great a change in his life. "Well, it happened this way," said Jack, as he tightened the grip of his little Bible under his arm, and thrust his hands deeper into the wide pockets of his navy blue trousers. "About two years ago when we were stationed at Malta, I became greatly concerned and troubled about my soul's eternal welfare. I knew that were I to die at that time I would be eternally lost, and I tried in vain to banish the thought from my mind; but the more I tried the more I was plunged into misery. This
went on for some time, till I could endure it no longer; so one day I resolved to put an end to my soul trouble by getting drunk. With this object in view I sallied forth, but had not gone far when my attention was drawn to singing which proceeded from an open-air meeting that was being held on our garrison island, by a few Christians. I listened, and there came to me the words of the song they sang:
"Jesus is tenderly calling thee home—
Calling to-day, calling to-day!
Why from the sunshine of love wilt thou roam,
Farther and farther away?"
"For a few moments I seemed to falter, but swift as an arrow came the voice of Satan, 'Right ahead! Quick march!' I was about to march off to a neighbouring public house, but was arrested by the wonderful words of life which came floating through the clear air:
"Jesus is waiting, oh, come to Him now—
Waiting to-day, waiting to-day!
Come with thy sins, at His feet lowly bow;
Come, and no longer delay!"
"Going over to where the little band of singers were, I listened as I had never done before. Could it be true, I wondered within myself, that Jesus was waiting and ready to receive me with all my sins? I could hold myself back no longer, and that night, at the open-air meeting, amid the scoffs and jeers of my shipmates who were loitering around, I knelt down and accepted Christ as my Saviour." His story told, the young bluejacket bid me a cheery "Good-night," as he sped off to some secluded prayer meeting, unashamed to carry under his arm, his Bible and bright red Sankey hymn book.
How many stories circle round William Cowper’s famous hymn, “There is a fountain filled with blood.” This hymn was written about the year 1770, and was based on the text: “In that day there shall be a fountain opened to the house of David and to the inhabitants of Jerusalem for sin and for uncleanness” (Zech. xiii. 1). Friendless, and without a penny in the world, a middle-aged man made his way down Water Street, New York. In his pocket was his sixth dismissal from the United States Navy on the charge of chronic alcoholism. So utterly hopeless and dejected was he, that he turned in the direction of the East River, determined on suicide. As he trudged along he observed a shaft of light streaming from a partly opened door on the other side of the street, and with that light came the strains of the familiar hymn:
“There is a fountain filled with blood,
Drawn from Immanuel’s veins;
And sinners plunged beneath that flood,
Lose all their guilty stains.”
He stopped and listened, and as the hymn proceeded he murmured to himself, “Why, that is the hymn my mother used to sing when I was a youngster. I will go over and see who sings it before I kill myself.” He slipped through the swinging doors, and immediately found himself in the midst of about a hundred wrecks of humanity like himself. Jerry McAuley, a popular figure, and the friend of the fallen, was standing before them with open Bible, talking in a loving, familiar way. “Now boys,” said he, “if you mean business you may be saved from your sins just where you are. Christ can save a river pirate, a drunkard and a thief, for He saved me.” At that moment there came flickering into the mind of the profligate a little ray of hope, and he cried out, “Why, if that’s true, I’ll try it.” At the
invitation of Jerry he came forward and knelt down at a front seat. Jerry knelt by him and whispered into his ears wonderful words of life. Broken down in spirit, the poor drunkard then and there accepted Christ as his Saviour, and rose to his feet, the shackles of sin forever broken, a free man.
Of Major Whittle’s popular Gospel song, “Come Believing,” there is a story told of a lawyer from the West, who had sunk so low as to become a beggar in the streets of New York. Homeless and penniless he stumbled by a mission hall, the windows of which were open, and he stopped a moment to listen to the singing
“Once again the Gospel message
From the Saviour you have heard;
Will you heed the invitation?
Will you turn and seek the Lord?”
He had been brought up in a Christian home, and as he listened to the singing, his past life rose before him. He decided to go in, and as he took his seat the audience was singing the second verse:
“Many summers you have wasted,
Ripened harvests you have seen;
Winter snows by spring have melted,
Yet you linger in your sin.”
The words of the hymn went straight home to his heart, and he realised that what they sang was indeed a true picture of his own wasted life. Then came the third verse which ended:
“While the Spirit now is striving
Yield and seek the Saviour’s side.”
Broken down, and unable to restrain his emotion, as he pictured his lost condition, the poor man jumped
to his feet as the hymn closed, and cried out, "I will yield; I will seek the Saviour's side!" That night the wanderer found rest and peace for his troubled soul, by believing and trusting in the Lord Jesus Christ, and was happily restored to his wife and children.
It was the singing of Charles Wesley's hymn:
"Depth of mercy! can it be
Mercy still reserved for me?
Can my God His wrath forbear?—
Me, the chief of sinners, spare?"
that brought about in a wonderful way the conversion of an actress, who was appearing at a theatre in a provincial town. Passing along a quiet street one evening, she heard singing which proceeded from a house, the door of which stood partly open. Attracted by the sweet song, she looked in and there saw a number of people sitting together singing this hymn. For a moment or two she lingered, and when the song finished there followed a simple but earnest prayer. So impressed was the young actress with the hymn, that she procured a book containing it. Seeking the quietude of her room, she read over and over again the words of the hymn she had heard at the cottage meeting. That night she gave her heart to God, and resolved to leave the stage. The manager of the theatre, hearing of her decision, pleaded with her to continue to take the leading part in the play she had made famous in other cities, and eventually she promised to appear on the following night. When the time came, the curtain rose, and the orchestra began to play the accompaniment to the song she was expected to sing. There she stood, facing the great audience, but her thoughts were far away. Supposing that she had become temporarily embarrassed, the band played the prelude over a
second and third time. Then, with clasped hands, and eyes suffused with tears, she stepped forward, and sang with deep emotion:
"Depth of mercy! can it be
Mercy still reserved for me?"
And amid a silence, the solemnity of which had never before been experienced in that theatre, the curtain dropped and the performance was brought to a sudden close.
I heard a somewhat similar incident related by the gentleman himself, whose remarkable experience was the means of not only diverting his professional career into another channel, but in entirely transforming his life. While travelling in Australia, he was one evening passing along a quiet street in the town where he was staying, when his attention was suddenly arrested by singing. He was an eminent musician and composer, and accustomed to hearing only the best music the world could produce. What he heard at that moment when brought to a standstill, was not the performance of some classic piece of music, or the latest opera song, but the sweet strain of an old-fashioned hymn:
"My Jesus, I love Thee, I know Thou art mine!
For Thee all the pleasures of sin I resign."
Surely that melody had rung in his heart in days of long ago, and afresh his soul began to drink it in:
"My gracious Redeemer, my Saviour art Thou!
If ever I loved Thee, my Jesus, 'tis now!"
Drawn to the spot from whence the singing proceeded, his curiosity was aroused and he entered the humble building. It was a Salvation Army barracks.
There were but a handful of people, and as he entered they were pouring out their hearts to God in song. The musician remained during the simple service which followed; he realised for the first time that he was a sinner, and though high in the scale of this world's popularity he must needs come down to the level of those zealous Salvationists. That night he trusted in Jesus, and to-day the name of that convert is known throughout the world. From that little Salvation Army meeting there went out to sing the songs of salvation in almost every country of the globe, the famous preacher musician, Mr. W. H. Jude.
This beautiful and tender little hymn was written in 1858, by a Canadian boy, William Ralph Featherston, when he was only about sixteen years of age. The hymn first appeared in the London Hymn Book, without the author's name, and was for many years incorrectly attributed to Dr. A. J. Gordon, who composed the music to which it is sung. Mr. Featherston died in Montreal in 1870, aged twenty-eight. The original copy of the hymn in the author's handwriting, is still a cherished treasure in the family.
Many wonderful instances are on record where the singing of a hymn, wafted from an open-air service, has been the means of arresting more than one wanderer from the paths of sin. By me, as I write, there are two striking testimonies to the power of sacred song, which I received in connection with this subject. One is from a railway guard, who writes: "Standing at the corner of a street, listening to a band of Gospel workers who were singing the hymn:
'God is calling the prodigal,'
"I yielded myself to God and was saved."
Another railwayman testifies: "While standing at
an open-air meeting, under the influence of drink, I heard a converted comedian sing that fine old hymn:
'I was once far away from the Saviour,
As vile as a sinner could be,
And I wondered if Christ the Redeemer,
Could save a poor sinner like me.'
"Realising my sinfulness, and knowing that I was indeed far, far away from God, I decided that from that night I would serve the Lord. Since then I have been kept by the power of God—Hallelujah!"
Still another, sent by a London correspondent, from which I take the following: "On Sunday afternoon we were holding an open-air meeting and were singing the hymn, 'Have you any room for Jesus?' When we came to the second verse:
'Room for pleasure, room for business;
But for Christ the crucified—
Not a place that He can enter
In the heart for which He died?'
—two men came along on their way to fly pigeons. 'Jack,' said one, as he came near, 'that's me; room for pleasure, room for business, but no room for Christ. You can fly the pigeons if you like, but I'm not going.' Bill returned home—if such could be called home—with his pigeons, and coming back soon after, entered the Gospel meeting, at the door of which he had been attracted by the singing. His wife, at a loss to understand what was taking place, followed Bill along the street into the chapel, and sat down beside him. That night, husband and wife found peace by trusting in the Lord Jesus Christ. 'The home to-day," continues my
friend, "is one of the brightest in the great city of London."
Instances could be multiplied, vividly illustrating how the arrow of conviction from the quiver of the Almighty has gone home to the heart of many a wayfarer, by the singing of a hymn or chorus at an open-air service. On the occasion of Sankey's last visit to this country I heard him relate a very impressive incident. When Moody and Sankey returned to America after their first visit to Great Britain, their first meeting was held on a Sunday, in front of the old Congregational church in the village of Northfield, Mr. Moody's home, the building being too small to accommodate the great numbers who had come to hear the evangelists. It was on this occasion that Sankey sang his famous hymn, "The Ninety and Nine," from the steps of the church, which was the means of the conversion of a man who heard him singing it across the Connecticut River, quite half a mile away. He had refused to attend the service, and was quite angry because his family and neighbours had all gone to the meeting. It was a calm summer evening when the song was sung, and Mr. Sankey, sitting at his small cabinet organ, with the church behind him acting as a sounding board to send his voice so great a distance, rang out that impressive story of the lost sheep so clearly and distinctly, that the man sitting on his doorstep on the opposite bank of the river, caught the message of the song, awakening him to the fact that he was one of the lost sheep, and that the Good Shepherd was seeking him. Two weeks later he was led to attend a meeting at a small schoolhouse near his home, with the result that the lost one was found and brought into the fold. He removed to Northfield and became an active member of the church, from the steps of which the sweet song was sung that
reached his heart. Some years afterwards at the laying of the corner stone of the new Congregational church, at Northfield, Mr. Moody requested Mr. Sankey to stand on the corner stone and sing once more "The Ninety and Nine," as he hoped that the new church would be one whose mission it would be to seek the lost ones. "While I was singing," said Mr. Sankey, "Mr. Caldwell, the man who had heard the song across the river, lay dying in his cottage, which stood not far from the new church. Calling his wife to his bedside, he asked her to open the south window, as he thought he heard singing. Together they listened to the same song which had been used to lead him into the way of life." Surely a remarkable coincidence, and one that must have awakened in the mind and heart of the dying man, very precious memories.
How often have there been wonderful cases of conversion from the singing of some old-fashioned hymn, heard under peculiar circumstances and in unexpected places. One striking instance happened during the Crimean War. It was during a period when our soldiers were passing through days of bitter hardship. In the life of Duncan Matheson, the Scottish evangelist, who so faithfully ministered to the spiritual welfare of the men, we read that one night, weary and sad, returning from Sebastopol to the old stable at Balaclava, where he lodged, his strength gone, sickened with the sights he had seen, depressed by the thought that the siege seemed no nearer an end, so, trudging along in mud knee-deep, he looked up and noticed the stars shining calmly in the clear sky. Instinctively his weary heart mounted heavenward, thinking of "the rest that remaineth for the people of God," and began to sing aloud:
THE ROMANCE OF SACRED SONG
"How bright those glorious spirits shine!
Whence all their white array?
How come they to the blissful seats
Of everlasting day?
Lo! these are they from sufferings great,
Who came to realms of light,
And in the blood of Christ have washed,
Those robes which shine so bright."
Next day, though wet and stormy, he went out and came upon a soldier in rags, standing under an old verandah for shelter; his naked toes were showing through worn-out boots. Matheson, speaking words of encouragement, gave him half a sovereign to purchase shoes. The soldier thanked him and said: "I am not what I was yesterday. Last night as I was thinking of our miserable condition, I grew tired of life, and said to myself. . . . I can bear this no longer, and may as well put an end to it. So I took my musket and went down yonder in a desperate state, about eleven o'clock; but as I got round the point, I heard some person singing, 'How bright those glorious spirits shine;' and I remembered the old tune and the Sabbath school where we used to sing it. I felt ashamed of being so cowardly, and said: Here is somebody as badly off as myself, and yet he is not giving in. I felt, too, he had something to make him happy which I had not, but I began to hope I might too, get the same happiness. I returned to my tent, and to-day I am resolved to seek the one thing." "Do you know who the singer was?" asked the evangelist. "No," was the reply. "Well," said the other, "it was I." Tears rushed into the soldier's eyes, and handing back the half sovereign, he said: "Never, sir, can I take it from you after what you have been the means of doing for me."
Charles Wesley's immortal hymn, "Jesus, Lover of my soul," has almost since its birth, two hundred years ago, been the centre of countless stories, but few are
more impressive than that told by a lady who formed one of a party of tourists on board an excursion steamer, which sailed down the Potomac River in America, one lovely summer evening in the early eighties. There was on board a gentleman who delighted the passengers gathered on deck by the happy rendering of some of the old hymns so much beloved the world over. He had just finished singing "Jesus, Lover of my soul," having given the first two verses with much feeling and a peculiar emphasis upon the concluding lines, which thrilled the hearts of every one present, when, for a few seconds a profound hush fell upon all on deck, as the vessel moved slowly through the waters. Presently, from the outskirts of the crowd, a gentleman pressed forward to the side of the singer:
"Beg your pardon, stranger," he said, "were you actively engaged in the late war?"
"Yes, sir," the man of song courteously replied, "I fought under General Grant."
"Well," said the first speaker, "I did my fighting on the other side, and I think—indeed, am quite sure—I was very near you one bright night eighteen years ago this very month. It was a night such as this. If I am not mistaken, you were on guard duty. We, of the South, had sharp business on hand, and you were one of the enemy. I crept near your post of duty, my murderous weapon in my hand; the shadows hid me. As you paced back and forth you were humming the tune of the hymn you have just sung. I raised my gun and aimed at your heart; but at that moment there rang the words:
'Cover my defenceless head
With the shadow of Thy wing.'
"Your prayer was answered. I couldn't fire after that, and there was no attack made upon your camp."
that night. I felt sure, when I heard you sing this evening, that you were the man whose life I was spared from taking.”
As the Southerner finished, the singer, visibly touched, grasped the outstretched hand of the other, and said with much emotion: “I remember that night very well, and also the feeling of depression and loneliness with which I went forth to duty. I knew my post was one of great danger, and I was more dejected than I remember to have been at any other time during my military service. I paced my lonely beat, thinking of home and friends, and all that life holds dear. Then the thought of God’s care for all that He had created came to me with peculiar force. If He could so care for the sparrow, how much more for man created in His own image! With this comforting thought, I sang the prayer of my heart and ceased to feel lonely. How that prayer was answered I never knew until this evening.”
Few stories demonstrating the wonderful power of sacred song are more remarkable and inspiring than what has been recorded of the famous hymn:
“All hail the power of Jesus’ name!”
A missionary stationed in India, met one day in the village street, a strange-looking native, who proved to belong to an interior tribe, living in barbarism, entirely ignorant of the Gospel. On hearing that no one had yet carried the “Good News” to this isolated tribe, a desire filled his soul to obey the Lord’s command: “Go ye into all the world and preach the Gospel to every creature.” Though his friends tried to persuade him not to go alone, he was determined to go and tell these savages of “Jesus and His love.” Soon after
reaching the habitation of the tribe, the savages immediately surrounded him, armed with spears, in a threatening attitude. The missionary could not speak their language, but felt that Christ was with him, and, closing his eyes in silent prayer for protection, he raised his violin—which he invariably carried with him—and began to sing and play:
"All hail the power of Jesus' name,
Let angels prostrate fall;
Bring forth the royal diadem,
And crown Him Lord of All!"
With eyes still closed he continued to lift his heart and voice in song, and on reaching the verse:
"Let every kindred, every tribe,
On this terrestrial ball,
To Him all majesty ascribe,
And crown Him Lord of all!"
He opened his eyes, and found the scene completely changed. Every spear was lowered, the warriors made signs, took the stranger to their huts, and gave him food and shelter. The singing of the hymn, though the savages could understand not a word, touched the hearts of those uncivilised tribesmen in a wonderful way, and was the means in God's hands, of opening the door for the preaching of the Gospel in that dark region.
"All Hail the Power of Jesus' Name," was written by Edward Perronet, about the year 1780, and first appeared in the *Gospel Magazine*. The tune, by W. Shrubsole, is as famous as the hymn itself, and was written about the same time as the words, receiving its name "Miles Lane" from the chapel in Miles Lane, London, where Shrubsole was for many years organist.
But the words of Perronet’s hymn have inspired many a composer since Shrubsole’s setting first appeared. “Diadem” is an old tune that is still sung with great heartiness and fervour. It had its origin in a typical Lancashire village named Droylsden, about three miles east of Manchester. Nearly a century ago the majority of the inhabitants divided their time between hand-loom weaving and hat-making in the daytime, and singing and practicing their instruments in the evening in preparation for the next Sunday’s services at the Wesleyan chapel. In 1837 the leader of the choir was a young musical enthusiast named James Ellor, then in his eighteenth year. Under his fostering care the services acquired more than a local reputation, and when anything special was advertised to take place, people came from far and near to attend the performance of Ellor’s famous village choir. The young choir-leader was always on the look-out for something fresh, and one day in 1838, he went into a neighbour’s workshop and, flourishing a piece of music paper in his hand, called out, “Look here, lads! What d’ye think o’ this?” “This” was a piece of new music! The men wiped their hands and the old ones put on their glasses and, crowding round Ellor, promptly “solf’a’d” the new tune over two or three times. “That’s good, lad,” said one, “an’ where d’ye get it from?” “It’s aht o’ my own yed, an’ it goes to ‘Crown Him Lord of All,’ and we’ll have it next anniversary,” said James, all in a breath.
Such was the birth of this popular tune, and as the anniversary drew near, Ellor made copies of the various parts for the players. These took their copies away with them, and thus the tune got spread about, until within a very short time “Diadem” became the leading feature at all anniversaries for miles round. Ellor subsequently gave up hat-making, and got employment on the new railway then being constructed between Manchester and Godley Junction. In 1843 he conducted his last anniversary, and shortly afterwards emigrated to America. Little is known of him during his later years, except that he worked for some time at his old occupation of hat-making. For many years before his death he was nearly blind. James Ellor died in 1899, in his eightieth year.
Described by James Montgomery as one of the finest hymns in our language, "Jerusalem, my happy home," has been a favourite almost since its birth in the reign of Queen Elizabeth. The version by David Dickson, a native of Edinburgh, is perhaps the most beloved; snatches of which might often be heard in days gone by, not only among the hills and glens of Scotland, but in lands beyond the seas. A young Scotsman lay dying in America, and a Presbyterian minister of Scottish parentage, who happened to be in the vicinity, was called to his bedside. On his arrival, the good man observing that the end was not far distant, and learning that the poor fellow was not prepared to die, tenderly laid his hand on the brow of the sufferer, and sought to point him to the Saviour. But the more he endeavoured to accomplish his object, the more determined appeared the dying man to thwart all the good man's efforts to reach his heart. After many attempts, the minister, almost in despair, left the bedside, walked toward the window, and half unconsciously began to sing:
"Jerusalem, my happy home,
Name ever dear to me:
When shall my labours have an end
In joy, and peace, and Thee?"
This effectually attracted the attention of the dying
Scotsman. A tender chord had been touched, and with a quivering voice he cried out, "My dear mither used tae sing that hymn." His softened spirit was now upon his Redeemer, and bursting into tears he acknowledged his sinfulness and enquired the way of salvation—which it was hoped he indeed found. Many years had passed away since he had heard that hymn sung in far away Scotland; but its words recalled all the scenes and feelings of home, and produced results which, it is probable, that mother had never thought of.
CHAPTER VI.
Hymns Sung in Tragic Moments
Strange indeed are the places in which hymns have been sung. In divers perils, and under tragic circumstances, have people derived consolation and courage in times of need by lifting up their hearts in sacred song.
A few hours before the heroic Nurse Cavell was executed, she was visited by Mr. Gahan, the British chaplain. Having read to her a portion of Scripture, and commended the brave woman into God's gracious keeping, she joined him in repeating Lyte's hymn, "Abide with me." As he said good-bye, she smiled, and said, "We shall meet again; 'Heaven's morn shall break, and earth's vain shadows flee.'"
Lord Kitchener ordered this hymn to be sung at the thanksgiving service at Khartoum, to commemorate the victory of Omdurman, and to acclaim that country for Christianity. When, during the Great War, Lord Allenby took Jerusalem without firing a single shot, "Abide with me" was afterwards sung amid impressive scenes.
A beautiful story comes from Uganda, during the violent persecutions of the Christians there. Some of the native lads were taken by their enemies, and after inflicting dreadful torture they bound them to a scaffold, under which a fire was kindled. As the flames and smoke gradually rose they were cruelly mocked by
their persecutors, but through it all the brave Christian lads clung with great tenacity to their faith, and with their latest breath they lifted up their voices to God in the sweet hymns the good missionaries had taught them to sing.
When the Armenian Christians were being brutally massacred by the Turks, we are told that they sang a translation of "Rock of Ages."
It is also related that Bishop Marvin, wandering homeless in Arkansas, during the American Civil War, found himself marvellously cheered when in the wilderness he overheard a widowed old woman singing, "Nearer, my God to Thee," in the midst of a dilapidated log cabin.
Another incident of the Civil War, also relating to this hymn, concerns a little drummer boy who lay dying on the battlefield of Fort Donelson. The poor lad had lost an arm, and as his life-blood ebbed away, he was heard to raise his voice as he sang his favourite hymn, rendering with his latest breath:
"Nearer, my God to Thee,
Nearer to Thee;
E'en though it be a cross
That raiseth me!"
Possibly the most tragic incident associated with this hymn occurred in the spring of 1912, when the Titanic, "that greatest of all ships, so glorious in strength and beauty," sank in mid Atlantic, with the loss of over sixteen hundred lives. Of the awful calamity which befell the mighty monarch of the deep, on that ill-fated Sunday morning, much has been written; but through the dimness of time's fog the vision of that mighty ship remains. Faintly there appear the lights gleaming across the dark waters, tier above tier, street above street. Ever and anon might be heard gay
HYMNS SUNG IN TRAGIC MOMENTS
music from her decks, and the laughter of women borne lightly on the wind. Then suddenly there is heard the sound of a dreadful shock, the ripping of steel plates by a hidden iceberg, the stagger of that mighty vessel as she receives her death-blow. Then there comes the last plunge forward, the cries of agony in one awful chorus of despair, as the great floating city disappears into the ocean's dark abyss. Amid that awful scene, there remains vividly portrayed in the memory an event which will always be associated with this dreadful catastrophe. When hope of salvation began to wane, in order to avoid a panic even at the last, the heroic bandsmen assembled on deck and played lively tunes to buoy up the spirits of the passengers. But, as the last moments of that great and mighty sea monster drew nigh, and when all hope of salvation had vanished, there was a pause in the music, then out in the clear air of that fatal Sunday morning, there arose to heaven the plaintive strains of that immortal hymn, "Nearer, my God, to Thee, Nearer to Thee!" It is not generally known that the person who suggested the hymn was the Rev. John Harper, of Walworth Baptist Church, London. He was on his way to America with his daughter. It is said that he got together a number of passengers who knelt with him in prayer as the vessel sank, and at his request the band struck up the sweet and familiar hymn. A sailor who was present, and was afterwards rescued, told the story of the good minister whose name was then unknown.
"Nearer, my God, to Thee," was a great favourite of President McKinley, and as he lay dying by the hand of an assassin, he was heard to sing faintly the words of that beautiful hymn. It was also a favourite of our late King Edward, who once wrote: "Among serious hymns, there is none more touching nor one that goes more truly to the heart than 'Nearer, my God, to
THE ROMANCE OF SACRED SONG
Thee!’” This hymn was written by Sarah Adams, the daughter of a couple who first met in Newgate Gaol, where her father had been sent to lie for six months for defending the French Revolution, and criticising the political conduct of a certain Bishop Watson.
A remarkable story is related by one of the Fisk University singers, who was on board the ill-fated steamer Seawanhaka, when she took fire. Driven by the flames, he, along with the greater number of passengers, flung himself into the sea. Swimming to where his wife was struggling in the water, he bade her place her hands firmly on his shoulders, while he endeavoured to reach some wreckage to which they might cling. This, his wife did until almost exhausted, she murmured, “I can hold on no longer!” “Try a little longer,” was the response of the wearied and agonized husband; “let us sing ‘Rock of Ages.’” And as the sweet strains floated over the turbulent waters, it was heard by the sinking and dying. One after another they raised their heads above the sullen waves, joining a feeble effort in this sweet pleading prayer:
“Rock of Ages, cleft for me,
Let me hide myself in Thee.”
As the song rose, their strength was renewed; another and yet another received fresh courage. And now in the distance could be seen a boat approaching. Could they hold out till it arrived? Still they sang; and ere long, with superhuman strength, laid hold of the life-boat, upon which they were borne in safety to land. The survivor, who was inspired in those desperate moments to sing the prayer of his heart, believes that Toplady’s “Rock of Ages” saved many another besides himself and his wife.
Charles Wesley’s well-known hymn, “Jesus, Lover
HYMNS SUNG IN TRAGIC MOMENTS
of my soul," has also proved a source of comfort and consolation under similar circumstances. The Ocean Queen was wrecked in the English Channel, and a steamer cruising along in the darkness soon after heard the sound of singing coming across the water. The tune was familiar, and they could faintly catch the words:
"Jesus, Lover of my soul,
Let me to Thy bosom fly,
While the nearer waters roll,
While the tempest still is high;
Hide me, O my Saviour, hide,
Till the storm of life is past;
Safe into the haven guide;
Oh, receive my soul at last."
The captain bore his vessel down in the direction from whence came the singing, and presently an object was observed in the sea. As he approached, the captain discovered it was a woman with a child hugged to her breast, clinging to some wreckage belonging to the ill-fated ship in which she had sailed, and he immediately lowered a boat. Guided by the singing, which now and again almost died away, the sailors soon reached the brave woman, who, with death's dark waters all around, was singing her song of trust to her faithful Lord and God.
In the summer of 1912, the sailing ship Criccieth Castle, of Carnarvon, met with disaster off Cape Horn, the most southerly point of South America, and the crew had to abandon the ship. The captain, his wife, and four-year-old son, the second officer, and thirteen of the crew got into the large lifeboat, while the other officers and men left in a smaller boat, which was never again heard of. The weather was intensely cold, and owing to a bad leak in the boat there was never less than eighteen inches of water in it, although some
of the men were continually baling it out. For the last seven days of that awful journey of two hundred miles to the Falkland Islands they were entirely without food, and six men died. One night, when all hope seemed gone, one of the starving sailors suddenly burst into a hymn:
"Light in the darkness, sailor, day is at hand!
See o'er the foaming billows fair Haven's land,
Drear was the voyage, sailor, now almost o'er;
Safe within the lifeboat, sailor, pull for the shore!"
Its effect was magical. The exhausted men, still toiling doggedly at the oars with what strength remained to them, sat bolt upright as the first few notes floated out over the silent waste of waters. The hymn inspired them, put new heart into them, and they bent to their work again with hope rekindled in their breasts. Every night, until land was reached, the sailor sang the hymn to cheer his comrades. It was an old hymn, written in the language of the sea, urging all voyagers on the sea of life to get a passage in the lifeboat of salvation, so as to ensure a safe journey to the shores of the Better Land. The refrain is specially appropriate and inspiring:
"Pull for the shore, sailor; pull for the shore!
Heed not the rolling waves, but bend to the oar;
Safe in the lifeboat, sailor, cling to self no more;
Leave the poor old stranded wreck, and pull for the shore!"
P. P. Bliss is the writer of both words and music of this hymn, which, because of its nautical appeal, has always been a special favourite amongst mission workers in our various sea ports. It would, no doubt, be in this connection that the sailor who sang the song to cheer the sinking spirits of his shipmates first heard the
joyful melody. The following narrative suggested the writing of the hymn: "We watched the wreck with great anxiety. The lifeboat had been out some hours, but could not reach the vessel through the great breakers that raged and foamed on the sandbank. The boat appeared to be leaving the crew to perish. But in a few minutes the captain and sixteen sailors were taken off, and the vessel went down.
"'When the lifeboat came to you, did you expect it had brought some tools to repair your old ship?'" I asked one of the rescued men.
"'Oh, no, she was a total wreck. Two of her masts were gone, and if we had stayed mending her only a few minutes, we must have gone down, sir.'"
"'When once off the old wreck, and safe in the lifeboat,'" I continued, "'What remained for you to do?'"'
"'Nothing, sir,'" was the ready reply, "'but just to pull for shore.'"
How the singing of hymns did indeed soothe the savage breasts of a cruel band of Chinese brigands, who had captured and imprisoned a number of our missionaries, was told by Mr. R. W. Porteous at a special meeting recently held in the Kingsway Hall, London, under the auspices of the China Inland Mission. Roped together, the unfortunate missionaries were compelled to march for miles along a rough mountainous road till they reached one of the brigands' strongholds, into which Mr. Porteous and his wife were flung. "We were not anticipating death," said Mr. Porteous, "but the joy and glory beyond it. Then one of the guards brought me my concertina, and asked me to play it. The thumb-straps had been taken off; but after tying it up with tape and string I managed to play them some tunes; and then they asked us to sing to
them. Together we stood there, my wife and I, well knowing that at any moment the order for our execution might be carried out, and we sang:
'Some day the silver cord will break,
And I no more as now shall sing;
But oh, the joy when I shall wake
Within the palace of the King!'
One moonlight night, soon after, they were taken out to a lonely spot on a hill. "This is the place," said the officer, and one of the brigands who carried the executioner's knife took it down from his shoulder. "We stood looking up at the stars," said Mr. Porteous, as he told the story, "and God's peace came into our hearts, and we could not restrain ourselves from singing praise to Him. We sang:
'Face to face shall I behold Him,
Far beyond the starry sky;
Face to face in all His glory,
I shall see Him by and by.'
One of the officers in charge turned to the native prisoners and said, "Listen to these foreigners singing. They are not afraid to die." "We do not know," continued Mr. Porteous, "whether it was the singing that touched the hearts of the guards and kept them from carrying out their threat to kill us, but we do know it was the restraining hand of God." They were preserved that night, and after five weeks' close confinement were miraculously released.
Though more than a decade has passed since the days when Europe was rent from corner to corner with the ravages of the Great War, there are still echoes of those dark days, when, amid the thunder of guns, could be heard the singing of our gallant lads. Songs they had; and yet in the hour when facing death, how often
HYMNS SUNG IN TRAGIC MOMENTS
would rise to their lips the hymns they learned at Sunday school or in the old home. How the singing of a simple chorus at a critical moment gave new strength and courage, was told by a young R.A.M.C. officer. It happened during one of the landings of our men at Gallipoli. Detachments had landed and were scaling the rocks, when the Turks attacked in overwhelming force. The confusion was appalling, and to add to the difficulties of the situation, the British battleships could not support their comrades for fear of mowing as many of them down as the enemy. The seamen, with perspiration pouring down their faces at the agony of restraint, stood to their guns, longing to fire, and yet unable to do so. They could see their brave comrades fighting and falling, now gaining, now losing, and yet they dared not send off a single shell to defend them. Then from somewhere a voice rang out, in the words of the well-known chorus:
"For you I am praying,
I'm praying for you."
It was instantly caught up and rolled out till the volume of song reached the fighting men on shore, and all at once a change was seen. The men drew together with a firmer front, and succeeded in pressing back the enemy. Now came the sailors' longed-for opportunity, and the guns immediately belched forth from all the ships. One terribly wounded boy afterwards gave his testimony that the chorus, which some unknown voice had struck up amid the din of battle, had helped him in a perfectly wonderful way, and how it had penetrated the deafening noise, giving new strength and courage. In spite of having both legs shot away, this young hero smiled bravely, and whilst his wounds were being dressed, he could only speak of all that the chorus had
meant to him and many others, during those terrible moments.
How often during the great struggle, has the plaintive strain of Newman’s famous hymn, “Lead, kindly Light,” been heard, under strange and touching circumstances. At the battle of the Marne, two comrades took part in an attack upon the enemy trenches, when one of them—a fine Christian lad—was mortally wounded, and as he lay in the trenches dying, he sang that hymn through. “And as I lay at his side, firing at the advancing Germans,” said his comrade—who afterwards told the story—"it seemed as if the angels in heaven were listening. I can never forget the feelings I had. The last verse seemed to rise and rise until it flooded the trenches. Strength was given him to sing in his last moments as he had never sung before.”
And what was the verse?
“So long Thy power hath blessed be, sure it still
Will lead me on,
O’er moor and fen, o’er crag and torrent, till
The night is gone;
And with the morn those angel faces smile
Which I have loved long since and lost awhile.”
The brave soldier lad had greeted his loved ones in the midst of battle. He died singing and entered the land of cloudless love. The night had gone, death was swallowed up in victory.
A chaplain, who accompanied the British Expeditionary Force to the Balkans, gives a thrilling account of the remarkable change which suddenly came over the assembled troops on board a large transport, when, in the darkness of the night, some of the lads struck up an old-fashioned hymn. “The night before we landed at Salonica,” he writes, “we arranged a great farewell concert on board the transport. Two
thousand five hundred officers and men, including the general and the ship’s captain, were present. We were anchored in what, perhaps, is the most beautiful bay in the world. During the day we had feasted our eyes on the wonderful old world city, with its minarets, spires, palaces, and monasteries. We had gazed at the golden sunset, and the purple mountain ranges of the Balkans, and we wanted to sing. The programme was of a varied character, and had to be carried out in the darkness, as no lights were allowed. We had some wonderful talent on board. A choir of Welsh soldiers sang in a remarkable manner ‘The Comrades’ Song of Hope,’ and ‘Land of My Fathers.’ They thrilled us, but they did something more—they inspired our souls and lifted us to heaven. I asked them to sing again, and Welshmen are as a rule deeply religious and love hymns. They did not respond with any rollicking chorus, but sang the grand old hymn:
‘There’s a land that is fairer than day,
And by faith we can see it afar,
For the Father waits over the way,
To prepare us a dwelling place there.
In the sweet... by and by...
We shall meet on that beautiful shore.’
“Can I describe the effect? Impossible! A solemn hush stole over the ship—a silence which grips the soul. You could feel the influence of heaven at work. Here were over two thousand weary travellers, thousands of miles from home, and they were thinking of dear ones left behind. Would they ever return to the old village? Would they ever see their loved ones again? Many, alas, never; but, stealing over the waters of the Eastern ocean, and echoing over the old mountains of Serbia and Greece, was a hymn that possibly had never before been sung in those parts, and it gave just the message
those brave, and yet troubled lads of ours wanted. What if they should fall in battle, and be buried in a strange land? To those whose trust for salvation was in Christ there would be a glad reunion some day, in the sweet by and by, in that heaven of love where all sorrow and parting would be over. The General, a fine Christian, gave the lads a parting message, and good seed was sown. How all this would have rejoiced the hearts of St. Paul and his followers, the ancient Thessalonians! The apostle carried to those parts the message of life after death through the resurrection of our Lord from the dead, and two thousand years afterwards British soldier lads, on their way to take part in a great struggle, sing of the same hope on the eve of their landing. Wonderful and inspiring thought!"
One night, a crowd of our soldier boys in France had gathered together in one of the army huts of the Christian Soldiers' Association, at a rest camp behind the British lines. Tired and hungry after a long and trying spell in the trenches, many of the lads were seated about, enjoying a brimming cup of freshly made tea, while others had sought a secluded spot to read, or to write a letter to loved ones at home. Suddenly, amid the babel of voices there came three short, sharp blasts of the sentry's whistle, the signal for "Lights out," and almost immediately could be heard the distant whirr of approaching enemy aircraft. In a moment the hut was cleared—all but about a dozen Christian young fellows, who remained where they were. "I happened to be seated at the piano when we were plunged into darkness," said my friend, when telling me the story, "and as I heard the bombs dropped in quick succession, carrying death on every hand, my fingers instinctively trembled over the keys, as I softly sang the prayer that rose to my lips:
HYMNS SUNG IN TRAGIC MOMENTS
'Other refuge have I none,
Hangs my helpless soul on Thee:
Leave, oh leave me not alone,
Still support and comfort me:
All my trust on Thee is stayed,
All my help from Thee I bring;
Cover my defenceless head
With the shadow of Thy wing.'
"At the first note, my companions joined in the singing of that verse from Charles Wesley's immortal hymn, which long before that dreadful hour had proved a comfort and solace to countless thousands. We then knelt down and committed ourselves into the safe keeping of our Heavenly Father."
Thus, through that dark hour, that little band of soldiers were wonderfully preserved from harm, while many of their less fortunate comrades who had taken refuge elsewhere were killed. Amongst them was the poor sentry nearby, who, but a few minutes before had sounded the alarm.
In the year 1894, the gallant Major Wilson, with a reconnoitring party of thirty-three troopers, were suddenly attacked in Matabeleland by a horde of three thousand natives, who surrounded them in a forest. They fought from early morning till after mid-day. Early in the engagement all their horses had been killed, and behind their dead bodies the brave fellows kept up a desperate fight for several hours, till their ammunition was exhausted and there was not one man left to stand or fire.
When nearly all were wounded or killed, the Induna, in relating the incident, says that Wilson's party took off their hats, and sang something, the kind of song he had heard missionaries sing to the natives. Soon only one man was left, but the Matabele had such a dread of them, that even then they did not rush in and assegai
THE ROMANCE OF SACRED SONG
them until the last man had fallen. And so impressed were Lobengula’s warriors with the bravery of the white men, that when at last they did scramble over the human barricade, they did not mutilate them in any way. It was of this incident that Mary Georges wrote:
"They sang—the white men sang—
Sang in the face of death,
And the forest echoes rang
With their triumphant breath."
I have not been able to discover what it was that those gallant men sang in the hour of death—possibly it may never be known. We are told it was the kind of song the missionaries sang to the natives, and our thoughts instinctively hover round some of the grand old hymns, which have come to us from generation to generation: hymns whose melody has haunted the ear amid the storm and stress of life’s long battle.
Incidents associated with the lifting of the voice in sacred song in strange and tragic moments, whether in the field of battle, or on the high seas, have always a glamour of romance, which never fails to rivet the circumstance on the mind, in a peculiar and ineffaceable way. There are, however, in the quiet moments of everyday life, frequent incidents no less fascinating, and worthy of sympathetic note. A touching story was related to me quite recently. A young man suffering from cancer was about to undergo an operation in one of our large Scottish infirmaries. When the surgeon had made a careful examination he discovered that the disease was so deeply rooted that the only hope of saving the life of the patient was the removal of his tongue. The young man was already in the operating theatre, when the surgeon, in a tender and sympathetic way, explained to the sufferer that even though the operation should be successful, he would never again
be able to speak, and he was asked whether there was anything he wished to say before the operation commenced. For a moment a shadow crossed the brow of the young Christian, at the thought that he would never again be able to testify in song or speech for the Master whom he loved. But soon the shadow passed, and a smile lit up his face. He sat up, and lifting up his voice he sang the hymn:
"There is a fountain filled with blood,
Drawn from Immanuel's veins,
And sinners plunged beneath that flood
Lose all their guilty stains."
Ere the third verse was reached, not an eye of those who stood round was dry. How he sang; for his heart was in the song. Then came the last verse:
"Then in a nobler, sweeter song
I'll sing Thy power to save,
When this poor lisping, stammering tongue
Lies silent in the grave."
The anaesthetic was administered, the operation performed, but, alas, the patient never regained consciousness; thus his last song on earth would be his first in heaven.
CHAPTER VII.
Children's Songs
"Give me a bairnie's hymn,
For I want no earthly lore,
But the sweet refrain of some childlike strain,
On the brink of Canaan's shore.
* * * * * *
Give me a bairnie's hymn;
I feel like a child to-night,
Lying down to rest on its father's breast,
To awake at morning light."
These verses, from a poem by Mrs. Battersby, which appeared in the Sunday School Teacher, in 1874, were suggested by a request for a "bairn's hymn" by Dr. Guthrie, the preacher-philanthropist, as he lay on his death-bed. And it is said that the simple strain soothed and strengthened him during his last moments. Not infrequently these tender melodies, among the first that infant voices learn to lisp, are often among the last whispered by dying saints in their latest hour. With most of us, some of our earliest religious awakenings were in connection with hymns sung at the Sunday school. How often a little hymn would quiet us, and beget within the heart a seriousness and longing! And yet we have not infrequently heard those who wholly understand and appreciate the type of music commonly
termed classic, ridicule a certain class of hymns and tunes, altogether forgetting that these very compositions may exert on other minds a holy and happy influence.
A clergyman was one day busily engaged in his study, when a rough looking man, apparently under the influence of drink, very unceremoniously entered and handed him a note, which he said was from the teacher of the infant class of the Sunday school. The note informed the clergyman that the bearer was the father of one of her scholars who had met with an accident, and that the child lived in such a locality that the teacher dare not visit. "What is your name, and where do you live?" asked the good man. "My name is Peter O'More," answered the other, with a rough Irish brogue, "and I live on an ould canal boat at the foot of Harrison Street."
"And what is the matter with your child?"
"Och! and is it Kitty, my own little darling Kitty, the only child of six that has been born till me? She was playing about on a ship where I was to wark, and she fell down the hatchway and broke her leg, and poor Kitty's leg is not set right, your riverence, for I have no money to pay a docther. Och! poor Kitty! and I've nothing to give her to ate, your riverence."
The minister went down and found a dreadful state of things. The poor little suffering child was overjoyed to see him, whom she at once recognised. She lay upon the "locker" or side seat of an old canal boat, which had been laid up for the winter. There was no fire, though it was a bitterly cold day, no food, and scarcely any article of furniture or any comfort whatever. The minister did what he could to relieve the wants of the little sufferer. The parents had both been drunk the previous night, and in a quarrel had unintentionally knocked the child off the seat, and broken the limb again after it had been set. Having
obtained the services of a surgeon who again set the limb, the good man sat down on the locker to talk to little Kitty, while he fed her with some nourishing food which he had procured. He asked her if she could read. "No," she could not read a word; "but I can sing something I learned in the Sunday School," said she.
"Well, what can you sing, Kitty?"
In a moment her sweet voice broke out:
"There is a happy land,
Far, far away,
Where saints in glory stand,
Bright, bright as day,
O how they sweetly sing
'Worthy is our Saviour King!'
Loud let His praises ring,
Praise, praise for aye."
Tears gathered in the listener's eyes as the sweet song was lisped by the little sufferer.
"Well, Kitty, that is beautiful. Where do you think that Happy Land is?" he asked.
"Oh," came the ready reply, "I suppose it is up in the sky, where God is and where the angels live."
"Do you think you will ever go there, Kitty?"
"If I'm good, and love God, I think I shall."
Poor Kitty could not read, nor could either of her parents. She knew nothing of heaven and divine things, except what she had been taught at the Sunday School; and most of what she remembered was associated with such words and sentiments as we have quoted. Eternity alone will unfold the power of such simple truth, and simple yet sweet tunes, upon infant minds.
"There is a Happy Land," was one of the "bairns' hymns" called for by Dr. Guthrie. How this little hymn came into being is an interesting story.
In 1838, when Andrew Young, at that time headmaster of Niddry Street School, Edinburgh, was on a visit to Rothesay, he happened to spend an evening at the home of one of his pupils, when the lady of the house entertained her visitor by playing several musical compositions on the piano. Among these was a sweet and tender air which charmed him exceedingly. It was a simple little Indian melody called "Happy Land," and Mr. Young, who was passionately fond of music, requested his friend to play it over and over again. As he listened to the music, he remarked that such a melody would make a capital children's hymn if wedded to appropriate words. All that night the melody kept ringing in his ears. Early in the morning he rose, and walking in the garden, wrote the hymn which has since spread over all the world.
Many beautiful tributes to "There is a Happy Land" have been recorded. "One day Thackeray, the novelist, was passing through a London slum, and heard a few ragged children in a gutter singing something. He stopped to listen. It was 'There is a Happy Land.' The contrast between the squalor of the poor waifs and the splendour of the subject of their song struck him so forcibly that he burst into tears." The Rev. J. C. Carrick, of Newbattle, writing in *Life and Work*, in 1890, mentions having had a letter from Mr. Young—only a week before his death—in which he said: "I have just noticed that in Dr. Paton's most interesting book on his missionary labours, there is a notice how a chief was converted through my hymn."
One of the most widely known and best loved of all children's hymns is:
"There's a Friend for little children,
Above the bright blue sky;
A Friend who never changes,
Whose love can never die."
THE ROMANCE OF SACRED SONG
Unlike our friends by nature,
Who change with changing years,
This Friend is always worthy,
The precious name He bears."
It was written by Albert Midlane, on February 27th, 1859, at Newport, Isle of Wight, and first appeared in a children's periodical edited by C. H. Mackintosh, under the title, "Above the bright blue sky." The hymn at once attained great popularity. To-day it has found its way into almost every corner of the world, and has been translated into nearly fifty languages. The first line, as originally published, read, "There's a rest for little children," the word "Friend" being substituted later. Mr. Midlane commenced writing when he was in his teens, his first contribution to hymnody, "God bless our Sunday School," which was sung to the tune of the National Anthem, appeared on May 24th, 1844.
"There's a Friend for little children" was written in a house, practically within a stone's throw from the home of Thomas Binney, who wrote the well-known hymn "Eternal light, eternal light." In his boyhood days, Albert Midlane was brought into close touch with Binney, from whom he may have received some encouragement and impetus in his poetical pursuits. In this connection, however, Mr. Midlane himself has said that it was his Sunday School teacher who did so much to shape his early life, and who prompted him to poetic efforts. The hymn was first scribbled in the author's notebook, and the original manuscript is still preserved.
A writer who was personally acquainted with the author of "There's a Friend for little children," gives the following as the true story of its origin: "Mr. Midlane's mind had been musing on its outline during the day—and that day a very busy day with other matters—and in the evening, his family having retired
to rest, he set himself to arrange and complete the idea. But time stole on, and morning came. Every one who has ever felt an intense interest in an undertaking knows that, at all costs, it must be pursued to the end, and so it was in Mr. Midlane's case. The end came at length. Alarmed at his absence, his wife came downstairs, only to find her husband in a state of unconsciousness, with head resting upon the now finished hymn. Restoratives and rest being ordered, consciousness was restored, but the result was that such night occupations were strictly forbidden for the future." Such was the origin of the hymn. On its first appearance, like many other questions of right, its authorship became a question of dispute, and not until a newspaper controversy was it finally settled, and its first signature of "A. M." became the fully recognised name of the now famous hymn writer.
Considering his busy life in connection with an ironmongery business which he conducted, Mr. Midlane was a fairly voluminous writer, well over a thousand poems and hymns having come from his pen. The greater number of these pieces have been published, including the popular revival hymn "Revive Thy work," which is possibly the next best known hymn, followed by "Passing onward, quickly passing." Though few of his hymns manifest a high degree of literary merit, nevertheless, the tone is good, and there is an unmistakable loyalty to the Word of God beautifully expressed, which, perhaps, in hymn writing is of no inconsiderable importance.
It is of interest to learn of the circumstances under which the standard tune, *In Memoriam*, to which the hymn is usually sung, was written: "The committee engaged on the music of *Hymns Ancient and Modern* were meeting in Langham Hotel, London, and when the hymn came up for consideration it was found that
though they had several tunes before them, none were considered satisfactory. It was suggested that a new tune might be written by one of the committee, and Sir Henry Baker proposed that Sir John Stainer should retire to his (Sir Henry's) bedroom, and try what he could do. Sir John complied with the suggestion, and in a very short time returned with the present tune which was at once adopted." It was in connection with the death of Sir John's young son, Frederick Henry Stainer, which took place about this time, that the tune *In Memoriam* received its name.
The author had the pleasure of witnessing the celebration of the jubilee of his famous hymn, when 3,000 children assembled in St. Paul's Cathedral, London, and joined in the singing of "There's a Friend for little children." Mr. Midlane passed away on February 28th, 1909, and as his body was laid to rest in Carisbrooke Cemetery, a pathos was added to the scene by a number of children, assembled at the graveside, blending their voices to the sweet strains of his immortal hymn.
Another hymn, reminiscent of sweetest memories of childhood's days, is:
"I think when I read that sweet story of old,
When Jesus was here among men,
How He called little children as lambs to His fold:
I should like to have been with Him then."
Written in pencil on the back of an old envelope, while the authoress was travelling in a stage coach nearly a hundred years ago, this little hymn, which quickly rose into favour, is still among our best loved children's hymns. The story of its composition has often been told, but will bear retelling. From an account contained in a charming book of reminiscence, written by Mrs. Jemima Luke, the authoress of the hymn, when
she had passed her eightieth birthday, we obtain a delightful pen-picture of the origin of this notable children's song.
In 1841, Miss Thompson (as she then was) proceeded to the Normal Infant School, Gray's Inn Road, to obtain some knowledge of the system. Mary Moffat, who afterwards became the wife of the famous Dr. Livingstone, was there at the time. In the course of their duties the teachers had to march up and down the schoolroom singing the marching pieces provided for their future use. Amongst these was a Greek air, the pathos of which at once attracted the young woman's fancy, and she searched several Sunday School hymn books for words to suit the measure, but in vain. Having been recalled home, she went one day on some missionary business to the little town of Wellington, five miles from Taunton, in a stage coach. It was an hour's ride. As she sat there, with no other inside passenger with whom she might converse, her thoughts recalled the pretty little tune with which she had been struck, and taking from her pocket a pencil she wrote on the back of an old envelope the first two verses now so well known. "The child's desire," as the composition was originally called, consisted of only two verses, and a third was afterwards added to make it a missionary hymn. "My father superintended the Sunday School at the little chapel belonging to the estate," writes Mrs. Luke. "He used to let the children choose the first hymn themselves. One Sunday they struck up their new hymn. My father turned to my younger sister who stood near him, and said, 'Where did that come from? I never heard it before,' 'Oh, Jemima made it,' was the reply. On the day following he asked me for a copy of the words and tune. This he sent with the name and address in full, to the Sunday School Teachers' Magazine, where it appeared the
THE ROMANCE OF SACRED SONG
following month. But for my father's intervention the hymn would in all probability never have been preserved."
In her early life, Mrs. Luke had decided to go out to India to engage in missionary work, but was prevented owing to a breakdown in health. Jemima Luke was born in Islington, August 19th, 1813. In 1843 she was married to the Rev. Samuel Luke, a Congregational minister. Though the authoress of "I think when I read that sweet story of old," had a ready pen, she did not produce another hymn of equal merit. Mrs. Luke died at Newport, Isle of Wight, February 2nd, 1906, in her ninety-third year.
A popular children's hymn, which is still a favourite with many of us who are no longer young, is:
"Tell me the old, old Story
Of unseen things above;
Of Jesus and His glory,
Of Jesus and His love."
It was written by Miss Katherine Hankey, over sixty years ago. Originally, "The old, old story" was a poem running to about fifty verses, of four lines each, in two parts. "The story wanted," formed Part I, and was made up of eight stanzas, which formed the hymn as we now know it. Part II is the answer to the request contained in the first part, and is entitled "The story told." It is a remarkable circumstance that the first part, which is really the request for the story, should become the favourite hymn instead of the story itself. Possibly because the introductory verses, which far surpass the second part, so beautifully express the feelings which are experienced by most of us. "I wrote Part I towards the end of January, 1866," says the writer, "I was unwell at the time—just recovering from a severe illness—and the second
verse really indicates my state of health, for I was literally 'weak and weary.' When I had written the first part, which consisted of eight verses, I laid it aside, and it was not until the following November that I completed the whole hymn." It is claimed that "Tell me the old, old story" has been translated into more languages than any other children's hymn. Its great popularity is due in no small measure to its close identification with Ira D. Sankey's Sacred Songs and Solos.
In Dr. Doane's arrangement, the original four-line verses have been turned into eight-lined stanzas, with the now familiar chorus added. Miss Hankey had already set a simple little tune to the words, and did not at first favour the setting of the American composer, as she considered that each four-lined verse was complete in itself, there being no connecting links between any two of the verses. There is no doubt, however, that but for Dr. Doane's tune, this hymn would not have attained the world-wide popularity it enjoys to-day. Writing to his friend Sankey, some years later, Dr. Doane has this to say about the music, and the occasion on which he composed it: "In 1867 I was attending the International Convention of the Young Men's Christian Association, in Montreal. Among those present was Major-General Russell, then in command of the English forces during the Fenian excitement. He arose in the meeting and recited the words of this song from a sheet of foolscap paper—tears streaming down his bronzed cheeks as he read. I wrote the music for the song one hot afternoon while on the stage-coach between the Glen Falls House and the Crawford House in the White Mountains. That evening we sang it in the parlour of the hotel. We thought it pretty, although we scarcely anticipated the popularity which was subsequently accorded it."
Many remarkable stories have been told of the wonderful influence "Tell me the old, old story" has exerted. One striking illustration is given in Sankey's volume in which he tells of a young stock-broker, utterly broken in life through gambling and drunken dissipation, who was brought to Christ through hearing a vast audience singing:
"Tell me the story softly,
With earnest tones and grave;
Remember! I'm the sinner
Whom Jesus came to save."
A hymn which was specially written over 70 years ago, by a Sunday School teacher for the very tiny members of her class, is:
"Jesus loves me! this I know,
For the Bible tells me so;"
The writer is Miss Anna B. Warner, an American lady. It was her invariable custom to write for her scholars a new hymn every month. Selecting a tune with which the children were familiar, Miss Warner would write words to suit the melody. "Jesus loves me" was one of these hymns. Soon after it was written, the hymn came into the hands of William B. Bradbury, who composed the tune to which it has since been sung. This hymn very quickly came into favour, not only in America but in this country, and it would be difficult to find a hymn book with a section for children's hymns which does not contain Anna B. Warner's little hymn. The Rev. Dr. Jacob Chamberlain, who spent a number of years among the Hindus, relates that many years ago he translated into Telugu, "Jesus loves me," and taught it to the children of the day school. "Scarcely a week later," he writes,
“as I was going through the streets of the native town on horseback I heard singing, which came from a side street. I stopped to listen, cautiously drawing to a corner, where, unobserved, I could look down the street and see and hear. And there was a little heathen boy, with heathen men and women standing around him, singing at the top of his voice:
“Jesus love me! this I know,
For the Bible tells me so.
Yes, Jesus loves me!
The Bible tells me so!”
As he completed the verse, some one asked the question: ‘Sonny, where did you learn that song?’ ‘Over at the missionary school,’ was the answer. ‘Who is that Jesus, and what is the Bible?’ ‘Oh, the Bible is the book from God, they say, to teach us how to get to heaven; and Jesus is the name of Him who came into the world to save us from our sins. That is what the missionaries say.’ And so the little boy went on—heathen himself, and singing to the heathen—about Jesus and His love. ‘That is preaching the Gospel by proxy,’ I said to myself, as I turned my pony and rode away, well satisfied to leave my little proxy to tell his interested audience all he himself knew, and sing to them over and over that sweet song of salvation.”
As a writer of children’s hymns, no name has attained so high a position as that of Mrs. Cecil Frances Alexander. Choicest, and best of all her many beautiful compositions may be placed that tender and inspiring hymn:
“There is a green hill far away,
Without a city wall,
Where the dear Lord was crucified,
Who died to save us all.”
Like the authoress of "Jesus loves me," Mrs. Alexander usually wrote her children's hymns for her Sunday School class, who first heard the verses read over to them before they were given out to the world. Mrs. Alexander was the daughter of Major Humphries of the Royal Marines, and was born in Ireland, in 1823. In 1850 she married William Alexander, at that time rector of a country parish in the county of Tyrone. During the five years spent here, Mrs. Alexander's great joy was to visit the poor and needy of her husband's parishioners. Speaking of these days, the old parish clerk was wont to tell of the lady who went with comforts for the sick and sorrowful in all weathers, "when it was not fit for the likes of her to be out!"
Of a kind and benevolent disposition, Mrs. Alexander was beloved by all with whom she came in contact. "From one poor home to another," wrote her husband in his biography of her, "from one bed of sickness to another, from one sorrow to another, she went. Christ was ever with her, and in her, and all felt her influence." It is related that "There is a green hill far away" was composed while Mrs. Alexander sat by the bedside of a sick girl, as she hovered between death and life. It was written in 1847, and first appeared in *Hymns for the Little Children*, a little volume published by the authoress. In 1867, Mrs. Alexander's husband was appointed Bishop of Derry and Raphoe, and ultimately Primate of all Ireland. But their social and ecclesiastical elevation did not deter her from ministering to the needs of the deserving poor in the new sphere to which she had been called. After a long and useful life Mrs. Alexander went to her rest in 1895, at the age of seventy-two. Though the authoress of "There is a green hill" wrote many hundreds of hymns and poems, it is as a writer of hymns for children that the name of Mrs. C. F. Alexander will always be remembered and
loved. Other children’s hymns from the same facile pen, to be found in most hymnals, which have won their way to the hearts of the young, are, “Do no sinful action,” “Day by day the little daisy,” and the beautiful hymn on the birth of Christ:
“Once in royal David’s city,
Stood a lowly cattle shed,
Where a mother laid her baby
In a manger for a bed.
Mary was that mother mild,
Jesus Christ her little child.”
Also the following little hymn, which is characterised by the same simplicity and beauty:
“We are but little children weak,
Nor born in any high estate,
What can we do for Jesus’ sake,
Who is so high and good and great?”
Julia A. Carey, an American hymnist, gave to us that simple little hymn, in language eminently suited for the very tiny young folks:
“Little drops of water,
Little grains of sand,
Make the mighty ocean,
And the beauteous land.”
Written in 1845, while the authoress was a teacher in the Boston Primary Schools, U.S.A., it was first published in the Gospel Teacher. A sixth verse, completing the sentiment of the hymn, was added by Bishop Bickersteth:
“Little ones in glory,
Swell the angels’ song:
Make us meet, dear Saviour,
For their holy throng.”
"Lord a little band and lowly," is a delightful little hymn, fragrant with sweet memories. It was written by Martha Evans Shelly (née Jackson), and first appeared in the *Child's Own Hymn Book* in 1844. It has passed into a large number of collections for children, but does not appear to attract present-day compilers in the measure it fully deserves. How this hymn came to be written is told by Mrs. Shelly: "At a Sunday School meeting in Manchester, the Rev. John Curwen, one evening, gave a lecture on singing. He sang a very pretty and simple tune, to which he said he had no suitable words, and wished that some one would write a hymn to it. I wrote these verses and gave them to him at the close of the meeting."
The tune which Mr. Curwen sang was a German composition, and was given in his *Child's Own Tune Book* under the name of "Glover." Written in a moment of inspiration, the words convey in simple language, sublime thoughts of praise and prayer.
"Lord, a little band and lowly,
We are come to sing to Thee;
Thou art great and high and holy;
O how solemn we should be!
Fill our hearts with thoughts of Jesus,
And of heaven, where He is gone;
And let nothing ever please us
He would grieve to look upon.
For we know the Lord of Glory
Always sees what children do,
And is writing now the story
Of our thoughts and actions too.
Let our sins be all forgiven;
Make us fear what e'er is wrong;
Lead us on our way to heaven,
There to sing a nobler song."
A hymn of exquisite beauty, and a special favourite with the little ones as an evening prayer is, "Jesus,
tender Shepherd, hear me.” It was written by Mrs. Mary Duncan, a beautiful and accomplished young lady. She was the daughter of Robert Lundie, the parish minister of Kelso, and was born at the manse there, April 26th, 1814. In July, 1836, she was married to William Wallace Duncan, minister of Cleish, Kinross-shire. This tender hymn was written for her own little children,—doubtless with no thought of publicity—about three years after her marriage, when Mrs. Duncan was barely twenty-five, a fact which gives a special interest to the composition. Toward the close of December, 1839, she contracted a severe cold which developed into pneumonia, and on January 5th, just a few months after the hymn was written, the young and beautiful life ended. Though not so widely known as other children’s hymns, “Jesus, tender Shepherd, hear me,” is universally loved all over Scotland, and the words of the first verse are still lisped by the little ones as an evening prayer. There is an unmistakable beauty in its simplicity:
“Jesus, tender Shepherd, hear me;
Bless Thy little lamb to-night;
Through the darkness be Thou near me;
Watch my sleep till morning light.”
This beautiful little prayer, in poetic form, reminds us of a similar composition, perhaps more widely known. Charles Wesley, whose prolific pen gave to the world the immortal “Jesus lover of my soul,” also wrote the sweet and tender hymn, which for more than a century has been used in countless homes by children of many generations:
“Gentle Jesus, meek and mild,
Look upon a little child,
Pity my simplicity,
Suffer me to come to Thee.”
Fain I would to Thee be brought;
Blessed Lord, forbid it not;
In the kingdom of Thy grace
Give a little child a place."
This was one of Wesley's earliest hymns, having been written about the year 1740; and though it has been repeatedly stated that the hymn was expressly written for his own children, such is not the case, as Charles Wesley was not married until many years after its composition.
F. A. Jones tells the story of an old man, over eighty years of age, who, when he lay dying, endeavoured in vain to recall a single prayer or hymn which might help to comfort him in his journey into the unknown. Since the age of twenty he had lived a godless life, forgetting the truths imbibed in his earlier days. Suddenly his vision cleared, and he saw himself a little lad again, kneeling at his mother's knee, repeating his evening hymn; unconsciously from his lips issued these tender words which for nearly seventy years he had neither uttered nor heard—"Gentle Jesus, meek and mild." But is it not the same with many of us? Many summers may have passed since those bygone childhood days, and yet with us there still linger cherished memories of the time when we, too, knelt by a mother's knee, and repeated the same old familiar lines.
CHAPTER VIII.
How we got our Popular Gospel Song
WHILE psalms and hymns stretch far back across the centuries, one department of it is of fairly recent origin. Gospel singing, as has already been observed, is a modern institution, and America can rightly be claimed as its birth-place. In the days of long ago the negroes of the Southern States chanted their own simple melodies, and in their own crude way were expressing the feelings and aspirations of the heart toward God. This particular type of negro is naturally a lover of song, and when, on occasions, he came to express his emotions, words and music followed with remarkable smoothness. From these old melodies, musicians of a later day have drawn an inspiration, as in the case of Ira D. Sankey's famous hymn "The ninety and nine," the melody of which is said to have had its origin in an old Southern plantation song, called "A wonderful stream is the river of time." The extraordinary way in which Sankey composed it anew to Miss Clephane's little poem is referred to in another chapter.
The question may arise: What constitutes an acceptable and useful Gospel song? The views of the writer are so well delineated by the graphic pen
of Charles H. Gabriel of America, the King of Gospel Hymn Writers, that we shall borrow his language: "First, the text must be systematically constructed, be spiritual and devotional; it should begin with an immediate declaration of subject, followed by an explication presented in a logical and intelligent manner. Gospel music is the language of the heart; the expression of hope, trust, longing, sorrow, joy, and even despair of the soul. It is the spontaneous overflow of happiness and a healing balm for the wounds of life; it is both sermon and song, praise and prayer, oblivion and remembrance."
No one would venture to assert that American hymnology contains anything to be compared with the masterpieces of English collections. Philip Phillips, popularly known as the "Singing Pilgrim," in the early seventies first brought to this country from America this new class of sacred music, which was subsequently used by Moody and Sankey in the early days of their British campaign. It may not be generally known that Newcastle-on-Tyne was the birth-place of the now famous Sankey hymn book. The hymns and tunes used at that period in the various places of worship in this country did not appear to be adapted for evangelical services, and the American evangelists adopted Philip Phillips' book, which contained many American hymns and some English tunes. This collection started in a very modest form, but in course of time, as new hymns were being written, fresh editions were published, until *Sacred Songs and Solos* gained a popularity among the masses that few of its successors enjoyed.
Writing at a later period, Sankey gives some interesting information bearing on the early days of his hymn book. The evangelists had only been a short time in England, and were conducting meetings
HOW WE GOT OUR POPULAR GOSPEL SONG
in the provinces, when Sankey wrote to London offering to give his selection of songs to the publishers of Philip Phillips' hymn book, *Hallowed Songs*, free of charge if they would print them. This they respectfully declined. About this time, Mr. R. C. Morgan, of *The Christian*, on hearing of the refusal of the other publishers to accept the hymns, offered to take them and publish them in pamphlet form. So Mr. Sankey cut from his scrap book twenty-three pieces, rolled them up, and wrote on them the words "Sacred Songs and Solos, sung by Ira D. Sankey at the meetings of Moody of Chicago." "This book, together with the edition of words only," says Mr. Sankey, "has now grown into a volume of twelve hundred pieces, and up to the present time has possibly the largest sale of any book except the Bible."
It is interesting to learn that, even in these days of many hymn books, *Sacred Songs and Solos* has not diminished in public favour, for up to the end of last year (1930), the sale had reached the unparalleled total of more than seventy million copies.
The visit of the American evangelists to this country, when the famous Sankey hymn book was first introduced to the world, received tremendous publicity. Wherever they went their reception was usually of the most cordial nature, while their hymns were sung and referred to in the most unlikely places. For instance, during a circus performance in Dublin, one clown, with a pretended air of dejection, said to another, "I say, I feel quite Moody to-night; how do you feel?" "Oh," responded the other, "I'm rather Sankeymonious." Contrary to what was expected, this by-play was not only met with hisses, but the whole audience rose to their feet and joined with tremendous effect in singing, "Hold the fort for I am coming!"
A story is told connected with a visit, about this
particular time, of the late Professor Blackie to Carlisle. Upon his departure from the Border city, the professor was given by his hostess a parcel, with strict injunctions not to open it until well on his journey; when he did so, he discovered it contained some bread and cheese, and one of Moody and Sankey's hymn books. Professor Blackie proved equal to the occasion, and replied to his hostess:
"For the body, cheese,
For the soul, Sankey;
For both of which, madam,
I heartily thank ye:
And blessed be she
Who did what she could,
To make a lean man fat,
And a bad man good."
Sankey's faith in the power of sacred song was fully rewarded, for he lived to see these songs make their way into the hearts of millions of people.
Among writers who laid the foundation of American Gospel hymnody the following names are familiar: W. B. Bradbury, P. P. Bliss, Philip Phillips, Ira D. Sankey, James McGranahan, W. J. Kirkpatrick, Robert Lowry, George C. Stebbins, H. R. Palmer, D. W. Whittle, T. C. O'Kane, J. R. Sweney, W. H. Doane, Fanny Crosby, E. O. Excell and Charles H. Gabriel. Of this group of sweet singers, whose songs have been carried to the ends of the earth, Mr. Stebbins and Mr. Gabriel alone remain.
Foremost amongst the hymn writers of half a century ago, when this new type of Gospel song came into being, consequent upon a wave of spiritual awakening throughout the United States of America, comes the name of Philip Paul Bliss, author of "Hold the Fort," "Whosoever will," "Almost persuaded," "Man of sorrows," "Free from the law," and a host of other
MAJOR W. D. WHITTLE.
W. J. KIRKPATRICK. P. P. BLISS.
C. H. GABRIEL. J. McGRANAHAN.
J. R. SWEENEY.
popular songs sung the world over. Of Puritan ancestry, Mr. Bliss was born in the forest and mountain region of Northern Pennsylvania on July 9th, 1838. His parents were poor and he had little help in the battle of life, but he won the victory. From the beginning, a love of song grew with his years, and instinctively his childish ear was caught by any note of nature. Thus we find Philip drawing forth notes from the reeds which grew near his father's house, and at the age of seven he was making for himself crude instruments in a most original fashion. When a boy of ten summers, he heard a piano for the first time. In later years, at a musical conference which he addressed, Mr. Bliss, speaking of his early days, told a very impressive story. "A barefooted mountain lad had gone, as was his custom, to the little village with his basket of fresh vegetables, which he peddled from door to door. One day, having sold his stock, he was on his way home, when the sound of music was wafted to his ear through the open door of a house by the way; he paused; the music continued, and drew him nearer, and nearer, until, unconsciously, he had entered the room where a lady was playing a piano accompaniment to the song she was singing. Entranced, he stood listening, his very soul lost in a sea of delight; such music he had never before heard. Some movement of his attracted the lady's attention; she turned, and seeing the boy, with a little scream of surprise cried out: 'What are you doing in my house? Get out of here with your great bare feet.'" Looking down at his feet as he told the story, Mr. Bliss continued, slowly: "Yes, my feet are large—but God gave them to me; and how I wish that that lady's children were here that I might sing to them."
Towards the end of 1857, when but a youth, Mr. Bliss attended a Musical Convention at Rome in Pennsylvania. It was here that he made the acquaintance of Mr. William B. Bradbury, a pioneer of American Gospel hymnody, and the author of many popular pieces of sacred song. The Convention turned the thoughts of Mr. Bliss to the subject of writing hymns, and we find it was about this time that he produced his first composition, which was to be followed by so many powerful sermons in song to be used for the glory of his Master in days to come. In the year 1869 an event occurred which Mr. Bliss regarded as the most important in his life; this was his meeting with Mr. D. L. Moody, who was then holding Gospel services in Chicago. Being possessed with a sweet, sympathetic bass voice of splendid tone and quality, Mr. Bliss's powerful singing at once attracted the attention of the evangelist. This memorable meeting constituted an epoch, for, from the time Mr. Moody met with Mr. Bliss, dates his impression of the unmistakable power of solo-singing in these evangelistic labours. Why should it be considered strange, he reasoned, that singing by a man of taste and musical ability, should soften the heart, and by God's blessing, break the hard crust of worldliness that may have gathered over the soul? It was only logical to suppose that if the voice of powerful speaking can arouse the conscience, why may not singing do the same? And thus, the vivid impression of the power of Gospel song which Mr. Moody received when he met Mr. Bliss, forms an epoch in the history of a movement that has been among the most blessed and remarkable during the last half-century, and has, to a great extent, changed the nature of religious meetings in all parts of Christendom. From a letter written by Mr. Bliss about this time, we take the following: "This singing and talking about the Good News of a present, perfect, free salvation, and justification by faith, is so popular and attractive, I do not
believe I shall ever find time for anything else. It seems to me it is needed. How much of everything else we hear preached, and how little Gospel!" Soon after this Mr. Bliss made the acquaintance of Major D. W. Whittle, an earnest evangelist, with whom he laboured for some time in the city of Chicago. The major took Mr. Bliss to stay with him at his house, number 43, South Street, and it was here that Mr. Bliss wrote the words and music of the popular hymn:
"I am so glad that our Father in heaven
Tells of His love in the Book He has given;
Wonderful things in the Bible I see;
This is the dearest that Jesus loves me."
—and of that stirring martial strain, familiar now to tens of thousands, "Hold the fort, for I am coming!"
Mr. Bliss realised more than any one the importance of storing the young mind with Gospel song, and his ability to teach children to sing a new hymn was amazing. From the moment he named a piece he seemed to inspire all with his enthusiasm, and while he led in some bright song, not an eye would wander from him, nor a face be dull. "He would say a few pithy words," writes one who was present at one of these memorable services, "explaining the sentiment of the song, a few more, possibly about the music and how to render it; sing a strain or two alone, then, after two or three repetitions, the school would march through and ring it out as if they had been familiar with it for months. It was as if he had the gift of infusing music into everybody. No matter how little musical culture or skill teachers and scholars had, no matter how out of key or out of time, they were naturally inclined to sing. Somehow, when Mr. Bliss led, the difficulties and irregularities and discords seemed to disappear, and
there was one grand thrill of feeling, one royal burst of harmony."
Altogether, Mr. P. P. Bliss wrote seven books of Sacred Songs, besides various contributions to musical journals. To him song-writing was a spontaneous outflow of the emotions and melody with which his soul was filled. When he found that God was using his songs to bring out some precious truth of the Gospel of love, or the exaltation of Christ his Lord, his heart would overflow with joy. Often he would come to his wife with the theme of a hymn, with his face shining and his eyes moist with tears, and would ask for prayers that God would bless the song.
Mr. Bliss had indeed a wonderful gift, and not only could he write stirring Gospel songs, but with very little effort he could compose appropriate and attractive music for them, which had much to do with their immediate popularity. As we have already seen, the songs were first introduced to the British public by Mr. Sankey during the great revival services held throughout the United Kingdom in 1873-74. Indeed, the greater number of the Gospel songs and solos sung by Mr. Sankey in this country were the compositions of Mr. Bliss.
Dr. F. W. Root, a celebrated composer and friend of Mr. Bliss, gives a very graphic pen-picture of this notable hymn writer. "He was a poet-musician," he writes, "and if ever a man seemed fashioned by the Divine hand for special and exalted work, that man was P. P. Bliss. He had a splendid physique, a handsome face, and a dignified, striking presence. It sometimes seemed incongruous, delightfully so, that in one of such great size and masculine appearance there should also appear such gentleness of manner, such perfect amiability, such conspicuous lack of self-assertion, such considerateness and deference to all,
and such almost feminine sensitiveness. He had not had opportunities for large intellectual culture, but his natural mental gifts were wonderful. His faculty for seizing upon salient features of whatever came under his notice amounted to an unerring instinct. Mr. Bliss's voice was always a marvel to me. He used occasionally to come to my room, requesting that I would look into his vocalization with a view to suggestions. At first a few suggestions were made, but latterly I could do nothing but admire. Beginning with E flat, or even D flat below, he would, without apparent effort, produce a series of clarion tones, in an ascending series, until, having reached the D (fourth line tenor clef), I would look to see him weaken and give up, as would most bass singers; but no, on he would go, taking D sharp, E, F, F sharp, and G, without weakness, without throatiness, without sound of straining, and without the usual apoplectic look of effort. I feel quite sure in saying that his chest range was from D flat below to A flat above, the quality being strong and agreeable throughout, and one vowel as good as another. He would have made a name and fortune on the dramatic stage."
On December 29th, 1876, Philip P. Bliss met his death in a railway disaster. He was travelling toward Chicago, when at Ashtabula, Ohio, a bridge gave way and the whole train was thrown into the stream below. Mr. Bliss might have escaped, but in an endeavour to rescue his wife from the flaming car he lost his life. Had it not been for that eventful wreck, what songs he might have given to the world, for he then was but 38 years of age.
CHAPTER IX.
Hymn Writers of America
In the previous chapter the endeavour has been to set forth as simply and lucidly as possible, the story of how we got our popular Gospel song. To do so, one of the pioneers of the work in this particular realm of hymnody was singled out and used as a vehicle by which, as we look back and see this wonderful movement spreading from country to country, and being used of God, we can, in a greater measure, comprehend its remarkable development and far-reaching results. This chapter will be devoted to sketching a brief history of some of the best known Gospel hymn writers of America. As a great number of this particular type of hymn were written during the period with which we have already dealt, this chapter, while running in sequence affords another opportunity to gather still further from the fertile fields of Gospel hymnody of America.
Among the sweet singers of the last two generations, the name of Frances Jane Crosby is a familiar one on both sides of the Atlantic. Her hymns are sung everywhere, and by Christians of all denominations, because they so beautifully express the wonderful spiritual influence upon the heart and life. She was born in the town of Southeast, Putnam County, New York, on March 24th, 1820, and her parents greatly rejoiced because God had sent them a babe. But in a few weeks their joy was turned to sorrow, for the little girl’s
eyes began to inflame; the physician made a mistake in his treatment of the trouble, and as the result, Fanny Crosby became hopelessly blind. Fanny, however, was an optimist, dwelling little on her affliction, and not one word did she ever speak of blame for the physician who had been responsible for her sad plight. "It may have been a blunder on the physician's part," she said in later years; "it was no mistake of God's." Among her greatest blessings she has always counted instruction in the Bible, and when only ten years old, she was able to recite the first five books of the Old Testament, and the first four in the New Testament. At the age of fifteen she was taken to the New York school for the blind, the first of its kind in the country. Always passionately fond of poetry, Fanny began writing verses when quite young, and when she was twenty-four years of age she timidly gathered up a few of her early poems and had them published under the title, *The Blind Girl, and other Poems*. The book attracted the notice of the celebrated American poet, William Cullen Bryant, who visited Fanny and spoke encouragingly to the young writer. Other volumes followed in quick succession, but it was not till 1864 that she began the great life-work for which she had been unconsciously preparing, the writing of hymns. "I verily believe," she wrote, "that it was God's intention that I should live my days in physical darkness, so as to be better prepared to sing His praise and incite others so to do. I could not have written thousands of hymns if I had been hindered by the distractions that would have been presented to my notice." The year 1850 was a memorable one, for it was the year of her conversion and consecration to God's service. Gospel meetings were being held in a Methodist Church near by. "Some of us," she writes, "went every evening, but although I sought peace, I could not
find the joy I craved until one evening—November 20th, 1850—I arose and went forward alone. After prayer the congregation began to sing the grand old consecration hymn of Dr. Isaac Watts:
‘Alas and did my Saviour bleed?
And did my Sovereign die?
Would He devote that sacred Head
For such a worm as I?’
And when they reached the third line of the last verse:
‘Here, Lord, I give myself away;
'Tis all that I can do’
—I surrendered myself to the Saviour, and my very soul was flooded with celestial light. I sprang to my feet, shouting ‘Hallelujah!’” Miss Crosby was married in 1858 to Mr. Alexander Van Alstyne, a teacher in the institution for the blind, but her husband wisely and unselfishly urged her to retain her maiden name in connection with her work, and always took a great interest in the genius of his wife.
Fanny Crosby, by which name she has always been known, first met Ira D. Sankey early in his career, and he set to music several of her hymns and used them with remarkable success in his evangelistic missions with D. L. Moody, which stirred the whole world for so many years. The singing evangelist she revered as God’s messenger of living song. They worked together with unison of soul, for the uplift of the downcast by means of sacred song. Mr. Sankey was very fond of the blind poetess, and to the day when he himself lost his sight, there were few who brought more joy and peace to his heart than Fanny Crosby.
Altogether she has written about seven thousand hymns, and some of the best known are: “Safe in the
arms of Jesus,” “Rescue the perishing,” “Jesus keep me near the Cross,” “Blessed Assurance,” “I shall know Him,” “I am Thine, O Lord,” “Saved by Grace.”
A remarkable circumstance, not generally known, in connection with this gifted authoress’ hymn writing, is that she rarely composed her verses without a small book or Testament held open before her eyes.
When Fanny Crosby was forty-eight years of age, Dr. W. H. Doane, who has written so many beautiful hymn tunes and was a constant friend of the blind hymn writer, came one day into the office of Biglow and Main in New York, and finding Fanny Crosby there in conversation with Mr. W. B. Bradbury, he said to her, “Fanny, I have just written a tune, and I want you to write a hymn for it.” “Let me hear it,” she replied. After he had played it over for her on a small organ, she exclaimed, “Why, that tune says, ‘Safe in the arms of Jesus,’ and I will see what I can do about it.” She at once went into an adjoining room, and in half an hour returned and repeated to him the words of the hymn:
“Safe in the arms of Jesus,
Safe on His gentle breast,
There by His love o'er shaded,
Sweetly my soul shall rest.
Hark! 'tis the voice of angels
Borne in a song to me,
Over the fields of glory
Over the jasper sea.”
It was first published in Dr. Doane’s book, entitled Songs of Devotion, in 1868. The hymn obtained great favour at once, and is said to be one of the first hymns of its kind the words of which were translated into a foreign language.
There is also her sweet consecration hymn which seems to make the deepest chords of one’s spiritual
nature vibrate, as with a touch of Heaven's own influence:
"I am Thine, O Lord, I have heard Thy voice,
And it told Thy love to me;
But I long to rise in the arms of faith,
And be closer drawn to Thee.
Draw me nearer, nearer, blessed Lord,
To the Cross where Thou hast died,
Draw me nearer, nearer, nearer blessed Lord,
To Thy precious bleeding side."
At the age of seventy-one Fanny Crosby wrote the hymn "Saved by Grace" which has now become so widely known. The story of how it came into being is worthy of relating here. While visiting Mr. Sankey at Northfield, Mass., where the summer conferences were held, Mr. Sankey asked her to make a short address to her many friends, as a message had been sent in by some of those present that they wished to hear her speak. She at first begged to be excused, but on further persuasion, consented to speak a few words. Mr. Sankey led her forward to the desk on which lay the Bible, and after speaking eloquently for a short time, Fanny closed her remarks by reciting this beautiful hymn, beginning:
"Some day the silver cord will break,
And I no more as now shall sing,
But O, the joy when I shall wake
Within the palace of the King!"
The chorus is all the more affecting remembering the blindness of the writer:
"And I shall see Him face to face,
And tell the story, saved by grace."
When she had finished, Mr. Sankey turned to her and said, "Why, Fanny, where did you get that beautiful hymn?" "You ought to know," was her reply
IRA D. SANKEY AND FANNY CROSBY.
HYMN WRITERS OF AMERICA
"These are the verses I gave you three years ago, and which I suppose have been put away in the safe with all my other unpublished hymns." The poetess had a wonderful memory, and although she had composed several hundred hymns since writing the words of "Saved by Grace," she was able to recite the hymn, word for word, exactly as she had penned it. Mr. George C. Stebbins set the words to an attractive tune, with the result that "Saved by Grace" has become one of the most useful and popular of modern hymns. The last verse is especially beautiful:
"Some day, when fades the golden sun
Beneath the rosy-tinted west,
My blessed Lord will say, 'Well done,'"
And I shall enter into rest."
That "Some day" came on February 11th, 1915, when the sweet singer of America entered into "The Palace of the King" at the age of ninety-five.
Mention has already been made of Philip Phillips, known in the early days as the "Singing Pilgrim." Though he laid no claim to being an educated musician, nevertheless he has composed many beautiful tunes, and being possessed of a voice of peculiar influence, he was able to impress his earnestness upon his hearers. As has been noted elsewhere, Philip Phillips was among the first to broadcast the Gospel by song, previous to the world-wide work of Moody and Sankey. Probably the hymn most closely associated with his name is, "The home of the Soul." There is a striking incident related of him when on a visit to the Holy Land, which will illustrate the wonderfully magnetic influence of Philip Phillips' powerful singing. "One day, as they wandered amid scenes of sacred memory, they were startled by the cry of 'Unclean! unclean!' Looking up the mountain side they plainly saw the
lepers. After watching them a moment Mr. Phillips began singing that peerless song:
‘I will sing you a song of that beautiful land
The far-away home of the soul,
Where no storms ever beat on the glittering strand
While the years of eternity roll.’
In silence those helplessly afflicted human beings stood listening as he sang stanza after stanza. They understood not a word he said, possibly, but the soul of the singer seemed to speak in a language that even they could comprehend, and in respectful attention they stood until the song ended, and the little party of American tourists passed from their view.”
The words of “The Home of the Soul” were written by Mrs. Ellen H. Gates, an American lady, after reading the narrative of Christian and Hopeful at the gate of Heaven in Bunyan’s Pilgrim’s Progress. “When the verses were forwarded to me, in 1865,” wrote Mr. Phillips, “I seated myself in my home with my little boy on my knee, and with Bunyan’s immortal dream-book in my hand, and began to read the closing scenes where Christian and Hopeful entered into the city; wondering at Bunyan’s rare genius, and like the dreamer of old, wishing myself among them. At this moment of inspiration I turned to the organ, with pencil in hand, and wrote the tune. This hymn seems to have had God’s special blessing upon it from the very beginning. One man writes me that he has led in the singing of it at a hundred and twenty funerals. It was also sung at the funeral of my own dear boy, who had sat on my knee when I wrote the tune.” This hymn was sung by Mr. Sankey over the remains of his beloved friend, Philip Phillips, the composer, at Fredonia, New York.
A hymn to be found in most mission hymn books has this chorus:
HYMN WRITERS OF AMERICA
"Moment by moment I'm kept in His love;
Moment by moment I've life from above;
Looking to Jesus till glory shall shine;
Moment by moment, O Lord, I am Thine."
It was written by Major D. W. Whittle, a close friend and colleague of P. P. Bliss, and the author of a considerable number of hymns in popular use to-day. A remark by Henry Varley, the evangelist, to the effect that he did not very much like the hymn, "I need Thee every hour," "Because," said he, "I need Him every moment of the day," gave the idea of the hymn "Moment by moment" to Major Whittle. His daughter, May Whittle, who afterwards became the wife of Will R. Moody, composed the music.
Daniel Webster Whittle was born at Chicopee Falls, Massachusetts, November 22nd, 1840. At the outbreak of the Civil War, in 1861, he enlisted in the 72nd Illinois Infantry, and saw much active service. At one time he was Provost Marshal on General Howard's staff, and was with Sherman in his famous march to the sea. It was after a severe engagement at Vicksburg in which he lost his right arm, and was taken prisoner by the enemy, that he was awakened to see his need of a Saviour. While he was recovering from his wound, and having a desire for something to read, he felt in his haversack and found the little Testament his mother had placed there on the morning of his departure for the war. For the first time he opened its pages. He read right through the book several times. Every part was interesting to him, and he found to his surprise that he could understand it in a way that he never had before. He understood the presentation of the truth in Paul's letter to the Romans and had it plainly before his mind that God gave Jesus, His Son, to be our Substitute, and that whoever would confess their sins and accept Him, should be saved. While in
this state of mind, yet with no fixed purpose to repent and accept the Saviour, he was awakened one midnight by an orderly, who said: "There is a boy in the other end of the ward who is dying. He has been begging me for the past hour to pray for him, but I'm a wicked man and cannot." "Why," said Whittle, "I can't pray. I never prayed in my life. I am just as wicked as you are." "Can't pray!" said the orderly. "Why, I thought sure from seeing you read the Testament, that you were a praying man. I can't go back there alone. Won't you get up and come and see him at any rate?" Moved by his appeal young Whittle arose from his cot and went with the orderly to the far corner of the room. A fair haired boy of seventeen or eighteen lay there dying. There was a look of intense agony upon his face, as he cried: "Oh, pray for me! Pray for me! I am dying. I was a good boy at home in Maine and went to Sunday School. But since I became a soldier I have learned to be wicked. And now I am dying, and I am not fit to die! Oh, ask God to forgive me! Ask Christ to save me!" "I dropped on my knees," said Major Whittle when telling the story, "and held the boy's hand in mine, as in a few broken words I confessed my sins, and asked God for Christ's sake to forgive me. I believe right there that he did forgive me, and that I was His child. I then prayed earnestly for the boy. He became quiet and pressed my hand as I pleaded God's promises. When I arose from my knees he was dead. A look of peace was upon his face, and I can but believe that God, who used him to bring me to the Saviour, used me to get his attention fixed upon Christ, and to lead him to trust in His precious blood. I hope to meet him in Heaven."
But for that little Testament placed in her soldier boy's haversack by that praying mother, the realm of
HYMN WRITERS OF AMERICA
hymnody would have been poorer to-day. For, though some of Major Whittle's hymns may to a certain extent lack literary merit, yet they ring true, and have been wonderfully used of God.
A testimony song, the joyous strains of which rarely fail to bring back happy memories of our first love, is still as heartily sung both at mission services and in the home as the day it was written:
"Come sing, my soul, and praise the Lord,
Who hath redeemed thee by His blood;
Delivered thee from chains that bound,
And brought thee to redemption ground."
Then there is that song of the joyous anticipation of our Lord's coming again, the chorus of which is so familiar to the ear:
"Oh, the crowning day is coming!
Is coming by and by!
When our Lord shall come in 'power'
And 'glory' from on high!
Oh, the glorious sight will gladden
Each waiting, watchful eye,
In the crowning day that's coming
By and by."
A favourite at Prayer Meetings and Revival Services and one which alone would have caused the author's name to be remembered is the hymn beginning:
"There shall be showers of blessing:
This is the promise of love;
There shall be seasons refreshing,
Sent from the Saviour above.
Showers of blessing,
Showers of blessing we need;
Mercy-drops round us are falling,
But for the showers we plead."
This hymn owes much of its popularity to an appropriate tune by Mr. James McGranahan, a colleague of Sankey, and one of the foremost composers of
Gospel music of his day. Other familiar hymns by Major Whittle, include: "The love that gave Jesus to die," "Jesus is coming," "I know whom I have believed," "I looked to Jesus," "There's a royal banner," and "Come believing." Many of Major Whittle's hymns were written over the *nom de plume* "El Nathan." He passed away at Northfield, March 4th, 1901.
Among mission hymns there is one that never seems to grow old. For well nigh half a century it has been sung at Gospel meetings, in church and chapel, in mission hall and at open-air services, wherever the Gospel has been proclaimed. At home, or in lands beyond the sea, its joyful strain has ever been in the ascendant:
"Have you been to Jesus for the cleansing power?
Are you washed in the blood of the Lamb?
Are you fully trusting in His grace this hour?
Are you washed in the blood of the Lamb?"
Elisha A. Hoffman, the composer of both words and music, was born at Orwigsburg, Pa., on May 7th, 1839. His father was a minister of the Gospel in the Evangelical Association, and the son followed actively in his footsteps. Mr. Hoffman was both poet and musician, and more than two thousand hymns have come from his pen, some having been translated into many languages and sung all round the world. Almost every one is familiar with the following hymns by Mr. Hoffman: "The Lord is coming by and by," "What a wonderful Saviour," "Abundantly able to save," "Resting in the everlasting arms," "Where will you spend Eternity?" and that sweetest of all prayer meeting hymns:
"'Tis the blessed hour of prayer, when our hearts lowly bend,
And we gather to Jesus, our Saviour and Friend;
If we come to Him in faith, His protection to share,
What a balm for the weary! Oh, how sweet to be there."
HYMN WRITERS OF AMERICA
A hymn better known in America than in this country is: "I must tell Jesus." Regarding its origin, Mr. Hoffman relates a very interesting story: "While I was pastor at Lebanon, Pa., I called one day at the home of a parishioner and found the lady in great distress and sorrow. Wringing her hands she cried: 'What shall I do, what shall I do?' I replied: 'You cannot do better than take it all to Jesus—you must tell Jesus.' For a moment she seemed abstracted in meditation, then her face glowed, her eyes lighted up, and with animation she exclaimed: 'Yes, I must tell Jesus, I must tell Jesus!' As I went from that sorrow-filled home a vision walked before me, a vision of a joy-illumined face, of a soul transformed from darkness into light, and I heard all along my pathway the echo of a tender voice saying, 'I must tell Jesus.'" Immediately on reaching his study Mr. Hoffman wrote both the words and music of the hymn "I must tell Jesus," which has brought comfort and solace to many a weary heart. Here is the first verse:
"I must tell Jesus all of my trials;
I cannot bear these burdens alone,
In my distress He kindly will help me,
He ever loves and cares for his own."
One of my earliest recollections of Gospel song was hearing my mother sing in the quiet of our home one of the hymns from Sankey's first hymn book, which was ever a favourite:
"Oh what a Saviour—that He died for me!
From condemnation He hath set me free;
'He that believeth on the Son,' saith He,
'Hath everlasting life.'
'Verily, verily, I say unto you;'
'Verily, verily,' message ever new!
'He that believeth on the Son'—'tis true!—
'Hath everlasting life.'"
THE ROMANCE OF SACRED SONG
Incongruous though it may seem, this chorus was used as a lullaby by which we children were put to sleep in the old rocking chair, to the song’s easy flowing measure.
It was comparatively new then. Since that far off day “Verily! Verily!” has been sung wherever Christ is preached, and now takes its place amongst the best known Gospel hymns. James McGranahan, the writer of the words and music of this hymn, was born at Adamsville, Pa., in 1840, and while a lad, learned to read music and sing alto. One of the pioneers of Gospel song, Mr. McGranahan was closely associated with Ira D. Sankey during his many Gospel campaigns. He succeeded P. P. Bliss in evangelistic work with Major Whittle, and for years his matchless tenor voice was heard in Gospel songs, both in America and England. Mr. McGranahan’s hymns were, and still are in great favour, and that too, after a generation has passed. As an illustration, an incident occurred while Mr. Moody was conducting a ten days’ mission in Spurgeon’s Tabernacle, London, toward the end of 1892. During these meetings, Mr. Moody received a request from Queen Mary, who was then Princess Mary of Teck,—who, with her mother, the Duchess of Teck, were in the audience—that Mr. McGranahan’s beautiful hymn “Sometime we’ll understand” should be sung. The request was, of course, readily complied with. This little incident is interesting for two reasons, first, as illustrating the favour with which not only Mr. McGranahan’s hymns, but the compositions of other writers of this class of hymn were received in this country; and second, as an indication that they have found their way into the palaces of royalty as well as the homes of the people, where one cannot doubt they have proved a blessing to many. Mr. McGranahan is also the composer of words and music of: “I am
HYMN WRITERS OF AMERICA
the way," "If God be for us," "They that wait upon the Lord."
A splendid musician, Mr. McGranahan is the composer of a great many well-known hymn tunes to be found in almost every evangelistic hymn book in present use. He fell asleep a few days after his sixty-seventh birthday, at his home at Kinsman, Ohio, resting upon his favourite verse "Verily, verily, I say unto you, he that believeth on Me hath everlasting life" (John vi. 47).
Among the writers of Sunday School, evangelistic and devotional music, there are few names better known than that of Robert Lowry, the author of the popular hymn "Shall we gather at the river?" He was born at Philadelphia, Pennsylvania, on March 12th, 1826, and became a true follower of the Lord Jesus at the age of seventeen. In earlier years he prepared for and entered the ministry of the Baptist denomination, his first charge being at West Chester, Pennsylvania. After many years in different pastorates, he was appointed Professor of Rhetoric in Lewisburg University. Dr. Lowry has been associated with some of the most popular Sunday School hymn books published in America. A musician of considerable ability, most of his hymns are set to music by himself, and are extremely popular. Although not by any means his best production it is in connection with the hymn "Shall we gather at the river?" that the name of Robert Lowry will always be associated. Having received the story from the lips of the author himself, Sankey tells us how the hymn came into being.
"On a sultry afternoon in July, 1864, Dr. Lowry was sitting at his study table in Elliot Place, Brooklyn, when the words of the hymn, "Shall we gather at the river?" came to him. An epidemic was raging through the city at the time, and he had been pondering the question,
‘Why do hymn writers say so much about the river of death, and so little about the pure river of the water of life?’ He hastily recorded the words, and then sat down before his parlour organ and composed the tune which is now sung in practically all the Sunday Schools of the world.”
When visiting London in 1880 on the occasion of the Raikes Centenary, when Sunday School workers from various parts of the world had come together, Dr. Lowry received a tremendous oration when introduced to the vast audience as the author of this hymn; and for some minutes it was impossible for him to speak; surely an eloquent testimony to the value of a hymn, beloved the world over. He is also the composer of the music to that martial strain: “We’re marching to Zion,” and that tender hymn, “I need Thee every hour.”
Dr. Lowry wrote the words and music of that popular mission hymn:
“What can wash away my stain?
Nothing but the blood of Jesus!”
Also that glorious resurrection hymn, the chorus of which is so well known:
“Up from the grave He arose,
With a mighty triumph o’er His foes;
He rose a Victor from the dark domain,
And He lives for ever with His saints to reign;
He arose! He arose!
Hallelujah! Christ arose!”
He is also the composer of that pathetic hymn, “Where is my wandering boy to-night?” which is said to have arrested many a wanderer on the downward road; of which the following story is a striking instance. The principal speaker at a meeting which filled to
HYMN WRITERS OF AMERICA
overflowing the Athenaeum at Bury, Lancashire, was a lady who held the vast audience enthralled by her earnestness. At the close of her discourse she said, "Before I came to Bury this afternoon, I made a call at Haslingden to see a dear old woman. I told here I was coming and for why; when, with tears in her eyes, she said, 'I wonder whether you can find my wandering boy?'""
Then came the words:
"Go for my wandering boy to-night;
Go, search for him where you will;
But bring him to me with all his blight,
And tell him I love him still."
"I wonder," continued the speaker, pausing for a moment, "whether that wandering boy is here?" "Yes!" exclaimed a voice, and from the back of the hall, through crowded aisles, a young man made his way to the front. Reaching forward, the speaker grasped the outstretched hand of the youth, and after exchanging a few words, the lady raised her hand, and amid a tense silence said, "Yes, it is that mother's boy. Do you wonder, when I tell you that before I came away we knelt down in that cottage, and prayed that this boy might be found?" The prayer was answered. That night the wandering boy returned home.
Many of Dr. Lowry's hymns were written after the Sunday evening service, when his mind refused to rest. He passed away at Plainfield, New Jersey, on November 25th, 1899, at the age of seventy-three, but will continue to preach the Gospel by his hymns long after his sermons have been forgotten.
Dr. George F. Root, the author of a favourite hymn for young folks commencing:
is much more widely known as a composer of popular music than as a hymn writer. A born musician, it is said that at the age of thirteen he could play a tune on as many instruments as he was years old. He was born at Sheffield, Mass., on August 30th, 1820. When yet in his teens he went to Boston to study music, where he associated with the celebrated Dr. Lowell Mason, composer of the music to "My faith looks up to Thee," and in a comparatively short time reached the top of his profession. Dr. Root was a voluminous writer and was contemporaneous with W. B. Bradbury in writing Sunday School music. While he wrote much for the people, as indicated in the character of his hymns, he was a musician honourably recognised by the profession, and was given the degree of Doctor of Music in 1873 by the Chicago University. Previous to devoting his musical talents to the writing of hymns, Dr. Root was everywhere recognised as America's foremost writer of war songs. The majority of these which gained popularity during the Civil War, and which in later years were adapted by other nations, were his composition. Dr. Root was a man of singularly gracious and engaging personality, and of spiritual convictions, aiming always to inspire others with high ideals both in character and art. His sacred songs, many of which are to be found in almost all present-day Gospel hymn books, have lifted and strengthened the fallen the world over. Dr. Root died on August 6th, 1895, and at his request nothing was sung at his funeral but the Doxology. Amongst his best known compositions are, "When He cometh," "Ring the bells of Heaven," "Knocking, knocking." He also
wrote words and music of "Altogether lovely," "Why do you wait?" "Narrow and strait," "Behold the Bridegroom Cometh," and that miracle pen picture:
"She only touched the hem of His garment
As to His side she stole,
Amid the crowd that gathered around Him;
And straightway she was whole.
Oh, touch the hem of His garment,"
Etc., etc.
Edgar Page Stites, the writer of the popular hymn, "Beulah Land," was for many years a prominent business man in Cape May, New Jersey. His pen name of "Edgar Page" has hidden for well nigh half a century, the real authorship of many a song that has voiced the religious enthusiasm of countless thousands. "It was in the year 1876 that I wrote 'Beulah Land,'" said Mr. Stites in a letter written after he had passed three score years and ten, "I could write only two verses and the chorus when I was overcome, and fell on my face. I could only weep and write no more. That was on Sunday. A week later I wrote the third and fourth verses, and again I was so influenced by emotion that I could only pray and weep. The first time it was sung was at the regular Monday morning meeting of Methodist ministers at Arch Street, Philadelphia, when Bishop McCabe sang it to the assembled ministers. Since then its story is known wherever religious people congregate."
"Beulah Land" is a song brimful of new found joy. Here is the chorus:
"O Beulah Land, sweet Beulah Land,
As on thy highest mount I stand,
I look away across the sea,
Where mansions are prepared for me,
And view the shining glory shore;
My heaven, my home for evermore."
THE ROMANCE OF SACRED SONG
During some correspondence in 1913, Mr. Stites sent me a favourite verse of "Beulah Land" in his own handwriting, which I am able to have reproduced here.
The Saviour comes and walks with me
And sweet communion here I see
He gently leads me with his hand
For this is heaven's border land
Edgar P. Stites
He was then seventy-six years of age, over six feet tall, and straight as an arrow. Remarkable, too, was the fact that the eyes with which he "looked away across the sea," had never needed glasses. At that advanced age he could see to read without them.
Edgar P. Stites is also the author of:
"Simply trusting every day,
Trusting through a stormy way."
A hymn which has been translated into about a dozen different languages, and was perhaps a greater favourite a generation ago than it is to-day, is the one beginning:
"Yield not to temptation,
For yielding is sin."
Horatio R. Palmer, the writer of this hymn, was born at Sherburne, N.T., on April 26th, 1834, and is the
author of several standard works on music. Referring to this hymn a friend of Mr. Palmer writes: "I travelled in Palestine with the author of the words and music of 'Yield not to temptation.' He told me that when he was thinking of the temptation around the young, the idea of the hymn flashed upon him. The first two verses came to him without any effort, but the third verse cost him some trouble."
The hymn was written in 1868, and first published in the *National Sunday School Teachers' Magazine*. When the famous prison at Sing Sing, New York, had women as well as men within its walls, a lady missionary was a regular visitor to the women's department. Every Sunday afternoon the prisoners were permitted to come into the corridor to hear her talk, and to join in the singing of hymns. One day some of the women rebelled against an order of the matron, and a scene of grave disorder followed. Screams, threats and profanity filled the air. Assistance was hastily summoned, when suddenly a voice rose clear and strong above the tumult, singing a favourite song of the prisoners:
"Yield not to temptation,
For yielding is sin;
Each victory will help you
Some other to win.
Fight manfully onward,
Dark passions subdue;
Look ever to Jesus,
He'll carry you through."
There was a lull; then one after the other joined in the singing of the sacred song; and presently, with one accord, all formed into line and marched quietly to their cells.
Besides composing music for a large number of popular Gospel hymns, including that appropriately
appealing tune to "Come sinner, come," H. R. Palmer is the writer of words and music of several other familiar sacred songs, the best known being, "Shall I let Him in?" and "There is a home eternal."
Although belonging to another school of American hymn writers to that at present under consideration, this chapter would not be complete without the inclusion of the author of one of the most beautiful hymns in the language. I refer to:
"My faith looks up to Thee,
Thou Lamb of Calvary,
Saviour Divine;
Now hear me while I pray;
Take all my guilt away;
O let me from this day
Be wholly Thine."
Dr. Ray Palmer, the writer of this hymn, was born at Rhode Island in 1809. "My faith looks up to Thee," was written when the author was just twenty-two years old, and was his first composition. Remarkable though it may seem, this initial performance in hymnody was by far his most successful, for though Dr. Palmer is the author of many other hymns, not one of them has attained the popularity enjoyed by his first production. To-day, after well nigh a century it is to be found in the hymnals of almost all denominations. The author says, concerning its composition: "I gave form to what I felt by writing, with little effort, these stanzas. I recollect I wrote them with very tender emotion, and ended the last line with tears." Some time afterwards, Dr. Lowell Mason, meeting Mr. Palmer in Boston, asked him for a contribution for a new hymn book he was preparing, whereupon he produced this hymn from his pocket book. Dr. Mason was so much impressed with it that he at once wrote for it the famous tune "Olivet," to which it has since been sung. When
next he met the author, Dr. Mason said to him: "Mr. Palmer, you may live many years and do many good things, but I think you will be best known to posterity as the author of 'My faith looks up to Thee.'"
On the evening preceding one of the most terrible battles of the American Civil War, a number of Christian young men met together in one of their tents for prayer. After spending some time committing themselves to God, it was suggested by one of the number that they should draw up a prayer expressive of the feelings with which they went forward to stand face to face with death, and all to sign it as a testimony to the friends of such of them who might fall in the impending battle. After consultation it was decided that a copy of "My faith looks up to Thee" should be written out, and that each should subscribe his name to it, so that father, mother, brother or sister might know in what spirit they laid down their lives. They did not all meet again, and this incident was related afterwards, by one who survived the great battle. The last verse is most beautiful and affecting, especially when this touching incident is recalled:
"When ends life's transient dream—
When death's cold sullen stream
Shall o'er me roll—
Blest Saviour, then in love,
Fear and distrust remove;
Oh, bear me safe above—
A ransomed soul."
Possibly the next best known of Ray Palmer's hymns is, "Jesus, Thou joy of loving hearts." He died at New Jersey in 1887.
Referring to Sankey's popular Songs and Solos, a hymnal editor of note writes: "It should be frankly recognised that whatever be our estimate of the literary value of the hymns, or the musical merits of the tunes
with which they are associated, they have obtained such general acceptance amongst the masses that it is almost impossible to dispense with them in mission work. Probably it is their unpretentious simplicity that constitutes their charm, and accounts for their extraordinary popularity, while it provokes the criticism of those who set art before utility."
And thus, looking back to the early days when Philip Phillips set out on his pilgrimage of song, we see, with the march of time, a great and wonderful development in hymnody; a development which has vastly improved and enriched the hymnody of the Church. Hymns have been written that will be sung until time shall be no more.
CHAPTER X.
Sacred Song Composers
ELOQUENT tribute deserves to be paid to the various composers who contributed in no small measure to the immense popularity and usefulness of Gospel song. As has already been observed, Luther, more than three centuries before, was one of the first to fully realise the great truth that the tune was of equal importance with the words. As a matter of fact, with him the tune was first, the words second. But though Luther did not scruple in some degree, to do violence to the language to fit it to the exigencies of the music, yet he had a good notion of what a hymn tune should be, for his doctrine would have fallen comparatively flat had not his hymns given wings to his teaching.
"The merits of a hymn," once said the late Lord Balfour, "lie chiefly in the tune and associations, so that the editor of a hymn book who divorced old words from their accustomed setting, is an iconoclast of the worst order." Most of us will agree that there is a great deal of truth and common sense in this statement. Constantly in touch with hymns new and old one cannot but observe the variety of tunes, appropriate and otherwise, which find a place in our hymn books. There are, for instance, the words of some of our very old hymns which are inseparably wedded for all time to their own particular tune. Link on another tune and the beauty of the piece is gone. An old precentor in one of our remote Scottish kirks was of the same
opinion after a disastrous attempt at a "new" tune. He had held the post of leader of praise in the kirk for fully fifty years, and for well nigh fifty-two times fifty he had led the singing of the 100th Psalm to the Old Hundredth tune. One Sunday, having had occasion to be absent from the kirk, Andra was greatly concerned about how the service would be carried through without him. Soon after, meeting one of the elders—who by the way, boasted of no small knowledge of music—the precentor enquired how they had got on at the kirk. "Oh," replied the elder, with a touch of conscious pride, "we got on fine; mon, we had a new tune to the 100th Psalm." "Oh, and what was't?" Andra asked; at which the elder struck up the words of the familiar Psalm to the tune of the old Scottish song, "Ye Banks and Braes o' Bonnie Doon."
"Ay, mon," thoughtfully responded the old precentor, "I never thocht o' that."
Andra determined not to be behind. On the following Sunday he was in his accustomed place, and when the 100th Psalm was announced from the pulpit, there was no uncertain sound in the old precentor's voice as he feelingly struck up the Psalm to the new tune:
"All people that on earth do dwell,
How can ye bloom——!"
It is not recorded how far the worthy precentor was led astray in his enthusiasm for the old familiar song tune, before he discovered his mistake. There is undoubtedly a danger in more ways than one by clashing a new tune with an old hymn. On the other hand, many an excellent hymn has found a premature grave because of an inappropriate tune.
About the middle of the nineteenth century we find in America a group of composers, many of outstanding ability, collaborating in the self imposed
task of introducing what might at that period be considered a new type of sacred music. Up to that particular time there had been, to a great extent, a lamentable lack of charm in the hymn and psalm tune in general use, and so they set out to give to the world something more musically effective. Thus was brought into being the now popular sacred song.
Possibly the earliest composer in this particular sphere of music was William B. Bradbury, an American celebrity. He has been affectionately referred to as the father of sacred song. Born in 1816, musical impulses came to him when working on a farm as a boy, and he made a journey to Boston, where, for the first time, he heard an organ. He is credited with being the originator of Sunday School music, which has had such an influence upon the life of the Church during the last half century and more. He was also the pioneer in publishing this class of music books, and was among the first to promote musical conventions in America, which in later years became so popular in bringing together many of the leading hymn writers, whose names are so familiar to-day. Mr. Bradbury's love for music and young people made these occasions very popular and helpful. Amongst his many musical compositions are, "He leadeth me," "Solid rock," "I love to think of the heavenly land," "Sweet hour of prayer," "Jesus loves me," and "The Golden Shore."
Composers of sacred song were drawn from various walks in life. John R. Sweeney, the composer of the music of "More about Jesus," "Beulah Land" and "Sunshine in my soul," which have been translated into many languages, has written over one thousand sacred songs. He was born in West Chester, Pa., in 1837. His first composition was produced when he was yet a boy, and at the age of twenty-two he held a
responsible position as teacher in the musical profession. During the American Civil War he entered the army as leader of a military band, which position he held until the government discontinued the use of military bands in the army. On his retiring from the service he was made Professor of Music at the Pennsylvania Military Academy. It was in 1871, about the time of the great spiritual awakening when Gospel song was yet in its infancy, that Mr. Sweeney turned his attention to writing sacred music, and during the remainder of his life devoted his talents to the production of hymns. It is said that he was the editor or associate editor of over sixty hymn books. Mr. Sweeney's music has the stamp of originality and what he contributed to the cause to which he was so closely devoted, ranks with the best writers of his day. He was a popular leader at conventions and camp meetings; and his love for the young led Mr. Sweeney to associate himself with one of the largest Sunday Schools in Philadelphia, where for ten years he led the singing. He passed away on April 10th, 1899.
The name of Dr. W. H. Doane will always be associated with Fanny Crosby. He was a great friend of the blind hymn writer and wrote the music of many of her best known hymns. Dr. Doane was born at Preston, Conn., in 1831, and, though educated for the musical profession, he followed the occupation of a manufacturer of wood-turning machinery, and had taken out more than eighty patents for his inventions. The writing of hymn tunes was, therefore, an employment of his leisure. Another of his hobbies was the collection of quaint musical instruments, a number of which he presented several years ago to the Cincinnati Art Museum. Dr. Doane was superintendent of a large Sunday School at Cincinnati, Ohio, and edited no less than thirty-five collections of hymns and tunes.
for Sunday School use, besides composing several cantatas and anthems. Among his musical compositions that came into general use in the churches, were a number that were found to be well adapted for evangelistic use, and were, therefore, incorporated in books issued for that purpose, the very mention of which will call to mind the best known and best loved hymns for evangelistic and devotional purposes. For instance, the following: "Safe in the arms of Jesus," "Rescue the perishing," "Jesus, keep me near the cross," "Pass me not, O gentle Saviour," "Tell me the old, old story," "I am Thine, O Lord," and "Saviour, more than life to me."
Though keenly devoted to music and writing, which amounted to well-nigh a life work, he did not allow it to interfere in any way with his making a large and successful business, and with his giving a great deal of thought and practical help to missionary and philanthropic work. Dr. Doane received the degree of Doctor of Music by the Dennison University, an honour he well deserved, for he served the cause of sacred song with conspicuous ability. He died at his home in Orange, N.J., on December 23rd, 1915, in his eighty-first year.
Possibly the most famous hymn tune composed by George C. Stebbins is the one written to Fanny Crosby's hymn, "Saved by grace." Amongst others by the same composer are "Jesus is tenderly calling," "Take time to be holy," "Ye must be born again." Mr. Stebbins was born in New York State in 1846, from whence he moved in 1869 to Chicago. It was here he began his life-long acquaintance with Messrs. Moody and Sankey. It is not generally known that more evangelists, both singer and preacher, rose to religious fame from Chicago, than from any other city or country. Among such were Moody, Sankey, Bliss, Whittle,
McGranahan, Case, Excell, Gabriel and Stebbins. In 1876 Mr. Stebbins engaged in evangelistic work with Mr. Moody, and was more or less associated with him for a long period, going three times abroad and twice to the Pacific coast to assist him. It was during his visit to Great Britain when assisting Moody and Sankey in the memorable campaign of 1883-85 that Mr. Stebbins composed the tunes of many of his popular hymns, including, "Jesus is tenderly calling." In his reminiscences, which appeared in *The Gospel Choir* some years ago, Mr. Stebbins graphically describes the great mission conducted in London, ably assisted by many of the leading people in the religious world of that time, including the Earl of Shaftesbury, Lord Kinnaird, and Sir George Williams, founder of the Young Men's Christian Association. Describing the closing days of the winter's mission, Mr. Stebbins explained that it was Mr. Moody's custom to have an all day's meeting at each of the various centres. "I remember," he writes, "at one of these meetings an incident occurred that caused a good deal of amusement. Messrs. Moody and Sankey, Major Whittle and Mr. McGranahan, Dr. Pentecost and myself had lunched with some friends, and on the way to the afternoon meeting, where Major Whittle and Dr. Pentecost were to speak, a remark was made to the latter, illustrating one of Mr. Moody's habits of speech that amused him, and which found a lodgment in his mind. When it came his turn to speak, Mr. Moody whispered to him, saying, 'Pentecost, be short; be short.' He began his address by saying, 'Mr. Moody has asked me to be short. I notice that he will have three quarters of an hour of enthusiastic singing, and then he will get up and speak for half an hour, but it is not every one who can do that; but any one who can pronounce Jerusalem in two syllables
can do most anything.” The remark caused a good deal of laughter by the congregation, and considerable notice by the Press of England. Spurgeon’s comment upon the incident, or rather upon Mr. Moody’s pronunciation of the word, was not only characteristic of the great preacher, but very happy and appropriate. He said, “I thank God there is one man in such hot haste to get the Gospel to the people that he does not stop to pronounce all the syllables of every word.”
Extending over a long period, Mr. Stebbins was an intimate friend of Fanny Crosby, and composed the music for a considerable number of her hymns. He was also joint editor with Messrs. Sankey and McGranahan in the production of a large number of hymn books to which he contributed many beautiful Gospel hymns. Speaking at a reception given in his honour at Chicago, when he was in his seventy-seventh year, Mr. Stebbins gives this beautiful testimony, eminently characteristic of him:
“One of the things for which I thank God is that I was diverted from my profession to sing the Gospel, and for having something of the gift for writing music. It has been the greatest privilege of my life to be in the work and of using the gift God has given me, in producing the hymns you have sung.”
In response to my request, Mr. Stebbins kindly wrote me a few bars of one of his favourite hymns. “I take pleasure in enclosing a brace of ‘In the secret of His presence,’” he wrote, “but do not wish to imply, by so doing, that it is my favourite, for I have never been able to satisfy my mind in that regard. Each hymn taken as a whole, both words and music, has its own individuality, its appeal and sphere of service, as you can well understand; and indeed, its own claim to that distinction. However, considering the rather unique place this hymn has occupied, the sentiment in
the words, their authorship, their setting and all, it is perhaps as worthy to be called a favourite as any hymn that bears my name."
The words of this beautiful hymn are by Ellen Lakshmi Goreh, a Mahratta Brahmin lady, and the hymn coming into the hands of Mr. Stebbins, he composed the music to which it has since been sung. It may be of interest to remark that it had its first introduction to the public in London, during the all-winter's mission already referred to, conducted by Moody and Sankey in 1883-84. Mr. Sankey sang it from the original manuscript, and Mr. Stebbins himself sang it as a solo in the latter part of that mission.
Mr. Stebbins, who is now in his eighty-sixth year, is spending the evening of a long and useful life at Brooklyn, New York.
William J. Kirkpatrick is another composer of a large number of popular Gospel hymn tunes which made their first appearance in Sankey's early collections. He was born in Ireland in 1838, and when he was yet a child his parents emigrated to America and settled in Pennsylvania, in which State his life was spent. He came of a musical family, and at an early age was quite proficient on the flute, fife, violin and 'cello.
ROBERT LOWRY.
GEORGE C. STEBBINS.
PHILIP PHILLIPS.
For a few years he worked as a carpenter and for some time during the Civil War he served as principal musician, fife major of the ninety-first regiment Pennsylvania Volunteers. He was returned to Philadelphia and was assigned to work as a ship-builder until near the close of the war. During the twelve years that followed, Mr. Kirkpatrick was connected with a furniture manufacturing company, but in 1878 he abandoned all commercial pursuits and gave his undivided attention to the writing of sacred music, gradually gaining the ear and admiration of the English-speaking world.
Mr. Kirkpatrick was a voluminous writer, and it would be difficult to place an estimate on what has been attained in the Lord’s vineyard by his thousands of compositions. As editor his name has been associated with over one hundred different collections of sacred music. Among the compositions that have made his name famous are the following well-known favourites: “Jesus saves,” “Wait and murmur not,” “He hideth my soul,” “’Tis so sweet to trust in Jesus,” and “When love shines in.” Possibly his best known hymn, of which he is the writer of both words and music, is the one beginning:
“I’ve wandered far away from God:
Now I’m coming home;
The paths of sin too long I’ve trod;
Lord, I’m coming home.
Coming home, coming home, never more to roam;
By Thy grace I will be Thine; Lord, I’m coming home.”
Mr. Kirkpatrick laid down his pen on September 29th, 1921, and passed into the presence of the King at the advanced age of eighty-three, while resting in his favourite chair. On the floor, at his feet, lay a slip of
paper on one side of which was written with lead pencil, scarcely legible, the following lines:
"Just as Thou wilt, Lord, this is my cry:
Just as Thou wilt, to live or to die.
I am Thy servant; Thou knowest best:
Just as Thou wilt, Lord, labour or rest."
A second stanza was written on the reverse side of the slip of paper in still more indistinct writing:
"Just as Thou wilt, Lord,—which shall it be,
Life everlasting waiting for me,
Or shall I tarry here at Thy feet?
Just as Thou wilt, Lord, whate'er is meet."
In a private letter to a friend of the writer, Mr. George C. Stebbins says, in writing of the passing of Mr. Kirkpatrick: "To voice such words of resignation, then to close his eyes and open them again in a moment's time, is as striking and impressive as it is beautiful. What an awakening he must have had! Well might every child of God covet such an ending of his life."
As a composer of music for evangelistic purposes, Dr. D. B. Towner, who wrote the tune to "Trust and obey," occupies a prominent position among writers, as his hymns will amply testify.
In the year 1885, Dr. Towner felt the call of God to enter His service and he gave himself wholly to evangelistic work. Some years later Mr. Moody arranged that he should become the director of the musical department of what has since become known as the "Moody Bible Institute" in Chicago, which was then in its infancy. From that time until he was called home, Dr. Towner carried on an important work, and through his teaching and personal contact with the students that have passed through the institution, he has impressed himself upon thousands of young men and women, who have gone from there, more or less
imbued with the spirit of consecration, he had ever manifested, and with the increased knowledge of the importance which music has in all kinds of Christian activities.
Dr. Towner, like the other singers who were more or less under Mr. Moody's direction, occasionally assisted him in some of his meetings when Mr. Sankey was not with him, but he was usually associated with some of the evangelists whose movements Mr. Moody had at his disposal. It is said that his voice has been heard in almost every State in the Union.
Dr. Towner was born in Pennsylvania in 1850, and passed away in his seventieth year.
Other familiar compositions of his are: "Anywhere with Jesus," "Full surrender," "Redeemed," "Saving grace," and "Grace is greater than our sin." His songs are still widely used in evangelistic work and his name is familiar to all who sing or love Gospel music.
The unparalleled distinction attained by Ira D. Sankey in the realm of Gospel song is world-wide, and it seems superfluous to write in this connection at any great length. His famous *Sacred Songs and Solos* is his monument. And yet, writing on this absorbing subject, one feels it incumbent to make more than a passing reference of one whose name is still a bright luminary in the sphere of Gospel hymnody. He was born in the village of Edinburgh in Western Pennsylvania on August 28th, 1840. At the age of sixteen he was converted while attending revival meetings, and at once associated himself with Sunday School work, in connection with which, at an early age, he was elected superintendent and leader of the choir. It was here that Mr. Sankey's voice began to attract attention, and before long the Sunday School overflowed with people who came to hear the singing. In this way, though unconsciously, he was making
preparation for the work in which he was to spend his life. It was in 1870, at a convention held at Indianapolis that Mr. Sankey first met Mr. Moody, where the latter was announced to lead a morning prayer-meeting at seven o'clock. Arriving rather late, Sankey took a seat near the door alongside one of the delegates, a Presbyterian minister, who immediately turned to him and whispered, "Mr. Sankey, the singing here has been wretched; I wish you would start up something when that man stops praying, if he ever does." Sankey promised to do so, and when the opportunity came he struck up the familiar hymn, "There is a fountain filled with blood." The congregation joined in heartily and a brighter aspect seemed to pervade the whole atmosphere. At the close of the meeting, Moody immediately sought out the singer. Of that memorable meeting, which was willed of God to have such far-reaching effects, even to the carrying of the glorious Gospel to the ends of the earth, I will let Sankey tell his own story: "Moody's first words to me, after my introduction, were, 'Where are you from? Are you married? What is your business?' Upon telling him that I lived in Pennsylvania, was married and had two children, and was in Government employ, he said in his characteristic manner, 'You will have to give that up.'
"I stood amazed, at a loss to understand why the man should tell me that I would have to give up what I considered a good position. 'What for?' I exclaimed.
"'To come to Chicago and help me in my work,' was the answer. When I told him that I could not leave my business, he retorted, 'You must; I have been looking for you for the last eight years.'"
Sankey promised to think the matter over, and would seek the Lord's guidance. The next day he received a card from Mr. Moody asking him to meet the evangelist at a certain street corner that evening, at six
o'clock. At the appointed hour Sankey was at the meeting place with some friends. In a few minutes Moody came along, but without stopping to speak he passed on into a store near by, and asked permission to use a large box. The permission being granted, he trundled the box into the street, and calling Sankey aside, asked him to get up on the box and sing something. Sankey climbed up and sang, "Am I a soldier of the Cross?" which soon gathered a considerable crowd. After the song, Mr. Moody got up and began to speak. The workmen were just going home from the factories and in a short time there was an open-air service of huge dimensions. The people stood spellbound as the words fell from Moody's lips, with wonderful force and rapidity. When he had spoken some time, he announced that the meeting would be continued in the Opera House. He asked Sankey to lead the way and sing some familiar hymns. This he did, and marched along the street singing, "Shall we gather at the river?" The crowd immediately followed. That night the Opera House was packed to the doors, so completely were the men carried away with the singing and the sermon from the store-box. Thus did Ira D. Sankey hear in no uncertain voice God's call to service; nor did he longer wait, but from that moment consecrated his life to the Lord. From that time on till the death of Mr. Moody in 1899, they were associated in their great life work of saving souls. The wonderful results of that world-wide service of song, which to-day is still making itself felt, will only be revealed in a coming day.
In 1873 Mr. Sankey sailed for England with Mr. Moody. The story of that great mission, when there swept a mighty wave of revival from shore to shore of the British Isles, has often been told and will remain indelibly imprinted in letters of gold on the pages of
history, so long as the Gospel in song and story is told forth.
It is a remarkable circumstance that before Mr. Sankey entered upon his career as an evangelist he had never attempted to write music suited to evangelistic work, but soon after his work with Mr. Moody assumed such proportions in 1873, he began this phase of his work which from that time forward gave him a place among the foremost writers of Gospel song. His first attempt to write a hymn tune was during their mission in Edinburgh, when the music he then wrote was his admirable setting to Dr. Horatius Bonar’s beautiful hymn, “Yet there is room.” Mr. Sankey’s second composition was his music to the well-known hymn, “I’m praying for you,” which has proved to be one of the most useful hymns in all the range of evangelistic hymnody, and which has been blessed to uncounted multitudes. His subsequent work as a composer along that line brought him into prominence as a writer of music of a devotional character that possesses strength and permanent value, and which undoubtedly places Ira D. Sankey among the most gifted writers of evangelistic and devotional music. Among his best known compositions which to-day are being used in many parts of the world, the following may be mentioned: “The ninety and nine,” “Hiding in Thee,” “Simply trusting,” “There’ll be no dark valley,” “A shelter in the time of storm,” and “When the mists have rolled away.”
But the one hymn which will always be associated with Mr. Sankey is “The ninety and nine.” I have a vivid recollection, which I shall always cherish, of hearing Sankey, in his own inimitable way, sing this inspiring song. It was during his last visit to this country in the winter of 1898–99, where he conducted services of “Sacred Song and Story.” In describing
the origin of the hymn Mr. Sankey related that he and Mr. Moody were travelling from Glasgow to Edinburgh during their Scottish mission, when he chanced to see the words of the hymn, which appeared in the poet’s corner of a newspaper he was reading. So impressed was he that he called Mr. Moody’s attention to the little poem, suggesting that it would make a useful hymn for evangelistic work. Having cut the verses out and placed them in his musical scrap-book, Sankey put it away in his pocket, and for the time the poem was forgotten. At the noon meeting held at the Free Assembly Hall, Edinburgh, on the second day, the subject presented was “The Good Shepherd.” It was an impressive address, and at the close, Mr. Moody turned to his colleague and asked him to sing a solo in keeping with the subject. Mr. Sankey had nothing suitable in his mind, and was greatly troubled to know what to do. But I will let Sankey tell his own story, as I heard him relate it. “At this moment,” he said, “I seemed to hear a voice saying: ‘Sing the hymn you found in the train!’” But I thought this impossible, as no music had ever been written for that hymn. Again the impression came strongly upon me that I must sing the beautiful and appropriate words I had found the day before, and placing the little newspaper slip on the organ in front of me, I lifted my heart in prayer, asking God to help me so to sing that the people might hear and understand. Laying my hands upon the organ I struck the chord of A flat, and began to sing.
“Note by note the tune was given, which has not been changed from that day to this. As the singing ceased, a great sigh seemed to go up from the meeting, and I knew that the song had reached the hearts of my Scottish audience. Mr. Moody was greatly moved. Leaving the pulpit, he came over to where I was seated.
Leaning over the organ, he looked at the little newspaper slip from which the song had been sung, and with tears in his eyes said: 'Sankey, where did you get that hymn? I never heard the like of it in my life.' I was also moved to tears and arose and replied: 'Mr. Moody, that's the hymn I read to you yesterday on the train, which you did not hear.' Then Mr. Moody raised his hand and pronounced the benediction, and the meeting was closed. Thus 'The ninety and nine' was born."
In the art of singing, Mr. Sankey possessed gifts of an extraordinary character, although he had no professional training previously. Having a high baritone voice of exceptional volume, purity and sympathy, he had quite unconsciously acquired the habit of correct tone production, which enabled him to preserve it uninjured to the end of a strenuous career. Sankey's interpretation of his sacred songs was his own conception; and in his rendering of them he ever kept before him the importance of clearly emphasising the subject of the hymn, even to the risk of doing violence to the accepted rules of musical phrasing. But the singer's heart was in the song, and he sang his messages to the hearts and consciences of the people with sincerity of soul, that made him, under God, the great evangel of song that he was.
A story, reminiscent of the sweet singer's visit to Great Britain, is worthy of relating here. One day Sankey drove out from London to Epping Forest. A shock-headed boy climbed up one of the wheels of the carriage, and with wondering and longing eyes gazed on the stranger. Sankey rested his hand lovingly on the lad, and expressed the hope that some day he would preach the Gospel. Years afterwards, when the sweet singer was old and frail and blind, and just before he entered the Homeland, the evangelist was
almost overcome with joy when Gipsy Smith told him that he was the boy in the glades of Epping Forest on whose curly head Sankey had laid his prophetic hand.
The strenuous work connected with his last mission in Great Britain proved too much for his strength, and he never recovered from the strain. A few years before his death a decline in health set in, which was followed by total blindness.
Of the sweet singer's closing days I will set down what has been written by his constant friend and fellow-composer, Mr. George C. Stebbins: "During the last two years of Mr. Sankey's life I visited him every few days when in the city, and had delightful times of talking over experiences in the past with him. In spite of his total blindness he was ever the same cordial and companionable friend he had always been. His humour would often manifest itself in recalling some amusing experiences in the past, laughing as he told them, and apparently enjoying them to the full as he lived them over again.
"But it was plain to be seen that his mind and heart had long been set on his home-going, for that subject would so often intrude itself in our conversations. Once he said to me, 'George, you will find me on Spurgeon Street, when you get up there.' And for well nigh a year before his going, every time I called upon him he would say before my leaving, 'George, I want you to be at the church next Sunday (the church known as Dr. Cuyler's, of which he had been a member for a good many years) for I'll be there, as I am going home.'
"He had so longed for the two last years to be 'absent from the body and present with the Lord,' that his passing on had become an obsession with him.
"The time so longed for came on the 14th of August, 1908. I was at the time at Northfield conducting the singing at the annual conference there, which he and I had done yearly from the beginning of the conferences in 1880, but I was 'at the church' when he was taken there to receive the last marks of affection and love from his host of friends, and looked for the last time upon the face of the great singer who had gone to join the choir of the Redeemed on High."
The services Ira D. Sankey rendered the cause of Christianity cannot be fully estimated; services which even after the passing of well nigh half a century are a blessing and inspiration to untold millions the world over—truly a monument more enduring than granite.
Among the composers of Gospel song tunes to be found in present-day hymn books, the name of E. O. Excell is frequently met with. Possibly his best known compositions are, "Count your blessings," and "Let Him in." Born in Stark County, Ohio, in 1851, he served an apprenticeship to the trade of bricklayer and plasterer, but when twenty-two years of age, the love of music drew him away from his humble calling, and he finally laid aside the trowel and hammer for the more congenial occupations to be found in the many avenues of music. He was always keenly interested in Sunday School work, an evidence of which makes itself known in his many compositions for children, notably the two favourite songs, "Jesus bids us shine," and "Jesus wants me for a sunbeam," which are included in almost every recently published hymn book for young folk.
Mr. Excell was the last one of the old line of singing evangelists, and laboured with many of the most famous evangelists from the early days of D. L. Moody up to the time of his death. His was a life of devotion in the service of the Master, and for twenty years he was the colleague of the Rev. Sam P. Jones, the
celebrated American evangelist. As a singer, Mr. Excell had few equals. Possessing a voice of remarkable sweetness and power, he was able to sway his audience with the earnestness and spirit of his expression. His last work was with Gipsy Smith, in Louisville, Kentucky, in the midst of which he was obliged to quit his labour and return home. His voice was heard for the last time in the chapel of the Wesley Hospital, Chicago, on Sunday afternoon, January 16th, 1921, where, sitting in a wheel-chair, he sang as a solo "It is Jesus," with wonderful effect. At the close of the service his old friend and fellow-composer, Charles H. Gabriel, sang with him, "Beautiful Isle of Somewhere." No longer to him is that land a vague "somewhere," for, a few months later, his barque sighted the haven of rest, and as the sun went down on the evening of June 10th he landed upon that heavenly shore.
A recognised authority on all matters pertaining to sacred music, Mr. Excell, in his day, edited over forty different hymn books, besides having written many Gospel songs that will be loved and sung by many generations to come. He has also written the words of a considerable number of popular hymns, including, "Scatter sunshine," "Grace enough for me," and "I'm happy in Him."
Perhaps no person in the history of Gospel song has attained the position occupied to-day by Charles H. Gabriel. A native of the State of Iowa, where he was born in 1856, Mr. Gabriel has been privileged to have personal acquaintance with many, if not all the noted hymn writers, from the days of Philip Phillips, P. P. Bliss, and others, to the present time; and is thus the last living link of an illustrious group of sweet singers in the world of American hymnody. Commonly known as the "King of hymn writers," Mr. Gabriel has written the words or music of more popular hymns.
than has any other person living to-day. His most famous hymn, of course, is "The glory song," which has been translated into almost every leading language of the world, and which has probably been printed more than a hundred million times. It has been said that no Gospel song in history ever attained the international popularity of "The glory song" in so short a time. It was written in 1900, and in less than five years it was sung around the world.
This hymn alone is enough to make Mr. Gabriel famous; but he has written a number of others almost as famous. Among these are: "Send the Light," "Calling the prodigal," "He is so precious to me," "Where the gates swing outward never," "To the harvest field away," "Awakening chorus," "He lifted me," "Sweeter every day," "Evening prayer," and others equally well known.
In all of those the words of which he was the author, there is very manifest a happy blending of the poet and musician, and along with it rare judgment and deep spiritual insight into the needs of presenting the saving truths of Scripture in clear and singable form. His work in both fields is worthy to be recognized as an ideal to be followed by writers of to-day who desire sympathetic and appropriate musical settings to hymns.
He has also written the music for many of the most popular hymns of recent date, including: "Higher
ground,” “The Way of the Cross leads home,” “Let the sunshine in,” “Hail Immanuel,” “The sparrow song,” “A Sinner made whole,” “Brighten the corner,” “Glory in my Soul,” as well as many others to be found in almost every evangelical hymn book in present use.
Scores of hymns have passed through his hands for final retouching, polishing and finishing. He is supposed to know better than any other living person just how to put the finishing touches to a hymn. At an early age he began teaching singing schools in his own State, and later on, his remarkable musical abilities made him a popular leader of song in the great religious conventions all over America.
Mr. Gabriel is a self-made man, and although he has reached the very pinnacle of his profession, he never received a single music lesson in his life. Untaught, he taught himself, because there were no teachers of music on the wild prairies, where his father’s cabin stood and where the boy spent his early years.
He was born in a little shanty built of boards, and plastered both inside and outside. It stood on the virgin plains of Iowa. There was not a tree or shrub to shelter it from the terrific storms and snows of winter, or shield it from the blazing suns of summer. The wagon roads were trails over the rolling prairies. Deer, wolves, and prairie chickens were plentiful. Few settlers had horses, ox-teams being mostly used.
By the time Charles Gabriel was five years of age, a schoolhouse had been built three-quarters of a mile west of his home, where boys and girls received instruction during the winter months when there was no work to do on the farms. The benches used were made of logs, split in half, with two legs at each end inserted in auger holes. They were too high for little feet to touch the floor, nor had they a support for the
back, and more than once, in a sleepy moment, some young hopeful turned turtle over that hard, hard seat.
"I never saw a musical instrument," he writes, "until I was about nine years old, and to this day I couldn't tell the name of that one, as nothing like it has ever come under my observation since. For use it was placed upon a table, as a dulcimer. It had bellows which the performer pumped with his left elbow, while, with the fingers of both hands, he played keys something like those of a concertina. The next musical instrument I saw was a melodeon of that day and style. I rode ten miles to see and hear it, and no music since then has sounded to me more divine. I heard it as I ploughed in the field; it sang in my ears as I did my 'chores' *; in my dreams it floated over the hills of weariness down into the valley of rest, where I lay asleep."
His experiences as a teacher of music in the early days were of a unique and varied character. In the winter of 1882 we find him teaching a class composed entirely of coloured people; the following year he is in Texas, his pupils being cowboys, who attend the practices carrying their lariats and "guns" with them; while over in Muskogee his singing class consisted largely of Indian girls, who were usually accompanied by adult Indians in all their beaded and blanketed originality. He also taught a class of Japanese in San Francisco.
Mr. Gabriel is a man of simple tastes, is a great lover of the common people, and he is especially fond of children. His great aim in hymn and music writing is to produce simple, direct words and music, which can be felt and sung by the masses. Like all other great composers he is fond of classic music, but he judiciously
* Odd jobs about the farm.
avoids all technicalities in writing his hymn tunes. This is perhaps one great secret of his success as a hymn writer.
It was his inviolable rule to write something every day, and his work, covering as it did so wide a field, made the task comparatively easy. Still the proverbial midnight oil burned very frequently into the small hours, for his work was always done at home, and at night time when all the world about him slept.
For a period extending over twenty years it has been the present writer's inestimable privilege to have had an intimate acquaintance with the eminent hymn writer, and although he is now beyond the allotted span of three score years and ten, his letters are still brimful of joyous youth, for his soul is full of music, and his heart still throbs with love for the Master. Mr. Gabriel has now retired from active business life, mostly spent in the city of Chicago, where the greater number of his hymns have been written, and with his devoted wife, is now spending the evening of his days on the Western shores of America, in a pretty little bungalow at Berkley in California. But though the e'ntide shadows of a long and useful life may be gathering on the near horizon, yet his pen is never idle not his harp silent; for Mr. Gabriel is still writing hymns and hopes to continue to do so till called home to join in the grand eternal song.
"My sixty years of Gospel song," he writes, "have been eventful and tolerant, interesting and tedious, hopeful and discouraging. Failure more often than success marks the path I have travelled. And now, since the years have led me up the eastern slope and over the mountain top of life, and I am hurrying down toward the silent sea that lies shimmering before me, I begin to realise that my work has not been so much a failure as I had concluded, for seldom do I appear on
the platform that I do not meet some one I have met or known in the years gone by; and to feel a substantial slap on the shoulder and to hear a voice as from out of the past say: 'Hello! old man; glad to see you once more," is like a benediction, while to clasp the hand of a friend and feel that pressure which proves itself genuine is worth more than all the gold that was ever mined."
The theme of his famous "Glory Song," written thirty years ago, is to-day more precious to its author than ever before, and were we permitted to listen to the outpouring of a heart full of joyful song, I doubt not we would catch the strains of the same glorious anticipation:
"When all my labours and trials are o'er,
And I am safe on the beautiful shore,
Just to be near the dear Lord I adore,
Will through the ages be glory for me.
Oh, that will be glory for me,
Glory for me, Glory for me,
When by His grace I shall look on His face,
That will be glory, be glory for me!"
CHAPTER XI.
Popular Writers of To-day
Since the advent of Sankey's *Sacred Songs and Solos*, there have been many imitators in this particular sphere of hymn production, but few, if any, have approached anything near the high water mark of popularity this initial experiment attained. Inspired by the overwhelming success of the American singer's venture in the production and development of sacred songs, others have launched out on similar lines, until to-day the number of Gospel song books is legion. Considering the world-wide publicity given to the Sankey hymns, it is only to be expected that a great number will live to be used and blessed of God for time to come. Still, with each passing generation, new writers are born, whose productions in some measure are supplanting many of the old favourites whose day and generation is past. The purpose of this chapter is to refer briefly to those writers—nearly all of whom are still living—whose hymns have attained no small measure of popularity, and are familiar to the hymn lover of the present generation.
It is worthy of note that by far the greater number of popular Gospel songs in use to-day have their origin in America, possibly because of the immense quantity of sacred song books published annually, thus giving unlimited scope to contributors of this particular type of hymn.
THE ROMANCE OF SACRED SONG
One of the newer songs which sprang into favour soon after leaving the pen of the author, begins:
"When upon life's billows you are tempest tossed,
When you are discouraged, thinking all is lost,
Count your many blessings, name them one by one,
And it will surprise you what the Lord hath done."
Like a beam of sunshine it sped forth, brightening the dark places of the earth, bringing joy and gladness to the heart wherever it was sung. The author is the Rev. Johnson Oatman, jr., a Wesleyan minister, whose hymns, to-day, will be better remembered than his sermons. He was born in America, in 1856, and is said to have written about two thousand hymns, his first being, "I am walking with my Saviour." Other well-known hymns by the same writer, are: "I'm depending on the Blood," "I know He is mine" and "There's no Friend like Jesus." Possibly his most popular song—a song which has carried his name to many lands—is: "There's not a Friend like the lowly Jesus, No, not one! No, not one!" In less than a year it was reprinted in thirty-five different hymn books, and has since been translated into many languages and dialects.
Among recent writers of sacred song, few names are more familiar than that of Mrs. C. H. Morris, many of whose hymns rank with the compositions of Fanny Crosby and Frances Ridley Havergal. She is the composer of the words and music of:
"Nearer, still nearer, close to Thy heart,
Draw me, my Saviour, so precious Thou art;
Fold me, O fold me close to Thy breast,
Shelter me safe in that 'Haven of Rest.'"
Born at Ohio, in 1862, Mrs. Morris was converted at the age of ten. Since the day of her new birth it
has been her one aim to consecrate her life to the service of the Lord, believing at the beginning of her career that if she would write a noble poem, she must first live a noble life. Mrs. Morris is not only a poetess, but she is also a gifted musician, and the music of many of her best known hymns are her own composition. Over nine hundred of her pieces have been published, a good number of which have found a place in many present-day mission hymn books, including: "Let Jesus come into your heart," "The fight is on," "Who is this Man of Sorrows?" and "The Stranger of Galilee."
In the prime of life Mrs. Morris became totally blind, but this affliction does not deter her from exercising the ministry of song with which she is endowed, the writing being done by dictation to her daughter.
In the remote hamlet of Westwood, about a dozen miles from the city of Boston, a young Baptist minister, while conducting open-air evangelistic services in the village square, was inspired to write a Gospel song, which has been one of the favourites of the last two generations, and has been used of God in saving many a lost soul. The story of how this hymn "Throw out the Life-line," came to be written, is worthy of recounting here. The Rev. Edward S. Ufford was pastor of the old village church, in the parsonage of which he penned this hymn. Out at sea, not many miles distant, could be seen at low water the remains of an old wreck embedded in the sand. "As I trod the shore on summer days," wrote Mr. Ufford, when telling the story, "my imagination strove to picture what the storm did on the fateful night when it tossed the craft ashore, where it was soon dashed to pieces in the gale. While my heart was thus yearning for an effective interposition, a thought came to me. 'Why not hold an open-air meeting in the village next Sunday afternoon, and warn all who might pass by of their danger?' This was
in the fall of 1886. I carried my small organ out into the square and began to sing. There soon gathered around me a group of listeners. On returning home, the imagery of the sea came before me. In my mental eye I could see a storm, a spar, a shipwrecked sailor drifting out beyond human reach, where he might sink. Taking a sheet of paper I wrote the four verses of the hymn in fifteen minutes. They came as if by inspiration. Then sitting down to my little instrument, I played a melody without mental effort, apparently, and so the song was born."
"Throw out the Life-line across the dark wave,
There is a brother whom some one should save;
Somebody's brother! Oh, who then will dare
To throw out the Life-line, his peril to share?"
Mr. Ufford came of a musical family, his father and grandfather having been choir leaders in various cities. He was born at Newark, N.J., in 1851, was converted to Christ in young manhood, and at once began to work for the Master. A life of D. L. Moody which he read with deep interest, was the means of inspiring him to devote his life to the ministry of the Gospel. The popularity of "Throw out the Life-line," which had spread over the country on the wings of enthusiasm, fitting into the Christian Endeavour and missionary movements, gave the author the idea of a tour around the world to sing the song and to follow in its wake. This he undertook in 1902, and found that "Throw out the Life-line" proved to be a passport wherever he went. In Honolulu, it was in the native hymn book there in the Hawaiian dialect, and the author had the unique experience of singing his song in the old church, where the congregation sang it back to him in their native tongue. Mr. Ufford was called home in 1930 in his seventy-ninth year.
POPULAR WRITERS OF TO-DAY
Precisely the same year that "Throw out the Lifeline" was written, Mrs. Jessie Brown Pounds gave us:
"Anywhere with Jesus I can safely go;
Anywhere He leads me in this world below;
Anywhere without Him dearest joys would fade;
Anywhere with Jesus I am not afraid."
Set to music by Dr. Towner, it was popular first as a Sunday School hymn, and afterwards became a favourite with the Christian Endeavour movement. A band of Endeavourers were in the habit of singing on Sunday afternoons at the Sing Sing Penitentiary, and, incongruous though it may seem, this hymn was a favourite with the prisoners. Among the latter there were two young men who had been sentenced to death for a murder committed by them in a house they had entered for the purpose of burglary. Under the ministry of the Christian young people who visited them, they were converted. On their last day on earth, when they were brought forth for execution, the condemned men made a public confession of their sin, saying, however, that though they merited the death they were about to suffer, they believed they had God's forgiveness, and that through His grace they could go "Anywhere with Jesus."
Mrs. Pounds was born at Hiram, a suburb of Cleveland, Ohio, August 31st, 1861, and began to write to various weeklies when she was fifteen years old. A few years later, an editor who published some of her verses, referred to them as being well adapted to use as hymns. This note caught the eye of Mr. J. H. Fillmore, a hymnal editor, who wrote the authoress asking her to write some hymns for a new book upon which he was working. In this way Jessie Brown Pounds began hymn writing. She is the authoress of several books of hymns. Among her best known are: "The way
of the Cross leads home,” “Scatter seeds of loving deeds,” “I know that my Redeemer liveth,” “The touch of His hand in mine,” and “Beautiful Isle of Somewhere.” The latter hymn, which was sung at President McKinley’s funeral, was written one Sunday morning, in 1897, when a slight ailment kept Mrs. Pounds from church.
Miss Eliza Edmunds Hewitt, the authoress of “Sunshine in my soul,” ranks among the foremost writers of popular Gospel songs in recent times. Born at Philadelphia, U.S.A., in 1851, in early life she developed a spinal malady, which caused her to be a shut-in sufferer for many years. In course of time a gradual improvement came about, and, during a slow convalescence, she began writing poems which attracted the attention of John R. Sweeney, the noted composer of sacred music.
Miss Hewitt’s hymns are the result of inspiration, the origin of which, to her, is often a mystery; she has never cared to keep a record, but their number has long since passed into thousands. “Sunshine in my soul,” has been sung far and near, many beautiful stories coming back to the authoress, telling of its wonderful use, and the happiness brought into many a weary heart.
Other favourite hymns by E. E. Hewitt are: “Sweeter as the days go by,” “No one like my Saviour,” “Will there be any stars,” “More about Jesus,” and “When we all get to Heaven,” a hymn of joyous anticipation:
“Sing the wondrous love of Jesus,
Sing His mercy and His grace;
In the mansions, bright and blessed,
He’ll prepare for us a place.”
One day, when passing down one of the poorer
R. H. DANIEL.
JOHNSON OATMAN, Jnr.
ADA R. HABERSHON.
JESSIE BROWN POUNDS.
J. W. VAN DE VENTER.
ANNIE SHERWOOD HAWKS.
INA DUDLEY OGDON.
streets of his native town of Williamsport, Mr. James H. Black saw a little girl sitting on a doorstep of a dilapidated house. Her ragged clothes, torn shoes and dejected appearance, told its own story of a drunken father and mother, and a dwelling unworthy of the name of home. He spoke to the girl and asked her if she would like to come to Sunday School. At the question, a wistful expression crept into the child's eyes as she softly answered, "Yes, I would like to go, but—" Mr. Black understood the longing in the heart of the ragged girl, and the following day a parcel arrived containing a new dress, shoes and hat. Bessie was at Sunday School the next Sunday, and many other Sundays. One day the roll was being called. Each one responded until Bessie's name was being called. There was no response. Again the name was called. Still there was no response. Mr. Black learned that the girl was ill, too ill to be present, and the thought came to him like a flash, "What if this girl should never answer again? What if she should die? What would her answer be when the final summons came?" Almost unconsciously he found himself saying, softly:
"When the trumpet of the Lord shall sound, and time shall be no more,
When the morning breaks eternal bright and fair;
When the saved of earth shall gather over on the other shore,
And the roll is called up yonder, I'll be there."
No sooner had he said the words than his trained ear told him that he had created something euphonious, and going to the piano, he struck off the music almost as spontaneously as he had the words. In a remarkably short time he had finished the hymn which has been changed very little, if any, since that night.
The little girl died shortly afterwards, but the hymn which her illness inspired will never die. It has been
sung round the world, and has been translated into half a dozen foreign tongues. Writing about this time, Charles H. Gabriel says, "While assisting Dr. J. F. Berry in the preparation of a collection of songs, he took from his desk a roll of manuscript which had been submitted for use in the book, and handed it to me, saying, 'See if you can find anything in this bunch.' One from that lot (written in green ink) attracted my attention. The composer's name was not familiar to me, and the title of his offering was a new thought: 'When the Roll is called up yonder.'" Thus the hymn was sent out to carry its message over land and sea.
Mr. J. M. Black is also the author of a small but popular collection of songs, entitled *Songs of the Soul*, which was favourably received, over half a million copies being sold. He also composed the music of "Where Jesus is 'tis Heaven," and "I remember Calvary," and wrote the words of "Safe in the Glory Land."
Mrs. Frank A. Breck has written fourteen or fifteen hundred hymns, and one of the best loved is:
"Face to face with Christ my Saviour,
Face to face—what will it be;
When with rapture I behold Him,
Jesus Christ who died for me?"
Brought up by God-fearing parents, she cannot remember when Bible reading and prayer were not her daily home custom. From the days of her youth Carrie E. Breck wrote verse and prose for religious and household publications. For a number of years after her marriage, in 1884, family duties superseded those literary, and only on rare occasions did she encourage the muse. Her first published hymn was "You ought to do something for Jesus." Writing of this memorable occasion, when new desires for this particular form of service filled her soul, Mrs. Breck says, "It
was a great joy to me, and as opportunity offered, I pencilled verse under all sorts of conditions—over a mending basket, with a baby in arms, and sometimes even when sweeping up or washing dishes, my mind moved to metre."
Some of her best known hymns are "Everybody should know," "Nailed to the Cross," "Help somebody to-day," and "Never give up trusting."
Among popular Gospel songs of the last decade which suddenly sprang into favour, is "I walk with the King." Homer Rodeheaver, the singing evangelist, whose voice has been heard in most of the States of America, relates an incident in connection with this song, which happened during a mission he was conducting. "I sang this song to a great crowd of coloured folks one night," he says, "and as I finished it, one of the good old-fashioned aunties got up from the back row, taking off her sun-bonnet, waving it in the air, and stepping high down the aisle, she exclaimed, 'Hallelujah! I walk wid Him too, brudder!' Then there came the chorus from all over the house, 'Yeah! we all walk wid Him down here!' This," continues Mr. Rodeheaver, "is the real purpose of the song, to get folk to walk with Him."
Here is the verse that seemed to stir her enthusiasm:
"O soul near despair in the lowlands of strife,
Look up and let Jesus come into your life;
The joy of salvation to you He would bring—
Come into the sunlight and walk with the King."
This song is from the versatile pen of James Rowe, who is said to have upwards of eight thousand hymns and poems to his credit. Born in England, in 1866, at an early age he entered the Government Survey Department, where he continued till 1890, when he
emigrated to America. Mr. Rowe began writing hymns and poems about six years later, his first song being "Speak it for the Saviour."
A hymn by the same writer, which has gained considerable favour of recent years in this country, and is a special favourite of Gipsy Smith, has this chorus:
"Be like Jesus, this my song,
In the home and in the throng;
Be like Jesus, all day long!
I would be like Jesus."
Many of the best songs of Mr. Rowe owe much of their popularity to the attractive musical settings of Mr. B. D. Ackley.
Few composers of recent years have attained the position among Gospel song writers occupied to-day by Mr. Ackley, whose first composition, "Somebody knows," was published in 1912. It became a favourite, and was soon followed by many popular pieces, among others, "I shall dwell for ever there," "How you will love Him," and "I am coming Home." For ten years Mr. Ackley held the position of organist in many important churches of New York and Philadelphia, and for a considerable period was pianist to the Rev. W. A. Sunday, the noted American evangelist.
Miss Ada Blenkhorn, the authoress of "Let the sunshine in," and many other familiar pieces, began writing hymns in 1892. Of this hymn, a prison chaplain said, "It has done our prisoners more good than all the sermons preached to them." Another said, "'Let the sunshine in' brought the first ray of light to a condemned criminal, who was converted, afterwards pardoned, and who has for several years been preaching the Gospel." Some years ago Miss Blenkhorn had almost decided to give up hymn writing, when one day a lady, whom she happened to meet,
said to her, "May some soul be converted through a hymn that you shall write, who would not be converted if you do not write it!" "Those beautiful and inspiring words," writes Miss Blenkhorn, "seemed an invisible and mighty chain that held me fast and would not let me give up."
It was during the Torrey-Alexander mission that the consecration song, "I surrender all," first came into favour in this country, and singularly enough it was the first hymn of the author, J. W. Van de Venter, to become popular. He has written over a hundred hymns, and a fair percentage of them, prominent among which are "Looking this way," "Sunlight," and "My mother's prayer," have found a place in present-day mission hymn books, and have been wonderfully blessed.
The eminent Dr. T. De Witt Talmage, in the course of one of his striking sermons related this story: A mother having lost her only child, sat in a darkened room, day after day, grieving for the little one the Reaper had bound with his sheaves, when the servant entered and said, "My dear Mistress, why do you grieve? Do not sit in the darkness—let's open the window and look toward the light." It was by reading this story, that Ina Dudley Ogden received the inspiration to write her first hymn, "Open wide the windows," which was set to music by Charles H. Gabriel. This was followed by a great number, which to-day are known and loved the world over. Among the many are, "Living where the healing waters flow," "Could I tell it," "Carry your cross with a smile," "Jesus will," and "Brighten the corner where you are."
Of this authoress, Mr. Gabriel, who has furnished me with so much information relating to present-day hymn writers with whom he has been constantly in touch, writes: "Mrs. Ogden has always been an
intense lover of Gospel songs, and their influence on her early life was the controlling incentive that gave to the world that which only she could give. The object of every song seems to have been the winning of souls. Loved by thousands who have sung her hymns, she shrinks from celebrity in the knowledge that her songs are God-given, and that without Him she could do nothing, that in this way He has chosen to use her in the work of His vineyard."
Dr. D. B. Towner, who wrote the tune to "Trust and obey," relates how this popular hymn came into being. Mr. Sankey, assisted by Dr. Towner, was conducting evangelistic meetings at Brockton, Mass. One night a young man rose in a testimony meeting and said, "I am not quite sure—but I am going to trust and obey." Dr. Towner jotted down the sentence, and sent it with the little story to the Rev. J. H. Sammis, a Presbyterian minister, suggesting it would be suitable for a hymn. Soon after the verses came back by post, and the tune came spontaneously. Since then:
"Trust and obey,
For there's no other way,
To be happy in Jesus,
But to trust and obey,"
has been sung from continent to continent.
Mr. Sammis found the Saviour when quite a young man, and for many years took an active part in Christian work. He afterwards gave up business to take over the duties of Y.M.C.A. secretary, which later led him to devote himself entirely to the ministry of the Gospel.
Mr. Sammis has written over a hundred hymns. Among this number special mention deserves to be made of "Jesus is a friend of mine," and "Glory all the way," which have gained considerable favour. It is, however, as the author of "Trust and obey," that the name of J. H. Sammis will always be associated.
POPULAR WRITERS OF TO-DAY
One of the newest songs, whose melody like a mighty wave has swept across the Atlantic, and to-day is sung at almost every mission service, is R. H. Daniel's "Since Jesus came into my heart." Mr. Daniel has written comparatively few hymns, and is to-day known to the world by a single composition, which, undoubtedly, owes much of its immense popularity to the bright and attractive melody set to it by Gabriel.
At the close of a big mission, many hundreds of people gathered to bid the missioner good-bye. A great number of these people had been converted during the campaign, and as they lifted their voices in song they changed the words of the hymn to "Since Jesus came into my home." The scene was memorable, and as the joyous note of praise arose from the assembled throng of people, the noise of traffic for some moments was silenced by the mighty wave of song:
"Floods of joy o'er my soul
Like the sea billows roll,
Since Jesus came into my heart."
CHAPTER XII.
Hymns That Have Helped
WANDERERS across the wilderness of life have told how they have been helped and cheered, at one time or another, by some almost forgotten hymn, which has been to them a well of refreshing, and the shadow of a great rock in a weary land. Such remarkable incidents are indeed worthy of recounting, testifying as they do, to the experience of the deeper matters of the soul.
How the singing of "Saviour, breathe an evening blessing," not only helped but stimulated confidence, was told by one who passed through the night of peril. During the Boxer outbreak in China, when many of the Lord's servants were cruelly put to death, a company of beleaguered missionaries gathered together at the close of a day through which they had lived in constant fear lest they should have to suffer the fate of so many of their fellow labourers. Separated from home and friends, facing death in a far off land, and full of tenderest feelings, they lifted up their hearts in song:
"Though destruction walk around us,
Though the arrows past us fly;
Angel guards from Thee surround us:
We are safe if Thou art nigh."
"Out of the storm," writes Miss Helen Knox Strain, one of the missionaries present that night, "each soul, renewing its strength, mounted up with wings as eagles and found peace in the secret of His presence. We
went through the hymn until we came to the last verse, 'Should swift death this night o'ertake us.' We stopped at that line, and thought we would rest in the promise that the angel of the Lord would protect us. And so it proved."
In humble life, our hymns are not without their beams of sunshine. Passing through a narrow alley one day, the attention of a mission worker was attracted by a woman's voice in cheerful song. The words of the refrain upon which she lingered seemed strangely out of place in such squalid surroundings. She sang:
"And I shall see Him face to face,
And tell the story, 'Saved by grace!'""
Looking toward the place from whence the song came, he saw a poor old woman, down on her knees, scrubbing the doorstep of her humble dwelling. In a moment, that sweet song took a depth and beauty of meaning, and a charm unimagined before. That poor old woman, in her weary toil, was a "daughter of the King," and heir of eternal glory, though for a short time away from her heavenly home. What she sang, while she scrubbed, was to her a glorious anticipation, and the sunshine of her soul burst forth in the song of her heart.
"I will gladly take my turn in testifying," once wrote the late Mr. W. T. Stead, "conscious though I am that the hymn which helped me most can lay no claim to pre-eminent merit as poetry. It is Newton's hymn which begins, 'Begone unbelief.' I can remember my mother singing it when I was a tiny boy, barely able to see over the book-ledge in the minister's pew; and to this day, whenever I am in doleful dumps and the stars in their courses appear to be fighting against me, that one doggerel verse comes back clear as a blackbird's note through the morning mist:"
THE ROMANCE OF SACRED SONG
"His love, in time past,
Forbids me to think
He'll leave me at last
In trouble to sink;
Each sweet Ebenezer
I have in review,
Confirms His good pleasure
To help me quite through."
The rhyme is bad enough, no doubt; the logic may or may not be rational; but the verse as it is, with all its shortcomings, has been as a lifebuoy, keeping my head above the waves when the sea raged and was tempestuous, and when all else failed." It was John Newton, the converted slave dealer, who wrote these lines away back about the middle of the eighteenth century.
It is rarely that a circumstance, so remarkable as the following, attends the casual recital of a hymn, for, though it brought the desired joy and consolation to the heart of one individual, it had somewhat the reverse effect on the conscience of another. In the old coaching days, a lady was seated on the outside of a stage-coach reading. During the journey she had been intently engaged over one particular page of a little book which she consulted from time to time, with evident enjoyment. Turning to her fellow passenger, a gentleman, who she perceived was well acquainted with the subject of religion, she held the open page towards him, and pointing to the hymn she had been reading, asked his opinion of it. He glanced at the first few lines:
"Come, Thou Fount of every blessing,
Tune my heart to sing Thy grace:
Streams of mercy never ceasing,
Call for songs of loudest praise."
He read no further, and turning away, waived the subject, endeavouring to direct the lady's attention to
HYMNS THAT HAVE HELPED
some other topic. She, however, ventured another appeal, describing the great benefits she had derived from the hymn, and expressing her strong admiration of its sentiments. At length, overcome beyond the power of controlling his feelings, the stranger burst into tears. "Madam," he said, "I am the poor unhappy man who wrote that hymn many years ago, and I would give a thousands worlds, if I had them, to enjoy the feelings I then had." The stranger was none other than Robert Robinson, who, sad to relate, had fallen on evil days.
Born in 1735 of lowly parents, his widowed mother sent the boy to London to learn the trade of barber. Here he came under the influence of George Whitfield, the eminent preacher, was converted, and began to study for the ministry. At the age of twenty-five he was called to the pastorate of the Baptist Church at Cambridge, where he attained great popularity. In later years he lapsed into careless ways, indulging in frivolous habits. Poor Robinson ran a zigzag course to the end of his days and died suddenly on June 9th, 1790.
From time to time remarkable stories reach me of the wonderful influence one hymn or other has had upon many of the inmates of our large prisons. In contrast to this, one cannot but be amazed at the sheer incongruity exhibited in one particular instance, as shown in the following, which is taken from the annual report of the Howard Association, published a few years ago. Mr. Thomas Holmes contributes an article on Sunday in a London Prison. He makes reference to the vesper hymn sung by prisoners at an afternoon service he was conducting. "There were," he says, "those thousand men, locked, bolted and barred in prison, with strong warders to keep guard over them; there they were on their knees singing a vesper:
THE ROMANCE OF SACRED SONG
'Lord, keep us safe this night,
Secure from all our fears,
May angels guard us while we sleep,
Till morning light appears.'
I could have called out—I almost did—"Locks, bolts and bars will keep you safe, and your warders will watch over you!" It seems very strange to me," he continues, "that in many of our prisons the one and only vesper hymn selected for the prisoners to sing should be this one."
How a hymn, heard at an unexpected moment and under strange circumstances, helped in a marvellous way, was related to me quite recently by an intimate friend of the subject of the story. He was a Christian worker and had recently lost two sons, both fine young men. One died very suddenly, and a short time afterwards the other was seriously injured in a motor accident. The young man was conveyed to hospital, where he succumbed to his injuries. His father stood by the bedside watching the last struggles ere life left that young body, and when all was over he seemed to give way to bitterness of heart and rebellion against God for the great affliction which had come upon him. Leaving the ward, he went into an adjoining room where his wife waited for him. As he entered, his wife, realising their son had passed away, but ignorant as to her husband's bitter thoughts, said to him, "Mattha, there's a young fellow doon there in the street whistling 'Will your anchor hold?'" The bereaved father broke down at the significant words of his wife, and could only say, "Weel, the storm is very high just now." And indeed it was. But he afterwards confessed that the strains of that hymn, whistled by a young plumber going to work in the early morning, wafted to the top storey of the Glasgow Royal Infirmary, was the means of overcoming the evil of
doubt and placing faith in the sure Anchor, Christ Jesus.
Under somewhat similar circumstances, "Art thou weary?" was the means of bringing peace and rest to a burdened soul. A young woman who had formerly been a Roman Catholic, was brought to the knowledge of the Saviour, and became a faithful and consistent Christian. Some time after, she was laid aside with a serious illness, and became very depressed in spirit. On one occasion when she had been left alone for the night, a cloud came over her spirit, the sense of loneliness grew upon her and she seemed forsaken of God. All looked so black that she dreaded the coming of the long lone night. Just then, the silence of the night was broken by the sound of footsteps on the stone flags of the pavement outside. A man, wearing the clogs of the factory worker was coming along. His soul was full of joy. As he approached the house where the sufferer lay awake, he suddenly raised his voice in song:
"Art thou weary, art thou languid,
Art thou sore distressed?
'Come to Me,' saith One; 'and coming,
Be at rest!'""
The herald of peace went on his way singing the song of his heart, little dreaming that deep down in the heart of a young woman went the words "Be at rest!" And as she listened to the message of song borne to her in that silent hour of the night, she cast herself upon the Lord. The sun had pierced the dark clouds, peace and rest filled her heart, and she doubted no more.
A touching narrative is related by a worker at the Evangelistic Services held in Glasgow by Major Whittle and Mr. James McGranahan, reminiscent of those stirring days in the early eighties. The hymn which wrought so great an impression, as recorded in the
THE ROMANCE OF SACRED SONG
following incident, was written by Ellen M. H. Gates. Here is the first verse:
"Oh, the clanging bells of Time!
Night and day they never cease;
We are wearied with their chime,
For they do not bring us peace;
And we hush our breath to hear,
And we strain our eyes to see,
If thy shores are drawing near:
Eternity! Eternity!"
The narrator observed in Bethany Hall, one Lord's Day evening, an old fellow-workman of his. Knowing that he had been a very irreligious man, he determined to call at the workshop to have a word with his old mate. A day to two later when he called upon John, he soon found that something was working in his mind altogether different from the old things. "Look here," said John, "I didna think there was muckle truth in religion, but I'm a wee bit staggered aboot it jist noo!"
"I was glad to see you in the Hall," said his friend; "but tell me what has staggered you."
"Weel, ye see, I've a sister, ye ken, an' a wee while ago she was hearin' aboot the meetin's in Bethany Hall. So somehow she an' her companion—jist like hersel', but gey fond o' singin'—gaed to the meetin'. Aweel, when she cam' hame, she jist put past her things, an' sat doon by the fire, nae speakin' a word. Syne, the wife noticed her een were fu' o' tears. 'What's the maither, Aggie?' Nae answer. 'Gang tae bed, there's a guid lass; ye'll hae to be up sune the morn.' The tears cam' faster. 'Oh, Mary, I canna, I canna gang to bed. I've been hearin' a hymn the nicht I'll niver forget. Oh, I seem to hear the sound o' bells from somewhere, callin' "Eternity! Eternity!" Oh, I'm gaun into ETERNITY; an' oh, how dark it is jist noo! Gang to my bed! Na; I'll gang to my knees.' An' so she
HYMNS THAT HAVE HELPED
did. The wife tauld me this," continued John, "an I gaed ben awhile, but I only glowered at her. Weel, the next night she gaed again, an' she sune came hame wi' her companion, an' they baith seemed sae glad, sae happy the gither, an' talked aboot 'I am the Door; by Me if any man enter in he shall be saved.' They declared they had entered in. Anyhow, they were happy. Next nicht the wife gaed tae, an' noo the hale hoose is like a kirk! I've been gaun, an' I want tae ken mair aboot these things; so I an' Wullie here, are comin' on Sabbath nicht, an' Aggie an' some mair o' her companions; an' mither an' me would like tae hear that song Aggie heard."
In his interesting little book on the subject of "Hymns that have helped," the late Mr. Stead gives a remarkable testimony, received from a Scotsman, relating to "Lead, kindly Light." "My spiritual experience has been varied," he writes, "I was baptised in the Roman Catholic church, brought up in the Congregational Independent, and at length I was fascinated by the history, energy, and enthusiasm of the Wesleyans. I was at one time a local preacher in that body with a view to the ministry. But my fervid fit of exaltation was evoked with the dusty facts of life, and smouldered down into a dry indifference. I sought nourishment in secularism and agnosticism, but found none. I was in the slough of despond, at the centre of indifference, with the everlasting NO on my lips, when 'Lead, kindly Light, amid the encircling gloom,' came to my soul like the voice of angels. Wandering in the wilderness, 'O'er moor and fen, o'er crag and torrent,' Newman's hymn was to me a green oasis, a healing spring, the shadow of a great rock. Through the light and power of God I was led to light and love in Christ in a way I had never before known or experienced."
From the same source we cull a striking testimony
of the wonderful help and consolation received from an old-fashioned hymn, by a working man. Speaking of his experience, he says he passed through a period of much tribulation, seeking peace and finding none: "I thought I had done my best, but still that was unsatisfactory. Something always seemed to be kept back; something that ought to have come out and did not, or rather, perhaps I should say that was not fully understood by the one to whom it was told. I had no doubt of my wish to repent, no doubt of my willingness to make every reparation in my power, but still peace would not come. At last, I took it all straight to Jesus, and the burden rolled away from my heart." The old-fashioned hymn which brought such consolation is not to be found in many present-day hymnals. It is by Helen H. Willis, and appeared in Sankey's earlier collections. Here is the first verse:
"I left it all with Jesus long ago;
All my sin I brought him and my woe;
When by faith I saw Him on the tree,
Hear His still small whisper, 'Tis for thee.'
From my heart the burden rolled away—
Happy day!"
"Some of my earliest religious awakenings," once wrote Dr. Pentecost, the celebrated preacher, "were in connection with the hymns for children that were just beginning to be sung in the Sabbath Schools when I was yet a little boy. I mention one beginning:
'I think when I read that sweet story of old,
When Jesus was here among men,
How He called little children as lambs to His fold:
I should like to have been with Him then.'
That little hymn would always quiet me and beget within my heart seriousness and longing. When, as a child, I used to hear or sing it, I would wonder if there
was any blessing that I might have from Jesus that would correspond to his calling little children to Him, and laying His gentle, loving hands on their heads and blessing them. And in after years, when I had grown to be a young man, away from home, and far from God by wicked works, that little hymn of my childhood would come to my memory; and more than once I have sung it with choking voice and tearful eye, and with emotions of real penitence in my heart. It is true that these effects were transient, but they were real and mighty; and I doubt not that God used that child’s hymn to keep my heart from becoming perfectly hardened against His ‘gentle voice.’
A friend of the late Mr. Charles M. Alexander happened to find in a magazine a little song entitled, “Tell mother I’ll be there,” which he posted to the noted Gospel singer, with the suggestion that it might be useful in connection with his evangelistic work. Mr. Alexander pasted it in his scrap-book and carried it around for a year before he found a suitable opportunity of using. One night in Newton, Kansas, Mr. Alexander was called on to sing a solo. “I saw in the audience a great crowd of railway men,” said the singer, “and I wondered what would reach those men. With some doubt, I finally decided to try this touching song, and was surprised at the extraordinary result. Many of the men confessed Christ immediately. When the meeting was over, one big burly engineer came up to me and said, ‘Mr. Alexander, I promised my mother on her death-bed that I would become a Christian, but instead of that I have been going to the devil faster than ever. Preaching never touched me, but that song did.’ I used that song every night,” continued Mr. Alexander, “and I’ve been using it ever since. I have seen as many as a hundred and fifty men at a single meeting rise and confess Christ during the singing of that hymn,
before the sermon began. It reaches all classes, because everybody has a mother. It has been criticised from a musical and a literary standpoint. I hesitated a long time before I would use the song, 'Tell mother I'll be there.' I have been criticised all over the world for singing it, but you would not criticise it if you knew what it had done, and what letters and testimonies I have received about it."
This song had an interesting origin. When President McKinley was in office, his mother lay dying in Canton, Ohio, several hundred miles away. She sent word that she wanted to see her boy once more before she died. The President chartered a special train, and telegraphed, "Tell mother I'll be there." A Gospel song writer caught up the idea and wrote the song.
How the singing of a Sunday School hymn brought solace to a young Highlander, as he lay dying in a foreign land, was told by a lady who was permitted to visit one of our military hospitals, soon after some wounded soldiers had been brought in. The young fellow had lost a limb, and the doctor said he could not live through the night. As he lay with closed eyes, his lips moved, and the lady, bending over him, the words "Mother, mother," came in a gentle whisper. Dipping her handkerchief in cold water she tenderly bathed his burning brow, and as she did so he caught her hand and kissing it, cried, "Oh, that is good! Thank you, lady, it minds me o' mother." "Can I write to your mother?" she asked. "No," he said, "the surgeon has promised to do that, but, oh, will ye no' sing to me?" For a moment she hesitated. Looking out of the window her eye caught the gleam of a distant stream, and there came to her the thought of that river "the streams of which shall make glad the city of God," and she began to sing softly the hymn, "Shall we gather at the river?" Eager heads were raised to
HYMNS THAT HAVE HELPED
listen to the sweet song, while bass and tenor voices, weak and tremulous, joined in the chorus:
"Yes, we'll gather at the river,
The beautiful, beautiful river;
Gather with the saints at the river
That flows by the throne of God."
When the song was ended, the lady looked into the face of the boy—for he was not yet twenty—and asked, "Will you be there?" "Yes, I'll be there, through what the Lord Jesus Christ has done for me," came the ready reply, as his face suddenly lit up with a smile. Tears gathered in the lady's eyes as she thought of the mother, in her far-off Scottish home, watching and waiting for tidings of her soldier boy, who was breathing his last in a foreign land. Next day the good lady returned, but she did not find her Scottish laddie, for, ere the bugle sounded the reveille he had crossed the river.
Hymns have ever been a comfort and consolation in time of war as well as in days of peace. From the disastrous battlefield of Magersfontein, in the South African War of thirty years ago, there was addressed to me a letter, written by a Highlander who had been dangerously wounded, early in the battle. "I am thankful to say God has been very good to me," he wrote. "The twenty-four hours I lay on the battlefield unattended was the happiest time I ever spent in my life. All the day and night the words of that hymn were floating through my mind:
"My Jesus, I love Thee, I know Thou art mine!
For Thee all the pleasures of sin I resign;
My gracious Redeemer, my Saviour art Thou;
If ever I loved Thee, my Jesus, 'tis now!"
I had neither doubt nor fear. The Lord was verily my Shepherd. His rod and staff comforted me. Christ
compassed me about; He eased my pain; He quenched my thirst; He appeased my hunger. The devil could not get in edgeways.” Outstretched upon the burning veldt, his life’s blood slowly ebbing away, the young Highlander’s consolation was in Jesus, a theme so beautifully expressed in the words of William R. Featherston, the Canadian boy’s little hymn. Facing death during those terrible hours, he was able to sing from his heart:
“I will love Thee in life, I will love Thee in death,
And praise Thee as long as Thou lendest me breath;
And say when the death-dew lies cold on my brow,
‘If ever I loved Thee, my Jesus, ’tis now!’”
Fanny Crosby’s popular hymn “Rescue the perishing,” which has been the battle-cry for the great army of Christian workers throughout the world, has been blessed to thousands of souls. It was very extensively used in the great Moody and Sankey campaigns, during which, abundant testimony to its power to reach the hearts of wanderers was amply demonstrated. Sankey tells a story of how this hymn was the means of bringing peace and happiness to one who had sunk to the lowest depths of degradation. One bitterly cold night, a middle-aged man staggered into a Mission Hall in New York. He was under the influence of drink, his face unwashed and unshaven, and his clothes in rags. He sank into a seat near the door, but was aroused by the hymn which was being sung. The words were strangely familiar, and seemed to have a sobering effect on his dulled senses:
“Rescue the perishing, care for the dying,
Snatch them in pity from sin and the grave;
Weep o’er the erring one, lift up the fallen,
Tell them of Jesus the mighty to save.”
As he listened, the hymn seemed to recall some memory of his youth long since forgotten. It went straight
to his heart. As the preacher told the simple story of the Gospel, and how the Lord had come to seek and to save that which was lost, the stranger listened eagerly. In his younger days, the preacher had been a soldier during the American Civil War, and in the course of his address, mentioned several incidents which had occurred in his military experience, giving the name of the company in which he served. At the close of the meeting the man staggered up to the preacher, and in a broken voice, said:
"When were you in that company you spoke of?"
"Why, all through the war," was the reply.
"Do you remember the battle of _____?"
"Perfectly."
"Do you remember the name of the captain of your company at that time?"
The preacher mentioned the name.
"You are right! I am that man. I was your captain. Look at me to-day and see what a wreck I am. Can you save your old captain? I have lost everything I had in the world through drink, and I don't know where to go. The hymn you have just sung seems to tell me there is hope, even for a wretch like me."
He was saved that night, and was soon helped by some of his former friends to get back his old position. He never tired of telling the story of how a soldier saved his captain, and how the hymn "Rescue the perishing" was used of God in taking him out of the pit of iniquity, and setting his feet upon redemption ground.
Pathetic, indeed, was the death in America of Dr. John Watson (better known by his pen-name "Ian Maclaren"), and appropriate to a degree was his call for the words of the Scottish hymn, "My ain countrie." His physician secured a copy of the verses, and the dying preacher and author found consolation by their recitation. This is how the hymn opens:
THE ROMANCE OF SACRED SONG
"I am far frae my hame, an' I'm weary often whiles,
For the lang'd-for hame-bringin', an' my Father's welcome smiles;
An' I'll ne'er be fu' content until my e'en do see
The gowden gates o' heaven, an' my ain countrie."
On the last occasion that Sankey visited this country, I heard him sing "I'm far frae my hame," and I have still a distinct recollection of how the American singer thrilled the large audience by his remarkable rendering of this beautiful Scottish song. There is a touch of pathos in its origin, but I will let Sankey tell the story in his own words: "Many years ago John Macduff and his young bride left Scotland on a sailing vessel for America, there to seek his fortune. After tarrying a few weeks in New York they went West, where they were successful in accumulating a good competence. By and by his wife's health began to fail. The anxious husband said that he feared she was homesick.
"'John,' she replied, 'I am wearying for my ain country, will ye no' tak' me to the sea, that I may see the ships sailing to the homeland once more?'
"Her husband's heart was moved with compassion. In a few weeks he sold their Western home and took his wife East to a pleasant little cottage by the sea, whose further shores broke on the rocks that line the coast of Scotland. She would often sit and gaze wistfully at the ships sailing from the bay, one after another disappearing below the horizon on their way to her ain countrie. Although she uttered no complaint, it was evident that she was slowly pining away. John was afraid that she would die in a foreign land; and as an effort to save her he sold his New England home, and took her back across the ocean. She speedily recovered by the keen mountain air, the sight of purple heather, nodding bluebells, and hedge-rows white with fragrant hawthorn blossoms in bonnie Scotland, her
own dear native land. To her it was home. And there is no sweeter word in any language than 'home.'"
Mary Lee, when a young woman of twenty-three, wrote "My ain countrie," after hearing the story of John Macduff and his wife. She was born at Cronton Falls, New York, in 1838. At an early age she lost her mother, and was left in the charge of a Scottish nurse, from whom she learned something of the Scottish dialect. It was a special favourite of Mr. Sankey, and the warmth and zeal with which he rendered it never failed to captivate the good folks north of the Tweed. Possibly it may have been on such an occasion that the words of "My ain countrie" were impressed on the memory of Ian Maclaren, to be recalled years after as he lay dying far from his native land, bringing to him in his last moments, the peace and consolation his weary heart longed after.
INDEX OF AUTHORS’ NAMES
A.
Ackley, B. D., 210.
Adams, Sarah, 20, 116.
Alexander, Cecil Frances, 139, 140.
Alexander, Chas. M., 61, 62, 223.
Alford, Dean, 79, 80.
Arnold, Matthew, 34, 36.
Auber, Harriet, 20, 51, 52.
B.
Bachelor, Mary A., 58.
Balfour, Lord, 177.
Baring-Gould, S., 81, 82.
Baxter, Lydia, 59.
Bennett, S. F., 86.
Black, J. H., 207, 208.
Blenkhorn, Ada, 210, 211.
Bliss, P. P., 83, 118, 148, 149, 150, 151, 152, 153, 161, 166, 181.
Bickersteth, Bishop, 87, 88, 141.
Bonar, Dr. Horatius, 14, 15, 40, 41, 42, 43, 44, 45, 92, 190.
Bradbury, W. B., 138, 148, 150, 157, 170, 179.
Breck, F. A., 208.
Bryant, William C., 155.
Burns, Robert, 32, 52.
Byrom, John, 76, 77.
Byron, Lord, 33.
C.
Carey, Julia A., 141.
Clephane, Elizabeth C., 55, 145.
Codnor, Elizabeth, 95, 96.
Cousin, Anne R., 56, 57.
Cowper, William, 20, 31, 71.
Crosby, Fanny, 20, 148, 154, 155, 156, 157, 158, 180, 181, 226.
D.
Daniel, R. H., 213.
Dickson, David, 111.
Doane, W. H., 137, 148, 157, 180, 181.
Doddridge, Philip, 20, 26, 27, 28, 29, 31, 40.
Duncan, Mary, 143.
Dykes, Dr., 39.
E.
Ellerton, Canon, 37, 38, 39.
Elliott, Charlotte, 46, 47, 48.
Ellor, James, 110, 111.
Excell, E. O., 148, 182, 194, 195.
F.
Fawcett, John, 70, 71.
Featherston, William R., 102.
G.
Gabriel, Chas. H., 84, 146, 148, 182, 195, 196, 197, 198, 199, 208, 211, 213.
Gates, Ellen H., 160, 220.
INDEX OF AUTHORS’ NAMES
Gladstone, W. E., 66.
Gordon, A. J., 102.
Goreh, Ellen L., 58, 184.
Guthrie, Dr., 128.
H.
Habershon, Ada R., 60, 61, 62.
Hankey, Katherine, 136.
Hamilton, Eliza H., 86.
Havergal, Frances Ridley, 20, 48, 49, 50, 51.
Hawks, Annie S., 53, 54.
Heber, Reginald, 20, 25, 73, 74, 75.
Hewitt, E. E., 206.
Hoffman, E. A., 164, 165.
J.
Jude, W. H., 102.
Julian, J., 33, 38, 39, 40, 66, 87.
K.
Keble, John, 34, 35, 37, 38.
Kelly, Thomas, 20.
Ken, Bishop, 17, 18, 20.
Kirkpatrick, William James, 148, 184, 185, 186.
L.
Lee, Mary, 230.
Livingstone, Dr., 28, 135.
Lowry, Robert, 148, 167, 168, 169.
Luke, Jemima, 135, 136.
M.
Mackay, Margaret, 52, 53.
Mason, Dr., 175.
Matheson, Dr., 89.
McGranahan, James, 148, 163, 166, 167, 182, 183, 219.
McKinley, President, 115, 206, 224.
Midlane, Albert, 132, 133, 134.
Montgomery, James, 31, 32, 33, 34, 72, 111.
Morgan, R. C., 147.
Morris, C. M., 202, 203.
Moody, D. L., 14, 15, 18, 43, 59, 60, 61, 90, 91, 96, 104, 105, 146, 147, 148, 150, 156, 159, 166, 181, 182, 183, 184, 187, 188, 189, 190, 191, 192, 194, 204, 226.
N.
Neal, J. M., 19, 30, 35, 36, 37.
Newman, J. H., 34, 35, 77, 78, 122, 221.
Newton, John, 20, 29, 30, 31, 72, 215.
O.
Oakeley, Sir H., 35.
Oatman, Johnson, 202.
Ogden, Ina D., 211.
O’Kane, T. C., 148.
INDEX OF AUTHORS’ NAMES
P.
Palmer, H. R., 148, 172, 173, 174.
Palmer, Ray, 174, 175.
Paton, John G., 57, 131.
Peace, Dr. A. L., 89.
Pentecost, Dr., 93, 94, 182.
Perronet, E., 20, 109, 110.
Phillips, Philip, 146, 148, 159, 160, 176, 195.
Pounds, Jessie Brown, 205.
R.
Robinson, Robert, 217.
Rodeheaver, Homer, 209.
Root, F. W., 152, 169, 170.
Rowe, James, 209, 210.
S.
Sammis, J. H., 212.
Sankey, Ira D., 14, 15, 17, 18, 43, 44, 57, 59, 60, 61, 85, 90, 96, 104, 105, 137, 138, 145, 146, 147, 148, 152, 156, 158, 159, 160, 165, 166, 175, 181, 182, 183, 184, 187, 188, 189, 190, 191, 192, 193, 194, 212, 226, 229, 230.
Shelly, Martha E., 142.
Shrubsole, W., 109, 110.
Smith, Gipsy, 193, 195, 210.
Spurgeon, C. H., 183.
Spafford, H. G., 90, 91.
Stead, W. T., 215, 221.
Stebbins, Geo. C., 61, 148, 159, 181, 182, 183, 184, 186, 193, 194.
Stites, Edgar P., 171, 172.
Sullivan, Sir A., 81.
Sunday, W. A., 210.
Sweeney, J. R., 148, 179, 180.
T.
Talmage, Dr., 21, 211.
Toplady, A. M., 20, 65, 66, 116.
Torrey, Dr., 61, 62.
Towner, D. B., 186, 187, 212.
U.
Ufford, E. S., 203, 204.
V.
Venter, J. W., Van de, 211.
W.
Wainwright, J., 76, 77.
Warner, Anna B., 138.
Watts, Isaac, 17, 18, 23, 24, 25, 26, 31, 40, 156.
Watson, Dr. J., 228.
Wesley, Chas., 18, 20, 24, 30, 40, 65, 77, 100, 106, 116, 125, 143, 144.
White, Kirke, 20.
Whittle, W. D., 83, 99, 148, 151, 161, 162, 163, 164, 166, 181, 182, 219.
Willis, Helen H., 222.
Y.
Yates, John H., 84.
Young, Andrew, 131.
104 783 |
DELHI-MEERUT RIOTS
Analysis, Compilation and Documentation
DELHI-MEERUT RIOTS
ANALYSIS, COMPILATION AND DOCUMENTATION
ASGHAR ALI ENGINEER
1988
Ajanta Publications
No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system without permission in writing from the Publishers.
ISBN 81-202-0198-1
First Published 1987
by
AJANTA PUBLICATIONS (INDIA)
7255/8 Prem Nagar, Shakti Nagar, Delhi-110007
Distributors:
AJANTA BOOKS INTERNATIONAL
1 U B Jawahar Nagar Bungalow Road Delhi-110007
Printed by M.R. Prints, at Pasricha Press, Delhi-110031
## CONTENTS
1 Introduction 1
**PART I**
WHAT HAPPENED DESCRIPTION AND ANALYSIS
2 On Returning From Meerut —Asghar Ali Engineer 11
3 Meerut—Shame of the Nation —Asghar Ali Engineer 16
4 Muslim Politics The Awful Price of Sbhababuddin Bukhari Line —M J Akbar 33
5 The Agony of Meerut —Inderjit Bodhwar, Tana Midha and Others 37
6 Hitlerite Barbarities in Meerut —Nikhil Chakravarty 46
7 A Politician’s Riot —Ajay Base and Pankaj Pachauri 49
8 Inept Local Govt Caused the Riots —Askari H Zaidi 56
9 Politicians Poured Oil on Fire 61
10 Old Delhi Too in the Grip of Communal Frenzy —Asghar Ali Engineer 65
11 Delhi—Walled City Riots 74
12 Imam Climbs Down to Victory —Bharat Bhushan 109
**PART II**
THE ROLE OF PAC AND MALIANA MASSACRE AND THEREAFTER
13 When the So called Guardians Turn into Communal Armies Order Crumbles —M J Akbar 115
14 117 Missing in Mahana, Presumed Dead —Pankaj Pachauri 120
15 A Massacre that Wasn’t —Anil Saxena 125
16 Leaning of Live with Official Lies
— Neena Vyas and Nilanjan Mukhopadhyay 128
17. The Cover-up
— Sankarshan Thakur 132
18. Maliana Massacre Evidence of the Government’s Cover-up
— Sankarshan Thakur 138
19. Blood-Hounds without Leash
— K.M. 142
20 Ganga Canal of Death
— Sanjay Suri and Rohul Pathak 145
21. Tale of Two Survivors
— Somito Gupta and Rahul Bedi 147
22 PAC Jawan’s Shot Dead 200
— Sankarshan Thakur 150
PART III
WHAT EDITORS WROTE AND READER’S OPINED
23 Holocaust in Meerut
Editorial from Times of India 155
24 When Will We Learn (Indian Express) 157
25. State and Communalism
(Economic and Political Weekly) 159
26. A Deadly Confrontation (India Today) 162
27 A Day of Sadness (Telegraph) 165
28 Warning of Meerut (Blitz) 167
29 Enforce the Law (Indian Express) 169
30. Hiding the Crime of Meerut (Telegraph) 171
31. Meerut Massacre—A Fore-Taste of Fascism
(Urdu Times, Bombay) 173
32. Is Chief Minister of U.P. A Democrat?
Foizol (Jodid), Delhi 181
33. Reader’s Opinions 183
PART IV
INTERVIEWS
34 We Can Reform the PAC But When the Intelligentsia Turns Communal, Where Do We Go —Syed Shahabuddin 203
35 "We Never Expected This" —Bir Bahadur Singh 209
PART V
WHAT APPEARED IN THE LANGUAGE PRESS
36 Alam Bhai Slap on the Face of Delhi Communalists 215
37 An Attempt of Hide PAC's Crime —Virendra Kumar 218
38 What We Saw in Meerut? What We Saw in Sayhana—Peace March by Sadiya Dehlavi 222
39 Ten Important Questions About Meerut 224
PART VI
DOCUMENTATION
40 "I Regret Your Leadership" —Arif Mohammad Khan 229
41 Political Parties 235
42 Resolution Passed in the Two Days National Executive in Meeting of All India Qaumi Tanzeem Held in New Delhi on 28th and 29th June 1987 253
43 Text of Memorandum Submitted to the Prime Minister by the Muslims of Maliana —Ejaz Ali Siddiqui and all the Muslims of Maliana 256
44 List of Those Killed by PAC in Maliana Firing 258
45 Action Plan for Communal Amity —K R Malkani 260
INTRODUCTION
The Meerut riots must make us all sit up and think seriously what must be done to strengthen the forces of secularism in our country if it is to be saved from not only destabilization but also from disintegration. The ruling politicians are very fond of talking about 'destabilization' these days and thereby imply that certain elements within the country is at this game in collusion with 'foreign elements'. This may or may not be so but thing is certain the way ruling politicians in U.P. directly or indirectly colluded with rank communalists in Meerut and allowed the administration, the police and the PAC to be communalised. It is a very serious matter. Such things, and done so bluntly, can, and do, undermine stability of the country.
Before we discuss other aspects, it is highly necessary in my opinion, to pin down the administration's responsibility for controlling a riot. An ex-chief minister of Maharashtra as well as some senior administrator from Calcutta told me that no riot can continue more than a couple of hours, if the authorities mean business. This claim is certainly borne out from situation in West Bengal today. Hardly any riot has taken place there during the rule of the Left Front Government. It has issued strict instructions to the police and the administration not to allow any riot to spread. It should be curbed immediately. This policy has paid good dividends in West Bengal which has been by and large free of communal trouble.
In contrast if we see what happened in Meerut under the Cong-I Government led by Shri Vir Bhabadur Singh, is simply
unbelievable. For five days the PAC and police led mobs from 19th (morning) to 23rd (evening) looted, burned and killed. All this happened unchecked during the curfew hours. The PAC bullets protected fanatic mobs committing arson during broad daylight. And all this was happening with full knowledge of the authorities comfortably ensconced in their offices. Moreover, it happened when the Chief Minister was present in Meerut. If this be the attitude of the authorities how can one expect rioting to be controlled?
Something more than worse happened in Meerut in those five fateful days from 19th to 23rd May, 1987. The PAC shot several people dead mostly those sitting in the 'security' of their houses in Malyana, a village, on the outskirts of Meerut, on 23rd May, afternoon. And no action was taken against these blood thirsty-marauders and their commander Tripathi, but it was not even withdrawn from there. What was even more shameful was that the administration is now trying to cover up the whole incident saying that no such massacre ever took place and that of the 110 people reported missing most of them have returned and that one person who wanted to make false claims for compensation manufactured this list.
This is of course a shameful lie. The massacre did take place in which no less than 78 persons were killed or burnt alive. This writer investigated the whole incident and included in the report "Meerut—Nation's Shame". Also some other journalists belonging to Telegraph, Calcutta and Sunday Observers, Bombay, investigated the incident after this attempt at the cover up and found administration's version quite untenable. What greater shame for a government than to protect those of PAC and the Police who killed scores of people. Can such a government have any right to rule? A government which encourages its forces to massacre innocent people. The very same PAC took around 40 people from Hashimpura, Meerut, near Muradnagar Canal, shot them and threw them into it. The floating bodies were discovered after a few days and also two survivors told the gory tale. The evidence for this has been so overwhelming that the Vir Bahadur Singh Government has so far not mustered courage to cover it up although initially it had denied it. No wonder the noted journalist Mr. Nikhil Chakravarty called it
"Huliente Barbarities in Meerut"
If the government has such attitude one can never hope to see the end of communal violence in this country. In fact most of the political parties in India today have acquired vested interest in communal politics. It helps them generate, or so they think, vote banks. Of all these parties the Congress (I) has now much greater stakes in resorting to the politics of communalism. We do not wish to analyse this here. We have done it elsewhere. Suffice it to say that the Governments in Gujarat, UP etc are not very much interested in strongly putting down those elements which find it politically profitable to generate communal hatred and openly provoke communal violence, be it on Babri mosque Ram janambhoomi issue or any other. No wonder than that one finds increasing projection of religious festivals, ratha yatras and personalities on Government media like the radio and TV.
It is for no other reason that even National Integration Council has been reduced to a farce. After every major riot a ritual convening of the Council takes place where some shibboleths are repeated ad nauseum and the same reported in the papers. That's about all as far as this body is concerned. The Government never took seriously its suggestion to set up riot squads which will consist of people from different religions and that minorities will be given twenty five percent representation on it. So many major riots have taken place after the NIC resolved to set up such squads but not a single state has taken lead in this direction. Even the Gujarat Chief Minister Shri Amarsingh Chaudhari developed cold feet on the suggestion for recruiting more in the police force from minority communities. His reason for taking this measure is that it would lead to communal reaction in the state as well as in the police ranks.
Now Gujarat is one of the most communally sensitive states today and is often rocked with communal violence. Ahmedabad and Baroda besides many other cities have been witnessing major communal flare-ups from time to time for last several years and still the Government refuses to give adequate representation to minorities—a step strongly recommended by the National Integration Council. This is all because the
Government has no political courage to take such a step displeasing certain elements from the majority community. The Haksar Committee set up by NIC also made some highly useful suggestions to checkmate outbreak of communal violence. The recommendations, as expected, were hardly taken seriously.
The Indian Express editorially points out, "The fact is that the state is already armed with adequate powers to put down communal miscreants with a strong hand. It just does not use them. The Haksar Committee of the National Integration Council has brought this out forcefully. After noting how religious establishments have flouted various existing laws on public peace and order with impunity, the committee has aptly observed: 'Even a token enforcement of one single issue like this would bring a sense of credibility to the debates of the National Integration Council.'" The editorial further points out very pertinently, "Indeed the very fact that the Prime Minister, the Home Minister and the other eminent leaders who attended the meeting failed to highlight the report of the Committee and instead, were content to repeat the worn-out cliches about strengthening the secular fabric etc., shows that they remain stuck as if in a groove."
This is as far as the National Integration Council is concerned. What can one expect from administration when there is no political will to curb communalism and communal violence. The petty political interests have led to developing what can be described as politics of confrontation. Both the Hindu as well as Muslim leaders are resorting to it most unabashedly. The Muslim fundamentalists took very aggressive postures on the Shah Bano controversy. The Rajiv Government went down on its knees and enacted a law on maintenance of Muslim woman which was neither secular nor Islamic. The whole movement was inspired by political, not religious considerations. In order, therefore, to please Hindu communalists who were highly displeased by the enactment of the Muslim Women's Bill, the Rajiv Government allowed the Babri mosque to be unlocked thus raking up another very bitter controversy.
The Babri mosque/Ram Janambhoomi controversy, like that of the Shah Bano case, has done greatest harm to the
secular atmosphere in the country. From both sides again deepest emotions are being stirred by the interested politicians thus bringing about confrontation between the two principal communities. This controversy has become the most major issue though it must be treated a non-issue, thanks to the pre-eminence the politics of communalism and confrontation has acquired in our country today. There is nothing wrong, I must say, in debating the issue academically. It is for leading historians of the country to debate it. The noted historian R.S. Sharma maintains that Ayodhya where Ram was supposedly born could not have existed during that period (i.e., around sixth century B.C.) as there was dense jungle around there and iron was yet to be invented to provide implements for clearing such dense forest.\(^{10}\) Prof. Sharma's view could be contested by other historians. But it should certainly not be allowed to be politicised. It greatly weakens the secular fabric of our society.
The Muslims point out (I must say in all fairness that even some Hindus have taken this view) that in Tuzk-e-Babri, Babar has advised his son and successor Humayun not to hurt religious feelings of the Hindus and not to permit cow slaughter, if he wants to rule over India. How can then be himself commit the sacrilege of demolishing such a sacred temple supposed to be the birthplace of Lord Rama, one of the most sacred religious figures of the Hindus? Moreover, they argue, Babar was a good Muslim and a good Muslim is required by injunction of his religion not to touch the sacred places belonging to other religions. It would be outrageous to the Quranic teaching as the Quran goes on to say: "Our Lord is Allah. And if Allah did not repel some people by others, cloisters, and churches, and synagogues, and mosques in which Allah's name is much remembered, would have been pulled down. And surely Allah will help him who helps Him. Surely Allah is Strong, Mighty."\(^{11}\)
Thus anyone who demolishes any place where Allah's name is remembered, he must be replaced by one who can protect it.
Thus the argument could go on historical as well as religious level but it should not be dragged into the political arena as it has been done today. The issue would not be resolved even if
say Hindus have their way and are able to convert the mosque finally into a temple. It would leave a bitter taste in the mouth of Muslims and the controversy could be raked up by any demagogue in future whenever it suited his interest. The best thing in my opinion would be for truly religious minded people on both sides (and those who do not have any political interest) to sit together and resolve the issue showing full respect for each other's religious sensibilities. However, I admit, there are powerful political interests involved from both the sides and they may not permit such an ideal solution.
The Government, Central or State, would be of no help either. It is the government, as pointed out earlier, which is no less culpable in this controversy. It may either pretend to be neutral or may tilt to one or the other side depending on the political pressures or exigency of the situation. What is then to be done? One thing is to be borne in mind. We will be confronted with many more Meerut riots, if this controversy is not resolved sooner than later. I am of the opinion that there is no dearth of goodwill on either side despite such bitterness generated by the vested interests. There are quite a few, both among the Muslims as well as the Hindus who sincerely wish this controversy to be resolved peacefully. Their voice seems to be feeble yet but given strong will, it can be made much more effective. They must come together, pool their resources and organize an active campaign. It would certainly gather momentum. It would make the communalists on either side defensive. By now they have overreached themselves. All that is needed is to shake our complacency and act.
I would also like to make a point here that the politics of communalism is very tempting but not really so very rewarding. Even a quick glance at some election results would bear me out. People at large have tended to vote on secular issues rather than on caste and communal considerations. The communalists in the recent state assembly elections in Kerala, West Bengal and Jammu and Kashmir were on the run. They had adopted very aggressive postures but drew meagre results, if not altogether blank. I am aware of the percentage of votes they polled by them. In some cases they even polled little more than before. But even this percentage rise is by no means proportional.
to the aggressiveness and barrage of their propaganda. No communal party has captured power and formed government of its own. At best it has shared power like the BJP or Muslim League. Apart from this menacing proportions communalism has acquired recently see no prospects for any communal party to be able to turn it into electoral advantage.
But how does one explain this contradiction? A human being behaves at different levels. At one level and under certain circumstances, our religious susceptibilities become predominant and we respond to that situation. At another level our secular interests seem to be at stake and we respond to that situation as best as we can within the given parameters or choices. Even a scientist behaves differently while in the laboratory with his apron on and quite differently while in an ordinary human situation with his apron off.
We should also remember that the Hindu-Muslim conflict is historically conditioned and has deeper emotional causes. Someone or the other is bound to exploit it for the perceived benefits, political, social or even anti-social and economic. My submission, therefore, is that while we should fight resolutely against such forces we should not get disheartened as the integrative and secularising processes are also at work in our society. The determined fight against the forces of communalism and fundamentalism is highly necessary to raise the consciousness of people, to make them realise the importance of shedding hatred against other communities in the process of nation building and widening secular spaces. Such resolute struggle against forces of communalism would also help keeping away the menace of fascism which becomes real in certain situations.
Looking to the overall situation of socio-economic and socio-political forces in the country today it seems difficult to do away completely with communal violence. The politics of confrontation, and that of communalism and fundamentalism, let us remember, is not due only to our communal conditioning and religious sensibilities but also on account of increasing assertion for share in power and economic resources in developing situation. In other words, left to itself ethnicity is no more than prejudice and awareness of being different but when pitted in a competitive, developing, dynamic situation, it acquires sharp
political complexion. It is for this reason that no developing country (so also developed countries) is free of ethnic conflict in some or the other degree.
It is only when we begin to see the problem in this perspective that we would understand it properly. At one level we have to condemn strongly riots in Meerut as great crime against humanity and yet on the other level we have to understand them as an integral part of development process, perhaps an inevitable price of the development process. From Biafra in Africa to the Philippines in the Far East Asia it is the same story. How can the Indian subcontinent be different?
Notes and References
1. See Times of India Report "Massacre that wasn't", dated 17-6-87.
2. It is included in the book vide page ... It can also be seen in Economic Political Weekly in its issue of 20 June 1987.
3. See in Telegraph, 17 June 1987, Sankarshan. Thakur's "Nobody has returned to Maliana Village" and in Sunday Observer, 21 June 1987, Pankaj Pachauri, "Operation Cover Up in Maliana".
4. See his political commentary in Sunday Review of Times of India, dated 14th June 1987.
5. See Asghar Ali Engineer, "Hindu Muslim Relations" Mainstream May 9 & 16, 1987.
6. See news item to this effect in Bombay Samachar, Bombay, dated 23-6-87.
7. See Indian Express, 25-6-1987 Editorial "Enforce the Law",
8. Ibid.
9. For detailed analysis see Asghar Ali Engineer, Introduction to the Shah Bano Controversy (Orient Longman, 1987).
10. See R.S. Sharma's interview "Historians are looking at the past from a new angle", Times of India, 12-4-87.
11. The Quran, 22.40.
PART 1
WHAT HAPPENED:
DESCRIPTION AND ANALYSIS
ON RETURNING FROM MEERUT
Asghor Ali Engineer
Yes, I just returned from Meerut. I am writing these lines in sheer anguish. I have no words to describe the agony of the people of that unfortunate city. What I saw is unbelievable. With all that has happened in Meerut, I don't think we can call ourselves civilised, let alone citizens of a great nation with rich heritage of past. We are simply barbarians who take pleasure in killing each other. It is sheer sadistic pleasure. Loss of life and destruction of property has been unprecedented. I have investigated many riots but I have not seen the things I saw in Meerut. Believe me, I am not exaggerating.
What is sad, the Hindus told me how many Hindus have been killed and the Muslims how many Muslims were slaughtered. No one was genuinely concerned with human life as such. Life is, and must be treated as sacred. It cannot be compartmentalised. The malady begins with such compartmentalization. I do not say religious identities should be abolished. It is, at best, a distant ideal. All I am saying is that we must learn to respect life per se, not Hindu or Muslim life. We ought to have learnt at least this much by the end of twentieth century.
I always feel proud of that India that produced Gautam Buddha and Mahavir Jain, Nanak and Kabir, apostles of peace, non violence and tolerance. But here in Meerut, I am crossing rivers of blood, trampling here and there over human bones, smell burning flesh, see nothing but hatred and intolerance in the eyes of people. Will this hatred cease to be? Will the Hindus and Muslims of Meerut stop slaughtering each other? I have no such message of optimism. Tension can be felt,
hatred can be experienced, PAC is round the corner to slaughter, provocation or no provocation.
I try to discover the causes. It is not easy. Again, each community has its own story to tell which confuses rather than enlightens. Fortunately there are few who have retained their objectivity, it is they who help me arrive at some tentative conclusion. I can clearly see neither community is totally blameless though Muslims have suffered immensely. But Hindu losses are not insignificant though much less.
Talking to the Hindus and Muslims it appears Babri Masjid/Ram Janambhoomi controversy has played havoc. The politics of confrontation has wrought this destruction. Vishwa Hindu Parishad and Babri Action Committee are real culprits.
The common Hindus and Muslims understand the real game behind this non-issue, that is what it is. It is neither mosque nor temple which is aimed at; it is only the politics of certain vested interests. Earlier we understand this simple truth, better it is both for the country as well as for the security of the masses of people.
Let Muslims clearly understand that the rallies and marches for Babri mosque would succeed only in repeating Meerut. Let them know few irresponsible Muslim leaders are endangering the safety and security of thousands of innocent Muslims. Why after all sacrifice so many precious lives for a mosque? Let my Muslim brothers understand mosque cannot be more sacred than human life, it is the creation of God; mosque, on the other hand, is the creation of man. It can be built and rebuilt. Under constraint the holy Quran permits even eating of kbinzir (i.e. pork). After saying that eating of dead (animal), flesh of swine and blood is forbidden, the Quran goes on to say, "Then whoever is driven by necessity, not desiring, nor exceeding the limit, no sin is upon him." (2:173)
The message is very clear. Life is above everything else. To save life in certain circumstances, one can even eat forbidden food provided of course intention is not to defy. By analogy one can say Babri mosque, even if more sacred than human life (which it is not), in the given circumstances today, can be surrendered in order to save human life. The massacre cannot be allowed to go on. Those Muslim leaders who still insist on
holding rallies and marches for salvaging Babri mosque are playing to say the least, with innocent lives, destroying what Allah has created, to save what man has built. If they are so concerned about the Babri mosque, why are they silent about seven thousand mosques converted into stables and residences in and around Punjab during the partition riots? Does it not clearly show that it is Politics, rather than mosque which matters?
I would also like to appeal to my Hindu brothers not to be misled by the propaganda of the Vishwa Hindu Parishad. Such campaigns would take country nowhere. It would only result in weakening of the nation and in tarnishing of its name. The process of nation building can never be complete without active and meaningful contribution of minorities. Our democracy and constitution can have no meaning if a section of Indian population feels insecure and paralysed. India cannot have the kind of international prestige it is entitled to if communal violence rocks it again and again. Is it not sheer hypocrisy to be proud of philosophy of non-violence and at the same time conniving at killing? Does this killing and communal frenzy lead us anywhere near any solution? Even Hitler's 'final solution' did not work.
Even if Muslims are not behaving, they should be won over, rather than killed. My own personal experience shows that little sympathy and understanding create a lot of goodwill among Muslims. Killing would only generate more hatred and bitterness and would result in killing even on larger scale. Repeated occurrence of communal riots clearly corroborates this simple observation. I earnestly hope my Hindu brothers would show maturity and wisdom and take initiative in stopping this bloodshed. We need more Ramas, not Ram Janambhooms. It is his sacrifice for the sake of his younger brother which is needed most.
I would also like to address the Central Government as well as the Government of UP. Both have not covered themselves with glory. Rajiv Gandhi has deliberately neglected this sensitive issue. He does not want to earn odium from either community. He surrendered to the Muslim fundamentalists on the question of Muslim Women's Bill and to the Hindu fundamentalists on the question of Ram Janambhoomi. The
short term interests override long term interests of the country and thus push it on the precipice of instability, violence and communal madness. The UP Government has of course been the real culprit. It has played most unscrupulous role in this whole controversy.
What is most shameful for it is that the Chief Minister Mr Vir Bahadur Singh is no less involved in the massacre by the PAC in Meerut. All evidence goes to show that it has happened with his connivance. The least he can do is to immediately resign. He was unwilling to withdraw the PAC when it was on its killing spree. He was against handing over the town to the army. Had it been handed over to the army many precious lives could have been saved. Now that all this massacre has taken place he is not willing to withdraw the PAC, let also taking action against it. The order suspending the commandant was withdrawn. Can such a Government have faith of people? We will have more Malianas if we have more Vir Bahadur Singhs in our country.
The way PAC killed people in Meerut it raises serious question of human existence in our country. The human rights organizations should seriously ponder over this matter. Where human life is decimated in this arbitrary fashion by the state security forces themselves where is the question of human rights. Human rights go with human life. In my opinion the recently constituted human right's commission should hold an inquiry into the Meerut killings by PAC and other civil rights organizations should launch an intensive campaign for disbanding PAC and trying those culprits from amongst them who are responsible for such monstrocities.
I would also like to appeal to the intelligentsia in our country to launch an intensive campaign with all honesty and integrity against communal and fundamentalist forces whichever community they belong to. Let us realise that communalism is the very negation of our composite nationalism and also of democracy. There can be no compromise with it. Let us put as much pressure on the ruling and other parties to stop the communal and caste arithmetic in winning the elections. Let them face the people with real issues. Let us also launch campaigns among people not to vote on such parochial
considerations but on the basis of the work the candidates or the parties sponsoring them do for the uplift of the people. It is only our unflinching commitment to the cause of the people that would put communalism on defensive. Communism is on the offensive today because we have consented to give importance to the obscurantist issues, forgetting the real issues.
With all that has happened our faith in secularism should not be shaken. It is our only hope. Greater the adversity, more severe the test of our faith in secularism. Also, let us explain to the people that no religion can be above humanism. Human concerns should be valued more than the religious concerns. Applying same principle in the political arena, democratic concerns should be valued more than the party concerns. Today these concerns have been reversed and hence Ahmedabads, Bhiwandis, Moradabads and Meeruts in our country.
Will genuine secularists and humanists arise? They have nothing to loose except communalism.
MEERUT—SHAME OF THE NATION
Asghar Ali Engineer
Meerut erupted again within five years. It was rocked with riots in 1982 when the PAC (Provincial Armed Constabulary), a communal outfit had shot not less than 45 persons in Firoz building, an area where mostly bidi-workers lived. The 1982 riot was no ordinary occurrence, but the recent one has far surpassed it in its ferocity and intensity. What I saw during my stay in Meerut can hardly be described in words. The words 'barbarism' and 'monstrosity' are not enough to describe what happened there. Communal riots are shame of the nation; but the riot in Meerut would put even Hitler's acts to shame. It is rather painful for me to compare Nazi Germany of yester years with democratic India of today. May be it is a bit exaggerated and way of saying it. But those who visit Meerut and spend few days hearing the horror stories would not completely rule out this comparison.
The occurrence of communal violence has become so frequent that we have almost stopped reacting to it. I was in Chiangmai (Thailand) when I first read about Meerut riots in newspapers. When I drew attention to this of an Indian colleague he spontaneously said what is new about it. Though I was not surprised by this reply, it did shock me. We have taken communal violence in our stride, something of an ordinary occurrence not worth much bothering about. It is surely a dangerous signal. We have stopped resisting and challenging the monster who is swallowing, day after day, all our human values and national ideals. We have surrendered to it and learnt
to live with it. Who knows who would be its victim next? In 1982 Prof Harpal Singh, a thoroughly secular soul deeply committed to humanitarian values had done his best to establish peace and harmony along with his equally committed wife and son. In the recent riots his son Dr Prabhat became the victim of this monster. Dr Prabhat was his only son. Prof Singh and his wife were distraught, but not despaired, when I met them. This is perhaps the only sign of hope in otherwise hopeless situation.
II
Meerut is an important town in UP, some sixty kilometers from Delhi. Its population is around one million of which Muslims constitute 45 lakhs, i.e., around 45%. It is known for its weaving, brassband and scissors' industries. These industries are controlled both by Hindus as well as by Muslims though the Hindus can be said to be senior partners. Nevertheless, the Muslims in Meerut had made it and could be said to have had economically arrived. But all this now lay in shambles.
To begin with, it is important to search for the roots of the communal trouble in Meerut. As we know, the problem is extremely complex. There is no single cause which can fully explain it. Nevertheless, we have to try and unearth various causes. It can be said that the nature of the trouble is primarily politico-economic. I am using these words rather cautiously, though not very precisely. It is very difficult to be precise in such a highly confused situation as it prevails today in that unfortunate town.
I would throw light on what I mean by the primary or most significant cause of communal trouble being politico-economic. Muslims in Meerut, primarily the Ansaris who control power looms—and there are 40,000 powerlooms in the city—had begun to become quite prosperous. These Ansaris, having made some money in cloth weaving and printing industry—branched out in some other economic areas, transport, for example. They took to other miscellaneous economic activities as well. Also, many Muslims set up auto repair workshops while large number of trucks pass through the town as it is on an important
trading route. Besides this, the Muslims also have a share in the brassband and scissor manufacturing industry. Nadir Ali & Sons are the biggest manufacturers and exporters of brassband in Meerut, though not the only one.
Many prosperous Muslims, as is quite natural, have developed political aspirations. In these days of increasing fundamentalism many of them, like their Hindu counterparts in the Vishwa Hindu Parishad, BJP and RSS, find it very easy to fulfill their political aspirations through politicised religion. In fact the politicisation of religion is all about this. Motivation for this comes less from religious fanaticism and more from political ambitions. It is for the same reason that the middle castes among Hindus which have acquired economic clout are communalised. There is another side of the coin as well. Religion, on account of its high emotive power, is used by those also who are getting increasingly sidelined in politics. Due to rise of middle castes, the Brahmins are beginning to lose their grip over the political power. Hence they are floating more religio-political organizations like the Vishwa Hindu Parishad to register their political presence.
Among the Muslims one more factor, psychological in nature, operates powerfully, i.e. their being in minority. They tend to cling more to religion both due to minority psychology and their comparative backwardness. Oil revolution in the middle east, the revolution in Iran and increasing communalization at home of certain sections of the Hindu society, have further intensified their tendency to lean on religion not as a faith but an instrument of political mobilization. This becomes great irritant for a section of Hindus, especially the communal and religious fanaticism among the Muslims and this perception feeds their own (i.e. Hindus') fanaticism. One can feel this very much in Meerut today.
If this is coupled with economic prosperity, as it has happened in Meerut, it has multiplying effect on the Hindu psyche. The ferocity with which the business establishments have been destroyed there bears testimony to this observation. Entire rows of shops including the isolated ones on Hapur road, Sadar, P.L. Sharma Road, Begum Bridge, Western Kachehri Road, Mawana stand (where nine buses and many shops
belonging to Muslims) were reduced to ashes. It is very difficult to estimate how many powerlooms were destroyed. Even barber shops and laundries were not spared. Around 50 motor mechanic shops within one kilometer of Mawana bus stand were burnt. On Hapur road 30 shops, on Azad road 28 shops (on both sides), at Samudra Gate 9 shops and 8 workshops on the opposite side became the targets of mob fury. And all this was done in the full presence of PAC on 19 May in the morning. Most of the Muslims whose economic establishments were burnt or looted told me that economically the Muslims in Meerut have gone back by at least 25 years. They had achieved a measure of prosperity after great deal of efforts. Again they are where they were a decade and half ago.
Add to this the fact that Muslims are 45% in Meerut. This adds to their political clout. They become quite decisive in election arithmetic of any party which seriously contends elections. Even very senior Congress leaders from the area were heard saying we will not allow Meerut to become another Khalistan. And let it be noted that recently the Muslims have, through their utterances and activities created such an impression. For example Sahim Ahmad Ansari likes himself to be described as 'Muslim Bhimdranwale'. He feels if one crore Sikhs can become headache for the Hindus why can't 20 crore Muslims (unfortunately like the Hindu communalists the Muslim communalists also exaggerate their population).
Here it must be emphasised that the Babri Mosque/Ram Janambhoomi issue is playing havoc with the situation specially in UP. It has become highly emotive for both the communities. One does not know where this politics of confrontation would lead to. Demagogues from both the sides are having fields day. In my opinion the leaders of the Babri Action Committee who have been arousing Muslim feelings on a non-issue for their short term political interests are no less responsible for what happened in Meerut. The rally held in Delhi and the speeches made there by the Shabi Imam, Syed Shahabuddin and others have greatly angered Hindus. Meerut being very close to Delhi a large number of Muslims from there went to participate in the rally. Some of them even wore shroud. Let Mr. Shahabuddin and others understand that their demagogy on
Babri mosque issue is putting thousands, if not lakhs of Muslims in danger. The whole of U.P today is sitting atop a volcano.
Here it is also important to draw attention to the fact that the Central Government is equally responsible for the hotting up of the Ram janambhoomi/Babri Mosque controversy. It is deliberately allowing it to be prolonged. This enables it to manipulate votes. In fact the Government gave it a new dimension by allowing the lock on the mosque to be removed and letting in the Hindu worshippers. First it surrendered to the Muslim fundamentalists by passing the Muslim Women's Bill and then to Hindu fundamentalists by unlocking the mosque to mollify their feelings aroused by this Bill. In the process it strengthened the hands of communalists from both the communities. What if secularism went by complete default. It requires vision, dedication and commitment to uphold secular and humanistic principles at one's own cost. In this sense struggle for freedom is qualitatively different from efforts at economic development. The former demands sacrifice and the latter encourages corruption. One necessitates suffering while the other promotes grabbing the fruits of development by the middle and upper classes when the development is within the capitalist framework as in India. Also, the democratic political processes bring in ever greater political consciousness and increasing degree of assertion on the part of weaker sections and minorities on one hand, and, more ruthlessness and greater degree of corruption among the privileged upper and stronger sections of society. And democracy today in India has become an art of manipulating for purposes of votes the consciousness of one section or the other. This is how the ruling Congress is also surviving in power. There is no longer any question of giving a definite direction to developmental process with a will to solve people's problems and resolve issues based on value premises.
The dispute about the Babri mosque/Ram janambhoomi should also be seen in this light. The Central and the State Governments have neither will nor political courage to resolve the issue on the basis of principles. On the contrary, they are deliberately prolonging the conflict with a view to extract maximum political mileage out of it. They are allowing the
conflict to sharpen, the communities to get embroiled in such non-issues so that they may rule. Our democratic rulers are no less apt in following the policy of divide and rule, a political legacy inherited from our erstwhile British rulers. Hindus and Muslims in contemporary India are as sharply divided as they ever were.
III
This forms backdrop to the Meerut riots. The question is how it all began and how? No definite answer is forthcoming. There are as many opinions as the people you talk to. Everyone has a theory. One has to listen patiently to all these theories to construct some credible picture. Communal tension in Meerut had very much been there for long. It has all the ingredients often propounded by the theorists of communal riots for communal violence to break out—history of communal violence (communal riots having taken place there in 1938, 1946, 1949, 1962, 1968, 1973, 1982, 1987), higher proportion of Muslim population (40 to 45%), economic competition (mainly in powerloom industry there being 40,000 powerlooms in the town) and sharp political polarization (presently on the Ram Janamabhoomi/Bahri Masjid issue) and struggle for power (within ruling Congress itself between Hindus and Muslims).
The situation had become highly explosive after the Babri Masjid rally held in Delhi. As already pointed out a large number of Muslims had participated in it from Meerut. The militancy among the Muslims was growing on this question. The Hindus were by no means less aggressive, perhaps a degree more. If Muslims raised the slogan "ham hein Muslim tees crore, lenge tera khoon nichod" (we Muslims are 30 crore, we will wring blood out of you), the Hindus shouted 'Hindu Sikh bhar bhar, Muslim quam kahan se aar' (Hindu-Sikh are brothers, where did Muslim community come from) and 'Hindustan men rahna hai to Hindu hankar rahna hoga' (if you want to live in India, you have to become Hindu). These slogans were painted on the walls of Meerut.
Also, the power struggle within and without the Congress has sharpened. The high percentage of Muslim population makes it difficult for the Hindu Congress candidates to get
ticket for the assembly and parliamentary seats from here. This, and the Ram Janambhoomi controversy has brought about a sharp polarization within the Congress. It has been vertically divided between Hindus and Muslims. The riot further sharpened this communal division. Also, a large section of Muslims is totally alienated from the ruling party. Even Muhsina Qidwai was elected largely by non-Muslim votes in the last parliamentary elections. The Congress bosses including the Chief Minister, it is alleged, wanted to teach these Muslims a lesson for not voting Congress. The Chief Minister, it is a fact, was not well disposed towards Muslims during the riot. He even brushed aside a Congress M.P.'s suggestion to withdraw the PAC during the riot as it was playing havoc with the Muslims. What is worse, number of major incidents of killing and burning of Muslim houses and shops took place when the Chief Minister was present in the town. He hardly pulled up the administration for its inability to control violence. Many openly allege it all happened with his connivance, though obviously it is difficult to verify.
It is also said by many that the Chief Minister was more interested in arranging men for the Prime Minister's May 16th rally in Delhi than in taking steps to tone up administration for the possible outburst of communal violence which most observers felt was a distinct possibility. The Chief Minister was asked to arrange at least for 750 busloads of people and he is reported to have arranged for 1000 in his enthusiasm to please the P.M. It is also alleged that many of these buses returned loaded with lethal weapons and they were allowed to go to the respective mohallahs unchecked. All this arsenal obviously came for the riot. The police which was keeping vigil since the last riot in April was withdrawn and engaged in the mobilization efforts for the Delhi rally. This also left the field free for trouble-makers.
It was also reported to me by some well informed persons within the Congress circle that the Chief Minister was bent on preventing meetings of Shri V.P. Singh and hence he wanted some excuse to ban his meetings. Communal tension came handy to him and a limited flare up suited his purpose. However, things went out of control soon. Anyway, Mr. Singh could not
address the meetings in Meerut district.
The 16th April communal incident also was a contributing factor. It was on the occasion of a Muslim festival *Shab Barat*. As already pointed out, a section of Muslims have achieved a measure of prosperity in Meerut. It is also known that newly emergent prosperous sections, especially those belonging to lower social origins spend part of their earnings on celebrating religious festivals with much greater fanfare. The Muslims of Meerut also tried to excel each other by display of fireworks on this occasion. There was competition between two groups of Muslims. It is said a police dargha on duty was struck with a firework piece and as he was drunk he fired and as a result two Muslims were killed on spot. Another incident is also reported to have taken place on the same night. Muslims had arranged religious sermon near Hashmpura chauraiba. Nearby there was mundan function in a Hindu family. Some Muslims objected to film songs being played on loudspeaker and asked them to stop. There was argument followed by quarrel. Some body from the Hindu side allegedly fired. The Muslims then went on rampage setting some Hindu shops to fire. Hindus largely suffered damage. In all 12 persons were reported to have been killed, both Hindus as well as Muslims. Curfew was imposed and the situation brought under control. However, tension prevailed and preparations continued for revenge.
IV
The current round of violence began on 18 May after 9 p.m. Again, there is a lot of confusing reports about how it all began. The exact sequence of events is difficult to establish as there are many versions even by those who were present on the scene of events. Whatever the case, the scene of action was Hashmpura locality and near by Imliyan mosque—a locality where mostly Ansari Muslims live (i.e., those belonging to the weavers' community).
It is alleged a Khola (stall) belonging to a Muslim was set on fire by some miscreants. Thereafter electricity suddenly went off. It is difficult to say whether it was a deliberate act or the usual tripping. This was followed by a loud bang. Some thought it was transformer which exploded.
Anyway a crowd collected, Hindus as well as Muslims collected in large numbers. Then there began stoning on the crowd from Surajkund side. The crowd went berserk and Muslims are reported to have set afire a sweetmeat shop belonging to a Hindu and killing its owner Brij Bhushan.
The PAC came and surrounded Hashimpura after 11 P.M. and asked the Muslims to come out of their houses so that a search may be conducted. Nobody came out. The PAC thereupon entered Hashimpura for making arrests. The inmates stoned PAC and it had to retreat under stoning. While running a PAC jawan dropped his rifle (some say it was snatched). The PAC now came in much greater strength after midnight and announced that all should come out otherwise there will be firing. The people of Hashimpura and Imliyan came out along with their women and children. There were more than three thousand people who came and sat on the road outside the Imliyan mosque. It was now that the light came. The rifle which was dropped (or snatched) was found outside a house. The PAC entered that house and shot dead two brothers inside.
A jeep was parked nearby the Imliyan. A police officer wanted to drive away the jeep with some arrested people. The women resisted but the jeep drove off killing a girl and injuring three women (according to another version a girl and three women were killed and yet another version has it that no one was killed, only some women were injured). There was great commotion and some people started stoning PAC which fired in turn killing several people (six according to one version). Thereupon some people entered the mosque, took possession of the mike and appealed to the Muslims to gather together to save their brethren from the police and the PAC. They also raised emotional slogans which further added to the gravity of the situation. They may have done that out of fear but nevertheless it aggravated the matter. My investigations elsewhere also show that the Muslims have lost all faith in the law and order machinery. They prepare to defend themselves and the police takes this as a clear proof of their intention to create trouble. It thus arrests large number of Muslims which leads to further panic.
Same thing seemed to happen in Meerut on that fateful night. Large number of Muslims were arrested. It was really indiscriminate arrest. This angered the Muslims and added to this the news about a girl having been killed by the police jeep. The Muslim crowd now went berserk and started burning, looting and killing the Hindus. Many Hindu shops were burnt. Also, some factories (for weaving and printing cloth) at Pillo khdi were completely burnt. Ten to twelve persons (both Hindus as well as Muslims) inside these factories were burnt alive. On Hapur road at about 5:30 a.m., Dr. Prabhat, son of Prof. Harpal Singh who was proceeding to perform an emergency operation in his Fiat car was caught and burnt alive along with his car, Prof. Harpal Singh, his wife who is an educationist and Dr. Prabhat. All three are highly committed to communal harmony and have been working untiringly for communal peace in Meerut for a number of years. It was indeed a great tragedy that such a person became victim of communal monster. However, Prof. and Mrs. Harpal Singh had maintained their poise, were sad but not bitter about the incident which snatched away from them their only son.
From 19th morning curfew was clamped on the entire town and now the PAC and Hindu mobs took charge of the situation. They started burning and looting during the curfew hours which went on up to 23rd May quite unchecked. It is said quarter of the town was burning. Fire could be seen anywhere and everywhere. Even the cantonment area where all top government officials live was raging with fire. It was here that the stadium and Mawana bus stand are situated. There are a number of shops belonging to Muslims near the stadium. All were reduced to ashes along with 9 buses parked at the Mawana Stand. According to Muhammad Hamid and Nawab Tyerwala who owned buses told me that the arsonists came around early morning on 19th May and set fire to their 7 buses, two trucks and three tractors along with many shops around the stadium. They claimed that total loss due to burning of the buses and trucks was around a crore of rupees.
The devastation caused during these three days is to be seen to be believed. The worst sufferers were the residents of Shashtri nagar and Miyan Mohammad nagar. Miyan Mohammad
nagar presents a sight of total devastation. There were in all 228 houses belonging to mostly poor Muslims like rikshaw pullers, vegetable vendors, mechanics, bricklayers, etc. Not a single house were left intact. All were demolished, burnt and looted. It must have been a major operation indeed. The residents, mostly women I talked to, alleged that on 19th May at about 2 p.m. PAC came and suddenly began to fire which continued up to 5 p.m. Soon the PAC was joined by the Bangis and chamaris nearby and they started looting and setting fire to the houses along with the PAC. Many families lost their bread-earners along with everything they had. I listened to helpless crying women who had nothing left except clothes on their bodies. It was heart rending scene. Unfortunately the story of devastation of Mian Mohammad nagar has not been reported in the press perhaps because very poor Muslims live here. Many of them were burnt alive and their houses could still be seen when visited and many were killed in the unprovoked PAC firing.
Shashtri nagar, a middle class and upper middle class locality, was another scene of devastation. Generally in riots poor people get killed or their houses burnt. But since late this is no longer true. Many middle class Muslims were killed or their houses burnt last year in Ahmedabad and earlier in May 1984 in Bombay-Bhiwandi riots, those Muslims who were living in cosmopolitan localities. Same thing happened in Meerut too. Shashtri nagar is a mixed locality with Muslim houses strewn among Hindu houses in different sectors. Many of these Muslims here believed in living with other communities to promote better inter-communal relationship. They feel sadly mistaken now. Even the houses of a high income-tax official, a chief engineer, an executive engineer and a noted Urdu poet and a rich businessman who had recently migrated from Africa were not spared. Not less than 250 houses were burnt or looted. Here too nothing remained in these houses except four walls and the inmates either fled simply with clothes or were killed or burnt alive.
A young girl Kahkshan Shaukat who could speak English fluently and whose house was looted in sector 4 of Shashtrinagar told me that there was no tension in their area. The members of
peace committee assured them on the morning of 19th May that they need not be afraid and that in the event of trouble they would be protected. However, the trouble began soon thereafter. One Shiv Kumar Sharma living in house no 379 was provoking trouble, she alleged. Some 20 to 25 boys set fire to Muslim houses. It all began at 10 a.m. on 19th May morning. The peace committee members then expressed their helplessness and we had to flee. Our house was completely looted and nothing remained, she said.
According to Mohammad Aslam, also a resident of Shashtri nagar, one Mohanlal Kapur, an ex-MLA (BJP), Shakuntala Kaushik who runs a school in the area, Arun Agarwal and OP Khanna, the President of Vyapar Sangh Meerut (who has VIP connection) were openly inciting the mob to loot and burn Muslim houses. Many others also, especially those owning business establishments too complained about the complicity of OP Khanna but he has not been touched. He enjoys all the patronage of administration.
Many were burnt alive in Shashtri nagar. In two houses alone in which many people had taken refuge, some 33 persons were burnt alive. Some were killed in firing in the area. Isolated instances of burning have also been reported. The above residents claimed not less than 100 persons were killed in Shashtri nagar alone. But there is no way to check this figure. There may be some exaggeration which is understandable. Many girls are also missing from here. It is feared they may have been raped and killed. A shrine of Najafshah was also completely demolished. In civil lines area too five shrines have been destroyed and one has been converted into a temple. They have written "OM" in bold letters on that temple. What is most shocking is that many Congress MLAs were leading the mobs to loot and burn. Many eye-witnesses corroborated this charge.
Another tragedy of unprecedented nature occurred on the night of May 22. More than 300 persons were arrested by the PAC on that fateful night from Hashimpura, the main trouble spot and were carted away in trucks. Most of them were young and middle-aged. Some were taken in jail but around hundred who were not shown on record, were taken in several trucks to
Moradnagar near Ganga canal. They were lined up near the canal and shot one by one. The horrified story was narrated by three boys who miraculously survived by feigning death. They escaped. Two of them are in a Meerut hospital under heavy guard of CRPF and are not allowed to talk to anyone. Their statement, however, has been recorded by a magistrate and is sealed. The contents are not known. The third somehow reached Delhi and took shelter with Janata Party leaders. He was presented to the press by Janata Party President Shri Chandra Shekhar.
More than twenty-three bodies have so far been found floating in the canal. More are being discovered. Again Heaven alone knows how many were killed. Many allege not less than 100 were shot by the PAC near the canal, as many are reported still missing.
Massacre in Malyana
Another tragedy struck this sleeping village some 6 klo meters from Meerut, in fact it is situated on its outer periphery. What happened here is no less shocking than what happened near Ganga canal. Perhaps more. The brutal massacre, once again by the murderous PAC would put any civilised nation to shame. Malyana has been described by some papers as Mai Lai of India. Perhaps the comparison is not inapt.
The survivors of the massacre in Malyana narrated the harrowing story to me when I visited them on 9th June in the afternoon. It all began, according to Mr. Zaheer Ahmad, an inmate of relief camp, at about ten in the morning of 23rd May. A liquor shop was looted by the PAC belonging to a Hindu and the liquor was distributed to the Harijan population of the village. Then suddenly the PAC started shooting at about 2-30 p.m. They positioned themselves on the terraces of the Muslim houses and began to shoot inside the houses. There are around 5,000 Muslims in all out of twenty-five thousand population of the village. Most of them are poor agricultural workers or small land holders. They live together in clusters around a pond.
There was no tension in the town before the PAC descended on it totting its guns. Even the Hindu inmates of the village
admitted There was no Muslim conflict either, a Scheduled Caste resident told me. The PAC came to the village on the pretext that the Muslims of Malyana were hiding some who fled from Meerut evading arrest. This charge was vehemently denied by the Muslims of Malyana. Anyway, the shooting spree of the PAC began at about 2:30 p.m. and continued up to 5:30 p.m. The Harijans of the village joined the PAC to loot and burn (Liquor from the vand looted earlier by the PAC was supplied to them to make them drunk, alleged the Muslims). Almost all the houses were looted. Many got burnt alive inside the houses. Most pathetic case was that of Mr Mahmood's family. He, his wife and 4 children, six in all, were trapped inside and burnt to death. Another child, a girl, was away with grandmother and thus survived. In all, according to the list prepared by the village Muslims, 78 died in firing and burning operations. Several are still missing, more than hundred, some of whom have been subsequently traced. Dead bodies are still being discovered. I saw the graves of the dead near the dirty pond. Two to three persons have been buried in each grave for want of space.
The DIG Moradabad Range was also present during the shooting spree and subsequently negotiated with the Muslims of Malyana who had come out of their houses with their hands up and pleading either to be arrested or shot. The CRPF arrived at 2 a.m. and Muslims went back to their houses (whatever was left of them) and felt little more secure. A judicial inquiry has been ordered.
Everywhere in Meerut I was told that Chamars and Bhangis (Scheduled Castes) joined hands with upper caste Hindus in looting and burning. It is highly disturbing phenomenon and must be seriously taken by those who advocate Dalit-Muslim alliance. Probably these were given liquor and money in most cases to do the dirty job and also got the loot. Their political consciousness seems to be very low and have basically Hindu ethos. This is fully exploited by the upper caste Hindu communal leadership. It is for the BSP and other Dalit leaders to ponder over the matter and take steps to raise political consciousness of these people.
VI
It is really difficult to estimate how many died in these
unprecedented riots in Meerut. It is the worst of all other riots I have so far investigated. It is rumoured that more than twenty thousand died. But it is ridiculously high figure. This belief gets strengthened by discovery of bodies every day from nullahs, wells and obscure corners. Two other more moderate estimates of the dead are 1500 and 350. The Government admits only little over hundred deaths which again seems far below the actual number. My guess is about 400 deaths in these riots. The Government gives its estimate on the basis of post mortem performed but that is not the reliable way, at least in the present case. Firstly, dead bodies are being discovered every day until now. Secondly, there were many powerloom workers from Bihar who lived alone in Meerut without family. Many of them are reported missing. While some have fled to their native place, some have been certainly killed. It would be definitely known only after some time.
The arrests were equally arbitrary. Most of the arrested are innocent people with no previous criminal record of any kind. The PAC and the police arrested whomever they could lay their hands upon. When I visited a jail in Meerut there were scores of children of 12 to 15 years old. Even the three survivors of the Moradnagar canal massacre are of 20 years or below. I also talked to some people whose near and dear ones were arrested. It was not difficult to guess that the arrested ones were innocent and had played no role in inciting riots or taking part in it in any other way. Only a few of them may have been involved in stoning PAC which was the only way to defend themselves. This is not to say that there are no mischievous or anti-social elements among the Muslims who took part in the riot. But there are. But most of the arrested are not. The total number of arrests made according to the Government sources is between 2,400 and 2,500 but actual number is said to be around 4,000. Needless to say most of the arrested are Muslims though Hindu mobs were openly indulging in looting, burning and killing. Some prominent among them have been identified. They include MLAs and ex-MLAs also. However, none of them has been touched.
VII
My investigations lead me to certain conclusions which others also seem to agree with. Some journalists have also written about them while reporting on Meerut riots. Firstly, the Chief Minister of UP is certainly not above board. He had his own political motives in not quelling the riots quickly, if not initiating them. He was certainly not sympathetic towards loss of life and property of minorities. Secondly, the administration has been thoroughly communalised. Though there were some officers of proven integrity but they were made ineffective. Radhey Shyam Kaushik, the district magistrate, is not communal but he lacks independence and tocs the line. He does the bidding of his political bosses. In fact, if the administration is determined, and is not lax in any way, it checks communal violence in no time. No administration worth its salt would allow the mobs to indulge in loot and arson openly during curfew hours, much less shooting down people inside their houses.
Fourthly, the PAC's role was most shameful. It openly indulged in mass murders unchecked. No democratic government can tolerate such a rank communal force. But the force has not even been withdrawn from Meerut, let alone being disbanded and severely punished for its unpardonable crimes. Its commandant Tripathi was suspended for a day and again reinstated next day 'for fear of revolt in PAC'. The PAC's composition itself is castest and communal in character. There are Brahmins, Yadavs, Ahirs and Rajputs in its ranks. Despite assurances given after every major riot to recruit from amongst lower castes and minorities up to 25%, nothing happens and the assurance is once again repeated after yet another riot. So much for the honesty and integrity of the Government both Central as well as the State. If the Government has any respect for the human life it must immediately disband this monstrous force. The communal character of the force is obvious from the fact that while minorities dread it, it is eulogised by the majority communists, i.e., the communalists amongst it.
Fifthly, the Babri Masjid/Ram Janambhoomi controversy must be immediately resolved. If early steps are not taken in that direction, for lack of political courage, or for manipulating
votes, there will be many more Meeruts. The whole of North India is boiling with rage on this issue. The Central Government has primary responsibility in this matter. It should not evade it.
If the Government does not fulfill its responsibility in this direction, the secular intelligentsia should assertively come forward and put pressure on the Government to resolve the issue. The intellectuals from amongst the minority community should also appeal to their fellow religionists not to be misled by the hyperbole of the irresponsible members of the Bahri Action Committee. They are only serving their political interests at the cost of innocent lives of Muslims. No words can be minced about it.
Time has come for all secular elements to join hands together to resist the menace of communalism. The Fascism is surely on the march in this garb. Without strengthening secularism, let us remember, democracy has no future in our country.
MUSLIM POLITICS: THE AWFUL PRICE OF THE SHAHABUDDIN-BUKHARI LINE
M J Akbar
Meerut, June 1 Salim Ahmed Ansari loves the nhas he himself has promoted he is know in Meerut as Salim urf (the Urdu for alias) Bhindranwale Around 40 years of age, he edged his way into politics through the Muslim Youth Conference organised by Javed Hashb When the Babri Masjid issue came up, he at once hitched his wagon to this cause His road to "salvation" for Muslims, repeated in conversation after conversation with the city's youth, was, "larkha, marna hai (we will fight and die)" And he would buttress this line with the example of Punjab "Ek karor sardar hain, aur yeh to samhal nahin rahe Rajiv Gandhi sey, ham 20 karor hain, hamen kaise Samhalega (There are one crore Sikhs, and Rajiv Gandhi cannot tackle them, how will he tackle 20 crore Muslims?)" The accuracy of the figures is not the point, the direction of the message is Other, more famous, Muslim leaders of this ilk have been giving their support to such preaching Ansari organised a meeting of the Hyderabad MP Salahuddin Owaisi, and Owaisi's message to the Muslims of Meerut was "Taakat pada karo Taakat maa ke gaatlon se samjhauta karwa deti hai (Build your strength It is strength which creates a settlement with the killers of your mother)" It does not take much intelligence to appreciate the reference points
The new politics of these modern extremists is actually an echo of the old Muslim League politics. "Turn nationalism into
a bargaining point. Threaten the state with revolt in order to win your demands. And arm yourself to prove how strong you can be."
This provocative challenge to India was given a kind of formal status at the Boat Club rally in support of the Babri Masjid demands. There is no proper understanding yet of the impact the speeches made there have had on excitable Muslims, particularly the youth. Meerut, being a town close to Delhi, sent thousands to that rally. Imam Bukhari's speech there, declaring that this country could not be called a legitimate home of the Muslims, has been reported, but not the more dangerous speech of Syed Shahabuddin. This correspondent has a video cassette of that speech, and the implicit and explicit call to violence that exists there has to be heard to be believed. Syed Shahabuddin warned India of the "consequences" of Muslim anger.
The reason why Shahabuddin is more dangerous than the Imam is because he provides the legitimacy of a respectable politician to these provocative arguments. When the Imam says all this, well he was a loudmouth extremist even when everyone from H.N. Babugunn to Mrs Indira Gandhi kept tripping over one other to purchase his support. But Shahabuddin represents what secular Indian Muslim leadership should be all about: his career in the foreign service was interrupted when eminent leaders like Atal Behari Vajpayee and Chandra Shekher brought him to politics to become the Muslim leader of the Janata ranks. Today, he is an important spokesman of the Janata Party, and not the Muslim League. And if a senior Janata MP can give a barely disguised call for insurrection, then why blame the Muslim youth in a depressed mohalla from believing that perhaps there is no answer except violence? And if upstarts parading themselves as Muslim Bhindranwales find patronage from the Babri Masjid committee (whose most famous member is Shahabuddin), then what justice is there in picking up a poor and frustrated young man from mohalla like Hashimpura and condemning him as an antisocial and antinational element?
For the last two years, and in particular after the Babri Masjid controversy, this message of salvation through confrontation has been drilled into the Muslim mind in *mohalla* after *mohalla*, through meetings small and large, addressed by extremist leaders minor and major. A whirlwind was being deliberately sown, and we have begun reaping it. Opposition parties like the Janata condoned it because they were convinced that only Rajiv Gandhi would suffer (although the one Janata chief minister, Ramakrishna Hegde, was wise enough not to allow his own party MP Shahabuddin to hold meetings in Bangalore).
There is great political profit in death of course. Those who light the fires typically run away from the arson. Salim Ansari has spent the whole period of the riots in Delhi. Imam Bukhari can take comfort in the fact that the administration did not allow him to enter Meerut. Shahabuddin went for a couple of hours, and extracted every ounce of publicity with photographs sent to the newspapers (for some unknown reason Shahabuddin was laughing in the picture—no one has understood what there was or is to laugh about).
The tactic is a transparent one: build up passions, and provoke the community to the point of the riots which, when they come, will justify your original call for militancy.
And it is these leaders like Imam Bukhari and Syed Shahabuddin who have given fresh currency to the charge of antimationalism. That there is a powerful section within India always ready to hammer Muslims with the partition stick is not a secret. But this belief is spreading today, thanks to the speeches being made by the most prominent leaders of the Muslim community. The problem is accentuated by the fact that the traditional political machinery in small towns has been replaced these days by an aggressive lumpen element eager to fill the vacuum.
We have reached then a circle of vicious and horrible distress on the one side political parties like the Janata giving legitimacy to leaders like Shahabuddin, and on the other criminal and communal forces like the PAC using this excuse to indulge in the massacre of innocents. And in this process, the people are getting communalised as never before. There is no Congress and Janata and Lok Dal during riots in Meerut, there is only Hindus and Muslims. Even so called "leftist" Congress
leaders of Meerut have become great devotees of the PAC.
And the man who has used the anti-national stick against Muslims in Meerut is Vir Bahadur Singh, the chief minister of Uttar Pradesh, and even when Rajiv Gandhi came to Meerut, V B Singh organised demonstrations in favour of the PAC, and even as non-controversial a leader as Mrs Mohsina Kidwai was insulted and abused with the tacit consent of the chief minister.
All through, the administration in Meerut has adopted an anti-Muslim attitude. And if there was any more evidence needed to prove the point then it came with the transfer of two senior police officers who had the reputation of integrity, VKB Nair and Nathu Sai. The officers who allowed or participated in brutality of the most horrendous kind are being protected by the state administration. And the line is being spread that it was the Muslims who were the real aggressors.
There is truth in the charge that it was the Muslims who first provoked the riots "Inspired" by the Bukhari Shahabuddin loyalists, they began the trouble on the 18th. But after that the revenge has been severe. As we reported yesterday, the police and the PAC even indulged in arbitrary killing. There was Maliana. And today survivors are talking of a figure of around 200 in one incident alone near the Hindon canal.
These weavers from Hashimpura became the "sala Pakistanis" whom the PAC sent to their death in cold blood, all pawns in a game being played for stakes as high as power. Votes are being counted in terms of corpses, and by every player—both the Shahabuddins and the Vir Bahadur Singhs. After 40 years, our nation's acceptability level for murder and mayhem has risen to unprecedented levels—that is one of the "achievements" of our system. And yet there comes a point when the redcoated mind must demand a halt, an accountability.
Imam Bukhari can perhaps be left alone in the petty world he has created. But at least leaders who claim that they feel a sense of responsibility towards the country must take action—Chandra Shekhar against his communal henchman Syed Shahabuddin, and Rajiv Gandhi against his communal chief minister, Vir Bahadur Singh. That will not solve the problem, which is already a monster much larger than its progenitors. But it will be a very small beginning towards that end.
THE AGONY OF MEERUT
Inderjit Badhwar, Tanu Midha and Others
As almost always the case, last fortnight's bloody communal carnage in Meerut began with an insignificant event. On May 16, Ajay Sharma, a resident of Kotwah, was gunned down in a land dispute—something hardly unusual in that area. A case was registered and the police rounded up four people. Following the leads they provided, a police posse, with some back up from the Provincial Armed Constabulary (PAC), went on the night of May 18, to Yameen from Hashumpura.
The resistance to the arrest was stiff, particularly since it was the Ramzan period and the police had ill timed their entry to reach around Iftar, the time when Muslims break their roza for the day to eat the evening meal. The police tried to force their way in, and in the melee three PAC men were dragged into the lane and their rifles snatched. Initially, it was only a fight between the police and the Muslims. But the affair soon assumed communal colour as a roadside stall—owned by a Muslim, but rented by a Hindu—was set ablaze and a shop owner nearby, Shashi Bhushan was stabbed to death.
In the confusion during the melee, somebody came onto the loudspeakers of the Imiliyan mosque calling all "believers of the faith" to come out and protect their religion. Predictably, this brought not just the Muslims out but also the Hindus. According to District Magistrate Radhey Shyam Kaushik, echoes of this announcement were heard from masjids in
neighbouring localities too, and soon it was a free for-all on the roads. Nobody knew who was killing whom and who was burning what in the dark. All we could see was fire at night and smoke for in the morning," recalled Mehboob Mian, as he huddled in his room in Hashimpura.
In that first round alone, 15 people were killed and more than 350 shops and three petrol pumps reduced to ashes. Indefinite curfew was clamped in the area and the army conducted a flag march later that afternoon. Yet the arson had spread. By the next day the rioting had affected nearly every part of the city. Even places like Shastri Nagar, Saket and Civil Lines, which had never been affected by riots, reported incidents. Curfew was clamped throughout the city.
Soon enough, the violence spread to the walled city of Old Delhi, where eight people were killed in 24 hours and scores injured. And all of Uttar Pradesh, where the majority of India's Muslims live, became a high tension communal wire threatening to spark off conflagrations of unimaginable dimensions. Over the next few days, Modinagar witnessed an exchange of fire in which two people were killed, four people died in a bomb explosion in Bulandshahr, another blast sent the army rushing into Hapur, and reports of tension flowed in from Gaziabad, Muradnagar, Muzaffarnagar and Moradabad. Prohibitory orders were enforced in 36 districts, and a complete ban on public meetings was imposed throughout the state.
For the next 10 days, the only movement at daybreak in Meerut was that of men in helmets and war fatigues with sten-guns slung over their shoulders, inching their way out of their shelters and climbing aboard convoys of trucks. At night, the deserted streets were taken over by stray dogs chasing police vans and jeeps. Meerut's six lakh terror-stricken residents remained shuttered and bolted within their dwellings, as more than 13,000 members of the army, paramilitary forces and armed police laid siege to their city in the hope of stamping out the violence in every mohalla and bylane.
What had turned one of the busiest and most frenetic towns in North India into a ghost town, that resounded day and night to the rumbling of police trucks and the tramping of military boots, was the unprecedented communal frenzy in
which more than 150 people died and 1,000 were wounded. More than Rs 20 crore worth of property lay destroyed, and business losses until the end of the fortnight were estimated at Rs 50 crore. A city which had thrived on jobs provided by tailors' scissors, handlooms, and sporting goods had almost overnight, it seemed, abandoned its peaceful tools of trade and taken up guns, bombs and knives.
The gruesome and bloody carnage suggested parallels with New Delhi's holocaust in 1984 and Ahmedabad's homicidal war of 1985. Barely two months earlier, police battalions had patrolled Meerut's streets following riots in which 12 people were killed. But in the first week of May, in what turned out to be a fatal blunder, the local administration, believing the situation to be normal, withdrew the extra police force. When the fighting broke out barely 10 days later, on May 18, the district administration found itself virtually crippled in its reaction because of the shortage of reinforcements. As heavily armed and slogan shouting mobs belonging to both communities began fighting pitched battles, the 500 strong district police, immediately available, found itself hopelessly unequal to the task of containing the violence. By the time reinforcements were deployed, the damage had been done.
The withdrawal of the extra police contingents was particularly perplexing as, during the previous two months, intelligence reports had indicated that members of both communities had begun to stockpile huge quantities of arms. The state intelligence bureau had been constantly feeding reports to the Home Ministry that the threatened show of force over the never ending Ram Janambhoomi Bahi Masjid issue could precipitate communal trouble in sensitive districts, particularly in Meerut, Rampur, Moradabad, Kanpur, and Allahabad.
The administration had also done absolutely nothing about local intelligence reports that had warned of a communal colour being put on Ajay Sharma's death, and which asked for intensive patrolling of areas affected by the April riots. But not only did the administration ignore these warnings, it even released those detained for disturbing peace in the previous riots. The administration had no clue about the level of preparations that were on "in case of a spark". Indeed, the
evidence, both mental and physical, was overwhelming. There was a relay of announcements from places of worship, organised mobs in hundreds on the streets, and plentiful use of firearms. Yes the police and administration appeared oblivious of all the action.
The excuses continued till Home Minister Buta Singh, after a tour of the riot affected areas, castigated the administration for not being geared for the riots. At that point, State Home Secretary Mata Prasad conceded "It is no doubt our failure." Union Home Ministry sources confirmed that the state administration was taken by total surprise when the "unprecedented and unanticipated violent activity in May" broke out.
One major cause of the flare up, say Home Ministry sources, was the incompetence of the district administration. Its police official were suspect, most of its inspectors and sub-inspectors had been sent to Meerut under political pressure from other parts of the state and even the city Superintendent of Police B K Chaturvedi—who was finally transferred out and replaced by B K Gupta along with Varanasi's senior superintendent of police, G L Sharma, flown in as additional senior superintendent—was considered a controversial official. Although no changes were made at the civil administration level, the district magistrate was rendered virtually redundant when Minister of State for Home, Surender Singh Chauhan, pitched camp in Meerut and proceeded to take charge of the situation.
But as the riots continued even after the mobilisation of additional forces Prime Minister Rajiv Gandhi summoned Chief Minister Bir Bahadur Singh to New Delhi for emergency consultation. Talking to newsmen at the Iftar dinner hosted by Union Tourism Minister Musfi Mohammed Sayed Rajiv admitted that "communal conflicts were posing a grave danger to the country" and disclosed that he would visit Meerut the next day to take stock of the situation. But the following day it was Buta Singh who heli-ported to Meerut for a three hour visit to the city. He ordered the confiscation of all arms, house-to-house searches and the immediate arrest of "bad characters".
The administration was also asked to re-establish communication with community leaders, and political leaders were instructed to form peace committees. These directions were
given only after Union Minister for Urban Development Mohsna Kidwai complained to the Prime Minister—following her own visit to Meerut—that the district administration was treating the ordeal as a law and order problem and had failed to take such steps as convening regular peace meetings between community leaders.
For the first time in Meerut house to house searches were conducted in all the sensitive localities, including Sohrab Gate, Jali Kothi, Shahpur Gate, Gudri and Kotla. These searches, conducted on a massive scale, involved a complete cordonning off of entire localities by the army, which would position itself in every lane, on rooftops and at every entrance to the locality. Then announcements would be made over the loudspeaker asking all male members of the house to come out with their licensed arms. Hundreds poured out and were escorted to nearby compounds. Then the CRPF women companies would move into houses and search the entire area.
There were large scale seizures of firearms, including telescopic guns, pistols, single and double barrel guns and rifles with plenty of ammunition. Within a week, more than 2,500 people were arrested in Meerut alone, which has a total jail capacity of just 800. Detenus had to be shunted off to jail in Fatehgarh and Agra.
But just as things appeared to be returning to normal, there occurred a massacre that will probably go down in police history as one of the worst cases of police brutality. The news of this caused fresh eruptions in neighbouring areas in the state. Mahana, a small cluster of villages 5 km from Meerut, had not yet come under curfew. A PAC platoon had been sent there to keep the tension under control. The area has a population of around 35,000, of which 4,000 are Muslims. At 2.45 p.m. in the afternoon of Saturday, May 23, the PAC contingent arrived. Joined by some Hindu residents, they began burning down Muslim houses. It took them precisely two hours to complete the task they had assigned themselves. At the end of those two hours nothing stirred in Mahana as people sat huddled together, gripped by fear as the stench of smoke and human bodies rose around them. Those who ran out to escape the inferno were showered with gunfire by PAC men positioned on
rooftops. Scores fell to the bullets.
This happened just 70 km from the capital, but news of the massacre began to filter through only on Sunday evening. Monday morning's newspapers carried the official toll of "seven killed in police firing". In silent outrage the survivors pointed to charred parts of bodies in burnt houses. They pointed to blood trails and marks showing that bodies had been dragged out of houses down the lane. The PAC unit was quietly withdrawn and army sent in, but officially the toll remained seven. Army officers admitted on condition of anonymity that they had personally counted over 50 bodies when they had arrived in Maltana.
While officially Meerut remained calm for the third day running and life passed off peacefully, the remainders of the brutal killings in the town were still floating down the Ganga canal 30 km away. The Government said 16 bodies had been recovered from the canal, but once again the unofficial toll was far higher. All the bodies were naked, many had bullet injuries in their backs. In a two-hour drive along an 8 km stretch of the canal from Muradnagar to Massoorie, India Today reporters saw two bodies floating down the canal.
Shahheed of Nahal village said the bodies had been floating past for the previous eight days. He had counted 40 bodies in one day. Yusuf Mian, an irrigation department worker of Jhal Massoorie claimed to have seen 10 bodies a day in the canal for four days. But the most gruesome account came from Mehdi Hussain, a contractor who runs six *atta chakies* at the canal weir at Jhal Dasna. Hassan, who supervises his *chakkies* by the canal the whole day, said that on May 26 he had counted 9 bodies and many more in the preceding days. At times, four or five bodies together would come floating down and get stuck in the iron grating across the channel weir. They would get the Irrigation Department's *bhangi* to pull the bodies out with a hook and dump them in the canal beyond the weir gate.
What made this potentially explosive was the fact that the Ganga canal runs through predominantly Muslim territory through Ghaziabad district which adjoins Delhi. Inevitably, communal hatred was spreading into the rural countryside too.
But in Lucknow, M C Dwivedy, PAC deputy inspector general, the de facto chief of the force, insisted that he knew nothing of any massacre at Mahana. And there was no official information on which company from which battalion, and under the charge of which officer, was sent to Mahana. But as the bodies started floating down the canal and the tethered cord of tolerance came close to snapping once again, the chief minister finally announced a judicial inquiry.
But the PAC brutality was not the only element that cast its deathly pall over Meerut. In many ways Meerut today resembles what many Delhi localities looked like in the aftermath of the November 1984 riots. The town is pock marked with burnt out shops and houses charred skeletons of dozens of buses and trucks still litter the main roads. And there were clear signs that some organisation and preparation had gone into fuelling the uncontrollable communal frenzy which swept the town.
But when compared to Delhi in 1984, there was one significant difference. In Meerut, both communities seem to have been gripped by a junoon (madness) and struck out at each other with equal ruthlessness. Selected Muslim shops in Hindu areas were looted and burnt down while Hindu shops in Muslim areas met the same fate. In past riots, the fringe elements of both communities and goondas were involved. This time almost all the Muslims are ranged against all the Hindus. "Political party labels have ceased to have any meaning in Meerut. Today you are either Hindu or Muslim", said Mirajuddin Ahmed, 28, president of the Meerut Youth Congress, drawing a picture of a communally polarised society.
There was little scope left for humane behaviour, as two doctors belonging to the two communities discovered to their cost. Dr Atul Bhatnagar is a well known surgeon who runs a nursing home in the town. Many of his Muslim patients were awaiting treatment when a Hindu mob surrounded the nursing home and demanded that the doctor hand over all the Muslim inmates of the nursing home. Bhatnagar refused and spirited away his Muslim patients to a hospital. The mob vented its anger on the nursing home itself, causing extensive damage.
A less drastic fate met Dr Sirajuddin Ahmed, who donated blood, a rare AB group, to save the life of a young Hindu, Rishi Ram, who had been grievously injured by a police bullet. The blood saved Rishi Ram's life but Ahmed was abused and pilloried by his own community members for saving the life of a Hindu.
Hakim Saifuddin Ahmed, a recipient of the Padma Shri and an honorary physician to many presidents, including Zail Singh said most of the nafsat (hatred) between the two communities had been caused by the Babri Masjid and Ram janambhoomi action committees. "It has become so bad that if two buffaloes get into a fight, one will be labelled Hindu and the other Muslim," he said.
But why Meerut? There were several answers. With its proximity to Delhi, the town is particularly susceptible to national communal politics emanating from here. Every ripple in the capital is first felt in Meerut. "Any activist from Delhi with a few hours to spare, slips in to spread his message," said a political worker.
There is also acute trade rivalry in Meerut, known for its textile trade as well as local crafts like scissor-making. In textiles, it has traditionally been the Hindus who have owned the factories, and Ansari Muslims who have provided the labour force. The friction between the owners and workers adds fuel to the fire each time there is a communal flare up. Demographically, the six lakh population of the city is divided almost equally between Hindus and Muslims. And they are both equally prosperous. The tensions are inevitable as each is intolerant of the other's well-being. Add to all this a weak administration, and the weight of the Babri Masjid dispute. A flare up was inevitable. After last carnage, confidence building was clearly important.
And the postponement of Rajiv Gandhi's planned visit to Meerut was a negative signal. "Why didn't Rajiv come here? Why does he only go visiting tribal areas?" asked a caustic official in Meerut. But prime ministerial visits apart, there was little doubt that communal tension could not be eased in the State and several parts of the country without a way being found to untangle the Babri Masjid issue, or taking cover
behind the fact that it is a court matter. And there was equally little doubt that unless determined efforts were made to turn the tide against communalism in politics, with political parties taking the lead, the communal problem would continue to fester.
—Excerpt from India Today,
June 15, 1987
What we normally understand by a communal riot is only half the story of what happened in Meerut last month. A communal riot has come to be known in common Indian parlance as the outbreak of a clash between members of two communities—very often between Hindus and Muslims—in a city or in a part of it, which often results in breaking of heads, a number of stabbing incidents, some looting of shops and sometimes arson.
It may last a few days or a few weeks, and then by the exercise of an extra dose of police enforcement combined with some efforts at mutual understanding initiated by the local elite of the two communities, normalcy returns.
But the full story of the Meerut happenings in May is something which far oversteps this traditional dimension of a communal riot. Once the curfew cordon was relaxed and the place ceased to be totally out of bounds, the news of the ghastly happening began to come out; the special feature of Meerut 1987 has slowly begun to seep into our consciousness. But it is still a slow process, and few have yet grasped its fearsome implications.
By the run-of-the-mill journalist's yardstick, there is perhaps nothing extraordinary in the Meerut happenings, however deplorable these may be. There is evidence of inflammatory incitement. There was the collection of small arms, not just brickbats. There were the goonda gangs interested in looting shops and settling scores with their rivals and adversaries. There were the usual stories of harassment of innocent citizens, the curfew, etc.
But something more has happened in Meerut which has never happened in this country. It has now come to be known that at a certain place the PAC gangs suddenly descended, knocked and burst into the hutsments of poor Muslims, picked up whomsoever they could grab (mainly young), packed them off into PAC trucks, took them far away and pushed them out of the trucks and lined them up. They then shot them down and threw the dead bodies into the river.
All this happened not at the height of a battle royal, nor were the goondas rounded up and shot down, though even that would have been reprehensible and untenable under any rule of law. The points to note are: First, those picked up and packed off were not treated this way because they were culprits involved in rioting and killing but were victims of an indiscriminate round up of all and sundry. Secondly, there was absolutely no pretence of bringing them to justice or detaining them under extraordinary laws meant to deal with large scale communal violence, but to kill them in cold blood and dispose of their corpses by throwing them into the river.
Nowhere was there any attempt whatsoever to implicate those picked up with any specific allegations of any criminal act. The ghastly operation of shooting them was meant to strike terror and nothing but terror in a whole minority community—not to single out those suspected to be involved in riots or in inciting riots. This is the stark, simple fact. The message is unambiguous.
The serious implications of this PAC killing seem to have been missed by both the government and political parties. Here is a clear case of an organ of the state going out with cold-blooded calculation to raid and round up a whole group of citizens, whisk them away, shoot them while in custody and then throw their bodies into the river. Such fiendish acts have never taken place before. Let it be repeated—these persons were not arrested while actually engaged in any violent act, the police produced no charges against them, they were not told where they were being taken and then they were liquidated. Doesn't this remind one of the Nazi pogroms against the Jews?
There are, of course, innumerable cases of the police
humping off individual prisoners in cold blood and then passing them off as having been "killed in an encounter." This has become common practice against Naxalite prisoners in some parts of the country. There are other cases of a large number having been gunned down for having put up resistance against arrest. But the PAC in UP has excelled in all this.
And what has the government done to meet this situation? The UP chief minister was camping at Meerut. The prime minister after putting off his trip once did visit Meerut. But the PAC was not reprimanded, nor the officer-in-charge suspended. Rather, the chief minister was at pains to disabuse the press that the PAC commandant had been suspended as reported earlier. A judicial enquiry has at least been ordered. But can the government give the minority community its sense of security back by holding a protracted judicial enquiry? For any government aware of its responsibility of ensuring the citizen's security, the least it can do is to disarm and reprimand on the spot this terror gang in the employ of the government.
The PAC of Uttar Pradesh has of course a shameful record of communal brutalities. More than once the government both in UP and at the Centre had stated they would discipline it and reorganise it along secular lines. Nothing of the kind happened. Instead the PAC this time has broken its own record in Hitlerite barbarities.
Political leaders have condemned communal violence in Meerut. Some have tried to prove that the Muslims started the riot, others have pointed out some of the slogans raised such as "Pakistan Zindabad." It is also said that the total damage due to loot and arson hit the Hindus more than the Muslims. This is because the Hindus there are better off than the Muslims.
Some have of course criticised the PAC's high handedness. But the serious qualitative change brought about by the PAC picking up citizens from their homes, herding them into trucks and killing them, has not yet been sharply brought into focus. If this is not nipped in the bud, there will be these deadly pogroms. It will be not the death of secularism but the birth of fascism.
This is the danger signal from Meerut.
--Excerpts from Times of India
As Meerut and adjoining riot-affected areas limped back to normalcy this week, there was growing concern at the realisation of the ominous and unprecedented nature of the 10 day-long orgy of murder, rape, loot and arson, which slowly came to light.
On all counts, the riots in Meerut this month stand out in the entire history of communal violence in the country. They have been surpassed in sheer savagery only four times in the past—the Partition holocaust, the Assam massacres, the anti-Sikh carnage after Indira Gandhi’s assassination, and the Ansari Baug massacre in Bhiwandi in 1984.
There are few takers in Meerut and elsewhere for the official death toll of 133, as reports show that in Maliana village alone, the Provincial Armed Constabulary (PAC) along with Hindu hoodlums slaughtered over 100 Muslim men, women and children in cold blood. While unofficial estimates of more than 20,000 dead may well be exaggerated (not unusual in the aftermath of such a bloody riot), it would not be surprising if the actual death toll is well over the four-figure mark.
Yet, the number of victims, however gruesome the toll, is less horrific than the criminally manifest negligence of the Central as well as state administration in checking the wave of communal violence. Equally horrendous has been the blatantly communal role of political parties, most notably the
Congress (I) in the riots. Most frightening of all, normal law-abiding citizens, both Hindu and Muslim, for the first time actively participated in the frenzy of killing and depredation.
The colossal administrative failure to stop the spiral of violence in Meerut is highlighted by the fact that top ministers, politicians and officials were palpably ineffective despite being on the spot. Home Minister Buta Singh, Minister of State for Home Affairs P. Chidambaram, Urban Development Minister Mohsina Kidwai (whose Parliamentary constituency is Meerut), Uttar Pradesh Chief Minister Veer Bahadur Singh, UP Home Minister Gopinath Dixit and State Director General of Police S.K. Bhatnagar were all present in the city and had first-hand knowledge of the orgy of violence. Yet, the riots continued.
Chidambaram for instance, reached Meerut on the third day of the riots. He has visited the city twice since but has still failed to provide a correct assessment to the Central government about the extent of the rioting and what needed to be done to bring it under control. Buta Singh arrived on the fifth day when the official toll had already reached 54. Despite repeated requests by local officials he felt it was unnecessary to hand the town over to the army. The day after he told journalists that the situation in Meerut had been brought under control, the Maliana massacre took place.
Mohsina Kidwai who normally should have played a key role in defusing communal tension, proved to be a virtual liability and in fact a source of added tension herself. As soon as she arrived, a large number of prominent Hindu citizens including many local Congress (I) leaders warned that there would be more violence if the MP from Meerut was allowed to visit curfew bound areas. Mrs. Kidwai was greeted with filthy abuse by her own partymen in front of Chidambararam and chief minister Veer Bahadur Singh. She was later sidelined completely and not allowed to either go around Meerut or participate in official meetings between Central leaders and Local administrators.
If the Central leaders proved wholly inadequate in dealing with the situation, the state level leadership was embroiled in its own political games. Several officials in Meerut complained about
the way the chief minister showed amazing complacency about the mounting violence in the city. Many believe that he deliberately soft pedalled the taking of drastic measures against rioters because the violence fitted in well with his bid to divert attention from the public enthusiasm which VP Singh was arousing.
The chief minister is also believed to have taken particular pleasure in carrying out a slander campaign against Mohsina Kidwai and reportedly egged on local Hindu Congressmen to humiliate and insult her in front of Chudambaran Singh, who is believed to be an old political rival of Kidwai and has been seeking an opportunity to cut the Central minister down to size.
As far as ministers and officials like home minister Gopinath Dixit and S K Bhatnagar are concerned, they seemed to be unaware of even elementary details of the law and order situation. Persistent attempts by reporters to get their assessment of the riots met with the standard answer, "The District Magistrate is the spokesman of the administration, you talk to him."
The District Magistrate, RS Kaushik, overwhelmed by the presence of such top Central and state dignitaries and pressurised by local communal lobbies, looked a tired and broken man by the first few days. Soon his lack of authority was apparent from the way he was being openly pushed around by junior police officers and small time local politicians.
It is not surprising, therefore, that despite the deployment of the army from the second day of the riots, the violence continued for more than a week. Army officers privately complained that troops had received no clear cut orders and sounded confused on the extent of their authority. They were also aghast at the level of communalisation of the local administration.
An army Major, preparing his men to launch a house-to-house search in the Suraj Kund of Meerut was so exasperated by the behaviour of a Sub District Magistrate who kept insisting that only the Muslims were responsible for the violence, that he finally told the latter to shut up. "Even educated, responsible officials here have been affected by the communal madness," the Major remarked bitterly.
Further, a lack of coordination between the troops, paramilitary forces, and the local police was evident. Central Reserve Police Force Commandant AS Khan waited in his hotel for four days before he was called for a coordination meeting for the first time. "The riots could have been contained within the first few days had the local administration gone about it in the proper way," he said.
In the confusion, the true communal colour of the PAC, notorious for its anti-Muslim bias all over Uttar Pradesh, was exhibited with gruesome consequences. Apart from the ghastly outrages committed at Mahana, the PAC was involved in innumerable other communal attacks all over Meerut and in many cases it was an open confrontation between them and Muslims. The communal frenzy of the PAC did not even spare itself - according to one report, a PAC unit turned against its solitary Muslim member and butchered him on the spot in the Shastri Nagar area of Meerut.
As far as political parties and groups are concerned, almost all have reportedly played a dirty role in either directly or indirectly whipping up passions. The Congress (I) which has traditionally controlled the city's politics by winning every single Assembly election from Meerut since Independence, is the most guilty.
The party is divided vertically on communal lines. Mohsina Kidwai, for instance, has been actively encouraging Muslim groups and leaders in Meerut for several years now. Her own assistant, personal secretary Anwar Ahmed, is a close friend of some notorious anti-social Muslim activists in the city and has often pulled strings to get them out of trouble.
The Congress(I) Rajya Sabha member Shanti Tyagi, also from Meerut, and the local MLA Jai Narain Sharma on the other hand, have been active in fomenting Hindu militancy in the city. It is under their protection that the Shiv Sena, the Hindu Sena, Manav Dharam Samaj and other fundamentalist organisations have mushroomed in Meerut in the past few years.
Hindu and Muslim communalists, both claiming Central and State patronage, have not surprisingly gained increased legitimacy amongst their respective communities in recent
months, even though Muslim fanatics have been distributing guns and bombs in various Muslim areas their Hindu counterparts, trishuls and other deadly weapons in theirs.
This mounting tension had already resulted in a round of communal clashes in April leaving 10 dead and scores injured. These, however, had been largely manipulated by a few hoodlums who took advantage of the Muslim festival Shab-e-Baraat. This time the riots succeeded in sucking large sections of ordinary citizens from both the communities.
Loudspeakers in mosques (which number around 120 in Meerut) were used for communal propaganda, which thereby reached all classes of Muslims immediately. The sloganeering during the first few days of the riots was nerve chilling. Dr PP Sharma and Dr K Gopal, owners of city clinics, have tape recorded what they claim are excerpts of some of the slogans issued from mosques: "Mar kar liya Pakistan, hams kar lenge Hindustan", "Hindustan mein rehna hoga ta Allah a Akbar kahna hoga" (a reversal of the Hindu fanatics' slogan "Hindustan mein rahna haga ta Vande Matram kahna haga") "Police dogs, you have got a hashing in Punjab and now you are trying to show your teeth to us", "Muslims, Islam is in danger, if you die for religion you will go to heaven, if you win, you would have reached Mecca" and "Let rivers of blood flow" are just some of the less obscene slogans.
Muslims on the other hand claim they have tape recorded filthy communal abuses hurled at them by PAC jawans and local Hindus threatening them with among other things, complete genocide. They alleged that a common rallying call of the PAC on the rampage in Meerut was, "Orders from Veer Bahadur Singh are to kill every Muslim in sight."
What frightens an outsider is the fact that the ugly communal mood in Meerut continues even after the worst of the rioting is over today. Groups of Hindus and Muslims huddle in separate groups cursing each other's community. In one shocking instance, a reporter of a Delhi-based weekly newspaper was told to shut up by agitated Hindu local journalists and officials when he tried to ask the District Magistrate at a press conference what the death toll at Mahana was. "Why are you people so bothered when Muslims are killed, why don't..."
you ask about the death toll of Hindus?" they shouted.
In this communal madness, the voices of sanity have been increasingly stifled. One of the most tragic instances of religious fanaticism during the Meerut riots was the murder by Muslim mobs of Dr Prabhat Kumar, son of Dr Harpal Singh, a highly respected local Communist leader who saved the lives of countless Muslims during the 1982 Meerut riots, the professor, who with his wife, had helped noted human rights activist Asghar Ali Engineer compile an analytical study of the 1982 riots, said sadly, "My personal loss is nothing compared to the loss of Meerut and the nation; thanks to the mixing of religion and politics."
Another progressive poet Basheer Badr, whose house was burnt by Hindu mobs, lamented, "The communal polarisation is so complete in Meerut that it is difficult to sustain your secularism. You have to hide in the lap of one communalism or the other to live in this city."
Meerut has been a hotbed of communal tension for the past many decades. There is little doubt however, that since the Rajiv Gandhi government decided to blow hot and cold over the communal question, there has been a qualitative change for the worse.
The backing given by Rajiv Gandhi to Muslim communalists through the Muslim Women's Bill has to a large extent led to the respectability of Islamic fundamentalism not only in Meerut but elsewhere in Uttar Pradesh too. Fundamentalist preachers, intellectuals and politicians have come to New Delhi and been greeted with great fanfare by the prime minister. They have had their photographs taken with him and gone back to parade their increased clout to their community.
Rajiv Gandhi's ad hominem in Punjab is also one of the reasons behind the rising tide of communalism. Many Muslim fanatics have sought to use the example of Sikh militancy and what they see as New Delhi's capitulation in the face of it, as a model for their own community. In Meerut and other places, the most frequent slogans being used by Muslim fundamentalists is, "If the Sikhs can hold the government to ransom, why can't the Muslims, with their larger numbers do the same?"
On the other hand, Hindu fundamentalists have been cleverly playing on the same Punjab theme to whip their community into a communal frenzy. Masquerading under the guise of nationalism, Hindu militancy has been taking cruder forms. The mushrooming of organisations like the Shiv Sena all over the Northern India and the increasing use of the trishul as a symbol of 'Hindu national unity' is a proof of this.
Significantly, the Congress (I) which has been encouraging Muslim fundamentalism on the plea that this would promote communal harmony, is at the same time egging on this newly spread Hindu chauvinism as well. So much so, that the RSS, which earlier had a virtual monopoly over such communal propaganda, is being sidelined by local Congress (I) bosses!
Thus, whatever the immediate catalyst to the communal madness at Meerut this month, there is little doubt that its roots lie in the basic policy of the Government in appeasing communalism of whatever hue.
Said Ved Aggarwal, a veteran freedom fighter in Meerut, "Rajiv Gandhi is trying to walk a tightrope by appeasing both the communities at the same time. It is a dangerous game. Once people see through this game they will stop cutting each other's throats and turn against the government."
Excerpts from Sunday Observer
31st May, 1987
INEPT LOCAL GOVT. CAUSED THE RIOTS
Askari H. Zaidi
Talking to the newspapers on May 22, Mr. Buta Singh, Union Minister for Home, admitted that the local administration was not fully geared to deal with a situation of this dimension.
A candid confession indeed. But Mr. Singh did not say why the administration was not geared even when the city was communally charged and there was every possibility of a major communal flare-up.
One day later in Lucknow, the principal home secretary of the U.P. government, Mr. Mata Prasad asserted that the intelligence network had not failed. This means the administration had been alerted. In fact, there was no need of intelligence report to make an assessment of the situation. That isolated incidents of communal nature had been taking place in the city was common knowledge.
A senior civil servant, who requested anonymity, said the local activists of the Babri Masjid action committee and the Vishwa Hindu Parishad had been stoking religious passions for quite some time.
However, there was hardly any effort on the part of the administration to prevent the communal polarisation. The local leaders also chose to exploit the situation to further their political interests, instead of helping the authorities in containing the increasing communalism.
The result was that on the day of Shahe Barat on April 14, riots broke out over the playing of loudspeakers. The disturbances took the toll of over 10 lives, besides destruction of property worth lakhs of rupees. This should have been alarming enough for the administration to take stern measures to curb the growing communalism and activities of the mischievous elements.
However, it did the opposite, the curfew was lifted on April 27. The charged atmosphere warranted police presence and strict enforcement of section 144 for at least the statutory three months. But the authorities thought otherwise. Not only this, most of the 800 "miscreants" who were arrested during the riots, were also released on bail.
"Meerut has been a communal trouble spot. The district administration should have considered this before releasing them," said Mr Ajit Singh Sethi, the Congress MLA from the Meerut cantonment constituency.
Giving details of the sequence of the events which later developed into one of the worst riots this country has ever witnessed, the district magistrate, Mr R S Kaushik, said on May 20 that the killing of a resident of Mohalla Kaichiyan, Mr Ajay Sharma, over tenancy dispute was interpreted in communal terms as the FIR lodged by the brother of the deceased charged six persons of the minority community. Four of them were arrested within two days.
The next day passed off without any trouble. However, on May 18, Mr Kaushik said, around 8 p.m. the explosion of a country bomb in Chipiwada sparked off a clash between two groups belonging to two different communities. About the same time, a "khokha" was set on fire in Surajkund area, about two km from Chipiwada. "Police contingents were rushed to diffuse the situation at both the places," Mr Kaushik said.
The incident should have alerted the authorities to increase the vigil in the sensitive areas. "Police pickets should have been posted at all the trouble spots as during the month of 'Ramzan' when Muslims gather in large numbers in mosques for collective 'namaaz' and 'tarahi' (collective recitation of the Quran), a resident of Mausampura, Mr Mohammad Azhar said.
However, the police made the mistake of going to make
some preventive arrests from the Mohalla Imliyan at the time of 'Sahari' about 3 o'clock in the morning of May 19 Mr Kaushik told me that the arrests were made between 2:30 am and 5:00 am at 'Sahari' when Muslims have their meals before starting the next day's fast. "The coming of the police at the time of 'Sahari' was enough provocation for the residents to come out of their houses to defend the attack on their faith", a resident of Imliyan said at the Medical College.
Moreover, "due to our bad experience while under police custody during the April riots, we all decided to prevent the police from making the arrest," he added.
Lanes were blocked with drums etc. At some places, the residents lied down on the road to prevent police entry. Mr Kaushik admitted at the press briefing on May 20 that at least one child was crushed to death by a police vehicle.
The Imliyan resident said, that rumours were heard that Hindus had surrounded the area. This led to announcements from the loudspeaker fitted at a mosque, calling upon the residents to come out to fight in the name of religion.
A small police contingent which had come to make the arrests, soon took to heels as the crowd kept swelling. Then started the large scale destruction and rioting which soon engulfed a large part of the city.
This historic city had never witnessed violence of this magnitude. Besides the official death roll of more than 100, over 600 houses and commercial establishments have been destroyed in the riots. The judicial inquiry ordered into the Malyana "carnage" will establish the chronology of the events and the circumstances leading to, among other things, the recovery of the dead bodies from the Hindon and Ganga canal.
But the larger question is, can the Meerut riots be explained away merely as a "total failure of administrative machinery"? Doing so would amount to skirting the basic issues. Because, besides local factors like business rivalry in the handloom trade and lapses of the administration, there are many other issues which emerge from the Meerut experience which are pointers to an extremely grim future scenario.
A major part of Uttar Pradesh has witnessed tremendous communalisation ever since the issue of the Babri Masjid—Ram
janambhoomi surfaced. Both the minority and the majority communities have used the issue to whip up passions. The recent rally in Delhi was indicative of the growing aggressiveness of the minorities.
The Bahri Masjid Action Committee and the Vishwa Hindu Parishad have not lost a single opportunity to stoke the fire. Attempts were made during the Meerut riots to use mosques in the same way as Gurudwaras are used in Punjab to spread political messages. Most ominously, following the unlocking of the disputed shrine by the order of a judicial magistrate, the propaganda of the leaders of the minority community that the judiciary too had lost all credibility has found increasing adherents.
With both communities emotionally charged, the present month of Ramzan has been the most tense in recent history. Thus, when violence erupted in Meerut and later spilled over to other parts of western UP and Delhi, observers were not exactly surprised because tension had been prevailing in this part of the country for quite some time.
The significance of the Meerut riots lies in the changing nature of the communal violence in the country. As was witnessed during the November 1984 riots in Delhi and elsewhere, several persons were burnt alive in Meerut. Houses were set on fire with mobs waiting outside to shoot down those who tried to run out of the house. Inquiries at the hospitals in Meerut revealed that most victims had suffered gun shot injuries. Such brutality is a recent trend.
Another new and ominous development was the active involvement of middle class youth in acts of violence and arson. In fact, the riots would not have spread to middle class localities like Shastri Nagar, had the educated middle class not participated in actual violence. This reporter is a witness to many such incidents in State Bank colony, Chhupi Tank and Shastri Nagar. Houses of a particular community were selected in these locality for attacks. The famous Urdu poet, Mr Bashir Badr, is one such victim whose house in Shastri Nagar was burnt to ashes. Fortunately, he was away in Delhi, and his family members were given shelter in the house of a Hindu neighbour.
These riots were different also in that they spilled over to the
adjoining rural areas. Except for the Malyana incident, in which the PAC apparently went berserk and killed many, all other incidents reported from the villages like Abdulpur, Jataoli and Kaseru Khera were of a minor nature. Yet they were the grim reminders of the partition riots which had affected the rural areas as much as they did the cities.
*Times of India*, 26 & 29 May, 1987
'POLITICIANS POURED OIL ON FIRE'
Riot Analysis
The mayhem of Meerut caused by a weak-kneed administration lacking proper perspective and comprehension of a dangerous situation has been compounded by a panicky political leadership characterised by its immaturity and inexperience.
Meerut, is a ghost city today and it would take a long time for its people to recover from the trauma and at last a decade for its shattered economy to recover from the colossal loss it had suffered.
As and when peace is restored, it would remain fragile because of the unprecedented loss of human lives and property in communal riots which assumed dimensions of a communal war—overtaking even rural areas where communal harmony was not eroded before inspite of periodical communal eruptions in the town.
The enormity of the bloodcurdling communal violence has been attempted to be matched by a deployment of military, paramilitary forces and the police on a scale not witnessed ever before in any other riot torn town. However, the situation could not be described as completely under control. Hardly a day passes which is not marked by case of arson, murders, looting or firing.
Senior officers of the Government, serving and retired and experienced politicians who have occupied positions of responsibility just feel flabbergasted by the caravan of the Chief Minister, Minister of State for Home and more ministers and their cronies along with the Director General of Police, camping in Meerut.
In between, Central Ministers including Mr Buta Singh, Mrs Mohsna Kidwai, MP from Meerut and Mr Chidambaram visited the town and each of them spent two days.
Never before in the history of communal riots in Uttar Pradesh were the Chief Minister, Home Minister or any other Minister advised to visit any town on the eruption of violence. They had been prevented from going there as their presence is considered a hindrance rather than help by those entrusted with responsibility of controlling riots and restoring peace. "They have no business to be there. They do politics and not administration when they are there. They divide the attention of officers who are in the thick of battle," says Mr S C Dixit, a former Director General of Police and now president of the Vishwa Hindu Parishad (eastern region).
More than a dozen administrative and police officers who had the experience of controlling communal riots expressed their bewilderment over the continued presence of political leaders. They blamed them for the escalating trouble. "The job of political leaders is to go and do the healing after we have crushed the riots ruthlessly and restored normalcy," most of them said on condition of anonymity.
It is the politically motivated refusal to face facts which has marred Meerut town and threatens to trigger off a bigger conflagration by stoking the powder keg, which Uttar Pradesh has become after the deep communal divide over the Ayodhya shrine.
Communal violence flared up in Meerut on the night of April 14 over competition of fireworks on the occasion of Shub-e-barat.
A row started over the use of a loudspeaker by a Hindu family but violence erupted when a sub inspector hit by a cracker fired three shots. Twelve persons were killed and 22 others injured and the curfew which was imposed the same night was completely lifted after 12 days on April 27. Neither was any post mortem held nor the police or administration set up reorganised. All those arrested in connection with violence were released and the incident was forgotten as usual by the powers that be, after normalcy was restored.
The Delhi rally on May 16 became the preoccupation of the
powers that be and it was decided that about 5 lakh people from Meerut alone would participate. The administrative machinery was geared up to fulfill the commitment of the Chief Minister to his party bosses. There was no time to give the healing touch to those who had fought a pitched battle barely fortnight back.
The intelligence set up kept on informing the administration of incidents, and the tension building up in some localities, which have been the scene of savage communal flare up of last month. However, the local administration preoccupied with the Congress (I) rally on the Boat Club lawns preferred to ignore the warning and the government in Lucknow was on a holiday in Delhi.
Two bombs were exploded by Muslims resulting in injury to one person on May 15. In the same locality a young man was assaulted on May 16 in Tirthraj locality of Kotwali police station when he was returning from the house of his brother Mr J P Sharma in Chhupiola. He was removed to hospital where he died around 1:15 am. This resulted in serious tension in three localities. His body was taken in a procession to Surajkund for cremation the next day—May 17. Most of the mourners were young men and according to intelligence reports "anger was visible and some of them even talking of revenge." The tension further mounted.
The administration took both incidents lightly. The bomb explosion were dismissed as bursting of crackers and the killing of Mr Sharma was linked to a dispute between his brother and his landlord over tenancy.
The only step taken was to impose prohibitory orders under Section 144, a ritual repeated off and on, without increasing the deployment of force. In the wisdom of the top brass, police was reduced to seven companies only where as it should have been around 15 companies during the month of Ramzan and soon after the lifting of curfew last month. Similarly, no preventive arrests were made.
The result was that next day (May 18) two crackers were exploded around 8:30 pm in Dhohiwada locality. A Muslim khokha was set on fire near Saraswati Mandir of Surajkund by hurling a burning tyre around 10:15 pm. Muslims who had gathered for their night prayers in a nearby mosque came out
and indulged in arson, setting fire to two Hindu khokas (kiosks), brickbattling, hurling molotov cocktails, and kerosene soaked cloth missiles. A boy who was injured died around 1.30 a.m. in the hospital. The police reached the scene of occurrence a bit late and controlled the situation.
The administration ordered round up of criminals, anti-social elements and those who were involved in the April rioting late in the night. Around 2.30 a.m. policemen entered the Imliyan locality for searching and effecting arrests. To this there was strong resistance. Muslims pelted brickbats and fired from rooftops forcing the police force to beat a hasty retreat after firing a few shots.
Loudspeakers fitted on a mosque exhorted Muslims to come out on roads as Islam was in danger. It was Shahri time (taking meals before the sun rise for fasting during the day).
Armed Muslim mobs came out on the roads as if ready for the signal and started an orgy of arson, loot, brickbattling, stabbing and firing. Similar calls of "Har Har Mahadev" were given from a few temples but the Hindus were not prepared for the type of resistance they encountered.
The local administration collapsed and by the time reinforcement was rushed and army staged a flag march, Hindus had suffered a terrible loss of life and property. Mr. Chidambaram and Mrs. Mohisina Kidwai and the Union Home Secretary reached Meerut but Lucknow had only a sketchy picture of "one killed and five injured". When the Chief minister and Home Minister left for Meerut late in the afternoon, the Home Secretary was already informing pressmen that eight persons were killed, 22 injured and 200 houses and shops were destroyed.
Since then violence has been spreading and death rapidly rising. Dead bodies removed by the police to keep casualty figures down are yet to be recovered and accounted for, after discovery of 17 bodies in the Muradnagar canal.
Violence continued unabated infiltrating into rural areas because now it was the turn of the majority community with its greater striking power to settle its scores.
Indian Express, May 27, 1987
OLD DELHI TOO IN THE GRIP OF COMMUNAL FRENZY
Asghar Ali Engineer
In big cities, old walled areas are usually more riot prone for variety of reasons. It is true of Delhi, Hyderabad Baroda and several other places. Firstly, these areas are very thickly populated with a maze of narrow winding lanes and bye-lanes. It becomes easier for mischief makers to operate in these narrow lanes. Secondly, low income group people with high level of unemployment, low level of education, self employed petty traders with closely knit communal relations live in these areas. Mostly there are separate communal enclaves, with less scope for wide range of inter communal relationship (however, recent experiences show that now even areas with much wider scope for inter communal relationship are becoming riot-prone in a number of places). These conditions are comparatively more congenial for breeding communal violence. Thirdly, such highly populated areas with narrow lanes and bye lanes prove secure havens for anti-social elements to operate which are always on the increase in fast expanding urban areas. Crime is integral part of urbanization specially in capitalist societies and crime tends to become integral part of politics, both secular and communal, specially the latter as it also helps legitimise crime as well as criminals. All this explains to some extent why walled parts
of certain cities easily become riot prone, given certain other conditions.
Walled part of old Delhi, especially around Jama Masjid falls into the above category. Some parts of this area have separate communal enclaves whereas some have mixed population. Most of the houses and shops are located in very narrow lanes and bye lanes with high density of population. Of late there have been communal clashes in these areas and thus this part of Delhi has its own history of communal riots. Also, communal tension in Delhi, especially in the walled area, has been on the increase since the formation of religious senas like the Shiv Sena, Bajrang Sena, Adam Sena etc. The Ram Janambhoomi/Babri Masjid controversy has added its own bitterness between the two principal communities. The Sita Ram Bazar in this area is a hotbed of Hindu communalists and trishuldhari Shivsainiks whereas the Shahi Imam holds sway in the Jama Masjid area with his son's Adam Sena though it is not as coherent and strong as the Shiv Sena. However, it makes its own contribution to communal tension.
The current riots in Delhi started on 22 May 1987 in the aftermath of riots in nearby Meerut. All sorts of rumours were afloat about what happened in that unfortunate town. These rumours acquired added significance as large number of Muslims from this area have close interaction with the Muslims of Meerut both on business as well as family level. Hindus too were seething with anger as they were fed on rumours that a large number of Hindus have been killed in Meerut by Muslims. The Delhi riots must be seen in this backdrop.
II
In such a situation any event, small or big, was good enough to trigger off the events. Two days before the riot started there was a rumour that a Muslim molested a Hindu girl in Ballimaran. It is difficult to verify this but rumours have their own explosive role in communally tense situation. Main incidents took place on 22nd May, the last Friday of Ramadan, the Muslim month of fasting and prayer. However, some incidents preceded the incidents on Friday.
There is a small mosque near the Hauz Qazi Police Thana.
This mosque was attacked on 19 May at night. The Imam of the mosque was killed by a mob of about 20/25 persons. The nephew of the Imam, however, escaped and lodged a complaint with the Police but it is alleged, the Police took no notice of the incident. The Imam had died on the way to hospital. There was another incident in a place called Charkhewalan on the same night. Fifteen to twenty people attacked a mosque when the late night prayer was going on. The mob tried to break open the door. The Muslims tried to escape by jumping off the other wall. They ran towards Ballimaran and heard a firing shot. A boy called Ashok was killed when the police fired to control the situation.
The Muslims set fire to a number of paper godowns belonging to the Hindus in Baradari Sher Afghan. One press was also set alight by some Muslim miscreants. In these disturbances, places like Sita Ram Bazar, Zeenat Mahal, Lai Kuan which is a Harijan colony, were mainly affected. Like Meerut here too the Harijans were used by the upper caste Hindus against the Muslims. Here too many Muslims alleged that the upper caste Hindus give liquor and money to do the dirty job. They do not come directly into the picture. In many areas of Old Delhi Muslims and Harijans live close by as both the communities live in poverty and are faced with high levels of unemployment.
III
The main incidents, as pointed out above, took place on 22 May, the last Friday of the sacred month of Ramzan. It all started after the Muslims came out of the Jama Masjid after saying their afternoon prayer (Namaze Zuhr). It is alleged that the Sahi Imam delivered a provocative sermon and provoked Muslims came out of the mosque and began to stone Hindu areas. This was unfortunately uncritically accepted by various newspaper reporters. The police also dished out a highly biased version of the events. However, close investigation shows that this does not seem to be the case. In any case it was not as simple as that. The Shiv Sena and its followers do not seem to be that innocent as they make it out to be.
Many Muslims who were present in the Jama Masjid on
that fateful day told me that far from delivering a provocative sermon the Imam had implored the Muslims to remain peaceful and go home with their eyes downcast, if they are true Muslims. This is born out by sources who are by no means sympathetic to him. The Muslims allege that when they came out of the mosque at about 2.30 p.m. they were surrounded by the Hindu mob from two sides near Hauz Qazi and were stoned. Many were injured in this incident. Similarly there was incident in the Chowri Bazar area from where too Muslims were passing after prayer. In Chitli Qabar area some Muslims tried to loot a Hindu shop but some other Muslims foiled the attempt. Nevertheless some damage was done. The other four Hindu shops in the area remained safe.
In the same area police resorted to lathi charge but some Muslim boys snatched lathis from two police constables and beat them up. The policemen ran away but one constable remained behind and fired in the reverse direction. Two boys were injured and all the shops in the area downed their shutters. Similarly in Kuch Chalan some Muslims tried to break open the Hindu shops which had been closed down earlier but they could not break them. The Rajastban Police fired (there were 10/12 of them) on round and then threw tear gas shell. Then again firing began at about 3.30 p.m. and continued upto 7.00 p.m. The firing was continuous and many fires were shot at those looking from the windows but here (i.e. in Kucha Chalan) no one was fortunately injured. Again from 7 p.m. to midnight there was intermittent firing. Many Muslims were arrested while going for their evening prayer.
In Gali Mochiyan the police shot dead a Muslim boy on 22 May after afternoon prayer. The name of the boy killed was Zahid Ali. He was around 16 years old. In fact, Zahid Ali was not even peeping out of window. He was inside the room behind the cooler. The bullet struck him after piercing through the cooler. This incident took place around 6 p.m., his elder brother Majid Ali told me.
The worst incidents took place in the Chandni Mahal police station area. All the Muslims told me unanimously that the SHO of this police thana Mr. Ram Kumar Sharma is highly prejudiced against Muslims. He made indiscriminate arrests of
250 Muslims from his area, most of whom were quite innocent and had nothing to do with rioting. He also got almost all the gates which were installed in 1947 for protection of mohallahs removed without any reason or rhyme. No explanation was given. Mr Mohammad Ahid, the president Yuva Janta, Chandni Chowk District, also told me that Ram Kumar Sharma is highly anti Muslim and he makes no bones about it. In the Chandni Mahal Police Station area there are many Hindu shops but not a single one was damaged and even then Mr Sharma harassed Muslims. He has ordered firing thrice in last fifteen days, he said. Even Hindu-Muslim relations are quite cordial in this area. There was no tension of any kind in this locality. It was nothing but anti Muslim action by the police at the instance of the SHO Mr Sharma, he alleged. Not a single Hindu has suffered in this locality and yet maximum number of Muslims have been arrested from this area. He is not being transferred under pressure from some communal Hindus, Mohammad Ahid maintained.
It was in Chandni Mabal area again that one Rashid Ahmad, aged about 21, was killed by the police bullet. He was on the roof of the house when he was hit at about 3 p.m. on 22 May. Rashid was removing his shirt when he was hit. The shot was fired with an intention to kill. His brother Rafiq told me that for last several days there was so much terror among Muslims that such terror was not witnessed even during 1947 riots and all this despite the fact that there is no Hindu-Muslim tension in the area. It was all doing of the SHO Mr Sharma.
One old Urdu journalist Raziullah told me that Hindus in fact helped us a lot in this vicinity. They are like our near and dear ones. We have no complaints against them. It was all due to Mr Sharma. Rounds (1700) were fired during a couple of days, he alleged. Even on 19 May 30 rounds were fired from the corner of Chandni Mahal to Sui wakan. In the Chandni Mahal police station area 4 young boys were killed on 22 May disturbances, thanks to Mr Sharma, Raziullah said. The names of the boys killed are Muhammad Said, Rashid Ahmed, Qamarul Hasan and Yasir Allah Rakha. On the same day one boy Mohammad Shafiquddin, aged 23, was killed in Kucha Anarkhan. It is also in the Chandni Mahal police station.
jurisdiction
Thus there is overwhelming evidence of anti-Muslim bias against the SHO, Mr Sharma and yet he remains unharmed. Repeated complaints and Urdu Paper reports did not draw the notice of the authorities. This has made Muslims very bitter indeed. Those arrested were not released until the day of our inquiry, i.e. 7th June. They were subsequently released on 12th June under pressure from Shahi Imam who kept the Jama Masjid closed from 28th May onwards in protest and opened it only after assurance to meet his demands. We will talk about the closure of the mosque a little later.
On the day of the disturbances, i.e. 22 May extensive damage was done to Sima Lodge in Churiwalan where mostly those coming from Pakistan stay. It was special target of wrath. Many Pakistani inmates were injured and were evacuated to other places nearby. Each and every room in the lodge was damaged. However, the Muslims in this area retaliated by burning Hindu shops. More than 13 shops were burnt, among them were the shops of toodgrams, pan cigarette, gram roaster and milk vendor. These shops were also badly damaged.
Kucha Pandit is highly congested, narrow street and has mixed population of Hindus and Muslims. It was in this lane that firing took place for several hours and people were struck with terror. It is really miraculous that despite such prolonged firing no one was injured or killed. It was being talked among some sections of Muslims that "God sent His angels to intervene and take the bullets on their bodies." This folk-talk about miracle is indicative of the kind of firing that took place there that day. A few shops were looted and damaged in this lane. Shamshad Ali, a fruit vendor and few other shops were damaged. He sustained a loss of around Rs. 10,000. One Hindu shop was also set ablaze in this area. About two hundred thirty-eight Muslims were taken into custody under section 238 from Kucha Pandit, some of them have been charged under 15 different sections and were not released until 11 June when I was investigating.
In Turkman Gate area too Hindu and Muslims fought against each other. There was mutual stoning. A Hindu shop selling silver ornaments was burnt in Sita Ram Bazar area by
some Muslim miscreants. Police resorted to firing. One Muslim boy was standing on the terrace of fire house. He was warned by the police to go into the house but he did not move and was shot dead. These incidents took place on 22 May evening.
One similar incident took place near Gali Kudla Kashan which is adjacent to Katra Ghulam Mohammad Gali Kundla Kashan is mostly inhabited by the Hindus. One Mr Arif alleged that there was incessant stoning and pelting of other missiles for more than two hours from this Gali on the afternoon of 27 May. There is big oven for roasting vermicelli (siwaiyao) in Katra Ghulam Mohammad. Many Hindus from Eastern UP work here. All of them fled when stoning started. When they were running away a canthoo of ghee (clarified butter) fell down into the oven and it caught fire. There is the house of one Mr Qamarul Hasan adjacent to this oven. It caught fire too and was filled with smoke. Qamarul Hasan, in order to save his family used a window overlooking a mosque and brought out through the widow several members of his family. He was the last to come out. While he was descending from the window an armed police on the terrace of a house opposite to his house took aim at him and shot him dead. He was quite innocent had nothing to do with the disturbances around and was the sole bread earner in the family. His tragic death left shattering impact on the Muslims around and mourned by all. He was just 21 year old. It was alleged that a Shiv Sena leader from the area was actively associated with the police and was instrumental in getting many Muslims arrested. About 18 Muslims were taken into custody from this area.
III
In these disturbances in Delhi on 22 May in the wake of Meerut riots some 15 persons were killed, twelve of whom died by police bullets and were mostly young. The remaining died due to stabbing, three or four of whom were Hindus. One Hindu boy was killed in the police firing.
There was an atmosphere of terror among the Muslims when I was investigating around the Jama Masjid area. These widespread disturbances, no doubt, have left shattering impact.
on the Muslim mind. Most of those killed in police firing were quite innocent. The arrests were also one-sided and quite indiscriminate. Hundreds were in the police custody until 11th June, many of whom were tortured. The Muslims in general have come to believe that law and order machinery is not protecting them but for killing them. They have no faith in it any longer. It leads them to defend themselves and they tend to collect some crude weapons towards this end. This becomes evidence of 'preparation for rioting' in police searches and leads to large scale arrests. It has really become a vicious circle for them. They do not know how to break out of it.
A group of social activists led by Shri Purushottam Agarwal of Delhi University was also carrying on investigations in the area. They fully corroborated what many Muslims told me. Purshottam Agarwal described to me the sufferings of the Muslims of Katra Ghulam Mohammad who had been arrested in large numbers. It was impossible for them to get bail. The whole police machinery is prejudiced. The political authorities are equally indifferent. Their complaints fall on deaf ears.
The Shahi Imam closed down the mosque on 28 May by way of protest. In the face of these police excesses the measure proved tremendously popular among the Muslims in the disturbed areas of old Delhi. Most of the Muslims in the area applauded it. A large banner outside the mosque proclaimed in bold letters:
"Protest against extreme atrocities & barbarism Jama Masjid, Delhi shall remain closed till guilty police officials are severely punished, all innocents arrested are immediately released. Not verbal but practical assurance for the security of life and property for future is made May 28, 1987."
The measure of course proved highly controversial in different circles. Some Muslims also criticised it and opposed conversion of place of worship into a political arena. Some fatwas (religious opinion) from the theological seminary of Deoband were obtained against the closure. There was heated debate among the Muslims even about these fatwas. The affected Muslims of course despised those "sarkari maulvis" who
issued these fatwas. The opinion in favour of the closure was so strong that those who issued the fatwas had to retreat and explain why they issued the fatwa.
The closure of the mosque proved highly embarrassing for the Government. Its image was getting tarnished in the Muslim countries. Ultimately, it accepted these demands and released those arrested on 13th June. The mosque was opened after getting the assurance for the release of the Muslims. Here no one would dispute the fact that places of worship should not be misused for political ends and hence in that sense the closure of the mosque was not justified. But the important question is why the Government turned deaf ear to the Muslim demand to release the innocent. If the Government had taken timely action there would have been no need to close the mosque and the Imam would not have acquired the popularity he did.
The politics of communalism is already very deeply embedded. It is no longer peripheral to our polity, in fact it is tending to become an integral part of our mainstream politics, thanks to opportunistic manipulations of communal sentiments for mobilising votes by secular parties as well. It is for this reason that Bahrí Mosque and Ramjanamhoomi has acquired such tremendous significance, something which normally should have been treated as non-issue. I would venture to say that these are wages of sins of secular parties. Had they honestly fought for the real issues facing the people and worked for their solution our polity would not have been communalised to such an extent. Though it is quite late in the day, still it is not too late not to learn any lesson.
DELHI WALLED CITY RIOTS
I Introduction
The immediate trigger for the riots was a series of innocuous local incidents, often blown out of all proportion by fast travelling rumours. These initial sparks carried the potentialities of a conflagration since they were taking place in an atmosphere which had already been surcharged by communal dispute over the Babri Masjid—Ram Janambhoomi controversy, and widespread communal riots in Ahmedabad, Allahabad, Baroach and other places. The outbreak of riots in Meerut a day earlier (on May 18), also contributed to the fuelling of tensions in the walled city. As a result, while in normal circumstances, a local dispute would have been settled amicably and faded into oblivion, in the present situation it always threatens to assume lasting communal dimensions. Our investigations into the incidents which sparked off the riots on the first day—May 19—confirm the belief that several extraneous factors at the national level like the Babri Masjid dispute, the Shah Bano controversy and the Ahmedabad riots, have sharpened the alienation between the two communities even in those localities of the walled city, where there have never been communal riots on this scale since 1947. Thus, the soil was already fertile enough to breed the riots.
Our findings also indicated other political and socio-economic factors which could have interrupted the traditional harmonious relations between Hindus and Muslims in these
areas. The rise of a younger, articulate generation of Muslims, who voice their grievances against discrimination in educational facilities and jobs, some of whom have made good by competing with the Hindus in modern business enterprises, is often perceived as a threat by Hindus. The traditional Hindu mercantile community which dominates in the walled city is found to be resenting the Muslim intrusion into its commercial enclave, and the Muslim assumption of symbols of upward mobility. Hindus tend to raise their eyebrows at the assertion of an equal status by a community which they have been used to look down upon as their inferiors in the post-Independence era. The Muslims, quite rightly, are offended by this.
In the absence of conscious efforts by secular political forces in the locality to reconcile these growing differences and integrate the two communities, members of both the communities in their political choices are tending to fall back on their respective religious fundamentalist platforms. Hindu communal organizations like the Vishva Hindu Parishad and their Muslim counterparts like the Adam Sena have been successful to a large extent in rallying their respective followers through the technique of demagogic intimidation, and their ability to ascribe nefarious motives to the other community. We found elderly Muslim politicians and intellectuals who had been traditionally Congress men, pushed to the wall by the indifference of the ruling party to the genuine grievances of the Muslim population. Similarly, we found elderly Hindu Congressmen, including freedom fighters, who had become disgusted with and alienated from the ruling party because of the growing incorporation into the party of anti-social elements and Hindu communalists. Their retirement from the scene left the walled city an ideal breeding ground for communal and religious fanatics as well as the local underworld—the touts, drug peddlers, hired killers who swoop down upon any chaos to settle old scores. In such a situation, the same and secular elements in both the communities appeared to be rendered immobilized by their inner frustrations as well as by the overwhelming communal frenzy that prevailed outside.
We feel that the mutual distrust and hostility between the Hindu and Muslim residents of the walled city, built up all these
days by the changes in the socio economic scene in the localities as well as by the estrangement reinforced by the happenings outside, played a large part in the eruption of communal riots that enveloped the walled city during the period (May 19-24) under our review. At the same time, we cannot ignore the role of the administration, including the police, in augmenting the communal tensions by their behaviour during the riots—a behavioral pattern which was marked by a deliberate harassment, persecution and provocation of the Muslim minority.
II The Background and the Sequence of Events
The May 1987 communal flare-up in the walled city occurred within a span of little over a year, after a similar outbreak in the same area, which indicates the continuity of communal tensions for quite some time in the walled city.
On February 14, 1986, which was a Friday, Muslims at the call of their religious leaders observed the day as a day of protest against the court judgement on the Babri Mosque Ram Janam Bhoomi dispute. After the prayers at the Jama Masjid, the devotees were returning to their respective homes, when a dispute arose over the display of a Hindu banner on the top of a temple at Lal Kuan which congratulated the Hindus on the liberation of Ram Janambhoomi. As the Muslims shouted slogans against it, the police overtook them and chased them into Gali Qasim Jan. An SHO, Jag Pravesh Kaushal was alleged to have shot down two Muslim youths—Mohammed Zakir and Subhan Ullab—who were actually trying to persuade the people to go back to their homes to prevent a communal confrontation. At some spots, members of both the communities fought each other and indulged in arson and looting. The indiscriminate firing by the police leading to the death of the above-mentioned youths, combined with the authorities refusal to punish the guilty police officer, had left a bitter trail in the minds of the residents. (See Delhi PUCL report on the incidents, in PUCL Bulletin, April 1986)
Tensions again ran high in the area in September, 1986, when curfew had to be imposed in Chandni Mahal, Jama Masjid and Hauz Qazi, following resentment among the Muslims against the reported entry of some Suoday picketers, with their shoes
on, into the Mosque at Ferozeshah Kotla, on September 7. At least one person was killed in the flare up. In September again a Bajrao Dal rally, a tribal armed display of Hindu militancy, was held which turned violent, assaulting people on roads and houses. On the 4th of October, mobs again went on the rampage in the walled city when the Ram Lila procession was not allowed to pass through its traditional route due to communal tensions in that area. At least 50 persons were injured.
In the meantime, the controversy over Babri Masjid—Ram Janambhoomi had been hotting up all over India. A massive gathering of Muslims in the Boat Club lawns was addressed by the Shahi Imam of Jama Masjid, Syed Shahabuddin and other Muslim leaders on March 30, 1987. Shahi Imam was reported to have urged the Muslim Ministers and MP's in a highly emotional speech, to open their eyes and ears to the danger faced by Islam and warned that otherwise their houses would be looted and burnt. Syed Shahabuddin was found trying to restrain the latter.
Fast on the heels of the Muslim rally at Boat Club came the Ram Navami procession on April 7, organised by the Hindu communal groups. What used to be a peaceful religious demonstration in other years was turned this year into an occasion for raising Hindu communal slogans and displaying gestures of militancy containing all the potentialities of a violent outbreak.
As passions ran high on both sides in an atmosphere overloaded with mutual suspicion and distrust, Prime Minister Rajiv Gandhi chose to address a rally of his partymen in the Boat Club lawns on May 16 this year, where he cautioned his followers against efforts of 'destabilization' by foreign powers and their agents in India. Referring to these agents and the US arming of Pakistan, he thundered "Unko hum karake ka jawab denge, aisa jawab denge ki unke jo maa baap bane huay hain unki naani yaad kara denge hain" (We will give them such a fitting reply that those who are their godfathers will remember their grandmother).
(In the course of our investigations and interviews with the residents of the walled city, while the Hindus felt that the Shahi Imam's speech at Boat Club inflamed Muslim passions, the
Muslims said that the Prime Minister's Boat Club speech was interpreted as a war cry against the Muslims by the Hindu miscreants.
In the meantime, in the walled city, religious bigots of both the communities continued to step up their propaganda. A few weeks before the outbreak of the May riots during the Ramzan a three wheeler scooter was observed moving around Nai Sarak area with a loudspeaker blaring forth the following slogans "Ab ki Id Hogi Kaise? Moradabad Jaise" (How will you observe your Id? As it happened in Moradabad, i.e warning the Muslims about attacks on them during the Moradabad riots), "Agar Hindustan me rahna hai, To Vande Matram kahna hai" (if you want to live in India, you will have to chant Vande Mataram), "Jo mangeka Babri, Uska wakt hai aakhri" (Whoever wants Babri (masjid) will find his days numbered).
A large number of leaflets were distributed by the Indraprastha Vishwa Hindu Parishad questioning the patriotic loyalties of the Muslims and urging the establishment of a Hindu Raj.
The May 19 riots came as a climax to these sustained efforts, made by both Hindu and Muslim fundamentalists to stoke communal passions.
Our investigations into the sequence of events on May 19, and later, revealed how rumours, or distorted versions of an incident in a surcharged atmosphere could provoke communal violence.
The first incident occurred in the evening of May 19 when a dispute took place between a cyclist and a motorcycle driver—both Hindus—on Nai Sarak in the walled city. The wife of Vijay Kumar Sharma, the motor cycle driver, complained that the cyclist, Balakrishna, had misbehaved with her, which led to a scuffle between Sharma and Balakrishna. Two passersby travelling in a rickshaw, who happened to be Muslims, intervened to separate the two. Soon the police arrived and took all of them to the Town Hall police station. The woman was reported to have told the police that the Muslim passersby had intervened in the dispute to protect her. Within the police station a reconciliation was brought about. But on Sharma's complaint Bala Krishna was detained. Meanwhile, outside the
rumour had spread that Muslims bad misbehaved with a Hindu woman. Mobs descended on the streets, and soon sporadic clashes between the two communities broke out in different parts of the walled city.
At around 9:30 p.m. in Ballimaran, a Hindu, Om Prakash Kasbyap was shot. Twenty-two-year-old Om Prakash was standing on the third floor of Dharamshala Bissomal when he received bullet wounds in his stomach and arms. He came down running, and according to his relatives, he cried "Ali Muhammad ne mujhe goli mara." Om Prakash lived with his mother and five brothers in a small two-roomed house opposite the Dharamsala. His brother, Naunthal informed the police and took him to JP Hospital where he died at 1:30 a.m. the next morning. Ali Mohammad, the person named by Om Prakash is the son of Fazal Ilahi, the landlord of Om Prakash's house, with whom he had a long-standing dispute over rent. Subsequently, Fazal Ilahi, his four sons including Ali Mohammad and his brother Ahsan were arrested. But according to the local residents, they were released at the intervention of Shahi Imam of Jama Masjid. When we later asked the DCP (Crime & Railways), Mr. Amod Kanth, about the incident, he said that Om Prakash could have been hit by a police bullet, since the police at that time were firing in the air to disperse warring crowds.
Here again, what was presumably to start with a landlord-tenant dispute, took on the ominous colours of a communal discord, the popular mood being what it was.
The next serious incident of the day was, however, of a more direct communal nature. At around 11:45 p.m., Shauqat Mia, a watchmaker and Imam of masjid Hakim Baqi in Hauz Qazi, was stabbed repeatedly by assailants who raided the masjid. While Shauqat died, and two of his associates were injured, one of them escaped and informed the police, who came and arrested five local Hindus.
As intermittent clashes and arson continued in Ballimaran, the police clamped curfew in the area. The curfew was extended to the Turkman Gate area also where till then there had not taken place any serious incident. Residents of the area later in the course of our interviews with them, blamed the police for "such a reckless action which not only created unnecessary
harassment for the inhabitants, but also encouraged to a large extent the riots that broke out there on Friday". It is possible that the curfew was imposed as a precautionary measure in view of the tension already present in other areas. We shall deal later with the Turkman Gate area incidents.
On May 20, which was Wednesday, the army staged a flag march through the walled city.
Both Wednesday and Thursday (May 20 and 21) were comparatively calm, although tension prevailed. The curfew was relaxed for four hours—two in the morning and two in the evening—on Thursday. Continuing arrests of people, many among whom were described innocent by the neighbours, news of deaths of those who were injured on May 19, as well as of the escalation of riots in Meerut, fuelled the tensions that were building up among the residents of the walled city, who were already feeling the pinch of the curfew. Shops had remained closed for the last two days. Whether little foodstuff was available for sale when the curfew was relaxed on Thursday, were sold at nearly twice their usual prices; children in many areas had to go without milk.
Meanwhile, late on Thursday night, at a meeting with the Lt. Governor of Delhi, where political leaders, local leaders of the riot-affected areas, and police officials were present, it was decided to lift the curfew for the whole day on Friday to allow the Muslims to go out to offer prayers at mosques on Jamat-ul-Vida, the last Friday of Ramzan. This appeared to be a crucial decision in view of what happened later on Friday.
Although the Bharatiya Janta Party later claimed that they had opposed the lifting of the curfew at the meeting with the Lt. Governor, our interviews with many among those who were present at the meeting revealed that the BJP representatives never made any such observation. While one observer at the meeting told us that the BJP representatives might have had expressed some "feeble" opposition to the decision to lift the curfew, another observer said that the BJP was in favour of entrusting the police with the decision to maintain, or lift the curfew.
From our talks with those who attended the meeting it was evident that the decision to lift the curfew was to a great extent
influenced by the pressure of the walled city's trading community, which consisted of both Hindus and Muslims, which had been suffering economically because of the disruption in commercial activities due to continuous curfew. The BJP, apparently did not want to lose their traditional support among the trading community (mainly Hindu) by opposing the lifting of the curfew at the time of the meeting with the Lt Governor. But later on, when violence flared up in the streets, the BJP politicians found it convenient to dissociate themselves from what obviously was a unanimous decision at the Lt Governor's meeting. Taking up an 'I-told-you so' pose, the BJP picked up the curfew issue as a handle in its political fight against the ruling party.
A few residents of the walled city with whom we talked later, felt that the authorities should have relaxed the curfew for a few hours on Friday to enable the devotees to go to the mosques and return home, instead of lifting it throughout the day. Some others felt that the curfew should have been lifted in phases, instead of continuing it for days together which brought hardship to the residents.
Friday saw the worst of the riots. From what we could piece together after interviewing a cross-section of residents of Ballimaran, Jama Masjid, Baradari and Turkman Gate areas, it appears that the first wave of violence began after the end of the Jamat ul Vida prayers in the afternoon, and its epicentre was the Hindu dominated Chawrihazar area. As the 'namaazis' (the devotees who offered prayers) were returning from Jama Masjid, trouble broke out in that part of the city. According to one eye witness, Md Zahir, a spinning machine dealer, as the crowd of 'namaazis' neared the Standard Sweets shop in Churiwafan, stones were hurled upon them from its top. In the melee that followed, Seema Lodge, a Muslim hotel, was set on fire. Lathis, spears and bottles were freely used by both sides.
While most of the Muslims we interviewed more or less corroborated the above account, the Hindus had a different version. A resident of a Daft Slum in Gali Magazine in Churiwalan who claimed to be an eye witness, said that it all began when one of the 'namaazis' picked up a stool outside
Baha Dudh Bhandar and banged it, which acted as a trigger. Several shops and restaurants—Rajeev Restaurant, Golden Restaurant, Shivdhan Dudh Bhandar—were attacked, and many were injured.
According to another account, the disturbances started in Hauz Quazi, where the 'narmaazis' became violent when they reached the spot where the Imam of Masjid Baqi was killed on Tuesday. The mob looted and burnt shops while the police stood watching, according to Tarachand Chandelwal of the BJP.
The trouble soon spread to the adjoining lanes and bye-lanes, where shops were looted and burnt and people stabbed and beaten-up. Members of both the communities were involved in the incidents. But in our interviews, while the Muslims sometimes acknowledged the participation of hooligans from their own community, the Hindus invariably stuck to the position of being innocent victims. While the Muslims' main target of criticism was the police, the Hindus were usually all praise for the police. The failure of both the communities to provide an impartial account—even by those who claimed to be eye-witnesses—of the origins of the trouble, indicates to what extent the wall of prejudice and suspicion separates the two communities. Dalits whom we visited in their quarters in Katra Khawaspura under the Chandni Mahal police station, repeatedly asserted that they had never had any contact with their Muslim neighbours who lived cheek by jowl in Kuchachalan. Similarly, Muslims in the Turkman Gate area described to us the Dalits as "chamars" who did not come from a "good caste". Caste-Hindus were often too eager to describe the Muslims in the walled city as Pakistani agents. In such an atmosphere of communal divide, versions of incidents were bound to be influenced by the particular bias of respective communities.
We, therefore, tried to check and cross-check accounts of some of the major incidents that took place during the May 19-25 period, and visited those spots to assess the nature of handling of the problems by the administration and the police, and to examine the impact of police action and curfew.
We began our investigations on May 23, Saturday, when
the injured and the dead were being brought to hospitals from riot-affected areas. We interviewed the doctors who were attending them and the injured who narrated their experiences.
We followed this up by visiting the area of major incidents on Sunday, May 24. One such incident took place at Kucha Nahar Khan, a twisting lane behind the Golcha cinema house at Daryaganj. A predominantly Muslim residential area, it has a couple of Hindu families and one Sikh family. The residents told us that on Friday, following the disturbances after the prayers, the retreating mob entered the area and set fire to two Hindu shops, belonging to Pannalal and Shambunath Zamiruddin Khan, who is a prominent Congress (I) leader of the area, and a special police officer, told us that the miscreants poured petrol from a scooter to set fire. Zamiruddin intervened and saved the Hindu shop owners by removing them from the burning shops. He rang the police, but they arrived half an hour late.
We visited Baradari where two godowns and a Press belonging to Hindu Businessmen were burnt on Tuesday. We earlier interviewed the Hindu residents of Pandit Kucharam who had been stabbed. These included several women. While the victims of stabbing and arson had been mainly Hindus this time, the victims of police firings and high handedness were Muslims.
On Saturday, at about 1:45 p.m., a 34-year-old Muslim, Shafiquddin who was washing his hands in the ground floor covered verandah of his house (2060 Kucha Nahar Khan), was shot dead by the CRPF. We visited the house of the victim, who was the only earning member of a family of one handicapped brother, one unmarried sister and their mother. The relatives and neighbours told us that before Shafiquddin got hit, police appeared to be chasing someone in the streets. Shafiquddin came out in the verandah and went to the water-tap to wash before reading his 'namaz', when the CRPF personnel posted on the roof-top of a school building opposite took aim at him and fired five rounds. One bullet entered his body, and he died. We saw the holes made in the ground by the other bullets, blood stains and a pair of slippers still lying on the floor.
Residents of the area told us that the police had behaved in an extremely high-handed fashion with the Muslims. They accused a police officer of shouting at the Hindu neighbours: "Why didn't you kill a few Muslims?" The police made a house-to-house search, but could not find any firearms. To prevent the police from entering the lane and terrorizing the people, Zamiruddin and the other residents locked from inside the massive iron gate that stood at the entrance of Kucha Nahar Khan at its crossing with Kucha Chalan—the road where police pickets were placed. They reassured the Hindu and Sikh families of protection.
On Sunday, after we had finished our investigation in Kucha Nahar Khan and had come out, within minutes the CRPF and the Delhi armed police arrived at the spot and with welding instruments took down the gate. We saw gun-toting policemen entering the lane and beginning a house-to-house search. They aimed their rifles at us and asked us to leave the area immediately.
On the same day we visited the Turkman gate area. Residents told us that on Friday between 2.30 and 3.30 p.m., a crowd of 'Dalits' came from the direction of Kali Masjid, Bazar Sitaram (a Hindu area), and started looting shops, owned by both Hindus and Muslims. They included a shop owned by one Ram Kumar Vilas, and 'Shahana Tailors'. The residents could recognize among the miscreants, sons of a former Deputy Superintendent of Police, who were also now employed in the police.
Although the local Muslims told us that the nearby Derwale Masjid had been set on fire by the miscreants, we checked and found that it was not burnt. In response to our later inquiries, while some said that it was stoned, others said that attempts were made to set it on fire.
As the looting continued, some of the residents rang up the police. But the police did not arrive for the next three hours. Around 6 p.m., people from the top of Hindu houses next to blocks A, W, X and V, started stoning the Muslims. Soon after that the police arrived, with a crowd of Hindus following them. They fired several rounds, though no one was injured. The residents described to us in graphic terms how
when they were just about to break their day long fast, the police started firing. "Women and children started screaming. We remembered those days of 1975, (when during the Emergency, the police demolished houses in the area). Then they began the arrests. They just barged into homes, beat up men and women and children. The ASI, Mabender Singh of Turkman Gate chowki took the children and shouted at them saying they were 'Pakistanis'. Although most of the children were released on the intervention of a senior police officer, we were told that there were still some children in jail. One of them was 12 year old Mohammad Sakir. 'This is how the future of our children is ruined,' commented one of the residents.
The residents also complained to us that the police indulged in looting. In one case, the police raided the house (No A-13) of Islamuddin, arrested him and his son, and took away a VCR set. Islamuddin is a heart patient, who had just been released from Bansal Hospital.
We met several people who had been beaten up by the police. An old man showed us his back and his hands swollen with marks of rifle butts. We saw women and children who had been victims of police assault. From the FIRs we examined we found that all the 115 persons arrested from the area were Muslims.
When we went to Gali Wazir Beg in the Turkman Gate area, we met residents who had the same story to tell us—the unprovoked attack by police on men, women and children, their indulging in looting, and their blatant communal bias directed against the Muslims. We were told that when on Friday, people gathered to offer prayers at Masjid Pulaowali in the Gali Wazir Beg, the police suddenly entered the mosque, broke two windows and the loudspeaker. They challenged the 'namazis' and said 'Tum me se jo bahadur hai, wo bahar aye, hum shikar karega'” (Anyone from you who is brave enough, come out, we'll play hunting). They then began a rampage. They opened the lock of Rafiq Tea Stall and took away Rs 2500, threw out all the things, drank all the soft drinks and broke the bottles on the road, and carried away about a hundred butter packets.
During our talks with the victims of police atrocities in the Turkman Gate area, a few names kept cropping up repeatedly. The residents named them as the main perpetrators of the violence. Among the police they identified the ASI Mahender Singh of Turkman Gate chowki and two constables Rajinder Singh and Mahesh. They also mentioned the names of Ashok Jain, Metropolitan Councillor of Turkman Gate belonging to Congress (I), Swaroop Singh who runs a dairy in the area, and Yadeshwar of the Vishva Hindu Parishad as those responsible for the arrests of innocent Muslims.
While the police remained tight lipped about our inquiries, Ashok Jain and Swaroop Singh agreed to talk to us. Mr Jain denied being anywhere near Turkman Gate at the time of arrests there. "If arrests were made," he said, "it was wrong because local people were not involved. It was all the work of outsiders." He informed us that on a previous occasion, the name of Sardar Swaroop Singh, a Sikh, had been linked with riots. His son Surender, in particular, was alleged to have been involved in the past. Mr Jain asserted this time they were not involved.
In spite of the past record of Swaroop Singh and his son, they appeared to be important enough in the eyes of the Lt Governor's Secretariat to be invited at the Thursday meeting at his residence where the crucial decision of lifting the curfew was taken. When we met Swaroop Singh, he informed us that his son attended the meeting.
On Monday, the 25th, we went for a visit to Katra Khawaspura, inhabited by Dalits. The previous day during our investigations in Kucha Nahar Khan, we were told by the Muslim residents that the Dalits from Khawaspura (which is adjacent to Kucha Nahar Khan) had pelted stones on the Muslims after the Friday prayers.
The Dalit families of Khawaspura, however, had a different story to tell. They said that on that eventful Friday at around 3:30 p.m., a woman of their community, while standing on the rooftop of her house, was suddenly attacked by a knife-wielding Muslim, who had climbed onto the Dalit house top from the adjoining Muslim school which is in Kucha Chalan. The woman came running down and alerted the rest of the colony (a wide area with a courtyard surrounded by double storied
houses) The menfolk were then out on work. The Dalit residents showed us broken utensils and charred remains of other household things on their terraces, which they said were results of intermittent stoning and throwing of kerosene soaked burning rags by their Muslim neighbours from Kucha Chalan. The Dalit residents, however, all praise for the local police, who they said were prompt in intervening in stopping further attack—in marked contrast to what we heard about the police from the Muslim in Kucha Nahar Khan. There was one lone dissenter in the Dalit colony who complained about the late arrival of the police. Others, however, shouted him down. When we asked the residents why the broken utensils and burnt household property had been kept on the terraces, we were told that the police had asked them to keep it there.
From our interviews with the Dalit families of Katra Khawaspura, and the Muslims of neighbouring Kucha Nahar Khan and the Kalan Mahal area, it appeared that the two communities, although neighbours, and their houses neck to neck, had little contact with each other. Further, there were frequent sources of friction. The Dalits told us that the Muslims objected to their going up and sitting on their terraces, since the terraces on this side overlooked the inner courtyard of the Muslim houses on the other side, where the privacy of the Muslim women seemed to be always threatened by outsiders. The Muslims of the adjoining streets on the other hand, complained to us that the Dalits, after consumption of liquor, often tossed the empty bottles into the courtyard of the neighbouring Muslim houses. Thus tension appeared to have been built up all these years over private altercations which are inevitable in any crowded locality where the residents always feel threatened by incursions into their tightly hound world of privacy.
From Katra Khawaspura, we moved into Gali Jamunwali, a predominantly Muslim inhabited lane, where we found the residents highly agitated over the recent dismantling by the CRP of the old iron gate at the opening of the lane. (Incidentally, such gates, known as 'phataks' are a part of the traditional town planning of Old Delhi. They are locked from inside during the night to prevent the incursion of anti-social
elements from outside). The Muslim residents of the lane complained to us that while the gates elsewhere in the Hindu areas were left intact, the CRP people were systematically breaking down the gates of the Muslim-inhabited lanes. On being asked why they were doing so, the police came up with the laconic reply: "we are working under orders from above."
We made inquiries, and found that the lone Hindu family residing in Gali Jamunwali was safe and being looked after by their Muslim neighbours. Thus, there did not appear to be any immediate need for evacuating the Hindu family which could have been an excuse for the police to break down the gate.
We felt, therefore, that the way in which the old iron gates were pulled down in the Muslim areas—in Kucha Nahar Khana and Gali Jamunwali—amounted to unnecessary provocation to Muslim sentiments and further aggravated their feelings of victimization.
Our next stop was at Haveli Azam Khan in the Chitli Qabar area opposite Jama Masjid. We visited the house of Zahir Ali, a 17-year old boy who was shot dead by the police on Saturday. Zahir was sitting in a room on the first floor of the house, when a bullet flew into the room, penetrating the wall and killed him. We saw the hole made by the bullet at the corner of an air-cooler fixed on to the wall. We peeped down from the window of the room, and found that the CRP police were still stationed in the street down below and were pointing their rifles at the houses above. The houses in the street were all inhabited by Muslims.
Residents of the area complained to us that while the police failed to apprehend the real culprits, who came from outside and indulged in rampage, they harassed the residents and terrorized them by indiscriminate firing. They showed us the few Hindu shops and a Hindu temple in the area, which had remained unharmed. Here also an iron gate, 40 years old, was pulled down by the CRP men.
The riots left 88 injured and nine dead. Our inquiries at the hospital and interviews with the injured, threw further light on the composition of the victims, the methods of assault adopted by the rioters, and the nature of police retaliation.
The Saturday evening—by when the maximum number of
casualties had taken place—we traced 14 patients in the JP Hospital. There were three women victims. The majority suffered stab wounds in the abdomen, chest and back, acid burns, lacerated wounds as well as injuries from lathis on heads.
From our interview with injured we gathered the impression that they were innocent victims of both the communal frenzy and indiscriminate firing of the police. The preponderance of Hindus among those who suffered stab injuries and non-bullet inflicted wounds, indicates that they were victims of direct assaults by Muslim miscreants (which in many cases was confirmed by our talks with the injured Hindus) who might have attacked them with daggers, lathis or acid. Among the injured Muslims also there were some victims of similar stab and lathi attack—inflicted by Hindu miscreants. Our later inquiries revealed that many Muslims did not go to hospital but got treated by local doctors. But a substantial portion of the injured Muslims sustained bullet injuries which suggests that they were direct targets of police firing. We should add here that among the injured we found at least eight police personnels—most of them constables, including one Muslim. The nature of their injuries indicated that they might have been hit by stones or brickbat.
From our interviews with the injured, we could form an idea of the methods of assault used by the rioters (both Hindus and Muslims), and the methods used by the police, ostensibly to control riots. Let us give a few illustrations.
We interviewed two patients in Ward No. 6 of JP Hospital. One was Bhagwan Singh, a 16 year old Hindu, and the other was Islamuddin, a 28 year old Muslim. Ironically, both were sharing the same bed (No. 28) at the time of our interview, since there was no extra bed available in the Ward because of the rush of victims.
Bhagwan worked in a sweets shop in Churiwallan. While he was working in the shop on Friday, the assailants entered the shop and hit him with lathis on his head and shoulders. Islamuddin, who came from Purnea in Bihar, worked as a tailor in a shop by the side of a mosque at Sitaram Bazar. He and his three colleagues were attacked in the shop. The shop was set on fire. His two colleagues were burnt alive. Kerosene
was sprinkled on them before the fire was set. Islamuddin managed to escape with stab injuries in his thigh and a broken hand.
In the same ward, 19 year old Jameer was lying with a bullet injury in his back on bed number 21, while just by his side on bed number 20 was lying middle-aged Mulk Raj with five stab injuries. Jameer told us that on Saturday, he was standing on the doorstep of his house in a lane, when he heard sounds of firing. As he turned to go back inside, he received a bullet in his back, another victim of reckless police firing. Mulk Raj on the other hand, was a victim of Muslim communal frenzy. He owned the Amar Printing Press in Lal Quan area. On the first night of the riots (May 19) a mob attacked his press, beat up the workers and cut off the telephone wires. On the fateful Friday, May 22, again the mob came. This time they destroyed his press and stabbed Mulk Raj. He described his assailants as "Mussalmans of the same moballa with whom I everyday eat lunch, and have tea together." They shouted "Kafir hai, kafir hai, maro, maro" (He is an infidel: Beat him up).
In Ward No. 5, we came across 23-year old Abdullah of 2398 Gali Gudriyan of the Turkman Gate area, who was lying with bullet wounds on bed number 6. He told us that on Saturday, the fifth day of the curfew, having been without food all these days, he wanted to get some bread. From the rooftop of his house he asked the policeman on duty in the streets, if he could go out to buy bread. The policeman gave his approval. But as soon as Abdullah came out on the street, the same policeman took aim at him and he received bullet wounds on his waist and left hand.
Abdullah's experience once again confirms the allegations of police bias against Muslims that were voiced wherever we went during our investigation.
III Police Terrorization and Impact of Curfew
From our interviews with the residents of the walled city, as well as from what we ourselves saw during our tour of the curfew-bound parts, it was evident that the police had behaved in an extremely discriminatory manner, making the Muslims in
the main, face the brunt of their persecution. A sadistic desire to humiliate the Muslims in every way—from bullying and abusing elderly people to arresting and cold-blooded killing of younger people—marked the behaviour of both the local police and the CRPF.
While it is yet to be proved how many among those arrested were the real culprits, it is obvious from our findings so far that those who were shot by the police were innocent, like the two who were killed—24-year old Shafiquddin of Kuch Nahar Khan, and 17 year old Zahir Ali of Haveli Azam Khan (described in the previous section).
In similar cases of cold-blooded murder by the police in the past, the guilty police personnel were let off without any punishment. As we have mentioned earlier (in the previous section), the SHO Jag Pravesh Kaushal, who killed two innocent Muslims at Gah Qasim Jan in February 1985, was not even suspended in spite of repeated demands by the residents for his proper punishment.
This time also, the residents of Haveli Azam Khan named certain policemen who, they alleged, had behaved brutally, and demanded that they be punished. We were told that the SHO of Chandni Mahal police station, Ram Kumar Sharma, took particular delight during the curfew to patrol the streets with his men, firing every now and then in all directions. On Friday, he was seen by the local people waving his revolver and shouting "I am going to teach every Muslim a lesson." We heard from 18-year old Md. Asim, who was lying wounded on Bed No. 5 and ward No. 9 of J.P. Hospital that on Friday when he was coming out from a mosque after his prayers, it was this same SHO of Chandni Mahal police station who fired at him and injured him in his right leg.
We should add, however, that SHO Ram Kumar Sharma seemed to enjoy a different reputation in Gah Jamunwali of Kalan Mahal, where a prominent Muslim gentleman, Dr. Khalif Anjum gave him a clean chit and added that he had helped to quell riots there.
Later, on May 27, when a Hindu was shot dead in the Chandni Mahal area SHO, Sharma intervened to assert that the dead man was not a victim of communal clash but of a family
feud, and claimed that he had saved the situation from taking on communal turn. (See Indian Express, May 28, 1987).
Because of the controversy surrounding this particular SHO, we feel that there should be an inquiry into his actions—relating to both his role in different areas during the riots and his intervention in the May 27 shooting incident (which still remains a mystery as it is yet to be found out who shot the man—his relatives, Muslim or the police?).
One of our team members happened to be present as an observer at a meeting in the office of the SHO of Turkman Gate on Sunday, when the police officer met the Muslims of the area to take stock of the situation. Almost with one voice, the Muslims there complained about police partisanship and brutalities. They alleged that the police sided with the Hindus quite openly so much so that they shot in directions pointed out by the Hindus. There were cases of theft right in front of the police who did nothing to protect either the Muslim shops or the mosques. Muslim women and children were beaten up and innocent people were arrested. Because of police terrorization, many families from Blocks A, F and W had left their homes.
After having listened to their complaints, the SHO pleaded innocence about the culprits, whom he described as outsiders and anti-socials. Evading the responsibility of his own police force, he advised the Muslims to hand over trouble creators and guard their areas. In reply to his advice, a Muslim gentleman assured the SHO that the Hindu neighbours and their properties were safe in their area, and that the Muslims would pay with their lives to protect them. But, he added, if any unpleasant situation arose the Muslims wanted to be assured that the police would not be partisan and that they would get justice. To this the SHO, however, did not have any reply.
When we later met Mr. Amod Kanth DCP (Crime and Railways), and drew his attention to the allegations of looting by the police in the Turkman Gate area (see previous section), he dismissed it as "nonsense", and asserted that police had never behaved in a communal fashion anywhere. Defending the arrest of Muslims only in that area (see previous section), he said that it was because "most of the damage had been done
by the Muslims"
About the allegations against the Chandni Mahal police station SHO Ram Kumar Sharma, Mr Kaush said that those were attempts to malign Sharma because of his drive against the drug peddlars of the area.
It appears from the above that senior police officers in Delhi are unwilling ever to consider allegations brought against their colleagues and subordinates by the citizens. Yet, it is not unknown that the police have a long tradition of communalism, as corroborated by the strictures passed on them by several judicial commissions set up in the wake of communal riots. The National Police Commission in its Sixth Report in March 1981, noted several instances where "police officers and men appear to have shown unmistakable bias against a particular community while dealing with communal situations," and referred in this connection to the composition of the police force, "which is heavily weighted in favour of the majority community".
Given this unequal composition of the police force we feel that it is all the more necessary for senior police officers to be responsive to allegations of communal partnership against the police, and to take immediate action against the guilty to correct the mistakes. The National Police Commission in its Sixth Report recommended "whenever allegations are made they should not be brushed aside by senior officers, but should be carefully examined and if specific instances of misbehaviour on the part of any policemen are found to be true, such policemen should be punished in an exemplary manner so that the credibility of the police organization, as such, with the people, is not impaired."
But judging by both past and present experiences, when policemen found indulging in criminal activities like cold-blooded murder of innocent people, or looting of shops, are allowed to go scot-free, it appears that the senior authorities in the Police are bent on 'impairing the credibility of the police organization' with the citizens at large.
The police terrorization not only led to direct physical injuries like fatal wounds, but also to indirect sufferings to thousands of residents of the walled city. The continuous
curfew became a matter of considerable hardship to the residents all the more because of the rigid way in which the police imposed it. Here also a policy of discrimination was followed. While in the Hindu areas like Charkewalan, Dariba Kalan and Esplanade Road (which were also curfew bound), the people were allowed to come out on the lanes and chat in the Muslim areas like Chitli Qabar, Matta Mahal, Ballimaran, the streets were completely deserted with the armed police patrol being the only conspicuous sight. Anyone venturing even to peep out from the roof top could be a ready target for the trigger-happy policemen of the streets. A number of casualties in these Muslim areas could have been the results of such innocuous curiosity on the part of the residents, particularly the younger members of these households.
Discrimination was also noticed in the issue of curfew passes. While Ashok Jain, a Congress (I) Metropolitan Councillor told us that it was quite "easy" to get a pass, we found that Mirza Mohammad Osman, a Janta Party Municipal Councillor from the Hauz Qazi area, was denied a curfew pass.
But the worst effects of the curfew were on the civic services of the area. When we visited Pahari Bhojla on Monday the 25th—the sixth day of the curfew—we found it almost impossible to enter the lanes because of the stench from the garbage that had accumulated. Since the conservancy lorries did not enter the lanes, and the conservancy workers would not come to remove the garbage as often they were not allowed by the police to enter the lanes, the residents heaped them up at the opening of the lane. They feared that if it continued to accumulate, there was every possibility of epidemics breaking out.
The problem was aggravated by the dearth of water supply and lack of essential commodities. The curfew was relaxed initially for one hour between 5 a.m. to 6 a.m., and later for two hours only in the morning. But no private shops were allowed to be open during those hours. Curfew was relaxed in the evenings only after May 28.
By imposing and retaining the curfew for more than a week without making any arrangements or earmarking any agency to provide the trapped residents with essential services and
commodities, the authorities betrayed an extremely irresponsible and callous attitude. It appeared to be a form of collective punishment inflicted on an entire population of an area, for no fault of theirs, but for the unwillingness or failure of the police to make a distinction between the real criminals and the law abiding common citizens.
Even the curfew pass often did not help the residents. We were told of cases of CRP personnel picketing the streets often tearing off curfew passes issued by the Delhi police. They did not appear to recognize the authority of the local police, which again underlines the perennial tussle between para military force and the police creating problems for the ordinary citizens in the process. In this connection we should mention the case of Md. Inam, Lecturer in Commerce in Atmaram Sanatan Dharam College. On Friday, May 22 at about 7 p.m. he was returning home in Suwallan next to Chandni Mabal. He had a curfew pass and his college identity card. As he reached Chutti Qabr, he was attacked by the police from behind. In spite of showing his curfew pass, he was badly beaten up and was kept in the lock up. The FIR says that Inam was indulging in rioting and was arrested at 2.45 p.m. on Friday.
IV Living Conditions in the Walled City
The impact of police action and the curfew on the residents of the walled city cannot be fully measured unless we have an idea of the socio economic conditions that prevail in this part of the city.
The external symptoms of the living conditions in the walled city are overcrowded and deteriorated houses, an environment polluted by industrial units dens of criminals, drug peddlars and smugglers and high rates of unemployment and disease. The internal symptoms are resentment, hostility, despair, apathy, self depreciation.
Built in the 17th century to house a population of 60,000 the walled city of Shahjahanabad today contains at least ten times the original population. In the total area of 1,240 acres that it spans, the density of population is one of the highest in the world. The majority of households live in single-room dwellings, primarily because of low income levels. Rent paying
capacity is extremely low, and some families have been living in rented houses for decades paying the same amount as rent that their grandparents paid. This often leads to tensions with the landlords who want to hutch up the rents, or oust the old tenants. Om Prakash, who was killed on the first day of the riots (see Section II), used to pay Rs 8 for the two roomed house where he stayed with his family—a hone of contention with his landlord.
Most of the households do not have basic amenities like water or toilets. Long queues before public tuhewells are a common feature in this part of the city. Scramble for water often leads to disputes among the residents.
The walled city's environment is polluted by the numerous small industries (a large number of which are chemical industries) which are concentrated here. This area also serves as a major commercial centre, both wholesale and retail, for entire northern India. Over the years, there has been a large scale invasion of traditional residential areas by the commercial activities. The narrow and congested streets prevent administrative detection both of illegal industrial growth and of criminal activities.
The emotional ill-health that stems from these environmental and economic problems of the walled city is manifest in a continuum that ranges from the anxious resident seeking an escape to some other area, to the frustrated unemployed becoming criminally psychotic. The entry of all sorts of persons—either through commercial activities or in search of an ideal haven for criminal activities—into the walled city, has disturbed the old social cohesiveness of the various localities.
What is the role of the police in all this? The close association, collaboration, and at times identity, of criminals and the police is the pattern of day to day life in the straggling streets and lanes of the walled city. The young people come to know and accept it. Not only do they not respect the police but they see the police as part of the entire nexus of the underworld.
Living in a kind of unstable equilibrium, the younger members of the lower income groups, find it difficult to resolve the conflicts in which they are caught up. The shadow of persecution and suspicion that always hovers over the Muslim youth, particularly if he is unemployed, makes it even more
difficult for them to adjust to the environment. Some among them find means of escape in directive and assertive anti-social behaviour, the underworld gang to which they might adhere themselves, providing them with the necessary protection. Others gain satisfaction through temporary identification with the verbal militancy of charismatic Islamic fundamentalists.
Given this socio-economic backdrop of the walled city, one can understand how the miserable lot of the residents can get worse at moments of crisis, like the outbreak of a communal riot bringing in its trail police harassment, curfew, withdrawal of civic amenities among other things.
V The Role of Politicians and Political Parties
We had suggested earlier that the gradual withdrawal of secular political forces from the social and political scene of the walled city had created a vacuum which is being filled up fast by the increasingly powerful religious fundamentalists of both Hindu and Islamic varieties.
Our findings revealed that the changes in the policies of the Congress (I) leadership and in the composition of the party's local cadres in recent times, had to a large extent alienated old Congressmen (both Hindus and Muslims) who at one time wielded considerable influence on the population of the area, and could intervene effectively at times of communal tensions.
We found that the position of Muslim Congressmen in particular was extremely pathetic. Typical is the case of Dr Khaliq Anjuman of Gali Jamunwali in Kalan Mahal, who is a former government officer and is a prominent personality of the area. He told us that his repeated telephone calls to the Congress (I) leaders to stop the dismantling of the 'phatak' in front of his lane did not yield any results. None of the central Congress (I) Muslim leaders cared to visit the riot affected area and find out the problems that the residents were facing. Narrating his experiences with Congress (I) leaders of his community, a frustrated Dr Anjuman said that they were hesitant to voice the grievances of the Muslims before the party's central leadership. "They are more afraid of Rajiv than 'Khuda,'" he said. When someone suggested that the present riots could blacken Rajiv's reputation, he exploded, "One has to have a clean face to
get it blackened" (Kalkk un par parti hai, jinka muh safed ho).
The feelings of being let down by the Congress (I) leaders of their own community were quite widespread among the Muslims of the walled city. While some like Zamiruddio Khan, of Kucha Nahar Khan, gave somewhat mounted expression to the feeling, others like Dr. Anjum were more outspoken. The latter appeared to move to the conclusion that the Congress (I) was becoming a Hindu communal organisation—a feeling shared by both Hindus and Muslims.
We found that a number of Muslim politicians of the walled city who were once with the Congress (I) had left it in recent years and had chosen to join some other party. Mirza Mohammad Osman, an ex-Congressman who is now the Janata Party Municipal Councillor, squarely blamed the Congress (I) for the riots. He said that the main representatives of the Congress (I) to the area was Swaroop Singh (mentioned in Section III in connection with the Turkman Gate incidents), who was a known 'goonda' in the area. Mirza Siddique Ali, another prominent Muslim leader of the Janata Party felt that there was a deep conspiracy to provoke the Muslims to violent actions so that it helped the Government to prove its theory of destabilization. Syed Hamid Hussain Khizer, another ex-Congressman who is now the Muslim League Municipal Councillor, said: "It is unfortunate that the people who have joined Congress (I) are such that no good Hindu or good Muslim would join the Congress (I) now." He also accused Swaroop Singh of supplying bottles from his godown for use as missiles during riots. "Congress (I) is now like a 'nakli' (adulterated) vanaspati," he said.
The estrangement of traditional Muslim Congressmen from the present Congress (I) is not only because of their growing perception of the Congress (I) as a Hindu communal organisation, but also because of the increasing attraction towards it of anti-social elements. This was evident to us when these Muslim leaders repeatedly drew our attention to the local 'goondas' who were dominating the Party organisation in their areas.
This revulsion against the new breed of Congress (I) men was also shared by old Hindu Congressmen. We met Jawala Prasad of Naya Bans, who is in his late sixties and was at one time a
Congressman, but now belongs to no political party. When asked whether he perceived any change in the functioning of the Congress today, he narrated an experience as an example: "During the riots in 1947, Nehru heard that there were riots at Jama Masjid. He rushed there in his car. Today this does not happen." His son at this point interrupted and came up with a comment that summed up the difference between then and now: "In those days, there were 'jan-priya' (popular) leaders. Now we have bullet proof leaders."
Elderly people like Jawala Prasad still retain the spirit of communal harmony of the past. During the present riots, elders of both the communities of the lane where he lives, got together and decided to prevent any disturbances. Old Jawala Prasad demanded that he be carried on his 'charpoy' to the gate of the lane so that he could guard it.
But the influence of these elders on their respective communities is already on the wane. Among the Muslims in particular, the increasing feeling of injustice fuelled by discrimination in jobs and the communal partisanship of the police, is straining the patience of the younger generation. As an elderly Muslim gentleman put it bluntly at the meeting with the Turkman Gate SHO (described earlier in Section II): "If shots are fired into your house, and a hundred year old man comes and asks you to take it easy, even I will not listen to him."
The decreasing influence of the elderly leaders of the communities was also corroborated by other political leaders, like Premchand Gadodia of Kucha Sanjogiam, Naya Bans, who is the senior joint secretary of the Mercantile Association and member of the Home Guard. He said that in the past coordination between the various communities in the locality was maintained through regular meetings of the community representatives. But such meetings had stopped for the last 15 years. There was no regard or respect for the old leaders among the new generation.
Besides the Congress (I), and the Janata Party, the other major political force represented in the area is the Bharatiya Janata Party (BJP). The BJP member elected to the Delhi Metropolitan Council from the Jama Masjid area is a Muslim—Ragum Khurshid Kidwai. But the Muslim residents of the area
did not appear to have any confidence in her ability to intervene in their favour. They felt that the religious orientation of the BJP (pro-Hindu) could restrict her capabilities.
BJP leader of the walled city, Vaid Vidyaratna, expressed the view during our interview with him that the riots were pre-planned. But he would not elaborate on 'who planned it', beyond saying that it all started because 'Rajivji got embroiled in factional squabbles'.
But Vidyaratna's son, Atul, who is an RSS supporter, was more explicit in his views. He said that the Muslims were pro-Pakistani, and felt that in 1947, Indian politicians should have said that if there were no Hindus in Pakistan, no Muslims should be allowed to live in India. About the recent riots he said that the Muslims wanted to scare away the Hindus from the old Delhi areas and capture them for themselves. His opinions and militancy were echoes of the numerous leaflets that were distributed by the Vishwa Hindu Parishad on the eve of the riots.
The first time the BJP came out with a public statement was on May 20 when its leader Madan Lal Khurana said that the riots appeared to be pre-planned, arms stocked in houses were freely used in the riots and that the failure of the Delhi administration and the intelligence agencies were responsible for the riots. We have already referred to the next step of the BJP--its ambiguous role at the meeting with the Lt. Governor the next day (May 21) on the issue of lifting the curfew (see Section II). After that we find the BJP Delhi unit holding a protest demonstration on May 25 demanding the resignation of the Lt. Governor for his decision to relax curfew on the previous Friday. Led by its Delhi unit president Madan Lal Khurana, the demonstrators originally decided to stage their protest action at the Prime Minister's residence, but they were prevented from going beyond Fernze Shah Road, where about 115 persons courted arrest.
It is significant that on the same day (May 25), several other Hindu organizations like the Bajrang Dal, the Indra Prastha Vishwa Hindu Parishad, the Hindu Manch, the Sanatan Dharam Mahasabha and the Arya Samaj held demonstrations at various places in the capital, voicing the same demand—"Resignation of the Lt. Governor for allowing the relaxation of curfew".
It appears that the BJP, for certain reasons, is unwilling or unable to identify the actual sources of the conflagration and the real culprits. It has not yet been able to substantiate its allegation about the stocking of arms in houses.
We next tried to ascertain the views of left and secular political parties whose members are active to some extent in the walled city. Among them the major parties are the CPI and the CPI(M). A CPI spokesman with whom we talked explained the riots in terms of the increasing number of unemployed people who, he said, formed the anti-social elements in both the Hindu and Muslim communities. About the role of the Congress (I), he felt that it was basically a secular party but sometimes it went "astray". He attributed the causes of the riots to forces interested in breaking up the unity of the country. But when asked to identify those forces he would not name them. A CPI leaflet released soon after the riots, blamed 'imperialist forces' for helping the communal elements in the country to break up the unity of the country (An Appeal to the Delhi citizens from the CPI, signed by Prem Sagar Gupta Secretary, Delhi State Unit, CPI, Delhi, no date given).
The CPI(M) appeared to be less ambiguous in identifying those responsible for the riots. In a press handout released on May 20 (which was not carried in full by the national newspapers), it attempted to trace the roots of the inflammatory communal propaganda indulged in by the religious fundamentalists in both the communities. It released to the press photostat copies of communal handbills brought out by the Indraprastha Vishva Hindu Parishad, to which it had earlier drawn attention of the Delhi Administration which, it said, "chose to turn a blind eye". These handbills described the Muslims as unpatriotic and extolled the superiority of a Hindu Raj. The CPI(M) at the same time, referred in its handout to the 'fatwa' (mandate) issued by the Shahi Imam of the Jama Masjid to keep eating shops closed during the day in the month of Ramzan, and to attempts to forcibly implement the 'fatwa'. (According to information collected by our fact-furnishing team this was a meeting of Muslim eating house owners at the Jama Masjid, where they collectively took a decision to keep their shops closed during the Ramzan hours of fasting—a
decision which was announced in posters carrying the signatures of the eating-house owners. We request the CPI(M) to verify the matter). The CPI(M) press handout described this as a "direct interference in the rules governing commercial activity in the walled city and other Muslim majority areas". Describing the role of the "two big political parties, viz. the Congress(I) and the BJP", the handout accused the former which controls the administration, of being totally indifferent to the communal menace, and the BJP of trying to encash it for its "opportunistic ends".
From an analysis of the role of the major political parties in the recent Delhi riots, it seems that neither the leaders nor the cadres of these parties effectively intervened at any stage to stem the deterioration of communal relations in the walled city. Residents told us again and again about inflammatory speeches and distribution of provocative leaflets by religious fundamentalists on the eve of the outbreak. But few appeared to be aware of any efforts having been made by secular parties or groups to counteract effects of such communal propaganda. Although the CPI(M) in its press statement (referred to earlier) claimed to have held street corner meetings and public rallies before the riots, in the lanes and by-lanes that skirt Jama Masjid and Turkman Gate, few residents were aware of such meetings. The CPI spokesman whom we interviewed said that his party could not hold meetings because of the curfew but had distributed leaflets. But we fail to understand why the secular-minded political parties could not organize a peace rally (with the permission of the administrative authorities, through the affected areas on Wednesday and Thursday (May 20 and 21) which were comparatively calm days, and from local peace committees with members of both the communities in vulnerable spots, for the observance of a peaceful Jamat-ul-Vida the next day. At least, we are not aware of any efforts that were made by these political parties in such a direction. And yet, such a preparatory step could have gone a long way in preventing the violent clashes that erupted on Friday.
We agree at the same time that mere efforts of the political parties could not have been enough to contain the wide spread conflagration. The responsibility of Delhi administration
assumes importance in this context. It allowed for days together, right under its nose, the uninterrupted flow of open communal propaganda which bred mutual hostilities that were awaiting a small spark to explode into a communal riot. It did not take any action against those who were coming out daily with provocative leaflets which were not clandestine but carried the names of publishers and press.
We were also amazed to find the Delhi administration legitimizing the position of such blatantly communal organizations like the Vishwa Hindu Parishad. Among those who were invited to attend the meeting at the Lt Governor's place on Thursday, May 21, was Yadeshwar, a leading member of the Vishwa Hindu Parishad.
VI Conclusion
While we do not claim that the above account is an exhaustive analysis of the causes and a comprehensive record of the events during the Delhi riots of May 19-25, we do assert at the same time that our investigations have thrown light on certain major factors that contributed to the riots and aggravated tensions between the two communities both during and after the riots.
In this connection, we squarely blame the Delhi administration for ignoring the open communal propaganda that had been going on both within and outside the walled city for months together before the riots and for refusing to take firm steps against those indulging in such propaganda among both the communities.
We also hold the Delhi police and CRP responsible for antagonizing the Muslim minorities by discriminating against them and terrorizing them while tackling the riot situation, and later while implementing the curfew. In this context we would like to draw the attention of all to the warning we sounded in December 1984, in our report on the anti-Sikh riots in Delhi at that time, entitled "Who Are the Guilty? Referring to the apathy and hostility of Delhi administration and police against the minorities who were forced to live under a shadow of perpetual suspicion and terror, we warned "Muslims and Christians fear that a similar outrage can be perpetrated
against them also at any provocation." We find that instead of heeding to this warning and correcting its mistakes, the Delhi administration had remained apathetic towards the rising communal tensions in the walled city, and when the riots broke out the police acted in a way that was definitely hostile against the Muslims.
In the course of our investigations we noticed certain alarming trends which have grave political implications for the future of our country in general, and Hindu Muslim relations in particular. We found that over the years, attitudes on both sides—Hindus and Muslims—had hardened due to several factors. First, among the Muslims there is a feeling of discrimination against them, sometimes overt, sometimes subtle, in service and education—a feeling that was voiced by many Muslims in the walled city. Refusing to accept such discrimination, a new rising generation of Muslims are becoming increasingly articulate against such behaviour of the majority community. A number of young Muslims are venturing into business in the area. Both the articulations and the initiatives of the Muslims are often perceived by the traditional and orthodox Hindus of the area as signs of aggression. The Hindu hostility is further reinforced by the constant propaganda by communal organizations like the Vishva Hindu Parishad and irresponsible Hindu politicians of various parties which had built up a Muslim stereotype in their minds. This stereotype is that of a pro Pakistani Muslim. In the eyes of these Hindus, every Muslim is a potential Pakistani agent.
Such prejudices of the majority community are also often boosted by irresponsible and indiscreet circulation of news by the Centre. At the height of the Delhi riots, the Union Home Ministry chose to release figures of Pak nationals who had come to India and were untraceable, implying that they were all Pakistani agents who had gone underground and fomented the latest riots in Meerut and Delhi (see Indian Express, May 24, 1987). One could well imagine the impact of such news on both the communities especially when widely read evening newspapers like Veer Arjun, known for its communal bias, carried editorials and articles blaming "Pakistani agents" and "pro-Pakistan Muslims" for the riots in Delhi and elsewhere.
The Muslim grievances, against the majority community and the latter's suspicion about the Muslims—each feeding upon the other—have been further aggravated by the social distance between the two communities in the walled city. Although living side by side for many years, Hindu and Muslim families hardly intercommunicate. Women from a Hindu family in Nayas Bans told us categorically that they had no contact with the Muslim women next door. Whatever little relationship there might have existed between the male members of the communities, is also likely to disappear after the recent riots. A Dalit daily wage earner from Gali Magazine in Churiwalan told us that they used to play cards with their Muslim neighbours. But now, that would be over. Old residents of the area regretted that the past custom of holding occasional meetings of the elders of both the communities in the localities was no longer followed, resulting in widening the distance between the two.
In a situation already beset by mutual suspicions and social distance, some recent developments at the national level led to further complications in Hindu-Muslim relations in the walled city. The Shah Banoo judgement and the Babri Masjid-Ram Janamabhoomi controversy were two such developments. In the absence of a secular-minded strong leadership among the Muslims, the Islamic fundamentalists sprang into the scene and raised the hokey that Islam was in danger. For the common Muslim, the hokey gained more and more ominous dimensions as he saw the mushrooming of numerous Hindu militant organizations like the Vishva Hindu Parishad, Bajrang Dal, Hindu Shiv Sena, etc who began to claim almost every Muslim mosque or monument as the site of some old Hindu temple. The explosion of communal riots, one after another in Moradabad, Aligarh Ahmedabad, Meerut, drove the Muslims increasingly into the lap of Islamic fundamentalists who were perceived as their only protectors against the 'trishul dharts'. The growing identification of the Congress (I) run administration with Hindu communalism (evident during the 1984 anti-Sikh disturbances and in communal riots in Congress I ruled states), as well as the failure of the secular political parties to win the confidence of the Muslim masses, put the
secular-minded Muslim leaders to a difficult situation. They became more and more isolated from the mainstream that was developing in their community. Their place has been taken over by the fundamentalists who by their flamboyant speeches can rouse religious frenzy, by local ambitious politicians who assume postures of militancy, and by anti-social elements who are out to settle old scores and are bent on personal aggrandizement.
These trends could explain to some extent the pattern of the Delhi riots this time, where the Muslims often made the first offensive, as evident from the large number of Hindu casualties on the first day. But we should hasten to add that the figures in the hospital, from which we came to the above conclusion, may not give us the total picture. During our tour of the walled city we learnt that many wounded Muslims did not report to the hospitals, either out of fear of being harassed by the police, or for some other reasons. They got treated by their local doctors. Some cases of death of Muslims may not even have been reported, like the burning alive of two tailor shop workers in Sitaram Bazar, as disclosed to us by the lone survivor of the arson, Islamuddin (see Section 11).
The real victims of the riots, as usual, were the poor of both communities. Om Prakash, Shaukat Mian—the Imam of Masjid Hakim Baqi—and Shafiquddin were the sole earning members of their respective families. Their families now face economic ruin which the measly Rs. 20,000 compensation announced by the government can hardly mitigate. The trading and business community, mostly Hindu, lost a lot of property, but the large number of rickshaw pullers, thelewalas and the daily wage earners of both the communities suffered worse, not being able to earn even the Rs. 8 to 10 a day while the walled city was under curfew.
The impact of the riot has been to further sharpen the communal divide. Our impression is that the communal divide is becoming total, not only in the walled city, but outside also. Even people in responsible positions among the majority community appear to have developed closed minds, as evident from the biased news reports and editorial commentaries in some of the national dailies, and the venomous outpourings indulged in
by many against the Muslims in their day to day conversations. Typical is the attitude of Amrish Saxena, a young Hindu trader whom we met while waiting for our curfew passes at the Daryaganj police station. His comment on the Muslims who had made good 'Yeh kal ke gadhe jo khachar par chadha karte the aaj dhai dhai lakh ki jadad bann kar baathe hai' (Asses of yesterday who used to ride on mules only the other day, and now they are sitting on lakhs of rupees).
We cannot at this stage venture to suggest ways and means to tackle the wider communal problems that are besetting our nation as a whole, since it is beyond the purview of the present investigation. But on the basis of our findings from the inquiries about the recent Delhi riots, we propose to make a few recommendations.
First, we feel that there must be a sustained campaign by political parties and secular minded groups, like civil liberties or social welfare organizations demanding the punishment of the police personnel who are found to be responsible for the killing of innocent people during the riots. This is urgent in view of the common practice among the police to resort to indiscriminate firings whenever there are riots anywhere, leading to the killing of citizens, often children and women, who had nothing to do with the riots. Since they have never been punished for such acts of criminal irresponsibility (barring occasional suspension followed by reinstation after a certain interval, or transfer to some other police station), they can indulge in these murderous shooting outs with impunity. It is about time that they are made accountable to the public for their actions.
Secondly, we urge the secular forces—political, social and cultural groups—in both the communities to come together and evolve regular mechanisms for positive intervention in inter-community disputes, and for lessening tensions in the walled city. A vigorous campaign to isolate the religious fundamentalists in both the communities is of vital importance, in this connection. Permanent bodies like peace committees, or 'mohalla' defence squads consisting of both Hindu and Muslim residents could go a long way in preventing the communal elements and anti-social outsiders from disturbing communal harmony. It is important for secular-minded, democratic
elements within both the communities to have firm control over such committees or squads so that they do not fall into the hands of the local politicians or gangsters (which we understand from our inquiries, had often been the fate of similar bodies in the past). Given the communalization of the police and administration, the insecure citizens instead of going to communal organisations, may find an alternative avenue of seeking redress from these local citizens' committees.
IMAM CLIMBS DOWN TO VICTORY
Bharat Bhushan
Immediately after the Shahi Imam ordered the reopening of Jama Masjid, there was a mad rush into the mosque. Having been denied the opportunity of praying at the mosque for two weeks, people seemed pleased that they could offer prayers there this Friday at least.
They were laughing, shaking hands and congratulating each other. One of the devotees shouted to a bearded man at the door, perhaps an employee of the mosque, "Kyon Bhai Khomeini, kya haai ham (So, how are you, Khomeini)?"
"Khomeini" smiled, shook hands with his friend and went inside with him to pray.
The very fact that even friendly banter is so immersed in the religious idiom is so indicative of the influence of religion on the people. And no one knows this better than Syed Abdullah Bukhari, who became the first Shahi Imam of Jama Masjid to close down the mosque.
Ever since the Jama Masjid was completed in 1656, it has been closed only twice. To crush the local resistance to them and as a punishment to the Muslims who were at its forefront, the British closed it down from 1957 to 1862. But this time the mosque was closed by its own keeper to "punish" the government.
The Imam has indeed managed to force the government to accede to his demands. Action, including a departmental inquiry is being initiated against the station house officer of Chandni Mahal police station and all the Muslim youngsters.
arrested during the riots are to be released on bail.
Most of them are expected to be freed by this evening and if there is any hitch, a vacation magistrate or a special magis-
tra will expedite the bail proceedings tomorrow. Apparently,
a committee would also be set up to examine each case to decide whether the charges are fake.
The Imam seems unsure in victory—still suspecting that
the government may be up to some tricks even in agreeing to his demands. Hence, his claim that "I am only 75 per cent satisfied" and the declaration that the black flags and banners put up in the Masjid and the black cloth covering some of the domes and minarets would not be pulled down just as yet.
And hence also the threat delivered minutes before reopening the mosque: "Aai mere quam too sun le! Agar choti-choti baton mein dhoka hua to mein kal subah phir se Jama Masjid band karwa doonga (Listen O members of my community, if there is any betrayal) on the small assurances given, then I will close down the Jama Masjid tomorrow once again.)
One reason why the Imam is unable to stage a complete climb-down is that only till yesterday evening he was busy rejecting government overtures. He dismissed the setting up of government inquiries as "the cold-storage approach," which he rejected completely. His cronies were feeding his belligerence and he was describing the issue as "international" claiming that "15 to 20" journalists from London and "10 to 15" journalists from the U.S. were coming to see him every day. And perhaps it is not easy to accept the overnight resolution of "international" issues.
In any case, the issue had certainly become national. The Muslim clergy had got divided on the legitimacy of the Shahi Imam's action and a poster war had begun. Two muftis of Deoband had issued fatwas (edicts) against the action of the Imam, another from Madarsa Amcensia nf Delhi had done the same but then beaten a retreat under pressure.
Posters appeared condemning the Deoband muftis. They were asked in one poster whether it was not true that they were issuing fatwas against the same Muslims of Delhi who give them donations, 40 per cent of which the muftis use for eating murg-mussallam at home?
They were also shamed by quoting the instance of a similar fatwa in Spain which had apparently resulted in the martyrdom of the Muslim General Tariq Bin Ziyad.
If "history" was not being bandied about, then it was sympathy. Here was a cause no one could question (indeed, even avowedly secular social workers in the walled city confirm the charges against the police and brand its behaviour in the recent riots as partisan).
It was a cause that all kinds of elements could cash in. Thus even the redoubtable Haji Mastan put up a poster only this morning describing himself as a dardmand (sympathiser), claiming that blood drops instead of tears were flowing from his eyes because "the blood of Muslims had become cheaper than dirty water."
This poster juxtaposed the picture of Haji Mastan, Sadar, Dalit Muslim Minority Suraksha Mahasangh, to that of the Shahi Imam. Thus clearly, even those desirous of becoming a part of the sympathy wave were not hopeful of a solution so soon. So how did it come about?
Although no one questioned the path adopted by the Imam in public, many, including some influential younger elements close to him, thought that his unilateral decision to close the Jama Masjid may become counterproductive if the situation persisted. And it would persist because the Imam was leaving no room for compromise. These forces seem to have prevailed upon the Imam to give up his intransigence which finally made the solution possible.
Thus if the resolution of the Jama Masjid crisis is outwardly a victory for the Imam, it is also in fact a victory for the younger elements who have his ear. This is significant because it shows a movement although small at present, towards the emergence of a new and younger political tendency.
It is perhaps too early to comment on the various elements of this incipient tendency. But it is significant that for the first time this year all the Muslim hotels and restaurants in the vicinity of Jama Masjid remained closed during the day in the month of Ramzan. A number of youngsters claimed that they helped in appealing to the owners to close shop.
It is equally significant that most political observers believe
that the Naib Imam Syed Ahmad Bukhari, the Shahi Imam's son, may be coming out into his own as an important leader. His claim that he would be the first to disband the Adam Sena if the government banned all the religious senas and dals—was described by an old inhabitant of the walled city.
From Times of India, 13 June 1987
PART II
THE ROLE OF PAC AND MALIANA MASSACRE AND THEREAFTER
WHEN THE SO-CALLED GUARDIANS TURN INTO COMMUNAL ARMIES, ORDER CRUMBLES
M J Akbar
Meerut, May 31 This is the same story that one wrote from Varanasi and Pune and Jamshedpur and Ahmedabad and Moradabad and anywhere else Two communities start a slow pirouette of confrontation which gradually builds up to the moment when the tension must explode into violence Then the beat of a drum near a mosque, or a stone thrown on a procession, or a knife slipped into the side of a stranger in the darkness of the night The frenzy of battle, and then a rogue section of the police takes sides, presiding over murder, loot and worse to teach these "sala Pakistanis" a lesson, amidst the familiar echoes in the background of treachery and sinister motive Passion spent, the frenzy abates the same elements of the state machinery reassert themselves, the rogues are brought under some control, relief is distributed amid the familiar complaints of discrimination, rumour and the anger of revenge keep hearts simmering, and only slowly, very slowly, the living stop counting the bodies of the dead All that remains at the end is memory, the clutching sob of the simple, bewildered, ageing mother unable to understand why her 25 year-old son is dead today when he was living yesterday, and the liquid sparkle of an old father's near blind eyes, bearing a permanent
pool of tears, which will neither course down the white stubble on a dark face nor go back to the wells inside his heart from which they sprang.
But news is often nothing but the latest twist in an old tale. So, for better or for worse, read how the bodies from the town of Meerut were found floating down the Hindon canal.
The silhouette is easily recognisable despite the drapes, but the statue of the last Prime Minister at the Indira Chowk remains unveiled because the official ceremony has not been organised yet. The high street curves at the Chowk towards the Gulshan Cinema on one side and a new and rather pretty mosque on the other. Small lanes, like quicksilver rivulets, run off from the main road, and in these lanes survive as best as they can those who serve the needs of the middle class and the poor - the weavers and lower income wage earners. It is appropriate that the colony of Bihari Ansari weavers called Hashimpura should be near the Indira Gandhi Chowk, because it was during her memorable 1969-71 phase that they were first given loans and concessions which helped these weavers to reach the status which is best described in Hindi: "Khaata-peeta". It was the welfare programmes of those days which rescued this particular lot from destruction. They have pukka houses now, each with its set of looms on which the family works to earn its livelihood.
Abdul Majid's house is atypical for only one reason; there is a very beautiful neem tree rising at the centre of the courtyard. The old man and his wife were not particularly upset when the Provincial Armed Constabulary came to their house at five that Friday afternoon: after all, they were doing the same search in every Muslim house along this lane. They told the police they would sit near the tree while they looked wherever they wanted. There was nothing to find in any case, in the two dark rooms of the house. In the end, the only thing which the police could find in this home was a young man, the son, Din Mohammad, in his early twenties, and now the principal bread-earner of the family. They accepted their fate when the son was arrested. Once again, virtually every able bodied young man in the lane was taken away as a potential or actual rioter.
When six nights later on May 28, the police came to Abdul Majid's house and told him that his son in jail was anxious to meet his father they could barely understand this spasm of generosity, there was a tinge of anxiety, the fear that their son might be in pain after torture They reached the jail at around nine that night It was also the night of the Id moon, the happiest evening in the Muslim's calendar, but thoughts of a celebration were far away as the old man waited for the police to let him know what fate they had in store for him He learnt the truth finally at around 2:30 in the morning His son Din Mohammed was dead And the father had only been called to take away the body and bury it before dawn The orders were strict they had to hurry the body right away, before Id the next day, and a grave had even been dug As best as he could, Abdul Majid, with the help of a friend who had come with him, offered the last prayers and interred the putrefying body of his son in the darkness He was not alone There were three other fathers come to bury bodies that night
These four Muslim youths were not victims of riots but of cold-blooded murder by the police in lock-up
The district authorities have now given up trying to deny these four deaths In answer to how, the police say they have no explanation And the civil service administrator for Meerut airily waves away charges of discrimination by saying that it is always the administration, is it not, which becomes the goat in such matters (it slipped his burdened memory that the word is scapegoat, but no matter)
Arbitrary Killings
A fear and horror is running through the Muslims of Meerut that such deaths in police custody could be substantial More, and gruesome reports are surfacing about how the brutal and communal force known as the PAC has been taking its revenge on Muslims
More than two dozen young Muslims picked up from the Hashimpura area by the PAC on the bank of Hindon canal in Ghaziabad that night were shot dead and dumped into the water where their bodies floated downstream till they were fished out next morning And it might never have been known that these
men had died at the hands of the Provincial Armed Constabulary rather than in communal clashes were it not for a miracle. Two of those young men who were shot and thrown into the canal survived. Mujibur Rahman and Baduruddin (whose relatives were in Hashimpura when we visited it) were shot in the chest and thrown into the canal, but they did not die. When the Muradnagar police picked them up 10 kilometres downstream they were still breathing. They are in the Narinder Mohan hospital now, under heavy security, but with breath enough to narrate their awful story.
Who knows how many became victims of the PAC that night.
The administration says calmly that 2,568 people were picked up from about 70 mohallas. Is that the figure which was picked up or only the number which was finally lodged in a number of jails dotted across Uttar Pradesh from Agra to Fatehgarh (which is 400 kilometres from Meerut)?
The riots, and the toll they have extracted, were indescribably bad, but perhaps nothing compared to the arbitrary murder committed by the PAC.
There is a point beyond which it is impossible, at least for this correspondent, to describe death: you have to leave it as a statistic and be done with it. And though it is far from adequate, the only available answer we have for crime is punishment. It may not bring back the dead, but it does become a message for the living. In the first wave of revulsion after news of the PAC's behaviour spread, there was talk that the superintendent had been suspended, that the PAC might be withdrawn. Only days later, the gestures are being heavily diluted. Because the Provincial Armed Constabulary has become the hero of the Hindus. One slogan which greeted Rajiv Gandhi continually during his visit to Meerut on Saturday was "PAC zindabad". This morning as we were walking through the deserted, curfew-struck streets of Meerut, we met a khadi wearing chief warden who described himself as a Congressman doing his duty for the nation in these arduous times. There was one thing which we must never forget to write, he said. "Everyone in Meerut wants the PAC to stay." He insisted, "Ask anyone." Had it not been for the fact that the Prime Minister decided to drive himself and stopped wherever he noticed groups of Muslims on the streets, he would have returned to Delhi with only the orchestrated part of the story.
The truth about a communal riot has so many shades that it is rarely possible to define anything in black and white. As far as the battles between the people are concerned, there was enough guilt on both sides, the Muslim no less than the Hindu. But the partnership of the involved people is to be expected that, after all, is why the riots began in the first place. It is when the so-called guardians turn into communal armies that order begins to crumble. To allow guilty policemen to escape is to condone a police state. And the consequences of that will be far, far larger than just death in Meerut.
*Telegraph*, 1 June 1987
117 MISSING IN MALIANA, PRESUMED DEAD
Pankaj Pachauri
Corpses have no religion. They either have names, or as happened in Meerut in the last fortnight, they only have numbers. But in the sleepy hamlet of Mahana on the Meerut-Bagpat Road, khaki-clad men from the 44th battalion of the Provincial Armed Constabulary (PAC) tried their best to snatch even these numbers from the dead.
Survivors of Malliana are trying their best to identify those who were butchered by communalists in khaki or others, on the afternoon of last Saturday. The task is heart-rending, gruesome, and almost impossible.
Impossible because the two sackfuls of bones extricated from a kuchcha well in Abdul Sattar's courtyard cannot tell you whom they once belonged to. The PAC men dumped a gunny-bagful of salt into the well along with 11 family members and two young girls who had come to attend the marriage of Sattar's daughter.
After the rotting flesh and bones were retrieved from the well this Wednesday, Meerut's district magistrate, Radheyshyam Kaushik, announced at his press conference: "One more body was recovered from a well in Malliana today. The toll now is 15."
The administration counts the dead by the postmortems done at the Medical College mortuary where bones cannot be sent for autopsy. Likewise, the rotting dead bodies which floated...
downstream on the Upper Ganga Canal and the Hindon river were also never sent for autopsy.
On Friday evening, Salim Ali, an electrician, was doing the rounds of the deserted, grave like village with a list of 117 people who have been missing since the madness descended on Islam nagar, Sanjay nagar and Kalyan nagar, where the Muslims of the village were concentrated.
Eighteen persons from the missing list have been identified and buried in the middle of the village under army vigil, the rest are presumed dead. "No one got the chance to flee", Salim said. Kaushik does not entertain questions on the subject. "Now the enquiry commission will give its final word on Maliana", he hedges.
The massacre at Mallana was almost clinical in its swiftness. Within four hours the PAC executed their multi-pronged operation of besieging the Muslim hamlet, moving down villagers with their 303 rifles motting Harijans from the nearby Multan nagar area to burn the houses with the corpses in them, and finally, dump bodies in the canals about eight kilometres away from the village.
After their operation, the PAC men and their commandant, RD Tripathi, vanished from the scene. No one, including the state home minister Gopinath Dixit and the Uttar Pradesh Director General of Police, SK Bhatnagar, knows where the battalion under RD Tripathi is stationed now. "They have been removed from Maliana", is all they are willing to disclose.
The PAC men started entering Maliana on Saturday morning. Initially there was one truck posted at Transport nagar, about a kilometre from the hamlet. The auto parts and mechanic shops lining the main road had been burning since Thursday night. Some of the shops here belonged to the Malana residents, a majority of whom are petty businessmen and mechanics.
With the first truck reaching the village came rumours that "PAC is coming to arrest people". The village, especially the Muslims, became paranoid at the mention of the PAC—the force has a reputation of being anti-Muslim.
Interview with a number of survivors and those nursing
bullet wounds at the medical college surgical ward provided eyewitness accounts of what happened at Maliana that fateful afternoon.
The teose villagers had gathered at the main Maliana mosque for their afternoon namaaz at one o'clock. Fleeing was out of the question, as Meerut was under curfew. During the namaaz the villagers heard shouts from nearby villages inhabited by Hindus.
Soon after namaaz four more PAC trucks were seen winding their way into Maliana through the dusty kuchcha approach road. A jeep was also seen leading the cavalcade.
One truck was stationed about 200 meters from the main road. PAC men left the trucks, spread out in the hamlet in pickets of four and six. One such picket headed towards the mosque after learning that most of the males in the village were congregated there. That was at 2:30 p.m.
The first shouts were heard from near the mosque accompanied by the cries of people falling prey to the gunfire. As if on cue, the other PAC men started firing indiscriminately at anyone in sight. The village reverberated with bullet-shots and screams.
"I saw from the rooftop, about 2,000 people coming down from the Multan nagar side. The PAC men were calling them and waving to them," 14-year old Gulsaam said. He saw nothing after a bullet tore through his elbow and he crumpled into an unconscious heap.
Aizaz Ali Siddiqui took over where Gulsaam left off. PAC men led the lynch mobs to the houses belonging to Muslims, goaded them to kill people and set fire to the houses. For the next three hours, the madness continued.
Fearing that the entire Muslim community in the village would soon be wiped out, some of the villagers who owned trucks or were drivers, took their vehicles out and asked the people to pile into them. "We wanted to leave the villages with our lives. Let them set fire to our houses, we could always build more", Vakil Ahmed, a tailor reasoned.
But the people who had huddled into the trucks were asked to disembark at gunpoint by the PAC men. They ran again towards their houses, only to be intercepted by the bloodthirsty mobs
These houses, their occupants dead, are strewn with tell tale signs of what occurred caked, dried blood blood soaked stones half burnt clothes in the courtyards and ravaged, gutted rooms with fires still smouldering on the third day after the massacre.
The trucks which were to take the villagers away from the carnage were ironically used by the PAC men to carry their bodies to be dumped into the canals. All these trucks are now untraceable. There were at least five of them in the village according to survivors.
When the Senior Superintendent of Police reached the village around 7:30 p.m., an eerie silence met him. About 2,500 villagers had gathered in an open place near the village pond, surrounding the 10 dead bodies they had refused to hand over to the PAC. "You either kill us or take us prisoners", villagers told the policemen who reached there later.
At first, the PAC men had the nerve to justify their deed on the pretext that their men were fired at from the mosque, but they soon realised that the excuse was a very feeble one. None of them could show any injury, nor any arms were recovered from the village.
The PAC modus operandi of dumping the dead into the canals shows striking similarity to the behaviour of the administration during the fortnight long riots which ravaged Meerut. To keep the number of the deaths low, the police and PAC have been frequently dumping dead bodies on the outskirts of the district.
This is substantiated further by the fact that some of the bodies recovered from the canals were declared ten days old at the autopsy. A home guard's jeep was seen leaving the Jah Kothi area with three dead bodies. When questioned, the driver replied that he was going to the mortuary, but enquiries at the mortuary revealed that the jeep had never reached there. It was last seen heading towards Bagpat Road in the opposite direction.
The police have evolved a code name for this operation 'Samman taiyyar hai?' (Is the luggage ready?) The police men ask each other on the wireless network. If the answer is
"readymade maal taiyaar hai" (the readymade luggage is ready), vehicles are despatched from the Kotwali head-quarters to dispose of the bodies. Some wireless operators jokingly referred to the goings on as "operation readymade luggage". District control room chief, SS Yadav dismisses the allegations as mere "rumours".
In such a scenario, it is difficult to even start counting number of deaths in Maliana, let alone the rest of Meerut.
From Sunday Observer, 31 May 1987
A MASSACRE THAT WASN'T
Anil Saxena
This is a story of a massacre that wasn't. The toll in the so-called Mahana massacre of May 23, it is now clear, was only 15. And of the 15 only three persons died of bullet injuries, the rest succumbing to burns and other injuries.
The initial reports, based on a list put out by Mahana residents had hinted that anything between 78 and 111 people had been killed in a holocaust ostensibly perpetrated by UP's provincial armed constabulary (PAC).
The figure of 111 had been provided to UP's home minister, Mr Gopinath Dixit, and to media persons on May 24 by two residents of Mahana, Mr Basbir Khan and Mr Salim Akhtar Siddiqui. However, the list given to the Prime Minister, Mr Rajiv Gandhi, when he visited Mahana on May 30, mentioned only 76 names—22 of whom had been buried while the others were missing. The Meerut police continued to dispute these figures but did not give any definite figure.
According to the new senior superintendent of police, Mr Girdhari Lal Sharma, of the initial list of 111 people 84 have been located at Kalanjari village about 14 km from Mahana off the Meerut Baghpat road. Further, 12 others had escaped to Kotri, a village adjacent to Kalanjari. Thus brings the toll down to 15, which, according to Mr Sharma, is the exact figure of the post mortems done of Mahana's dead.
Mr Sharma's version was confirmed by this correspondent.
who visited Kalanjar yesterday. Some of the reportedly missing persons that this reporter met were: Abdul Aziz, Khatoon, Hakim, Zafri, Imran, Nizam and Asmat. According to them, they were evacuated from Maliana by the local police on May 23 and were later brought to Kalanjar.
In fact, according to Mr. Chand Prakash Singh, a Kalanjari resident, the group of 84 were originally inhabitants of Kalanjar, who had migrated to Maliana in search of work. When trouble broke out in Maliana on May 23, they simply went home.
The police have confirmed the genuineness of those who have settled in Kalanjar and their names have been verified by the DSP, Mr Awaz Ahmed, and other policemen. The signatures of these people have also been taken.
It is learnt from reliable sources that the initial FIR lodged by the people of Maliana had no mention of the killings by the PAC and the police. The persons named in the FIR were mostly Harijans belonging to the same village.
Mr Bashir Khan, officials say, was motivated purely by the lure of grabbing the compensation. Mr. Chand Khan, among those presumed dead, has complained to Mr. George Joseph, commissioner, relief and rehabilitation, that Mr Bashir Khan had claimed a close relation with Mr. Chand Khan and received Rs. 20,000 as compensation. Mr. Bashir Khan had been absconding for several days.
What really happened at Maliana on May 23? The exact sequence of events will be known only when the report of the judicial inquiry ordered by the government comes out. But, according to government officials and residents of Maliana, the deaths that day were largely a consequence of a communal riot in which the PAC's role was more in the nature of providing protection to one community.
Maliana is a large village (population: approx. 30,000) with a significant Muslim population (approx. 3000). Though on the out-skirts of Meerut, Maliana has never had any communal violence even as Meerut has been repeatedly wrecked by riots.
But tension had been building up in Maliana after communal riots again erupted in Meerut on May 18. Loudspeakers from mosques were repeatedly issuing calls to the faithful to come to the aid of an embattled faith. It was then that a PAC company
headed by Mr R D Tripathi was posted there.
The tension boiled over on May 23, and following a clash between two communities, a constable of the PAC was hit by a bullet. From all available sources, it appears that the PAC did then take sides. For one thing, all the dead belong to one community.
But, as the number of deaths due to bullet injuries makes out, there is no evidence to suggest that the PAC indulged in a large scale massacre.
As Meerut city which still has curfew during the nights, lumps back to normal, Mahana wears an abandoned look even during the day. It is unlikely that the deserters will come back to this village in the near future.
*Times of India*, 17 June 1987
LEARNING TO LIVE WITH OFFICIAL LIES
Neena Vyas and Nilanjan Mukhopadhyay
Even while the relatives of those killed in the riots and atrocities by the Provincial Armed Constabulary at Maliana, a hamlet near Meerut town, mourn for the dead, the Uttar Pradesh Government and the Central Government are trying their best to hush up the truth. The graves are there for all to see, but the State Government claims that the people supposed to have been killed have returned to the village. They have not.
On June 23, the villagers submitted to the district authorities a list of 67 people who, they say, were killed during the riots in Maliana. Of these, 36 died during the rampage by PAC jawans on May 23. When these two reporters visited Maliana, the villagers pointed out to a cluster of graves near the village pond. They said 22 victims had been buried there including 19 "whose bodies were handed over to us by the district authorities". Yet the district administration says that only 15 persons were killed in Maliana. "The graves are not a figment of our imagination," the villagers counter.
If the Union Home Minister, Mr. Buta Singh, is to be believed, only 10 persons were killed in Maliana. He told the National Integration Council meeting in Delhi on Tuesday that "86 of the 96 people alleged to have been massacred in Maliana have been traced."
The following day Mr. D.S. Yadav, additional district
magistrate, said "Fifteen people were killed in Mahana. After post mortem, the bodies were handed over to the relatives." In another office, 15 yards away from Mr Yadav's room, sits the deputy commissioner for relief and rehabilitation, Mr Govindan Nair. According to him, 19 people were killed in Mahana. "The list was given to me after verification by Mr G L Sharma senior superintendent of police," he added. In fact, the Rs 20,000 compensation ordered by the government for the next of kin of those killed in last month's violence has already been paid to the relatives of 19 victims.
The death toll in Mahana has become a game of numbers. The government is desperately trying to dilute the enormity of the atrocity in which PAC jawans went berserk, firing on unarmed villagers and setting fire to property after looting valuables. The police has claimed that the villagers had fired upon the PAC jawans when they had gone to conduct a search. But this is an untenable excuse as till date neither the police nor the district administration has been able to provide the names of PAC jawans injured in the "firing."
"Operation Cover up", to hide the truth from the public, has been going on for a month now. The incident took place on May 23. It was not mentioned in the day's official briefing. The next day curfew passes were not issued to reporters, neither did the officials mention the incident. On May 25, the Prime Minister called a meeting in Delhi after reports of the massacre appeared in a section of the national and foreign Press and a judicial inquiry was ordered.
Four days after the massacre, the State Home Minister, Mr Gopinath Dixit visited Mahana and was handed over a list of 111 people "feared killed". He was told that the people were "missing". The list was passed on to the police through the district administration and the police "traced" 85 of "missing" people at Kalinjri and Sewal Khas villages, a short distance from Mahana. That would mean 26 still remain "untraced" and are presumably, dead.
Sixty seven? Fifteen? Ten? Nineteen? Twenty six? What is the actual toll? Mahana's residents told these reporters that the initial list of 111 names was "incorrect because hundreds of people had fled the village and the names of the missing or dead
were collected from hearsay." They insist that the fresh list of 67 names is based on special FIRs and investigations by village elders.
The villagers also feel that the onus of tracking down the number of those killed should rest on the government. Islamuddio, a villager, said: "Isn't it the government's responsibility to find out how many people died?"
The government does not feel so. Disinformation was fed to newspapers and part of it was carried. One such bit of disinformation claimed that the compensation for a person called Chaod Khan, said to have died in the rioting, had been claimed by his relatives. But later, according to the "source" who fed the "news", Chand Khao was found alive. When confronted, both Mr Yadav and Mr Govindan Nair denied that anything like that had occurred. They said Chand Khan was injured and a compensation of Rs 500 was handed over to him. But Mr. Yadav cannot explain who fed the "news" and why the government did not bother to deny it.
Who are the 83 people "traced" by the police? These people were not among the 111 mentioned in the list submitted to the administration. They are not even residents of Maliaoa. For instance, Muhammad Anes Saifi, one of the 83 "traced" by the police, is a resident of Multao Nagar. He fled the village along with others after "some houses were burnt on May 22-23".
He said: "All of us ran away to our ancestral homes in Kalinjari. I come to work at Multao Nagar every day now, but return to Kalinjari for the night. No one from Maliaoa is in Kalinjari". Saifi was never on the list of 111 missing people to begin with. The 83 people "traced" are not the people who were missing at the time the list was submitted.
The fresh list of 67 names which has been submitted to the administration does not include anyone from Multan Nagar. Even earlier the residents of Maliaoa had never claimed that there were residents of Multan Nagar among the victims, though one person is said to have been killed in that locality on May 22.
The 67 names include those of 36 killed in the PAC rampage on May 23 and those who died in the rioting preceding the atrocity. The 36 victims of the PAC rampage were residents of
Maliana, Islamanagar and Sanjay Colony which are all adjacent to each other.
Till the Maliana massacre took place, the district administration had claimed that no incidents of violence had taken place in the Transport Nagar police station area, which includes the greater Maliana area. The truth is that even before the May 23 rampage, arson and violence had taken place in Multan Nagar, Chanderlok Nai Basti, Kishanpur and Shekhpur. According to a special FIR lodged with the police in Chanderlok Nai Basti, an entire family was killed by a mob on the night of May 21. Only three children had survived. Killings had taken place in Shekhpur and Kishanpur too. But curfew was clamped on the Transport Nagar police station area only after the Maliana massacre.
Statesman, 27 June 1987
THE COVER-UP
Sankarshan Thakur
If the Uttar Pradesh government's judicial inquiry team ever visits Mahana, it will find this village deceptively commonplace. A massive white-wash operation is underway; an army of labourers is working overtime to rebuild the rows of demolished or gutted houses, bullet marks on the walls are being erased and painted over. In the guise of relief work, all evidence of the May 23 massacre is being cleverly and quickly removed. No wonder the administration is so keen on the "repair work" and has deputed so many of its officials to finish off the job.
But then, the residents of Mahana will always be there to tell the story. There is Parvez Ahmed Siddiqui, just 28, who saw fellow villagers being gunned down by the men of the Provincial Armed Constabulary and the accompanying hordes who looted and set fire to houses; there is Naseem Bano, a broken woman of 55, whose house was ravaged by the PAC-led mob; there is Wahidan, a widow whose son Ashraf was killed in the firing and now goes around the labyrinthine streets of Mahana looking for him behind every door. The people of Mahana have given her up as mad. "Her son is dead, we saw him die, where will he return from?" people ask.
The Meerut administration claimed last week that 85 of those reported missing from Mahana had "returned" to the village and with this, "new facts were emerging on the alleged
massacre." Police officials also began asserting that "no massacre took place in Maliana, only 15 people were killed."
Indeed, it was made out that the entire Maliana operation was "planned to defame the PAC." Officials called newsmen individually and fed them stories of how those presumed dead in Maliana had been "discovered alive and well in village Kalenjari" and how the "malicious exaggerations" of the press had been exposed. Boasted the additional district magistrate of Meerut, Mr D S Yadav, "We have found 85 people, the others will also be found soon and the real story will be out."
Investigations, however, reveal a situation quite to the contrary. Sure, there are 85 persons, including several women and children, taking refuge in Kalenjari, about 10 km from Maliana, off the Meerut Baghpat highway. But they do not belong to Maliana. They come, instead, from Multannagar which is some distance from Maliana. And the reason they are hiding in Kalenjari is that they have heard what the PAC did in Maliana.
"We escaped from Multannagar after the attack on Maliana. The police and the PAC tried to harass us on the way but we were somehow able to reach here because we have relatives in Kalenjari and they came to our rescue," says Nandu Mian, one of those taking shelter. All the (15 or so) families in Kalenjari are from Multannagar. This fact they make very clear: "We belong to Multannagar, not Maliana village," says Ikram, "and we are not returning unless the conditions become all right."
These people neither belong to Maliana and nor have they "returned" as the district administration contends. It is an unabashed and obvious cover up. The Meerut administration's executives should have known that they would be caught—people would go to Maliana and Kalenjari to check. But the cover up was, perhaps, more mischievous than it was obvious. Their basic attempt was to project Multannagar as a part of Maliana, and then say that people from Maliana were alive and returned.
Maliana is part of a cluster of semi-urban settlements. Some 10 km west of Meerut, and since it is the biggest locality in the area, even other settlements like Sanjay Colony, Indira Nagar
Islamnagar, Shekhpura and Multannagar, are often referred to as Maliana—a "greater Maliana" of sorts. What is crucial is that in the May 23 incident, the PAC action was concentrated on Maliana village and not in the entire area referred to as Maliana. Some other settlements like Sanjay Colony and Kishanpura were also affected but not so badly. Multannagar was spared the wrath of the PAC but as the fugitive residents in Kalenjari have affirmed they ran away apprehending trouble.
The administration also made a big show of the fact that 33 families of Maliana had "returned" to their houses. They made it sound as if members of these families too, had been proclaimed dead by the "sensationalist" press. The truth about the return of these 33 families is very different. These were the people whose houses had either been broken down or burnt, and who had been forced to live in a makeshift relief camp in the middle of the village. A few weeks after the May 23 incident, they just returned to their houses.
What has actually happened in the aftermath of the Maliana massacre is that the administration has been aided in muddying the water while some unscrupulous elements who have been politicking with variously exaggerated lists of the dead and the missing—one Basheer Ahmed Khan is even said to have collected a huge compensation in the name of people who do not exist (although it is not clear how the administration release the compensation amount if it had no record of the dead person). Some of the lists of the dead prepared by self-proclaimed guardians run into several hundred names. Thus has made the administration's complaint about "exaggerated" accounts of Maliana seem convincing.
In this welter of accusations and counter-accusations, the real issues have got lost. If indeed, as the administration says, there was no massacre in Maliana, if indeed the entire operation was "planned to desame the PAC", how is it that close to 70 houses were either burnt down or demolished in the village? Why did the administration resort to such a huge repair operation in Maliana? Were the Army officials, who said they counted close to 50 bodies when they came into Maliana, lying? And the most clinching, if there was no massacre in Maliana, why did the Uttar Pradesh government hurriedly have to order a judicial inquiry?
It is, perhaps, not a coincidence that the Uttar Pradesh government has chosen a time when the PAC is under severe censure to leak a portion of the inquiry commission report of the 1980 Moradabad riots which it had kept under wraps for several years. The substance of the two lines leaked is that the PAC was not to blame for the Moradabad riots. Consider now what a report on the Moradabad riots in Sunday (September 1980) had to say:
"Once again this was a case of police brutality pure and simple. The forces of law and order unleashed their fury on thousands of Muslims who had come to offer prayers on their day of joy and celebration. Hundreds died, including many children. Both the Hindus and Muslims of Moradabad are telling anyone who will listen that what happened on August 13 in their city was a straight clash between the police and Muslims. But the police in order to throw a cover on their brutal misdeeds, are telling lies about what precisely happened and building up false trails that might shift the focus of attention."
What was said of Moradabad seven years ago fits well into what has happened in Mahana and other parts of Meerut. The same PAC acting with the same brutality and resorting to the same devious means of cover up. The PAC lobby in Uttar Pradesh is working overtly and covertly to somehow shift the blame. The shrill and stage managed pro PAC demonstrations in the presence of the Prime Minister and the volley of pro PAC statements from various quarters are examples.
The shrill and stage managed pro PAC demonstrations in the presence of the Prime Minister in Meerut and the volley of pro PAC statements from the local MLA, Mr Jai Narain Sharma (there will be devastation in Meerut if the PAC is removed, he has said) are only part of an organised campaign. So is the leakage of "telling" parts of the report on the 1980 Moradabad riots.
It is not strange that while the administration readily allows unfettered pro PAC campaigns, it blindly refuses to entertain even a word against it, much less take action? There was an announcement from Lucknow immediately after the Mahana story broke that Mr R D Tripathi, commandant of the PAC battalion which was involved in the massacre, had been suspended. But this was promptly and firmly denied.
Instead, forthright officials like Mr. B.K.V. Nair, the senior superintendent of police in Meerut, who had played a positive role during the riots, were sent packing. The chief secretary of Uttar Pradesh, Mr. Mata Prasad, who quoted higher figures of the dead and missing in his briefings in Lucknow than what the Meerut administration was claiming, is now awaiting transfer orders. Professor S.A.P. Rizvi, chief proctor of Meerut's D.N. College, who asked the district magistrate to give details of the number of dead and missing in a peace committee meeting last week, was arrested under the National Security Act (NSA). His only fault was that he had asked uncomfortable questions.
While the administration is making bogus noises about the "return of those presumed dead" it has kept silent on several crucial questions about Mahana as well as PAC operations in other parts of Meerut. For instance.
* What was the provocation for the PAC to raid Mahana which has never seen communal disturbances and which was calm even as Meerut burned?
* Where are the missing people from Hashimpura and other neighbouring localities who are alleged to have been picked up by the PAC during the search operations?
* Where did the dead bodies in the Hindon and Ganga canals come from?
* Is it true that there were two major massacres in and around Meerut for, the bodies were recovered from both canals and there is no linking channel between the two?
* What is the truth in the allegations of Zulfiqar Nasir (who said he had a providential escape) about the coldblooded PAC shooting on the banks of the Ganga canal killing dozens of persons near Muradnagar?
* What do Babuddin and Mujibur Rehman, two other survivors of similar operations, have to say in their report to the Ghaziabad police?
* Why are the two being kept incommunicado under heavy guard at the Mohanagar hospital? Is there a threat to their life? If so, from whom?
* Is it a mere coincidence that the headquarters of the 41st battalion of the PAC (which is said to have been involved in the
killings) is just a few minutes drive from the scene of the massacre which Babuddin and Mujibur Rehman survived?
A large number of people from Hashimpura and its environs are still missing—their exact number is impossible to pinpoint as there has been a proliferation of lists of missing people giving varying estimates. The administration, despite repeated inquiries from relatives, says it has no record of them. But to name just a few: Ashraf (20), Qamaruddin (24), Siraj Ahmed (22), Naeem (15), Islamuddin (21) and Ayub (17). Nobody seems to know what happened to them nearly a month after the PAC trucks took them away.
Some of those who were in different jails of Uttar Pradesh have been released on bail but they have come with broken bodies and horrifying tales to tell. Merajuddin, whose arm and leg were fractured in the police beating, said he was shifted from one police station to another before finally being thrown into the Fatehgarh prison. "At every place we were beaten up mercilessly. I would faint, they would give me water and then beat me again."
According to Ahmed's account—his arm is broken and his fingernails are blue from beating—the PAC and the police never offered any food or medical aid to the prisoners. Neither were they produced in any court. "They had just taken us away to teach us a lesson, that is what they kept saying as they beat us," Syed Ahmed despairs.
Dr Zahiruddin Asri, a Hashimpura resident, who has been treating the victims says very few escaped with minor injuries. "It appears the torture has been very severe. Some have come back with near fatal head injuries and are still serious. I would not be surprised if some died in the jails," he fears.
The administration and the police, of course, have an obvious alibi. These people were not beaten up by the police, they were injured in the riots and the police saved them. That is the line they will take. That is the kind of cover they have always found and got away with.
*The Telegraph*, 21 June 1987
MALIANA MASSACRE: EVIDENCE OF THE GOVERNMENT'S COVER-UP
Sankarshan Thakur
This story is for the benefit of Mr. Buta Singh, Union home-minister, who told a meeting of the National Integration Council (NIC) that the allegations of a massacre at Maliana by the PAC was "disinformation" and that the "missing" were returning home.
This story is meant for the eyes of the chief minister of Uttar Pradesh, Mr. Vir Bahadur Singh, who has launched a massive cover-up of one of the most shameful episodes in our independent history.
This story is meant for the district magistrate of Meerut, Mr. Radheyshyam Kaushik, a government officer who has been lying blatantly in an effort to hide a brutal massacre of innocent people.
This story is meant for the conscience of those bureaucrats who claim that they are now confused about what happened in Maliana after the Uttar Pradesh government's "clarification".
All these men of power and authority have been trying to persuade the nation that nothing much happened in Maliana. And the government of Uttar Pradesh, through its information officer in Delhi, has been planting stories on the media to spread such an impression.
We print here the death toll—a list of people who were
killed by the murderous jawans of the Provincial Armed Constabulary and the mobs that accompanied them only in Malana (The massacre of those picked up from Hashimpura is another story. So far, this callous administration has not tried to cover up that accusation, at least.) We have reason to believe that the number of the dead in Malana could be higher than this.
This is a list of 67 names given by relatives and friends who witnessed the massacre and, therefore, the margin of error is minimal. The final tally could well be bigger because many bereaved have run away in fear and have not yet returned. On page five we print pictures of individuals who have lost close relatives, and who will be eager to provide witness of their horror to an honourable and independent body.
Here is the list of the dead: Mudassar Ali (26), son of Liaqat Ali, Mohammed Asraf (40), son of Karim Baksh, Gulam (22), son of Anwar Ahmed, Tasleem (20), son of Hanif Ahmed, Nasir Ahmed alias Muina (24), son of late Bashur Ahmed, Kaniz Fatima (42), wife of Mohammed Yaseen, Bahlu (6), son of Ali Hasan, Shahjehan (12), daughter of Umar Deen, Mohammed Ali (70), relatives unknown, Ibrahim (11), son of Abdul Sattar, Yusuf (22), son of Abdul Sattar, Yunus (15), son of Abdul Sattar, Noorjeban (10), daughter of Abdul Sattar, Shabnum (8), daughter of Abdul Sattar, Zehunnisa (20), daughter of Abdul Sattar, Ballo (24), daughter of Abdul Sattar, Asgari Begum (40), wife of Abdul Sattar (45), son of Mohammed Ali, Munir (20), relatives unknown, Sanjeeda (10), daughter of Shareef Ahmed, Mehmood Ahmed (35), son of Aziz, Naseem Bano (30), wife of Mehmood Ahmed, Mohammed Arif (15), son of Mehmood Ahmed, Mohammed Waris (8), son of Mehmood Ahmed, Aftaab (3), son of Mehmood Ahmed, Taranum (12), daughter of Mehmood Ahmed, Mehmood Ahmed (60), relatives unknown, Mohammed Akbar (65), son of Itwari, Bilkees (17), daughter of Mange Khan, Mange Khan (40) son of Kabul Khan, Raissa (13), daughter of Mange Khan, Kashmiri (38), wife of Mange Khan, Rashid (45), son of Mohammed Shakura, Eedan (40), wife of Rashid Ahmed, Akbari Begum (50), wife of Khairati Junna (50), son of Ahmed Karim, Saeeda (30), wife of Ismail Ahmed, Shakoora (80), son of Alla Mehr, Rafiq Ahmed (26), son of Bashur Ahmed, Kallu
(45), son of Hussain Baksh, Bashiran (40), wife of Almuddin; Zeenat (35), wife of Naseem Khan; Razia (40), wife of Sageer Ahmed, Sageer Ahmed (45), son of Ismail Khan, Usman (28), son of Subhan Ali, Jumma (age unknown), son of Abdul Karim, Rafiq Ahmed (48), son of Allahdiya, Husnoo (45), wife of Rafiq, Nafees (26), son of Rafiq, Rahisu (20), son of Rafiq; Neemo (22), wife of Nafees, Sahra (18), wife of Rahisu, Gudiya (5), daughter of Nafees, Bahlu (3), daughter of Nafees, Munni (2), daughter of Nafees, Zulfiqar (1), son of Rahisu; Farooq (35), son of Kallu, Muneeza (32), wife of Farooq, Jullo (14), daughter of Farooq, Mumtaz (12), daughter of Farooq, Sartaj (9), son of Farooq, Ahsa (4), daughter of Farooq, Guddo (5), daughter of Farooq; Rukhsana (12), daughter of Anis Ahmed, Shehzad (2), son of Anis Ahmed, Reshma (1), daughter of Farooq, Bahlu (4) son of Munni Begum
Special FIRs have been registered by relatives and village elders for each of the 67 dead Islamuddin, a Maliana resident, who has been maintaining records of the PAC massacre and its aftermath, showed this correspondent copies of the complaints, duly stamped and signed by the authorities. "If the government says nothing happened here and very few people died, we are ready to agree. Only, they should return our people. The government is lying."
The Telegraph could verify the death of 49 persons by speaking to close relatives and neighbours of the dead in the Maliana area. Several affected families have abandoned Maliana in fear after what happened on May 22 and 23—their houses, the ones that still stand, lie empty.
The Meerut administration, in its report, is believed to have said that nothing much happened in Mahana. The district magistrate, Mr Radheeshyam Kaushik, himself boasts that "the administration has come clean with the return of 85 of the missing persons. There was no massacre in Mahana as reported in a section of the press." The simple question to ask him is: If nothing much happened in Mahana, how were 70 houses in the village, which his own administration is now repairing at a cost of over Rs 2 lakhs, burnt down or demolished?
And the claim that 85 of the missing persons from Mahana had "returned", or been "traced", too, is a cover-up for the
crime. The 85 persons who have taken shelter in Kalenjari village are not from Mahana hut Multannagar. And Islamuddin of Mahana makes it very clear that the list of 67 dead has nothing to do with the people in Kalenjari. "They are different people, we never named them in our list. Perhaps the government prepared its own fake list and then said people had been found."
The government's attempt to confuse the entire affair is only too evident from the different figures they have been quoting on different occasions. Officially the administration says only 15 persons died in Mahana but the same administration has given compensation claims for the deaths of 19 people.
The Uttar Pradesh government claimed that '85 of the 111 missing persons have returned,' and yet, some days later, the home minister, Mr Buta Singh, tells the standing committee of the NIC that "86 of the 96 persons have been traced." How did 111 become 96?
The people in Mahana are still there in fear and dread, because the PAC is still around, and the chief minister is still around, and the lying district magistrate is still around. And this story is, in the end, an appeal to the conscience of India to stand up and cry against the horror and fascism.
*Telegraph*, 28 6-87
BLOOD-HOUNDS WITHOUT LEASH
K.M.
Over two decades ago, when Pravir Chandra Bhanjdeo was killed in his Jagdalpur palace, it was stated that the SAF deployed there went berserk when its Naik fell to an Adivasi arrow, stormed the palace and killed the Raja.
Since this story of armed constabulary running amok has been repeated at many places in the country during these two decades it would be better to remember that the common man who is not paid or trained to be disciplined has no right to run berserk if his near and dear ones are butchered—all that the law permits him is to lodge a complaint with the police and wait for two decades or more for justice to be done. Why should a trained, armed, disciplined force should get away with this story, is beyond comprehensions. But, all the same, this force has so far got away with it.
An IAS officer had stated privately, at the time of the Bastar killing that, that was the price one had to pay for keeping a bull dog (SAF) as a pet.
But to keep a bull dog as a pet, one has to train it and keep it and keep it in leash. At least two Prime Ministers—Morarji Desai and Indira Gandhi—publicly articulated the urgent need to re-organise the PAC of Uttar Pradesh to train it in a fashion that it acquired not only a human face but human behaviour as well.
In fact, after the Moradabad 'riots' about six years ago, the UP Government stated that it would raise a new battalion of PAC under a new recruitment policy (with liberal representation of the minority communities to the force).
The State Government is now content with saying that the minority representation in the police force of the State is gradually looking up. But if an IAS officer is to be believed (again in private conversation), the prospect of being posted in
every riot torn town raised the "recruitment fee" (bribe) immediately. If a person had to pay Rs 10,000 to be taken on the police force, he had to pay Rs 20,000 to get into this PAC battalion.
The argument is simple. A civil police constable has an opportunity everyday of the year to make a fast buck. He can be posted at a road crossing and collect his "fee" from every that passes that way. He could collect his "hastia" from all the street vendors. He collects from both sides in every case of crime. But these opportunities are just not available to a man in the PAC because he comes into direct contact of the people only when there is a breakdown of the law and order machinery —like a riot. That provides him with the only opportunity to earn a little on the sly (the looting spree in every riot torn area). If a battalion was to be rushed to every disturbed area its opportunities to earn would be immense.
Jamshedpur, Basti Varanasi, Barabanki, Moradabad, Aligarh, Meerut—wherever there have been communal flare ups, almost the unanimous demand from the minority community has been for the immediate removal of the armed police from the area. At many places Muslims have publicly stated that they had been protected by their Hindu neighbours from the police atrocities.
Why is the Armed Constabulary an anathema to the minorities? Perhaps, its composition would afford a clue. The bulk is semi educated young men from the Backward communities—the first generation literates. The strength of these constabularies was raised when the Jana Sangh was in power in the States in northern India.
The situation could perhaps have been retrieved later by training the PAC all over again—not in the use of their heads. Perhaps the authorities could realise that mere chest measurements are not enough for recruitment—that is the outlook, the behaviour, commitment to secularism, the need to keep calm in moments of stress were also important. But something held the authorities back.
Was it the spectre of 1973? There was a PAC revolt in UP there and poor Kamla pati Tripathi, the then Chief Minister of UP, had to pay the political price for it. Since then, the
official policy about the armed constabularies in the States seems to be to handle with care with gloves—as though it was combustible material. There must be sound reasons for not listening to two Prime Ministers, representing almost the entire political spectrum in the country.
Whatever the reasons, the price is being paid now. Meerut has had a qualitatively different one from other riots. For a few years, communal riots were quelled expeditiously and peace was resorted everywhere in a couple of days. In Meerut, however, the members kept smouldering and occasionally throwing up flames for a long time. It was different in another aspect. For the first time, perhaps, men were picked up from their houses during the curfew hours, taken outside the locality and shot down. A friend of the minority community said bitterly—the PAC is the sword arms of Hindu communalism.
There were other symptoms of the difference in Meerut. One such was that men like, Shanti Tyagi, blamed only the Muslim communalism. Then Congressmen openly accused their own MP of a bias. These are the wages of inaction of the secular forces in the country. For years, now, communal sentiments have been roused systematically and political parties have looked at the development merely as a law and order problem.
Above five years ago, a commentator had written about the rising divide between different communities. He had said that somehow a Hindu is more of a Hindu now, a Muslim more of a Muslim. They tended to look at the problems from the communal angle. It was because there was no other angle before them. Because nobody had shown them the Indian angle.
Precious years have been lost since then—precious years since Indira Gandhi suggested the reorganisation of the PAC, since the Vishwa Hindu Parishad took out Ekamata Rath and Ram Janaki processions, since communal venom began to be spouted from public platforms, since secular forces went into hibernation.
Now the blood-hounds are in the street without the leash.
Excerpts from Mainstream,
13 June 1987
GANGA CANAL OF DEATH
Sanjay Suri and Rahul Pathak
It has become a canal of death. Not a life giver, but a grave for Maliana's dead. The Ganga canal is not holy today, it is deathly. It brings no purification, only putrefaction.
The canal became the carrier of corpses four days ago. Estimates of the number vary from 50 to a few hundred. Everyone has seen the bodies, no one has counted.
Everyone has seen the bodies, except the police, so they say. The District Magistrate, Mr. Nazim Zardi said, 'I have no information about any bodies found in the canal.'
The police at Munsoorie say it's all rumour. "We have not seen any body in the canal," said a sub-inspector at the police station. "Nor has anyone else. They are just making stories."
People from the village said they saw several bodies today. "The bodies sometimes get stuck at a dam downstream," they said. We drove along the bank downstream to Dehra.
Armed policemen were patrolling the bridge. "They have just cleared the crowd," said an elderly resident. "A body floated down a moment ago," he said. "It got stuck at the dam, and the police had to push it along."
A policeman rose to stop us as we went further, but we found the body. It was the body of a man with his legs hunched up. Two crows were perching on the body, while pecking at the rotting flesh.
The sub inspector from Munsoorie had followed us on his
motorcycle. He drove off as we walked up to approach him. "There are no bodies here," he said, as he went.
Half an hour's drive along the bank upstream and we reached Barkha village. Here, as a dairy worker was talking about the bodies in the canal, another body floated up. It was the body of a woman, floating upside down. A blue sari was wrapped around the bloated body.
Everyone in the area was talking of the bodies, except the police. The canal flows down to Bulandshahr before it splits into smaller canals. At Bulandshahr, the town was full of talk of the bodies. On Monday night, several bodies had got stuck at a dam. They were pushed out by the police in the morning.
From Muradnagar to Bulandshahr, people stood on bridges, watching for bodies. The police stood with them, to make sure that the bodies did not get stuck, and to clear the bridges of the crowds that collected to watch.
Policemen were posted at every dam down the canal. The instructions were clear. They were to get the bodies out of their area. "We have been instructed not to try to stop them," said an official at a dam. "We are told that if we do, it will lead to problems for us."
At Samota, near Sikandarahad, an old man was given the task of clearing bodies that got stuck in the dam there. "I am frightened," he said. "I cannot look at the bodies, I cannot bear the stench. And still I have to push them through the gaps."
The police, who do not let people enter Meerut, are now circulating the remains of Meerut's dead along the countryside. The result is visible. Tension has spread to all these areas, up to Bulandshahr and beyond.
—From Indian Express, 27 May 1987
TALE OF TWO SURVIVORS
Samita Gupta and Rahul Bedi
Twenty-nine persons of Hashimpura in Meerut, all members of a community, were rounded up by the police on May 22, driven 15 km to the Gangnahar, shot one by one at point blank range and hurled into canal.
Only two of the 29 are alive today—in the male surgical ward of Narinder Mohan Hospital here, an hour's drive from Meerut.
The two are Babu ud Din, 17, and Mujibur Rehman, 25, powerloom workers in Hashimpura. The authorities, fearing another attack on these, the only surviving witnesses to the pogrom by the Provincial Armed Constabulary (PAC), perpetrators of the massacre, have placed two guards each inside and outside their room.
According to their statements, the two men, along with 27 others, were rounded up at around 4 pm on May 22 in riot torn Meerut, loaded into a truck and driven to Jam, on the banks of the Gangnahar. Shortly, after nightfall, all 29 were lined up, shot and thrown into the swift flowing waters.
Mujib, shot in the right lung from the front and Babu ud Din, shot in the left lung from the back, were chanced upon by the Muradnagar police at around 4 am on May 23, about 10 km downstream from Jam. Around 6 am, the blood-spattered men were admitted to the hospital, at least 10 hours after being shot.
The two survivors' injuries indicate that the shots fired upon them were well-aimed and passed clean through their lungs. The weapon was reportedly a small bore handgun.
Initially, the two were guarded by the Uttar Pradesh police. When word of their being alive spread, the U.P. police were replaced by the CRPF, on May 24 evening, as the Union Home Ministry feared attempts to eliminate them. The panic-stricken Uttar Pradesh Government, desperate to blank out the Jani carnage, also wants to ensure the survivors' silence.
The U.P. Government is also scared that evidence of carnages other than the one at Jani may surface, since the number of bodies fished out of the Gangnahar and the parallel Hindon river are conservatively estimated by the Union Home Ministry at 38.
Given the amply corroborated fact that bodies were seen continuously for six days, starting May 21, it is plausible that there were carnages similar to the one at Jani in Meerut district, aimed at a minority community. The number of victims of such PAC barbarities could be substantially higher than the conservative Union Home Ministry count of 38.
Till May 26, the State Government refused to admit that bodies had been sighted in the two waterways. But surprisingly, a CID inquiry into the bodies floating in the Gangnahar and Hindon river was ordered a day earlier, on May 25.
Not having any survivors to depose would definitely help the Chief Minister, Mr Bir Bahadur Singh, and his Home Minister, Mr Gopinath Dixit, to continue in office.
Another survivor of PAC excesses, Mohammad Usman, similarly shot and dumped into the Gangnahar, is in the All-India Institute of Medical Sciences (AIIMS). He, too, is under heavy guard and not permitted to talk to anyone.
At the Narendra Mohan Hospital, the authorities were evidently desperate to keep the identities of the two injured men secret. Even the doctors and nurses attending on them have to enter their names in a special register each time they visit them. These reporters were aggressively turned away by the CRPF men at the door when they attempted to meet Mujib and Babu-ud Din. Requests to photograph the guards were greeted by the snatching of the camera. And the window
overlooking the courtyard outside was hastily closed.
A stream of senior police officers armed with fruits have been paying solicitous visits to the only closely guarded room in Mohan Hospital. The two injured are reported to be on the mend.
Along the Gangnahar, from Muradnagar, where the two injured were found, to Jani, the scene of the crime, there is fear and hostility.
In the villages along this 10 km stretch of canal, residents have been told not to talk about the bodies flowing past for six days till May 26. Since the bodies began appearing on May 21, the inhabitants here have organised vigilant groups to protect themselves. The insecurity was apparent when the escort of a burka clad woman, obviously Muslim, identified himself as a Hindu.
Queries regarding the floating corpses were met either with suspicion or directions to seek information upstream. All, however, seemed to be well aware of the bodies. At the water headworks in Jani, there was open hostility. A clutch of residents belonging to the majority community not only claimed ignorance, but told gory tales of the massacre of Hindus in Meerut.
The village Pradhan, accompanied by henchmen and the waterworks keeper, turned up to prevent any of the locals from giving these reporters information about the bodies. Shooing away villagers eager to tell stories of the body count, he claimed that corpses floating in the canal were "common". "An average of at least 25 floated past every month from Haridwar", he said.
—From Indian Post, 30 May 1987
PAC JAWANS SHOT DEAD 200
Sankarshan Thakur
At least 170 to 200 men belonging to a community who were picked up by the Provincial Armed Constabulary (PAC) from Hashimpura in Meerut on May 22 are feared dead. Most of them were packed off in PAC trucks, driven to the Ganga canal near Muradnagar, shot at point blank range and thrown into the water.
The exact number of those suspected killed is not clear—nobody seems to have taken count of those rounded up by the PAC in Hashimpura and loaded into six trucks—but now that they have been untraced for more than 10 days, it is assumed that they are dead. More than 50 bodies with bullet injuries were recovered from the Ganga canal during the course of last week but survivors have affirmed that the PAC had taken away no less than 170 to 200 old and young men from Hashimpura, separating them from the women and children after the house-to-house searches on May 22. None of them has returned home.
The Janata Party MP, Syed Shahabuddin, who is sheltering Zulfikar Nasir, a 17-year-old survivor of the massacre, wrote a letter to the Prime Minister today expressing his apprehension that close to 200 people had been killed by the PAC in what seems to be the most cold-blooded and gruesome police massacre. He demanded that either the government declare the whereabouts of the missing or admit guilt. He is also filing
a habeas corpus writ in the Supreme Court on the missing men of Hashimpura
Feigning death
Zulfikar Nasir, who was in one of the PAC trucks taken to the Ganga canal but escaped by pretending to be dead after a bullet pierced his right armpit, says he was in the second of the six trucks brought to Hashimpura. There were 30 to 35 men in each truck and when we were driven away, the other trucks were still being loaded. They were taking away all the young and old men of our locality."
The trucks left Meerut around seven in the evening but its occupants had no idea where they were being driven to or why because they had been instructed to squat with their heads down. "The only thing I know is that we were brought down near a canal late in the night. The dozen or so PAC men then discussed something, loaded their rifles and began shooting people one by one." Yaseen, an old man and Ashraf, a young boy of Zulfikar's age, were shot dead in front of him. He himself was third in the queue but had a miraculous escape.
"I did not see the others being killed with my own eyes as I lay, feigning death, in the thicket besides the canal. I could hear the shots ringing out continuously and the sound of the bodies splashing into the canal," says Zulfikar. He managed to trudge to the main Muradnagar road with two other boys—Arif and Kamruddin—who too had lived. They drew some water from a handpump near Muradnagar but the clank of the handpump woke up people in the neighbourhood who said they would call the police. A frightened Arif ran away. Kamruddin, who had injuries nearer the heart, died there and Zulfikar hid in a urinal near the Muradnagar checkpost where he remained till 6 pm the next evening. "When I came out, I was lucky to find some relatives in Muradnagar who took me to a hakim for first aid."
The Ghaziabad district president of the Janata Party, Mr Ansari, then brought him to Mr Shahabuddin in Delhi.
Zulfikar is one of the five known survivors of the massacre. Two, Mujibur Rahman and Babuddin, are struggling for life under heavy CRPF guard at the Narendra Mohan hospital at
Mohan Nagar near Delhi. Another, Mohammed Usman is in the casualty ward of the AIIMS in the capital and the fifth, whose name is not yet known is reported to be in hiding in Ghaziabad.
Mujibur Rahman and Babuddin have been able to file firsthand reports which corroborate Zulfikars' version of how the PAC massacre were conducted: people lined up, shot and dumped into the canal. But obviously, the two were brought in another truck for Zulfikar does not remember having heard of them. Nobody, apart from policemen and officials has been allowed to see the two.
Don't Trouble him
Mohammed Usman, the stitches raw on his stomach which was riddled with bullets, reclines in obvious pain in the surgical ward of AIIMS. A plain clothesman who describes himself as "an official from thana," stand guard, eyeing Usman sternly. Usman is eager to speak about what happened to him but the plainclothesman buts in to say, "He is not in a position to speak. Don't trouble him. At the most you can ask him his address." Usman gets the message and shuts up, obviously terrorised. "I am from Bakre ka Path in Hashimpura. Please inform my parents. Please," he pleads. He looks at the plainclothesman, perhaps to steal a moment and tell his tale but the policemen has been staring hard at him all along. It will take some time yet for the full Hashimpura story to be told.
From Telegraph, 2 June 1987
PART III
WHAT EDITORS WROTE AND READERS OPINED
The continuing communal conflagration in Meerut, just 80 kilometres from the capital, must send alarm bells ringing all over the country. Meerut speaks of the tragedy that is threatening to overwhelm the country. And the explosion of hate comes in the wake of what has been happening in Ahmedabad again and again. Meerut has, of course, not been an oasis of communal harmony. Quite the contrary. The last communal riot took place barely a month ago. The district administration was myopic in withdrawing a major part of the para military contingents called in then. But there is no guarantee that their presence could have prevented the present disaster. The army has now been out there for nearly a week, and yet the killings continue. The official death count has already crossed 75. The true count must be much higher. The holocaust in Meerut is not an isolated development. It has taken place against the background of a grave deterioration in the relations between the majority community and the minority community not only in Meerut but the whole of UP after 1947-48. Both the Hindus and the Muslims have convinced themselves that they are in a siege, they do not need much provocation to get at one another. Apart from long standing cause for communal tensions, the Babri Masjid Ram Janambhoomi controversy has seriously aggravated the sense of insecurity and aggressiveness among both communities.
Meerut had been tense following a stabbing incident, but the conflagration followed a face-to-face encounter between the Muslims and the police in the wake of a bomb explosion. Unwisely perhaps the police decided to launch combing operations in the early hours of the morning when Muslims congregate for the pre-dawn meal ("sahari") in the Ramzan period. Determined to prevent arrests, and angry at the alleged partisan role of the police in last month's riots, the Muslims offered resistance the police had apparently not anticipated. The police force, denuded by last month's withdrawals, was soon on the defensive and could do little as Muslim mobs went on the rampage. Their targets? Mostly the business establishments of the Hindus. The revenge of U.P.'s infamous Provincial Armed Constabulary (PAC) was to come later.
Two more facts need to be noted. This time, unlike in the past, more deaths have been caused by sniper fire than by stabbings. Communal killings spread from the walled city area to the middle class suburbs. In a situation such as the one in Meerut, it is facetious to suggest solutions. All we can do is to limit the damage. For that limited and realistic objective, it is necessary to post good administrators in communally sensitive places, withdraw the PAC immediately from Meerut and other similar places, replace it with Central para-military forces, set up the proposed specially-trained force to deal with communal violence and extend the duration of the ban on religious processions and use of loudspeakers.
—From *Times of India*, 27 May 1987
WHEN WILL WE LEARN?
In just a few days the violence in Meerut, has already claimed over 110 dead. The extent and pattern of killing, of arson, of hatred reflect the ever deepening problems of our state, of our society. Meerut has often witnessed communal flare ups in recent years. But the acts of violence hitherto used to be confined largely to the Walled City. We all assumed that it was easy for rapid elements to set it aflame at the slightest provocation. But this time the communal frenzy that first held familiar parts of the city in its grip spread quickly to the outskirts. New residential colonies, by no means congested, soon became the targets of arsonists and looters. And now, what is even more ominous, the adjoining rural areas, too, are being sucked into this vortex of violence. The most disturbing symptom has been the traumatic experience of the hamlet of Malian, ten kilometres from Meerut city. Its 2,000 inhabitants have had a hard time escaping from the clutches of a platoon of an arm of the state itself, the Provincial Armed Constabulary. The discovery of sixteen bodies in the Hindon Canal in Ghaziabad, Noida and the Delhi suburb of Kalyanpuri is further proof of the widening trail of death and destruction in the wake of the outbreak of violence in Meerut. And as reports of mounting tension in the neighbouring towns of Muzaffarnagar, Modinagar and Hapur come in, no one knows when the authorities will manage to restore a semblance of order. It is not that this time rood or in general the killings
and riots are "pre-planned". But hundreds seem to be perpetually prepared for them with arms, with networks of associates beyond the reach of, in many ways more "efficient" than the machinery of the state.
The things that need to be done have been listed ad nauseum. The police should be reformed and reculturated; people should feel involved in the affairs of the community; we should prevent politicians and priests from misusing places of worship; the ever-widening nexus of criminals across cities should be smashed; the nexus of criminals and politicians, and now priests, must be broken: our leaders must be such that the community listens and responds to them; as one cannot set a policeman to check every person and a policeman to check every policeman, and therefore, as Governments can run only when the people hold them in esteem, our Governments must be legitimate. All this has been listed time and again, but we have almost given up the attempt to attend to these fundamentals. Our horizons have shortened; at each moment we are relieved if only we can get through the latest, the present crisis. And yet these fundamentals are what have to be attended to. And that can be done only by more and better persons devoting themselves to public and community life. How much greater travail will the country have to bear before that happens?
—From Indian Express, 27 May 1987
Meerut town in western UP has been twice engulfed by communal violence within a month. The death toll in the first bout which lasted six days from April 15 was officially put at 0 while other estimates are that the figure was more likely 14. In the latest round of mayhem, which began on May 16 and is hopefully just being brought under control as we go to press, the number of deaths is estimated to have exceeded 200 already though, true to form, the government's count is still less than half that figure. Communal tension and violence in UP have been growing since the unlocking of the Bahrí masjid/Ram Janamhoomi temple in Ayodhya in January last year. More than a dozen episodes of communal violence have occurred this period in different parts of the state and a number of new communal organisations, both Hindu and Muslim, have emerged and gathered strength. The Bahrí Masjid Action Committee organised its massive rally of Muslims on the Boat Club lawns in New Delhi on March 30 to demand that possession of the place of worship in Ayodhya be handed over to the Muslims. Just the previous day the Vishwa Hindú Parishad had called a UP wide bandh by Hindus which it followed up with a mass pledge taking rally (sankalpa samaroh) in Ayodhya on April 5 to "protect" Ram Janambboomí and erect a temple there at a cost of Rs 25 crore.
It would be natural, therefore, to see the communal holocaust in Meerut against the above background of mounting communal tensions. However, what comes through most strikingly in the reports of the tragic events of the last few weeks
in Meerut is not so much their undoubted connection with the larger communal tensions as the part played in them by the administration and the police forces. Most infamous, of course, have been the doings of the notorious Provincial Armed Constabulary (PAC). On May 23, in the most brutal episode of the Meerut riots, personnel of the PAC—according to one report, inebriated after consuming country liquor earlier looted from a shop—launched a murderous attack on the Muslim section of Malana village in the vicinity of Meerut.
Huts were set on fire with entire families inside them and people fleeing for their lives were indiscriminately fired upon. This carnage is said to have lasted about three hours. Many of the dead bodies were apparently dumped into the Ganga canal and have been spotted floating in the canal at different places, some near Delhi. In Malana certainly there had been no pre-existing communal enmity and the violence was entirely the handiwork of the PAC. Malana was not however, the PAC's only contribution to the Meerut mayhem. Other reports speak "PAC jawans joined the mobs in looting and burning shops. When shopowners tried to protect their establishments, the PAC even provided fire cover to the arsonists." The PAC carried out house-to-house searches. "There were complaints that it was going about them in a most brutal manner, beating up men, abusing and molesting women."
The PAC's past record in regard to communal riots—in Aligarh and Moradabad, for instance—should have prepared everyone, including the government, for its bloody performance in Meerut. And yet as many as 30 companies of the PAC were deployed in Meerut and it is on this force that the government has relied the most for containing the communal violence. So the culpability is not of just the PAC but of the administration as a whole. In any case, the other branches of the administration have been scarcely more exemplary in their non-partisanship or in their earnestness in controlling the violence. (Incidentally, a former director general of police of the state is now president of the Vishwa Hindu Parishad, eastern region.) What for instance, is one to make of the provocation explicit in the despatch of the police to make preventive arrests from a Muslim locality of the town at around 3 o'clock in the morning at the
time of 'sabari' when during Ramzan Muslims have their meals before starting the next day's fast? And after the April riots in Meerut, the UP home minister lost no time in declaring, pointing by implication to the Muslims, that the disturbances had been carefully planned. More recently, through a widely published announcement the state government banned the entry of eight well known Muslim leaders into Meerut. The insinuation is once again plain enough, just as it is in the speculation, attributed to the Union Home Ministry, about the role of Pakistani agents in the Meerut riots.
The relevance of these facts extends beyond the apportioning of blame for the killings and violence in Meerut. Their significance really lies in the questions they raise about the hitherto much-vaunted secular credentials of the present Indian state which has to be a central issue in any consideration of the phenomenon of communalism and growing communal tensions in our society.
—From *Economic and Political Weekly*, 30 May 1987
A DEADLY CONFRONTATION
Aligarh in 1978, Jamshedpur in 1979, Moradabad in 1980, Hyderabad in 1981, Meerut in 1982, Bhiwandi in 1984, Ahmedabad in 1985, now Meerut again. And the canker is spreading. All of Uttar Pradesh has been a communal powder keg these last several months. Gujarat has been transformed from a peaceful industrious state, to a place where communalism has taken deep and seemingly permanent root. Punjab has a problem of a different sort, fundamentalism has raised its head in Kashmir, unlikely states like Kerala see communal forces test their strength at the polls in one state after another.
It is not a new problem—the first recorded communal riot in the country dates back to 1714—and Meerut is not the worst that the country has seen. As many as 1,500 died, for instance, in the 1969 riots in Ahmedabad, and even more in the anti-Sikh riots in Delhi in 1984. But never before has the country witnessed such a prolonged period of tension of this nature, nor such a dangerous and rapid spread of the communal virus. Religious places have become major foci of dispute, and religious issues have come to the forefront in several ways (the Muslim Women Bill, for instance). The 'Hindu backlash' has been a political talking point for three years and more. Large rallies have been organised by bodies associated with one community or another. As the tension mounted and was sustained, a Meerut almost had to happen. And if temperatures aren't lowered, it will almost certainly be another town.
tomorrow
Several questions now need to be asked, and answered. What should be done to reverse the surging tide of communalism before it swamps the country? Why did the administration fail in Meerut, although notice had been served (so to speak) with a smaller riot a month earlier? What are emerging contours of the problem?
It was easy to argue, before Independence, that the occasional communal riots were the result of the British policy of divide and rule. Now the reality is known to be more complex, taking in both the communalising of politics and the politicisation of religion, the jostling by different communities for a better share of the cake, the tensions created by cases of mass conversions, and the emergence in the forefront of organisations and leaders who speak only for their community. And a country that prides itself on the secular character of its politics has no effective answer to the menace.
Certainly, it is the mishandling of issues by the Government that has partly resulted in today's situation. The Ramjanambhoomi Bajri Masjid issue should not have been allowed to fester in the manner that it has, and the Muslim women's bill brought about a further divide, with a reliance on fundamentalist opinion as opposed to more forward looking schools of thought. For good measure, secular political parties have not hesitated to team up with communal counterparts, lending them a certain respectability. And although it was obvious even a year ago that tension was building up, little was done to defuse the situation.
Some of the old responses to the problem are now obviously inadequate. Tired bodies like the National Integration Council (though well-intended) have proved to be ineffective, and traditional strategies like peace marches have fallen into disuse. Periodic reviews at the highest levels have resulted in decisions to draw up lists of sensitive places, to form special peace squads and so on. All of this, quite apparently, has had little effect. Even in Meerut, a town with a history of violence of this kind and a riot as recently as a month ago, the administration withdrew police forces too quickly, a mistake that was repeated in Delhi.
A solution must of course include gearing the administration to be more sensitive to such flashpoints, and to take advance action. But corrective measures must go well beyond this. Culprits who indulge in rioting have so far been getting away scot-free, despite the ritual appointing of commissions of inquiry. This record in enforcing the law must change. Religious processions (which often provide the flashpoint for riots) cannot be banned in all cases, but they should be discouraged. Public address systems should be prohibited in places of worship, and the encroachment on public land for building places of worship prevented. The police itself is often affected by the communal virus, and the only effective solution is to have a proper mix of communities in the recruitment of policemen. In the field of education, textbooks and syllabi must be carefully screened to weed out any overlap of communalism. And in politics, the Communist Party of India (Marxist) has shown the way in Kerala by refusing to have truck with any of the communal parties, and other should do the same. They also must choose candidates for elections carefully so as to pack Parliament and the assemblies with people who have a secular outlook.
The initiative for all this must come from the top. And the leadership must be seen to be serious, not issue empty threats as Rajiv Gandhi did on Meerut, only for the riots to continue for five more days. Chief Minister like Bir Bahadur Singh must accept the responsibility of failure. The country needs to be reassured that the Government is serious about tackling the communal curse.
—From India Today, 15 June 1987
Every minority has its problems, and India's Muslims have more than their legitimate share, thanks to the many tragedies which have marked their history. But it is, by and large, not a community which allows its problems to depress it beyond a certain level—they may have not one but a hundred reasons for anger, and they may have not one but a hundred complaints about economic discrimination, but they still have resources within them which prompt a smile as easily as a frown. For 40 years, since partition, they have suffered from suspicion, and a wide range of its consequences. They have learnt to live with the accusation that they are traitors or (more fashionable these days) agents of foreign interests. But no matter what their dilemmas and frustrations, there has been one day of the year on which they have chosen to forget their problems and enjoy what should indeed be considered a normal requisite of existence—happiness and good cheer. This is the day of Id, marking the end of the holy month of prayer and fasting, Ramzan. Id is a day which every Muslim, no matter how far his intellect or his environment may have taken him away from his religious roots, keeps aside for the enjoyment of brotherhood and love which are such important elements of the faith. This year, even the happiness of Id has been snatched away from Muslims. As city after city of the north broods under the weight of curfew, and as memories of yet another brutal and arbitrary massacre by the hooligans of the Provincial Armed Constabulary turn the heart bitter, Id loses all its traditional flavour of gaiety and joy. It is not the thought of just riots—bad enough as that may be—that destroys the Muslim's confidence; no one is so romantic as to believe that in a country like ours social tensions will not erupt into violence. After all, there are enough mischief-makers on both
sides of the communal divide. It is more the thought of the completely inhuman irresponsibility and brutality of the guardians of law and order that shakes the community. Nor is this the first time that the thoroughly notorious PAC of Uttar Pradesh has behaved in this manner. It has a justified reputation of being a rogue force. The problem is that no government, neither at the state nor at the central level, has ever been courageous enough to take action against a force which not only is criminal in behaviour but so utterly harmful to the national interest and unity. Why? What great constituency or interest does this PAC represent? Why are politicians so afraid of this body of scoundrels in uniform? We learn now that the superintendent of the force which committed such crimes in Maliana has been suspended; is that sufficient? The PAC is being accused of nothing less heinous than deliberate murder. Is one officer's suspension all that will be done?
Home Minister Buta Singh has now virtually admitted to the Opposition leaders he met that the Uttar Pradesh government ignored the warning that the Centre had given to the state about impending communal riots. Will the chief minister of the most populous state in India be allowed to get away without an answer to this charge of sheer irresponsibility? New Delhi cannot afford to become a victim of a paralysis of will: it must do what it has been elected to do—rule. Silence is not a solution in this the gravest of times. The Prime Minister does not have a magic wand by which he can make all communal riots disappear, but he does have the authority, vested in him by the people, to administer the land and protect the lives of its citizens. There is no more time really for the luxuries of longwinded commissions: what we need is the firm hand of administrative justice which can restore the confidence of the people in the rule of law, and make the potential culprit apprehensive and wary and afraid of the consequences of his mischief. When a state government has failed, the Centre must step in to protect the weak. The time has come for Delhi to tell Lucknow that this misrule cannot go on.
—From Telegraph, 29 May 1987
WARNING OF MEERUT
The DRIVE TO DESTABILISE INDIA is not any nearer its end with the relaxation in tension in the relations between the President and the Prime Minister following reports that the President, though still unwilling to accept the Cabinet's interpretation of the constitutional provision regarding his right to be informed and the Prime Minister's duty to keep him informed, would rather leave the issue for his successor than take a precipitate action.
For one thing, even now it is not clear whether the reports emanating from Rashtrapati Bhavan, which have induced a feeling of relaxation, should be taken to mean what they seek to convey, or they represent a change of tactics. If uncertainty persists, it is because of the apprehension that Giani might strike once again and take his adversary by surprise. The Prime Minister's side is, therefore, not lowering its guard. And rightly so.
For another, destabilisation operates at many levels and is not a uni-dimensional operation. Differences between the head of state and the head of government were being manipulated by the Hershmans and the Swedish Radio operators from far away. But on the ground, here in this country, communal conflicts have been so manipulated that today the entire fabric of our society is faced with the danger of being torn asunder. Developments in Fiji show how operations at these two levels can converge to oust a government within a month of its election.
FIJI is a small country. Ours is much bigger, and our society is far more complex with deep-rooted internal tensions and far greater variety of diversities. And yet, our freedom
struggle was disrupted, our country partitioned, and an imperialist base built on Indian soil in the shape of Pakistan.
Our Independence itself was bedevilled with destabilisation. Destabilisation works most effectively, even if in the manner of slow poisoning, at the level of our thinking as individuals, as social groups and as a nation-to-be-making, and it speaks volumes of our shameful lack of vigilance that despite the traumatic experience of the Partition we are again faced with the danger of Khalistan, and worse, almost a civil war on communal lines in parts of the country—in the UP and Gujarat, in particular.
It would be wrong to dismiss what is happening in Meerut as a local affair or to explain the communal riots in Old Delhi as an extension of the Meerut disturbances. As in Gujarat, where the communal virus has permeated into all parts of the state, in UP communal violence has erupted in one town after another during the last one year.
Whatever the reasons for the simmering communal volcano that Gujarat has become, in the UP it all started with the Ram-Janaki Yatras organised by the Vishwa Hindu Parishad, which is led by the RSS and patronised by opposition (Lok Dal and Janata) as well as Congress politicians. Wherever the yatras spent the night, whether in a big village or a small town, Nairang Dals were set up consisting of professional bullies and unemployed peasant youth. In no time, rival Adam Senas of the Muslims emerged. The small sector illegal arms industry has since then been in a state of boom which knows no end.
Both in the case of Meerut and in that of Delhi, the authorities are unjustified in claiming that they were taken by surprise. Even a casual passerby through the lanes and by-lanes of the walled city could sense not only the tension in the air but also the preparations going on for a confrontation. This is also true of Meerut.
To both places, what we are witnessing is organised violence, and if UP authorities do not wake up in time there will be a Meerut in every big and small town of the state. That precisely is what destabilisation is aimed at.
—From Blitz, 30 May 1987
ENFORCE THE LAW
The frequent and heated debates on communalism have been of little avail in combating the evil. This has been one of the more dispiriting aspects of the problem. There have been prescriptions galore. The Government apparatus in turn has responded to each wave of violence by demanding more powers. The fact is that the state is already armed with adequate powers to put down communal miscreants with a strong hand. It just does not use them. The Haksar Committee of the National Integration Council has brought this out forcefully. After noting how religious establishments have flouted various existing laws on public peace and order with impunity, the Committee has aptly observed "Even a token enforcement of one single issue like this would bring a sense of credibility to the debates of the National Integration Council." The Central Government itself has still not realised the importance of concentrating on effective enforcement of the relevant laws instead of organising meetings and seminars on communal harmony. Otherwise, it would have taken good care to give the widest publicity to the report of the Haksar Committee of the National Integration Council. Characteristically enough, an official spokesman was more interested in publicising the speeches of the members of the Standing Committee of the NIC at its meeting on Tuesday. Indeed, the very fact that the Prime Minister, the Home Minister, and the other eminent leaders who attended the meeting failed to highlight the report of the Committee and, instead, were content to repeat the worn out cliches about strengthening the secular fabric etc shows that they remain stuck as if in a
As the Committee has pointed out, Section 153A of the Indian Penal Code empowers the Government to punish a person who promotes feelings of enmity, hatred and ill-will between religious groups or castes. This Section also deals with misuse of places of worship. Action can also be taken under this Section against organisers of and participants in processions if any of them by word or signs attempts to create disharmony or a feeling of enmity among various communities. Yet, all these offences continue to be committed with sickening regularity as though there is nothing like a Government in existence. More often than not it is because Ministers and party leaders find it politically rewarding to shield the religious leaders who violate the laws that the police can take neither preventive nor punitive action. Again, Ministers have obliged some religious leaders with all kinds of favours including lucrative business arrangements. This is yet another reason why the religious leaders think nothing of indulging even in such offences as asking the faithful to encroach upon public land in the name of religion. It is because of the political patronage enjoyed by these law-breakers that the police force, too, finds it safe to play at best a passive role. To make matter worse, as has been revealed particularly in Gujarat, the police have developed close links with the leaders of the underworld who find it profitable to kindle communal disturbances on the slightest pretext.
The Haksar Committee's Report ought to be placed before Parliament and the State Legislatures. Extensive public discussion on its contents will create an atmosphere in which the allies of the diehard communalists entrenched with and outside the Government will find it difficult to treat the laws on the maintenance of public order as though they were mere scraps of paper. Simultaneously, the Prime Minister and the Chief Ministers should ensure that neither they nor their Cabinet colleagues attend religious functions as guests, chief or ordinary. They must also ask the official media not to publicise them. Last but not least, there should be no hesitation in taking firm and prompt action under the law against newspapers that are found to be wilfully fanning communal hatred.
From *Indian Express*, June 25 1987
HIDING THE CRIME OF MEERUT
All good things, as we have learnt, must come to an end. So, hopefully, must also all bad things. The Chief Ministership of Mr Vir Bahadur Singh too, consequently, will come to an end some day. But if Delhi has any instinct for self-preservation then this will happen sooner rather than later. The damage that he is doing to the party and to Prime Minister Rajiv Gandhi is so enormous that not checked at once it will be impossible to retrieve the situation after a while. If people like Vir Bahadur Singh are going to be the guardians of the Congress interest till the next elections then the party should be prepared for a Haryana style wipeout. He began his term with an error whose consequences have seared the nation badly—the reopening of the Babri Masjid. The price the country has had to pay in terms of communal tension for this one bit of mischief should have been sufficient to drive anyone from office. But he remains. And his last bit of idiocy will go down as one of the most inexcusable ever. The chief minister of Uttar Pradesh had no business stopping Mr VP Singh from addressing his rally at Gorakhpur. It is irrelevant whether Mr VB Singh agrees with the former finance minister or not, it is even more beside the point that he might be wanting to buy an extra lease of life through sycophancy. He simply has no business to try and fetter any political voice. This is a free nation, and everyone has the right to be heard. And those who tamper with this right, either through manipulation of the
law or the manipulation of administration, do so at huge risk to themselves.
But of all the acts of omission and commission which taint the present government in Lucknow nothing is more heinous than the enormous cover-up operation that is going on to hide the guilt of Meerut. What happened in Maliana and Hashimpura should be a lesson in shame and humiliation. Any decent chief minister would not have waited to be asked to resign, but would have gone voluntarily. But then we cannot expect any standards from the present incumbent. If, however, that was bad, what is going on now is far, far worse. Stories are being planted that the "so-called" dead are "returning" to Maliana (incidentally, no one is even trying to suggest the same for Hashimpura, because the extent of the brutal killings from this area are really beyond imagination). And advertisements, which can only be called cruel, are being placed in newspapers to express "concern" over the welfare of minorities. The government of Uttar Pradesh is trotting out lies in a desperate attempt to save its skin. Worse, a few journalists are participating in this exercise. But while we cannot expect any better from Lucknow, surely Delhi should step in if it has some respect for the norms of civilised existence left. The exigencies of the present mood, and the impending presidential elections may hold Mr. Rajiv Gandhi's hand at the moment, but the restraints should get over after the presidential elections. To allow Mr. Vir Bahadur Singh to get away with this would be to condone fascism. And the price for that will be far higher than anyone can assess at the moment.
—From *The Telegraph*, 20 June 1987
MEERUT MASSACRE—A FORE-TASTE OF FASCISM
Meerut is an extreme case of communal polarisation which encompasses the elite and the masses alike, pervades the state machinery and distorts the political will. Fascism thrives on falsehood. The campaign of disinformation and misinformation has made us deaf and blind, it has paralysed our hands and feet, it has brutalised our hearts and frozen our souls. In the face of bestial and inhuman brutalities, women have been demonstrating in favour of the killers, youth have been shouting 'PAC Zindabad', the administration has proclaimed its pride in the PAC, the Central Government has been hesitant to order the PAC out, the political parties have by and large taken little or no notice of the atrocities, the conscience has gone to eternal sleep. Does the Nation have a conscience?
This is Genocide
Mass deportation and mass liquidation should remind us of Nazi Germany, they were the strategic objective of German Fascism. When citizens are taken out of their homes at gun point, entire localities are lined up or made to squat with their hands up, when all able bodied males are forced into trucks, when some trucks proceed to secluded areas and their occupants are taken out one by one and systematically shot in cold blood, their bodies flung into a river or a canal, are we far behind Nazi Germany? Yes, the scale varies but the impulse is the same. The Hashimpura youth, the Malana people—all citizens of the Republic of India—were sought to be liquidated.
not because they were criminals or had committed a specific offence but because they professed a particular faith. This is genocide, nothing more, nothing less. This is the negation of the Rule of Law. But the defenders of human brotherhood and rule of law do not open their mouths!
PAC has Political Backing
Could the PAC have committed these atrocities, acted so blatantly, threatened to turn Mahboob into Jalianwala Bagh, openly professed its objective of massacre, without political support? Where does the chain of command lead us? To the Collectorate in Meerut, to the State Secretariat in Lucknow or to the Central Secretariat in New Delhi? Where does the trail of blood end? Let those who, in the darkness of night, oblivious of their oath of office, ordered the PAC to 'teach the bloody rascals a lesson', have the courage to own their responsibility. Let those in high positions who gave clearance to the Gauleiter of Uttar Pradesh speak out. But these ruthless rulers, these small men, will not speak. They have no courage; they will launch a campaign of disinformation, browbeat the recalcitrant officials; intimidate the victims, destroy evidence and conceal the crime. They may even shed a few tears, parade their benevolence, make some cosmetic changes but behind the facade, this band of killers pat each other on the back and plan yet another holocaust.
One day history will catch up with them and tear off the mask of hypocrisy they habitually wear. When the book of their plots and conspiracies shall be widely read and when all of us will have to answer the question, what were you doing then?
Alibis, Excuses, Lies
Is there any limit to self-delusion? The Muslims rose in revolt. Well, then why did they choose the last days of Ramzan? To spoil their Id! The Muslims planned simultaneous attack on 19 May morning. Why did the thickly populated Muslim areas of Meerut remain quiet? The Muslims launched a premeditated attack on the Hindus. Why are then four Hindu families in Hashumpura still safe? But one may ask why did the police force its way into Imliyan in the early hours
of the morning when, because of Ramzan, the entire mobhalla is awake and crush a child under its jeep, kill woman and arrest hundreds?
There is a law to regulate search and detention. Which law sanctions police firing without magisterial order? Which law sanctions beatings and humiliations in police custody? Which law sanctions putting dwellings to torch and looting them? Which law sanctions police firing without warning before lathi-charges or tear gas, to scatter a rambous mob? They speak of snipers. How many policemen were killed or even injured in the "operations"? They speak of communal violence. Yes, but that is a very limited view. There was the Muslim mob on Hapur Road and the Hindu mob in Shastri Nagar. Where else? PAC did not invade Hasbimpura or Maliyan, did not burn and loot Sadar Bazar and Kutchery Road shops and Mian Mohammad Nagar huts to save Hindu lives or property. PAC invaded those Muslim localities and laid them desolate to strike terror in the Muslim community. They behaved as if they were in a war zone and dealing with an enemy. The PAC acted with the felt blessings of their masters. It was not a wayward force in mutiny. It was a disciplined force carrying out orders.
Spontaneous Muslim Reaction to PAC Raid
Yes, the unprovoked, inexplicable early morning raid on Maliyan (where no Hindu lives) was followed by the formation of a Muslim mob from neighbouring colonies, after they not only heard the SOS call but also the firing. They could do nothing to save the people of Imliyan. In their anger they burnt down vehicles, attacked shops and showrooms, damaged factories, big and small in Hapur Road and in Pilokhri. They also killed some innocent persons including Prabhat, the son of that great fighter for communal harmony, Prof Harpal Singh. Within hours, there was a massive Hindu retaliation fuelled by rumours spread by the Administration then about the State Bank Colony being ravaged by Muslim mobs. The raging fire then consumed all decencies, all moral restraint, all Shastri Nagar. The anti-social elements were herded by the police and deployed against Muslim shops and houses in a systematic campaign, a helping hand and providing fire cover, as an
effective response to any resistance.
I do not justify the acts of the Muslim mob in Hapur Road and Pilokhri. I consider such acts not only morally unjustifiable but also stupid and counterproductive but I can see the how and why of it and I can say that it was a spontaneous reaction. And there is no other evidence of depredation by a Muslim mob thereafter anywhere in Meerut; there might have been sporadic acts of violence. Had the PAC acted promptly in Hapur Road and killed some rioters that dark morning of 19 May, they would have been justified. But how does one justify their crimes in Muslim Mohallas from 25th onwards?
State violence not just communal
Meerut is the only marginally a case of communal violence; it is a case of state violence against a people. Some sections of the Hindu community may have applauded, some may have rejoiced, some may have felt satisfied, some may have been misled, some may have cooperated, some may have given moral support but few Hindus actually participated in the killing. And many saved their neighbours.
Truth - the only antidote
This is the only silver lining in the dark cloud. If only the people assert themselves, if only they know the truth, if only they know that contrary to common belief, 90% of the victims are Muslims, that 90% of the over 3000 arrested are Muslims, that 90% of the dwelling units or shops burnt and looted are Muslim, they would be forced to change their false impression created by official propaganda. Had Mahana not broken into headlines, had the Ganga Canal floated away the dead bodies, the crime would have gone unnoticed. All visiting dignitaries were shown the devastated Hapur Road where all shops were depicted as Hindu property; they were told of the Muslim attack, without being told of the PAC attack on Imliyan and of a woman killed and of the child crushed under a jeep. They were not told that not one Hindu shrine or place of worship has been touched in Meerut; while at least a dozen or more mosques and shrines have been destroyed or damaged, that Muslims have been killed by police bullets in Muslim mohallas.
or beaten to death in police or judicial custody. No Hindu has fallen to police bullets in Muslim mohallas, no Muslim in a Hindu mohalla, why?
Terrorism or Social Peace
Yes, there is anger, there is bitterness, there is desperation in the Muslim mind, this anger, bitterness and desperation can easily ignite the fire of revenge and retaliation. That may be a natural reaction but would that be wise reaction? Violence must be contained, violence must be stopped, the area of violence must not be widened, the cycle of violence must not be prolonged. That is fine but what if violence is adopted by the State as a policy? What, if Hashimpura and Mahana are sought to be re-enacted all over the country? What, if Hindu fascism descends on our benighted land? What, if racial memories are systematically revived to burst into volcanic fury? In despair and helplessness, does a human being care whether he lives or dies? Terrorism is born out of oppression, when there is no court of appeal, no mechanism of redress. Will the cruel regime rethink its strategy? Does it want war or peace? Does it want a civil war? Does it want the country to be dotted with forts and fortresses in constant conflict with each other? Does it understand the meaning of terrorism on a national scale?
India: A Moral Wilderness
Meerut is not only a test of our secular profession but of our human sensitivity. History will record that in a moment of trial no Gandhi undertook a fast unto death, no Nehru ran into a mob, no JP jumped into the fire, no Gandhian, no liberal, no democrat, no humanist, no leftist, made it a question of conscience or even of political priority. Brutalised and power-hungry, our generation has let down the spirit of India, the land of Mahavira, Buddha and Gandhi, of Chishti, Kabir and Nanak. Our land has turned into a moral wilderness, the landscape littered with the charred bones of the innocent.
Our secular intellectuals are unfortunately looking at the problem through tinted glasses and from the wrong end. Hindu communalism, parading as nationalism is on the offensive, Muslim communalism there is, but on the defensive. To seek
social peace through economic or educational development of the Muslim, even if sincere, indicates a wrong and myopic diagnosis. Those who equate the killer and the killed are morally blind and politically motivated. They say that Meerut is a sequel to Muslim protest over conversion of Babri Masjid into a temple. So the Muslim should surrender his mosque to save his life. In free India, is it a fair choice?
Message of Meerut
What should be the message of Meerut, of Maliana and Hashimpura for Indians, for Muslim Indians and for Hindu Indians, for the Government, for the political parties? That it shall not happen again. But I see no sign of repentence, of remorse. This is worse than the original sin, more deadly than the crime. Then no power on earth, not all our guns and missiles, our tanks and submarines, can save the country. And who lives if India dies?
To save India, a reaffirmation of the basic values of our Freedom Movement and of our Constitution, is needed. The creation of a truly secular state and the establishment of a non-partisan administrative machinery, which have the will and the means to curb Hindu Fascism—the biggest threat to the integrity of the country today—and protect the minorities and the weaker sections from its constant onslaught. But the ruling elites preoccupied as they are with projecting a Hindu face for electoral reasons, are unfortunately only fond of mouthing professions of secularism. The Question that Meerut poses is: Do they really mean them? The Challenge that Meerut constitutes is; Will they rise above short-term gains?
—From Muslim India, July 1987
ONE MORE EXAMPLE OF GUILTY ACT
One more incident of criminal neglect of the ill reputed administration and police of Meerut. In Meerut in Chandralok locality 23 persons in a house are attacked on 31 May night and out of them 20 persons are burnt alive and the three which somehow survive go to the Police Station for lodging complaint next morning of this barbaric act of murder and arson, the police which itself is part of the crime not only refuse to take down the complaint but detain them in the lock-up for two days and then give them "grand amount" of Rs 15 and advise them to go to Muzaffarnagar at their relatives' house.
When the Police was not ready even to note down the complaint what these hapless people struck with calamity could have done. Obviously all the three (Anis Ahmad, Mehbub and Mumina) went to Muzaffarnagar. When the condition returned to a bit normalcy, they returned and once again tried to lodge complaint against the murderers but this time also they failed. They knew the murderers of their 20 family members because they were none other than their own neighbours. But the Police perhaps did not want that its office should become storehouse of criminal record in Meerut and it should become for it difficult to own the responsibility. Perhaps the filth of communalism has penetrated their minds and they might be avoiding arrest of members of the majority community. Otherwise it is difficult to explain such a conduct on their part.
The Indian Muslims do not raise hue and cry without any reason that in the communally sensitive areas and districts their life and property are not secured. From Moradabad to Mahana the administration and Police and the PAC's criminal neglect and inefficiency and their turning away the eyes from
crimes and more often their own participation have become so common and the hands which were given guns to protect the lives of weaker sections and minorities, if those very hands are smeared with blood of the innocents, then not only Muslims but all secular minded and justice loving people protest and demand that the police be representatives of all sections of people.
Now it is no longer a secret that as far as the Meerut riots are concerned the character of the Chief Minister of U.P. Veer Bahadur Singh and his Home Minister Gopinath Dixit is not above board and these two should not continue to occupy for a minute these high posts in secular India (if India has at all remained secular until now). But one does not know what are the compulsions of the Rajiv Government at the Centre that it is being compelled to tolerate this notoriety. But the Rajiv Government will have to think when the Chief Minister of U.P and its Home Minister have committed such crimes that they cannot be exonerated then why should they not be removed and why it does not accept the demand for dissolving the PAC for ever? Now not Vir Bahadur Singh but Rajiv Gandhi will be answerable for this.
When Rajiv Gandhi was introducing the Russian people to the Indian culture, art and artists in Moscow through Festival of India that time many parts of Meerut were again burning due to communal riots and the greater part of the city had come under the shadow of curfew. Before this we were given good news that the Rath Yatra has passed off peacefully but within three hours of this that little peace which was established with great difficulty also vanished in thin air...
The Muslims and minorities are quite restless and this restlessness can go only when political changes take place in U.P. but also the administrative machinery is thoroughly overhauled and the home ministry and all its departments are not purged of all shortcomings. Earlier Rajiv and his party does this, better it is for them because tomorrow may be decisive not only for Meerut or U.P. but for the destiny of whole of the country.
Translated from Editorial Board of Urdu Times
Bombay, 4 July 1987.
IS CHIEF MINISTER OF UP A DEMOCRAT?
The Chief Minister of UP Mr Vir Bahadur Singh has in order to hide his sins and restoring to outright fraud has caused Mr Agarwal, the Director General of Police to retire. By retiring him he wants to show that we are quite democratic as we have suspended the Director General of Police in the matter of Meerut and Maliana.
(1) Whereas the reality is that he was already due for retirement.
(2) Even if the time was not over and he was removed for his responsibility in Meerut riots we do not want to go into this aspect responsible he is, but not only the Director General of Police, responsible is also the Commandant of PAC under whose supervision the PAC and the Police functioned in Meerut and at whose instance innocent Muslims were martyred by the official bullets.
First of all they (i.e. the PAC) deserve punishment and punishment not the one meted out to the Police Chief Mr Agarwal but punishment according to the Constitution and Law of the country. They should be suspended and tried under section 302.
Justice demands that those who violated the law, are guilty of dereliction of duty and massacring innocent Muslims under incitement of communalism be hanged, nothing less would do.
If our Government and Vir Bahadur Singh are sincere, they should not hesitate to discharge their duty. (Mere suspension would not do) as if they are suspended today, they can be promoted tomorrow. This is the old game the administration has been playing.
Is it not a joke that if some communalist belonging to the majority community is killed, the Government goes mad, the national press goes astray, but (even) if thousands belonging to the minority are massacred, no one is disturbed. What kind of democracy is this? What kind of Constitution and Law is this? It is something for the majority and something else for minority. This dualism should end and the Government must fulfill its responsibility in a proper manner. If Vir Bahadur Singh wants to hide his crime he must, without any inquiry, put to trial the PAC and the Police which was on duty in Meerut and Mahana under section 395-97 and 302 and also they should be relieved of duty and sent to prison. Nothing less than this would inspire confidence among Muslims nor would they tolerate anything else.
—From Faisal (Jadid), Delhi, 11 June 1987
Communal Riots
Sir,—What we have witnessed in Meerut, Gujarat and Bihar is nothing but the ugly face of religious lunacy. These so-called religions, on whose altar thousands of innocents are sacrificed, are paradoxically professing and propagating the message of love and not hatred. An analytical and impartial approach would show that the so-called communal riots are the result of various other factors which have no connection with any religious faith.
In India one's communal feelings can very easily be instigated and provoked. The victims are normally innocent and gullible people who happen to belong to a particular religion or caste. A glaring example of this is the riot in Broach (Gujarat).
There it started with the Prohibition Squad's surprise raids on well-organised liquor dens, on whose payroll the local police force and anti-social elements are included. And the immediate consequence was a riot with a communal colour. The driving force behind this riot was, of course, the underworld kings. The nexus between these underworld kings and politicians is well-known.
In this situation the need is to educate the masses and enlighten them on the modus operandi of these anti-social elements who have been taking undue advantage all through these years. Some of the following steps could be considered by the concerned authorities.
To identify the pockets where frequent communal riots occur and find out the root cause of the same.
To form a citizens' cell consisting of people of high esteem from various walks of life to educate the masses.
To eliminate the nexus between law-enforcing authorities and anti-social elements at all levels and effect periodical transfers of police officials.
To set up a special police force exclusively trained for effectively tackling communal riots. While deploying these forces in riot-affected areas, care should be taken not to include policemen of that area. This will ensure proper and impartial exercise of their duties. As soon as normalcy is attained, these forces must be withdrawn.
All religious leaders should abstain from making inflammatory statements which would provoke a particular religion or caste.
Religious leaders should propagate virtues like peaceful co-existence, tolerance, love, amity and cooperation among their followers.
The above steps, I am sure, can create a congenial atmosphere where the sacred teachings of the Geeta, Quran, Bible and other religious books will flourish and thus pave the way for a new India, free from communal riots.
—A.G. Augustine
Communal Riots
Sir,—It is clear that all secular-minded Indians are gravely concerned about the recent happenings in Meerut, Delhi and elsewhere. However, such riots will continue to occur and may soon become as common place as any other piece of statistics—so many dead, so many arrested, so much loss to property—unless the opinion leaders and the government show the necessary will and determination to do the following things:
1. Stop canvassing for votes on the basis of religion and caste—no more Muslim votes, Jat votes, Harijan votes etc. for election to any office.
2. Ban all communal parties immediately.
3. Ban all religious processions—in a secular society religion should be confined within the home or place of worship.
4. Arrest and detain under N.S.A. all communal leaders
who provoke and incite their communities against others
5 Create Anti Riot squads comprising specially trained personnel from all communities, who will be able to quickly defuse any communal build-up
6 Act quickly upon the recommendations made by the various enquiry commissions, such as Justice Madon, Justice Saxena, etc
Finally, one can only hope for better all-round education through which understanding, peace and harmony will result
Bharati Jaffrey
Hameed Jaffrey
Riot Remedy
Sir,—Any conflict has two perpetrators and on each side invariably, there are faults which cause and worsen conflicts.
If one party withdraws, the other will have none to fight against and the conflict will disappear. On the other hand, if each side waits for peace initiatives to come from the other, before making its own conciliatory moves, then mutual mistrust grows and the inevitable result is escalation.
Hindu-Muslim communal riots are examples of such conflicts. They can be ended only by unilateral action. There are examples in the life of the Prophet of Islam which show that it is Muslim who must take the initiative.
Rivalry and conflict between Muslims and non-Muslims the world over means that the latter are unable to see Islam in its true light and see it through the tainted vision of their own prejudices. The overriding concern for Muslims should be that the true message of Islam should reach people in all its purity and in an atmosphere conducive to dispassionate consideration. This atmosphere cannot be generated where there is conflict and mistrust. Hence, to end conflicts, Muslims must take unilateral steps for peace, without waiting for the initiative to come from the other side.
This is exactly what the Prophet did at Hudaybiya in 6 A.H. by refusing to be provoked by the Qurash's harassment and ending a 20 year old conflict by accepting all the demands. In doing so, the Prophet defused tensions between Muslims and others. This seemingly capitulatory action, the Quran tells us,
was a "clear victory" for Muslims.
If Muslims are to defuse the bomb of communal riots, they should, like the Prophet, refuse to be provoked, even in the face of grave provocation.
(Maulana) Wahiduddin Khan
Editor-in Chief, Al Risala, New Delhi
An Appeal
Sir,—We are deeply disturbed by what is happening in Delhi, Meerut and other places in UP and Gujarat. This is one of the gravest crises the country has had to face since 1967 and the genocide that has taken place at Malwa by the PAC is one of the worst holocausts. We do not think we are using excessive or sensational language. A judicial enquiry commission has been appointed. At this grave moment, we suggest the following:
(1) We appeal to all religious and political leaders to refrain completely from issuing communally inflammatory statements. This applies particularly to utterances from all places of worship and other platforms.
(2) We appeal to the police to exercise utmost tact, humanity and discretion while searching places where hoarding of arms is suspected.
(3) The pattern of recent riots has shown that small incidents have been blown out of proportion and given a communal colour; it should be the duty of the residents of the locality to meet immediately and defuse tensions.
(4) We appeal to editors, editorial staff and reporters not to put one particular community in the dock, while commenting on or reporting riots.
(5) We appreciate that Ramzan I'd is one of the two holiest feasts of Muslims. However, we would earnestly request Muslim religious leaders to consider whether large assemblies this year for the I'd are in the interests of Muslims themselves. (This applies equally to the religious festivals of other communities in the near future).
(6) The government of Maharashtra, at present, is preoccupied with the octroi dispute. We feel the communal situation in Bombay and in the districts of the state is much
more menacing than the octroi. We request the state government to concentrate on this matter and work out short term and long term measures to resolve tensions and protect the minorities. In particular, we would recommend that it set up a protective machinery which could be activated by the would-be victims of the riots before the riots take shape. This protective machinery should be made accountable for failing to respond to appeals for help.
(7) We would make a general appeal to all genuine secular-minded citizens to come together in a positive manner and set up citizens' organisations geared towards both prevention of a riot and resolution of a riot situation in the course of its occurrence. We feel that we have had enough of post riot reports and analyses.
Jean Kalguiker, Anand Patwardhan,
Sagar Sarhadi, Iqbal Masud
Cure for Communalism
Any conflict has two perpetrators, and there are invariably faults on each side which cause and exacerbate it. It takes two to make a fight. If one party withdraws itself from the conflict, the other will have none to fight against and the conflict will disappear.
If, on the other hand, each party waits for peace initiatives to come from the other before undertaking conciliatory moves of its own, then the mistrust between the two will continue to grow.
Hindu Muslim communal riots, which have become a regular feature of Indian life, are an example of such a conflict. They can only be ended by unilateral action from one side. There are examples in the life of the Prophet of Islam which show that it is the Muslims who should take this initiative. Worldly rivalry and conflict between Muslims and non-Muslims means that the latter view Islam not in its true light, but through the tainted vision of their own prejudice. Muslims are their enemies so they adopt an antagonistic posture towards Islam as well.
This is a situation which should be intolerable to Muslims, whose overriding concern should be for the true message of Islam.
to reach other people in all its purity, and in an atmosphere conducive to objective and dispassionate consideration. Seeing that such an atmosphere cannot be generated where there is conflict and mistrust, they should ensure an end to conflicts with other people; they should take unilateral steps for peace, without waiting for the initiative to come from the other side.
This is exactly what the Prophet of Islam did at Hudaihiya (6 AH). By refusing to be provoked in the face of harassment from the Quraish, and accepting all their demands, he put an end to conflict which had been raging for 20 years. In doing so, he defused the tension which had marked relations between Muslims and their non-Muslim compatriots. The result of his seemingly capitulatory action, as the Quran tells us and history verifies, was a "clear victory" for the Muslims.
If the Muslims are to defuse the sitting bomb of communal riots, as it is their duty to do, they can only do so by following the example of the Prophet, and refusing to be provoked, even in the face of provocation from the other side. Failure to do this can only result in further escalation in a conflict which serves to distort Islam in the eyes of others, especially their adversaries.
Maulana Wahiduddin Khan, Editor-in-chief,
Al-Risala, New Delhi
While in one part of the world the "Butcher of Lyon" is brought to trial, in India the butchers of Hashimpura are roaming scot free. Our leaders undertake padayatras to restore peace and amity in Punjab, yet none is undertaken for the Muslims of Meerut. We express our sympathy for the Tamils in Sri Lanka and then go about killing our fellow citizens in India. Even after four decades of Independence we still behave like savages. We have always dreamt of living in Bapu's India, but at present we are living in Hitler's India.
Faiz Ahmed Faizi, Calcutta
We share the pain and anguish expressed by Mr M.J. Akbar, in "When the so-called guardians turn into communal armies, order crumbles" (June 1). We particularly agree with the comment, "To allow guilty policemen to escape is to condone
a police state." The consequences of such partisanship can be nothing short of disastrous.
Ashis Bardhon, Asok Ghosh Alauddin Ahmed,
Aitghara, 24 Parganas (N)
As a peace loving citizen of India, I am ashamed that despite 40 years of Independence we are still divided. We still kill our brothers in the name of caste, creed and religion. We must do something to stop communal violence before it is too late.
Shibli, Fatmi, Jamshedpur
We seem very concerned over happenings in South Africa and Sri Lanka but care little when Muslims in our own country are butchered and their hearths and homes destroyed by trigger-happy policemen. Our preaching of non violence is nothing but propaganda to hoodwink the world. Our conscience is blunted and hearts steelied, which is why not a year passes without the blood of Muslims being spilled in one part of India or another.
In order to restore the confidence of Muslims, the Vir Bahadur Singh ministry must be dismissed and President's rule imposed in Uttar Pradesh. The PAC must be disbanded and in its place, a new force consisting of secular minded people should be formed in which the minorities must be given adequate representation. Those involved in the Meerut killings should be tried for murder. And the next of kin of each riot victim must be given adequate compensation and suitable arrangements must be made for their employment.
These steps would not bring back the dead, but would wipe the tears from the eyes of the riot victims to some extent.
Dr Sayeeduedin, Calcutta
Communalism cannot be brought under control in Uttar Pradesh so long as a man like V.B. Singh and a force like the PAC continue to be the guardians of law and order in the state. A mere inquiry will not help ease the problem. The Centre should instead take drastic steps to curb communalism.
Md Khalil Calcutta
I would like to express my heartfelt sympathy to the Muslims who have suffered at the hands of the PAC in Meerut M.J. Akbar's report, "When the so called guardians turn into communal armies, order crumbles" (June 1) was hardhitting. It is vitally necessary that the PAC is punished for its inhuman deeds. How can we think of advancing into the 21st century without settling our communal problems? We were better off under the British rule, where we were all treated equally—for we were equal in our sorrows. Our only hope lies in our children, who shall grow up and learn to laugh at our follies.
Abhay Sobti, Calcutta
You have rightly expressed the feelings of the aggrieved Muslim community in your editorial, "A day of sadness" (May 29). I am puzzled as to why the central and state governments have not taken stern action against the culprits.
There is no dissimilarity between the PAC and the Punjab terrorists. In fact, the terror unleashed by the PAC is the worst form of terrorism. The new law against terrorism must be enforced against those involved in looting, carnage and arson in Meerut.
Syed Ejaz Ahmad, Burnpur
Your editorial, "A day of sadness" (May 29), deserves praise. I agree with you that what we need is a firm administrative hand, not any paperwasting inquiry, to quell riots and restore law and order.
The Uttar Pradesh government has proved totally inefficient in containing communal violence. The Centre should take prompt action against the V.B. Singh ministry. Those behind the massacre should not be spared.
Md Masud, Calcutta
Thank you for your timely editorial "A day of sadness" (May 29). It is time that the V.B. Singh ministry in Uttar Pradesh is dismissed and the PAC and those involved in the Meerut and Delhi riots are punished. The government should also provide adequate compensation to those who lost their
relations in the riots and extend loans to those whose property and business were destroyed
Syed Abul Hayat, Burnpur
The Adolf Hitlers of this world do not die easily. Some are reborn in the shape of Mr Botha, a few are reincarnated in the form of President Jayewardene, the lucky ones emerge as Mr Vir Bahadur Singh and the power hungry Syed Shahabuddin. And, of course, the less fortunate ones are recruited by the PAC.
The recent massacre of 200 men belonging to a community in Hashimpura only goes to show that the horrors of Nazism are not a thing of the past after all.
With successors as worthy as Mr V B Singh and Syed Shahabuddin, Adolf Hitler must be really resting in peace.
Abhik Dutta, Calcutta
We have been plagued for a long time by the problem of communal riots. Yet, the greatest culprit, the PAC—the so-called peace keeping force—has always been let off without punishment. In fact, now a days the government seems to support its heinous crimes, like in Malana where PAC jawans indulged in a killing spree.
Ashish Datt, Calcutta
Three Points
Sir,—This refers to Ms Asha Rao’s letter (IE, June 5) titled “One sided advice” which was in response to our “Earnest appeal” (IE, May 30). Ms Rao has specifically asked us to clarify three points in our appeal.
We have been asked why we did not direct our appeal to the Shahi Imam Bukhari who “fired the first salvo”. Our reply is that in a developing situation like a communal riot it is entirely wrong to raise questions of this type.
Without doubt, the Imam should not have made the intemperate statements that he did at the Boat Club rally. However, this rally was held quite some time before the Meerut and Delhi riots, and no press report has traced a nexus between the two events. Tension was building up on both sides and there
was calculated negligence on the part of the police. In this situation instead of appealing to any one individual we appealed to all concerned, including those speaking from places of worship and other platforms.
Basically, our appeal was not a post mortem (which Ms Rao seems to require), but an appeal to all communities and the Government to avoid the recurrence of such riots and specifying remedial measures.
With regard to the second point—police searches—raised by Ms Rao, we submit that Doordarshan in its evening bulletin on May 28 carried a report of seven people having been killed at Maliana during house to house searches made by the PAC. A day earlier, the national dailies (including this paper) carried a report of the massacre of more than 50 people at Maliana, also by the PAC. Our reference was specifically to this incident.
The atrocities committed by the PAC in Maliana have been by now well documented by several journals, including "India Today," and, therefore, our appeal for "tact, humanity and discretion" in these searches is justified.
The third point raised by Ms Rao is the apportioning of blame for the riots. She places the minority community squarely in the dock. We feel this is a side stepping of the basic issue, which is whether there was any political will on the part of the administration in Meerut and Delhi to stop the situation from developing into a large scale riot.
All press reports, without exception, argue that no such resolve was evident, neither was there any evidence of political will on the part of the Central Government to solve the Ram Jaumabboomti Bahri Masjid controversy to the satisfaction of both communities, which is clearly the major cause of communal tension in Meerut, Delhi and other places.
We maintain that in view of this conclusion, to try and blame one community is to play into the hands of communalists.
Almost all riots in post Independent India are the result of collusion between the administration and communal elements, the most recent and obvious case being the anti-Sikh riots of November 1984.
The response to our appeal has been very positive and heartening—judging by the letters received directly by us. With all
due respect to Ms Rao, her letter should not be taken as an indication of the general feeling among more thoughtful and concerned citizens.
One more flash point may be reached when, according to reliable reports, the Rath Yatra will be taken out in early July in Gujarat. We would request the State Government to take every step possible to avoid a recurrence of riots—a blood bath in the case of Gujarat—which are generally sparked off by religious processions.
Jean Kalgutkar, Sagar Saradhu,
Anand Patwardhan, Iqbal Masud
45, Nair Road, Bombay 400008
Painful Silence
Sir—The death of seven people in the police lock up at Meerut is not an isolated case. Between that the police had many other similar achievements, the most infamous being the Bhagalpur blinding case. What was shocking was the timing of the crime (on the eve of Id), not to speak of the painful silence of our leaders whose compassion for the community spills over on the slightest pretext and who leave no stone unturned to champion their cause in the Shah Bano and personal law issues.
I did not find the women's wing of any party voicing sympathy for the mothers who received the bodies of their sons on the eve of the happiest day of the year, with a warning to bury them before dawn. I did not find any Arif Muhammad Khan resigning or Mr Justice Masood or Dr Anwara Khatoon condemning the brutality. Perhaps they want to keep intact their "holier than thou" image.
Communalism is diffused throughout India. There is no place where we might hope it would cease to exist. Something should be done to check this and save our country from ruin.
—Yours, etc.,
M Jameel, Calcutta
Deny Publicity to Communalism
Sir—On May 24, Mark Tully, the BBC correspondent in India, gave a vivid and detailed description of the riots in Meerut and Delhi. According to him, they are due entirely to
the Babri Masjid-Ram Janmabhoomi controversy. But linking a motive to problems and events does not seem to be justified.
True, leaders of the two communities have not yet come to a settlement. But if the media goes on identifying the Mandir-Masjid issue as the origin of all communal conflicts, that could be interpreted as an attempt to give it more publicity, to the detriment of national unity.
The Ayodhya dispute is a test case which, if settled carefully and generously, will bring glory to both communities. Fanatic segments should realize that violence to end this controversy will only prolong it indefinitely.
Those searching for an amicable solution may look for a precedent in Kerala where a site was claimed by both Hindus and Catholics. Hindus took their stand on the basis of a temple nearby while Catholics pointed to an ancient Cross found standing there.
Communal feelings ran high and violence was expected any day, any time. But the Catholic Bishops were able to convince their people that they had a moral duty to show greater generosity to opponents who belonged to the majority community. The dispute was settled peacefully by stipulating that the church would be given another site. This way politicians were prevented from playing their dirty game.
Leaders of other communities in India would do well to keep this good example in mind. The media should also try to suppress communalism.
Yours etc.,
S.G. Mampilli, New Delhi
Communal Riots
Sir,—Mr. Rajiv Gandhi and Mr. Buta Singh have come out with the expected condemnation of the communal riots in Meerut and Delhi. As usual, the administration has reacted late to the situation. Mr. Buta Singh acknowledges as much in his statement that "the happenings were of such a dimension that the administration was not fully geared for it".
Are the authorities ever properly geared to react appropriately to any of the "happenings" that disturb the peace and progress of the nation? I seem to recall that the "restructured"
20-Point programme had made much of the concluding point, which spoke of "a responsive administration". I wonder why the well advertised "managerial approach" to governance which Mr Gandhi once used to trumpet (and which is presumably being used) has not yet created an administration that is properly responsive to the signs and portents of such incidents in a community.
A properly responsive administration can, in most cases, nip such happenings in the bud or at least minimize their severity. Perhaps Mr Gandhi is using an ineffective version of such a "managerial approach"—which is why his Government is so incomplete in everything it has tried, whether the issue be to eradicate corruption, or to create an effective "new educational policy", or to help the drought stricken in Orissa, or properly to manage relations with the President.
G S Chandy, Bangalore
Drastic Steps Needed
Your coverage of the recent riots in Delhi and Meerut was excellent. The question now is, what are the steps that the government is taking, or at least contemplating against those responsible for the gruesome violence? And when is chief minister Vir Bahadur Singh going to resign?
Both the violence in Meerut and the recent massacre in Bihar call for drastic steps from the Centre. But while Mr Bhandesbware Dubey, Bihar chief minister, has reportedly been put under pressure to relinquish his office, no similar action seems to have been initiated against the Uttar Pradesh chief minister. There is no doubt that the general state of lawlessness in Bihar is assuming very serious dimensions. But the massacre by a part of the law enforcing machinery in Meerut was far more dastardly. Yet the government responded only with empty promises of inquiry.
The government cannot hope to heal the wounds of the affected people with long winded inquiries that ultimately serve no purpose. Such callousness only increases feelings of insecurity among the minority community.
Md. Adnan, Salmaan Ali, Jamshedpur
Communalism can never be expected to be brought to an end in the State of Uttar Pradesh as long as a man like Mr V B Singh remains at the helm of affairs and a communally motivated force like the PAC continues to be the guardian of law and order there. It is puzzling as to why the Union government has not initiated stern action against those responsible for the massacre of innocent people in Meerut and Mahana.
It is obvious that the state government has bungled in its duties towards the people of Uttar Pradesh judging from the way it has handled the riots. It is high time that Mr V.B. Singh was dismissed by the Prime Minister and deserving punishments were meted out to the culprits. The failure of Mr Rajiv Gandhi to do so would only mean that gruesome violence had his tacit support.
N Ahmed, Jamshedpur
Casteist ethnic and communal tensions are threatening to tear our country apart. People are not only getting used to but also growing impervious to news of violence emanating from Bihar, Punjab, Meerut or Delhi. This is most unfortunate considering that India was once held in high esteem the world over for its unique characteristic of unity in diversity.
Before independence, people of 11 communities—Hindus, Muslims, Sikhs, Christians, Parsis, et al—stood up unitedly under the leadership of Mahatma Gandhi to achieve freedom from foreign rule. In independent India, this bond of communal amity has been greatly undermined.
If things continue going this way there is no future for this country. It is now for the present generation to rebuild the bridges between the different communities.
Gurvinder Singh Dhullon, Calcutta
Over 150 innocent Muslims were reportedly killed by PAC jawans during the recent disturbances in Uttar Pradesh. The only punishment that was meted out for this was that a PAC officer was called back by the state government and asked to provide an explanation.
In contrast, when eight Hindus were burnt to death in mob violence in Ahmedabad, the Gujarat Sessions court judge,
sentenced five Muslim residents of the disturbed locality to death. It remains to be seen how many of the guilty PAC men treated the same way.
Y Patel, Calcutta
Apropos of your editorial "V B Singh must be sacked" (June 2), I would like to point out that riots are not new to Meerut and its surrounding areas. Prior to independence there were riots in and around Meerut which did, to some extent, contribute to the waning of the British power in India. However, the riots that take place now only contribute to the destabilisation of our nation.
The recent riots which took a huge toll of innocent lives can leave no one in doubt regarding the level of efficiency of V B Singh government. Instead of performing its assigned duty of maintaining law or order, the PAC went on the rampage killing innocent people and destroying property. Yet the Uttar Pradesh government, for all practical proposes, just sat back and watched. The failure to tackle the situation swiftly amply proves that the administration in the state has virtually collapsed. It is time that President's rule be imposed in Uttar Pradesh.
Sudhir Kumar Mishra
Tarbha, Orissa
Your report regarding the recent riots in Meerut as well as the editorial, 'V B Singh must be sacked'" (June 2), were eye openers. It is most shocking that state government officials should be involved in such a gruesome carnage, either through direct assistance or indirectly by closing their eyes to the crime and trying to hush it up.
Since the Uttar Pradesh government will obviously not take any action on its own, the press must take it upon itself to ensure that proper and immediate action is initiated against the PAC hooligans in uniform. For murder so foul, mere suspension or transfers would be too lenient a punishment.
R Bhattacharyi, Shillong
—From Telegraph, 23 June 1987
PAC not Guilty
Syed Shahabuddin has, while talking to Jyoti Malhotra (May 31), poured venom on the PAC. But it was he and
Bukhari who ordered the Muslim youths to fight on the streets. No armed forces are expected to remain silent when confronted with sophisticated arms like sten-guns, machine-guns and bombs. The PAC deserves all praise for restraint under heavy pressure.
Tatyā Tendulkar, Kolhapur
— Interviews with communal persons like Shahabuddio do not serve any purpose.
Nirmal Singh, Bhopal
— Shahabuddin's suggestion that recruitment in the police force be made on a communal basis is a remedy worse than the disease itself. It must be recalled that towards the end of the British rule in India, all government services were communalised with the Muslim representing around 50 percent, which eventually brought about the partition of the country.
V V. Phatak, Bombay
Disband PAC
I refer to Shankarshan Thakur's report "Terrified villagers recount rampage by PAC jawans" (May 26).
The role of this notorious force has always been a terrifying one, be it at Moradabad, Allahabad, or now Meerut. Even children, the sick and the very old are not spared by these organised bunch of thugs.
Armmen posted in Meerut have severely criticised the role of the PAC jawans, who were actively involved in killing and looting. Has the time not come for the authorities to seriously think of disbanding the PAC before it is too late?
Imran Ahmed, Calcutta
I write with deep regret and shock on the role of the Provincial Armed Constabulary (PAC) in the recent riots in Meerut. Do they deserve the title "constable" when they kill, loot houses, molest women and remain a silent spectator to the sufferings of the common people?
The Uttar Pradesh government was quick to restrict the movement of the Shahi Imam and other Muslim leaders, but is satisfied with the PAC. The virus of communalism is being
allowed to spread by the administration and helpless people are falling victims to this communal game
Mrs Shamim Ahmed, Jamshedpur
There is no need for us to feel threatened by the Islamic bomb India is being destroyed by its own communal bomb
B R Rao, Jamshedpur
Nothing is more harmful to our national integrity and secular status than the rapidly increasing communal forces I appreciate the Prime Minister's warning of firm action against the rioters and anti-socials in Meerut and Delhi Drastic action is certainly needed against those involved in looting, arson and butchering of the innocent But I would like to know what our Prime Minister thinks of the negligence of the Uttar Pradesh government and the despicable game of the infamous PAC in the riot torn areas Is Mr Gandhi contemplating action against them?
The Prime Minister should immediately take firm action against the troublemakers and also punish the lax administration of Uttar Pradesh He should also take steps to reorganise the PAC
Syed Ejaz Ahmed, Burnpur,
The Telegraph
PART IV
INTERVIEWS
Syed Shahabuddin
WE CAN REFORM THE PAC, BUT WHEN THE INTELLIGENTSIA TURNS COMMUNAL, WHERE DO WE GO?
Jyoti Malhotra interviewed Janata MP, Syed Shahabuddin, on Friday on the Meerut riots.
Do you feel the Meerut riots were in any way different from the other riots that have taken place since partition?
Over the years the phenomenon of communal violence in our country has changed. And, therefore, not only Meerut but in most places riots have taken on the colour of a confrontation between the Muslim minority and the state police force. Thirty/forty years ago riots were primarily confrontations between Hindu mobs and Muslim mobs. Or you had Muslim mobs attacking Hindu localities or Hindu mobs attacking Muslim localities. The law and order machinery, with exceptions, was more or less neutral and they used the force at their command to quell violence. I noticed the change in Moradabad. I found there that the police had used its bullets but the casualties in Muslim mohallas were Muslims. So I asked this simple, commonsense question: I said, obviously it was not Muslim mobs which were attacking Muslim mohallas, it must have been a non-Muslim mob. And if the police came there to suppress violence, then the victims largely should have been non-Muslims (in Muslim mohallas, as in Hindu mohallas under Muslim attack the casualties would have been Muslims), but no, here there are Muslim casualties in Muslim mohallas, how do you explain that?
From there I began probing and came to the conclusion that basically it is not the Hindu masses or the Muslim masses which are involved today in communal violence. There are anti-social elements in every community and they avail of every opportunity of looting, killing, etc. But basically the casualties in a riot arise when the police supports anti-social elements, when the police provides the fire power....
The anti-social elements may be Hindu or Muslim....
Of course, I cannot rule out the possibility of Muslim anti-social elements attacking wherever they have a chance. For example, in Meerut itself, on Hapur road most of the arson has been committed by Muslims. But in most instances today, the casualties in riots as a whole are due to police firing, and they are largely one-sided. It is the police vs the Muslim community.
Look at Maliana. The victims are hundred per cent Muslim. And you ask them—after all human life is valuable and in every civilized society you give information about who has been killed. When? Where? How? This information is simply not available here.
So you ask a basic question, have the riots undergone a change? My answer is that largely the configuration of the riots today is not mob vs mob or mob vs locality. It is largely a community faced with state violence. And this is a very horrifying state of affairs because if a community loses confidence in the law and order machinery, if it looks upon the police as an enemy, then it is bound to make its own preparations. And this is what is happening all over the place. The PAC, particularly in Uttar Pradesh, has acquired such a notoriety—in 1972 Mr. Madhu Dandavate inquired into the PAC's role in the Ferozabad riots, in 1982 he along with a parliamentary team went to Meerut and denounced the PAC, in 1980 I enquired into the Moradabad riots and it was wholly the PAC's brutality—and yet it remains as it is, untouched, unrepentant.
The National Integration Council (NIC) passed a resolution that the armed constabulary should be restructured so that it reflects the population in the area. It did not mean that anybody was calling for a reservation, but that the people were conscious of the fact that a police force should be a composite force. You cannot bring peace in Punjab by having purely a Sikh force. You cannot bring peace in UP by having purely a Hindu force, it must have a Muslim face also.
What did you find in Meerut when you went there?
You know the real trouble started on 18th morning. The police went to arrest some people in connection with the April disturbances. I don't know why they chose to do so now when the situation had become normal. But anyway, they went at 4 a.m during sehri. They arrested some people, it created a commotion. Women and children came out of their homes and lay down before the police jeeps, the jeeps over-ran them. One child was killed on the spot, women were fired at.
After this, naturally when there is lack of confidence and the Muslims say, 'This PAC has come to burn us, to loot us' — and the PAC enters Muslim mohallas as if it is entering a war zone, it is invading a foreign country — the result is confrontation. And then there was an announcement from the mosque addressing the police force, "We are warning you, don't enter our area, if you do so then we shall also do whatever we can." And that was the beginning. So where was the Hindu mob?
Then within a day they sealed off the Muslim mohallas, and in the mixed bazaar area the police got the anti-social elements to collect together to pick and attack, burn and loot Muslim shops.
Is this a riot? This is not a riot. So how do I blame my Hindu brothers? I cannot blame my Hindu brothers. I only blame this government or its machinery.
But is this state machinery synonymous with the Hindu fanatic element?
Well, the government or the party in power may have a political purpose, but if after the decision of the NIC or after Mrs Gandhi's 15 point directive which was issued after the 1982 Meerut riots that the minority should be better represented in the police force etc., if nothing happens, what conclusion can you draw? I know nothing can be done overnight, but five years
have passed. Obviously then the government wants to say one thing and do another. Everybody says the PAC behaves in a communal manner, in a criminal manner and yet you do nothing about it. So I can only draw the conclusion that you had a certain purpose.
Look at Maliana. What the PAC has done is now well-known. Is there any civilized country in the world where such a force would not have been disbanded and asked to go back to the barracks, placed under arrest and then tried? We talk about atrocities in South Africa, in Sri Laoka. You tell me, does human life have no value? There are eyewitness reports that people were brought in trucks to the Hindon river and shot. One person who fled into the bushes is alive and I will produce him before the press. Dead bodies are coming out all over the place in two canals, the Hindon and Ganges canals and the government knows nothing, sees nothing.
How can we be proud of being Indians. Nobody has been permitted to go into the Muslim mohallas until day before yesterday. In Meerut do you know that the Muslim women I met in the collectorate told me that they had gone without food or water for seven days. May be Maliana shall touch our conscience and bring into sharp focus what needs to be done. It is a mere hour's drive from here, but so was Nellie only 35 km from Guwahati and yet 3000 people were killed.
So you feel that there should be a Muslim component to the police force?
I'm not talking about a Muslim component, I'm talking about a composite force. In UP, for example, the tradition has become that the force does not comprise only of Hindus, but even among the Hindus it is made up of a particular caste. So when the PAC goes to a Harijan locality it commits the same atrocities.
We cannot afford this segmentation. When our Hindu brothers were fleeing from Punjab I spoke to Barnala. I said that the only way you can prevent this from happening is to take all the Hindu policemen and put them to the thanas in the four most affected districts—the Hindus will automatically have confidence in them.
There were three Central ministers in Meerut Buta Singh, P. Chudamaram
(Interrupting) But what did they do there? That's what I want to ask. I have charged that every time Mr Veer Bahadur Singh goes to where a riot has taken place his visit exacerbates the situation. Instead of quenching the fires he gives the police a pat on the back and says, "You are great, go ahead, give them a good, sound heating."
But Mrs Mohsina Kidwai is the MP from Meerut
I don't want to speak about Mrs Mohsina Kidwai. Do you know that she wasn't even allowed to participate in consultations. If she had the slightest sense of self respect she should have resigned from the Cabinet. But these Congress men and women are devoid of self respect. It is her constituency, it is her voters who are being killed and what has she done? Has she gone to any locality which has been not affected? When she went there she went to the Circuit House, from there she apparently went to the hospital and then she came to Delhi. And I am receiving call after call from Meerut that we are besieged.
Apart from the PAC, who else was significant in Meerut
One, that people were killed, some of them burnt, in localities like Shastri Nagar, which is not a locality inhabited by anti-social elements. These are upper middle, upper class people, educated intelligentsia. I don't mind if Hindu goondas fight Muslim goondas. I don't even mind if the PAC commits atrocities because tomorrow we can reform the PAC. But after all the communal virus, like war, begins in the minds of men and if our intelligentsia is affected by this virus then what, where do we go from here? This is the essential question of Meerut. How could a poet like Bashir Badar's house be burnt how could a son of Dr Harpal Singh (a doctor who has lived there for decades) be killed? This is the question.
The other aspect of Meerut is that when people are in distress, they have nowhere to go to. So that is why in a statement I appealed to the secular forces, 'put yourself between the people and the administration.'
There have been charges that in Meerut the Muslims have hoarded weapons over a period of time and used them in the riot?
I don't buy this theory of large-scale hoarding, but, if I am afraid, if I face a threat and if I don't have trust in the forces of law and order, then what shall I do? Shall I not make a defensive preparation? Hindus told me that every house in Meerut has weapons, whether you search Hindu or Muslim localities. The problem is that if you search only Muslim areas then obviously you blame only Muslims.
Tell me, if out of desperation and out of frustration, the police is hitting you from one side and you don't know where to go and some men in anger say, 'All right if we have to die, we shall die but we shall burn up this place'. Do you blame them? After all, anger and desperation comes out of frustration. And this is what happened. On the 19th morning after Imliyan Mohalla was attacked the Muslim youth came out on the other side and burnt up Hapur road. There is no doubt in my mind that was done by Muslims.
What should you do now, with Meerut?
Firstly, you should open up the place, because there is no better disinfectant than sunlight. And there is no better curative than truth. And then in every locality make it a rule that it is the responsibility of the local majority community to be responsible for the life and property of the minority. You must create a vested interest in social peace. Secondly, please take away this monster of a PAC, its very sight inspires revenge. I tell you, then Meerut will have peace. And if you want to inspire confidence at least catch the culprits of Maliana because you must know who they are, disarm them and try them in a proper judicial manner. Fourthly, let the government be generous, go and rebuild their homes, give the widows a job.
From Sunday Observer, 31 May 1987
Bir Bahadur Singh
"WE NEVER EXPECTED THIS"
The 52 year old Uttar Pradesh Chief Minister Vir Bahadur Singh is known for his political acumen and ability to survive crises. But last fortnight when communal violence in Meerut, for the second time in less than eight weeks threatened to spread to other parts of the state he was faced with the worst political crisis since he took office 20 months ago. Singh who has recently been flying all over the country in search of new capital, found himself confined to the riot torn city for a week. Curiously, the high death toll hardly seemed to have affected his morale. He spoke to visiting politicians and journalists with confidence giving the impression that nothing new or serious had happened there. Senior Editor, Prabhu Chawla talked to him for more than an hour. Excerpts from the interview.
Q What do you think was the cause of the riots?
A I can't give you a specific answer on this now. I have ordered a CID inquiry. But I must clear the impression that it is the worst ever riot. In 1982, the riots continued for over two months and more people were killed then. We have been able to restore normalcy in Meerut in less than two weeks.
Q How many were killed then?
A I don't have the exact figures with me but the toll was definitely higher.
Q. Do you also agree with the Union Home Minister Buta Singh's view that the riots were pre-planned?
A. I can only say one thing. Both the communal and anti-social elements exploited the religious sentiments of the people to create a law and order problem.
Q. It is believed there was tension in Meerut since the last riot. What preventive steps did you take?
A. We never expected this. It had cooled down. But now it appears that plans were quietly being made. The CID and the Administration always maintained that nothing would happen. They had arrested the anti-social elements.
Q. The State is acquiring a bad image now?
A. I don't know why only Uttar Pradesh is making headlines. It is not the only State which is affected.
Q. On the Ramjanambhoomi-Babri Masjid issue, you have been accused of following a pro-Hindu line.
A. This is an absurd allegation. Did I take any personal or administrative decisions in this case? Courts in the country are independent and can't be forced to work on communal considerations.
Q. What have you done as Chief Minister in this regard?
A. I have given directions to restructure the entire police force, including the PAC, in the state. Their training programme is being reworked so that the policemen acquire a secular outlook. Special battalions are being raised in which all the communities will be proportionately represented. We are also reviewing school-books to see that they are free of communal content.
Q. But politicians always pander to communal elements during elections.
A. That is not true of our party. We don't encourage communal elements. We believe in the protection of the rights of the minorities.
Q. But your state police is full of communal elements.
A. The entire force is not communal. We are looking into some specific complaints and will take stern action.
Q. It seems that you are losing grip not only on the administration but on your cabinet colleagues as well.
A I don't think this is true. I enjoy the full support of my party and the Cabinet.
Q Don't you feel that your image as a business like chief minister suffered?
A No, I meant business when I took over. I took disciplinary action and dismissed senior officials. I directed all my cabinet colleagues to be in office by 10 a.m. and not to leave the city without my permission.
Q But the law and order situation has worsened.
A Absolutely not. In fact, it is the best during the last five years. There have been no gang killings, roadside robberies, daylight murders and dacoities. Ours is the most peaceful state today.
Q As far as communal situation is concerned, is the worst over?
A Yes. But we have to guard ourselves against any eventuality. It can be done through dialogue and persuasion. We are going to sort out the Ramjanambhoomi-Babri Masjid conflict with the help of local people.
Q Is there no immediate threat to your leadership?
A Yes, I believe so. I enjoy the full support of the party high command and have complete freedom to handle any political and administrative situation. There is no political crisis.
PART V
WHAT APPEARED IN THE LANGUAGE PRESS
ALAM BHAI'S SLAP ON THE FACE OF DELHI COMMUNALISTS
Those who hold any one community responsible for the Old Delhi riots and spread communal poison should meet Kanchan, a seven year old girl and her 40 year old mother Jaya Tiwari in the emergency ward of Jay Prakash Narayan Hospital (Irwin Hospital). They should be sent to Rekabganj, behind Delite Cinema where on the night of 22 May during curfew hours 100,125 Hindus and Sikhs enjoyed the hospitality of Muslims who had sought refuge in the mosque and spent the night there.
What Vijaya Tiwari of Kucha Pandit says with the head of her injured daughter in her lap and what Maulana Muhammad Yaqub Qureshi and Kashmirilal say is a big slap on the face of those who want to create cleavage among those families who have been living together for generations.
Kanchan was hit with a bullet on the head during the disturbances. The person who took this Hindu child to hospital risking his life amidst the barrage of soda water bottles, stones and bullets, his name is Alamhbai. Kanchan's mother also does not know what is the full name of Alamhbai but she says he is like my father-in-law, or my father. There is no relation greater than this. Had Alamhbai's family not saved us, none of us would have survived. He called us to his house the moment disturbances began. Our whole family (Vijaya's family) is still in their house. They eat there and sleep there and Alamhbai after getting the girl admitted to the hospital has been
living here and there. He sleeps in the mosque and prays for Kanchan's recovery. Curfew has prevented him from going home. On the day of disturbances the police arrested five brothers, sons and nephews of Alambhai. He is searching for members of Corporation so that they may help him getting them released from the clutches of police. But the leaders and corporators are busy issuing statements against the riots. Alambhai every day comes to the hospital to inquire about the girl. He consoles her mother and asks her to have trust in God.
Kanchan's family lives in house no. 3453 above Amar Press in Kucha Pandit. Just behind it is Alambhai's house, said Kanchan's mother. "When the Friday prayer was over peacefully on 22 May, we heaved a sigh of relief. But the crowd passing through the lane did not seem to be well-intentioned. Suddenly we heard the sound of attack and hue and cry. People were running helter-skelter. We did not know what to do. Then Alambhai jumped into our house from the side-terrace and told us, 'Outside miscreants are rioting. You immediately come to our house otherwise you may not be safe.' And then Alambhai hurriedly took Kanchan and her 17 years old sister Deli and her mother and Father Amarnath Tiwari to his house via his terrace. He also, at his own risk, took 8 employees of Amar Press below to his house. Kanchan's mother said that when we reached Alambhai's house his daughter-in-law told us to remove sari and put on shalwar kamiz (North Indian Muslim dress) so that the rioters should not recognize.
"We were all sitting with Alambhai's family behind closed door. On the terrace were the men of armed police and CRP. Suddenly a bullet fired from a terrace pierced the door and fell down after hitting Kanchan's head. Her face was covered with blood. We were sitting with our injured daughter for more than one hour. Police was not allowing us to go out. After one hour the police knocked at the door and as soon as the door opened they arrested sons, nephews and brothers of Alambhai. I kept on pleading that they have done nothing. They are all sitting inside. This family has saved our lives. But no one listened to me." The suffering mother said that Alambhai took injured Kanchan to hospital and did not allow
anyone to accompany him. He reached the police van hiding himself here and there and got her admitted into the hospital. When she reached there (i.e. Kanchan's mother) she was being operated.
Many such Shams Arif and Alamhais are also found in Rekabgaoj of Daryaganj mohalla who (i.e. the Qureshis of the area) gave refuge to hundreds of Hindus, Sikhs and Muslims in the nearby mosque. They used to bring food from their houses to feed these distressed people. The Imam of Badi Masjid, Muhammad Yaqub Qureshi, Muhammad Shafi Qureshi and others are witnesses to the fact that how many Hindu women were taken safely to their houses. Kashmiri Lal said that in this mosque people of all religions took refuge. The unity which we have would be hardly found elsewhere. We celebrate all festivals together. Imam said solemnly: "Don't be surprised. He is Hindu and is sitting inside the mosque at night despite tension outside. You would not like to know more than this perhaps. Kashmiri Lal has been living here for generations. He has his toy factory in Karol Bagh known as Rekha Toy Industries. He said that we have not seen any riot here after 1947. The Qureshi butchers and the Jats have set an example of communal harmony."
—From Bebak, Malegaon
(Reproduced from Hindi Daily Jansatta, 25 May 1987)
AN ATTEMPT TO HIDE PAC'S CRIME
Virendra Sengar
When PAC's butchery at Meerut could not be suppressed inspite of all the efforts of officers of UP government including those of its Chief Minister, efforts are being made to distort the facts. So much so that when the Prime minister Rajiv Gandhi visited Meerut, he had to face "PAC Zindabad" slogans. The "official butchers" are being given all protection and efforts are being made to suppress the voice of relatives of PAC victims of Hashimpura Muhalla.
When the facts could not be suppressed, the administration made all efforts to present a distorted picture of actual happenings. During this campaign the investigating journalists were fed with the information by the administration that all those killed by PAC were Pakistanis who were in Meerut to create troubles and some of them might not be Pakistanis but were very dangerous.
This reporter after investigation learnt that there is no truth in official report. All those killed were the residents of Hashimpura area and belonged to one community.
Janata Party President Shri Chandra Shekhar presented before the reporters at Delhi a young man named Zulfikar Nasir on June 1st who was carried on the truck at Hashimpura and he was thrown in the Ganga Canal after being shot. Fortunately he was slightly injured and alive. He narrated the entire incident.
Next day when the above news was flashed in newspapers, the District collector of Meerut Mr. Radheshyam Kaushik contradicted the report and said to the local newspaper reporters
on 2 June that Zulfiqar is not a resident of Hashimpura and the person presented by Shri Chandra Shekar is not real.
The District Collector's statements proved completely false after investigations. Zulfiqar is a resident of Hashimpura and this Reporter met his family members. Nasir's mother kept on weeping and his father Abdul Jabbar is behind bars. In this context, the District Collector could not be contacted despite several telephone calls.
I met several persons who lost their near and dear ones in this massacre. Let us hear Zulfiqar Nasir's own version of the happenings before we reckon the report of other affected families.
The 17-year-old Zulfiqar has studied up to High School. Even after a lapse of 12 days, his eyes are full of fright. He said that PAC Jawans fired for three hours on 22nd May in our Hashimpura Mohalla. The residents of our Mohalla at first retaliated for some time by throwing stones—thereafter they ducked out of fear.
Few hours after the firing stopped, the Army and the PAC made people sit outside their houses under the pretext of taking search. About 500 people were made to sit on the road. Then they screened the old, the sick, and the children. Then the remaining, about 200, were loaded on five PAC trucks. The truck which was loaded last, I was made to board it.
The truck had started in the evening. I do not know along which path it proceeded but when the truck was parked near Ganga Canal some people had said that this is the Murad Nagar Canal. First Yasin (50 years old) was made to get down and was shot dead. Second Ashraf was shot and I was third to be shot when they were lining me up for being shot. I was shivering. For this reason they missed the shot. The bullet passed bruising the lower part of my hand. I fell down, immediately after I was hit and I tried to hold my breath to save myself. As soon as I fell down I was thrown into the canal. Therefore, I could not see who among others were shot though I was hearing the sound of firing and cries of the people.
When I started flowing in the canal I caught its side wall and kept on hanging on to it. I remained in that position for half an hour and kept on waiting so that I could come out after
the truck leaves
When the truck left I crawled out. Blood was flowing from my band. I had seen many people being thrown into the water but I was not in a position to do anything. After sometime my fellow boy Arif came near me. He was not hit with the bullet because when people were being shot he had jumped from the truck into the canal. Many bullets were fired at him after he jumped but he remained safe. We both were greatly afraid. When we were coming out of the canal we saw that injured Qamruddin was crying with pain at some distance... Qamruddin was hit with many bullets. Lot of blood had flown from his body. He was saying again and again, "Nasir you run and save your life. I will not survive." At last I ran away seeing an opportunity. I began hiding when people followed me. And then I hid myself in a urinal near an electric house on Murad Nagar Road. One Darogha kept on searching for me until morning. I was hearing his voice 'Sala Kaba Bhag gaya' (Where the swine has run away). I hid myself for full fifteen hours... A Janata party leader Shri Ansari took me to Delhi.
Among the survivors there is another youth called Naim. He had also jumped into the canal from the truck. He stayed with the Uncle in Razapur near Ghaziabad after the incident. His statement about the incident fully corroborates Zulfikar's statement. He says for sure that he saw 15-16 people being shot. When he saw there was no way to survive, he jumped into the canal from the truck. Naim gave the information that along with Arif in all three persons jumped but, one was hit with the bullet in the water itself. Naim has also given the name of the boy hit with the bullet.
Naim is so terrified that he is not sure whether he will be able to survive in future.
In order to know what happened to Kamruddin who was left behind by Arif and Zulfikar, Saleem Ahmed kept on visiting Ghazijahad police Station for 10 days. But no one gave any information to him. At last a Darogha from Muradnagar police station took pity and told him that when the police party was taking injured Kamruddin to Hospital he died and the body was sent for postmortem. It was from
police that Jamaluddin, father of Kamruddin got photo of his body and the last ray of hope was also extinguished.
When the hapless Jamaluddin went home with the confirmed news of his son's death to Hashimpura, this reporter was present there. A pall of gloom fell on the whole lane and a wave of cry went up hearing the news of death.
One D M Khan of this Mohalla said that since this PAC massacre all those who are missing from their homes are being suspected to have died.
On enquires it came to light that many who were loaded on trucks on 22nd May, many of them are in jails. A list has also been sent from the jail. But those people who are not in jail or police station, it is difficult to know the exact number of such people but there are about hundreds of such persons about whom nothing is known. People suspect that the PAC has slaughtered more than 40 youth.
From Chauthi Dunya Weekly, 13 June 1987
WHAT WE SAW IN MEERUT? WHAT WE SAW IN SAYHANA?
TWO SIDES OF COIN
Peace March by Sadiya Dehlavi
It was the fifth day of riots in Meerut. There was dreadful and painful silence. There was no one on the roads for miles except the police. There was bullet marks on the walls of the houses. It was the scene of day of reckoning inside the mohallahs. We saw with our own eyes the shattered pieces of human heads struck with bullets. Innocent widows were crying and asking why did it happen. Who will look after their children now?
The powerlooms inside the houses were burnt. One 80 year old man told us weeping that his young son and grand daughter both became victims of bullets. We could not even go out to bury them due to curfew. He hurried his children in the lane outside. There were strewn bodies of dead animals inside lanes. They died because of starvation. People had nothing to eat inside mohallahs; people were not venturing to go out of their houses due to fear.
Seeing such soul searing conditions, neither could I swallow anything nor could I sleep at nights. I felt totally helpless. We are killing each other even after 40 years of independence. Alas, we have kept our identity as Hindu, Muslim and Sikh rather than as Indians. Why such aggressive communalism spreading in a country for whose freedom all fought together? It is weakening the very roots of our country.
We have resolved to fight against this fast spreading communalism. To revive Gandhiji's ways we decided to go on peace march to Meerut. We collected youth of different religions and started from the Boat Club, Delhi, for Meerut, on foot. In the searing sun and intolerable heat we kept on walking. For the first time we saw some regions of Jamnana. It was difficult to believe Delhi has spread so much. It was evening before we left the boundaries of Delhi. On the way people kept on walking with us for some distance. We reached Mohan Nagar at about 6:30 PM where we stayed for the night. We got up at 4 in the morning and came on the highway after bathing at 5. We passed through some villages also. We felt very happy to visit Sayhana in Ghaziabad district. We felt as if we had reached an oasis after loosing our way in a desert. Hindus and Muslims live in perfect unity there. Hindus contribute for the madrasas there. With great pride the people of that village told us that not a single person has been killed here since 1947. In such a village we see bright future for India and we tend to be convinced that humanity is still alive.
In cities where fire of prejudice and communalism is spreading, in villages people's conscience is still like a minaret of light. There is fraternity, sympathy and love among them. There is still sincerity and simplicity among the people in villages. We were pleased to see the political consciousness of these village people. Everywhere we gave the message that we should adopt non violent ways of Gandhiji and should unite to fight against injustice and oppression.
That day we walked 23 kilometers in scorching sun. Due to intensity of heat our lips went dry and our feet developed sores. But we neither felt pain nor the scorching sun. We reached Meerut at about 1 PM. There many people had collected to welcome us. In the evening we went to many mohallahs with the message of peace. We were relieved to see that tension was decreasing. We were also convinced that if people have noble motive, God gives them courage, fortitude and energy.
Translated from Qaumi Tarjuman, Lucknow
TEN IMPORTANT QUESTIONS ABOUT MEERUT
1. When tension had spread in Meerut city on the death of Ajay Sharma, why efforts were not made to defuse it by showing it as a consequence of mutual enmity and why arrangements were not made to stall the possibility of riot?
2. Why the 38 companies present in Meerut for riots which continued upto 17 April were withdrawn when the possibility of eruption of riots was very much there?
3. Every child in Meerut knows that usually there is communal tension there. Why then no arrangement was made for the same?
4. Meerut city and district have been under communal tension for the past several years. Why then licences for arms were issued? Presently there are 17 thousand licensed arms in Meerut. One additional district magistrate issued maximum number of licenses for money. Complaint was lodged against him with the Government. What action was taken against him?
5. When it has been decided in the National Integration Council not to appoint PAC in riot-torn areas, why was it then posted?
6. Despite strong possibility of riots why only three fire-brigade vehicles were in fit condition. It was for this reason so many incidents of fire took place. Last year 70 vehicles of fire-brigade used to be in ready condition. Why all were not in working condition?
7. Why the Meerut district administration did not emphasise reinforcement of police force?
8 Why the district administration did not take suitable steps when the local intelligence unit had warned about the possibility of communal riot and bringing in of weapons from Delhi? Were vehicles coming from Delhi checked? Not only this but even after the rioting broke out, only the vehicles on Delhi Meerut road were checked whereas all the vehicles coming from outside should have been checked.
9 Knowing fully well that in Ramadan at the time of sahri (meal just before dawn), there is always large crowd of Muslims, why the police went to arrest the killers of Ajay Sharma at that hour?
Apart from the above questions there have been many other faults on the level of district administration and Meerut city became victims of lawlessness and disorder. Akashwani and Doordarshan (television) also were presenting the news about Meerut in such an exaggerated way that it was felt the Government media are deliberately taking interest in riots. There was an explosion in Bulandshahar in a fire cracker factory in which four persons died. Similar incident took place in Aligarh and in Ghaziahad, there was quarrel due to mutual enmity. But the Television presented these news along with the news about Meerut riots as if in these places also riots have broken out.
—From *Ravivar Weekly*, Calcutta
PART VI
DOCUMENTATION
"I REGRET YOUR LEADERSHIP"
Here is a translation of the entire text of the letter, which has created a commotion in Congress-I circles, with its sharp attack on the leadership and its style of functioning.
My dear Rajivji,
I am writing about the communal disturbances which have recently occurred in various parts of the country. I hope you will consider seriously the facts which I have mentioned in this letter and with national concurrence you will take necessary steps against communalism and separatism.
During communal riots, such as the one in Meerut, the violence and arson shocked all Indians who had faith in secularism, communal affinity, national unity and integrity. Once again the country which has been propagating the message of truth, non-violence and universal brotherhood has had to bear the agony of witnessing its principles being mutilated. Innocent persons have been butchered and feelings of hatred are being spread among brethren, resulting in an atmosphere of violence which seems to be a nightmare. Such violence has been condemned throughout the country. You too in your statement expressed your anxiety about ascertaining the basic responsibility of the government so as to protect the minority as well as the majority community.
But the question arises, how can we solve the problem by merely showing concern and condemning the communal violence? The intensity of communal violence has awakened the national conscience. Even while Meerut is limping back to normalcy, suitable steps must be taken to maintain communal
harmony, before such an incident is repeated. I understand that communal riots are subjects of shame and anxiety for the whole nation. But the communal riots are only symptoms of the disease which is called communal feeling. If we want to be free from communal violence, we must get rid of communalism. If we want to protect the innocent people, we must act against communal forces and communal politics. Communal violence is a communal disease. It is an administrative duty to quell communal disturbances. But actions against communalism can be taken only on a political basis.
If you take a look at the history of the last 40 years, you find that the Congress governments have been doing such work. Pandit Jawaharlal Nehru, Shastriji and Mrs Indira Gandhi always cautioned the nation against the dangers of communalism and they fought tooth and nail against communal forces. During this period due to efforts made on various fronts, we have seen popular sentiments against communalism. That is why on the eve of the parliamentary elections in 1984, the whole country supported your determination to fight against the communal and divisive forces. But it is not known how you have distracted yourself from the popular policy of secularism and compromised with communal and divisive forces.
I don't want to remind you about old wounds. But to solve a problem, it is necessary to find out its cause. During the last two years, in the name of harmony, a compromise has been made with all kinds of communal destructive and divisive forces which threaten us with violence. I understand that behind all these compromises, you must have aimed for peace. But such bargaining failed to bring peace. On the contrary, communal and divisive forces gained a lease of life and authenticity and consequently increased their power to threaten us. That is a great challenge to us.
If we learn a lesson from the history of communal politics in India, it is not profitable for the nation to have any compromise with the communal forces, which become strong and destructive by mere encouragement. The evil of communalism cannot be controlled either by talk or compromise. We must face the challenge for the sake of national integrity, and effective confrontation with communalism can bring real national unity.
and integrity
I would like to bring to your notice your statement after the communal violence in Meerut. Your appeal to maintain communal cordiality was very much appreciated. It was proper to warn the violent communal forces. But over and above you mentioned the responsibility of the government to protect the minor and major communities. I understand in view of constitutional provisions, religion cannot be the basis to provide safety. Any citizen belonging to any caste and creed deserves to be protected by the government. I do not understand how long Indians will be seen as fragmented religious groups. The day is still awaited when every Indian will be equal as per rights and protection guaranteed by the Constitution and laws.
When there are communal disturbances, the capacity of the administration to maintain a peaceful atmosphere is very doubtful. Impartiality and duties of responsible persons are not without complaint.
In the case of Meerut, allegations against a section of the police are shocking. But according to knowledgeable sources, the administration and the police acted as per the political leadership. On the one hand, a central minister is alleged to have put political pressure to remove police bundobust from the sensitive areas, and on the other hand atrocities occurred after the disturbance in the presence of the chief minister.
According to my personal knowledge a senior police authority in Meerut expressed his inability to obey the orders of the chief minister at a meeting. Within four days he was transferred. I have come to know that the said authority has sent complete details of the meeting to the central home minister.
It is a great pity that a secular congressman who was considered a symbol of communal cordiality is now in the clutches of communalism. Allegations may be made against lower employees and authorities, but it cannot be denied that the ultimate responsibility lies with the political leadership for moulding the human attitudes in the country. We may have complaints against administrative authorities and the police but the fact is that we have allowed responsible partymen to encourage communalism. Action should be taken against government employees who have been partial.
But would you think of taking action against those who not only talked of communalism in Parliament but also ridiculed the Indian Constitution and the Judiciary? Effective action must be taken against a partisan and a communalist whether he is a minister or a government employee or an ordinary citizen. Only then can a repetition of Meerut be avoided.
If we decide to fight against communalism we have to change our methods of work and principles. Communalism is a great danger to our national unity and integrity. We have to see that our decisions and work do not encourage the communal and divisive forces.
If you see some of the decisions taken during your administration, it will be quite clear what the main cause is for the unexpected increase of communalism. Since independence till the death of Mrs. Gandhi, the government has tried to keep communal forces at a distance. Men involved in communal politics made tall promises but did not have the capacity to fulfill these promises. That's why they failed to enthuse their respective communities and gain their confidence. But this has been done during your administration and communal leaders have been able to fulfill their promises. The distance between the government and communal leaders has disappeared. Consequently, communal forces and communal politicians have achieved a new lease of life and authenticity.
The danger is that the problems are getting more and more difficult due to our mistakes and we do not have the courage to rectify them. It is necessary to keep a distance from communal forces and we should not have any pact with communal political parties such as the Muslim League in Kerala and Ittehadul Muslemin in Hyderabad. We should not encourage the formation of new organisations on a communal basis. I have recently seen a circular whereof as per your instructions, Mr. Tariq Anwar has organised a new group of Muslim members of Parliament and MLAs under the name of Kaumi Tanzim.
The group has been asked to give a solution to the Ram Janamboomi-Babri Masjid controversy. This controversy has created tension and uneasiness everywhere in the country, especially in Uttar Pradesh. The cause of this controversy is well-known. The solution to the problem must be found out immediately. But is it possible for an organisation comprising one community to give the solution to such a problem?
Still there is time. You must go deep into the matter; and believe me there is no dearth of such people who can bring about communal harmony between the two communities and they can help us to find a solution which will be acceptable to all of us. But for this, communal leaders must be left out of the limelight. Until the communal forces are disintegrated and weakened, the danger to national unity will persist.
Why are there communal riots? Why does communalism spread? Why are innocent people being butchered? Being the prime minister you are responsible for stopping such disturbances. Whatever may be the causes of violence—administrative inefficiency or foreign hands or increasing communalism—nobody is interested in clarifications. You must make your administrative set-up very effective. If it is a foreign hand, it must be broken. The struggle against communalism has been part and parcel of our national movement. Basically, communal riots and communal political leadership are symbols of fragmented popular faith. Inefficiency and failure is due to political leadership.
The following couplet of an Urdu poet is very logical in the present context:
Don't give me all these excuses
Tell me why the caravan has been looted
I have no concern with the robbers
It's a question of your leadership
Shall I tell you why the caravan was looted?
Because you were in league with the robbers
I have no complaint against the looters
But I regret your leadership.
I understand that we should accept the challenge in time to save national unity and we must start a war against communal forces so that we shall be able to protect our national unity and integrity. Communal politics and national unity cannot be maintained together. To keep national unity intact we must be free of communal poison. To do so we may face political inconvenience for the time being. We may have to offer some
sacrifices, but as a result of this the nation will benefit. In this context I would like to draw your attention to an interview given by Jawaharlal Nehru to the Bombay Chronicle on 26 August 1936. He said, 'Success is not merely winning a few seats in an election, by compromising principles and ideals. Success is following your ideals, even if you have to pay a price for it.' I believe that in future the principles and ideals of Panditji will inspire us to fight against communalism.
With thanks.
Yours sincerely
Arif Mohammad Khan
POLITICAL PARTIES
CPI(M) Politbureau Statement on Meerut and Delhi Riots
The Polit Bureau of the Communist Party of India (Marxist) expresses grave concern at the riots that have rocked Meerut city and which have affected the capital city of Delhi also. The protracted nature of the riots, the intensity of violence and the break down of the administrative machinery in Meerut is truly disturbing. Scores of persons of both communities killed, large scale arson and organised looting, all point to the failure of the U P Government to restore a semblance of order and security for citizens since the last round of serious rioting in April. The army has had to be called in to control the situation. In Delhi also, the relaxation of curfew in a tense situation has led to the recurrence of violence with arson and looting and deaths of four persons.
These riots are taking place in an overall situation of growing communal tension in the country. Gujarat continues to be gripped by communal violence with the latest round taking place in Broach. Various centres in Uttar Pradesh and Maharashtra have also witnessed communal incidents. Every time such violence breaks out, pious determinations are made by the leaders of the ruling and some other political parties. The authorities solemnly pledge themselves to curb the communal forces. These, however, do not lead to practical actions. Hence the riots recur.
Once again, the Prime Minister has declared that the communal forces which are out to create tension and organise riots will be "firmly dealt with". The question is not of
administrative measures alone, but a combination of administrative and political measures. The preparations by communal elements continue for a long time before the actual outbreak of communal violence. A firm and uncompromising fight against these forces and a giving up of the opportunistic policy of appeasing them on explosive issues is necessary on the part of the ruling party. Only this can help to a united campaign against both majority and minority communalism (23 May 1987).
CPI: Statement of Indrajit Gupta, M.P. & N.E. Balram, M.P.
We visited Meerut on Saturday (23rd) but were refused curfew passes which would have enabled us to meet the people in some of the riot affected areas of the city. The Collector claimed, on the one hand, that since the night of the 22nd May no fresh incidents of violence had occurred in the city; on the other, he obviously did not want Members of Parliament to see things for themselves.
Our queries as to how press reporters were free to go about and even to publish inflammatory accounts of incidents and casualties while MPs were being discriminated against, were not replied to.
We found Meerut a dead city under total curfew, which is probably effective only along the main thoroughfares but flagrantly violated inside the congested localities. A virtual civil war is being waged with a surfeit of fire arms, and rioters of both communities running amok.
The indescribable agony and suffering of the common people are further compounded by the depredations and communal ferocity of the notorious PAC, about which we heard many stories from people who had come to the collectorate.
Meerut is in the grip of a callous, bureaucratic and inefficient administration, which is responsible for the huge loss of life and property. The Congress Party and its government at the centre as well as in Lucknow, will have to pay dearly for their irresponsible and opportunist game of appeasing the avowedly communal and feudal-motivatist forces for the sake of so-called electoral advantages.
Despite the tension which was in the air ever since the April disturbances in Meerut, the administration did not take
even the most elementary precautions like prior rounding up of well known criminals and anti socials seizure and confiscation of unlicensed arms, setting up of joint citizens' committees, etc., the intelligence system was paralysed. What happened to the oft repeated assurance given in Parliament that District Magistrates and Police superintendents will be held responsible for communal outbreaks in their areas?
Innocent persons of both communities are being tortured and massacred while the government goes on issuing futile "appeals" and warnings". Those forces responsible for planned and organised rioting are playing the game of national destabilisation and deserve to be suppressed with an iron hand. But Rajiv Gandhi fiddles, while the country burns.
New Delhi, 25 May 1987
New Delhi, 9 June, 1987
CPI Rajeshwar Rao's Letter to Prime Minister on Communal Situation
The communal situation in the country is alarming. The ghastly incidents that have taken place in Meerut, Delhi and Gujarat are a grim warning to all of us who hold our country's secular democratic set up and unity dear to our heart.
The rapid communal and religious fundamentalist parties and organisations of various hues are out to fan communal riots and polarise the people on the basis of religion in order to destroy our secular democratic set up and unity of the country.
Some of the self appointed religious leaders have the gump tion to issue hukumnamas/firman and edicts which challenge the very basis of our secular democracy and impose them on the people and even on the elected representatives of the people and high dignitaries.
If effective steps are not taken, the communal conflagration is bound to spread and engulf the whole country.
The imperialists are encouraging these forces in order to destabilise and balkanise our country.
The background of the present wave of communal violence is, of course, the Babari Masjid Ramjanambhoomi controversy.
which had erupted with the unlocking of the Bahari Masjid complex as a result of the Faizabad district judge's verdict.
I must say that the central government and the UP government cannot escape responsibility for this.
Now it is necessary that all patriotic and secular minded parties, organisations and forces should unite, setting aside petty partisan considerations, to defeat the game of the religious fundamentalists.
A two-pronged approach is needed to meet the menace of communalism. A united political campaign is needed to counter the communal poison, along with stringent administrative measures against the mischief-makers.
In conclusion, I have to bring to your notice the utter failure of the Meerut district administration in controlling the situation and allowing the rapid communal forces to run riot and cause immense damage to the lives and property of the people.
25 May 1987
CPI (ML) : Statement on Communal Violence
The CPI(ML) expresses its grave anxiety at the spurt of communal violence in large parts of the country. The foci of the Meerut and Delhi where the worst crimes against humanity have been committed in the form of ghastly murders, rapes, burning and looting. The police and administration have not only stood as silent spectators but have actively participated in the massacre of Muslims. In Meerut the PAC was responsible for murder of hundreds of people through firing and burning.
The Muslims all over the country are asking the question: Is any Muslim safe in his person in today's India? People who care for the unity and integrity of the country and boast of our great secular tradition must come forward at this gravest moment of India's history with a positive reply to the above question and accompany action. While Rome was burning, Nero was fiddling. Today, when Meerut and Delhi are burning, the third chieftain of the Nehru dynasty is busy in covering up his corrupt regime. At this critical juncture, the people should come forward to say a firm "NO" to the communal programme of Rajiv and his regime.
The CPI(ML) demands that the UP Government should resign immediately bearing moral responsibility for the massacre in Meerut.
The CPI(ML) also demands that the Central Government should appoint a commission led by a Supreme Court judge to enquire into the recent riots in Delhi and Meerut.
Sd/- Santosh Rana, General Secretary
25 May 1987
Indian People's Front (Uttar Pradesh) Statement on Communal Violence
Indian People's Front (IPF) expresses its grave concern over the continued communal violence in Meerut and holds the communal policies of Congress(I) governments in the Centre and Uttar Pradesh responsible for this holocaust.
Last year a conspiratorial decision to unlock Babri Masjid-Ram janambhoomi was taken at the behest of the Central government and with the active cooperation of the State's Chief Minister, Bir Bahadur Singh. This decision swept the whole country with a new communal frenzy, and Meerut has become its victim.
To divert also the attention of the people from the issues of corruption and anti-national activities of the political leaders sitting at the top of power in the centre, Congress(I) has engineered the riots in Meerut with the help of communal forces.
The IPF demands the dismissal of the State government and appeals to all democratic forces to combat unitedly the divisive policies enforced by the collaboration of the government and communal forces.
Lucknow, 25 May 1987
CPI(M) Politbureau Resolution on Meerut Disturbance
The Polit Bureau of the Communist Party of India (Marxist) views with serious concern the fact that the communal conflicts and tensions are not subsiding in Meerut but also spreading to neighbouring areas. The State Government and the Centre have obviously failed to come to grips with the serious situation. They have allowed the PAC a free hand in Meerut and this force has once again revealed its partisan and vicious character.
On the other hand, it is shocking that the authorities have prevented opposition leaders from visiting the affected areas to get a proper report of the developments.
In this situation, it is strange that the Prime Minister has not found it necessary to convene a meeting of the National Integration Council. This narrow approach must be given up. The CPI(M) demands the immediate holding of a National Integration Council meeting to take stock of the situation.
27 May 1987
BSP President Kanshi Ram on Muslim Community
Q: What is your relationship with the Muslim community?
Ans.: We feel that they are also becoming victims these days. They are also having problems. So those victims need our help and we will be able to help them and protect their constitutional, social, human and political interests.
A staunch follower of Dr B R Ambedkar, Mr Kanshi Ram believes that the root of all ills in this country is brahmaniyood—the domination of 15 per cent Brahmins, Thakurs and Baniyas over the 85 per cent scheduled castes, backwards and minorities.
(Source: The Telegraph, Calcutta)
Janata Party Joint Statement of Central Parliamentary Board
"The terrible communal holocaust in Meerut has continued for over a week resulting in the killing of hundreds. Even more tragic are the atrocities in which the PAC has indulged in Meerut culminating in the Malwa massacre and the killings on the Hindon river. Delhi has also been convulsed by riots. Such tragic happenings bring profound agony to the people and put the whole nation to shame.
"We are constrained to say that the Government has compounded its irresponsibility by arresting the prominent Opposition MPs who went to visit the disturbed areas. Thereby it has sought to draw a veil on the realities, and has only exposed itself.
"As the situation continues to worsen and casts its ominous shadow on other parts of the State, the Central Government
must intervene immediately and take effective steps to save further loss of life and property. We would also suggest that the PAC should be reorganised expeditiously in accordance with the National Integration Council.
"We would humbly urge all citizens to help ensure the peace and work for communal amity and brotherhood."
New Delhi, 28 May 1987
Chandra Shekhar Press Conference, New Delhi,
1 June 1987
The Meerut tragedy is the "most shocking incident in my political life". The police arrested 100 to 400 people on May 22 took many of them in five trucks to different places from the Hashimpura locality and some of them have been reported killed on the canal bank and their bodies thrown into the canal.
The law of the jungle prevails. "The tyranny of the PAC is evident. How can a police force inspire confidence if it works in a partisan manner?"
"If Mr Rajiv Gandhi has any compassion or any sense of duty he should have the cold blooded murder of scores of youths investigated. We cannot be silent spectators to the tyranny and callousness of UP Government and the Centre."
I request the Bar and People's Union of Civil Liberties to "assure the people that this sort of thing will be looked into, so that it will not be allowed to happen again."
If the Pakistanis were there in Meerut and responsible for the riots the Government has a duty to find the spies and foreigners and expel them. If they cannot, the Home Minister and the Chief Minister have no right to continue in office."
I got to know of the shocking killings along the canal when I met 17 year old Zulfiqar Nasir, one of the survivors. I then tried in vain to get in touch with the Home Minister, Mr Buta Singh and Mrs Kidwai. The incident occurred ten days ago, bodies have floated down the Ganga & Hindon canals. The Government is yet to explain the circumstances in which these persons were killed. Both the Union and state government are silent on the complaints made against the PAC.
If the whole country is being managed by foreign forces, what is the Prime Minister managing? The forces of the Government of India are seen only in Fair fax and Bofors affairs"
IUML Memorandum of 5 June 1987 to President
2 It is abundantly clear that the responsibility of the communal violence at Meerut in U.P and the inhuman heartrending massacre at Maliana rests on the local administration and the U P Government.
There are innumerable complaints of shocking excesses indulged in by the Provincial Armed Constabulary (PAC) While at several places the rioters enjoyed the patronage of the PAC at various other places like Hashimpura, Imliyan and at several others, the PAC itself broke open the doors of houses and indulged in a spree of loot and arson. Many were beaten up and there were mass arrests of Muslims dragged from their houses in various mohallas. The massacre at Maliana and the dead bodies with gun shots flowing down the Hindon river and the Ganganahar for nearly five days, have no parallel in the history of violence. There are reports that the dead were those who had been rounded up by the PAC.
4 It is a matter of grave concern that the arrested lodged in jails have been humiliated and mercilessly beaten, in some cases by the other prisoners in jail. A number of them had then to be taken to hospitals while some have died in jail.
5. In all fairness the present Chief Minister and the Council of Ministers be dismissed.
6. It is absolutely necessary that the PAC be totally withdrawn from Meerut, etc and the army and the central forces be put in total and effective charge.
A complete restructuring of the PAC with heavy participation of Muslims and other minorities be undertaken without any further delay.
10 The shocking, unfortunate communal holocaust needs to be discussed fully in all its aspects by the Parliament. We, therefore, request that the President be advised to summon an immediate session of Parliament.
Musawarat Demands withdrawal of PAC from Meerut, Induction of Army and Dismissal of UP Government
Statement by Mr Shababuddin (M P), Working President of AIMMM
The massive communal conflagration in Meerut has resulted from the initial failure of the administration and its subsequent inability to check the excesses of the PAC. It is surprising that peace-keeping has so far been left largely to the same unreformed and unrepentent force. It is well known that the PAC, notorious as a brutal and communal force, has lost the confidence of the Muslim community which has, therefore, good reasons to suspect its intentions when it moves towards or operates in a Muslim locality. It is not understood why the CRPF and other central forces have been deployed on peripheral duties and the PAC has the charge of sensitive areas. If the continuing cycle of confrontation, retaliation and repression is to stop, the PAC must be immediately withdrawn and replaced by the CRPF or BSF, and the Army must not remain on stand by orders but should take control of the riot affected areas.
There is a strange coincidence between the visits of the Chief Minister of UP, he at Allahabad or Meerut, and the aggravation of the communal situation. Perhaps they serve to make his police force still more brutal and partisan in their behaviour. In any case his incompetent and inefficient government has failed to control the situation and is responsible for the loss of hundreds of lives and crores worth of property. This government must be immediately dismissed.
Secondly, the communication gap between the administrative machinery and the Muslim community in Meerut is near total. Their SOS receive no response. Their leaders have no access to visiting dignitaries. Their localities have been sealed. Their lines have been cut and they cannot even reach the DM or SSP. The secular forces must immediately establish a permanent presence in Meerut by deputing a team of responsible leaders to act as a buffer as well as a media and to secure prompt and effective action when Muslim areas are under attack and plead for help and assistance.
A section of the local press is fanning communal fire by publishing concocted and exaggerated reports. Immediate steps should be taken to control such irresponsible writings.
The Mushawarat demands immediate visits to Meerut by the Prime Minister and the Home Minister. They should talk to the leaders of both the communities and all political parties and take effective steps not only to restore law and order but to lay firm foundations for communal peace and social harmony.
S. Shahabuddin, New Delhi, 22 May 1987
Statement on Meerut Situation
The All India Muslim Majlis e Mushawarat strongly denounces the brutal massacre of the Muslim community in Meerut and its environs by the PAC and holds the Government of Uttar Pradesh wholly responsible for the Meerut situation and demands its dismissal and the proclamation of Presidential rule in the State.
The Mushawarat demands that the PAC be immediately withdrawn from riot affected areas; its past behaviour should be inquired into by a high power Commission; it should be disarmed till criminal elements are screened out and it is reorganised in accordance with the provisions of the National Integration Council.
The Mushawarat demands immediate relief measures for the victims of police brutalities in Meerut, beginning with adequate supply of essential provisions to the curfew bound areas, in accordance with the guidelines of the Central Government.
The Mushawarat also demands adequate precautionary measures for the protection of the Muslim community on Id-ul-Fitr, which falls on 29 May 1987.
The Mushawarat calls upon the Muslim community all over the country to observe Id ul-Fitr in a simple and dignified manner and refrain from usual festivities as a mark of sorrow after the Meerut massacre.
The Mushawarat Central Council shall meet immediately after Id ul-Fitr to take stock of the situation.
New Delhi, 27 May 1987
NUBC Statement by Chandrajit Yadav
A Delegation of the National Union comprising of Ch Brahm Perkash, Chandrajit Yadav, Syed Shahabuddin MP, Maulana S A Hashmi Sh K R Sunda, N G Uke and Prof M K Sami visited the riot affected areas of Meerut City and met hundreds of people belonging to all sections of the society in several riot affected areas. The Delegation also met the District Authorities including the District Magistrate and SSP Meerut.
It was admitted by all that since Independence unfortunately Meerut City has seen the biggest communal riot which has completely shattered the secular fabric and created a deep communal divide. Life of the city has been so seriously affected that it will take years to restore the harmony between the different communities and also the economic life of the common people. Hundreds of people have been killed and several crore worth of property has been destroyed. The traditional industries, such as, handloom, sports good manufacturing units, scissors have been severely affected. Hundreds of units have been fully destroyed and burnt to ashes. Thousands of residential houses and shops have also been burnt. It is a great tragedy that the district which was the centre of our freedom struggle and which had produced thousands of valiant patriots has, unfortunately, now become the centre of communal riots.
There is a big gap between the number of deaths given by the District administration and the people of Meerut City. The general assessment is that not less than 500 people have been killed. Some say the figure might reach 1000. About 3 thousand people have been arrested since the riot started and only 60 of them have been released, mostly children in the age group of 8 to 12 years. A few old people whom we met were severely beaten by police in police lock ups. All of them were above 60 years of age.
Rumours as usual, in every communal riot played havoc in Meerut city. That affected both the communities and later on invariably in all cases it was found that all rumours were baseless and no effective measures have been taken even till today. In our opinion the Administration dealt with the situation in the
most unimaginative manner. The mishandling of law and order situation was one of the major factors which aggravated the situation and angered the people.
The District authorities deliberately did not seek the cooperation of the local leaders and respectable people in Mohallas. This, in our opinion, has become the accepted norms in UP with law and order authorities that they shun the cooperation of the public and depend only on their administrative machinery which, unfortunately, is looked upon with suspicion as it has lost the image of impartiality. Their rude behaviour with common people and unbecoming behaviour with women folk and indiscriminate beating creates an allergy against the UP PAC when they reach the spot of the trouble. This is what exactly happened in Meerut. In Meerut even opposition leaders and MPs were prevented from visiting the city including Syed Shababuddin MP. This is highly deplorable.
UP is sitting on a volcano according to the admission of UP Government. 32 Districts have been identified as sensitive districts. One must give serious thought to why the communal virus has spread over almost two-third area of U.P. We feel that for many years the ruling party leaders and administrators have indulged in creating a communal atmosphere for their political advantage. This infected the entire atmosphere and provided a golden opportunity for anti-social, rowdy, communal and biased elements in the administration to create this unfortunate condition.
During our visit, we came across several shining examples which instil hope in secular values and harmony between various communities. There were people in both the communities who came forward and saved the lives of the people belonging to other communities even at the risk of their own lives. One such shining example is of Prof. Harpal Singh whose son, by profession a doctor, became the victim of rowdy elements. Prof. Singh and his wife, the entire family even after the loss of their only son, has great faith in secular values and are anxious for restoration of peace and normalcy in the city. Without even an atom of bitterness even after this great tragedy, they offered their service for restoring communal harmony in the city. Such people must be publicly honoured.
Suggestions
Our suggestions are as under
1 In Meerut City and District Peace and Harmony Committees should immediately be formed from Mohalla City to District Levels where respectable people of all sections of the society should be taken. They should meet and tour their respective areas and appeal to the people for restraint and calm.
2 The list of arrested people should be displayed at the Police Station and also the important people in the Mohalla should be invited and shown the names so that there is no scope for rumours that people are missing or killed. This will go a long way in pacifying the people and creating normalcy. We suggested it to the District Magistrate and they promptly agreed to one suggestion. It is hoped that it will be done.
3 For a few more days the Army should be allowed to stay in the City and PAC should remain in the background. There were serious complaints against the PAC Platoons and we were told that even two PAC Platoons were withdrawn on the basis of these complaints.
4 UP Government has announced two enquiries—one Judicial Inquiry of the incidents in Maliana area and another Gyan Prakash Inquiry Committee for Meerut Communal riots. In our opinion only one High Level Judicial Enquiry Committee should be appointed and asked to submit its report within 2 months.
5 Facts have been brought to our notice that people arrested by PAC platoon in Hashimpura Mohalla were taken outside the City and shot dead by PAC Jawans. Some survivors have harrowing tales. We demand that Hashimpura incidents should also be included in the Judicial Enquiry.
6 There is a big gap between the figures of deaths as given by the District authorities and as given by the people of Meerut City. We demand that on the pattern of Moradabad if the kith and kin submit an affidavit that some one has been killed in his/her family in the riot affected areas, they also furnish an indemnity bond and in all such cases compensation should be given to those families.
7. On the pattern of 1984 Delhi riots, compensation and aid should be given to all affected families that will also help rapid rehabilitation of industries and business.
8. Our demand is that PAC should be overhauled and re-organised. It should rather be replaced by 'Anti riot Force,' with a composite particularly having adequate representation from all sections of society particularly minorities, Scheduled Castes, Scheduled Tribes and Backward Classes.
9. During the last few years there has been mushroom growth in the formation of Senas and Dals on the communal divide. We demand that such organisations, Senas, Dals should be immediately banned.
New Delhi, 4 June 1987
A Joint Statement of Intellectuals
Like most of our compatriots we feel deeply disturbed by the recent chain of the communal riots. The orgy of violence has destroyed innocent lives and inflicted an all round misery.
The brutal conduct of some men in uniforms has disturbed us the most. Their mad spree of killings, torture and disgracing the citizens should cause concern both on humanistic grounds and endangerment to stability of the secular state. If administration gets divided on communal and caste basis and the paramilitary and police forces are permitted to endanger the lives and property of the people, then how do we sustain the Nation's cohesiveness and unity?
The political leaders must tell the nation as to why they did not neutralise well in time, the communalists and the fundamentalists who were organising the yatras and the rallies and arming the Senas and rousing communal passion?
Unless some credible and preventive steps are urgently taken, we may still see repetitions of these ghastly deeds.
Foremost, the government must prosecute all those members of the P A C., and police who have disgraced their uniforms. Their misdeeds must be treated at par with treason and tried by special courts. The judicial enquiries will, as in the past, take long time and end in nothing.
Even at this late hour recommendations of the National Integration Council regarding creation of composite anti-riot
forces be implemented to infuse confidence.
The Ram Janam Bhoomi-Babri Mosque syndrome has been exploited by both sides to rouse passions. This must stop and the issue be settled by negotiations, arbitration or judicial process. In the meantime all demonstrations must be banned.
This is the moment when all secular minded citizens and institutions must assert and unmask the communalists and the divisive elements in all communities who misuse the name of religions and cause havoc to our unity. They are the ones who are destroying our Nation.
I K Gujaral, former Union Minister
P N Haksar, former Chairman Planning Commission
Rajinder Sachar, former Chief Justice
Subhadra Joshi, Former M P
Kuldip Nayar, Journalist
Badar Ud din Tyahji, Retd Secretary
M K Rasgotra, Retd Foreign Secretary
M Farooqi, Secretary C P I
D R. Goyal, Journalist
Brij Mohan, former President D P C C
Inder Mohan, President P U C I
Yunus Dehlvi, Journalist
Qamar Rais, Professor
Joginder Sharma, CPI(M)
A R. Shervani, Journalist
Rashid Talib, Journalist
National Federation of Progressive Writers:
Statement on Communalism
After studying the first-hand reports of current riots in Meerut and Delhi the Secretariat of the National of Progressive Writers notes with profound grief and indignation that the governments at the Centre and in Uttar Pradesh must squarely bear the blame for unprecedented loss of lives and property in the communal carnage.
For the last many months, Hindu and Muslim communalists were not only openly building up communal tensions, but were also preparing for launching murderous crusades. In many towns and cities, the communalists were collecting firearms and
other weapons, and bandit groups were being organised in the name of various senas.
The government knew about it, but took no preventive steps either to physically disarming the miscreants or to diffuse the situation by solving the Babri Masjid problem through political and also the deteriorating economic between the two sides in dispute.
The secretariat of the NF PW is constrained to view this criminal negligence as a deliberate attempt of the circles to create a diversion for the people who were quite agitated by some cases of corruption at the highest level and also by deteriorating economic conditions, thanks to the ruling groups' commitment to the ill-intentioned advices of the US controlled World Bank.
It is now revealed that even when the riots lurked round the corner, the law and order machinery was knowingly weakened in the riot prone areas in the name of sending additional police force for the ruling party's rally in Delhi. Afterwards when the riot broke out there was an unexplained delay in refurbishing the law and order machinery.
The ill-famed PAC was posted in Meerut in predominantly Muslim areas without taking into account its past communal and predatory record. With PAC around, a tragedy like that of Maliana, could have been very well predicted.
The secretariat of the NF PW considers the imperialists conspiratorial threats of destabilisation quite serious, but warns the government that it cannot hide its efficiency and sins behind it and the people would not allow to use it as a shield to deprive the working people of their bread and to enrich the exploiters.
The secretariat of the NF PW calls upon the government to at least return to its professed policy of secularism and restore its credibility in controlling the law and order situation. It must ensure that Meerut is not enacted elsewhere. It calls upon the government to immediately put an end to the Babri Masjid-Ram janambhoomi controversy by declaring it a protected National Monument and putting it under the Archaeological Department.
It appears to all sections of the secular people to energetically support our demand and join the struggle to bring sanity
to victory and defeat the humanism culture that is being spread in the name of religion.
The NF PW has always considered the fight against communalism as a primary task of all cultural workers and reiterates its demand for ensuring separation of religion from politics and ban on formation of communal political parties and their participation in elections.
(Source New Age, 14 June 1987)
Extracts from Union Minister For Tourism Mufti Mohammad Syed's Letter to the Prime Minister
Dear Prime Minister,
The presidential election is over. As a congress worker, who has been extremely worried about the recent development in the country I feel a sense of pride that our party has demonstrated exemplary unity at this juncture. Every congressman and woman has risen to the occasion to meet a serious challenge. But there is no room for complacency. The question remains Where do we go from here?
After coming to Delhi, I have been observing more closely than earlier, the events in the rest of the country. The monster of communalism is eating into the vitals of national unity. After my visit to Meerut, I had occasion to apprise you of the horror I witnessed there. If victims of communal carnage lose hope of safety from the administration charged with the duty to protect them, if Congress workers become victims of communal passion, if the government is dilatory and negligent in dealing firmly with communalism, how long will we be able to safeguard national unity? If the Congress does not uphold the banner of secularism with seriousness and resolute determination, who else will do? What I saw in Meerut has shaken me to the bones. What has saddened me more is the inadequacy of our response. For us in Jammu and Kashmir, secularism has been the most cherished ideal. To see the secular fabric of the country being torn asunder is an unbearable agony for workers like me. In fact, because of our acts of omission and commission in Jammu and Kashmir and other parts of the country, the evil of communalism is gaining strength even in our state.
As a nation and as a party, we are standing at the crossroads of history. The legacy of socialism, secularism, nationalism and democracy bequeathed to us by our great leaders is in peril. We cannot successfully meet the present challenges unless we act with the sagacity and dedication to build a strong and united party which will carry credibility with the people.
RESOLUTION PASSED IN THE TWO DAYS NATIONAL EXECUTIVE IN MEETING OF ALL INDIA QAUMI TANZEEM HELD IN NEW DELHI ON 28th & 29th JUNE 1987
Qaumi Tanzeem has noted with great satisfaction that the meeting of the standing committee of the National Integration Council headed by Prime Minister, Shri Rajiv Gandhi felt deep concern over the deterioration in the communal situation. Prime Minister rightly pointed out that a multi-pronged approach was needed to deal with the menace.
It suggested a more active role for political parties in promoting national integration, strict adherence to the three language formula and rigorous enforcement of laws to punish communal and disruptive elements.
Qaumi Tanzeem hails the outcome of the meeting and calls for the implementation of those decisions and suggestions which might go a long way in settling the issues so sensitive like communal, sectarian and regional.
Even in the past the council has made concrete suggestions, but unless some serious thoughts are given and within that framework some corrective and remedial measures are adopted, nothing could help the situation improve.
Hence it is necessary that efforts are made on administrative, social, educational and legal fronts to promote national integration and to eradicate communalism.
The Tanzeem endorses the stand taken by the council that all-out effort should be made to implement that 15 Point Programme for the minorities. We also appreciate that it was felt necessary to create job opportunities for the minorities to which due representation should be given in the services and
the police forces
The Tanzeem appreciates the concern shown by Prime Minister, who is also the Chairman of N.I.C. who stressed the need of concentrating on communalism along in that meeting and while referring to the role of administration, he called for infusing a national outlook in the regional services. He emphasised that inculcating a national outlook among the services was a very important part of national integration.
The Qaumi Tanzeem feels perturbed over the role of law enforcing agencies and a part of administration specially in the context of Meerut not and none other than the Home Minister, Shri Buta Singh has admitted that the administration had proved itself unable to cope with the situation.
It is also disturbing to note that communalism is assuming serious proportions and challenging the fundamentals of the Constitution. We support the view-point of Mr. Rajeshwar Rao that no religious leaders should be treated above the constitution and the laws of the land.
The Tanzeem feels that the sense of insecurity has overtaken the Muslim community and they feel quite shaky now and if this continues they will get alienated more and more from all secular parties. In order to create confidence in them there may be short term and long term plans.
There is no doubt about that the tension being built around Ram Janambhoomi-Bahri Mosque over the period of two years has played its role in Meerut riots. This can not be ignored for a long time and some political solution must be found. This will expose the real game of those political parties who arouse passions but can never sit to find out the lasting solution.
Some structural changes could be effected in P.A.C. and other administrative wings of the Govt. Immediate relief should be provided to those who need help. Large scale arrest and harassment must be stopped and those found innocent should be released forthwith.
Members from both the communities with secular credentials should be brought together and let them march together in disturbed areas, which would help remove mistrust.
Those completely uprooted should be rehabilitated. There should be survey of the extent of loss and a comprehensive
rehabilitation programme should be chalked out. The support of genuine voluntary organisations wedded to secularism must be enlisted and they should be put into action to help create atmosphere congenial to peace and undertaking.
The Tanzeem has noticed that gossips and rumours cause great harm and play with human lives and properties. There should be some permanent arrangement to nip such mischievous game in the bud at the right moment and deterrent punishment should be meted out to those found spreading rumours. We would rather suggest a kind of permanent machinery be set up to deal with this mischief mongering.
Communalism and communal violence are interrelated but not the one and the same thing. Communalism is an ideology and those who propagate it with sustained efforts are the generators and the carriers of the communal germs. They impress upon people and get communal followers but these followers are just the victims, their social condition also propel them in that direction. These people need to be educated to overcome their communal thinking and prejudices. On the other hand, the generators of communal germs should be branded as the enemies of the society and to be given no quarter.
The Tanzeem considers both the communities—Hindus and Muslims ethnically the same and have strong bonds of a historical and geographical past. They have to live together and work collectively for their progress and livelihood in India. Any breach attempted to divide them must be filled—for none is an aggressor and none is outsider. They are all sons and daughters of India.
The Tanzeem suggest that separation of religion from politics must be given substance and political will is required to do away with the forces exploiting religion for narrower ends.
LIST OF THOSE KILLED BY PAC IN MALIANA FIRING
1. Mudassar Siddique S/o Aji Liyakat Ali Age 26
2. Mohammad Ashraf S/o Karim Baksh 40
3. Gulfam S/o Anwar Ahmed 22
4. Tasleem S/o Arif Ahmed 20
5. Rafiq Ahmed S/o Basheer Ahmed 26
6. Nasir Ahmed S/o Vazeez Ahmed 24
7. Kaniz Fatima W/o Md. Yameen 42
8. Mehmood Ahmed S/o Aziz 35
9. Akbari W/o Khairati 50
10. Naseem Bano W/o Mohammad Ahmed 30
11. Mohamed Arif S/o Mohammad Ahmed 15
12. Mohammed Waris S/o Mohammad Ahmed 3
13. Aftab S/o Mohammad Ahmed 3
14. Bablu S/o Ali Hasan 3
15. Tarannum S/o Mohammad Ahmed 12
16. Shahjahan Begum D/o Amar Deen 5-6
17. Mohammad Ali S/o Unknown 70
18. Ibrahim S/o A. Sattar 10-11
19. Yusuf S/o A Sattar 22
20. Asghari W/o Abdul Sattar 40
21. Yunus S/o Abdul Sattar 15
22. Banu D/o Abdul Sattar 24
23. Nurjahan D/o Hbdul Sattar 10
24. Shabnam D/o Abdul Sattar 3
25. Munni D/o Unknown 22
26. Zaibunnisa D/o Sattar Ahmed 21
27. A. Sattar S/o Md. Ati —
28. Mahmud Faish Imam Masjid, Sanjay Colony 60
29. Mohammad Akbari W/o Itwari 65
30. Mage Khan S/o Kabul Khan 40
| No. | Name and Details | Age |
|-----|---------------------------------------------------------------------------------|-----|
| 31. | Bilkis D/o Mage Khan | 17 |
| 32. | Raeesa D/o Mage Khan | 13 |
| 33. | Kashmiri W/o Mage Khan | 35 |
| 34. | Rasheed S/o Md Shakur | 45 |
| 35. | Idan W/o Rasheed Ahmed | 40 |
| 36. | Jumma S/o Unknown | 50 |
| 37. | Kalu S/o Hussain Baraksha | 45 |
| 38. | Bashiran S/o Azmuddin | 46 |
| 39. | Zattun W/o Chandkhan | |
| 40. | Zinat (divorcee) W/o Navabkhan | |
| 41. | Raziya W/o Sagheer Ahmed | 40 |
| 42. | Sagheer Ahmed S/o Ismail | 45 |
| 43. | Sakhavat S/o Unknown | |
| 44. | Chimman S/o Unknown | |
| 45. | Anwar S/o Unknown | |
| 46. | Rafique S/o Afsahdeen | 45 |
| 47. | Nafis S/o Rafique | |
| 48. | Khalifa S/o Rafique | |
| 49. | Raeesuddin S/o Rafique | |
| 50. | Munni D/o Rafique | |
| 51. | Husna D/o Rafique | |
| 52. | Three daughters in-law of above family have also been killed whose names are not known | |
| 53. | Mustaqum S/o Unknown | |
| 54. | Akhtar S/o Unknown | |
| 55. | Saeeda W/o Ismail | |
| 56. | Shakura S/o Ajjab Mehr | |
| 57. | Iqbal S/o Shakura | |
| 58. | Rehana D/o Abdul | |
| 59. | Mohammad Ismail S/o Hameed | |
| 60. | Nawab S/o Md Ismail | |
| 61. | Shahzad S/o Md, Ismail | |
| 62. | Naseem D/o Md Ismail | |
| 63. | Vakeela W/o Name not readable | |
| 64. | Zulfiqar S/o Unknown | |
| 65. | Mubin | |
| 66. | Akhtar | |
| 67. | Akbar | |
| 68. | Asghar | |
When India was tragically partitioned in 1947, it was hoped that the country had seen the last of the Muslim problem. Unfortunately, today that problem is as acute as ever. Only thinking men and saints who view the totality of life can comprehend the problem, and it is only they who can help us resolve this problem amicably, sensibly and finally.
First, we must clearly understand that there is nothing called Indian Muslims—that is, Muslims who happen to be living in India. What we have instead are Muslim Indians—Indians, who happen to be Muslim. Muslim Indians are nearer to Hindu Indians than to Muslims in West Asia. Less than 0.1 per cent of Muslim Indians can honestly trace their ancestry to outside India. Muslim Indians and Hindu Indians are blood brothers.
Over the centuries, many Hindus embraced Islam as a result of force, temptation and casteism. That state power was not the only factor in conversion is proved by the fact that Uttar Pradesh, where Muslims ruled the longest, has a huge Hindu majority whereas East Bengal where they ruled for a much shorter time, has a Hindu minority. However, even after conversion, Muslims retained many of the attitudes and institutions of the old faith, complete with the ideas of right and wrong and high and low, of amulets and pirs in life and of chadars and lamps in death. For Sheikh Nizamuddin Aulia of Delhi, "a command of the Pir is like a command of the Prophet" and a
visit to the Pir's tomb is "spiritually more exhilarating than a pilgrimage to Mecca"
Bahns Ban
Kings fought each other and they even invoked religion. But these were basically power struggles and not religious wars. The rulers slaughtered people but ordinary Hindus and Muslims did not go about killing each other. Hindu Muslim riots are a phenomenon of the British period.
Many Muslim rulers gave preferential treatment to Muslims. But they also gave special concessions to 'Multanis' and Hindu traders in the interests of economy. While the orthodox kiazis sided with the ruling elite, the sufi saints criticized both—kings and kiazis—for their corruption and hypocrisy.
Babur, the first Mughal ruler, banned cow slaughter and this ban continued until the British took over. Just like old Hindu kings, the Mughals drank only Ganga jal and Akbar and his successors were not circumcised. Jahangir was half Hindu and Shah Jahan was three fourths Hindu by birth. They called themselves Jahangir (Keeper of the World) and Shah Jahan (World Ruler) to the intense irritation of the Caliph and the Turks. Aurangzeb removed the Kalma from coins on the ground that Hindus also touched them, and Akbar made it clear that "India cannot be ruled by Arab rules of a thousand years ago."
Aurangzeb was a violent fanatic. But even he was more of a king than a momin. He snubbed the Caliph in Turkey and neither informed him of his accession to the throne nor asked for his recognition and blessings. Indeed, the Mughal rulers regarded themselves as also Caliphs. Aurangzeb imposed the jiziya on Hindus in the name of Islam but the real reason was that the treasury had been drained because of his wasteful wars. He fought to keep Central Asia attached to Delhi and made the revealing remark that while other Indian princes fought for their respective principalities, he was the only one to think of "All India." Here was the national core of the Mughal empire.
In the Mughal and provincial courts, Dussera and Diwali, Basant and Holi were celebrated as national festivals. While the Hindu saints rejected fanaticism and resisted tyranny, they had
nothing against Islam as such. Tukaram said: "Never forget to take the name of Allah first." Even immigrant Muslims developed an intense love for India. For Amir Khustau, Delhi was more important than Mecca. Mecca, he said, should go round Delhi in reverence.
The popular impression that Muslims were pro-British, is only a half-truth. Muslims resisted British power more than the Hindus. Tippu Sultan was the only Indian ruler who died fighting the British before the 1857 rising. And Muslims resisted even British culture and the English language much more than the Hindus. For all these reasons, for a long time the British were more anti-Muslim than anti-Hindu. The Muslims took a much more active part in the 1857 uprising than the Hindus. Until then, Hindus and Muslims were recruited to the same army units, after the revolt, they were put in separate regiments. The policy of divide et impera had begun!
After 1857, the British cursed every Muslim rebel as "badmash" (rogue) while the Hindu rebel was only dubbed a "Pandy" (after Mangal Panday). It was only after 1905, when the Hindus began to talk of swaraj, and to practise swadeshi, that the British became decisively anti-Hindu and pro-Muslim—and the Muslim became increasingly pro-British.
That being the historic perspective, it should not be difficult for Hindus and Muslims to resolve their problems amicably and finally on the basis of good sense, good faith and common destiny.
True History
First of all the history of India, written by the British from the imperial angle and mechanically regurgitated by the present authorities should be replaced by one written by Indians from the Indian perspective. Such a history should be people-oriented and not elite-centred. This history should not suppress unpleasant facts, but it should not overlook pleasant facts. Such a true and balanced history of the Indian people will go a long way in improving community relations.
The life, limb, property and honour of all citizens must be safe. District officers must be held fully responsible for law and order. In cases of violence there must be a prompt inquiry.
The findings of such inquiries must be made public, the guilty punished, and the victims compensated.
A feeble Minorities Commission does no good to anybody. We should instead have a human Rights Commission at all levels to deal with complaints of discrimination on the basis of caste, creed or community. Once that is done, Muslims should refrain from mouthing all kind of grievances real or imaginary. Simultaneously, Hindus should stop eyeing Muslims with suspicion and reproach.
Hindus nurse a grievance over the excesses of invaders who happened to be Muslim. Muslim Indians should look upon these scourges of humanity as invaders, and not as fellow Muslims (Muslims in Sind, Pakistan, have already shown the way by acclaiming Raja Dahir Sen, the last Hindu ruler of Sind, as a great hero, and denouncing Mohammed Bin Qasim as an invader and enslaver). By way of goodwill Muslim Indians should return the more important temples known to have been forcibly converted by tyrannical kings into mosques to Hindus.
Muslim Indians should not try to be more Muslim than even the Mughals. They should heartily celebrate Dussera, Diwali and other Hindu festivals and Hindus should not hesitate to observe Id and Muharram in mixed localities as was the practice until only 50 years ago. If non Christians can exchange gifts and greetings at Christmas why can't Hindu and Muslim friends do the same for Id and Diwali? Incidentally Tazias are only a Muslim adaptation of Rathyatras and Aurangzeb had actually banned them as "un Islamic".
Urdu is 75 per cent Hindi. Just as Braj Avadhi, Bhojpuri, Maithili and Rajasthani are accepted as forms of Hindi, such Urdu works as are not highly Persianised and, therefore, not easily understood should be included in the Hindi curriculum of schools and colleges. This should not, however, affect the position of Urdu in the Persian script.
Polygamy should be banned in India, as it is already banned in Pakistan and several other Muslim countries, to protect Muslim women. In addition the step would also allay Hindu fears of a disproportionate increase in the Muslim population.
New Laws
Muslim members of the Constituent Assembly had urged proportional representation in the legislatures to ensure equitable representation of all communities. Muslim leaders still feel strongly about it. And they are right. Outside of Anglo-America most democratic countries have proportional representation and Indian electoral laws should also be reformed accordingly.
Indians—Hindu and Muslim—should help each other wherever possible. For example, Muslim Indians wishing to visit Pakistan and Muslim Pakistanis wishing to visit India, should be given passports and visas without inordinate delay.
Likewise Muslims should try to help Hindus living in Muslim countries. For example, Hindus working in Arab countries are not allowed to build temples. A Hindu dying in these countries cannot be cremated there. Muslim Indians should plead the case of Hindu Indians on these issues. It is possible that Arab countries will take their time to do the needful, but Muslim Indians should champion the cause of their Hindu brothers. Gestures like these generate considerable goodwill at the personal level.
Above all Hindus and Muslims must learn to respect each other's religious leaders. As Gandhiji rightly said "There is in Hinduism room enough for Jesus, as there is for Mohammed, Zoroaster and Moses." Hindus should not hesitate to respect Mohammed as a great man—even as they admire a Lenin or a Napoleon. And Muslim Indians should not hesitate to adore Rama and Krishna. Let them remember that Muslim Iran prides itself on its pre-Islamic heroes like Rustom and Sobrab; Muslim Egypt glories in pre-Islamic Pharoahs, Rameses and Cleopatra; and Muslim Indonesia has Sri Rama as its main national hero.
If Hindus and Muslims ponder over these matters dispassionately and respond to suggestions positively, the ground could be prepared for a great reconciliation in the subcontinent. That will be a great day for all mankind.
—From The Statesman, 30 May 1987 |
YORK NORTHWEST AREA ACTION PLAN
Progress Report, incorporating Issues and Options Consultation Summary
Summary
1. This report reviews progress to date on the York Northwest Area Action Plan, and gives an overview of comments and headline results arising from consultation on the Issues and Options Report between November 2007 and January 2008. Members are asked to note the progress to date and the use of the consultation findings to inform the preparation of the Preferred Options. Members are further asked to agree the approach and programme for the preparation of the Preferred Options, which are also set out in this report.
Background
2. The Area Action Plan (AAP) for York Northwest will be one of a suite of Development Plan Documents in the Local Development Framework. The AAP will be used to guide and control the development of this major area of change. The timetable for preparation of the AAP was agreed at a meeting of the Executive on 27 February 2007 and has been incorporated into the Local Development Scheme (LDS).
3. At a meeting of the Executive on 25th September 2007 Members agreed the content of the Issues and Options report, the preparation of associated documents and for public consultation to take place.
Progress Report
Progress to Date
4. Good progress has been made on the initial Issues and Options stage of producing the Area Action Plan. Programme milestones identified in the Local Development Scheme have been met, with issues and options consultation carried out between November 2007 and January 2008.
5. A range of evidence base documents, including both city wide studies and specific York Northwest studies have informed the preparation of the Area Action Plan to date. Outstanding evidence base studies are now also well
advanced; it is anticipated that the councils Retail Capacity Study will be reported to members in June, the PPG17 Open Space Study in July, and the Employment Land Review in September 2008. Other work, including transport modelling, specific retail work and a specific York Northwest Open Space Study is also being progressed. Viability assessment work will be required on the emerging Preferred Options to ensure deliverability considerations are addressed.
6. Transport modelling will look at the infrastructure required to accommodate general growth in the city and the additional infrastructure needed to deliver development of York Northwest. The work will inform the development of a tool to allocate costs of infrastructure to different stakeholders and mechanisms for securing funding. The Council is also progressing a number of transport initiatives which will inform the transportation requirements and funding for the Area Action Plan, including, Access York (park and ride and outer ring road projects), new growth point infrastructure bid, cycling city status, tram train and the Hopgrove roundabout.
Stakeholder Activity
7. York Central stakeholders are making progress in a number of areas in anticipation of seeking approval from their various boards to enable the site to be marketed. It is hoped that a developer will be appointed in advance of the Councils Preferred Options consultation, which will enable developer involvement in the plan making process as early as possible. Preferred Options Consultation is scheduled in the Local Development Scheme for January 2009.
8. Associated British Foods have appointed contractors who are due to commence demolition on the site on 6th May. The Councils Countryside and Landscape Officers, alongside consultants appointed by Associated British Foods, will ensure that the works are carried out in a way that does not compromise onsite ecological interests. It is anticipated that demolition works will be completed before the end of this year. A team of consultants, headed by Rapleys Planning Consultants, are carrying out background study work to inform the development of the site.
Forthcoming Work
9. The Local Development Scheme allocates the period between February and December 2008 for “consideration of representations on Issues and Options and preparation of Preferred Options document” in respect of the York Northwest Area Action Plan. This report summarises the first stage of the Preferred Options work. A detailed three stage program of information analysis, development of conceptual options and Preferred Options report has been produced to structure work over the coming year. Greater detail on this program is given at paragraph 39.
Issues and Options Consultation
Consultation Plan
10. The consultation was carried out in accordance with a plan and timetable agreed with Members. The plan, set out at Appendix 1, took into account the views expressed by the public on a pilot consultation strategy for York Central in 2006. Given the wide range of interests involved in the project at local, city wide and sub-regional/regional level, a variety of consultation methods and techniques were used. Consultation on the Issues and Options Report took place between 5th November 2007 and 14th January 2008.
Consultation Information
11. A range of information was produced and tailored to reflect the needs of the different consultation events and diverse interests in the area. A leaflet was prepared to raise awareness of the consultation and to tell people how they could get further information. A summary report was used to give a short overview of the matters raised in the full Issues and Options report. A short questionnaire was attached to the leaflet, and a more detailed public comment form was distributed with all copies of the full and summary report (Appendix 12). A Background and Frequently Asked Questions leaflet was also produced to provide a guide to the planning process. Presentational material was also produced for the workshops and exhibitions.
12. All documents and details of the consultation were outlined on the York Northwest website and an on-line survey was available for comments to be submitted. In addition, a full set of documents was also made available at all libraries and at the Guildhall/ St Leonard’s Place planning reception.
Consultation Events and Response
13. A considerable level of interest was shown in the consultation, and a wide range of comments have been made. The analysis section of this report provides an overarching summary of the headline points from the various consultation events (see Appendix 3). Summary reports of individual consultation events and responses are available on the Councils web site and on request (Appendices 6–11).
14. A total of 470 individuals and organisations were sent consultation documents, including specific consultation bodies and individuals and organisations with a direct interest in the area. 294 responses were received, including 245 responses to the questionnaire and leaflet. A plan giving a detailed breakdown of community responses is attached at Appendix 2.
15. Around 365 people participated in workshop, public exhibition and focus group events. A further 53 residents attended the presentation at Holgate ward committee meeting and 37 residents attended the Acomb ward committee meeting.
1) Workshops: Transport, business and commerce, and design/environment workshops were externally facilitated by an independent public engagement company, Icarus. A further workshop was held for the York Central Task Group
and was facilitated by the council. 101 representatives with a specialist knowledge, or interest, in the topic for the session attended the workshops. Individual reports for each workshop were prepared and circulated to all attendees and are available on request. Icarus also prepared a report summarising the findings of the 3 consultation workshops, which is available on request. An overarching report summarising the findings of all the workshops is available on request (Appendix 9).
2) **Public Exhibitions**: Four public exhibitions were held – one in a city centre location, and three others at locations accessible to existing communities living near the York Central and British Sugar sites. The exhibitions were advertised at each of the three local ward committees, on the leaflet, in the local press and on the Council’s web site. Around 230 people attended the exhibitions. Future involvement by young people in the consultation process was raised as an issue.
iii) **Focus Group**: A focus group of residents of York was facilitated by Icarus, an independent public engagement company. Ten people were recruited through the Talkabout Panel, which is a representative panel of 2000 residents who live within the city and make comments on a wide range of issues. A report setting out the views of the focus group is available on request (Appendix 10).
iv) **Presentations**: Presentations were made to the Environment Forum, the Inclusive York Forum and the York Central Steering Board. Responses from the Environment Forum and partners to the York Central Steering Board, Yorkshire Forward and CBRE (representing Network Rail/National Museum of Science and Industry) are included in the Summary of written representations (Appendix 8). A report setting out the views of the Inclusive York Forum, a forum of hard to reach groups, is available on request (Appendix 11).
**Analysis**
16. The Issues and Options report is structured around the key areas of vision and development objectives, sustainable community, employment, housing, social infrastructure, transport, culture and tourism, open space and built sporting facilities, urban design and delivery. The design of the questionnaire/public comment form and all the events were structured around finding out views on the key theme areas. In addition, the views on the boundary for the York Northwest area were also sought.
17. This section of the report gives an overview of the headline points arising from all the various consultation events and representations received in relation to each of these areas. Further detail can be found in the overarching summary report, attached in Appendix 3. Due to the significant quantity and detail of feedback given, for the purposes of this report, summary reports of individual consultation events and responses have been prepared and are available on request (and on the Councils web site), in Appendices 6–11. The full range of comments will be used to inform the preparation of the Preferred Options for York Northwest.
Boundary
18. Various sites were suggested for inclusion in the Area Action Plan boundary. These include the Royal Mail sorting office, the Monkhill Confectionary (Craven) site, the Acomb Water Treatment Works, part of Clifton Ings and various other open spaces around British Sugar, including Poppleton Lane Allotments, the Civil Service Sports ground and the current Manor Field Sports Ground. A plan of the proposed additional sites and rationale for inclusion, is attached at Appendix 5. It is anticipated that a report will be brought to members later this summer for comment on the possible Area Action Plan boundary.
Vision and Objectives
19. There was overall positive support for the draft vision with between 64% and 79% of people supporting the statement. Between 10-13% of people objected to one or more aspects of the draft vision. Particular support was registered for the themes of sustainability, integration, innovation, quality, vitality (liveliness) and quality of life. Several specific issues were considered to be under-represented, and were recommended to be included in the vision, including vitality (liveliness), provision of green transport provision, representation of leisure and retail activities and the importance of York’s historic character. Viability was also mentioned at the stakeholder workshops, but not by public consultees. There were several requests for the vision to be simplified and to reflect the specific purpose of the individual sites.
20. The objectives were prioritised in the public feedback and at the workshops. Two of the highest objectives from both were for ‘integration with the city/surrounding area’ and ‘meeting housing needs’. The public response also gave high priority to the provision of ‘high quality health, education, retail and community facilities’, whilst the need to provide ‘exemplar high quality development in terms of building design and spaces’ was given high priority by the workshop participants. It is anticipated that a report will be brought to members later this summer for comment on the possible Area Action Plan vision and objectives. The full quantitative feedback received is outlined in Appendix 4.
Creating a Sustainable Community
21. Creating sustainable communities is an overriding theme for the development; key issues are highlighted in this section of the report and in each theme area.
22. Overall, a broad level of agreement was expressed over issues presented on creating a sustainable community, though some reservations were expressed on most issues, and in particular on flood mitigation, design/housing and contamination. All participants were in full agreement that York’s unique characteristics should be protected. Some objections were raised to the principle of focusing development on transport nodes. Specific suggestions were made in the workshops for ecological/open space provision, environmental issues, transport and design. The Focus Group called for sustainable design and transport solutions, drawing on best practice. Representations supported the concept of a sustainable community as a key overriding principle in developing the area, although it was felt that there should be greater detail in terms of specific standards and facilities.
Employment
23. High levels of agreement were expressed at workshops on the issues of employment need, and quality and integration of employment uses. A need was identified in all events to relate the type of business to the particular circumstances of York. Thus linkages between office uses and business tourism were made, high tech business employing university graduates were promoted and small-scale manufacturing responding to particular ‘niche’ markets in the City were all mentioned.
24. The particular characteristics of the two sites were also emphasised at the workshop events with strong support for offices/research and development uses at York Central and general/small scale engineering at British Sugar. There was significant support for a Central Business District at York Central. Warehousing received only moderate support at British Sugar and was not supported at York Central. The need for training and recognition that jobs were needed for those with low skill levels was also made at both the Focus Group and Inclusive York Forum.
Housing
25. In general, a broad range of housing types, sizes and densities were supported. Comments on the options suggested that the densities were too prescriptive. High quality open space and linkages to public transport and services were considered important, particularly for higher densities. Whilst opinion on the density options varied between events/representations received, Option 2 (concentrating high densities around transport nodes) was given a higher level of support. Concerns about flood risk areas were highlighted at workshops, the Focus Group and in the representations received.
26. Broadly speaking feedback supported a housing mix in accordance with the Housing Market Assessment as well as higher proportions of housing. Provision of a higher proportion of flats was only supported by a small minority, although York Central was felt to be better suited to accommodate flats than British Sugar. Specialist housing types, for example, sheltered housing, ‘extra-care’ and ‘social’ housing were promoted by the Inclusive York Forum. At the workshop events the concept of exemplar housing for both energy and water conservation was put forward with highly sustainable housing design.
Social Infrastructure
27. The provision of a range of community, health and education facilities were supported as promoting vitality within the site, as well as providing employment. At the workshop events it was felt that provision should reflect the different areas of the sites with the scale of facilities relating to location. York Central was seen to relate to the City Centre, with British Sugar to the local community.
28. Provision of comparison retailing around the station received a mixed response (with some high levels of objection) with concern over town centre viability and necessity for provision but support for the sustainable location and connections with the NRM/train station. Support for the different options of scale/location of centres was also mixed with some supporting a district centre at British Sugar,
others small scale facilities throughout the sites. There was, however, significant support for local centres in the public leaflet response. Provision of the following facilities were supported, a health centre, indoor sports centre, community hall, youth centre, library, swimming pool, social club and live music venue. Specialist facilities were supported by the Inclusive York Forum. All Saints Secondary School, Gillygate Surgery and Lidgett Grove Scout Group, registered interest in relocating within the area.
**Culture and Tourism**
29. The inclusion of additional high quality cultural facilities in the vicinity of the station/NRM was widely supported. Improved attractions, infrastructure and accommodation were also held to be important in the representations received. The provision of high quality open space linked to these facilities was seen to be a key point with opportunities for a new ‘place’ for tourists and others to be provided. High quality hotels/conferencing facilities were also considered to be important to the role of York in the business/tourism sectors. The location of such facilities in the vicinity of the station were widely supported in the representations and workshop events, but received a more mixed response in the public comments. Generally however people disagreed with the location of a hotel at the British Sugar site.
30. There was general agreement for the provision of a new pedestrian and cycle bridge link to the city. Access/integration with the city centre was identified as being of critical importance in the workshop events. Opportunities were also highlighted for riverside improvements and a ‘riverside’ corridor.
**Transport and Accessibility**
31. There was wide support for increasing the priority of public transport, cyclists and pedestrians over car use. Representations stressed the need for a full detailed transport study, which would consider York Northwest within the surrounding wider highway network. Comments made by the Focus Group highlighted effective traffic management as critical to reducing congestion.
32. The provision of park and ride facilities was widely supported, with linkages to a rail halt/local interchange. A linking bridge over the River Ouse was also considered fundamental to the integration of the area with the city centre. Provision of a local interchange on British Sugar was supported, particularly if linked to any tram train halt and local centre. Provision of new, linked pedestrian cycle routes within the green infrastructure networks was also supported to promote more walking and cycling.
33. The provision of tram train was strongly supported, although it was recognised that this would be a medium to long-term project. It was noted that whilst provision should be made in any plans for a safeguarded route through the area, the feasibility of tram train was still being investigated and it may not come forward.
34. The quantitative analysis of responses to the access options (workshop events) generally indicates more agreement with pedestrian and cycle access options than the vehicular/public transport options given, which had objections to all but
the Water End and the tram train options. Difficulties in many of the options have been highlighted in the comments given in the feedback received.
**Open Space and Built Sporting Facilities**
35. Generally, public realm/green infrastructure was seen to be very important to the development of the area and should be maximised. The retention of existing facilities and open spaces on, or near, the sites (e.g. the Ings flood plain) were also considered key aspects of any new development. The integration of spaces, accessibility, availability and affordability for all age groups and for people with disabilities were also mentioned in many of the representations/events. The importance of the Railway Institute facilities and the need to protect/relocate these within the area was also highlighted. Other suggested facilities included a swimming pool, community centres and a concert hall. The public feedback on the types of open space facilities showed a higher preference for natural/semi natural green space, green corridors, parks and play areas, although generally there was a high level of support for all facilities. The Yorkshire Wildlife Trust offered management expertise for ecological and greenspaces in exchange for workspace within a development.
36. The siting of a stadium within the area received a mixed response. Whilst the opportunity to locate a city-wide facility in a sustainable location next to the station was recognised, difficulties in terms of deliverability, maintenance, traffic congestion, viability and poor use of brownfield land were also noted. Possible tensions with the facilities provided by the Railway Institute and the Central Business District were also raised. In terms of the location of a stadium within York Northwest, feedback from the workshops supported provision adjacent to the rail station, whilst the public comments were more supportive of a facility linked to a new district centre on British Sugar.
**Urban Design**
37. High quality, innovation and excellence in design were generally considered to be essential in the feedback received. The need to respond to the existing character of areas with bespoke design was supported. The Focus Group identified the opportunity for development to be ecologically pioneering, contemporary and daring, which would contrast with the historic city context. Building height was highlighted as an important issue, together with the potential creation for new views across the city. Building future requirements into the design of buildings and spaces was also mentioned with reference to climate change and ‘exemplar’ low carbon living and working to be taken into account.
**Delivery**
38. There was relatively little comment/feedback on this section, although this is probably because no options were put forward. The representations received did highlight issues of phasing and the delivery of social and environmental infrastructure.
Next Steps – Preferred Options
39. Work has started to prepare the second next stage of the Area Action Plan, the Preferred Options. This work will be carried out in-house with consultant support as required. It is anticipated that the preparation of the Preferred Options will take place in three phases: information analysis; development of conceptual options; and preparation of Preferred Options report. It is anticipated that a report will be brought to Members outlining progress on the emerging options this summer, for their consideration and comment.
Information Analysis
40. The first phase will include work to develop a mechanism for identifying and apportioning transport infrastructure that is required to deliver the development. Work will also be undertaken to update baseline information and prepare spatial plans which will outline constraints and opportunities. The vision for the area and objectives of the AAP will be reviewed and amended as appropriate. The analysis and testing of the Options contained within the Issues and Options report, will include development and application of criteria to evaluate the options. These criteria will include sustainability appraisal and consultation feedback. This will enable conclusions to be reached on whether to reject options or continue work on them to a conceptual stage. This work will be incorporated into a background report to the Preferred Options Report.
Develop Conceptual Options
41. It is intended to produce conceptual options on the broad directions for size/scale and location of uses within the area. These will take into account the work carried out in paragraph 40 above and will include work to establish a boundary for the area. An analysis of costs of developing the area eg. contamination, transport and service infrastructure will be undertaken. This will inform high level viability testing of emerging conceptual options. This will necessitate close liaison with York Central and ABF consultant teams. Sustainability appraisals of the options will also inform this stage of work on an iterative basis.
Preferred Options Report
42. The conceptual options will be analysed and refined to inform the Preferred Options. It is anticipated that the Preferred Options will comprise a limited number of comprehensive spatial strategy options. These will include broad land use locations with policy directions. This work will be accompanied by a sustainability appraisal to assess the implications of the proposals put forward. This work will form the basis for a further period of public consultation to allow public feedback on the developing options.
43. It is apparent from the guidance emerging on Area Action Plans that to be ‘sound’ it must be clear what the plan will deliver, the feasibility of this, by whom it will be delivered and the timescales for this. The Preferred Options will therefore be subject to high-level viability testing to ensure the deliverability of options being developed.
Options
44. There are no options relating to this report.
Corporate Priorities
45. The York Northwest area provides large brownfield development opportunities adjacent to the city centre. Development of this area will help to protect and enhance York’s existing built and green environment and provides an opportunity for a flagship sustainable development. The regeneration of this area will support the following corporate priorities,
- Increase the use of public and other environmentally friendly modes of transport
- Improve the quality and availability of decent affordable homes in the City
- Improve the contribution that Science City York makes to economic prosperity
Implications
46. Implications are as listed below:
- Financial None.
- Human Resources (HR) None
- Equalities None
- Legal None
- Crime and Disorder None
- Information Technology (IT) None
- Property None
- Other None
Risk Management
47. In compliance with the Councils risk management strategy, there are no risks associated with the recommendations of this report.
Recommendations
48. Members are asked to:
1) Note the comments received on the Issues and Options documents and support their consideration in informing the Preferred Options stage for the Area Action Plan.
Reason: To ensure that the Area Action Plan can be progressed to the next stage of work, in accordance with the Local Development Scheme.
2) Note the progress made to date on production of the Area Action Plan and agree the approach for the development of the Preferred Options stage of work.
Reason: To ensure that Members views are included in the generation of Preferred Options put forward.
Contact Details
Author:
Sue Houghton | Tel: (01904 551375)
York Northwest Project Manager
Ann Ward Tel: (01904 552409)
York Northwest Project Officer
Ben Murphy Tel: (01904 551415)
York Northwest Project Officer
City Strategy
Chief Officer Responsible for the report:
Bill Woolley
Director of City Strategy
Report Approved √ Date 30/04/08
Specialist Implications Officer(s): None
Wards Affected: Holgate, Micklegate, Acomb and Rural West York
For further information please contact the author of the report
Background Papers:
Report to Executive 25 September 2007
Report to LDF Working Group 24 July 2007
Report to Executive, 27 February 2007
Appendices
Appendix 1: Consultation Plan
Appendix 2: Consultation Response Analysis
Appendix 3: Overarching Summary
Appendix 4: Summary of Vision and Objectives
Appendix 5: Summary of Comments on Boundary
Appendix 6: Summary of Report of Public Comments Forms (available on website)
Appendix 7: Written Representations: Index of Consultees (available on website)
Appendix 8: Summary of Written Representations (available on website)
Appendix 9: Summary of Report of Workshop Events (available on website)
Appendix 10: Focus Group Report (available on website)
Appendix 11: Inclusive York Forum Summary (available on website)
Appendix 12: Copies of Public Comment Form/Leaflet/Poster (available on website) |
Matters raised with permission—
Reported Statement of Home Minister to U.S. Ambassador in 2009 about slow economic development of India (pages 1-2)
Scarcity of anti-Rabies vaccine in Hospitals (page 2)
Revelations made by Shungulu Committee about instances of corruption in Commonwealth Games (pages 2-5 and pages 6-9)
Damage being done to crops by the blue bucks in the country (pages and pages 9-10)
Message from Lok Sabha—
The Institutes of Technology (Amendment) Bill, 2011 – Laid on the Table (Page 5)
Welcome to Parliamentary Delegation from Korea (Page 9)
Special Mentions
Need to put ban on the sale of Chinese items to protect the Small Scale Industries (pages 10-11)
Concern over large number of transfers of faculty members of the Pandit Sunderlal Sharma Central Institute of Vocational Education (PSSCIVE) (page 11)
Demand to provide financial assistance to Chhattisgarh for developing infrastructure to fight naxalism (pages 11-12)
Demand to initiate urgent steps to extradite Warren Anderson, former Chairman of the Union Carbide Corporation (pages 12-13)
ii
Demand to maintain regular supply of coal to thermal power plants of the country, particularly in Maharashtra (page 13)
Demand to improve the working of LIC of India and give adequate facilities to its agents (pages 13-14)
Need for effective mechanism for solid waste management (pages 15-16)
Papers Laid on the Table (pages 15-40)
Reports of the Department Related Parliamentary Standing Committee on Agriculture – Laid on the Table (page 40)
Statements of the Department Related Parliamentary Standing Committee on Social Justice and Empowerment – Laid on the Table (pages 40-41)
Statement by Minister
Status of implementation of recommendations contained in the Thirtieth Report of the Department-related Parliamentary Standing Committee on Finance – Laid on the Table (page 41-42)
Valedictory Remarks (page 42-43)
National Song (page 43)
RAJYA SABHA
Friday, 25th March, 2011/6th Chaitra, 1932 (Saka)
The House met at eleven of the clock,
MR. DEPUTY CHAIRMAN in the Chair.
MATTERS RAISED WITH PERMISSION
MR. DEPUTY CHAIRMAN: Prof. Ram Gopal Yadav...(Interruptions)... Mr. Venkaiah Naidu, your name is there. I will call you.
Reported Statement of Home Minister to U.S. Ambassador in 2009 about slow economic development of India
पूरे . राम गोपाल यादव (उत्तर प्रदेश ) : श्रीमन् , उत्तर भारत देश के बहुत महत्वपूर्ण अंग हैं। उत्तर भारत ने देश की आजादी से लेकर अब तक इस देश की तरक्की में बहुत महत्वपूर्ण भूमिका अदा की है। वह चाहे पंजाब रहा हो या देश के अन्य हस्सिये , इसने पश्चिम जवाहरलाल नेहरू , राजेंद्र बाबू , जय पुरकाश जी और डॉ. लोहिया जैसे बहुत बड़े लोग देश को दाश। लेकिन , मुझे बहुत अफसोस के साथ कहना पड़ रहा है कि आजके कुछ समाचारपत्रों में यह खबर छपी है कि हाफ्टुस्तान के गृह मंत्री ने कहा कि भारत का वॉक्सास तेजी से होता , अगर इसके केवल दक्षिणी और पश्चिम में हस्सी ही होती। ... (वयवधान ) ... यह बात हाफ्टुस्तान के गृह मंत्री ने अमेरिका के राजदूत से कही। यह बहुत ही चट्टिस को बात हैं , देश के लिए बहुत ही शर्मनाक बात हैं। ... (वयवधान ) ... मैं आपके माध्यम से यह चाँद्या कि देश के गृह मंत्री अपने इस बयान को वंद्रियों करे , इसका खंडन करे , क्योंकि यह पूरे भारत की एकता और अखंडता के लाये खतरनाक है , जबकि देश का गृह मंत्री देश की एकता और अखंडता का परस्परीक होता है। उनके मुँह से इस तरह की बात शेषा नहीं देती और उत्तर भारत के लोगों में इसके प्रति जबरदस्त रोष है , असंतोष है। मैं आपसे अनुरोध करूंगा कि आप डायरेक्ट करें और देश के गृह मंत्री से कहें कि वे इसका खंडन करें।
श्री शविनन्द तविरी (बहिर) : सर, इसके बारे में गृह मंत्री जी को सफाई दें। चाहिए। ... (वयवधान ) ... यह बहुत ही खतरनाक बात है। ... (वयवधान ) ...
श्री उपसभापति : इसकी authenticity क्या है? Now, Shri Narendra
Kumar Kashyap...(Interruptions)...
श्री शविन्द्र तकिया: सर, यह जो खबर छपी है, यह सही है या गलत है, ... (व्यवधान ) ...
श्री उपसभापति: इसमें क्या है? ... (व्यवधान ) ... उन्होंने Zero Hour दिया है। उन्होंने बोला है: ... (व्यवधान ) ... Zero Hour में गवर्नमेंट का response नहीं आता। ... (व्यवधान ) ...
श्री माननीय सदस्य: सर, यह देश की अखंडता ... (व्यवधान ) ... की ही बात नहीं है ... (व्यवधान ) ...
श्री उपसभापति: यह सब अखबार की खबर हैं: ... (व्यवधान ) ...
श्री राम कृपाल यादव (बहिर) : अगर देश के गृह मंत्री को यह भावना रहेगी, तो... (व्यवधान) ... यह तो बहुत दुःखायपूर्ण हैं... (व्यवधान) ...
श्री वीर पाल सहि यादव (उत्तर प्रदेश) : यह देश की एकता और आखड़ा की बात है... (व्यवधान) ...
श्री उपसभापति : श्री नरेन्द्र कुमार कश्यप।
Scarcity of anti-Rabies vaccine in Hospitals
श्री नरेन्द्र कुमार कश्यप (उत्तर प्रदेश) : मान्यवर, में जीरो ऑवर के माध्यम से देश की एक गंभीर समस्या सदन के माध्यम से सरकार के संजान में तामा चाहता हूँ।
मान्यवर, हमारे देश में हर दिन हजारों लोग कुल्ले के काटने से गंभीर रूप से बीमार होते हैं। उससे एक ऐसी बीमारी हो जाती है, जसिका इलाज अगर समय से न किया जाए, तो कश्मी भी आदमी को मौत नारिचित हैं।
THE MINISTER OF HUMAN RESOURCE DEVELOPMENT (SHRI KAPIL SIBAL): Sir, this is a clear case of discrimination. आपने कहा कि "बालकों के काटने से" यह तो बहुत gender discrimination हैं। मैं आपत्ति जताना चाहता हूँ... (व्यवधान) ...
श्री राजपाल सहि सेनी (उत्तर प्रदेश) : एक तो कुल्ले के काटने से लोग मर रहे हैं और... (व्यवधान) ...
श्री उपसभापति : बेठिए, बेठिए।
श्री नरेन्द्र कुमार कश्यप : मान्यवर, ये हालात उस समय और जयदा मुश्किल हो जाते हैं, खास तौर से हमने गाँवों, देहातों और शहरों में देखा है, जब कुल्ले के काटने के बाद अस्पतालों में उस मरीज को दाँवी नहीं मिलती, उसके इंजेक्शन नहीं मिलते और अगर इंजेक्शन मिलते भी हैं तो उसकी कीमत इतनी जयदा होती है कि एक गरी ब आदमी उसको खरीद नहीं पाता, लहिंजा हजारों लोग हर साल बच्चे इलाज के मर जाते हैं। मैं सरकार को ध्यान इस गंभीर बीमारी की तरफ दबिना चाहता हूँ। सरकार इस पर कोई न कोई ऐसा कानून बनाये जसिसे पंचायत से लेकर जिला, तहसील और मंडल सतर पर इंजेक्शन का इंतजाम हो या...
Revelations made by Shunglu Committee about instances of corruption in Commonwealth Games
SHRI M. VENKAIAH NAIDU (Karnataka): Mr. Deputy Chairman, Sir, it is a very serious issue that I am going to raise here, a matter of urgent public importance.
The V.K. Shungulu Committee probing the Commonwealth Games scam submitted three reports to the Government. Those reports have not been made public so far. It may be recalled, Sir, that this Committee was appointed by the Prime Minister himself with an assurance that immediate action will be taken once the enquiry is completed. Now, the enquiry has been completed and three reports have been submitted. The reports indict the Delhi Government, Lieutenant Governor and the DDA. All these three agencies were named in the reports for their irregularities, involvement in corruption as also cost overrun and abnormal delay.
Sir, the first report said that there was a loss of Rs.200 crores to the exchequer. As per the second report, the builder at the Games Village declared an advance tax of Rs.780 crores, leaving another Rs.500 crores. Many unwanted works were taken up. They had missed their deadlines and suffered from tendering and procurement maladies. It is a very serious matter, Sir. The entire country was agitated about the irregularities in the Commonwealth Games.
It may be recalled that in the backdrop of that issue, the Prime Minister had volunteered himself and asked us to wait till the enquiry was over and said that action will be taken. So far, no action has been taken. More than Rs.60,000 crores of public money has been looted and cheated. ... (Interruptions) ...
MR. DEPUTY CHAIRMAN: After he has...
SHRI M. VENKAIAH NAIDU: Please try to understand. It is a very serious issue. ... (Interruptions) ...
MR. DEPUTY CHAIRMAN: I am not saying anything to you. But please tell this to the others. ... (Interruptions) ...
SHRI M. VENKAIAH NAIDU: Sir, they may be requesting the same. It is such a big fraud committed on the nation. It has really brought shame to the nation in the international community in conducting the Commonwealth Games.
MR. DEPUTY CHAIRMAN: All that cannot be discussed during Zero Hour.
SHRI M. VENKAIAH NAIDU: Sir, the purpose of conducting the
Commonwealth Games, at that time, was to create a common wealth for the country. It has become individual wealth and loot for some individuals. Now, the Committee appointed by the Prime Minister himself has indicted the Delhi Chief Minister. She has no right to continue in office. ... (Interruptions) ... She must quit forthwith. Then, action should be taken against the Lieutenant Governor of Delhi who is also involved in this. Also, action must be taken against the DDA officials who are also involved in this massive scam and scandal. This requires immediate attention of the House. We demand
that the Government should come forward, place the report of the Shungulu Committee on the Table of the House and, then, we should discuss it. The Government should also inform us of the action taken in this matter. ... (Interruptions) ... The other day, when the report came and the matter was raised in the Delhi Assembly, all the Opposition Members were taken out of the Delhi Assembly. Are we in a democracy?
MR. DEPUTY CHAIRMAN: We cannot discuss Delhi Assembly here.
SHRI M. VENKAIH NAIDU: I am not discussing it, Sir. We are the Council of States. I am just bringing to your notice what is happening right under our nose in the National Capital. Such a massive loot has taken place and there is a serious indictment by the official committee appointed by the Prime Minister. What is the response of the Government? What action does the Government want to take on this? ... (Interruptions) ...
श्रीमती माया सहि (मध्य प्रदेश): सर, इस वक़िय पर मैं भी कुछ कहना चाहती हूँ ... (व्याख्यान) ...
MR. DEPUTY CHAIRMAN: You only associate yourself.
श्रीमती माया सहि: सर, मैं सख्य को इस वक़िय के साथ सम्बद्ध करती हूँ।
श्री श्रीगोपाल व्यास (छत्तीसगढ़): महोदय, मैं सख्य को इस वक़िय के साथ सम्बद्ध करता हूँ।
श्री जय पुष्काश नारायण सहि (झारखंड): महोदय मैं सख्य को इस वक़िय के साथ सम्बद्ध करता हूँ।
श्री पुरकाश जावडेकर (महाराष्ट्र): महोदय, मैं सख्य को इस वक़िय के साथ सम्बद्ध करता हूँ।
श्री बलविंदर सहि भुंडर (पंजाब): महोदय, मैं सख्य को इस वक़िय के साथ सम्बद्ध करता हूँ।
श्री युद्धनारायण पाणि (उड़ीसा): महोदय, मैं सख्य को इस वक़िय के साथ सम्बद्ध करता हूँ।
श्री पुरुषोत्तम खोड़ाभाई रूपाला (गुजरात): महोदय, मैं सख्य को इस वक़िय के साथ सम्बद्ध करता हूँ।
कुछ माननीय सदस्य: महोदय, हम सख्य को इस वक़िय के साथ
SHRI S. S. AHLUWALIA (Jharkhand): What is the response of the Government? ... (Interruptions) ... We want a response from the Government, Sir. ... (Interruptions) ...
श्री पुस्तोतम खोड़ामाई रघला: दौरंक यहाँ कांग्रेस की मुख्य मंत्री है, इसलिए कुछ नहीं काया गया। ... (स्वयंचन) ...
MR. DEPUTY CHAIRMAN: By raising an issue during Zero Hour, you cannot seek the response of the Government. ... (Interruptions) ... You should have raised this issue in some other form. ... (Interruptions) ... But the point is you have raised it in Zero Hour. ... (Interruptions) ... You have raised it in Zero Hour. ... (Interruptions) ...
SHRI M. VENKAIH NAIDU: We had requested for a JPC. It was not agreed upon. ... (Interruptions) ... Then, the Supreme Court ... (Interruptions) ...
MR. DEPUTY CHAIRMAN: Venkaiah Naiduji, if you wanted to raise this issue, you should have raised it in some other form. You have raised it in Zero Hour. What is it that I can do? ... (Interruptions) ...
MR. DEPUTY CHAIRMAN: The House is adjourned for fifteen minutes.
The House then adjourned at eleven minutes past eleven of the clock.
The House re-assembled at twenty-six minutes past eleven of the clock,
MR. DEPUTY CHAIRMAN in the Chair.
SHRI RAJIV PRATAP RUDY (Bihar): Sir, Shri Venkaiah Naidu raised the issue... (Interruptions) ...
श्री उपसभापति : वह हो गया हैं ... (व्यवधान )...
SHRI M. VENKAIH NAIDU: Sir, I hope the Government will take this seriously and do something.
MR. DEPUTY CHAIRMAN: The Government will definitely... (Interruptions) ...
SHRI M. VENKAIH NAIDU: ...in pursuance of the Report. Nobody is there, Sir.
MR. DEPUTY CHAIRMAN: The Government is there.
SHRI M. VENKAIH NAIDU: Sir, it is not an ordinary issue like a Zero Hour Mention. It is a very important issue, burning issue for the entire country. ... (Interruptions) ...
MR. DEPUTY CHAIRMAN: Now, Message from the Lok Sabha.
________________
MESSAGE FROM THE LOK SABHA
The Institute of Technology (Amendment) Bill, 2011
SECRETARY-GENERAL: Sir, I have to report to the House the following message received from the Lok Sabha, signed by the Secretary-General of the Lok Sabha:
"In accordance with the provisions of rule 96 of the Rules of
Procedure and Conduct of Business in Lok Sabha, I am directed to enclose the Institutes of Technology (Amendment) Bill, 2011, as passed by Lok Sabha at its sitting held on the 24th March, 2011.”
Sir, I lay a copy of the Bill on the Table.
Revelations made by Shungulu Committee about instances of corruption in Commonwealth Games – (Contd.)
THE MINISTER OF STATE IN THE MINISTRY OF PARLIAMENTARY AFFAIRS; THE MINISTER OF STATE IN THE MINISTRY OF PERSONNEL, PUBLIC GRIEVANCES AND PENSIONS AND THE MINISTER OF STATE IN THE PRIME MINISTER’S OFFICE (SHRI PAWAN KUMAR BANSAL): Sir, the Shunglu Committee was set up by an Executive Order. Sir, the Report has to be studied by the Government. And, Sir, I can only assure that whatever action is ever required, the Government would never be found wanting in that. So, we will not ever deter from taking any action whatever is necessary. But, Sir, it is absolutely essential...(Interruptions)...
SHRI PRAKASH JAVADEKAR (Maharashtra): That is your promise. ...(Interruptions)...
SHRI M. VENKAIAH NAIDU: Will you act on the recommendations? That is all. ...(Interruptions)...
SHRI PAWAN KUMAR BANSAL: Sir, it is imperative that the Report...(Interruptions)... No, no, no. ...(Interruptions)...
SHRI M. VENKAIAH NAIDU: Do not act like a lawyer. Will you act on the Shunglu Committee Report? It is simple. Just one line. ...(Interruptions)...
SHRI PAWAN KUMAR BANSAL: Sir, I said that the Government will not be found wanting in anything that the Government has to do. The Government understands its responsibility. The Report will have to be studied. Wherever any action is required and the Government feels that action is required, the action would be taken. Sir, I would only submit...(Interruptions)...
SHRI M. VENKAIAH NAIDU: That means so many adjectives. ...(Interruptions)... Why did you appoint that Committee? ...(Interruptions)...
SHRI PAWAN KUMAR BANSAL: Sir, I would only submit that the leaders of the Opposition should also give us some time to study the Report.
SHRI M. VENKAIAH NAIDU: That I can understand, Sir. But saying...
...(Interruptions)...
SHRI PAWAN KUMAR BANSAL: But he should not pass judgment on that.
...(Interruptions)...
SHRI M. VENKAIAH NAIDU: But saying, if the Government feels it necessary, he was trying to cover up...(Interruptions)...
SHRI S.S. AHLUWALIA: Sir, when we demanded the JPC on the CWG, at that time they said that the Shunlu Committee Report is there.
...(Interruptions)...
SHRI M. VENKAIAH NAIDU: What action are you going to take on
...(Interruptions)...
MR. DEPUTY CHAIRMAN: Now, the Minister has said it.
...(Interruptions)... You wanted a response from the Government, the Government has responded.
...(Interruptions)...
SHRI M. VENKAIAH NAIDU: Sir, no Report has come.
...(Interruptions)... No Report has come. ...(Interruptions)... No inquiry has taken place. ...(Interruptions)...
MR. DEPUTY CHAIRMAN: Venkaiah Naiduji,...
...(Interruptions)...
SHRI S.S. AHLUWALIA: Sir, it was not a normal CAG Report.
...(Interruptions)... That it will go to the President and then...
...(Interruptions)...
MR. DEPUTY CHAIRMAN: You mentioned the issue. ...(Interruptions)... You wanted a response from the Government, the Government has responded.
...(Interruptions)... You referred to a newspaper, and they have responded. As and when ...
...(Interruptions)...
SHRI M. VENKAIAH NAIDU: Sir, we understand your problem and also the problem of the House because today is the last day. But, at the same time, it is a Committee appointed by the Prime Minister with an assurance that prompt action will be taken. ...(Interruptions)...
SHRI S.S. AHLUWALIA: It is the Prime Minister’s...
...(Interruptions)...
SHRI M. VENKAIAH NAIDU: Sir, the hon. Minister is just adding adjectives to it. ...(Interruptions)...
MR. DEPUTY CHAIRMAN: Now, Shri Avtar Singh Karimpuri. ठीक हैं, ठीक हैं, कृपया आप बेचिए। ...(वयवधान )...
SHRI M. VENKAIH NAIDU: Mr. Shukla ... (Interruptions) ...
श्री उपसभापति : कृपया आप बेठिए। ... (व्यवधान )... Shri Shukla, please sit down. ... (Interruptions) ... Now, Shri Avtar Singh Karimpuri. ... (Interruptions) ... कृपया आप बेठिए। ... (व्यवधान )...
Damage being done to crops by the blue bucks in the country
श्री अवतार सहि करीमपुरी (उत्तर प्रदेश ) : उपसभापति महोदय, मैं आपकी इजाजत से एक महत्वपूर्ण मुद्दा सदन में उठा रहा हूँ। पूरे देश में कस्बियों को नुकसान करने में नील गाय का बहुत बड़ा योगदान है।
कशिन आजदेश में आत्महत्याएं कर रहे हैं, लाखों की संख्या में कशिनों ने आत्महत्याएं की हैं। जब देश का कशिन आत्महत्या करता है, उस देश के मजदूर की हालत भी बद से बदतर हो जाती है। मजदूर और कशिन को बचाने के लिए, जो कशिनों की फसलें नील गाय द्वारा बर्बाद की जा रही हैं, उनको रोकना बहुत जरूरी है। मैं आपके माध्यम से सरकार से यह कहना चाहता हूँ, अपील करना चाहता हूँ कि नील गाय से कशिनों की फसलों की बर्बादी को बचाने के लिए सरकार कोई पहल करे। यह सरकार हमेशा यह कहती है कि हम कशिनों के बड़े हतियार हैं। एक तरफ जहां बहुत बड़े पैमाने पर उन्हें नकली बीज और नकली दवाइयां दी जारही हैं, वहीं दूसरी ओर कशिनों के उनकी फसल की सही कीमत नहीं दी जा रही है। महोदय, कशिन पहले से ही बहुत बड़े आर्थिक संकट से ज़हर रहा है, उसके ऊपर यह जो नील गाय है, इसको रोकने के लिए आजादी के 63 साल के बाद भी कोई इंतज़ाम नहीं कार्रिग गए हैं। मैं यह कहना चाहता हूँ कि सरकार इस पर नोटिस ले और इसके संबंध में पूरे देश में कोई नीति तैयार करे, ताकि कशिनों की फसल को बचाया जा सके। इसके अलावा जो रोड ऐक्सीडेंट्स होते हैं, उनमें भी लोगों की बहुत सी जान गयी हैं। मेरा आपके माध्यम से अनुरोध है कि इस विषय को गंभीरतापूर्वक लेले हुए इसके संबंध में कोई असरदार नीति तैयार करके कशिनों और आमजनमानस को बचाने के लिए पूर्णायम तैयार करिया जाए। धन्यवाद।
श्री जय पुरकाश नारायण सहि (झारखंड): महोदय, मैं माननीय सदस्य के वक्तव्य से सुविध को सबृह करता हूँ।
श्री ओम पुरकाश माधुर (राजस्थान): महोदय, मैं माननीय सदस्य के वक्तव्य से सुविध को सबृह करता हूँ।
श्री सुनेद्र मोतीलाल पटेल (गुजरात): महोदय, मैं माननीय सदस्य के वक्तव्य से सुविध को सबृह करता हूँ।
श्री पुरुष शोक्तम खोड़ाभाई रूपाला (गुजरात): महोदय, मैं माननीय सदस्य के वक्तव्य से सुविध को सबृह करता हूँ।
SHRI K.N. BALAGOPAL (Kerala): Sir, I want to raise an important matter. ... (Interruptions) ... One MP of this house ... (Interruptions) ...
SHRI SITARAM YECHURY (West Bengal): Sir, please allow him. ... (Interruptions) ...
MR. DEPUTY CHAIRMAN: No, no, there is no permission.
...(Interruptions)... I cannot allow like that. ...(Interruptions)...
Nothing will go on record. ...(Interruptions)... We have not
permitted. ...(Interruptions)... Special Mentions.
...(Interruptions)... There is no notice. ...(Interruptions)... It is
just an allegation. ...(Interruptions)...
SHRI SITARAM YECHURY: It is an important matter, Sir.
...(Interruptions)... Please allow him, Sir. ...(Interruptions)...
MR. DEPUTY CHAIRMAN: You should give your notice.
...(Interruptions)... If he is a Member, there should be notice.
...(Interruptions)... I cannot take up like this.
...(Interruptions)... There should be notice. ...(Interruptions)...
Even if he is a Member of this House, we have to give notice.
...(Interruptions)... If you are raising anything against a Member,
you have to give notice to him. ...(Interruptions)... How can I allow
it? ...(Interruptions)...
SHRI SITARAM YECHURY: You please understand the situation. ... (Interruptions) ...
MR. DEPUTY CHAIRMAN: I understand the situation but I cannot allow violation of the rules. ... (Interruptions) ... It is our fundamental duty to give notice to the concerned Member. ... (Interruptions) ... We have to give notice to the Member. ... (Interruptions) ... Now, the Member is not here. ... (Interruptions) ... I cannot allow it in his absence. ... (Interruptions) ...
SHRI SITARAM YECHURY: Sir, the problem is that today is the last day of this Session. ... (Interruptions) ... We are seeking permission of the Chair. ... (Interruptions) ...
MR. DEPUTY CHAIRMAN: You can give notice during the next Session. ... (Interruptions) ... What I can do, please tell me. ... (Interruptions) ... One minute. ... (Interruptions) ... We have a foreign delegation.
WELCOME TO PARLIAMENTARY DELEGATION FROM KOREA
MR. DEPUTY CHAIRMAN: Hon. Members, I have an announcement to make.
We have with us, seated in the Special Box, Members of a parliamentary delegation from the Republic of Korea, currently on a visit to our country under the distinguished leadership of His Excellency Mr. Park Hee-Tae, Speaker of the National Assembly of the Republic of Korea.
On behalf of the Members of the House and on my own behalf, I take pleasure in extending a hearty welcome to the leader and other members of the delegation and wish our distinguished guests an enjoyable and fruitful stay in the country. We hope that during their stay here they would be able to see and learn more about our parliamentary system, our country and our people, and that their visit to this country will further strengthen the friendly bonds that exist between India and the Republic of Korea.
Through them we convey our greetings and best wishes to the Parliament and the friendly people of the Republic of Korea.
Damage being done to crops by the blue bucks in the country — (Contd.)
SHRI SITARAM YECHURY: Sir, please allow me to raise an issue. I am perfectly aware of the rules but the point is that, it is a news item that has appeared after the time for the notice was over. Today is the last day. I am beseeching the wisdom of the Chair on how this matter can
be raised. It is very serious. A Member of our House was caught with Rs. 57 lakhs on the eve of elections. That is the point.
...(Interruptions)...
MR. DEPUTY CHAIRMAN: Without giving a notice, you cannot name a Member. ...(Interruptions)...
SHRIMATI BRINDA KARAT (West Bengal): We are not naming the Member.
SHRI SITARAM YECHURY: I am not naming anyone.
MR. DEPUTY CHAIRMAN: When I allow you, then, you have to name him.
SHRI SITARAM YECHURY: I am asking your wisdom...
MR. DEPUTY CHAIRMAN: I cannot because it is a violation of his Fundamental Right, the Member’s Right. ...(Interruptions)...
SHRIMATI BRINDA KARAT: Sir, the TMC MP, a defector MP to the TMC, to the Trinamool Congress...(Interruptions)... He is caught with Rs. 57 lakhs. How is it possible?
MR. DEPUTY CHAIRMAN: No, no, if he is a Member, then please try to understand.
SHRIMATI BRINDA KARAT: We are not taking the Member’s name. We are taking the Party’s name. Who is sending that money to buy votes in Assam?
MR. DEPUTY CHAIRMAN: We are not discussing a political party here. Now, we will take up Special Mentions. Shri Jai Prakash Narayan Singh. ...(Interruptions)... Nothing will go on record. ...(Interruptions)... Nothing will go on record. I am helpless. Please...(Interruptions)... You have other platform, not here. ...(Interruptions)...
SPECIAL MENTIONS
Need to put ban on the sale of Chinese items to protect the Small Scale Industries
श्री जय प्रकाश नारायण सहि (झारखंड ): महोदय , होशी का:
त्योहार अभी हाल ही में मनाया गया है। इस त्योहार के अवसर पर बहुत बड़ी मात्रा में चीन में बनी पचिकारिया, रंग और अन्य पदार्थ बाजार में बक़ि रहे थे। महोदय, दक्षिण के अवसर पर पटाके और रक्षा बंधन के अवसर पर रखियां भी चीन से आ रही हैं। हमारी मंडियों में चीन की सब्जियां और फल भी उपलब्ध हैं। इसके अतिरिक्त चीन में बने खलिने, घड़ियां, ताले और न जाने कतिपय और भी पदार्थ हमारे बाजार में बक़ि रहे हैं, जस्ते कारण हमारा लघु उदयोग लगभग बन्द होता जाता है। मैं माननीय वाणिज्य मंडी जी से नक्कल करना चाहूँगा कि लघु उदयोग को बचाने।
के लिए भारत सरकार अति आवश्यक जरूरी काम करे, ताकि हमारी डेश की आवश्यक वस्तुएं हमारे देश में नामित लघु उद्योग के माध्यम से ही बाजार में बकी। मैं सरकार से यही आग्रह करना चाहता हूँ। धन्यवाद।
श्री कलराज मन्थ (उत्तर प्रदेश): महोदय, मैं इसका समर्थन करता हूँ।
श्री रघुनंदन शर्मा (मध्य प्रदेश): महोदय, मैं भी इसका समर्थन करता हूँ।
Concern over large number of transfers of faculty members of the Pandit Sunderlal Sharma Central Institute of Vocational Education (PSSCIVE)
SHRI SHADI LAL BATRA (Haryana): Sir, I draw your attention to the large number of the transfers of the faculty members recruited for working in the area of vocational education at the apex body i.e. Pandit Sunderlal Sharma Central Institute of Vocational Educational (PSSCIVE), Bhopal. Sir, the Government had permitted NCERT to set up PSSCIVE as apex body for growth and development of vocational education in India in pursuance of the implementation of the National Policy on Education, 1986 and Programme of Action 1992. Later, advertisements appeared in national newspapers for recruitment of faculty in vocational education to be posted at PSSCIVE, Bhopal. The job description is to work in the area of vocational education. Consequently, the faculty was recruited and posted at PSSCIVE, Bhopal. They were working in PSSCIVE. A few transfers of faculty members in commerce, home science and agriculture of vocational stream have been made in last 8-10 years from PSSCIVE to other constituent units of NCERT. This has greatly hampered the development of vocational education and manpower development programmes in the country. I request the Government to call for a detailed performance report of those transferred faculty members from PSSCIVE, Bhopal and make an assessment of what work they had performed in the area of vocational education as part of their current and past assigned duties besides taking steps to post them back at PSSCIVE, Bhopal for which they had been recruited so that the success of the vocationalisation of education programme in the country and delivery system of instructions and support service could be achieved in a big way.
Demand to provide financial assistance to Chhattisgarh for developing infrastructure to fight naxalism
श्री श्रीरामपाल वृयास (छत्तीसगढ़) : उपसभापति महोदय, सभी लोग जानते हैं कि नया राज्य छत्तीसगढ़ सर्वधिक नक्सल प्रभावित हैं। वहां तेजी से हो रहे वाक्स के पर्यायों में यह सबसे बड़ी बाधा है। राज्य सरकार लंबे समय से पुलिस बल आधुनिकीकरण योजना के अंतर्गत 'ए' श्रेणी की मांग कर रही है। इस योजना के केवल एक वर्ष बढ़ाने से छत्तीसगढ़ का काम नहीं चलेगा क्योंकि इन क्षेत्रों में अत्यधिक सुखा उपकरण, बैरक, थाना, चौकी नक्सिण आदि कराया जाना आवश्यक है। वास्तव में छत्तीसगढ़ को 'ए' श्रेणी में रखते हुए 100 परिस्थिति केंद्रों देने की आवश्यकता है। राज्य ने शहरों के बड़े परिस्थितियों, धार्मिक, सांस्कृतिक संस्थानों इत्यादि की सुरक्षा के लिए सटेट इंडस्ट्रियल सक्योरिटी फोर्स के गठन का निर्णय किया है। इस हेतु भी इंडियन राज्य बटालियन के अनुरूप केंद्र द्वारा सहयोग प्रदान करना चाहिए।
Demand to initiate urgent steps to extradite Warren Anderson, former Chairman of the Union Carbide Corporation
SHRI K.N. BALAGOPAL (Kerala): Sir, the Delhi Chief Metropolitan Court yesterday accepted the plea of the CBI for extradition of Warren Anderson, the key accused in Bhopal Gas Tragedy. The court said, “There is no bar in law seeking the extradition of Anderson and there is no ground or reason for the US authorities for not cooperating with the CBI.”
The court also asked the CBI to approach the Ministry of External Affairs which will send the extradition request to the US authorities. Technically, a court’s nod is required for such a request to be moved, and that is fulfilled now.
The CBI was not pursuing Anderson’s extradition since 2004, when their last such request was rejected by the US authorities without proper reason. The Union Carbide Gas Tragedy was the worst pesticides/chemical factory disaster in the world. It is a well settled fact that the people were massacred due to the guilty
negligence on the part of the management of Union Carbide. More than 15,000 people died when toxic gas leaked from Union Carbide plant on the nights of 2nd and 3rd December, 1984. The number of seriously injured were more than 50,000. But the country has witnessed a judicial process which did not brought in the real culprits with appropriate punishment. The kingpin of the massacre was not brought before the process of law.
Finally, now, the judiciary has again directed the CBI for extradition of Anderson. This is to be done immediately with the bold intervention of the External Affairs Ministry and the hon. Prime Minister of India.
Hence, I urge upon the Government to take urgent steps on this issue for protecting the belief of the common people in rule of law.
DR. BARUN MUKHERJI (West Bengal): Sir, I associate myself with the Special Mention made by my friend.
SHRIMATI BRINDA KARAT (West Bengal): Sir, I also associate myself with the Special Mention made by my colleague.
**Demand to maintain regular supply of coal to thermal power plants of the country, particularly in Maharashtra**
DR. BHARATKUMAR RAUT (Maharashtra): Sir, I wish to bring to the notice of the House the grave situation of coal supply to various power generation units all over the country that is likely to create acute shortage of power generation during the ensuing summer, thus resulting in further staggering of electricity supply. Considering the demand for electricity increase during summer, the shortage of power generation will further worsen the situation.
The Union Power Minister, Shri Sushilkumar Shinde, in his letter to the PMO has said that the Coal India Limited has failed to meet the supply target. This will result in reduction of some 17,000 MW generation capacity that is in the pipeline from being switched on.
In India, as much as 70 per cent of electricity is generated by coal-based thermal stations. These plants proved to be the backbone of the power supply, particularly during the summer when hydro production is much less. Sir, Maharashtra being the largest power generation State, it is the worst sufferer due to the coal shortage. Out of total 16,000 MW installed capacity, Maharashtra produces 7,295 MW power through its thermal plants, while only 2,344 MW is generated through hydro-means. In this case, it is necessary to ensure that coal supply to Maharashtra remains intact. Out of seven thermal power plants in Maharashtra, four plants are undergoing renovation for extra generation capacity. If the coal supply declines, all efforts will go waste. Considering that Maharashtra is already facing shortage of over 3,000 MW, the upgradation of thermal power plants is an immediate necessity.
I, therefore, request the hon. Power Minister to look into the issue on urgent basis.
and reinstate coal supply to thermal power plants, particularly those in Maharashtra.
Thank you.
Demand to improve the working of LIC of India and give adequate facilities to its agents
श्री उद्धारायण पाणि (उड़ीसा) : उपसभापति महोदय, भारत सरकार द्वारा इन दिनों लागि जा रहे कई पदक्षेपों से ऐसा लगता है जैसे राष्ट्रीयकृत बीमा उद्योग, भारतीय जीवन बीमा नबिम नकारात्मक रूप से प्रभावित होता जा रहा है। आजभी देश का आम बीमा-धारक भारतीय जीवन बीमा नबिम अथवा लाइफ
इंश्योरेंस कॉर्पोरेशन ऑफ़ इंडिया पर पूर्ण वशि वास्तव रखता है। LIC के पास 21 करोड़ वृद्धिक्षमता बीमा-धारकों को 34 करोड़ परिलिखित हैं। पूरे वशि व में केवल चार देशों अध्यक्षता इंडोनेशिया, अमेरिका, भारत एवं चीन की जनसंख्या है। इससे अधिक हैं। दावा भुगतानों में LIC वशि व में पृथ्वी स्थान पर है। वर्ष 2009-10 में इस नवी में कुल 2.11 करोड़ दावों के तहत 53,954 करोड़ राशि का भुगतान किया है। बकाया दावों का पूर्तित भाग 1.14 पूर्तित है। LIC ने आम आदमी बीमा योजना के तहत अभी तक 1.63 करोड़ भूमिहीन परिवारों का बीमा किया है। इस संबंधन ने वर्ष 2009-10 में आयकर के रूप में 3,257.38 करोड़ एवं सेवाकर के रूप में 1,371.84 करोड़ उपर का भुगतान किया है। इस सार्वजनिक कार्यक्रम ने मार्च, 2010 तक 1,48,197 करोड़ उपर सामाजिक एवं दायागत कार्यक्रम अध्यक्षता ऊर्जा, हाड़सिंग, सचिवाई, रोड, पोर्ट और पुल आदि में नवीकरण किया है। LIC देश में सबसे बड़ा संस्थागत नवीकरण है, जस्ते वर्ष 2009-10 में 61,000 करोड़ उपर से अधिक संयोजक मार्केट में नवीकरण किया है। इसमें संयोजक कुल 13.59 लाख अभिकर्ता कार्यरत हैं। जो संख्या के दृष्टि से वशि व में सर्वाधिक हैं।
महोदय, आज की आवश्यकता है कि इन्हीं अभिकर्ताओं की सुविधा के पश्चिम LIC ध्यान दें एवं भारत सरकार LIC के पश्चिम ध्यान दें जस्ते देश के आम बीमा-धारकों को अधिक सुविधा प्राप्त हो। सरकारी नीति वादिशी बीमा कंपनियों के अनुकूल नहीं होनी चाहिए। सरकार के अब ऐसा कोई कदम नहीं उठाना चाहिए, जस्ते इस राष्ट्रीय उद्योग भारतीय जीवन बीमा नवीम को नुकसान पहुंचे और वादिशी बीमा कंपनियों को लाभ मिले, बल्कि यह सुनिश्चित किया जाना चाहिए कि LIC के तमाम अभिकर्ताओं के अच्छी से अच्छी सामाजिक सुरक्षा प्राप्त हो, ताकि ये आम बीमा-धारकों की दया से सेवा कर सकें।
श्री पुरकाश जावडेकर (महाराष्ट्र): महोदय, जो वक्त्य माननीय सदस्य ने उठाया है, मैं अपने को उससे सम्बंध करता हूँ।
श्री श्रीगोपाल वयास (छत्तीसगढ़): महोदय, जो वक्त्य माननीय सदस्य ने उठाया है, मैं अपने को उससे सम्बंध करता हूँ।
श्री कप्तान सहि सोल्की (मध्य प्रदेश): महोदय, जो वक्त्य माननीय सदस्य ने उठाया है, मैं अपने को उससे सम्बंध करता हूँ।
SHRIMATI BRINDA KARAT (West Bengal): Sir, I associate myself with what the hon. Member, Shri Rudra Narayan Pany, has said.
SHRI MOHAMMED AMIN (West Bengal): Sir, I associate myself with what the hon. Member, Shri Rudra Narayan Pany, has said.
SHRI T.K. RANGARAJAN (Tamil Nadu): Sir, I associate myself with what the hon. Member, Shri Rudra Narayan Pany, has said.
SHRI R.C. SINGH (West Bengal): Sir, I associate myself with what the hon. Member, Shri Rudra Narayan Pany, has said.
**Need for effective mechanism for solid waste management**
SHRI THOMAS SANGMA (Meghalaya): Sir, I rise to bring forth my concerns about the lack of effective policy on waste management. There is no Indian policy document which examines
waste as part of a cycle of production-consumption-recovery or perceives waste through a prism of overall sustainability. In fact, interventions have been fragmented and are often contradictory. The new Municipal Solid Waste Management Rules, 2000, which came into effect from January, 2004, fails even to manage waste in a cyclic process. Waste management is still a linear system of collection and disposal, creating health and environmental hazards. Urban India is likely to face a massive waste disposal problem in the coming years. A closer look at the current and future scenario reveals that waste needs to be treated holistically, recognizing its natural resource roots as well as health impacts. Critical issues, such as, industry responsibility, a critical paradigm to enable sustainable recycling and to catalyze waste reduction through, say better packing, have not been touched upon. Any new paradigm should include a cradle-to-grave approach, with responsibility being shared by many stakeholders, including product manufacturers, consumers and communities, the recycling industry, trade municipalities and the urban poor. While holistic solutions are available, municipalities have struggled to implement them without proper planning and support from various Ministries.
Realising the need for a rethinking over waste management in the country, I urge upon the Government of India to pay the required attention towards this neglected issue and take appropriate steps in improving the waste management in the country.
SHRI ASHOK GANGULY (Nominated): Sir, I associate myself with the Special Mention made by the hon. Member.
SHRI RANJITSINH VIJAYSINH MOHITE-PATIL (Maharashtra): Sir, I also associate myself with the Special Mention made by the hon. Member.
SHRIMATI B. JAYASHREE (Nominated): Sir, I associate myself with the Special Mention made by the hon. Member.
PAPERS LAID ON THE TABLE
Quarterly review of trends in relation to Budget (2010-11)
THE MINISTER OF STATE IN THE MINISTRY OF FINANCE (SHRI NAMO NARAIN MEENA): Sir, I lay on the Table, under sub-section (1) of Section 7 of the Fiscal Responsibility and Budget Management Act, 2003, a copy (in English and Hindi) of the Statement on Quarterly Review of the trends in receipts and expenditure in relation to the Budget at the end of the third quarter of the financial year 2010-11.
[Placed in Library. See No. L.T. 4397/15/11]
I. Report and Accounts (2009-10) of NACIL, New Delhi and Pawan Hans Helicopters Ltd., New Delhi and related papers
II. Report and Accounts (2009-10) of IGRUA, Raebareli and related papers.
THE MINISTER OF STATE OF THE MINISTRY OF CONSUMER AFFAIRS, FOOD AND PUBLIC DISTRIBUTION (PROF. K.V. THOMAS): Sir, on behalf of Shri Vayalar Ravi, I lay on the Table—
I. (1) A copy each (in English and Hindi) of the following papers under sub-section (1) of Section 619A of the Companies Act, 1956:—
(i) (a) Annual Report and Accounts of the National Aviation Company of India Limited (NACIL), New Delhi, for the year 2009-10, together with the Auditor’s Report on the Accounts and the comments of the Comptroller and Auditor General of India thereon.
(b) Review by Government on the working of the above Company.
[Placed in Library. See No. L.T. 4401/15/11]
(ii) (a) Annual Report and Accounts of the Pawan Hans Helicopters Limited, New Delhi, for the year 2009-10, together with the Auditor’s Report on the Accounts and the comments of the Comptroller and Auditor General of India thereon.
(b) Statement by Government accepting the above Report.
(2) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4546/15/11]
II. A copy each (in English and Hindi) of the following papers:-
(a) Annual Report and Accounts of the Indira Gandhi Rashtriya
Uran Akademi, Fursatganj, Raebareli, for the year 2009-10, together with the Auditor’s Report on the Accounts.
(b) Statement by Government accepting the above Report.
(c) Statement giving reasons for the delay in laying the papers mentioned at (a) above.
[Placed in Library. See No. L.T. 4241/15/11]
Report and Accounts (2009-10) of Centre for Studies in Civilizations for implementation of the Project of History of Indian Science, Philosophy and Culture, New Delhi and related paper
THE MINISTER OF HUMAN RESOURCE DEVELOPMENT AND THE MINISTER OF COMMUNICATIONS AND INFORMATION TECHNOLOGY (SHRI KAPIL SIBAL): Sir, I lay on the Table, a copy each (in English and Hindi) of the following papers:--
(a) Annual Report and Accounts of the Centre for Studies in Civilizations for implementation of the Project of History of Indian Science, Philosophy and Culture, New Delhi, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Centre.
(c) Statement giving reasons for the delay in laying the papers mentioned at (a) above.
[Placed in Library. See No. L.T. 4402/15/11]
Report and Accounts (2009-10) of various Hotel Management and Catering Institutes and related papers
THE MINISTER OF TOURISM (SHRI SUBODH KANT SAHAY): Sir, I lay on the Table, a copy each (in English and Hindi) of the following papers:-
(a) Annual Report and Accounts of the National Council for Hotel Management and Catering Technology (NCHMCT), Noida, Uttar Pradesh, for the year 2009-10, together with the Auditor’s Report on the Accounts.
(b) Annual Report and Accounts of the Institute of Hotel Management, Catering Technology and Applied Nutrition, Srinagar, Jammu and Kashmir, for the year 2009-10, together with the Auditor’s Report on the Accounts.
(c) Annual Report and Accounts of the Institute of Hotel Management, Catering and Nutrition, Pusa, New Delhi, for the year 2009-10, together with the Auditor’s Report on the Accounts.
(d) Review by Government on the working of the above Council/Institutes.
[Placed in Library. See No. L.T. 4403/15/11]
I. Report and Accounts (2009-10) of various Shipping companies and related papers.
II. Report and Accounts (2009-10) of Kolkata Port Trust and related papers.
THE MINISTER OF STATE IN THE MINISTRY OF PLANNING; THE MINISTER OF STATE IN THE MINISTRY OF PARLIAMENTARY AFFAIRS; THE MINISTER OF STATE IN THE MINISTRY OF SCIENCE AND TECHNOLOGY AND THE MINISTER OF STATE IN THE MINISTRY OF EARTH SCIENCES (SHRI ASHWANI KUMAR): Sir, on behalf of Shri G. K. Vasan, I lay on the Table:-
I. (1) A copy each (in English and Hindi) of the following papers, under sub-section (1) of section 619A of the Companies Act, 1956:
(i) (a) Thirty-fourth Annual Report and Accounts of the Dredging Corporation of India Limited (DCI), Delhi, for the year 2009-10, together with the Auditor’s Report on the Accounts and the comments of the Comptroller and Auditor General of India thereon.
(b) Review by Government on the working of the above Corporation.
[Placed in Library. See No. L.T. 4164/15/11]
(ii) (a) Annual Report and Accounts of the Shipping Corporation of India Limited, Mumbai, for the year 2009-10, together with the Auditor’s Report on the Accounts and the comments of the Comptroller and Auditor General of India thereon.
(b) Annexure to the Directors’ Report of the Shipping Corporation of India Limited, Mumbai, for the year 2009-10.
(c) Review by Government on the working of the above Corporation.
(2) Statement (in English and Hindi) giving reasons for the
delay in laying the papers mentioned at (1) (ii) above.
[Placed in Library. See No. L.T. 4163/15/11]
II. (1) A copy each (in English and Hindi) of the following papers, under sub-section (2) of Section 103 of the Major Port Trusts Act, 1963:-
(a) Annual Accounts of the Kolkata Port Trust, Kolkata, for the year 2009-10, and the Audit Report thereon.
(b) Review by Government on the Annual Accounts and Audit Report of the above Port Trust.
(2) Statement (in English and Hindi) giving reasons for the delay in laying the papers maintained at (1) above.
[Placed in Library. See No. L.T. 4245/15/11]
**Report and Accounts (2009-10) of JNARDDC, Nagpur and related papers**
SHRI ASHWANI KUMAR: Sir, on behalf of Shri Dinsha J. Patel, I lay on the Table, a copy each (in English and Hindi) of the following papers:
(a) Annual Report and Accounts of the Jawaharlal Nehru Aluminium Research Development and Design Centre (JNARDDC), Nagpur, for the year 2009-10, together with the Auditor’s Report on the Accounts.
(b) Review by Government on the working of the above Centre.
(c) Statement giving reasons for the delay in laying the papers mentioned at (a) above.
[Placed in Library. See No. L.T. 4384/15/11]
**Notifications of the Ministry of Environment and Forests**
THE MINISTER OF STATE OF THE MINISTRY OF ENVIRONMENT AND FORESTS (SHRI JAIRAM RAMESH): Sir, I lay on the Table:
(i) A copy each (in English and Hindi) of the following Notifications of the Ministry of Environment and Forests, under Section 26 of the Environment (Protection) Act, 1986:
(1) S.O. 1333 (E), dated the 7th June, 2010, regarding Constitution of the State Level Environment Impact Assessment Authority (SEIAA), West Bengal.
(2) S.O. 1803 (E), dated the 23rd July, 2010, regarding Constitution of the State Level Environment Impact Assessment Authority (SEIAA), Gujarat.
(3) S.O. 2402 (E), dated the 4th October, 2010, regarding
Constitution of the State Level Environment Impact Assessment Authority (SEIAA), Karnataka.
(4) S.O. 2551 (E), dated the 13th October, 2010, regarding Constitution of the State Level Environment Impact Assessment Authority (SEIAA), Uttar Pradesh.
(5) S.O. 2646 (E), dated the 26th October, 2010, regarding Constitution of the State Level Environment Impact Assessment Authority (SEIAA), Andhra Pradesh.
(6) S.O. 254 (E), dated the 7th February, 2011, regarding Constitution of the State Environment Impact Assessment Authority, Punjab.
(7) S.O. 255 (E), dated the 7th February, 2011, regarding Constitution of the State Environment Impact Assessment Authority, Bihar.
(8) S.O. 2780 (E), dated the 12th November, 2010, amending Notification No. S.O. 49(E), dated the 8th January, 2008, to substitute certain entries in the original Notification.
[Placed in Library. See No. L.T. 4339/15/11]
(9) G.S.R. 951 (E), dated the 4th December, 2010, publishing the Wetlands (Conservation and management) Rules, 2010.
[Placed in Library. See No. L.T. 3978/15/11]
(ii) A copy (in English and Hindi) of the Ministry of Environment and Forests Notification No. S.O. 473 (E), dated the 3rd March, 2011, notifying species on the verge of extinction for the State of Tamil Nadu, under sub-section (3) Section 62 of the Biological Diversity Act, 2002.
[Placed in Library. See No. L.T. 4407/15/11]
Outcome Budget (2011-12) of the Ministry of Youth Affairs and Sports
THE MINISTER OF STATE IN THE MINISTRY OF COAL (SHRI PRATIK PRAKASHBAPU PATIL): Sir, on behalf of Shri Ajay Maken, I lay on the Table, a copy (in English and Hindi) of the Outcome Budget, for the year 2011-12, in respect of the Ministry of Youth Affairs and Sports.
[Placed in Library. See No. L.T. 4408/15/11]
Notification of the Ministry of Consumer Affairs, Food and Public Distribution
PROF. K.V. THOMAS: Sir, I lay on the Table, under sub-section (4) of Section 52 of the Legal Metrology Act, 2009, a copy (in English and
Hindi) of the Ministry of Consumer Affairs, Food and Public Distribution (Department of Consumer Affairs) Notification No. G.S.R. 202 (E), dated the 9th March, 2011, publishing the Legal Metrology (Packaged Commodities) Rules, 2011.
[Placed in Library. See No. L.T. 4409/15/11]
I. Notifications of the Ministry of Human Resource Development
II. Reports and Accounts (2008-09 and 2009-2010) of various Universities, NIT’s, Navodaya Vidyalaya Samiti, IITs, IIMs Research Institutes, Engineering Institutes, Boards, Sarva Siksha Abhiyan and Councils and related papers.
THE MINISTER OF STATE IN THE MINISTRY OF HUMAN RESOURCE DEVELOPMENT (SHRIMATI D. PURANDESHWARI): Sir, I lay on the Table:-
I. (i) A copy each (in English and Hindi) of the following Notifications of the Ministry of Human Resource Development (Department of School Education and Literacy), under Section 33 of the National Council for Teacher Education Act, 1993, along with delay statement on the Notifications:-
(1) F. No. 51-1/2009/NCTE (N&S), dated the 26th July, 2010, publishing the National Council for Teacher Education (Recognition Norms and Procedure) Second Amendment Regulations, 2010.
[Placed in Library. See No. L.T. 4458/15/11]
(2) No. F. 47-49/2009/NCTE/CDN (Pt.), dated the 20th August, 2010, regarding nomination of Chairperson of the Eastern Regional Committee, Bhubaneswar.
[Placed in Library. See No. L.T. 4458/15/11]
(3) No. F. 47-49/2009/NCTE/CDN, dated the 25th August, 2010, regarding termination of Northern Regional Committee, Jaipur.
[Placed in Library. See No. L.T. 4459/15/11]
(4) F. No. 51-3/2009/Pt./NCTE (N&S), dated the 25th August, 2010, publishing the National Council for Teacher Education (Guidelines for regulation of tuition fees and other fees chargeable by unaided teacher education institutions) Amendment
(5) F. No. 61-03/20/2010/NCTE (N&S), dated the 25th August, 2010, regarding minimum qualifications for a person to be eligible for appointment as teacher.
[Placed in Library. See No. L.T. 4458/15/11]
(ii) A copy each (in English and Hindi) of the following Notifications of the Ministry of Human Resource Development (Department of Higher Education), under Section 24 of the All India Council of Technical Education Act, 1987:-
(1) No. F. 37-3/Legal/2008, dated the 22nd February, 2010, repealing All India Council for Technical Education (Constitution and functions of National Board of Accreditation) Regulations, 2008.
[Placed in Library. See No. L.T. 4417/15/11]
(2) No. F. 37-3/Legal/2010, dated the 6th-12th February, 2010, publishing the All India Council for Technical Education (Grant of Approvals for Technical Institutions) Regulations, 2010.
[Placed in Library. See No. L.T. 4416/15/11]
(iii) A copy each (in English and Hindi) of the following Notifications of the Ministry of Human Resource Development (Department of Higher Education), under subsection (2) of Section 40 of the Indira Gandhi National Open University Act, 1985:-
(1) No. IG/Admn. (G)/Ord.13/2005/1948, dated the 16th-22nd October, 2010, amending Clause 2.4 of the Ordinance on Research Degree Programmes of the Indira Gandhi National Open University.
[Placed in Library. See No. L.T. 4264/15/11]
(2) No. 1G/Admn. (G) /Ord.5/2009/1978, dated the 6th-12th November, 2010 regarding amendment to the Ordinance on Recognition of Examinations/Degrees of the University.
[Placed in Library. See No. L.T. 4460/15/11]
(3) No. IG/Admn. (G)/Ord.9/2000/1998, dated the 27th November, 2010 to 3rd December, 2010, adding sub-clause (5) of
Clause 9 of the Ordinance 9 on the Conduct of Examination and Evaluation of Student Performance of the Indira Gandhi National Open University.
[Placed in Library. See No. L.T. 4264/15/11]
II (i) (1) A copy each (in English and Hindi) of the following papers under sub-section (3) of Section 30 and sub-section (5) of Section 31 of the Assam University Act, 1989:
(a) Seventeenth Annual Report of the Assam University, Silchar, for the year 2009-10.
(b) Annual Accounts of the Assam University, Silchar, for the year 2009-10 and the Audit Report thereon.
(c) Review by Government on the working of the above University.
(2) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4447/15/11]
(ii) (1) A copy each (in English and Hindi) of the following papers under sub-section (3) of Section 13 and sub-section (3) of Section 13A of the Banaras Hindu University Act, 1915:-
(a) Annual Report of the Banaras Hindu University, Varanasi, for the year 2009-10.
(b) Annual Accounts of the Banaras Hindu University, Varanasi, for the year 2009-10, and the Audit Report thereon.
(c) Review by Government on the working of the above University.
(2) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4448/15/11]
(iii) A copy each (in English and Hindi) of the following papers, under sub-section (3) of Section 30 and sub-section (4) of Section 31 of the Central Universities Act, 2009:-
(i) (a) First Annual Report of the Central University of Karnataka, Gulbarga, for the year 2008-09.
(b) Annual Accounts of the Central University of Karnataka, Gulbarga, for the year 2008-09 and the Audit Report thereon.
(c) Review by Government on the working of the above University.
[Placed in Library. See No. L.T. 4448/15/11]
(ii) (a) Second Annual Report of the Central University of Karnataka, Gulbarga, for the year 2009-10.
(b) Annual Accounts of the Central University of Karnataka, Gulbarga, for the year 2009-10, and the Audit Report thereon.
(c) Review by Government on the working of the above University.
(d) Statements (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4451/15/11]
(iv) (1) A copy each (in English and Hindi) of the following papers, under sub-section (2) of Section 30 and sub-section (4) of Section 31 of the Mahatma Gandhi Antarrashtriya Hindi Vishwavidyalaya Act, 1996:-
(a) Annual Report and Accounts of the Mahatma Gandhi Antarrashtriya Hindi Vishwavidyalaya, Wardha, for the year 2009-10, together with the Auditor’s Report on the Accounts.
(b) Statement by Government accepting the above Report.
(2) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4254/15/11]
(v) A copy each (in English and Hindi) of the following papers, under sub-section (4) of Section 29 of the University of Hyderabad Act, 1974:-
(a) Annual Accounts of the University of Hyderabad, Hyderabad, for the year 2009-10 and the Audit Report thereon.
(b) Statement giving reasons for the delay in laying the papers mentioned at (a) above.
[Placed in Library. See No. L.T. 4056/15/11]
(vi) A copy each (in English and Hindi) of the following papers under sub-section (3) of Section 29 of the Pondicherry University Act, 1985:--
(a) Twenty-fourth Annual Report of the Pondicherry University, Pondicherry, for the year 2009-10.
(b) Review by Government on the working of the above University.
(c) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4130/15/11]
(vii) (1) A copy each (in English and Hindi) of the following papers, under sub-section (3) of Section 30 of the Mizoram University Act, 2000:-
(a) Annual Report of the Mizoram University, Aizawl, for the year 2009-10.
(b) Review by Government on the working of the above University.
(2) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4438/15/11]
(viii) (1) A copy each (in English and Hindi) of the following papers, under Section 32 and 33 of the English and Foreign Languages University Act, 2006:-
(a) Annual Report of the English and Foreign Languages University, Hyderabad, for the year 2008-09.
(b) Annual Accounts of the English and Foreign Languages University, Hyderabad, for the year 2008-09 and the Audit Report thereon.
(c) Statement by Government accepting the above Report.
(d) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
(ix) A copy each (in English and Hindi) of the following papers, under sub-section (4) of the Section 34 of the Tripura University Act, 2006:-
(a) Annual Accounts of the Tripura University, Suryamaninagar, Tripura, for the year 2009-10, and the Audit Report thereon.
(b) Statement giving reasons for the delay in laying the papers mentioned at (a) above.
[Placed in Library. See No. L.T. 4439/15/11]
(x) A copy each (in English and Hindi) of the following papers, under sub-section (4) of Section 22 of the National Institutes of Technology Act, 2007:-
1. (a) Annual Report and Accounts of the National Institute of Technology, Jaipur, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
[Placed in Library. See No. L.T. 4422/15/11]
2. (a) Fiftieth Annual Report of the National Institute of Technology, Durgapur, for the year 2009-10.
(b) Fiftieth Annual Accounts of the National Institute of Technology, Durgapur, for the year 2009-10 and the Audit Report thereon.
(c) Review by Government on the working of the above Institute.
[Placed in Library. See No. L.T. 4440/15/11]
3. (a) Annual Report and Accounts of the National Institute of Technology (formerly Regional Engineering College), Warangal, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
[Placed in Library. See No. L.T. 4429/15/11]
4. (a) Annual Report and Accounts of the National Institute of Technology, Jamshedpur, for the
year 2007-08, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
[Placed in Library. See No. L.T. 4421/15/11]
5. (a) Annual Report and Accounts of the Visvesvaraya National Institute of Technology, Nagpur, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
[Placed in Library. See No. L.T. 4420/15/11]
6. (a) Annual Report and Accounts of the National Institute of Technology, Agartala, for the year 2008-09, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
[Placed in Library. See No. L.T. 4431/15/11]
7. (a) Annual Report and Accounts of the National Institute of Technology, Karnataka, Surathkal, Mangalore, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
[Placed in Library. See No. L.T. 4452/15/11]
8. (a) Annual Report and Accounts of the National Institute of Technology, Calicut, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
[Placed in Library. See No. L.T. 4443/15/11]
9. (a) Annual Report and Accounts of the National Institute of Technology, Rourkela, for the year 2009-10, together with the Audit Report.
(b) Review by Government on the working of the above
(c) Statements (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4441/15/11]
(xi) A copy each (in English and Hindi) of the following papers, under sub-section (3) of Section 27 and sub-section (4) of Section 28 of the Jamia Millia Islamia Act, 1988:-
(a) Annual Report of the Jamia Millia Islamia (JMI), New Delhi, for the year 2009-10.
(b) Annual Accounts of the Jamia Millia Islamia (JMI), New Delhi, for the year 2009-10 and the Audit Report thereon.
(c) Review by Government on the working of the above University.
(d) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4455/15/11]
(xii) A copy each (in English and Hindi) of the following papers, under, under sub-section (3) of Section 31 of the Nagaland University Act, 1989:-
(a) Fifteenth Annual Report of the Nagaland University, Kohima, for the year 2009-10.
(b) Review by Government on the working of the above University.
(c) Statement giving reasons for the delay in laying the papers mentioned at (1) above.
[Placed in Library. See No. L.T. 4445/15/11]
(xiii) A copy each (in English and Hindi) of the following papers:-
(1) (a) First Annual Report and Accounts of the Central University of Rajasthan, Ajmer, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above University.
(c) Statement giving reasons for the delay in laying the papers mentioned at (i) (a) above.
[Placed in Library. See No. L.T. 4446/15/11]
(2) (a) First Annual Report of the Central University of Kashmir, Srinagar, for the year 2009-10.
(b) Annual Accounts of the Central University of Kashmir, Srinagar, for the year 2009-10, and the Audit Report thereon.
(c) Review by Government on the working of the above University.
(d) Statement giving reasons for the delay in laying the papers mentioned at (ii) (a) and (b) above.
[Placed in Library. See No. L.T. 4454/15/11]
(3) (a) Annual Report of the Navodaya Vidyalaya Samiti, New Delhi, for the year 2009-10.
(b) Annual Accounts of the Navodaya Vidyalaya Samiti, New Delhi, for the year 2009-10, and the Audit Report thereon.
(c) Review by Government on the working of the above Samiti.
(d) Statement giving reasons for the delay in laying the papers mentioned at (iii) (a) and (b) above.
[Placed in Library. See No. L.T. 4418/15/11]
(4) (a) Annual Report and Accounts of the School of Planning and Architecture (SPA), Vijayawada, for the year 2008-09, together with the Audit Report.
(b) Review by Government on the working of the above School.
(c) Statement giving reasons for the delay in laying the papers mentioned at (iv) (a) above.
[Placed in Library. See No. L.T. 4457/15/11]
(5) (a) Annual Accounts of the Indian Institute of Technology (IIT), Chennai, for the year 2009-10 and the Audit Report thereon, under sub-section (4) of Section 23 of the Institutes of Technology Act, 1961.
(b) Statement giving reasons for the delay in laying the papers mentioned at (v) (a) above.
[Placed in Library. See No. L.T. 3924/15/11]
(6) (a) Annual Report of the Indian Institute of Technology (IIT), Hyderabad, for the year 2009-10.
(b) Annual Accounts of the Indian Institute of Technology (IIT), Hyderabad, for the year 2009-10, and Audit Report thereon,
under Sub-section (4) of Section 23 of the Institutes of Technology Act, 1961.
(c) Statement by Government accepting the above Report.
(d) Statement giving reasons for the delay in laying the papers mentioned at (vi) (a) and (b) above.
[Placed in Library. See No. L.T. 4056/15/11]
(7) (a) Annual Accounts of the Indian Institute of Technology (IIT), Kanpur, for the year 2009-10, and Audit Report thereon, under sub-section (4) of Section 23 of the Institutes of Technology Act, 1961.
(b) Statement giving reasons for the delay in laying the papers mentioned at (vii) (a) above.
[Placed in Library. See No. L.T. 4051/15/11]
(8) (a) Annual Report of the Indian Institute of Technology (IIT), Gandhinagar, for the year 2009-10.
(b) Statement by Government accepting the above Report.
(c) Statement giving reasons for the delay in laying the papers mentioned at (viii) (a) above.
[Placed in Library. See No. L.T. 4545/15/11]
(9) (a) Annual Accounts of the Indian Institute of Technology (IIT), Kharagpur, for the year 2009-10, and the Audit Report thereon, under sub-section (4) of Section 23 of the Institutes of Technology Act, 1961.
(b) Statement giving reasons for the delay in laying the papers mentioned at (ix) (a) above.
[Placed in Library. See No. L.T. 4048/15/11]
(10) (a) Annual Report and Accounts of the Indian Institute of Management (IIM), Indore, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
(c) Statement giving reasons for the delay in laying
the papers mentioned at (x) (a) above.
[Placed in Library. See No. L.T. 4419/15/11]
(11) (a) Annual Report and Accounts of the Indian
Institute of Management (IIM), Lucknow, for the
year 2009-10, together with the Audit Report on
the Accounts.
(b) Review by Government on the working of the above
Institute.
(c) Statement giving reasons for the delay in laying
the papers mentioned at (xi) (a) above.
[Placed in Library. See No. L.T. 4461/15/11]
(12) (a) Thirty-sixth Annual Report of the North
Eastern Hill University, Shillong, for the year
2009-10.
(b) Annual Accounts of the North Eastern Hill
University, Shillong, for the year 2009-10 and
the Audit Report thereon.
(c) Review by Government on the working of the above
University.
(d) Statement giving reasons for the delay in laying
the papers mentioned at (xii) (a) above.
[Placed in Library. See No. L.T. 4450/15/11]
(13) (a) Annual Report and Accounts of the Indian
Institute of Science Education and Research (IISER),
Thiruvananthapuram, for the year 2009-10,
together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above
Institute.
(c) Statement giving reasons for the delay in laying
the papers mentioned at (xiii) (a) above.
(14) (a) Annual Report of the Indian Institute of Science Education and Research (IISER), Bhopal, for the year 2009-10.
(b) Annual Accounts of the Indian Institute of Science Education and
(c) Review by Government on the working of the above Institute.
(d) Statement giving reasons for the delay in laying the papers mentioned at (xiv) (a) and (b) above.
[Placed in Library. See No. L.T. 3974/15/11]
(15) (a) Annual Report and Accounts of the Indian Institute of Science Education and Research (IISER), Pune, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
(c) Statement giving reasons for the delay in laying the papers mentioned at (xv) (a) above.
[Placed in Library. See No. L.T. 4424/15/11]
(16) (a) Annual Report of the Indian Institute of Science (IISc), Bangalore, for the year 2009-10.
(b) Annual Accounts of the Indian Institute of Science (IISc), Bangalore, for the year 2009-10 and the Audit Report thereon.
(c) Review by Government on the working of the above Institute.
(d) Statement giving reasons for the delay in laying the papers mentioned at (xvi) (a) and (b) above.
[Placed in Library. See No. L.T. 4425/15/11]
(17) (a) Annual Report and Accounts of the National Institute of Technical Teachers’ Training and Research (NITTTR), Bhopal, for the year 2009-10, together
with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
(c) Statement giving reason for not laying the papers mentioned at (xvii) (a) above.
[Placed in Library. See No. L.T. 4436/15/11]
(18) (a) Annual Report and Accounts of the North Eastern Regional Institute of Science and Technology (NERIST), Itanagar, Arunachal Pradesh, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
(c) Statement giving reasons for the delay in laying the papers mentioned at (xviii) (a) above.
[Placed in Library. See No. L.T. 4437/15/11]
(19) (a) Annual Report and Accounts of the Sant Longowal Institute of Engineering and Technology (SLIET), Longowal, Punjab, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
(c) Statement giving reasons for the delay in laying the papers mentioned at (xix) (a) above.
[Placed in Library. See No. L.T. 4435/15/11]
(20) (a) Annual Report and Accounts of the National Institute of Technical Teachers’ Training and Research (NITTTR), Chandigarh, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
(c) Statement giving reason for not laying the
papers mentioned at (xx) (a) above.
[Placed in Library. See No. L.T. 4434/15/11]
(21) (a) Annual Report and Accounts of the National Institute of Technical Teachers’ Training and Research (NITTTR), Kolkata, for the year
2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
(c) Statement giving reason for not laying the papers mentioned at (xxi) (a) above.
[Placed in Library. See No. L.T. 4433/15/11]
(22) (a) Annual Report and Accounts of the Central Institute of Technology (CIT), Kokrajhar, Assam, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
(c) Statement giving reasons for the delay in laying the papers mentioned at (xxii) (a) above.
[Placed in Library. See No. L.T. 4426/15/11]
(23) (a) Annual Report of the National Institute of Industrial Engineering (NITIE), Mumbai, for the year 2009-10.
(b) Annual Accounts of the National Institute of Industrial Engineering (NITIE), Mumbai, for the year 2009-10, and the Audit Report thereon.
(c) Review by Government on the working of the above Institute.
(d) Statement giving reasons for the delay in laying the papers mentioned at (xxiii) (a) and (b) above.
[Placed in Library. See No. L.T. 4428/15/11]
(24) (a) Annual Report and Accounts of the Board of Apprenticeship Training (BOAT), Western Region, Mumbai, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Board.
(c) Statement giving reasons for delay in laying the papers mentioned at (xxiv) (a) above.
[Placed in Library. See No. L.T. 4427/15/11]
(25) (a) Annual Report and Accounts of the Jharkhand Education Project Council, Ranchi, for the year 2008-09, together with the Auditor’s Report on the Accounts.
(b) Statement by Government accepting the above Report.
(c) Statement giving reasons for the delay in laying the papers mentioned at (xxv) (a) above.
[Placed in Library. See No. L.T. 4456/15/11]
(26) (a) Annual Report and Accounts of the National Council of Rural Institutes (NCRI), Hyderabad, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Review by Government on the working of the above Institute.
(c) Statement giving reasons for the delay in laying the papers mentioned at (xxvi) (a) above.
[Placed in Library. See No. L.T. 4263/15/11]
(27) (a) Annual Report and Accounts of the Primary Education Development Society of Kerala (PEDSK) in respect of the Sarva Shiksha Abhiyan, Kerala, for the year 2008-09, together with the Auditor’s Report on the Accounts.
(b) Statement by Government accepting the above Report.
(c) Statement giving reasons for the delay in laying the papers mentioned at (xxvii) (a) above.
[Placed in Library. See No. L.T. 4260/15/11]
(28) (a) Annual Report and Accounts of the Sarva Shiksha Abhiyan Rajya Mission, Tripura, for the year 2008-09, together with the Auditor’s Report on the Accounts.
(b) Statement by Government accepting the above Report.
(c) Statement giving reasons for the delay in laying
the papers mentioned at (xxviii) (a) above.
[Placed in Library. See No. L.T. 4431/15/11]
(29) (a) Annual Report and Accounts of the Sarva Shiksha
Abhiyan Society, U.T. Chandigarh, for the year
2009-10, together with the Auditor’s Report on
the Accounts.
(b) Statement by Government accepting the above Report.
[Placed in Library. See No. L.T. 4430/15/11]
(30) (a) Annual Report and Accounts of the Central Institute of Classical Tamil (CICT), Chennai, for the year 2009-10, together with the Audit Report on the Accounts.
(b) Statement by Government accepting the above Report.
(c) Statement giving reasons for the delay in laying the papers mentioned at (xxx) (a) above.
[Placed in Library. See No. L.T. 4253/15/11]
I. Report and Accounts (2009-10) of NEPA Ltd., Nepanagar, Madhya Pradesh and related papers
II. Outcome budget (2011-12) of the Ministry of Heavy Industries and Public Enterprises
THE MINISTER OF STATE IN THE MINISTRY OF HEAVY INDUSTRIES AND PUBLIC ENTERPRISES (SHRI A. SAI PRATHAP): Sir, I lay on the Table –
I. A copy each (in English and Hindi) of the following papers, under sub-section (1) of Section 619A of the Companies Act, 1956:-
(a) Sixty-third Annual Report and Accounts of the NEPA Limited, Nepanagar, Madhya Pradesh, for the year 2009-10, together with the Auditor’s Report on the Accounts and the comments of the Comptroller and Auditor General of India thereon.
(b) Review by Government on the working of the above Company.
(c) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned (l) above.
[Placed in Library. See No. L.T. 4474/15/11]
II. A copy (in English and Hindi) of the Outcome Budget, for the year 2011-12, in respect of the Department of Heavy Industry in the Ministry of Heavy Industries and Public Enterprises.
Notifications of the Ministry of Health and Welfare
THE MINISTER OF STATE IN THE MINISTRY OF HEALTH AND WELFARE (SHRI DINESH TRIVEDI): Sir, I lay on the Table, under Section 38 of the Drugs and Cosmetics Act, 1940, a
copy each (in English and Hindi) of the following Notifications of the Ministry of Health and Family Welfare:
(1) G.S.R. 82 (E), dated the 10th February, 2011, regarding prohibition of manufacture, sale and distribution of certain drugs.
[Placed in Library. See No. L.T. 4481/15/11]
(2) G.S.R. 101 (E), dated the 18th February, 2011, publishing the Drugs and Cosmetics (2nd Amendment) Rules, 2011.
[Placed in Library. See No. L.T. 4487/15/11]
I. Report and Accounts (2009-10) of BECIL, NOIDA and related papers
II. Outcome budget (2011-12) of the Ministry of Information and Broadcasting
SHRI ASHWANI KUMAR: Sir, on behalf of Dr. S. Jagathrakshakan, I lay on the Table:
I. (1) A copy each (in English and Hindi) of the following papers, under sub-section (1) of Section 619A of the Companies Act, 1956:
(a) Fifteenth Annual Report and Accounts of the Broadcast Engineering Consultants India Limited (BECIL), NOIDA, for the year 2009-10, together with the Auditor’s Report on the Accounts and the comments of the Comptroller and Auditor General of India thereon.
(b) Review by Government on the working of the above Company.
(2) Statement (in English and Hindi) giving reasons for the delay in laying the papers mentioned at (1) above.
II. A copy (in English and Hindi) of the Outcome Budget, for the year 2011-12, in respect of the Ministry of Information and Broadcasting.
[Placed in Library. See No. L.T. 4543/15/11]
संचार और सूचना पुरौद्योगिकी मंत्रालय की अधिसूचना
(श्री सचिन पायलट) मंत्रालय में राज्य मंत्री
डाकघर अधिनियम, 1898 की धारा 74 की उप-धारा (4) के अधीन भारतीय डाकघर (दक्षितीय संशोधन) नायिम, 2011 को प्रकाशित करने वाली संचार और सूचना पुरौद्योगिकी मंत्रालय (डाक विभाग) की अधिसूचना सं. साक.नि. 37 (अ), दक्षिण 19 जनवरी, 2011 की एक पुस्ति (अंबेजी तथा हिंदी में) सभा पतल पर रखता हैं।
[Placed in Library. See No. L.T. 4387/15/11]
Reports and Accounts (2009-10) of various companies and related papers
THE MINISTER OF STATE IN THE MINISTRY OF PETROLEUM AND NATURAL GAS AND THE MINISTER OF STATE IN THE MINISTRY OF CORPORATE AFFAIRS (SHRI R.P.N. SINGH): Sir, I lay on the Table –
I. A copy each (in English and Hindi) of the following papers under sub-section (1) of Section 619A of the Companies Act, 1956:-
(i) (a) Sixth Annual Report and Accounts of the Indian Strategic Petroleum Reserves Limited (ISPRL), New Delhi, for the year 2009-10, together with the Auditor’s Report on the Accounts and the comments of the Comptroller and Auditor General of India thereon.
(b) Review by Government on the working of the above Company.
[Placed in Library. See No. L.T. 4544/15/11]
(ii) (a) Fifty-first Annual Report and Accounts of the Oil India Limited (OIL), Dibrugarh, Assam, for the year 2009-10, together with the Auditor’s Report on the Accounts and the comments of the Comptroller and Auditor General of India thereon.
(b) Review by Government on the working of the above Company.
[Placed in Library. See No. L.T. 4391/15/11]
(iii) (a) Seventeenth Annual Report and Accounts of the Oil and Natural Gas Corporation Limited (ONGC), New Delhi, for the year 2009-10, together with the Auditor’s Report on the Accounts and the comments of the Comptroller and Auditor General of India thereon.
(b) Review by Government on the working of the above Corporation.
[Placed in Library. See No. L.T. 4390/15/11]
Notifications of the Ministry of Power
SHRI ASHWANI KUMAR: Sir, on behalf of Shri K.C. Venugopal, I lay on the Table, a copy each (in English and Hindi) of the following Notifications of the Ministry of Power:
(1) No. JERC-14/2010, dated the 15th December, 2010, publishing the Joint Electricity Regulatory Commission for State of Goa and Union Territories (Procurement of Renewable energy) Regulations, 2010.
(2) No. JERC-19/14/2010, dated the 19th February, 2011, publishing Corrigendum to Joint Electricity Regulatory Commission (Procurement of Renewable Energy) Regulations, 2010.
[Placed in Library. See No. L.T. 4393/15/11]
Various reports (2010–11) of CAG
SHRI NAMO NARAIN MEENA: Sir, I lay on the Table, under clause (1) of article 151 of the Constitution, a copy each (in English and Hindi) of the following Reports:-
(i) Report of the Comptroller and Auditor General of India for the year ended March, 2010 (No. 23 of 2010–11): Union Government (Indirect Taxes – Central Excise – Compliance Audit);
[Placed in Library. See No. L.T. 4462/15/11]
(ii) Report of the Comptroller and Auditor General of India for the year ended March, 2010 (No. 24 of 2010–11): Union Government (Indirect Taxes – Customs – Compliance Audit);
[Placed in Library. See No. L.T. 4463/15/11]
(iii) Report of the Comptroller and Auditor General of India for the year ended March, 2009 (No. 25 of 2010–11): Union Government (Indirect Taxes – Service Tax – Performance Audit);
[Placed in Library. See No. L.T. 4464/15/11]
(iv) Report of the Comptroller and Auditor General of India for the year ended March, 2010 (No. 29 of 2010–11): Union Government (Indirect Taxes – Service Tax – Compliance Audit);
[Placed in Library. See No. L.T. 4465/15/11]
(v) Report of the Comptroller and Auditor General of India for the year ended March, 2009 (No. 30 of 2010–11): Union Government (Indirect Taxes – Central Excise and Service Tax – Performance Audit of Cenvat Credit Scheme).
[Placed in Library. See No. L.T. 4466/15/11]
(vi) Report of the Comptroller and Auditor General of India for the year ended March, 2010 (No. 35 of 2010–11): Union Government (Defence Services) (Performance Audit of Defence Estates
Management); and
[Placed in Library. See No. L.T. 4466/15/11]
(vii) Report of the Comptroller and Auditor General of India for the year ended March, 2010 (No. 36 of 2010-11): Union Government (Direct Taxes) – Taxation of assesses engaged in the Film and Television industry.
[Placed in Library. See No. L.T. 4467/15/11]
REPORTS OF THE DEPARTMENT RELATED PARLIAMENTARY STANDING COMMITTEE ON AGRICULTURE
श्री सत्यप्रत चनुवंदी (उत्तराखंड): महोदय, मैं वक्षिग – सर्बधित कृषि सबधी ससदीय सथायी समिति (2010-11) के नक्सिलिखित परितिवेदनों की एक-एक परिति (अच्छी तथा हल्दी में) सभा पटल पर रखता हैः:
(i) Seventeenth Report on the Action Taken by the Government on the Observations/Recommendations contained in the Sixth Report (Fifteenth Lok Sabha) of the Committee on ‘Demands for Grants (2010-11)’ of Ministry of Agriculture (Department of Agriculture and Co-operation);
(ii) Eighteenth Report on the Action Taken by the Government on the Observations/Recommendations contained in the Seventh Report (Fifteenth Lok Sabha) of the Committee on ‘Demands for Grants (2010-11)’ of Ministry of Agriculture (Department of Animal Husbandry, Dairying and Fisheries);
(iii) Nineteenth Report on the Action Taken by the Government on the Observations/Recommendations contained in the Eighth Report (Fifteenth Lok Sabha) of the Committee on ‘Demands for Grants (2010-11)’ of Ministry of Agriculture (Department of Agricultural Research and Education); and
(iv) Twentieth Report on the Action Taken by the Government on the Observations/Recommendations contained in the Ninth Report (Fifteenth Lok Sabha) of the Committee on ‘Demands for Grants (2010-11)’ of Ministry of Food Processing Industries.
STATEMENTS OF THE DEPARTMENT RELATED PARLIAMENTARY STANDING COMMITTEE ON SOCIAL JUSTICE AND EMPOWERMENT
SHRI PRAVEEN RASHTRAPAL (Gujarat): Sir, I lay on the Table, a copy each (in English and Hindi) of the following Statements of the Department-related Parliamentary Standing Committee on Social Justice and Empowerment (2010-11):
(i) Action taken by the Government on the recommendations contained in Chapter-I and final replies in respect of recommendations contained in Chapter-V of the Thirtyeighth Report of the Standing Committee on Social Justice and Empowerment (2008-2009) on Action Taken by Government on the recommendations contained in Thirty-second Report of the Standing Committee on Social Justice and Empowerment (2007-08) on “Voluntary Organizations Engaged in Implementing Integrated Programme for Welfare of Older Persons”;
(ii) Action taken by the Government on the recommendations contained in Chapter-I and final replies in respect of recommendations contained in Chapter-V of the Forty-first Report of the Standing Committee on Social Justice and Empowerment (2008-2009) on Action Taken by Government on the recommendations contained in Thirty-third Report of the Standing Committee on Social Justice and Empowerment (2007-08) on ‘Demands for Grants (2008-09)’ of the Ministry of Social Justice and Empowerment;
(iii) Action taken by the Government on the recommendations contained in Chapter-I and final replies in respect of recommendations contained in Chapter-V of the Seventh Report of the Standing Committee on Social Justice and Empowerment (2009-10) on Action Taken by Government on the recommendations contained in Third Report of the Standing Committee on Social Justice and Empowerment (2009-10) on ‘Demands for Grants (2009-10)’ of the Ministry of Minority Affairs; and
(iv) Action taken by the Government on the recommendations contained in Chapter-I and final replies in respect of recommendations contained in Chapter-V of the Eighth Report of the Standing Committee on Social Justice and Empowerment (2009-10) on Action Taken by Government on the recommendations contained in Second Report of the Standing Committee on Social Justice and Empowerment (2009-10) on
‘Demands for Grants (2009-10)’ of the Ministry of Tribal Affairs.
STATEMENT BY MINISTER
Status of implementation of recommendations contained in the Thirtieth Report of the Department-related Parliamentary Standing Committee on Finance
THE MINISTER OF STATE IN THE MINISTRY OF PETROLEUM AND NATURAL GAS AND THE MINISTER OF STATE IN THE MINISTRY OF CORPORATE AFFAIRS (SHRI R.P.N. SINGH):
Sir, I make a statement regarding status of implementation of recommendations contained in the Thirtieth Report of the Department-related Parliamentary Standing Committee on Finance on Demands for Grants, for the year 2010-11.
[MR. CHAIRMAN in the Chair.]
VALEDIRCTORY REMARKS
MR. CHAIRMAN: Hon. Members, the 222nd Session of the Rajya Sabha comes to a close today. Being a shorter Budget Session, it was principally devoted to the financial business of the Government.
The Session began with the Address of the respected Rashtrapatiji to both the Houses of Parliament assembled together on 21.02.2011 and was followed by a discussion on the Motion of Thanks on the President’s Address. In addition to the discussion on the Railway and General Budgets as well as return of the concerned Appropriation Bills and the Finance Bill, an important legislative business transacted during this Session related to the passing of The Orissa (Alteration of Name) Bill, 2010 along with The Constitution (One Hundred and Thirteenth Amendment) Bill, 2010.
Several statements were made or laid by the Prime Minister and other Members of the Council of Ministers during this Session. A large number of Special Mentions on matters of public importance as well as matters raised with the permission of the Chair were made during this Session. Two Calling Attention notices on ‘unprecedented and untimely heavy rains in Andhra Pradesh’ and ‘utilization of agricultural land for non-agricultural purposes’ were discussed. A Short-duration Discussion on the statement made by the Prime Minister on 18.03.2011 on newspaper report on ‘Cash-for-vote’ was also taken up. I am requesting the Secretary-General to make available the statistical information relating to this Session.
In order to optimize the output of the Question Hour and, at the same time, allow Members regular opportunity to mention matters of urgent concern with the permission of the Chair, it was decided to commence the working day at 11:00 a.m. with ‘Matters raised with the permission of the Chair’ and shift the Question Hour to the first hour.
after lunch. The Chair appreciates the cooperation in the matter of the hon. Members. Initial results have been encouraging. I have also received some feedback which will help fine-tune the new arrangement.
Six new Ministers were introduced to the House by the Prime Minister. Shrimati Hema Malini, elected in a by-election from the State of Karnataka, joined the House during the Session.
I take this opportunity to thank the Leader of the House, the Leader of the Opposition, the leaders of various parties and groups and the hon. Members for the cooperation extended by them for the overall smooth functioning of the House. I also thank the Deputy Chairman, the Members of the panel of Vice-Chairmen and the officers and staff of the Secretariat for their help and cooperation.
Now, before we adjourn the House sine die, we will have the National Song.
(The National Song, "Vande Mataram", was then played.)
MR. CHAIRMAN: The House stands adjourned sine die.
The House then adjourned sine die at twelve of the clock. |
A divide-and-conquer approach to the pairwise opposite class-nearest neighbor (POC-NN) algorithm
Thanapant Raicharoen *, Chidchanok Lursinsap
Advanced Virtual and Intelligent Computing Center (AVIC), Department of Mathematics, Faculty of Science, Chulalongkorn University, Bangkok 10330, Thailand
Received 20 June 2004; received in revised form 15 October 2004
Available online 17 February 2005
Communicated by R.P.W. Duin
Abstract
This paper presents a new method based on divide-and-conquer approach to the selection and replacement of a set of prototypes from the training set for the nearest neighbor rule. This method aims at reducing the computational time and the memory space as well as the sensitivity of the order and the noise of the training data. A reduced prototype set contains Pairwise Opposite Class-Nearest Neighbor (POC-NN) prototypes which are close to the decision boundary and used instead of the training patterns. POC-NN prototypes are obtained by recursively iterative separation and analysis of the training data into two regions until each region is correctly grouped and classified. The separability is determined by the POC-NN prototypes essential to define the locations of all separating hyperplanes. Our method is fast and order independent. The number of prototypes and the overfitting of the model can be reduced by the user. The experimental results signify the effectiveness of this technique and its performance in both accuracy and prototype rate as well as in training time to those obtained by classical nearest neighbor techniques.
© 2005 Elsevier B.V. All rights reserved.
Keywords: Pattern classification; Nearest neighbor rule; Prototype selection; Prototype replacement
1. Introduction
The Nearest Neighbor (NN) rule was originally proposed by Cover and Hart (Cover and Hart, 1967; Hart, 1966) in 1966 and has been shown to be very effective in many applications of pattern recognition. One reason for the use of this rule
is its conceptual simplicity, which is easy to implement. Moreover, under some continuity assumptions on the underlying distributions, the asymptotic error rate of this rule is at most twice Bayes’ probability of errors (Cover and Hart, 1967; Duda et al., 2001). However, it suffers from various drawbacks. Firstly, it requires a large memory space as the entire training data set has to be stored and each test pattern has to be compared with every training pattern. Secondly, it requires a large computational time for finding the neighbors. Thirdly, it is sensitive to the noisy and/or outlier patterns. To alleviate these drawbacks, two approaches, prototype selection and prototype replacement are introduced (Kuncheva and Bezdek, 1998). Both approaches aim at modifying an original training pattern (prototype) in order to reduce its size as well as to improve classification performance. One of the first and most popularly used techniques of prototype selection is the Condensing Nearest Neighbor (CNN) proposed by Hart (1968). The main goal of the condensing method is to obtain the reduced and consistent set of prototypes (Hart, 1968) to be used with the $1$-$NN$ rule without error in the training set or in with the $k$-$NN$ rule without significantly degrading its performance. The condensing method proceeds by repeatedly selecting the prototypes whenever they cannot be correctly classified by the currently selected set. The whole process is iterated until there are no changes in a complete pass through the initial training set. However, the method does not, in general, yield a minimal size of the consistent subset, and the final size as well as composition of the final condensed set may strongly depend upon the order of presentation data. Since the development of the CNN, other methods were proposed successively, such as the Reduced Nearest Neighbor (RNN) rule proposed by Gates (1972). An algorithm for a Selective Nearest Neighbor (SNN) decision rule was introduced by Ritter et al. (1975). Tomek presented the Two Modifications of the CNN by growing the condensed set using only patterns close to the decision boundary (Tomek, 1976). The way in which pairs of prototypes are selected makes the algorithm very good at preserving the original decision boundaries. Chidananda Gowda and Krishna introduced the concept of mutual nearest neighbor neighborhood for selecting patterns close to the decision boundaries (Chidananda Gowda and Krishna, 1979). The position of a prototype in the ordered list of neighbors of its nearest neighbor from an opposite class is used as a way to measure the closeness to boundaries. Several theoretical results on CNN have been obtained in (Devroye et al., 1996).
Another well known used technique of prototype selection is the editing method (Devijver and Kittler, 1982) proposed by Wilson (1972). The main goal of the editing method is to improve the performance by discarding outliers and possible overlapping among classes rather than prototype reduction. However, the drawbacks of the editing method are that it still leaves too many prototypes in the edited set, and the complexity of computing the edited subset is very high. Therefore, Sanchez proposed the $k$-Nearest Centroid Neighbors ($k$-NCN) in order to identify and eliminate mislabeled, noisy and atypical training patterns (Sanchez et al., 2003). Several editing experiments are carried out and comparative results are presented in (Ferri et al., 1999). The exploration and exploitation of the synergy among the NN editing and condensing methods in order to facilitate the use of NN techniques in real-world applications was studied in (Dasarathy et al., 2000).
For the prototype replacement approach, one of the first methods, proposed by Chang (1974), repeatedly merges the two nearest neighbors of the same class as long as this merger does not increase the error rate on the training set. One drawback of this method is that it may yield the prototypes that do not characterize the training set well in terms of generalization (Toussaint, 2002).
In order to obstruct the undesirable property of the order dependence of presentation data, several attempts, for example in (Devi and Murty, 2002; Alpaydin, 1997), were suggested to obtain selected prototypes that are less sensitive to this property. However, most of the improvements of NN methods cannot avoid overfitting for noisy and/or overlapping data, and do not consider any statistical properties of training data.
This work is focused on developing a new method for obtaining a set of selection prototypes as well as replacement prototypes. Unlike all of these above mentioned methods, our proposed method is based on a divide-and-conquer approach. That is the analogy to partition original training patterns into smaller regions by finding $POC-NN$ prototypes for the regions, and then combine $POC-NN$ prototypes for the regions into a set of selection prototypes. The rest of this paper is organized into five sections. Sections 2 and 3 present our proposed $POC-NN$ method for prototype selection and replacement, respectively, including its analysis. In Section 4, some experiments are described and the performance of our method is evaluated. Section 5 concludes the paper.
2. The methodology of the $POC-NN$ for prototype selection
The idea is to isolate a subset of the training set that suffices for nearest neighbor predictions, and throw away the remaining data. Intuitively, it seems important to keep the training points (patterns) that are close to the decision boundaries and on the correct side of those boundaries, while some points far from the boundaries should be discarded. The bottom line of this idea is to find a $POC-NN$ pattern for a given two-class training data set and then the remaining patterns can be discarded.
2.1. Finding $POC-NN$ pattern algorithm
Let $S$ be a training set of $n$ patterns composed of two subsets $S^{(1)}$ and $S^{(2)}$ whose sizes are $|S^{(1)}| = n^{(1)}$ and $|S^{(2)}| = n^{(2)}$. Both $S^{(1)}$ and $S^{(2)}$ are in different classes, namely classes 1 and 2, respectively, and $S^{(1)} \cap S^{(2)} = \emptyset$. The algorithm to find a $POC-NN$ pattern is given as follows.
**Function FINDING-POC-NN ($S$: Dataset)**
1. Let $S^{(1)}$ and $S^{(2)}$ be two training sets with classes 1 and 2, whose sizes are $S^{(1)}$ and $S^{(2)}$, respectively.
2. **If** $n^{(1)} \geq n^{(2)}$
**Then**
3. $x_m =$ mean of $S^{(1)}$.
4. **Else**
4.1. Let $x_{p1} \in S^{(1)}$ be the nearest pattern to $x_m$.
4.2. Let $x_{p2} \in S^{(2)}$ be the nearest pattern to $x_{p1}$.
**Endif**
5. **Return** $(x_{p1}, x_{p2})$ as a $POC-NN$ pattern.
Fig. 1(a) shows an example of how finding $POC-NN$ algorithm works. There are two classes, 1 and 2. Each pattern in class 1 is denoted by the symbol “+” and each pattern in class 2 is denoted by the symbol “*”. The mean of patterns ($x_m$) in class 1 is denoted by the symbol “◇”. The $POC-NN$ prototypes ($x_{p1}, x_{p2}$) are enclosed in circle symbols “○”. This $POC-NN$ prototype is performed during the training process and all considered patterns in all classes are considered as the training patterns. The identical $POC-NN$ patterns as prototypes are always obtained independent from the different reordering of the training patterns. Moreover, the $POC-NN$ patterns have the desirable property similar to the Support Vectors (Cortes and Vapnik, 1995; Vapnik, 1998) which induce the optimal separating hyperplane. This is because if all the training vectors are linearly independent, the two closest patterns of opposite classes are support vectors. This proof can be found in (Roobaert, 2000). However, the $POC-NN$ algorithm is not guaranteed to find these support vectors. Fig. 1(b) and (c) show the prototypes created by using the $CNN$ method (Hart, 1968) which does not guarantee the optimal solution. The condensed prototypes are enclosed in square symbols “□”. The obtained prototypes strongly depend on the order of presentation of the training patterns. In this case, it depends on the first pattern of each class. Fig. 1(d) also shows the prototypes by using the $CNN$ method which creates the problems of redundant prototypes and also slow convergence.
Once $POC-NN$ prototypes are found, a separating linear hyperplane is generated and orthogonally placed in the middle of the distance between these $POC-NN$ prototypes. This hyperplane acts as a decision boundary (similar to a Voronoi Diagram). The boundary of each region
is defined by the corresponding hyperplane generated by the $POC-NN$ prototype lying in each region. The following is the detail of prototype selection by $POC-NN$ algorithm.
### 2.2. Prototype selection by $POC-NN$ algorithm for two-class classification problem
In this section, we present our proposed algorithm for two-class classification problem and in the later subsection, we extend our algorithm for multi-class classification problem.
**Prototype selection by $POC-NN$ algorithm for two-class classification**
Let $S$ be a training set of $n$ patterns composing of two classes, and $POC-NN-SET$ initially be an empty $POC-NN$ prototypes set.
**Function SELECTING-POC-NN ($S$: Dataset)**
1. Find a $POC-NN$ prototype in $S$ by using $(x_{p1}, x_{p2}) = FINDING-POC-NN (S)$.
2. Determine the center point $c = \frac{x_{p1} + x_{p2}}{2}$.
3. Create a separating hyperplane $H: \{x | w \cdot x - b = 0\}$,
where $w = \frac{x_{p1} - x_{p2}}{\|x_{p1} - x_{p2}\|}$ and $b = w \cdot c$.
4. Save $(x_{p1}, x_{p2})$ and coresponding $H$ into the $POC-NN-SET$.
5. Divide all patterns of $S$ into two regions, namely $R1$ and $R2$, where
$R1 = \{x_i \in S | w \cdot x_i - b \geq 0\}$, and
$R2 = \{x_i \in S | w \cdot x_i - b < 0\}, \forall i, i = 1, \ldots, n$.
6. Find any misclassification in both regions.
7. If any misclassification exists in region $R1$
Then
8. Consider all data in this region $R1$ as a new data set
Call $SELECTING-POC-NN (R1)$.
Endif
9. If any misclassification exists in region $R2$
Then
10. Consider all data in this region $R2$ as a new data set
Call $SELECTING-POC-NN (R2)$.
Endif
11. If no more misclassification exists
Then
12. Return $POC-NN-SET$ as a set of selected prototypes.
Stop.
Endif
Fig. 2 shows an example of how the algorithm works for non-linearly separable problems. The initial separating line ($H_1$) created by a $POC-NN$ ($x_{p1}, x_{p2}$) is shown in Fig. 2(a). However, this line still creates misclassified training patterns lying on the right side of this line. All the training patterns lying on this side are considered as a new data set. The second line ($H_2$) as shown in Fig. 2(b) is introduced to resolve the previously misclassified patterns. After performing the algorithm, a set of selecting prototypes corresponding to the $POC-NN-SET$ consisting of two pairs of $\{(x_{p1}, x_{p2})\}$ is obtained. The patterns from $S^{(1)}$ are separated into two regions (regions 1 and 2). Each region has a prototype ($\{x_{p1}\}$) representative class labeled 1. The patterns from $S^{(2)}$ are not separated, however, they have two prototypes ($\{x_{p2}\}$) representative class labeled 2.
2.3. Reducing complexity and sensitivity to noise
In order to reduce the complexity and sensitivity to noise as well as to avoid overfitting for overlapping data, we relax the separating condition. A separating hyperplane can be considered as a slab of width alpha ($\alpha$), called acceptance interval. All patterns lying within the acceptance interval of the slab are considered as correctly classified patterns, and also considered as noisy and/or outlier patterns which can be ignored or discarded. Our acceptance interval ($\alpha$), is defined as follows.
Suppose $S$ be a training data set of $n$ points with two classes and $x_i \in S$.
**Definition 1.** Let $\{x | w \cdot x - b = 0\}$ be a hyperplane ($H$). $\alpha$ is an acceptance interval if $|w \cdot x_i - b| < \alpha$, and $\alpha > 0$ and all $x_i$ are correctly classified.
The acceptance interval is proportional to the distance $d$ between two $POC-NN$ prototypes, and is defined priori to the training process. In the other word, $\alpha$ is defined as $\alpha_r \times d$ for $0 < \alpha_r < 1$. This $\alpha_r$ is called $\alpha$-ratio.
Fig. 2. This example shows how the proposed algorithms works: (a) initial separating line ($H_1$); (b) second separating line ($H_2$).
These $\alpha$ values for each hyperplane are not necessary the same values, but are dependent on $\alpha$-ratio and the distance between corresponding $POC-NN$ prototypes.
Fig. 3 shows an example of an acceptance interval defined by giving the $\alpha$-ratio 1:5 or 0.2 (a), and 1:10 or 0.1 (b) and (c). In Fig. 3(a), our algorithm will stop after one iteration and generate two $POC-NN$ prototypes. Each $POC-NN$ is a prototype representative for each class, even though there are three misclassified patterns lying within the acceptance interval. They are considered as noisy and/or outlier patterns which are ignored. In Fig. 3(b), there are only two instead of three misclassified patterns lying within the acceptance interval and considered as noisy and/or outlier patterns. One misclassified pattern is still there, therefore, the algorithm will continue and generate the other two $POC-NN$ prototypes shown in Fig. 3(c).
2.4. Analysis of the selecting $POC-NN$ algorithm
There are some interesting consequences in selecting $POC-NN$ prototype worth discussing. Given a training set $S$ of $n$ patterns ($n > 1$) in $d$ dimensions composing of two subsets $S^{(1)}$ and $S^{(2)}$, whose sizes are $n^{(1)}$ and $n^{(2)}$. Both $S^{(1)}$ and $S^{(2)}$ are in different classes, namely classes 1 and 2, respectively, and there never exist two
patterns with different class label on the position \((S^{(1)} \cap S^{(2)}) = \emptyset\).
**Property 1.** The selecting \(POC-NN\) algorithm converges after \(m\) iterations, where \(m\) is the number of \(POC-NN\) patterns having the values between 1 and \(n - 1\).
**Proof.** We start our proof with \(n^{(1)}\) and \(n^{(2)} = 1\), so these two patterns are \(POC-NN\) pattern to each other and its hyperplane can be constructed to make them linearly separable. If we add a new pattern into this space, we always have two cases:
Case 1: A new pattern lies on the correct side, so all data are still linearly separable and correctly classified. The algorithm stops after one iteration.
Case 2: A new pattern lies on the wrong side. Here, we need to consider only the data lying on the wrong side. A new pattern will become a new \(POC-NN\) pattern and a new hyperplane that makes these data linearly separable, is introduced. The algorithm will stop after two iterations.
If we gradually add a new pattern until there are \(n\) patterns of training set \(S\), we still always have two cases as before. The algorithm will stop after \(n - 1\) iterations.
Assume that the algorithm does not converge. This would only be possible if no hyperplane could be constructed. But note that the hyperplane cannot be constructed if the nearest patterns of the two classes in data group cannot be found. It would only be possible if the data group has only the same class data and that is also the stop condition of our algorithm. This is a contradiction and hence non-convergence does not occur. \(\square\)
The consequence of this property yields that all training patterns are separated into the regions of correctly classified classes after convergence. Therefore, a set of patterns in each region is a *consistent set*. However, the whole set of the selected \(POC-NN\) prototypes may not be a *consistent set*. This proof can be shown by giving a counterexample that the obtained set of prototypes generated by our algorithm does not give 100% accuracy on the training set by using 1-\(NN\) rule. Similar to the Tomek’s algorithm, the consistency is not guaranteed. A counterexample has been shown in (Toussaint, 1994). However, there is no theoretical evidence on how the consistency of the condensed set relates to the generalization abilities. It may have an arbitrarily poor performance when applied to unseen patterns as shown by some experimental examples in Section 4.
**Property 2.** The time complexity of the selecting \(POC-NN\) algorithm is \(O(dn^2m)\).
**Proof.** The upper bound of the time complexity is derived step by step in the following analysis. In step 1 of the selecting \(POC-NN\) algorithm, the calculation for finding a \(POC-NN\) prototype takes in worst case \(O(dn^2)\), and in steps 5 and 6, the calculation takes \(O(dn)\). The time complexity is bounded by \(O(dn^2 + dn)\) after one iteration. If the algorithm does not converge, step 1 of the algorithm will continue with at most in worst case \((n - 1)\) patterns. The time complexity is bounded by \(O(d(n - 1)^2 + d(n - 1))\) after two iterations. From the property 1, our algorithm converges after \(m\) iterations, therefore, the total time complexity after \(m\) iterations is bounded by \(O(dn^2 + dn) + O(d(n - 1)^2 + d(n - 1)) + \ldots + O(d1^2 + d1)\) which is equal to \(O(dn^2m)\). \(\square\)
Comparison with *Condensing and Editing NN* algorithm, the complexity of computing the condensed subset and edited subset are \(O(dn^3)\) and \(O(d^3n^{[d/2]}\ln n)\), respectively (Toussaint, 2002; Duda et al., 2001). Both time complexities are higher than the selecting \(POC-NN's\). In addition, the comparison with the other two modified condensing methods which attempt to keep only prototypes close to the decision boundary, Tomek’s algorithm (Tomek, 1976) and Gowda–Krishna’s algorithm (Chidananda Gowda and Krishna, 1979), the computational complexity of both algorithms is considerable higher than the original \(CNN\) algorithm. In case of Tomek’s algorithm, only the step for finding a special subset, called *Gabriel Neighbors* (Toussaint, 1994), already takes \(O(n^3)\) time. For a large number \(n\) such as that in OCR applications shown by USPS data (USPS, 1994) experimental example in Section 4, it is not feasible. Gowda–Krishna’s algorithm takes more time than Tomek’s algorithm in some cases as shown in (Chidananda Gowda and Krishna, 1979), since it requires a preprocess for finding
the Mutual Neighborhood Value (MNV) of each pattern to order the patterns according to MNV, and a post-process for reducing the number of condensed set of prototypes.
**Property 3.** The selecting POC-NN algorithm is order independent.
**Proof.** The training set can have $n!$ permutations order of how to present the data. In steps 3 and 4 of the finding a POC-NN pattern algorithm, the centroid or mean of a set of points (patterns) is calculated. The addition remains invariant to the order of the data elements. Thus, the mean operation is order independent. Finding the nearest point to the centroid is also order independent, since the centroid and the distances between the centroid and all points from different classes are fixed values. The shortest distance is found and the corresponding point is selected as the representative prototype. If there is a tie between two or more points for the closest distance from the centroid, the point with the smallest value of the mean of its feature is selected. If this is the same, the point with the smaller value of feature 1 is selected. If this is still the same, the point with the smaller value of feature 2 is selected etc. So the distance measure is order independent and the minimum of these distances remains the same no matter what the order of the presentation data is. The other steps in the algorithm are obviously order independent, thus the selecting POC-NN algorithm is order independent. $\square$
### 2.5. Algorithm for multi-class classification problem
In this section, we extend our proposed algorithm for multi-class classification problem by using a combination of many two-class classifiers into a multi-class classifier. In contrast to the standard approach to the $k$-class problem by a ‘one-against-rest’ $(1 - v - r)$ like scheme, we use the alternative approach, so called ‘one-against-one’ $(1 - v - 1)$ scheme using pairwise classification with $\frac{k(k-1)}{2}$ binary classifiers.
We classify all $k$-class training data into pairwise two-class training sets and perform our selecting POC-NN algorithm on each two-class training set. It is possible that the same pattern becomes a POC-NN prototype more than once, when it has been selected by another combination of pairwise two-class training set. However, it is still the same representative prototype for its class label. After performing our algorithm for each two-class pair, we obtain POC-NN prototypes representative for each class. The time complexity for multi-class classification will be increased to $O\left(\frac{k(k-1)}{2} (dn^2m)\right)$, where $n$ is the number of patterns with two classes and $m$ is the number of POC-NN prototypes.
### 3. The methodology of POC-NN for prototype replacement
After performing our prototype selection by POC-NN algorithm on a training set, this training set is separated into many regions of correctly classified classes. Each region has representative POC-NN prototypes being elements of a reference set. In this section, the original training data set are replaced by a number of prototypes that do not necessarily coincide with any pattern in training set. Our POC-NN method for prototype replacement is applied by replacing all POC-NN prototypes in each region with the mean of patterns in its region. Since, in the statistical principle, the mean of a data set is the best understood measure of central tendency for a quantitative data set (McClave et al., 2001), therefore the mean can be a good representative pattern for all patterns in its region. In consequence, the number of replacement prototypes in general will be decreased when compared with the number of selection prototypes.
#### 3.1. Prototype replacement by POC-NN algorithm for two-class classification problem
We propose our prototype replacement by POC-NN algorithm by modification of the prototype selection by POC-NN algorithm for Two-Class Classification as follows.
Let MOR-NN-SET initially be an empty set of mean of region prototypes.
**Function** \( REPLACING-POC-NN \) (\( S \): Dataset)
1. Find a \( POC-NN \) prototype in \( S \) by using \((x_{p1}, x_{p2}) = FINDING-POC-NN(S)\).
2. Determine the center point \( c = \frac{x_{p1} + x_{p2}}{2} \).
3. Create a separating hyperplane \( H \):
\[
\{x | w \cdot x - b = 0\},
\]
where \( w = \frac{x_{p1} - x_{p2}}{\|x_{p1} - x_{p2}\|} \) and \( b = w \cdot c \).
4. Save \((x_{p1}, x_{p2})\) and corresponding \( H \) into the \( POC-NN-SET \).
5. Divide all patterns of \( S \) into two regions, namely \( R1 \) and \( R2 \), where
\[
R1 = \{x_i \in S | w \cdot x_i - b \geq 0\},
\]
and
\[
R2 = \{x_i \in S | w \cdot x_i - b < 0\}, \forall i, i = 1, \ldots, n.
\]
6. Find any misclassification in both regions.
7. If any misclassification exists in region \( R1 \)
Then
8. Consider all data in this region \( R1 \) as a new data set
call \( REPLACING-POC-NN(R1) \).
Else
9. Save \( x_{mor} \) being the mean of \( R1 \) into the \( MOR-NN-SET \).
Endif
10. If any misclassification exists in region \( R2 \)
Then
11. Consider all data in this region \( R2 \) as a new data set
call \( REPLACING-POC-NN(R2) \).
Else
12. Save \( x_{mor} \) being the mean of \( R2 \) into the \( MOR-NN-SET \).
Endif
13. If no more misclassification exists
Then
14. Save \( x_{mor} \) being the mean of \( R1 \) into the \( MOR-NN-SET \), and
Save \( x_{mor} \) being the mean of \( R2 \) into the \( MOR-NN-SET \).
15. Return \( MOR-NN-SET \) as a set of replacement prototypes.
Stop.
Endif
Fig. 4 shows an example of how the prototype replacement algorithm works on the same data set shown in Fig. 2. After the algorithm converges, the training set is separated into three regions of correctly classified classes. Therefore, we obtain three \( MOR-NN \) prototypes \((x_{mor})\) used to replace the previous four \( POC-NN \) prototypes shown in Fig. 2(b). Normally, the number of \( MOR-NN \) prototypes is less than the number of \( POC-NN \) prototypes.
### 3.2. Analysis of the replacing POC-NN algorithm
**Property 4.** The replacing POC-NN algorithm is order independent.
**Proof.** As shown in Property 3, the selecting \( POC-NN \) algorithm is order independent. The replacing \( POC-NN \) algorithm is almost identical except that it additionally calculates the mean of each region. Since the mean operation is order independent, our replacing POC-NN algorithm is also order independent. \( \square \)
### 4. Experimental results
Our algorithms are tested and evaluated on a number of standard classification datasets of benchmarks, both artificial and real. These data sets are taken from UCI Repository of machine learning databases (Murphy and Aha, 1994)
Table 1
Properties of the data sets used
| Datasets | No. of features | No. of classes | No. of trg. patterns | No. of test patterns |
|----------|-----------------|----------------|----------------------|----------------------|
| 1. USPS | 256 | 10 | 7291 | 2007 |
| 2. DNA | 180 | 3 | 2000 | 1186 |
| 3. Sonar | 60 | 2 | 104 | 104 |
| 4. Vowel | 10 | 11 | 528 | 462 |
| 5. Wine | 9 | 7 | 118 | 60 |
| 6. Liver | 6 | 2 | 230 | 115 |
| 7. Thyroid | 5 | 3 | 143 | 72 |
| 8. Iris | 4 | 3 | 100 | 50 |
| 9. Spiral | 2 | 2 | 129 | 65 |
except for the first and second ones which are the USPS data (USPS, 1994) and DNA from Statlog (Michie et al., 1994). The properties of the datasets are given in Table 1. In the first four data sets, the training and test sets are separated. In the others, we separated the original data sets into training and test sets (2:1 ratio) by using the following criteria. The first pattern belongs to the test set, and the second and third patterns belong to the training set. The fourth pattern again belongs to the test set, and the fifth and sixth patterns belong to the training set, and so on. We use Euclidean distance ($L_2$-norm) to measure the dissimilarity and $1 - NN$ rule to test the results in all experiments.
The results obtained on these datasets are reported in Table 2. The prototype rate (PR%) computed from the percentage of prototypes to all training patterns, and the accuracy rate (AR%) computed from the percentage of correctly predicted test patterns to all test patterns obtained using the selecting (S-) and replacing (R-) POC-NN, the Condensed Nearest Neighbor (CNN), the Tomek’s algorithm (Tomek), the Gowda–Krishna’s algorithm (GKA), and the accuracy rate (AR%) obtained using all prototypes (NN) are given for each data set. The number of prototypes corresponding to the PR% obtained using the R-POC-NN algorithm normally is less than the number of prototypes obtained from using the S-POC-NN algorithm. N/A (for not applicable/available) in the Table 2 signifies that the learning scheme did not finish training. If learning could not be completed within the time period of two weeks then it was terminated and marked N/A.
Our results are mostly better than the results obtained by the CNN, especially, the accuracy rates using POC-NN are better than CNN in all cases, and better than Tomek and GKA in all cases except Liver and Thyroid. However, the prototype rates using POC-NN are better than the other algorithms in some cases. In the cases of wine, thyroid and iris data, the selecting POC-NN shows better results in both accuracy and prototype rates than the results from the CNN. In the cases of sonar, thyroid and spiral data, the replacing POC-NN also shows better results in both accuracy and prototype rates than results from the CNN. Moreover, in case of vowel and thyroid
Table 2
The comparison results of S-POC-NN, R-POC-NN, CNN, Tomek, GKD, and NN
| Datasets | S-POC-NN | R-POC-NN | CNN | Tomek | GKA | NN |
|----------|----------|----------|-----|-------|-----|----|
| | PR% | AR% | PR% | AR% | PR% | AR%|
| 1. USPS | 28.20 | 93.42** | 29.79 | 93.57*** | 11.84 | 91.58 | N/A | N/A | N/A | 100.00 | 94.37 |
| 2. DNA | 60.70 | 74.37 | 58.65 | 87.77*** | 44.30 | 72.30 | N/A | N/A | N/A | 100.00 | 76.39 |
| 3. Sonar | 60.58 | 80.77 | 51.92 | 90.38 | 62.50 | 90.38 | 56.73 | 88.46 | 50.00 | 87.50 | 100.00 | 91.34 |
| 4. Vowel | 50.19 | 54.11** | 56.82 | 57.14*** | 19.69 | 47.19 | 20.83 | 51.52 | 18.56 | 55.84 | 100.00 | 56.28 |
| 5. Wine | 43.22 | 71.67 | 38.13 | 58.33 | 44.07 | 66.67 | 42.37 | 68.33 | 38.98 | 68.33 | 100.00 | 68.33 |
| 6. Liver | 58.69 | 54.78 | 43.47 | 53.04 | 56.52 | 53.91 | 55.65 | 58.26 | 50.00 | 57.39 | 100.00 | 59.91 |
| 7. Thyroid | 20.27 | 88.89 | 15.38 | 91.67 | 20.97 | 88.89 | 16.08 | 95.83 | 15.38 | 90.28 | 100.00 | 91.67 |
| 8. Iris | 17.00 | 98.00 | 16.00 | 96.00 | 20.00 | 98.00 | 18.00 | 98.00 | 16.00 | 98.00 | 100.00 | 98.00 |
| 9. Spiral | 56.59 | 95.38* | 35.66 | 92.31 | 37.98 | 89.23 | 37.98 | 93.85 | 34.11 | 89.23 | 100.00 | 96.92 |
Best accuracy rates (AR%) among these algorithms except NN are bold-faced. Tomek and GKA are not applicable (N/A) on USPS and DNA, because they require unacceptable time to run on these data. The symbols “***”, “**” and “*” indicate the statistically significant difference to CNN at the level 99%, 95% and 90%, respectively.
Table 3
The results comparisons of $POC-NN$ and the other algorithms by conducting the three-fold ($K = 3$) and five-fold ($K = 5$) cross-validation
| K-Fold | Datasets | $S-POC-NN$ | $R-POC-NN$ | CNN | Tomek | GKA |
|--------|----------|------------|------------|-----|-------|-----|
| | | PR% | AR% | PR% | AR% | PR% | AR% | PR% | AR% | PR% | AR% |
| $K = 3$| Wine | 43.82 | 73.05 | 37.67 | 58.43 | 41.86 | 71.37 | 40.45 | 73.05 | 38.20 | 72.50 |
| | Liver | 57.35 | 60.64** | 44.83 | 57.76 | 59.81 | 58.07 | 57.78 | 57.19 | 50.06 | 57.46 |
| | Thyroid | 18.15 | 90.25 | 14.42 | 88.38 | 19.53 | 90.24 | 16.05 | 93.49 | 14.42 | 90.23 |
| | Iris | 20.67 | 96.00 | 17.67 | 98.67 | 16.00 | 94.00 | 14.67 | 98.00 | 13.00 | 92.00 |
| | Spiral | 58.50 | 96.39 | 37.88 | 91.75 | 38.40 | 92.79 | 38.14 | 93.82 | 35.56 | 91.78 |
| $K = 5$| Wine | 42.42 | 71.86 | 37.08 | 64.60 | 40.45 | 73.59 | 37.93 | 73.57 | 36.52 | 73.57 |
| | Liver | 58.91 | 60.29** | 46.74 | 55.65 | 58.70 | 59.42 | 58.26 | 59.13 | 53.12 | 60.00 |
| | Thyroid | 16.05 | 89.77 | 13.02 | 91.63* | 16.86 | 89.77 | 16.05 | 94.42 | 14.19 | 93.49 |
| | Iris | 8.42 | 99.00* | 7.17 | 98.67 | 7.50 | 96.33 | 13.83 | 98.67 | 12.00 | 94.00 |
| | Spiral | 60.17 | 97.41 | 36.46 | 90.74 | 33.89 | 96.90 | 34.79 | 97.41 | 34.67 | 94.35 |
Best accuracy rates (AR%) are bold-faced.
Table 4
The results comparisons of $S-POC-NN$ by using without $z_r$-ratio ($z_r = 0$) and with different-ratio ($z_r < > 0$), and using $CNN$
| Datasets | $S-POC-NN$ ($z_r = 0$) | $S-POC-NN$ ($z_r < > 0$) | $CNN$ |
|----------|------------------------|--------------------------|-------|
| | Prototypes | AR% | $z_r$ ratio | Prototypes | AR% | Prototypes | AR% |
| | # | % | | # | % | | % |
| 1. USPS | 2056 | 28.20 | 93.42** | 0.25 | 631 | 8.66 | 89.16 | 863 | 11.84 | 91.58 |
| 2. DNA | 1214 | 60.70 | 74.37* | 0.05 | 793 | 39.65 | 73.19 | 886 | 44.30 | 72.30 |
| 3. Sonar | 63 | 60.58 | 80.77 | 0.05 | 58 | 55.77 | 80.77 | 65 | 62.50 | 90.38 |
| 4. Vowel | 265 | 50.19 | 54.11** | 0.8 | 103 | 19.51 | 50.22 | 104 | 19.69 | 47.19 |
| 5. Wine | 51 | 43.22 | 71.67 | 0.3 | 49 | 41.53 | 71.67 | 52 | 44.07 | 66.67 |
| 6. Liver | 135 | 58.69 | 54.78 | 0.5 | 95 | 41.30 | 54.78 | 130 | 56.52 | 53.91 |
| 7. Thyroid | 29 | 20.27 | 88.89 | 0.25 | 22 | 15.38 | 88.89 | 30 | 20.97 | 88.89 |
| 8. Iris | 17 | 17.00 | 98.00 | 0.5 | 14 | 14.00 | 98.00 | 20 | 20.00 | 98.00 |
| 9. Spiral | 73 | 56.59 | 95.38* | 0.1 | 72 | 55.81 | 95.38* | 49 | 37.98 | 89.23 |
Best accuracy and prototype rates between $S-POC-NN$ and $CNN$ are bold-faced. The symbol “**” and “*” indicate the statistically significant difference to $CNN$ at the level 95% and 90%, respectively.
Table 5
The time comparisons of $POC-NN$, and the others in second
| Datasets | Training time is second |
|----------|-------------------------|
| | $S-POC-NN$ | $R-POC-NN$ | $CNN$ | Tomek | GKA |
| 1. USPS | 297,650 | 298,700 | 560,370 | N/A | N/A |
| 2. DNA | 109,277 | 11,048 | 74,474 | N/A | N/A |
| 3. Sonar | 8.94 | 9.03 | 90.20 | 925.32 | 882.68 |
| 4. Vowel | 55.18 | 55.78 | 81.29 | 14,045.00 | 2873.77 |
| 5. Wine | 3.29 | 3.32 | 20.15 | 101.51 | 189.63 |
| 6. Liver | 4.49 | 4.57 | 38.07 | 620.41 | 1021.83 |
| 7. Thyroid | 1.38 | 1.42 | 6.65 | 56.54 | 24.89 |
| 8. Iris | 0.80 | 0.81 | 2.56 | 30.42 | 8.87 |
| 9. Spiral | 0.66 | 0.69 | 8.77 | 41.81 | 33.12 |
Tomek and GKA algorithm are not applicable (N/A) of USPS and DNA, since they require unacceptable time (more than two weeks) to run on these data set.
data, the replacing $POC-NN$ even gives better accuracy rate than $NN$ by using all the training set as prototypes. So the higher accuracy rate of the training data cannot guarantee to obtain a better accuracy rate of test (untrained) data. The consistent set is not necessary related to the generalization abilities.
From these results, we formulated the hypothesis that using the $POC-NN$ algorithm is more effective than using $CNN$ algorithm. To determine if this hypothesis was statistically significant, we analyzed the results using a test of significance involving differences of proportions (Hayter, 2002). In the Table 2, this test shows that the level of significance of the hypothesis is 0.01, 0.05 and 0.1 indicated by symbols “***”, “**” and “*”, which provide a 99%, 95% and 90%, respectively, confidence level according to one-tailed proportions of the normal curve that it is correct.
To increase statistical significance of the results on the data set whose training and test sets are not separated, we conduct the $K$-fold cross-validation technique which probably is the simplest and most widely used method for estimating prediction error.
Fig. 5. Prototype rate (PR%) and accuracy rare (AR%) as a function of $\alpha$-ratio: (a) USPS, (b) DNA, (c) sonar, (d) vowel, (e) wine, (f) liver, (g) thyroid, (h) iris and (i) spiral.
(Hastie et al., 2001), and report the average cross-validation estimate of prototype and accuracy rate as shown in Table 3. Comparison with the results in Table 2, for the three-fold cross-validation \((K = 3)\), the \(S\)-POC-NN still shows better results in accuracy rate than results from the CNN in all cases, and shows the best results in accuracy rate in three cases. For the five-fold cross-validation \((K = 5)\), the \(S\)-POC-NN still shows better results in accuracy rate than the others in all cases except the Wine. The symbols “**” and “*” also indicate 95% and 90%, respectively, confidence interval for estimating the difference between accuracies of \(S\)-POC-NN, \(R\)-POC-NN and CNN using a one-tailed paired \(t\)-test (Mitchell, 1997).
In order to reduce the number of prototypes, the concept of acceptance interval \((z)\) is considered. Table 4 shows the results performed on the same datasets by using the \(S\)-POC-NN with different \(z\)-ratio and CNN. Most accuracy rates remain the same by reducing the number of prototypes. In many cases, the \(S\)-POC-NN with \(z\)-ratio gives better both accuracy and prototype rates than the CNN. In the case of DNA, sonar and vowel, there is a significant reduction in the number of prototypes used. So the reduced set of prototypes is very useful. By using the concept of acceptance interval \((z)\), we can regulate the value of \(z\) in order to control the complexity and avoid the overfitting of our model. Fig. 5 shows that the number of POC-NN prototypes and also the prototype rate that depend on choosing a priori value of an \(z\)-ratio. The \(z\)-ratio can also be employed to define the prototype and accuracy rates. In most cases, the accuracy and prototype rates decrease when the \(z\)-ratio increases. However, in cases of Sonar and Liver, we obtain the better accuracy rate with lesser prototype rate, even though the \(z\)-ratio is increased.
We also compare the training and testing time required for the results shown in Table 2. The less require computational time the more efficiency for an algorithm, even though it is carried out offline. The time comparisons are summarized in Table 5. All experiments were done on the same Pentium III-1GHz computer with 256 MB RAM, and all algorithms were implemented in MatLab version 6.5. We implemented these five different training algorithms in different training programs used to obtain the prototype set in each data set. The selecting POC-NN algorithm has the best training time for all datasets. In case of sonar, the POC-NN’s training time is approximately 10 times faster than the CNN’s training time, and approximately 100 times faster than the Tomek’s and the GKA’s training time. The Tomek and the GKA algorithms require very high the computational time and are not applicable for some real world problems, such as USPS and DNA. The time and space complexity of our POC-NN method are very competitive.
5. Conclusion
A new POC-NN method based on divide-and-conquer approach to the selection and replacement of a set of prototypes has been proposed. For a finite number of training data, our algorithm always converges with all patterns separated into the regions of correctly classified classes. The time complexity required to execute our algorithm is in the order of \(O(dn^2m)\). Our method is fast as well as simple, and moreover, it can overcome the undesirable property of the order dependence and the sensitivity of noisy data. The prototype and accuracy rates can also be regulated by a user-defined parameter, called an acceptance interval ratio \((z_r)\). The prototype and accuracy rates obtained have been compared with CNN, Tomek, GKA, and NN with all the training patterns. For most of the results, our method showed better performances in both prototype and accuracy rates than the results from the others. For all the above mentioned results, our method showed the best training time. The relationship between the \(z_r\) and the average distance among data as well as its variance are not studied here. The value of \(z_r\) should be locally adaptive, according to the distribution nature of the data. This issue is essential and worth for further investigation.
References
Alpaydin, E., 1997. Voting over multiple condensed nearest neighbors. Artificial Intell. Rev. 11, 115–132.
Chang, C.L., 1974. Finding prototypes for nearest neighbor classifiers. IEEE Trans. Computer 23 (11), 1179–1184.
Chidananda Gowda, K., Krishna, G., 1979. The condensed nearest neighbor rule using the concept of mutual nearest neighborhood. IEEE Trans. Inform. Theory 25 (4), 488–490.
Cortes, C., Vapnik, V.N., 1995. Support Vector Networks. Mach. Learn. 20 (3), 273–297.
Cover, T.M., Hart, P.E., 1967. Nearest neighbor pattern classification. IEEE Trans. Inform. Theory 13, 21–27.
Dasarathy, B.V., Sanchez, J.S., Townsend, S., 2000. Nearest neighbor editing and condensing tools-synergy exploitation. Pattern Anal. Appl. 3, 19–30.
Devi, V.S., Murty, M.N., 2002. An incremental prototype set building technique. Pattern Recognit. 35, 505–513.
Devijver, P.A., Kittler, J., 1982. Pattern Recognition: A Statistical Approach. Prentice-Hall, Englewood Cliffs, NJ.
Devroye, L., Györfi, L., Lugosi, G., 1996. A Probabilistic Theory of Pattern Recognition. Springer-Verlag, Inc, New York.
Duda, R.O., Hart, P.E., Stock, D.G., 2001. Pattern Classification. John Wiley and Sons, Inc.
Ferri, F.J., Albert, J.V., Vidal, E., 1999. Considerations about sample-size sensitivity of a family of edited nearest neighbor rules. IEEE Trans. Systems Man Cybernet. B 29 (5), 667–672.
Gates, G.W., 1972. The reduced nearest neighbor rule. IEEE Trans. Inform. Theory 18, 431–433.
Hart, P.E., 1966. An asymptotic analysis of the nearest-neighbor decision rule. Stanford Electron. Lab., Stanford, Calif., Tech. Rep., SEL-66-016, 1828-2.
Hart, P.E., 1968. The condensed nearest neighbor rule. IEEE Trans. Inform. Theory 14 (3), 515–516.
Hastie, T., Tibshirani, R., Friedman, J., 2001. The Elements of Statistical Learning. Springer Series in Statistics, Springer-Verlag.
Hayter, A.J., 2002. Probability and Statistics for Engineers and Scientists. Duxbury.
Kuncheva, L.I., Bezdek, J.C., 1998. Nearest prototype classification: Clustering, genetic algorithm, or random search? IEEE Trans. Systems Man Cybernet. C 28 (1), 160–164.
McClave, J.T., Benson, P.G., Sincich, T., 2001. Statistics for Business and Economics. Prentice Hall.
Michie, D., Spiegelhalter, D.J., Taylor, C.C., 1994. Machine learning, neural and statistical classification. Available from <ftp://ftp.ncc.up.pt/pub/statlog>.
Mitchell, T., 1997. Machine Learning. McGraw Hill.
Murphy, P.M., Aha, D.W., 1994. UCI Repository of machine learning databases [http://www.ics.uci.edu/mlearn/MLRepository.html]. Department of Information and Computer Science, University of California, Irvine, CA.
Ritter, G.L., Woodruff, H.B., Lowry, S.R., Isenhour, T.L., 1975. An algorithm for a selective nearest neighbor rule. IEEE Trans. Inform. Theory 23, 1179–1184.
Roobaert, D., 2000. DirectSVM: A simple support vector machine perceptron. In: Proceedings of IEEE International Workshop on Neural Networks for Signal Processing, Sydney, Australia.
Sanchez, J.S., Barandela, R., Marques, A.I., Alejo, R., Badenas, J., 2003. Analysis of new techniques to obtain quality training sets. Pattern Recognit. Lett. 24, 1015–1022.
Tomek, I., 1976. Two modifications of CNN. IEEE Trans. Systems Man Cybernet. Cybernet. 2, 769–772.
Toussaint, G.T., 1994. A counterexample to Tomek’s consistency theorem for a condensed nearest neighbor. Pattern Recognit. Lett. 15, 797–801.
Toussaint, G., 2002. Proximity Graphs for Nearest Neighbor Decision Rules: Recent Progress. In: Proceedings of 34th Symposium on Computing and Statistics, Montreal, Canada.
USPS, 1994. Machine learning, neural and statistical classification. Available from <ftp://ftp.kyb.tuebingen.mpg.de/pub/bs/data/>.
Vapnik, V.N., 1998. Statistical Learning Theory. Wiley, New York.
Wilson, D.L., 1972. Asymptotic properties of nearest neighbor rules using edited data. IEEE Trans. Systems Man Cybernet. Cybernet. 2, 408–420. |
This select bibliography draws on printed material in class B464 (Public Health) and sources from our manuscript collection. Additional material can be found in classes B468 (Hospitals and Hospices), D126 (Benevolent Societies) and D244 (Local Government). These sources are intended as an aid to researching the occurrence of outbreaks of infectious diseases on the Island. Cholera appeared in 1832/3 and again in 1866 with additional scares about its return; smallpox, typhus fever, dysentery and tuberculosis were other illnesses to be combatted. The reading material can help in the study of how the Island sought to prevent the spread of disease and why public health reforms emerged in the nineteenth century. Increased government regulation and improvements to water supplies, public sanitation, housing, medical provision, preventative vaccinations, diet and general public knowledge can be documented. The effectiveness of these reforms is difficult to trace accurately given that mortality statistics reported by the Registrar-General are available from 1880 onward only.
Additional library sources can be helpful in setting the scene to nineteenth century developments. For instance, the Island had been subjected to an 1759 Order in Council ordering a quarantine of forty days on all vessels from the Mediterranean to prevent the spread of the plague (see Atholl Papers 60-18). Amongst the listed sources an 1883 tussle between two men, Haviland and Talbot, brings some colour - they argue about the relative merits of the Isle of Man as a potential health resort. Also of note is the emergence of a nursing heroine in the person of Nellie Brennan (1792-1859) whose selfless efforts during the 1832/3 cholera outbreak have been lavishly praised in print.
**PRINTED SOURCES**
**BYE-LAW REGULATIONS**
There is a series of such bye-laws at Library reference D244. They are arranged as follows:
| Reference | Location |
|-----------|----------------|
| D244/1/2 | General |
| D244/1/3 | Parishes |
| D244/1/4 | Castletown |
| D244/1/5 | Douglas |
| D244/1/6 | Onchan |
| D244/1/7 | Port St Mary |
| D244/1/8 | Peel |
| D244/1/9 | Port Erin |
| D244/1/10 | Ramsey |
In addition section D244 contains bound volumes of indexed Local Government Acts (1886-1901) at D244/4a.
**Bye-Laws, Rules and Regulations framed and made by the Douglas Town Commissioners and confirmed by the Tynwald Court**
Douglas: James Brown, ‘Times’ Office, 1865
Library Ref: F71/5
Bye-Laws with respect to new streets and buildings made by Douglas Town Commissioners
Douglas: 1893
Made under powers conferred by the Local Government Act of 1886 and approved by Tynwald Court, 17 November 1893, these by-laws include stipulations for building drainage, ventilation, water closets, privies and cesspools.
Model Bye-Laws for the regulation of Ashpits, New Streets, Buildings, &c made under the Provisions of the Local Government Acts, 1900
42 page pamphlet
Vaccination Amendment Bill, Isle of Man: Reply of Dr Walter R Hadwen to the Medical Evidence and Arguments of Members of The Legislative Council presented in favour of the Continuance of Legal Compulsion
Gloucester: F J Brooke, 1913
Anti vaccination responses to evidence arguing for compulsory vaccination. Published by The British Union for the Abolition of Vivisection.
STATUTORY PROVISIONS
The statutes are indexed and available on the Reading Room shelves. The following is not intended as a comprehensive summary of relevant Manx legislation.
Act for better supplying the town of Douglas with water, 1834
Public Health Acts, 1884 & 1884
Douglas Town Improvement Act, 1889
Local Government Act, 1889
CHOLERA AND SMALLPOX PREVENTION
Letter from Government Office
[to Commissioners of Town and Village Districts] noting the wishes of the Lieutenant Governor for contingency planning given the outbreak of cholera in Europe.
31 August 1892
Church pamphlet
‘A Special Form of Prayer to Almighty God to be read...in all Churches and Chapels in the Isle of Man on Sunday 19th August 1866...for continued preservation from the cattle plague and for protection against the cholera.’
Directions and Regulations for the prevention of the spread of smallpox in the town of Douglas
Order issued by Lt-Governor Henry Loch, September/November 1877
Report of Committee of Tynwald Court, appointed on 14 April 1893 to enquire into the working of the Vaccination Acts, 30 May 1893
Douglas, 1893
The report notes an unsatisfactory state of affairs concerning the registration of vaccinations and declares:
‘...it is manifest that the Vaccination Act [of 1878] is very inefficiently carried into effect’.
SANITATION
Sanitary Matters in the Isle of Man
The Sanitary Record, 14 March 1879, pp 161-165
Critical account of sanitary affairs on the Island, by the journal’s ‘Special Commissioner’. Compares Manx with British legislation and gives a detailed eye witness account of sanitary conditions.
Lancet, The Medical Notes in the Isle of Man by a Frequent Visitor
21 February 1891 pp 459-460
This reviews medical provision, climate and sanitary reforms on the Island.
MEDICAL PROVISION
Medical Aid and Poor Relief Commission (Isle of Man): Report and Evidence with Appendices presented to the Tynwald Court by Order of His Excellency the Lieutenant-Governor
Douglas: James Brown and Son, 1879
Sets out the testimonies of numerous individuals and the commissioners’ conclusions about existing provisions for providing medical aid and hospital accommodation to the poor with recommendations for improvements and the prevention of the spread of epidemic or contagious diseases.
White Hoe Isolation Hospital, BraddanReport on Infectious Diseases and Isolation Hospital Administration for the ten years ending 31 March 1896
Douglas, 1896
Compiled by Town Clerk Thomas H Nesbitt, this 56 page report contains useful tabulated statistics and notes on the response to infectious diseases as well as a history of the beginning of White Hoe.
GENERAL SOURCES
Davies, Charles. A Dissertation on the Causes of Consumption with special reference to its Prevalence in the Isle of Man / by Charles Davies
Douglas: Brown & Sons Ltd, 1899
Medical degree thesis submitted to the Victoria University, 1899.
Haviland Alfred. *The essential requisities of a sea-side health resort and the requirements of a health seeker; with the physical geography and climate of the Isle of Man: A lecture / by Alfred Haviland*
Douglas: Brown and Son, 1883
Delivered at the Masonic Hall, Douglas on 13 June 1883 this lecture compares the Island's climate favourably to Cos in the Mediterranean Sea. For a severe critique of the lecture see Talbot, below.
Radcliffe, Constance K. *Ninety Years Young: A short illustrated account of Ramsey District Cottage Hospital 1907-1997 / by Constance Radcliffe*
Douglas: Nelson Press Co., 1997.
Library Ref B468/16
Shining by the Sea: *A History of Ramsey 1800 - 1914 / by Constance Radcliffe*
Douglas: Nelson Press Co. Ltd, 1989
pp 45 - 48 and 71-73 cover public health
Library Ref F72/14 Reading Room
*Consumption: The Social and Geographical Causes Conducing to its Prevalence / by Alfred Haviland*
Douglas: J Brown and Son, [1883]
Lecture delivered by the author at the Victoria Hotel, Douglas in December 1883 in the presence of the Lt-Governor. Includes a coloured map illustrating the geographical distribution of consumption on the Isle of Man. Additional pamphlets by the same author can be found at this reference.
Library Ref B464/3; looseleaf map at B464/1x
*Here is the News: an illustrated Manx History / by Terry Cringle with features by R E Forster and G N Knivetton*
Douglas: The Manx Experience, 1992
ISBN 1873120 07 9 (Softback)
ISBN 1873120 09 5 (Casebound)
See especially page 88 'A matter of life or death' on the cholera outbreak of 1832. Note also pages 124 and 132.
Library Ref F64/180 Reading Room
Talbot, Rev Theosophilus. *Mr Haviland's Lecture Reviewed I-VI / by Rev T Talbot*
1883
Six newspaper articles of October-December 1883 criticizing the content and claims of an 1883 lecture on the 'Essential requisities of a health resort' (at B464/3) and querying the island's suitability as a health resort for those afflicted with rheumatism, nervous disorders and consumption given its actual climate of 'chilling fogs, excessive rain and boisterous winds'.
Library Ref B464/2xf
**NELLIE BRENNAN SOURCES**
McHardy, Mary. *The Cholera Epidemic in the Isle of Man / by Mary McHardy*
Journal of the IOMFHS, 10th Anniversary issue, May 1989, pp 29-34
Article tracing the spread of the disease to the British Isles, the 1832 epidemic on the island and the 1866 outbreak.
Library Ref L6/FAM
McHardy, Mary. *The Story of Judith Crinnal / by Mary McHardy*
Journal of the IOMFHS Vol.XIII No.2. May 1991, pp 57-58
Account of the death from cholera of this 30 year old woman and the associated information passed down by word of mouth concerning her burial and the location of her grave.
Library Ref L6/FAM
*Newspaper Index to Isle of Man papers (to ca. 1850)*
This contains plentiful references to cholera for the years 1826 - 1833 under the heading 'Health-General'. Smallpox, dysentery and typhus fever are also reported upon in this section.
See also: MD 312 (accession 6417) for typescript extracts from the Manx Sun for July - October 1832 on the cholera epidemic.
Pantin, Guy. *A study of Maternal Mortality and Midwifery on the Isle of Man, 1882 to 1961*
Offprint from Medical History, 1996, 40: 141-172
Library Ref B464/23
Radcliffe, Constance K. *Life in Ramsey 1765-1865 / by Constance Radcliffe*
Proceedings of Isle of Man Natural History and Antiquarian Society Vol IX, No 2 pp 183-196. See especially p 192.
Library Ref L6 Counter
Cowin, Hilda. *A Manx Heroine / by Hilda Cowin*
Douglas: Isle of Man Times reprint, 1932
The story of Nellie Brennan; episodes of the cholera outbreak 100 years ago.
Library Ref G88/5xf & G88/1x
Forrest, Katherine A. *Manx Recollections - Memorials of Eleanor Elliott / by Katherine Forrest*
London: James Nisbet and Co, 1894
Douglas: James Brown and Son, 1894
Chapter 4, 'The Cholera', contains a considerable amount of reference to Nellie Brennan.
Library Ref F65/83
Johnson, Joseph. *Noble Women of our Time / by Joseph Johnson*
London: T Nelson and Sons, 1883
See chapter XI - 'Miss Brennan among the Manx Poor', pp 246-263
Library Ref G88/B7/3
*Miss Eleanor Brennan: a Manx Methodist Worthy / by Joseph Johnson*
Douglas: 1895
Serialized in the Manx Wesleyan Methodist Church Record, Vol 3, pp 2, 11, 18, 28
Library Ref L6 MW 1/1xq
Stowell, Bellanne. *Labours of Love, A Manx Story, being a Memorial of Eleanor Brennan / by Bellanne Stowell*
Douglas: John Mylrea, 1859
Douglas: Hospice Care, 1988
For biographical details of the authoress see Manx Worthies, p 107 at G4. The reprint contains a commentary by Dr Guy Pantin.
Library Ref G88/B7/1-2
**MANUSCRIPT SOURCES**
**Atholl Papers**
See Reading Room card index to the deposit under subheading 'Health'. Includes correspondence on smallpox and procedures for vaccination (1817 -1824)
**Cholera Deaths in Castletown, 1832**
List of persons buried in Malew, with dates, in the cholera epidemic of 1832
Library Ref MD 287 (accession 6378)
See also: a published list of these names in IOMFHS Journal Vol XVII No 1 February 1995 pp 13-14
'Cure for cholera'
'One teaspoonfull of salt - One teaspoonfull cayenne peprr - Half a pint of boiling water mixed and taken soon as possible'. Undated handwritten remedy, probably compiled prior to the 1866 outbreak.
Library Ref Ms 1592 A
**The Dorcas Society of Douglas / by Mrs N M Wilson**
Paper read at the Methodist History Society meeting of October 1971 with illustrative record appendices. The address charts the history of the Society which was formed on 1 December 1834, a day set aside by Wesleyan Methodists and others of Douglas for public thanksgiving for the exemption of the town from cholera during the past year. The object of the Society was to administer relief through the distribution of 'plain and necessary articles of clothing to needy men, women and children'.
Library Ref MD 551 (accession 6778)
**Government Office Papers**
Various references, especially folders GO 37 and GO 39. Examples given below are by no means exclusive. The reading room card index should be consulted for further information and clarification.
GO 9/5 Draft of letter from the Lt-Governor to the Secretary of State informing him of the outbreak of cholera symptoms in Peel, 1866. Poor legibility.
GO 37/2 Printed open letter from W Trench (Medical Officer, Liverpool.), advising on the prevention of cholera, 1866.
GO 37/9 Request to the High Bailiffs of various districts asking for information about suitably equipped houses in their areas to be used at short notice for cholera patients; also about the availability of people willing to nurse, 1866.
GO 37/11-14 Responses to the above, 1866.
GO 37/18 Proclamation of precautions against cholera, 1873
**The Health of the Manx: Advances during 150 years / by Dr Guy Pantin**
Photocopy of text of lecture delivered in 1974. Previous attitudes of the Manx to health and medicine are explored and the causes and events of the 1832 cholera epidemic and smallpox outbreaks of the nineteenth century are well covered.
Library Ref MD 15123 (accession 8502)
**The Inquiry of Samuel Harris, 1878 / by Dr. Guy Pantin**
Photocopy of text of lecture delivered by Dr Guy Pantin in 1986 based on the conduct, findings and results of the Medical Aid and Poor Relief Commission chaired by Harris. Touches on the prevention of cholera in Douglas by supplying piped water from a protected source (page 4) and notes 'It was not so much the diseases which killed the people as the overcrowding and poor feeding that caused them to succumb to illnesses which better fed succeeding generations successfully withstood'.
Library Ref MD 1036 (accession 8616)
The Manx National Heritage Library holds a wealth of various and extensive source material. Staff at the library and throughout Manx National Heritage, are always pleased to hear from anyone with an interest in Manx history.
WT rev. ed. 6.00
Manx National Heritage,
Douglas, Isle of Man IM1 3LY.
Tel: (01624) 648000 Fax: (01624) 648001
E-mail: firstname.lastname@example.org
Website: www.gov.im/mnh
Opening times: Monday to Saturday, 10am to 5pm
Closed Sunday and for one week in January |
Body Temperature, Oxygen Uptake and Heart Rate during Walking in Water and on Land at an Exercise Intensity Based on RPE in Elderly Men
Kazutaka Fujishima\textsuperscript{1)} and Tomihiro Shimizu\textsuperscript{2)}
\textit{1) Institute of Health Science, Kyushu University}
\textit{2) Joetsu University of Education}
Abstract The purpose of this study was to clarify the characteristics of the physiological response that occurs while walking in water and on land at an exercise intensity based on the rating of perceived exertion (RPE) in elderly men. Nine elderly men ranging from 66–70 years of age participated in this study as subjects. The actual trials consisted of walking for 20 minutes in 31°C and 35°C water on an underwater treadmill. The water depth of the treadmill corresponded to the level of the xiphoid process in the subject. The same subjects performed on-land walking using a moving belt treadmill for 20 minutes at a room temperature of 27°C. The exercise intensity during walking in the two water trials and the on-land trial was the same “somewhat hard” measured on the basis of the subject’s RPE rating of 13. There was no significant difference between the subjects’ rectal temperatures among the three trials. The mean skin temperature and mean body temperature while walking for 20 minutes in 35°C water were significantly higher ($P<0.01$) than in 31°C water and on land. There were no significant differences in oxygen uptake and heart rate among the two trials in water and the on-land trial. The above results suggest that the exercise intensity based on a subject’s RPE may be an effective index for the prescription of thermoneutral water walking in the same way that it is for land walking in the elderly. \textit{J Physiol Anthropol} 22 (2): 83–88, 2003 http://www.jstage.jst.go.jp/en/
Keywords: water walking, body temperature, HR, RPE, elderly men
Introduction
Recently, walking and jogging in water have become popular nonswimming aerobic activities. These activities have been used as a part of rehabilitative, therapeutic and general conditioning programs, and are thought to be particularly useful for people with lower extremity injuries (Cureton, 1997). As indicated by Evans et al. (1978), the effects of water buoyancy and resistance make a high level of energy expenditure possible with relatively little movement or strain on the lower extremity joints. For this reason, water walking and jogging are also effective exercises for individuals with a body-weight problem as well as for people of middle or advanced age. Previous studies have reported metabolic and cardiorespiratory responses during walking or jogging in a pool (Evans et al., 1978). It has been difficult, however, to fix the physical and physiological intensity of walking and jogging in a pool.
Various studies of the cardiorespiratory and metabolic responses while walking on an underwater treadmill have been conducted (Migita et al., 1996; Hotta et al., 1993, 1994; Shimizu et al., 1998; Takaoka et al., 1999; Shono et al., 2000, 2001; Kato et al., 2002). Generally, heart rate (HR) and rating of perceived exertion (RPE) are used to establish the intensity of exercise in water (Pollock et al., 1994; American College of Sports Medicine: ACSM, 1998). HR during water exercise is lower than that during treadmill work on land (McArdle et al., 1976; Shimizu et al., 1998). Therefore, it is important to examine the relationship between HR while walking in water and HR while walking on land at an exercise intensity based on RPE.
Takeshima et al. (1997) reported the relationship of HR and oxygen uptake ($\text{VO}_2$) during water-walking in 30°C water and land-walking at 28°C at an exercise intensity at level 13 on the RPE scale for the elderly. These researchers carried out the trial using one water temperature and a 6-minute exercise duration in a pool. Therefore, it is necessary to clarify the finding with respect to different water temperatures, a given walking speed and the exercise duration of a practical exercise prescription.
On the other hand, it is known that HR during exercise in water is influenced by body temperature (Holmér and Bergh, 1974; Galbo et al., 1979). However, no such study has been performed on the relationship of cardiorespiratory responses including body temperature while walking in water to those same responses while walking on land at an exercise intensity based on RPE for elderly people.
The purpose of this study was to evaluate the interrelationships among body temperature, $\text{VO}_2$ and HR while
walking in water and on land in order to clarify the effectiveness of RPE for prescribing water walking for elderly people.
**Methods**
**Subjects**
Nine healthy male volunteers participated in this study. Their mean (SD) age, height, body mass, body mass index and percentage of body fat (%Fat) were 67.9 (1.7) years, 163.2 (9.0) cm, 61.6 (8.7) kg, 23.0 (1.5) kg·ht (m²)⁻¹ and 22.3 (4.8) %, respectively. %Fat was predicted based on the triceps and subscapula skinfold thickness of the subjects (Brozek et al., 1963; Nagamine and Suzuki, 1964).
This study was approved by the Ethics Committee of the Institute of Health Science, Kyushu University, and informed written consent was obtained from each subject.
**Exercise intensity**
To determine the intensity of exercise for each subject, preliminary tests were conducted to standardize RPE rating of 13 (somewhat hard) for each subject while walking in water and on land using Borg’s RPE scale (Borg, 1970). This rating scale was used because exercise of a moderate intensity was recommended by ACSM (1998), and Feriche et al. (1998) reported that the ventilatory threshold is equivalent to the intensity of RPE rating scale of 13.
**Procedures**
Subjects wore swimming trunks and rested in a sitting position for 40 minutes at a room temperature of 27 (0.9)°C before walking. The actual trials consisted of walking for 20 minutes in 31 (0.18)°C and 35 (0.33)°C water on an underwater treadmill (Aqua Ciser II, Model 100R, Ferno Washington Inc., USA). The depth of the water was the level of the subject’s xiphoid process. Subjects wearing swimming trunks rested in a sitting position for 40 minutes and then performed on-land walking using a moving belt treadmill (Treadmill 8700, Landice, Inc., USA) for 20 minutes at a room temperature of 27 (0.5)°C. The exercise intensity achieved was the same “somewhat hard” measured on the basis of the subject’s RPE while walking in water and on land. Three trials were performed at least one day apart at approximately the same time of day.
**Measurements and instruments**
Rectal temperature (Tre) and skin temperatures of the arm (Tarm), chest (Tchest) and thigh (Tthigh) were measured every minute with thermistor probes beginning with a pre-exercise rest period in air and continuing until the end of all the exercises. Tre was measured by inserting the thermistor probe 12 cm deep into the rectum of a subject. Adiabatic covers for the skin (Nihon Koden Co., Ltd., Japan) and transparent tape (3M Co., Ltd., USA) were used to prevent water infiltration into the sites of the skin temperature probes during each water-walking trial. A portable data recording machine (VMM-67, VINE Co., Ltd., Japan) was used to record the temperature minute-by-minute at the measurement site. Mean skin temperature (Tsk) was calculated as 0.25Tarm+0.43Tchest+0.32Tthigh (Roberts et al., 1977). Mean body temperature (Tb) was calculated from the equation (Gagge and Nishi, 1977) Tb=0.67Tre+0.33Tsk. VO₂ was calculated every 15 seconds during the experiment by an automatic breath aeromonitor (AE-300S, Minato Medical Science Co., Ltd., Japan). This analyzer was calibrated against a known standard calibration gas before each experiment. HR was continuously monitored using a Life Scope 8 (BSM-7105, Nihon Koden Co., Ltd., Japan). The waterproof electrode (Vitrode; D-90, Nihon Koden Co., Ltd., Japan) was used to measure HR during each water-walking trial.
**Statistical analysis**
The difference between the mean values of the parameters measured in water and in air at 27°C was analyzed by one-way repeated-measures ANOVA. A value of P<0.05 was taken as the limit for statistical significance. Descriptive statistics included means and standard deviations (SD).
**Results**
The changes in Tre during the two trials in water and the on-land trial are shown in Fig. 1. The value of Tre during walking for 20 minutes in 31°C water, 35°C water and on land did not change in the three trials. There was no significant difference in the value of Tre among the three trials. The changes in Tsk during the two trials in water and the on-land trial are shown in Fig. 2. Tsk and Tb during walking for 20 minutes in 31°C water and during walking on land decreased slightly in both trials. Tsk and Tb during walking for 20 minutes in 35°C water increased significantly compared with that measured before walking. Tsk and Tb during walking for 20 minutes in 35°C water were significantly higher (P<0.01) than in 31°C water or on land (Fig. 2 and Fig. 3).
VO₂ during the three walking trials increased gradually during the first 5 minutes of each exercise and then maintained a steady level (determined by ANOVA) until the end of the exercise. VO₂ for the two water-walking trials and the land-walking trial during a period from 5 minutes after the beginning of the exercise to the end of the exercise was 903(149) ml·min⁻¹ in 31°C water, 881(139) ml·min⁻¹ in 35°C water and 794(80) ml·min⁻¹ on land (Fig. 4). There was no significant difference in VO₂ between the two trials in water and the on-land trial.
HR during the two trials in water and the on-land trial increased gradually during the first 4-5 minutes of each exercise and then maintained a steady level until the end of the exercise according to an ANOVA analysis (Fig. 5). The increases in HR from 5 minutes after the start of the exercise to the end of the exercise in comparison with the pre-exercise measurement were 20(7) beats·min⁻¹ in 31°C water,
18(9) beats·min\(^{-1}\) in 35°C water and 15(10) beats·min\(^{-1}\) on land (Fig. 5). There was no significant difference in HR between the two trials in water and the on-land trial.
**Discussion**
In the present study, there was no significant difference in Tre among the two trials in water and the on-land trial. Shimizu et al. (1998) reported similar results, finding that there were no significant differences in Tre while walking in 30°C water, 35°C water and in air at 25°C with an exercise intensity equivalent to approximately 50% VO\(_2\) max for men. They found, however, that Tre during walking in 25°C water was lower than during walking in air at 25°C. Nadel et al. (1974) and Galbo et al. (1979) showed that core temperature increased in 33°C water at approximately 70% VO\(_2\) max. Therefore, core temperature during exercise in water obviously depends on the water temperature and exercise intensity.
In the present study, the value of Tsk during walking in 35°C water was higher than that measured during walking in 31°C water and that measured in on-land trials (Fig. 2). These findings suggest that Tsk during walking in water will be close to the water temperature. Nielsen and Davies (1976) reported similar results, finding that Tsk during exercise in water at
Fig. 3 Change in mean body temperature (Tb) between two trials, one in water and the other on land. *P<0.05 and †P<0.05 show that the temperatures are significantly different from those recorded while walking in 31°C water and on land.
Fig. 4 Change in oxygen uptake (VO₂) between two trials, one in water and the other on land.
30°C and at 33°C was close to the two respective water temperatures. Tb during the two water trials showed the same patterns of change as shown by Tsk. These findings suggest that Tsk was directly affected by the water temperature, and consequently a higher Tb resulted in 35°C water.
In this study, VO₂ based on each subject’s RPE was approximately the same between the two trials in water and the on-land trial. These findings were consistent with the report of Takeshima et al. (1997) with respect to VO₂ level during walking in water and on land based on RPE rating of 13.
In this study, HR for the two water-walking trials and the land-walking trial during a period from 5 minutes after the beginning of the exercise to the end of the exercise was 97(7) beats·min⁻¹ in 31°C water, 94(14) beats·min⁻¹ in 35°C water and 92(10) beats·min⁻¹ on land (Fig. 5). No significant difference in HR was recorded in the two trials in water and the on-land trial. The results were not consistent with the result that HR during walking in 30°C water was significantly lower than HR during walking on land (Takeshima et al., 1997). This inconsistency might be caused by the differences in the water depth; that is, the depth of water in the present study was at the level of a subject’s xiphoid process, and the study of Takeshima et al. (1997) was performed with the subjects immersed to the level of the axilla.
It seems reasonable that a lower HR during water exercise may have caused a greater venous return, a higher stroke
volume and a higher cardiac output in water immersion (Farhi and Linnarsson, 1977). Furthermore, many studies (Holmér and Bergh 1974; McArdle et al., 1976; Gleim and Nicholas, 1989; Shimizu et al., 1998) have indicated that HR during exercise in water decreased with decreasing water temperature. The results in the present study were not consistent with other studies, which may be due to the thermoneutral water temperatures used (31°C and 35°C). As is the standard, the values of the RPE scale (Borg, 1970) were chosen to be as close as possible to one-tenth of the corresponding HR. In the present study, the exercise intensity was conducted so as to use RPE rating of 13 (somewhat hard); however, HR during each exercise trial was lower than 130 beats·min⁻¹ (Fig. 5). It appears that the reason for the lower HR may be the decrease in maximal HR with aging. Therefore, caution should be exercised when using RPE with elderly people.
In conclusion, the above results suggest that the exercise intensity judged on the basis of the subject’s RPE may be an effective index for the prescription of thermoneutral water walking in the same way that it is for land walking.
Acknowledgements The authors wish to thank the Physical Therapist, Masao Sato of Niigata Chuo Hospital and Takahiro Adachi of the Graduate School of Human-Environment Studies, Kyushu University, for supporting this research. The authors also wish to thank Prof. Dr. Takeshi Ueda of Fukuoka Prefectural University for his invaluable guidance concerning statistical analysis.
References
American College of Sports Medicine (1998) The recommended quantity and quality of exercise for developing and maintaining cardiorespiratory and muscular fitness, and flexibility in healthy adults. Med Sci Sports Exerc 30: 975–991
Borg G (1970) Perceived exertion as an indicator of somatic stress. Scand J Rehabil Med 2: 92–98
Brozek J, Grande F, Anderson JT, Keys A (1963) Densitometric analysis of body composition: revision of some quantitative assumptions. Ann NY Acad Sci 110: 113–140
Cureton KJ (1997) Physiologic responses to water exercise. In: Ruoti RG, Morris DM, Cole AJ, eds. Aquatic Rehabilitation. Lippincott, Philadelphia, 39–56
Evans BW, Cureton KJ, Purvis JW (1978) Metabolic and circulatory responses to walking and jogging in water. Res Quart 49(4): 442–449
Farhi LE, Linnarsson D (1977) Cardiopulmonary readjustments during graded immersion in water at 35°C. Respir Physiol 30: 35–50
Feriche B, Chicharro JL, Vaquero AF, Perez M, Lucia A (1998) The use of a fixed value of RPE during a ramp protocol: Comparison with the ventilatory threshold. J Sports Med Phys Fitness 38: 35–38
Gagge AP, Nishi Y (1977) Heat exchange between human skin surface and thermal environment. In: Lee DHK, ed. Handbook Physiology. Reactions to Environment Agents. Am Physiol Soc, Bethesda, 69–72
Galbo H, Houston ME, Christensen NJ, Holst JJ, Nielsen B, Nygaard E, Suzuki J (1979) The effect of water temperature on the hormonal response to prolonged swimming. Acta Physiol Scand 105: 326–337
Gleim GW, Nicholas JA (1989) Metabolic costs and heart rate responses to treadmill walking in water at different depths and temperatures. Am J Sports Med 17: 248–252
Holmér I, Bergh U (1974) Metabolic and thermal response to swimming in water at varying temperatures. J Appl Physiol
Hotta N, Muraoka Y, Ogaki T, Kanaya S, Fujishima K, Hong JP, Masuda T (1993) Cardiovascular responses to treadmill (flowmill) walking in water. Kurume Journal of Health & Physical Education 1: 19–23 (in Japanese with English abstract)
Hotta N, Ogaki T, Kanaya S, Fujishima K, Hagiwara H (1994) Exercise prescription for aged men and women in water. Descente Sports Science 15: 78–83 (in Japanese with English abstract)
Kato T, Onishi S, Kitagawa K (2002) Physiological responses during underwater walking and running using underwater treadmill. Research J Phys Educ Chukyo Univ 43(2): 1–8 (in Japanese with English abstract)
McArdle WD, Magel JR, Lesmes GR, Pechar GS (1976) Metabolic and cardiovascular adjustment to work in air and water at 18, 25, and 33°C. J Appl Physiol 40: 85–90
Migita T, Hotta N, Ogaki T, Kanaya S, Fujishima K, Masuda T (1996) Comparison of the physiological responses to treadmill prolonged walking in water and on land. Japan J Phys Educ 40: 316–323 (in Japanese with English abstract)
Nadel ER, Holmér I, Bergh U, Åstrand P-O, Stolwijk JAJ (1974) Energy exchanges of swimming for man. J Appl Physiol 36: 465–471
Nagamine S, Suzuki S (1964) Anthropometry and body composition of Japanese young men and women. Hum Biol 36: 8–15
Nielsen B, Davies CTM (1976) Temperature regulation during exercise in water and air. Acta Physiol Scand 98: 500–508
Pollock ML, Graves JE, Swart DL, Lowenthal DT (1994) Exercise training and prescription for the elderly. Southern Med J 87: S88–S95
Roberts MF, Wenger CB, Stolwijk JAJ, Nadel ER (1977) Skin blood flow and sweating changes following exercise training and acclimation. J Appl Physiol 43: 133–137
Shimizu T, Kosaka M, Fujishima K (1998) Human thermoregulatory responses during prolonged walking in water at 25, 30, and 35°C. Eur J Appl Physiol 78: 473–478
Shono T, Fujishima K, Hotta N, Ogaki T, Ueda T, Otoki K, Teramoto K, Shimizu T (2000) Physiological responses and RPE during underwater treadmill walking in women of middle and advanced age. J Physiol Anthropol 19: 195–200
Shono T, Fujishima K, Hotta N, Ogaki T, Masumoto K (2001) Cardiorespiratory response to low-intensity walking in water and on land in elderly women. J Physiol Anthropol 20: 269–274
Takaoka I, Ohnishi T, Okamura S, Suzuki D (1999) The physiological responses and RPE to different water flow and belt velocities during “flowmill” walking in water. Jour Health Sports Sci Juntendo Univ 3: 61–67 (in Japanese with English abstract)
Takeshima N, Nakata M, Kobayashi F, Tanaka K, Pollock ML (1997) Oxygen uptake and heart rate differences between walking on land and in water in the elderly. JAPA 5: 126–134
Received: July 29, 2002
Accepted: December 18, 2002
Correspondence to: K. Fujishima, Institute of Health Science, Kyushu University 11, Kasuga 816–8580, Japan
Tel. & Fax: +81–092–583–7849
e-mail: firstname.lastname@example.org |
Lin Zhong
Professor
Department of Electrical & Computer Engineering
Rice University, Houston, TX 77005
Email: email@example.com URL: http://www.linzhong.org
RESEARCH INTERESTS
Mobile & embedded systems, wireless networking
EDUCATION
Ph.D. in Computer Engineering, Princeton University, 2005
M.S. in Electronic Engineering, Tsinghua University, 2000
B.S. in Electronic Engineering with Honors, Tsinghua University, 1998
HONORS & AWARDS
• ACM SIGMOBILE RockStar Award, 2014
• Charles Duncan Award for Outstanding Academic Achievement (Rice University), 2014
• Best Paper Award, ACM MobiSys, 2014
• Best Paper Award, ACM ASPLOS 2014
• Best Paper Award, ACM MobiSys, 2013
• Best Paper Award, ACM PhoneSense Workshop, 2011
• Best Paper Award, ACM MobiSys, 2011
• National Science Foundation CAREER Award, 2011
• Brown Teaching Award, 2010
• Best Demo/Poster Award, ACM HotMobile, 2010
• Outstanding Faculty Associate, Jones College 2009
• Mark Weiser Best Paper Award, IEEE PerCom 2009
• Best Paper Award, ACM MobileHCI, 2007
• One of the 30 most influential papers published in the first decade of *IEEE Design Automation & Test in Europe Conference* (DATE), 2007
• Harold W. Dodds Honorific Fellowship, Princeton University, 2004-2005
• AT&T Asian Pacific Leadership Award, 2000
PROFESSIONAL EXPERIENCE
• Professor: July 2016 to present, Dept. of ECE, Rice University
• Associate Professor: July 2011 to June 2016, Dept. of ECE, Rice University
• Visiting Researcher: May to Aug. 2011 and Mar. to Dec. 2012, Microsoft Research, Redmond
• Assistant Professor: September 2005 to June 2011, Dept. of ECE, Rice University
• Summer Research Intern, 2004 and 2005, Microsoft Research, Redmond, WA
• Summer Research Intern, 2003, NEC Labs, America, Princeton, NJ
• Graduate Research Assistant: Fall 2000 to Spring 20005, Princeton University
PUBLICATIONS
Refereed Conference & Workshop Proceedings
1. Lin Zhong, Jiong Luo, Yunsi Fei and Niraj K. Jha, "Register binding based power management for high-level synthesis of control-flow intensive behaviors," in Proc. *IEEE Int. Conf. on Computer Design (ICCD)*, Sept. 2002.
2. Lin Zhong and Niraj K. Jha, "Interconnect-aware high-level synthesis for low power," in Proc. *IEEE/ACM Int. Conf. on Computer-Aided Design (ICCAD)*, Nov. 2002.
3. W. Wang, T. K. Tan, J. Luo, Y. Fei, L. Shang, K. S. Vallerio, Lin Zhong, A. Raghunathan, and N. K. Jha, "A comprehensive high-level synthesis system for control-flow intensive behaviors for low power," in Proc. *IEEE Great Lakes VLSI Symp.*, Apr. 2003.
4. Lin Zhong and Niraj K. Jha, "Graphical user interface energy characterization for handheld computers," in Proc. *IEEE/ACM Int. Conf. on Compilers, Architectures & Synthesis for Embedded Systems (CASES)*, Oct. 2003.
5. Pallav Gupta, Lin Zhong and Niraj K. Jha, "A high-level interconnect power model for design space exploration," in Proc. *IEEE/ACM Int. Conf. on Computer-Aided Design (ICCAD)*, Nov. 2003.
6. Lin Zhong and Niraj K. Jha, "Dynamic power optimization for interactive systems," in Proc. *IEEE Int. Conf. on VLSI Design*, Jan. 2004.
7. Rui Zhang, Pallav Gupta, Lin Zhong, and Niraj K. Jha, "Synthesis and optimization of threshold logic networks with application to nanotechnologies," in Proc. *IEEE Design Automation & Test in Europe Conf (DATE)*, Feb. 2004. *(selected as one of the most influential papers in the first decade of DATE)*
8. Yunsi Fei, Lin Zhong, and Niraj K. Jha, "An energy-aware framework for coordinated dynamic software management in mobile computers," in Proc. *IEEE/ACM Int. Symp. on Modeling, Analysis, and Simulation of Computer & Telecommunication Systems (MASCOTS)*, Oct. 2004.
9. Lin Zhong, Srivaths Ravi, Anand Raghunathan, and Niraj K. Jha, "Power estimation for cycle-accurate functional descriptions of hardware," in Proc. *IEEE/ACM Int. Conf. on Computer-Aided Design (ICCAD)*, Nov. 2004.
10. Lin Zhong and Niraj K. Jha, "Energy efficiency for handheld computer interfaces: Limits, characterization, and practice," in Proc. *USENIX/ACM Int. Conf. Mobile Systems, Applications, and Services (MobiSys)*, June 2005.
11. Le Yan, Lin Zhong, and Niraj K. Jha, "User-perceived latency based dynamic voltage scaling for interactive applications," in Proc. *ACM/IEEE Design Automation Conf. (DAC)*, June 2005.
12. Le Yan, Lin Zhong, and Niraj K. Jha, "Towards a responsive, yet power-efficient, operating system: A Holistic approach," in Proc. *IEEE/ACM Int. Sym. on Modeling, Analysis, and Simulation of Computer & Telecommunication Systems (MASCOTS)*, Sept. 2005.
13. Lin Zhong, Mike Sinclair and Niraj K. Jha, "A Personal-area network of low-power wireless interfacing devices: System & hardware design," in Proc. *ACM Int. Conf. Human Computer Interaction with Mobil Devices and Services (MobileHCI)*, Sept. 2005.
14. Lin Zhong, Mike Sinclair, and Ray Bittner, "A Phone-centered body sensor network platform: Cost, energy efficiency, and user interface," in Proc. *IEEE Workshop on Body Sensor Networks*, April 2006.
15. Lin Zhong, Bin Wei, and Mike Sinclair, "SMERT: Energy-efficient design of a multimedia messaging system for mobile devices," in Proc. *ACM/IEEE Design Automation Conf. (DAC)*, July, 2006.
16. Ahmad Rahmati and Lin Zhong, "Context-for-Wireless: context-sensitive energy-efficient wireless data transfer," in Proc. *ACM/USENIX Int. Conf. Mobile Systems, Applications, and Services (MobiSys)*, June 2007.
17. Ahmad Rahmati, Matti Hiltunen, Lin Zhong, and Rittwik Jana, "Reliability techniques for RFID-based object tracking applications," in Proc. *IEEE/IFIP Int. Conf. Dependable Systems and Networks (DSN)*, June 2007.
18. Le Yan, Lin Zhong, and Niraj Jha, "Energy comparison and optimization of wireless body-area network technologies," in Proc. *Int. Conf. Body Area Networks (BodyNets)*, 2007.
19. Lin Zhong, Diana El-Daye, Brett Kaufman, Nick Tobaoda, Tamer Mohamed, and Michael Liebschner, "OsteoConduct: Wireless body-area communication based on bone conduction," in *Proc. Int. Conf. Body Area Networks (BodyNets)*, June 2007.
20. Quming Zhou, Lin Zhong, and Kartik Mohanram, "Power signal processing: a new perspective for power analysis and optimization," in *Proc. ACM/IEEE Int. Symp. Low Power Electronics Design (ISLPED)*, August 2007.
21. Ahmad Rahmati, Angela Qian, and Lin Zhong, "Understanding human-battery interaction on mobile phones," in *Proc. ACM Int. Conf. Human Computer Interaction with Mobile Devices and Services (MobileHCI)*, September 2007. *(received Best Paper Award)*
22. Ahmad Rahmati and Lin Zhong, "Usability evaluation of a commercial Pocket PC phone: a pilot study," in *Proc. ACM Int. Conf. Mobile Technology, Applications and Systems (Mobility)*, September 2007.
23. Nilanjan Banerjee, Ahmad Rahmati, Mark Corner, Sami Rollins, and Lin Zhong, "Users and batteries: interactions and adaptive energy management in mobile systems," in *Proc. Int. Conf. Ubiquitous Computing (Ubicomp)*, September 2007.
24. Mian Dong and Lin Zhong, "Logic synthesis with nanowire crossbar: Reality check and standard cell-based integration," in *Proc. IEEE Design Automation & Test in Europe Conf (DATE)*, March 2008.
25. Jiayang Liu and Lin Zhong, "Micro power management of active 802.11 interfaces," in *Proc. ACM/USENIX Int. Conf. Mobile Systems, Applications, and Services (MobiSys)*, June 2008.
26. Mian Dong and Lin Zhong, "Challenges to crossbar integration of nanoscale two-terminal symmetric memory devices," in *Proc. IEEE Int. Conf. Nanotechnology*, August 2008.
27. Guangming Hong, Ahamd Rahmati, Ye Wang, and Lin Zhong, "SenseCoding: Acceleration-assisted motion estimation for efficient video encoding," in *Proc. ACM Int. Conf. Multimedia*, October 2008.
28. Ahmad Rahmati*, Clayton Shepard*, and Lin Zhong, "NoShake: Content stabilization for shaking screens of mobile devices," in *Proc. IEEE Int. Conf. Pervasive Computing and Communication (PerCom)*, March 2009. *(Equal contribution)*
29. Jiayang Liu, Zhen Wang, Lin Zhong, Jehan Wickramasuriya, and Venu Vasudevan, "uWave: Accelerometer-based personalized gesture recognition and its applications," in *Proc. IEEE Int. Conf. Pervasive Computing and Communication (PerCom)*, March 2009. *(received the Mark Weiser Best Paper Award)*
30. Husheng Li, Lin Zhong, and Kun Zheng, "Drowsy transmission: physical layer energy optimization for transmitting random packet traffic," in *Proc. IEEE Conf. Computer Communications (INFOCOM)*, April 2009.
31. Stephen So, Ardalan Amiri Sani, Lin Zhong, Frank Tittel, and Gerard Wysocki, "Laser spectroscopic trace-gas sensor networks for atmospheric monitoring applications," in *Proc. IPSN Wrkshp. Sensor Networks for Earth and Space Science Applications (ESSA)*, April 2009.
32. Yunxin Liu, Ahmad Rahmati, Yuanhe Huang, Hyukjae Jang, Lin Zhong, and Yongguang Zhang, "XShare: supporting impromptu sharing of mobile phones," in *Proc. ACM Int. Conf. Mobile Systems, Applications, and Services (MobiSys)*, June 2009
33. Mian Dong, Kevin Choi, and Lin Zhong, "Power modeling of graphical user interfaces on OLED displays," in *Proc. ACM/IEEE Design Automation Conf. (DAC)*, July 2009.
34. Mian Dong, Kevin Choi, and Lin Zhong, "Power-saving color transformation of mobile graphical user interfaces on OLED-based displays," in *Proc. ACM/IEEE Int. Symp. Low Power Electronics and Design (ISLPED)*, August 2009.
35. Xiaohan Ma, Mian Dong, Lin Zhong, and Zhigang Deng, "Statistical power consumption analysis and modeling for GPU-based computing," in *Proc. ACM Wrkshp. Power Aware Computing and Systems (HotPower)*, Co-located with SOSP, October 2009.
36. Xiaoming Chen, Zhendong Zhao, Ahmad Rahmati, Ye Wang, and Lin Zhong, "Sensor-assisted motion estimation for efficient H.264/AVC video encoding," in *Proc. ACM Int. Conf. Multimedia*, October 2009.
37. Jiayang Liu, Lin Zhong, Jehan Wickramasuriya, and Venu Vasudevan, "User evaluation of lightweight user authentication with a single tri-axis accelerometer," in *Proc. ACM Int. Conf. Human Computer Interaction with Mobile Devices and Services (MobileHCI)*, September 2009.
38. Hang Yu, Lin Zhong, and Ashutosh Sabharwal, "Adaptive RF chain management for energy-efficient spatial-multiplexing MIMO transmission," in *Proc. ACM/IEEE Int. Symp. Low Power Electronics and Design (ISLPED)*, August 2009.
39. Ahmad Rahmati, Lin Zhong, Jehan Wickramasuriya, and Venu Vasudevan, and Daniel Stewart, "Enabling pervasive mobile applications with the FM radio broadcast data system," in *Proc. ACM Int. Workshop on Mobile Computing Systems and Applications (HotMobile)*, February 2010.
40. Clay Shepard, Ahmad Rahmati, Chad Tossel, Lin Zhong, and Phil Kortum, "LiveLab: measuring wireless networks and smartphone users in the field," in *Proc. Workshop on Hot Topics in Measurement & Modeling of Computer Systems (HotMetrics)*, June 2010
41. Ardalan Amiri Sani, Hasan Dumanli, Lin Zhong, and Ashutosh Sabharwal, "Power-efficient directional wireless communication on small form-factor mobile devices," in *Proc. ACM/IEEE Int. Symp. Low Power Electronics and Design (ISLPED)*, August 2010
42. Ardalan Amiri Sani, Lin Zhong, and Ashutosh Sabharwal, "Directional antenna diversity for mobile devices; characterizations and solutions," in *Proc. ACM Int. Conf. Mobile Computing and Networking (MobiCom)*, September 2010.
43. Chad Tossel, Phillip Kortum, Clayton Shepard, Ahmad Rahmati, and Lin Zhong, "Assessing the performance of common tasks on handheld mobile computers," in *Proc. Human Factors and Ergonomics Society (HFES)*, September 2010.
44. Xiaozhu Lin, Ahmad Rahmati, and Lin Zhong, "Dandelion: A framework for transparently programming phone-centered wireless body sensor applications for health," in *Proc. ACM Wireless Health*, October, 2010.
45. Zhen Wang, Xiaozhu Lin, Lin Zhong, and Mansoor Chistie, "Why mobile browsers are slow," in *Proc. ACM Int. Workshop on Mobile Computing Systems and Applications (HotMobile)*, 2011
46. Xuan Bao, Trevor Narayan, Ardalan Amiri Sani, Wolfgang Richter, Romil Roy Choudhury, Lin Zhong, and Mahadev Satyanarayanan, "The case for context-aware compression," to appear in *Proc. ACM Int. Workshop on Mobile Computing Systems and Applications (HotMobile)*, March 2011.
47. Dong Mian and Lin Zhong, "Self-constructive, high-rate energy modeling for battery-powered mobile systems," in *Proc. ACM Int. Conf. Mobile Systems, Applications, and Services (MobiSys)*, June 2011
48. Dong Mian and Lin Zhong, "Chameleon: a color-adaptive web browser for mobile OLED displays," in *Proc. ACM Int. Conf. Mobile Systems, Applications, and Services (MobiSys)*, June 2011 (*received Best Paper Award*)
49. Hang Yu, Ahmad Rahmati, Ardalan Amiri Sani, Lin Zhong, Jehan Wickramasuriya and Venu Vasudevan, "Data broadcasting using mobile FM radio: design, realization and application," in *Proc. ACM Int. Conf. Ubiquitous Computing (UbiComp)*, September 2011.
50. Hang Yu, Lin Zhong, Ashutosh Sabharwal, and David Kao, "Beamforming on mobile devices: a first study," in *Proc. ACM Int. Conf. Mobile Computing and Networking (MobiCom)*, September 2011.
51. Robert LiKamWa, Yunxin Liu, Nicholas D. Lane, and Lin Zhong, "Can your smartphone infer your mood?" in *Proc. ACM Workshop on Sensing Applications on Mobile Phones (PhoneSense)*, co-located with ACM SenSys, November 2011. *(received Best Paper Award)*
52. Xiaozhu Lin, Zhen Wang, Robert LiKamWa, and Lin Zhong, "Using low-power processors in smartphones without knowing them," in *Proc. Int. Conf. Architectural Support for Programming Languages and Operating Systems (ASPLOS)*, March 2012.
53. Ning Ding, A. Pathak, D. Koutsounikolas, C. Shepard, Y. C. Hu, and Lin Zhong, "Realizing the full potential of PSM using proxying," in *Proc. IEEE INFOCOM Mini-Symposium*, March 2012.
54. Zhen Wang, Felix Xiaozhu Lin, Lin Zhong, and Mansoor Chishtie, "How far can client-only solutions go for mobile browser speed?" in *Proc. the World Wide Web Conference (WWW)*, April 2012.
55. Chad Tossell, Phil Kortum, Ahmad Rahmati, Clayton Shepard, and Lin Zhong, "Characterizing web use on smartphones," in *Proc. ACM SIGCHI Conf. Human Factors in Computing Systems (CHI)*, May 2012.
56. Clayton Shepard, Hang Yu, Narendra Anand, Li Erran Li, Tom Marzetta, Yang Richard Yang, and Lin Zhong, "Argos: practical base stations with many antennas," in *Proc. ACM Int. Conf. Mobile Computing and Networking (MobiCom)*, August 2012.
57. Ahmad Rahmati, Chad Tossell, Clayton Shepard, Phil Kortum, and Lin Zhong, "Exploring iPhone usage: the influence of socioeconomic differences on smartphone adoption, usage and usability," in *Proc. ACM Int. Conf. Human Computer Interaction with Mobile Devices and Services (MobileHCI)*, September 2012.
58. Xiaozhu Lin, Zhen Wang, and Lin Zhong, "Supporting distributed execution of smartphone workloads on loosely coupled heterogeneous processors," in *Proc. Wrkshop. Power-Aware Computing and Systems (HotPower)*, October 2012.
59. Clayton Shepard, Narendra Anand, and Lin Zhong, "Practical performance of MU-MIMO precoding in many-antenna base stations," in *Proc. ACM CellNet Workshop* (collocated with MobiSys), June 2013.
60. Robert LiKamWa, Bodhi Priyantha, Matthai Philipose, Lin Zhong, and Paramvir Bahl, "Energy characterization and optimization of image sensing toward continuous mobile vision," in *Proc. ACM Int. Conf. Mobile Systems, Applications and Services (MobiSys)*, June 2013. *(Received Best Paper Award)*
61. Robert LiKamWa, Yunxin Liu, Nicholas D. Lane, and Lin Zhong, "MoodScope: building a mood sensor from smartphone usage patterns," in *Proc. ACM Int. Conf. Mobile Systems, Applications and Services (MobiSys)*, June 2013.
62. Xiaozhu Lin, Zhen Wang, and Lin Zhong, "K2: A mobile operating system for heterogeneous coherence domains," in *Proc. ACM Int. Conf. Architectural Support for Programming Languages and Operating Systems (ASPLOS)*, March 2014. *(Received Best Paper Award)*
63. Ardalan Amiri Sani, Kevin Boos, Shaopu Qin, and Lin Zhong, "I/O paravirtualization at the device file boundary," in *Proc. ACM Int. Conf. Architectural Support for Programming Languages and Operating Systems (ASPLOS)*, March 2014.
64. Ardalan Amiri Sani, Kevin Boos, Min Hong Yun, and Lin Zhong, "Rio: A system solution for sharing I/O between mobile systems," *Proc. ACM Int. Conf. Mobile Systems, Applications and Services (MobiSys)*, June 2014. *(Received Best Paper Award)*
65. Robert LiKamWa, Zhen Wang, Aaron Carroll, Xiaozhu Lin, and Lin Zhong, "Draining our Glass: an energy and heat characterization of Google Glass," in *Proc. ACM SIGOPS Asia-Pacific Workshop on Systems (APSys)*, June 2014.
66. Chao Xu, Xiaozhu Lin, and Lin Zhong, "Device drivers should not do power management," in *Proc. ACM SIGOPS Asia-Pacific Workshop on Systems (APSys)*, June 2014.
67. Mian Dong, Tian Lan, and Lin Zhong, "Rethink energy accounting with multi-player game theory," in *Proc. ACM Int. Conf. Mobile Computing and Networking (MobiCom)*, September 2014.
68. Hang Yu, Oscar Bejarano, and Lin Zhong, "Combating inter-cell interference in 802.11ac-based multi-user MIMO networks," in *Proc. ACM Int. Conf. Mobile Computing and Networking (MobiCom)*, September 2014.
69. Chao Xu, Xiaozhu Lin, Yuyang Wang, and Lin Zhong, "Automated OS-level device runtime power management," in *Proc. ACM Int. Conf. Architectural Support for Programming Languages and Operating Systems (ASPLOS)*, March 2015.
70. Robert LiKamWa and Lin Zhong, "Starfish: efficient concurrency support for computer vision applications," in *Proc. ACM Int. Conf. Mobile Systems, Applications and Services (MobiSys)*, May 2015.
71. Kevin A. Boos, Ardalan Amiri Sani, and Lin Zhong, "Eliminating state entanglement with checkpoint-based virtualization of mobile OS services," in *Proc. ACM SIGOPS Asia-Pacific Workshop on Systems (APSys)*, July 2015.
72. Clayton Shepard, Abeer Javed, and Lin Zhong, "Control channel design for many-antenna MU-MIMO," in *Proc. ACM Int. Conf. Mobile Computing and Networking (MobiCom)*, September 2015.
73. Robert LiKamWa, Yunhui Hou, Julian Gao, Mia Polansky, and Lin Zhong, "RedEye: Analog ConvNet image sensor architecture for continuous mobile vision," in *Proc. ACM/IEEE Int. Symp. Computer Architecture (ISCA)*, June 2016.
74. Clayton Shepard, Jian Ding, Ryan E. Guerra, and Lin Zhong, "Understanding real many-antenna MU-MIMO channels," in *Proc. of IEEE Asilomar Conference*, November 2016.
75. Min Hong Yun, Songtao He, and Lin Zhong, "Reducing latency by eliminating synchrony," in *Proc. the World Wide Web Conference (WWW)*, April 2017.
76. Kevin Boos, Emilio Del Vecchio, and Lin Zhong, "A characterization of state spill in modern operating systems," in *Proc. ACM EuroSys Conference*, April 2017.
77. Kevin Boos and Lin Zhong, "Theseus: a state spill-free operating system," in *Proc. ACM PLOS Workshop*, October 2017.
78. Clayton Shepard, Rahman Doost-Mohammady, Jian Ding, Ryan Guerra, and Lin Zhong, "ArgosNet: a multi-cell many-antenna MU-MIMO platform," in *Proc. of IEEE Asilomar Conference*, November 2017.
**Refereed Journal Articles**
1. Jiong Luo, Lin Zhong, Yunsi Fei and Niraj K. Jha, "Register binding based RTL power management for control-flow intensive designs," *IEEE Trans. on Computer-Aided Design of ICs & Systems*, Aug. 2004.
2. Rui Zhang, Pallav Gupta, Lin Zhong, and Niraj K. Jha, "Threshold network synthesis and optimization and its application to nanotechnologies," *IEEE Trans. on Computer-Aided Design of ICs & Systems*, Jan. 2005.
3. Lin Zhong and N. K. Jha, "Interconnect-aware low power high-level synthesis," *IEEE Trans. on Computer-Aided Design of ICs & Systems*, Mar. 2005.
4. K. S. Vallerio, Lin Zhong and N. K. Jha, "Energy-efficient graphical user interface design," *IEEE Trans. on Mobile Computing*, July 2006.
5. Lin Zhong, Srivaths Ravi, Anand Raghunathan and Niraj K. Jha, "RTL-aware cycle-accurate functional power estimation," *IEEE Trans. on Computer-Aided Design of ICs & Systems*, October 2006.
6. Lin Zhong and Niraj K. Jha, "Dynamic power optimization targeting user delays in interactive systems," *IEEE Trans. on Mobile Computing*, November 2006.
7. Jun Yao, Lin Zhong, Douglas Natelson, and James M. Tour, "Etching-dependent reproducible memory switching in vertical SiO$_2$ structures," in *Applied Physics Letters*, vol. 93, issue 25, 2008
8. Yunsi Fei, Lin Zhong, and Niraj K. Jha, "An energy-aware framework for dynamic software management in mobile computing systems," *ACM Trans. on Embedded Computing Systems*, April 2008.
9. Mian Dong and Lin Zhong, "Nanowire crossbar logic and standard cell-based integration," in *IEEE Trans. on Very Large Scale Integration Systems*, August 2009.
10. Ahmad Rahmati and Lin Zhong, "Human-battery interaction on mobile phones," in *Pervasive and Mobile Computing*, vol. 5, no. 5, October 2009.
11. Jun Yao, Zhong Jin, Lin Zhong, Douglas Natelson, and James M. Tour, "Two-terminal nonvolatile memories based on single-walled carbon nanotubes," in *ACS Nano*, 3 (12), pp. 4122-4126, 2009.
12. Jiayang Liu, Zhen Wang, Lin Zhong, Jehan Wickramasuriya, and Venu Sudevane, "uWave: Accelerometer-based personalized gesture recognition and its applications," in *Pervasive and Mobile Computing*, vol. 5, issue 6, pp. 657-675, December 2009.
13. Jun Yao, Lin Zhong, Zengxing Zhang, Tao He, Zhong Jin, Patrick J. Wheeler, Douglas Natelson, and James M. Tour, "Resistive switching in nanogap systems on SiO$_2$ substrates," in *Small*, 5 (24), pp. 2910-2915, 2009.
14. Jun Yao, Zhengzong Sun, Lin Zhong, Douglas Natelson, and James Tour, "Resistive switches and memories from silicon oxide," in *ACS Nano Letters*, 2010.
15. Yunxin Liu, Ahmad Rahmati, Yuanhe Huang, Hyukjae Jang, Lin Zhong, Yongguang Zhang, and Shensheng Zhang, "Design, realization, and evaluation of xShare for impromptu sharing of mobile phones," in *IEEE Trans. Mobile Computing*, December 2010.
16. Clayton Shepard, Ahmad Rahmati, Chad Tossell, Lin Zhong, and Phillip Kortum, "LiveLab: measuring wireless networks and smartphone users in the field", in *ACM SIGMETRICS Perform. Eval. Rev.*, vol. 38, no. 3, December 2010.
17. Ahmad Rahmati and Lin Zhong, "Context-based network estimation for energy-efficient ubiquitous wireless connectivity," in *IEEE Transactions on Mobile Computing*, January 2011.
18. Xiaoming Chen, Zhendong Zhao, Ahmad Rahmati, Ye Wang, and Lin Zhong, "SaVE: Sensor-Assisted Motion Estimation for Efficient H.264/AVC Encoding," in *IEEE Trans. Circuits and Systems for Video Technology*, March 2011
19. Jun Yao, Lin Zhong, Douglas Natelson, and James Tour, "Intrinsic resistive switching and memory effects in silicon oxide," in *Applied Physics A* (Materials Science & Processing), 102(4), 2011.
20. Jun Yao, Lin Zhong, Douglas Natelson, and James Tour, "Silicon Oxide is a non-innocent surface for molecular electronics and nanoelectronics studies," in *Journal of the American Chemical Society*, 133(4), 2011.
21. Jun Yao, Lin Zhong, Douglas Natelson, and James M. Tour, "In situ imaging of the conducting filament in a silicon oxide resistive switch," *Nature Scientific Reports*, 2:242, 1-5, 2012.
22. Chad Tossell, Phillip Kortum, Clayton Shepard, Laura Barg-Walkow, Ahmad Rahmati, and Lin Zhong, "A Longitudinal Study of Emoticon Use in Text Messaging from Smartphones", in *Computers in Human Behavior*, vol. 28, no. 2, March 2012.
23. Mian Dong and Lin Zhong, "Chameleon: a color-adaptive web browser for mobile OLED displays," in *IEEE Transactions on Mobile Computing*, May 2012.
24. Chad Tossell, Phillip Kortum, Clayton Shepard, Ahmad Rahmati, and Lin Zhong, "Getting real: A naturalistic methodology for using smartphones to collect mediated communications", in *Advances in Human-Computer Interaction*, June 2012.
25. Hang Yu, Lin Zhong, Ashutosh Sabharwal, “Power Management of MIMO Network Interfaces on Mobile Systems,” in *IEEE Trans. VLSI*, July 2012.
26. Mian Dong and Lin Zhong, “Power modeling and optimization of OLED displays,” in *IEEE Trans. Mobile Computing*, September 2012.
27. Chad Tossell, Phillip Kortum, Clayton Shepard, Ahmad Rahmati, and Lin Zhong, "An Empirical Analysis of Smartphone Personalization: Measurement and User Variability", in *Behavior & Information Technology*, vol. 31, no.10, October 2012.
28. Jun Yao, Jian Lin, Yanhua Dai, Gedeng Ruan, Zheng Yan, Lei Li, Lin Zhong, Douglas Natelson, and James M. Tour, "Highly transparent nonvolatile resistive memory devices from silicon oxide and graphene," *Nature Communications*, October 2012.
29. Xiaohan Ma, Mian Dong, Lin Zhong, and Zhigang Deng, "Performance and power consumption characterization of 3D mobile games," in *IEEE Computer*, April 2013.
30. Ahmad Rahmati and Lin Zhong, "Studying smartphone usage: lessons from a four-month field study," in *IEEE Transactions on Mobile Computing*, July 2013.
31. Ardalan Amiri Sani, Zhiyong Tan, Peter Washington, Mira Chen, Sharad Agarwal, Lin Zhong, and Ming Zhang, "The wireless data drain of users, apps, and platforms," in *ACM SIGMOBILE Mobile Computing and Communications Review*, October 2013.
32. Christopher Hunter, Lin Zhong, and Ashutosh Sabharwal, "Leveraging physical layer cooperation for energy conservation," in *IEEE Transactions on Vehicular Technology*, January 2014.
33. Ahmad Rahmati, Clayton Shepard, Chad Tossell, Lin Zhong, Phillip Kortum, Angela Nicoara, and Jatinder Singh, "Seamless flow migration on smartphones without network support," in *IEEE Transactions on Mobile Computing*, March 2014.
34. Xiaozhu Lin, Zhen Wang, and Lin Zhong, "K2: A mobile operating system for heterogeneous coherence domains," in *ACM Transactions on Computer Systems*, June 2015.
35. Ahmad Rahmati, Clayton Shepard, Chad Tossell, Lin Zhong, and Phil Kortum, "Practical context awareness: measuring and utilizing the context dependency of mobile usage," in *IEEE Transactions on Mobile Computing*, September 2015.
36. Evan Everett, Clayton Shepard, Lin Zhong, and Ashutosh Sabharwal, "SoftNull: Many-antenna full-duplex wireless via digital beamforming, in IEEE Transactions on Wireless Communications, December 2016.
**ACTIVE FERDERAL GRANTS**
- *National Science Foundation*, SaTC: CORE: Medium: Collaborative: Defending against Compromise and Manipulation of Mobile Communities, $499,999, 2017-2020, Rice PI (Leading PI Ben Y. Zhao).
- *National Science Foundation*, CCSS: Programmable Mixed-Signal Vision Sensor for Continuous Mobile Vision, $350,000, 2016-2019, PI.
- *National Science Foundation*, NeTS: Large: Practical Foundations for Networking with Many-Antenna Base Stations, $2,400,000, 2015-2020, PI (Co-PI: Edward Knightly, Ashutosh Sabharwal and Ness Shroff).
- *National Science Foundation*, NeTS: Large: Collaborative Research: Foundations of Hierarchical Full-duplex Wireless Networks, $1,800,000, 2013-2018, Co-PI (PI: Ashutosh Sabharwal; other Co-PI: Edward Knightly)
PAST FEDERAL GRANTS
• National Science Foundation, CSR: Small: I/O Virtualization at the Device File Boundary and its Applications, $413,000, 2014-2017, PI.
• National Science Foundation, II-NEW: A Reconfigurable Multi-Cell Research Platform for Massive Multiple Input Multiple Output (MIMO) Networks, $1,000,000, 2014-2017, PI (Co-PI: Ashutosh Sabharwal and Edward Knightly)
• National Science Foundation, EARS: Enhanced Spectrum Availability and MU-MIMO Coordination for High Spatial-Spectral Efficiency, $597,125, 2014-2017, Co-PI (PI: Edward Knightly)
• National Science Foundation, MRI: Development of ScaleMed: A Platform for Scalable mHealth Research and Development, $400,000, 2014-2017, Co-PI (PI: Ashutosh Sabharwal; other Co-PI: Ashok Veeraraghavan)
• National Science Foundation, CSR: Small: Per-Process Energy Accounting in Mobile Systems, $440,000, 2012-2015, Leading PI (Other PI: Tian Lan from George Washington University)
• National Science Foundation, CNS- 1054693 CAREER: Reinventing Smartphones for Sensing, $450,000, 2011-2016, PI.
• National Science Foundation, NeTS: Small: LAWN: Scaling Up Cellular Data Networks using a Large Number of Antennas, $450,000, 2012-2015, Leading PI (Other PI: Richard Yang from Yale University)
• National Science Foundation, MRI: Development and Deployment of an Operational and Programmable Diverse-Spectrum Access Network, $2,940,000, 2011-2015, Co-PI (PI: Edward Knightly, other Co-PIs Ashutosh Sabharwal and Will Reed)
• National Science Foundation, NetSE: Large: Urban-Scale Polymorphic Wireless Networks: Community-Driven Assessment, Design, and Access, $1,800,000, 2010-2015, Co-PI (PI: Edward Knightly, other Co-PIs Bob Stein and Will Reed)
• National Science Foundation, CSR: Medium: Collaborative Research: System Solutions for Context-aware and Energy-Efficient Mobile Displays, $700,000, 2011-2014, Leading PI (other PI Massoud Pedram from USC)
• National Science Foundation, NSF/FDA SIR: System Solutions for Improved Image Quality of Context-aware Mobile Displays, $105,000, 2011-2012, PI.
• National Science Foundation, MRI: Development of mobileWARP-A Platform for Next-Generation Wireless Networks and Mobile Applications, $2,000,000, 2009-2013, Co-PI (PI: Sabharwal, other Co-PIs: Aazhang, Cavallaro, Knightly, Dacso)
• National Science Foundation, CRI/IAD: Programmable and At-Scale Infrastructure for Wireless Access, Mobile Computing, and Health Sensing, $1,500,000, 2008-2012, Co-PI (PI: Edward Knightly)
• National Science Foundation, ECCS: Multi-Layer Integrated Resource Management for Mobile Wireless Systems, $350,000, 2009-2012, Co-PI (PI: Cavallaro)
• National Science Foundation, HCC-Medium: Collaborative Research: Understanding and Optimizing Wireless Mobile Computing for Underserved Urban Communities, $692,000, 2008-2011, PI (Co-PIs: Knightly, Crowder, and Kortum)
• National Science Foundation, CSR---HCC: Human Factors in Energy-Efficient Mobile Computing System Design, PI, $492,970, 2007-2011.
• National Science Foundation, CSR---EHS: Coordinated Energy Optimization of Mobile Embedded Systems with User Information, PI, $178,000, 2007-2010.
• National Science Foundation, MRI: Development of Open-access Photonic Networked Sensors (PHOTONS) for Security, Industrial and Environmental Applications, $362,000, 2007-2010, Co-PI (PI: Sabharwal, other Co-PIs: Koushanfar, Tittel, and Wysocki)
• National Science Foundation, NeTS-WN: Collaborative Research: Mesh Networks for Under-Served Urban Communities: Engaging Users and Integrating Mobile Access and Health Sensing, $378,000, 2007-2009, Co-PI (PI: Knightly)
UNIVERSITY SERVICES
• Chair, Faculty Search Committee, 2015 and 2016
• Chair, Computer Engineering Undergraduate Curriculum Committee 2013
• Member, ECE Department Visibility Committee 2005-2006, 2014-Present
• Member, ECE Department Graduate Committee 2006-2007
• Member, ECE Department Undergraduate Committee, 2007-Present
• Member, ECE Department Faculty Search Committee, 2009-2012
• Faculty Associate, Hanszen College 2005-2008
• Division Advisor (Engineering), Jones College, 2008-2012
PROFESSIONAL SERVICES
• Program Committee Co-Chair: WiNTECH 2008, ACM MobiSys 2012, HotPower 2013, IPSN 2014, MCS 2014, ACM APSys 2016.
• Member, Program committees for MobiSys 2007, 2008, 2010-2015, 2017 and 2018; MobiCom 2009, 2012-2015, 2017; HotMobile 2009, 2010, 2012, 2013, and 2015; ASPLOS 2013 and 2014; DSN 2013 and 2014; SigComm 2015; USENIX ATC 2015; SenSys 2013 and 2014; NSDI 2017 and 2018
• Panelist, NSF proposal review panels
PROFESSIONAL SOCIETIES
IEEE (Senior Member); ACM (Senior Member), Society for Information Display (SID)
TEACHING
• ELEC513/COMP513: Complexity in modern systems, Spring 2013-2015, Rice
• ELEC528 Mobile & Embedded System Design and Applications, Spring 2006, Rice
• ELEC518 Energy Efficiency in Modern Systems, Spring 2009-2011, Rice
• ELEC424 Mobile & Embedded System, Fall 2006-2011, 2013, 2014, 2016 Rice
• ELEC527 Nano computing, Spring 2007-2008, Rice (co-taught with Jim Tour)
• ELEC101 Elements of Electrical Engineering, Spring 2010-2011, Rice (co-taught with Ashutosh Sabharwal)
STUDENTS MENTORED
Current
Min Hong Yun PhD student, Rice University
Kevin Boos PhD student, Rice University
Jian Ding PhD student, Rice University
Nisal Menuka PhD student, Rice University
Wenqiu Yu PhD student, Rice University
Namitha Liyanage PhD student, Rice University
Ramla Ijaz PhD student, Rice University
Graduated
Clayton Shepard (Ph.D., 2017) (first job: CTO of Skylark Wireless)
Robert LiKamWa (Ph.D., 2016) (first job: Assistant Professor of EE and Media, Arizona State University)
Ardalan Amiri Sani (Ph.D., 2015) (first job: Assistant Professor of CS, University of California-Irvine)
Hang Yu (Ph.D., 2015) (first job: Schlumberger; now MobiSport)
Xiaozhu Lin (Ph.D., 2014) (first job: Assistant Professor of ECE, Purdue University)
Mian Dong (Ph.D., 2013) (first job: Samsung Research-America; now MobiSport (co-founder))
Ahmad Rahmati (Ph.D., 2012) (first job: Nokia Research; now Apple Inc.)
Jun Yao (Ph.D., 2011, co-advised with James Tour and Douglas Natelson) (first job: Postdoc Harvard; now Assistant Professor of ECE, University of Massachusetts-Amherst)
Jie Liao (M.S., 2016) (first job: Facebook)
Eddie Reyes (M.S., 2015) (first job: Startup)
Chao Xu (M.S., 2014) (first job: Google)
Zhen Wang (M.S., 2013) (first job: Google)
Siqi Zhao (MS, 2013) (first job: Amazon)
Hasan Dumanli (MS, 2009) (first job: Schlumberger)
Jiayang Liu (MS, 2008) (first job: Microsoft) |
A multi-position ratchet mechanism has a retractor holder which receives a retractor stem connected to a retractor. A pawl selectively engages the retractor stem to allow for the retractor stem to be positioned relative to the retractor holder. The retractor holder has a passage and is connected at a base pivot to a base having a toothed hub. The toothed hub extends into the passage and a locking mechanism connected to the retractor holder allows for the selection of a pivot angle of the retractor holder relative to the base. The base rotates relative to a ring holder in a first configuration. In a second configuration a rotation angle between the base and the ring holder is set and locked with a locking actuator. The locking actuator is a cam actuator with drives a cam surface about a locking pivot to allow contact surface between the base and the ring holder to contact one another. A retaining mechanism selectively maintains a ring in a slot in the ring holder.
23 Claims, 6 Drawing Sheets
FIG. 1
FIG. 2
U.S. Patent May 4, 2004 Sheet 3 of 6 US 6,729,205 B2
Diagram of mechanical components labeled with numbers and arrows indicating direction.
Fig 4
Fig. 5
MULTI-POSITION RATCHET MECHANISM
CLAIM OF PRIORITY
This application claims the benefit of U.S. Provisional Patent Application No. 60/279,395 filed Mar. 28, 2001.
FIELD OF THE INVENTION
The present invention relates to a multi-position ratchet mechanism for holding a surgical retractor blade and, more particularly to a quick-release ratchet mechanism which permits the retractor blade to be tilted and rotated into the wound so that the organ may be retracted and lifted at the same time.
BACKGROUND OF THE INVENTION
A surgical retraction systems is described in U.S. Pat. No. 4,254,763. This patent, incorporated by reference, describes the ability to utilize a retractor ring to retract organs in the plane of the ring.
An improvement for use with the ring of U.S. Pat. No. 4,354,763 was disclosed in U.S. Pat. No. 4,424,724. The '724 Patent, also incorporated by reference, discloses a multi-position ratchet mechanism for holding a surgical retractor blade allowing the blade to be pivoted into the wound from a ring and retracted at the same time. This device allowed the retractor to be tilted, or pivoted, into the wound and retracted with one motion.
The ratchet mechanism of the '724 patent only allows for pivoting in a single plane about the ratcheting mechanism perpendicularly to the ring to which the mechanism is secured. A need exists for rotation in other planes as well, preferably in a plane substantially perpendicular to the plane of the operation of the ratcheting mechanism, i.e., the plane of pivoting, or the plane of the ring.
Furthermore, the ratchet mechanism of the '724 patent is not believed to adequately secure the retractor to the ring. Specifically, the ratchet holder relies on tension of the retractor blade to pull the holder forward so that the edge of a dowel pin engages indentations of the ring to prevent the circumferential sliding of the ratchet holder and to prevent the ratchet holder from falling off the ring. However, there are instances when this necessary tension may not be present or adequate and the ratchet holder would have little, if any, resistance to prevent coming off the ring. Furthermore, torque applied to the ratchet holder may also cause problems. Accordingly, a need exists to restrain the ratchet holder from moving away off of the ring.
SUMMARY OF THE INVENTION
The multi-position ratchet mechanism of the preferred embodiment provides a multi-position ratchet mechanism for holding the retractor blade for permitting it to be pivoted into a wound and retracted at the same time which may be thought to simulate the natural “toe-in” method of retraction one could achieve with the hand. The mechanism of the present invention also permits the position of the ratchet mechanism to be moved circumferentially about the support ring and allow for rotation about the selected position relative to the support ring. Furthermore the ratchet mechanism is preferably configured to securely connect with the support ring.
The multi-position ratchet mechanism includes a ring holder, which is mounted onto the support ring, such as an oval support ring. Selective positioning of the ratchet holder relative to the ring may be achieved through the use of a transverse slot in the ratchet holder which receives a thickness of the ring therein. A retention mechanism is utilized to prevent the ring from inadvertently coming out of the transverse slot.
The retention mechanism restrains the ring in the transverse slot of the ring holder. Specifically, the ring typically has a width which terminates proximate to the end transverse slot when the ring is installed in the transverse slot. The retention mechanism is moved into position to prevent movement of the ring out of the transverse slot. In the preferred embodiment, the retention mechanism is pivoted in and out of position with a lever.
The multi-position ratchet also preferably includes the ability to rotate relative to the ring as well as the ability to be secured in a fixed radial orientation relative to the ring upon rotation to a desired angle. Opposing serrated washers which cooperate to resist rotation when engaged maybe utilized. The serrated washers maybe engaged or disengaged utilizing a cam, or otherwise.
BRIEF DESCRIPTION OF THE DRAWINGS
The particular features and advantages of the invention as well as other objects will become apparent from the following description taken in connection with the accompanying drawings in which:
FIG. 1 shows a top elevational view of the multi-position ratchet mechanism of the present invention;
FIG. 2 shows a top elevational view of the multi-position ratchet mechanism of FIG. 1 connected to a retractor stem and a ring, and showing both horizontal and vertical plane rotation; operation of the cam actuator is shown in phantom;
FIG. 3 shows a partially exploded perspective view of the mechanism of FIG. 1;
FIG. 4 shows a top elevational view of the mechanism of FIG. 1 showing vertical plane rotation;
FIG. 5 shows an exploded cutaway top elevational view of an alternative embodiment of a portion of a locking mechanism utilized in a ring holder; and
FIG. 6 shows a side cross sectional view of the alternative embodiment of FIG. 5 in use.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
The multi-position ratchet mechanism 10 of FIGS. 1–4 maybe utilized in place of the mechanisms illustrated in FIG. 1 of U.S. Pat. No. 4,424,724. However, the multi-position ratchet mechanism 10 of the presently preferred embodiment is believed to improve on the basic design of the prior art ratchet mechanism.
The ratchet mechanism 10 has a ring holder 12 connected to a retractor holder 14. The retractor holder 14 may be pivoted up and down to place a retractor blade (not shown) at a desired depth. Furthermore, the ring holder 12 may be rotated relative to the ratchet holder 14. Specifically, the illustrated ring holder 12 rotates about axle 15 shown in FIG. 3. A pair of opposing washers 16, 18 oppose one another, and in a first configuration are spaced from one another, and in a second configuration contact one another as described in more detail below. The washer 16 may be press fit into well 13 as illustrated in FIG. 3. Washer 18 may similarly be connected to base 26.
In FIGS. 1, 2 and 4 the rotational element is located in the first configuration. The washers 16, 18 have serrations 20, 22 on opposing faces (as illustrated in FIG. 3) which cooperate
to prevent the ring holder 12 from rotating relative to the retractor holder 14 when in the second configuration shown in phantom in FIG. 2. In FIGS. 1, 2, and 4 the washers 16, 18 are spaced apart from one another a sufficient distance to allow the retractor holder 14 to rotate about the axle 15 relative to the ring holder 12. A spring 24 may be utilized to preferentially maintain the washers 16, 18 in either the first or the second position. In the preferred embodiment, the washers 16, 18 are biased in the first configuration. When the radial alignment of the retractor holder 14 relative to the ring holder 12 is to be changed, the operator moves the washers 16, 18 apart to the first configuration such as by locating the cam actuator 17 as illustrated in FIGS. 1 and 4, and in solid line in FIG. 2. The retractor 14 may then be rotated relative to the ring holder 12 about the axle 15. The operator may then move the washers 16, 18 into contact with another to allow the serrations 20, 22 to cooperate as illustrated to the second configuration by rotating the cam actuator 17 about clamping pivot 19 as shown in phantom in FIG. 2.
If the spring biases the washers 16, 18 into the second position in other embodiments, it may be desirable to allow for the radial alignment of the retractor holder 14 relative to the ring holder 12 to be selected in another manner. Specifically, the operator may pull the washers 16, 18 apart, and rotate to the desired position, and then allow the spring to pull the washers 16, 18 back together.
The mechanics of the rotation and locking about the axle 15 may best be understood by referring to FIG. 3. The clamping pivot 19 is illustrated as parallel to the axle 15. The axle 15 contains groove 21 which preferably has a groove surface 23 parallel to groove surface 25. The groove surfaces 23, 25 are in planes, perpendicular to the axis of the axle 15. The axis of the axle 15 may also be referred to as a rotation axis. The axle 15 is also connected to or comprised of spring member 24. The base 26 is placed over the axle 15 with groove 21 accessible through slot 27. The cam actuator 17 is then placed so that cam surface 29 and clamping pivot 19 extend into the slot where the cam surface 29 is operable with the groove 21. The clamping pivot 19 is then secured with a retaining member such as Add a pin 11 shown in FIG. 2. As the cam actuator 17 is moved as shown in FIG. 2 (in phantom), the cam surface 29 acts against the groove surface 25 causing the bottom surface 31 to exert a downward force against the slot bottom 33 and to force the washers 16, 18 together in the second, or locked, configuration. Movement of the cam actuator 17 in the opposite direction releases the force. A spring member 24 may assist in separating the washers 16, 18. Accordingly, the spring member 24 biases the rotational element in the first configuration.
In other embodiments, the rotational element may not be biased in either of the first or second configurations, or the rotational element maybe biased in the second configuration. Furthermore, in other embodiments, the rotational element may not have a second configuration, and remain in the first configuration allowing the retractor holder 14 to instantly rotate relative to the ring holder 12 without releasing a locking mechanism.
In the preferred embodiment, one of the opposing washers 16 is connected to the ring holder 12 at a top surface of the ring holder 12 and the other washer 18 is connected to a base 26. The base 26 also preferably includes a toothed hub 28 which assists the retractor holder 14 in being pivotally positioned relative to the ring holder 12. The toothed hub 28 is press fit into the base 26. The base 26 has a base pivot 30 which connects the retractor holder 14 to the base 26 while providing a pivot point for directing the blade of a retractor (not shown) to a specific depth in an opening. The toothed hub 28 fits within passage 37 in the retractor holder 14 and assists in selecting a pivot angle of the retractor holder 14 relative to the base 26.
The retractor holder 14 preferably includes a lock mechanism 32 which is moveable between a locking position and an unlocked position. In the locking position at least one grip 34 engages the toothed hub 28 to provide a desired angular position of pivot of the retractor holder 14 relative to the ring holder 12. An actuation lever 76 may be utilized to pivot the grip 34 in and out of engagement with a portion of the toothed hub 28 about pivot 78. A spring is preferably utilized to bias the locking mechanism 32 in locking engagement with the toothed hub 28 as shown in FIG. 4. The retractor holder 14 also preferably contains a rectangular bore 36 for receiving the stem 38 of a retractor blade as shown in FIGS. 2 and 4. The retractor stem 38 may be held in fixed relation to the retractor holder 14 by the pawl blade 40 of the pawl 42. The pawl 42 is preferably spring biased to engage the retractor stem 38 by a spring. The pawl 42 has an actuation lever 80 which pivots about pivot 82 to engage and disengage the pawl blade 40.
In addition for allowing for rotation of the retractor holder 14 relative to the ring holder 12, the preferred embodiment also employs a ring holder 12 which secures the mechanism 10 to the ring 46 with a retention mechanism 48 as shown in FIGS. 1–3. The retention mechanism 48 is preferably actuated by a lever 50 which is spring biased to retain the ring 46 within a transverse slot 54. Although the preferred embodiment has a retention mechanism 48 attached to a lever 50 for retaining an end 56 of the width of the ring 46 within the transverse slot 54, other locations along the width of the ring 46 could also be retained by a retention mechanism 48. Furthermore, the retention mechanism 48 need not necessarily be equipped with a release lever 50 in all embodiments. The retention mechanism 48 may also include one or more surfaces in the transverse slot 54, such as opposing slot surfaces, which move to secure the ring 46 in the slot 54.
The retention mechanism 48 illustrated rotates about retention pivot 60 and is preferably spring biased in the retaining position shown in FIGS. 1 and 2. Depression of the lever 50 moves the retention mechanism 48 about the retention pivot 60 to allow the ring 46 to be inserted into the transverse slot 54. An arm 62 extending between the retention mechanism 48 and the retention pivot 60 is at least partly contained housed within notch 64 in the retaining position. The spring 84 may be received within a spring bore 66. An access port 68 allows access within the ring holder 12 to secure a pin through the retention pivot 60. Springs for the locking mechanism 32 and pawl mechanism 42 may be similarly located on those structures.
The locking mechanism 32 and the pawl 42 operate in a similar manner as the retention mechanism, however in alternative embodiments, different designs may provide for different operation of these members.
In order to ensure a snug fit for the ring 46 in the transverse slot 54, a spring member 70 is placed at least partially through a bore (obscured from view) in the ring holder 12. A retaining plate 72 keeps the spring member 70 in position so that plunger 74 contacts the inserted ring member 46 and exerts a force so that the ring member 46 is contacted by the plunger 74 and the retention mechanism 48. The spring member 70 is preferably under compression when the ring member 46 is installed to positively retain ring against retention mechanism 48 having lip 86.
FIGS. 5 and 6 show an alternatively preferred retention mechanism 100. In this embodiment, a pivoting arm 102
with a stop 104 retains ring 46 in a transverse slot 106 which is similar to the transverse slot 54 of the preferred embodiment. The arm 102 is moveable about pivot 108 to allow a ring 46 to be inserted into the slot. The pivot 108 is preferably connected to slide 110 which is spring biased by spring 112 to a first end 114 of the ring holder 116. The arm 102 illustrated is biased into a locking position by spring 118. The slide 110 is moveable within housing 120 which is connected, and preferably an integral portion of the ring holder 116.
The slide 110 allows the retention mechanism 100 to accommodate a plurality of ring widths. FIG. 6 shows the operation of the arm 102 in a downward direction to release the ring 46 from the slot 106. A top surface 122 of the arm 102 contacts the ring 46 in compression to retain the ring 46 in the slot 106. Furthermore, the stop 104 retains the ring 46 from sliding out of the slot 106.
Numerous alternations of the structure herein disclosed will suggest themselves to those skilled in the art. However, it is to be understood that the present disclosure relates to the preferred embodiment of the invention which is for purposes of illustration only and not to be construed as a limitation of the invention. All such modifications which do not depart from the spirit of the invention are intended to be included within the scope of the appended claims.
What is claimed is:
1. A multi-position ratchet mechanism comprising:
a ratchet holder having a passage within the ratchet holder and a retractor slot extending through the retractor holder;
a locking mechanism connected to said ratchet holder and having a grip actuation lever extending outside of said passage for ready actuation by a user, said locking mechanism having a grip normally biased into said passage, and operation of the grip actuation lever moving the grip out of the passage;
a base pivotably connected at a base pivot to the ratchet holder about a pivot axis;
a ratchet extending from said base into said passage of said ratchet holder and operatively engaging said grip whereby a pivot angle of the ratchet holder with respect to said base may be selectively varied;
a pawl connected to said ratchet holder and having a pawl actuation lever extending outside of said retractor slot for ready actuation by a user, said locking mechanism having a pawl blade normally biased into said retractor slot, and operation of the pawl actuation lever moving the pawl blade out of the retractor slot; and
a ring holder having a ring slot, said ring holder rotatable relative to the base in a first configuration; and
a plunger spring biased into the ring slot.
2. The multi-position ratchet mechanism of claim 1 wherein the ring holder further comprises a retaining mechanism configured to selectively positively retain a ring within the slot.
3. The multi-position ratchet mechanism of claim 2 wherein the retaining mechanism is connected to an arm and a lever, whereby actuation of the retaining mechanism moves the retaining mechanism into and out of a holding position.
4. The multi-position ratchet mechanism of claim 1 wherein the ring holder has a top surface and the base has a bottom surface, the top surface of the ring holder rotating parallel to the bottom surface of the base in the first configuration, and the ratchet holder pivots about the pivot axis orthogonally to the rotation axis.
5. The multi-position ratchet mechanism of claim 1 further comprising a locking actuator, said locking actuator having a locking position wherein said base and said ring holder are selectively locked at a rotation angle relative to one another in a second configuration.
6. A multi-position ratchet mechanism comprising:
a ratchet holder having a passage within the ratchet holder and a retractor slot extending through the retractor holder;
a locking mechanism connected to said ratchet holder and having a grip actuation lever extending outside of said passage for ready actuation by a user, said locking mechanism having a grip normally biased into said passage, and operation of the grip actuation lever moving the grip out of the passage;
a base pivotably connected at a base pivot to the ratchet holder;
a ratchet extending from said base into said passage of said ratchet holder and operatively engaging said grip whereby a pivot angle of the ratchet holder with respect to said base may be selectively varied;
a pawl connected to said ratchet holder and having a pawl actuation lever extending outside of said retractor slot for ready actuation by a user, said locking mechanism having a pawl blade normally biased into said retractor slot, and operation of the pawl actuation lever moving the pawl blade out of the retractor slot;
a ring holder having a ring slot, said ring holder rotatable relative to the base about a rotation axis in a first configuration; and
a plunger spring biased into the ring slot.
7. The multi-position ratchet mechanism of claim 6 further comprising a retention mechanism located at least partially external to the ring slot, said retention mechanism selectively operable between a lock and unlock position, where when in said lock position, an inserted ring contacts the plunger and the retention member.
8. A multi-position ratchet mechanism comprising:
a retractor holder having a retractor slot extending through a portion of the retractor holder;
a pawl operably coupled to the retractor holder having a retaining and a release position, wherein when said pawl is placed in said retaining position, said pawl retains an inserted retractor stem relative to the retractor holder;
a base pivotably connected to the retractor holder at a base pivot; and
a ring holder having a top surface portion and a slot configured to receive a ring member, said ring holder operably coupled to the base, said base rotatable about a rotation axis relative to the ring holder in a first configuration along the top surface portion of the ring holder, said base fixed at an angular position relative to the ring holder in a second configuration; said retractor holder pivots through an arc in a pivot plane perpendicular to the base pivot, the base rotates through a plane of rotation perpendicular to the rotation axis, and said pivot plane is substantially orthogonal to said plane of rotation in the first configuration.
9. The multi-position ratchet mechanism of claim 8 further comprising a locking mechanism configured to selectively set a pivot angle of the ratchet holder relative to the base.
10. The multi-position ratchet mechanism of claim 8 wherein the base has a planar bottom surface portion and the ring holder has a planar top surface portion, and said planar
bottom surface portion of said base rotates parallel to the planar top surface portion of the ring holder in the first configuration.
11. The multi-position ratchet mechanism of claim 8 further comprising a locking actuator, said locking actuator having a locking position wherein said base and said ring holder are selectively locked at a rotation angle relative to one another in a second configuration.
12. A multi-position ratchet mechanism comprising:
a retractor holder having a retractor slot extending through a portion of the retractor holder;
a pawl operably coupled to the retractor holder having a retaining and a release position, wherein when said pawl is placed in said retaining position, said pawl retains an inserted retractor stem relative to the retractor holder;
a base connected to the retractor holder; and
a ring holder having a top surface portion and a slot configured to receive a ring member, said ring holder operably coupled to the base, said base rotatable about a rotation axis relative to the ring holder in a first configuration along the top surface portion of the ring holder, said base fixed at an angular position relative to the ring holder in a second configuration;
a locking actuator having a locking position wherein said base and said ring holder are selectively locked at a rotation angle relative to one another in a second configuration, said locking actuator further comprising a cam surface rotatable about a locking pivot, and rotation of the cam surface in a direction of rotation about the locking pivot locks the base relative to the ring holder at the rotation angle.
13. A multi-position ratchet mechanism comprising:
a retractor holder having a retractor slot extending through the retractor holder;
a pawl operably coupled to the retractor holder having a retaining and a release position, wherein when said pawl is placed in said retaining position, said pawl retains an inserted retractor stem in a fixed position relative to the retractor holder;
a base pivotably connected to the retractor holder at a base pivot about a pivot axis;
a ring holder having a ring slot configured to receive a ring member, said ring holder connected to the base, said pivot axis parallel to the ring slot; and
a retention mechanism connected to the ring holder and selectively operable to retain and release an inserted ring relative to the ring slot.
14. The multi-position ratchet of claim 13 wherein the ring holder is operably coupled to the base, wherein the base is rotatable relative to the ring holder in a first configuration and held at a fixed rotation angle relative to the ring holder in a second configuration.
15. The multi-position ratchet of claim 13 further comprising a spring member operably coupled to the ring holder having a plunger at least partially located within the ring slot, said plunger and said retainer mechanism cooperating to retain an inserted ring in the ring slot.
16. A multi-position ratchet mechanism comprising:
a retractor holder having a retractor slot extending through the retractor holder;
a pawl operably coupled to the retractor holder having a retaining and a release position, wherein when said pawl is placed in said retaining position, said pawl retains an inserted retractor stem in a fixed position relative to the retractor holder;
a base pivotably connected to the retractor holder at a base pivot about a pivot axis;
a ring holder having a ring slot configured to receive a ring member, said ring holder operably coupled to the base, wherein the base is rotatable relative to the ring holder in a first configuration and held at a fixed rotation angle relative to the ring holder in a second configuration; and
a retention mechanism connected to the ring holder and selectively operable to retain and release an inserted ring relative to the ring slot; and
a locking mechanism selectively operable to select the fixed rotation angle through a range of at least ninety degrees.
17. The multi-position ratchet of claim 16 further comprising an axle coupling the base to the ring holder, said axle extending along a rotation axis with the base rotatable about the rotation axis.
18. The multi-position ratchet of claim 17 further comprising contact surfaces located between the base and the ring holder said contact surfaces separated in a first configuration allowing rotation of the ring holder relative to the base about the axle.
19. The multi-position ratchet of claim 18 wherein the contact surfaces contact one another in a second configuration to prevent rotation of the ring holder relative to the base.
20. The multi-position ratchet of claim 19 further comprising a cam actuator operably coupled to one of the base and ring holder, said cam actuator movable between a locked and an unlocked position wherein when in said locked position, said contact surfaces are moved in contact with one another.
21. The multi-position ratchet of claim 20 further comprising a slot in the base having a slot bottom, and said cam actuator having a cam surface and a bottom surface pivotable about a locking pivot, said slot receiving the cam surface and bottom surface, and said axle further comprising a groove, said groove receiving said cam surface wherein when said cam actuator moves from the first position to the second position, the bottom surface of the cam actuator applies a force against the slot bottom thereby forcing the contact surfaces in contact with one another.
22. The multi-position ratchet of claim 21 wherein said contact surfaces are biased in one of the first and second configurations.
23. A multi-position ratchet mechanism comprising:
a retractor holder having a retractor slot extending through the retractor holder;
a pawl operably coupled to the retractor holder having a retaining and a release position, wherein when said pawl is placed in said retaining position, said pawl retains an inserted retractor stem in a fixed position relative to the retractor holder;
a base pivotably connected to the retractor holder at a base pivot;
a ring holder having a ring slot configured to receive a ring member, said ring holder connected to the base; and
a retention mechanism connected to the ring holder and selectively operable to retain and release an inserted ring relative to the ring slot; and
a toothed hub connected to one of the base and retractor holder, said toothed hub cooperating with a locking mechanism having a grip, said grip operatively engaging the toothed hub whereby a pivot angle of the ratchet holder with respect to said base may be selectively varied. |
Public Service Act 1974
ALTERATION OF DAY APPOINTED A PUBLIC HOLIDAY IN THE SHIRES OF BACCHUS MARSH AND GISBORNE
PROCLAMATION
By His Excellency the Governor of the State of Victoria and its Dependencies in the Commonwealth of Australia, &c., &c., &c.
Whereas by sub-section (1) of section 71 of the Public Service Act 1974, the fourth Thursday in September shall be observed as a public holiday throughout the municipalities mentioned in the Fifth Schedule to the said Act, including the Shires of Bacchus Marsh and Gisborne.
And whereas it is made to appear to me expedient that the said day should not be a public holiday throughout the Shires of Bacchus Marsh and Gisborne.
Now therefore I, the Governor of the State of Victoria, in the Commonwealth of Australia, &c., &c., and with the advice of the Executive Council of the said State, and in pursuance of the provisions of sub-section (3) of section 71 of the Public Service Act 1974, do by this my Proclamation declare that the said day should not be a public holiday throughout the Shires of Bacchus Marsh and Gisborne, and appoint Monday, 21 September 1981, to be a public holiday throughout the Shires of Bacchus Marsh and Gisborne.
Given under my Hand and the Seal of the State of Victoria, aforesaid at Melbourne, this eleventh day of August in the year of our Lord, One thousand nine hundred and eighty-one, and in the thirtieth year of the reign of Her Majesty Elizabeth the Second, Queen of Australia.
(L.s.) HENRY WINNEKE
By His Excellency's Command
L. H. S. THOMPSON
Premier
GOD SAVE THE QUEEN!
PUBLIC HOLIDAYS
PROCLAMATION
By His Excellency the Governor of the State of Victoria and its Dependencies in the Commonwealth of Australia, &c., &c., &c.
In pursuance of the provisions contained in section 71 (2) of the Public Service Act 1974, I, the Governor of the State of Victoria, in the Commonwealth of Australia, by and with the advice of the Executive Council of the said State, do by this my Proclamation appoint the days and dates hereunder mentioned to be observed as Public Holidays or Public Half-Holidays (as the case may be) at the places respectively specified, viz.:
Public Holidays
MONDAY, 21 SEPTEMBER 1981, throughout the Shire of Newham and Woodend.
THURSDAY, 8 OCTOBER 1981, throughout the Shire of Warracknabeal.
WEDNESDAY, 14 OCTOBER 1981, throughout that part of the Shire of Walpeup westerly from the Township of Linga to the Shire Boundary.
Public Half-Holidays from the Hour of Twelve o'clock noon
TUESDAY, 1 SEPTEMBER 1981, throughout the North Riding of the Shire of Arapiles.
WEDNESDAY, 9 SEPTEMBER 1981, throughout the Shire of Seymour.
THURSDAY, 1 OCTOBER 1981, throughout the City of Horsham.
TUESDAY, 6 OCTOBER 1981, throughout the Borough of Kerang.
TUESDAY, 3 NOVEMBER 1981, throughout the Township of Wedderburn within the Shire of Korong.
THURSDAY, 25 FEBRUARY 1982, throughout the Shire of Seymour.
Given under my Hand and the Seal of the State of Victoria aforesaid, at Melbourne, this eleventh day of August in the year of our Lord One thousand nine hundred and eighty-one, and in the thirtieth year of the reign of Her Majesty Elizabeth the Second, Queen of Australia.
(L.s.) HENRY WINNEKE
By His Excellency's Command
L. H. S. THOMPSON
Premier
GOD SAVE THE QUEEN!
BANK HOLIDAYS
PROCLAMATION
By His Excellency the Governor of the State of Victoria and its Dependencies in the Commonwealth of Australia, &c., &c., &c.
In pursuance of the provisions of the Bank Holidays Act 1958, I, the Governor of the State of Victoria, in the Commonwealth of Australia, do by this my Proclamation appoint the days and dates named hereunder special days to be observed as Bank Holidays or Bank Half-Holidays (as the case may be) at the places respectively mentioned, that is to say:
Bank Holidays
THURSDAY, 24 SEPTEMBER 1981, at Warburton and Yarra Junction within the Shire of Upper Yarra.
TUESDAY, 3 NOVEMBER 1981, at Warburton and Yarra Junction within the Shire of Upper Yarra.
No. 85—40460/81—Price 60 cents, delivered $1.10. Subscription rate: $72 per annum.
LOCAL GOVERNMENT DEPARTMENT
ORDER CONFIRMED—CITY OF CHELSEA
The Minister of the Crown administering the Local Government Act 1958 on 14 August 1981 confirmed the Order hereinafter referred to in pursuance of section 514 of the said Act namely:
An Order of the Council of the City of Chelsea made on 6 April 1981 directing the compulsory taking of the land described hereunder for road purposes.
All that piece of land being part of the land described in Certificate of Title Volume 470 Folio 914 commencing at the north-western corner of Lot 3, section F, Titles Office Lodged Plan No. 5211; thence by a line bearing 65 deg. 54 min. for 6-93 metres; thence by a line bearing 164 deg. 41 min. for 6-17 metres; thence by a line bearing 246 deg. 49 min. for 0-94 metres; thence by a line bearing 155 deg. 54 min. for 9-14 metres; thence by a line bearing 245 deg. 54 min. for 3-04 metres and thence by a line bearing 335 deg. 54 min. for 15-24 metres to the point of commencement.
L. S. LIEBERMAN
Minister for Local Government
Local Government Department
Melbourne (81/2623)
Forrests Act 1958
NOTICE OF INTENTION TO DEDICATE AREAS OF CROWN LAND AS RESERVED FOREST
In accordance with the provisions of Section 46 of the Forrests Act 1958 notice is hereby given that it is intended to recommend to His Excellency the Governor in Council that pursuant to Section 45 of the said Act, the area of Crown Land described in the schedule hereunder be dedicated as reserved forest.
SCHEDULE
47170 hectares more or less of unoccupied Crown Land in the Parishes of Bindi, Eucambene, Gelantipy West, Glenmore, Marroo, Mellieha, Murrindal West, Nappa, Nowong, Tung and Wallapoot as shown hatched on plan No. Vic. F.C. 13 lodged in the Central Plan Office. (A copy of this plan can be inspected at the Forests Commission Offices at Mansfield, Nowa Nowa and Swifts Creek.)
T. L. AUSTIN
Minister of Forests
Note—This notice was published first in the Government Gazette of 12 August 1981.
Forrests Act 1958
NOTICE OF INTENTION TO DEDICATE AREAS OF CROWN LAND AS RESERVED FOREST
In accordance with the provisions of Section 45 of the Forrests Act 1958 notice is hereby given that it is intended to recommend to His Excellency the Governor in Council that pursuant to Section 45 of the said Act, the area of Crown Land described in the schedule hereunder be dedicated as reserved forest.
SCHEDULE
46500 hectares more or less of unoccupied Crown Land in the Parishes of Baw Baw, Binnum, Bullung, Matlock, Moolpah, Stander, Telbit and Toombon as shown hatched on plan No. Vic. F.C. 10 lodged in the Central Plan Office. (A copy of this plan can be inspected at the Forests Commission Offices at Erica and Neerim South.)
T. L. AUSTIN
Minister of Forests
Note—This notice was published first in the Government Gazette of 12 August 1981.
Transport Regulation Act
TRANSPORT REGULATION BOARD
HEARING OF APPLICATION
Notice is hereby given that the following joint application by Eldred Coach Tours Pty. Ltd. and M. A. & J. E. Kneebone of Lang Lang to operate a bus service between Yarram and Melbourne to replace VicRail passenger services on the Yarram—Langwarrin railway line will be heard at the Korumburra Court House, Bridge Street, Korumburra on 3 September 1981 commencing at 10.30 a.m.
Application to license four commercial passenger vehicles, each with seating capacity for 45 persons, to operate a bus service between Yarram and Melbourne as detailed below:
**ROUTE (i)—From Yarram Railway Station via Yarram Street, Commercial Road and South Gippsland Highway to Alberton (Town Stop) thence along the South Gippsland Highway to Welshpool (Town Stop), (Town Stop) return to Foster Street, (Town Stop) via Toora Road, then via Main Street, Bridge Road, Hoddle Road to Fish Creek (Town Stop) and Buffalo (Town Stop) thence to Meeniyang (Town Stop) thence to Leongatha via Langwarrin Gippsland Highway, Langwarrin Road and Long Street to Leongatha Railway Station thence via Anderson Street and the South Gippsland Highway to Korumburra Railway Station via Main Road, Bridge Street and Station Street, thence along the South Gippsland Highway to Loch (Town Stop) thence to Lang Lang (Town Stop) thence to Lang Lang (Town Stop) thence to Main Street to Lang Lang (Town Stop) thence along the South Gippsland Highway to Koo Wee Rup (Town Stop) via Rossitors Road, Gardner Street, Moodie Street, returning to the South Gippsland Highway via Main Road, Bridge Street, Rossitors Road thence to Tooradin (Town Stop), thence along the South Gippsland Highway to Dandenong Railway Station, thence via Foster Street, Clow Street, Short Road, Mulgrave Freeway, Mulgrave Road, Waverley Road, Mulgrave Road, Tooronga Road, South Eastern Freeway, Alexandra Avenue, City Road, Clarendon Street, Spencer Street to Spencer Street Railway Station.
**ROUTE (ii) EXPRESS Bus—From Yarram Railway Station via Yarram Street, Commercial Road, and South Gippsland Highway to Alberton (Town Stop) thence along the South Gippsland Highway to Welshpool (Town Stop), Toora (Town Stop), Foster (Town Stop) and Meeniyang (Town Stop) thence via Kenna-warra Road and Long Street to Leongatha Railway Station thence via Anderson Street and the South Gippsland Highway to Korumburra Railway Station via Main Road, Bridge Street and Station Street thence along Station Street, Bridge Street, Commercial Road and then express along the South Gippsland Highway, South Gippsland Freeway, Mulgrave Freeway, Waverley Road, Waverley Road, Mulgrave Road, Tooronga Road, South Eastern Freeway, Alexandra Avenue, City Road, Maffra Street, City Road, Clarendon Street, Spencer Street to Spencer Street Railway Station.
**NOTE—(a) Passengers to and from intermediate locations along the routes between towns may be picked up and set down at any point between Yarram and Tooradin at the discretion of the coach driver.
(b) On express services no passengers will be picked up or set down between Melbourne and Korumburra.
(c) After 1.00 p.m. on Saturday and all day Sunday coaches will stop to pick up and set down passengers at Cranbourne.
**PARCELS**
The following arrangements shall apply in relation to carriage of parcels:
(a) Parcels and VicRail correspondence to be conveyed between Spencer Street and/or Dandenong Railway Station and Railway Stations at Korumburra, Leongatha and Yarram, also Toora township.
(b) Outside business hours at any of the above locations inwards parcels are to be placed under security in a locked area.
(c) Parcels to be conveyed shall not exceed 40 kilograms mass and/or one cubic metre in size.
**TIMETABLE**
| | Mon.—Fri. | Mon.—Sat. | Mon.—Mon. | Fri. | Sat. | Fri. | Sun. | Sun. |
|----------------|-----------|-----------|-----------|------|------|------|------|------|
| | a.m. | a.m. | a.m. | p.m. | p.m. | p.m. | p.m. | p.m. |
| Yarram | dep. | ... | 7.05 | ... | ... | ... | 3.55 | ... |
| Fish Creek | dep. | ... | 7.25 | ... | ... | ... | ... | ... |
| Leongatha | dep. | 6.10 | 6.15 | 8.25 | 8.26 | 3.10 | 3.25 | 5.00 | 5.30 | 4.20 |
| Korumburra | dep. | 6.27 | 6.32 | 8.45 | 8.45 | 3.28 | 3.43 | 4.17 | 5.48 | 4.38 |
| Dandenong | dep. | 6.44 | 6.49 | 8.59 | 8.59 | 3.44 | 3.59 | 4.34 | 5.04 | 4.44 |
| Melbourne | arr. | 9.05 | 8.50 | 10.40| 11.15| 5.35 | 6.05 | 8.10 | 8.20 | 6.30 |
| | Mon.—Sat. | Mon.—Fri. | Mon.—Sat. | Fri. | Sat. | Fri. | Sun. | Sun. |
| | a.m. | p.m. | p.m. | p.m. | p.m. | p.m. | p.m. | p.m. |
| Melbourne | dep. | 9.50 | 12.25 | 3.35 | 4.05 | 6.30 | 6.50 | 7.10 | 8.55 |
| Dandenong | dep. | 10.15 | 1.25 | 4.40 | 5.10 | 7.23 | 7.43 | 8.00 | 9.45 |
| Korumburra | dep. | 11.54 | 2.44 | 6.10 | 6.44 | 8.44 | 9.22 | 10.42|
| | p.m. | | | | | | | |
| Leongatha | arr. | 12.10 | 3.00 | 6.25 | ... | 8.58 | ... | 11.00|
| | Mon.—Thur. | Fri. and Sat. |
| | dep. | 6.30 | 6.27 | 9.02 | 9.00 | 9.38 |
| Fish Creek | dep. | 7.05 | 9.35 | 10.10 | 11.10|
| Yarram | arr. | 7.30 | 10.23 | 11.10 | |
**FARES**
As prescribed by the Victorian Railways Board.
Notice of any objections should be forwarded to the Secretary of the Board not later than 25 August 1981.
19 August 1981
C. J. V. SMITH, Secretary
Corner Lygon and Princes Streets, Carlton, 3053, Wednesday, 19 August 1981
Transport Regulation Act
TRANSPORT REGULATION BOARD
HEARING OF APPLICATIONS
Notice is hereby given that the following applications will be considered by the Transport Regulation Board on 8 September 1981.
MURRAY, G. C., Balwyn. Application to license one commercial passenger vehicle with seating capacity for 10 persons, to be purchased, to operate as a Touring Omnibus for the carriage of parties on specialised walking tours, rock climbing and several outdoor activities throughout the State of Victoria to forest, mountain and river areas.
Duration of Tours—1 to 8 days.
Note—Vehicle operations shall only be over main or two wheel drive access roads.
McAULEY, S. & H. (On Behalf of Detector & Industrial Services, Belmont). Application to licence one commercial passenger vehicle with seating capacity for 9 persons to operate for the carriage of the applicants' clients, free of charge, from Geelong to various Victorian Goldfield areas and return.
STORY, N. I. & V. M., Bendigo. Application to licence one commercial passenger vehicle with seating capacity for 28 persons, to be purchased, to operate as an Urban Special Service Omnibus from within a 20-km pick up radius of Bendigo.
Stroax, N. I. & V. M., Bendigo. Application to license one commercial passenger vehicle with large seating capacity to be purchased, to operate: (a) Under the same terms and conditions as licence T.S.1780 in the name of the applicants. (b) As an Urban Special Service Omnibus from within a 20-km pick up radius of Bendigo.
Note—This application is subject to the cancellation of licence T.S.1780.
Application for renewal of licences as shown by the persons listed hereunder to operate under the same terms and conditions.
L. C. Dyson's Bus Services Pty. Ltd., Reservoir; M.O.4, M.O.84, M.O.103, M.O.150, M.O.162, M.O.189, M.O.270, M.O.309, M.O.351, M.O.362, M.O.409, M.O.410, M.O.411, M.O.518, M.O.681, M.O.687, M.O.1038, M.O.1040.
O'Berer, J., Penola; T.S.1493.
Notice of any objections should be forwarded to reach the Secretary of the Board not later than 2 September 1981.
Applicants are advised that it will not be necessary to appear on the hearing date specified, unless advised in writing by the Board.
C. J. V. Smith, Secretary
Corner Lygon and Princes Streets, Carlton, 3053, Wednesday, 19 August 1981
NOTICE TO MARINERS
[No. 29 of 1981]
AUSTRALIA—VICTORIA
MALLACOTA
Lights Established
Reference Position—Mallacotta Inlet Light. (Lat. 37 deg. 33-7 min. S., Long. 149 deg. 45-7 min. E. approx.) Lights have been established in the following positions bearings and distances from reference position:
(a) Captain's Point; 015 deg. 29 min.; 72 metres (0-04 miles).
(b) Mallacoota Strait (South); 333 deg. 31 min.; 3938 metres (2-13 miles).
(c) Mallacoota Strait (North); 329 deg. 20 min.; 5264 metres (2-84 miles).
(d) Genoa River; 322 deg. 38 min.; 9928 metres (5-36 miles).
Characteristics—(a) Flash Red every 6 seconds. Flash 1-0 sec.: Eclipse 5-0 sec.
(b) Flash White; Green every 5 seconds. Flash 1-0 sec.: Eclipse 4-0 sec. White from 115 degrees through North to 295 deg.; Green elsewhere.
(c) Flash White, Red every 3 seconds. Flash 0-5 sec.: Eclipse 2-5 sec. White from 134 degrees through South to 314 degrees; Red elsewhere.
(d) Flash white every 4 seconds. Flash 0-5 sec.: Eclipse 3-5 sec.
Elevation—(a) (d) 5 metres. (b) (c) 4 metres.
Visibility—(a) (b) (c) (d) 3 miles.
Abridged Description—(a) Fl. R. 6s. 5m. 3M. (b) Fl. W.G. 5s. 4m. 3M. (c) Fl. W.R. 3s. 4m. 3M. (d) Fl. 4s. 5m 3M.
Charts Affected—AUS 806.
Publication Affected—Sailing Directions, Victoria 1970, pages 562, 563 and 675.
A. J. WaggleN
Port Officer
Public Works Department
Ports and Harbours Division
168 Exhibition Street
Melbourne, 3000, 7 August 1981
Country Fire Authority
PERMISSION TO HOLD FIRE BRIGADE DEMONSTRATIONS
In pursuance of the provisions of section 103 of the Country Fire Authority Act 1958, the Country Fire Authority has granted permission for the holding of fire brigade demonstrations as under:
URBAN FIRE BRIGADES
At Castlemaine on Sunday, 18 October 1981.
At Tongala on Sunday, 1 November 1981.
At Geelong on Sunday, 1 November 1981.
At Queenscliff on Sunday, 1 November 1981.
At Boronia on Saturday, 7 November 1981.
At Bacchus Marsh on Sunday, 8 November 1981.
At Shepparton on Sunday, 8 November 1981.
At Sunbury on Saturday, 14 November 1981.
At Swan Hill on Sunday, 15 November 1981.
At Wendouree on Saturday, 15 November 1981.
At Berri on Sunday, 15 November 1981.
At Tullamarine on Saturday and Sunday, 21 and 22 November 1981.
At Dunolly on Sunday, 22 November 1981.
At Marysville on Saturday, 28 November 1981.
At Wodonga on Sunday, 29 November 1981.
At Geelong on Sunday, 29 November 1981.
At Nyah West on Sunday, 29 November 1981.
At Cohuna on Sunday, 4 December 1981.
At Portarlington on Saturday, 5 December 1981.
At Pakenham on Saturday, 5 December 1981.
At Creswick on Sunday, 6 December 1981.
At Kyneton on Sunday, 6 December 1981.
At Maffra on Saturday, 12 December 1981.
At Portland on Saturday, 12 December 1981.
At Portarlington on Sunday, 13 December 1981.
At Pyramid Hill on Sunday, 13 December 1981.
At Goulburn on Saturday, 15 December 1981.
At Casterton on Sunday, 10 January 1982.
At Bellarat on Saturday, 16 January 1982.
At Echuca on Sunday, 17 January 1982.
At Shepparton on Saturday, 23 January 1982.
At Bairnsdale on Saturday, Sunday and Monday, 30 and 31 January and 1 February 1982.
At Wangaratta on Sunday, 31 January 1982.
At Mildura on Sunday, 31 January 1982.
At Kyneton on Sunday, 31 January 1982.
At Boort on Monday, 1 February 1982.
At Pakenham on Friday, 5 February 1982.
At Kyabram on Friday, 5 February 1982.
At Maryvale on Saturday, 6 February 1982.
At Bacchus Marsh on Saturday, 6 February 1982.
At Belgrave on Sunday, 7 February 1982.
At Mooroopna on Sunday, 7 February 1982.
At Warrnambool on Sunday, 7 February 1982.
At Drysdale on Sunday, 7 February 1982.
At Arthurs Creek on Saturday, 13 February 1982.
At Dandenong on Saturday, 13 February 1982.
At Geelong on Saturday, 13 February 1982.
At Gisborne on Sunday, 14 February 1982.
At Shepparton on Sunday, 14 February 1982.
At Ballarat on Sunday, 14 February 1982.
At Sebastopol on Sunday, 14 February 1982.
At Traralgon on Saturday and Sunday, 20 and 21 February 1982.
At Springvale on Saturday, 27 February 1982.
At Echuca on Sunday, 28 February 1982.
At Cobden on Sunday, 28 February 1982.
At Warrnambool on Saturday, Sunday and Monday, 6, 7 and 8 March 1982.
At Heathcote on Sunday, 21 March 1982.
At Chelsea on Saturday, 3 April 1982.
RURAL FIRE BRIGADES
At Whittlesea on Sunday, 8 November 1981.
At Newbridge on Sunday, 8 November 1981.
14 August 1981
L. T. D'Arcy
Secretary
Milk Board Act 1958
SCHEDULE XXI—REGULATIONS
CONSIDERATION OF APPLICATION FOR MILK CARRIER'S LICENCE
Notice is hereby given that the application made by the person named hereunder for a Milk Carrier's Licence to operate upon the route and in the area set out opposite the name of the applicant will be considered by the Victorian Dairy Industry Authority at its offices at Denison Avenue, Hawthorn, on Wednesday, 26 August 1981, commencing at 10.00 a.m.:
| Name and Address of Applicant | Route and Area |
|-------------------------------|----------------|
| De Cicco Industries Pty. Ltd. (trading as Siracusa Cheese), 70 Dawson Street, Coburg North. | From dairy farms in Gippsland to the Company's premises at Coburg North. |
12 August 1981
BRIAN BUXTON, Secretary
Victorian Dairy Industry Authority
Education Act 1958
NOTICE OF THE MAKING OF AN ORDER UNDER SECTION 13 (1) OF THE ACT
Pursuant to section 13 of the Education Act 1958 I hereby give notice that an Order of the Governor in Council was made on 11 August 1981 under sub-section (1) of the said Act constituting a council for the Cranbourne Meadows Technical School.
A. J. HUNT
Minister of Education
Education Act 1958
NOTICE OF THE MAKING OF AN ORDER UNDER SECTION 13 (4) OF THE ACT
Pursuant to section 13 of the Education Act 1958 I hereby give notice that an Order of the Governor in Council was made on 11 August 1981 under subsection (4) of the said Act that certain provisions relating to the membership of the Narracan Primary School Council be amended.
A. J. HUNT
Minister of Education
COMPANIES ACT 1961
Notice is hereby given in pursuance of sections 308 (2) and 308 (3) of the Companies Act 1961 that at the expiration of three months from the date hereof the names of the following Companies will, unless cause is shown to the contrary, be struck off the Register and the said Companies will be dissolved.
Dated 6 August 1981
E. B. MITCHAM
Deputy Commissioner for Corporate Affairs
Corporate Affairs Office
Melbourne
Companies Above Referred To
| Name of Company | Number of Registration |
|-----------------|------------------------|
| Hastings Deering (Commercial Vehicles) Pty. Ltd. | C38326V |
| Mirboor Service Station Pty. Ltd. | C10085L |
| Lloyd's Business Service Pty. Ltd. | C15058W |
| Tsurc Pty. Ltd. | C16260A |
| Blackshaw Woollen Mills Pty. Ltd. | C23363R |
| Atherton Films Pty. Ltd. | C31292R |
| Garelath Pty. Ltd. | C34521P |
| Balcombe Auto Supplies Pty. Ltd. | C35764C |
| Fyn-Art Jewellery Company Pty. Ltd. | C36091R |
| Chris Building Company Pty. Ltd. | C36222R |
| Holsher Investments Pty. Ltd. | C37278C |
| R. J. Thornton Pty. Ltd. | C38369U |
| A.P.B. Constructions Pty. Ltd. | C39329F |
| Bellamy Estates Pty. Ltd. | C40006Y |
| Thornton Road Constructions Pty. Ltd. | C40257C |
| Charles Trainor and Sons Pty. Ltd. | C41636Y |
| Christchurch Finance Corporation Pty. Ltd. | C41903G |
| M.F.P. Nominees Pty. Ltd. | C45487E |
| Lancelot Investments Pty. Ltd. | C46663P |
| Alfand Investments Pty. Ltd. | C51013P |
| Kenmax Traders Pty. Ltd. | C51485P |
| Proton Pty. Ltd. | C57874D |
| J. Hall Pty. Ltd. | C58696H |
| Peter Wright Motors Pty. Ltd. | C60173U |
| J. & J.W. Pty. Ltd. | C62308H |
| Edensmith Pty. Ltd. | C63691R |
| Australian Water Service Pty. Ltd. | C69337D |
| Milton Hammond Motors Pty. Ltd. | C67242R |
| The McEwan's Group Construction Co. Pty. Ltd. | C71056H |
| Universal Stainless Steel of Australia Pty. Ltd. | C73852H |
| R. A. Dickson Pty. Ltd. | C74861S |
| Westfield Agencies Pty. Ltd. | C75115 |
| Jarrett Coles Pty. Ltd. | C75408Y |
| Mullard Aviation Pty. Ltd. | C77557H |
| Roy Davkin Motors Pty. Ltd. | C78123A |
| H. & M. Furniture Pty. Ltd. | C80276R |
| Zey Pty. Ltd. | C81345U |
| L.H.L. Investments Pty. Ltd. | C82757L |
| Corporate Resources Australia Pty. Ltd. | C83447F |
| A. J. Inglis Pty. Ltd. | C84580A |
| Galway Investments Pty. Ltd. | C85035 |
| Seaview Hill Nominees Pty. Ltd. | C89222Y |
| Customline Homes Pty. Ltd. | C89799N |
| Artanale Pty. Ltd. | C90256R |
| Sedopitna Securities Pty. Ltd. | C91107Y |
| Westgate Pty. Ltd. | C91709K |
| Burns Foundries Pty. Ltd. | C91874C |
| Burns Smelters Pty. Ltd. | C91875E |
| Jondel Pty. Ltd. | C92803R |
| Dow White Pty. Ltd. | C92331N |
| Gibb Holdings Pty. Ltd. | C96621G |
| John Kontek & Son (Vic.) Pty. Ltd. | C97674M |
| Tannin Family Holdings Pty. Ltd. | C97869P |
| Scaffolding Services Pty. Ltd. | C98595Y |
| M.B. Plastic Containers Pty. Ltd. | C99950D |
| Connewarre Holdings Pty. Ltd. | C100961X |
| McKenzie Court Pty. Ltd. | C102166E |
| Woodworking and Engineering Supplies Pty. Ltd. | C104527F |
| Frisina's Properties Pty. Ltd. | C109616U |
| Select Towing Service Pty. Ltd. | C109826C |
| Webley Pty. Ltd. | C109936S |
| Mason's Realty Pty. Ltd. | C11386B |
| Vanzella Investments Pty. Ltd. | C11654D |
| Gilardi Trading Pty. Ltd. | C11810Z |
| M. & F. Trading Pty. Ltd. | C12887N |
| Dastac Nominees Pty. Ltd. | C13383K |
| Bruce S. Robertson & Associates Pty. Ltd. | C14376W |
| Valjaj Properties Pty. Ltd. | C14774K |
| Bangalow Nominees Pty. Ltd. | C15326Y |
| Lucian Nominees Pty. Ltd. | C15797J |
| Stonily Holdings Pty. Ltd. | C15844S |
| Victorian Vibratory Vibrations Pty. Ltd. | C17031V |
| Pigato Investments Pty. Ltd. | C17522E |
| Emerald Developments Pty. Ltd. | C17528T |
| Sigal Industries (Victoria) Pty. Ltd. | C18025S |
| Riepin Lenik (Autoguard) Pty. Ltd. | C18486K |
| Shira Nominees Pty. Ltd. | C19192L |
| Gordon Terrace Developments Pty. Ltd. | C19979Y |
| Laharum Nominees Pty. Ltd. | C20300S |
| Dalby Nominees Pty. Ltd. | C20302S |
| Reeves Manufacturing Company Pty. Ltd. | C31213R |
| Fairwatch Nominees Pty. Ltd. | C31859Z |
| Crumby Pty. Ltd. | C33704G |
| Pacific International Air Charter Pty. Ltd. | C38030U |
| Roktus Nominees Pty. Ltd. | C36451W |
| Coef Finance Corporation Pty. Ltd. | C40298J |
| South Eastern Forests Ltd. | C40911D |
| Beel Specialised Services Pty. Ltd. | C41408X |
| J. & M. Davies Enterprises Pty. Ltd. | C45808Y |
| Murray Enterprises Pty. Ltd. | C51139T |
| Etondale Pty. Ltd. | C51888F |
| Musicimport Pty. Ltd. | C53162N |
Private Agents Act 1966
NOTICE OF RECEIPT OF APPLICATIONS FOR LICENCES UNDER THE PROVISIONS OF THE PRIVATE AGENTS ACT 1966
The Clerk(s) of the Magistrates' Court(s) as listed hereunder hereby give notice that applications, as under, have been lodged for hearing by the said Court(s) on the date(s) specified.
Any person desiring to object to any of such applications must—
(a) lodge with the respective Clerk of the Magistrates' Court a notice in the prescribed form of his objection and of the grounds thereof; or
(b) cause a copy of such notice to be served personally or by post upon the applicant at least three days before the hearing of the application; and
(c) send or deliver—
(i) where the objection is not made by the officer in charge of the police district in which the Court is situated—a copy of the notice to such officer; and
(ii) where the objection is not made by the Registrar or Deputy Registrar—a copy to the Registrar.
| Full Name of Applicant* | Place of Abode of Applicant or Nominee | Name of Firm or Corporation | Address for Registration | Type of Licence | Date of Hearing |
|-------------------------|----------------------------------------|----------------------------|--------------------------|----------------|----------------|
| Kenyon, John Arthur | 4 Osbourn Ave, Kangaroo Flat | Castlemaine and District Armoured Car and Security Services | Abel St, Bendigo | Watchman | 21.8.81 |
| | | | | | |
| | Dated at Bendigo 7 August 1981 | | | | |
| | R. HOLLIS, Clerk of the Magistrates' Court | | | | |
| Crowley, Christopher | 49 Princes St, Flemington | Mayne Nickless | 88 Bakers Rd, North Coburg | Watchman | 7.9.81 |
| | | | | | |
| | Dated at Coburg 5 August 1981 | | | | |
| | P. WENDEN, Clerk of the Magistrates' Court | | | | |
| Bounds, Jeffrey | 1 Laughlin Ave, Nunawading | Armaguard | 88 Bakers Rd, North Coburg | Watchman | 7.9.81 |
| | | | | | |
| | Dated at Coburg 7 August 1981 | | | | |
| | P. WENDEN, Clerk of the Magistrates' Court | | | | |
| Baker, Frederick, Arthur| 1/20 Sandown Rd, Ascot Vale | T.N.T. Group 4 Total Security Pty. Ltd. | 213 Arden St, North Melbourne | Watchman | 15.9.81 |
| Libertore, John Antonio | 408 Albion St, West Brunswick | " | " | " | " |
| Kneale, John Robert | 39 Ninevah Cres, Mulgrave | " | " | " | " |
| Kalos, George | 12 Little St, Port Melbourne | " | " | " | " |
| Lincoln, William John | 6/19 Kell Crt, Greensborough | Transurety Ltd. | Cnr. Arden and Lothian Sts, North Melbourne | " | " |
| Carrell, Richard John | 1/7 Moray St, Fawkner | " | " | " | " |
| Splatt, Mark Anthony | 69 William St, Newmarket | " | " | " | " |
| King, David Verne | 9 Benghazi Ave, Ashburton | T.N.T. Group 4 Total Security Pty. Ltd. | 213 Arden St, North Melbourne | " | " |
| Salmon, Cyril Alfred | 21 Addison St, Elwood | " | " | " | " |
| O'Connor, Patrick | 587 Victoria St, West Brunswick | " | " | " | " |
| Cizmazia, Julius | 21 Harrigan Ave, Clayton | " | " | " | " |
| Buller, John Frederick | 1 Eva Pl, Epping | Transurety Ltd. | Cnr. Arden and Lothian Sts, North Melbourne | " | " |
| | Dated at Melbourne 5 August 1981 | | | | |
*Or in the case of a firm or corporation, of the Nominee
| Full Name of Applicant* | Place of Abode of Applicant or Nominee | Name of Firm or Corporation | Address for Registration | Type of Licence | Date of Hearing |
|------------------------|--------------------------------------|----------------------------|--------------------------|----------------|----------------|
| Elliott, Clifford William | 81 Austin Cres, Pascoe Vale | Armagnard | 88 Bakers Rd., North Coburg | Watchman | 7.9.81 |
| Wierda, John Willen | 13 Wallowa Cres, Frankston | County Security Services | 67 Wallace St., Cranbourne | Provisional Watchmans | " |
| Pratt, Michael Kenneth | 4 Yonde Crt, Greensborough | " | 140 Byrne St, North Fitzroy | Process Server | 31.8.81 |
| Hansen, Phillip Charles | Lot 13 McArthur Rd, Upper Beaconsfield | " | 80 Paisley St, Footscray | Inquiry Agent | 2.9.81 |
| White, Ian Francis | 217 Union Rd, North Albury | Mayne Nickless | 94 York St, South Melbourne | Watchman | 1.9.81 |
| Squire, Christopher Leslie | 1/494 Murphy St, Lavington, N.S.W. | " | 94 York St, South Melbourne | Watchman | 1.9.81 |
| Wiltshire, Keith Charles | High St, Barnawatha | Mayne Nickless | 94 York St, South Melbourne | Watchman | 8.9.81 |
| Eastmure, Joseph William | 11 West Hill Dr, Mt Evelyn | " | 11 West Hill Dr, Mt Evelyn | Guard Agent | 10.9.81 |
* Or in the case of a firm or corporation, of the Nominee
SITTINGS OF THE SUPREME COURT FOR THE HEARING OF CRIMINAL TRIALS IN MELBOURNE FOR THE YEAR 1982
The Judges of the Supreme Court have appointed that the Sittings of the said Supreme Court for the hearing of Criminal Trials in Melbourne shall, during the year 1982 be held on the days and dates indicated on the Schedule following, that is to say:
| February | March | April | May | June | July | August | September | October | November | December |
|----------|-------|-------|-----|------|------|--------|-----------|---------|----------|----------|
| Mon. 1 | Mon. 1| Thur. 1| Mon. 3| Tue. 1| Thur. 15| Mon. 2| Wed. 1| Fri. 1| Mon. 1| Wed. 1|
Dated 10 August 1981
P. S. MALBON
Prothonotary of the Supreme Court
Police Regulation Act, Section 122
SALE OF UNCLAIMED MOTOR VEHICLE
A 1968 model Valiant sedan former Reg. No. VA 8646 (Tas.) engine no. LV307 7189; now in the possession of Police on 10 January 1981 and by Court Order to be sold by Public auction at the Frankston District Headquarters Complex, Davey Street, Frankston, at 10.00 a.m. on Friday, 21 August 1981.
S. I. MILLER
Chief Commissioner
DEPARTMENT OF MINERALS AND ENERGY
Subject to any necessary exclusions, &c., it is proposed to grant the following Mining Leases:
No. 355-1; Mildura Plaster Mills Pty. Ltd.; 237-2 ha, Parish of Goongerah.
No. 626; Wattle Gully Gold Mines No Liability; 86 ha, Parish of Chewton.
APPLICATION FOR MINING LEASE DECLARED ABANDONED
No. 679; Rory O'Briain Bell; 260 ha, Parish of Kosciusko.
APPLICATIONS FOR EXPLORATION LICENCES REFUSED
No. 1028; C.R.A. Exploration Pty. Limited; 792 km², County of Tanjil.
No. 1029; C.R.A. Exploration Pty. Limited; 792 km², County of Tanjil.
No. 1030; C.R.A. Exploration Pty. Limited; 528 km², Counties of Buln Buln and Tanjil.
EXPLORATION LICENCES GRANTED
No. 870; Dampier Mining Company Limited; 525 km², comprising Graticular Blocks Nos. 2153, 2223, 2224, 2225, 2294, 2295, 2296 and 2297 Hamilton Map Sheet.
No. 871; Dampier Mining Company Limited; 462 km², comprising Graticular Blocks Nos. 2082, 2083, 2084, 2154, 2155, 2226 and 2298 Hamilton Map Sheet.
No. 939; Hardrock Exploration Pty. Ltd. 66 km², comprising Graticular Block No. 1514 Melbourne Map Sheet.
No. 975; C.R.A. Exploration Pty. Limited; 198 km², comprising Graticular Blocks Nos. 1093, 1094 and 1166 Melbourne Map Sheet.
TERM OF EXPLORATION LICENCE EXTENDED
No. 554; Pennzoil of Australia Limited and Picon Explorations Pty. Limited; 396 km², Counties of Hampden, Ripon and Villiers.
TERM OF EXPLORATION LICENCE EXTENDED AND AREA RELINQUISHED
No. 690; C.R.A. Exploration Pty. Limited; extended for twelve months. Area retained 132 km² comprising Graticular Blocks Nos. 324 and 332 Melbourne Map Sheet. Area relinquished 192 km², comprising Graticular Blocks Nos. 283 and 335 Melbourne Map Sheet. This area will become available again for Exploration Licence on 17 November 1981.
APPLICATION FOR SEARCH LICENCE DECLARED ABANDONED
No. 1945; Norman Wright; 40 ha, Parish of Elphinstone.
APPLICATIONS FOR SEARCH LICENCES REFUSED
Nos. 1827, 1828, 1829, 1830 and 1831; New Yankee Metals Pty. Limited; 40 ha's each, Parish of Blackwood.
Nos. 1926 and 1927; Sue Conners; 38 ha's each, Parish of Warrowtue.
Nos. 2144, 2145, 2146, 2147 and 2148; Kenneth Inglis Lindner, Peter Frederick Bull; 40 ha's each, Parish of Yehrip.
SEARCH LICENCES GRANTED
No. 1609; John Christian Shimmen, Robert Olsson; 40 ha, Parish of Neerim East.
No. 1637; Eric Clifford Braumann; 18 ha, Parish of Nerrigundah.
No. 1752; Castlemaine Engineering and Mining Co. Pty. Ltd.; 40 ha, Parish of Chewton.
Nos. 1954 and 1955; Lionel Michael Gross; 40 ha's each, Parish of Bullumwaal.
Nos. 1958, 1959, 1960 and 1961; Vidor Limited; 40 ha's each, Parish of Korweinguboora.
No. 2028; David George Morgan; 40 ha, Parish of Bullumwaal.
No. 2035; Australian Coal and Gold Holdings Limited; 40 ha, Parish of Cooma.
SEARCH LICENCE EXPIRED
No. 1477; Beryl Lucinda Howard; 40 ha, Parish of Bogong North.
TAILINGS LICENCES GRANTED
No. 4452; Shire of Maldon; to remove tailings over that area of Crown Land situated east of Porcupine Creek west of the Porcupine Flat Recreation Reserve Extension, and north of Crown Allotment 18, section F, in the Parish of Maldon.
No. 4496; Gerald Rodger Gill; to remove tailings situated at the Old Battery one mile in a north-westerly direction from the top of Egerton Hill situated in the Parish of Drummond.
No. 4497; Gerald Rodger Gill; to remove tailings situated at the Humbolt Mine situated in the Parish of Drummond.
No. 4625; Condon Mines, Pty. Ltd.; to remove tailings located immediately to the north of the State Battery at Egerton situated in the Parish of Bungala.
APPLICATION FOR EXTRACTIVE INDUSTRY LEASE REFUSED
No. 176; Reginald Donald Evans; 54-6 ha, Parish of Angahook.
APPLICATIONS FOR EXTRACTIVE INDUSTRY LEASES DECLARED ABANDONED
No. 253; Alcoa of Australia Limited; 62 ha, Parishes of Angahook and Jan Juc.
No. 254; Alcoa of Australia Limited; 53 ha, Parishes of Angahook and Jan Juc.
EXTRACTIVE INDUSTRY SEARCH PERMIT EXPIRED
No. 88; Sydney John Treasure, Warren John Grant, David John Tilson, John Giannarelli; 20-2 ha, Parish of Bulgarback.
EXTRACTIVE INDUSTRY SEARCH PERMITS GRANTED
No. 103; The Readymix Group Limited; 13-4 ha, Parish of Loy Yang.
No. 104; The Readymix Group Limited; 10 km², Parish of Loy Yang.
D. G. CROZIER
Minister for Minerals and Energy
NOTICE OF APPLICATION FOR A PIPELINE LICENCE
Hematite Petroleum Pty. Ltd. and Esso Exploration and Production Australia, Inc. of 148 William Street, Melbourne, and care of Price Waterhouse Nominees (Victoria) Pty. Ltd., 447 Collins Street, Melbourne, respectively have applied for a pipeline licence in respect of the construction of a 324 mm pipeline along the route and in the position described hereunder for the conveyance of petroleum between the West Kingfish A platform situated as described below in Licence Area VIC/L7 and the Kingfish A platform situated as described below in Licence Area VIC/L7 of which they are registered holders.
ROUTE AND POSITION OF THE PIPELINE
Between the West Kingfish A platform proposed for construction at or about Latitude 38 deg. 35 min. 40-6 sec. South, Longitude 148 deg. 06 min. 15-1 sec. East and the Kingfish A platform situated at or about Latitude 38 deg. 35 min. 51-8 sec. South, Longitude 148 deg. 08 min. 35-6 sec. East along the sea bed.
Dated 17 July 1981
Made under the Petroleum (Submerged Lands) Act 1967 of the Commonwealth of Australia,
Made under the Petroleum (Submerged Lands) Act 1967 of the State of Victoria.
DIGBY CROZIER
Designated Authority
COUNTRY ROADS BOARD
RESOLUTIONS OF THE COUNTRY ROADS BOARD
The Country Roads Board, in pursuance of the provisions of the Country Roads Act 1958, has passed Resolutions the dates whereof and the terms of which are scheduled hereunder:
SCHEDULE
Unclassified Road
Resolution dated the Twentieth day of July, One thousand nine hundred and eighty-one made pursuant to section 21 and 21A of the Country Roads Act 1958 declaring the road in the City of Collingwood as shown hatched on plan numbered G.P.14553A hereunder to be a road within the meaning and for the purposes of the said Act.
Roadside Reserve
Resolution dated the Twentieth day of July, One thousand nine hundred and eighty-one, made pursuant to section 112A of the Country Roads Act 1958 declaring the area of land in the City of Collingwood as shown hatched on plan numbered G.P.14553B hereunder to be a roadside reserve within the meaning and for the purposes of the said Act.
Freeway
Resolution dated the Third day of August, One thousand nine hundred and eighty-one, made pursuant to sections 21 and 101 of the Country Roads Act 1958 varying the declaration of the Tullamarine Freeway in the City of Keilor which was published in the Government Gazette of the Nineteenth day of December, One thousand nine hundred and seventy-nine, on page 3961, by inclusion of the area shown diagonally hatched and exclusion of the area shown cross hatched on plan numbered G.P.14073A hereunder.
Unclassified Road
Resolution dated the Third day of August, One thousand nine hundred and eighty-one, made pursuant to sections 21 and 110 of the Country Roads Act 1958, declaring the widening of Matthews Avenue in the City of Keilor shown hatched on plan numbered G.P.14073C hereunder to be part of a road within the meaning and for the purposes of the said Act.
COUNTRY ROADS BOARD
RESOLUTIONS OF THE COUNTRY ROADS BOARD
The Country Roads Board, in pursuance of the provisions of the Country Roads Act 1958, has passed Resolutions the dates whereof and the terms of which are scheduled hereunder:
SCHEDULE
State Highway
Resolution dated the Twenty-seventh day of July, One thousand nine hundred and eighty-one, made pursuant to sections 21 and 74 of the Country Roads Act 1958 declaring the widening of the Kiewa Valley Highway in the Shire of Bright as shown hatched on plan numbered G.P.13276 hereunder to be part of a State highway within the meaning and for the purposes of the said Act.
Main Roads
Resolution dated the Twenty-seventh day of July, One thousand nine hundred and eighty-one, made pursuant to sections 21 and 58 of the Country Roads Act 1958 declaring the deviation from the Whittlesea-Kinglake Road in the Shires of Eltham and Yea as indicated by diagonal hatching on plans numbered G.P.13238a, G.P.13238b, G.P.13238c, G.P.13239a, G.P.13239b, G.P.13240a, G.P.13240b and G.P.13240c hereunder to be part of a main road within the meaning and for the purposes of the said Act, and also declares that such deviation shall be in lieu of the existing road as shown thereof indicated by cross hatching on plans numbered G.P.13238a, G.P.13238b, G.P.13238c, G.P.13239a and G.P.13239b.
Resolution dated the Twenty-seventh day of July, One thousand nine hundred and eighty-one, made pursuant to section 21 of the Country Roads Act 1958, declaring the widenings of the Kilmore-Kilmore East Road in the Shire of Kilmore as shown hatched on plans numbered G.P.13043a and G.P.13043c hereunder to be part of a main road within the meaning and for the purposes of the said Act.
Resolution dated the Twenty-seventh day of July, One thousand nine hundred and eighty-one, made pursuant to section 21 of the Country Roads Act 1958, declaring the widenings of North Road in the Cities of Moorabbin and Oakleigh as shown hatched on plan numbered G.P.15155 hereunder to be part of a main road within the meaning and for the purposes of the said Act.
Unclassified Road
Resolution dated the Twenty-seventh day of July, One thousand nine hundred and eighty-one, made pursuant to sections 21 and 110 of the Country Roads Act 1958, declaring that the whole of Patrick Street in the Shire of Kilmore, as shown hatched on plan numbered C.P.1787 hereunder to be part of a road within the meaning and for the purposes of the said Act.
6 August 1981
G. K. COX
Secretary
Teaching Service Act 1958
TEACHING SERVICE (TEACHERS TRIBUNAL)
REGULATIONS
AMENDMENT No. 570 (T.T.109)
The Teachers Tribunal, in pursuance of the powers conferred by the Teaching Service Act 1958, hereby amends Regulation 23 of the Teaching Service (Teachers Tribunal) Regulations as follows:—
REGULATION 23
In the second proviso to sub-clause 8 (a) delete the figure "28" and substitute therefor the figure "35".
J. J. KENNEDY, Chairman
L. D. KEEN, Acting Secretary
Office of the Teachers Tribunal
Melbourne, 31 July 1981
Cemeteries Act 1958
SCALE OF FEES OF THE BALLAN PUBLIC CEMETERY
In pursuance of the powers conferred upon them by the Cemeteries Act, the trustees of the Ballan Public Cemetery hereby make the following scale of fees, which shall come into operation upon publication in the Government Gazette, and from and after such publication every scale of fees heretofore made by the said trustees shall be and is hereby rescinded to the extent to which it conflicts with this scale.
Private Graves
$
| Description | Fee |
|--------------------------------------------------|-------|
| Land, 2-44 m x 1-22 m (Lawn Section) | 120.00|
| Land, 2-44 m x 1-22 m (Monumental Section) | 50.00 |
| Own Selection of Land | 10.00 |
Sinking Charges for Private Graves
| Description | Fee |
|--------------------------------------------------|-------|
| Sinking grave 1-83 m deep | 100.00|
| Each additional 0-3 m | 15.00 |
| Sinking oversize grave | 25.00 |
| Cancellation or order to sink (if commenced) | 13.00 |
Reopening Charges
| Description | Fee |
|--------------------------------------------------|-------|
| Reopening grave (no cover) | 100.00|
| Reopening grave (with cover) | 110.00|
Extra Charges
| Description | Fee |
|--------------------------------------------------|-------|
| Interment outside prescribed hours, or on Saturdays, Sundays or Public Holidays | 30.00 |
| Interment in a private grave without due notice | 30.00 |
Miscellaneous Charges
| Description | Fee |
|--------------------------------------------------|-------|
| Interment fee | 20.00 |
| Permission to erect a headstone or monument—5 per cent of cost with a minimum of $10.00 | 9.00 |
| Permission to construct a headstone grave or to erect any stone kerb, brick tile-work or concrete | 9.00 |
| Exhuming the remains of a body (when authorized) | 150.00|
| Interment of ashes in a private grave | 20.00 |
Approved by the Governor in Council, 11 August 1981
—TOM FORRISTAL, Clerk of the Executive Council
Cemeteries Act 1958
SCALE OF FEES OF THE BOX HILL PUBLIC CEMETERY
In pursuance of the powers conferred upon them by the Cemeteries Act the trustees of the Box Hill Public Cemetery hereby make the following scale of fees, which shall come into operation upon publication in the Government Gazette, and from and after such publication every scale of fees heretofore made by the said trustees shall be and is hereby rescinded to the extent to which it conflicts with this scale.
Land for Private Graves
$
| Description | Fee |
|--------------------------------------------------|-------|
| Plots 2-44 m x 1-22 m | 150.00|
Sinking or Re-opening Private Grave
| Description | Fee |
|--------------------------------------------------|-------|
| Sinking grave, 2-13 m x 0-83 m deep (standard measurement) | 180.00|
| Re-opening grave (without slab or cover) | 180.00|
| Re-opening grave (with slab or cover) | 205.00|
Note—Where the removal and replacement of slab is not performed by cemetery employees or where the holder of the Right of Burial so wishes, all replacements must be made by the holder, prior to the opening of the grave, for such removal and replacement to be performed by a monumental mason at the expense of the holder of the Right of Burial.
Approved by the Governor in Council, 11 August 1981
—TOM FORRISTAL, Clerk of the Executive Council
Cemeteries Act 1958
SCALE OF FEES OF THE BUMBERRAH PUBLIC CEMETERY
In pursuance of the powers conferred upon them by the Cemeteries Act the trustees of the Bumberrah Public Cemetery hereby make the following scale of fees, which shall come into operation upon publication in the Government Gazette, and from and after such publication every scale of fees heretofore made by the said trustees shall be and is hereby rescinded to the extent to which it conflicts with this scale.
Lawn Section
$
| Description | Fee |
|--------------------------------------------------|-------|
| Grave site, 2-44 m x 1-22 m | 130.00|
| Interment | 100.00|
| Bronze plaque and flower vase | 120.00|
N. S. STEWART, Trustee
P. STEWART, Trustee
V. WYATT, Trustee
Approved by the Governor in Council, 11 August 1981
—TOM FORRISTAL, Clerk of the Executive Council
Cemeteries Act 1958
SCALE OF FEES OF THE COBRAM PUBLIC CEMETERY
In pursuance of the powers conferred upon them by the Cemeteries Act the trustees of the Cobram Public Cemetery hereby make the following scale of fees, which shall come into operation upon publication in the Government Gazette, and from and after such publication every scale of fees heretofore made by the said trustees shall be and is hereby rescinded to the extent to which it conflicts with this scale.
$
| Description | Fee |
|--------------------------------------------------|-------|
| Land for private grave, 2-44 m x 1-22 m | 70.00 |
| Sinking grave, 1-83 m deep (extra) | 55.00 |
| Each additional 0-3 m | 10.00 |
Cemeteries Act 1958
SCALE OF FEES OF THE MOOROOPNA PUBLIC CEMETERY
In pursuance of the powers conferred upon them by the Cemeteries Act the trustees of the Mooroopna Public Cemetery hereby make the following scale of fees, which shall come into operation upon publication in the Government Gazette, and from and after such publication every scale of fees heretofore made by the said trustees shall be and is hereby rescinded to the extent to which it conflicts with this scale.
Public Graves
$
Interment in grave without exclusive right—stillborn child 22.00
Interment in grave without exclusive right—others 44.00
Number peg or label 7.00
Private Graves
Land 2-44 m x 1-22 m 82.00
Own selection of site (extra) 23.00
Sinking Charges for Private Graves
Sinking grave 1-83 m deep 72.00
Each additional 0-3 m 14.00
Sinking oversize grave (extra) 27.00
Sinking oversize grave for American Type casket (extra) 27.00
Cancellation of order to sink (if commenced) 14.00
Reopening Charges
Reopening grave (no cover) 62.00
Reopening grave (with cover or kerb) 69.00
Miscellaneous Charges
Interment fee 23.00
Certificate of right of burial 4.00
Number plate or brick 7.00
Inspection of plan or register 5.00
Annual maintenance (single grave)—optional 24.00
Permission to erect a headstone or monument—5 per cent of cost with minimum of $7.00 14.00
Permission to construct a brick grave or to erect any stone kerb, brick tilework or concrete 10.00
Grave renovations or additional inscription 9.00
Exhuming the remains of a body (when authorized) 100.00
Interment of ashes in a private grave 22.00
Memorial wall niche and plaque 69.00
Extra Charges
Interment not in the prescribed hours or on Saturdays, Sundays or Public Holidays 28.00
Interment in private grave without due notice 27.00
Late fee (per half-hour or thereof in excess of fifteen minutes) 4.00
Lawn Section
Lawn Grave Site, 2-44 m x 1-22 m 125.00
Interment fee 125.00
Plaque 80.00
Flower container 20.00
Interment on weekends and public holidays (extra) 30.00
Approved by the Governor in Council, 11 August 1981—
Tom Forristal, Clerk of the Executive Council
Cemeteries Act 1958
SCALE OF FEES OF THE SANDFORD PUBLIC CEMETERY
In pursuance of the powers conferred upon them by the Cemeteries Act the trustees of the Sandford Public Cemetery hereby make the following scale of fees, which shall come into operation upon publication in the Government Gazette, and from and after such publication every
scale of fees heretofore made by the said trustees shall be and is hereby rescinded to the extent to which it conflicts with this scale.
$20.00
Private grave, 2·44 m x 1·22 m
J. F. WIDDICOMBE, Trustee
T. L. DOYLE, Trustee
S. CARTER, Trustee
C. H. BROOKE, Trustee
Approved by the Governor in Council, 11 August 1981—
TOM FORRISTAL, Clerk of the Executive Council
Cemeteries Act 1958
SCALE OF FEES OF THE WODONGA PUBLIC CEMETERY
In pursuance of the powers conferred upon them by the Cemeteries Act the trustees of the Wodonga Public Cemetery hereby make the following scale of fees, which shall come into operation upon publication in the Government Gazette; and from and after such publication this scale of fees heretofore made by the said trustees shall be and is hereby rescinded to the extent to which it conflicts with this scale.
Lawn Area
$140.00
Land 2·44 m x 1·22 m
$35.00
Sinking a Grave to 2·13 m
$65.00
Sinking a Grave to 1·37 m
$90.00
Reopening a Grave
$105.00
Desk and Single Plaque
$145.00
Desk and Double Plaque
$40.00
Interment on Saturday, Sunday or Public Holiday (extra)
$45.00
Sinking a Grave for Stillborn Child
Children's Section
$30.00
Land 2·44 m x 1·22 m
$45.00
Sinking a Grave for Stillborn Child
$30.00
Land for Stillborn Child
$65.00
Sinking a Grave to 1·37 m
$105.00
Desk and Single Plaque
Monumental Section
$80.00
Land 2·44 m x 1·22 m
$30.00
Land for Stillborn Child
$95.00
Sinking a Grave to 2·13 m
$65.00
Sinking a Grave to 1·37 m
$45.00
Sinking a Grave for Stillborn Child
$25.00
Permit to plant Headstone
$90.00
Reopening a grave (without headstone)
$110.00
Reopening a grave (with headstone)
$90.00
Interment of ashes in Niche Wall including a Bronze Plaque
$40.00
Interment on Saturday, Sunday or Public Holiday (extra)
$30.00
Casket (unusual size) (extra)
G. GEHRIG, Trustee
G. W. EMBURY, Trustee
W. MADEX, Trustee
Approved by the Governor in Council, 11 August 1981—
TOM FORRISTAL, Clerk of the Executive Council
Town and Country Planning Act 1961
MELBOURNE METROPOLITAN PLANNING SCHEME
AMENDMENT No. 115, PART 2B
Notice of Approval
In pursuance of the powers conferred by the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981 approved a planning scheme entitled the Melbourne Metropolitan Planning Scheme, Amendment No. 115, Part 2a, modifying the ordinance to vary the use of setback areas in industrial zones for car parking and loading facilities and changes to the location of car parks and such planning scheme comes into operation on the date this Notice of Approval is published in the Government Gazette.
A copy of the planning scheme as approved may be inspected free of charge during office hours, at the office of the Department of Planning, 235 Queen Street, Melbourne, at the office of the Melbourne and Metropolitan Board of Works at 625 Little Collins Street, Melbourne, and when available, at the Office of Titles, Melbourne, and at the Central Plan Office of the Department of Crown Lands and Survey, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
MELBOURNE METROPOLITAN PLANNING SCHEME
AMENDMENT No. 139, PART 2
Notice of Approval
In pursuance of the powers conferred by the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, approved a planning scheme entitled the Melbourne Metropolitan Planning Scheme, Amendment No. 139, Part 2, in respect of the municipal district of the Shire of Bulla and Cities of Kellor and Brighton and such planning scheme comes into operation on the date this Notice of Approval is published in the Government Gazette.
A copy of the planning scheme as approved may be inspected free of charge, during office hours, at the office of the Department of Planning, 235 Queen Street, Melbourne, at the office of the Melbourne and Metropolitan Board of Works at 625 Little Collins Street, Melbourne, and when available, at the Office of Titles, Melbourne, and at the Central Plan Office of the Department of Crown Lands and Survey, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
MELBOURNE METROPOLITAN PLANNING SCHEME
NOTICE OF A PROPOSED AMENDMENT TO THE MELBOURNE METROPOLITAN PLANNING SCHEME
Amendment No. 177
Notice is hereby given that pursuant to sub-section 7 of section 52 of the Town and Country Planning Act 1961, the Minister has required that the Melbourne and Metropolitan Board of Works give notice of and deposit for inspection by the public for a period of one month a proposed amendment to the Melbourne Metropolitan Planning Scheme.
The amendment, known as Amendment No. 177 is to rezone to Restricted Business land in Sandringham Road and Trentham Street, Sandringham, and to amend Clause 22AA of the Ordinance.
The proposed amendment is available for inspection at the office of the Melbourne and Metropolitan Board of Works, 625 Little Collins Street, Melbourne, at the office of the Department of Planning, 235 Queen Street, Melbourne, and at the office of the City of Sandringham, and will be open for inspection during office hours, free of charge.
Any persons wishing to make a submission in respect of the proposed amendment should do so in writing direct to the Minister for Planning at 500 Collins Street, Melbourne, to reach the Minister on or before 19 September 1981.
Dated 18 August 1981
O. T. W. COSGRIFF
Secretary
Melbourne and Metropolitan Board of Works
625 Little Collins Street, Melbourne, Vic. 3001
Town and Country Planning Act 1961
CITY OF ARARAT PLANNING SCHEME
AMENDMENT No. 22
Notice of Approval
In pursuance of the powers conferred by the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, approved a planning scheme entitled the City of Ararat Planning Scheme, Amendment No. 22, in respect of the municipal district of the City of Ararat and such planning scheme comes into operation on the date this Notice of Approval is published in the Government Gazette.
A copy of the planning scheme as approved may be inspected free of charge during office hours, at the office of the Department of Planning, 235 Queen Street, Melbourne, at the office of the Council of the City of Ararat at Municipal Offices, Ararat, and when available, at the Office of Titles, Melbourne, and at the Central Plan Office of the Department of Crown Lands and Survey, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
CITY OF ARARAT PLANNING SCHEME
AMENDMENT No. 27
Notice of Amendment
In pursuance of the powers conferred by sub-section 8 of section 32 of the Town and Country Planning Act 1961, the Governor in Council, on 11 August 1981, amended the City of Ararat Planning Scheme insofar as it makes a number of amendments to the Planning Scheme Map.
A copy of the amendment may be inspected during office hours at the office of the Department of Planning at 235 Queen Street, Melbourne, at the office of the Council of the City of Ararat and when available at the Office of Titles, Melbourne and at the Central Plan Office of the Department of Crown Lands and Survey, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
REVOCATION OF THE CITY AND SHIRE OF ARARAT JOINT PLANNING AUTHORITY
INTERIM DEVELOPMENT ORDER
Notice of Revocation
In pursuance of the powers conferred by section 26 of the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, revoked the City and Shire of Ararat Joint Planning Authority Interim Development Order.
A copy of the revocation may be inspected, free of charge, during office hours at the office of the Department of Planning, 235 Queen Street, Melbourne, and at the office of the City and Shire of Ararat Joint Planning Authority at the office of the Council of the City of Ararat, Municipal Offices, Ararat.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act, 1961
PORTLAND PLANNING SCHEME
(TOWN OF PORTLAND)
AMENDMENT No. 21 (2)
Notice of Approval
In pursuance of the powers conferred by the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, approved a planning scheme entitled the Portland Planning Scheme (Town of Portland), Amendment No. 21 (2) in respect of the municipal district of the Town of Portland and such planning scheme comes into operation on the date this Notice of Approval is published in the Government Gazette.
A copy of the planning scheme as approved may be inspected free of charge during office hours, at the office of the Department of Planning, 235 Queen Street, Melbourne, at the office of the Council of the Town of Portland at Portland, and when available, at the Office of Titles, Melbourne, and at the Central Plan Office of the Department of Crown Lands and Survey, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
SHIRE OF ARARAT PLANNING SCHEME
INTERIM DEVELOPMENT ORDER
Notice of Approval
In pursuance of the powers conferred by the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, approved an Interim Development Order made by the Council of the Shire of Ararat for the whole of the municipal district of the Shire of Ararat.
The Interim Development Order provides that land use and development for the whole of the Shire shall be subject to consent by the Responsible Authority.
A copy of the Interim Development Order may be inspected, free of charge, during office hours at the office of the Council of the Shire of Ararat, Ararat, and at the office of the Department of Planning, 235 Queen Street, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
SHIRE OF BACCHUS MARSH PLANNING SCHEME
AMENDMENT No. 12
Notice of Amendment
In pursuance of the powers conferred by sub-section 6 of section 32 of the Town and Country Planning Act 1961, the Governor in Council, on 11 August 1981, amended the Shire of Bacchus Marsh Planning Scheme insofar as it amends the Planning Scheme Map in Bacchus Street, Bacchus Marsh.
A copy of the amendment may be inspected during office hours at the office of the Department of Planning at 235 Queen Street, Melbourne, at the office of the Council of the Shire of Bacchus Marsh, Bacchus Marsh, and when available at the Office of Titles, Melbourne, and at the Central Plan Office of the Department of Crown Lands and Survey, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
SHIRE OF COHUNA PLANNING SCHEME
INTERIM DEVELOPMENT ORDER
Notice of Approval
In pursuance of the powers conferred by the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, approved an Interim Development Order made by the Council of the Shire of Cohuna for the whole of the municipal district of the Shire of Cohuna.
A copy of the Interim Development Order may be inspected free of charge, during office hours at the office of the Shire of Cohuna, King Edward Street, Cohuna, and at the office of the Department of Planning, 235 Queen Street, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
SHIRE OF KILMORE PLANNING SCHEME 1973
AMENDMENT No. 21, 1977
Notice of Approval
In pursuance of the powers conferred by the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, approved a planning scheme entitled the Shire of Kilmore Planning Scheme 1973, Amendment No. 21, 1977, in respect of the municipal district of the Shire of Kilmore, until such planning scheme comes into operation on the date this Notice of Approval is published in the Government Gazette.
A copy of the planning scheme as approved may be inspected free of charge, during office hours, at the office of the Department of Planning, 235 Queen Street, Melbourne, at the office of the Shire of Kilmore, Town Hall, 16 Sydney Street, Kilmore, and when available, at the Office of Titles, Melbourne, and at the Central Plan Office of the Department of Crown Lands and Survey, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
CITY OF KNOX PLANNING SCHEME 1965
AMENDMENT No. 223, PART 1, 1980
Notice of Approval
In pursuance of the powers conferred by the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, approved a planning scheme entitled the City of Knox Planning Scheme 1965, Amendment No. 223, Part 1, 1980, in respect of the municipal district of the City of Knox, until such planning scheme comes into operation on the date this Notice of Approval is published in the Government Gazette.
A copy of the planning scheme as approved may be inspected free of charge, during office hours, at the office of the Department of Planning, 235 Queen Street, Melbourne, at the office of the City of Knox at the Civic Centre, 511 Belwood Highway, Knoxfield, and when available at the Office of Titles, Melbourne, and at the Central Plan Office of the Department of Crown Lands and Survey, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
SHIRE OF LILYDALE PLANNING SCHEME
AMENDMENT No. 39, 1973
Notice of Approval
In pursuance of the powers conferred by the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, approved a planning scheme entitled the Shire of Lilydale Planning Scheme 1958, Amendment No. 39, 1973, in respect of the municipal district of the Shire of Lilydale, until such planning scheme comes into operation on the date this Notice of Approval is published in the Government Gazette.
A copy of the planning scheme as approved may be inspected free of charge, during office hours, at the office of the Department of Planning, 235 Queen Street, Melbourne, at the office of the Shire of Lilydale, Anderson Street, Lilydale, and when available, at the Office of Titles, Melbourne, and at the Central Plan Office of the Department of Crown Lands and Survey, Melbourne.
N. G. HAYNES, Secretary
Department of Planning
Town and Country Planning Act 1961
SHIRE OF MANSFIELD INTERIM DEVELOPMENT ORDER
AMENDMENT No. 6
Notice of Amendment
In pursuance of the powers conferred by section 26 of the Town and Country Planning Act 1961, the Governor in Council on 11 August 1981, amended the Shire of Mansfield Interim Development Order.
A copy of the amendment may be inspected, free of charge, during office hours at the office of the Department of Planning, 235 Queen Street, Melbourne, and at the office of the Council of the Shire of Mansfield at Mansfield.
N. G. HAYNES, Secretary
Department of Planning
PUBLIC TRUSTEE ACT 1958, SECTION 17
I hereby give notice that on 30 July 1981, the Public Trustee held Elections to Administer the following deceased persons' estate in accordance with section 17 of the Public Trustee Act 1958.
WILLIAMS, EDWARD JAMES, also known as Edward Henry Zahner and Edward William Laing, late of Mosman Hall, Charters Towers, North Queensland, pensioner, died 14 February 1981.
P. T. SPENCER
Public Trustee
168 Exhibition Street, Melbourne, 3000
12 August 1981
Creditors, next of kin and others having claims against the estate of any of the under-mentioned deceased persons are required to send particulars of their claims to the Public Trustee, 168 Exhibition Street, Melbourne, Vic. 3000, the personal representative, on or before 26 October 1981; after which date the Public Trustee may dispose of distribute the assets having regard only to the claims of which the Public Trustee then has notice:
FOOT, FLORENCE MAUD, late of c/- St. Leeor Private Hospital, 31 Thanet Street, Malvern, spinster, died 22 May 1981.
WILLIAMS, EDWARD JAMES, also known as Edward Henry Zahra and Edward William Laing, late of Mosman Hall, Charters Towers, North Queensland, pensioner, died 14 February 1981.
Melbourne, 12 August 1981
P. T. SPENCER
Public Trustee
Evidence Act 1958
APPROVAL OF MACHINES FOR MICRO-FILMING DOCUMENTS
I, Haddon Storey, Her Majesty's Attorney-General for the State of Victoria, pursuant to the provisions of section 53c of the Evidence Act 1958, hereby approve for micro-filming documents in the ordinary course of business the following photographic copying machine—
RECORDAK RELIANT 750-L MICROFILMER
Dated at Melbourne, 9 August 1981
HADDON STOREY
Attorney-General
Stamps Act 1958
STAMPS (EXEMPTIONS) REGULATIONS 1971
NOTICE
Pursuant to the provisions of the Stamps (Exemptions) Regulations 1971, I hereby declare—Lloyds International Limited to be, for the purpose of sub-division (13A) of Division 3 of Part II. of the Stamps Act 1958, a dealer in the unofficial short-term money market.
S. VINEY
Comptroller of Stamps
Stamp Duties Office
Melbourne
Co-operation Act 1958
CORIO STATE SCHOOL CO-OPERATIVE LIMITED
D.K.A. RECREATION CLUB LIMITED
FOURTH DONCASTER BOY SCOUT CO-OPERATIVE LIMITED
YARRAMBAT RECREATION CO-OPERATIVE CLUB LIMITED
Notice is hereby given in pursuance of section 78 (7) of the Co-operation Act 1958 and section 305 (2) of the Companies Act 1961 that, from the expiration of thirty months from the date hereof, the names of the aforementioned societies will, unless cause is shown to the contrary, be struck off the register and the societies will be dissolved.
Dated 5 August 1981
P. J. ROGAN
Acting Deputy Registrar
Co-operation Act 1958
BURWOOD HEIGHTS STATE SCHOOL CO-OPERATIVE
LAKESIDE HIGH SCHOOL CO-OPERATIVE LIMITED
MARCELLIN COLLEGE P. & F. CO-OPERATIVE LIMITED
PADUA COLLEGE HALL CO-OPERATIVE LIMITED
PINES SOCCER CO-OPERATIVE LIMITED
NOTICE OF DISSOLUTION OF SOCIETIES
Notice is hereby given that I have this day registered the dissolution of the abovenamed societies and cancelled their registration under the abovenamed Act.
Dated at Melbourne, 5 August 1981
P. J. ROGAN
Acting Deputy Registrar
CONTRACTS ACCEPTED—(Series 1981–82)
AMENDMENTS
| Schedule Number | Item Number | New Rate | Effective Date |
|-----------------|-------------|----------|----------------|
| | | | $ |
Antiseptics, Disinfectants, Deodorants, &c.
Minimum order Quantities
1/01 13 50 units† 1.7.81
14 10 units
15 1 unit
† Change of description:
From: 5 litre tins
To: 5 litre plastic containers
Electrical Goods
1/05 61 Olex List 28.7.81 Less 35% 11.8.81
Soaps and Soap Mixtures
1/15 20 Purchase Reg. 84 14.8.81
Piping and Fittings
1/34 13 Change of Description:
From: ½ inch
To: 1½ inch 1.7.80
Motor Spirit, Kerosene, &c.
1/53B 37 0.6256 3.8.81
Photographic and X-ray Equipment
1/55 289 120.95 17.8.81
290 62.36
291 50.38
294 138.72
295 143.04
296 45.53
297 185.34
298 150.33
299 87.56
Hand Tools (General)
1/56 332 1.85 (5 mm.)* 1.9.81
1.85 (6 mm.)*
2.07 (3-½ mm.)*
3.07 (2-5 mm.)*
4.89 (2 mm.)*
*Trading terms unchanged
Stationery (General)
1/64 58 1982 Invicta List Less 30% Nett 1.8.81
158 0.1950 22.4.81
159 0.2555
160 0.2990
Provisions—Hobson Park Hospital, Traralgon
2/24 (2) 1 0.4325 1.8.81
2 0.4765
Provisions—Warrnambool District
2/25 (4) 3 1.05 1.8.81
Provisions—French Island Recreation Camp
2/30 (4) 85 Change of brand:
From: Heinz
To: San Remo
Change of unit of measurement:
From: per tin
To: per carton of 8
Change of contractor:
From: Hastings Continental
Rd., 24 High Street, Hastings
To: Peninsula Meat Co., 24 High Street, Hastings
J. M. PAWSON, Secretary to the Tender Board
CONTRACTS ACCEPTED—(Series 1981-82)
PUBLIC WORKS
His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof, has by Order made on 11 August 1981, approved of the acceptance by the Minister of Public Works of the undermentioned offer without public tenders being invited, viz.:
Offer of Bruce R. Mitchell Sales Pty. Ltd. for supply of furniture at Travancore Developmental Centre, Flemington, for the sum of ten thousand eight hundred and twenty-two dollars and fifty cents ($10,822.50)—(C.184416–1/003).
TOM FORRISTAL
Clerk of the Executive Council
At the Executive Council Chamber
Melbourne, 11 August 1981
Members of the Advanced Dental Technicians Qualifications Board
CLIVE GEOFFREY DENNIS, D.D.Sc.,
GEOFFREY NEIL HALLAM, T.D.L.C.,
DOUGLAS CHARLES MADSEN, L.D.S., B.D.Sc.,
LOYD GEORGE O'BRIEN, M.D.Sc., L.D.S., D.D.S.,
F.R.A.C.D.S.,
WILLIAM PATRICK STROWNIX, and
MONTAGUE WARDMAN.
to be Members of the Advanced Dental Technicians Qualifications Board, pursuant to the provisions of section 16 (1) of the Dental Technicians Act 1972 for a period of four years commencing 21 August 1981.
Member of Committee of Management of Hospital
WENDY RYAN,
to be a Member of the Committee of Management of the Tawonga District General Hospital, pursuant to the provisions of section 63r (1) of the Hospitals and Charities Act 1963 for a period of three years commencing 11 August 1981.
APPOINTMENTS AND RESIGNATIONS
APPOINTMENTS
His Excellency the Governor of the State of Victoria by and with the advice of the Executive Council thereof has by Orders made on 11 August 1981 been pleased to make the under-mentioned appointments, viz.:
HEALTH COMMISSION
Trustees of Public Cemeteries
ROBERT HAROLD PONTING,
GERALD JOSEPH O'DEA, and
IVON ALLAN ENGLISH,
to be Trustees of the Nagambie Public Cemetery, additional trustees.
BRYAN EDWIN HAYES,
to be a Trustee of the Violet Town Public Cemetery, vice E. D. Hogan, deceased.
PATRICK STEWART,
to be a Trustee of the Bumberrah Public Cemetery, additional trustee, pursuant to section 3 (1) of the Cemeteries Act 1958.
Deputy Superintendent of Mental Hospital
MARCUS KINGSLY BENJAMIN, M.B., B.S., D.P.M.,
F.R.A.C.P., M.C.P., M.R.A.N.Z.C.P.,
to be Deputy Superintendent of E. C. Dax House, pursuant to the provisions of section 26 (1) of the Mental Health Act 1959 as from 2 August 1981.
Superintendent of Mental Hospital
MARCUS KINGSLY BENJAMIN, M.B., B.S., D.P.M.,
F.R.A.C.P., M.C.P., M.R.A.N.Z.C.P.,
to be Superintendent of E. C. Dax House, pursuant to the provisions of section 26 (1) of the Mental Health Act 1959 for the period 6 September 1981 to 14 November 1981, vice D. K. Bhaskaran, on leave.
Superintendent of Mental and Psychiatric Hospital
JOHN REYS FORDYCE,
to be Superintendent of the Bendigo Mental Hospital and Psychiatric Hospital, pursuant to the provisions of section 26 (1) of the Mental Health Act 1959 for the period 18 August 1981 to 23 August 1981, vice J. P. Bromford, on leave.
Member and Chairman of the Advanced Dental Technicians Qualifications Board
JOHN LEONARD McARDLE,
to be a Member and Chairman of the Advanced Dental Technicians Qualifications Board, pursuant to the provisions of section 16 (1) of the Dental Technicians Act 1972 for the period 21 August 1981, to 5 November 1983.
LAW DEPARTMENT
Justices of the Peace
CARL RUDOLPH EVRES, Bridge Road, Richmond, and
STEVEN JOSEPH CLEMENT, 1 Windsor Crescent, Altona North,
to be Justices of the Peace for the State of Victoria.
Commissioners for Taking Declarations, &c.
DOUGLAS ALLEN,
CHARLES CARME, MARIO JOSEPH AQUILINA,
JOHN CHARLES ARCHER,
RICHARD CECIL AVERY,
PETER KEITH BELL,
ALLAN ALFRED BLABY,
IAN JOHN CLARKE,
THOMAS CHARLES CRAY,
HELEN JOY CHISHOLM,
ANDREW CHLEBICA,
DENIS MICHAEL CLARKE,
BEVERLEY ANNE CLYNE,
NICHOLAS JOHN CONSTANCE,
BEVERLEY DUSTAN COOPER,
TERENCE ALBERT FARLEY,
KEVIN WILLIAM FOLEY,
PAUL JAMES FREETH,
WILLIAM BRUAN GAY,
JOHN VICTOR GELLER,
PETER STANLEY GOWERS,
ROBERT GLENN GREENING,
IAN RICHARD HALI,
BARBARA KATHLEEN HALLIDAY,
FREDERICK RALPH HEBB,
JOHN BROADWAY HERBERT,
LEANNE MAREE HIBBLE,
JANE ANN HOBSON,
LEO WILLIAM HOARE,
WILLIAM JAMES HOEY,
MICHAEL ROY JINKS,
PAUL JAMES KELLY,
RICHARD WILLIAM JONES,
DIANNE FAY KNEEBONE,
PETER JAMES KNOWLER,
BERNARD JOHN MALCOLM LAWSON,
GORDON WYLIE LINDSAY,
PETER ROXBURGH MARTIN,
ROBERT CYRIL MCDONALD,
TREVOR DAVID MCDONALD,
NEIL FRANCIS MCQUINN,
NINO NUNZIO MENOLASCINA,
ROBERT EMERSON MCGHILL,
GEOFFREY WILLIAM IRVINE,
MAX STEPHENSON PEART,
RAY HENRY PITCHER,
GRAEME JOHN PIDGEON,
ANNEETTE MARIE PRICE,
MICHAEL DENNIS QUINN,
LESLIE WILLIAM RANTALL,
SUSAN ELIZABETH ROBERTSON,
WILLIAM BRUCE RONALD,
JOHN ANDREW PATRICK ROSS,
BRIAN WILLIAM ROWE,
No. 85—19 August 1981
IAIN BRUCE SEIDMAN,
ALISON LESLIE STOTT,
PAUL JOHN TAYLOR,
BRENDA VERES,
DAVID DEAN WHITE,
JULIE ANN WILLIAMS, and
RAYMOND VERNON WOOD, corner 'Lygon' and Princes Street, Carlton.
LINDSAY JAMES BROWN, 3 Bowen Crescent, Melbourne,
JOHN CHEONG HUNG CHU, 21 Pratt Street, Moonee Ponds, and
PETER LAURENCE DAVIS, 23 Hammond Street, Altona,
HERBERT WILLIAM EDWARDS, 51 McIntyre Road, North Sunshine,
HENRY TIMOTHY KONG, 620 Bourke Street, Melbourne,
DOUGLAS JOHN LOVELESS, 323-351 Canterbury Road, Ringwood,
GEORGE ANDERSON MATTIN, 70-72 Langtree Avenue, Mildura,
PEPPER JOSEPH ANTHONY NARDELLA, 1 Exeter Road, North Croydon, and
CORAL LENORE ROBERTS, Main Street, Poowong,
to be Commissioners for taking Declarations and Affidavits under the Evidence Act 1958.
**Bailiff of the Magistrates Court**
OWEN VICTOR DALY, Magistrates' Court, Kew,
to be Bailiff of the Magistrates' Court at Kew pursuant to the provisions of section 31A of the Magistrates' Court Act 1971.
TOM FORRISTAL
Clerk of the Executive Council
At the Executive Council Chamber
Melbourne, 11 August 1981
---
**Marine Act 1958**
**APPOINTMENT OF WHARF MANAGER**
His Excellency the Governor of the State of Victoria by and with the advice of the Executive Council thereof has by Order made on 28 July 1981 approved the appointment of Senior Sergeant Kenneth Herbert Blay, No. 13750, as Wharf Manager at Warneet to carry out that portion of Part II of the Marine Act 1958 which relates to the management of Public Wharfs and to be an officer under section 19 of such Act to levy and collect wharfage rates thereat, vice Senior Sergeant Norman Lewis Baker, No. 11718, transferred.
TOM FORRISTAL
Clerk of the Executive Council
---
**Inflammable Liquids Act 1966**
**Liquefied Gases Act 1968**
**APPOINTMENT OF INSPECTOR**
The Public Service Board hereby appoints Noel S. Clough, Inspector Inflammable Liquids, Third Division Department of Minerals and Energy, as an Inspector of Inflammable Liquids and an Inspector of Liquefied Gases, pursuant to section 3 (1) of the *Inflammable Liquids Act* 1966 and section 4 (1) and the *Liquefied Gases Act* 1968; without additional salary.
---
**Inflammable Liquids Act 1966**
**Explosives Act 1960**
**APPOINTMENT OF INSPECTOR**
The Public Service Board hereby appoints Stephen J. Dobie, Chemist Class "SO-2", Second Division, Department of Minerals and Energy, as an Inspector of Inflammable Liquids and an Inspector of Explosives, pursuant to section 3 (1) of the *Inflammable Liquids Act* 1966 and section 5 (1) of the *Explosives Act* 1960; without additional salary.
**TERMINATION OF APPOINTMENT**
Nick Awdejew, Industrial Science Officer, Class "E-3A", Second Division, ceases acting as an Inspector of Inflammable Liquids and Inspector of Liquefied Gases effective from the date of his resignation.
Duilio Francis Dambrosi, Chemist, Class "SO-2", Second Division, to cease acting as an Inspector of Inflammable Liquids and Inspector of Explosives, effective from the date of his resignation.
---
**Liquor Control Act 1968**
**APPOINTMENT OF LICENSING INSPECTORS**
In accordance with the authority conferred upon me by Section 6 of the *Police Regulation Act* 1958, I, John Roderick Hall, Deputy Commissioner of Police, do hereby appoint under sub-section (1) of Section 22 of the *Liquor Control Act* 1968, the following Officers of Police as Licensing Inspectors for the Divisions of the Police Districts as shown:
| Division Number | Police District | Rank and Name |
|-----------------|-----------------------|----------------------------------------------------|
| 1 | Prahran | Inspector Graeme Francis Crawford (from 2, 8, 9, 10, 31, 10, 81) |
| 1 | Upper Goulburn | Inspector Michael John O'Hagan (vice Inspector C. F. Side) |
| 2 | Upper Murray | Chief Inspector George Bruce |
13.8.1981 J. R. HALL
Deputy Commissioner (Administration)
---
**RESIGNATIONS**
His Excellency the Governor of the State of Victoria by and with the advice of the Executive Council thereof has by Orders made on 11 August 1981 accepted the resignations of the persons named hereunder of the offices mentioned, viz.:
**Health Commission**
**Member of the Health Advisory Council**
ROLAND McNAUGHTON BRAZIER,
as a Member of the Health Advisory Council as from 2 June 1981, in accordance with the provisions of section 23 (8) of the *Health Commission Act* 1977.
Members of the Committee of Management of Hospitals
WILFRED HERBERT MISSEN,
as a Member of the Committee of Management of the Becac and District Hospital as from 22 June 1981,
VICTOR JAMES ROTHACKER,
as a Member of the Committee of Management of the Inglewood Hospital as from 14 July 1981,
DONALD JAMES HANSON,
as a Member of the Committee of Management of the Macarthur and District Memorial Hospital as from 15 July 1981, and
SAMUEL JAMES KREBS,
as a Member of the Committee of Management of the Southern Peninsula Hospital as from 21 July 1981, in accordance with the provisions of section 63c (1) of the *Hospitals and Charities Act* 1958.
---
**Law Department**
**Justice of the Peace**
WILLIAM DONALD McCREDIE,
as a Justice of the Peace for the State of Victoria.
**Commissioners for Taking Declarations, &c.**
STEVEN JOSEPH CLEMENT,
CARL RUDOLPH EYRES,
BRIAN HORACE RIPPER,
HUGH JAMES STEWARTS, and
ARTHUR JOHN TAYLOR
as Commissioners for taking Declarations and Affidavits under the Evidence Act 1958.
TOM FORRISTAL
Clerk of the Executive Council
At the Executive Council Chamber
Melbourne, 11 August 1981
ORDER IN COUNCIL
RAILWAY CONSTRUCTION AND PROPERTY BOARD ACT 1979
At the Executive Council Chamber, Melbourne the fourth day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Maclellan Mr Lieberman
Mr Kennett Mr Jenkins
VESTING OF HOUSES IN THE RAILWAY CONSTRUCTION AND PROPERTY BOARD
Whereas it is provided by section 24 (1) of the Railway Construction and Property Board Act 1979, that where the Victorian Railways Board by resolution determines that a dwelling house and appurtenant land vested in the Victorian Railways Board that is or has been used for the purposes of a residence for persons employed in the Railway Service should be transferred to the Railway Construction and Property Board for the purposes of Part V of that Act the Governor in Council may, by order published in the Government Gazette, vest the land described in the resolution in the Board for the purposes of Part V, subject to all existing leases, tenancies and other rights or interests of or in that land.
And whereas the Victorian Railways Board by resolution dated 15 June 1981, determined that the dwelling houses and appurtenant lands described in the Schedule hereunder should be transferred to the Railway Construction and Property Board for the purposes of Part V of the Railway Construction and Property Board Act 1979.
Now therefore, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof, and in pursuance of the provisions of section 24 (1) of the Railway Construction and Property Board Act 1979 by this Order vests in the Railway Construction and Property Board the dwelling houses and appurtenant lands described in the Schedule hereunder subject to all existing leases, tenancies and other rights or interests of or in that land.
SCHEDULE
| Address | Lot No. | L.P. No. | Title Particulars | Folio |
|-------------------------------|---------|----------|------------------|-------|
| 6 Showers Ave, Chelsea | 6 | 6709 | 4520 | 979 |
| 2 Fenton St, Warragul | 39 | 40848 | 8084 | 710 (part) |
| 60 Oak St, Seymour | 15 and 28 | 2861 | 4688 | 436 (part) |
| 5 Munro Crt, Korumburra | 12 | 26054 | {7455 014} {7869 022} {7875 114} (part) |
| 8 Enderby Crt, Braybrook | 113 | 44101 | 9233 | 934 |
| 7 Hotham St, Braybrook | 174 | 44101 | 9233 | 995 |
| 4 Hughes St, Braybrook | 26 | 44101 | 9233 | 847 |
| 5 Pritchard Ave, Braybrook | 223 | 44101 | 9234 | 045 |
| 35 Pritchard Ave, Braybrook | 201 | 44101 | 9234 | 023 |
| 1 Rochester St, Braybrook | 89 | 44101 | 9233 | 910 |
| 6 Showers St, Braybrook | 81 | 44101 | 9233 | 902 |
| 27 Showers St, Braybrook | 106 | 44101 | 9233 | 927 |
| 19 Hotham St, Braybrook | 168 | 44101 | 9233 | 989 |
| 35 Beachley St, Braybrook | 119 | 44101 | 9233 | 940 |
| 40 Lorensen Ave, Merlynston | 16 | 29274 | 2049 | 788 (part) |
| 2 Beachley St, Braybrook | 109 | 44101 | 9233 | 930 |
| 37 Pritchard Ave, Braybrook | 200 | 44101 | 9234 | 022 |
| 29 Showers St, Braybrook | 107 | 44101 | 9233 | 928 |
| Dillon St, Ultima | 8 Section 12 | 6006 | 5140 | 933 |
And the Honourable Robert Roy Cameron Maclellan, Her Majesty's Minister of Transport for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
RAILWAY CONSTRUCTION AND PROPERTY BOARD ACT 1979
At the Executive Council Chamber, Melbourne, the twenty-eighth day of July, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Dixon | Mr Ramsay
Mr Lacy
DEVELOPMENT OF RAILWAY LAND
Whereas it is provided by section 20 (4) of the Railway Construction and Property Board Act 1979 that the Railway Construction and Property Board may make recommendations to the Minister for or with respect to the use or development of railway land.
And whereas it is provided by section 20 (5) of the said Act that where the Board has made recommendations to the Minister under section 20 (4) the Minister may submit the recommendations to the Governor in Council for approval.
And whereas the Board has made the following recommendation in respect of Railway land situated as Sale Railway Station.
RECOMMENDATION
That the land which is surplus to railway requirements be used to facilitate the relocation of the Sale Railway Station and Goods Yard out of the main township. This operation would be undertaken under Agreement with VicRail, the Sale City Council and the Railway Construction and Property Board who would be charging a fee for its services. The Sale City Council will make available the land required for a new station and will fund the cost of the road works. The balance of the land will be set aside by the VicRail Board for the station improvement.
In conjunction with the Sale City Council it is proposed to jointly develop the vacated property as a Retail/Residential complex. The expansion of the central business district of Sale would incorporate:
(i) part retail use including a major department store and supermarket plus associated carparking;
(ii) the reinstatement of Desailly Street as a through road in accordance with the Sale Planning Scheme of 1978;
(iii) part business and/or civic offices and associated civic functions, e.g. theatre or cinema; and
(iv) part medium density residential use and provision of parkland/open space.
Now therefore His Excellency the Governor of the State of Victoria by and with the advice of the Executive Council thereof, doth hereby approve of the foregoing recommendation.
And the Honourable Robert Roy Cameron Maclellan, Her Majesty's Minister of Transport for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
HOUSING ACT 1958
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
DECLARATION OF PRIVATE STREETS AS PUBLIC HIGHWAYS WITHIN THE CITY OF TRARALGON
Whereas pursuant to section 107 of the Housing Act 1958 it is among other things enacted that "where works have been carried out in accordance with an agreement for the full construction of any roads, pavements, culverts or drains" made pursuant to the provisions of the said section 107 such roads, pavements, culverts or drains shall be under the care and management of the Municipality which shall be liable for the cost of any maintenance, repair, alteration or reconstruction thereof and the Governor in Council may by order published in the Government Gazette declare any road so constructed to be a public highway.
And whereas by Order dated 2 April 1974 the Governor in Council consented to an agreement between the Housing Commission and the City of Traralgon regarding street and drainage construction in Oak Avenue, Myrtle Crescent, Poplar Avenue, Ash Street, Cedar Street and the foot road between Lots 82 and 83 in the Heathcote Estate situate in the municipality of the City of Traralgon and the carrying out of the works enumerated in the said Agreement.
And whereas the works of road pavements, culvert and drainage construction have now been carried out in accordance with the said Agreement between the Housing Commission and the City of Traralgon.
Now therefore His Excellency the Governor of the State of Victoria by and with the advice of the Executive Council of the said State doth in pursuance of the powers conferred by section 107 of the Housing Act and upon recommendation of the Housing Commission doth by this order declare the streets more particularly delineated and shown hatched on the plan hereto annexed to be dedicated to the public as Public Highways within the meaning of any law now or hereafter in force and that the Council of the Municipality in which the streets are situate shall hereafter be liable for all cost of any maintenance, repair, alteration or reconstruction thereof.
PART OF CROWN ALLOTMENT 49
PARISH OF TRARALGON
COUNTY OF BULN BULN
L.P. 97154
And the Honourable Jeffrey Gibb Kerrett, Her Majesty's Minister of Housing for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
HOUSING ACT 1958
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
DECLARATION OF PRIVATE STREETS AS PUBLIC HIGHWAYS WITHIN THE CITY OF SHEPPARTON
Whereas pursuant to section 107 of the Housing Act 1958 it is among other things enacted that where works have been carried out in accordance with an agreement for the full construction of any roads, pavements, culverts or drains made pursuant to the provisions of the said section such roads, pavements, culverts or drains shall be under the care and management of the Municipality which shall be liable for the cost of any maintenance, repair, alteration or reconstruction thereof; and the Governor in Council may by order published in the Government Gazette declare any road so constructed to be a public highway.
And whereas by Order dated 14 November 1978 the Governor in Council consented to an Agreement between the Housing Commission and the City of Shepparton regarding street and drainage construction in Oxley Street, Olympic Avenue, Centre Road, Hugard Court, Sturtel Court, Grace Court, Gillam Street, McLean Court, Jones Court, Lonsdale Court and Lonsdale Court in the Parkside Estate situate in the municipality of the City of Shepparton and the carrying out of the works enumerated in the said Agreement.
And whereas the works of road pavements, culvert and drainage construction have not been carried out in accordance with the said Agreement between the Housing Commission and the City of Shepparton.
Now therefore His Excellency the Governor of the State of Victoria by and with the advice of the Executive Council of the said State and in pursuance of the powers conferred by section 107 of the Housing Act 1958 do hereby recommend that the Housing Commission doth by this order declare the streets more particularly delineated and shown hatched on the plan hereunder to be absolutely dedicated to the public as Public Highways and to be subject of any new municipal by-laws in force and that the Council of the Municipality in which the streets are situate shall hereafter be liable for all cost of any maintenance, repair, alteration or reconstruction thereof.
PART OF CROWN ALLOTMENTS 79 & 80*
PARISH OF SHEPPARTON
COUNTY OF MIGRA
LP. 123795
LP. 123796
And the Honourable Jeffrey Gibb Kennett, Her Majesty's Minister of Housing for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
Drainage of Land Act 1975
STATE RIVERS AND WATER SUPPLY COMMISSION
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
VARIATION OF A FLOOD-PRONE AREA—SALE
Whereas under Section 35 of the Drainage of Land Act 1975, the Governor in Council may at the request of the Drainage Authority by Proclamation published in the Government Gazette declare any specified area or areas of land to be liable to flooding by any river or group of rivers and may by the like proclamation revoke or vary any such proclamation.
And whereas as a result of such request land at Sale below the 1% probability flood level and more particularly detailed on Plan No. 132641 lodged in the Plan Room of the State Rivers and Water Supply Commission at 590 Orrong Road, Armadale was proclaimed on 10 February 1981.
And whereas a request has now been received from the State Rivers and Water Supply Commission to vary the above proclamation to include that area being the Crown Reserve fronting the Thomson and Latrobe Rivers adjacent to the existing proclaimed freehold land.
And whereas under Section 33 of the Act the State Rivers and Water Supply Commission may be cited as the Drainage Authority in relation to the land.
Now therefore the Governor of the State of Victoria with the advice of the Executive Council of the said State and in pursuance of the above recited powers do by this order vary the proclamation of land at Sale as liable to flooding to include all that land being the Crown Reserve fronting the Thomson and Latrobe Rivers adjacent to the existing proclaimed freehold land.
And the Honourable Glyn Jenkins, Her Majesty's Minister of Water Supply for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
SUPERANNUATION ACT 1958
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
Pursuant to the powers conferred by the provisions of paragraph (ia) of subsection (1) of Section 10 of the Superannuation Act 1958, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof doth by this Order declare that the provisions of the Superannuation Act shall apply to the officer named in the schedule set out hereunder.
SCHEDULE
Thomson, Melanie, Gippsland Institute of Advanced Education.
And the Honourable Lindsay Hamilton Simpson Thompson, Her Majesty's Treasurer for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
COUNTRY ROADS BOARD
Act No. 6229.
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
ORDER APPROVING OF LAND BEING ACQUIRED AND ROADS DEVIATIONS OR WIDENINGS BEING MADE TOGETHER WITH ALL ANCILLARY WORKS REQUIRED TO BE EXECUTED IN CONJUNCTION THEREWITH
His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof, being satisfied that there are funds legally available for acquiring such land, hereby approve the acquiring of the land described in the schedule hereto and as the making of new roads and deviations from and widenings of existing roads, together with all ancillary works required to be executed in conjunction therewith, referred to in the said schedule.
SCHEDULE
Widenings of Existing Roads
The land shown hatched on plan numbered G.P.15476 hereunder required for the Buchan-Orbost Road in the Shire of Orbost.
The land shown hatched on plan numbered G.P.15486 hereunder required for Traralgon Creek Road in the Shire of Traralgon.
Deviations from an Existing Road
The land shown hatched on plan numbered G.P.15437 hereunder required for the Stawell-Pomonal Road in the Town of Stawell.
And the Honourable Robert Roy Cameron Maclellan, Her Majesty's Minister of Transport for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
CROWN LAND (RESERVES) ACT 1978
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
CROWN LANDS PERMANENTLY RESERVED
His Excellency the Governor of the State of Victoria, in the Commonwealth of Australia by and with the advice of the Executive Council of the said State, pursuant to the provisions of sections 4 and 7 of the Crown Land (Reserves) Act 1978, are hereby the opinion that the Crown lands contained are required for the purposes respectively mentioned, doth hereby permanently reserve and also except from prospecting or from occupation for mining purposes under any miner's right, the Crown lands hereinafter described, viz:
KORNOONG—For the Management of Wildlife—88 hectares more or less, being Crown allotment 58A, Parish of Kornong, as indicated by hatching on plan hereunder—(£110¹) (Rs:11396).
TERRINALLUM—For the Management of Wildlife—174 hectares, more or less, being Crown allotment 39B, Parish of Terrinallum, as indicated by hatching on plan hereunder—(£137²) (Rs:11385).
And the Honourable William Vasey Houghton, Her Majesty's Minister of Lands for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
CROWN LAND (RESERVES) ACT 1978
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
CROWN LAND SPECIFIED AS BEING PERMANENTLY RESERVED FOR PUBLIC RECREATION
His Excellency the Governor of the State of Victoria, in the Commonwealth of Australia, by and with the advice of the Executive Council of the said State, pursuant to the provisions of section 4 (5) of the Crown Land (Reserves) Act 1978, doth hereby specify that the land hereinafter described, being permanently reserved for unspecified purposes, is so permanently reserved for Public Recreation, viz:
MODEWARRE—The Crown land comprising 526 hectares, more or less, in the Parish of Modewarre, being the land permanently reserved for Public Purposes by Order in Council of 14 July 1965, published in the Government Gazette dated 21 July 1965 (page 2333)—(Rs:8457).
And the Honourable William Vasey Houghton, Her Majesty's Minister of Lands for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
CROWN LAND (RESERVES) ACT 1978
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
CROWN LANDS PERMANENTLY RESERVED
His Excellency the Governor of the State of Victoria, in the Commonwealth of Australia, and with the advice of the Executive Council of the said State, pursuant to the provisions of sections 4 and 7 of the Crown Land (Reserves) Act 1978, and being of the opinion that the Crown lands contained within the area for the land respectively mentioned, both hereby permanently reserve and also except from prospecting or from occupation for mining purposes under any miner's right, the Crown lands hereinafter described, viz.:
NERRIN NERRIN—For the Management of Wildlife—77 hectares, more or less, being Crown allotment 123A, Parish of Nerrin Nerrin, as indicated by hatching on plan hereunder—(N.86(*) (Rs.11314).
And the Honourable William Vasey Houghton, Her Majesty's Minister of Lands for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
CROWN LAND (RESERVES) ACT 1978
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
LAND DECLARED TO BE PERMANENTLY RESERVED FOR THE PROTECTION OF THE COASTLINE
His Excellency the Governor of the State of Victoria, in the Commonwealth of Australia, by and with the advice of the Executive Council of the said State, pursuant to the provisions of section 4 (6) of the Crown Land (Reserves) Act 1978, doth hereby permanently reserve Crown land hereinafter described to be permanently reserved for the protection of the coastline, viz.:
ANGLESEA AND JAN JUC—The reserved Crown land in the Township of Anglesea and the Township of Jan Juc being the land shown colored blue on plan numbered C.P.R.8 lodged in the Central Plan Office of the Department of Crown Lands and Survey—(A.183(*) (J.7(* and *) (Rs.11041)).
And the Honourable William Vasey Houghton, Her Majesty's Minister of Lands for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
CROWN LAND (RESERVES) ACT 1978
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
CROWN LANDS TEMPORARILY RESERVED
His Excellency the Governor of the State of Victoria, in the Commonwealth of Australia, by and with the advice of the Executive Council of the said State, pursuant to the provisions of sections 4 and 7 of the Crown Land (Reserves) Act 1978, and being of the opinion that the Crown lands contained within the areas on the plans opposite respectively mentioned, doth hereby temporarily reserve and also except from prospecting or from occupation for mining purposes under any miner's right, the Crown lands hereinafter described, viz:
BAMBRA—For Water Supply, 7.317 hectares, being Crown allotments 52A and 54A, Parish of Bambra, as indicated by hatching on plan hereunder—(B.90(*) (Rs.11234).
GERANGAMETE—For Supply of Stone, 4.634 hectares, being Crown allotment 23F, Parish of Gerangamete, as shown on Certificate Plan No. 104143 lodged in the Central Plan Office—(G.170(*) (Rs.11628).
MYALL—For Water Supply, 2.487 hectares, being Crown allotment 23J, Parish of Myall, as indicated by hatching on plan hereunder—(M.585(*) (Rs.11490).
BOORT—For Plantation Purposes, 2694 square metres, being Crown allotment 9, section 1, Township of Boort, as indicated by hatching on plan hereunder—(B.654(*) (Rs.5764).
WANGARATTA—For Public Park and Public Recreation—3.236 hectares, being Crown allotment 5, section 71A, Township of Wangaratta, as shown on Certified Plan No. 105086 lodged in the Central Plan Office—(W.85(*) (Rs.11818).
WANGARATTA—For Public Buildings, 8000 square metres, being Crown allotment 6, section 71A, Township of Wangaratta, as shown on Certified Plan No. 105086 lodged in the Central Plan Office—(W.85(*) (Rs.11816).
WANGARATTA—For the Promotion of Tourism—3449 square metres, being Crown allotment 4, section 71A, Township of Wangaratta, as shown on Certified Plan No. 105086 lodged in the Central Plan Office—(W.85(*) (Rs.11817).
And the Honourable William Vasey Houghton, Her Majesty's Minister of Lands for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
CROWN LAND (RESERVES) ACT 1978
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
CROWN LANDS TEMPORARILY RESERVED
His Excellency the Governor of the State of Victoria, in the Commonwealth of Australia, by and with the advice of the Executive Council of the said State, pursuant to the provisions of sections 4 and 7 of the Crown Land (Reserves) Act 1978, and being of the opinion that the Crown lands concerned are required for the public purposes respectively mentioned, do hereby temporarily reserve and also except from prospecting or from occupation for mining purposes under any miner's right, the Crown lands hereinafter described, viz.:
KRAMBRUK—For Navigational Beacon—400 square metres, being Crown allotment 12A, section 3, Parish of Krambruk, indicated by hatching on plan hereunder—(K.139(12)) (Rs.11458).
And the Honourable William Vasey Houghton, Her Majesty's Minister of Lands for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
CROWN LAND (RESERVES) ACT 1978
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
CROWN LAND PERMANENTLY RESERVED
His Excellency the Governor of the State of Victoria in the Commonwealth of Australia, by and with the advice of the Executive Council of the said State, pursuant to the provisions of sections 4 and 7 of the Crown Land (Reserves) Act 1978, and being of the opinion that the Crown land concerned is required for the purpose mentioned, do hereby permanently reserve and also except from prospecting or from occupation for mining purposes under any miner's right, the Crown land hereinafter described, viz.:
ANGLESEA—For Jan Juc—For the protection of the coastline—the land in the Township of Anglesea and the Parish of Jan Juc as shown colored red on plan numbered C.P.R.8 lodged in the Central Plan Office of the Department of Crown Lands and Survey—(A.183(*) (J.7(*) and 4) (Rs.11041).
And the Honourable William Vasey Houghton, Her Majesty's Minister of Lands for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
LAND CONSERVATION ACT 1970
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
APPOINTMENT OF A MEMBER OF THE LAND CONSERVATION COUNCIL
In pursuance of the provisions of the Land Conservation Act 1970, and all other powers him thereunto enabling His Excellency the Governor of the State of Victoria by and with the advice of the Executive Council thereof doth hereby appoint Thorold Harvey Gunning, being a person who has experience in industry and commerce to be a member of the Land Conservation Council during the period 11 August 1981 to 10 August 1985 both dates inclusive.
And the Honourable William Vasey Houghton, Her Majesty's Minister for Conservation for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
SOIL CONSERVATION AND LAND UTILIZATION ACT 1958
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
APPOINTMENT OF A PERSON TO ACT AS A MEMBER DURING THE ABSENCE OF THE DEPUTY CHAIRMAN
Whereas leave has been granted to Dennis Neville Cahill, a Member and Deputy Chairman of the Soil Conservation Authority, to be absent from the State of Victoria so that he may act as a consultant in India to the Food and Agriculture Organization of the United Nations. And whereas provision is made in the Soil Conservation and Land Utilization Act 1958 for the appointment for a period of not more than six months of a qualified person to act in the place of any member who is absent from Victoria. Now therefore I, the Governor of the said State by and with the advice of the Executive Council thereof, do hereby appoint James Frederick Allinson to act as a Member of the Soil Conservation Authority for a period of six months from and inclusive of 16 August 1981, or until the return to Victoria of Dennis Neville Cahill, whichever first happens.
And the Honourable William Vasey Houghton, Her Majesty's Minister for Conservation for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
SOIL CONSERVATION AND LAND UTILIZATION ACT 1958
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
APPOINTMENT OF A MEMBER TO ACT AS DEPUTY CHAIRMAN
Whereas leave has been granted to Dennis Neville Cahill, a Member and Deputy Chairman of the Soil Conservation Authority, to be absent from the State of Victoria so that he may act as a consultant in India to the Food and Agriculture Organization of the United Nations. And whereas provision is made in the Soil Conservation and Land Utilization Act 1958 for the appointment by the Governor in lieu of one of the Members of the Soil Conservation Authority to act as Deputy Chairman if the Deputy Chairman is so absent. Now therefore I, the Governor of the said State by and with the advice of the Executive Council thereof, do hereby appoint John Samuel Chilcott to act as Deputy Chairman of the Soil Conservation Authority for a period of six months from and inclusive of 16 August 1981, or until the return to Victoria of Dennis Neville Cahill, whichever first happens.
And the Honourable William Vasey Houghton, Her Majesty's Minister for Conservation for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
RAILWAY CONSTRUCTION AND PROPERTY BOARD ACT 1979
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
DEVELOPMENT OF RAILWAY LAND
Whereas it is provided by section 20 (4) of the Railway Construction and Property Board Act 1979 that the Railway Construction and Property Board may make recommendations to the Minister for or with respect to the use or development of railway land.
And whereas it is provided by section 20 (5) of the said Act that when the Board has made recommendations to the Minister under section 20 (4) the Minister may submit the recommendations to the Governor in Council for approval.
And whereas the Board has made the following recommendation in respect of Railway land situated at Culgoa.
RECOMMENDATION
That the land which comprises a large excavation used in the past by the RACV for water storage is surplus to railway requirements but could be used for swimming purposes by residents of Culgoa and district.
Now therefore, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof, doth hereby approve of the foregoing recommendation.
And the Honourable Robert Roy Cameron Maclellan, Her Majesty's Minister of Transport for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
VARIATION OF THE POWERS OF THE HOSPITAL AND BENEVOLENT HOMES BOARD AND THE RESIDENTIAL CHILD CARE WORKERS BOARD
In pursuance of the powers conferred by section 20 of the Labour and Industry Act 1958, and all other powers him thereunto enabling, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof, doth hereby—
1. Vary the powers of the Hospital and Benevolent Homes Board so that in substitution for the powers heretofore conferred on the said Hospital and Benevolent Homes Board, it shall have the power to determine any industrial matter in relation to the process, trade, business or occupation of—
(1) a person or persons or class or classes of persons employed in (or about and) doing any kind of work connected with the carrying on of—
(a) a hospital;
(b) a hospital, home or service for the care or treatment of intellectually or physically handicapped persons;
(c) a sanatorium for the treatment of persons suffering from any form of tuberculosis;
(d) a hostel giving residential care, including a benevolent home, a special accommodation house, a convalescent home, an orphanage, a babies' home and a children's home (but excluding a private dwelling home);
(e) a laboratory attached to a hospital;
(f) a bank, a bank's radiography, diagnostic or pathological clinic or service;
(g) a medical, dental or nursing clinic or service; or
(h) an ambulance service by any hospital or hospitals.
(2) a first-aid attendant employed in connection with an industrial or commercial undertaking: but excluding—
(i) any professional employee;
(ii) any child care or family welfare worker, executive or superintendent employed in connection with premises providing residential care for children;
(iii) any person employed in a child minding centre of the type required to be registered under the Health Act 1958;
(iv) any occupation subject to any other Wages Board.
2. Vary the powers of the Residential Child Care Workers Board so that in substitution for the powers heretofore conferred on the said Residential Child Care Workers Board, it shall be appointed for the occupation of any child care or family welfare workers, executives and superintendents, employed in connection with premises providing residential care for children up to and including 17 years of age, but excluding foster parents and any occupation subject to any other Wages Board.
And the Honourable James Halford Ramsay, Her Majesty's Minister of Labour and Industry for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
PLASTER OF PARIS BOARD—NUMBER OF MEMBERS REDUCED
In pursuance of the powers conferred by section 22 of the Labour and Industry Act 1958, and all other powers him thereunto enabling, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof, doth hereby vary the Order made on 2 February 1977, insofar as ordered that the Plaster of Paris Board should consist of eight members and a Chairman and doth hereby order that the said Plaster of Paris Board shall consist of six members and a Chairman.
And the Honourable James Halford Ramsay, Her Majesty's Minister of Labour and Industry for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
VICTORIA LAW FOUNDATION ACT 1978
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
APPOINTMENT OF A MEMBER OF THE VICTORIA LAW FOUNDATION
In pursuance of the provisions of section 4 of the Victoria Law Foundation Act 1978, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof, doth by this Order appoint the CAMERON FUTRELL being a nominee of the Law Institute of Victoria, to be a member of the Victoria Law Foundation for the period of five years commencing on 11 August 1981.
And the Honourable Haddon Storey, Her Majesty's Attorney-General for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
FORESTS ACT 1958, No. 6254
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
ORDER SETTING ASIDE AND DECLARING THE LAKE ELIZABETH SCENIC RESERVE
Under the powers conferred by the Forests Act and all other powers enabling him in that behalf, His Excellency the Governor of the State of Victoria, by and with
the advice of the Executive Council of the said State, doth by this order set aside and declare the area of Reserved Forest described in the Schedule hereto to be a scenic reserve known as the Lake Elizabeth Scenic Reserve.
SCHEDULE
The Reserved Forest in the Parish of Kaanglang containing 99 hectares shown by shading on the plan hereunder.
And the Honourable Thomas Leslie Austin, Her Majesty's Minister of Forests for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
LOCAL GOVERNMENT DEPARTMENT
At the Executive Council Chamber, Melbourne, the eighteenth day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Crozier | Mr Dixon
Mr Lacy
APPOINTMENT OF MEMBERS OF THE LOCAL AUTHORITIES SUPERANNUATION BOARD
Pursuant to the provisions of section 20 of the Local Authorities Superannuation Act 1958, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof, doth by this Order appoint the following persons to be members of the Local Authorities Superannuation Board for the period ending on 21 August 1986:
WILLIAM THWAITES nominated by the governing body of the Municipal Association of Victoria, who shall be Chairman;
FREDERICK JAMES MADDERN nominated by the governing body of the Municipal Association of Victoria;
HAROLD JAMES HOPKINS nominated by the Treasurer of Victoria;
GARY ARTHUR WEAVER nominated by the governing body of the Municipal Officers' Association of Australia, Victoria State Branch; and
PAUL SLAPE nominated by the governing body of the Federated Municipal and Shire Council Employees' Union of Australia, Victoria Division.
And the Honourable Louis Stuart Lieberman, Her Majesty's Minister for Local Government for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
INDUSTRIAL TRAINING ACT 1975
At the Executive Council Chamber, Melbourne, the eleventh day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Houghton | Mr Ramsay
Mr Austin
APPOINTMENT OF DEPUTY CHAIRMAN AND ACTING DEPUTY CHAIRMAN OF TRADE COMMITTEES
In pursuance of the provisions of section 13 of the Industrial Training Act 1975, and all other powers him thereunto enabling, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof, doth by this Order appoint
KENNETH GILFRETH an officer of the Public Service, being a suitable person, to be Deputy Chairman of Trade Committees; and
WILLIAM FORDMAN an officer of the Teaching Service, being a suitable person, to be Acting Deputy Chairman of Trade Committees—
for a period of six months from and including 12 August 1981.
And the Honourable Brian James Dixon, Her Majesty's Minister for Employment and Training for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
SOIL CONSERVATION AND LAND UTILIZATION ACT 1958
At the Executive Council Chamber, Melbourne, the eighteenth day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Crozier | Mr Dixon
Mr Lacy
APPOINTMENT OF A MEMBER OF THE SOIL CONSERVATION AUTHORITY
In pursuance of the powers conferred by the Soil Conservation and Land Utilization Act 1958, and all other powers him thereunto enabling, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council thereof doth by this Order appoint JOHN SAMUEL GILMORE to be a member of the Soil Conservation Authority during the period from and including 21 August 1981 up to and including 20 August 1984.
And the Honourable William Vasey Houghton, Her Majesty's Minister for Conservation for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
GEELONG WATERWORKS AND SEWERAGE TRUST
At the Executive Council Chamber, Melbourne, the eighteenth day of August, 1981
PRESENT:
His Excellency the Governor of Victoria
Mr Crozier
Mr Lacy
Mr Dixon
CONSENT TO BORROWING $1 700 000
Under the powers conferred by the Geelong Waterworks and Sewerage Act and all other powers enabling him to that behalf, His Excellency the Governor of the State of Victoria, by and with the advice of the Executive Council of the said State, hereby consents to the Geelong Waterworks and Sewerage Trust borrowing from the Crown One million seven hundred thousand dollars ($1 700 000), two separate loans of One million five hundred thousand dollars ($1 500 000) and Two hundred thousand dollars ($200 000) to meet the cost of water supply works.
And the Honourable Glyn Jenkins, Her Majesty's Minister of Water Supply for the State of Victoria, shall give the necessary directions herein accordingly.
TOM FORRISTAL
Clerk of the Executive Council
LANDS DEPARTMENT NOTICES
APPROACHING LAND SALES
Sale of Crown Land, will be held at the under-mentioned places and dates, viz.:—
| No. of Gazette | Date |
|----------------|--------------------|
| Bacchus Marsh | Thursday, 8 October 1981 |
| Daylesford | Thursday, 15 October 1981 |
| Red Cliffs | Friday, 8 November 1981 |
SALE OF CLOSER SETTLEMENT LAND BY AUCTION
Red Cliffs—Friday, 23 October 1981
PROPOSED REVOCATION OF TEMPORARY RESERVATION OF LAND BY ORDER IN COUNCIL
In pursuance of the provisions of section 10 of the Crown Land (Reserves) Act 1978, and is hereby given, that it is the intention of the Governor in Council to revoke temporary reservation of land by Order in Council hereunder referred to, viz.:
FRAMLINGHAM EAST—The temporary reservation as a site for Public purposes and the withholding from sale, leasing and licensing by Order in Council of 3 May, 1875 of 8.498 hectares more or less of land in the Parish of Framlingham East (north of allotment 39)—(F.52(*) (C.17281).
W. V. HOUGHTON
Minister of Lands
Department of Crown Lands and Survey
Melbourne
PUBLIC WORKS DEPARTMENT
Tenders are invited for the purposes indicated hereunder and will be received at 2 Treasury Place, Melbourne, 3002, until TWO p.m. on dates shown.
Tenders must be addressed to the Minister of Public Works with the envelope marked "Tender for ...". Hand-delivered tenders must be placed in the Department's tender box in the foyer, Ground Floor, 2 Treasury Place, Melbourne.
Tenders documents are available from the Contracts Office, Room 29, 2 Treasury Place, Melbourne and where indicated, at offices of Inspectors of Works.
★Tenders for works identified with star (*) will only be considered from Contractors qualified for repairs and painting works. Any Tenderers wishing to qualify may obtain the application form from the Contracts Office.
Tuesday, 25 August 1981
Building, Electrical and Mechanical Works
LAVERTON—(Re-advertisement)—Provision of fumigating and drying chamber, Public Records Office Repository.
MOE—Erection of family group home, Mental Health Division. (W.O., Traralgon and Warragul.)
Miscellaneous
BALLARAT—Maintenance cleaning, period 1 October 1981 to 30 September 1984, State Government Offices. (W.O., Ballarat.)
BOX HILL—Supply of chairs and tables, Technical College.
PORT MELBOURNE—Purchase of one (1) only thirty tonne tri-axle low loader, Depot.
SOUTH MELBOURNE—Maintenance cleaning, period 1 September 1981 to 31 August 1984, Court House.
TERANG—Maintenance cleaning, period 1 October 1981 to 30 September 1983, Agriculture Department. (W.O., Camperdown.)
VARIOUS—Removal of waste from Government buildings, period 31 August 1981 to 31 August 1983, Various.
WESTERN REGION, MELBOURNE—Window cleaning, period 1 September 1981 to 31 August 1982, Various.
Tuesday, 1 September 1981
Building, Electrical and Mechanical Works
MELBOURNE—Alterations and fitting out works, Public Buildings, 140 Bourke Street.
MELBOURNE—Fitting out works on Levels 21, 22 and 23, Nourse House.
MOE—(Re-advertisement, amended specification)—Building works, Stage 2 tenancy accommodation for Health Commission, Enterprise House.
SYNDAL—Energy conservation modifications, Technical School.
WILLIAMSTOWN—Installation of intruder detection system, High School.
Miscellaneous
MOE—Maintenance cleaning, period 1 October 1981 to 30 September 1984, Community Welfare Services. (W.O., Traralgon.)
RINGWOOD—Maintenance cleaning, from 1 September 1981 to 31 August 1984, Community Welfare Services.
TRARALGON—Maintenance cleaning, period 1 October 1981 to 30 September 1984, Agriculture. (W.O., Traralgon.)
Tuesday, 8 September 1981
Building, Electrical and Mechanical Works
ALTONA NORTH—Internal and external painting and repairs, Technical School.
JANEFIELD—Fire escape stairs to Wards F2, F3 and F4, Training Centre.
ST. ALBANS—(Re-advertisement, amended specification)—External and internal renovations, painting and repairs, building works, Primary School.
Miscellaneous
MELBOURNE—Maintenance cleaning, period 1 November 1981 to 31 October 1984, Public Offices, 1 Treasury Place.
ALAN WOOD
Minister of Public Works
Public Works Department
Melbourne, 18 August 1981
CITY OF BALLAARAT
NOTICE OF INTENTION TO TAKE LAND COMPULSORILY
Whereas the Council of the City of Ballaarat deems it expedient to exercise its powers of taking compulsorily the lands described hereunder for the work or undertaking of providing public open space located within the City Baths Reserve and whereas the Council has caused to be prepared maps and other papers setting out the general description of the work or undertaking for which the lands proposed to be taken are to be used, the description of the lands proposed to be taken and the names of the owners or reputed owners, lessees or reputed lessees, mortgagees and occupiers of these lands so far as these names are known to or can be ascertained by the Council and whereas the said maps and other papers are deposited at the office of the City Council at Ballaarat and shall be open for inspection by all persons interested at all reasonable hours for the space of forty clear days after the publication of this notice in the Government Gazette now notice is hereby given to all persons affected by the proposed taking of the said lands to set forth in writing addressed to the Council or the Town Clerk within forty clear days of the publication of this notice in the Government Gazette all objections which they may have to the taking of the said lands.
The Land Referred To:—
All that piece of land being Crown Allotment 8a Section 11 Township of Ballaarat North, Parish of Ballaarat, County of Grenville and being land more particularly described and delineated in Crown Grant Vol. 381, Fol. 059.
Dated 11 August 1981
By order of the Council
4793 I. C. SMITH, Town Clerk
CITY OF BROADMEADOWS
LOAN No. 125
Notice of Intention to Borrow the Sum of $500 000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the City of Broadmeadows proposes to borrow the principal sum of $500 000 secured by a charge over the general rates of the municipality, such sum to be raised by the grant of a mortgage in accordance with the provisions of the Local Government Act 1958.
1. The maximum rate of interest that may be paid is 14-7 per centum per annum.
2. The purpose for which the loan is to be applied is:
| Description | Amount |
|----------------------------------------------------------------------------|--------|
| Reconstruction of Glenroy Road—Pascoe Vale Road to Argyle Street (balance of provision) | 100 000 |
| C.R.B.—Main Roads— | |
| (i) Construction of Bridge—Mickleham Road (Hackett Street) | 4 500 |
| (ii) Construction of Mickleham Road, Hackett Street | 21 000 |
| C.R.B.—Unclassified Roads— | |
| (i) Reconstruction of Camp Road (Merlyn's Creek to Sydney Road) | 30 000 |
| (ii) Reconstruction of Johnston Street (Pearcedale Parade to Ripplebrook Drive) | 110 000 |
| Reconstruction of View Street (completion) | 35 000 |
| Reconstruction of Beckett Street | 52 000 |
| Reconstruction of Anselm Grove | 38 000 |
| Main Town Hall roof replacement (balance of provision) | 87 500 |
| Electrical wiring Town Hall (part provision) | 22 000 |
Total: 500 000
3. The period of the loan shall be fifteen (15) years.
4. The moneys borrowed shall be repayable by providing out of the Municipal Fund half-yearly instalments of approximately $41 719.06 each, payable on the first and last days of May and November during the currency of the loan. The first instalment shall be payable on 1 May 1982.
5. Such moneys shall be repayable to the Commissioners of The State Bank of Victoria at the Office of the said Commissioners, 180 Collins Street, Melbourne.
The plans and specifications and the estimate of the cost of the proposed works and a statement showing the proposed expenditure of the moneys to be borrowed are open for inspection at the Offices of the Council, Pascoe Vale Road, Broadmeadows.
4807 BRUCE D. McGREGOR, Town Clerk
CITY OF BRUNSWICK
LOAN No. 119
Notice of Intention to Borrow the Sum of $200 000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the City of Brunswick proposes to borrow the principal sum of $200 000 secured by a charge over the general rates of the municipality, such sum to be raised by the grant of a mortgage in accordance with the provisions of the Local Government Act 1958.
1. The maximum rate of interest that may be paid is 14-7 per cent. per annum.
2. The purpose for which the loan is to be applied is:
Electricity Supply Department—Works and Stores Depot, Stewart Street (part cost)—$200 000.
3. The period of the loan shall be 10 years.
4. The loan shall be repaid by the creation of a sinking fund and an amount of $15 173.59 being set aside annually out of the municipal fund for such purpose as determined by the State Auditor General.
5. Such moneys shall be repayable to State Insurance, 49 Collins Street, Melbourne.
The plans and specifications and the estimate of the cost of the proposed works and a statement showing the proposed expenditure of the moneys to be borrowed are open for inspection at the office of the Council of the City of Brunswick, at 233 Sydney Road, Brunswick.
K. D. WILSON, Town Clerk
Town and Country Planning Act 1961
CITY OF COLAC PLANNING SCHEME 1963
Notice that a Planning Scheme has been Prepared and is Available for Inspection
Amendment No. 16
Notice is hereby given that the City of Colac in pursuance of its powers under the Town and Country Planning Act 1961 has prepared a Planning Scheme for the purpose of rezoning the following land—
(a) Area west of Forest Street between Barongarook Creek and Hearn Street—Present zoning Agricultural, proposed zoning Industrial "A" and Public Open Space proposed—Stream marginal reserve.
(b) Area south of Wilson Street east of Barongarook Creek—Present zoning Industrial "A", proposed zoning Residential and Public Open Space proposed—Stream marginal reserve.
(c) No. 1 Bromfield Street—present zoning Public Purposes, proposed zoning Commercial "B".
(d) Area west of Barongarook Creek north of Dennis Street—Present zoning Public Open Space—Stream marginal reserve, proposed zoning Public Purposes Reserves Existing—Scout Hall.
(e) Area south of Hearn Street west of Talbot Street—Present zoning—Industrial "A", proposing zoning Agricultural.
(f) Old railway line south of Hearn Street—Present zoning Public Purposes proposed, proposed zoning Public Purposes Existing.
(g) No. 53 Murray Street East—Present zoning Residential—Proposed zoning Public Purposes Reserves Existing—Primary School.
A copy of the Scheme has been deposited at the Municipal Offices, 4-6 Rae Street, Colac, and at the offices of the Department of Planning, Victoria, 235 Queen Street, Melbourne, and will be open for inspection during office hours by any person free of charge.
Any persons affected by the Scheme are required to set forth in writing any submissions they may wish to make with regard to the Scheme addressed to the Town Clerk, City of Colac, Municipal Offices, 4-6 Rae Street, Colac, 3250, by 19 November 1981, and to state whether they wish to be heard in respect of their submissions.
6 August 1981
B. R. GREAVES, Town Clerk
CITY OF DONCASTER AND TEMPLESTOWE
LOAN NO. 239
Notice of Intention to Borrow the Sum of $250 000 for Permanent Works and Undertakings
Notice it hereby given that the Council of the City of Doncaster and Templestowe proposes to borrow the principal sum of $250 000 secured by the grant of a mortgage in accordance with the provisions of the Local Government Act 1958.
1. The maximum rate of interest that may be paid is 14-7 per cent per annum.
2. The purpose for which the loan is to be applied is:
(i) Road Construction—
(a) Beverley Street private street construction—Council contribution (part cost) 140 000
(b) Balwyn Road (Manningham Road to Nirvana Crescent) private street construction—Council contribution (part cost) 10 000
(c) Atkinson Street (Williamsons Road to Anderson Street). Full construction (part cost) 15 000
(ii) Bulleen Park, pavilion (part cost) 35 000
(iii) Bulleen Park—Provision of services associated with pavilion 50 000
3. The period of the loan shall be fifteen years.
4. The moneys borrowed shall be repayable by providing out of the municipal fund half-yearly instalments of approximately $20 859.52 each including principal and interest on 1 October and 1 April during the currency of the loan. The first instalment shall be payable on 1 April 1982.
5. Such moneys shall be repayable to The National Bank of Australia, 900 Doncaster Road, Doncaster East.
The plans and specifications and the estimate of the cost of the proposed works and a statement showing the proposed expenditure of the moneys to be borrowed are open for inspection at the Municipal Offices at 699 Doncaster Road, Doncaster, during office hours.
G. J. MORTON, Town Clerk—Manager Administration
CITY OF ESSENDON
LOAN NO. 148
Notice of Intention to Borrow the Sum of $191 000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the City of Essendon proposes to borrow the principal sum of One hundred and ninety-one thousand dollars ($191 000) secured by a charge over the general rates of the municipality, such charge to be raised by the grant of a mortgage in accordance with the provisions of the Local Government Act 1958 (as amended).
1. The maximum rate of interest that may be paid is 14-70 per cent per annum.
2. The purpose for which the loan is to be applied is as follows:
(A) Depot Construction Completion of Stage 2 60 000
(B) Golf Links Development 131 000
191 000
3. The period of the loan shall be 20 years.
4. The moneys borrowed shall be repayable by providing out of the municipal fund half-yearly instalments of approximately $14 912.41 each including principal and interest on 1 October and 1 April during the currency of the loan. The first instalment shall be payable on 1 April 1982.
5. Such moneys shall be repayable to The Commissioners of the State Bank of Victoria at the office of the said Commissioners, 385 Bourke Street, Melbourne.
The plans and specifications and the estimate of the cost of the proposed works and undertakings showing the proposed expenditure of the moneys to be borrowed are open for inspection at the office of the Council of the City of Essendon, State of Victoria at the Essendon Civic Centre, corner of Kellaway Avenue and Pascoe Vale Road, Moonee Ponds.
B. C. BEATTIE, City Manager/Town Clerk
CITY OF ESSENDON
LOAN No. 149
Notice of Intention to Borrow the Sum of $65,000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the City of Essendon proposes to borrow a principal sum of Sixty-five thousand dollars ($65,000) secured by a charge over the general rates of the municipality, such sum to be raised by the grant of a mortgage in accordance with the provisions of the Local Government Act 1958.
1. The maximum rate of interest that may be paid is 14-7 per cent per annum.
2. The purpose for which the loan is to be applied is as follows:
For and towards the cost of:
| Description | Amount |
|--------------------------------------------------|--------|
| (a) Clifton Park Pavilion (Aberfeldie Football Club) | 45,000 |
| (b) Clifton Park Bowling Club (Clubhouse building Alterations) | 20,000 |
3. The period of the loan shall be 10 years.
4. The moneys borrowed shall be repayable by providing out of the municipal fund half-yearly instalments of approximately $3,034.44 each, including all interest on 1 April and 1 October during the currency of the loan. The first instalment shall be payable on 1 April 1982.
5. Such moneys shall be repayable to The Commissioners of The State Bank of Victoria at the office of the said Commissioners, 180 Bourke Street, Melbourne.
The plans and specifications and the estimate of the cost of the proposed works and a statement showing the proposed expenditure of the moneys to be borrowed are open for inspection at the office of the Council of the City of Essendon, Suite 10, Victoria at the Essendon Civic Centre, located in Kellaway Avenue and Pascoe Vale Road, Moonee Ponds 3039.
B. C. BEATTIE, City Manager/Town Clerk
CITY OF FRANKSTON
LOAN No. 211
Notice of Intention to Borrow the Sum of $80,000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the City of Frankston proposes to borrow the sum of $80,000 secured by a charge over the general rates of the Municipality, such sum to be raised by the grant of a mortgage in accordance with the provisions of the Local Government Act 1958.
In connection therewith the following information is stated:
(a) The amount of principal which it is proposed to borrow is $80,000.
(b) The maximum rate of interest that may be paid is 14-7 per cent per annum.
(c) The times when moneys borrowed are to be repayable are 10 April and 10 October during currency of the Loan, the first instalment being payable on 10 April 1981.
(d) The place such moneys shall be repayable is at the office of the Commercial Savings Bank of Australia Limited, Nepean Highway, Frankston.
(e) The purpose for which the loan is to be applied is:
Town Planning—Land Purchase.
(f) The manner in which the loan is to be liquidated is by provision out of the Municipal fund of 30 half-yearly instalments of approximately $6,675.05 including principal and interest.
The plans and specifications and estimates of the cost of the works, and a statement showing the proposed expenditure of the moneys to be borrowed are open for inspection at the Civic Centre, Frankston.
A. H. BUTLER, Town Clerk
Civic Centre, Frankston
CITY OF ST. KILDA
BY-LAW No. 209
A By-Law of the City of St. Kilda made and passed under the provisions of the Local Government Act 1958 and numbered 209 for amending By-Law No. 205 of the said City.
In pursuance of the powers conferred by the Local Government Act 1958 and any and every power thereunto enabling the mayor, councillors and citizens of the City of St. Kilda do enact as follows:
By-Law No. 205 of the City of St. Kilda shall be altered as follows:—
In clause 3 relating to Order of Business for the words commencing with "(i) Apologies" and concluding with "(xi) General and Urgent business" there shall be substituted the words:—
(i) Opening Prayer.
(ii) Apologies.
(iii) Declarations by Councillors of any direct or indirect pecuniary interest in any items on the Agenda.
(iv) Confirmation of the Minutes of previous meeting and the signing thereof.
(v) Consideration of Inward Correspondence.
(vi) Presentation of reports and recommendations by Council committees.
(vii) Reports by delegates appointed by the Council to other bodies.
(viii) Questions with Notice to Senior Officers.
(ix) Reception and reading of petitions, joint letters and memorials and deputations.
(x) Orders of the Day.
(xi) Motions of which notice has been given in accordance with the provisions of the Local Government Act."
This By-Law shall come into operation on its publication in the Victoria Government Gazette.
Resolution for passing this By-Law agreed to by the Council of the City of St. Kilda on 25 May 1981, and confirmed on 6 July 1981.
The common seal of the Mayor, Councillors and Citizens of the City of St. Kilda has been hereto affixed by Order of the Council of the said City, in the presence of:
C. J. J. KING, Mayor
M. L. JELBART, Councillor
W. A. SISSON, Town Clerk
Town and Country Planning Act 1961 (Twelfth Schedule)
CITY OF SHEPPARTON—CITY OF SHEPPARTON PLANNING SCHEME 1953
NOTICE THAT A PLANNING SCHEME HAS BEEN PREPARED AND IS AVAILABLE FOR INSPECTION
Amendment No. 57, 1981
Notice is hereby given that the Council of the City of Shepparton in pursuance of its powers under the Town and Country Planning Act 1961, has prepared a planning scheme for the purpose of:
Rezoning, from Residential to Commercial "B" and Commercial "A" property at the south-east corner of Numurkah Road and Pine Road, providing a 10 metre road widening on the east side of Numurkah Road, for a distance of 54-88 metres north of Pine Road.
A copy of the scheme has been deposited at the office of the Council, Municipal Offices, Civic Centre, Welsford Street, Shepparton, and also at the office of the Department of Planning, 500 Collins Street, Melbourne, and will be open for inspection during office hours by any person free of charge.
Any persons affected by the Scheme are required to set forth in writing any submissions they may wish to make with respect to the Scheme, addressed to the Town Clerk, City of Shepparton, P.O. Box 989, Shepparton, 3630, by 20 November 1981, and to state whether they wish to be heard in respect of their submissions.
I. L. GILBERT, Town Clerk
Local Government Act 1958
CITY OF SPRINGVALE
Notice of Intention to Take Land Compulsorily
Notice is given that it is the intention of the Council of the City of Springvale to exercise in pursuance of the powers conferred on it by the Local Government Act 1958 to take compulsorily the following land—all that piece of land being a 2-40 m by 2-40 m splay from the south-west corner of Parcel No. 5 Plan of Subdivision 6156 and being more particularly described in Certificate of Title Volume 6281 Folio 014.
The said land is required for and being taken for the following work or undertaking by the said Council—the proposed work or for removing the fence at the southwest corner of the property to enable better traffic movement around the Noble Park Elderly Citizens Club within its municipal district.
The Council has caused to be prepared map and other papers showing the position and extent of such work or undertaking, and more particularly describing the said land and the exact site and measurements thereof and stating that the names of the owners or reputed owners of the said land are Carmelo A. Mastoran and Maria Tripi, Shop 14a Women's Club of 15 Stuart Street, Noble Park, and the names of the occupiers thereof are A Mastoran and G. Giannoulakis of Shop 14a Leonard Avenue, Noble Park, and F. Sacco and E Bus of Shop 14a Leonard Avenue, Noble Park, and the name of the mortgagee is the Commercial Banking Company of Sydney Limited of 257 Collins Street, Melbourne.
The said map and other papers have been approved by the Council and are now deposited for inspection by all persons interested at the offices of the City of Springvale situate at Springvale Road, Springvale, and may be inspected there during office hours.
All persons affected by the proposed work or undertaking are hereby required to set forth in writing any objections they may wish to make to the Municipal Clerk within forty (40) clear days from the publication of this notice in the Government Gazette, all objections which they may have to such work or undertaking.
Dated 19 August 1981
By order of the Council
K. D. MOODY, Town Clerk
CITY OF WILLIAMSTOWN
Loan No. 88
Notice of Intention to Borrow the Sum of $400 000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the City of Williamstown proposes to borrow the sum of Four hundred thousand dollars ($400 000) secured by a charge over the general rates of the Municipality, such sum to be raised by the grant of a mortgage in accordance with the provisions of the Local Government Act 1958 (as amended).
1. The maximum rate of interest that may be paid is 14-7 per cent per annum.
2. The purpose for which the loan is to be applied:
| Description | Amount |
|------------------------------------|----------|
| (a) purchase of land Newport Quarry| 260 000 |
| (b) Mason Street drainage works | 140 000 |
3. The period of the loan shall be ten years.
4. The moneys borrowed shall be repayable by providing out of the Municipal Fund twenty half-yearly instalments of $38 790.38 each, including principal and interest on 23 September and on 23 March during the currency of the loan. The first instalment shall be payable on 23 March 1982.
5. Such moneys shall be repayable at the Commonwealth Savings Bank of Australia, Williamstown.
6. The plans and specifications and the estimate of the cost of the proposed works and a statement showing the proposed expenditure are open for inspection at the office of the Council of the City of Williamstown, Municipal Offices, 104 Ferguson Street, Williamstown.
J. T. SLOAN, Town Clerk
BOROUGH OF EAGLEHAWK
Loan No. 57
Notice of Intention to Borrow the Sum of $80 000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the Borough of Eaglehawk proposes to borrow the principal sum of Eighty thousand dollars ($80 000) secured by a charge over the general rates of the Municipality, such sum to be raised by the grant of a mortgage in accordance with the provisions of the Local Government Act 1958.
1. The maximum rate of interest to be paid is 14-50 per centum per annum.
2. The purpose for which the loan is to be applied is for the purchase of the property at 80–82 Victoria Street for use as municipal offices.
3. The period of the loan shall be four (4) years.
4. The moneys borrowed shall be repayable by providing out of the Municipal Fund seven half-yearly instalments of $6178.66 each including principal and interest on 20 April and 20 October during the currency of the loan, and a final half-yearly instalments of $82 256.59. The first instalment shall be payable on 20 April 1982.
5. The moneys shall be payable at the C.B.C. Savings Bank Limited, Melbourne.
6. The plans and specifications and the estimates of the cost of the proposed works and a statement showing the proposed expenditure of the money to be borrowed are open for inspection at the office of the Council of the Borough of Eaglehawk, at the Town Hall, Eaglehawk.
Dated 11 August 1981
A. J. SMARK, Town Clerk
Twelfth Schedule
PORT FAIRY PLANNING SCHEME 1959
BOROUGH OF PORT FAIRY
Notice that a Planning Scheme has been Prepared and is Available for Inspection
Amendment No. 15—1981
Notice is hereby given that the Council of the Borough of Port Fairy in pursuance of its powers under the Town and Country Planning Act 1961 has prepared a planning scheme for the area on the east side of Sackville Street, Port Fairy between Barron and Regent Streets, more commonly known by the M.U.I.O.Q.F. Hall. The purpose of the scheme is to rezone the land from Special Use to Residential.
A copy of the scheme has been deposited at the Borough Chambers, Barron Street, Port Fairy and at the office of the Department of Planning, 235 Queen Street, Melbourne, and will be open for inspection during office hours by any person free of charge.
Any persons affected by the scheme are required to set forth in writing any submissions they may wish to make with respect to the scheme addressed to the Town Clerk, Borough of Port Fairy, P.O. Box 51 Port Fairy, by 21 September 1981, and to state whether they wish to be heard in respect of their submissions.
IAN C. PATIENCE, Town Clerk
Twelfth Schedule
PORT FAIRY PLANNING SCHEME 1959
BOROUGH OF PORT FAIRY
Notice that a Planning Scheme has been Prepared and is Available for Inspection
Amendment No. 16—1981
Notice is hereby given that the Council of the Borough of Port Fairy in pursuance of its powers under the Town and Country Planning Act 1961 has prepared a planning scheme for the north-west corner of the intersection of Cox and Giles Streets, Port Fairy. The purpose of the scheme is to rezone the land from Railway Purposes to Residential.
A copy of the scheme has been deposited at the Borough Chambers, Barron Street, Port Fairy and at the office of the Department of Planning, 235 Queen Street, Melbourne, and will be open for inspection during office hours by any person free of charge.
Any persons affected by the scheme are required to set forth in writing any submissions they may wish to make with respect to the scheme addressed to the Town Clerk, Borough of Port Fairy, P.O. Box 51, Port Fairy, by 19 November 1981, and to state whether they wish to be heard in respect of their submissions.
IAN C. PATIENCE, Town Clerk
SHIRE OF EUROA
APPOINTMENT OF POUNDKEEPER
Notice is hereby given that Ronald Alfred Saunders of 10 Graham Street, Euroa, was appointed Poundkeeper by Council resolution on 18 February 1980.
B. MORAN, Shire Secretary
SHIRE OF GLENELG
APPOINTMENT OF PROSECUTING OFFICER
Notice is hereby given that Sergeant Terry Graeme Kiely, No. 15635, has been appointed Prosecuting Officer for the Shire of Glenelg in lieu of Sergeant William Arthur Johnson, No. 10342, retired.
D. A. TICKELL, Shire Secretary
Town and Country Planning Act 1961 (Twelfth Schedule)
SHIRE OF HEALESVILLE—SHIRE OF HEALESVILLE
INTERIM DEVELOPMENT ORDER 1976
NOTICE THAT AN AMENDMENT TO AN INTERIM DEVELOPMENT ORDER HAS BEEN PREPARED AND IS AVAILABLE FOR INSPECTION
Amendment No. 3
Notice is hereby given that the Shire of Healesville in pursuance of its powers under the Town and Country Planning Act 1961 has prepared an amendment to the Shire of Healesville Interim Development Order 1976 for the purpose of reserving land for proposed public open space to provide for the extension of Gallipoli Park, Marysville.
A copy of the amendment has been deposited at the Shire Office, 231 Maroondah Highway, Healesville, at the Upper Yarra Valley and Dandenong Ranges Authority Old Melbourne Road, Lilydale West and at the Department of Planning, 235 Queens Street, Melbourne and will be open for inspection during office hours by any person free of charge.
Any persons affected by the Scheme are required to set forth in writing any submissions they may wish to make with respect to the Scheme addressed to the Shire Secretary, Shire of Healesville, 231 Maroondah Highway, Healesville Vic. 3777, by 25 September 1981 and to state whether they wish to be heard in respect of their submissions.
Dated 18 August 1981
R. E. HARDISTY, Shire Secretary
SHIRE OF KILMORE
By-Law No. 30
Notice is hereby given that the Council of the Shire of Kilmore on 5 August 1981 made a By-Law known as the Shire of Kilmore By-Law No. 30 for the purpose of regulating the management and control of library services provided by the Council jointly with the Moonee Valley Regional Library and for imposing penalties for breaches of the said By-Law.
A copy of the By-Law is open for inspection free of charge during office hours at the office of the Council, Civic Centre, Sydney Street, Kilmore.
JIM ELVEY, Shire Secretary
Town and Country Planning Act 1961
SHIRE OF MAFFRA—SHIRE OF MAFFRA PLANNING SCHEME
MAFFRA TOWNSHIP
NOTICE THAT A PLANNING SCHEME HAS BEEN PREPARED AND IS AVAILABLE FOR INSPECTION
Amendment No. 20, 1981
Notice is hereby given that the Shire of Maffra in pursuance of its powers under the Town and Country Planning Act 1961 has resolved to amend the Shire of Maffra Planning Scheme (Maffra Township) to—
(a) introduce a Transport Uses Zone and re-zone land bounded by Landy and Morison Street and the Maffra Sewerage Farm into Rural Zone to Transport Uses Zone and by including Transport Depot as a new land use; and
(b) introduce junk yards and panel beating as uses which the Council may allow in the General Industry (Industrial B) Zone.
A copy of the Scheme has been deposited at the office of the Shire of Maffra, Johnson Street, Maffra and at the office of the Department of Planning, 235 Queen Street, Melbourne and will be open for inspection during office hours by any person free of charge.
Any persons affected by the Planning Scheme are required to set forth in writing all objections that they have addressed to the Shire Secretary, Shire Office, Maffra on or before 19 September 1981 and state whether they wish to be heard in respect of their objections.
Dated 19 August 1981
J. RENNICK, Municipal Clerk
SHIRE OF MANSFIELD SEWERAGE AUTHORITY
(BONNIE DOON URBAN DISTRICT)
Sewer Construction, Bonnie Doon
The Authority hereby gives notice of its intention to construct sewers, manholes and associated works in the Township of Bonnie Doon in Phillip Street, Arrot Street, Maroondah Highway Service Road and crossing the railway line to the Caravan Park.
Plans are available for inspection at the Authority Office during office hours.
G. D. PAYNE, Secretary
Shire of Mansfield Sewerage Authority, 48 High Street, Mansfield, 3722
SHIRE OF ROSEDALE
NOTICE OF CHANGE OF NAME OF ROAD
At the Ordinary Meeting of the Council of the Shire of Rosedale held on 20 July 1981, the Council resolved to change the name of the road previously known as Nowak Road (in the Parish of Willang) to Piera Road.
J. L. MITCHELL, Shire Secretary
Town and Country Planning Act 1961 (Twelfth Schedule)
SHIRE OF SEYMOUR—SHIRE OF SEYMOUR PLANNING SCHEME
NOTICE THAT A PLANNING SCHEME HAS BEEN PREPARED AND IS AVAILABLE FOR INSPECTION
Amendment No. 57, 1981
Notice is hereby given that the Shire of Seymour in pursuance of its powers under the Town and Country Planning Act 1961 has prepared an amending Planning Scheme for the purpose of re-zoning part of Crown allotment 9, in the Parish of Seymour from Rural A Zone to Residential Zone.
A copy of the Scheme has been deposited at the Shire Office, Seymour and at the office of the Town and Country Planning Board, 235 Queen Street, Melbourne and will be open for inspection during office hours by any person free of charge.
Any persons affected by the Planning Scheme are required to set forth in writing all objections they may have, addressed to the Shire Secretary, Shire Office, Seymour on or before 19 September 1981 and to state whether they wish to be heard in respect of their objections.
Dated 10 August 1981
J. MATHEWS, Municipal Clerk
Town and Country Planning Act 1961 (Twelfth Schedule)
SHIRE OF SEYMOUR—SHIRE OF SEYMOUR PLANNING SCHEME
Notice that a Planning Scheme has been prepared and is available for inspection
Amendment No. 58, 1981
Notice is hereby given that the Shire of Seymour in pursuance of its powers under the Town and Country Planning Act 1961 has prepared an Amendment Planning Scheme for the purpose of re-zoning part of Crown Allotment 46, Section 1, Parish of Seymour from Light Industrial Development Zone to Rural B Zone.
A copy of the Scheme has been deposited at the Shire Office, Seymour and at the Victorian Town and Country Planning Board, 235 Queen Street, Melbourne and will be open for inspection during office hours by any person free of charge.
Any persons affected by the Planning Scheme are required to set forth in writing all objections they may have, addressed to the Shire Secretary, Shire Office, Seymour on or before 19 September 1981 and to state whether they wish to be heard in respect of their objections.
Dated 10 August 1981
J. MATHEWS, Municipal Clerk
SHIRE OF SEYMOUR
Declaration of Result of Poll on a Proposal to Levy Rates on the Basis of the Unimproved Capital Value of Rateable Property
I hereby give notice that at a poll taken on Thursday, 6 August 1981 on the above-mentioned proposal, pursuant to the provisions of section 317 of the Local Government Act, the following votes were recorded:
Number of valid votes recorded for the proposal 1215
Number of valid votes recorded against the proposal 1158
Informal votes 116
Majority of votes for the proposal 57
Number of persons whose names are inscribed on the Municipal Roll 4285
As a majority of the valid votes recorded are for the proposal, which is thereby constituted, the number being not less than one-third of the total number of ratepayers included on the Municipal Voters' Rolls, I therefore declare the proposal to use Site Values to be carried.
J. W. MATHEWS, Returning Officer
SHIRE OF WERRIBEE
Loan No. 136
Notice of Intention to Borrow the Sum of $200 000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the Shire of Werribee proposes to borrow the principal sum of Two hundred thousand dollars ($200 000) secured by a charge over the general rates of the Municipality, such sum to be raised by the grant of a mortgage in accordance with the provisions of the Local Government Act 1938.
1. The maximum rate of interest that may be paid is 14·7 per cent per annum.
2. The purpose for which the loan is to be applied is:
$ Part construction Recreation Centre—Stage 2 280 000
Part reconstruction of Lantana Avenue and part reconstruction of Strathmore Crescent, Werribee 15 000
Reconstruction of Railway Avenue, Laverton 10 000
Purchase of property for road widening purposes—corner Market Road and Railway Avenue, Werribee 20 000
Reconstruction of Eynesbury Road, south of Exford Road 25 000
Construction of Loop Road, between Watton and Station Streets, Werribee 120 000
Drainage works—Little River Road, Little River 30 000
3. The period of the loan shall be ten (10) years.
4. The moneys borrowed shall be repayable by providing out of the Municipal Fund, 19 half-yearly instalments of $39 037·72, each including principal and interest on 1 April and 1 October during the currency of the loan, with a final instalment of $441 888·20 at the end of ten years. The first instalment shall be payable on 1 April 1982.
5. Such moneys shall be repayable to the Commonwealth Savings Bank of Australia, Head Office, 8–26 Elizabeth Street, Melbourne 3000.
The plans and specifications and the estimates of the cost of the proposed works and a statement showing the proposed expenditure of the moneys to be borrowed are open for inspection at the office of the Council of the Shire of Werribee, Civic Centre, 45 Princes Highway, Werribee
Dated 28 July 1981
J. T. KERR, Shire Secretary
SHIRE OF WERRIBEE
Loan No. 137
Notice of Intention to Borrow the Sum of $200 000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the Shire of Werribee proposes to borrow the principal sum of Two hundred thousand dollars ($200 000) secured by a charge over the general rates of the Municipality, such sum to be raised by the grant of a mortgage in accordance with the provisions of the Local Government Act 1938.
1. The maximum rate of interest that may be paid is 14·7 per cent per annum.
2. The purpose for which the loan is to be applied is part construction of the Laverton Community Complex.
3. The period of the loan shall be twenty (20) years.
4. The moneys borrowed shall be repayable by providing out of the Municipal fund, half-yearly instalments of approximately $6 615·09, each including principal and interest on 1 May and 1 November during the currency of the loan. The first instalment shall be payable on 1 May 1982.
5. Such moneys shall be repayable to the Commissioners of the State Bank of Victoria at the Office of the said Commissioners, 385 Bourke Street, Melbourne 3000.
The plans and specifications and the estimates of the cost of the proposed works and a statement showing the proposed expenditure of the moneys to be borrowed are open for inspection at the office of the Council of the Shire of Werribee, Civic Centre, 45 Princes Highway, Werribee 3030.
Dated 28 July 1981
J. T. KERR, Shire Secretary
SHIRE OF YARRAWONGA
Loan No. 68
Notice of Intention to Borrow the Sum of $175 000 for Permanent Works and Undertakings
Notice is hereby given that the Council of the Shire of Yarrawonga proposes to borrow the principal sum of $175 000 secured by a Charge over the General Rates of the Municipality, such sum to be raised by the granting of a Mortgage in accordance with the provisions of the Local Government Act 1938.
1. The maximum rate of interest that may be paid is 14·7%.
2. The purposes for which the loan is to be applied are:
Purchase of Plant, namely:
| Item | Cost |
|-----------------------|--------|
| Grader | 991 000|
| Trucks (3) | 25 800 |
| Utility Vehicle | 7 500 |
| Grid Roller | 24 700 |
| Tractor (Secondhand) | 28 000 |
Total: 175 000
3. The period of the Loan shall be for 10 years.
4. The moneys borrowed shall be repayable by providing out of the Municipal Fund half-yearly instalments of $16,670.79 each, including Principal and Interest on 15 October and 15 April each year, during the currency of the Loan. The first instalment shall be payable on 15 April 1982.
5. Such moneys shall be repayable to the A.N.Z. Savings Bank Ltd., Melbourne, or such other places as the lender may direct.
6. Plans and specifications and the estimates of the costs of the proposed works and a statement showing the proposed expenditure of the moneys to be borrowed are open for inspection at the office of the Shire of Yarrawonga, Shire Hall, Belmore Street, Yarrawonga, during office hours.
D. J. PRESLEY, Shire Secretary
Town and Country Planning Act 1961 (Twelfth Schedule)
SHIRE OF YEA—SHIRE OF YEA PLANNING SCHEME
NOTICE THAT A PLANNING SCHEME HAS BEEN PREPARED AND IS AVAILABLE FOR INSPECTION
Notice is hereby given that the Council of the Shire of Yea, in pursuance of its powers under the Town and Country Planning Act 1961 has prepared a planning scheme for the purpose of amending the Ordinance of the Shire of Yea Planning Scheme.
A copy of the scheme has been deposited at the Civic Centre, Yea, and at the office of the Department of Planning, Victoria, 235 Queen Street, Melbourne, and will be open for inspection during office hours by any person free of charge.
Any persons affected by the planning scheme are required to set forth in writing any submissions they may wish to make with respect to the scheme, addressed to the Shire Secretary, Shire of Yea, Civic Centre, Yea 3717, on or before 27 November 1981, and to state whether they wish to be heard in respect of their submissions.
F. F. BERKERY, Shire Secretary
MOOROOPNA WATERWORKS TRUST
EXTENSION OF WATERWORKS DISTRICT, PROCLAMATION OF URBAN DISTRICT AND CONSTRUCTION OF WORKS FOR SUPPLY OF WATER TO THE TOWNSHIP OF TOOLAMBA
Notice is hereby given that the Mooroopna Waterworks Trust has made application to the Honourable The Minister of Water Supply for the extension of Waterworks Districts, proclamation of the Toolamba Urban District and for the construction, maintenance and continuance of Water Supply Works within those Districts under the provisions of the Water Act.
A general plan and description of the proposed works have been submitted with the application, and copies of same may be seen at the Trust's Office at McLennan Street, Mooroopna.
Dated at Mooroopna 4 August 1981
R. F. FORSTER, Secretary
GEELONG WATERWORKS AND SEWERAGE TRUST
Pursuant to section 60 of the Geelong Waterworks and Sewerage Act 1958 (No. 629) the Trust has prepared plans to indicate the particulars of surface and underground works, the construction of which the Trust intends to begin, at a date not less than one month after publication of the notices, in or adjacent to the following localities within the Drainage Area:
Tanner Street, Breakwater, City of Geelong.
Chapel and Hickey Streets, Whittington, Shire of Bellarine.
Notice is hereby given that the plans indicated are open for public inspection at the Trust's Offices, 61-67 Ryrie Street, Geelong between the hours of 8.30 a.m. and 5.20 p.m. from Monday to Friday (public holidays excepted) by the owners or occupiers of land or premises within the Drainage Area.
L. C. SPITTY, Secretary
LAKES ENTRANCE SEWERAGE AUTHORITY
GENERAL NOTICE MADE PURSUANT TO THE PROVISIONS OF THE SEWERAGE DISTRICTS ACT 1958 AND AMENDMENTS THERETO
The Lakes Entrance Sewerage Authority having made provision for carrying off the sewage from each and every property which or any part of which is within the sewerage area hereinafter described doth hereby declare that on and after 1 September 1981, each and every property which or any part of which is within the said sewerage area shall be deemed to be a sewered property within the meaning of the Sewerage Districts Act 1958.
The boundaries of the sewerage areas hereinbefore referred to are:
Sewerage Area No. 9
Commencing at the point of intersection of the shoreline of the North Arm and the prolongation of the line of the north-western rear boundaries of the allotments on the north-western side and facing the Shore below the point on the boundary of Sewerage Area No. 1, thence following the boundary of Sewerage Area No. 1 north-easterly, south-easterly and south-westerly to its point of intersection with the shoreline of the North Arm; thence in a westerly direction along the shoreline of the North Arm to the point of commencement.
Sewerage Area No. 10
Commencing at the north-west corner of Lot 3 on L.P.94280 being a point on the boundary of Sewerage Areas No. 1 and No. 2, thence following the boundary of Sewerage Area No. 2 north-east and northerly to the southern corner of Lot 121 on L.P.92681 being a point on the boundary of Sewerage Area No. 6; thence following the boundary of Sewerage Area No. 6 northerly to the south-western corner of Lot 120 on L.P.93324 being a point on the boundary of Sewerage Area No. 1; thence following the boundary of Sewerage Area No. 1 easterly and southerly to the point of commencement.
By order of the said Sewerage Act
I. R. BULMER, Chairman
W. J. HOBSON, Secretary
MOE SEWERAGE AUTHORITY
Plans have been prepared for the laying of sewer mains to serve the following streets in the City of Moe.
Job 111/169 30 John Street.
The plans are available for inspection during normal working hours at the office of the Authority, Albert Street, Moe.
Dated 17 August 1981
A. DEWAR, Secretary
DIOCESE OF WANGARATTA
Diocesan Synod
Notice is hereby given that the Bishop of the Diocese of Wangaratta has convened the Diocesan Synod for Saturday 19 September 1981 at 9.45 a.m. at the Cathedral Church of the Holy Trinity, Wangaratta.
M. NELSON, Registrar
ACT No. 391—AMENDMENT ACT 1925
First Schedule
Whereas by virtue of the powers conferred by the Act of the Parliament of Victoria numbered 391 the Governor on 13 April 1963 allowed an application for leave to dispose of the land described in the said allowance: Now therefore the Alexander Robert Gillies Trustee authorises the appropriation of the denomination known as The Uniting Church in Australia Property Trust (Victoria) trustees of the said land and of Rev. John Beasleyman The Uniting Church in Australia Property Trust (Victoria) being any person or persons entitled to minister in or occupy a building or buildings upon the said land do hereby apply
to the Governor of the State of Victoria to amend or vary the said allowance in manner hereinafter mentioned: And I certify—
That the only trustee of the said land resident in Victoria is Rev. John Beautyman.
That the only buildings upon the said land are nil and that the only persons entitled to minister in or occupy the same are the above-named The Uniting Church in Australia Property Trust (Victoria).
A. R. GILLESPIE, Secretary
The common seal of the Uniting Church in Australia Property Trust (Victoria) was hereto affixed in pursuance of a resolution passed at a meeting of the members of the Trust in the presence of:
A. R. GILLESPIE, Member
N. EATON, Member
JOHN W. BEAUTYMAN, Minister
Existing Statement of Trusts
Description of Land—Biregurra Wesleyan Church. Site for Wesleyan Church purposes temporarily reserved by Order in Council of 13 April 1863. 6070 square metres. Township of Biregurra Parish of Biregurra County of Polwarth, being allotment 8A, section E, commencing at the north-eastern angle of north alignment of Jenner Street and the western alignment of Beal Street; bounded thence by Beal Street bearing 195 deg. 28 min. 100-58 metres; by Skene Street bearing 285 deg. 28 min. 60-35 metres; by a line bearing 15 deg. 28 min. 100-58 metres, and thence by Jenner Street bearing 105 deg. 28 min. 60-35 metres to the point of commencement.
Name of Trustees—The Uniting Church in Australia Property Trust (Victoria).
Powers of Disposition—Such powers of disposition including powers of sale, lease or mortgage as are given to the Trust by the Uniting Church in Australia Property Trust (Victoria) Act Number 9021 of 1977.
Purposes to which Proceeds of Disposition are to be Applied—To such Uniting Church in Australia purposes as shall be approved by the Committee of the Victorian Synod of The Uniting Church in Australia.
ACT No. 391—FIRST SCHEDULE
I, Alexander Robert Gillespie head or authorized representative of the denomination known as The Uniting Church in Australia with the consent of The Uniting Church in Australia Property Trust (Victoria) trustees of the land described in the subsequent statement of trust, and of Rev. J. Paul Hudson, 85 Peter Street, Geelong, as being the person entitled to minister in or occupy a building or buildings upon the said land, hereby apply to the Governor of the State of Victoria for leave to dispose of the said land by sale, lease or mortgage as herein mentioned in the said statement of trusts, and I hereby certify that the said land was reserved by Letters on 9 May 1856, for the purpose of a Wesleyan Church. The only trustee of the said land resident in the State of Victoria are The Uniting Church in Australia Property Trust (Victoria). That the only buildings upon the said land are Nil and that the only persons entitled to minister in or occupy the same are the above-named Rev. J. Paul Hudson.
A. R. GILLESPIE
The common seal of The Uniting Church in Australia Property Trust (Victoria) was hereto affixed in pursuance of a resolution passed at a meeting of the Members of the Trust in the presence of:
A. R. GILLESPIE, Member
N. EATON, Member
REV. J. PAUL HUDSON, Minister
Description of Land—Mt. Duneed Wesleyan Church. Site set apart for Wesleyan Church purposes reserved by Letters dated 9 April 1856. 8094 square metres. Parish of Duneed, County of Duneed, being allotment 12, section 21. Commencing at a point bearing 90 deg. 00 min. 402-3 metres from the north-eastern angle of Crown allotment L, section 21; bounded thence by a road bearing 90 deg. 00 min. 80-47 metres; and thence by lines bearing 180 deg. 00 min. 100-58 metres, 270 deg. 00 min. 80-47 metres and 0 deg. 00 min. 100-58 metres to the point of commencement.
Name of Trustees—The Uniting Church in Australia Property Trust (Victoria).
Powers of Disposition—Such powers of disposition, including powers of sale, lease or mortgage as are given to the Trust by the Uniting Church in Australia Property Trust (Victoria) Act Number 9021 of 1977.
Purposes to which Proceeds of Disposition are to be Applied—To such Uniting Church in Australia purposes as shall be approved by the Committee of the Victorian Synod of The Uniting Church in Australia.
Notice is hereby given that Lang Lang Golf Club has applied for a Lease pursuant to section 134 of the Land Act 1958 for a period of twenty-one (21) years in respect of Allotments 90A and 90C Parish of Lang Lang, County of Mornington containing 141 acres 2 roods 31 perches more or less as a site for Amusement and Recreation (Golf Club).
Notice is hereby given that the partnership conducted by Klaus Eckmeyer and Marie Eckmeyer and Byron Paul Blakeway carrying on business as bakery and pastry cooks at 45 High Street Hastings was dissolved on 4 June 1981.
STRAUN WRIGHT-SMITH & SON, solicitors, 13 Rosebud Parade, Rosebud
Notice is given that Phillip Francis Skehan has retired from the partnership carried on under the business name of "Welspool Abattoirs" at Midland Highway, Welshpool as from 21 October 1980 and that the said business is carried on from such date by the remaining partner Neville George Buckland.
PHILLIP FRANCIS SKEHAN
NEVILLE GEORGE BUCKLAND
NOTICE OF DISSOLUTION OF PARTNERSHIP
Notice is hereby given that the Partnership heretofore subsisting between us the undersigned Martin John O'Connor and Philip Leslie Read carrying on business as Accountants, Business Agents and Consultants, Business and Financial Consultants at Suite 3, 172 Bourke Road, Boronia in the State of Victoria under the style or firm name of "O'Connor Read & Partners" is being dissolved by mutual consent as from 30 June 1981.
Dated 12 August 1981
MARTIN JOHN O'CONNOR, PHILIP LESLIE READ, by their Solicitors, Ridgeway Clements, 461 Bourke Street, Melbourne
Notice is hereby given that the partnership heretofore subsisting between the undersigned James Grace Connelly, Graham John Connelly, Thomas Harold Connelly, Kenneth John Connelly and Clive Rodger Connelly all o' Oneo carrying on business as Farmers and Graziers under the name of "J. E. & A. G. Connelly & Sons" has been dissolved as from 30 June 1981 and that the said James Grace Connelly, Thomas Harold Connelly, Kenneth John Connelly and Clive Rodger Connelly shall continue the said business under the name of "J. E. & A. G. Connelly & Sons".
Dated 30 July 1981
A. G. CONNELLY
G. J. CONNELLY
T. H. CONNELLY
K. J. CONNELLY
C. R. CONNELLY
Notice is hereby given that the partnership heretofore subsisting between Segundo Jose Ortiz, Augustina Ortiz, Nelson Vezzosso and Tania Vezzosso carrying on business in panel beating and spray painting at 4 Burton Crescent South Melbourne in the style or firm of "South America Smash Repairs" has been dissolved by mutual consent as from 24 June 1981, the said Nelson Vezzosso and Tania Vezzosso having retired from the said firm.
Notice is hereby given that the partnership heretofore subsisting between John Kavadis, Hairdresser, Andrew Kavadis, Hairdresser, both 11 Inverness Way North Balwyn and Trevor Woods, Hairdresser of Flat 1, 26 Southey Street Elwood carrying on business as Ladies Hairdressers at 181 Exhibition Street, Melbourne under the style or firm of "John Kavadis of Melbourne" has been dissolved by mutual consent as from 10 March 1981.
JOHN KAVADIS
ANDREW KAVADIS
TREVOR WOODS
D. Condon & Co., solicitors, of 83 William Street, Melbourne
Co-operative Housing Societies Act 1958
BEAUMARIS (No. 2) CO-OPERATIVE HOUSING SOCIETY LIMITED
Notice is hereby given that at a special general meeting of the abovenamed society duly convened and held at Damoor House, 1 Central Avenue, Moorabbin, on Tuesday, 11 August 1981, at 12.30 p.m. the following special resolution was duly passed:—
"That the society having successfully completed its objectives 24 months ahead of its expected term, be wound up voluntarily, and that John Patrick Chaundy of 126 Wellington Parade, East Melbourne, be appointed liquidator for the purposes of the winding up."
Dated 11 August 1981
I. J. DACK, Chairman
N. S. FINN, Secretary
Co-operative Housing Societies Act 1958
GREATER MOORABBIN (No. 4) CO-OPERATIVE HOUSING SOCIETY LIMITED
Notice is hereby given that at a special general meeting of the abovenamed society duly convened and held at Damoor House, 1 Central Avenue, Moorabbin, on Tuesday, 11 August 1981, at 12.30 p.m. the following special resolution was duly passed:—
"That the society, having successfully completed its objectives 28 months ahead of its expected term, be wound up voluntarily, and that John Patrick Chaundy of 126 Wellington Parade, East Melbourne, be appointed liquidator for the purposes of the winding up."
Dated 11 August 1981
I. J. DACK, Chairman
N. S. FINN, Secretary
MULLER SERVICES PTY. LTD.
Notice is hereby given in pursuance of section 254 (2) (b) of the Companies Act 1961 that at an extraordinary general meeting of members of Muller Services Proprietary Limited held at 170 Amanda Street, Beaumaris, on 8 August 1981 the following special resolution was passed:—
"That the company be voluntarily wound up."
MAX O. MULLER, Director
Companies Act 1961
BURGMAN NOMINEES PTY. LTD., of 33 Melrose Street, Sandringham
Notice is hereby given that a Petition for the winding up of the abovenamed Company by the Supreme Court was on 5 August 1981 presented by Breamar Appliances Pty. Ltd. and that the said Petition will be heard before the Court sitting at Melbourne at the hour of 10.30 o'clock in the forenoon on Thursday, 10 September 1981 and any creditor or contributory of the Company desiring to appear or oppose the meeting of Creditors on the said Petition may appear at the time of hearing by himself or his Counsel for that purpose; and a copy of the Petition will be furnished to any creditor or contributory of the said Company responsible for any amount assigned on payment of the regulated charge for the same.
The Petitioner's address is 400 Princes Highway, Noble Park.
The Petitioner's solicitor is Anthony M. Simons, of 6th Floor, 60 Albert Road, South Melbourne.
ANTHONY M. SIMONS, Solicitor for the Petitioner
Note.—Any person who intends to appear on the hearing of the said Petition must give one day's notice in writing of his intention so to do. The notice must state the name and address of the firm and must be signed by the person or firm or his or their Solicitor (if any) and must be served or if posted must be sent by post in sufficient time to reach the abovenamed not later than 4 o'clock in the afternoon on 9 September 1981.
MYER CO-OPERATIVE HOUSING SOCIETY LTD. (IVL)
NOTICE OF FINAL MEETING IN THE VOLUNTARY WINDING UP
Take notice that the affairs of the abovenamed Society are now being wound up and that, in pursuance of section 272 (1) of the Companies Act 1961 and the Cooperative Housing Societies Act 1958, a General Meeting of the Society will be held at the offices of Arthur Andersen & Co., 330 Collins Street, on 18 September 1981, at 3.00 p.m. for the purposes of:
(i) laying before it an account showing how the winding up has been conducted and the property of the Society disposed of and giving any explanations thereof; and
(ii) passing a resolution that the books and papers of the said Society and all the liquidator relevant to the affairs of the Society be delivered after a period of twelve months from the date of the meeting.
Dated 13 August 1981
GEORGE A. COHEN, Liquidator
In the matter of the Companies Act 1961, as Amended; and in the matter of H. L. LEE INVESTMENTS PTY. LIMITED
At an Extraordinary General Meeting of Shareholders of H. L. Lee Investments Pty. Limited, duly convened and held at 109 Pitt Street, Sydney, on 7 August 1981, the following Special Resolution was duly passed:
"That the Company be wound-up voluntarily and that Ian Gordon Bruce was appointed Liquidator for the purpose of the winding-up."
Dated 10 August 1981
S. M. NUNAN, Liquidator
Companies Act 1961—In the matter of HARRISON HANDBURY PTY. LTD.
Notice is hereby given that at a meeting of shareholders of Harrison Handbury Pty. Ltd. held on 15 July 1981 the following resolution was passed as a Special Resolution:
"That the company be wound up voluntarily and that Mr Scott Maurice Nunan of Curtain & Nunan, Chartered Accountants of 486 Sydney Road, Coburg be appointed Liquidator."
S. M. NUNAN, Liquidator
QUEST PROJECTS PTY. LTD. (IN LIQUIDATION)
At a general meeting of members of Quest Projects Pty. Ltd. duly convened and held at First Floor, 53 Kingsway, Glen Waverley, on 13 August 1981 the following special resolution was duly passed:
"That the Company be wound up voluntarily on the ground that the Company has ceased trading and that Miss Elaine Prentice of 710 Ferntree Gully Road, Mulgrave, be appointed Liquidator.
NORRIS, NORRIS & WATERS, solicitors for the liquidator, of Shop 14t Station Street, Bayswater, 3153
BIDVAH NOMINEES PTY. LTD.
Notice is hereby given that at an Extraordinary General Meeting of the members of the abovementioned Company held on 17 August 1981 it was resolved that the Company be wound up voluntarily and that Michael Woolf Naphtali of 5th Floor, 71 Queens Road, Melbourne, 3004, be appointed liquidator.
Notice is also given that after 21 days from this date, I shall proceed to distribute the assets. All creditors having any claims against the Company should furnish particulars of same by that date otherwise I shall proceed to distribute the assets without regard to their claim.
Dated 17 August 1981
M. W. NAPHTALI, Liquidator
Companies Act 1961
NOTICE OF MEETINGS
Notice is hereby given that pursuant to Section 271 of the Companies Act an Annual General Meeting of the creditors and members of the undermentioned companies will be held at the offices of Bent & Cough Public Accountants, 144 Jolimont Road, East Melbourne on Wednesday, 2 September 1981 at the time stated.
10.30 a.m. Mietti Constructions Pty. Ltd. (in Liquidation).
11.00 a.m. Hilton Acceptance Pty. Ltd. (in Liquidation).
Dated 12 August 1981
D. J. COUGLE, Liquidator
Bent & Cough, public accountants, 144 Jolimont Road, East Melbourne, 3002
SOUTHLAND PLUMBING PTY. LTD.
Notice is hereby given of a Meeting of Creditors of Southland Plumbing Pty. Ltd. to be held at City of Frankston Library Conference Room, Central Park, Frankston on Friday 28 August 1981 at 2.00 p.m.
The Company has convened an Extraordinary General Meeting of its Members for the same day for the purpose of considering the passing of the following resolutions:
"That the Company be wound up voluntarily."
"That Edward Peter Taylor, Chartered Accountant, of 583 Hampton Street, Hampton be appointed Liquidator of the Company."
The Creditors of the Company will be asked to consider the following matters:
"The appointment of a Liquidator."
"The remuneration of the Liquidator."
"Appointment of a Committee of Inspection, if required."
Please note that only creditors who have lodged a form of Proof of Debt will be entitled to vote at the meeting of Creditors.
Completed forms may be forwarded to the office of E. P. Taylor, Chartered Accountant, 583 Hampton Street, Hampton, 3188, or be handed to the Chairman on the day of the Meeting.
DAVID KEITH COLE, Director
The Companies Act 1961, Section 271 (1)
Companies Regulations
Regulation 26 (2) (b)
BRUTUS CLOTHING HOLDINGS PTY. LIMITED (IN LIQUIDATION)
NOTICE OF MEETING OF CREDITORS AND MEMBERS
Notice is hereby given that a meeting of the creditors and members of Brutus Clothing Holdings Pty. Limited (in liquidation) will be held at the offices of Saunders, 19th floor, 114 William Street Melbourne on 4 September 1981 at 9.15 a.m.
Agenda
To lay before the meeting the annual accounts of the liquidator's acts and dealings and of the conduct of the winding up.
Dated 12 August 1981
NEIL W. CURWOOD, Liquidator
The Companies Act 1961
HALLMARK INVESTMENTS PTY. LTD. (IN VOLUNTARY LIQUIDATION)
Notice is hereby given that pursuant to Section 272 of the Companies Act 1961, a General Meeting of the members of the Company will be held at 401 Bourke Street, Melbourne on Wednesday 16 September 1981 at 9.30 o'clock in the forenoon for the purpose of having laid before them an account showing how the winding up has been conducted and of hearing proof that the Company disposed of and of hearing any explanation that may be given by the liquidator.
Dated 5 August 1981
W. A. K. A'BECKETT
Companies Act 1961—In the matter of Cambria Timber and Trading Pty. Ltd.
Notice is hereby given that a Scheme of Arrangement between the company, its members and creditors was approved by an Order of the Supreme Court of Victoria on 6 August 1981.
The commencement date of the Scheme is 12 August 1981.
Pursuant to Clause 7 (b) of the Scheme, notice is hereby given that if you claim to be a scheme creditor, you are required to lodge your claim with me no later than 25 August 1981, together with a Statutory Declaration specifying the amount of a claim against the company at the Fixed Date which was 27 February 1981.
Dated 12 August 1981
JOHN M. SPARK, Scheme Trustee
Orr, Martin & Waters, chartered accountants, 460 Bourke Street, Melbourne, 3000. Telephone 602 1644 4855
In the Supreme Court of Victoria—Co. No. 12081—in the matter of the Companies Act 1961; and in the matter of BAMCO TRANSPORT PTY. LTD.—Memorandum of Advertisement.
Notice is hereby given that a Petition for the winding up of the abovenamed company by the Supreme Court was on 23 July 1981 presented by Mobil Oil Australia Limited and that the said Petition is directed to be heard before the sitting at the Supreme Court, Law Courts, Lonsdale Street, Melbourne at the hour of 10.30 o'clock in the forenoon on 3 September 1981 and any creditor or contributory of the said company desiring to support or oppose the making of an order on the said Petition may appear at the time of hearing by himself or his Counsel for the purpose; and a copy of the said Petition will be furnished to any creditor or contributory of the said Company requiring the same by the undersigned on payment of the regulated charge for the same.
The Petitioner's official address is 2 City Road, Melbourne.
The Petitioner's Solicitors are Messrs Arthur Robinson & Co. of 447 Collins Street, Melbourne
ARTHUR ROBINSON & CO.
Note—Any person who intends to appear on the hearing of the said Petition must serve or send by post to the abovenamed Arthur Robinson & Co., notice in writing of his intention so to do. The notice must state the name and address of the person, or, if a firm, the name and address of the firm and must be signed by the person or firm or firm or his or their solicitor (if any), and must be served, or, if posted, must be sent by post in sufficient time to reach the abovenamed not later than 4.00 o'clock in the afternoon of 2 September 1981.
The Companies Act 1961—In the matter of Albert Harvey Engineering Company Pty. Ltd. (in Liquidation)—Members' Winding Up.
Notice is hereby given that at an Extraordinary General Meeting of members of the company duly convened and held at 500 Collins Street, Melbourne in the State of Victoria on 7 August 1981, the following resolution was proposed and passed as a Special Resolution:—
"That the Company be wound up voluntarily in accordance with the provisions of the Act relating to a members voluntary winding up."
Dated 10 August 1981
J. D. ADAMS, Liquidator
Care of Parkhill Lithgow & Gibson, 500 Collins Street, Melbourne 3000
The Companies Act 1961, Section 260
Gourlay Constructions Pty. Ltd.
Notice of Meeting of Creditors
Notice is hereby given that a Meeting of Creditors of the abovenamed Company will be held at the Boardroom of the Institute of Chartered Accountants, 9th Floor, 140 Queen Street, Melbourne on Thursday, 27 August 1981 at 10.00 a.m.
Agenda
1. Appointment of Chairman.
2. Report and Presentation of Statement of Affairs.
3. Discussion.
4. Nomination of Liquidator.
5. Return of the Liquidator.
6. Committee of Inspection.
7. Other Business.
Dated 13 August 1981
By order of the Board
D. M. GOURLAY, Director
Messrs Venn Milner & Terracall, chartered accountants, Suite 2, 497 Burke Road, Camberwell, Victoria 3124 4861
Form 92
Companies Act 1961, Companies Regulations 26 (2) (b)
Aquilla Plywood & Timber Supplies Pty. Ltd. (in Liquidation)
Notice of Meeting of Members
Notice is hereby given that pursuant to section 241 of the Companies Act 1961, a meeting of the members of Aquilla Plywood & Timber Supplies Pty. Ltd. (in Liquidation) will be held at the offices of Messrs Duesbury's, 114 William Street, Melbourne, Vic. 3000, on 27 August 1981 at 10.00 a.m., for the purpose of appointing a Committee of Inspection to act with the Liquidator.
Dated 11 August 1981
WARICK A. LEEMING, Official Liquidator
Duesbury's, chartered accountants, 114 William Street, Melbourne, Vic. 3000 4862
Debair Pty. Ltd. (in Voluntary Liquidation)
Notice of Final Meeting of Members, Pursuant to Section 272
Notice is hereby given in pursuance of section 272 of the Companies Act 1961 that the final general meeting will be held at the offices of Messrs Hungerford Hancock & Offner, 29th Floor, 367 Collins Street, Melbourne on 21
September 1981 at 9.00 a.m. to receive the liquidator's account showing how the winding up has been conducted and the property of the company has been disposed of and to hear any explanations that may be given by the liquidator.
K. R. HALL, Liquidator
Companies Act 1961, Section 260
COOROONA NOMINEES PTY. LTD.
Notice is hereby given that at a meeting of the above company to be held for voluntary winding up it is to be proposed and a meeting of its creditors will be held at the offices of John C. Barnes & Co., 470 Bourke Street, Melbourne at 11 a.m. on the same day, 26 August 1981.
J. F. McARDLE, Director
John C. Barnes & Co., 470 Bourke Street, Melbourne 3000 Telephone 67 8579
The Companies Act 1961, Section 260
GAMMA TOOLS PTY. LTD.
Pursuant to Section 260 of the Companies Act 1961
Notice is hereby given that a Meeting of Creditors of Gamma Tools Pty. Ltd. will be held at the offices of Ferrier Hodgson & Co., Chartered Accountants, 24th Floor, 459 Collins Street, Melbourne on Friday, 28 August 1981 at 3.00 p.m. for the purpose of considering the affairs, the Company having convened an Extraordinary General Meeting of its Members to be held on Friday, 28 August 1981, and for the purpose of considering and, if thought fit, passing a Special Resolution that the Company be wound up voluntarily.
Dated 13 August 1981
H. ZSETEY, Director
Ferrier Hodgson & Co., chartered accountants, 459 Collins Street, Melbourne 3000
Form 92
Companies Act 1961, Companies Regulations 26 (2) (b)
AQUILLA PLYWOOD & TIMBER SUPPLIES PTY. LTD. (IN LIQUIDATION)
NOTICE OF MEETING OF CREDITORS
Notice is hereby given that pursuant to section 241 of the Companies Act 1961, a meeting of the creditors of Aquilla Plywood & Timber Supplies Pty. Ltd. (in Liquidation) will be held at the offices of Messrs Duesbury, 114 William Street, Melbourne, Vic. 3000, on 27 August 1981 at 10.00 a.m. for the purpose of appointing a Committee of Inspection to act with the Liquidator.
Important for Voting
A person shall not be entitled to vote as a creditor at the meeting unless he has made known to the Liquidator the amount which he claims to be due to him from the company.
A creditor entitled to attend and vote at the meeting may appoint another person as his proxy to attend and vote. A proxy form is enclosed, which may be lodged with the Chairman of the meeting.
Dated 11 August 1981
WARICK A. LEEMING, Official Liquidator
Duesbury, chartered accountants, 114 William Street, Melbourne, Vic. 3000. Telephone 67 8331
STYLECRAFT MODES PTY. LTD. (IN VOLUNTARY LIQUIDATION)
Notice is hereby given in pursuance of section 272 of the Companies Act 1961, that a general meeting of the members of the abovenamed Company will be held at the offices of E. L. Vail and McBain, 570 Lonsdale Street, Melbourne on 25 September 1981 at 11.00 a.m. for the purpose of having the liquidators' accounts laid before them showing the manner in which the winding up has been conducted and the property of the Company disposed of and of hearing any explanation that may be given by the liquidators, and also of determining the manner in which the books, accounts and documents of the Company and of the liquidators thereof shall be disposed of.
Dated 14 August 1981
G. H. RIPPER AND D. T. RIPPER, Joint Liquidators
FUTURITY INVESTMENTS PTY. LTD. (IN VOLUNTARY LIQUIDATION)
Notice is hereby given, in pursuance of section 272 of the Companies Act 1961, that a general meeting of the members of the abovenamed Company will be held at the offices of E. L. Vail and McBain, 570 Lonsdale Street, Melbourne on 25 September 1981 at 10.15 a.m. for the purpose of having the liquidators' accounts laid before them showing the manner in which the winding up has been conducted and the property of the Company disposed of and of hearing any explanation that may be given by the liquidators, and also of determining the manner in which the books, accounts and documents of the Company and of the liquidators thereof shall be disposed of.
Dated 14 August 1981
G. H. RIPPER AND D. T. RIPPER, Joint Liquidators
B. E. WATTS PTY. LTD. (IN VOLUNTARY LIQUIDATION)
Notice is hereby given, in pursuance of section 272 of the Companies Act 1961, that a general meeting of the members of the abovenamed Company will be held at the offices of E. L. Vail and McBain, 570 Lonsdale Street, Melbourne on 25 September 1981 at 10.30 a.m. for the purpose of having the liquidators' accounts laid before them showing the manner in which the winding up has been conducted and the property of the Company disposed of and of hearing any explanation that may be given by the liquidators, and also of determining the manner in which the books, accounts and documents of the Company and of the liquidators thereof shall be disposed of.
Dated 14 August 1981
G. H. RIPPER AND D. T. RIPPER, Joint Liquidators
CEENAN DELMOS PTY. LTD. (IN VOLUNTARY LIQUIDATION)
Notice is hereby given, in pursuance of section 272 of the Companies Act 1961, that a general meeting of the members of the abovenamed Company will be held at the offices of E. L. Vail and McBain, 570 Lonsdale Street, Melbourne on 25 September 1981 at 10.45 a.m. for the purpose of having the liquidators' accounts laid before them showing the manner in which the winding up has been conducted and the property of the Company disposed of and of hearing any explanation that may be given by the liquidators, and also of determining the manner in which the books, accounts and documents of the Company and of the liquidators thereof shall be disposed of.
Dated 14 August 1981
G. H. RIPPER AND D. T. RIPPER, Joint Liquidators
VICTORIAN CELLARS PTY. LTD. (IN VOLUNTARY LIQUIDATION)
Notice is hereby given, in pursuance of section 272 of the Companies Act 1961, that a general meeting of the members of the abovenamed Company will be held at the offices of Touche Ross and Co., 10th Floor, 440 Collins Street, Melbourne on Friday, 25 September 1981 at 10.00 a.m. for the purpose of having the liquidators' accounts laid before them showing the manner in which the winding up has been conducted and the property of the Company disposed of and of hearing any explanation that may be given by the liquidators, and also of determining the manner in which the books, accounts and documents of the Company and of the liquidator thereof shall be disposed of.
Dated 14 August 1981
M. L. WEINERT, Liquidator
MERINDIE HOLDINGS PTY. LTD. (IN VOLUNTARY LIQUIDATION)
Notice is hereby given, in pursuance of section 272 of the Companies Act 1961, that a general meeting of the members of the abovenamed Company will be held at the offices of Touche Ross and Co., 10th Floor, 440 Collins Street, Melbourne on 25 September 1981 at 10.30 a.m. for the purpose of having the liquidators' accounts laid before them showing the manner in which the winding up has been conducted and the property of the Company
disposed of and of hearing any explanation that may be given by the liquidators, and also of determining the manner in which the books, accounts and documents of the Company and of the liquidators thereof shall be disposed of.
Dated 14 August 1981
R. A. NASH AND J. A. NASH, Joint Liquidators
In the matter of the Companies Act 1961, Section 260; and in the matter of NORTH WESTERN HARDWARE SUPPLIES PTY. LTD.—Notice of Meeting of Creditors
Notice is hereby given that a meeting of the creditors of North Western Hardware Supplies Pty. Ltd. will be held at the offices of Hungerford Hancock & Offner, 8th Floor, 44 Market Street, Melbourne in the State of Victoria on Wednesday, 26 August 1981 at 11.30 in the forenoon.
AGENDA
To consider a special resolution:
"That Mr Douglas Orson Oldfield (Chartered Accountant) be and is now appointed Liquidator of the Company and that the provisions of section 277A (1A) of the Companies Act 1961 shall not operate to validate the appointment of the abovenamed as Liquidator."
A creditor is not entitled to vote at the meeting unless he had lodged with the Chairman of the meeting a proof of the debt which he claims to be due to him from the Company.
Dated 11 August 1981
ARTHUR DONALD HILLGROVE, Director
Hungerford Hancock & Offner, chartered accountants,
44 Market Street, Melbourne, 3000. 614 1200
In the Supreme Court of Victoria—1981 Co. No. 12105—in the matter of the Companies Act 1961; and in the matter of a Petition for the Winding Up of HARNEY & RUDD PTY. LTD.
Notice is hereby given that Petition for the winding up of Harney & Rudd Pty. Ltd. by the Supreme Court was on 13 August 1981 presented by Mathes Court Pty. Ltd. (trading as managers) and that the said petition is directed to be heard before the Court sitting at the Practice Court, Law Courts, William Street, Melbourne, at the hour of 10.30 o'clock in the forenoon on 17 September 1981 and any creditor or contributor of the said Company desiring to support or oppose the making of an Order of the said Petition may appear at the time of hearing himself or his Counsel for that purpose and a copy of the Petition will be furnished to any creditor or contributor of the said Company requiring the same by the undersigned upon payment of the regular charge for the same.
The Petitioner's address is 44-46 Crown Avenue, Mordialloc, in the State of Victoria.
The Petitioner's Solicitor is Mr John Bennison of Messrs Michael Richmond & John Bennison, 493 Main Street, Mordialloc, in the State of Victoria.
JOHN BENNISON of the firm of Michael Richmond & John Bennison
Note.—Any person who intends to appear on the hearing of the said Petition must serve on or send by post in sufficient time to John Bennison of the firm of Michael Richmond & John Bennison, notice in writing of his intention so to do. The Notice must state the name and address of the person or, if a firm, the name and address of the firm, and must be signed by the person or firm or his or their Solicitor (if any), and must be served or posted, must be sent by post in sufficient time to reach the abovenamed not later than four o'clock in the afternoon of 16 September 1981.
PETER R. KEMPSON & CO., of 51 Hardware Street, Melbourne, as town agents for Kempson & Gay, of Suite 6, 79 Mahoneys Road, Forest Hill.
In the Supreme Court of Victoria—1981 Co. No. 12105—in the matter of the Companies Act 1961; and in the matter of a Petition for the Winding Up of HARNEY & RUDD PTY. LTD.
Notice is hereby given that Petition for the winding up of Harney & Rudd Pty. Ltd. by the Supreme Court was on 13 August 1981 presented by Mathes Court Pty. Ltd. (trading as managers) and that the said petition is directed to be heard before the Court sitting at the Practice Court, Law Courts, William Street, Melbourne, at the hour of 10.30 o'clock in the forenoon on 17 September 1981 and any creditor or contributor of the said Company desiring to support or oppose the making of an Order of the said Petition may appear at the time of hearing himself or his Counsel for that purpose and a copy of the Petition will be furnished to any creditor or contributor of the said Company requiring the same by the undersigned upon payment of the regular charge for the same.
The Petitioner's address is 44-46 Crown Avenue, Mordialloc, in the State of Victoria.
The Petitioner's Solicitor is Mr John Bennison of Messrs Michael Richmond & John Bennison, 493 Main Street, Mordialloc, in the State of Victoria.
JOHN BENNISON of the firm of Michael Richmond & John Bennison
Note.—Any person who intends to appear on the hearing of the said Petition must serve on or send by post in sufficient time to John Bennison of the firm of Michael Richmond & John Bennison, notice in writing of his intention so to do. The Notice must state the name and address of the person or, if a firm, the name and address of the firm, and must be signed by the person or firm or his or their Solicitor, and must be served or if posted, must be sent by post in sufficient time to reach the abovenamed not later than four o'clock in the afternoon of 16 September 1981.
In the matter of the Companies Act 1961; and in the matter of C.C.W. INVESTMENTS PTY. LTD.
Notice is hereby given that at a meeting of the members of C.C.W. Investments Pty. Ltd. on 14 August 1981 the following resolution was passed as a Special Resolution:
"That the Company be wound up voluntarily and that Edward John Bates of 361 Nepean Highway, Mordialloc, be and is hereby appointed liquidator to act for the purpose of such winding up."
E. J. BATES, Liquidator
In the matter of the Companies Act 1961; and in the matter of JAYSOUND INVESTMENTS PTY. LTD.
Notice is hereby given that at a meeting of the members of Jaysound Investments Pty. Ltd. on 17 August 1981 the following resolution was passed as a Special Resolution:
"That the Company be wound up voluntarily and that Frederick John Quinton and Edward John Bates, both of 361 Nepean Highway, Mordialloc, be and are hereby appointed liquidators to act jointly or severally for the purpose of such winding up."
F. J. QUINTON, Liquidator
E. J. BATES, Liquidator
Companies Act 1961, Section 254 (2) (b)
AGAR HOLDINGS PTY. LTD.
Notice is hereby given that at an Extraordinary General Meeting of the members of the abovenamed Company held on 17 August 1981 it was resolved that the Company be wound up voluntarily and that Bruce Herbert Whitaker of 8 Station Street, Ringwood, be appointed Liquidator.
Notice is also given that after 21 days from this date I shall proceed to distribute the assets of creditors having any claims against the Company should furnish particulars of same by that date, otherwise I shall proceed to distribute the assets without regard to their claim.
Dated 17 August 1981
B. H. WHITAKER, liquidator, 8 Station Street, Ringwood, Vic. 3134
The Companies Act 1961, Section 272 (2)
NOTICE OF FINAL MEETING
Notice is hereby given that a Final Meeting of the members of the undermentioned companies will be held at the offices of O. G. Parkinson & Son, Suite 904, 450 Little Collins Street, Melbourne at the time stated on Wednesday 30 September 1981 for the purpose of having laid before them an account showing how the winding up has been conducted and the property of the Company disposed of and of bearing any explanation that may be given by the liquidator:
W. J. Manufacturers Pty Ltd. (in Voluntary Liquidation)—10.00 a.m.
Childcraft Pty. Ltd. (in Voluntary Liquidation)—10.15 a.m.
Dated 17 August 1981
4897 PATRICK M. WOOD, Liquidator
The Companies Act 1961, Section 272 (2)
NOTICE OF FINAL MEETING
Notice is hereby given that a Final Meeting of the members of the undermentioned companies will be held at the offices of O. G. Parkinson & Son, Suite 904, 450 Little Collins Street, Melbourne at the time stated on Wednesday 30 September 1981 for the purpose of having laid before them an account showing how the winding up has been conducted and the property of the Company disposed of and of bearing any explanation that may be given by the liquidator:
W. J. A. Stationery Pty. Ltd. (in Voluntary Liquidation)—9.30 a.m.
H. W. Rice Pty. Ltd. (in Voluntary Liquidation)—9.45 a.m.
Dated 17 August 1981
4898 PHILLIP G. HOWARD, Liquidator
Companies Act 1961—In the matter of YALAM NOMINEES PTY. LTD., trustee for Rostrevor Welsh Class Trust, trading as "Real Price", 12 Malvern Street, Melbourne.
Notice re Meeting of Creditors Pursuant to Section 260
Notice is hereby given that a Meeting of Creditors of the abovenamed Company will be held at 144 Jolimont Road, East Melbourne on Wednesday, 9 September 1981 at 10.30 a.m. The Company will lay before the Meeting its Members for the same day for the purpose of considering a Special Resolution that the Company be wound up voluntarily.
Dated 17 August 1981
B. J. WALSH, Director
Bent & Cougle, public accountants, 144 Jolimont Road, East Melbourne, 3002
Companies Act 1961—In the matter of LILITH INSURANCES PTY. LTD. (in Liquidation)
Notice is hereby given that at an extraordinary meeting of the Members of the abovenamed Company held on 17 August 1981, it was resolved that the Company be wound up voluntarily and at a Meeting of Creditors held on the same day pursuant to Section 260 it was resolved that for such purpose Mr D. J. Cougle of 144 Jolimont Road, East Melbourne, Public Accountant, be appointed Liquidator.
Notice is also given that the Liquidator has fixed 30 September 1981 as the date on which creditors of the Company are to prove their debts or claims and to establish any title they may have to priority under Section 292 of the Companies Act.
Dated 17 August 1981
D. J. COUGLE, Liquidator
Bent & Cougle, public accountants, 144 Jolimont Road, East Melbourne, 3002
4900
GREY LODGE PTY. LTD. (IN VOLUNTARY LIQUIDATION)
NOTICE OF MEETING OF CREDITORS AND MEMBERS
Notice is hereby given that a joint Meeting of Creditors and Members of Grey Lodge Pty. Ltd., will be held at 68 High Street, Wodonga, Vic., on 22 September 1981, at 11 o'clock in the forenoon.
Agenda
Liquidator to present a report showing how the winding up of the Company has been conducted and how the property of the Company has been disposed of.
Dated 14 August 1981
4901 W. A. GOULD, Liquidator
The Companies Act 1961
P. J. SELETO & COMPANY PTY. LTD. (IN LIQUIDATION)
Notice is hereby given that a first dividend is intended to be declared in the above matter. Creditors who have not proved their debts by 4 September 1981 may be excluded from this dividend.
Dated 18 August 1981
R. M. H. COLE, Liquidator
Touche Ross & Co., chartered accountants, 440 Collins Street, Melbourne, 3000
4917
The Companies Act 1961, Rule 99—In the matter of OGLE S. & M. PTY. LTD. (IN LIQUIDATION)
Take notice that the Liquidator of the abovenamed Company has fixed 18 August 1981 as the day on or before which creditors of the Company may lay to prove their debts or claims and to establish any title they may have to priority under section 292 or be excluded from the benefit of a first dividend made before such debts are proved or as the case may be from objecting to such a dividend.
Dated 17 August 1981
A. M. HORSBURGH, Liquidator
Wallace, McMullin & Small, chartered accountants, 499 St. Kilda Road, Melbourne, Vic. 3004
4918
The Companies Act 1961
MIDLANDS HARDWARE PTY. LTD. (IN LIQUIDATION)
NOTICE CONVENING FINAL MEETING OF THE COMPANY,
PURSUANT TO SECTION 272 (1)
Notice is hereby given that the final meeting of members and creditors of the abovenamed Company will be held at the offices of Touche Ross & Co., First Floor, 440 Collins Street, Melbourne, on 15 October 1981 at 10.30 a.m. for the purpose of laying before the Meeting an account showing how the winding up has been conducted and the property of the Company has been disposed of and giving any explanation thereof.
Dated 18 August 1981
G. O. HARRISON, Liquidator
Touche Ross & Co., chartered accountants, 440 Collins Street, Melbourne, 3000
4919
## Register of Unclaimed Moneys held by the—
### Unclaimed Moneys Act 1962
| Name of Owner on Books and Last Known Address | Total Amount Due to Owner | Description of Unclaimed Money | Date when Amount first became Payable |
|---------------------------------------------|---------------------------|---------------------------------|--------------------------------------|
| MARYBOROUGH HOLDINGS LIMITED | | | |
| Cairns, Andrew A., Tweedale St, Dunolly | 112.50 | Dividend | 3.8.79 |
| Cheshire, R. B., 159 Bayview Rd, McCrae | 22.50 | " | " |
| Chisholm, A. H. (estate of) | 189.61 | " | " |
| Craven, A., c/o C.B.A. Ltd., G.P.O. Box 1752, Adelaide, S.A. | 112.50 | " | " |
| Girvan, G. F. (estate of) | 17.46 | " | " |
| Hamilton, Margaret J., Bairnsdale | 32.40 | " | " |
| Hayward, E. 50 Bayview Rd, Maryborough | 10.98 | " | " |
| Looney, Jack, 108 Queen St, Melbourne | 11.25 | " | " |
| Looney, O. J. P., 1 Dana St, West Doncaster | 11.25 | " | " |
| Parker, Lindsay W., 10 High St, Maryborough | 10.08 | " | " |
| Parker, (estate of) | 17.28 | " | " |
| Porter, Olive S., 54 Booth St, Golden Square | 13.14 | " | " |
| Roberts, John R., Dayboro Rd, Petrie, Qld. | 11.25 | " | " |
| Shay, James L., c/o Mrs. A. Pike, R.S.S. Goldsborough | 33.75 | " | " |
| Burkinshaw, R. F., c/o Donald Bruceville Mail, Lockhart, N.S.W. | 15.40 | " | 4.7.80 |
| Chisholm, Robert B., 159 Bayview Rd, McCrae | 27.50 | " | " |
| Chisholm, Alexander H. (estate of), Pymble, N.S.W. | 231.77 | " | " |
| Duncombe, Isobel (estate of), Surrey Hills | 297.77 | " | " |
| Hamilton, Margaret J., Bairnsdale | 39.60 | " | " |
| Harris, Ethel M., 29 Ormond Rd, Avondale Heights | 13.75 | " | " |
| Hayward, Ellis, 10 Wellington St, Maryborough | 13.42 | " | " |
| Hood, Phyllis M., 11 Roger St, Kyabram | 10.45 | " | " |
| Looney, O. J. P., 1 Dana St, West Doncaster | 13.75 | " | " |
| Parker, Lindsay W., 10 High St, Maryborough | 12.32 | " | " |
| Parker, William (estate of), Ormond | 21.12 | " | " |
| Porter, Olive S., 54 Booth St, Golden Square | 16.06 | " | " |
| Roberts, John R., Dayboro Rd, Petrie, Qld. | 13.75 | " | " |
| Shay, James L., c/o Mrs. A. Pike, R.S.S. Goldsborough | 41.25 | " | " |
| Talbot, A. J. and K., 25 Bowen St, Warragul | 76.56 | " | " |
| Young, Maria Madge, 6 Penquite Rd, Launceston, Tas. | 11.00 | " | " |
4680
JOHNS CONSOLIDATED-LIMITED
| Name of Owner on Books and Last Known Address | Total Amount Due to Owner | Description of Unclaimed Money | Date when Amount first became Payable |
|-----------------------------------------------|---------------------------|---------------------------------|--------------------------------------|
| Ahern, Willis Charles, Post Restante G.P.O., Sydney, N.S.W. | 20.00 | Dividend | 31.10.79 |
| Donnelly, Charles William, 78/25 King St, Prahran | 27.00 | " | 27.4.79 |
| Kar, Betty, c/o Kevin C. L. Kar, 711 Hang Lung Centre, 2-20 Paterson St, Causeway Bay, Hong Kong | 54.00 | " | " |
| Taylor, Thomas K. (estate of), c/o Mr Hector L. Davis, 10 Hastings Crt, Heidelberg | 42.00 | " | 31.10.79 |
Companies Act 1961–1971
WERRIBEE' HOTEL PTY. LTD. (IN VOLUNTARY LIQUIDATION)
**NOTICE OF FINAL MEETING, PURSUANT TO SECTION 272**
Notice is hereby given pursuant to section 272 of the Companies Act that the Final General Meeting of members of the abovenamed company will be held at the office of Charman, Tierney & Associates, 1st Floor, 10-12 Prospect Street, Box Hill, on 2 September 1981 at 10.30 a.m., for the purpose of having an account laid before it showing the manner in which the winding up has been conducted and the property of the company disposed of and hearing any explanation that may be given by the Liquidator.
Dated 17 August 1981
4920
D. J. CHARMAN, Liquidator
Companies Act 1961–1971
LYNSMITH PTY. LTD. (IN VOLUNTARY LIQUIDATION)
**NOTICE OF FINAL MEETING, PURSUANT TO SECTION 272**
Notice is hereby given pursuant to section 272 of the Companies Act that the Final General Meeting of members of the abovenamed company will be held at the office of Charman, Tierney & Associates, 1st Floor, 10-12 Prospect Street, Box Hill, on 3 September 1981 at 11.30 a.m., for the purpose of having an account laid before it showing the manner in which the winding up has been conducted and the property of the company disposed of and hearing any explanation that may be given by the Liquidator.
Dated 17 August 1981
4921
D. J. CHARMAN, Liquidator
Companies Act 1961, Section 260
MENZIES PICTORIAL PTY. LTD.
**NOTICE OF MEETING OF CREDITORS**
Notice is hereby given that a meeting of Creditors of Menzies Pictorial Pty. Ltd. will be held at the Meeting Room of Hosking, Goldberg & Saffer, 1st Floor, 153 Park Street, South Melbourne on Wednesday, 26 August 1981 at 10.00 a.m., the Company having convened a meeting of its members for the same day for the purpose of considering a Resolution that the Company be wound up voluntarily.
Dated: 14 August 1981
VINCENT PEPE, Director
Hosking, Goldberg & Saffer, public accountants, 153 Park Street, South Melbourne 3205. Telephone 690 6466
4924
Companies Act 1961, Section 272 (2)—In the matter of AGGA LIMITED (in Liquidation)—Notice of Final Meeting
Notice is hereby given that pursuant to section 272 of the Companies Act, the final meeting of the members of the abovenamed company will be held at 11th Floor, 447 Collins Street, Melbourne at 10.00 a.m. on Tuesday, 22 September 1981 for the purpose of having an account laid before them showing the manner in which the winding up has been conducted and the property of the company disposed of, and hearing any explanation that may be given by the liquidator.
Dated 14 August 1981
JOHN EDWIN TOMLINSON, Liquidator
11th Floor, 447 Collins Street, Melbourne 3000 4922
Companies Act 1961, Section 272 (2)—In the matter of DENNYS LASCELLES FINANCE PTY. LIMITED (in Liquidation)—Notice of Final Meeting
Notice is hereby given that pursuant to section 272 of the Companies Act, the final meeting of the members and the creditors of the abovementioned company will be held at 11th Floor, 447 Collins Street, Melbourne at 10.00 a.m. on 21 September 1981 for the purpose of having an account laid before them showing the manner in which the winding up has been conducted and the property of the company disposed of, and hearing any explanation that may be given by the liquidator.
Dated 18 August 1981
J. E. TOMLINSON, Liquidator
11th Floor, 447 Collins Street, Melbourne 3000 4923
Companies Act 1961, Section 254 (2)
EMPROTEL PTY. LTD.
Notice is hereby given that at an extraordinary general meeting of Empotel Pty. Ltd. was convened and held at 5 Stradella Avenue, Vermont South on 14 August 1981 the following resolution was passed as a Special Resolution:
"That the Company be wound up voluntarily and that Douglas Ewart Tonkin, Chartered Accountant, 51 Queen Street, Melbourne 3000 be and is hereby appointed liquidator to act for the purpose of such winding up, with the right to distribute assets in specie."
Notice is also given that I shall proceed to distribute the assets after 28 days from 14 August 1981. All creditors should prove their debts by that date, otherwise I shall proceed to distribute the assets without regard to their claim.
Dated 17 August 1981
D. E. TONKIN, Liquidator
Marquand & Co., chartered accountants, 51 Queen Street, Melbourne, Vic. 3000 4925
Companies Act 1961, Section 254 (2) (b)
VICTORIA MILLS PTY. LTD.
Notice is hereby given that at an Extraordinary General Meeting of the members of the abovementioned company held on 11 August 1981 it was resolved that the company be wound up voluntarily and that Alan John Truett of 461 Bourke Street, Melbourne be appointed liquidator.
Notice is also given that after 28 days from that date I shall proceed to distribute the assets. All creditors having any claims against the company should furnish particulars of same by that date, otherwise I shall proceed to distribute the assets without regard to their claim.
Dated 11 August 1981
A. J. TRUETT, Liquidator
461 Bourke Street, Melbourne 3000 4926
In the Supreme Court of Victoria—Co. 12065—In the matter of the Companies Act 1961; and in the matter of J. & M. DAVIES ENTERPRISES PTY. LTD.—Notice of Winding Up Order
Winding Up Order: Made 13 August 1981.
Name and Address of Liquidator: Michael Joseph O'Keeffe C/o O'Keeffe & Deckker 66 High Street Ashburton Vic. 3147.
B. J. O'DONOVAN, Crown Solicitor for the Commonwealth and solicitor for the petitioner 4927
In the Supreme Court of Victoria—Co. 12064—In the matter of the Companies Act 1961; and in the matter of R. & N. FREIGHTERS PTY. LTD.—Notice of Winding Up Order
Winding Up Order: Made 13 August 1981.
Name and Address of Liquidator: Kevin James Browne C/o Bentley, Wheeler, Cartledge & Co. 1st Floor, Albert Square 37 Albert Road Melbourne Vic. 3004.
B. J. O'DONOVAN, Crown Solicitor for the Commonwealth and solicitor for the petitioner 4928
In the Supreme Court of Victoria—Co. 12063—In the matter of the Companies Act 1961; and in the matter of MICRON HOLDINGS PTY. LTD.—Notice of Winding Up Order
Winding Up Order: Made 13 August 1981.
Name and Address of Liquidator: John Martin Walsh C/o Wallace, McMullin & Smail 499 St Kilda Road Melbourne Vic. 3004.
B. J. O'DONOVAN, Crown Solicitor for the Commonwealth and solicitor for the petitioner 4929
In the Supreme Court of Victoria—Co. 12062—In the matter of the Companies Act 1961; and in the matter of WILBERFORCE MANAGEMENT SERVICES (Vic.) PTY. LTD.—Notice of Winding Up Order
Winding Up Order: Made 13 August 1981.
Name and Address of Liquidator: Ronald Dennis Widows Chartered Accountant 703 South Road Moorabbin Vic. 3181.
B. J. O'DONOVAN, Crown Solicitor for the Commonwealth and solicitor for the petitioner 4930
RITE-LINE UNIFORMS PTY. LTD.
NOTICE PURSUANT TO SECTION 254 (2) (b) OF THE COMPANIES ACT 1961
Notice of Resolution
At an Extraordinary General Meeting of members of Rite-Line Uniforms Pty. Ltd. held at 5 Plummer Road, Mentone on 17 August 1981 a Special Resolution was passed that the Company be wound up voluntarily and that Donald Alexander Neil McLardy of M. R. M. Smith, Peacock & Henshaw, Chartered Accountants, 5th Floor, 124 Exhibition Street, Melbourne, be appointed Liquidator.
Dated 17 August 1981
4931 D. A. N. McLARDY, Liquidator
ALEXANDER LANCELLOT HITCHINS, late of 308-310 Raymond Street, Sale, cycle manufacturer, Deceased
Creditors, next of kin and others having claims in respect of the Estate of the Deceased who died on 19 September 1980 at 308-310 Raymond Street, Sale, are required by Xavier Crofton Lee, solicitor of 17 Myerova Street, Pymble in the State of New South Wales by the Supreme Court of Victoria on 22 June 1981 and required by the said Francis Xavier Crofton Lee, and other particulars to take care of the undermentioned Solicitors by 26 October 1981 after which date the said Francis Xavier Crofton Lee may convey or distribute the assets, having regard only to the claims of which he then has notice.
Dated 5 August 1981
WARREN, GRAHAM & MURPHY, solicitors, 99 Raymond Street, Sale 3350 4813
Creditors, next of kin and others having claims in respect of the estate of Robert Simpson late of 58 Durham Street Heidelberg Retired deceased who died on 28 August 1980 the Estate of the Deceased will be distributed to Sarah Agnes Simpson of 58 Durham Street Eaglemont. We are required to send particulars of their claims to the said executrix care of the undermentioned Solicitors by 19 October 1981 after which date she will distribute the assets having regard only to the claims of which she then has notice.
PEARCE & WEBSTER, solicitors, 430 Little Collins Street, Melbourne 4814
Creditors, next of kin and others having claims in respect of the estate of Florence May McFarlane late of Melville Bacteriologist, Spinning doctor, who died on 10 April 1981 and Probate of whose Will has been granted to Charles Austin McFarlane of 7 Ash Grove Dandenong Retired are required to send particulars of their claims to the said executor care of the undermentioned Solicitors by 30 October 1981 after which date he will distribute the assets having regard only to the claims of which he then has notice.
PEARCE & WEBSTER, solicitors, 430 Little Collins Street, Melbourne
Creditors, next of kin and others having claims in respect of the estate of James Kenneth McCarroll also known as Kenneth James McCarroll late of 5 Steadman Street, Swan Hill Railway Employees deceased who died on or about 17 June 1981 are to send particulars of their claims to the administratrix Beverley Florence McCarroll of 75 Vine Street, Bendigo, care of the undermentioned Solicitors by 26 October 1981, after which date she will distribute the assets having regard only to the claims of which she then has notice.
EVERY & EVERY, solicitors, 71 Bull Street, Bendigo
Creditors, next of kin and others having claims in respect of the estate of Albert Hamilton Pitcher late of 50 Glencoe Street, Bendigo, Farmer deceased, who died on 19 March 1981, are to send particulars of their claims to the executor Albert Ambrose Pitcher of 39 Williamson Street, Bendigo, care of the undermentioned Solicitors by 26 October 1981, after which date he will distribute the assets having regard only to the claims of which he then has notice.
EVERY & EVERY, solicitors, 71 Bull Street, Bendigo
HAROLD JOHN MASON, late of Jindivick, farmer, DECEASED
Creditors, next of kin and others having claims in respect to the estate of the deceased who died on 15 June 1981 are required by the Trustees Elsie Irene Emily Mason Brackenbury, John and Heather Mary Taylor to send particulars of their claims to these care of the undersigned Solicitors by 23 October 1981 after which date the Trustees may convey or distribute the assets having regard only to the claims of which they then have notice.
GRAY, FRIEND & LONG, solicitors, Warragul
EDNA GUNN, late of 10 Victoria Street, Williamstown, widow, DECEASED
Creditors, next of kin and others having claims against the estate of the said deceased who died on 19 April 1981 are to send particulars of their claims to Walter Campbell Gunn and Diane Withers care of 536 Whitehorse Road Mitcham by 30 October 1981 after which date they will distribute the assets having regard only to the claims of which they then have notice.
MULCAHY MENELSON & ROUND, solicitors, 536 Whitehorse Road, Mitcham
Creditors, next of kin and others having claims in respect of the estate of Ann John Lintern late of 9 Penlyne Avenue Vermont in the State of Victoria Sales Assistant deceased who died on 17 December 1980 are to send particulars of their claims to National Trustees Executors and Agency Company of Australasia Limited of 95 Queen Street Melbourne in the said State and Ann Evelyn Lintern Married Woman and Henry Joseph Lintern both of 9 Penlyne Avenue Vermont in the said State care of the National Trustees Executors and Agency Company of Australasia Limited by 19 October 1981 after which date they will distribute the assets having regard only to the claims to which they then have notice.
REGINALD C. BUTLER & CO., solicitors, 312 Centre Road, Bentleigh
JEAN MARIE FORISTAL, late of Hillview Private Nursing Home, 764 Canterbury Road, Surrey Hills, in the State of Victoria, widow, DECEASED
Creditors, next of kin and others having claims in respect of the estate of the abovenamed deceased, who died on 13 January 1981 are required by the joint Executrices Jean Gladys Orchard of 15 Beckett Road, Donvale in the said State Married Woman and Elizabeth Mary Main Street, Abbotsford in the said State Widow to send particulars of their claims to the Executrices care of the undermentioned Solicitor by 30 October 1981 after which date the Executrices may convey or distribute the assets having regard only to the claims of which they then have notice.
WINIFRED McCOOK, solicitor, 173 Whitehorse Road, Ringwood, Victoria
HILDA RUBERTRA MAYO, late of the Convent of the Good Shepherd, Abbotsford, in the State of Victoria, sister in religion, DECEASED
Creditors, next of kin and others having claims in respect of the estate of the deceased who died on 25 May 1981 are required by the executor, National Trustees Executors and Agency Company of Australasia Limited of 95 Queen Street, Melbourne to send particulars to it by 25 November 1981 after which date the executor may convey or distribute the assets having regard only to the claims of which it then has notice.
Dated 12 August 1981
CORNWALL STODART & CO., of 380 Lonsdale Street, Melbourne, solicitors for the National Trustees Executors and Agency Company of Australia Limited
Creditors, next of kin and others having claims in respect of the estate of the deceased Henry Wright late of "Coomararo", Alawna Road, Toorak, Retired Company Director deceased who died on 24 February 1981 are required by the personal representative, Honoree Michael Wright of 3 Alphington Street, Fairlight Barrister-at-Law and Stephen Ronald Marriott of 124 Stanhope Street, Malvern Company Director to send particulars to them at the undersigned Solicitors by 23 October 1981 after which date the said personal representative may convey or distribute the assets having regard only to the claims of which they then have notice.
PAVEYS, solicitors, 360 Collins Street, Melbourne, 3000
ELLEN VERA CLARK, formerly of Lot 24 Steven Street, Hurstbridge, but late of 4 Ashleigh Road, Armadale, widow, DECEASED
Creditors, next of kin and others having claims in respect of the estate of the deceased who died on 10 March 1981 are required by the Executors Stuart James Clark of 15 Stedcombe Street Alawna via Dandenong North Territory Public Service and Geraldine Clark of 4 Ashleigh Road, Armadale, Widower to send particulars to the undermentioned firm by 30 September 1981 after which date the said executors may convey or distribute the assets having regard only to the claims of which notice has by then been given.
SETON WILLIAMS & SMYTH, solicitors, 409 Collins Street, Melbourne
Creditors, next of kin and others having claims in respect of the estate of Elizabeth Elvira McFarlane late of 195 Best Street, East Lake Married Woman deceased who died on 1 June 1981 are required by the Executors The Trustees Executors and Agency Company Limited of 401 Collins Street, Melbourne to send particulars of their claims to the said Trustee Company by 21 October 1981 after which date the Executors may convey the assets having regard only to the claims of which it then has notice.
MESSRS H. L. YUNCKEN & YUNCKEN, solicitors, 271 William Street, Melbourne, 3000
Creditors, next of kin and others having claims against the estate of Maile Morrissey formerly of 4 Hoyningen Square, Toorak in the State of Victoria but late of Unit 2 No. 69 Sackville Street, Kew aforesaid Widow deceased who died on 20 April 1981 are required to send particulars of their claims to the Executors, Mr and Mrs John Brendan Morrissey and Margaret Mary Nicholson of care of the undermentioned Solicitors on or before 30 October 1981 after which date the Executors will distribute the assets of the Estate having regard only to the claims of which he shall then have notice.
JOHN M. GALVIN & ASSOCIATES, solicitors, 343 Little Collins Street, Melbourne, 3000
Creditors, next of kin and others having claims in respect of the Estate of Dorothy Hazel Archer late of Casita Gardens, Little Wattle Bay, Toorak, in the State of Victoria deceased, who died on 6 March 1980, are required by the Administrator Richard Walton Archer of 4/66 Beresford Road Rose Bay in the State of New South Wales Airline Pilot to send particulars of their claims to him in care of the Administrator, 21 Collins Street Melbourne, after which date he will distribute the assets having regard only to the claims of which he then has notice.
McCAY & THWAITES, solicitors, 395 Collins Street Melbourne
Creditors, next of kin and others having claims in respect of the Estate of Denis Annie Fitzgibbon of Ficarra Private Hospital 47 Riverdale Road Hawthorn, in the State of Victoria Retired deceased, who died on 23 May 1981, are required by the executor William Andrew Thwaites of 7/395 Collins Street Melbourne in the said State Solicitor to send particulars of their claims to him in care of the undermentioned solicitors by 21 October 1981 after which date he will distribute the assets having regard only to the claims of which he then has notice.
McCAY & THWAITES, solicitors, 395 Collins Street Melbourne
JOHN BINNS THOMAS, late of 8 Wheeler Street, Berwick, retired, DECEASED
Creditors, next of kin and others having claims in respect of the Estate of the deceased who died on 5 May 1981 are required by the Trustees Ivan Binns Thomas Butcher of 8 Hopewell Street, Berwick and Therese Jean Hopkins Married Woman of "Dhoo-ba" Lyndhurst Road, Lyndhurst to send particulars to them at 376 Collins Street, Melbourne by 21 October 1981 at which date the Trustees may convey or distribute the assets having regard only to the claims of which they then have notice.
Dated 19 August 1981
JOHN P. RHODEN, solicitors, 376 Collins Street, Melbourne
ALFRED LLEWELYN EDWARDS, late of 1372 Dandenong Road, Oakleigh, in the State of Victoria, retired, DECEASED
Creditors, next of kin and others having claims in respect of the Estate of the deceased who died on 29 May 1981 are required by the trustee Emily Edwards of 1372 Dandenong Road Melbourne in the said State to send particulars to the undermentioned Solicitors to send particulars to her by 19 October 1981 after which date the trustee may convey or distribute the assets having regard only to the claims of which she then has notice.
PHILLIPS FOX AND MASEL, 461 Bourke Street, Melbourne
Creditors, next of kin and others having claims in respect of the estate of Robert Henry Bligh late of 24 Lyndhurst Road Hawthorn in the State of Victoria Gentleman deceased who died on 3 May 1981 are required by the executor National Trustees Executors and Agency Company of Australasia Limited on 25 Queen Street Melbourne in the said State to send particulars of their claims to the said Company by 20 October 1981 after which date the said Company will distribute the assets having regard only to the claims of which it then has notice.
DARVALL & HAMBLETON, solicitors, of 80 Collins Street, Melbourne, 3000
Creditors, next of kin and others having claims in respect of the estate of Phyllis Mary Gordon, Pale late of Glenore Nursing Home Berries Road Cookham Berkshire in the United Kingdom Widow who died on 30 December 1980 are required to send the particulars of their claims to The Trustees Executors and Agency Company Limited of 401 Collins Street Melbourne by 21 October 1981 after which date it will distribute the assets having regard only to the claims of which it then has notice.
HADEN SMITH & FITCHETT, solicitors, 423 Bourke Street, Melbourne
Creditors, next of kin and others having claims in respect of the Estate of Gwendolyn Alice Oliver Gordon late of 60 Jennings Road, Kyabram Widow deceased who died on 1 June 1981 are to send particulars of their claims to John Watson McCallum and Elizabeth Louise Marson care of the undersigned by 21 October 1981 after which date they will distribute the estate having regard only to the claims of which they then have notice.
COOK & MCCALLUM, solicitors, 422 Collins Street, Melbourne
Creditors next of kin and other persons having claims against the estate of Henry Richard Chamber late of 12 Meryl Street Armadale in the State of Victoria Retired Textile Worker deceased who died on 9 June 1981 are required to send particulars of their claims to the Executor National Trustees Executors and Agency Company of Australasia Limited in the said State to send particulars to them at 4-6 Howitt Street Melbourne by 20 October 1981 after which date the Executor will distribute the assets having regard only for the claims of which it then has had notice.
JAMES P. OGGE & CO. AND E. K. O'DONNELL, solicitors, 4-6 Howitt Street, South Yarra
Creditors next of kin and others having claims in respect of the estate of Stella Florence Smiles late of 3/14 Nunn Street Benalla Widow who died on 10 May 1981 are requested to send particulars of their claims in writing to the undermentioned Solicitors for the executors George Alfred Hagenauer and Pearl Lorraine Hagenauer by 9 November 1981 after which date the executors will distribute the assets having regard only to the claims of which they then have notice.
HAMILTON, CLARKE & BALKIN, solicitors, 55 Nunn Street, Benalla
Creditors next of kin and other persons having claims against the estate of Mary Seenie Hermine Walmsley late of Section Road, Greenvale in the State of Victoria deceased who died on 9 June 1981 are required to send particulars of their claims to the Executor The Trustees Executors and Agency Company Limited of 401 Collins Street, Melbourne by 15 October 1981 after which date the Executor will distribute the assets, having regard only for the claims of which it then has had notice.
THE TRUSTEES, EXECUTORS & AGENCY COMPANY LIMITED, of 401 Collins Street, Melbourne
JOSEPH MORLEY, late of 12 High Street, Coburg, in the State of Victoria, retired, DECEASED
Creditors, next of kin and others having claims in respect of the estate of the deceased, who died on 13 May 1980 whose Will was proved by John Francis Morley of 12 High Street, Coburg, in the Executors appointed therein on 16 April 1981 are required by the Executor to send particulars to him by 20 October 1981 after which date the Executor may convey or distribute the assets, having regard only to the claims of which he then has notice.
Dated 14 August 1981
LLOYD P. GOODE & CO., solicitors, 406 Lonsdale Street, Melbourne
Creditors next of kin and others having claims against the estate of Marie Elling Florence Twyford late of 16 Karinna Drive, Frankston in the said State Minister deceased who died on 21 March 1981 are required by the Executors Philip Stanborough of 4 Myrtle Grove, Blackwood Retired Public Servant and Francis Read Monotti of 13 Scott Street, Dandenong Solicitor to send particulars of their claims to the said Executors and of Messrs R. Monotti & Co. of 13 Scott Street, Dandenong, Solicitors, by 2 November 1981 after which date they will distribute the assets of the said estate having regard only to the claims of which they then have notice.
F. R. MONOTTI & CO., solicitors, 13 Scott Street, Dandenong
Creditors, next of kin and others having claims in respect of the Estate of Eva Catherine Bilton late of Parkhurst Private Hospital Lygon Street North Carlton Widow deceased who died on 10 February 1981 are required to send the particulars of their claims to David Thomas
Mottram, care of Richard Dunstan & Partners Solicitors of 396 Little Bourke Street, Melbourne by 30 October 1981 after which date he will distribute the assets of the Estate having regard only to the claims of which he then has notice.
RICHARD DUNSTAN & PARTNERS, solicitors, 396 Little Bourke Street, Melbourne
PERCY PIGGOTT ANDERSON, formerly of 9 Anderson Street, Kewburn, but late of 15 Wardale Road, Springvale South, retired, Deceased, intestate
Creditors, next of kin and others having claims in respect of the Estate of the deceased who died on 20 June 1980 are required by Kenneth Harold Anderson formerly of 128 Nicholson Street, Coburg but now of 20 Old Stuart Street, Caulfield Retired, the Administrator of the Estate of the said deceased to send particulars to him in the care of the undermentioned Solicitors by 31 October 1981 after which date the said Administrator may convey or distribute the assets having regard only to the claims of which he then has notice.
MACPHERSON AND KELLEY, solicitors, 229 Thomas Street, Dandenong
MARGARET ENID WYNNE DOWLING, late of "Bimbaden", Ocean Grove in the State of Victoria, widow, Deceased
Creditors, next of kin and others having claims in respect of the abovenamed deceased (who died on 14 August 1979) are required by the Executors (John Willoughby Kenny formerly of 10 Dawson Avenue, Elsternwick in the State of Victoria, now of 9 Bromby Street, South Yarra in the said State, Accountant and Mary Wynne Grove of 11 Melrose Avenue, Highton in the said State, Married Woman) to send particulars of such claims to the undermentioned Solicitors by 30 October 1981 after which date the said Executors may convey or distribute the assets having regard only to the claims of which it has notice.
MESSRS HARWOOD & PINCOTT, solicitors, 77 Moorabool Street, Geelong
Creditors, next of kin and others having claims in respect of the Estate of Jack Rexhep (also known as Jack Rexhep and Jack Rexhap), formerly of 179 Waverley Road East Malvern but late of Amahla House 1019 Glenhuntly Road Caulfield South in the State of Victoria formerly Wharf Lanehouse, late of 1019 Glenhuntly Road, who died on 14 March 1981 are to send the notices of their claims to Daut Feim formerly of 137 Park Drive Parkville but now of 78 Harding Street Coburg in the said State formerly Wharf Lanehouse, late of 78 Harding Street, Pensioner, Executor of the Estate of the said Deceased care of Leo Browne Solicitor of 180 Elgin Street Carlton in the said State by 30 November 1981 after which date the said Daut Feim will distribute the Estate having regard only for the claims of which he then has notice.
LEO BROWNE, solicitor, of 180 Elgin Street, Carlton
Creditors, next of kin and others having claims in respect of the Estate of John Blackburn, late of 368 Clarke Street, Northcote, Retired Foreman, deceased, who died on 13 June 1981 are required by the Executors The Trustees Executors and Agency Company Limited of 401 Collins Street, Melbourne to send particulars of their claims to the said Trustee Company by 21 October 1981 after which date it will distribute the assets having regard only to the claims of which it then has notice.
MESSRS H. L. YUNCKEN & YUNCKEN, solicitors, 271 William Street, Melbourne
JOHN COLIN MORRISON, late of 52 Russell Street, Grafton, in the State of New South Wales, retired, Deceased
Creditors, next of kin and others having claims in respect of the estate of the deceased (who died on 22 January 1981) are required by The Perpetual Executors and Trustees Association of Australia Limited of 50 Queen Street, Melbourne to send particulars of their claims to the said Company by 21 October 1981 after which date it will convey or distribute the assets having regard only to the claims of which the Company then has notice.
TURNER, NESS & DAVIS, solicitors, 8-12 Batman Street, Melbourne
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Wednesday, 7 October 1981 at 12.00 noon at the Police Station, Foster (unless process be stayed or satisfied).
All the Estate and Interest (if any) of Hedgeley Dene Friesian Farm Pty. Ltd., registered office, Southdown Farm, 141 Osborne Street, Yarraville, proprietor of an estate in fee simple in the land described in Certificate of Title Volume 8143 Folio 216 upon which is erected three brick veneer dwellings and numerous outbuildings and is situated on the northerly bank of Waratah River approximately 11 kilometres from Fish Creek, 7 kilometres from Sandy Creek, 22 kilometres from Foster, 45 kilometres from Leongatha, 16 kilometres from Walkerville and overlooks Waratah Bay and inlets. The property is known as Hedgeley Dene.
Registered Mortgage No. H.751334 affects the said estate and interest.
Terms—Cash only
J. GREEN, Deputy Sheriff
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Wednesday, 23 September 1981 at 11.30 a.m. at the Police Station, Wangaratta (unless process be stayed or satisfied).
All the Estate and interest (if any) of Michael William Mace, same representative, and Patricia Beryl Mace, married woman, both of 20 O'Keefe Street, Wangaratta as joint proprietors of an estate in fee simple in the land described in Certificate of Title Volume 8194 Folio 474 upon which is erected a weatherboard dwelling known as No. 20 O'Keefe Street, Wangaratta.
Registered Mortgage No. H.111408 affects the said estate and interest.
Terms—Cash only
J. GREEN, Deputy Sheriff
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Wednesday, 30 September 1981 at 12 noon at the Police Station, Colac (unless process be stayed or satisfied).
All the Estate and interest (if any) of Marian L. Egbert (owner on Certificate of Title as Marian Lois Egberts), trading as Noela Rose, proprietoress, of 118 Hart Street, Colac, as joint proprietor with Abraham Egberts, teacher, of an estate in fee simple in the land described in Certificate of Title Volume 6448 Folio 615 upon which is erected a brick veneer dwelling known as No. 118 Hart Street, Colac.
Registered Mortgage Nos. H.241820 and H.579930 affect the said estate and interest.
Terms—Cash only
K. R. MARTIN, Sheriff's Officer
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Thursday, 1 October 1981, at 2.30 p.m. at the Bailiff's Office, 2nd Floor, 300 Burwood Road, Hawthorn (unless process be stayed or satisfied).
All the Estate and Interest (if any) of Dimitrios Kassie, machine operator, of Lethbridge Street, Moonee Ponds, as joint proprietor with Evrikia Kassie, married woman, of an estate in fee simple in the land described in Certificate of Title Volume 4555 Folio 982 upon which is erected a weatherboard dwelling known as No. 9 Lethbridge Street, Moonee Ponds.
Registered Mortgage No. J.268268 affects the said estate and interest.
Terms—Cash only
4935 O. V. DALY, Sheriff's Officer
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Wednesday, 23 September 1981, at 2.30 p.m. at the Bailiff's Office, 2nd Floor, 300 Burwood Road, Hawthorn (unless process be stayed or satisfied).
All the Estate and Interest (if any) of A. Kyriakidis (shown on Certificate of Title as Antonios Kyriakidis), labourer, of 113 Union Street, Brunswick, as joint proprietor with Hrisoula Kyriakidis, married woman, of an estate in fee simple in the land described in Certificate of Title Volume 3213 Folio 557 upon which is erected a weatherboard house known as No. 113 Union Street, Brunswick.
Registered Mortgage No. F.545410 affects the said estate and interest.
Terms—Cash only
4936 O. V. DALY, Sheriff's Officer
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Wednesday, 23 September 1981, at 2.30 p.m. at the Bailiff's Office, 2nd Floor, 300 Burwood Road, Hawthorn (unless process be stayed or satisfied).
All the Estate and Interest (if any) of Mr R. Saunders (shown on Certificate of Title as Raymond Edward Saunders), builder, of 4 Tambo Court, Mentone as joint proprietor with Margaret Eileen Saunders, married woman, of an estate in fee simple in the land described in Certificate of Title Volume 9273 Folio 102 upon which is erected a brick veneer dwelling known as No. 4 Tambo Court, Mentone.
Registered Mortgage No. H.822493 affects the said estate and interest.
Terms—Cash only
4937 T. M. NICHOLLS, Sheriff's Officer
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Thursday, 1 October 1981 at 2.30 p.m. at Bailiff's Office, 2nd Floor, 300 Burwood Road, Hawthorn (unless process be stayed or satisfied).
All the Estate and Interest (if any) of Thomas King, sales representative and Joan King, married woman (Shown on Certificate of Title as Thomas Henry King and Jean King), both of 5 Thomas Street, East Doncaster as joint proprietors of an estate in fee simple in the land described in Certificate of Title Volume 8800 Folio 913 upon which is erected a brick dwelling known as No. 5 Thomas Street, East Doncaster.
Registered Mortgage Nos. E.107216, G.221537 and Caveat No. J.290318 affect the said estate and interest.
Terms—Cash only
4940 T. N. NICHOLLS, Sheriff's Officer
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Wednesday, 23 September 1981 at 2.30 p.m. at Bailiff's Office, 2nd Floor, 300 Burwood Road, Hawthorn (unless process be stayed or satisfied).
All the Estate and Interest (if any) of Paul T. Coppens (Shown on Certificate of Title as Paul Theodore Coppens), musical director, of 146 Bolton Street Etham, as joint proprietor of an estate in fee simple in the land described in Certificate of Title Volume 8269 Folio 412 upon which is erected a brick veneer dwelling known as No. 146 Bolton Street Etham.
Registered Mortgage No. F.921000, F.921001 and G.668347 affect the said estate and interest.
Terms—Cash only
4942 K. R. MARTIN, Sheriff's Officer
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Wednesday, 23 September 1981 at 2.30 p.m. at Bailiff's Office, 2nd Floor, 300 Burwood Road, Hawthorn (unless process be stayed or satisfied).
All the Estate and Interest (if any) of Samuel Gioia (Shown on Certificate of Title as Salvatore Gioia), bricklayer of 62 Medway Street, Box Hill as joint proprietor with Lynette Rosemary Gioia of an estate in fee simple in the land described in Certificate of Title Volume 3614 Folio 787 upon which is erected a dwelling house known as No. 62 Medway Street Box Hill.
Registered Mortgage No. C.288182 and Caveat Nos. H.110606 and H.790360 affect the said estate and interest.
Terms—Cash only
4943 K. R. MARTIN, Sheriff's Officer
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Wednesday, 23 September 1981 at 2.30 p.m. at Bailiff's Office, 2nd Floor, 300 Burwood Road, Hawthorn (unless process be stayed or satisfied).
All the Estate and Interest (if any) of Howard Ostle Borowski, electrician, 46 Service Street, Hampton as joint proprietor with Mary Julienne Lila Borowski, Married Woman of an estate in fee simple in the land described in Certificate of Title Volume 7303 Folio 600 upon which is erected a single storey weatherboard residence known as No. 46 Service Street, Hampton.
Registered Mortgage No. F.739809 affects the said estate and interest.
Terms—Cash only
4944 T. M. NICHOLLS, Sheriff's Officer
In the Supreme Court of the State of Victoria
SALE BY THE SHERIFF
On Wednesday, 23 September 1981 at 2.30 p.m. at Bailiff's Office, 2nd Floor, 300 Burwood Road, Hawthorn (unless process be stayed or satisfied).
All the Estate and Interest (if any) of K. Greenwell (Shown on Certificate of Title as Keith Edward Greenwell), truck driver of 107 Hotham Road, Sorrento as joint proprietor with Maria Denise Greenwell, married woman of an estate in fee simple in the land described in Certificate of Title Volume 7743 Folio 210 upon which is erected a dwelling known as No. 107 Hotham Road, Sorrento.
Registered Mortgage No. F.400820 affects the said estate and interest.
Terms—Cash only
4945 T. M. NICHOLLS, Sheriff's Officer
INSOLVENCY NOTICES
COMMONWEALTH OF AUSTRALIA
Bankruptcy Act 1966–1980
Re: IAN KEITH AND VALERIE CLAIR HANCOCK
both of 21 Toorak Drive, Dingley
NOTICE OF MEETING
Take notice that the abovenamed Ian Keith Hancock and Valerie Clair Hancock have on 13 August 1981, signed an authority under sub-section (1) of Section 88 of the Bankruptcy Act 1966–1980 authorising Barry Keith Taylor of 576 St. Kilda Road, Melbourne, to call a meeting of their creditors and to take control of their property, and that in pursuance of Section 194 of the Bankruptcy Act 1966–1980, a Meeting of Creditors of the abovenamed Debtors will be held at the Meeting Room, B. K. Taylor & Co., 3rd Floor, Dairy Industry House, 576 St. Kilda Road, Melbourne on Friday 4 September 1981, at 11.00 a.m.
Dated 13 August 1981
B. K. TAYLOR, Controlling Trustee
B. K. Taylor & Co., accountants, 3rd Floor, 576 St. Kilda Road, Melbourne, Vic., 3004
COMMONWEALTH OF AUSTRALIA
Bankruptcy Act 1966–1980
DAVID COLIN AND MARGARET HYDE
both of 68 Benga Avenue, Dandenong
PHILLIP ELLISEN AND LORAIN EVELYN HYDE
both of 16 Phillips Street, Dandenong North, formerly trading as Gem Kitchens
NOTICE OF MEETING
Take notice that the abovenamed David Colin Hyde, Margaret Hyde, Phillip Ellison Hyde and Loraine Evelyn Hyde have on 13 August 1981, signed authorities under sub-section (1) of Section 188 of the Bankruptcy Act 1966–80 authorizing Barry Keith Taylor of 576 St. Kilda Road, Melbourne to call a meeting of their creditors and to take control of their property and that in pursuance of Section 194 of the Bankruptcy Act 1966–1980, a Meeting of Creditors of the abovenamed Debtors will be held at the Meeting Room, B. K. Taylor & Co., 3rd Floor, 576 St. Kilda Road, Melbourne on Thursday 3 September 1981 at 11.00 a.m.
Dated 10 August 1981
B. K. TAYLOR, Controlling Trustee
B. K. Taylor & Co., accountants, 3rd Floor, Dairy Industry House, 576 St. Kilda Road, Melbourne, Vic., 3004
NOTICE OF MAKING OF STATUTORY RULES WHICH ARE NOT YET AVAILABLE
Notice is given of the making of the following Statutory Rules:
No. Stamps Act 1958
356/1981. Stamps Regulations 1981
Transport Regulation Act 1958
362/1981. Transport Consolidated (Revocation of Licence Fees) Regulations 1981
Local Government Act 1958
363/1981. Municipal Electrical Engineers Board (Amendment 2) Regulations 1981
Port of Melbourne Authority Act 1958 (No. 6312)
364/1981. Port of Melbourne Authority (Superannuation) Regulations 1981, Amendment No. 1
Labour and Industry Act 1958
365/1981. Labour and Industry (Foundries) (Amendment) Regulations 1981
Rural Finance and Settlement Commission Act 1961
366/1981. Soldier Settlement and Land Settlement (Crown Grant Fees) Regulations 1981
Films Act 1971
367/1981. Films (Amendment) Regulations 1981
Melbourne and Metropolitan Board of Works Act 1958
368/1981. By-Law No. 179: Fees for Restriction or Withdrawal and Restoration of Water Supply
NOTICE OF MAKING AND AVAILABILITY OF STATUTORY RULES
In pursuance of the provisions of the Subordinate Legislation Act 1963 and the Regulations made thereunder notice is given of the making and availability of the following Statutory Rules:
| No. | Title | Price |
|--------------|----------------------------------------------------------------------|-------|
| 357/1981. | Stock Medicines (Standards) Regulations 1981 | 20c |
| | West Moorabool Water Board Act 1968 | |
| 358/1981. | West Moorabool Water Board (Chairman's Fee) (Revocation) Regulations 1981 | 20c |
| 359/1958. | Road Traffic Act 1958 | |
| 359/1958. | Road Traffic (Amendment) Regulations 1981 | 20c |
| 360/1981. | Local Government Act 1958 | |
| 360/1981. | Building Qualifications Board (Building Inspectors) Regulations 1981 | 20c |
| 361/1981. | Local Government Act 1958 | |
| 361/1981. | Building Qualifications Board (Building Surveyors) Regulations 1981 | 20c |
| | Public Service Act 1974 | |
| | PSDI09/1981. Public Service Amendment Determinations (No. 109) 1981 | 20c |
Copies of these Statutory Rules are available and may be purchased at the Victorian Government Printer Shop, located at 7A Parliament Place, Melbourne 3002. If ordering by mail, remittance should be addressed to "The Government Printer, Box 203, P.O., North Melbourne, 3051", and should include 50c delivery and handling fee.
The annual subscription rates for Statutory Rules for the year commencing 1 January 1981, payable in advance, are as follows:
*Statutory Rules (other than Public Service Determinations), Pamphlet copies only $42.00
Public Service Determinations $30.00
*The Bound Volume of Statutory Rules is not included in the subscription rate.
F. D. ATKINSON
Government Printer
THE "VICTORIA GOVERNMENT GAZETTE"
PUBLICATION—A Victoria Government Gazette is published every Wednesday evening unless advertised otherwise.
SUBSCRIPTIONS—The annual subscription rate, including postage, is $72, payable in advance.
PRIVATE ADVERTISEMENTS—Advertisements will be received by the Government Printer not later than the day preceding publication except ordinary notices at or before one p.m. and at double rates between one p.m. and half past three p.m. The charge for advertisements is 90 cents per line in single column and $1.80 per line in double column. The charge for a full page is $10.00. The title forms one line as a heading. On an average ten words make a line of single column. Every signature must likewise be counted as a line. The final words of a paragraph though only portion of a line must be counted as one line. Signatures (in particular) and proper names must be written very clearly in the text; one side only of each sheet of paper should be written upon.
All documents NOT CLEARLY PREPARED will be returned unpublished. Where brands occur, not provided for by the ordinary letters of the alphabet, a written explanatory description also must be supplied.
PAYMENTS—Unless the advertiser has a Credit Account, all payments are required in advance. Remittances should be made by cheque, postal order or money order, payable to "GOVERNMENT PRINTER".
CORRESPONDENCE—All correspondence should be addressed to "The Victorian Government Printer, P.O. Box 203, North Melbourne, 3051".
CONTENTS
Appointments 2759
Bank Holidays 2739
Cemeteries—Scale of Fees 2753
Companies Act 1961—Notice 2743
Contracts 2758
Country Roads Board 2747
Courts 2745
Estates of Deceased Persons 2757
Government Notices 2740
Insolvency Notices 2793
Lands 2772
Minerals and Energy 2746
Notice to Mariners 2742
Notice of Making of Statutory Rules 2793
Orders in Council—Acts—Railway Construction and Property Board; Housing; Drainage of Land; Superannuation; Country Roads Board; Crown Land (Reserves); Land Conservation and Soil Conservation and Land Utilization; Railway Construction and Property Board; Labour and Industry; Victoria Law Foundation; Forests; Industrial Training; Local Government; Soil Conservation and Land Utilization; Geelong Waterworks and Sewerage. 2761 et seq
Police Sale 2746
Private Advertisements 2773
Proclamations 2739
Public Holidays 2739
Regulations—Act—Teaching Service 2752
Resignations 2760
Tenders 2772
Transport Regulation Board—Public Hearings 2741 |
Random Access to the Time Domain in the AMPLE language
Chris Jordan
Hybrid Technology Ltd
Cambridge
Great Britain
ABSTRACT
AMPLE is a powerful and versatile music programming language and environment for use on microcomputers. The language 'nucleus' includes a wide range of computation and sound control functions for advanced musical applications, linked to an innovative textual music notation for more traditional musical forms. Higher-level components of the environment cater for specific user requirements such as instrument design, real-time performance control and staff notation, and low-level software interfaces provide extended access to a variety of musical input and output devices, both via MIDI and directly via host-specific connections. Currently, the most popular use of AMPLE is in the Hybrid Music System for the BBC Microcomputer, including the Music 5000 Synthesiser.
AMPLE is a procedural language with a 'word' structure like that of LOGO - the user program is a hierarchy of words, each defined in terms of pre-existing user words and pre-defined system words. Words can be created and edited individually, interactively, and in a variety of forms, always including the traditional textual program form.
The user program runs in real time under the control of a time manager that merges and sorts events from all concurrent processes before playing. This allows random access to the time domain (over a range limited by system load and memory capacity), so that as well as the usual positive time intervals between events, negative time intervals can be accommodated. Thus, programs and scores are free to generate event sequences in non-time orders that better suit the algorithms of the program or score. The major reward is simpler programming of a variety of overlapping musical structures, both large and small scale, both via low-level sound control words and, in particular, AMPLE's textual music notation.
Simple applications of the random time access facility are described, along with related features of the AMPLE textual music notation.
NOTATION ELEMENTS
AMPLE music notation has words (symbols) for four basic types of musical event:
| word | function |
|------|----------|
| letters A-G & a-g | notes of different pitches |
| X | hit (pitchless note) |
| ^ | rest |
| / | hold (continue last event) |
Other words set the parameters of events, the most important here being ',' (comma) which sets the length in timebase units. In scoring, a basic 'beat' length is set with ',', and multiples of this are achieved by extension with the / word (hold). For example:
```
envelope
staff
AMPLE 12, X // // X // // X // X // // // // /
```
On playing, X (hit) expands into a 'gate on' event followed by a 'duration' (time interval). ^ (rest) expands into a 'gate off' and duration, and / into just a duration. In each case, the length of the duration is that set by the previous ',', in this case 12 units.
Hence, this simple music event sequence translates into a sound event sequence of just 'gate on' and 'gate off' events with separating time intervals.
The full set of AMPLE music words is described in the Music 5000 User Guide (1986).
NEGATIVE LENGTHS
If a negative length setting is in force, the duration of each music event moves the 'time pointer' back rather than forward, so the next event will play before rather than after. Hence, the example could be re-scored backwards as follows:
```
-12, // // X // X // X // X // X
```
The sounding result is the same, but there is a net backwards movement in time rather than a forwards one, so subsequent music events will be displaced.
BACK-HOLD
The back-hold word, \ , has the effect opposite to that of the hold word, /. It has the effect of a hold with the current length negated, so, for example, the sequence
```
/// \\\
```
has a zero net result.
One application is in the programming of the 'overhanging pick-up' musical structure - a short sequence of notes that leads-in to a section of a piece, for example a verse of a song. This can be a problem because the pick-up is temporally part of the previous section, but must be functionally attached to its own section if this is to be called-up as a procedure in different contexts.
The negative duration ability lets the pick-up be included in its rightful functional place by allowing back-spacing from the section start to the pick-up start. Using back-holds, this is quite clear, for example:
```
12, \\ def G/// //fed c///
```
The pickup is quite free to overlay events in the tail end of the preceding section, generated by the same or another process.
CHORDS
Round brackets are used to denote chords, in which subsequent musical events of a group play alongside the first on successive voices. The subsequent events appear as a bracketed group after the first ('main') event, for example:
```
tune: C C / D A
chord sequence: C(EG) C(EA) /(F/) D(FA) ^(**)
```
In this simple example, the C, E and G in the first group play together as a C major chord. In the third group; one voice starts a new note while others are held, and the final chord of rests silences all voices.
For various reasons not discussed here, it is desirable to have a syntax like this in which additional notes are added after an unchanged main note. The availability of negative lengths allows this to be implemented by simple definitions of '(' and ')', with no need for program look-ahead or retention of the last duration for possible retraction. The simplified actions of the bracket words are as follows:
| word | action | comment |
|------|--------|---------|
| ( | n, \ | move back to main event start |
| | 0, | make further events 'stay put' |
| ) | m, / | restore net duration of main event |
| | n, | restore original ',' setting |
where
n = ',' value of the main event
m = n + total duration inside brackets
Normally, m = n since the length of bracketed events is 0.
Additional actions (not shown) control voice selection.
BROKEN CHORDS
The bipolar duration implementation of chords allows more advanced chord-like groups to be scored very simply.
The programmer may explicitly use ',' inside the brackets to set a non-zero length so that there is a time interval between bracketed events. This gives broken (arpeggiated) chords:
```
normal chords broken chords
48, C(EGB) F(ACE) 48, C(8,EGB) F(8,ACE)
Voice 4 B-----E----- B-----E-----
Voice 3 G-----C----- G-----C-----
Voice 2 E-----A----- E-----A-----
Voice 1 C-----F----- C-----F-----
```
The brackets ensure that the net duration is always that of the main note. If the total length inside the brackets exceeds that of the main event, then a net negative duration is needed - the 'strum' overlaps the next main event:
```
48, C(24,EGB) F(24,ACE)
Voice 4 B-----E----- B-----E-----
Voice 3 G-----C----- G-----C-----
Voice 2 E-----A----- E-----A-----
Voice 1 C-----F----- C-----F-----
```
Alternatively, the arpeggiating ',' setting can itself be negative, so that the strum anticipates the main event:
```
48, C(-8,EGB) F(-8,ACE)
Voice 4 B-----E----- B-----E-----
Voice 3 G-----C----- G-----C-----
Voice 2 E-----A----- E-----A-----
Voice 1 C-----F----- C-----F-----
```
Echo simulated on successive voices is very similar to broken chords from the point of view of time domain access. The difference is that the events that are directed to successive voices with successive time displacement are copies of the main event, rather than additional scored events. The word 'Echo' (provided as a special affects extension) employs AMPL's music event vector to generate copies of each event and displace them in time using / and @. The interval between echoes (positive or negative) and the number of echoes can be specified.
GATE ARTICULATION
A particularly illustrative application of random access is simple control over articulation via the gate signal. Note-separating gaps of fixed length (non-legato or 'portando') are achieved by each note or hit reaching back to prematurely terminate its predecessor. Staccato style is achieved by the note or hit reaching forward to terminate itself a fixed time in the future.
Both these reaching actions have a zero net change on the time position, so can simply be added to the standard interpretation of music events using the event vector. The articulation actions can themselves be expressed in terms of simple music events - rest, hold and back-hold. This is more obvious when the articulations are thought of as rests added between notes:
gap (non-legato)
```
envelope
staff
AMPLE 12, X // *X// *X* X* X* // * // *
```
staccato
```
envelope
staff
AMPLE 12, X * // X*// X* X* X* // * // *
```
An interesting observation is that the articulation action is a simple form of 'macro' action; one that expands each music event into a sequence of music events with derived parameters, normally used for more advanced compositional processes.
The simplified definition of the 'gap' action is as follows:
```
ACT( % start vector sequence
4, \^ % add leading rest
ACT % perform default note action
)ACT % end vector sequence
```
Notice that the second line is identical in essence to the over-hanging pickup score - each separating gap can be considered a rest that over-hangs from the following note.
The staccato action is:
```
ACT( % start vector sequence
-4, \^ % add trailing rest
ACT % perform default note action
)ACT % end vector sequence
```
An important point to note here is that it is the sounding portion of the note that adopts the fixed length of the rest, and the silent 'rest' portion is the variable-length remainder - the interval between the note's 'gate off' event and the next note's 'gate on'. The duration of the rest is negative, serving to return the time position to the start of the note.
The music event sequence on the second line of the definition is not only the most compact representation (no back-rest is available) of the function, but also illustrates the commonality in structure of the two actions. In practice, both these actions are performed by a single word 'len' (supplied in a special effects extension), which takes the length setting as a simple, positive or negative numeric argument.
Since durations can pass over existing events, there is nothing to prevent a gap or staccato interval being longer than the notes themselves. Under these conditions, each note reaches beyond its immediate predecessor and successor to modify events further removed in the time stream. Complicated interactions of note length and sequence arise, giving interesting articulation and phrasing effects. There is clearly much scope for further experimentation along these lines, for example, the use of multiple actions controlling a range of parameters as the basis for musical sequence generation and transformation in advanced compositional processes. This is well within the capabilities of current implementations of the AMPLE language.
REFERENCES
Music 5000 User Guide (1986). Hybrid Technology Ltd, Cambridge, Great Britain |
Effect of Metal Pickling and Electroplating Industrial Sludge-Borne Heavy Metals on Wheat (*Triticum aestivum*) Seedling Growth
1Sudarshana Sharma, 2Parmanand Sharma, 3Sazada Siddiqui 2A. K. Bhattacharyya
1Department of Biochemistry, Bundelkahnd University Jhansi, India
2School of Environmental Sciences, Jawaharlal Nehru University, New Delhi, India
3Department of Botany, Bundelkahnd University Jhansi, India
firstname.lastname@example.org
Abstract:
A pot culture study has been undertaken to evaluate the effect of rolling and pickling industrial sludge amendments on growth response and bioaccumulation of heavy metal in wheat seedlings. Processed acidic waste was first treated with three doses of lime (0, 0.5 and 1%) and then mixed with two soils in different ratios (0, 10 and 20%). Samples were filled in earthen pots (2Kg/pot) one week before planting and seven days old wheat seedlings (3 per pot) were transplanted in these pots and pots were kept in glass house. Temperature of glass house was maintained at 22±2°C and moisture contained at 50% of water holding capacity.
DTPA extractable heavy metals and metals in seedlings increased with increasing doses of industrial sludge amendments. Biomass and growth has been also found to increase with increasing rate of sludge. Lime enhanced the biomass and reduced the heavy metal concentrations. Although 20% treatments in both soils showed a significant enhancement in shoot length but metals like Pb was found beyond permissible limit. The heavy metal in wheat seedlings follow the trend Zn>Pb>Cu>Cd. Lime has a negative correlation with availability and uptake of heavy metals. Results showed that application of lime treated industrial sludge to soil could be useful in order to increase crop growth in the glass house. [Nature and Science 2010;8(3):1-8], (ISSN: 1545-0740).
Key words: pot culture, industrial waste, bioaccumulation, DTPA extractable metals.
1. Introduction:
The concern about heavy metals is that they are non-biodegradable and may therefore accumulate in the environment. Thus, one of the development challenges facing this decade is how to achieve a cost effective and environmentally sound strategies to deal with the global waste crisis (Alloway and Aryes, 1997). The crisis has also threatened the assimilative and carrying capacity of the earth, which is our life support system.
Not only the developed countries but developing countries like India is also facing the problem of huge quantity of solid waste releasing per day in environment. Some of them like industrial and hospital waste come under hazardous and some time it is also carcinogenic due to its composition. The National Capital Territory of Delhi, India, with a population more than 14 million, covering an area of 1483 sq Km has emerged as one of the biggest centers of small-scale industries in the country. It has 21 industrial areas, generating an enormous quantity of industrial waste per day (Office of the Commissioner of Industries, Delhi, 1996). Disposal of this perpetual quantity of waste remains a major problem. Land disposal of waste is a very old method because organic wastes obtained from different sources (Municipal, Industrial, or zoo technical) has two advantages: it avoids accumulation of wastes in the environment and it provides organic matter and nutrients for soil (Narwal et al., 1983).
Although the nutrient content of wastes makes them attractive as fertilizers, land application of many industrial and sewage sludge are constrained by the presence of heavy metals, hazardous organic chemicals, salts, and extreme pH (Cameron et al., 1997). On the other hand land application of industrial/sewage sludge is the potential source for metals accumulation in crops, which in turn will transfer to food chain posing a potential health hazard to human being (Cameron et al., 1997). Investigations illustrate the benefits of sludge are extremely important, since there is still a general reluctance among agriculturists to recognize the economic value of the sludge in order to improve the soil organic status without contaminating the environment (Korentejar, 1991). From most of the research works it is proved that the industrial and municipal waste have high amount of N and P (Mohammad and Battikhi, 1997). So, there is an increasing interest in the agricultural application of sludge due to the possibility of recycling valuable components: N, P and other plant nutrients (Gupta and Sinha, 2007).
Soil is a vital resource for sustaining two human needs of quality food supply and quality environment. Plants grown on a land polluted with municipal, domestic or industrial wastes can absorb
heavy metals in the form of mobile ions present in the soil solution through their roots or through foliar absorption. These absorbed metals get bioaccumulated in the roots, stems, fruits, grains and leaves of plants (Fatoki, 2000). By considering above facts this study was proposed to assess the effect of industrial sludge on growth and yield of wheat seedling under glasshouse conditions. Heavy metal concentrations were determined in the sludge and soil to characterize plant soil interactions of the sludge borne heavy metals on different soil types. In the best of our knowledge a very few study have been undertaken for the phytoremediation of any type of industrial sludge, hence, it is the first attempt in this field.
2. Materials and methods:
2.1. Collection of samples and preparation
Industrial sludge samples from metal finishing industries were collected from Wazirpur industrial area of Delhi. This is situated northwest part of Delhi, covering an area of 210 acre, is a big source of solid wastes generation, producing more than 30% of total solid waste of Delhi (Office of the Commissioner of Industries, Delhi, 1996). We collected 30 waste samples from 3 blocks of study site in three seasons i.e. Monsoon, Winter and Summer.
Waste samples stored in a transparent poly bags and fresh samples were used to measure pH, Electrical Conductivity, Moisture Content (%) and Water Holding Capacity (WHC). Then processed samples were subjected to a proper mixing to form a single representative sample. Analyses were done for the potential bioavailable heavy metals (Lindsay and Norvell, 1978) and total heavy metal concentration (EPA 3050 method). To neutralize the acidity of waste it was treated with lime in different ratios (0, 0.5, and 1%).
Two different soil types were collected from New Delhi, India. The site 1 soil (S1) is sandy loamy farmland soil while site 2 (S2) is a sandy nursery soil. The lime treated sludge samples were mixed to soils in 3 different ratios (0, 10 and 20%) and used to fill earthen pots (2Kg/pot) one week before planting. Wheat (HD1553) seeds were collected from National Seed Corporation, New Delhi (NSC) and seven days old seedlings were transplanted, 3 per pot, into pots. The experiment was performed under glasshouse conditions.
2.2 Soil and plant analysis
Soil samples were analyzed at two different times during the experiment: at initial stage of experiment and after 28 days of growth. The soil samples, which were analyzed at the end of experiment, consisted of soil sample and wheat seedling roots. Samples were analyzed for heavy metal content (available and total heavy metal) of four metals (Pb, Cd, Zn and Cu).
Wheat seedlings were allowed to grow for 28 days in the glass house before the harvesting. The above-ground parts were separated from the roots, which were left in the soil for analysis. Plant materials were washed several times with tap water and then with ionized water. Now foliages were dried at 60°C to determine its dry mass. Above ground plant samples were analyzed for total heavy metal content (Allen et al., 1986).
2.3 Translocation factor (f)
The accumulation of metals in the plant parts when taken up from soil was determined as f factor, also known as transfer coefficient (Smith, 1996).
2.4 Statistical analysis:
All the data were analyzed using one way ANOVA test using GPIS software (1.13) (Graphpad, California, USA) and different correlations and regressions has been done for statistical analysis of data by using SPSS version 11.5.
3. Results and discussion:
3.1 General properties of waste and soils:
The dried sludge had a solid content of 93.5%. All the parameters like electrical conductivity, organic carbon and cation exchange capacity were more in industrial sludge in comparison to soils (data not shown), except pH i.e. very low in waste (3.05±0.03). Heavy metals were (Zn, Pb, Cu, and Cd) analyzed and compared to current heavy metal guidelines (USEPA, 1993) were depicted in Table 1. The heavy metal guidelines exceeded for Pb (440±0.36 mg/Kg) only. A high Zn, Cu and Cd concentration in this waste is because more than 70% industries are involved in metal pickling and electroplating. Giri and Bhattacharyya, (1999) also reported that industrial waste from metal finishing industries have a high quantity of macro and micronutrients as well as heavy metals, which is consistent with this study. In our study total and available heavy metals in waste have following trend; Zn>Pb>Cu>Cd
Soil pH is one of the major factors controlling the availability of heavy metal in soils (Brady, 2000). Background pH values of the site1 and site 2 soils were 8.76±0.03 and 8.22±0.04 respectively (data not shown). In control soils total metal concentrations were more in site1 soil than site 2 soil and showed the following trend; Zn>Pb>Cu>Cd. For high concentration of heavy metals in site 1 soil is might be due to its location. It is situated within 5Km radius of Badarpur Thermal Power Station (BTPS) and it is assumed that fly ash and other
pollutant are likely to contaminate this soil. Analytical results indicated that in control soils values of all heavy metals were below the literature levels of a typical soil except Pb (19ppm).
3.2 Heavy metal accumulation and translocation in wheat seedlings:
For both soils, total heavy metal as well as DTPA-extractable heavy metal contents increased linearly as the waste percentage increased in soil (Fig1). Rappaport et al., (1988) also reported that amount of DTPA-extractable Cd, Cu, Ni, Mn, and Zn increased linearly with rate of sludge application. Due to the acidic nature of waste, its amendment increased the pH and decreased the availability of metals as pH is also one of the soil factors that governs the availability of metals in soils (Norwal et al., 1983). Tables 2 showing that availability of all studied DTPA-extractable metals were also depend on soil pH, OC and CEC.
In the present study, correlation analysis (r) was performed in order to investigate the relationship between the DTPA extractable metals in different amendments of the electroplating sludge and the total accumulation of metal in the plant parts (Table 3). It was observed that all the metals in plants were highly correlated with the available metals of the soil at the end of experiment, except, Zn and Pb. Many other researchers were also reported the same and were consistent with this study (Kareka, 2004).
Copper is one of the most important essential elements for plants and animals (Alloway, 1995). Fig 1A showed that guideline limits of 25ppm for Cu doesn’t exceeded in any soil. Effect of industrial amendments and lime were found to follow the similar trend like Zn. Cu concentrations in seedling tissue did not reach phytotoxic levels 20-100ppm (Smith, 1996) in any treatments (Fig 2A). The normal transfer coefficient for Cu in plants is between 0.01-0.1 (Kloke et al., 1984) and it consisted with this study (0.03-0.12). So, Cu in this study is not responsible for any type of phytotoxicity.
In normal situation Zn acts as micronutrient but at higher concentration it becomes phytotoxic (Alloway, 1995). It is evident from the Fig 1B that Zn concentrations were above the guideline limits after the waste amendments in both soils. The value within the bar showed the availability of metal as the percent of total metal in soils. Waste amendments increased the Zn concentration in all treatments (McBride et al., 2003). But the increment was more in site 1 soil than site 2 soil due to background level in soils. Lime amendments significantly (P<0.05) reduced the availability of DTPA extractable metals in all treatments as compared to its counterpart without lime treatment. This can be explained by the increase in soil pH following lime amendment, and the reduced solubility of these metals in soil at high pH. Wong et al., (2001) and Vulkan et al., (2000) also reported that lime amendments reduced the availability of metals in amended soils.
Fig 2B showed the Zn concentration in wheat seedling tissue after the 28 days of growth and the value within the bar showed the transfer coefficient (f). Zn concentration didn’t reach phytotoxic levels (100-400ppm) in any treatment (Fifield and Hannies, 2002). Normal transfer coefficient (f) of Zn for the plants was found 1-10 (Kloke et al., 1984). But in our study it was below this value and it ranges 0.25-0.57. As evident from the Fig 2A transfer coefficient was also more in site 1 soil than site 2 soil and lime have negative effect on transfer coefficient. The higher Zn concentration of site 1 soil transfers more Zn and thus uptake of Zn by seedlings.
Cadmium is an important heavy metal pollutant, the presence of which in agricultural soils and crops is of great concerns. Naturally occurring Cd concentration in soils ranges from 0.001-3ppm (Alloway, 1995). Waste amendment also increased the total and bioavailable Cd in the soils, which supports the study of McBride, (2003). Cd availability to crops is affected by: soil pH, CEC, soil texture, crop species and interrelationship of Cd with other elements in soil (Table 3). In contrary Tsadilas (1995) reported that available Cd is only pH dependent and independent on other physiological properties of soil like CEC, OC etc.
Cd level decreased significantly in the soil types over the 28 days due to accumulation of Cd in plant tissue (Hooda et al., 1997). We observed that most of the Cd is retained by the roots and not transferred to the shoots which show that wheat is an accumulator for metals (Fig 2C). The uptake of Cd into the wheat seedling did not reach the phytotoxic levels of (5-30ppm) except seedlings from 20% waste amended (without lime) site 1 soil. The plant transfer coefficient for Cd in plants tissue is 1-10 (Kloke et al., 1984). So for transfer coefficient all the plants have less value (0.09-0.30) than this limits and transfer coefficient. Lime reduced the uptake of Cd as well as transfer coefficient (f).
Lead, being the zootoxic metal, needs to be monitored in plants parts used by humans and animals (Alloway, 1995). Lead occurs naturally in all plants. In the environment, lead exists almost entirely in inorganic form. Significant reductions were observed in Pb concentration for both soils during this study. Fig 1D showed that the total and bioavailable Pb concentration in both control and waste amended soils. Lead concentration of both soils were below the permissible limits (750-1200μg/g) set by the USEPA, 1993. Up to seedling
stage not much variation obtained in this study, which showed that Pb is not much more mobile in soil system, which was consistent with the findings of Kabata-Pendias and Pendias, (1991); and Valtcho et al., (2004).
Uptake of the Pb in the seedling was low and did not reach the phytotoxic level of 30-300ppm (Alloway, 1995) as depicted in Fig 2D. A normal plant transfer coefficient ($f$) value for Pb in wheat seedling is between 0.01-0.1 (Kloke et al., 1984), therefore the transfer coefficient for both soils was high only in without lime treated waste amendments. Lime reduces the transfer coefficient as well as uptake of the Pb by wheat seedlings. Higher transfer coefficient and uptake of Pb in the seedling tissue occurred in site1 soil due to its background level.
3.3 Growth:
The shoot length, fresh mass and dry weight of wheat seedlings were depicted in table 4. Industrial sludge amendments increased plant growth significantly as compared to the control ones. But 10% waste amendments were found more significant over 20% amendments in both soils. This emphasized the potential short-term beneficial effects of sludge to soils as an organic soil conditioner. From these findings it evident that soil amendment above 10% is not good because it also provides heavy metals to plants and growth become retarded. Plants from site 1 soil exhibited a good growth compare to its counterpart due to its background level.
Lime amendment demonstrated a positive effect on plants growth at each sludge amendment level, which was likely due to increased in pH and reduction in the availability of heavy metals. Most agricultural crops grow well when soil pH is between 6.0-7.0, because nutrients are more available at pH about 6.5 (Soummare et al., 2003). However heavy metal behavior still differs individually and other soil physical properties, like textures might play an important role in heavy metal behavior in soils (Maclean et al., 1987).
4. Conclusion:
In this study it was observed that the industrial sludge from metal finishing industries was highly acidic and had a high amount of heavy metals. Lime had the positive effect and reduced the availability, uptake as well as transfer coefficient in all treatments. The industrial sludge affected the yield of the wheat seedlings positively compared with control plants. However no phytotoxic effects could be proven, because phytotoxic levels were not exceeding in wheat seedling tissues. Although it was a pot culture study and could not draw the actual picture but may be used as baseline data for further analysis. For land application of sludge, needs to possibly take in to consideration the environmental conditions, crop plated, soil type and sludge type. This might lead to the unrestricted industrial sludge on agricultural land, causing a decrease in technological costs for wastewater treatments plants (and subsequent financial profit) to eliminate heavy metals in sludge.
ACKNOWLEDGEMENT
The authors like to express their thanks to Vice-Chancellor and dean (SES) Jawaharlal Nehru University, New Delhi for giving funds, necessary permission and facilities to carry out this research work.
Corresponding To:
Parmanand Sharma,
School of Environmental Science,
Jawaharlal Nehru University,
New Delhi, India-110067
Phone No. +91 9968897332
Fax No. +91 11 26741502
E-mail: email@example.com
REFERENCES
1. Allen S E, Grimshaw H M, Rowland A P, Chemical analysis. In: Methods in Plant Ecology. Oxford, Blackwell Scientific Publication, London, 1986; pp 285–344.
2. Alloway B J, Heavy metals in soils. Blackie academic and Professional, an imprint of Chapman and Hall. London, New York. 1995.
3. Alloway B J, Ayres D C, Chemical Principles of Environmental Pollution. Second Edition, Blackie Academic and Professional, Chapman and Hall, London, 1997; pp. 190-42.
4. Angelidis M, Gibbs R J, Chemistry of metals in anaerobically treated sludges. Water Research 1989; 23: 29-3.
5. Brady N C, The Nature and Properties of Soils. Cornell University and United States agency for international development 10th edition, Prentice Hall of India, Pvt. Ltd. New Delhi. 2000.
6. C.E.C. (Council of the European Communities) Directive of 12 June 1986 on the protection of the environment, and in particular of the soil, when sewage sludge is used in agriculture (86/278/CEE). Official Journal of the European Communities, L181, 1986; pp 6-12.
7. Cameron K C, Di H J, McLaren R G, Is soil an appropriate dumping ground for our
8. Fatoki, O S, Trace zinc and copper concentration in road side vegetation and surface soils: A measurement of local atmospheric pollution in Alice, South Africa. Int. J. Env. Stud., 2000; 57: 501-13.
9. Fifield F W, Haines P J, Environmental chemistry. Second Ed. Blackwell publishers, Oxford. 2000.
10. Giri A K, Bhattacharyya A K, Environmental problems associated with solid waste generated in Wazirpur industrial area. Indian J. Environ. Protection. 1999; 19(1): 915-20.
11. Gupta A K, Sinha S, Phytoextraction capacity of the plants growing on tannery sludge dumping sites. Bio. Technol. 2007; 98: 1788–94.
12. Hooda P S, McNulty D, Alloway B J, Aitken M N, Plant availability of heavy metals in soils previously amended with heavy application of sewage sludge. J. Sc. Food Agric. 1997;73: 446-54.
13. Kabata-Pendias A, Pendias H, Trace Elements in Soils and Plants, CRC Press Inc., Florida, 1991.
14. Karaca A, Effect or organic wastes on the extractability of cadmium, copper, nickel and zinc in soil. Geoderma 2004; 122: 297–3.
15. Kloke A, Sauerbeck DR, Vetter H, Changing metal cycles and human health, (ed. J.O. Nriagu). Spinger-Verlag, Berlin. 1984; pp.113-141.
16. Korentejlar L, A review of the agricultural use of sewage sludge: benefits and potential hazards. Water SA 1991; 17(3): 189-96.
17. Lindsay W L, Norvell W A, Development of a DTPA soil test for zinc, iron, manganese, and copper. J. Soil Sc. Soc. Am. 1978; 42: 421–28.
18. Maclean K S, Robinson A R, Macconnell H M, The effect of sewage sludge on heavy metal content of soil and plant tissue. Comm. Soil Sc. Plant Anal. 1987; 18 (11): 1303-16.
19. McBride M C, Cd uptake by crops estimated from soil total Cd and pH. Soil Sc. 2003; 167: 62-7.
20. McBride, Murry, Nibarger E A, Richards B K, Steenhuis T S, Trace metal accumulation by red clover grown on sewage sludge-amended soils and correlation to Mehlich’3 and caleium chloride-extractable metals. Soil Sc. 2003; 168: 29-8.
21. Mohammad A M, Battikhii A M, Effect of sewage sludge on some soil properties and barley plant Muwaquar area. Agriculture Sc. 1997; 24: 204-16.
22. Narwal R P, Singh B R, Panhwan B R, Plant availability of heavy metals in sludge treated soil: I. effect of sewage sludge and soil pH on the yield and chemical composition of rape. J. Environ Qual. 1983; 12: 358-65
23. Rapaport B D, Martens D C, Reneau Jr, R B, Simpson T W, Metal Availability in Sludge-amended Soils with Elevated Metal Levels. J. Environ. Qual. 1988; 17: 42-7.
24. Schmidt J P, Understanding phytotoxicity threshold for trace elements in land applied sewage sludge. J. Environ. Qual. 1997; 26: 4-10.
25. Smith S R, Agriculture recycling of sewage sludge and environment. CAB international. Wallingford, 1996.
26. Soummare M, Track F M G, Verlas M G, Effects of a municipal solid waste compost and mineral fertilization on plant growth in two tropical agricultural soils of Mali. Bio. Technol. 2003; 86: 15-20.
27. Tsadilas C D, Matsi T, Barbayiannis N, Dimoyiannis D, The influence of sewage sludge application on soil properties and on the distribution and availability of heavy metal fractions. Commun. Soil Sc. Plant Anal. 1995; 26(15-16): 2603-19.
28. USEPA, Clean Water Act Section 503. 1993; Vol. 58, No. 32. Washington. DC.
29. Valtcho D, Warman R, Phytoavailabilty and fractionation of Cu, Mn, and Zn in soil following application of two composts to four crops. Environ Pollut. 2004; 131:187-95.
30. Wong J W C, Lai K M, SU D S, Fang M, Availability of heavy metals for Brassica chinensis grown in an acidic loamy soil amended with domestic and industrial sewage sludge. Water, Air, Soil Pollution. 2001; 128: 339–53.
Figure 1: Showing the metals concentration (ppm) in control and industrial sludge amended soils at beginning (B) and (E) of experiment. Values between vertical bars indicate percentages availability of metals. Values are the mean of three data with SD ($\pm$)
Notation: e.g. S1/S2-10$_2$-0.5%$_3$: 1-Site1/2 soil; 2-percentage of the waste; 3-percentage of lime treatment
Figure 2: showing the total metal concentration (µg/g of plant material) in wheat seedling tissues in all treatments. f factor indicated within vertical bars. Values between vertical bars indicate percentages availability of metals. Values are the mean of three data with SD (±)
Notation; e.g. S1/S2/-10-2- 0.5%; 1-Site ½ soil; 2-percentage of the waste; 3-percentage of lime treatment
| Metals | Total metal content | Extractable metal content |
|--------|---------------------|---------------------------|
| | Industrial Sludge | Soil | Industrial Sludge | Soil |
| | Limits values in sludge’ (mg/Kg) | Site1 | Site2 | Limits values in soil’ (mg/Kg) | Site1 | Site2 |
| Zn | 1290±0.52 | 2800 | 42.25±0.58 | 22.25±0.48 | 130-200 | 44.25±0.27 | 10.63±0.02 | 4.69±0.02 |
| Cu | 410±0.44 | 1500 | 25.75±0.12 | 13.48±0.15 | 60-100 | 31.72±0.56 | 2.40±0.02 | 1.06±0.01 |
| Cd | 30.16±0.21 | 30 | 4.21±0.12 | 2.25±0.11 | 1-3 | 1.15±0.01 | 1.03±0.06 | 0.83±0.02 |
| Pb | 440±0.36 | 300 | 65±0.21 | 50±0.39 | 19 | 56.12±0.11 | 8.34±0.12 | 5.84±0.06 |
C.E.C. (1986)
Table 2: Linear regression coefficient values* ($R^2$) between soil DTPA-extractable metals and pH, organic carbon, CEC in soil at harvesting stage
| Metals | $R^2$ value in Site 1 soil | $R^2$ value in Site 2 soil |
|--------|---------------------------|---------------------------|
| | pH | OC | CEC | pH | OC | CEC |
| Cu | 0.72 | 0.65 | 0.53 | 0.85 | 0.89 | 0.94 |
| Zn | 0.09 | 0.08 | 0.10 | 0.09 | 0.06 | 0.02 |
| Cd | 0.96 | 0.91 | 0.80 | 0.81 | 0.70 | 0.82 |
| Pb | 0.25 | 0.15 | 0.05 | 0.39 | 0.14 | 0.09 |
*at 0.05 level of significance
Table 3 Correlation coefficients (r) between soil available metals at the end of experiment and metal concentrations in seedlings of wheat plants
| Metals | Site 1 Soil | Site 2 Soil |
|--------|-------------|-------------|
| Cu | 0.90 | 0.81 |
| Zn | 0.86 | 0.78 |
| Cd | 0.81 | 0.83 |
| Pb | 0.86 | 0.76 |
*at 0.05 level of significance
Table 4: Average shoot length, fresh and dry mass after 28d of growth
| Treatments | Shoot length (cm) | Fresh mass (g/plant) | Dry mass (g/plant) |
|------------|-------------------|----------------------|--------------------|
| S1-0-0 | 28.3 | 10.2 | 1.02 |
| S1-10-0 | 35.2 | 13.2 | 1.22 |
| S1-10-0.5 | 37.7<sup>c</sup> | 15.2 | 1.45<sup>ac</sup> |
| S1-10-1.0 | 40.1<sup>b</sup> | 17.5 | 1.61<sup>a</sup> |
| S1-20-0 | 39.8<sup>b</sup> | 14.2 | 1.31<sup>b</sup> |
| S1-20-0.5 | 41.3<sup>a</sup> | 18.5<sup>c</sup> | 1.68<sup>ac</sup> |
| S1-20-1.0 | 44.5<sup>ac</sup>| 21.2<sup>bc</sup> | 1.98<sup>ab</sup> |
| S2-0-0 | 25.2 | 9.8 | 1.0 |
| S2-10-0 | 30.3 | 12.3 | 1.14 |
| S2-10-0.5 | 33.5 | 13.2 | 1.28<sup>ac</sup> |
| S2-10-1.0 | 35.8<sup>c</sup> | 15.6 | 1.34<sup>a</sup> |
| S2-20-0 | 36.8<sup>b</sup> | 15.2 | 1.35<sup>b</sup> |
| S2-20-0.5 | 39.1<sup>b</sup> | 16.9<sup>c</sup> | 1.45<sup>ac</sup> |
| S2-20-1.0 | 42.5<sup>ab</sup>| 18.8<sup>bc</sup> | 1.69<sup>ab</sup> |
Notation; e.g. S1/S2,-10,-0.5%; 1-Site1/2 soil; 2-percentage of the waste; 3-percentage of lime treatment. <sup>a</sup>p<0.001; <sup>b</sup>p<0.01; <sup>c</sup>p<0.05 compared to control. |
§ 44-101.01. Purpose.
The purpose of this chapter is to set uniform, minimum standards of licensure for community residence facilities currently regulated under Chapter 34 of Title 22 of the District of Columbia Municipal Regulations and other facilities when they provide services that assist residents with the activities of daily living. This chapter creates a new category of licensure called "assisted living residence".
§ 44-102.01. Definitions.
For purposes of this chapter, the term:
(1) "Activities of Daily Living" or "ADLs" -- means activities including eating, bathing, toileting, grooming, dressing, undressing, mobility, and in place transfers.
(2) "Aging in place" -- means minimizing the circumstances which require a person to move to a different setting when his or her condition changes.
(3) "Assistant Living Administrator" or "ALA" -- means the licensee, or a person designated by the licensee, who oversees the day-to-day operation of the facility, including compliance with all regulations for licensed assisted living residences.
(4) "Assisted Living Residence" or "ALR" -- means an entity, whether public or private, for profit or not for profit, that combines housing, health, and personalized assistance, in accordance to individually developed service plans, for the support of individuals who are unrelated to the owner or operator of the entity. "Assisted Living Residence" or "ALR" does not include a group home for persons with mental retardation as defined in § 44-501(5) or a mental health community residence facility as that term is used in Chapter 38 of Title 22 of the District of Columbia Municipal Regulations.
(5) "Change of ownership" -- means the transfer of ownership by an individual, partnership, or association to another and includes transfers of the legal or beneficial ownership of 10% or more of the stock of a corporation that owns or operates an ALR.
(6) "Chemical restraint" -- means the use of a psychopharmacologic drug for a purpose other than to treat a standard psychiatric diagnosis whose criteria are set forth by the American Psychiatric Association.
(7) "Cognitive impairment" -- means the loss of those mental processes that orchestrate relatively simple ideas, movements, or actions into goal directed behavior including a lack of judgement, planning, organization, self-control, and the persistence needed to manage normal demands of the individual's environment. "Cognitive impairment" refers to a condition that interferes with decision-making skills or effective communication including Alzheimer's disease, multi-infarct dementia, stroke, Parkinson's disease, and other neurological conditions.
(8) "Functional assessment" -- means an assessment of a resident's ability to perform activities of daily living, instrumental activities of daily living, and the degree of assistance required, if any.
(9) "Health-Care Licensure Act" -- means subchapter I of Chapter 5 of this title.
(10) "Health-Care Protection Act" -- means Chapter 10 of this title.
(11) "Healthcare practitioner" -- means a person licensed as a physician or nurse practitioner.
(12) "Healthcare provider" -- means a healthcare practitioner, home health agency, hospice, rehabilitation agency, or health management organization.
(13) "In place transfer" -- means movements that involve changes in position in place. "In place transfer" includes an activity such as moving from a bed to a wheelchair or regular chair, moving from a wheelchair to a toilet, bathtub, shower, or car, and moving from a wheelchair, regular chair, or toilet seat to a standing position.
(14) "Individualized Service Plan" or "ISP" -- means a written plan developed by the provider, in conjunction with the resident and his or her surrogate, if appropriate, which identifies, among other things, services that the licensee will provide or arrange for the resident.
(15) "Instrumental Activities of Daily Living" or "IADL" -- means daily activities such as housekeeping, meal preparation, shopping, money management, and travel outside the ALR.
(16) "Licensee" -- means any person, association, partnership, or corporation to which a license is issued pursuant to this chapter.
(17) "Physical restraint" -- means any manual method or physical or mechanical device, material, or equipment attached to or adjacent to the resident's body, such as mitts or vests, that the individual cannot remove easily and which restricts freedom of movement or normal access to one's own body.
(18) "Physician's statement" -- means the form approved by the Mayor pursuant to § 44-108.02(b).
(19) "Resident" -- means an individual admitted to an ALR pursuant to subchapter VI of this chapter.
(20) "Resident agreement" -- means the admission agreement between the resident, the resident's surrogate, when appropriate, and the assisted living residence.
(21) "Shared responsibility" -- means a process by which the resident, or the resident's surrogate, and the ALR arrive at an acceptable balance between the resident's desire for independence and the facility's legitimate concerns for safety, where there is a disagreement. The purpose of "shared responsibility" is to provide complete information to the resident and
the surrogate so that the parties can arrive at an informed agreement of which services are to be provided and in what manner.
(22) "Shared responsibility agreement" -- means a formal written agreement that outlines the responsibilities and actions of all parties. The agreement is a process for resolving discrepancies between the individual resident's right to independence and the provider's concerns for the safety and well being of the individual and others.
(23) "Surrogate" -- means a person designated by a resident to act on the resident's behalf pursuant to law.
(24) "Trained Medication Employee" or "TME" -- means an individual employed to work in an ALR who has successfully completed the training program developed by the Mayor pursuant to § 44-109.06 and who is certified to administer medication to residents.
DC ST § 44-103.01
§ 44-103.01. Authority to operate an assisted living residence in the District of Columbia.
It shall be unlawful to operate an assisted living residence in the District of Columbia without being licensed and in compliance with the provisions of this chapter.
DC ST § 44-103.02
§ 44-103.02. Initial ALR Licensure.
(a) Applications for licensure shall be made in writing on a form or forms prescribed by the Mayor at least 60 days prior to the date needed.
(b) An applicant for ALR licensure shall pay a licensure fee as determined by the Mayor.
(c) An ALR license issued by the Mayor shall state the effective date and expiration date of the license and maximum number of residents allowed to reside in the ALR.
(d) An application for an ALR license shall include the following information:
(1) Identification of the owner and documentation supporting the fact that the ALR is owned or otherwise under the control of the applicant;
(2) Identification of the ALA and information concerning the ALA's qualifications;
(3) Criminal background check information pursuant to subchapter II of Chapter 5 of this title;
(4) Documentation and explanation of any prior denial, suspension, or revocation of license to provide care to third parties;
(5) Location of the ALR;
(6) Statement of program;
(7) Proof of solvency;
(8) Proof of insurance coverage;
(9) Statement of services to be offered;
(10) Maximum number of residents planned;
(11) Verification that the real property where the ALR is located is owned, leased, or otherwise under the control of the applicant; and
(12) Structure of applicant's organization and names of board members and officers.
(e)(1) The Mayor shall conduct an initial pre-licensure inspection of the premises of the ALR and of its records.
(2) An applicant for licensure shall provide the following information at the time of the pre-licensure inspection:
(A) Certificate of occupancy;
(B) Disaster plan;
(C) Staffing plan;
(D) Resident funds management system;
(E) Medication management system;
(F) Individual Service Plan policy and procedures;
(G) Admission, transfer, and discharge policies;
(H) Resident agreements, both financial and nonfinancial;
(I) Location of the ALR;
(J) Maximum number of residents to be served;
(K) Program statement;
(L) Proof of solvency; and
(M) Other reasonably relevant information required by the Mayor.
(f) Based on information obtained during the pre-licensure inspection required by subsection (e) of this section, the Mayor shall either approve the application unconditionally for 12 months or deny the application.
(g) The Mayor shall re-inspect an ALR within 6 months of the effective date of the initial licensure.
DC ST § 44-103.03
§ 44-103.03. Special ALR licensure for community residence facilities. [Expired]
§ 44-103.03. Special ALR licensure for community residence facilities. [Expired]
DC ST § 44-103.04
§ 44-103.04. Renewal of ALR License.
DC ST § 44-103.04
Chapter 1. Assisted Living Residence Regulation.
Subchapter III. Licensure and Inspection.
§ 44-103.04. Renewal of ALR license.
(a) An ALR license shall be renewed every 12 months.
(b) Applications for renewal of a license shall be made in writing on a form or forms prescribed by the Mayor at least 60 days prior to the expiration of the license. The Mayor shall renew a license after receiving an application containing the information required by § 44-103.02(d) and completion of an inspection of the premises, if the Mayor finds that the application meets the requirements of this chapter.
(c) The Mayor shall issue the renewal license prior to the expiration date of the existing license if the applicant submits an application for renewal within 90 to 60 days prior to the expiration of the license and the Mayor finds that the application meets the requirements of this chapter.
(d) An applicant for renewal shall pay a renewal fee as determined by the Mayor. The fee shall cover the costs involved in processing the renewal applications and conducting inspections of the premises.
(e) Based on information provided to the Mayor and by the on-site inspection, the Mayor shall:
(1) Renew the license for 12 months;
(2) Issue a provisional license for up to 12 months if the ALR is not in full compliance with the regulations but, in the opinion of the Mayor, the noncompliance does not constitute an immediate safety or health hazard and the licensee has submitted to the Mayor an acceptable plan of correction with specific time frames; or
(3) Suspend or revoke the license as specified in § 44-104.01.
DC ST § 44-103.05
§ 44-103.05. Changes in Licensee.
DC ST § 44-103.05
Chapter 1. Assisted Living Residence Regulation.
Subchapter III. Licensure and Inspection.
§ 44-103.05. Changes in licensee.
(a) The following changes occurring within an ALR shall require revision of the license:
(1) Change in the maximum number of residents for which the facility is licensed;
(2) Name change of the ALR;
(3) Change in ownership of the ALR;
(4) Change in location of the ALR; or
(5) Voluntary closure of the ALR.
(b) A request for changes which requires the reissuance of a license shall be made in writing to the Mayor at least 60 days in advance of the effectiveness of the changes. An application fee, as established by the Mayor, shall accompany a request for changes.
(c)(1) The licensee shall notify residents and their surrogates of any proposed changes set forth in its request for changes, in writing, 60 days before the effective date of the proposed changes.
(2) A licensee shall include the following information in its request for changes:
(A) The method for informing residents and their surrogates of its intent to make the requested changes; and
(B) The actions the licensee shall take to assist residents in securing comparable housing, if necessary.
(d)(1) Whenever there is a change of ownership, sale, assignment, or other transfer of an ALR
from the person or organization named on the license to another person or organization, the transferee shall apply for a new license.
(2) A transferee shall apply for a new license at least 60 days before the final transfer.
(3) The licensee named under the current license shall remain responsible for the operation of the ALR until a new license is issued to the transferee.
(4) The Mayor shall issue a new license to the transferee if the transferee meets the requirements for licensure under this chapter. Upon issuance of the new license to the transferee, the transferor shall return its license to the Mayor by certified mail.
DC ST § 44-103.06
§ 44-103.06. Inspections.
(a) In addition to the inspections required by § 44-103.02(e), the Mayor may inspect an ALR at the Mayor's discretion to ensure compliance with this chapter.
(b) The Mayor shall at all times ensure that any ALR licensed pursuant to this chapter is able to continually provide appropriate care to its residents. The Mayor shall notify an ALR if, at any time, the Mayor determines that the ALR is unable to provide appropriate care to any of its residents.
(c) Inspection of an ALR, or prospective ALR, for purposes of initial licensure or compliance after license renewal shall be conducted by the Mayor following the procedures set forth in § 44-505 and the requirements of this chapter.
(d) If, upon inspection, the Mayor determines that an ALR, or prospective ALR, is in violation of one or more of the requirements of this chapter, the Mayor shall give written notice of such violation to the ALR, or prospective ALR, within 15 working days of the inspection and may suggest a remedy for each violation.
(e) The violating ALR, or prospective ALR, shall submit a written response to the Mayor within 15 working days of receipt of the Mayor's notice. The violating ALR, or prospective ALR, may deny the alleged violation, accept the Mayor's suggested remedy, or propose its own remedy.
(f) If the Mayor accepts the ALR's, or prospective ALR's, written response, the Mayor may either issue a license to the ALR, or prospective ALR, if appropriate, or conduct a follow-up inspection to confirm compliance.
(g) If the Mayor and the ALR, or prospective ALR, cannot agree on an acceptable corrective
action, or if the ALR, or prospective ALR, fails to respond in writing within 15 working days of receipt of the Mayor's notice, the Mayor shall determine what action to take, including a penalty in accordance with this chapter and give the ALR, or prospective ALR, notice of his or her determination.
(h) The Mayor may inspect an ALR for the purpose of investigating a complaint.
**DC ST § 44-103.07**
§ 44-103.07. Restrictions.
**DC ST § 44-103.07**
Chapter 1. Assisted Living Residence Regulation.
Subchapter III. Licensure and Inspection.
§ 44-103.07. Restrictions.
(a) An ALR licensed pursuant to this chapter shall not use in its title the words "hospital," "sanitarium," "nursing," "convalescent," "rehabilitive," "sub acute," or "hospice."
(b) Only a licensed ALR may describe, market, and offer itself as such. No other entity, whether licensed or not by the District government, shall describe, market, or offer itself as an Assisted Living Residence without first obtaining an ALR license. Violation of this requirement shall constitute operation of an ALR without a license and shall be subject to penalties in accordance with this chapter.
**DC ST § 44-103.08**
§ 44-103.08. Appeals.
**DC ST § 44-103.08**
Chapter 1. Assisted Living Residence Regulation.
Subchapter III. Licensure and Inspection.
§ 44-103.08. Appeals.
Appeals under this subchapter may be taken pursuant to subchapter XII of this chapter.
§ 44-104.01. Sanctions.
(a) The sanctions set forth in § 44-509.
(b) On determining that a licensee has violated this chapter, the Mayor, in addition to the sanctions required by subsection (a) of this section, may:
(1) Restrict the number of residents the licensee may admit;
(2) Require the licensee to reduce the number of residents in its care;
(3) Require the licensee and any of its staff to receive remedial instruction or training in a specific area;
(4) Require the licensee to use the services of a management firm approved by the Mayor;
(5) Notify or require the licensee to notify a resident who is affected by the violation and his or her surrogate;
(6) Increase the frequency of monitoring visits during a specified period of time;
(7) Enter into an agreement with the licensee establishing certain conditions for continued operation, including time limits for compliance; and
(8) Petition a court of competent jurisdiction to appoint a receiver as provided in subchapter II of Chapter 10 of this title.
(c) If the Mayor determines that the licensee has violated a condition or requirement of an imposed sanction, the Mayor may suspend or revoke the license.
(d) Appeals under this section may be taken pursuant to subchapter XII of this chapter.
§ 44-104.02. Civil Penalties.
DC ST § 44-104.02
Chapter 1. Assisted Living Residence Regulation.
Subchapter IV. Sanctions and Penalties.
§ 44-104.02. Civil Penalties.
(a) The Mayor may impose one or more of the civil penalties authorized under § 44-509 against persons who:
(1) Maintain or operate an unlicensed ALR; or
(2) Otherwise violate provisions of this chapter.
(b) Notwithstanding any other provision of law, penalties authorized under § 44-509 shall not be imposed by the Mayor unless a violation, cited during an inspection:
(1) Is within the control of the ALR; and
(2) Poses an immediate or serious and continuing danger to the health, safety, welfare, or rights of residents.
(c) If during a follow-up inspection the Mayor determines that violations within the control of the facility which were cited in an immediately prior inspection have not been corrected or have recurred, the Mayor may impose penalties authorized under § 44-509.
(d) Appeals under this section may be taken pursuant to subchapter XII of this chapter.
DC ST § 44-104.03
§ 44-104.03. Criminal Penalties.
DC ST § 44-104.03
Chapter 1. Assisted Living Residence Regulation.
Subchapter IV. Sanctions and Penalties.
§ 44-104.03. Criminal penalties.
The criminal penalties set forth in § 44-509 shall apply to an ALR.
DC ST § 44-104.04
§ 44-104.04. Emergency Suspension.
DC ST § 44-104.04
Chapter 1. Assisted Living Residence Regulation.
Subchapter IV. Sanctions and Penalties.
§ 44-104.04. Emergency suspension.
(a) The Mayor may immediately suspend, as an emergency action, a license on finding that the licensee's premises are unsafe for human habitation or pose an immediate threat to the health and safety of its residents.
(b) The Mayor shall deliver a written notice to the licensee informing it of the emergency suspension, giving the reasons for the suspension, providing the provisions of law with which the licensee has failed to comply that form the basis for the emergency suspension, and notifying the ALR of its right to request a hearing and to be represented by counsel.
(c) The filing of a hearing request shall not stay the emergency suspension. If the licensee is dissatisfied with the emergency suspension, it may appeal the suspension as provided in § 44-503.
(d) When a license is suspended pursuant to this section, the licensee shall immediately return the license to the Mayor, notify the residents and surrogates of the suspension, and make every effort to assist them in making other assisted living arrangements.
DC ST § 44-105.01
§ 44-105.01. Standard of Care.
DC ST § 44-105.01
Chapter 1. Assisted Living Residence Regulation.
Subchapter V. Resident's Rights and Quality of Life.
§ 44-105.01. Standard of care.
(a) An ALR must care for its residents in a manner and in an environment that promotes maintenance and enhancement of the residents' quality of life and independence.
(b) In order to promote resident independence and aging in place in a residential setting, at a minimum, an ALR shall offer or coordinate for payment 24 hour supervision, assistance with scheduled and unscheduled activities of daily living, and instrumental activities of daily living as needed, as well as provision or coordination of recreational and social activities and health services in a way that promotes optimum dignity and independence for the residents.
DC ST § 44-105.02
§ 44-105.02. Self-Determination, Choice, Independence, Participation, and Privacy.
DC ST § 44-105.02
Chapter 1. Assisted Living Residence Regulation.
Subchapter V. Resident's Rights and Quality of Life.
§ 44-105.02. Self-determination, choice, independence, participation, and privacy.
(a) A resident shall have the right to be treated at all times as follows:
(1) Courteously;
(2) Respectfully;
(3) With full recognition of personal dignity and individuality; and
(4) With assurance of privacy and the opportunity to act autonomously and share in the responsibility for decisions.
(b) A resident of an ALR shall have the right to live in an environment that:
(1) Maintains and enhances the resident's dignity, independence, and respect in full recognition of his or her individuality and physical and mental capabilities;
(2) Is creatively designed to counter loneliness, depression, dependence, boredom, and designed to manage difficult behavior;
(3) Provides opportunities for socialization, social interaction, leisure activities, and spiritual and religious activities consistent with the preferences and background of the resident; and
(4) Facilitates participation by arranging for transportation and assisting with communication and social skills and other services.
DC ST § 44-105.03
§ 44-105.03. Dignity.
A resident shall have the right to the following:
(1) A safe, clean, comfortable, stimulating, and homelike environment allowing the resident to use personal belongings to the greatest extent possible;
(2) Control time, space, and lifestyle;
(3) Free access to visitors of his or her choice;
(4) To receive and send correspondence without any restrictions;
(5) To maintain personal possessions to the extent the health, safety, and well being of others is not disturbed;
(6) To remain in his or her living unit unless a change corresponds to the uncoerced preference of the resident or conforms to the obligations set forth in the resident's contract respecting discharge and is related to the resident's preference or to transfer conditions stipulated in his or her contract with the ALR;
(7) To approve his or her roommate whenever possible, if the resident is living in a semi-private unit;
(8) To attend or not attend religious services of his or her choice;
(9) To choose activities and schedules consistent with his or her interests, and physical, mental, and psychosocial well-being;
(10) To interact with members of the community inside and outside the facility and make choices about aspects of his or her life in the facility that are significant to the resident;
(11) To be free from mental, verbal, emotional, sexual and physical abuse, neglect, involuntary seclusion, and exploitation; and
(12) To participate in the development, implementation, and review of plans designed to provide services to residents, including the Individualized Service Plan.
DC ST § 44-105.04
§ 44-105.04. Accommodation of Needs.
DC ST § 44-105.04
Chapter 1. Assisted Living Residence Regulation.
Subchapter V. Resident's Rights and Quality of Life.
§ 44-105.04. Accommodation of needs.
A resident shall have the right to the following:
(1) To receive adequate and appropriate services and treatment with reasonable accommodation of individual needs and preferences consistent with their health and physical and mental capabilities and the health or safety of other residents;
(2) To have access to appropriate health and social services, including social work, home health, nursing, rehabilitative, hospice, medical, dental, dietary, counseling, and psychiatric services in order to attain or maintain the highest practicable physical, mental and psychosocial well-being;
(3) To remain in the current setting, forgoing a recommended transfer to obtain additional services as contracted for by the resident or secure additional services in a manner acceptable to the ALR;
(4) To engage in a shared responsibility agreement with the ALR which is acceptable to the resident and the ALR and does not violate any applicable law;
(5) To refuse to participate in any service once the potential consequences of such participation have been explained and a shared responsibility agreement has been reached, if necessary, between the resident, the surrogate, and the ALR;
(6) To be free of physical restraints at all times; and
(7) To be free of chemical restraints.
DC ST § 44-105.05
§ 44-105.05. Representation and Resolution of grievances and complaints.
§ 44-105.05. Representation and resolution of grievances and complaints.
(a) A resident shall have the right to the following:
(1) To designate a person as his or her surrogate or to have their guardian, or advance directives or surrogate health decision maker, act for them if for any reason the resident cannot act for him or herself;
(2) To uncoerced consent;
(3) To present grievances and complaints without fear of threat of retaliation and have them acknowledged and acted upon promptly with due respect to the provisions of this chapter;
(4) To have access to an internal grievance and complaint procedure for any denial of services or rights provided for under this chapter and to an external review process by an independent person or entity;
(5) To address grievances and complaints to representatives of the Office of the Long-Term Care Ombudsman of the District of Columbia pursuant to Chapter 7 of Title 7 or other representative;
(6) To appoint a specially designated person or attorney to represent the resident in any grievance or complaint, procedure, or appeal process who shall have access to all necessary and relevant books and records of the ALR; and
(7) To organize and participate in and hold meetings of resident groups in the ALR and invite staff or visitors to the meetings and have a designated staff person to assist and respond to written requests resulting from the meetings.
(b) The ALR shall maintain complete written records of the filing and disposition of all grievances, complaints, and appeals.
(c) Appeals under this section may be taken pursuant to subchapter XII of this chapter.
§ 44-105.06. Privacy and confidentiality.
DC ST § 44-105.06
Chapter 1. Assisted Living Residence Regulation.
Subchapter V. Resident's Rights and Quality of Life.
§ 44-105.06. Privacy and confidentiality.
(a) A resident shall have the right to the following:
(1) To access their ALA and healthcare records on demand;
(2) To have their records kept confidential and released only in accordance with their informed uncoerced consent in accordance with District and federal law;
(3) To have their records maintained during their residency;
(4) To have their records maintained for up to 3 years after discharge or death; and
(5) To have any case discussion, consultation, examination, or treatment of the resident be kept confidential.
(b) If, for any reason, a resident cannot act for him or herself, their consent shall be given on their behalf by their designated surrogate which consent shall also be uncoerced and informed.
DC ST § 44-105.07
§ 44-105.07. Full Disclosure.
DC ST § 44-105.07
Chapter 1. Assisted Living Residence Regulation.
Subchapter V. Resident's Rights and Quality of Life.
§ 44-105.07. Full disclosure.
A resident shall have the right to full disclosure of contract terms and billing practices that are fair and reasonable.
DC ST § 44-105.08
§ 44-105.08. Notice of Resident's Rights.
DC ST § 44-105.08
Chapter 1. Assisted Living Residence Regulation.
Subchapter V. Resident's Rights and Quality of Life.
§ 44-105.08. Notice of resident's rights.
An ALR shall place a copy of a document delineating the resident's rights, as set forth in this chapter, in a conspicuous location, plainly visible and easily read by residents, staff, and visitors and provide a copy to each resident and resident's surrogate upon admission and at
the time of any change to the resident's status, level of care, or services available to the resident.
DC ST § 44-105.09
§ 44-105.09. Abuse, Neglect, and Exploitation.
DC ST § 44-105.09
Chapter 1. Assisted Living Residence Regulation.
Subchapter V. Resident's Rights and Quality of Life.
§ 44-105.09. Abuse, neglect, and exploitation.
(a) An ALR shall develop and implement policies and procedures prohibiting abuse, neglect, and exploitation of residents.
(b)(1) An ALR, employee of an ALR, or other person who believes that a resident has been subjected to abuse, neglect, or exploitation shall report the alleged abuse, neglect, or exploitation immediately to the assisted living administrator who shall take appropriate action to protect the resident. The ALR shall report any allegation of abuse, neglect, or exploitation brought to its attention to the Mayor and the Adult Protective Services Program, administered by the Family Services Administration of the Department of Human Development.
(2) An ALR or employee of an ALR may be subject to a penalty imposed by the Mayor for failure to report an alleged incident of abuse, neglect, or exploitation pursuant to Chapter 19 of Title 7.
(3) An ALR shall thoroughly investigate any allegation of abuse, neglect, or exploitation and shall take appropriate action to prevent further incidents. The ALR shall report the results of its investigation and actions taken, if any, to the Mayor.
(c) An ALR shall post signs that set forth the reporting requirement of this section conspicuously in the employee and public areas of the ALR.
DC ST § 44-106.01
§ 44-106.01. Admissions.
DC ST § 44-106.01
Chapter 1. Assisted Living Residence Regulation.
Subchapter VI. Admissions; Residential Agreements; Quality of Care; Discharge; Transfer.
§ 44-106.01. Admissions.
(a) An ALR shall accept as residents only individuals for whom the ALR can provide appropriate services unless the ALR arranges for third party services or the resident does so with the
agreement of the ALR.
(b) Prior to admission of a resident, the ALA or designee shall determine that the resident is appropriate for admission to the ALR and that the resident's needs can be met in addition to the needs of the other residents.
(c) An ALR may only admit individuals who are at least 18 years of age.
(d) No individual may be admitted who at the time of initial admission, and as established by the initial assessment:
(1) Is dangerous to him or herself or others or exhibits behavior that significantly and negatively impacts the lives of others, where the ALR would be unable to eliminate such danger or behavior through the use of appropriate treatment modalities; or
(2) Is at high risk for health or safety complications which cannot be adequately managed by the ALR and requires more than 35 hours per week of skilled nursing and home health aide services combined, provided on less than a daily basis, according to section 2113.1 of HCFA Pub. 75 and 42 CFR, sections 409.32, 409.33, and 409.44.
(e) An ALR shall not admit individuals who require the following:
(1) More than intermittent skilled nursing care;
(2) Treatment of stage 3 or 4 skin ulcers;
(3) Ventilator services; or
(4) Treatment for an active, infectious, and reportable disease or a disease or condition that requires more than contact isolation.
(f) The ALR shall maintain records of all denials of admission.
(g) Nothing in this section shall automatically exclude persons with primary or secondary mental health issues from admission.
DC ST § 44-106.02
§ 44-106.02. Resident Agreements.
DC ST § 44-106.02
Chapter 1. Assisted Living Residence Regulation.
Subchapter VI. Admissions; Residential Agreements; Quality of Care; Discharge; Transfer.
§ 44-106.02. Resident agreements.
(a) A written contract must be provided to the resident prior to admission and signed by the resident or surrogate, if necessary, and a representative of the ALR. The nonfinancial portions of the contract shall include the following:
(1) The ALR's organizational affiliations (including parent or subsidiary organizations, religious or charitable affiliation, and management company);
(2) The specific nature of any special care that it holds itself out to provide, such as specialty in Alzheimer's disease or Parkinson's disease;
(3) An identification of services to be included and excluded, part of which is the ISP;
(4) A list of resident rights including grievance procedures;
(5) Unit assignment and procedures if changes occur;
(6) Admission and discharge policies which include clear and specific criteria for admission, transfer, and discharge;
(7) A description of responsibility for provision or coordination of healthcare, if any;
(8) An arrangement for notification in case of the resident's death; and
(9) A disposition of the resident's property upon discharge, transfer, or death of the resident.
DC ST § 44-106.03
§ 44-106.03. Financial Agreements.
DC ST § 44-106.03
Chapter 1. Assisted Living Residence Regulation.
Subchapter VI. Admissions; Residential Agreements; Quality of Care; Discharge; Transfer.
§ 44-106.03. Financial agreements.
(a) The written resident agreement required by § 44-106.02 shall include financial provisions which indicate the following:
(1) The obligations of the ALR, the resident, or the resident's surrogate as to performance of the following:
(A) The handling of the finances of the resident;
(B) The purchasing or renting of essential or desired equipment and supplies;
(C) The coordinating and contracting for services not covered by the resident agreement; and
(D) The purchasing of medications and durable medical equipment;
(2) Separate and accurate records of all funds and personal property deposited with or managed by the ALR for the benefit of a resident which include a receipt stating the date, amount, and purpose of any transaction and the current balance;
(3) Rate structure and payment provisions covering all rates to be charged to the resident, including the following:
(A) Service packages;
(B) Fee for service rates; and
(C) Any other nonservice related charges;
(4) Payment arrangements and fees, if known, for third-party services not covered by the resident agreement, but arranged for by either the resident, the resident's surrogate, or the ALR;
(5) Identification of the persons responsible for payment of all fees and charges and a clear indication of whether the person's responsibility is or is not limited to the extent of the resident's funds;
(6) A provision which provides at least 45 days notice of any rate increase except if necessitated by a change in the resident's medical condition;
(7) The procedures the ALR will follow in the event the resident or surrogate can no longer pay for services provided for in the resident agreement or for additional services or care needed by the resident; and
(8) The terms governing the refund of any pre-paid fees or charges in the event of a resident's discharge from the ALR or termination of the resident agreement.
(b) The ALR shall establish billing, payment, and credit practices that are fair and reasonable.
DC ST § 44-106.04
§ 44-106.04. Individualized Service Plans.
DC ST § 44-106.04
Chapter 1. Assisted Living Residence Regulation.
Subchapter VI. Admissions; Residential Agreements; Quality of Care; Discharge; Transfer.
§ 44-106.04. Individualized Service Plans.
(a)(1) An ISP shall be developed for each resident prior to admission.
(2) An ISP shall be developed following the completion of the "post move-in" assessment.
(3) The ISP shall be written by a healthcare practitioner using information from the assessment.
(4) The ISP shall be developed with the resident, or surrogate, as a full partner.
(5) The ISP shall be signed by the resident, or surrogate, and a representative of the ALR.
(6) The ISP shall include a shared responsibility agreement when necessary.
(7) The ISP shall be based on such factors as:
(A) The medical, rehabilitation, and psychosocial assessment of the resident;
(B) The functional assessment of the resident; and
(C) The reasonable accommodation of resident and, if necessary, surrogate preferences.
(b) The ISP shall include the services to be provided, when and how often the services will be provided, and how and by whom all services will be provided and accessed.
(c) During the ISP development process, the ALR shall confer with the prospective resident and, if necessary, the surrogate to arrive at a mutual agreement as to the responsibilities of each party in accessing care and achieving related outcomes.
(d) The ISP shall be reviewed 30 days after admission and at least every 6 months thereafter. The ISP shall be updated more frequently if there is a significant change in the resident's condition. The resident and, if necessary, the surrogate shall be invited to participate in each reassessment. The review shall be conducted by an interdisciplinary team that includes the resident's healthcare practitioner, the resident, the resident's surrogate, if necessary, and the ALR.
(e) An ALR shall facilitate aging in place to the best of its ability with the understanding that there may be a point reached where adequate and appropriate services can not be marshalled to support the resident safely, making transfer to another setting necessary.
DC ST § 44-106.05
§ 44-106.05. Shared Responsibility Agreements.
DC ST § 44-106.05
Chapter 1. Assisted Living Residence Regulation.
Subchapter VI. Admissions; Residential Agreements; Quality of Care; Discharge; Transfer.
§ 44-106.05. Shared responsibility agreements.
(a) Whenever disagreements arise as to lifestyle, personal behavior, safety, and service plans the ALR staff, resident or surrogate, and other relevant service providers shall attempt to develop a shared responsibility agreement.
(b) A shared responsibility agreements represents a tool for ALRs to recognize an individual resident's right to autonomy by respecting his or her right to make individual decisions regarding lifestyle, personal behavior, and ISPs. In some cases, a resident's decision may involve increased risk of personal harm and therefore potentially increase the risk of liability by the ALR absent an agreement between the resident and ALR concerning such decisions or actions. In such instances the ALR shall:
(1) Explain to the resident, or surrogate, why the decision or action may pose risks and suggest alternatives to the resident; and
(2) Discuss with the resident, or surrogate, how the ALR might mitigate potential risks.
(c) If, after consultation with the ALR as required by subsection (b) of this section, a resident decides to pursue a course of action, such as refusal of services, that may involve increased risk of personal harm and conflict with the ALR's usual responsibilities, the ALR shall:
(1) Describe to the resident the action or range of actions subject to negotiation; and
(2) Negotiate a shared responsibility agreement, with the resident as a full partner, acceptable to the resident and the ALR that meets all reasonable requirements implicated. The shared responsibility agreement shall be signed by the resident or surrogate and the ALR.
DC ST § 44-106.06
§ 44-106.06. Resident Records.
DC ST § 44-106.06
Chapter 1. Assisted Living Residence Regulation.
Subchapter VI. Admissions; Residential Agreements; Quality of Care; Discharge; Transfer.
§ 44-106.06. Resident records.
A record shall be maintained for every resident and include the following:
(1) The resident agreement required by this subchapter, including the "Resident's Rights" statement and any additional agreements;
(2) The functional assessment of ADLs;
(3) A physician's statement, including medical orders and rehabilitation plans;
(4) The ISP and any revisions thereto;
(5) All shared responsibility agreements; and
(6) Any note and comments added to the record by the ALR.
DC ST § 44-106.07
§ 44-106.07. Services to be Provided.
DC ST § 44-106.07
Chapter 1. Assisted Living Residence Regulation.
Subchapter VI. Admissions; Residential Agreements; Quality of Care; Discharge; Transfer.
§ 44-106.07. Services to be provided.
(a) An ALR shall provide or ensure the provision of the following:
(1) Twenty-four hour supervision and oversight to ensure the well-being and safety of its residents;
(2) Three nutritious and attractive meals and additional snacks, modified to individual dietary needs as necessary, on a daily basis;
(3) A variety of fresh and seasonal foods, adapted to the food habits, preferences, and physical abilities of the residents;
(4) At minimum, some assistance with ADLs and IADLs to meet the scheduled and unscheduled service needs of the residents; and
(5) Laundry and housekeeping service not provided by the resident or surrogate.
(b) An ALR shall facilitate access for a resident to appropriate health and social services, including social work, home health agencies, nursing, rehabilitative, hospice, medical, dental, dietary, counseling, and psychiatric services.
(c) An ALR shall provide or coordinate scheduled transportation to community-based services.
DC ST § 44-106.08
§ 44-106.08. Discharge and Transfer.
DC ST § 44-106.08
Chapter 1. Assisted Living Residence Regulation.
Subchapter VI. Admissions; Residential Agreements; Quality of Care; Discharge; Transfer.
§ 44-106.08. Discharge and transfer.
(a) When a resident wishes to be discharged from an ALR, the resident or surrogate shall give 30 days written notice to the ALR.
(b) When a sudden, unexpected, and life-threatening medical emergency arises necessitating the immediate transfer of the resident to an acute care facility, the ALR shall immediately notify the surrogate and the resident's healthcare provider of the transfer. The ALR shall provide the surrogate and healthcare provider with information concerning cause of the transfer and the name and location of the acute care facility.
(c) After a resident is transferred pursuant to subsection (b) of this section, his or her return to the ALR shall be determined by the renegotiation of the ISP with the resident or surrogate, the resident's healthcare provider, and the ALR. If, in renegotiating the ISP, the interested parties determine that the resident can no longer safely reside at the ALR, discharge planning shall take place in consultation with the resident or surrogate, the resident's healthcare provider, and the ALR. Under these circumstances the ALR shall waive the 30 day notice requirement.
(d) Before a resident may be discharged on an involuntary basis, the ALR shall provide 30
days written notice to the resident and surrogate of the planned discharge, and make arrangements for the discharge in consultation with the resident, the surrogate, and the healthcare provider. Any involuntary discharge shall conform to the notice and process established in subchapter III of Chapter 10 of this title.
(e) Although an ALR shall make every effort to avoid discharge, grounds for involuntary discharge may include the following:
(1) Failure to pay all fees and costs as specified in the contract; and
(2) Inability of the ALR to meet the care needs of the resident as provided in the ISP.
(f) An involuntary discharge shall be canceled upon the occurrence of one of the following:
(1) The payment of all monies owed at any time prior to discharge; or
(2) The negotiation of a new ISP.
DC ST § 44-107.01
§ 44-107.01. Staffing Standards.
DC ST § 44-107.01
Chapter 1. Assisted Living Residence Regulation.
Subchapter VII. Staffing and Training.
§ 44-107.01. Staffing standards.
(a) An ALR shall be supervised by an ALA who shall be responsible for all personnel and services within the ALR.
(b) The ALA shall ensure that each resident has access to appropriate medical, rehabilitation, and psychosocial services as established in the ISP and that there is appropriate oversight, monitoring, and coordination of all components of the ISP, including necessary transportation and the delivery of needed supplies.
(c)(1) An ALA shall be at least 21 years of age.
(2) An ALA shall possess at least a high school diploma or general equivalency diploma ("G.E.D.") or have served as an operator or administrator of a licensed community residence facility ("CRF") in the District of Columbia for at least one of the past 3 years in which the CRF has met minimum legal standards. An ALA employed on or after June 24, 2000 shall have at least a high school diploma or G.E.D. and have served as a direct care provider or administrator for at least one of the past three years.
(3) An ALA shall possess satisfactory knowledge of the following:
(A) The philosophy of assisted living;
(B) The health and psychosocial needs of residents;
(C) The resident assessment process;
(D) The development and use of ISPs;
(E) Medication administration, including cuing, coaching, and monitoring residents who self-administer medications, with or without assistance;
(F) The provision of assistance with activities of daily living and personal hygiene;
(G) Residents' rights;
(H) Fire and life safety codes;
(I) Infection control, including standard precautions to prevent infection;
(J) Food safety and sanitation;
(K) First aid and cardiopulmonary resuscitation (CPR);
(L) Emergency disaster plans;
(M) Human resource management, including staff employment, orientation and training, employee rights and protection against discrimination, harassment, and wrongful discharge; and
(N) Financial management.
(d) An ALA shall:
(1) Employ staff and develop a staffing plan in accordance with this chapter and based upon the following criteria to assure the safety and proper care of residents in the ALR:
(A) The health, mental condition, and psychosocial needs of the residents;
(B) The fulfillment of the 24-hours-a-day scheduled and unscheduled needs of the residents;
(C) The size and layout of the ALR;
(D) The capabilities and training of the employees; and
(E) Compliance with all of the minimum standards in this chapter;
(2) Assure that sufficient staff who know how to implement the ALR's evacuation plan and emergency management plan are on the premises at all times to implement emergency procedures;
(3) Assure that each person employed by the ALR maintains personal cleanliness and hygiene;
(4) Develop written job descriptions for staff who are responsible for providing personal services to residents and provide a copy of the job description to the employee;
(5) Assign duties to each staff member consistent with his or her level of education, preparation, and experience;
(6) Assure that there is at least one staff member within the ALR at all times who is certified in first-aid and CPR;
(7) Assure that all members of the staff are mentally and physically capable of performing their assigned duties;
(8) Assure that each employee has a background check pursuant to federal and District law executed at the time of initial employment;
(9) Assure that members of the staff appear to be free from apparent signs and symptoms of communicable disease, as documented by a written statement from a healthcare practitioner;
(10) Remove from duty any staff member who is found to have, or is suspected of having, a communicable disease or is mentally or physically incapable of performing her or his duties until the ALA determines that such impairment no longer exists;
(11) Maintain personnel records for each employee that include documentation of criminal background checks, statements of health status, and documentation of the employee's communicable disease status;
(12) Assure that, during periods of temporary absence of the ALA when residents are on the premises, a staff member who is at least 18 years of age and meets the staffing standards of the ALA required by this section shall assume responsibilities of the ALA; and
(13) Complete the training required by § 44-107.02 and 12 additional hours of training, annually, conducted by a nationally recognized organization that possesses experience in training staff in dementia care, such as the Alzheimer's Disease and Related Disorders Association, on managing residents who are living with cognitive impairments.
(e) Newly hired staff shall have 30 days to document their communicable disease status. For the purposes of this subsection, "newly hired staff" means any individual who is hired by an ALR regardless of the individual's previous work experience. An employee who is transferring from one ALR to another ALR that is under the same management or ownership, without break in service, shall not be considered newly hired staff.
(f) Employees shall be required on an annual basis to document freedom from tuberculosis in a communicable form.
(g) The staff shall:
(1) Be at least 21 years of age;
(2) Possess current and appropriate licensure and certifications as required by law;
(3) Possess sufficient skills, education, training, and experience to meet the needs of the residents;
(4) Subject to their assigned duties and responsibilities, possess a satisfactory knowledge of the resident assessment process, use of ISPs, resident health and psychosocial needs, and resident rights;
(5) Complete initial and ongoing training pertaining to the philosophy of care in ALRs and meeting the personal care needs of residents; and
(6) Not work unsupervised without satisfactory completion of the training required by § 44-107.02.
(h) Staff who have not completed the training required by § 44-107.02 must work at all times under the supervision of a staff member who has satisfactorily completed this training.
(i) For the purposes of subsection (g)(5) of this section "ongoing training" means a regularly scheduled program of staff training designed by the ALR to assure that all staff who have
direct resident contact possess the skills necessary to provide high quality services in a manner appropriate to the philosophy of assisted living and includes staff training in how to monitor changes in a resident's condition, including physical and cognitive assessments.
DC ST § 44-107.02
§ 44-107.02. Staff Training.
DC ST § 44-107.02
Chapter 1. Assisted Living Residence Regulation.
Subchapter VII. Staffing and Training.
§ 44-107.02. Staff training.
(a) All staff shall be properly trained and be able to demonstrate proficiency in the skills required to effectively meet the requirements of this chapter. Prior to the date of hire, an employee must meet or possess one of the following criteria:
(1) Be certified as a nurse's aide;
(2) Be certified as a home care aide as defined in the Medicare criteria in OBRA 1987;
(3) Be properly trained by virtue of holding current licenses in a healthcare related field;
(4) Be properly trained under a plan approved by the Mayor which covers the following topics, for a minimum of 40 hours:
(A) Delivering care for the bed-bound resident, including bathing, feeding, shampooing, dressing, positioning, and toileting;
(B) Use of the first aid kit and knowledge of its location;
(C) Confidential treatment of personal information;
(D) Procedures for detecting and reporting suspected abuse, neglect, or exploitation of residents;
(E) Managing difficult aggressive behavior;
(F) Advanced body mechanics;
(G) Communicating with adults, including those with communication deficits such as aphasia, hearing loss, loss of eyesight, and cognitive impairments;
(H) Recognizing the signs and symptoms of dementia;
(I) Caring for the cognitively impaired with such behaviors as wandering, repetitive questions, and confusion;
(J) Techniques for assisting residents in overcoming transfer trauma;
(K) Awareness of resident's "change in condition", including depression and ability to report changes to the appropriate staff according to the protocol of the ALR;
(L) Basic competence in housekeeping, laundry, food handling, and meal preparation; and
(M) Any specialized training for special needs not covered through the basic training.
(b) Within 7 days of employment, an ALR shall train a new member of its staff as to the following:
(1) Their specific duties and assignments;
(2) The purpose and philosophy of the ALR;
(3) The services provided;
(4) The daily routines;
(5) The rights of residents;
(6) The emergency procedures and disaster drills and techniques of complying, including evacuating residents when applicable;
(7) Elementary body mechanics, including proper lifting and in place transfer;
(8) Choking precautions and airway obstruction, including the Heimlich Maneuver; and
(9) Infection control.
(c) After the first year of employment, and at least annually thereafter, a staff member shall complete a minimum total of 12 hours of in-service training in the following:
(1) Emergency procedures and disaster drills;
(2) Rights of residents;
(3) Four hours covering cognitive impairments in an in-service training approved by a nationally recognized and creditable expert such as the Alzheimer's Disease and Related Disorder Association; and
(d) On an annual basis, the ALA shall complete 12 additional hours of training on cognitive impairments approved by a nationally recognized organization with expertise in dementia such as the Alzheimer's Disease and Related Disorders Association.
DC ST § 44-108.01
§ 44-108.01. General.
DC ST § 44-108.01
Chapter 1. Assisted Living Residence Regulation.
Subchapter VIII. Resident Assessment.
§ 44-108.01. General.
A resident assessment required by this subchapter shall form the basis for the development of the resident’s service plan, and shall be completed for all routine admissions to the ALR.
§ 44-108.02. Medical, Rehabilitation, and Psychosocial Assessment.
(a) A medical, rehabilitation, and psychosocial assessment of the resident shall be completed within 30 days prior to admission.
(b) The ALR shall maintain resident information obtained from a standardized physician's statement approved by the Mayor. The information shall include a description of the applicant's current physical condition and medical status relevant to defining care needs, and the applicant's psychological and cognitive status, if so indicated during the medical assessment.
(c) The assessment shall be based on an examination by the prospective resident's primary, licensed healthcare practitioner within 30 days prior to admission. The information obtained from the examination shall include at least the following:
(1) The individual's medical history with a recent evaluation;
(2) Any significant medical conditions affecting function, including the individual's ability for self-care, cognition, behavior, and psychosocial activities;
(3) Presence of allergies;
(4) Confirmation that the applicant is free from communicable TB and from other active, infectious, and reportable communicable diseases;
(5) Current medication profile and projected and other needed medications, treatments and service; review of nonprescription drugs and review of possible adverse interactions;
(6) Current dietary needs and restrictions;
(7) Medically necessary limitations or precautions; and
(8) Monitoring or tests that may need to be performed or followed up after admission.
§ 44-108.03. Functional assessment.
Within 30 days prior to admission, the facility shall collect, on a standardized form approved by the Mayor, the following information regarding each applicant:
(1) Level of functioning in activities of daily living including bathing, dressing, grooming, eating, toileting, and mobility;
(2) Level of support and intervention, including any special equipment and supplies, required to compensate for the individual's deficits in activities of daily living;
(3) Current physical or psychological symptoms of the individual requiring monitoring, support, or other intervention by the ALR;
(4) Capacity of the individual for making personal and healthcare related decisions;
(5) Presence of disruptive behavior or behavior which presents a risk to the physical or emotional health and safety of self or others;
(6) Social factors, including:
(A) Significant problems with family circumstances and personal relationships;
(B) Spiritual status and needs; and
(C) Ability to participate in structured and group activities and the resident's current involvement in such activities.
§ 44-108.04. Short-term residential care.
For individuals who will stay in the ALR no longer than 30 days, only the following information is required for admission to the ALR:
(1) An analysis of the individual's current physical condition, medical status, and functional assessment as set out in this subchapter; and
(2) A resident agreement in accordance with this chapter.
§ 44-108.05. Emergency placement.
A resident admitted as an emergency placement, not to exceed 14 days, must comply with admission, physical examination, and assessment requirements of this chapter.
§ 44-109.01. Responsibilities of the ALR Personnel in Medication Management.
An ALA shall ensure that an initial assessment identifies whether a resident:
(1) Is capable of self-administering his or her own medications;
(2) Is capable of self-administering his or her own medication, but requires a reminder to take medications or requires physical assistance with opening and removing medications from the container, or both; or
(3) Requires that medications be administered by a TME or a licensed nurse.
§ 44-109.02. Pre-Admission Medication Management Assessment.
Within 30 days prior to admission, the ALR shall consult with the prospective resident's healthcare practitioner regarding:
(1) The prospective resident's current medication profile, including a review of nonprescription drugs;
(2) Possible adverse interactions;
(3) Common expected or unexpected side effects; and
(4) The potential that such medications have to act as chemical restraints.
DC ST § 44-109.03
§ 44-109.03. On-Site Medication Review.
DC ST § 44-109.03
Chapter 1. Assisted Living Residence Regulation.
Subchapter IX. Medication Management.
§ 44-109.03. On-site medication review.
The ALR shall arrange for an on-site review by a registered nurse every 45 days to:
(1) Supervise the administration of medications by Trained Medication Employees;
(2) Assess the resident's response to medication; and
(3) Assess the resident's ability to continue to self-administer his or her medications.
DC ST § 44-109.04
§ 44-109.04. Medication Storage.
DC ST § 44-109.04
Chapter 1. Assisted Living Residence Regulation.
Subchapter IX. Medication Management.
§ 44-109.04. Medication storage.
(a) The ALA shall provide a secured space for medication storage with access to a sink and cold storage in the same area. Space for necessary medical supplies and equipment shall be provided.
(b) The storage area shall be kept locked when not in use.
(c) The storage area shall be used only for storage of medications and medical supplies.
(d) The key to the storage area shall be kept on the person of the employee on duty who is responsible for administering the medications.
(e)(1) All medications shall be kept in their original packaging and shall be properly labeled and identified.
(2) The label of each resident's prescription medication container shall be permanently affixed and contain the resident's full name, healthcare practitioner's name, prescription number, name and strength of drug, lot number, quantity, date of issue, expiration date, manufacturer's name, if generic, directions for use, and cautionary or accessory information. Required information appearing on individually packaged drugs or within an alternate medication delivery system need not be repeated on the label.
(3) All over-the-counter (OTC) medications repackaged by the pharmacy shall be labeled with an expiration date, name and strength of the drug, lot number, date of issue, manufacturer's name if generic, and cautionary or accessory labels, in accordance with U.S.P. regulations. Original manufacturer's containers shall be labeled with at least the resident's name. The name label shall not obstruct any of the aforementioned information.
(4) In the "unit of use" distribution system, each dose of medication shall be individually packaged in a hermetically sealed, tamper-proof container, and shall carry full manufacturer's disclosure information on each discrete dose. Disclosure information shall include product name, strength, lot number, expiration date, and the manufacturer's distributor's name.
(5) Single use and disposable items shall not be reused.
(6) No stock supply of prescription medications shall be maintained, unless prior approval is obtained from the Mayor.
(7) Discontinued or expired medications shall be destroyed within 30 days in the ALR, or, if unopened and properly labeled, returned to the pharmacy. All medication destroyed in the ALR shall be witnessed and documented by two persons, one of whom shall be the ALA or the ALA designee.
(8) Residents who self-administer may keep and use prescription and nonprescription medications in their units as long as they keep them secured from other residents.
DC ST § 44-109.05
§ 44-109.05. Medication Administration.
DC ST § 44-109.05
Chapter 1. Assisted Living Residence Regulation.
Subchapter IX. Medication Management.
§ 44-109.05. Medication administration.
(a) Licensed nurses, physicians, physician assistants, and TMEs may administer medications to residents or assist residents with taking their medications.
(b)(1) Each resident shall be identified prior to drug administration.
(2) Drugs prescribed for one resident shall not be administered to another resident.
(3) The TME shall report drug errors to the healthcare practitioner or licensed nurse and shall document the incident in the resident's record.
DC ST § 44-109.06
§ 44-109.06. Medication Management Training Program.
DC ST § 44-109.06
Chapter 1. Assisted Living Residence Regulation.
Subchapter IX. Medication Management.
§ 44-109.06. Medication management training program.
(a) The Mayor shall develop medication management training courses which shall be approved by the Board of Nursing. The medication administration training program shall include instruction in the following areas:
(1) Cuing, coaching, and monitoring residents who self-administer medications with or without assistance;
(2) Pharmacology;
(3) Terminology related to medication;
(4) Procedures and precautions in administering medication;
(5) Types of medication;
(6) Actions, interactions, and effects of medication;
(7) Administration of medication in emergency or life-threatening circumstances;
(8) Recordkeeping, storage, handling, and disposal requirements for medications;
(9) Rights of residents;
(10) Monitoring of vital signs;
(11) Federal and District of Columbia laws governing medication; and
(12) Reference sources related to medication.
(b) The Mayor shall maintain a list of approved medication administration courses for the training of persons to be certified by the District of Columbia as TMEs.
(c) In order to maintain certification, every 2 years a TME shall successfully complete a clinical update or refresher course approved by the Mayor.
(d) The ALA shall document completion of the medication training course in TME's personnel file.
**DC ST § 44-109.07**
§ 44-109.07. Medication control.
---
**DC ST § 44-109.07**
District of Columbia Official Code 2001 Edition *Currentness*
Division VIII. General Laws.
Title 44. Charitable and Curative Institutions. (Refs & Annos)
Subtitle I. Health Related Institutions.
Chapter 1. Assisted Living Residence Regulation.
Subchapter IX. Medication Management.
- **§ 44-109.07. Medication control.**
(a) Each resident shall be identified prior to drug administration.
(b) Drugs prescribed for one resident shall not be administered to another resident.
(c) Staff shall report drug errors and adverse drug reactions immediately to the ALA or ALA designee who shall report, as appropriate, to the doctor, prescriber, pharmacist, resident, and resident's surrogate and shall document the incident in the resident's record.
(d) A unit drug compliance package (blister or bubble or unit dose package) shall be developed and implemented where feasible.
(e) Medications shall be refrigerated separately from lab specimens and food.
---
**DC ST § 44-110.01**
§ 44-110.01. General Conditions.
---
**DC ST § 44-110.01**
Chapter 1. Assisted Living Residence Regulation.
Subchapter X. Facility Regulations.
§ 44-110.01. General conditions.
(a) An ALR shall meet applicable zoning, building, housing, sewer, water, fire prevention codes, rules, and regulations of the District of Columbia.
(b) An ALR shall maintain all structures, installed equipment, grounds, and individual living units in good repair and operable.
(c) An ALR may be classified as a residential occupancy and may be located in a single or multi-family dwelling.
DC ST § 44-110.02
§ 44-110.02. Fire Safety.
DC ST § 44-110.02
Chapter 1. Assisted Living Residence Regulation.
Subchapter X. Facility Regulations.
§ 44-110.02. Fire safety.
An ALR shall comply with the Life Safety Code of the National Fire Protection Association, NFPA 101, 1997 edition as follows:
(1) An ALR shall be in compliance with Chapter 22, New Residential Board and Care Occupancies, Life Safety Code of the National Fire Protection Association; and
(2) An existing community residence facility that is converting to an ALR shall be in compliance with Chapter 23, Existing Residential Board and Care Occupancies, of the Life Safety Code of the National Fire Protection Association.
DC ST § 44-110.03
§ 44-110.03. General Building Exterior
DC ST § 44-110.03
Chapter 1. Assisted Living Residence Regulation.
Subchapter X. Facility Regulations.
§ 44-110.03. General building exterior
(a) An ALR shall ensure that the exterior of its facility, including walkways, yards, porches, chimney, gutters, downspouts, paintable surfaces, and accessory buildings are maintained structurally sound, sanitary, and in good repair.
(b) An ALR that provides services to wheelchair-bound residents, shall make reasonable accommodations to render the ALR accessible to residents who are wheel-chair bound through the installation of a chair lift, curbcuts, an exterior ramp, or like accommodations.
§ 44-110.04. General building interior
(a) An ALR shall ensure that the interior of its facility including walls, ceilings, doors, windows, equipment, and fixtures are maintained structurally sound, sanitary, and in good repair.
(b) An ALR shall ensure that floors and stairways provide a clean, slip-resistant, and safe surface, free of tripping hazards.
(c) An ALR shall install and maintain assist handrails or grab bars, whenever practicable, on each side of interior stairways and on one side of corridors and in bathrooms.
(d) An ALR shall provide common areas for social and recreational use totaling at least 35 square feet per resident for living, dining, therapy, and recreational activities.
§ 44-110.05. Accessibility
An ALR that provides services for wheelchair-bound residents, shall insure that:
(1) Doorways and hallways provide a clear opening of at least 32 inches; and
(2) Thresholds exceeding 1/2 inch are modified to provide a 1:12 maximum slope.
§ 44-110.06. Bathrooms.
(a) An ALR shall ensure that there is one full bathroom, for every 6 residents, including live-in family or staff. Additional full or half baths shall be available to non-live-in staff. No resident shall be required to traverse more than one flight of stairs to access a bathroom and appropriate accommodations shall be made for residents who are unable to climb stairs.
(b) When applicable, bathrooms shall contain adequate space and strategically located grab bars to allow wheelchair-bound residents to utilize toilets, tubs, showers, and wash basins without traversing a stairway.
(c) An ALR shall insure that the temperature of the hot water at all taps to which residents have access is controlled by the use of thermostatically controlled mixing valves or by other means, including control at the source, so that the water temperature does not exceed 110 degrees Fahrenheit.
§ 44-110.07. Health, Light, and Ventilation.
(a) An ALR shall ensure that each facility is lighted and ventilated in accordance with Title 12 of the District of Columbia Municipal Regulations (District of Columbia Construction Codes Supplement of 1992). Artificial night lights for corridors and exterior security lighting shall be installed.
(b) Each room shall have either a functioning ceiling light fixture or another source of artificial light.
(c) An ALR shall ensure that heating and air conditioning equipment is maintained to ensure that:
(1) During waking hours, an interior temperature throughout the facility of at least 72 degrees Fahrenheit when outside temperatures are 65 degrees Fahrenheit or below, is maintained throughout the facility;
(2) During sleeping hours, an interior temperature of at least 68 degrees Fahrenheit, when outside temperatures are 65 degrees Fahrenheit or below, is maintained throughout the facility;
(3) For individual units, heating and air conditioning shall be maintained at a temperature which is comfortable for the individual resident, whenever practicable; and
(4) When inside temperature exceeds 85 degrees Fahrenheit, mechanically cooled air shall be used in areas of the building used by residents with no inside area used by the residents allowed to exceed 90 degrees Fahrenheit.
DC ST § 44-110.08
§ 44-110.08. Bedrooms.
DC ST § 44-110.08
Chapter 1. Assisted Living Residence Regulation.
Subchapter X. Facility Regulations.
§ 44-110.08. Bedrooms.
(a) An ALR located in an existing building shall ensure that bedrooms provide at least 70 square feet of habitable space for single occupancy resident units and 100 square feet of habitable space in double occupancy resident units. Each bedroom shall have adequate dresser and closet or wardrobe space for residents' seasonal clothing and personal belongings. A secure storage space in a resident's unit shall be made available if requested by the resident.
(b) Any ALR located in a building newly constructed or renovated after June 24, 2000 shall ensure that bedrooms provide at least 80 sq. ft. of habitable space for single occupancy and 120 sq. ft. of habitable space for double occupancy.
(c) An ALR shall ensure that each resident has an adult size bed with clean comfortable mattress and extra linens. Additional furnishings, such as night stand, desk, chair, mirror, waste basket, etc., shall be made available, subject to residents wishes and tastes. Beds in double occupancy bedrooms must be at least 3 feet apart. Residents may choose to provide their own furnishings after being made aware of the furnishings that the facility is required to provide. All furnishings must meet the Fire Safety Code and be maintained in good repair.
§ 44-110.09. Kitchen.
An ALR shall provide a kitchen that has the following:
(1) Storage, refrigerator, or freezer space for perishable and nonperishable foods;
(2) Food preparation areas with cleanable surfaces;
(3) Equipment to prepare and serve food at safe and palatable temperatures; and
(4) Sufficient equipment and staffing to be in compliance with section 1116 of Title 14 of the District of Columbia Municipal Regulations, as those regulations are applicable to boarding houses.
§ 44-110.10. Laundry.
An ALR shall provide an on-site laundry facility for use of staff for residents personal laundry. All laundry shall be processed and handled in a manner to prevent the spread of infection by:
(1) Separate processing and storage of incontinent items; and
(2) Sanitizing by hot water and appropriate chemical agents.
§ 44-110.11 Special requirements for ALRs with 17 beds or more.
(a) An ALR that provides sleeping accommodations for more than 16 residents shall comply with this subchapter.
(b) The ALR may be free-standing or a distinct part of an institutional occupancy.
(c) The ALR shall be responsible for providing or coordinating personalized care to individuals who reside in their own living units (which may include dually occupied units) which may or may not include a kitchenette or living rooms and which contain bedrooms.
(d) Living units may or may not include bathrooms; except that, no more than 4 residents shall share a common bathroom. Shared bathrooms shall be in close proximity and on the same floor as living units or bedrooms.
(e) Living units or bedrooms may be locked at the discretion of the residents, except when the resident's assessment documents indicate otherwise.
(f) An ALR shall have a central dining room, living room or parlor, and common activity center (which may also serve as living rooms or dining rooms).
(g) An ALR providing 17 beds or more shall be in compliance with section 512 of Title 12 of the District of Columbia Municipal Regulations, making the facility accessible to residents with physical disabilities and aged residents.
(h) An ALR shall ensure that all food is prepared and served in accordance with Chapters 20 through 24 of Title 23 of the District of Columbia Municipal Regulations and shall organize
plumbing facilities to insure that food is processed and served so as to be safe for human consumption.
DC ST § 44-110.12
§ 44-110.12. Certificate of Need.
DC ST § 44-110.12
Chapter 1. Assisted Living Residence Regulation.
Subchapter X. Facility Regulations.
§ 44-110.12 Certificate of Need.
A Certificate of Need shall not be required for Assisted Living Residences licensed under this chapter.
DC ST § 44-111.01
§ 44-111.01. Insurance for Assisted Living Residences.
DC ST DC ST § 44-111.01
Chapter 1. Assisted Living Residence Regulation.
Chapter XI. Insurance for Assisted Living Residence
§ 44-111.01 Search Term End Insurance for Assisted Living Residences.
(a) Each ALR shall carry insurance for at least the following:
(1) Hazards (fire and extended coverage) in the amount of $500.00 per resident to protect belongings with a minimum of $2,000 of coverage per facility;
(2) Premises, personal injury, and products liability at least in the following amounts:
(A) For one to 2 beds, $100,000 per occurrence;
(B) For 3 to 9 beds, $300,000 per occurrence; and
(C) For 10 or more beds, $500,000 per occurrence; and
(3) Incidental malpractice coverage specific to the duties required of an ALA manager or any staff member in the amount of at least $100,000.
(b) If an ALR is not owned by the operator or manager, the operator or manager shall obtain proof of the owner's premises liability coverage, such as a certificate of standard landlord coverage, or shall place the owner on the operator's or manager's policy as an additional named insured.
§ 44-112.01 Appeals from actions of the Mayor.
A person or licensee aggrieved by an action of the Mayor under this chapter may appeal the Mayor's action by filing a request for hearing as provided in § 44-505.
§ 44-113.01 Rulemaking by the Mayor.
The Mayor shall promulgate proposed rules where necessary to supplement the provisions of this chapter. The proposed rules shall be submitted to the Council for a 45-day period of review, excluding Saturdays, Sundays, legal holidays, and days of Council recess. If the Council does not approve or disapprove the proposed rules, in whole or in part, by resolution within the 45-day review period, the proposed rules shall be deemed approved.
§ Implementation. [Repealed] |
A PROBABLE FAULT NEAR BRETZVILLE, DUBOIS COUNTY, INDIANA
GEORGE WHITLATCH, Indiana University
During the past summer (1930), while engaged in the investigation of the clay resources of Indiana, the writer found a structure just north of Bretzville, southern Dubois county, that exhibits very peculiar relations. At the moment, believing that the structure was probably well known in geologic literature, only casual interest was aroused. Later, however, on finding that the literature contains only a single meager description of the structure, and that at variance with the writer's observations, it was felt that further consideration should be accorded this abnormality. Consequently, a second visit to Bretzville was made in order to study the structure in greater detail. This additional study has further emphasized the very peculiar structural conditions existing at this place and has added much valuable evidence for the view that the structure is a result of faulting.
**Review of Literature.** The railroad cut just north of Bretzville, in which the above structure is exposed, has been described by J. A. Price in the report on the coals of Dubois county. He describes it as follows: "In the cut north of Bretzville, the coal measured 32-34', with 2-4" of discontinuous shale above, 6-8" of light reddish shale to shaly sandstone above the shale and shaly sandstone, and above this last mentioned shaly sandstone occurs 6-8' of shale. The coal has a decided dip to the west at this point, dipping probably 4-5' in 50 or 60 yards. At the west end of the cut the strata are somewhat distorted. Coal II has dipped beneath the surface and 7' above the railroad track is an exposure of a broken ledge of whitish sandstone, 12-15" thick. In this sandstone ledge occur streaks of coal sometimes 2 inches thick. This coal outcrop is best seen along the south side of the cut where it is exposed for 40' or more with a dip to the east, 3 feet or more in 15 feet. Above the coal at one place is a broken ledge of whitish sandstone, 1-3' thick and 20' long. Below the coal is an outcrop of 6' of shale to shaly sandstone. Just east of the center of the cut the rocks form a trough, as seen on the south side, and on the west side of the trough, near the top of the cut, are two large impure limestone boulders."
**Location.** The above structure is located in a cut on the Southern Railway just north of Bretzville, a railroad station 3 miles east of Huntingburg, southern Dubois county. The railroad runs practically due east and west at this point and is crossed at right angles, by means of a wooden overhead bridge, by the north-south road to Ferdinand and Jasper. The cut in question is directly below this bridge. As the railroad passes approximately one-eighth of a mile north of the cross-roads known as Bretzville, the cut is thus located in the southwest quarter of the northeast quarter of Section 32, Township 2 South, Range 4 West.
**Topography and Geology.** The topography of southern Dubois county is fairly rugged with hills rising in places to heights of 75 to 150 feet or more. In passing through this area, the Southern Railway has been forced to make numer-
---
1 Proc. Ind. Acad. Sci. 40: 251-257. (1930) 1931.
223rd Ann. Report, Ind. Dept. Geol. Nat. Resources, 1898. (Geo. H. Ashley), pp. 1123-24.
ous cuts, some of considerable depth. Most of the more rugged portions of the area are characterized by long ridges of fairly uniform slope, as would be expected in a region of sandstones and shales. Southern Dubois county has not been glaciated and outcrops of Pennsylvanian (Pottsville) rocks, which underlie the area, are numerous. The normal dip of this region is to the west.
**Description of Cut.** The north-south road to Ferdinand and Jasper follows the crest of a long ridge in the vicinity of Bretzville. The top of this ridge is slightly flattened, with the west side somewhat the higher. The slopes of the ridge are fairly uniform. The Southern Railway passes through this ridge, by means of a cut, in an east-west direction for a distance of approximately 675 feet. The cut is about 30 feet deep at the point of maximum depth.
Of the two sides of the cut, the north side is freest of debris and the position of the strata is plainly visible. The structural conditions exhibited on the north side are almost startling. The eastern half of the cut (Section A—Figure 1), measured from the eastern end of the cut to the bridge, is about 320 feet in length. For practically half of this distance westward from the eastern extremity of the
**Fig. 1.**
**Fig. 2.** View of east end of Southern Railway cut near Bretzville, showing steep dip of coal bed. Top of coal indicated by man's hand—coal disappears at overhead bridge due to faulting.
**Fig. 3.** View of north side of cut at overhead bridge, showing fault relations of "block" and eastern half of cut. Light portion directly beneath bridge is horizontal strata of "block." West dipping coal bed and its associated strata are to the right of "block."
cut, the strata are obscured by heavy drift. At a point about 150 feet east of the bridge, a coal, 32-34 inches thick, and 2-3 feet of associated dark gray underclay are exposed and are unobscured for the remaining 160 feet to the bridge where they disappear. This vein of coal has a west dip of 4-5 feet in 160 feet. Above the coal,
and conforming to the west dip of the coal is a succession of thin sandstones and shales. Six to eight inches of gray shale lie directly over the coal; above the shale is three feet of decidedly disconformable sandstone; overlying this sandstone is 15-20 feet of brown, sandy shale. The three feet of thin sandstones of the above section become more shaly in character as the bridge is approached. The south side of the eastern half of the cut is practically identical with the north side in dip and sequence of strata.
The above succession of west dipping strata end abruptly against a "block," some 50 feet in width in which the strata lie practically horizontal. Except for a small mass of thin sandstones directly under the bridge, at a height of six feet above the railroad bed, this block of horizontal strata is composed of shales similar to those in the eastern half of the cut, i.e., the shales are sandy and brown in color. The strata of this "block" meet the west dipping strata of the eastern portion of the cut in a very definite vertical line. (See Section B—Figure 1).
The west side of this "block" ends in a more or less well defined line, bringing up suddenly against 25 feet of similar shales that dip to the east at an estimated dip of 20-25 degrees. The highly inclined strata of this western portion of the cut do not meet the strata of the "block" abruptly but seem to have a lessening dip over a distance of 4-5 feet as the horizontal strata of the "block" are approached. This flattening of dip ("E"—Figure 1) resembles a drag zone along the plane of movement of a fault. The ends of the strata of the block appear to have been dragged upward in this zone "E", and if such is actually the case, this drag would account for the apparent flattening of dip of the highly inclined strata to the west. A second factor which lends strength to the theory that this zone has been subjected to faulting is seen in the large number of vertical joints in the highly inclined shales. The vertical jointing has divided the shales into blocks, the effect of jointing being more pronounced towards the zone of junction of Sections C and B.
These highly inclined strata of Section C appear to lie upon a semi-triangular shaped mass of sandstones and shales. The base of this mass extends from the extreme west end of the cut to a point some 150 feet east where the sandstone grades into the highly inclined shales. The lower three feet of this mass consists of fairly massive, gray sandstone, highly crossbedded. The crossbedding of this lower three feet of the sandstone makes it difficult to determine the point at which the shale and sandstone meet. The upper portion of this triangular mass is composed of 5-10 feet of thin bedded, nearly horizontal, yellow sandstones and some
shales. The entire mass of shales and sandstones of this extreme western part of the cut seem to dip slightly to the east as the highly inclined strata are approached, although the junction of Sections C and D is rather indefinite. The line EF (Figure 1), however, does have some appearance of being a shear plane, i.e., that Section C might have sheared towards the east over the triangular mass D.
The stratigraphical relations on the south side of the cut are not so well exposed as those of the north side. (Figure 5). As noted above, the eastern half of the south side has practically the same relations as found on the north side, i.e., the coal and the overlying sandstones and shales dip to the west. At the bridge on the south side, opposite the "block" of the north side, the relations are much masked by drift, but there is apparently a block of nearly horizontal shales which abruptly meets the west dipping beds of the eastern half of the cut. This "block" of the south side is of considerably less width than the "block" on the north side; its width probably does not exceed 10 feet, whereas the north side "block" had a width of some 60 feet. The west side of this smaller block, similar to the conditions on the opposite side of the cut, meets a mass of shales, highly inclined to the east, which extends west for about 30 feet. The true conditions are so masked by drift that it is impossible to determine whether there is a drag zone at the point of contact of the "block" and the highly inclined strata. A small ravine has cut down through the steeply dipping strata and exposed the shale along its course. Shear planes are evident in the shale at that point. The planes dip steeply toward the southeast indicating some movement in that direction. The jointing noted in the highly inclined shales of the north side is equally well developed on this south side. Fifty feet west of the bridge the relations are less obscured. From this point west to the western end of the cut, the entire section is revealed as brown sandy shale, except for two lenses of sandstone and three small veins of coal at the extreme west end. The shale has a fairly uniform dip of approximately 10 degrees to the east.
The sandstone lenses mentioned above are about 100 feet from the west end of the cut. The lower lens lies about 8 feet above the railroad bed; it is composed of whitish sandstone, two feet thick in its eastern extension, and is underlain with 3-4 inches of coal and 5 to 6 feet of shale. The lens has a dip to the east, conforming to the dip of the shale that encompasses it. This lower lens is exposed for a distance of 30-40 feet, and as it is traced to the west, the sandstone splits into two small lenses which are separated by streaks of coal, 1-2 inches thick. This
portion of the lower lense is overlain by 2 feet of gray underclay, 1 inch of coal, and a second lense of whitish sandstone, 2-3 feet thick, that has a slight dip towards the west. These sandstones and coals lense out to the east, and although it is impossible to trace their westward extensions to the end of the cut due to the thickness of drift, they apparently pinch out in that direction also.
**Discussion and Conclusions.** It would be unwise to attempt to draw any final conclusions in regard to the above feature at Bretzville, since the present data are too meager and insufficient to justify such procedure. Practically anyone who has had occasion to make correlations based upon data from the Pennsylvanian strata will realize the danger of drawing conclusions based upon the study of small areas. The same is true in a measure for all conclusions, as generalizations require study of a structure both as an isolated unit and as a unit of the regional structure; but the rule is particularly applicable to Pennsylvanian structures. Change of character of sediments due to depositional factors is, at times, bewildering in the Coal Measures. It is not unusual to find shales changing into sandstone along the same horizon within incredibly short distances. Much of the coal of this period was laid down in small basins, many of which were steep sided, the result being that often the strata of these basins have steep dips. Again, horizons are not continuous over wide areas but are often interrupted, due to erosion periods during times of emergence. Consequently, it is evident that any conclusions regarding the structure at Bretzville must necessarily involve further investigation of the surrounding area adjacent to the structure.
Despite the inadequate nature of the present data, the writer does feel that certain assumptions based upon the observed facts are justifiable. In making such assumptions in an attempt to give a possible explanation of the origin of the Bretzville feature, the writer wishes it to be clearly understood that both the assumptions and the suggested explanation given below are not necessarily true. They are merely an attempt to explain the structure on the basis of present knowledge; future study may reveal facts out of harmony with the following explanation.
In the first place, the conditions existing in the cut at Bretzville may be ascribed, at least in part, to the presence of a small coal basin and its attendant depositional features. The high dip of the coal and its associated strata in the eastern end of the cut may be due to the original dip of the basin in which these sediments were laid down. The east dip of the western half of the cut may also be partially explained in this manner, particularly D of Figure 5 of the south side. Possible supporting evidence for this assumption is seen in the thin veins of coal at the west end of the cut. The basin may have been very steep sided on the western side since some known coal basins of Indiana are of this type. However, the idea of a coal basin does not explain the "block" of horizontal strata found at B on both sides of the cut, the drag of this block on the north side and the shear planes on the south side, the vertical jointing of the shales just west of the "blocks," and the portions of highly inclined shales. Were the structure due to deposition alone, one would expect more or less uniformity of relations between the north and south sides of the cut—relations that are conspicuous by their absence.
We are thus forced to seek for other factors that may have aided in producing the present features. The cut, when viewed in its entirety, presents the appearance of having been subjected to faulting. It is, therefore, a logical assumption to ascribe the present abnormalities to faulting within a coal basin. The block
diagram, Figure 6, is a hypothetical attempt to illustrate the probable faulting within a portion of this coal basin and the relation of the form of the basin to the present features now exposed in the railroad cut near Bretzville. This diagram is based upon the structural relations of the north side of the cut.
The coal basin depicted in Figure 6 is assumed to have a fairly uniform but rather high dip to the west in its eastern extension, but to rise sharply to the west. This assumption is based upon the presence of the nearly horizontal sandstone mass, D, that is exposed at the west end of the cut, the mass being considered a part of the steep side of a coal basin. A further assumption is made that this steep side, being a part of an irregular basin, does not extend north and south but runs northeast-southwest at this point; hence the absence of a similar mass on the opposite (south) side of the cut. The influence of this steep side in developing the present relations of the cut is important. The sediments deposited in the northwest part of the basin were laid down over a steeply inclined surface. As a consequence of such deposition, the strata of shales at this point, previous to the faulting, were already dipping strongly to the east. (C—Fig. 6). The shales a short distance to the south (those exposed at D—Fig. 5), being deposited on a more gentle slope of the west side of the basin, possessed a lower dip to the east than the corresponding strata to the north. These differences of dip due to original deposition upon high and lower portions of the basin floor would seem also to account for the coal streaks found at the west end of the cut on the south side and the absence of such on the north side. To the north and west the coal probably pinched out rapidly, due to the steep dip of the basin, whereas the coal deposit thickened to the south, although even there one sees the rapid thinning of the coal to the west.
The initiation of stresses which resulted in the faulting of this area is not pertinent to our problem; it is sufficient to say that movement likely began along the plane of X-Y, Fig. 6, and a little later, along the plane of O-P. The plane of X-Y is sharp, with no notable evidence of excessive drag. We can assume from such criteria that this eastern side (X-Y) of the "block," B, suffered a sharp fracture along its entire vertical plane of movement. The result of such a clean break would be more rapid displacement of the east side of the block than along the west side (O-P) in which the displacement was retarded by excessive drag. The drag zone along O-P indicates such a relation attending the down-faulting of B. The ultimate effect of this differential movement of the two sides of the block was
a slight up-tilting of the west side of the block until the strata were practically horizontal as now exposed.
The displacement of B and the consequent drag of the strata along O-P set up definite stresses, principally of tensional character, the components of which were downward and horizontal (eastward). The effect of this tension is well exhibited by the vertical jointing of the shales in the portion of the cut marked "C." The strata of C, being already highly inclined, were easily affected by these stresses and sheared eastward as well as downward along the plane of E-F. The net result of this shearing was an increase of the high dip of these strata in the northwest portion of the basin. The strata to the south, not possessing an original high dip like those to the north, did not shear so readily and were not disturbed over so great a lateral distance. An additional reason for the shales being inclined over a zone of less width on the south side is seen in the fact that the fault block, B, narrows to the south. The cause of this narrowing of the fault block to the south is not known.
Such are the probable conditions that prevailed during the faulting of the Bretzville coal basin. The above hypothesis of the origin of the present structural features of the cut near Bretzville will explain most of the features of the structure, but there are some factors still unknown that no doubt would throw much light upon the problem. The amount of displacement of B, the true direction of the faults, and their extent are among the factors which might be determined by study of the surrounding areas. The foregoing theory of the development of the structure is confined to the consideration of the structure itself. No consideration has been given the regional structure. This should be done and no theory of the above structure should be considered as final without such a study. For instance, the Bretzville feature may be a fault as suggested but may also be related to a major regional fault. The suggestions of certain oil geologists, who have worked in this territory, that "there is something wrong in the vicinity of Bretzville," would seem to support the possibility of major faulting in southern Dubois county. It is hoped that future study may be made of this region and that the true relations of the feature at Bretzville may be solved.
The following is a list of the books and pamphlets which have been published by the Society since its foundation in 1874.
A complete set of these publications may be obtained from the Secretary, at the price of 2s. 6d. per volume, or 3s. 6d. per pamphlet. |
Street Excavation Permit
Date received:
Work start date:
Work end date:
The following items must be submitted at the time of application:
- [ ] Surety bond of $10,000 for 2 years
- [ ] Copy of Contractor's License
- [ ] ACORD certificate of liability insurance)
- [ ] Traffic Control Plan
Drawing submitted: [ ] Yes [ ] No
Drawing approved: [ ] Yes [ ] No
Project Information
Project Name:
Excavation location:
Contractor Information
Company:
Representative:
Address:
Business Phone:
Contact Person:
Cell Phone:
Subcontracting Company:
Business Phone:
Sub. Contact person:
Cell Phone:
Nature of Work
Type of work: [ ] Trench [ ] Boring
Work Zone: [ ] Commercial [ ] Residential
Work will include (mark all that apply): [ ] Curb & Gutter [ ] Sidewalk [ ] Driveway [ ] Irrigation
[ ] Water [ ] Sewer [ ] Storm Drain [ ] Gas [ ] Power [ ] Phone [ ] CATV [ ] Fiber
Asphalt replacement required: [ ] Yes [ ] No
If so, patching will be done by:
Comments:
Provisions
● All construction must conform to Pleasant Grove City Standard Specifications and Drawings.
● This permit expires within 30 days of the date received.
● All asphalt must be patched within 5 days of opening the trench.
● The contractor is responsible for work-zone safety and traffic control.
● Upon completion of work, inspections must be scheduled with Pleasant Grove City Public Works.
● The contractor must provide and obtain approval from the City Engineer of improvement drawings and cut sheets for all utilities, curb, gutter, and sidewalk.
● If a road closure is required, the contractor must contact Pleasant Grove Department of Public Safety (801-785-3506), Public Works (801-785-2941), and the Alpine School District (801-763-7072).
● This permit must be approved and signed prior to work being started.
● A copy of this permit must remain with the contractor while work on the project continues.
Notice: Compliance with all laws and regulations is required to protect the public from damage, injury, or loss. All necessary safeguards shall be erected and maintained by the contractor for safety and protection. The granting of a permit does not presume to give authority to violate or cancel the provisions of any other state or local law regarding construction or the performance of construction.
I hereby certify that I have read and examined this application and know the information supplied to be true and correct. All provisions of laws and ordinances governing this type of work will be complied with whether specified or not.
_________________________ ________________________________
Pleasant Grove City Streets Division Contractor
PG Engineering Notes:
| Work Description | Fee |
|------------------------------------------------------|-------|
| Street Excavation Permit (6 months) | $125 |
| Permit Extension (6 month intervals) | $125 |
| Traffic Control Plan review fee | $40 |
| Asphalt Patch - Road PCI > 70 (100 SF MIN) | $250 |
| Asphalt Patch - Road PCI < 70 (100 SF MIN) | $100 |
| Asphalt Patch - Road PCI > 70 (EACH 100 SF ABOVE INITIAL) | $100 |
| Asphalt Patch - Road PCI < 70 (100 SF ABOVE INITIAL) | $40 |
| Asphalt Patch in Road with Sealcoat (per SF) | $0.25 |
| Directional Boring (LF) | $0.25 |
| Pothole | $25 |
| Existing Curb and Gutter Replacement (per LF) | $0.50 |
| Existing Sidewalk Replacement (per SF) | $0.20 |
| Driveway Inspection (Each Driveway) | $100 |
| After hours inspection | $60 |
| Working without a permit | $500 |
| Local Street Closure (per day/per lane/per block) | $50 |
| Arterial Street Closure (per day/per lane/per block)| $100 |
| Crack seal & Sealcoat per square yard (SY) | $2.50 |
7-8-13: INSURANCE REQUIREMENTS:
A. Before a permit is issued, the applicant shall furnish to the city evidence that such applicant has a comprehensive general liability and property damage policy that includes contractual liability coverage endorsed with the following limits and provisions:
1. A minimum of one million dollars ($1,000,000.00) combined single limit per occurrence for bodily injury, personal injury and property damage, and not less than one million dollars ($1,000,000.00) in the aggregate. The general aggregate limit shall apply separately to the permit, or the general aggregate limit shall be two (2) times the required occurrence limit. The coverage shall be in the nature of broad form commercial general liability coverage. The city attorney may increase or decrease minimum insurance limits, depending on the potential liability of any project.
2. All policies shall include the city, its employees, officers, officials, agents, volunteers and assigns as insureds. Any reference to the "city" shall include the city, its employees, officers, officials, agents, volunteers and assigns.
3. The coverage shall be primary insurance as respects the city, its employees, officers, officials, agents, volunteers and assigns. Any insurance or self-insurance maintained by the city, its employees, officers, officials, agents, volunteers and assigns shall be in excess of the permittee's insurance and shall not contribute to or with it.
4. Any failure to comply with reporting provisions of the policy shall not effect coverage provided to the city, its employees, officers, officials, agents, volunteers and assigns.
5. Coverage shall state that the permittee's insurance shall apply separately to each insured against whom claim is made or suit is brought, except with respect to the limits of the insurer's liability.
6. Underwriters shall have no right of recovery or subrogation against the city, it being the intent of the parties that the insurance policy so affected shall protect both parties and be primary coverage for any and all losses covered by the described insurance.
7. The insurance companies issuing the policy or policies shall have no recourse against the city for payment of any premiums due or for any assessments under any form of any policy.
8. Each insurance policy shall be endorsed to state that the coverage shall not be suspended, voided, canceled or reduced in coverage or in limits, except after thirty (30) days' prior written notice by certified mail, return receipt requested, sent to the city.
9. Each policy shall be endorsed to indemnify, save harmless and defend the city and its officers and employees against any claim or loss, damage or expense sustained on account of damages to persons or property occurring by reason of permit work done by the permittee, his/her subcontractor or agent, whether or not the work has been completed and whether or not the right of way has been opened to public travel.
10. Each policy shall be endorsed to indemnify, hold harmless and defend the city and its officers and employees against any claim or loss, damage or expense sustained by any person occurring by reason of doing any work pursuant to the permit, including, but not limited to, falling objects or failure to maintain proper barricades and/or lights as required from the time work begins until the work is completed and the right of way is opened for public use.
B. Insurance is to be placed with insurers with an AM best rating of no less than an A carrier, with a rating of "7" or higher.
C. The permittee shall furnish the city with certificates of insurance and original endorsements effecting coverage required by the permit. The certificates and endorsements for each insurance policy are to be signed by a person authorized by that insurer to bind coverage on its behalf. The city expressly reserves the right to require complete, certified copies of all required insurance policies at any time. Consequently, the permittee shall be prepared to provide such copies prior to the issuance of the permit.
D. If any of the required policies are, or at any time become, unsatisfactory to the city as to form or substance, or if a company issuing any such policy is, or at any time becomes, unsatisfactory to the city, the permittee shall promptly obtain a new policy, submit the same to the city for approval, and thereafter submit verification of coverage as required by the city. Upon failure to furnish, deliver and maintain such insurance as provided herein, the city may declare the permittee to be in default and pursue any and all remedies the city may have at law or in equity, including those actions outlined in this chapter.
E. The permittee shall include all subcontractors as insured under its policies or shall furnish separate certificates and endorsements for each subcontractor. All coverages for subcontractors shall be subject to all of the requirements stated herein.
F. Any deductibles or self-insured retentions shall be declared to and approved by the city. At the option of the city, either the insurer shall reduce or eliminate such deductibles or self-insured retentions as respects the city, its employees, officers, officials, agents, volunteers or assigns, or the permittee shall procure a bond, in a form acceptable to the city, guaranteeing payment of losses and related investigations, claim administration and defense expenses.
G. A property owner performing work adjacent to his/her residence may submit proof of a homeowner's insurance policy in lieu of the insurance requirements of this section.
H. A provider may be relieved of the obligation of submitting certificates of insurance under the following circumstances:
1. If such company shall submit satisfactory evidence in advance that:
a. It is insured in the amounts set forth in this chapter, or has complied with state requirements to become self-insured. Public utilities may submit annually evidence of insurance coverage in lieu of individual submissions for each permit; and
b. Said coverage provides to the city the same scope of coverage that would otherwise be provided by a separate policy as required by this chapter; or
2. The work to be performed under the permit issued to the applicant is to be performed by the city, in which case insurance or other risk transfer issues shall be negotiated between the city and the applicant by separate agreement. (Ord. 98-14, 6-9-1998)
ASPHALT SURFACE (ASPHALT PATCHING IN EXISTING ROADWAY)
CUT ASPHALT TO WIDTH OF TRENCH, CONSTRUCT PIPELINE & RESTORE SUB-SURFACE THEN CUT ASPHALT 2' WIDER THAN TRENCH ON EACH SIDE THAT ASPHALT EXISTS, REMOVE ASPHALT, EXCAVATE THE EXISTING MATERIAL FOR TON EACH SIDE OF THE TRENCH IN ORDER TO RECONSTRUCT THE FULL PAVEMENT SECTION. FINISH ASPHALT RESTORATION.
3" MIN. ASPHALT ON LOCAL ROADS
4" MIN. ASPHALT ON COLLECTOR ROADS OR MATCH THE EXISTING THICKNESS WHICHEVER IS GREATER
8" MIN. ROAD BASE
AS SPECIFIED
OTHER PAVEMENT LAYER(S) AS NEEDED TO MATCH EXISTING PAVEMENT STRUCTURE
VARIES
IMPORTED BACK FILL REQUIRED
LOCATOR TAPE (TAPE TO READ TYPE OF UTILITY)
12 GAUGE TRACER WIRE
PIPE EMBEDMENT INITIAL BACKFILL
SEE NOTE 6
4" MIN. FOUNDATION
AS SHOWN ON PLAN OR MINIMUM SPECIFIED FOR TYPE OF PIPELINE
ASPHALT SURFACE REPAIR FOR TRENCHING IN A NEWLY CONSTRUCTED OR REHABILITATED ROAD (LESS THAN 5 YEARS OLD)
CUT ASPHALT TO WIDTH OF TRENCH, CONSTRUCT PIPELINE & RESTORE SUB-SURFACE THEN CUT ASPHALT 6' WIDER THAN TRENCH ON EACH SIDE THAT ASPHALT EXISTS, REMOVE ASPHALT, EXCAVATE THE EXISTING MATERIAL FOR TON EACH SIDE OF THE TRENCH IN ORDER TO RECONSTRUCT THE FULL PAVEMENT SECTION. MILL THE ASPHALT EDGE FOR A MIN WIDTH OF 1' TO A DEPTH OF 1.5". FINISH RESTORATION.
3" MIN. ASPHALT ON LOCAL ROADS
4" MIN. ASPHALT ON COLLECTOR ROADS OR MATCH THE EXISTING THICKNESS WHICHEVER IS GREATER
2 3/4 " MIN. ROAD BASE
AS SPECIFIED
OTHER PAVEMENT LAYER(S) AS NEEDED TO MATCH EXISTING PAVEMENT STRUCTURE
VARIES
IMPORTED BACK FILL REQUIRED
LOCATOR TAPE (TAPE TO READ TYPE OF UTILITY)
12 GAUGE TRACER WIRE
PIPE EMBEDMENT INITIAL BACKFILL
SEE NOTE 6
4" MIN. FOUNDATION
AS SHOWN ON PLAN OR MINIMUM SPECIFIED FOR TYPE OF PIPELINE
NOTES:
1. SLOPE TRENCH TO MEET OSHA REQUIREMENTS (LATEST EDITION) OR USE TRENCH BOX.
2. FOUNDATION AND BEDDING MATERIAL AS REQUIRED.
3. INSTALL PIPELINES ON STABLE FOUNDATION WITH UNIFORM BEARING. FOR LENGTH OF BARREL, EXCAVATE IN BEDDING FOR ALL PIPE JOINTS.
4. WHERE A NEW TRENCH PATCH IS PROPOSED THAT IS PARALLEL TO AN EXISTING ONE, THE PROPOSED TRENCH IS LESS THAN 5' OF THE ORIGINAL ASPHALT PAVEMENT WILL REMAIN BETWEEN THE PATCHES, THE ENTIRE SECTION OF ASPHALT TO THE FAR END OF THE EXISTING OR ADDITIONAL TRENCH PATCH WILL BE REQUIRED TO BE MOVED AND REPLACED.
5. TACK COAT TO BE APPLIED ON ALL SURFACES PRIOR TO PLACING ASPHALT.
6. CULINARY WATER PIPE REQUIRES 6 INCHES OF BEDDING MATERIAL, 4 INCHES FOR ALL OTHERS. |
Assessing Trends in EC Development Policy
An Independent Review of the European Commission’s External Aid Reform Process
Jean Bossuyt
Terhi Lehtinen
Anne Simon
Geert Laporte
Gwénâelle Corre
ECDPM Discussion Paper No. 16
Assessing Trends in EC Development Policy
An Independent Review of the European Commission’s External Aid Reform Process
Jean Bossuyt
Terhi Lehtinen
Anne Simon
Geert Laporte
Gwénaelle Corre
This study was funded by the United Kingdom Department for International Development
ECDPM would like to thank the Commission officials interviewed for their open and constructive contribution to this paper.
# Table of Contents
**Executive Summary** ........................................................................................................... 3
- Background ......................................................................................................................... 3
- Designing the reforms ....................................................................................................... 3
- Assessing the reform process ............................................................................................ 3
- Priorities to improve impact ............................................................................................... 4
- Conclusions ....................................................................................................................... 4
**Reforming EC External Assistance: A Shared Priority** .................................................. 5
- A timely and complex reform ............................................................................................. 5
- Background to this paper .................................................................................................. 7
**Understanding the Crisis** ............................................................................................... 8
- Agreement on the Root Causes? ...................................................................................... 8
- Implications for the Current Reform Process .................................................................. 10
**Assessing the Reform Strategy** ..................................................................................... 11
- Apparent consensus on the main objectives .................................................................... 11
- Guiding principles for implementation ............................................................................ 13
- Main Tensions and Risks Arising ..................................................................................... 14
**Key Priorities to Improve Impact** .................................................................................. 17
- Refine the EC Policy Framework ..................................................................................... 17
- Transform the approach to programming ........................................................................ 19
- Use complementarity as a strategic management tool .................................................... 20
- Organize an effective implementation of EC aid ............................................................ 21
- Review control and evaluation .......................................................................................... 22
- Improve the human resource base ................................................................................... 24
**Conclusions** ..................................................................................................................... 25
Executive Summary
Background
Since the Santer Commission resigned in March 1999, reform has topped the EU’s agenda. In March 2000, the Commission adopted a White Paper, which aims to profoundly change the ‘culture’, organisation and management of the institution. Within this broad agenda, the highest priority is given to reforming the EC’s external assistance, both in terms of strategy (what is the added value of EC aid?) and management (how should the Commission organise itself?).
This study, commissioned by the UK’s Department for International Development (DFID), provides a ‘snapshot’ of the ongoing reform process. Based on desk review and consultations with EC officials (DG DEV, RELEX, Trade), it presents an independent assessment of the reform proposals from a broad developmental perspective and identifies key challenges in further improving the performance of EC aid.
Designing the Reforms
The design of the current reform process has to take account of the generally negative political climate towards the European Union, in which political support for a truly coherent and effective European aid strategy is questionable. It has to deal with a complex set of policy instruments (with political, aid and trade dimensions) and consider the appropriate pace and sequencing of the reforms. It also faces many vested interests, a legacy of parallel administrative structures and cultures, and an urgent need to re-motivate staff after successive rounds of reform.
The experiences of previous reforms underline the need to design reform proposals from an analysis of the root causes that currently prevent the EC from being a credible global player and donor. These include: The changing roles of the Commission (from being the ‘motor’ of the European integration process to becoming a ‘manager’ of large funds), growing fragmentation and compartmentalisation of the policy and institutional framework, poor priority setting with policies being guided by instruments instead of by clearly defined political objectives, the expanding policy agenda, a mismatch between responsibilities and management resources, a ‘disbursement culture’ with numerous centralised control layers, and contradictory pressures from Member States and the Parliament.
Consultations highlighted five lessons for the ongoing reform process: First, choose an integrated approach that tackles both ‘upstream’ issues (e.g. the quality of EC aid policies) and ‘downstream’ problems (e.g. complex procedures). Second, make clear reform choices (e.g. on structures) based on objective criteria instead of political compromises. Third, recognise the political nature of external assistance. Fourth, create the conditions for effective implementation. Finally, ensure support from Member States and the European Parliament through new forms of inter-institutional dialogue.
Assessing the Reform Process
There is a wide consensus among Commission officials on the key objectives that should drive the reform process. The reform aims to restore the political legitimacy and credibility of European aid, both as a donor and as a ‘global player’, to radically improve management performance, to reform financial and administrative control mechanisms, to ensure impact on the ground, and to make the Commission a learning organisation.
All Commission officials interviewed recognised the urgency of reform. However, important divergencies could be noted with regard to basic reform choices, the implementation strategy, and the impact of reform on the overall quality of EC aid. Risks in the reform process include the sidelining of the development perspective within the new structures, overlooking partner countries in the design of reforms, focusing too much on disbursement instead of quality, creating hybrid structures without a capacity for collective management, insufficiently addressing how different institutional cultures will be managed, and failing to build enough capacities to implement the reforms.
**Priorities to Improve Impact**
The paper details six priority areas to improve the impact of EC aid. First, refine the EC policy framework, second, introduce a radical new approach to programming, third, use complementarity as a strategic management tool, fourth, ensure that implementation is effective, fifth, review control and evaluation methods, and sixth, improve the human resource base.
**Conclusions**
Everyone is asking whether this reform will make a difference? Based on our consultation process, five tentative conclusions can be drawn:
- the reform is seen as an urgent and vital priority for the RELEX group. There is a momentum for effective change;
- the overall policy direction for the reform package seems quite logical and coherent while the choices on structures and related division of labour raise concerns;
- implementation will be the real test case. Taking into account the complexity and fragility of the reform process, it may be useful to adopt a ‘rolling implementation’ approach, thus ensuring a close monitoring of progress achieved, problems encountered and remedial action to be taken.
- several issues merit greater attention, namely the need (i) to safeguard a ‘development perspective’ in the new structures, (ii) to reconcile the search for a new ‘management logic’ with effective policies, with participation of different actors and with quality of aid, and (iii) to ensure an effective role for evaluation and (iv) to make effective progress in improving the human resource base.
- probably the most realistic vision for the EC in the next years is to clearly and unambiguously develop its own ‘specificity’ as a global player and a donor while promoting management reforms that are consistent with this aim.
Reforming EC External Assistance: A Shared Priority
“Reform is important for the Commission as a whole. But reform is a question of do or die for the Commission’s external aid” (Commissioner P. Nielson, Speech at the Society for International Development, Amsterdam, 27 March 2000)
A Timely and Complex Reform
1. Since the resignation of the Santer Commission in March 1999, reform has been on top of the EU’s political agenda. The Prodi Commission came into office in September 1999 with a mandate to modernise the ‘culture’, organisation and management of the institution. In March 2000, the Commission adopted a White Paper on Reform, presented by Vice-President Kinnock, which aims to profoundly change (i) the process of setting priorities and allocating resources, (ii) the human resource policy, and (iii) financial management. If all goes well, the reforms should be implemented by the second half of 2002, following an ambitious Action Plan.
2. Within this broad agenda, the highest priority is given to reforming the management of the EC’s external assistance\(^1\). EC aid has been widely criticised over the last years for its overall ineffectiveness. The RELEX\(^2\) Group of Commissioners has decided to turn this policy area – which has a direct and visible impact on the political priorities of Member States – into one of the “flagship reforms” of the Commission.
3. Despite recent improvements\(^3\), few will deny that urgent reforms are needed to the way the EC manages its abundant aid resources. The signs are on the wall. The EU is the world’s most generous provider of development assistance, yet its influence and impact as a global player are far below its potential. The distinctive added-value of European aid remains unclear, as the Commission finds it difficult to make coherent and effective use of all its instruments\(^4\) (foreign policy, trade, development, finance). There is a growing gap between policy ambitions and implementation on the ground. The Commission’s aid system is fragmented in terms of instruments, procedures and institutional mechanisms. Long delays appear in different parts of the system. Bureaucratisation and centralisation (a ‘visa culture’) are rampant. Human resources are too scarce for the tasks and budgets assigned. Staff de-motivation is a reality in Brussels and in the field. The net result is a growing ‘malaise’ and a risk that EC aid further loses its credibility and political legitimacy.
4. To reverse this downward trend, the new Commission set in motion an \textit{ambitious reform process}. This combines a strategic review of European aid (what are the tasks and functions of the Commission in the years to come?) with a wide range of internal management reforms (what sort of organisation should the Commission be to fulfil its tasks and functions?).
---
\(^1\) EU refers to the European Union, including the Commission and the Member States. EC refers to the European Commission and related Community competences. The EC Delegations are the Commission’s diplomatic missions in the field.
\(^2\) RELEX group includes four Directorates-general (DG) in charge of external relations, namely DG External relations (RELEX), DG Trade, DG Development (DEV) and DG Enlargement (ELARG).
\(^3\) A good example is the successor agreement to Lomé IV, which introduces a coherent package of reforms in political dialogue, mechanisms for cooperation (e.g. rolling programming) and a new trade régime.
\(^4\) The Treaty of Maastricht on European Union (1992) formally recognised the existence of a European Community development policy and gave it a legal foundation. Common objectives for EC external assistance were defined and explicit reference was made to the need to increase coordination, complementarity and policy coherence.
5. A wide range of initiatives has been taken to implement this agenda. A new (functional) division of labour among RELEX Commissioners was introduced with a view to reducing fragmentation and dispersion of responsibilities while ensuring greater coordination and coherence. A Communication presenting an overall EC Development Policy Statement is being prepared for submission to the Council and the Parliament. Proposals are circulating to change the role of Committees\(^5\) and to harmonise development-related budget lines. A Task Force on the Reform of the EU External Aid Programme has been working on a package of urgent measures (e.g. to clear the accumulated backlog of dormant commitments) and options to drastically improve the management of EC aid in terms of its efficiency, responsibility and effectiveness. On March 28, 2000, the RELEX Commissioners Group adopted draft conclusions on how the re-organisation will be carried out.
6. While these initial steps look promising, it is also clear that this will be a complex process. Many factors are likely to complicate the life of the reformers, including:
- The overall political climate towards the EU is not very conducive for progressive reforms. Enthusiasm for supra-nationalism has been declining across Europe and the Commission is on the defensive at many levels.
- All donor agencies are currently reviewing their approaches to development cooperation, to adjust to major changes in the international environment and to increase impact. The EC, therefore, is not unique in reforming its aid programmes. The specific nature of EC aid, however, complicates the review process. EC cooperation policy must embrace all aspects of the EU’s ties with the developing countries. The political, economic and trade dimensions are complementary to aid. Reconciling them in a way that takes into account both the priority needs of each partner country and the Union’s objective interests, is not easy.
- The reform agenda is huge, calling for a coherent set of measures in different (inter-related) areas. This raises complex issues of pace and sequencing of the reforms. Inevitably, a wide range of vested interests will need to be confronted. It will take clout, creativity and time to overcome the legacy of many years of vertical organisation and parallel administrative cultures.
- Commission staff needs to be motivated (after successive rounds of reform) and associated in the debate. Ownership is the key to make the reform work. This holds particularly true for actors in the field (such as the Delegations), who are increasingly frustrated by the multiplication of centralised controls that hamper effective implementation.
- Reformers need to come to grips with many political, institutional and technical questions for which there are no simple answers. How much political support is there for a truly coherent and effective European aid strategy, commensurate to the EU’s power and responsibilities as a global player? How can a common administrative culture be fostered within the RELEX family? How can development cooperation be properly integrated in a more politically oriented EU cooperation? What content and organisational form should be given to the principles of devolution (deconcentration) or externalisation? What needs to be done to ensure that the reforms result in changes in partner countries?
- The constructive support of all institutional actors in the complex EU system will be required to make the reform a success.
\(^5\) There are Committees in charge of European Development Fund (EDF) for cooperation with African, Caribbean and Pacific (ACP) countries, for cooperation with Asia and Latin America (ALA) and Mediterranean countries (MED).
Background to this Paper
7. The reform process is attracting considerable interest among different actors and stakeholders concerned with EC development policy. They are keen to engage constructively in this vital debate, to be informed about new trends in Community development policy, and to understand the scope and pace of the planned reforms.
8. Against this background, the UK’s Department for International Development (DFID) commissioned a short study of trends in EC development policy from the European Centre for Development Policy Management (ECDPM), an independent foundation specialising in EU cooperation with ACP countries. The purpose is to assess the development priorities of the new European Commission and the reforms that are envisaged to implement these priorities.
9. The study was carried out in two phases. In phase one, ECDPM drafted an “Issues Paper” on trends in EC development policy, based on desk review and existing documents. In phase two, ECDPM discussed the Issue Paper with Commission staff (DG- DEV, RELEX, Trade) in order to produce this final paper.
10. It is obviously a delicate exercise to assess an ongoing reform process. Some Commission staff had questions about the rationale, timing and added value of the initiative. As an independent foundation, ECDPM has no interest in the “politics” of the reform process, nor should it try to duplicate the detailed technical work done by the Task Force in its high quality reports. Taking into account the limited time available for this assignment, the paper’s ambition is to provide a ‘snapshot’, or a basic analysis of the reform. Its potential added-value lies in (i) reviewing the main proposed policy and institutional changes in EC aid, (ii) assessing the reform from a broad developmental\(^6\) perspective without making major statements on “how good or how bad” the reform proposals are, (iii) identifying some of the main challenges to further improve EC aid performance, and (iv) promoting an informed and constructive debate on the future of EC development policy.
11. This concern with added value is reflected in the structure of the paper. Chapter II tries to understand why EC aid has gradually lost its clout and why past reform attempts failed to make a dent in this problem. This may help to identify lessons for the current reform process. Chapter III describes and assesses the EC’s overall strategic approach to the reform process, by focusing on the key objectives, guiding principles for implementation and the main tensions and risks arising from the reform. Chapter IV analyses a number of key priorities for improved impact of EC programmes. The paper ends with some conclusions.
\(^6\) In this paper, European development policy is understood to include aid, trade and political aspects. Adopting a ‘developmental perspective’ means asking whether the proposed reforms will lead to an improved impact in developing countries and societies.
Understanding the Crisis
“Over the last decades, European external assistance gradually lost its overall unity of conception, action and structure (integrating the political dimension, trade and development cooperation). This explains, to a large extent, the loss of our capacity to produce added-value and impact” (EC official interviewed)
Agreement on the Root Causes?
12. Before assessing the development priorities and reform plans of the new Commission, it is useful to briefly analyse why the EC aid machinery has got itself stuck. Why is Europe confronted with a serious credibility problem in the field of international cooperation? What has made sound management of EC aid so difficult? Why have past reform attempts had little impact? These questions are not innocent. In the absence of an agreement on the true nature of the crisis, it is hard to identify an appropriate reform package, let alone to mobilise political support for effective change.
13. The experience of recent reform attempts is sobering in this respect. There is a large consensus that the creation of a Common Service for External Relations (SCR)\(^7\) was not the best solution for the overall management crisis affecting EC aid. Undoubtedly, the SCR brought a certain private sector logic into the management of EC aid. The SCR did fulfil a useful role in catching-up with the backlog of ‘sleeping projects’ and in streamlining the administration of Community aid. In practice, however, it soon became clear that it makes little sense to focus on ‘downstream’ problems in the delivery of aid without also looking at the more fundamental ‘upstream’ issues such as the political relevance and quality of EC aid policies, the need to drastically reduce the number of projects to be managed at central level, and the coordination and complementarity with Member States and other players. Furthermore, the 1998 reform failed to address the structural shortage of human resources and to take account of field realities in the design of the reform. This, evidently, also hampers SCR performance. Even more distortive was the decision to split the project cycle (at the wrong point). This compounded the compartmentalisation of Commission services, while increasing the potential for internal conflicts in the management of aid. This quickly led the SCR to become “the alibi for all the problems in the geographical departments”.
14. This experience underlines the need to design reform proposals on the basis of a down-to-earth analysis of the root causes\(^8\) that prevent the EC from being a credible global player and donor. These include:
- **Changing roles**. A less visible, but profoundly important factor to explain the current crisis, is the ‘existential transition’ in which the Commission finds itself, i.e. from acting as a ‘motor’ for the European integration process to becoming a ‘manager’ of large funds (this holds particularly true for the policy area of external relations). This requires a fundamental shift in management culture, a reform that may have been postponed for too long.
- **Growing fragmentation of the policy and institutional framework**. The initial unity of conception and action that characterised EC development policy in the 1970’s has gradually been lost. The different regional agreements grew increasingly apart, leading to a proliferation of institutional mechanisms, instruments, procedures and budget headings. Reform measures in the last 15 years, notably on structures and responsibilities, were often induced by political
---
\(^7\) The Joint Service for the management of Community Aid to Non-Member Countries (SCR) was set up by a Commission decision dated 15 October 1997, becoming operational by stages as from 1 July 1998. Its task is to manage all aspects - technical and operational, financial and accounting, contractual and legal - of the Community’s aid to non-member countries. It is also responsible for audits and evaluations.
\(^8\) Documented in external evaluations such as the OECD Development Review Series : European Community, 1998, No 30 or the global evaluations of ACP (951338), ALA (951401), MED (951405)
compromises or vested interests rather than by considerations of aid effectiveness. The net result was a growing compartmentalisation (‘cloisonnement’) of Commission services\(^9\) and a reduced capacity to formulate coherent policies. Other negative results include lost opportunities to benefit from economies of scale, an inability to make the best use of available expertise within the Commission, and continued difficulties in building institutional memory.
- **Poor priority setting.** The Commission largely failed to adopt a strategic approach to managing its external assistance programmes. Policies have been guided by instruments rather than by clearly defined political objectives, cooperation priorities and country, region or sector specific strategies. This exacerbated management problems, as ‘strategy’ often meant little more than an aggregation of ill-fitting projects. It also made it difficult to ensure coherence with stated policy objectives (e.g. the fight against poverty), to have political leverage on partner countries (in relation to available resources) and to build complementarity with Member States.
- **Expanding policy agenda.** Like other donor agencies, the EC has been confronted with an increasingly complex and diversified policy agenda for international cooperation. It is pressured to provide a specific contribution to an emerging international strategy for development. In addition to this, the EC is part and parcel of the EU’s external action, alongside trade policy and political dialogue. Over the last decade, the EU’s role and responsibilities as a global player have widened considerably. All this has led to a dramatic expansion of the Community’s aid programmes in terms of geographic outreach, priorities, tasks, instruments and financial resources. This rapid growth, however, took place without a commensurate adaptation of the Community’s strategies, organisational set-up, management systems and capacities.
- **Mismatch between responsibilities and management resources.** “An army cannot walk without boots”. In these words, a Commission official described the achilles heel of the EC aid system. The EC clearly lacks the necessary people, in numbers and in skills, to properly manage the many tasks and budgets under its responsibility. The lack of a comprehensive and efficient personnel policy that gives adequate attention to performance further compounds the problem. In the absence of political support to tackle this problem, the Commission has reverted to gap-filling solutions (mainly in the form of massive subcontracting to high-cost external expertise), which does little to help institutional learning.
- **Limited focus on implementation and poor feedback.** Some EC officials highlighted that past systems focused on control and commitment level instead of implementation capacities. EC Delegations tend to see “Brussels” as a main obstacle for effective implementation on the ground. Examples given include the tendency to design new policies without considering implementation requirements and capacity constraints, the obsession with spending money (the ‘disbursement culture’) and the use of centralised control layers instead of result-oriented management systems –that put a premium on strategic evaluations, impact and feedback.
- **Contradictory pressures.** Several interviewees consider Member States and the European Parliament to be “partners in crime”, sharing responsibility for the loss of credibility of EC aid. Recent evaluations support this allegation. For instance, Commission services have to cope with more than 80 legal bases reflecting the proliferation of budget lines voted by the European Parliament. There are divergent views on the eventual budgetisation of the EDF. The financial regulations are too rigid for effective multi-annual programming. The annual budget cycles pushes the Commission to commit resources without duly assessing the feasibility on the ground. Political control from Member States (or ‘comitology’) and the Parliament has become a complicated and time-consuming system, draining scarce human resources at the Commission level while providing little added value in qualitative terms.
---
\(^9\) Communication was perceived to be a problem, both ‘horizontally’ between services and ‘vertically’ between the Commissioner and his/her DG.
Implications for the Current Reform Process
15. What lessons for the current reform process can be drawn from this assessment? Commission officials interviewed agreed on the need to avoid the mistakes of the past by introducing reforms that would not tackle the structural problems affecting EC aid. As one interviewee put it: “this is not the time for careful medical treatment, but for surgery”.
16. In our assessment, five lessons should be considered in assessing the current reform process:
- **Avoid piecemeal approaches**. The reform should clearly adopt an integrated approach that aims to gradually eliminate the multi-layered bottlenecks to improved EC performance, both ‘upstream’ and ‘downstream’ and including ‘soft issues’ such as changes in management culture, attitudes, and incentives.
- **Make clear choices rather than political compromises**. A good reform is the result of clear choices in terms of objective criteria (efficiency, responsibility and effectiveness), and not because of vested interests or political compromises (as in the past).
- **Recognise the political nature of external assistance**. Over the past years, assistance based on political imperatives (i.e. Phare, Tacis, Obnova\(^{10}\)), natural disasters (i.e. Hurricane Mitch, Mozambique), and finally on international emergencies (i.e. Kosovo, East Timor) has become a predominant feature of the EC’s external assistance, thus reducing the proportion for traditional development aid. This increases the need to coordinate actions closely within the framework of the CFSP\(^{11}\), including with the High Representative. Even in ‘traditional’ developing countries, the scope of cooperation has shifted towards economic cooperation and regional trade agreements (MED, ALA, ACP) and reinforced political partnerships (reflected in EU-Latin America and EU-Africa summits).
- **Protect implementation**. This is the rallying cry that can be heard from officials with responsibility for “getting things done”, especially at the field level. In their view, the reform should start from what is needed to ensure effective implementation on the ground. ‘Protecting’ implementation means checking the relevance and feasibility of new management systems and procedures with those that will use them. It means that EC Delegations should be protected from an endless stream of interference from headquarters. It also challenges the EC to adapt its management systems to the requirements of much more participatory and decentralised forms of development cooperation, as promoted for example in the new post-Lomé agreement.
- **Ensure consistent political support from Member States and Parliament**. A constructive dialogue among the Council, the European Parliament and the Commission is a pre-requisite for strengthening the EC’s political profile, building complementarity, simplifying controls and ensuring greater flexibility for actors in the field. Organising such a new culture of inter-institutional dialogue is a shared responsibility, requiring effective communication flows and a willingness to openly discuss desirable reform options in different areas\(^{12}\).
---
\(^{10}\) PHARE provides assistance to Eastern European accession candidate countries, TACIS to the countries of the former Soviet Union and OBNOVA to countries in the western Balkans.
\(^{11}\) The Common Foreign and Security Policy of the European Union, the so called ‘second pillar’.
\(^{12}\) The geographical division of Commission services has largely corresponded to the geographic and thematic composition of Council working groups. Recent merging of I and II pillar geographic working groups (Asia, Latin America, Africa etc.) triggered a reform process that aimed to improve the coherence of the EU’s external actions. Furthermore, the new thematic approach in the Commission reform could contribute to discussions on the role of Development Council, whose de facto ‘geographic’ scope still seems to be limited mainly to the ACP, except in the case of humanitarian aid (such as Hurricane Mitch and East Timor).
Assessing the Reform Strategy
"The reform should primarily be geared at strengthening the political dimension of European external assistance. This, in turn, should dictate the choice of the most appropriate management structures" (EC official interviewed)
Apparent Consensus on the Main Objectives
17. From statements by relevant Commissioners, policy reports and interviews conducted in the context of this study, there seems to be a wide consensus among Commission officials on the key objectives that should drive the reform process. Evidently, this process should be seen as part of the wider process of administrative reform, as proposed in the White Paper.
18. In order to facilitate analysis, the main objectives pursued by the reform process are described below in five key clusters:
- **Restoring the political credibility and legitimacy of European aid**. This is the broader political ambition behind the reform process. The global context of development policies has changed significantly in the last decade. New forms of international cooperation - going far beyond traditional aid policies - are required to better manage globalisation and environmental interdependencies, to combat rising poverty levels, and to prevent, manage and resolve conflicts. The EU has a critical role to play in fostering this new agenda for international cooperation. Yet so far, its potential for action and impact remains "grossly under-exploited" both as a political power and as a donor. Hence, a key objective of the reform is to put Europe firmly back on the map as a credible and effective global player. In practice, this leads the Commission to propose a wide range of reforms, aimed at:
- setting clear strategic priorities for EC external assistance;
- aligning EC aid with internationally agreed development objectives (e.g. agreed upon at the level of DAC-OECD);
- making a much more coherent and effective use of policy instruments (political dialogue, development cooperation and trade);
- ensuring consistency with second pillar strategies and policies;
- searching in a proactive manner to build complementarities with other players (particularly EU Member States) on the basis of comparative advantages;
- adapting EC structures to this new policy agenda by (i) dividing the responsibilities of Commissioners by subject (trade, development, enlargement and CFSP) rather than by geographical region, (ii) re-allocating responsibilities for CFSP, human rights and democratisation to DG RELEX and responsibility for trade to DG Trade, (iii) giving DG DEV the overall responsibility for framing development cooperation policies for all developing countries, and (iv) entrusting the RELEX Commissioner with the mandate to coordinate the external relations activities of the Commission
- **Radically improving management performance**. In terms of priorities, this is clearly the most urgent reform ambition of the RELEX Commissioners. It makes little sense for the Commission to have major ambitions in international cooperation if it cannot deliver fast, reliable and high quality aid. The Task Force therefore concentrated on ways to get the EC aid machinery back on track. Based on its work, the RELEX Commissioners Group endorsed the following reform proposals at their meeting of 28 March:
---
13 See chapters 2.3 and 2.4 of Overall Policy Statement, version 24/02/2000.
- partial reunification of the project cycle from identification to implementation, to be placed in the hands of a single body (see further for more details),
- full reunification of the project cycle (including programming) for some ‘politically sensitive’ programmes (Phare by DG ELARG; human rights, CFSP and Rapid Reaction Fund by DG RELEX; management of partnership with NGO’s by DG DEV, management of humanitarian activities by ECHO\textsuperscript{14});
- development of an effective multi-annual programming system that sets out clear strategies, budgets, expected results and implementation guidelines for aid interventions, to be approved by SCR, and monitored under the responsibility of the RELEX Group of Commissioners;
- extensive devolution of authority to the Delegations, to ensure that the management of the project cycle – from identification to implementation – can take place, as much as possible, at local level.
• \textit{Reforming financial and administrative control mechanisms}. The Commission has suffered from a lack of credibility both in terms of financial and administrative management. In the past, reports on mismanagement and fraud reinforced control mechanisms instead of enhancing administrative capacities. The current reform proposes to review control mechanisms according to modern management methods. This should improve the effectiveness and quality of aid delivery and restore confidence in the Commission’s management capacities. Reform includes the following measures:
- redefinition of the role of “comitology” to speed up implementation and to re-focus control on sectoral and country-level strategies (especially at the programming stage), policy guidelines and ex-post reporting;
- creation of a new evaluation and monitoring system, characterised by independence and feedback;
- introduction of “results-based management” as a key element of control and the responsibilisation of staff;
- introduction of “activity-based budgeting” and the regrouping of multiple small budget lines under a few thematic budget headings to reduce the number of legal bases and burdensome reporting mechanisms;
- creation of an inter-departmental Quality Support Group (QSG)\textsuperscript{15} to improve the consistency and quality of programming;
- Decentralisation of financial control in the level of DGs and introduction of ex-post auditing.
• \textit{Ensuring impact on the ground}. This objective appears in a less prominent way in the different reform papers, but is certainly very \textit{present} in the mind of Commission officials interviewed. It is captured in reform proposals geared at:
- devolving authority to EC Delegations;
- deconcentrating and redeploying staff to EC Delegations;
- rationalising the use of technical assistance;
- strengthening regional EC Delegations;
- redefining the role of “Brussels” in the management of EC aid;
- reducing the number and interference of centralised, ex-ante control systems;
- exploring ways to decentralise management responsibilities to partner countries.
\textsuperscript{14} European Community Humanitarian Office.
\textsuperscript{15} The purpose of the QSG is to ensure the quality of programming documents, identify best practices that promote excellence and innovation, disseminate results, inform the RELEX group on the progress of a programming exercise, and to propose any necessary improvements.
• **Becoming a learning organisation.** The complexity and scope of the international cooperation agenda puts a premium on specialisation, pooling of resources and expertise, cross-fertilisation of best practices, and systematic evaluation and feedback. For reasons mentioned above (see par. 14), the EC has been rather weak on these points, despite its expert staff. The Commission now seems determined to gradually turn itself into a “learning organisation”. To this end, it has proposed:
- to harmonise, where possible, the policy and institutional framework for EC aid management;
- to break down the walls between Commission services and expertise as a precondition for improved effectiveness and impact;
- to promote collective decision-making processes among RELEX Commissioners, both at the level of programming and evaluation;
- to identify best practice inside and outside (“benchmarking”) so as to promote excellence and innovation;
- to turn evaluation and feedback into a tool for institutional learning and management;
- to disband the extensive use of external technical assistance offices in order to ensure a more cost-effective and accountable mobilisation of external expertise, internal learning and capacity building.
**Guiding Principles for Implementation**
19. So much for the overall objectives of the reform. The question is how the Commission plans to implement this ambitious reform agenda. What methodology has it selected for the reform process? What are the guiding principles for implementation? To what extent does the Commission’s implementation strategy take into account the lessons of past reforms and build upon positive elements of the past?
20. Obviously, perceptions on the nature and methodology of the reform vary according to the perspective of the different institutional actors involved in the reform process. What seems logical to a Cabinet member may be heresy to an EC Delegate in the field. What seems a ‘participatory process’ to the officials involved in the Task Force, may be seen as a ‘top-down process’ by other actors. Some interviewees saw issues related to structures and the division of responsibilities as “fairly neutral”, considering the reform’s intention to create a collective management culture. Others perceived coherence in structures as a key factor, largely determining the success of any collective management system. Attempts were made in this paper to properly reflect these divergencies of views.
21. At first sight, efforts were made to avoid some of the mistakes from the past:
- **Dialogue rather than a purely political negotiation.** Contrary to past reforms, there was a search for appropriate reform options through internal debate. This certainly holds true for the working methods of the Task Force. From the outset, it decided not to focus the reform agenda immediately on structures. Instead, it adopted a pragmatic approach, starting with practical matters (what urgent measures are needed to reduce the backlog?) and technical issues (how can the Commission best organise itself to deal with programming in a reunited project cycle?). Furthermore, the Task Force operated like a relatively open discussion forum where experiences, best practices and lessons from other donor agencies could be pooled and analysed on their merits.
- **Focus on the political dimension of development cooperation.** Another interesting feature of the reform methodology is the perceived need to get out of the dichotomy between ‘development cooperation’ and ‘external relations’. The Task Force, for instance, made a clear methodological choice to search for new management mechanisms that would allow EC aid to
be more firmly embedded in the broader political objectives of the EU as well as in the new priorities on the international cooperation agenda. It has sought to overcome the barriers between those allegedly involved with ‘development and poverty alleviation’ and those concerned with ‘politics and economic cooperation’.
- **Recognition of implementation needs.** The adoption of a pragmatic, down-to-earth approach to the reform process has facilitated a focus on conditions for effective implementation (much more than in the past). Thus, new joint SCR/geographical department working methods have been developed to deal with the urgent problems related to the backlog. The Task Force has insisted on the need to properly organise the interface between programming and implementation. A strong warning has been issued against devolving authority to EC Delegations without first tackling the issue of staffing.
- **Relation to Member States and Parliament.** While the Commission has institutional responsibility for the reform process, several interviewees agreed on the need to properly involve Member States and the Parliament in the reform process (so far, however, it is not clear how the Commission plans to translate this laudable intention into practice).
### Main Tensions and Risks Arising
22. All Commission officials interviewed recognised the urgency of reforms. However, important divergencies could be noted with regard to the basic choices of the reform process, the chosen implementation strategy and the likely impact of the reforms on the overall performance of the EC in the field of external assistance.
23. Our assessment of possible weak points in the proposed reforms can be summarised in the following six categories:
- **Development perspective sidelined?** Following a trend that can be observed in most OECD countries, EC development cooperation is becoming more strongly integrated in the overall EU foreign policy. This evolution arouses quite some controversy and concerns. Some interviewees, for instance, observed that the development needs of ACP countries might not be best served by a separate agreement, disconnected from mainstream EU foreign policy considerations. As mentioned before, the nature of EC development cooperation is getting increasingly ‘politicised’. Against this background, they argue, there is a logic in the current reform and proposed division of labour. Others, however, expressed major worries that the ‘development perspective’ may quickly lose ground in the new RELEX framework. For example, in the reform process, the former DG8 trade unit in charge of the trade regime with the ACP, has become the “trade and development” unit within the horizontal DG trade. There are concerns over a small unit’s capacity to defend the interests of developing countries within a larger horizontal DG in charge of European trade interests. Underlying these worries are divergent views on whether developing countries’ interests are best served within regional or multilateral arrangements\(^{16}\). Other examples could be provided to illustrate this potential danger. In essence, this raises the issue of how best to marry foreign policy concerns and development objectives. This is difficult in all EU Member States. In the European context, it is even more complex to find a balance between these two worlds.
- **Partner countries overlooked?** On the whole, little attention seems to be given to the role and place of partner countries in the reform process. Few interviewees touched on that subject. To some extent, this is quite logical, as reforms are inherently internal processes. However, it could be dangerous to underplay the partner country perspective in the design of reform
---
\(^{16}\) Commissioner Lamy has requested an evaluation of regional trade arrangements, including assessment on envisaged ‘Regional partnership agreements’ (REPA) with the ACP.
proposals. At the end of the day, implementation of many of the new proposals (programming, budget aid and sector approaches, decentralised management, quality control) will have to take place in the field, in close collaboration with partner countries. Another worry is the consistency between the planned reforms and the commitments in cooperation agreements. It is hard, for instance, to see how the Commission will reconcile its (legitimate) desire to concentrate its efforts on a limited number of areas with the principles of partnership and ownership that underpin ACP-EU cooperation. Another example is that the internal tendency to focus EC aid on “big projects” may contradict with political objectives to reorient EC aid to poverty reduction or to work more with decentralised actors, whose needs are often for smaller projects.
- **Pushing money or delivering quality aid?** A key objective of the reform is to dramatically improve the “efficiency” of EC aid (to speed up implementation). This laudable objective puts terrible pressure on the EC to increase disbursements, including with regard to the backlog. In the current political climate and taking into account the scarcity of human resources, this may lead to a situation where pushing money (or realising “*le chiffre d’affaires*”) becomes the key institutional incentive rather than the pursuit of quality aid. As one of our interviewees observed: “Success in implementation is not the same thing as ensuring better impact”.
- **Hybrid structures or collective management?** There is some tension regarding the proposed structural reforms. Three issues stand out. First, there is concern that the reform has again generated a hybrid structure in relation to the division of labour between geographic and thematic services. The choice for a functional approach is not consistently pursued at the level of the services involved. Portfolios and departments have not been matched. This may seriously erode the overall capacity and influence of DG DEV. There are doubts whether a collective management system within the RELEX family can work effectively under these conditions\(^{17}\). Second, certain thematic areas (such as CFSP and human rights) have already been unified, merging services from different DG’s. For example, the human rights unit of the former DG8 was merged into the overall human rights unit in DG RELEX. However, these two units had a slightly different scope – DG RELEX was a focal point for horizontal, “macro-level” human rights issues (UN, OSCE\(^{18}\), COHOM\(^{19}\), etc.), whereas the former DG8 unit managed concrete “micro-level” projects within the ACP region. There are worries that reconciling these different agendas may prove to be difficult. Third, the proposal to fully re-unify the project cycle (including programming) may appear logical for fairly ‘self-standing’ programmes (e.g. pre-accession instruments). The same does not necessarily hold true for issues such as human rights and NGO-funding, which are an integrated part of EU cooperation policies and programmes. Managing these programmes *outside* the general SCR (or Office) implementation framework may lead to ‘stand-alone’ and inconsistent approaches.
- **Multi-cultural management?** The fragmentation and isolation of different geographic services in the past have contributed to specific institutional “identities” inside the Commission bureaucracy, including privileged “partnerships” with each respective geographic area. The restructuring of services seriously “shakes” some established boundaries, and it will inevitably take time to change mentalities and working methods in order to adjust to new management requirements. Perceived insufficiency of staff consultations and neglect of “cultures” prior to regrouping services might slow down the pace of change.
---
\(^{17}\) For example, coherence of the thematic approach is hampered by different financing mechanisms between regions. On the one hand, ALA and MED cooperation frameworks are bound to annual budget cycles, whereas the EDF cooperation has a more flexible multi-annual perspective. However, harmonisation of procedures seems to lead towards the “budgetisation” of the EDF, even though there is no consensus about its future. Costly political priorities such as aid to Kosovo and East Timor inevitably lead to budget cuts in aid to other developing countries within the Community budget. The extra-budget nature of the EDF has so far allowed the ACP to conserve their shares.
\(^{18}\) Organisation of Security and Cooperation of Europe.
\(^{19}\) Council’s horizontal working group on human rights.
• *A capacity gap?* This is, undoubtedly, an ambitious reform programme that will fundamentally challenge the way in which the Commission manages its aid resources. Effective implementation will require a host of new capacities to be developed. A case in point is the proposed new system of programming, with its focus on a much more strategic and integrated approach to development cooperation. This implies building a strong capacity at the EC level for strategic analysis. Other examples could easily be provided. All point to the need to quickly bring in new skills and capacities. This promises to be a delicate issue in the sequencing of reforms.
Key Priorities to Improve Impact
“What have been the main problems with the EU’s approach to human rights? Have our policies been coherent? Has the EU in all its many emanations taken a consistent line? Have we been able to deliver? [...] Looking ahead, what should our objectives be? What do we want to achieve? [...] What is the EU’s added value in this area? [...] How should we organise ourselves?” (Commissioner C. Patten, Speech at the Human Rights Discussion Forum, Brussels, 30 November 1999)
24. This quote summarises many of the urgent policy and institutional questions to be addressed in the reform process. It focuses on the EU’s role in human rights, but could easily be extrapolated to other policy areas.
25. In this section, we focus on largely the same questions in the context of the overall reform process. Six key priorities for improving the impact of EC aid are identified. Each is assessed using a common analytical framework which (i) briefly explains the issue at stake, (ii) analyses the Commission’s plans, and (iii) identifies some of the key opportunities and challenges arising.
26. These priorities are clearly inter-related. They need to be addressed in a coherent manner, as a ‘package’. The best illustration of this is perhaps the issue of human resources. There was a wide consensus among Commission officials that it makes little sense to come up with ambitious new policy statements (e.g. on poverty alleviation) or plans for programming and complementarity, if the human resource problem is not addressed *at the same time* (or even in the first place).
Refine the EC Policy Framework
27. In his speech at the previously mentioned SID Conference, Commissioner Nielslon aptly captured the first priority at hand: “There is a pressing need for a more clearly defined development policy at the European level. The Commission must join the development donor mainstream. It needs a policy to do so”. In practice, this means clarifying the overall political objectives of EU external assistance, its ‘core business’ and specific added value.
28. To this end, the Commission is developing three main initiatives:
- **Definition of an overall EC Development Policy Statement.** This document, elaborated by the different RELEX services, proposes overarching strategic guidelines to EC external aid programmes, including the ‘mainstreaming’ of the poverty focus\(^{20}\) as a key element of the EC development cooperation. It stresses the need to relate Community assistance more strongly to internationally agreed strategies (e.g. recent UN summits, DAC strategies, the Comprehensive Development Framework, the Poverty Reduction Strategy Papers). The Commission also commits itself to produce annual reports on the implementation of EC development policy.
- **Improving policy coherence.** The re-organised structure for external action is seen to provide a better institutional setting to make progress on the thorny issue of EU policy coherence. The Commission also wishes to upgrade existing monitoring systems to ensure a ‘coherence check’.
---
\(^{20}\)The Treaty of Amsterdam (art. 177) incorporates the commitment to *refocus aid on the objective of poverty reduction*. A poverty focus goes beyond aid to LDC’s, it is also closely related to economic growth, macro-economic and trade policies, social sector development and fiscal policies. The operationalisation of a poverty focus includes: identifying suitable indicators to measure the evolution of poverty in a dynamic way (rather than on the basis of static and questionable indicators), balancing performance-based aid allocations with the needs of the poorest populations, most often living in non-performing countries, measuring social effects of macro-economic and trade policies, and linking up with major international undertakings on poverty.
• **Harmonisation.** The reform aims to rationalise the ‘patchwork’ of cooperation agreements and instruments that emerged over time. During the interviews, however, several Commission staff argued that the first priority should be to harmonise approaches to programming and procedures rather than policies (considering the growing differentiation among partner regions and countries).
29. How to assess these policy reforms? Will they make much of a difference? The following challenges and opportunities arise:
• **Relevance of an overall development policy statement?** The ‘negotiation’ of an overall policy framework for all developing countries is an interesting experiment, as it may serve as a laboratory to develop a new culture of consultation, joint decision-making and institutional learning between (compartmentalised) departments. However, it remains to be seen how this new policy framework will be used in practice. Will it serve as a guide to re-orient EC aid or will it be shelved, as many other statements and resolutions before? Critics could easily argue that the Policy Statement still sees a role for Community aid in too many areas and that the proposed criteria to define the EC’s ‘comparative advantage’ will not suffice to set ‘negative priorities’.
• **Towards a more sophisticated debate on poverty?** The nature of cooperation\(^{21}\) with ALA, MED, ACP, TACIS and PHARE countries is different, constituting a major challenge to any attempt to ‘harmonise’ policies through an overall strategic framework. Especially the proposal to mainstream a ‘poverty focus’ has triggered a semantic debate and negotiation among RELEX services on the meaning of ‘poverty-focus’ and ‘development’ in EC external assistance. This may help to clarify and enrich the concept of poverty. There are obviously areas where greater policy harmonisation should be possible across regions (e.g. on policies and approaches to micro-credit). Much more could be done, with due respect to the principle of differentiation, to translate the poverty objective more consistently in country strategies, sectoral policies and monitoring and evaluation systems. The challenge, however, will be to implement this objective in a creative and balanced way, which duly takes into account the multi-dimensional nature of poverty, the diversity of situations and needs in developing countries, as well as the Community’s specific role and added-value. The new system of programming should create opportunities for DG DEV to bring poverty concerns to bear in the elaboration for country strategies for non-ACP countries.
• **Need for flexible cooperation in new foreign policy areas.** The strengthened articulation of aid, trade and political dimension should improve the Commission’s internal coordination on new policy issues, such as non-military aspects of crisis management, which involves DG RELEX as a coordinator of the CFSP and DG DEV as focal point for humanitarian aid. Also, cooperation with the High Representative for CFSP is crucial. Recent crisis situations, such as Kosovo and Mozambique, have increased pressure to drastically improve the European dimension of external action.
• **Strategic priority setting for EC external assistance.** Strategic coordination among RELEX Commissioners becomes the key element in the future coherence of EC external assistance and the redefinition of the Commission’s role in the EU’s foreign policy. A particular challenge will be to better articulate EU interventions at different levels (global, regional, national and local). This requires new systems of programming and improved complementarity - the two next priorities on our list.
\(^{21}\) For example, EC external assistance to MED-countries is considered part of ‘economic cooperation’ to achieve a free trade zone by 2010. Similarly, cooperation with ALA countries combine development (in Central America) and trade (FTA negotiations with Mercosur) elements. It is argued that many middle-income countries (for example within the Mercosur) do not consider themselves in the category of ‘poor countries’ where a poverty focus as a key objective would be justified. These ‘political’ sensitivities are often coupled with the simplified idea that a poverty focus would exclusively mean aid to LDC’s.
Transform the approach to programming
30. If the EC is to become a global player with a coherent international cooperation policy, it must change the way in which priorities are set, overall budgetary resources are allocated and strategies are formulated at country and regional levels. In the current jargon of the RELEX reform, this process is called “programming”. It refers to the “upstream” part of the project cycle, i.e. the political process to define overall country and regional strategies (the so-called “strategic phase”) and the translation of these agreed strategies into concrete country or regional programmes (the so-called “programming phase”).
31. The Commission is keen to “radically overhaul its approach to programming”. The new system is meant to fulfil four key functions:
- **Set political priorities.** Multi-annual programming is the core of the new system. It implies making clear political choices, establishing aid strategies and budget allocations (at global, regional and country levels), ensuring the right policy mix for each beneficiary country covering political goals, financial assistance, sectoral priorities and other Community instruments (i.e. the coherence objective), defining criteria for the choice of projects, and agreeing on financial targets and results to be achieved on the basis of a regular feedback from the field.
- **Enhance a dialogue with partner country and Member States.** The EC wants to use programming as a tool for a fuller and more concrete dialogue with the governments and societies of partner countries (i.e. to organise conditionailities and increase policy leverage) and with Member States (i.e. to organise complementarity).
- **Act as a management tool.** Programming could serve as a management tool to set financial targets, to monitor progress in implementation, to identify and explain variances with stated targets, to reorient priorities, etc. This, evidently, implies a clear link between programming and budgetary procedures. It requires a well-functioning system of “rolling programming” (such as envisaged in the successor agreement to Lomé IV), so as to allow for flexibility and systematic fine-tuning. This management tool should not only serve internal purposes. It should be coherent with partner country priorities and be used in other coordination fora.
- **Provide a baseline for evaluation.** The key to an effective system to control quality and evaluate results is to organise at the level of programming. It provides an “upstream” tool to set targets and to define expected results in priority areas, to be monitored according to established criteria and indicators. The EC also wants to ensure that evaluation findings will feed into the next cycle of programming.
32. These proposed changes to the programming process look promising and merit support. But how can this new system be made to work in practice? What factors can accelerate this reform or block it? In our assessment, five main points merit to be further consideration:
- **Obligation to work together.** One of the perceived strengths of the new functional division of labour and related absence of a clear geographical split, is that it pushes RELEX Commissioners to dialogue and collaborate with each other. It creates a sort of obligation to jointly develop a coherent policy mix for a particular region or country in the context of the programming exercise. For this system to work, however, it will be essential to create new mechanisms to ensure that development concerns are taken seriously, especially in programming aid to non-ACP countries.
- **Feasibility.** This is clearly a bold and ambitious reform. It remains to be seen whether the Commission can generate the political and institutional support to get this process on track. It
will require a relaxation of financial regulations, a major shift in bureaucratic culture, a strategy to reconcile the new system of programming with the requirements of a partnership approach on the ground, and a wide range of new capacities to manage the new system of programming.
- **Linkage between programming and implementation.** It will be critically important to properly organise the interface between the functions of programming and implementation in the new system of project cycle management. Several interviewees were concerned that programming would go into too much detail, thus bringing rigidity into the next phases of the project cycle.
- **EC policy priorities.** The programming exercise will confront the EC with the question of its specific contribution and added value (in relation to Member States and other development players). How will it, for instance, determine areas of sectoral concentration? How can EU programming build on existing coordination efforts in a given country? How far should it go in relating its contribution to initiatives such as the Comprehensive Development Framework?
- **Partner country capacity for programming.** In order to move towards greater ‘ownership’ and ensure decentralised forms of programming, efforts should be made to strengthen the capacity of partner countries to better programme their priorities.
### Use Complementarity as a Strategic Management Tool
33. In its reform proposal, the Commission seems committed to make a much more pro-active and strategic use of complementarity\(^{22}\) to improve overall EU aid effectiveness and management.
34. The new approach to complementarity is based on the following principles:
- **Concentrating on core activities.** Recognising its capacity constraints, the Commission proposes to take a proactive leading role in limited priority areas in order to restore confidence on its capacity to deliver aid effectively. In areas that fall outside this concentration, the Community would contribute financially to the programmes of the Member States and other donors. Consequently, there would not be a total withdrawal from certain sectors or policy areas, instead, the EC involvement would take different forms according to the type of policy area and country under consideration (e.g. possible co-financing).
- **Making more effective use of country strategy papers.** The Commission is keen to explore ways to use the process of defining country strategy papers as a tool to promote complementarity with Member States. This would facilitate an open exchange of views on the priority areas and expertise of each donor within a specific country context and contribute to a strategic definition of the European dimension of aid to that country.
- **Evidently, this new approach to complementarity should be welcomed.** By moving beyond the principle of “voluntary” coordination, the Commission could use complementarity as a strategic management tool (i) to set its own priorities (including “negative priorities”) according to its resources, available expertise and perceived comparative advantage, (ii) to avoid the danger of having to respond to a broad range of demands from partner countries, (iii) to stop the tendency of being used as “the donor of last resort”, and (iv) to clarify the specific added value of European aid.
---
\(^{22}\) Ever since the Treaty of Maastricht invited the EC and Member States to ensure greater ‘complementarity’ in their aid programmes, laudable efforts have been made to clarify the concept and to see how it could best be operationalised. The EC Communication and the ensuing Council resolution on complementarity (May 1999) are the most recent initiatives in this context. Yet in practice, complementarity still appears to be a political slogan, rather than an operational reality.
35. There are also serious implementation problems:
- **Strategic programming with Member States.** This implies first and foremost a capacity from the Commission to set “negative priorities” for itself. Unless the current ‘patchwork’ of objectives and means is replaced by a clearly designed set of priorities, complementarity will remain an idle quest. The next step will be to negotiate a division of labour with Member States. This implies that partners clearly indicate where their priorities lie and candidly identify the areas in which they have a weakness. This will be a difficult exercise as for many Member States, aid is an integral part of their foreign policy, based on national strategic interests. Previous operational coordination efforts have focused on past experiences, but strategic programming would require exchange of information on future activities, including trade, aid, political and even geo-strategic dimensions, in a given country.
- **Reviewing co-financing procedures and the untying of aid.** The announced use of co-financing on a wider scale would require that co-financing procedures be reviewed and possibly a progressive “harmonisation” of tendering and contracting procedures be introduced among different Member States. Capacity-building and cross-cutting expertise would be needed to effectively channel community funds through bilateral programmes. Furthermore, the question of untying of aid, previously proposed by the Commission, would have to be reviewed, even though there are very divergent views on the question.
- **Question of visibility.** Direct financing of programmes has been perceived as the best way to ensure visibility of each donor on the ground. New co-financing procedures would have to ensure the visibility of each contributing actor and the “added value” of cooperation. Clear guidelines on visibility should be elaborated to convince all the Member States to channel community funds through one Member State or other donors.
**Organize an Effective Implementation of EC Aid**
36. Reforming the ways and means to implement the ambitious EC policy objectives is another key priority of the reform process. It is a pre-requisite to ensuring the quality of programmes as well as their development impact on the ground.
37. The proposed reforms build on past initiatives (like the creation of the SCR) and evaluations. They aim to gradually ‘externalise’ and ‘devolve’ certain implementation responsibilities to the Delegations and, where conditions permit, to the authorities in beneficiary countries.
- **Re-unification of the project cycle.** As mentioned before, the idea is to partially re-unify the project cycle (from identification of projects to implementation) and to entrust it to a single body. In a first phase, pending more radical changes, this responsibility would be entrusted to the present SCR (for about 80% of the annual external assistance from the budget and the EDF). For some areas of “particular political sensitivity”, the management of programming and implementation would remain united (pre-accession instruments, CFSP spending, NGO partnership, human rights)
- **Externalisation.** Gradually, a body for project identification and implementation will grow out of the present SCR, with a marked preference for an office-type\(^{23}\) structure by 2002. ‘Externalisation’ should speed up management, reduce staff shortfalls and rationalise the use of technical assistance. The Commission is keen to ensure “full political control” over that body through a Board composed of all RELEX Commissioners (with the Development Commissioner as Chief Executive and the External Relations Commissioner as Chairman). The legal and budgetary framework for the body in its final form –when it might be floated
\(^{23}\) The RELEX group has compared agency- and office-types of structures in their conclusions of March 28, 2000.
out of the Commission itself- will depend on the wider discussions on externalisation and the revision of the Financial Regulation.
- **Devolution of implementing responsibilities to EC delegations.** The realistic sequencing (first resources, clear guidelines and capacity-building, then devolution of responsibilities) of the reform is crucial.
- **Further harmonisation of procedures.** This is another key element of the reform and a task to which the SCR has committed much effort over the past year and a half. A progress report (as per 31.12.1999) has recently been presented to the Development Committee of the Parliament, proposing ‘a package’ of further improvements and simplifications with regard to EC procedures and management.
38. In assessing these reform proposals, the following questions merit further attention:
- **Office or agency?** The choice of final form of externalisation should take into account lessons learnt from recent experiences (such as ECHO or the Agency for Reconstruction of Kosovo) and make an analysis of staffing costs and financial feasibility under A or B parts of the budget.
- **Setting up a framework for a field-focused management.** The SCR has struggled with problems of payments and financial management. Consequently, it failed to pay attention to field realities and to establish a proper monitoring system. The devolution of competencies to Delegations creates a unique opportunity to set up a “field-focused management” system, and to link the programmes and projects more closely to the objectives of sustainable development, ownership and capacity building on the ground. This new culture of management should use local expertise extensively.
- **A ‘market’ based logic of harmonised procedures.** The increasingly market-based tendering procedures for aid implementation may sometimes undermine development objectives. For example, tendering procedures tend to favour the cheapest offers, not necessarily those with the best quality. The discontinuity of subcontractors in the field prevents the EC from building up institutional memory. Asking bank guarantees on subventions to NGOs acting in the South for development purposes is profiting Northern commercial banks. All this may prevent the full utilisation of available local expertise. These potentially negative consequences of new procedures should be screened.
- **Size of projects.** The increased focus on large projects is understandable from a management perspective. Yet many relevant development actors in the field often operate at the grassroots level, where small projects could have important development impact. The risk of putting exclusive priority on financial management, at the expense of adaptation to local realities, may hamper the effectiveness of envisaged implementation reforms.
**Review Control and Evaluation**
39. Evaluation and auditing constitute the main pillars of an improved quality control system. The reorientation of the ‘comitology’ system to a much more strategic role could reduce the pressure on management while speeding up implementation.
40. The key elements of the reforms proposed by the EC are geared to move (i) from a control of procedures to a control of results and (ii) from ineffective ex-ante control to ex-post control. To this end, the reform proposes the following concrete measures:
• **New system of** evaluation/monitoring. Feedback and independence\(^{24}\) are the key characteristics of an effective evaluation system. In future, evaluation will seek to articulate three levels: strategic evaluation, evaluation of projects, and monitoring. Programming and evaluation should be connected.
• **Results-oriented** management. The assessment of results is part of the programming phase. It implies constant re-adjustment of indicators, set up together with recipient countries.
• **More effective comitology.** As recent discussions in EDF committees have shown, committees could be used more to have a strategic dialogue on priorities and complementarity. The new Internal Agreement on the 6\(^{th}\) EDF goes in this direction, emphasising the need for a sectoral approach and focusing on larger programmes. However, several interviewees felt the proposed changes are not going far enough.
• Rationalisation of budget control.
• Systematic quality control.
• Decentralised and ex-post financial control and auditing.
41. Different opportunities, but also challenges, emerge from these reforms that aim to review control and evaluation mechanisms:
• **Evaluation as a key element of the new “cooperation cycle”.** Evaluation should be seen as a strategic tool for improved impact instead of a control system. It allows better adaptation of programming to the regional conte, better use of previous experiences in the design of new programmes (best practices, high quality xt partners) and possible improvement of cooperation with other donors based on most effective division of tasks. The participation of partner countries in evaluation and monitoring activities should be enhanced.
• **Need to restore confidence and constructive inter-institutional cooperation.** Comitology and budgeting should not be seen as exclusive control mechanisms, but more as institutional settings for improved inter-institutional dialogue. Especially annual reviews, foreseen in the new rolling programming of the EU-ACP cooperation, will require more flexible committee procedures.
• **Effective monitoring of aid quality and impact at local level.** In the past, monitoring has played a marginal role in the system. By contrast, some foresee it becoming a key tool, used mainly by Delegations to measure the impact of the chosen “policy mix” at local level. Others see its role more attached to independent evaluation and feedback functions.
• **Quality of delivery versus quality of aid?** The new system should avoid considering the quality of aid delivery exclusively in terms of financing and auditing, leading to a possible confusion between disbursed money and impact. The monitoring of quality should focus more on expected (and achieved) results than on formal procedures.
• **Balance between financial control and impact monitoring and evaluation.** Following the resignation of the previous Commission due to fraud and mismanagement, the reform efforts have mainly focused on improved financial control. However, the use of monitoring and evaluation needs to be complementary to auditing in order to achieve qualitative improvements to EC aid.
\(^{24}\) Feedback from evaluation reports ‘enrich’ programming and establishes linkages between strategies and lessons learned from implementation. Feedback is also needed to improve the management and “responsibilisation” of staff. The need for independent evaluation is widely recognized, including in the personnel policy of evaluation service. It has been suggested that the evaluation unit should be attached to the Secretariat-General or even become an external structure, like the Court of Auditors.
Improve the Human Resource Base
42. To increase EC aid effectiveness, adequate human resources must be devoted to run the size of the various programmes. This applies for improving the administration of EC aid (e.g. working away the backlog). It is also a pre-requisite for putting the new reforms into practice or for realising the full potential of the Post-Lomé IV agreement.
43. How does the Commission propose to tackle this longstanding bottleneck? Four main steps are considered:
- **Applying the White Paper.** The Kinnock reforms will also be applied to the EC external assistance programmes. This may instill the much needed ‘management culture’ while improving the overall personnel policy.
- **Activity based budgeting.** This system is perceived to be the lynchpin of a new management system that would allow the Commission to match tasks with available resources.
- **New recruitments.** The newly envisaged ‘Office’ could address some of the most critical in-house staff shortages.
- **Rationalisation of costs related to Technical Assistance Offices.**
44. Will these reforms suffice? Two main observations can be made:
- **Quantity or quality?** There seem to be divergences of views on the desirability of going for “more staff” (in quantitative terms). While some consider this to be an absolute priority, grounded in objective criteria reflecting under-staffing, others feel the first priority should be to drastically review (and reduce) the Community’s policy priorities and areas of concentration. In the view of the latter, the Commission should first specify its added value and “niche of specialisation”. This may help to re-deploy human resources, build-up a critical mass of expertise in a given area and generate “high quality products” (e.g. take a lead role on adjustment issues). Having done this, the Commission would find itself in a much stronger position to make its case for “more staff”.
- **Making a better use of existing human resources within the Commission.** The objective underlying this option would be to enhance the cross-fertilisation of EC expertise between the different RELEX DGs and even beyond the ‘RELEX family’. This would imply that the RELEX services would no longer operate as separate bodies with their own sources of expertise on various development related areas. Synergies would be sought with other DG’s that have certain sectoral competence within the Commission.
Conclusions
45. Since the coming into office of the new Prodi Commission, an ambitious reform process of EC external assistance has been put on track, aiming to make radical improvements to the speed, quality and impact of EC aid. This paper sought to better understand the rationale, content and feasibility of the reform proposals that are emerging from the RELEX Group of Commissioners. In this concluding chapter, the following questions are considered: Will the proposed reforms make a difference? Will the different measures suffice to drastically improve EC aid delivery and development impact? Will they give a boost to the credibility of the EU’s Common Foreign and Security Policy? How feasible are they? What weak spots can be identified at this early stage? What development performance can be expected from the EC in the next two years?
46. **First**, the reform is clearly seen as an *urgent and vital priority* for the RELEX Group. Despite a certain ‘reform fatigue’, there is a consensus on the need for quick and radical reforms in the way the EC plays its dual role of global player and donor. Both the credibility and the legitimacy of European external assistance are seen to be at stake. This sense of urgency suggests there is a *real momentum for effective change*. Urgency, however, can be a double-edged sword. While it can help to clarify reform choices (e.g. the setting-up of an ‘Office’ rather than an ‘Agency’), it can also lead reformers to pay insufficient attention to certain critical dimensions (e.g. the partner country perspective, the coherence of the new structures, the dialogue with Member States and Parliament). In addition to this, one should not forget that the reform takes place in a difficult political climate and a situation of institutional crisis at the Commission. Against this background, it is probably not realistic to expect the reform to come up with ‘ideal’ structures and management systems in the short term. The reform is more likely to generate ‘second-best-options’, reflecting what is feasible at this stage, and to move on as conditions allow.
47. **Second**, the *overall policy direction* of the reform package seems *quite logical and coherent*. In many ways, the current reform process marks a break with past attempts in that it adopts a comprehensive approach, linking political objectives with far-reaching management reforms. It seeks to improve ‘connectivity’ among policies, instruments, structures and sources of expertise. It aims to create a proper interface between people involved in programming and implementation. It gives staff greater responsibility for the financial consequences of their decisions. It recognises realities in the field by emphasising the need to devolve authority to (reinforced) EC Delegations. It seeks to cut through time-consuming bureaucracy, by reviewing control and audit functions. It stresses the need for a new culture of dialogue with Member States and the Parliament (through a radical review of ‘comitology’). Most importantly, the Commission seems decided this time to tackle the human resource problem. Taken together, these measures have the potential to improve EC aid performance and impact. However, with regard to structures and division of responsibilities among RELEX Commissioners, things are less clear. Major fears have been expressed that the current reform will not solve the problem of “hybrid structures” and may erode the overall capacity and influence of DG DEV in development matters.
48. **Third**, while the reform agenda looks rather solid on paper, *implementation will be the real test*. Managing the transition from the current crisis situation to the ‘new-style Commission’ will be a complex and fragile process. Complex, because it remains to be seen how the Commission will actually put into practice the ambitious reform programme (in terms of mainstreaming the poverty focus, applying the new system of multi-annual programming, using complementarity as a strategic management tool, speeding-up implementation without losing a focus on quality, and moving from ex-ante to ex-post controls). Fragile, because the reform process has to confront different forms of *internal political and bureaucratic resistance* that may quickly erode the substance of the reforms. Against this background, it might be useful to adopt a ‘rolling implementation’ approach to the reform process, thus ensuring a proper monitoring of the process (what works, what doesn’t work), a continuous dialogue (especially with Member States) and remedial action in due time. This could be done, for instance, through some sort of informal Task Force arrangement, associating the different stakeholders.
49. **Fourth**, there are some **critical dimensions** in the reform process that merit **greater attention**, including:
- **Safeguarding the “development” perspective.** Like many EU Member States, the EC is keen to better articulate development cooperation, political action and trade. The proposed new culture of “collective management” to EU external assistance (to overcome traditional barriers between RELEX and DG DEV) may respond to both the changing world of international cooperation and the need for greater development impact. However, it would appear that stronger institutional guarantees are needed to ensure that development concerns are not sidelined. An efficient operation of the new programming system is a key priority in this respect.
- **Reconciling different reform logics.** The reform is primarily grounded in a “management logic” (to speed-up EC aid delivery). This may be justified, taking into account the prevailing crisis. However, the Commission may need to think through how this logic will be reconciled with three other forms of logic that are essential to effective aid: (i) defining a set of appropriate policies that make a difference, (ii) ensuring that local actors (governments, civil society, private sector, EC Delegations) are at the head of the new management system, and (iii) ensuring that aid quality gets priority over aid disbursement.
- **Making evaluation the linchpin of the new management system.** If development impact is a central concern of the reform, it would be useful to clarify how evaluation, monitoring and feedback – as strategic tools for policy formulation and implementation – will be institutionalised in the new system. At this stage, the reform proposals remain rather vague on this.
- **Strengthening the human resource base.** This priority seems high on the political agenda. Different ideas are on the table, including activity based management (to match tasks to resources). Putting this into practice will not be evident. There is both a sense of urgency and a need for creative and flexible solutions. It is not simply a question of “more staff” but also of attracting new capacities to cope with an increasingly complex development agenda and with the specific nature of RELEX work. Close monitoring of progress achieved in this area will be essential.
50. **Fifth**, what can be expected from the EC in the next two years in the field of external assistance? Probably the most **realistic vision** is for the EC to gradually develop its own ‘specificity’ as a global player and a donor while promoting management reforms that are consistent with this aim. The challenge will be to link a strategic vision (what is the specific added-value of EC aid in a rapidly changing international context?) with concrete management reforms (what changes can help the EC to effectively use its different policy instruments while delivering quality aid?).
27
The European Centre for Development Policy Management (ECDPM) was created as an independent organisation by the Netherlands Government in 1986.
ECDPM aims to improve international cooperation between Europe and countries in Africa, the Caribbean, and the Pacific (ACP). It does this through capacity building for policy management, the promotion of policy dialogue between ACP countries and Europe, and the provision of information and facilities for knowledge exchange.
The Centre’s objectives are:
- to strengthen the institutional capacities of public and private actors in ACP countries to manage development policy and international cooperation
- to improve cooperation between development partners in Europe and the South
Working from a small base, the Centre relies on its cooperation with other organisations, partner institutions and relevant resource persons in the design and execution of its activities.
ECDPM Discussion Papers report on work in progress at the European Centre for Development Policy Management. They are circulated for review and comment by interested practitioners, researchers and policymakers. Discussion papers are not formally edited and reviewed and receive a limited circulation. Comments, suggestions, and requests for further copies should be addressed to the author at the address on the back cover. Opinions expressed in this paper do not necessarily represent the views of ECDPM or its partners.
Further information: Martine van Randeraat, Onze Lieve Vrouweplein 21, NL-6211 HE Maastricht, The Netherlands, E-mail: firstname.lastname@example.org, Fax: (31)-(0)43.350 29 02, http://www.oneworld.org/ecdpm |
A memory system having a plurality of memory cells each capable of storing more than two states also has a reference voltage generator which is used to determine the state stored in a selected memory cell. The plurality of memory cells are arranged in groups of columns, and in a preferred embodiment, the reference voltage generator is arranged in columns which are located between two of the groups of columns. The memory system uses voltage sensing as opposed to current sensing and therefore the only current used is to charge line capacitances. The voltage reference generator is only enabled during the period of time that a cell is being read from.
11 Claims, 3 Drawing Figures
ROW DECODER/DRIVERS
ADDRESS INPUTS
COLUMN DECODER
ADDRESS INPUTS
COLUMN SELECT
COMP
COMP
COMP
COMP
COMP
COMP
BIT 1
BIT 2
BIT 3
BIT 4
FIG /
FIG 2
FIG 3
MEMORY SYSTEM HAVING MEMORY CELLS CAPABLE OF STORING MORE THAN TWO STATES
CROSS REFERENCE TO RELATED APPLICATIONS
Patent application Ser. Nos. 238,183 and 238,178 are related co-pending applications by the same inventor, assigned to the same assignee, and filed on same date as the present invention.
BACKGROUND OF THE INVENTION
This invention relates, in general, to memory systems, and more particularly, to a memory system having a plurality of memory cells each capable of storing more than two states.
Solid state memories are well-known, and particularly solid state memories using metal oxide semiconductor field effect transistors. There is a continuous effort to make memory systems smaller, faster, and use less power. One way to make a memory smaller is to be able to store more information in each memory cell location. However, one of the problems with such a memory is the ability to distinguish which state is stored in a selected memory cell. Some reference source will be required, but in order to make the memory a competitive product it must remain relatively small in overall size and low in power consumption.
The present invention provides such a memory system. Each memory cell is capable of storing more than two states. The memory cells are arranged in a manner that permits them to be made relatively small in size, which results in fast operation and low power consumption. The invention is described as having memory cells capable of storing four states which provides two binary bits. It will be understood that a memory cell could be built employing the principles of the present invention having memory cells capable of storing more than four different states. The present invention also has a reference voltage generator which consumes a very minimum amount of power and is only energized when data is being read from the memory. The reference voltage generator is built in a manner to conserve silicon chip area and to maximize speed of operation of the memory.
SUMMARY OF THE INVENTION
In carrying out the above and other objects and advantages of the present invention, there is provided a memory system having a plurality of memory cells capable of storing more than two states. In order to read data out of the system and determine the stored state, a reference voltage generator is provided which generates an output only when the memory is being read from. The memory system also includes a unique sense amplifier and translator which translates the output of the sense amplifier into binary format. The voltage reference generator is arranged in columns which are spaced among the memory cell columns in a manner which minimizes reading time.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 represents a memory system in block diagram form in accordance with an embodiment of the present invention;
FIG. 2 schematically shows a portion of the memory system of FIG. 1; and
FIG. 3 illustrates in logic and block diagram form a portion of the memory system of FIG. 1.
The exemplification set out herein illustrates the preferred embodiment of the invention in one form thereof, and such exemplification is not to be construed as limiting in any manner.
DETAILED DESCRIPTION OF THE DRAWINGS
A portion of a memory array along with the peripheral circuitry required to provide a memory system is illustrated in FIG. 1. The memory system shown in FIG. 1 is capable of being integrated onto a single semiconductor chip. The memory system has a row decoder 11 which receives row address inputs and provides an output which drives a selected row 41. The row address inputs are decoded by row decoder 11 so that only the selected row 41 is enabled. Although only five row select lines 41 are illustrated, it will be understood that in an actual memory there are usually many more row select lines. Row decoder 11 will typically include drivers to drive the decoder output since the decode will have many transistors connected to its output. A column decoder 12 receives column address inputs and provides outputs for enabling the selected column select devices in column select 13 and 14. Two groups of columns are illustrated. The first group of columns is represented by columns 19, 20, 21 and 22. The dashed lines between columns 21 and 22 indicate that in an actual system many more columns than just the four illustrated would be controlled by column select 13. The second group of columns is illustrated by columns 23, 24 and 25. The dashed lines between columns 24 and 25 indicate that in an actual system many more columns would be controlled by column select 14. In a preferred embodiment, column select 13 will control the same number of columns that column select 14 controls. Located between the two groups of columns are reference columns 16, 17 and 18.
The memory system illustrated in FIG. 1 will have a plurality of memory cells 15, wherein each memory cell is capable of storing more than two states. The memory cells are more fully described in related patent application Ser. No. 238,183 which is hereby incorporated herein by reference. Voltage reference columns 16, 17 and 18 each provide a different voltage level. Since it will not be known ahead of time which state is stored in one of memory cells 15, each time a memory cell is selected the three references must be used to determine which state is stored. Each reference cell 10 will contain a single transistor which will be substantially the same size as the single transistor contained within each memory cell 15. The voltage reference generator is described more fully in related patent application Ser. No. 238,178 which is hereby incorporated herein by reference.
Reference columns 16, 17 and 18 are each directly connected to sensing circuits 42 and 43. Sensing circuit 42 has three comparators 36, 37 and 38. Sensing circuit 43 also has three comparators which are identical to each of the comparators in sensing circuit 42. Column select 13 provides an output on line 39 which is connected to each of the comparators 36, 37 and 38. Column select 14 provides an output on line 40 which is connected to each of the comparators within sensing circuit 43. Each comparator in sensing circuit 42 and 43
provides an output which is connected to a transistor or decoder 44. The operation of the sensing circuits and translator 44 will be described in greater detail hereinafter.
Translator 44 is illustrated as providing a four binary bit output. The memory system illustrated in FIG. 1, by way of example, provides a four bit output. However, it will be understood that most memories provide an eight bit output since most words contain eight bits. Those persons skilled in the art will recognize that it will be necessary to provide two more groups of columns controlled by two more column selects and each group having a sensing circuit for an eight bit output.
In a preferred embodiment, the reference columns are arranged so that approximately the same time delay will be experienced from the furthermost memory column in one group as the furthermost memory column in the next group. For example, the delay from reference column 17 to column 19 will be the same as the delay from reference column 17 to column 25 assuming that column 19 is the furthermost column away on the left hand side and column 25 is the furthermost column away on the right hand side of reference column 17. In a memory system wherein time delays are not critical then only one group of reference columns is necessary and the single group of reference columns can service as many sensing circuits as required. A memory system providing a two binary bit output would only require one sensing circuit. In such a memory system the data contained in each memory cell location could be read out individually.
All reference cells 10 in column 16 provide the same voltage, while all the reference cells in column 17 will provide the same voltage which will be different from the voltage provided by the reference cells of column 16. All the reference cells of column 18 provide the same voltage which is also different from the voltage provided by the reference cells of column 16 or 17. When a row is selected, each of columns 16, 17, and 18 will provide a reference voltage so that each time a cell or cells are read from all three of the reference voltages required will be generated. Since it was assumed that only four states are stored in any one cell, only three references are required to determine which state is stored. One of the states is assumed to represent logic signal 00 or 11 and therefore does not require a reference. The state not requiring a reference can either be represented by the highest voltage or the lowest voltage available from a memory cell.
FIG. 2 illustrates, in schematic form, portions of columns 19 and 22 of FIG. 1. Transistor 50 represents a memory storage location as do transistors 51, 52 and 53. Each of the transistors is connected in a source follower configuration having their drains connected to voltage terminal 54 and their sources connected to a column line. Transistors 50 and 52 have their sources connected to column 19 whereas transistors 51 and 53 have their sources connected to column 22. Row select line 41 is connected to the gate electrodes of transistors 50 and 51, and row select line 41' is connected to the gate electrodes of transistors 52 and 53. In the system illustrated in FIG. 1 each row line is connected to the gate electrodes of one transistor in each column that is intersected by the row select line including the voltage reference generators. Each column line has an associated parasitic capacitance. The parasitic capacitance for column 19 is illustrated by capacitor 58, and the parasitic capacitance for column 22 is illustrated by capacitor 59. Each column as well as output line 39 is connected to a voltage reference during a precharge phase. Transistor 62 connects column 19 to the voltage reference, illustrated as ground, when a precharge signal connected to its gate electrode enables transistor 62. Column 22 is coupled to ground by transistor 63 and transistor 64 couples line 39 to ground. Transistors 63 and 64 are also enabled by the precharge signal. Transistors 56 and 57 serve as column select devices. Transistor 56 couples column 19 to line 39, and transistor 57 couples column 22 to line 39. The gate electrode of transistor 56 is connected to line 60, and the gate electrode of transistor 57 is connected to line 61. Lines 60 and 61 represent output lines for column decoder 12 (shown in FIG. 1). Column select 13 and column select 14 each contain a plurality of transistors such as transistors 56 and 57. Each transistor connects one column to line 39 or line 40.
Each reference column 16, 17, and 18 will look like column 19 and 22 except that the reference columns will not have column select devices such as transistors 56 and 57. Another difference, which is not a visual difference, is that all the transistors in a given reference column 16, 17 or 18 will each have the same threshold voltage. The transistors in a memory column, such as columns 19 and 22, will most likely not all have the same threshold voltage. The threshold voltage will be adjusted to represent the state stored at a given location.
FIG. 3 shows sensing circuit 42 along with a portion of translator 44. Sensing circuit 42 has comparators 36, 37 and 38 which receive an input from line 39. Comparator 36 also receives a first reference input on column line 16, comparator 37 receives a second reference input on column line 17 and comparator 38 receives a third reference input on column line 18. The output of comparator 37 goes to an inverter 71. Inverter 71 provides an output which goes to an input of NAND gate 73 and to the input of inverter 72. The output of inverter 72 represents a binary digit shown as Bit 1. The output of comparator 36 goes to a second input of NAND gate 73. The output of NAND gate 73 goes to an input of NAND gate 74. The output of comparator 38 goes to an input of inverter 76. The output of inverter 76 goes to a second input of NAND gate 74. The output of NAND gate 74 is a binary digit shown as Bit 2. Inverters 71, 72, and 76 along with NAND gates 73 and 74 are part of translator 44 (FIG. 1).
The truth table for the circuitry illustrated in FIG. 3 is as follows:
| Memory Cell Output | Comparator Output | Bit 1 | Bit 2 |
|---------------------|-------------------|-------|-------|
| Unknown | C1 | C2 | C3 | | |
| 1 (highest) | 0 | 0 | 0 | 0 | 0 |
| 2 | 1 | 0 | 0 | 0 | 1 |
| 3 | 1 | 1 | 0 | 1 | 0 |
| 4 (lowest) | 1 | 1 | 1 | 1 | 1 |
The left hand column of the truth table indicates the four states that can be stored in the memory cells. Since the transistor in each memory cell is connected in a source follower configuration the output from the memory cells will be a voltage. Number 1, in the left hand column, is assumed to be the highest voltage with number 4 the lowest voltage. Comparator output C1 is the output for comparator 36, comparator output C2 is the output for comparator 37, and comparator output
C3 is the output for comparator 38. Through an arbitrary choice, when the highest output voltage from a memory cell is present, all the comparator outputs are zeros which represent output bits 1 and 2 as zeros. When the second highest voltage level is present in the selected memory cell then the output from comparator 36 is a logic level "1" while the output from comparators 37 and 38 are logic levels "0" which produce output Bit 1 as a logic level "0" and output Bit 2 as a logic level "1". When the third highest voltage level is present in a selected memory cell then the output of comparators 36 and 37 are logic level "1's" and the output of comparator 38 is a logic level "0" making output Bit 1 a logic level "1" and output Bit 2 a logic level "0". The lowest voltage level in a memory cell causes all the outputs from the comparators to be logic level "1's" making output Bit 1 a logic level "1" and output Bit 2 a logic level "1". From this it can be seen that the first reference voltage level provided by column line 16 can be equal to the second highest voltage level or could be some value between the first and second highest voltage levels. The second reference voltage level provided by column line 17 could be equal to the third highest voltage level or some voltage level between voltage levels 2 and 3. The third reference voltage level provided by column line 18 could be equal to the lowest voltage level or some value between the third and fourth voltage levels.
By now it should be appreciated that there has been provided a unique memory system having a plurality of memory cells which are capable of storing more than two states. The memory system includes a plurality of references which are used to determine a state stored in a selected memory cell. The memory system is arranged in such a manner to result in a small memory which consumes a minimum amount of power and operates with minimum time delay. Each of the reference voltages is generated by a transistor having a threshold voltage close to or the same as a threshold voltage of a transistor used in the memory array and connected in the same manner. Thus each of the reference voltages will correspond very closely to one of the four possible voltages out of the memory cell. The required voltage levels are not generated until a row is selected. The reference voltage devices are arranged in columns which are in a layout similar to the memory columns. Consequently, any circuit parameter which can affect performance and which might vary in processing will vary the reference devices exactly the same as the memory devices. Therefore, the reference voltages will track the memory signals and the memory array will work over a wide range of processing parameters. The memory system has been illustrated as a read only memory and utilizes voltage sensing rather than current sensing. Thus, the only current needed is that required to charge the line capacitance. In addition, the memory devices can all be made minimum size, and hence the memory system can be made to draw very little power.
I claim:
1. A read only memory having a plurality of rows and columns, comprising: a plurality of memory cells arranged into the rows and columns, the memory cells being capable of storing more than two states; row decoder for receiving row addresses and providing a plurality of outputs, each of the outputs being coupled to a row so that when a predetermined row is addressed the predetermined row is activated; column select means coupled to each of the columns; column decoder for receiving column addresses and providing outputs to the column select means so that when a predetermined column is selected the predetermined column is coupled to an output of the column select means; a plurality of reference sources arranged in columns and rows so that a row of reference sources is enabled each time one of the plurality of rows of memory cells is enabled, the columns of reference sources being positioned in between columns of the memory cells; comparator means having a first input coupled to the output of the column select means and a second input coupled to the reference sources, the comparator means providing an output indicative of the state stored in a selected memory; and means for translating the state into a binary output.
2. The read only memory of claim 1, wherein the plurality of reference sources is one less than the number of states which each memmory cell is capable of storing.
3. The read only memory of claim 2 wherein the comparator means includes a separate comparator for each one of the plurality of reference sources.
4. The read only memory of claim 1 wherein each memory cell includes a field effect transistor connected in a source follower configuration.
5. The read only memory of claim 1 wherein the column select means includes a field effect transistor for each column.
6. The memory array of claim 1 wherein the columns having the plurality of reference sources are adjacent each other and have an equal number of memory cell columns on either side.
7. A memory array comprising a plurality of memory cells arranged in columns and rows, each memory cell being capable of storing more than two states; a plurality of references located among the memory cells and arranged in rows and columns like the memory cells; row decoder means for receiving row addresses and providing a row select signal for selecting an addressed row, the row select signal also enabling at least one reference; column select means for selecting a predetermined column; means coupled to the column select means for receiving an output from the column select means and for comparing the output against the at least one reference, the at least one reference being coupled to the means coupled to the column select means; and a translator for translating an output from the means coupled to the column select means to a digital output.
8. The memory array of claim 7 wherein the plurality of references are arranged so that all the references in a given column are the same.
9. The memory array of claim 7 wherein the columns having the plurality of references are located so as to have an equal number of memory cell columns on either side.
10. A memory array having a plurality of rows and columns of memory cells wherein the columns are divided up into groups having an equal number of columns, the memory cells each being capable of storing more than two states; a plurality of columns of reference cells located between adjacent groups of columns of memory cells, the reference cells being substantially identical in physical size to the memory cells; comparator means for receiving memory cell data comparing such data to one of the reference cells; and a translator for translating an output from the comparator means to binary data.
11. The memory array of claim 10 wherein the memory cells each being capable of storing more than two states can actually store at least four states; and the plurality of columns of reference cells being at least three columns, each column further comprising a plurality of reference cells wherein each reference cell has a field effect transistor wherein each of the columns of reference cells has transistors of different threshold voltages than the other of the columns of reference cells.
* * * * * |
Directions to the Geek/Roth Lakehouse
The following are driving directions from Chicago, the Chicago suburbs and Dubuque. If you need directions from a different location, please let us know. Also, please be careful when driving through the small towns as the speed limits change frequently and there are lots of well-funded police departments with very bored policemen.
• From Downtown Chicago and Northwest Suburbs (about a 2-1/2 hour drive from Chicago and a 2-hour drive from suburbs):
1. Take 90 West towards Rockford for 70 miles. After the McDonald’s Belvedere Oasis and Chrysler Plant, look for Highway 20 West.
2. Take Highway 20 West around Rockford and towards Freeport for 35 miles.
3. Right before Freeport, Highway 20 splits. Take 20 Business into Freeport (do not follow Highway 20 towards Galena).
4. You will be driving on Highway 20 Business for about 3 miles. At the Highway 26 intersection (there is a Wal-Mart on the corner), turn left onto Highway 26 South.
5. Drive on Highway 26 South for 7 miles. Turn right onto Highway 72 West. Drive on Highway 72 West for 10 miles.
6. At the 4-way stop sign intersection of Highway 72 and Highway 73, drive straight. The road changes and it is bumpy/hilly. You’ll pass a couple marinas on the right and will stay on this road for about 3 miles.
7. You’ll pass the “Fast Stop” gas station on the right-hand side and will soon pass signs that indicate trail crossings. When you see a sign that says “Dangerous Intersection,” you’ll turn right at this intersection onto Lake Carroll Blvd.
8. Take Lake Carroll Blvd. for .5 mile (you will pass a ski lift on the left) and turn right onto Spring Creek Drive.
9. Take Spring Creek Drive for about .4 mile and turn left onto Briarwood Dr. (Briarwood Dr. is a little hidden; if you reach a “Y” in the road, you’ve gone too far.)
10. Drive straight on Briarwood Drive for .3 of a mile and turn right onto Lakewood Court. You will pass some streets on the left but do not turn into them.
11. We live in the blue-gray house with white trim at the bottom of the street. There is a wooded area on the left, a brown house on the right and the lake is in the back. Our address is 12-119 Lakewood Court.
• From Western Chicago Suburbs (about a 2-hour drive):
1. Take I-88 (toll road) west to I-39 North (I-39 will be about 15 miles past the $.95 toll on 88).
2. Take I-39 North (towards Rockford) for about 5 miles and exit at Highway 64.
3. At the stop sign at the end of the exit ramp, turn left to head west on Highway 64.
4. You’ll be on Highway 64 for about 45 miles and will travel through various small towns including Oregon and Mt. Morris. BE VERY CAREFUL TO FOLLOW THE SIGNS FOR HIGHWAY 64 AS THERE WILL BE SEVERAL TURNS ALONG THE WAY.
5. In Lanark turn right onto Highway 73.
6. You will drive on Highway 73 for about 5 miles. At the 4-way stop sign of Highway 72 and Highway 73, turn left onto a bumpy/hilly road. You’ll pass a couple marinas on the right and will stay on this road for about 3 miles.
7. You’ll pass the “Fast Stop” gas station on the right-hand side and will soon pass signs that indicate trail crossings. When you see a sign that says “Dangerous Intersection,” you’ll turn right at this intersection onto Lake Carroll Blvd.
8. Take Lake Carroll Blvd. for .5 mile (you will pass a ski lift on the left) and turn right onto Spring Creek Drive.
9. Take Spring Creek Drive for about .4 mile and turn left onto Briarwood Dr. (Briarwood Dr. is a little hidden; if you reach a “Y” in the road, you’ve gone too far.)
10. Drive straight on Briarwood Drive for .3 of a mile and turn right onto Lakewood Court. You will pass some streets on the left but do not turn into them.
11. We live in the blue-gray house with white trim at the bottom of the street. There is a wooded area on the left, a brown house on the right and the lake is in the back. Our address is 12-119 Lakewood Court.
• From Dubuque (about a 70-minute drive):
1. Take Highway 20 east to Highway 73. You will go through Galena, Elizabeth, Woodbine and Stockton. Highway 73 is in a 45 mph zone about 10 miles east of Stockton.
2. Go south on 73 for about 13 miles. You may see signs for Lake Carroll, don’t turn into the main entrance where the biggest sign is located.
3. At the 4-way stop sign of Highway 72 and Highway 73, turn right and continue west on a bumpy/hilly road. You’ll pass a couple marinas on the right and will stay on this road for about 3 miles.
4. You’ll pass the “Fast Stop” gas station on the right-hand side and will soon pass signs that indicate trail crossings. When you see a sign that says “Dangerous Intersection,” you’ll turn right at this intersection onto Lake Carroll Blvd.
5. Take Lake Carroll Blvd. for .5 mile (you will pass a ski lift on the left) and turn right onto Spring Creek Drive.
6. Take Spring Creek Drive for about .4 mile and turn left onto Briarwood Dr. (Briarwood Dr. is a little hidden; if you reach a “Y” in the road, you’ve gone too far.)
7. Drive straight on Briarwood Drive for .3 of a mile and turn right onto Lakewood Court. You will pass some streets on the left but do not turn into them.
8. We live in the blue-gray house with white trim at the bottom of the street. There is a wooded area on the left, a brown house on the right and the lake is in the back. Our address is 12-119 Lakewood Court.
Welcome to Lake Carroll!
We hope you enjoy your stay at Lake Carroll! To make your time here relaxing and enjoyable, we've put together the following information and tips. Also included is a Departure Checklist to follow when you leave, ensuring that our home is ready for future visits and visitors. We appreciate you reading through this, and hope that you'll remember to sign our guestbook before you leave.
**Emergency Information**
- Fire/Police/Ambulance: 911
- Lake Carroll Security: 815-493-2599
- Lakehouse Address: 12-119 (section 12, lot 119) Lakewood Court, Lanark, IL 61046
- Lakehouse Phone Number: 815-493-6901
- Nearest Hospital: Freeport Memorial Hospital, 815-235-4131; take Highway 72 east to Highway 26; go north (left) on 26. In Freeport turn left on Stephenson St. (just past Lincoln St.); hospital is at 1045 W. Stephenson St. in Freeport, approximately 25 miles from Lake Carroll
- Poison Information Center: 608-262-3702
- The cabinet above the small desk in the kitchen has a small first aid kit and flashlight.
- Tim’s contact info:
- Cell: 312-543-2626 (usually on)
- Home: 773-281-2435
- Joe and Kelly’s contact info:
- Cell: 630-263-4608 (rarely on)
- Home: 630-769-9329
- If you notice anything unusual, such as a leak, weather damage, unsightly pests/rodents, vandalism, etc., please call us.
**Locks and Keys**
- We move the outdoor key lockbox from time to time and change the combination of the lockbox, so call Tim or Kelly to get the most recent information.
- To use the lockbox: Punch in the combination and push down the “Open” button to lift off the top of the lockbox. If you make a mistake when entering the combination, push “Clear” and start over.
- When using the key, unlock the deadbolt and doorknob locks as they’ll both be locked when you arrive.
- Please replace the key and lockbox immediately after use. To put the key back in the lockbox, push “Clear” and enter the combination again. Push the “Open” button down and the top of the lockbox should slide/lock back into place. Please return the lockbox to the super-secret hiding spot after use.
- BE CAREFUL…GETTING LOCKED OUT OF THE HOUSE IS EASY. When the bottom doorknob lock is in the “locked” position, you can easily close a locked door behind you. Remembering to put the lockbox back and knowing the combination would probably be a good idea. Otherwise, it’s okay to leave the doorknob unlocked during your stay as long as you use the deadbolt at night and when you’re not at home.
- There is an extra key to use if you leave the house; the key is located on the key holder on the left wall when you walk in the front door.
Kitchen
You’ll find the kitchen stocked with everything you’ll need. Items to note include:
- Cookie sheets and a pizza pan are in the drawer underneath the oven.
- Cooking oil and non-stick cooking spray are in the cabinet above the microwave.
- The pantry to the right of the small desk has paper towels and other dry goods.
- The cabinet above the small desk has a flashlight, sunscreen, bug spray and first aid kit; the drawer below the cabinet has Tim and Joe’s fishing licenses and a map of Lake Carroll.
- Coffee, filters and mugs are in the cabinet above the coffeepot.
- The upper cabinets to the right of the sink contain acrylic barware that is *not* dishwasher safe; please hand-wash these pieces if you use them.
- The closet to the left of the stove is the “garbage closet”; it includes a garbage can (naturally!), as well as garbage bags and broom/mop/bucket.
- The cabinet to the right of the refrigerator has some liquor, which you are welcome to enjoy. There are also matches on the top shelf in the small cardboard boxes.
- Check out the drawers below the liquor cabinet for plastic bags, plastic wrap, foil, hot pads, indoor/outdoor placemats and kitchen towels.
- Dish soap and dishwasher detergent are under the sink.
Air Conditioning and Heat
The thermostat is located in the hallway next to the Roth’s bedrooms. Note: It’s usually much cooler on the lower level, so be careful not to freeze yourself by setting the air conditioner too low.
Water - drink it at your own risk!
We have well water, so drinking frequently from the tap is not recommended. There’s bottled water in the garage refrigerator for drinking, and gallons of water under the sink for use in making juice, etc.
Cleaning Supplies – use of these items is not required!
- A vacuum can be found in the front hall closet.
- A broom/mop/bucket, as well as garbage bags, are in the closet to the left of the stove.
- Cleaning products are in the cabinet above the washer.
Laundry
- Feel free to use the washer and dryer just off of the kitchen.
- Laundry soap and dryer sheets are on top of the dryer. An extra bottle of detergent can usually be found in the cabinet above the washer.
- Be sure to empty the lint trap on top of the dryer between loads. If the dryer starts beeping, it means that the lint trap is too full.
Garage
- After emptying the kitchen wastebasket, please store garbage in the big can in the garage until you leave (the garage door is right by the washer and dryer).
The garage also has an extra refrigerator in which we keep beer, pop, beer, bottled water and beer…help yourself.
**Fish Wine Rack**
Due to the fact that Tim the Wine Connoisseur (a.k.a. Wine Snob) stores his premier collection here, using the wine in the fish rack is probably not a good idea unless you have written consent and/or verbal approval from Tim. You can, however, use the wine skewer, which is dangling off the left side of the fish rack. You can also help yourself to the less snobby wine collection in the cabinet to the right of the refrigerator.
**TV and Stereo**
Joe has quite an elaborate set-up, so using the TV and stereo is a little complicated. Please follow Joe’s step-by-step instructions for your viewing/listening pleasure:
- The hub of the system is the tuner on top of the entertainment cabinet to the right of the fireplace.
- Turn on power on the tuner and the TV and press the appropriate button on the tuner:
- For VCR: Push “Video 1”, making sure the VCR power is on.
- For stereo: Push “Video 2” and use tuner to find radio stations.
- For DVD/CD player: Push “DVD”, making sure the DVD power is on.
- For Direct TV: Push “TV/Sat”, making sure the Direct TV satellite box (on top of the tuner) is on.
- Please note that the speakers stay on all the time, so don’t worry that the little green lights on the speakers are always lit.
- If you’re having trouble with the system, there is a possibility that the light switch that controls the power in the room has been accidentally turned off. Next to the back door, make sure that the left switch is in the on position.
- Videotapes, DVDs and CDs are located in and around the entertainment cabinet.
- If you can figure out how to use the six remote controls for this system, please leave us instructions.
**Fireplace**
Until we have the chimney cleaned (probably Fall 2001), please do no use the fireplace.
**Deck and Grill**
- Fortunately the low railing on the deck allows for great lake views, but unfortunately it isn’t well positioned for safety. Please watch children when on the deck – especially when on deck chairs – as well as unruly, drunk visitors.
- To use the gas grill:
- Open top of grill.
- Turn on gas using valve on propane tank (turn valve to the left).
- Turn lower dial to “Start”.
- Push “Ignite” button (you may have to press this button a couple times). *If it fails to light, follow the directions on the top of the grill.*
- Once front burner ignites, turn other dials to “High” to heat up grill.
- The grill will be ready to use in about five minutes; turn burners to lower temperatures as appropriate after grill is heated.
- To turn grill off:
- Turn all dials to “Off” and turn off gas using valve on propane tank (turn valve to the right).
- After grill is cooled off, replace cover.
If the propane tank runs out and you need to replace it, you can do so at the “Fast Stop” which you passed on the way to our home; it’s open from 8 am – 10 pm.
**Lights and Fans**
Using the lights and fans is easy once you know a simple trick: Turn on the bottom switch labeled “Power” first, then you can turn on the “Light” and/or “Fan” switches. To control the speed of the fan, keep the “Fan” button pressed until it reaches the desired speed. Use “Power” again to turn off both the light and fan.
**Burglar Alarm**
We don’t use the burglar alarm, but if for some reason it happens to go off when you enter the house, use code number 8151 on the terminal next to the interior garage door (in laundry area) to disable it.
**Baby Essentials**
- A portable crib is in the closet of the queen guestroom.
- A high chair can be found in the kitchen or front hall closet.
- There are kids cups, plates and flatware in the kitchen; just look by the adult stuff and you’ll find them.
**Lower Level**
The stairs to the lower level are to the right of the garage door in the laundry area. On the lower level you’ll find two guestrooms, a pull-out couch, a guest bathroom, darts, pool, ping-pong and foosball. After you come down the stairs, there’s a closet to your right that has board games.
**Guest Bedrooms**
- The upstairs bedrooms are Tim’s (master), Joe and Kelly’s (with Canoe quilt) and Charlie and Audrey’s (adjacent to Joe and Kelly’s). Since we’re there most weekends, we don’t wash the sheets between every visit. So, to ensure a clean bed and room, please use the guestrooms on the lower level. One room has two twin beds (in the closet of this room, you’ll also find extra pillows and blankets). The other room has a queen-size bed with a portable dresser in the closet.
- Additionally, if needed, the golf couch is a sofa-sleeper; if the sheets aren’t on the bed, try the closet in the twin guest room.
- Please clean and replace sheets after use so that the beds are ready for our next visitors.
**Guest Bathroom**
- The guest bathroom is a little hidden…follow the little hallway next to the queen guestroom. It’s small, but should have everything you need.
- There are extra bath towels below the sink, and hand towels and wash cloths in the cabinet above the toilet. There are extra supplies in this cabinet, as well, including toilet paper, toothbrushes, toothpaste, soap and Tylenol.
- Please clean and replace towels after use.
**Extra Linens and Toiletries**
Extra bath towels and beach towels can be found in the upstairs linen closet next to Charlie and Audrey’s room. You’ll also find another necessity in this closet – extra toilet paper – as well as miscellaneous toiletries.
Neighbors
When looking at the lake, our neighbors to our immediate left are Dave and Karen, who are very friendly and know lots about the lake and area. Our neighbors to our right are Dave and Cindy.
Lake Swimming and Water Toys
• The lake has a steep drop-off just off the shoreline, reaching from 4 – 6 feet in depth around the dock area, so be careful to supervise kids outside at all times, especially when by the water and dock.
• Due to inconsistent water depth around the dock, please do not dive off the dock.
• You’ll find various floaties/life jackets in the lower level storage room. After you come downstairs, this room is the third door on the right side.
• Our kayak/paddles and paddleboat will either be on the dock, lower level patio or storage room. When using the kayak and paddleboat, you must have the same number of life jackets or floatable seat cushions as you do people in the kayak/boat. Please make sure that kids wear life jackets at all times when in the kayak/paddleboat.
• When using the kayak and paddleboat on the lake, please travel counter clockwise and stay within 100 feet of the shoreline. If you need to cross the lake, do so only in the middle of the lake using the most direct route possible.
Supplies
• It’s a good idea to bring whatever you want to eat and drink with you as there isn’t a big grocery store nearby. But if you do forget something, the “Fast Stop” that you passed on the way to our home is a great place to find lots of supplies, including food, gas, propane, beer/liquor, etc. It’s open 5:30 am every day and closes at 10 pm on the weekends (it closes earlier on the weekdays); there’s a brochure with the phone number on the refrigerator if you want to call before driving over there. It also has a little miniature golf course right behind it - $5 for adults and kids; children under 3 are free.
• If we run out of something while you’re visiting, please make a note of it on the notepad on the refrigerator so that we can replace it next time.
Outdoor Toys and Games
You’ll find badminton, horseshoes, bocce ball, basketball, football, baseball and many other games on the shelves in the garage. Please clean off (if needed) and return toys/games to the garage after use.
Fishing
• We fish off the dock but have not caught anything to be too proud of yet. The neighbors say that 50 yards off of our dock is a well-known fishing hole, but until we have a fishing boat, you’ll have to be creative in getting there.
• If you need to buy an Illinois fishing license, you can do so at the “Fast Stop.” For your convenience, Tim and Joe’s fishing licenses are located in the desk drawer in the kitchen; please remember to return them if you use them.
• Info and supplies: There’s information on the refrigerator that describes the fish limits. Adult and kid’s fishing gear is in the garage and there’s usually bait in the garage refrigerator (or it can be purchased at the Fast Stop).
Boat Rentals
Stop by the marinas that you passed on the way in to ask about boat rentals. We’ve heard that you can rent a pontoon boat from 8 am – 8 pm for about $150 per day. Other rentals may be available, as well.
Pool and Tennis Courts
- Our guests are welcome to use the pool and tennis courts, which is near the entrance where you came in. To get there from our house, take Lakewood Court (our street) to Briarwood Dr. Go left on Briarwood Dr. to Spring Creek Dr. and turn right. Take Spring Creek Dr. to Lake Carroll Rd. and turn left. You’ll see ski lifts on your right and shortly after you’ll turn right on Plum Tree Drive by the Isenhart Realty sign. This will lead to a lodge and campgrounds; a pool is to the right of the lodge and tennis courts are to the left of the lodge.
- To sign in at the pool, use our name and address – Roth, 12-119 – and this should get you in. You’ll need to know this as there will be a lifeguard on duty that will have you put this info on the register (on the table just inside the fence entrance).
- Some goofy pool rules:
- Hours are 12 noon – 8 pm…better late than never.
- Adult swim is 5 – 6 pm every day; kids are not allowed inside the entire pool area during adult swim, including the kiddie pool…go figure.
- Food and drink are not allowed in the pool.
- Tennis rackets and balls are on the shelves in the garage.
Golf
There’s a clubhouse and 18-hole golf course located near the main entrance of Lake Carroll. The green fees are around $20 (not including cart). Call 826-463-2808 to make a tee time. The course is about 10 minutes away. Drive back to the “Fast Stop” and instead of driving to the right in front of the Fast Stop, go to the left onto Lake Carroll Blvd., which is behind the Fast Stop. You’ll also pass some storage garages. Drive on Lake Carroll Blvd. for about 3 miles. At the stop sign, turn left onto Lake Carroll Blvd. (Lake Carroll Blvd. circles around the entire development.) The Clubhouse is about ¼ mile on the left.
Winter Sports
Lake Carroll is great for winter sports, including ice skating, fishing, skiing, snowmobiling and sledding:
- To ensure that the lake is frozen enough for use, please follow the instructions on the refrigerator about required temperatures for the appropriate freezing of lake water depths.
- Skiing/sledding is available near the entrance where you came in. To get there from our house, take Lakewood Court (our street) to Briarwood Dr. Go left on Briarwood Dr. to Spring Creek Dr. and turn right. Take Spring Creek Dr. to Lake Carroll Rd. and turn left. You’ll see ski lifts on your right and shortly after you’ll turn right on Plum Tree Drive by the Isenhart Realty sign. This will lead to the ski lodge.
Restaurants
We are certainly making no endorsements here, but there are a few nearby restaurants in Shannon (about 15 miles east on Highway 72), and in Lanark (about 5 miles east on Highway 72 and five miles south on Highway 73). In the summer months, they also serve food at the Lake Carroll Clubhouse. To get to the Clubhouse, drive back to the “Fast Stop” and instead of driving to the right in front of the Fast Stop, go to the left onto Lake Carroll Blvd., which is behind the Fast Stop. You’ll also pass some storage garages. Drive on Lake Carroll Blvd. for about 3 miles. At the stop sign, turn left onto Lake Carroll Blvd. (Lake Carroll Blvd. circles around the entire development.) The Clubhouse is about ¼ mile on the left.
Garbage Removal
Since Joe and Tim were too cheap to pay for Lake Carroll’s very pricey garbage removal, you’ll have to take your trash with you when you leave. We’ve been using the dumpster by the campgrounds, and haven’t been caught yet! The dumpster is located next to the lodge by the pool. To get there from our
house, take Lakewood Court (our street) to Briarwood Dr. Go left on Briarwood Dr. to Spring Creek Dr. and turn right. Take Spring Creek Dr. to Lake Carroll Rd. and turn left. You’ll see ski lifts on your right and shortly after you’ll turn right on Plum Tree Drive by the Isenhart Realty sign. Pull up just behind the lodge, throw your bags on the heap and drive off quickly.
**Standard Disclaimer**
This lakehouse is exclusively for use by friends, relatives, acquaintances, friends of friends, friends of relatives and friends of acquaintances. The owners do not accept responsibility for hangovers (although Extra Strength Tylenol is provided free of charge; see “Guest Bathroom” section above), injuries or bad behavior. If you get yelled at by the neighbors or authorities, please don’t associate your stupidity with us (we already have enough of that in our family). If you get hurt, please don’t sue us.
Seriously, though, we want the house and lake to remain a safe, fun place to visit. Please be careful when visiting, especially with small children.
Departure Checklist
Yes, we know we’re a bit anal, but we see this checklist as a way to ensure that the house is ready for our upcoming visits and visitors. Make fun of us if you’d like.
- Sheets and towels cleaned and replaced?
- Air conditioning turned off, or heat turned down to 60 degrees?
- Oven and burners off?
- Coffee maker off and unplugged?
- Dirty dishes cleaned/returned to cabinets and dishwasher emptied?
- Perishable food removed from refrigerator and pantry?
- All lights off?
- All fans off?
- All toilets quiet and not running water into their tanks?
- Grill burners off and gas valve on propane tank turned all the way to the right?
- Deck umbrella down and tied closed?
- TV, satellite, stereo tuner, VCR and DVD turned off?
- Fireplace flu closed and fireplace gas turned off?
- Windows closed and locked?
- Kayak/paddleboat returned to dock, lower level patio or storage room and turned upside down?
- Floaties returned to basement storage room?
- Garbage removed from inside garbage cans and garage?
- Guestbook signed?
- Garage door closed?
- Doors closed, locked and dead-bolted?
- Key lockbox returned to super-secret hiding spot?
Thanks!
Tim, Joe and Kelly |
Resolution No.: 2022-D050
Title: Reunion of the Diocese of The Episcopal Church in North Texas with The Episcopal Diocese of Texas
Proposer: The Rev. Canon Christine Faulstich
Topic: Dioceses
Directly Related: (Attached)
2018-A114 Amend Canon I.10 [Of New Dioceses]
Indirectly Related: (Available in the Acts of Convention database, searchable by resolution number)
None
In preparing this report, the Archives researched the resolutions in the Acts of Convention database for the period 1973 through 2018, selecting “direct” resolutions that have a substantive bearing on the proposed legislation. The “direct” resolutions are attached and “indirect” resolutions are available in the Acts of Convention database. Committee members who require other research assistance should contact the Archives through the Research Request Form or call 800-525-9329.
The Acts of General Convention 1973-2018 * Research report provided by The Archives of the Episcopal Church
D050 - Reunion of the Diocese of The Episcopal Church in North Texas with The Episcopal Diocese of Texas
Final Status: Not Yet Finalized
Proposed by: The Rev. Canon Christine Faulstich
Endorsed by: The Rev. Bradley Dyche, Ms. Katie Sherrod
Requests New Interim Body: No
Amends C&C or Rules of Order: No
Has Budget Implications: No
Cost:
HiA: HD
Legislative Committee Currently Assigned: 04 - Governance & Structure
Completion Status: Incomplete
Latest House Action: N/A
Supporting Documents: No
Resolution Text
Resolved, the House of _______________ concurring,
That the 80th General Convention approve the reunion of the Diocese of The Episcopal Church in North Texas with and into The Episcopal Diocese of Texas pursuant to Canon I.10.6.
Explanation
Canon I.10.6 allows a diocese and another diocese that was formed by division therefrom to be reunited, subject to the approval of their respective diocesan conventions and General Convention.
When The Episcopal Diocese of Texas (EDOT) was created in 1849, its territory included the entire state. Over time, as the population of the state grew, EDOT ceded parts of its territory to create new dioceses. For example, it ceded the north central part of the state to create the Diocese of Dallas. In 1982, the Diocese of Dallas in turn ceded its western 24 counties to create the Diocese of Fort Worth.
In 2008, the Diocese of Fort Worth was torn apart when, led by a former bishop, the Diocese and many of its congregations left the Episcopal Church and joined the Anglican Church in North America (ACNA). In response, Presiding Bishop Katharine Jefferts Schori
convened a convention at which the congregations who wished to remain in the Episcopal Church as the Episcopal Diocese of Fort Worth elected a Provisional Bishop and populated other diocesan bodies. Lengthy litigation between the two groups ensued over the issue of which group owned the name “Episcopal Diocese of Fort Worth” and the property of the diocese and its parishes. In 2020 the Texas Supreme Court ruled in favor of the ACNA group and awarded them the name and title to almost all the diocesan and parish properties. When the Supreme Court of the United States declined to hear the appeal of the Episcopalian, the decision of the Texas Supreme Court went into effect.
As a result, the diocese that remained in The Episcopal Church was renamed “The Episcopal Church in North Texas” (ECNTX). In its 24 North Texas counties are 13 congregations, many of which have found new places to worship. The diocese remains headquartered in Fort Worth, and its Provisional Bishop is now The Rt. Rev. J. Scott Mayer, Bishop of the Diocese of Northwest Texas.
After the conclusion of most of the litigation, Bishop Mayer convened a group of clergy and lay leaders to identify and evaluate options for the future of ECNTX. This discernment group eventually recommended to their Standing Committee that ECNTX reunite with and become a part of EDOT.
EDOT, headquartered in Houston, consists of 153 congregations in 57 counties in east, central, and south Texas, 15 college ministries, 58 Episcopal schools, Camp Allen, several social service agencies, and five foundations dedicated to the health and growth of the Church. It is also home to the Seminary of the Southwest. Its Bishop is The Rt. Rev. C. Andrew Doyle.
After extensive discussion, in May the two dioceses entered into an Agreement of Reunion, which was approved by both Bishops and their respective Standing Committees. In June the Agreement of Reunion was then approved by the conventions of both dioceses.
Resolved, That the 79th General Convention amend Canon I.10 as follows:
Sec. 1. Whenever a new Diocese shall be proposed to be formed within the limits of any Diocese, or by the junction of two (2) or more Dioceses, or parts of Dioceses, and such action shall have been ratified by the General Convention, the Bishop of the Diocese within the limits of which a Diocese is formed, or in case of the junction of two or more Dioceses, or parts of Dioceses, the senior Bishop by consecration, shall thereupon call the Primary Convention of the new Diocese, for the purpose of enabling it to organize, and shall fix the time and place of holding the same, such place being within the territorial limits of the new Diocese; the Ecclesiastical Authorities and the Standing Committees of the Dioceses involved in the proposed new diocese shall submit for approval to the Conventions of each Diocese involved a joint agreement of union setting forth their agreements, including the manner of determining the Bishop Diocesan and other Bishops (if any), the provisions of the Constitution and Canons of the new Diocese, and such other matters as may be necessary or proper. Upon approval by the Conventions of each of the involved Dioceses, the joint agreement of union shall be submitted for ratification by the General Convention no less than ninety (90) days prior to the first legislative day of the next meeting of the General Convention.
Sec. 2. Promptly after ratification by the General Convention, the Ecclesiastical Authority of the new Diocese, as set forth in the joint agreement of union, shall call the Primary Convention of the new Diocese, for the purpose of enabling it to organize, and shall fix the time and place of holding the same, such place being within the territorial limits of the new Diocese. In case there should be no Bishop who can call such Primary Convention, pursuant to the foregoing provision, then the duty of calling such Convention for the purpose of organizing and of fixing the time and place of its meeting, shall be vested in the Standing Committee of the Diocese within the limits of which the new one is erected, or in the Standing Committee of the oldest of the Dioceses by the junction of which, or of parts of which, the new Diocese may be formed. And such Standing Committee shall make the call immediately after ratification of the General Convention.
Sec. 3. Whenever one Diocese is about to be divided into two Dioceses, the Convention of such Diocese shall declare which portion thereof is to be in the new Diocese, and shall make the same known to the General Convention before the ratification of such division.
Sec. 4. Whenever a new Diocese shall have organized in Primary Convention in accordance with the provisions of the Constitution and Canons in such case
made and provided, and in the manner prescribed in the previous Sections of this Canon, and shall have chosen a name and acceded to the Constitution of the General Convention in accordance with Article V, Section 1 of the Constitution, and shall have laid before the Executive Council certified copies of the Constitution adopted at its Primary Convention, and the proceedings preparatory to the formation of the proposed new Diocese, such new Diocese shall thereupon be admitted into union with the General Convention.
Sec. 5. In the event of the erection of an Area Mission into a Diocese of this Church, as provided in Article V, See. Section 1, the Convocation of the said Area Mission shall be entitled to elect Deputies to the succeeding General Convention, and also to elect a Bishop. The jurisdiction previously assigned to the Bishop in the Area Mission shall be terminated upon the admission of the new Diocese.
Sec. 6 (a) When a Diocese, and another Diocese one or more other Dioceses which has been that were formed either by division therefrom or by erection into a Diocese or a Missionary Diocese formed by division therefrom, shall desire to be reunited into one Diocese, the proposed reunion must be initiated by the approval of the Conventions of the involved Dioceses of a joint agreement of union setting forth their agreements, including the manner of determining the Bishop Diocesan and other Bishops (if any), provisions of the Constitution and Canons of the new Diocese, and such other matters as may be necessary or proper. A mutual agreement between the Conventions of the two Dioceses, consented to by the Ecclesiastical Authority of each Diocese. If the said agreement of the Dioceses is made and the consents of their Conventions are given more than three months before the next meeting of the General Convention, the fact of the agreement and consents shall be certified by the Ecclesiastical Authority and the Secretary of the Convention of each involved Diocese to all the Bishops of the Church having jurisdiction and to the Standing Committees of all the Dioceses; and when the consents of a majority of such Bishops and of a majority of the Standing Committees to the proposed reunion shall have been received, the facts shall be similarly certified to the Secretary of the House of Deputies of the General Convention, and thereupon the reunion shall be considered complete. But if the agreement is made and the consents given within three months of the next meeting of the General Convention, the facts shall be certified instead to the Secretary of the House of Deputies, who shall lay them before the two Houses; and the reunion shall be deemed to be complete when it shall have been sanctioned approved by a majority vote in the House of Bishops, and in the House of Deputies voting by orders.
(b) The Bishop of the parent Diocese shall be the Bishop, and the Bishop of the junior Diocese shall be the Bishop Coadjutor, of the reunited Diocese; but if there be a vacancy in the Episcopate of either Diocese, the Bishop of the other Diocese shall be the Bishop, and the Bishop Coadjutor if there be one shall be the Bishop Coadjutor, of the reunited Diocese.
(c) Sec. 7 When the union of two (2) or more Dioceses or portions of Dioceses or the reunion of the two (2) or more Dioceses shall have been completed, the facts shall be certified to the Presiding Bishop and to the Secretary of the House
of Deputies. Thereupon the Presiding Bishop shall notify the Secretary of the House of Bishops of any alteration in the status or style of the Bishop or Bishops concerned, and the Secretary of the House of Deputies shall strike the name of the juniorany Diocese that will cease to exist or is being renamed from the roll of Dioceses in union with the General Convention and, if appropriate, amend the name of the newly united Diocese on the roll of Dioceses in union with the General Convention.
Citation: General Convention, *Journal of the General Convention of...The Episcopal Church, Austin, 2018* (New York: General Convention, 2018), pp. 1114-1116. |
Heavy flavour and exotic production at LHCb
Jianqiao Wang\textsuperscript{a,b,*}
\textsuperscript{b}On behalf of the LHCb collaboration.
\textsuperscript{a}Center for High Energy Physics, Tsinghua University, Beijing, China.
E-mail: firstname.lastname@example.org
These proceedings summarise recent LHCb results on heavy-flavour and exotic production in high-multiplicity $pp$ and heavy-ion collisions, covering the measurement of $B^0_s/B^0$ production ratio, the study of the prompt $D^0$ nuclear modification factor, the first measurement of $\Lambda_c^+/D^0$ production ratio in peripheral PbPb collisions in the forward rapidity region, and the first measurement of prompt $\chi_{c1}(3872)$ production in $p$Pb collisions. These results indicate that other possible initial- or final-state effects beyond current theoretical models may exist in heavy-ion collisions, which is important for the investigation of cold nuclear matter effects. More works are in progress to further investigate these effects.
1. Introduction
Heavy quarks are produced in the early-stage hard processes of ultra-relativistic heavy-ion collisions and are strongly affected by the presence of quark-gluon plasma (QGP), as well as by cold nuclear matter effects. Measurements on heavy flavour provide unique opportunities to study these nuclear matter effects in heavy-ion collisions and may help to better understand high-multiplicity $pp$ collisions. In initial stages of heavy-ion collisions, the modified parton distribution function in the nuclear environment (nPDF) [1, 2], or the possible existence of a colour glass condensate (CGC) [3, 4], as well as initial and final-state energy loss [5–7], will lead to modification of heavy-quark production cross-sections with respect to extrapolation from $pp$ collisions. Though the fragmentation mechanism [8, 9] dominates the hadron quark hadronisation in $e^+e^-$ collisions, recent measurements have shown that charm quark hadronisation differs between $e^+e^-$ and $pp$ collisions [10]. The quark coalescence mechanism [11] can become important for heavy quarks when the number of quarks increases in the system. It will lead to enhanced production of baryons, exotic states and hadrons with strangeness, especially in heavy-ion collisions.
The LHCb experiment is designed for precision measurements of beauty and charm hadrons. A detailed description of the LHCb detector and operation performances can be found in Refs. [12, 13]. It has become a generally proposed detector for analysing $pp$, $pPb$ and PbPb data. The asymmetric $pPb$ data include two different beam configurations with final state particles in either proton (forward, positive rapidity) or lead beam (backward, negative rapidity) direction pointing into the LHCb acceptance from the interaction region. These proceedings present the measurements of $B_s^0/B^0$ production ratio in $pp$ at $\sqrt{s} = 13$ TeV [14], prompt $D^0$ production [15] and $\chi_{c1}(3872)/\psi(2S)$ production ratio [16] in $pPb$ at $\sqrt{s_{NN}} = 8.16$ TeV and the prompt $\Lambda_c^+/D^0$ ratio in peripheral PbPb collisions [17]. These measurements help to investigate the initial- and final-state effects in heavy flavour production in heavy-ion collisions.
2. $B_s^0/B^0$ production ratio in $pp$ at $\sqrt{s} = 13$ TeV
The production ratio of $B_s^0$ mesons relative to $B^0$ mesons is measured as a function of event multiplicity [14], using a sample of $pp$ collisions collected at a centre-of-mass energy of $\sqrt{s} = 13$ TeV, corresponding to an integrated luminosity of 5.4 fb$^{-1}$. Both particles are reconstructed via the common final states $J/\psi \pi^+\pi^-$. The multiplicity dependence $\sigma_{B_s^0}/\sigma_{B^0}$ is shown in three different intervals of the $B$ meson $p_T$ in Fig. 1. In the lowest $p_T$ interval, $0 < p_T < 6$ GeV/$c$, the $\sigma_{B_s^0}/\sigma_{B^0}$ ratio is consistent with results from $e^-e^+$ collisions while increases with multiplicity. The significance of the slope from linear fit differs from zero by 3.4 standard deviations. The measurements in higher $p_T$ intervals are consistent with data from $e^-e^+$ and display no multiplicity dependence. These measurements are qualitatively consistent with expectations based on the emergence, in such high-multiplicity environment, of quark coalescence as an additional hadronisation mechanism, rather than fragmentation [8, 9] alone.
3. Prompt $D^0$ production in $pPb$ at $\sqrt{s_{NN}} = 8.16$ TeV
The $D^0$ production cross-section is measured in $pPb$ collisions at a center-of-mass energy per nucleon pair of $\sqrt{s_{NN}} = 8.16$ TeV [15]. The prompt $D^0$ candidates are distinguished from those from
beauty decays by the distribution of the charm-hadron impact parameter. The nuclear modification factor $R_{pPb}$ is calculated by comparing the cross-section with that in $pp$ collisions. The data points, as well as LHCb results at $\sqrt{s_{NN}} = 5.02$ TeV [18] and theoretical calculations [1–4] are displayed in Fig. 2. A significant suppression compared to unity is observed in the forward rapidity, which is consistent with the LHCb 5 TeV result and theoretical calculations. The $D^0$ $R_{pPb}$ shows a slight suppression compared to nPDF calculations at high-$p_T$ in the backward rapidity region, indicating possible initial- or final-state effects other than anti-shadowing.
### 4. Prompt $\Lambda_c^+/D^0$ ratio in peripheral PbPb collisions
The baryon-to-meson production ratios are excellent probes to study hadronisation in heavy-ion physics. This measurement gives the first result on $\Lambda_c^+/D^0$ production ratio $R_{\Lambda_c^+/D^0}$ in peripheral PbPb collisions in the forward rapidity region [17]. The mean value of the ratio $\langle R_{\Lambda_c^+/D^0} \rangle$ is about 0.27 and the results are given as a function of $p_T$ and rapidity in Fig. 3, in a centrality region of 65 – 90%. The data are compatible with Pythia 8 [19] calculation incorporating colour
reconnection mechanism for $p_T > 5$ GeV/$c$. However, discrepancies are observed with predictions from the statistical hadronisation model [20]. The new result also indicates a dependence of $R_{\Lambda_c^+ / D^0}$ on rapidity, where the suppression with respect to Pythia 8 originates from the difference in $p_T < 3$ GeV/$c$.
### 5. Prompt $\chi_{c1}(3872)/\psi(2S)$ ratio in $pPb$
There is still no consensus on the structure of the exotic hadron $\chi_{c1}(3872)$ states. The behaviour of these exotic states in heavy-ion collisions provides new ways to probe their properties. This is the first measurement of $\chi_{c1}(3872)$ states in $pPb$ collisions [16]. The production of $\chi_{c1}(3872)$ states is compared to that of $\psi(2S)$ mesons to study the final-state effects, while initial-state effects largely cancel in the $\chi_{c1}(3872)/\psi(2S)$ ratio. The ratio in both forward and backward rapidity regions are compared to that in $pp$ collisions [21] and PbPb collision [22], as shown in Fig. 4. A rising trend as a function of the collision system size is found, which is different from the observed decreasing trend versus multiplicity in $pp$ collisions [21]. This may indicate that the hadronic densities achieved in the forward and backward $pPb$ collisions allow the quark coalescence to become the dominant mechanism affecting $\chi_{c1}$ production. Future measurements of the nuclear modification factor of $\chi_{c1}(3872)$ and $\psi(2S)$, which are in progress at LHCb, can clarify these effects.
### 6. Summary
The LHCb detector has strong capabilities to study heavy flavour particles and exotic states in $pp$, $pPb$ and PbPb collisions. Evidence of enhancement of the $B_s^0/B^0$ production ratio with respect to event multiplicity is found at low-$p_T$ in $pp$ collisions, qualitatively consistent with expectations from the quark coalescence model. The $D^0$ nuclear modification factor is measured in $pPb$ collisions at $\sqrt{s_{NN}} = 8.16$ TeV. The result is consistent with theoretical calculations in forward rapidity regions, while a stronger suppression compared to nPDF calculations is seen at high-$p_T$ in backward rapidity regions. The measurement of $R_{\Lambda_c^+ / D^0}$ is first performed in forward regions. The points show general agreement with Pythia 8 calculation incorporating colour reconnection mechanism for $p_T > 5$ GeV/$c$, while the statistical hadronisation model does not reproduce the data. The $\chi_{c1}(3872)/\psi(2S)$ production ratio is measured in both $pPb$ and $PbPb$ collisions for the first time.
The ratio is compared across different collision systems and an increasing trend versus system size is observed, which indicates that quark coalescence may dominate the $\chi_{c1}(3872)$ production in $p$Pb collisions. More measurements are in progress with LHCb data to further investigate the nuclear matter effects and hadronisation mechanisms of heavy quark productions in heavy-ion collisions.
References
[1] K. J. Eskola, P. Paakkinen, H. Paukkunen, and C. A. Salgado, EPPS16: Nuclear parton distributions with LHC data, Eur. Phys. J. C77 (2017) 163, arXiv:1612.05741.
[2] K. Kovarik et al., nCTEQ15 - Global analysis of nuclear parton distributions with uncertainties in the CTEQ framework, Phys. Rev. D93 (2016) 085037, arXiv:1509.00792.
[3] B. Ducloué, T. Lappi, and H. Mäntysaari, Forward J/$\psi$ and D meson nuclear suppression at the LHC, Nucl. Part. Phys. Proc. 289-290 (2017) 309, arXiv:1612.04585.
[4] Y.-Q. Ma, P. Tribedy, R. Venugopalan, and K. Watanabe, Event engineering studies for heavy flavor production and hadronization in high multiplicity hadron-hadron and hadron-nucleus collisions, Phys. Rev. D 98 (2018) 074025, arXiv:1803.11093.
[5] I. Vitev, Non-Abelian energy loss in cold nuclear matter, Phys. Rev. C75 (2007) 064906, arXiv:hep-ph/0703002.
[6] Z.-B. Kang et al., Multiple scattering effects on heavy meson production in p+A collisions at backward rapidity, Phys. Lett. B 740 (2015) 23, arXiv:1409.2494.
[7] F. Arleo, G. Jackson, and S. Peigné, Impact of fully coherent energy loss on heavy meson production in pA collisions, JHEP 01 (2022) 164, arXiv:2107.05871.
[8] B. Andersson, G. Gustafson, G. Ingelman, and T. Sjostrand, *Parton Fragmentation and String Dynamics*, Phys. Rept. 97 (1983) 31.
[9] B. R. Webber, *A QCD Model for Jet Fragmentation Including Soft Gluon Interference*, Nucl. Phys. B 238 (1984) 492.
[10] ALICE collaboration, S. Acharya et al., *Charm-quark fragmentation fractions and production cross section at midrapidity in pp collisions at the LHC*, Phys. Rev. D 105 (2022) L011103, arXiv:2105.06335.
[11] R. J. Fries, B. Muller, C. Nonaka, and S. A. Bass, *Hadronization in heavy ion collisions: Recombination and fragmentation of partons*, Phys. Rev. Lett. 90 (2003) 202303, arXiv:nucl-th/0301087.
[12] LHCb collaboration, A. A. Alves Jr. et al., *The LHCb detector at the LHC*, JINST 3 (2008) S08005.
[13] LHCb collaboration, R. Aaij et al., *LHCb detector performance*, Int. J. Mod. Phys. A30 (2015) 1530022, arXiv:1412.6352.
[14] LHCb collaboration, *Evidence for modification of b quark hadronization in high-multiplicity pp collisions at $\sqrt{s} = 13$ TeV*, arXiv:2204.13042.
[15] LHCb collaboration, *Measurement of the prompt $D^0$ nuclear modification factor in pPb collisions at $\sqrt{s_{NN}} = 8.16$ TeV*, arXiv:2205.03936.
[16] LHCb collaboration, *Modification of $\chi_{c1}(3872)$ and $\psi(2S)$ production in pPb collisions at $\sqrt{s_{NN}} = 8.16$ TeV*, .
[17] LHCb collaboration, *Measurement of the $\Lambda_c^+ \to D^0$ production cross-section ratio in peripheral PbPb collisions*, arXiv:2210.06939.
[18] LHCb collaboration, R. Aaij et al., *Study of prompt $D^0$ meson production in pPb collisions at $\sqrt{s_{NN}} = 5$ TeV*, JHEP 10 (2017) 090, arXiv:1707.02750.
[19] J. R. Christiansen and P. Z. Skands, *String Formation Beyond Leading Colour*, JHEP 08 (2015) 003, arXiv:1505.01681.
[20] M. He and R. Rapp, *Hadronization and Charm-Hadron Ratios in Heavy-Ion Collisions*, Phys. Rev. Lett. 124 (2020) 042301, arXiv:1905.09216.
[21] LHCb collaboration, R. Aaij et al., *Observation of Multiplicity Dependent Prompt $\chi_{c1}(3872)$ and $\psi(2S)$ Production in pp Collisions*, Phys. Rev. Lett. 126 (2021) 092001, arXiv:2009.06619.
[22] CMS collaboration, A. M. Sirunyan et al., *Evidence for X(3872) in Pb-Pb Collisions and Studies of its Prompt Production at $\sqrt{s_{NN}}=5.02$ TeV*, Phys. Rev. Lett. 128 (2022) 032001, arXiv:2102.13048. |
Rosario v. Rockefeller and Kusper v. Pontikes - Voters and other Strangers
Neal Taslitz
Follow this and additional works at: http://via.library.depaul.edu/law-review
Recommended Citation
Neal Taslitz, Rosario v. Rockefeller and Kusper v. Pontikes - Voters and other Strangers, 23 DePaul L. Rev. 838 (1974)
Available at: http://via.library.depaul.edu/law-review/vol23/iss2/13
This Case Notes is brought to you for free and open access by the College of Law at Via Sapientiae. It has been accepted for inclusion in DePaul Law Review by an authorized administrator of Via Sapientiae. For more information, please contact email@example.com, firstname.lastname@example.org.
ROSARIO v. ROCKEFELLER AND
KUSPER v. PONTIKES—
VOTERS AND OTHER STRANGERS
In Rosario v. Rockefeller,¹ petitioners² brought suit in federal court for declaratory relief³ alleging that the New York Election Law⁴ had unconstitutionally deprived them of their right to vote in New York's 1972 presidential primary and had also abridged their freedom to associate with the political party of their choice. The avowed purpose of the challenged section of this law was to prevent political party "raiding," whereby voters in sympathy with one party enrolled themselves in an opposing political party with the intention of voting for the weakest candidate in
---
1. 410 U.S. 752 (1973) [hereinafter cited in text as Rosario].
2. Petitioner, Pedro J. Rosario, was joined by William J. Freedman, Karen Lee Gottesman, and Steven Eisner. All of the petitioners registered to vote for the first time in December of 1971, and at that time completed an enrollment blank designating their affiliation with the Democratic Party. None of the petitioners were permitted to vote in the Democratic primary held in New York on June 20, 1972. Brief for Petitioner at 8 n.4, Rosario v. Rockefeller, 410 U.S. 752 (1973).
3. Petitioners sued under 42 U.S.C. § 1983 and 28 U.S.C. § 1343(3). 42 U.S.C. § 1983 (1970) provides:
Every person who, under color of any statute, ordinance, regulation, custom, or usage of any State or Territory, subjects, or causes to be subjected, any citizen of the United States or any other person within the jurisdiction thereof to the deprivation of any rights, privileges, or immunities secured by the Constitution and laws, shall be liable to the party injured in an action at law, suit in equity, or other proper proceeding for redress.
28 U.S.C. § 1343(3) (1970) grants original jurisdiction to the district courts in civil actions under 42 U.S.C. § 1983.
4. N.Y. ELECTION LAW § 186 (McKinney 1964) provides in pertinent part:
All enrollment blanks contained in the enrollment box shall remain in such box, and the box shall not be opened nor shall any of the blanks be removed therefrom until the Tuesday following the day of the general election in that year. Such box shall then be opened by the board of elections and the blanks contained therein shall be removed therefrom by the board, and the names of the party designated by each voter under the declaration, provided such party continues to be a party, as defined in this law shall be entered by the board, opposite the name of such voter in the appropriate column of the two copies of the register containing enrollment numbers for the election district in which such voter resides. . . . When all of the enrollments shall be transcribed from the blanks to the register, the board of elections shall make a certificate upon the form printed in such registers, to the effect that it has correctly and properly transcribed the enrollment indicated on the blank of each voter to such registers. Such enrollment shall be complete before the succeeding first day of February in each year.
the latter party's primary. In effect, the statute in question prohibited otherwise qualified voters from participating in a primary election unless their enrollments in a political party were completed at least thirty days prior to the general election preceding the primary or about eight months prior to the June primary.\(^5\)
In an unpublished opinion, Chief Judge Jacob Mishler of the United States District Court for the Eastern District of New York, declared section 186 of the New York Election Law unconstitutional in violation of the petitioners' rights to vote in a primary election, and in violation of their first and fourteenth amendment rights to associate freely with the political party of their choice.\(^6\) The Court of Appeals for the Second
---
5. The cut-off date for enrollment provided by section 186 occurs almost eight months before a presidential primary (held in June) and eleven months before a non-presidential primary (held in September). N.Y. ELECTION LAW § 187 (McKinney 1964) provides exceptions for certain persons who failed to meet the cut-off date prescribed by section 186. Section 187 includes, in pertinent part:
Application for special enrollment, transfer or correction of enrollment.
1. At any time after January first and before the thirtieth day preceding the next fall primary, except during the thirty days preceding a spring primary, and except on the day of a primary, a voter may enroll with a party, transfer his enrollment after moving within a county, and under certain circumstances, correct his enrollment, as hereinafter in this section provided.
2. A voter may enroll with a party if he did not enroll on the day of the annual enrollment (a) because he became of age after the preceding general election, or (b) because he was naturalized subsequent to ninety days prior to the preceding general election, or (c) because he did not have the necessary residential qualifications as provided by section one hundred fifty, to enable him to enroll in the preceding year, or (d) because of being or having been at all previous times for enrollment a member of the armed forces of the United States as defined in section three hundred three, or (e) because of being the spouse, child or parent of such member of the armed forces and being absent from his or her county of residence at all previous times for enrollment by reason of accompanying or being with such member of the armed forces, or (f) because he was an inmate or patient of a veterans' bureau hospital located outside the state of New York at all previous times for enrollment, or the spouse, parents or child of such inmate or patient accompanying or being with such inmate or patient at such times, or (g) because he was incapacitated by illness during the previous enrollment period thereby preventing him from enrolling.
6. Special enrollment under the classification set forth in clause (c) of subdivision two is hereby expressly limited to a voter otherwise qualified, who did not have the qualifications to vote at the previous general election and such special enrollment is restricted to the same county the voter resided in at the preceding year.
6. Rosario v. Rockefeller, Civil No. 71-1573 (E.D.N.Y. Feb. 10, 1972). In this decision Judge Mishler also concluded that section 186 established a durational residence requirement, and as such, was in violation of the 1970 amendments to the VotCircuit reversed the district court's opinion, holding the statute constitutional as serving a "compelling state interest." 7 Eleven months later, in a five-to-four decision, the Supreme Court affirmed the appellate court's opinion. 8 In its conclusion the Court stated:
New York did not prohibit the petitioners from voting in the 1972 primary election or from associating with the political party of their choice. It merely imposed a legitimate time limitation on their enrollment, which they chose to disregard. 9
The Supreme Court viewed the statute as meeting the requirements of the "compelling state interest" test on the grounds that it preserved the integrity of New York's closed primary electoral process. 10 The decision is unique in that it marked the first time that the Supreme Court had dealt with a statute designed to prevent voters from "raiding" another political party. 11 The Court's opinion is significant in that it examines a statute aimed at preventing "raiding" and sanctions it by applying the "compelling state interest" test. However, the decision is narrow since it is applicable only to statutory schemes that use a means which does not absolutely deprive a voter of his constitutional rights. Thus, while recognizing the "compelling state interest" in the prevention of "raiding," the Court has only sanctioned one particular kind of statute dealing with this problem.
This note is divided into three sections. The first section briefly sur-
7. Rosario v. Rockefeller, 458 F.2d 649 (2d Cir. 1972), aff'd, 410 U.S. 752 (1973).
8. Stewart, J., delivered the majority opinion of the Court, in which Burger, C.J., and White, Blackmun, and Rehnquist, JJ., joined. Powell, J., wrote the dissenting opinion, in which Douglas, Brennan, and Marshall, J.J. joined.
9. 410 U.S. at 762.
10. Cf. Dunn v. Blumstein, 405 U.S. 330, 345 (1972); Bullock v. Carter, 405 U.S. 134, 145 (1972).
11. Jordan v. Meisser, 405 U.S. 907 (1972) (mem.), dealt with the same issue as that found in Rosario but was dismissed for want of a substantial federal question on the ground that the plaintiff could have enrolled after the cut-off date under an exception to the challenged statute. See Rosario v. Rockefeller, 458 F.2d 649, 654 n.6 (2d Cir. 1972). Two other cases the Supreme Court has decided which have been cited as supporting statutes designed to prevent "raiding" are Jenness v. Fortson, 403 U.S. 431 (1971) and Lippitt v. Cipollone, 404 U.S. 1032 (1972), aff'd mem., 337 F. Supp. 1405 (N.D. Ohio 1971). However, both of these cases dealt with restrictions placed upon a person's right to run as a candidate for political office rather than upon a person's right to vote in a primary election.
veys the judicial development of a citizen's right to vote in both general and primary elections. The second section compares and contrasts the Supreme Court's decision in *Rosario* with recent cases involving challenges to similar "raiding" statutes. Particular emphasis will be placed on the decision of the United States Supreme Court in *Kusper v. Pontikes*, overturning an Illinois election law. The third section contains a critical analysis of the Supreme Court's opinion in *Rosario*. This analysis will include the author's viewpoint regarding the significance of *Pontikes* in relation to *Rosario*. Finally, the conclusions of both cases taken together will be examined to determine the possible effects they may have on future decisions involving statutes designed to protect the integrity of the American political process.
**THE CITIZEN'S RIGHT TO VOTE IN GENERAL AND PRIMARY ELECTIONS**
The right to vote, and the Supreme Court's interest in the protection of that right against state abridgement, occupies a fundamental position in this country's democratic form of government.
Nearly one hundred years ago, in *Ex parte Siebold*, the Supreme Court held that Congress may enact statutes which protect every citizen against state interference with the right to vote. In *Yick Wo v. Hopkins*, referring to "the political franchise of voting," the Supreme Court stated that it was a "fundamental political right, because preservative of all rights."
---
12. *Kusper v. Pontikes*, 42 U.S.L.W. 4003 (U.S. Nov. 19, 1973) [hereinafter cited in text as *Pontikes*].
13. *See, e.g.*, Dunn v. Blumstein, 405 U.S. 330 (1972) (durational residence requirement for voting found unconstitutional); Kramer v. Union Free School District No. 15, 395 U.S. 621 (1969) (a state may not deny the right to vote to a citizen because he does not own or lease taxable realty; Williams v. Rhodes, 393 U.S. 23 (1968) (state abridgment of individual's right to associate and right to vote must serve a "compelling state interest"); Harper v. Virginia Bd. of Elections, 383 U.S. 663 (1966) (a state may not impose a poll tax); Carrington v. Rash, 380 U.S. 89 (1965) (a state may not deny the right to vote to a bona fide resident because he is a member of the armed forces); Reynolds v. Sims, 377 U.S. 533 (1964) (a state may not dilute a citizen's vote through malapportionment); United States v. Classic, 313 U.S. 299 (1941) (the right to vote in a primary election is entitled to the same protection against state abridgment as in a general election); Nixon v. Herndon, 273 U.S. 336 (1927) (a state may not deny Negroes the right to vote in a primary election); Yick Wo v. Hopkins, 118 U.S. 356 (1886) (the right to vote is a fundamental right); United States v. Reese, 92 U.S. 214 (1875) (a state may not deny Negroes the right to vote in a municipal election). For additional cases see Rosario v. Rockefeller, 410 U.S. 752, 765 (1973).
14. 100 U.S. 371 (1879). *See Ex Parte Yarbrough*, 110 U.S. 651 (1884).
15. 118 U.S. 356 (1886).
16. *Id.* at 370.
During the past decade, judicial decisions have vigorously protected a citizen's right to vote against infringement of any kind. In 1964 the Supreme Court decided *Reynolds v. Sims*,\(^{17}\) in which it stated:
Undoubtedly, the right of suffrage is a fundamental matter in a free and democratic society. Especially since the right to exercise the franchise in a free and unimpaired manner is preservative of other basic civil and political rights, and any alleged infringement of the right of citizens to vote must be carefully and meticulously scrutinized.\(^{18}\)
In *Harper v. Virginia Board of Elections*,\(^{19}\) the Supreme Court declared the right to vote to be one of the fundamental rights protected by the equal protection clause of the fourteenth amendment, and concluded that "classifications which might invade or restrain them must be closely scrutinized and carefully confined."\(^{20}\)
The stringency of the language used in the above decisions laid the foundation for the "compelling state interest" test, introduced in *Williams v. Rhodes*.\(^{21}\) In that decision the Supreme Court said that state abridgments of an individual's first and fourteenth amendment rights were valid only if they served the purpose of a "compelling state interest" and did not constitute an "invidious" discrimination.\(^{22}\) Less than one year later, the Supreme Court decided *Kramer v. Union Free School District No. 15*,\(^{23}\) in which it applied the same test used in *Williams* to a New York education law which restricted the vote in local board elections. The opinion is an extremely significant one because of the distinction it made
---
\(^{17}\) 377 U.S. 533 (1964).
\(^{18}\) Id. at 561-62; Cf. Carrington v. Rash, 380 U.S. 89, 96 (1965) which held that: "[S]tates may not casually deprive a class of individuals of the vote because of some remote administrative benefit to the State."
\(^{19}\) 383 U.S. 663 (1966).
\(^{20}\) Id. at 670.
\(^{21}\) 393 U.S. 23 (1968). *Williams* involved a challenge to a set of Ohio statutes which made it almost impossible for a new political party to have its name and its candidates placed on the ballot in the general election. The Court overturned the statutes concluding that: "The State has here failed to show any 'compelling interest' which justifies imposing such heavy burdens on the right to vote and to associate." Id. at 31. Cf. NAACP v. Button, 371 U.S. 415 (1963) where the "compelling state interest" test was used in protecting first amendment rights of association. The "compelling interest" test was much more rigid than the traditional equal protection formula established in McGowan v. State of Maryland, 366 U.S. 420 (1961). The Court in *McGowan* stated that "[a] statutory discrimination will not be set aside if any state facts reasonably may be conceived to justify it." Id. at 426. The new test also differed from the traditional one by placing the burden of proof in determining whether or not a statute is supported by a "compelling state interest" on the state itself. 393 U.S. 23, 31 (1968).
\(^{22}\) 393 U.S. 23, 30-31 (1968).
\(^{23}\) 395 U.S. 621 (1969).
between the types of requirements which states place on an individual's right to vote. The Court classified these requirements into two categories. The first category includes the basic voting requirements of citizenship, age, and residency. The second category includes all requirements beyond the "basic voting requirements." These "additional" requirements are subject to the "compelling state interest" test.\textsuperscript{24}
One of the questions examined in \textit{Rosario} was whether or not this "compelling state interest" test was applicable to statutes imposing restrictions on the right to vote in a \textit{primary} election. The question was first dealt with in \textit{Newberry v. United States}.\textsuperscript{25} In that decision the Court concluded that the word "elections," as used in article I of the Constitution, did not include primary elections. The Court stated that: "General provisions touching elections in constitutions or statutes are not necessarily applicable to primaries—the two things are radically different."\textsuperscript{26} However, in \textit{Nixon v. Herndon},\textsuperscript{27} a Texas statute which denied Negroes the right to vote in a Democratic Party primary election was declared unconstitutional in violation of the equal protection clause of the fourteenth amendment. The statute was later re-enacted giving the state executive committee of a political party the power to prescribe the qualifications of its party members; but in \textit{Nixon v. Condon},\textsuperscript{28} this statute was also found to violate the Constitution.
The \textit{Nixon} cases were decided under the equal protection clause of the fourteenth amendment but did not go as far as to define "primary" as a part of the electoral process itself. However, in the landmark decision of \textit{United States v. Classic},\textsuperscript{29} the Supreme Court reversed the position established twenty years earlier in \textit{Newberry} and held that the right to vote in a primary election was entitled to the same amount of protection from state abridgment as in a general election.\textsuperscript{30} Three years af-
\begin{itemize}
\item \textsuperscript{24} [If a challenged state grants the right to vote to some bona fide residents of requisite age and citizenship and denies the franchise to others, the Court must determine whether the exclusions are necessary to promote a compelling state interest.]
\item \textsuperscript{25} 256 U.S. 232 (1921).
\item \textsuperscript{26} \textit{Id.} at 250.
\item \textsuperscript{27} 273 U.S. 536 (1927).
\item \textsuperscript{28} 286 U.S. 73 (1932). \textit{But cf.} Grovey v. Townsend, 295 U.S. 45 (1935).
\item \textsuperscript{29} 313 U.S. 299 (1941).
\item \textsuperscript{30} The Court stated that:
Where the state law has made the primary an integral part of the procedure of choice, or where in fact the primary effectively controls the choice, the right of the elector to have his ballot counted at the primary, is likewise included in the right protected by Article I, § 2. And this right of
\end{itemize}
ter Classic, in Smith v. Allwright,\textsuperscript{31} the Court finally concluded that:
[The right to vote in . . . a primary for the nomination of candidates without discrimination by the State, like the right to vote in a general election, is a right secured by the Constitution.\textsuperscript{32}
Thus, the decisions reached in the above cases make it clear that the "compelling state interest" test must be equally applied to "additional" restrictions placed upon a citizen's right to vote in a \textit{primary election} as well as in a general election. In Rosario, the enrollment requirement found in section 186 of the New York Election Law is clearly an "additional" requirement since it goes beyond the basic restrictions concerning citizenship, age, or residence. It was, therefore, subject to the scrutiny of the "compelling state interest" test. The reasoning behind the Supreme Court's decision to uphold the statute on the ground that it met the "compelling state interest" test lies in the specific purpose of the statute itself. The Supreme Court for the first time was dealing with a statute which had as its sole purpose the prevention of political party "raiding" by voters in a primary election.\textsuperscript{33} The Court viewed the delayed enrollment scheme as essential in order to prevent party "raiding" from occurring.\textsuperscript{34} The prevention of party "raiding" was believed to be a "compelling state interest" since it insured the preservation of the integrity of the electoral process.\textsuperscript{35}
\begin{itemize}
\item participation is protected just as is the right to vote at the election, where the primary is by law made an integral part of the election machinery, whether the voter exercises his right in a party primary which invariably, sometimes or never determines the ultimate choice of the representative.
\end{itemize}
\textsuperscript{31} 321 U.S. 649 (1944).
\textsuperscript{32} \textit{Id.} at 661-62.
\textsuperscript{33} See note 11 \textit{supra}.
\textsuperscript{34} The purpose for requiring enrollment prior to a general election was that: Few persons have the effrontery or the foresight to enroll as say, 'Republicans' so that they can vote in a primary some seven months hence, when they full well intend to vote 'Democratic' in only a few weeks. And, it would be the rare politician who could successfully urge his constituents to vote for him or his party in the up-coming general election, while at the same time urging a cross-over enrollment for the purpose of upsetting the opposite party's primary . . . . Allowing enrollment any time after the general election would not have the same deterrent effect on raiding for it would not put the voter in the unseemly position of asking to be enrolled in one party while at the same time intending to vote immediately for another.
\textsuperscript{35} \textit{See Bullock v. Carter}, 405 U.S. 134 (1972) where the Court stated that: "[A] state has an interest, if not a duty, to protect the integrity of its political processes from frivolous or fraudulent candidacies." \textit{Id.} at 145. But \textit{cf. Dunn v. Blumstein}, 405 U.S. 330 (1972) where the Court maintained that the "prevention of
RECENT CHALLENGES TO "RAIDING" STATUTES
Justifying the length of the time limitation for enrollment in *Rosario*, the Supreme Court referred to *Lippitt v. Cipollone* in which it affirmed without opinion a decision of the United States District Court for the Northern District of Ohio. The lower court had declared constitutional an Ohio statute which made it impossible for a person to be a candidate for nomination or election in a political party primary if he had voted as a member of another political party in a primary election within the preceeding four years. The district court held that the statute protected a "compelling state interest" since it sought to prevent candidates from switching parties for opportunistic reasons and therefore protected the integrity of the political process. *Rosario* can be distinguished from *Lippitt* since the petitioner in *Lippitt* was seeking to run for a political office, whereas the petitioners in *Rosario* were seeking to vote in a primary election. This difference is significant because the right to hold office is not a fundamental right, as is the right to vote in a primary election. Thus, the issue involved in *Rosario* is unique, since the Supreme Court sanctioned a statute designed to prevent political party "raiding" by voters rather than by candidates.
Within the past two years, there have been several lower court opinions dealing with statutes designed to prevent party "raiding" by voters. In September of 1971, a federal district court overturned a provision of a
[electoral] fraud is a legitimate and compelling government goal," but invalidated Tennessee's durational residence requirements for voting in a general election as not necessary to further a "compelling state interest." *Id.* at 345.
36. 404 U.S. 1032 (1972), *aff'g mem.*, 337 F. Supp. 1405 (N.D. Ohio 1971); *cf.* Ray v. Blair, 343 U.S. 214 (1952).
37. *Cf.* Brief for Lawyers for McGovern as Amicus Curiae at 9, Rosario v. Rockefeller, 410 U.S. 752 (1973), *citing* Bendinger v. Ogilvie, 335 F. Supp. 572, 576 (N.D. Ill. 1971).
The state's interest in limiting *candidates* from switching parties . . . is greater than its interest in limiting *voters* from switching parties. The state's interest in preserving a vigorous and competitive two-party system is fostered by the requirement that candidates demonstrate a certain loyalty and attachment to the party in whose primary they are running; the same cannot be said of voters, however, who should be freer to demonstrate their changes in political attitude by voting for popular candidates or against unpopular candidates in any party's primary election. Thus, it is not inconsistent to prevent candidates from switching parties from election to election and at the same time to permit voters to do so.
(Emphasis added.)
38. Bullock v. Carter, 405 U.S. 134, 142-43 (1972).
39. Gordon v. Executive Comm. of the Democratic Party, 335 F. Supp. 166 (D.D.C. 1971) (per curiam).
South Carolina election law\textsuperscript{40} which denied a voter his right to vote in a primary election if he refused to take an oath affirming that he had not voted in another party's primary election, convention, or precinct meeting, within the preceding year. The court found that "[n]o sound or compelling purpose can possibly justify 'locking' a citizen into a party and denying to him for a full year freedom to change parties."\textsuperscript{41}
Less than three months later in \textit{Fonthan v. McKeithen},\textsuperscript{42} a district court upheld a Louisiana statute\textsuperscript{43} which required a six month waiting period after a voter had changed party affiliation before he would be eligible to vote in a primary election. The statute was subjected to the traditional "rational relation" test instead of the "compelling state interest" test used in \textit{Kramer v. Union Free School District No. 15}.\textsuperscript{44} The Court distinguished \textit{Fonthan} from the situation in \textit{Kramer} on the theory that the Louisiana statute in \textit{Fonthan} was a temporary suspension of voter eligibility, whereas the New York statute in \textit{Kramer} acted as a permanent suspension of voting rights. However, since the Louisiana restriction goes beyond reasonable citizenship, age, and residency requirements it should have been categorized as an "additional" restriction, and as such, subjected to the "compelling state interest" test prescribed in \textit{Kramer}.\textsuperscript{45} The New York statute challenged in \textit{Rosario} is similar to the statute challenged in \textit{Fonthan}, because neither act provides for a permanent deprivation of voting eligibility. In \textit{Rosario}, however, the Supreme Court correctly recognized a temporary denial of voting rights as an "additional" restriction and therefore, applied the "compelling state interest" test to the statute.
In May of 1972, the United States District Court of New Jersey held that certain provisions of a New Jersey election law\textsuperscript{46} were an unconstitutional burden on a citizen's right to vote and on a citizen's right to associate freely with a political party.\textsuperscript{47} The court stated:
\begin{quote}
[Although the defendants have met their burden of showing a compelling interest by the state in preventing 'raiding,' the infringements on the
\end{quote}
\textsuperscript{40}. S.C. Code Ann. § 23-400.71 (Supp. 1971).
\textsuperscript{41}. 335 F. Supp. at 169.
\textsuperscript{42}. 336 F. Supp. 153 (E.D. La. 1971), appeal dismissed sub nom. Fantham v. Edwards, 409 U.S. 1120 (1973).
\textsuperscript{43}. La. Rev. Stat. Ann. § 18:270.204 (Supp. 1973).
\textsuperscript{44}. 395 U.S. 621 (1969).
\textsuperscript{45}. See note 24 supra.
\textsuperscript{46}. N.J. Stat. Ann. § 19:23-45 (1964).
\textsuperscript{47}. Nagler v. Stiles, 343 F. Supp. 415 (D. N.J. 1972).
right to vote and the right of association which the State selected go far beyond what is necessary for the protection of that interest.\textsuperscript{48}
Judge Fisher distinguished the New Jersey statute from the one found in \textit{Rosario} on the grounds that the latter accomplished its goal of inhibiting “raiding” without imposing “an excessively long commitment to a party by a voter.”\textsuperscript{49} The significance of this distinction is that it qualifies the statute in \textit{Rosario} as one which serves a “compelling state interest” without being overly broad or making unnecessary infringements on the right to vote and the right to associate freely. This same distinction was pointed out two months later in \textit{Yale v. Curvin}.\textsuperscript{50} In overturning a Rhode Island statute\textsuperscript{51} which placed a twenty-six month waiting period before persons who had switched their political affiliations could vote in a primary election the court noted that “[u]nlike the Rhode Island law, the New York law is narrow and does not disenfranchise a voter from a succeeding primary election.”\textsuperscript{52} The district court recognized the legitimate interest that the state has in the prevention of “raiding” but concluded that the twenty-six month restriction on voting was too broad and too drastic an infringement upon a citizen’s constitutional rights.
On March 7, 1972, exactly one month prior to the appellate court’s decision to uphold the statute in \textit{Rosario}, the District Court for the Northern District of Illinois decided \textit{Pontikes v. Kusper}.\textsuperscript{53} The court overturned an Illinois statute which prohibited a voter from participating in a political party primary if he had voted in the primary of another party within the preceeding twenty-three months.\textsuperscript{54} Judge Swygert found the statute to be a substantial burden on a citizen’s right to vote in a primary election as well as an infringement upon their right of free association. Since party “raiding” by voters in Illinois was not considered to be a serious threat, the statute designed to prevent it was not deemed as serving a “compelling state interest,” and therefore, the statute was declared invalid.\textsuperscript{55} The opinion of the court in \textit{Pontikes} was influenced by the abso-
\begin{itemize}
\item \textsuperscript{48} \textit{Id.} at 418.
\item \textsuperscript{49} 343 F. Supp. 415, 417-18 (D. N.J. 1972).
\item \textsuperscript{50} 345 F. Supp. 447 (D. R.I. 1972).
\item \textsuperscript{51} R.I. GEN. LAWS ANN. §§ 17-15-24, 17-16-8 (1969).
\item \textsuperscript{52} 345 F. Supp. 447, 451 (D. R.I. 1972).
\item \textsuperscript{53} 345 F. Supp. 1104 (N.D. Ill. 1972) (Marovitz, J., dissenting), \textit{aff’d}, 42 U.S.L.W. 4003 (U.S. Nov. 19, 1973).
\item \textsuperscript{54} ILL. REV. STAT. ch. 46, § 7-43(d) (1971).
\item \textsuperscript{55} 345 F. Supp. 1104, 1108 (N.D. Ill. 1972). Referring to the statute, Judge Swygert stated that:
\end{itemize}
The state’s interest . . . could be characterized as “compelling” only if
lute and drastic qualities of section 7-43(d). Under the Illinois law, those that choose to switch parties after voting in a primary election were deprived of their right to vote in any primary for a period of twenty-three months. The provision made it impossible for the petitioners to vote in one party's primary in March of 1972 because they had voted in another party's primary thirteen months earlier. Thus, the statute had the effect of absolutely depriving petitioners of their right to vote. However, in Rosario the Supreme Court pointed out that the New York Election Law did not absolutely deprive the petitioners of their right to vote in the June primary, but rather placed a cut-off date on their enrollment—which they had to meet in order to vote in the primary. Thus, the Illinois and New York statutes are distinguishable in the degree to which they restricted the electorate's voting rights. Although both statutes are temporary restrictions, the statute in Pontikes absolutely denies voters their right to vote in a primary election, whereas the law in Rosario denies the right to vote only to those persons who fail to enroll at least thirty days prior to the general election preceding the primary in which they wish to vote.
In November of 1973, the United States Supreme Court affirmed the district court's decision in Pontikes. The absolute denial of the appellee's right to vote in the Democratic primary election was the determining factor in the Court's opinion. Referring to the effect of the Illinois statute on the appellee's right to vote, Mr. Justice Stewart stated: "[T]he state law absolutely precluded her from participating in the 1972 Democratic primary." This distinguishing factor explains why the Court reached opposite conclusions in Rosario and Pontikes. As Justice Stewart stated:
Unlike the petitioners in Rosario, whose disenfranchisement was caused by their own failure to take timely measures to enroll, there was no action
... raiding constitutes a more important danger than the danger to constitutionally protected rights ... [and] [t]here is no evidence to indicate that raiding is more likely to take place than "honest" switches of affiliation.
Id. at 1108.
56. See note 54 supra.
57. 410 U.S. at 757.
58. See exemptions listed in note 5 supra.
59. Kusper v. Pontikes, 42 U.S.L.W. 4003 (U.S. Nov. 19, 1973). Stewart, J., delivered the majority opinion of the Court, in which Douglas, Brennan, White, Marshall and Powell, JJ., joined. Burger, C.J., concurred in the result. Blackmun, J., filed a dissenting opinion. Rehnquist, J., filed a dissenting opinion, in which Blackmun, J., joined.
60. Id. at 4006.
61. Id.
that Mrs. Pontikes could have taken to make herself eligible to vote in the 1972 Democratic primary.\textsuperscript{62}
Thus, the Court found that the Illinois election law "substantially abridged" the appellee's freedom to associate with the political party of her choice and therefore, overturned the statute.
\textbf{THE WEAKNESSES AND IMPLICATIONS OF Rosario}
Before reviewing the effects that the Supreme Court's opinion in \textit{Rosario} may have on future decisions, it is necessary to examine the opinion's inherent weaknesses. The most significant weakness in Mr. Justice Stewart's opinion concerns his failure to apply the "least drastic means" test\textsuperscript{63} to New York's statutory enrollment scheme. Petitioners had argued that section 332 of the New York Election Law\textsuperscript{64} offered a "less drastic means" to prevent "raiding."\textsuperscript{65} However, the Court reasoned that section 332 was "too cumbersome to have any real deterrent effect on raiding in a primary."\textsuperscript{66} Therefore, the Court concluded that section 186 "imposed a legitimate time limitation" on the petitioner's
\textsuperscript{62} \textit{Id.}
\textsuperscript{63} In Shelton v. Tucker, 364 U.S. 479, 488 (1960) Mr. Justice Stewart pointed out that "even though the governmental purpose be legitimate and substantial, that purpose cannot be pursued by means that broadly stifle fundamental personal liberties when the end can be more narrowly achieved. The breadth of legislative abridgment must be viewed in the light of less drastic means for achieving the same basic purpose." \textit{Accord}, Dunn v. Blumstein, 405 U.S. 330, 343 (1972).
\textsuperscript{64} N.Y. ELECTION LAW § 332 (McKinney 1964) provides in pertinent part:
1. The supreme court or a justice thereof within the judicial district, or the county judge within his county, in a proceeding instituted by a duly enrolled voter of a party, at least ten days before a primary election, shall direct the enrollment of any voter with such party to be cancelled if it appears either that any material statement in the declaration of the voter upon which he was enrolled is false. . . .
2. The chairman of the county committee of a party with which a voter is enrolled in such county, may, . . . after a hearing . . . determine that the voter is not in sympathy with the principals of such party. . . . [T]he enrollment of such voter [may] be cancelled if it appears . . . that such determination is just.
\textsuperscript{65} \textit{Rosario v. Rockefeller}, 410 U.S. at 762 n.10.
\textsuperscript{66} \textit{Id. But see In re Mendelsohn}, 197 Misc. 993, 99 N.Y.S.2d 438 (Sup. Ct. 1950); Zuckman v. Donahue, 191 Misc. 399, 79 N.Y.S.2d 169 (Sup. Ct.), \textit{aff'd}, 274 App. Div. 216, 80 N.Y.S.2d 698, \textit{aff'd mem.}, 298 N.Y. 627, 81 N.E.2d 371 (1948); Werbel v. Gernstein, 191 Misc. 275, 78 N.Y.S.2d 440 (Sup. Ct.), \textit{aff'd}, 273 App. Div. 917, 78 N.Y.S.2d 926 (1948); \textit{In re Newkirk}, 144 Misc. 765, 259 N.Y.S. 434 (Sup. Ct. 1931). Also the Supreme Court's opinion in \textit{Rosario} failed to mention the criminal sanctions available in Article 16 of the New York Election Law for violations of the elective franchise. These sanctions further enhance the state's ability to prevent "raiding." \textit{See N.Y. ELECTION LAW §§ 421, 436} (McKinney Supp. 1972).
enrollment.\textsuperscript{67} Thus, as pointed out in Mr. Justice Powell's dissenting opinion: "The Court . . . fails to address the critical question of whether [the state's] interest may be protected adequately by less severe measures."\textsuperscript{68} Furthermore, Mr. Justice Powell believes that a less drastic enrollment deadline (thirty to sixty days prior to the primary election) would prevent most "raiding" activity, yet at the same time allow "honest" voters a reasonable opportunity to switch parties prior to the primary election without denying them their right to vote.\textsuperscript{69} Surely, a less drastic time limitation could have accomplished the same end, without imposing such a heavy burden on those who choose to switch political parties for legitimate purposes.\textsuperscript{70}
The Court's reasoning for sanctioning the cut-off date thirty days prior to the preceding general election as meeting the "least drastic means" test was based on the alleged inconsistency a voter would face in planning to vote for one party in a general election while at the same time switching his political affiliation with the intention of "raiding" another political party eight months hence.\textsuperscript{71} The amount of cognitive dissonance involved in switching parties prior to a general election would, no doubt, be hard
\textsuperscript{67} 410 U.S. at 762. Mr. Justice Stewart's conclusion in \textit{Rosario} seems to contradict his opinion concerning the "least drastic means" test which he discussed in Shelton v. Tucker, 364 U.S. 479, 488 (1960). \textit{See} note 62 \textit{supra}. This apparent contradiction might be explained by the particular situation found in \textit{Rosario}. Justice Stewart emphasized the fact that the petitioners in \textit{Rosario} did not explain why they failed to enroll prior to the cut-off date. Justice Stewart also emphasized the fact that the petitioners did not claim that they were unaware of the cut-off date. By concluding that the petitioners could have enrolled, but chose not to, Mr. Justice Stewart put the blame not on the statute, but on the petitioners themselves. 410 U.S. at 758. This factor seems to have been decisive in Justice Stewart's decision to uphold the New York statute. It is quite possible that Justice Stewart believed that the petitioners did not comply with the cut-off date intentionally in order to subject the election law in question to the rudiments of a judicial test. If, in fact, the petitioners had stated a valid reason why they had failed to enroll on time, the Court's decision might have been in their favor.
\textsuperscript{68} \textit{Id.} at 770.
\textsuperscript{69} \textit{Id.} at 771.
\textsuperscript{70} The appellate court in \textit{Rosario} pointed out that New York "has a particular interest in preventing raiding," since
[i]n addition to the major parties, Democrat and Republican, two minority parties, Conservative and Liberal, are established throughout the state and usually present a full slate of candidates in the general election. Yet as there are only 107,000 enrolled Conservatives and 109,000 enrolled Liberals as opposed to 2,950,000 enrolled Republicans and 3,565,000 enrolled Democrats, successful raiding of these minority parties would present little difficulty on statewide basis absent § 186.
458 F.2d 649, 652 n.3 (2d Cir. 1972). Obviously, this factor had an influence on the Court's decision to uphold such a drastic means. However, even in this situation, there is no evidence that "raiding" occurs more frequently than "honest" switches of affiliation against "raiders."
\textsuperscript{71} \textit{See} note 34 \textit{supra}.
to measure. Yet, it would seem that such dissonance would be a greater deterrent to "honest" switches in party affiliation than to those bent on sabotaging another political party. Moreover, placing the cut-off date for enrollment eight months in advance of a primary deprives the voter of making a rational decision in choosing a political party, since many of the candidates, let alone the issues, are not even known at that time.
Thus, forcing the voter to enroll in a political party eight months prior to the primary election in order to be eligible to vote in that primary deprives the voter of the opportunity to make a rational decision and may distort the results of the primary by not allowing the voter to vote for the true candidate of his choice. Such distortion would seriously undermine the validity of the primary election results. A cut-off date one or two months prior to the primary election would undoubtedly aid the voter in making a more educated and rational decision in choosing which party to enroll in, while still acting as a preventative to party "raiding."
A second weakness in the Court's opinion can be found in its decision not to grant the petitioner's standing to argue that section 186 conflicts with the 1970 amendments to the Voting Rights Act of 1965. Avoiding the plaintiff's allegations, the Court has left open the question of whether or not the Voting Rights Act was meant to apply to primary elections as well as to presidential elections. 42 U.S.C. § 1973(c)(1) seems to indicate that section 1973 aa-1(d) is applicable to primaries. An in depth examination by the Court on this issue would have been helpful.
72. Petitioners alleged that 42 U.S.C. § 1973 aa-1(d) specifically prohibits an otherwise qualified voter from being denied the right to vote in a presidential primary for failure to comply with any type of durational residency requirement. See Rosario v. Rockefeller, 410 U.S. 752, 759 n.9 (1973).
73. 42 U.S.C. § 1973aa-1 (1970) provides that each State shall provide by law for registration or, other means of qualification of all duly qualified residents . . . not later than thirty days immediately prior to any presidential election, for registration or qualification to vote . . . in such election.
Cf. Allen v. State Bd. of Elections, 393 U.S. 544, 563-67 (1968). But cf. Rosario v. Rockefeller, 458 F.2d 649, 654 (2d Cir. 1972) where the appellate court argued that § 1973 aa-1(d) did not apply to votes cast in presidential primaries.
74. See note 6 supra. H.R. REP. No. 439, 89th Cong., 1st Sess. 32 (1965) states:
Clause (1) of [The Voting Rights Act of 1965, 42 U.S.C. § 1973 l(c)(1)] . . . contains a definition of the term "vote" for purposes of all sections of the act. The definition makes it clear that the act extends to all elections—Federal, State, local, primary, special, or general—and to all actions connected with registration, voting, or having a ballot counted in such elections. (Emphasis added).
In estimating the possible effects of *Rosario* on future decisions, the subjectivity of the "compelling state interest" test needs to be kept in mind. That is, the individual needs of a state will have to be taken into consideration before the Court can determine whether or not a statute designed to prevent harm to the integrity of the American political process meets that state's "compelling interest." Thus, a statute that meets the test in one state may fail in another. Moreover, the stringency or absolute effect of the statute itself will determine whether or not it is acceptable. This latter determination was clearly pointed out by the Supreme Court's decision in *Pontikes* to overturn the Illinois election law.\(^{75}\) The Court was careful to note that although Mrs. Pontikes was not absolutely precluded from associating with the Democratic party, she was absolutely precluded from voting in the Democratic party's election. As the Court said:
"[T]he Illinois statute deprived her of any voice in choosing the party's candidates, and thus substantially abridged her ability to associate effectively with the party of her choice."\(^{76}\)
Thus, the line of demarcation where one's freedom of political association becomes "substantially abridged" can be drawn at the point where the state absolutely denies one of his right to vote in any political party primary. This line of demarcation is the distinguishing factor between the New York and Illinois election laws.
**CONCLUSION**
The Court's failure to declare the statute in *Rosario* unconstitutional as "substantially abridging" one's freedom of political association (merely because the statute did not absolutely deny the petitioners of their right to vote) is unfortunate. State imposed obstacles on the road to the voting polls should not have to be insurmountable before they are declared unconstitutional. A weighing of the state's interest in the prevention of party "raiding" against a citizen's right to participate in the primary election of his choice would be a more reasonable manner in which to determine whether or not a person's right to vote and/or right of political association have been "substantially abridged." In the words of Mr. Justice Black:
No right is more precious in a free country than that of having a voice in the election of those who make the laws under which, as good citizens,
\(^{75}\) Kusper v. Pontikes, 42 U.S.L.W. 4003, 4006 (U.S. Nov. 19, 1973).
\(^{76}\) Id.
we must live. Other rights, even the most basic, are illusory if the right to vote is undermined.\textsuperscript{77}
Although the Supreme Court's opinion in \textit{Rosario} may presently seem to be a step backwards from Justice Black's words of wisdom, hopefully, it will eventually be viewed as only a side-step taken to avoid a particular situation which could have hindered the integrity of the political process. In the more immediate future, \textit{Rosario} will add to the ammunition of those states desiring to impose requirements under which the right of suffrage may be exercised.\textsuperscript{78} In the final analysis, the greater a state's need to impose requirements on the right to vote in order to protect the integrity of the political process, the greater the probabilities are that those requirements will pass the "compelling state interest" test provided, of course, that the requirements imposed do not \textit{absolutely} deprive citizens of their right to vote.
\textit{Neal Taslitz*}
\textsuperscript{77} Wesberry v. Sanders, 376 U.S. 1, 17 (1963).
\textsuperscript{78} See, e.g., Lassiter v. Northampton County Bd. of Elections, 360 U.S. 45 (1959); Pope v. Williams, 193 U.S. 621 (1904); Mason v. Missouri, 179 U.S. 328 (1900).
* The author dedicates this case note to the memory of the late Thomas George Krepps, a 1973 graduate of the De Paul University College of Law. |
Acquisition and evaluation of radiometrically comparable multi-footprint airborne LiDAR data for forest remote sensing
Ilkka Korpela, Department of Forest Sciences, POB 27, 00014 University of Helsinki, Finland. email@example.com.
Abstract
Forest inventories comprise observations, models and sampling. Airborne LiDAR has established its role in providing observations of canopy geometry and topography. These data are input for estimation of important forest properties to support forestry-related decision-making. The primary deficiency in forest remote sensing is tree species identification. This study examines the quite atypical option of using multi-footprint airborne LiDAR data. Features of such sensor design exist in recently introduced multispectral laser scanners. The first objective was to acquire radiometrically normalized, multi-footprint (11, 22, 44 and 59 cm) waveform (WF) data that characterize 1064-nm backscatter reflectance on the interval scale. The second objective was to analyze and validate the data quality in order to draw the correct conclusions about the effect of footprint size. Finally, the data were analyzed in different forest canopies. The experiment was carried out in Finland. Footprint variation was generated by acquiring data at different flying heights and by adjusting the transmitted power. The LiDAR campaign was successful and the data were of sufficient quality, except for a 1 dB trend due to the atmosphere. Significant findings were made concerning the magnitude of atmospheric losses, the linearity of the amplitude scale and the bandwidth characteristics of the receiver, the stability of the transmitter, the precision of the amplitude data, the transmission losses in canopies and power lines as well as the response of WF attributes to footprint size in forest canopies. Multi-footprint data is a promising approach although the species-specific signatures were weak.
Keywords: Canopy structure, Tree species, Forest inventory, Laser scanning, Waveform, Normalization, Linearity, Impulse response, Losses
1. Introduction
Airborne LiDAR data are used for many purposes, and the rationales for the present study originate from LiDAR remote sensing (RS) of forests (Vauhkonen et al., 2014). It has recently become an integral part of forest management planning systems in Finland (Maltamo and Packalén, 2014). LiDAR is an observation tool that has reduced the sampling intensity and provided entirely new observations for the estimation of forest attributes. In Finland, forest planning inventories consist of area-based estimation of stand boundaries, attributes and management proposals, using a combination of field reference, aerial image and LiDAR data, all finally followed by a field inspection. The data acquisition costs have decreased compared with field-work intensive systems, with both improvements and deficiencies in the deliverables. The primary deficiency in forest RS, not only in
Finland, is tree species identification. There is a need for observations that aid in solving this problem. Multi-footprint data may comprise one option.
Regarding sensors, foresters have mostly used data acquired by topographic small-footprint instruments. Narrow beams promote accurate geometry and reach the ground. While simulation studies have highlighted how sensor properties influence the available information (Disney et al., 2010; Hovi & Korpela, 2014), certain constraints pertain to sensor design. They include the need for extremely high dynamic range, consideration of eye-safety, receiver sensitivity and bandwidth, availability of powerful lasers at different wavelengths, data transfer and storage capacity requirements, power consumption, limitations on the size of the mirror and the aperture, and the final instrument cost. Advances in the last 10–15 years relate to improvements in pulse repetition frequency, capture of waveform (WF) data, fast direct georeferencing and the detection of multiple echoes, inter alia. Teledyne Optech (Vaughan, Ontario) recently released a three-band, dual-footprint WF-recording sensor called Titan, in which the divergence of the 532-nm pulses is wider than that of 1064-nm and 1550-nm pulses. The relative reflectance difference of green and woody vegetation between the 1064 and 1550 nm bands is likely beneficial for tree species identification as the green/woody silhouette visible from the above shows between-species differences owing to differences in leaf angle distributions, and in branch and crown morphology. On the other hand, WFs carry species-specific traits (Hovi et al., 2015), when tens of pulses per crown are evaluated. How the combination of WF and multi-band data will benefit species identification, is a very interesting topic. Although advances are foreseeable with multispectral LiDAR, this study examines the relatively unconventional concept of multi-footprint LiDAR to investigate how such data characterize canopy structure, and in general augment our understanding of the characteristics of LiDAR data. The rest of this section provides further background to the topic and introduces hypotheses that were tested empirically.
The radar equation (Eq. 1) applies to LiDAR as well, and states that the received power ($P_r$) is influenced by the aperture area ($D$), the atmospheric transmission losses ($T$), the optical efficiency of the system ($Q$), the transmitted power ($P_t$), the laser beam divergence ($\beta$), the range ($R$), and the backscattering cross-section of the target ($\sigma$):
$$P_r = \frac{P_t D Q}{4\pi \beta^2} T^2 \frac{\sigma}{R^4}$$
where $\sigma$ can be said to be comprised of target reflectance, geometry and the illuminated/silhouette area. Eq. (1) is ‘instantaneous’, while in reality the power terms and $\sigma$ are time- or distance dependent and the optical power of the beam has a Gaussian cross-section profile (cf. Wagner et al., 2006; Mallet and Bretar, 2009).
Eq. (1) states that $P_r$ is in linear dependence with target reflectance and the transmitted power. A large aperture increases $P_r$ and the signal-to-noise ratio (SNR), because the extra solar illumination reaching the detector, non-attenuated, is minor compared to the internal receiver noise. Atmospheric attenuation is 0.1–0.3 dB/km in weather conditions that are suitable for LiDAR RS.
The effect of $R$, i.e. the spherical losses, depend on the target, because large surfaces, linear targets and blob-like objects all exhibit different response. A sensor that provides ratio- or interval-scale measurements of $P_r$ will show a fourfold (6 dB) increase in well-defined surfaces, when $R$ is halved, whereas $P_r$ is 8 or 16 times (9 or 12 dB) larger for a wire or a small leaf. Tree crowns are volumetric and the structure influences the signal change due to $R$ (Korpela et al., 2010, Gatziolis, 2011), making range normalization ill-posed in canopies. In linear targets, such as cables, the distribution of $P_r$ has an exponential-even shape with a modest peak near the maximum, while for blob-like features, the distribution is inverse-exponential.
Divergence defines the angular spread of the beam’s Gaussian irradiance profile. Here it is defined by the $1/e^2$ point. A target may give rise to a measurable echo, only if it is illuminated by the pulse center. The backscatter cross-section may shrink for a tilted target, or at least the returning pulse is extended and dampened. Because sensors record meaningful data only, backscattering ($P_r$) needs to be strong enough to trigger an observation, and the real signal distributions are always truncated by some combination of $\sigma$ and the receiver noise. The sensor used in this study stores a single continuous and fixed-length WF, the recording of which is triggered by the first threshold-exceeding echo. Some sensors record variable-length WF samples with pauses in between and thresholding applies to this ‘piecewise WF-recording’ as well (e.g. Armston et al., 2010). In general, weak backscattering preceding the first echo may remain undetected, while such weak within-canopy or ground scattering may be observable, when the recording is on, and missed, while it is paused. This constitutes a vertical sampling bias. A somewhat similar effect causes selection bias in the analyses of discrete-return (DR) intensity data, when the separation between only and first-of-many echoes is often made (Ørka et al., 2010; Korpela et al., 2010b), because this instrument-driven division is influenced by the backscatter cross-section profile of targets following the first interaction, for example the ground flora in the case of tree crowns.
Regarding multi-footprint data, changing solely the beam width will not influence $P_r$ as long as the target is a well-defined surface filling the footprint, while for small targets, $P_r$ varies according to the intersection geometry, as stated. Multiple scattering and transmission losses are introduced instantly following the first pulse-canopy interaction (the subsequent targets comprise an apparent differential backscatter cross-section profile (Richter et al., 2015), high-order volumetric scattering can be significant, and backscattering occurs far from the beam path (Hovi and Korpela, 2014). Foliage orientation, size, density and silhouette area were shown to influence the recorded intensity in broadleaved and coniferous trees in Finland (Korpela et al., 2010b, 2013; Hovi and Korpela, 2014). If beam width increases by a factor of $i$, while $P_r$ is constant, the average $P_r$ decreases by $1/i^2$ and $1/i$ in blob-like and linear targets, respectively. A wide beam is more likely to find directional canopy gaps and the proportion of pulses that illuminate the ground is higher. Many small targets that trigger an echo in narrow beams remain undetected in large-footprint data and contribute to transmission losses, which explain why ground signal levels can be expected to be lower in large-footprint data. The wider the pulse, the smaller is the likelihood for the pulse to reach the ground with maximal energy. Thus, measuring the reflectance properties of the forest floor using single-return data is more accurate in
narrow footprint data as shown for understory lichens in Korpela (2008). The peak amplitude distribution of canopy echoes can therefore be expected to be shifted towards low values in large footprint data. It can also be expected that the return WFs are widened in large-footprint data, because a wider beam illuminates targets that also spread across a larger depth.
As stated, spherical losses depend on the target geometry and it is therefore evident that a fixed signal-to-noise-ratio (SNR) cannot be maintained for the canopy and the ground, when scanning from different heights (Goodwin et al., 2006, Ørka et al., 2010). When $R$ is doubled, a 6 dB increase in $P_t$ will maintain the SNR for well-defined surfaces. We cannot expect this to hold true for the forest floor, because canopy transmission losses will increase with increasing footprint size. The SNR of small targets will fall even if the power is increased by 6 dB, which explains the reduction in canopy echo counts in high-altitude LiDAR (Goodwin et al., 2006). Of course, atmospheric losses introduce additional signal loss. Varying the scan zenith angle changes $R$ and the incidence angles. Pulses reflecting from tilted surfaces will exhibit widened echoes, in which the peak amplitude is dampened, and both the widening and the dampening depend on the footprint size (e.g. Mallet and Bretar, 2009).
Increasing $P_t$ improves the SNR. Other options include enlarging the receiver aperture and mirror, or decreasing the receiver noise level. The latter comprises dark noise of the detector, electronic and digitization noise as well as natural light that reaches the detector. The energy captured by an aperture that is located in the footprint center, and focuses light to the retina, cannot exceed critical values, which in turn limits $P_t$ and/or beam divergence. Eye-safety is a more severe concern at 532 nm than at 1064 nm or 1550 nm. Regarding multi-band sensors, we can note that the reflectance of foliage at 532 nm is one tenth of that at 1064 nm. Given the same aperture, divergence and output power, the green band signals will thus be 10 times lower. In order to improve the SNR of well-defined targets (at 532 nm), it is possible to increase both the transmitted power and/or beam divergence, where the latter may be needed for eye-safety. Because multiple scattering in vegetation is weak in visible bands, the returning 532 nm pulses are probably less extended compared to NIR pulses of the same beam divergence, which may be exploited in target classification using multi-band LiDAR.
Radiometric observables in LiDAR data include intensity and WF amplitude data. The instantaneous spectral irradiance is captured by the aperture, reflected to a collimating lens that directs it to the photon detector through a band-pass filter. The detector and the circuits that follow, i.e. the receiver, have certain noise characteristics, impulse response and bandwidth (BW) properties. The latter two determine how the dynamics of the photon surge are captured. Receiver bandwidth is inherently limited and rapid changes in the photon surge are low-pass filtered in the WF. As long as the transmitted pulses are stable, minor deficits in bandwidth are not critical for reliable ranging, which is also shown in this study. Receiver performance may depend on signal strength and for example very strong echoes can temporarily increase the noise level or cause 'ringing', i.e. 'ghost echoes', in the WF (e.g. Armston et al., 2010). If the internal signal levels are too high, i.e., outside the linear (dynamic) range of the receiver circuits, the sensor of this study performs erroneous discrete ranging and the intensity data are distorted.
A one-nanosecond sampling interval is typical in WF-recording sensors, and the length (full width at half maximum, FWHM) of the transmitted pulse is below 10 ns. The power goes up to several kWs in some nanoseconds only, but the exact power envelope of the transmitted pulse is usually not known. Any within- or between-pulse variation directly influences the convolution of the transmitted pulse with the backscatter cross-section profile of the targets. Some sensors provide a digitized sample of the transmitted pulse to enable exact ranging and echo attribute retrieval in post processing (Wagner et al., 2006; Roncat et al., 2014).
Radiometrically quantitative analyses require that the properties of the transmitted pulse (directional target illumination) and the receiver’s response to instantaneous at-sensor irradiance (measurement of reflected radiance) are known, in addition to the properties of the medium. Such absolute calibration is typically unattainable for LiDAR sensors and vicarious calibration with reflectance targets becomes the only option (cf. Wagner, 2010). This in turn is challenging because reference measurements of zero phase-angle reflectance are hard to establish and are rarely done. A linear response of the receiver is usually assumed (with gain and offset), or alternatively a look-up-table is used. Some sensors have two receivers and the model is needed for both. DR sensors can have individual circuits for the detection of echoes and the intensity values can have different calibration. The sensor used in this study has a single receiver, in which the gain of the signal chain is dynamically adjusted up to 3 dB by an automatic gain control (AGC) circuit. In order to draw conclusions about the influence of footprint size, using the WF data, it is vital to establish a mapping between the amplitude data and the at-sensor power. Such mapping was performed in this study.
This study examines the relatively unconventional concept of multi-footprint LiDAR to investigate how such data characterize canopy structure, and in general augment our understanding of LiDAR data. Specific objectives were:
1. Plan and carry out WF LiDAR data acquisition with a fixed beam divergence sensor that results in radiometrically comparable multi-footprint data in well-defined targets.
2. Carry out radiometric normalization of peak amplitude data using well-defined surfaces to remove the effects of the AGC circuit, to estimate the offset of the amplitude scale and to identify any deficits in the acquisition.
3. Validate the normalization across the study site to ensure that the data qualifies for analyses on the influence of the footprint size.
4. Acquire independent data using the same sensor to investigate and verify whether the amplitude scale of the receiver is in linear dependence with the instantaneous at-sensor power.
5. Investigate the impulse response of the receiver to assure that WF attributes are influenced only by the pulse-target interaction.
6. Carry out empirical tests to show that the hypotheses presented in the introduction are confirmed by the measured real data.
2. Experiment and results
2.1 Research site
The experiments were carried out in Hyvijärvi (61°50′N, 24°17′E), southern Finland. The area is mostly covered by forest, but there are also fields, open and forested peatlands, built environment and a network of forest roads. Aerial photographs cover the time period of 1946–2015 and there are 19 airborne LiDAR datasets acquired in 2004–2015. Geometric reference data include close-range photogrammetric images, surveying points and profiles. These data as well as the redundant airborne image and LiDAR data were used for geometric quality control of the LiDAR data sets used here. (Hovi, 2015).
2.2 Waveform-recording LiDAR data
Objective #1 was to acquire multi-footprint data, in which the same energy was transmitted by varying the beam divergence. Sensors with the option of changing divergence were not available. To produce the same effect, a Leica ALS60 sensor was used at different heights as its transmitted power can be adjusted by almost 20 dB (1–100%). The nadir $1/e^2$ footprint diameters were 11, 22, 44 and 59 cm from the acquisition heights of 0.5, 1.0, 2.0 and 2.7 km, respectively. The pulse power was set to nominal levels of 100, 55, 14 and 4% (Tables 1 and 2) to account for the spherical losses in well-defined surfaces. Strip overlaps were increased for the higher altitudes to reach comparable pulse density. According to the manufacturer, the FWHM of the transmitted pulses is 4 or 9 nanoseconds, which showed as 7 and 10 ns in the recorded data (Fig. 1). The wavelength is 1064 nm, and Fig. 2 shows the difference of intensity data in different targets between the ALS60 and a Riegl LMS-Q680i sensor, which operates at 1550 nm. The Riegl data was acquired in 2013 (not tabulated) and was used for illustrations only. The same ALS60 sensor was used for datasets 2012_## and 2013_## (Table 2). The acquisition in 2013 was carried out to verify the linear response of the receiver, objective #4. It comprised three flying heights: 700 m, 800 m and 900 m. The sensor was operated in an unorthodox manner by deliberately fixing both the transmitter power and the receiver gain (AGC, see later). Appropriate gain and power settings were found by first acquiring the 700-m data with the AGC on. These datasets have ±12.5% range variation.
The receiver front-end in the ALS60 implements an AGC and the gain can vary by 0–3.2 dB (typically less) in response to scene brightness so that the internal signal levels are kept in the linear region of the range detection circuits. As said, the gain can also be fixed (2013_## datasets). The influence of the AGC can largely be removed, using an 8-bit ‘voltage value’ that is stored for each pulse, shown first in Korpela (2008).
Table 1. Sensor parameters.
| Parameter | Value |
|------------------------------------------------|------------------------|
| Wavelength, nm | 1064 |
| Pulse length, FWHM, ns | 4, 9 |
| System WF, FWHM, ns | 7.8, 10.2 |
| WF of transmitted pulse | No |
| Divergence, mrad, 1/e² | 0.22 |
| Mirror | Oscillating |
| Front-end amplifier | Automatic Gain Control |
| Triggering of WF-recording | Discrete-return circuit|
| WF-recording sequence | Single, 256 samples |
| WF Sampling rate, ns | 1 or 2 |
| LiDAR position, XYZ | yes |
| Discrete-return echoes | on-the-fly; 1, 2, 3 or 4|
| Echo attributes | range, XYZ, intensity |
| Acquisition height, m | 500, 1000, 2000, 2700 |
Table 2. Characteristics of the experimental LiDAR acquisitions. See also Table 1.
| Dataset | 2012_ (05, 10, 20, 27) | 2013b_ (07, 08, 09) |
|---------------|------------------------|---------------------|
| Purpose | Multi-footprint data | Linearity tests |
| Date, local | July 5, 2012 | June 15–16, 2013 |
| Time, GMT | 19–22 | 21–00 |
| Phenology | mid-season | early season |
| Sensor | ALS60 | ALS60 |
| Altitude, m | 500, 1000, 2000, 2700 | 700, 800, 900 |
| WF density per m² | 5, 5, 3, 2 | 6, 6, 5 |
| Scan angle, degrees | ±15 | ±15 |
| PRF, kHz | 152, 99, 59, 45 | 106 |
| System WF, FWHM, ns | 7.8, 10.1, 10.3, 10.4 | 7.8 |
| Footprint, 1/e², m | 0.11, 0.22, 0.45, 0.59 | 0.15, 0.18, 0.20 |
| AGC | on | off |
ALS60 is primarily a DR sensor and the WF storage calls for an optional digitizing module. The storage is triggered by the first DR echo. A sequence of 256 amplitude (A) values is recorded including a buffer of 30 samples (Fig. 1). Two oscilloscopes take turns in digitizing the pulses at 1 GS/s. Their bandwidth (BW) is 300 MHz and the rise time for a square signal from 10% to 90% is 1.2 ns. The BW and linearity of the preceding circuits, including the impulse response of the photon detector are unknown, as is the
exact stability of the transmitter. However, the large difference in the FWHM of the transmitted and received 4 ns pulses clearly suggests bandwidth deficits.
**Fig. 1.** The x-axis shows the range in meters centered (0 m) at the first echo. In the left figure there are three single-echo WFs. The 500-m echoes are from the ground, while the 1000-m echo was triggered 2 m above the ground despite the strong ground signal. Weak backscattering from the signs is seen even for the 500-m pulses in the preceding buffer. In the figure on the right there are three pulses that gave rise to an echo from the top of the chimney. The 1000-m pulse reflected 'entirely' from the top, while the 2/2.7 km pulses have scattered from the chimney and the roof. The 'Sum WF' is the 1000-m WF summed with its shifted (6 ns, 90 cm) copy.
**Fig. 2.** Averaged and normalized peak amplitude values (asphalt=100) in 1064 and 1550 nm pulses (not tabulated, Riegl LMS-Q680i sensor, 700 m AGL, amplitude data linearized using a look-up-table) in different surfaces. Major reflectance differences of green vegetation are seen between 1064 and 1550 nm. Backscattering from tree crowns is not only influenced by the reflectance properties but also by footprint size and the foliage orientation and crown structure. Footprint diameters here are comparable, 0.44 m and 0.40 m.
### 2.3 Waveform attributes
The WF attributes were computed in a straightforward way for what were called noise-exceeding amplitude sequences (NEAS, also referred to as an echo). More advanced methods for detecting individual echoes that may contribute to a NEAS, and for the derivation of echo attributes such as amplitude and width are discussed in e.g. Roncat et al. (2014). A detailed description of the calculations is given in Hovi et al. (2015). In Hovi et al. (2015) these were found important in tree species identification.
Noise was defined as the variation around a baseline signal, during which the receiver recorded background illumination only. The AGC was a particular challenge in defining the baseline, as the gain influenced both the baseline and the noise. The baseline varied by up to 0.3 amplitude units, which is only 0.25% of the maximum amplitude and neglecting it did not influence the results. The threshold for the start and end of a NEAS was three standard deviations of the observed variation around the baseline (Fig. 3).
In brief, the WF is first moderately low-pass filtered. The NEAS is found by tracing the (interpolated) points where the signal arises from and falls into the noise. The highest amplitude in a NEAS is the peak amplitude, $pA$, which is in strong dependence with discrete-return intensity, but only in well-defined targets. The sum of the amplitude values is energy, $E$. The width attributes were $FWHM$ and $L$. $FWHM$ was computed for the highest peak (defining the half maximum) and $L$ is the total length of a NEAS. The rise speed, $riS$, was assumed to describe ‘the porosity’ of the illuminated target, and was defined as the ratio of the rise time and that of a pulse reflecting from a well-defined surface having the same $pA$. The normalization is needed as the above-the-noise rise time depends on the signal strength (c.f. constant fraction discriminator in range detection).
**Fig. 3.** WFs of two adjacent and collinear pulses in a pine forest. The 59-cm WF is shifted up by one unit to be visible. The first echoes have reflected from a branch 24.7 m high and had an XYZ separation of 0.16 m, while the ground intersection points had an XY separation of 0.17 m. The scan zenith angle was 8 degrees. Ground echo can be seen for the wider pulse. The attribute values for the first NEASs are 50 and 37 for $pA$, 8.3 and 12.6 for $FWHM$, 1.16 and 1.16 for $riS$, and 486 and 518 for $E$.
2.4 Radiometric normalization of the multi-footprint data
As stated, the AGC amplifier is a particular property of the receiver in the ALS60. In the datasets used, the relative gain was 1–2.1, i.e. 0–3.2 dB in power, assuming that the intensity and amplitude values are in linear dependence with at-sensor power. The influence of the AGC was the same on amplitude and intensity data. The $AGC_{voltage} \times$ gain relationship resembles a sigmoid, and the parameters of a normalization function ($f_{AGC}$, a 2nd-degree polynomial) were estimated using homogenous surfaces of varying reflectance (Fig. 4). These surfaces were bitumen roof, old asphalt, fine sand and grass. Fig. 4 shows the uncorrected data for grass.
The normalizing model (Eq. 2) of a pulse $i$, acquisition height $j$ and sensor $k$, used the explanatory variables $R$, $AGC_{voltage}$ and the scan zenith angle. The parameters to be solved were the offset of the amplitude, the parameters of $f_{AGC}$, as well as coefficients $d_j$ for each acquisition height that absorb possible deviations of the transmitted power, recalling that the power could only be set at 1% nominal intervals.
$$A_{norm(i,j,k)} = \left(\frac{R_i}{R_{ref(k)}}\right)^2 \cdot \frac{1}{f_{AGC(k)}} \cdot (A_{obs(i)} - offset_k) \cdot \cos(\theta_i) \cdot d_j$$
(2)
The estimation of $f_{AGC}$ was carried out by first minimizing a weighted sum of coefficients of variation and trend coefficients of univariate regression, indicating removal of trends, over all surfaces and datasets. At this phase 80–95% of the within-surface variance was removed. Parameters offset and $d_j$ were finally estimated by minimizing per surface mean amplitude differences of different acquisition heights in each surface (Fig. 5). In the solution, $d_j$ were 1.00, 3.56, 12.94, and 22.03 and the offset of the $A$
scale was 11.4. If the power had been set perfectly and atmospheric attenuation was zero, the expected values for $d_i$ are 1, 4, 16, and 29.16. Unfortunately, it was realized at this phase that the flight planning software did not account for atmospheric losses. Using the $d_i$ values, attenuation was assessed to be $0.22 \pm 0.06$ dB/km, which is realistic for 1064 nm. The 0.06 dB/km residual standard deviation shows that losses explain the deviations only partly. Power adjustment was done at 1% intervals, which corresponds to 25% relative accuracy in the 500 m data. Overall, the data showed an approximately 1 dB trend from 0.5 to 2.7 km, which is also visible in Fig. 4 as ‘vertical layers’.
Parameter estimation (vicarious calibration) of Eq. 2 could not be done by constraining with the absolute reflectance differences of the calibration surfaces, as only HCRF (hemispherical-conical reflectance factor) measurements at 900 nm by a field spectrometer were available. These measurements were shown to match the 2012 AGC-corrected data reasonably well in Korpela et al. (2013). However, the between-surface differences in Fig. 5 are in the instrument scale. Thus, the linearity of the $A$ data, with respect to power entering the aperture, had to be tested (Section 2.5).
### 2.5 Evaluation of the receiver’s compliance with the radar equation
Concerning the linearity of the ALS50/60 sensors, Korpela et al. (2010) compared the intensity data of Optech ALTM3100 and Leica ALS50-ii sensors to assess the influence of vegetation structure in range correction, i.e., the value of $pwr$ in the range normalization coefficient $(R/R_{ref})^{pwr}$. The results with ALTM3100 were in line with those by Gatziolis (2011) and the optimal $pwr$ term was 2.1–2.8 for trees and could be logically linked to species-specific structure and echo type. Thus, the $pwr$ term was close to 2 for ‘only’ echoes that are strong and was highest, nearly 2.8, for ‘first-of-many’ echoes in Scots pine that have the most ‘diffuse’ crowns. However, the ALS50-ii intensity data did not comply with the radar equation, and the optima were
unclear with the pwr term even below 2. A simple linear model with zero offset was used, which for ALS50-ii (which is similar to ALS60) is now known to be incorrect (Fig. 6), while in ALTM3100 the zero intensity was closer to the no-signal level.
The amplitude and intensity values in ALS60 were stored as 8-bit signed integers. ‘No-signal sections’ of the WFs, i.e., the front and tail parts of the 256-sample-long WFs, showed variance, which also responded to the AGC. The no-signal amplitude values were about 10–12, while the first-return intensity data was ‘truncated’ at zero (Fig. 6). The zero point (offset) of the first-return intensity was negative, and was estimated to be about −24 using Eq. (2), while the offset of $A$ was 11.4 (Section 2.4). The estimation of the offset of the intensity scale was almost ill-posed, probably because the linear model (response to at-sensor power) is inaccurate, which is also indicated by Fig. 6.
The 2013_## datasets were acquired to verify the linear response of the ALS60 receiver. In these data, the receiver gain and transmitted power were fixed. The pwr term of the $R$-normalization coefficient was again estimated using surfaces of varying reflectance. Fig. 7 implies that the peak amplitude data (restricted to pulses that had triggered a DR echo) are in a non-linear dependence with the power entering the aperture. Fig. 8 shows a power $\rightarrow A$ mapping that creates the pattern in Fig. 7. It deviates only slightly from a linear model. The component(s) that cause this modest non-linearity in the fixed receiver gain mode remain unidentified. Possibilities include the Avalanche photodiode and its transimpedance amplifier, other amplifiers and the digitizer. Furthermore, the non-linearity of low and high signals (Figs. 6, 7, 8) may be due to different components.
Fig. 6. The relationship between first-echo WF peak amplitude and DR intensity in a scene comprising bitumen, tile and metal roofs, short grass, concrete and asphalt. Echo triggering requires a $pA$ of approximately 20.
Fig. 7. Estimated power term of range normalization (y-axis) for the $pA$ data (x-axis) in the 2013_ datasets. In a linear sensor the power factor is two. Analyses apply to horizontally aligned surfaces of varying reflectance in pulses that gave rise for an observation in 700/800, 700/900 and 800/900 m pulses. The optimal power term minimized the squared sum of normalized differences for a particular surface and data pair. The observation is placed at the $pA$ value corresponding to the 800 m acquisition.
Fig. 8. A non-linear sensor response curve that produces the pattern in Fig. 7. The x-axis depicts the power entering the aperture, while the y-axis shows the corresponding peak amplitude value.
2.6 Variation of the received waveforms due to sensor effects
The mean FWHMs in well-defined targets were 10.1–11.3 ns and 6.8–7.9 ns in the 22–59 cm and 11 cm footprint data, respectively. The dependence of FWHM on $pA$ was different for the 4 and 9 ns pulses. FWHM increased moderately with $pA$ in the 9 ns data, while it decreased considerably with increasing $pA$ in the 4 ns data (Fig. 9).
Fig. 9. Averaged, centered and normalized WFs of 4 (7) and 9 (10) ns pulses (transmitted, received). The response of FWHM to surface brightness differs with pulse length. In 4 ns pulses the recorded WF rises in 5–7 ns. The shape difference of the rising part of the 4 ns pulses is caused by the receiver’s impulse response, which differs for weak and strong signals. The shape of the 9 ns pulses is retained better in bright and dark targets.
These findings imply that the receiver’s response to the fast rising 4 ns pulses differed slightly between weak and strong signals. The CV (coefficient of variation) of FWHM was 1–1.5%. To compare, in some low (300 m) altitude datasets that were acquired in Hyytiälä with the Riegl LMS-Q680i sensor, the FWHM depends strongly on the signal strength, because of the saturation of the amplitude scale due to the low altitude, while 700 m data with the same sensor showed practically no FWHM × pA dependence. The echo width should depend only on the target properties, which was not exactly true in the ALS60 data used here.
2.7 Evaluation of the radiometric normalization
Radiometric normalization (Eq. 2) was validated across the landscape, using surfaces that were known from the field and/or identified in large-scale aerial images. Each surface was represented by more than 100 pulses per footprint size. The relative differences in Table 3 are normalized with respect to the 11-cm data. The differences were correlated between the 22, 44 and 59 cm data with $R^2$ of 0.74, 0.86 and 0.89, which implies that the 11 cm data did have offsets due to the within- or between-strip fluctuation of the low (4% of maximum) transmitted power. Using 9% of the maximum would have been at the eye-safety limit. Some of the 11-cm strips were repeated the following day with a 20% (0.8 dB) higher pA. Table 3 also shows that the relative
deviations were negatively correlated with the surface brightness, i.e., the $pA$ values in dark surfaces and in the 22–59 cm data were modestly higher than in the 11 cm data. This can be caused by the different receiver response as the 11-cm data had shorter pulses with a fast rise in the WF.
As stated, the 2012 acquisitions unintentionally did not account for atmospheric losses. The at-sensor threshold backscattering that triggers an echo depends on the instrument, not the acquisition height as such. The atmospheric losses however caused this threshold to be higher, in terms of the required target backscatter cross-section, in the high-altitude data. The effect was estimated to be 1.0 dB or 26% between the 500-m and 2700-m data (Section 2.4). Based on the 900-nm field HCRF measurements of different bitumen surfaces that were carried out in 2008/2009 (Korpela et al. 2011), the minimal echo-triggering ‘reflectance’ was about 0.03 in 11 cm data and 0.038 in 59 cm data. For this study I found an even darker bitumen roof, in which most of the 500 m pulses had produced an echo, while the probability reduced with acquisition height (** in Table 3).
This 0–1 dB ‘echo-triggering bias’ was thus present in all data. It could be observed in comparable (same LiDAR strips, adjacent targets) data that covered a dark and gently sloping bitumen roof and an adjoining meadow (** in Table 3). In comparison to the 11 cm data, the $pA$ in the 22–59 cm data were 15% lower in bitumen, while the values for meadow were 1–4% higher. The roof had a reflectance variation due to litter (nearby trees), mosses and lichens (north vs. south exposure). The 11-cm data had more no-echo pulses, which means that the 11 cm samples were biased towards brighter roof patches, while the larger footprints averaged the variation. Overall, the validation in different surfaces showed that apart from very dark surfaces, the mean $pA$ was not affected much by the footprint size (Table 4). Furthermore, the differences in the dark validation surfaces were in all likelihood caused by small scale reflectance variation. One of the dark surfaces was a pond with dense canopies of sub-emergent mosses next to a sedge fen. The same scale-related phenomenon was observed here. The mosses did not always give rise to an echo and the 11-cm targets probably comprise small, dense and drier moss patches.
Table 3. Relative differences (%) in peak amplitude in different validation surfaces. Values are normalized to the 11 cm data except for the first row. Values in which the peak amplitude was lower than in the 11 cm case are positive. Surfaces marked with * or ** are adjacent and qualify for pair-wise comparisons. Surfaces #1, #2 and #3 are presented in Section 2.10.
| Surface | pA | 11 | 22 | 44 | 59 |
|--------------------------------|------|------|------|------|------|
| Field | 107 | – | 0.0 | -2.2 | 1.5 |
| | 115 | 0.0 | 2.1 | 1.6 | 3.8 |
| Sedge fen | 98 | 0.0 | -4.2 | -6.7 | -5.2 |
| Low-sedge bog | 108 | 0.0 | 0.7 | -5.8 | -4.5 |
| Sedge fen * | 107 | 0.0 | -1.9 | -1.6 | -2.0 |
| Sub-emergent mosses * | 24 | 0.0 | 12.4 | 10.9 | 17.3 |
| Fine sand | 87 | 0.0 | 0.5 | 1.4 | 2.1 |
| Dark bitumen roof ** | 17 | 0.0 | 12.9 | 15.2 | 15.1 |
| Meadow ** | 115 | 0.0 | -4.2 | -3.3 | -1.2 |
| Forest roads | 66 | 0.0 | -2.0 | -2.4 | -1.9 |
| | 77 | 0.0 | 0.9 | 2.0 | 2.6 |
| #1 | 80 | 0.0 | 0.5 | 3.0 | 2.9 |
| #2 | 70 | 0.0 | 3.0 | 3.2 | 3.4 |
| | 73 | 0.0 | 1.1 | -4.4 | 2.2 |
| #3 | 42 | 0.0 | 3.7 | 3.2 | 5.9 |
| | 50 | 0.0 | 3.2 | 2.5 | 4.3 |
| | 77 | 0.0 | -3.1 | -5.7 | -5.2 |
| Asphalt roads | 51 | 0.0 | 4.8 | 4.2 | 6.4 |
| | 49 | 0.0 | 2.5 | 2.8 | 4.6 |
| | 42 | 0.0 | 0.5 | 0.4 | 2.9 |
2.8 Scale-dependent reflectance variation
Table 4 shows the CV of pA in surfaces of varying small-scale variation. Asphalt and bitumen (different from that in Table 3) showed a stable CV. The between-pulse variation reduced with increasing footprint size in hay and in the mire surfaces owing to their decimeter-scale spatial variation. The mire was a mosaic of 20–60 cm wide dry hummocks surrounded by wetter white mosses and sedges. The multi-footprint data provided logical results. The lowest CV was 2.8%. Bitumen showed relatively high CVs, because of the low relative measurement accuracy of weak signals. A comparison between the AGC on (and normalized) and AGC off acquisitions could be made with the 2013_07 dataset. The CVs were slightly lower in the AGC off data: 8.9→8.6%, 6.9→6.7%, 5.4→4.9% and 4.9→4.0% in bitumen, asphalt, sand and grass, respectively. These results suggest that the AGC
normalization was effective and, conditioned on the stability of the targets, that the precision of the instrument was comparable between 2012 and 2013.
**Table 4.** Coefficient of variation (CV, %) of peak amplitude in different man-made and natural surfaces. Surfaces are in decreasing order of brightness (About 0.45→0.04 in reflectance).
| Surface | Footprint, cm |
|------------------|---------------|
| | 11 | 22 | 44 | 59 |
| Hay, < 50 cm | 7.0 | 5.8 | 4.6 | 3.6 |
| Short grass | 4.9 | 4.3 | 4.0 | 3.5 |
| Mire surface | 9.7 | 8.3 | 6.3 | 5.4 |
| Fine sand | 4.5 | 3.5 | 3.0 | 2.8 |
| Asphalt, old | 7.4 | 7.3 | 6.9 | 7.2 |
| Asphalt, new | 5.1 | 4.9 | 5.0 | 5.1 |
| Gravel road | 11.4 | 9.7 | 7.7 | 6.9 |
| Bitumen | 9.4 | 9.4 | 9.0 | 8.6 |
### 2.9 Transmission losses in wires and tree crowns
Next, the data was evaluated using careful measurements of transmission losses caused by three parallel (1.2 m separation, 152 m in total) power line cables 11 m above a spring barley field. The (bright) insulators in the poles were positioned using forward ray-intersection point estimates obtained in large-scale aerial images and the LiDAR points and constrained by the 1.2 m separation. Line-to-line distances were used to find pulses that had potentially intersected a cable. The limiting distance was the 95% point of the footprint diameter, i.e. 0.83, 0.62, 0.31 and 0.16 m for the 59, 44, 22 and 11 cm footprint diameters, respectively. Assuming that the pulse (intersection, 3D LiDAR positions were available) geometry was correct, that the irradiance profile of pulse cross-section is Gaussian, that the intersection probabilities were equal and that the cables had a 1.2 m separation, the theoretical losses for a cable of 1 cm thickness are 2.8, 1.5, 0.7 and 0.8%. The observed relative differences of $\rho A$ were 2.6, 1.6, 0.7 and 1.1% for the footprint sizes of 11, 22, 44 and 59 cm. The results were thus in line with expectations.
The transmission losses were next examined in an approximately 100-year-old barren pine stand having one dominant canopy layer only. Fig. 10 shows results of relative transmission losses measured by the peak amplitude of ground returns (of single return pulses) as a function of the 2D trunk-pulse distance at the 60% relative tree height, which approximates crown base height rather well in this stand. The crown width estimates ($N=213$) were obtained by fitting a surface of revolution to the LiDAR point cloud as in (Korpela et al., 2011; Hovi et al., 2015). The ground flora comprises three moss species and patches of reindeer lichens. Tree heights are 15.9–26.5 m. In Korpela (2008), the ground flora in this stand was mapped using a 960 m$^2$ photo mosaic and no correlation was found in the spatial distribution of the bottom flora and the trees, i.e., the reflectance properties of the
forest floor are spatially stationary. The pA of ground returns of pulses that had passed through the crown decreased with increasing footprint size (Fig. 10). Thus, less energy reached the ground when the footprint became larger and the single ground echo was weaker. Fig. 11 shows how the losses, as measured by energy of the first echo, correlate negatively with the peak amplitude of the ground return in two pine bogs. The other site has Sphagnum species and tussock cottongrass in the understory, while the ground flora of the other bog comprises a dense canopy of 30–60-cm-high wild rosemary. The reflectance differences of the low vegetation are reflected by the peak-amplitude values of the second echoes.
**Fig. 10.** Relative peak amplitude of (single) ground returns below a Scots pine canopy as a function of the pulse-trunk distance (m) at the crown base height. The amplitudes were normalized to data 3–5 m away from the trunks. A distance of zero corresponds to the case in which the pulse intersected the crown perimeter at the crown base height. A negative value means that the intersection was inside the crown model.
**Fig. 11.** Joint distributions of the energy of the first echo and the peak amplitude of the second echo in 9 ns pulses (1, 2 and 2.7 km data) from two barren pine bogs that differ in understory flora. The pulses had a single WF peak in each of the two echoes.
2.10 WF features in single-species tree canopies
Next, the response of WF attributes to the varying footprint diameter was evaluated in tree canopies. Three 30-year-old planted, non-thinned, pure and dense stands of Scots pine, Norway spruce and Silver birch adjacent to a forest road were chosen. The road surface was essential for controlling between-dataset offsets in $pA$. The roads are marked as #1, #2 and #3 in Table 3. The small relative offsets of $pA$ were eliminated from the $pA$ and $E$ differences in Table 5. The energy ($E$) attribute varied least with footprint size. The relative variance was also lowest in $E$ with CVs ranging from 10 to 20%. Interestingly, $E$ increased by 8, 6 and 7% with footprint size in pine, spruce and birch, respectively. The differences between species were largest in $pA$ and $E$, which is in line with the findings in Hovi et al. (2015). The relative decrease of $pA$ with increasing footprint size is interesting, as it was lowest for birch (16.6%), followed by pine (19.4%) and spruce (25.9%). An extra stand of Black alder showed a decrease of 12.6%, but this stand did not have a control surface. These findings indicate weak species-specific ‘signatures’.
Table 5. Mean values and CV (%) of first-echo WF attributes in 30-year-old pure pine, spruce and birch stands. The LiDAR plots (1300 m$^2$) contained approximately 70 trees. FWHMs were normalized to the mean values from well-defined surfaces. The $E$ attributes of 11 cm (4 ns pulses) were multiplied by 1.307, which is the average ratio between 11 cm (4 ns) and 9 ns data in well-defined surfaces of varying reflectance.
| Species | Footprint | FWHM | Ri/S | pA | Energy |
|---------|-----------|--------|-------|-------|--------|
| Birch | 11 | 1.51 (38) | 1.35 (28) | 66.1 (34) | 1141 (19) |
| | 22 | 1.49 (38) | 1.30 (25) | 67.9 (30) | 1163 (14) |
| | 44 | 1.68 (40) | 1.51 (26) | 58.7 (29) | 1205 (13) |
| | 59 | 1.77 (39) | 1.63 (25) | 55.2 (28) | 1235 (12) |
| Spruce | 11 | 1.33 (40) | 1.21 (26) | 60.2 (33) | 941 (20) |
| | 22 | 1.40 (41) | 1.25 (28) | 57.2 (32) | 935 (16) |
| | 44 | 1.66 (47) | 1.45 (30) | 49.6 (34) | 982 (15) |
| | 59 | 1.84 (47) | 1.60 (30) | 44.6 (33) | 995 (15) |
| Pine | 11 | 1.45 (37) | 1.30 (27) | 53.4 (30) | 869 (16) |
| | 22 | 1.49 (35) | 1.32 (24) | 51.8 (25) | 875 (12) |
| | 44 | 1.68 (34) | 1.53 (23) | 46.1 (22) | 916 (10) |
| | 59 | 1.77 (37) | 1.63 (23) | 43.1 (25) | 932 (13) |
FWHM increased as footprint diameter increased, which is according to expectations. When the footprint size increased from 11 to 59 cm the relative increase in FWHM was 17% in birch, 38% in spruce and 22% in pine. The rise time attribute showed a similar pattern with smaller differences between species, compared to echo width, i.e. FWHM. Depending on the range detection algorithm, the increase of rise time can influence the height distribution of canopy echoes such that the wide footprint...
points have lower heights. This phenomenon was not tested here, because it would have required the implementation of a range detection algorithm and careful geometric calibration of the LiDAR data. It can also be seen that while $pA$ decreases (dampens) with footprint size, $E$ increases moderately as return pulses stretch.
3. Discussion
This study was a largely successful attempt to acquire multi-footprint LiDAR data in which the SNR was fixed for well-defined surfaces, and to carry out radiometric analyses with these data. Basic hypotheses concerning the influence of footprint diameter were verified, although comprehensive testing in tree canopies was left to the future. The datasets that have been acquired are unique as are the investigations performed using those data. Hopkinson (2007) studied the influence of footprint size and acquired dual-footprint data using a sensor in which the divergence could be altered. He reports that while all other parameters remained constant, the intensity values, instead of not changing in well-defined surfaces or changing by a factor of 0.38 in linear targets, reduced by a factor of 0.22 when the divergence was changed from 0.3 to 0.8 mrad.
The ALS60 sensor had a fixed beam divergence, but owing to the adjustable output power, footprint diameters ranging from 11 to 59 cm could be realized. The used power level was a compromise as over 50% more could have been used. The use of full power would have increased the signals by 3 dB, but in that scenario, only 11 dB of power adjustment margin would have been available and the maximum acquisition height would have been less than 2 km. Furthermore, it was recommended to reduce the power of the lowest acquisition height.
Unexpectedly the campaign planning software did not account for atmospheric losses and a 1 dB trend remained in the data. It caused the minimal echo-triggering target backscatter cross-section to increase with acquisition height and footprint size. A fairly realistic estimate of the atmospheric losses, 0.22 dB/km, was derived as a sub product of the radiometric normalization. Yet, it could not be ruled out that the observed trend was partially caused by systematic errors in the adjustment of the transmitted power.
Backscatter reflectance or peak amplitude normalization was done using natural surfaces assuming that the peak amplitude observations are in a linear relationship with the peak at-sensor power (irradiance). The linearity assumption had to be validated, using a separate multi-height LiDAR dataset, which was acquired for that purpose. The validation was based on the dependence between the power received and the spherical losses from well-defined surfaces. Fixed pulse power and receiver gain were applied and the signals from well-defined surfaces were, according to the radar equation, 1.6 (2 dB) times stronger from 700 m compared to 900 m. Setting the output power and receiver gain correctly was crucial. While operating, the AGC circuit in the receiver can adjust and compensate the internal signal levels by up to 3 dB. The brightest targets had pulses that were saturated and distorted in the 700 m DR intensity data. These targets were omitted from the calibration (Fig. 7), because it
remained unclear if the saturation applied to the WFs as well. For example, dense canopies of (planophile) raspberry showed saturated intensity data. Because of the 2-dB difference in power received, the darkest objects that still triggered an echo in the 700 m data would not show in the 900 m data, and the brightest targets were rejected as the dynamic range of the receiver was exceeded in the 700 m acquisition. Natural and man-made surfaces were used in testing the linear response. The theoretical differences in power received are 23.4% and 26.5% between 700 m vs. 800 m and 800 vs. 900 m, respectively. The within-surface CVs of peak amplitude ranged from 3 to 15%, which means that the estimation of the power term of the range normalization coefficient was susceptible to random errors. A 2% error the mean peak amplitude (of a surface, e.g. in 800 m data) results in a coefficient estimate that is off by 0.2. The influence of the incidence angle, or rather, the scan zenith angle, was weak and could be omitted in all analyses.
The resulting normalized peak amplitude data (Eq. 2) could be described as interval-scale observations of backscatter cross-section or reflectance in well-defined targets, at 1064 nm. The smallest values corresponded to a reflectance of about 0.03, which was observed in one very dark bitumen roof. This estimate was based on previous field measurements of HCRF carried out for research in the reflectance calibration of aerial images (Korpela et al., 2011). If full power had been applied (eye-safety, 3 dB higher), the echo-triggering reflectance would have been approximately 0.015–0.02. In Korpela et al. (2013) it was found, using the 2012_10 (22 cm) data, that the minimum relative silhouette area in pine and spruce branches that produced an echo was about 10%. Fig. 2 suggests that the ratio of mean peak amplitude in bitumen and pine crowns is about 2.5 and the 900 nm HCRF reference for the bitumen roof in Fig. 2 was 0.05. If the reflectance of pine needles is 0.3–0.4, the relative silhouette of 0.1 corresponds to a ‘well-defined-surface reflectance’ of about 0.03–0.04. These calculations are coarse and inaccurate, but suggest that some coherence exists between earlier findings in canopies and those that were made here about well-defined surfaces.
The observations were shown to be influenced by instability of the transmitted power, in the low-power, low-altitude, small-footprint data. The sensor manufacturer also pointed out this possibility and the repeated 500 m acquisitions in the following day also suggested this. It remained unclear what kind of temporal patterns these variations had, but at least a campaign-level offset was found. It may well be that the power fluctuations are smaller in sensors that apply a few fixed power levels and dual-channel receivers.
The impulse response and bandwidth characteristics were analyzed and only minor signal-strength dependent deformations of the WFs were observed in the 4 ns pulses. The deformations were neglected in the analyses here. Our experience with the Riegl LMS-Q680i sensor (cf. Armston et al., 2010) is worth noting here. The output power of LMS-Q680i has fixed levels and is tightly coupled with the pulse repetition frequency (PRF). Some low altitude (300 m) WF data were ‘overexposed’ and mapped to the non-linear part of the amplitude scale, which created an unwanted dependence of the echo width on signal strength, which was not present in a 700-m dataset acquired using the same PRF. LMS-Q680i has two receivers, and WF samples are stored from
both receivers when necessary. Such dual-receiver, high- and low-gain designs are common and for example the new Leica ALS80-series sensors employ such a design. The gain difference of the channels is much larger than the 3 dB gain range of the AGC in ALS60. Compared with having two gain levels, the AGC in ALS60 could be considered a nuisance. In all, managing the wide dynamic range of LiDAR data acquisitions calls for different solutions, which are implemented in the transmitter and/or in the receiver. It would be better for radiometric analyses of LiDAR data if the sensor properties, i.e., an absolute calibration, were known. The reverse-engineering type of analyses that were carried out in this study would then not be needed for carrying out radiometrically quantitative LiDAR remote sensing.
The presented analyses revealed that the peak amplitude data in ALS60 was not in a very exact linear relationship with the power entering the aperture. However, this finding was obtained with the AGC-amplifier set at fixed gain and it cannot be ruled out that the AGC amplifier, when operating, causes further problems. However, this is unlikely, because the BW of such amplifiers can be expected to be larger than the BW of the digitizers or the photon detector. In Hovi et al. (2015), no evidence was found of systematic differences in peak amplitude or WF shape that were due to the two oscilloscopes behaving differently. As said, the oscilloscopes take turns in processing pulse pairs into WF data.
The dependence of ‘unregistered’ transmission losses on footprint size, preceding the triggering of waveform recording, was shown in power line cables and in pine crowns. The losses were from 2.6 to 0.5% in the cables and from 5 to 16% in the trees, for the footprint diameters of 11–59 cm. Canopies are comprised of gaps and clumped scatterers, while the cable is a single reflector obstructing the energy from reaching the ground, where the losses were detected as reductions in the average peak amplitude of hundreds of pulses. The sites used for loss detection had low and homogenous or at least spatially stationary ground reflectance variation. The influence of canopy transmission losses can be compensated to some degree in discrete-return (Korpela et al., 2012) and WF data (e.g. Richter et al., 2015) by imposing constraints such as constant reflectance or configuration of needles in a tree crown, which reduces the ambiguities related to backscatter cross-section. This was demonstrated by the comparison of the two bogs.
The relative variation of peak amplitude was assessed in surfaces of varying small-scale structure (Section 2.8). Fine sand, which has a HCRF of about 0.2 at 1064 nm, showed the lowest CV, 4.5% in 11-cm data, and as low as 2.8% in the 59-cm data. The target was a volleyball field with a ‘moderately bumpy’ surface and a few dry leafs blended in the sand. Bitumen was least influenced by footprint size, but the CVs were higher, on the order of 9%, because of the low reflectance. The results were logical, concerning the influence of the footprint size and the averaging effect in large footprint data. The observed CVs suggest that the instrument is rather stable or precise as there were always data from two or more strips per test surface and footprint size.
Regarding the influence of the footprint size on the tested first-echo WF attributes in tree canopies, there were minor species-specific differences or ‘signatures’. For example, the peak amplitude decreased most in Scots pine stand, the structure of which
is less clumped or ‘more diffuse’ compared to Norway spruce. The reduction in peak amplitude with increasing footprint size was smallest in the case of birch, which have compact crowns at the age of 30 years. Actually these findings that relate to canopy structure are line with those concerning the response of intensity data to $R$ variation (Korpela et al., 2010a; Gatsiolis, 2011), because the decrease of power received that is due to an increase of divergence is also dependent on the structure (Section 1). As hypothesized, the echo width increased with footprint size and the relative increase was largest in the canopy of spruce, which is reasonable, given the ‘clumped structure’ and ‘layered compact branches’. The energy was the largest in birch, and this could possibly be due to the slightly higher reflectance of the leaves. Then again, backscattering in birch is strongly influenced by leaf arrangement as well (Hovi and Korpela, 2014). The tests in forest canopies were hampered by the remaining between-dataset SNR differences, which were subtracted using the roads adjacent to the test stands as control surfaces.
4. Conclusions
This study showed that it is possible to acquire multi-footprint data that is radiometrically comparable for well-defined surfaces, using a commercial topographic LiDAR sensor. The study has shown how to carry out careful analyses to verify that the WF amplitude observations are on the interval scale. The impulse response and bandwidth characteristics of the sensor were tested in a simple manner to verify that the WF features depend only on the target-pulse interaction, in order to draw correct conclusions from observations concerning the influence of the footprint size. Many of the outlined hypotheses regarding the influence of footprint size could be confirmed in real data. These included the widening of the echoes, the increase of return energy, and the decrease of maximal amplitude in tree canopies, all with increasing footprint size. These effects had species-specific traits. Transmission losses as small as 0.5% could be detected by an appropriate selection of the test site and by using a large amount of pulses to cancel out the random variation. In conclusion, multi-footprint data may offer some advances in the estimation of tree or forest canopy structure. However, the species-specific differences are likely quite weak. A thorough investigation is needed to see if multi-footprint LiDAR can enhance forest remote sensing. For example, the estimation of canopy closure might benefit from dual-footprint data. The experience that was gained here constitutes a good starting point.
Acknowledgements
Advice about the ALS60 sensor by Dr. Ron Roth at Leica Geosystems was indispensable. I also thank the personnel at Finnmap in Helsinki, especially Mr. Jonne Davidsson, Mr. Felix Rohrbach and Mr. Jussi Havia, who arranged the unique data acquisitions with the ALS60. Dr. Arame Hovi wrote the original WF attribute extraction code. Dr. Kyle Eyvindson, Dr. Reija Haapanen and Dr. Kim Östman kindly revised the text. The Academy of Finland provided the funding for this work.
Armston, J., Disney, M., Lewis, P., Scarth, P., Phinn, S., Lucas, R., Bunting, P. & Goodwin, N. (2013). Direct retrieval of canopy gap probability using airborne waveform lidar. Remote Sensing of Environment 134:24–38.
Disney, M.I., Kalogirou, V., Lewis, P., Prieto-Blanco, A., Hancock, S. & Pfeifer, M. (2010) Simulating the impact of discrete-return lidar system and survey characteristics over young conifer and broadleaf forests, Remote Sensing of Environment 114(7): 1546–1560.
Gatziolis, D. (2011). Dynamic range-based intensity normalization for airborne, discrete return LiDAR data for forest canopies. Photogrammetric Engineering & Remote Sensing. 77(3): 251–259.
Goodwin, N.R., Coops, N.C. & Culvenor, D.C. (2006). Assessment of forest structure with airborne LiDAR and the effects of platform altitude. Remote Sensing of Environment 103(2): 140–152.
Hopkinson, C. (2007). The influence of flying altitude, beam divergence, and pulse repetition frequency on laser pulse return intensity and canopy frequency distribution. Canadian Journal of Remote Sensing 33: 312–324.
Hovi, A. & Korpela, I. (2014). Real and simulated waveform-recording LiDAR data in juvenile boreal forest vegetation. Remote Sensing of Environment 140: 665–678.
Hovi A., Korhonen L., Vauhkonen J. & Korpela I. (2015). LiDAR waveform features for tree species classification and their sensitivity to tree- and acquisition related parameters. Remote Sensing of Environment 173: 224–237.
Hovi, A. (2015). Towards an enhanced understanding of airborne LiDAR measurements of forest vegetation. Dissertationes Forestales 200. 69 p.
Korpela, I. (2008). Mapping of understory lichens with airborne discrete-return LiDAR data. Remote Sensing of Environment 112(10): 3891–3897.
Korpela, I., Ørka H.O., Heikkinen V., Tokola T., & Hyypä J. (2010a). Range- and AGC normalization of LIDAR intensity data for vegetation classification. ISPRS Journal of Photogrammetry and Remote Sensing 65(4): 369–379.
Korpela, I., Ørka, H.O., Maltamo, M., Tokola, T. & Hyypä, J. (2010b). Tree species classification using airborne LiDAR – effects of stand and tree parameters, downsizing of training set, intensity normalization and sensor type. Silva Fennica 44(2): 319–339
Korpela, I., Heikkinen, V., Honkavaara, E., Rohrbach F. & Tokola, T. (2011). Variation and anisotropy of reflectance of forest trees in radiometrically calibrated airborne line sensor images – implications for species classification in digital aerial images. Remote Sensing of Environment 115(8): 2062–2074.
Korpela I., Hovi A. & Morsdorf F. (2012). Understory trees in airborne LiDAR data - Selective mapping due to transmission losses and echo-triggering mechanisms. Remote Sensing of Environment 119: 92–104.
Korpela I., Hovi A. & Korhonen L. (2013). Backscattering of individual LiDAR pulses explained by photogrammetrically derived vegetation structure. ISPRS journal of photogrammetry and remote Sensing 83: 81–93.
Maltamo M. & Packalén P. (2014). Species-specific management inventory in Finland. In: Maltamo M., Næsset E., Vauhkonen J. (eds.). Forestry applications of airborne laser scanning: Concepts and case studies. Springer, The Netherlands, p. 241–252.
Mallet, C. & Bretar, F. (2009). Full-waveform topographic lidar: State-of-the-art, ISPRS journal of photogrammetry and remote sensing 64(1):1–16.
Ørka, H-O., Næsset, E., Bollandsås, O., M. (2010). Effects of different sensors and leaf-on and leaf-off canopy conditions on echo distributions and individual tree properties derived from airborne laser scanning. Remote sensing of environment 114(7): 1445–1461.
Richter, K., Blaskow, R., Stelling, N., & Maas, H. G. (2015). Reference Value Provision Schemes for Attenuation Correction of Full-Waveform Airborne Laser Scanner Data. ISPRS Ann. Photogramm. Remote Sens. Spat. Inf. Sci. 1: 65–72.
Roncat, A., Briese, C., Jansa, J. & Pfeifer, N. (2014). Radiometrically calibrated features of full-waveform lidar point clouds based on statistical moments. IEEE Geoscience and remote sensing letters, 11(2): 549–553
Wagner W., Ullrich A., Ducic V., Melzer T., Studnicka N. (2006), Gaussian decomposition and calibration of a novel small-footprint full-waveform digitising airborne laser scanner. ISPRS Journal of Photogrammetry and Remote Sensing 60: 100–112.
Wagner, W. (2010). Radiometric calibration of small-footprint full-waveform airborne laser scanner measurements: Basic physical concepts, ISPRS Journal of Photogrammetry and Remote Sensing 65 (6): 505–513.
Vauhkonen, J., Maltamo, M., McRoberts, R.E., & Naesset, E. (2014). Introduction to forestry applications of airborne laser scanning. In: Maltamo M., Næsset E., Vauhkonen J. (eds.). Forestry applications of airborne laser scanning: Concepts and case studies. Springer, The Netherlands, p. 1–16. |
The maquila industry is a major source of employment in Mexico, providing jobs for millions of people. However, many workers in the industry face significant challenges and issues related to their working conditions and rights.
One of the main concerns for maquila workers is the low wages they receive. Despite the fact that the industry is one of the largest employers in Mexico, many workers still struggle to make ends meet. This is particularly true for women, who often earn less than men and are more likely to be employed in lower-paying positions.
Another issue facing maquila workers is the lack of job security. Many workers are employed on a temporary basis, with no guarantee of continued employment or benefits such as health insurance or retirement plans. This can create a sense of uncertainty and instability for workers and their families.
In addition to these economic concerns, maquila workers also face a range of safety and health risks. The factories where they work are often crowded and poorly maintained, increasing the risk of accidents and injuries. Workers may also be exposed to hazardous chemicals and other workplace hazards without adequate protection.
Despite these challenges, maquila workers have been able to organize and advocate for their rights. In recent years, there have been several high-profile strikes and protests by workers demanding better wages, safer working conditions, and improved benefits. These efforts have helped to raise awareness about the issues facing maquila workers and put pressure on companies and government officials to address them.
Overall, while the maquila industry has brought significant economic benefits to Mexico, it is important to recognize the challenges faced by workers in the industry. By addressing these issues and promoting fair labor practices, we can help ensure that maquila workers are able to achieve a better quality of life and a more secure future.
closed—to any worker who agreed to withdraw from the suit and accept roughly ten percent of the compensation owed.
"Mexico," in the words of one organizer, "has one of the most progressive labor codes in the world. It is enforcing the law that is the problem." Most of the Empaque workers, skeptical of the possibility of enforcing their rights, accepted the settlement. Those who remained involved in the suit—and their relatives—were blacklisted from all industrial employment in Matamoros. Their only means of survival was to take in laundry or ironing, or cook tamales to sell in the street. "It was very hard," confessed one woman. "The schools did not care whether we had good jobs or not—we still had to buy shoes and uniforms and books for our children. With this kind of work, if it rains you cannot do anything. Sometimes we did not even have enough to eat." By the end of 1985, a scant 20 workers continued to press for their legitimate compensation, decreasing to 15 by the time the settlement was made in 1987.
In mid-1985, the American Friends Service Committee (AFSC), a Quaker service organization with various educational and advocacy programs along the border, responded to requests to assist the Empaque holdouts.
Research coordinated by AFSC revealed that Booth Fisheries, which owned Empaque, was in turn a subsidiary of the giant Sara Lee Corporation, headquartered in Chicago. And Sara Lee Corp., according to activist lawyer Pat de Carlo, who helped AFSC investigate the case, "is very sensitive to publicity—they want to keep a clean corporate image." After meetings with AFSC representatives and a letter writing campaign, Booth Fisheries management agreed to settle with the workers. Although the firm never acknowledged any legal obligation, they did express their "genuine admiration" for the courage of the small group of women who refused to accept less than they were owed. Still pending is a suit the women filed against Agapito, charging him with defrauding the workers.
**An Experience for All the Workers**
Not long before the settlement was paid, Gloria, a former Empaque worker, commented that "if our problem is solved there will be even more workers who don't just submit to what they are told. This experience is not just for us, but for all the workers."
Since the early 1980s, women in many border towns have been coming together to discuss their immediate problems and, where possible, take joint action to solve them. Hoping to avoid the notice
of management and union officials alike, the women begin by meeting in their neighborhoods, away from the factory. Often they organize informal study groups where they learn the details of Mexico's labor code.
Such groups can also be mobilized to pressure management to obey labor laws and call on the union to back them up. "Before," recalled one activist worker, "we might be suspended from work for 29 days if we did something management did not like. Now we will no longer accept this treatment." Women's committees have forced management to provide exhaust fans in one plant and safety equipment in another.
In Matamoros, Reynosa, and other border towns, the women's committees have won dozens of such victories and have begun to coalesce into a regional network over the past year. Recently, when 130 workers went out on strike in Piedras Negras, women from Matamoros raised $1000 to buy food for their families. Yet despite their string of successes, the women still prefer to keep a low profile, fearing blacklisting or even violence if they work too openly.
Health and safety issues are of paramount importance to the maquila women. Of more than a dozen women interviewed for this article, not one was free of chronic pain or illness caused by her working conditions. The Empaque workers told of standing in ice water and working all day with their bare hands in ice while packing shrimp. Women under 30 who suffer from arthritis or rheumatism are common. Electronics workers told of working with toluene, PCBs, and unknown solvents without any safety precautions or equipment. Chloracne, skin cancer, and leukemia are widely reported. But, the workers say, the government Social Security clinics will not classify their complaints as work-related, in one case diagnosing a problem as due to "reluctance to work."
One of the most haunting stories is told by Dr. Isabel de la O Alonso, head of the Matamoros School of Special Education. In the late 1970s, Dr. de la O began to notice a series of children arriving in her school who appeared to be suffering from a birth defect. They were mildly to profoundly retarded and shared a common appearance, "in some ways like that of children with Down's Syndrome," the doctor said. "Their condition does not correspond to anything that has ever been described in medical literature. But clearly they are all the same." In taking histories from parents, the school learned that all of the mothers had worked at a single maquila—Mallory Capacitors—during their pregnancy. Twenty such children have been found. Mallory left Matamoros in 1975 and its parent firm, P.R. Mallory, no longer
exists, swallowed up in a series of corporate takeovers. The women do not know the names of the chemicals they worked with, and an exhaustive search has failed to turn up similar cases anywhere in the world. "At this point," said Dr. de la O, "we see no possibility of seeking compensation." The doctor does what she can to prevent another Mallory, cautioning women about workplace chemical exposure and helping them obtain information about specific hazards.
One successful health project took place in 1985, when a support organization known as SEDEPAC brought maquila workers together with medical students from the Universidad Autonoma Metropolitana of Mexico City. The two groups collaborated on a series of three booklets on workplace hazards, medical rights, and related legal issues. The booklets have been widely used in educational workshops.
The drive to find out what chemicals they are working with has proved to be one of the greatest incentives to cross-border organizing. As one organizer explained, "There is no right-to-know law in Mexico. Management does not give out any information about hazards. In the warehouses, sometimes the Mexican workers will find data sheets on the materials they are unpacking, but these sheets are in English. This is where U.S. unions can help us tremendously."
In the past, mainly churches have been involved in cross-border campaigns—sometimes with very useful results. Some months before the Bhopal disaster, for example, representatives of the United Church of Christ met with Union Carbide officials, conveying to them the concerns of workers at Kemet, as Union Carbide is known in Mexico. As a result, Union Carbide's U.S. management instituted strict safety procedures in the Matamoros plant.
As the workers become better organized, they are increasingly seeking to expand their U.S. contacts to include unions. "We have met with representatives from ACTWU, UAW, and IUE [International Union of Electronics Workers]," commented one organizer. "Workers on both sides of the border have many concerns in common. They may be working in different factories but for the same corporation—this is often the case with garment and textile plants. We met some people from UAW in Buffalo when Trico, which makes windshield wipers, closed its Buffalo plant and moved to Matamoros. Expanding these contacts is a top priority for us."
In another sign of their growing maturity, the women's groups are launching a campaign around utilidades, or legally mandated
profit-sharing benefits. Because of their financial structure, maquilas, in the words of one corporate promoter, are "cost centers rather than profit centers"—that is, their account books are set up so that the Mexican subsidiaries show virtually no profits. By permitting this practice, the Mexican government allows the maquilas to evade its own law requiring 10% of all corporate profits to be distributed among workers. A leaflet from Matamoros points out that workers in a Mexican-owned chemical plant in the border region received between $2000 and $4000 in utilidades last year—in comparison to $1.20 received by workers at a Zenith plant. Notes the leaflet, "if they divide $4000 among 3000 workers in the plant as their share of the profits for making a million televisions in 1986, this means that Zenith made only four cents of profit on each television. If this is true, maybe Zenith should fire all its managers as incompetents!"
Using the Official Unions
Union democracy is another key issue in maquila organizing. The official CTM unions are closely tied in with Mexico's ruling party, the PRI, which has dominated the country's politics since the 1910 revolution. In the isolated border towns, union officials also hold great political power. Workers are overwhelmingly cynical about union leaders, accusing them of corruption and pandering to management.
At the same time, many see the unions as the best structure available for representing their interests. While the Empaque workers defied the official union and won, most activists in Matamoros and other border towns have focused on making the union more responsive to their needs. "The union uses us," Lupe,
a worker activist, said bluntly, "and we should use them." Union officials generally discourage individual workers from taking action, but the union is far more likely to back a group of workers, fearing to lose its hold on the workforce if it stands aside.
In a recent example at a GM plant, women learned that a majority of them were to be temporarily laid off, without required notification of labor authorities. [Although more male workers are entering the maquila workforce as auto plants move south, women are employed in these plants in soldering and similar jobs.] When a small group of women visited the secretary general of their local union, he dismissed their complaint as insignificant; when 100 women returned the next day, the official negotiated an agreement giving the women half their normal salary during the layoff.
Recently in Matamoros, the union has begun training workers to participate in the Comisiones Mixtas, joint commissions with worker, union, and management representatives that are legally mandated to monitor health and safety within each plant. Making these commissions more effective has been a key focus of worker activism for several years, and "our work is having an effect," noted a border organizer. "The union probably figures it should climb on the bandwagon. In the past, they always opposed our efforts."
The question of union affiliation is a controversial one throughout Mexico's labor movement. The border is far more isolated than other industrial areas, and the power of the official confederation is correspondingly greater. There are a few independent unions in Piedras Negras, where maquilas are newer and the CTM is less entrenched. Women there have organized independently in an electrical assembly plant as well as plants assembling toys and novelties. For the most part, though, in the words of one organizer, "working through the CTM is the only feasible approach. Many workers would like to have independent unions, but we are a long way from being strong enough to do that."
The most sustained struggle for union democracy has taken place in Reynosa, where in 1983 more than 12,000 workers staged a spontaneous wildcat strike when one of their number was fired and then arrested on trumped-up charges. Replacing the union leadership emerged as a central issue in the conflict, and after a complicated struggle a new secretary-general was installed and a large number of rank-and-file activists were permitted to take office.
Clara is a Zenith worker in Reynosa who became shift representative for her plant through the 1983 strike. "While we were outside the factory," she recalled, "cars full of masked men with machine guns came up and told us to get out, but we stood our ground. When we went back to work, management told us to forget about our struggle or we would be fired. Later they offered supervisory positions and raises to those who were most active in the movement. But I said no, I prefer to stay with the union."
The large-scale movement of 1983 has died away, Clara said, but some things have changed permanently. "Now, as soon as there is a problem, people come to me to discuss it, and we go together to the union office to try to resolve things. Now we have union meetings—we never had that before. We are talking about having classes and study programs. People are much more ready to protest—for example, if they are short-changed in their pay. Before they would just take it, there was nothing they could do. Now they will come to the union."
"This Is How It Starts"
Clara spoke for many women when she recounted the changes that activism has brought to her life. "Before," she said, "I was silent. I might listen and take notes, but I could not talk. Even when they asked my opinion I would not say anything. No one ever knew who I was. But then I went to some meetings with my friend Mari where she and I were the only ones [from the rank and file]. I had to speak up to support her. And I found that the more I talked, the more people liked it. I couldn't believe it. This is how it starts: you begin to look at your own problems and it wakes you up so that no one can make you turn back."
Like many activist women, Clara had to overcome strong opposition from her family. "My mother was very angry when I joined the strike. She said I would lose my job and gain nothing. She said, you're mixing in this without knowing anything. I told her, well, I'm learning—we all have to start somewhere. I would come home from union meetings at 2:00, 3:00 in the morning and she would accuse me of walking the streets. I said if she didn't like it I would move out, and she relented. By now she accepts that this is who I am." Today, Clara attends school as well as holding union office.
Once started, the courage to speak up keeps growing. When the workers won their demand for a new union leader, José Morales was sent up from Mexico City to take over. "He met with the shift representatives," Clara recalled, "and started telling us what to do. I told him, you forgot to take us into account. First ask us if we agree—and then if we like your ideas we will follow them. We
are the ones who will pay the price if your ideas are bad ones."
In one way or another, Clara's comments were echoed by virtually every woman who was interviewed for this article. Time and again, women spoke of how they had gained the confidence to stand up to authority, to trust their own judgement and voice their own opinions. Through coming together and organizing, their understanding of themselves as women and as workers is undergoing a profound transformation. And that, as much as any concrete gain, is a sign that conditions at the border are changing irreversibly.
References
Maria Patricia Fernandez-Kelly, *For We Are Sold, I and My People: Women and Industry in Mexico's Frontier*, State University of New York Press, 1983. Classic study on women and maquilas.
Costureras' Struggle Continues
The struggle to build an independent union in Mexico's huge garment industry began in the aftermath of the 1985 earthquake that devastated Mexico City, killing 1,000 garment workers and leaving 40,000 more unemployed, most of them women. After more than two years, the September 19 Garment Workers Union—which takes its name from the date of the quake—continues its efforts amid deeply altered, and more difficult, circumstances.
The earthquake, in ripping the facades from buildings housing garment sweatshops, had literally revealed to public view the conditions of extreme exploitation that had endured in the industry unremarked for decades. Because their suffering was so great, the women garment workers, or costureras, came to symbolize the trauma of the quake, and the first uprush of their movement was extensively and sympathetically covered in the Mexican press. The Mexican government, facing heavy criticism from many quarters for its handling of relief efforts, lent its active support, and material and moral support came pouring into the young union from every quarter—feminists, leftist and popular organizations, ecumenical groups, and other independent unions. These people's movements have remained committed to the garment workers' cause, but the media and government have not. A mantle of invisibility once again cloaks the Mexican garment industry.
September 19 now officially represents workers in 12 factories, but
Annette Fuentes and Barbara Ehrenreich, *Women in the Global Factory*, Institute for New Communications, 1983; distributed by South End Press.
*The Global Assembly Line*, a one-hour film by Lorraine Gray showing the impact of the global assembly line on women workers in Mexico, the Philippines and the U.S. Available as video (VHS) or 16mm film, for rent or purchase, from New Day Films, 22 Riverview Drive, Wayne, NJ 07470.
*Listen Real Loud: News of Women's Liberation Worldwide*, a quarterly bulletin on grassroots women's movements in the U.S. and around the world, published by the American Friends Service Committee, 1501 Cherry St., Philadelphia, PA 19102.
June Nash and Maria Patricia Fernandez-Kelly, eds., *Women, Men and the International Division of Labor*, State University of New York Press, 1983.
Its organizing has been hampered by opposition from the official CTM labor federation and by the failure of the government to fairly enforce the labor code. In Mexico, independent unions like September 19 are generally perceived as a threat to government domination of labor through the CTM, which has strong ties to the government party. In many plants, garment workers have been officially represented, without their knowledge or consent, by company unions affiliated with the CTM. Sexual harassment and violence have been used to dissuade women from voting for September 19.
In response, September 19 members are taking their case directly to the public, staging sit-ins and leafletting actions in Mexico City's central plaza. They also are building their organization internally, providing training and education for their members and recently opening a permanent day care center for members' children. Support from other groups inside and outside the country continues to play an important role. Funds for the day care center, for example, came from Mexican and U.S. ecumenical groups.
September 19 organizers will rally support in the U.S. with a speaker-movie tour and photo exhibit in late April of 1988. For more information, contact the Women & Global Corporation's Project, 1501 Cherry Street, Philadelphia, PA 19102—(215/241-7160). |
Evaluation of Unsupervised Emotion Models to Textual Affect Recognition
Sunghwan Mac Kim
School of Electrical and Information Engineering
University of Sydney
Sydney, Australia
firstname.lastname@example.org
Alessandro Valitutti
Department of Cognitive Science and Education
University of Trento
Trento, Italy
email@example.com
Rafael A. Calvo
School of Electrical and Information Engineering
University of Sydney
Sydney, Australia
firstname.lastname@example.org
Abstract
In this paper we present an evaluation of new techniques for automatically detecting emotions in text. The study estimates categorical model and dimensional model for the recognition of four affective states: *Anger*, *Fear*, *Joy*, and *Sadness* that are common emotions in three datasets: SemEval-2007 “Affective Text”, ISEAR (International Survey on Emotion Antecedents and Reactions), and children’s fairy tales. In the first model, WordNet-Affect is used as a linguistic lexical resource and three dimensionality reduction techniques are evaluated: Latent Semantic Analysis (LSA), Probabilistic Latent Semantic Analysis (PLSA), and Non-negative Matrix Factorization (NMF). In the second model, ANEW (Affective Norm for English Words), a normative database with affective terms, is employed. Experiments show that a categorical model using NMF results in better performances for SemEval and fairy tales, whereas a dimensional model performs better with ISEAR.
1 Introduction
Supervised and unsupervised approaches have been used to automatically recognize expressions of emotion in text such as *happiness*, *sadness*, *anger*, etc… Supervised learning techniques have the disadvantage that large annotated datasets are required for training. Since the emotional interpretations of a text can be highly subjective, more than one annotator is needed, and this makes the process of the annotation very time consuming and expensive. For this reason, unsupervised methods are normally preferred in the realm of Natural Language Processing (NLP) and emotions.
Supervised and unsupervised techniques have been compared before. (Strapparava and Mihalcea 2008) describe the comparison between a supervised (Naïve Bayes) and an unsupervised (Latent Semantic Analysis - LSA) method for recognizing six basic emotions.
These techniques have been applied to many areas, particularly in improving Intelligent Tutoring Systems. For example, (D’Mello, Craig et al. 2008) used LSA but for detecting utterance types and affect in students’ dialogue within Autotutor. (D’Mello, Graesser et al. 2007) proposed five categories for describing the affect states in student-system dialogue.
Significant differences arise not only between these two types of techniques but also between different emotion models, and these differences have significant implications in all these areas. While considering emotions and learning, (Kort, Reilly et al. 2001) proposed (but provided no empirical evidence) a model that combines two emotion models, placing categories in a valence-arousal plane. This mixed approach has also been used in other domains such as blog posts where (Aman and Szpakowicz 2007) studied how to identify emotion categories as well as emotion intensity. To date, many researchers have, however, utilized and evaluated supervised methods, mainly based on the categorical emotion model.
In this study, the goal is to evaluate the merits of two conceptualizations of emotions (a *categorical model* and a *dimensional model*) in which an unsupervised approach is used. The evaluation incorporates three dimensionality reduction methods and two linguistic lexical resources.
The rest of the paper is organized as follows: In Section 2 we present representative research of the emotion models used to capture the affective states of a text. Section 3 describes the techniques of affect classification utilizing lexical resources. More specifically, it describes the role of emotion models and lexical resources in the affect classification. In addition, we give an overview of the dimension reduction methods used in the study. In Section 4 we go over the affective datasets used. Section 5 provides the results of the evaluation, before coming to our discussion in Section 6.
2 Emotion Models
There are two significantly different models for representing emotions: the categorical model and dimensional model (Russell 2003).
The categorical model assumes that there are discrete emotional categories such as Ekman’s six basic emotions - anger, disgust, fear, joy, sadness, and surprise - (Ekman 1992). There are a number of primary and unrelated emotions in the model. Each emotion is characterized by a specific set of features, expressing eliciting conditions or responses. Some researchers have argued that a different set of emotions is required for different domains. For instance, the following emotion classes are used in the field of teaching and education: boredom, delight, flow, confusion, frustration, and surprise. The advantage of such a representation is that it represents human emotions intuitively with easy to understand emotion labels.
A second approach is the dimensional model, which represents affects in a dimensional form (Russell 2003). Emotional states are related each other by a common set of dimensions (e.g. valence or arousal) and are generally defined in a two or three dimensional space. Each emotion occupies some location in this space. A valence dimension indicates positive and negative emotions on different ends of the scale. The arousal dimension differentiates excited vs. calm states. Sometimes a third, dominance dimension is used to differentiate if the subject feels in control of the situation or not.
The categorical model and the dimensional model have two different methods for estimating the actual emotional states of a person. In the former, a person is usually required to choose one emotion out of an emotion set that represents the best feeling. On the other hand, the latter exploits rating scales for each dimension like the Self Assessment Manikin (SAM) (Lang 1980), which consists of pictures of manikins, to estimate the degree of valence, arousal, and dominance.
3 Automatic Affect Classification
3.1 Categorical classification with features derived from WordNet-Affect
WordNet-Affect (Strapparava and Valitutti 2004) is an affective lexical repository of words referring to emotional states. WordNet-Affect extends WordNet by assigning a variety of affect labels to a subset of synsets representing affective concepts in WordNet (emotional synsets). In addition, WordNet-Affect has an additional hierarchy of affective domain labels. There are publicly available lists relevant to the six basic emotion categories extracted from WordNet-Affect and we used four of the six lists of emotional words among them for our experiment.
In addition to WordNet-Affect, we exploited a Vector Space Model (VSM) in which terms and textual documents can be represented through a term-by-document matrix. More specifically, terms are encoded as vectors, whose components are co-occurrence frequencies of words in corpora documents. Frequencies are weighted according to the log-entropy with respect to a $tf-idf$ weighting schema (Yates and Neto 1999). Finally, the number of dimensions is reduced through the dimension reduction methods.
The vector-based representation enables words, sentences, and sets of synonyms (i.e. WordNet synsets) to be represented in a unifying way with vectors. VSM provides a variety of definitions of distance between vectors, corresponding to different measures of semantic similarity. In particular, we take advantage of cosine angle between an input vector (input sentence) and an emotional vector (i.e. the vector representing an emotional synset) as similarity measures to identify which emotion the sentence connotes.
3.2 Dimension Reduction Methods
The VSM representation can be reduced with techniques well known in Information Retrieval: LSA, Probabilistic LSA (PLSA), or the Nonnegative Matrix Factorization (NMF) representations.
Cosine similarities can be defined in these representations, and here, as other authors have done, we use a rule that if the cosine similarity
does not exceed a threshold, the input sentence is labeled as “neutral”, the absence of emotion. Otherwise, it is labeled with one emotion associated with the closest emotional vector having the highest similarity value. We use a predetermined threshold ($t = 0.65$) for the purpose of validating a strong emotional analogy between two vectors (Penumatsa, Ventura et al. 2006).
If we define the similarity between a given input text, $I$, and an emotional class, $E_j$, as $\text{sim}(I, E_j)$, the categorical classification result, CCR, is more formally represented as follows:
$$\text{CCR}(I) = \begin{cases}
\arg \max_j \left( \text{sim}(I, E_j) \right) & \text{if } \text{sim}(I, E_j) \geq t \\
"neutral" & \text{if } \text{sim}(I, E_j) < t
\end{cases}$$
One class with the maximum score is selected as the final emotion class.
Dimensionality reduction in VSM reduces the computation time and reduces the noise in the data. This enables the unimportant data to dissipate and underlying semantic text to become more patent. We will review three statistical dimensionality reduction methods (LSA, PLSA, and NMF) that are utilized in a category-based emotion model.
Latent Semantic Analysis (LSA) is the earliest approach successfully applied to various text manipulation areas (Landauer, Foltz et al. 1998). The main idea of LSA is to map terms or documents into a vector space of reduced dimensionality that is the latent semantic space. The mapping of the given terms/document vectors to this space is based on singular vector decomposition (SVD). It is known that SVD is a reliable technique for matrix decomposition. It can decompose a matrix as the product of three matrices,
$$A = U\Sigma V^T \approx U_k\Sigma_k V_k^T = A_k$$
where $A_k$ is the closest matrix of rank $k$ to the original matrix. The columns of $V_k$ represent the coordinates for documents in the latent space.
Probabilistic Latent Semantic Anlaysis (PLSA) (Hofmann 2001) has two characteristics distinguishing it from LSA. PLSA defines proper probability distributions and the reduced matrix does not contain negative values. Based on the combination of LSA and some probabilistic theories such as Bayes rules, the PLSA allows us to find the *latent topics*, the association of documents and topics, and the association of terms and topics. In the equation (2), $z$ is a *latent class variable* (i.e. discrete emotion category), while $w$ and $d$ denote the elements of term vectors and document vectors, respectively.
$$P(d, w) = \sum_z P(z)P(w|z)P(d|z)$$
where $P(w|z)$ and $P(d|z)$ are topic-specific word distribution and document distribution, individually. The decomposition of PLSA, unlike that of LSA, is performed by means of the likelihood function. In other words, $P(z)$, $P(w|z)$, and $P(d|z)$ are determined by the maximum likelihood estimation (MLE) and this maximization is performed through adopting the Expectation Maximization (EM) algorithm. For document similarities, each row of the $P(d|z)$ matrix is considered with the low-dimensional representation in the semantic topic space.
Non-negative Matrix Factorization (NMF) (Lee and Seung 1999) has been successfully applied to semantic analysis. Given a non-negative matrix $A$, NMF finds non-negative factors $W$ and $H$ that are reduced-dimensional matrices. The product $WH$ can be regarded as a compressed form of the data in $A$.
$$A \approx WH = \sum W H$$
$W$ is a basis vector matrix and $H$ is an encoded matrix of the basis vectors in the equation (3). NMF solves the following minimization problem (4) in order to obtain an approximation $A$ by computing $W$ and $H$ in terms of minimizing the Frobenius norm of the error.
$$\min_{W,H} \|A - WH\|_F^2, \quad s.t. \quad W, H \geq 0$$
where $W, H \geq 0$ means that all elements of $W$ and $H$ are non-negative. This non-negative peculiarity is desirable for handling text data that always require non-negativity constraints. The classification of documents is performed based on the columns of matrix $H$ that represent the documents.
### 3.3 Three-dimensional estimation with features derived from ANEW
Dimensional models have been studied by psychologists often by providing a stimulus (e.g. a photo or a text), and then asking subjects to report on the affective experience. ANEW (Bradley and Lang 1999) is a set of normative emotional ratings for a collection of English words ($N=1,035$), where after reading the words, subjects reported their emotions in a three dimensional representation. This collection provides the rated values for valence, arousal, and dominance for each word rated using the Self Assessment Manikin (SAM). For each word $w$, the normative database provides coordinates $\bar{w}$ in an affective space as:
\[
\bar{w} = (valence, arousal, dominance) \\
= ANEW(w)
\]
(5)
The occurrences of these words in a text can be used, in a naïve way, to weight the sentence in this emotional plane. This is a naïve approach since words often change their meaning or emotional value when they are used in different contexts.
As a counterpart to the categorical classification above, this approach assumes that an input sentence pertains to an emotion based on the least distance between each other on the Valence-Arousal-Dominance (VAD) space. The input sentence consists of a number of words and the VAD value of this sentence is computed by averaging the VAD values of the words:
\[
\overline{\text{sentence}} = \frac{\sum_{i=1}^{n} \bar{w}}{n}
\]
(6)
where \( n \) is the total number of words in the input sentence.
Since not many words are available in this normative database, a series of synonyms from WordNet-Affect are used in order to calculate the position of each emotion. These emotional synsets are converted to the 3-dimensional VAD space and averaged for the purpose of producing a single point for the target emotion as follows:
\[
\overline{\text{emotion}} = \frac{\sum_{i=1}^{k} \bar{w}}{k}
\]
(7)
where \( k \) denotes the total number of synonyms in an emotion. Anger, fear, joy, and sadness emotions are mapped on the VAD space. Let \( A_c, F_c, J_c, \) and \( S_c \) be the centroids of four emotions. Then the centroids, which are calculated by the equation (7), are as follows: \( A_c = (2.55, 6.60, 5.05), F_c = (3.20, 5.92, 3.60), J_c = (7.40, 5.73, 6.20), \) and \( S_c = (3.15, 4.56, 4.00). \) Apart from the four emotions, we manually define neutral to be \((5, 5, 5)\). If the centroid of an input sentence is the most approximate to that of an emotion, the sentence is tagged as the emotion (with the nearest neighbor algorithm). The centroid \(\overline{\text{sentence}}\) might be close to an \(\overline{\text{emotion}}\) on the VAD space, even if they do not share any terms in common. We define the distance threshold (empirically set to 4) to validate the appropriate proximity like the categorical classification.
### 4 Emotion-Labeled Data
Three emotional datasets, with sentence-level emotion annotations, were employed for the evaluation described in the next section. The first dataset is “Affective Text” from the SemEval 2007 task (Strapparava and Mihalcea 2007).\(^1\) This dataset consists of news headlines excerpted from newspapers and news web sites. Headlines are suitable for our experiments because headlines are typically intended to express emotions in order to draw the readers’ attention. This dataset has six emotion classes: anger, disgust, fear, joy, sadness and surprise, and is composed of 1,250 annotated headlines. The notable characteristics are that SemEval dataset does not only allow one sentence to be tagged with multiple emotions, but the dataset also contains a neutral category in contrast to other datasets.
We also use the ISEAR (International Survey on Emotion Antecedents and Reactions) dataset, which consists of 7,666 sentences (Scherer and Wallbott 1994), with regard to our experiments.\(^2\) For building the ISEAR, 1,096 participants who have different cultural backgrounds completed questionnaires about experiences and reactions for seven emotions including anger, disgust, fear, joy, sadness, shame and guilt.
The annotated sentences of the third dataset are culled from fairy tales (Alm 2009). Emotions are particularly significant elements in the literary genre of fairy tales. The label set with five emotion classes is as follows: angry-disgusted, fearful, happy, sad and surprised. There are 176 stories by three authors: B. Potter, H.C. Andersen, and Grimm’s. The dataset is composed of only sentences with affective high agreements, which means that annotators highly agreed upon the sentences (four identical emotion labels).
| Emotion | SemEval | ISEAR | Fairy tales | Total |
|---------|---------|-------|-------------|-------|
| Anger | 62 | 2,168 | 218 | 2,448 |
| Fear | 124 | 1,090 | 166 | 1,380 |
| Joy | 148 | 1,090 | 445 | 1,683 |
| Sadness | 145 | 1,082 | 264 | 1,491 |
Table 1: Number of sentences for each emotion
In our study, we have taken into account four emotion classes (Anger, Fear, Joy and Sadness) which are in the intersection among three datasets (SemEval, ISEAR and Fairy tales). The number of sentences for each emotion and each
---
\(^1\) The dataset is publicly available at http://www.cse.unt.edu/~rada/affectivetext.
\(^2\) Available at http://www.unige.ch/fapse/emotion/databanks/isear.html
| Data set | SemEval | ISEAR | Fairy tales |
|----------|---------|-------|-------------|
| | Prec. | Rec. | F1 | Prec. | Rec. | F1 | Prec. | Rec. | F1 |
| Anger | MCB | 0.000 | 0.000 | - | 0.399 | 1.000 | 0.571 | 0.000 | 0.000 | - |
| | CLSA | 0.089 | 0.151 | 0.112 | 0.468 | 0.970 | 0.631 | 0.386 | 0.510 | 0.749 |
| | CPLSA | 0.169 | 0.440 | 0.244 | 0.536 | 0.397 | 0.456 | 0.239 | 0.313 | 0.650 |
| | CNMF | 0.294 | 0.263 | 0.278 | 0.410 | 0.987 | 0.579 | 0.560 | 0.650 | 0.392 |
| | DIM | 0.161 | 0.192 | 0.175 | 0.179 | 0.286 | 0.604 | 0.290 | 0.392 | |
| Fear | MCB | 0.000 | 0.000 | - | 0.000 | 0.000 | - | 0.000 | 0.000 | - |
| | CLSA | 0.434 | 0.622 | 0.511 | 0.633 | 0.038 | 0.071 | 0.583 | 0.640 | |
| | CPLSA | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | |
| | CNMF | 0.525 | 0.750 | 0.618 | 0.689 | 0.029 | 0.056 | 0.704 | 0.741 | |
| | DIM | 0.404 | 0.404 | 0.404 | 0.531 | 0.263 | 0.351 | 0.179 | 0.255 | |
| Joy | MCB | 0.309 | 1.000 | 0.472 | 0.000 | 0.000 | - | 0.407 | 0.579 | |
| | CLSA | 0.455 | 0.359 | 0.402 | 0.333 | 0.061 | 0.103 | 0.847 | 0.727 | |
| | CPLSA | 0.250 | 0.258 | 0.254 | 0.307 | 0.381 | 0.340 | 0.555 | 0.436 | |
| | CNMF | 0.773 | 0.557 | 0.648 | 0.385 | 0.005 | 0.010 | 0.802 | 0.781 | |
| | DIM | 0.573 | 0.934 | 0.710 | 0.349 | 0.980 | 0.515 | 0.661 | 0.789 | |
| Sadness | MCB | 0.000 | 0.000 | - | 0.000 | 0.000 | - | 0.000 | 0.000 | - |
| | CLSA | 0.472 | 0.262 | 0.337 | 0.500 | 0.059 | 0.106 | 0.704 | 0.642 | |
| | CPLSA | 0.337 | 0.431 | 0.378 | 0.198 | 0.491 | 0.282 | 0.333 | 0.370 | |
| | CNMF | 0.500 | 0.453 | 0.475 | 0.360 | 0.009 | 0.017 | 0.708 | 0.760 | |
| | DIM | 0.647 | 0.157 | 0.253 | 0.522 | 0.249 | 0.337 | 0.408 | 0.240 | |
Table 3: Emotion identification results
The dataset used in our experiment is shown in Table 1. In addition, sample sentences from the annotated corpus appear in Table 2.
| Dataset | Sentences tagged with Sadness/Sad |
|----------|-----------------------------------|
| SemEval | Bangladesh ferry sink, 15 dead. |
| ISEAR | When I left a man in whom I really believed. |
| Fairy tales | The flower could not, as on the previous evening, fold up its petals and sleep; it dropped sorrowfully. |
Table 2: Sample sentences labeled with sadness/sad from the datasets
5 Experiments and Results
The goal of the affect classification is to predict a single emotional label given an input sentence. Four different approaches were implemented in Matlab. A categorical model based on a VSM with dimensionality reduction variants, (LSA, PLSA, and NMF), and a dimensional model, each with evaluated with two similarity measures (cosine angle and nearest neighbor). Stopwords were removed in all approaches. A Matlab toolkit (Zeimpekis and Gallopoulos 2005), was used to generate the term-by-sentence matrix from the text.
The evaluation in Table 3 shows Majority Class Baseline (MCB) as the baseline algorithm. The MCB is the performance of a classifier that always predicts the majority class. In SemEval and Fairy tales the majority class is joy, while anger is the majority emotion in case of ISEAR. The five approaches were evaluated on the dataset of 479 news headlines (SemEval), 5,430 responses to questions (ISEAR), and 1,093 fairy tales’ sentences. We define the following acronyms to identify the approaches:
- CLSA: LSA-based categorical classification
- CPLSA: PLSA-based categorical classification
- CNMF: NMF-based categorical classification
- DIM: Dimension-based estimation
The measure of accuracies used here were: Cohen’s Kappa (Cohen 1960), average precision, recall, and F-measure. While the kappa scores are useful in obtaining an overview of the reliability of the various classification approaches, they do not provide any insight on the accuracy at the category level for which precision, recall, and F-measure are necessary.
5.1 Precision, Recall, and F-measure
Classification accuracy is usually measured in terms of precision, recall, and F-measure. Table 3 shows these values obtained by five approaches for the automatic classification of four emotions. The highest results for a given type of scoring and datasets are marked in bold for each individual class. We do not include the accuracy values in our results due to the imbalanced proportions of categories (see Table 1). The accuracy metric does not provide adequate information, whereas precision, recall, and F-measure can effectively evaluate the classification performance with respect to imbalanced datasets (He and Garcia 2009).
As can be seen from the table, the performances of each approach hinge on each dataset and emotion category, respectively. In the case of the SemEval dataset, precision, recall and F-measure for CNMF and DIM are comparable. DIM approach gives the best result for joy, which has a relatively large number of sentences. In ISEAR, DIM generally outperforms other approaches except for some cases, whereas CNMF has the best recall score after the baseline for the anger category. Figure 1 indicates the results of 3-dimensional and 2-dimensional attribute evaluations for ISEAR. When it comes to fairy tales, CNMF generally performs better than the other techniques. Joy also has the largest number of data instances in fairy tales and the best recall ignoring the baseline and F-measure are obtained with the approach based on DIM for this affect category. CNMF gets the best emotion detection performance for anger, fear, and sadness in terms of the F-measure.
Figure 2 and Table 4 display results among different approaches obtained on the three different datasets. We compute the classification performance by macro-average, which gives equal weight to every category regardless of how many sentences are assigned to it.\(^3\) This measurement prevents the results from being biased given the imbalanced data distribution. From this summarized information, we can see that CPLSA performs less effectively with several low performance results across all datasets. CNMF is superior to other methods in SemEval and Fairy tales datasets, while DIM surpasses the others in ISEAR. In particular, CPLSA outperforms CLSA and CNMF in ISEAR because their performances are relatively poor. The result implies that statistical models which consider a probability distribution over the latent space do not always achieve sound performances. In addition, we can infer that models (CNMF and DIM) with non-negative factors are appropriate for dealing with these text collections.
Another notable result is that the precision, recall, and F-measure are generally higher in fairy tales than in the other datasets. These sentences in the fairy tales tend to have more emotional terms and the length of sentences is longer. The nature of fairy tales makes unsupervised models yield better performance (see Table 2). In addition, affective high agreement sentence is another plausible contributing reason for the encouraging experimental results.
In summary, categorical NMF model and dimensional model show the better emotion identification performance as a whole.
5.2 Cohen’s Kappa
The kappa statistic measures the proportion of agreement between two raters with correction for chance. The kappa score is used as the metric to compare the performance of each approach. Figure 3 graphically depicts the mean kappa scores and its standard errors obtained from the emotion classification. Comparisons between four approaches are shown across all three datasets. MCB is excluded in the comparison because the mean kappa score of MCB is 0.
Let \(MK_{CLSA}\), \(MK_{CPLSA}\), \(MK_{CNMF}\), and \(MK_{DIM}\) be the mean kappa scores of four methods. The highest score (\(MK_{CNMF} = 0.382\)) is achieved by the CNMF when the dataset is SemEval. In fairy tales, the CNMF method (\(MK_{CNMF} = 0.652\)) also displays better result than the others (\(MK_{CLSA} = 0.506\), \(MK_{DIM} = 0.304\)). On the contrary, the achieved results are significantly different in the case of the ISEAR dataset in comparison with the aforementioned datasets. The DIM (\(MK_{DIM} = 0.210\)) clearly outperforms all methods. The kappa score of the CPLSA approach (\(MK_{CPLSA} = 0.099\)) is quantitatively and significantly higher than the CLSA (\(MK_{CLSA} = 0.031\)) and CNMF (\(MK_{CNMF} = 0.011\)). Kappa score for the NMF-based methods is remarkably lower than the other three approaches.
According to (Fleiss and Cohen 1973), a kappa value higher than 0.4 means a fair to good level of agreement beyond chance alone and it is
\(^3\)Macro-averaging scores are defined as:
\[
P_m = \frac{1}{C} \sum_{i=1}^{C} p_i, \quad R_m = \frac{1}{C} \sum_{i=1}^{C} r_i, \quad F_m = \frac{1}{C} \sum_{i=1}^{C} f_i
\]
where \(C\) is total number of categories, and \(p_i\), \(r_i\), and \(f_i\) stand for precision, recall, and F-measure, respectively, for each category \(i\).
Figure 1: Distribution of the ISEAR dataset in the 3-dimensional and 2-dimensional sentiment space. The blue ‘x’ denotes the location of one sentence corresponding to valence, arousal, and dominance.
Figure 2: Comparisons of Precision, Recall, and F-measure: (a) SemEval; (b) ISEAR; (c) Fairy tales.
| Data set | SemEval | | | ISEAR | | | Fairy tales | | |
|----------|---------|----------|----------|---------|----------|----------|-------------|----------|----------|
| | Prec. | Rec. | F1 | Prec. | Rec. | F1 | Prec. | Rec. | F1 |
| MCB | 0.077 | 0.250 | 0.118 | 0.100 | 0.250 | 0.143 | 0.102 | 0.250 | 0.145 |
| CLSA | 0.363 | 0.348 | 0.340 | 0.484 | 0.282 | 0.228 | 0.662 | 0.640 | 0.630 |
| CPLSA | 0.189 | 0.282 | 0.219 | 0.260 | 0.317 | 0.270 | 0.282 | 0.307 | 0.280 |
| CNMF | 0.523 | 0.506 | 0.505 | 0.461 | 0.258 | 0.166 | 0.747 | 0.731 | 0.733 |
| DIM | 0.446 | 0.422 | 0.386 | 0.528 | 0.417 | 0.372 | 0.530 | 0.404 | 0.419 |
Table 4: Overall average results
Figure 3: Comparisons of Mean Kappa: (a) SemEval; (b) ISEAR; (c) Fairy tales.
an acceptable level of agreement. On the basis of this definition, the kappa score obtained by our best classifier ($MK_{CNMF} = 0.652$) would be reasonable. Most of the values are too low to say that two raters (human judges and computer approaches) agreed upon the affective states. However, we have another reason with respect to this metric in the experiment. We make use of the kappa score as an unbiased metric of the reliability for comparing four methods. In other words, these measures are of importance in terms of the relative magnitude. Hence, the kappa results are meaningful and interpretable in spite of low values. We can observe that the NMF-based categorical model and the dimensional model both experienced higher performance.
### 5.3 Frequently occurring words
The most frequent words used in fairy tales for each emotion are listed in Table 5. We choose this dataset since there are varying lexical items and affective high agreement sentences, as mentioned in Section 5.1. Stemming is not used because it might hide important differences as between ‘loving’ and ‘loved’. CNMF and DIM were selected for the comparison with the Gold Standard because they were the two methods with the better performance than the others. Gold Standard is the annotated dataset by human raters for the evaluation of algorithm performance. The words most frequently used to describe anger across all methods include: *cried*, *great*, *tears*, *king*, *thought*, and *eyes*. Those used to describe fear include: *heart*, *cried*, *mother*, *thought*, *man*, and *good*. Joy contains *happy*, *good*, and *cried* whereas sadness has only *cried* for three methods.
There is something unexpected for the word frequencies. We can observe that the association between frequently used words and emotion categories is unusual and even opposite. For instance, a ‘joy’ is one of the most frequent words referred to for *sadness* in the Gold Standard. In CNMF and DIM, a ‘good’ is employed frequently with regard to *fear*. Moreover, some words occur with the same frequency in more categories. For example, the word ‘cried’ is utilized to express *anger*, *fear*, and *joy* in the Gold Standard, CNMF, and DIM. In order to find a possible explanation in the complexity of language used in the emotional expression, some sentences extracted from fairy tales are listed below:
“The cook was frightened when he heard the order, and said to Cat-skin, You must have let a hair fall into the soup; if it be so, you will have a *good* beating.” – which expresses *fear*
“When therefore she came to the castle gate she saw him, and *cried* aloud for joy.” – which is the expression for *joy*
“Gretel was not idle; she ran screaming to her master, and *cried*: You have invited a fine guest!” – which is the expression for *angry-disgusted*
From these examples, we can observe that in these cases the affective meaning is not simply propagated form the lexicon, but is the effect of the linguistic structure at a higher level.
### 6 Conclusion
We compared the performances of three techniques, based on the categorical representation of emotions, and one based on the dimensional representation. This paper has highlighted that the NMF-based categorical classification performs
| Model | Emotion | Top 10 words |
|---------|---------|------------------------------------------------------------------------------|
| Gold Standard | Anger | king, thought, eyes, great, cried, looked, joy, mother, wife, tears |
| | Fear | great, cried, good, happy, thought, man, heart, poor, child, mother |
| | Joy | thought, mother, good, cried, man, day, wept, beautiful, back, happy |
| | Sadness | cried, fell, father, mother, back, joy, dead, danced, wife, tears |
| CNMF | Anger | great, cried, eyes, mother, poor, joy, king, heart, thought, tears |
| | Fear | cried, king, happy, good, man, heart, thought, father, boy, mother |
| | Joy | mother, thought, cried, king, day, great, home, joy, good, child |
| | Sadness | thought, cried, good, *great*, looked, mother, man, time, king, heart |
| DIM | Anger | eyes, fell, heart, tears, cried, good, stood, great, king, thought |
| | Fear | king, cried, heart, mother, good, thought, looked, man, child, time |
| | Joy | eyes, man, children, danced, cried, good, time, happy, great, wedding |
| | Sadness | cried, thought, *great*, king, good, happy, sat, home, joy, found |
Table 5: Most frequent 10 words from fairy tales
the best among categorical approaches to classification. When comparing categorical against dimensional classification, the categorical NMF model and the dimensional model have better performances. Nevertheless, we cannot generalize inferences on which of these techniques is the best performer because results vary among datasets. As a future work, we aim at performing a further investigation on this connection in order to identify more effective strategies applicable to a generic dataset. Furthermore, we aim at exploring improvements in the methodology, employed in this work, and based on the combination of emotional modeling and empirical methods.
**Acknowledgments**
This research is partially sponsored by a Norman I. Price Scholarship from the University of Sydney.
**References**
C. O. Alm (2009). *Affect in Text and Speech*, VDM Verlag Dr. Müller.
S. Aman and S. Szpakowicz (2007). Identifying expressions of emotion in text. *Text, Speech and Dialogue*.
M. M. Bradley and P. J. Lang (1999). Affective norms for English words (ANEW): Instruction manual and affective ratings. University of Florida: The Center for Research in Psychophysiology.
J. Cohen (1960). A coefficient of agreement for nominal scales. *Educational and psychological measurement* 20(1): 37-46.
S. D’Mello, A. Graesser, and R. W. Picard (2007). Toward an affect-sensitive AutoTutor. *IEEE Intelligent Systems* 22(4): 53-61.
S. D’Mello, S. Craig, A. Witherspoon, B. McDaniel, and A. Graesser (2008). Automatic detection of learner’s affect from conversational cues. *User Modeling and User-Adapted Interaction* 18(1): 45-80.
P. Ekman (1992). An argument for basic emotions. *Cognition & Emotion* 6(3): 169-200.
J. L. Fleiss and J. Cohen (1973). The equivalence of weighted kappa and the intraclass correlation. *Educational and psychological measurement* 33: 613-619.
H. He and E. A. Garcia (2009). Learning from Imbalanced Data. *IEEE Transactions on Knowledge and Data Engineering* 21(9): 1263.
T. Hofmann (2001). Unsupervised learning by probabilistic latent semantic analysis. *Machine Learning* 42(1): 177-196.
B. Kort, R. Reilly, and R. W. Picard (2001). An affective model of interplay between emotions and learning: Reengineering educational pedagogy-building a learning companion. *IEEE International Conference on Advanced Learning Technologies*, 2001. Proceedings.
T. K. Landauer, P. W. Foltz, and D. Laham (1998). An introduction to latent semantic analysis. *Discourse processes*, Citeseer. 25: 259-284.
P. J. Lang (1980). Behavioral treatment and bio-behavioral assessment: Computer applications. *Technology in mental health care delivery systems*: 119-137.
D. D. Lee and H. S. Seung (1999). Learning the parts of objects by non-negative matrix factorization. *Nature* 401(6755): 788–791.
P. Penumatsa, M. Ventura, A.C. Graesser, M. Louwerse, X. Hu, Z. Cai, and D.R. Franceschetti (2006). The Right Threshold Value: What Is the Right Threshold of Cosine Measure When Using Latent Semantic Analysis for Evaluating Student Answers? *International Journal on Artificial Intelligence Tools*, World Scientific Publishing.
J. A. Russell (2003). Core affect and the psychological construction of emotion. *Psychological review* 110(1): 145-172.
K. R. Scherer and H. G. Wallbott (1994). Evidence for universality and cultural variation of differential emotion response patterning. *Journal of Personality and Social Psychology* 66: 310-328.
C. Strapparava and R. Mihalcea (2007). Semeval-2007 task 14: Affective text. *Proceedings of the 4th International Workshop on Semantic Evaluations*, Association for Computational Linguistics.
C. Strapparava and R. Mihalcea (2008). Learning to identify emotions in text. *SAC ’08: Proceedings of the 2008 ACM symposium on Applied computing*, Fortaleza, Ceara, Brazil, ACM.
C. Strapparava and A. Valitutti (2004). WordNet-Affect: an affective extension of WordNet. *Proceedings of LREC*.
R. B. Yates and B. R. Neto (1999). Modern information retrieval. *ACM P*.
Zeimpekis D. and E. Gallopoulos (2005). TMG: A MATLAB toolbox for generating term-document matrices from text collections. *Grouping multidimensional data: Recent advances in clustering*: 187-210. |
Integrating Business Process and User Interface Models using a Model-Driven Approach
Renata Dividino*, Veli Bicer†, Konrad Voigt‡ and Jorge Cardoso‡
* ISWeb Research Group University of Koblenz-Landau
Email: firstname.lastname@example.org
† Forschungszentrum Informatik an der Universität Karlsruhe (TH)
Email: email@example.com
‡ SAP Research CEC Dresden
Email: {konrad.voigt,firstname.lastname@example.org
Abstract—Business services are complex entities that encompass descriptions about different aspects including business processes and user interfaces. Typically, the modeling of business services results on several correlated models. On the one hand, there is the need to keep these models apart in order to attain the levels of abstractions needed to model different aspects of a service. On the other hand, it is important to maintain the consistency and integrity of models. In this paper, we show how to maintain the consistency and integrity of models and present a use case for the integration of user interface design and business process models.
Index Terms—Service Engineering, User Interface Design, Model-driven Engineering, Business Process Integration.
I. INTRODUCTION
The Internet of Services brings a set of new challenges for the research community. One of the challenges that needs to be handled is the complexity associated with the modeling of real-world or business services. Services are complex entities that encompass descriptions about business processes, user interfaces, pricing models, legal constraints, knowledge assets, technological infrastructures, business rules, etc. One approach to manage the lifecycle of a service multifaceted structure is to distill its core characteristics into formal models and afterwards align and integrate the models. This paper describes how a model-driven approach can be used to model and maintain the consistency and integrity of business processes models (which describe services’ internal behavior) with service user interface models. While a service is typically composed by several models (up to 20 in our research [3]), we demonstrate the feasibility of our approach using only two models, which takes parts on the traditional design of the user interface (UI) and business processes, since it provides a concrete list of the major steps to take.
As a use case, we deploy a business service and model its underlying business process model. The business service is called EcoCalculator Service (ECS) and its main objective is to calculate the ecological value (eco value or EV) of products (e.g. car parts such as car sits) according to European directives and worldwide environmental regulations. We show the business process modeled with BPMN [1] showing the set of activities that needs to be carried out when the ECS is invoked by a customer, and the UI dialogue modeled with DIAMODL [2] which illustrates the user interaction. In addition, to provide support for business services modeling integrated with user interface design, we extend the modeling languages BPMN and DIAMODL, and implement integration rules in order to provide support for business process modeling integrated with UI design.
II. ISE METHODOLOGY FOR SERVICE ENGINEERING
In this section we cover our approach for service engineering, called ISE (an acronym for Inter-enterprise Service Engineering). First we introduce the service engineering discipline, describing its main aspects. This is followed by an overview of ISE, describing how the methodology applies Model Driven Engineering (MDE) in order to provide an integrated framework and implementing workbench.
A. Service Engineering
Service engineering (SE) [3] is a new approach to the design and implementation of services. Typically services evolve in a common ecosystem in which organizations and IT provide value in form of services. SE provides methodologies to cope with the complexity of several business actors and their interaction. Furthermore, SE specifies tools for implementing and deploying services, covering both, IT and business perspectives. Consequently, SE is a structured approach for creating a new (e-)service [3]. Approaches should translate an initial abstract description (e.g. natural language), through a sequence of representations, to a technical standard-based representation. The ISE methodology, described in the following section, addresses exactly the creation of new services using a guiding methodology.
B. ISE Methodology
In the context of the TEXO project\(^1\), the ISE methodology [3] and accompanying workbench for service engineering was implemented. Compared to other approaches, the ISE methodology considers a technical perspective as well as a business and operational perspectives. Since the notions
\(^1\)TEXO Business Webs in the Internet of Services (http://thesesprogramm.de/scenarios/en/txeo)
of abstraction (perspective) and dimensions (entities) were important for our approach, we have followed a solution based on the Zachman framework [4]. ISE combines this business approach with the concepts of Model Driven Engineering (MDE) [5].
Figure 1 shows the ISE framework which is part of ISE methodology. Inspired by Zachman framework and following the separation of concerns paradigm raised by [6], it divides and structures services into four main perspectives and five dimensions. The dimensions are: service description, workflow, data, people and rules. Figure 1 presents a reduced view, showing only the models relevant for this paper. Furthermore, each of these dimensions is divided into four perspectives (layers) of abstraction. Each of the perspectives of the ISE methodology can be regarded as a phase in the development (model refinement) of services. Thus, the models which are assigned to each layer support the development from different viewpoints (i.e., scope, business, logical, and technical). Additionally, models at different dimension but belonging to the same layer are integrated/binded to others in order to form the complete business service model at the respective level of abstraction. For all cells of the matrix, we have defined formal models which should be considered in the service development. Examples of models include UML, mind maps, BPMN, BPEL, OWL, DIAMODL, etc.
Therefore, ISE act as an integration platform for several models stored in cells of the framework. Throughout one dimension, models are created with respect to different views and refined until they conform to a technical specification. This leads to multiple representations of information on different layers of abstraction in the corresponding dimensions. Changes in one model have to be propagated into related models holding overlapping information (depicted by arrows in Figure 1). We divide the dependencies between models into two classes: (1) Vertical dependencies cover the synchronization of dependencies between models on different layers of abstraction in one dimension. (e.g. BPMN and BPEL) and (2) Horizontal dependencies define the synchronization of models on the same layer of abstraction. (e.g. BPMN and DIAMODL).
These dependencies require the integration of models. We have adopted a model transformation from MDE to allow for an automatic support of integration. Based on a common formal representation (e.g. MOF), a domain specific language for model transformation can be used to define rules and apply them to the models, in order to generate elements from a source model to a target model automatically. ISE uses for model integration the Query, View and Transformation (QVT) technique. We have chosen to rely on QVT because of matured concepts, well established infrastructure for model management and transformation, and available OMG standards. The integration is again depicted in Figure 1 using arrows. In the following sections we will show how such integration can be done.
III. BUSINESS PROCESS MODELING WITH UIs
In this section, we describe the theories for UI design and business process modeling used in this work, and introduce the DIAMODL and BPMN modeling languages.
A. User Interface Design and DIAMODL
A widely accepted principle of methodologies on UI design [7], [8], [9] is the separation (so called modeling in layers) of concerns with respect to the modeling of the structural, behavior and presentational aspects of UIs. In terms of ISE, structural aspects are included at the scope, business dimensions; behavior aspects at the logical dimension; and presentation aspects at the technical dimension. The main idea of the traditional delineation between structure, behavior and presentation (with additional layers of granularity) is to describe the full set of issues and considerations involved in more complex forms of interactive UIs.
Behavioral aspects mainly describes interactive qualities of the UI, i.e. how users and systems cooperate to perform the desired operations. Interactions can be modeled as a dialogue where operations that allow users to exchange message with the interface are defined. The UI is used as an communication channel for the user to exchange message with the system. Note that this layer acts as a glue bringing together system and interface (users).
Within the ISE methodology, behavior aspects of an UI are modeled using the DIAMODL notation. The DIAMODL notation proposed by [2] offers interactors-with-gates (i.e. states-with-transitions) interface component abstraction and the hierarchical state formalism statecharts for describing the UI’s behavior. DIAMODL’s five core elements are:
- **Interactors**. Interactors represent a distinct context for behaviors of the UI, such as an “input” state and an “output” state.
- **Gates**. Gates determine how to interpret states i.e. they model the actions associated with a state, and each state can be associated with another state through connections.
- **Connections**. Connections are directed associations between any two states that dictates under what conditions (“triggers”) the UI shifts from the first context (state) to the second.
- **Variables**. Variables are used for holding a resource (data value) and its contained structure.
- **Computations**. Computations enable message exchanges among UI components, and between a UI component and an interactive system.
Figure 2 presents a simplified DIAMODL model for the ECS example. The main input of the ECS is the identification of the part (product) for which the EV will be calculated. The user inputs the PID-value (\textit{textInteractorPID}) which is internally stored in a variable (\textit{PDI}). When the user clicks on the OK button (\textit{buttonInteractor}), the PID-value is sent to the system. Since the calculation of EV is highly dependent on directives and environmental regulations holding for the country in where the product is built and/or exported to, the country name has to be specified by the user (\textit{comboInteractorCountry}) – as default value the European directives are considered. This information is forwarded to the system which calculates the respective EV and returns it to the user (\textit{textInteractorEV}).
**Fig. 2.** DIAMODL model for the ECS
### B. Business Service Choreographies and BPMN
Modeling business processes, at a high level, requires the specification of the involved parties and process flow according to business objectives of the service. Within ISE methodology, the Business Process Modeling Notation (BPMN) is adopted to address this need. It is one of the well-known languages that aims to specify business processes at a high-level, rather than as executable workflows. It provides a modeling mechanism with a standard visual representation to create business process models and abstract the complexity of the process-level service design. According to BPMN specification [1], there are four types of core elements that help to define processes:
- **Flow Objects.** These are the main graphical elements to define the behavior of a process model. They mainly include events, activities, and gateways. Events indicate the triggers that occur during the course of the process. BPMN distinguishes start events and intermediate events as two kinds of event consumption whereas end events are only used for event production. Start events lead to process instantiation and intermediate events are consumed by a running process instance. Furthermore, activities are generic constructs to represent particular actions (e.g. tasks and sub-processes) to be performed in process. Gateways, on the other hand, are used to control splits and joins in the process flow.
- **Connecting Objects.** Connecting objects are used to link flow objects and other elements. There are three ways to connect elements which are: sequence flow, message flow and association.
- **Swimlanes.** Swimlanes are elements to represent the boundaries that exist in a process model due to the inter-organizational nature of business processes. The topmost swimlane element is the pool which categorizes and encapsulates the activities performed by an organization entity or a system. A pool can then be partitioned into lanes.
- **Artifacts.** Artifacts are used to provide additional information about the process such as annotations or data objects.
Figure 3 shows core BPMN elements used to model the business process of the ECS. The service starts with the PID-value input. The first two activities to be executed in parallel are the \textit{Loader} (Data Service) and \textit{Process Logging}. The \textit{Loader} retrieves the \textit{BOM} (bill of material) based on the \textit{PID} supplied by the consumer. In parallel, the \textit{Process Logging} activity writes log information indicating that a process instance has started. Once those two activities are completed, the process model checks if the \textit{BOM} has information describing all the subparts. Additionally, it checks if the name of a country is specified in which regulations will be used for the computation of the \textit{EV}. If no information is missing, the \textit{EV} is computed and \textit{Billing}, \textit{Payment}, \textit{Performance Analysis} and \textit{Logging} activities are executed. For reader’s convenience, UI part is also shown as illustrations of the components. Intuitively, this graphical notation of BPMN essentially provides a special kind of flowchart incorporating constructs tailored to business process modeling. Although it is quite useful in this sense, there is a lack of elements to model constructs specific to UI models.
### C. DIAMODL and BPMN Synchronization Requirements
In view of service engineering requirements, both service business process and user interaction contribute to the execution of a service to perform its functionality for a desired outcome. They contribute to the design from different angles, but need to be in synchronized in order to result in a coherent execution after service deployment. The common action units for both models are the activities. In DIAMODL, they correspond to the \textit{interactors} and \textit{computations} considering their assignment to a user or system, respectively. In BPMN, activities are either \textit{tasks}, or \textit{subprocesses}. However, BPMN does not explicitly state by whom an activity is performed, and accordingly no distinction of activity types (i.e. user or system). As expected, each model assess the other’s activities in a black-box manner – i.e. DIAMODL \textit{computation} represents the whole business process activities and BPMN \textit{tasks} standing for user activities and interactors.
Intuitively, the execution of activities from start to end in both models are \textit{event-driven}. Although the activities lead to accomplishing overall service goal separately in business process and UI, their flows need to be aligned to fit each
other. In Figure 3, this is illustrated using dashed lines between the UI components and business process elements. First, the UI needs to be aware of the state of the business process in order to set its behavior. This includes switching from one interactor to another depending on the state. More importantly, the business process is bound exactly to the events triggered by UI with the intention of avoiding non-deterministic behavior. In other words, the business process must ensure that an event occurring during its execution is properly handled by the corresponding UI elements in the case of a human involvement is needed. Therefore, we consider an event-based coordination to achieve the synchronization between two models.
However, utilizing such a synchronization (by means of model transformations as detailed in the next section) requires both models to be semantically equivalent in terms of modeling events. Considering the current state of DIAMODL and BPMN specifications, we distinguish three important points that need to be addressed:
- **Events in DIAMODL.** DIAMODL treats events implicitly as parts of other components such as computation objects or interactors. Computations support communication and synchronization with internal and external entities. However, even though computations explicitly state who are the participants of the dialogue, they do not define explicitly which are the events and event-types to be caught/thrown and the event-conditions/rules holding in the dialogue. In order to utilize their consumption and production with business process events, events triggered within computations need to be explicitly defined as a part of a model.
- **Events in BPMN.** BPMN comes with a rich set of event types. However, it does not have the capability to express different levels of events, i.e., specifying events only to be consumed by UI. This is required since a business process can interact not only with UI but also with other business processes in a BPMN model.
- **Two-way synchronization.** According to ISE methodology, both models are designed separately as a part of different service aspects. This confronts the requirement of capturing the common information in both models and adapt them for synchronization. A declarative technique (e.g., QVT) is needed to be implemented for model transformations.
### IV. INTEGRATION OF UI ORIENTED METHODOLOGY
We extend two well-accepted existing languages, the BPMN and DIAMODL, in order to have equivalent well-defined semantic in terms of modeling events. Furthermore, we show the synchronization between both models by means of models transformation algorithms.
**Fig. 4. Matching interactions between DIAMODL and BPMN**
#### A. Extension for the DIAMODL Notation
Following the requirements described in Section III, we propose a three refinements of the computation component. First, we differ between *internal computation* and *external computation* components. Internal computations are value-based mapping functions holding among UI components. External computation, however, are value-based mapping functions holding between UI components and an interactive system (as network services). For this purpose, external computations must include the description of endpoints and their messages regardless of what message formats or network protocols.
are used to communicate. In addition, when more than one external computation hold a dialogue with a service in which inputs and outputs are dependent from each other, external computation should define the message sequence flow, e.g. define which message should be processed first (start-message), which are the intermediate ones (intermediate-message), and the last one (end-message) before closing the communication channel. The symbols for internal and external computation are shown in Figure 4.
Second, computation components include event-triggers. We define three types of event messages: response, notification and error. Based on these message types, we specify a categorization of event triggers into catching and throwing events. Additionally, event trigger are sub-divided in interaction types. Throwing events’ interaction types are: (1) One-way - an operation that only sends an input (2) Request-response - An operation sends a request, then waits for a response or an error (3) Request-notification - An operation sends a request, then waits for the message notification or an error. Catching events are responsible for the reception of messages. Analogous to the throwing events, we define one-way, request-response, and request-notification interaction types for catching events.
At last, computations are extended with Event-Condition-Action (ECA) rules which automatically perform actions in response to events provided that stated conditions hold. ECA rules introduces reactive functionality to computations, and thus allowing computations to determine the direction of the dialogue flow based on a single rule. Example 4.1 shows a ECA rule for a notify-response–like message exchange. One direction is specified for the dialogue flow when the event-message is of type notification, and a different one for error messages.
Example 4.1: ECA rule for notify-response–like message exchanges
```
ON input-event
IF input-event.type == NOTIFICATION
DO goAssociation.sourceNode = compEx
```
B. Extension for Business Process Modeling Notation (BPMN)
The aforementioned BPMN events (e.g. start, intermediate and end) provide basic functionalities that allows to interact with the UI during the execution of a process. However, a BPMN event only provides one specific functionality such as receiving or sending a message (differentiated by black-white colors in Figure 4). Therefore, only one-way interactions can be represented with the current semantics of events stating that an event is either receiving or sending a message.
Since we want to match different event-type computations with BPMN events for a seamless interaction, there is a need to associate more than one event activity to the other types of interaction. For this, we use the idea of event patterns and introduce an event conjunction pattern as a container to include more than one event for a new abstract event (more patterns for complex events can be found in [10]). In Figure 4, this is shown using a dashed-line around two sending and receiving message events. Strictly speaking, an event conjunction pattern is just a container intended to be represented in formal syntax of the model, because the events can be interleaved. For example, the start and end events of the process can be in the same container, matching the whole process to a computation (request-response) in DIAMODL.
C. Synchronization between Models
Finally, a QVT-based model transformation is used to synchronize the models during service design. Figure 5 shows a fragment of the transformation using the QVT syntax. It should be noted that the transformation is defined based on the corresponding meta-models defined according to Eclipse Modeling Framework\(^2\) (EMF) specification. The actual mapping between the computations and events are handled by top-level Comp2Activity relation. It is specified on two domains of types Computation \((c)\) and Activity \((a)\). A pattern is applied on Computation to check whether it is external or internal – i.e. only external ones are mapped. Additionally, we correlate the matching elements by their names with an assumption that elements are uniquely labeled in both models.
In BPMN meta-model, all elements belong to a generic activity class, that is further specialized by their activityType attributes. This is achieved using three additional sub-relations that are called in the where clause based on the interactionType of the Computation. Due to space limitations, we only show the transResponse relation that maps a request-response type
\(^2\)http://www.eclipse.org/modeling/emf/
Computation to an event conjunction pattern. It is introduced as a new activity type (i.e. EventConj) in BPMN meta-model which involves one or more sub-events.
V. RELATED WORK
The need of modeling human user interaction within process modeling reflect the increase of recent work on this area. In [11] and [12], the authors use the standard process modeling language BPMN for both modeling processes and task modeling, and DIAMODL for modeling the UI structure and behaviour. The authors propose a refactoring process of the original BPMN model as an intermediary step in order to be synchronized with the DIAMODL model. This refactoring process aims to make explicit what each user does and how it interacts with its environments. However, they not specify how the exchange of message-event and data is conducted. In [13] and [14], the authors use a less formal UI model and couple to the BPMN when the business process is used as a starting point for the UI design. UI models and business process models are alignments with business requirements. On the one hand, it is not clear how the business requirements are collected, neither how the alignment of the models is carried out. In our approach we propose the alignment of the models following business requirements which, in turn, are defined in perspectives/dimensions following a formal methodology. We still keep the independence among models, so that we assure, for instance, that one business process model can be coupled with different UI behaviour models (which focus on the interaction regardless the environment considered). In [15], the authors apply a model-driven approach that derives UIs from business processes in order to keep the traceability of the business process and the UI. The authors make use of task models to bridge business process and UI design. In [16], the authors describes the design of user interface for a Web service based on the WSDL (Web Service Description Language) description. For the UI design, they have developed a new UI specification language and the associated authoring environment, which can be used for the abstract and concrete UI definition. Operations and data types defined in WSDL are bound to the UI model.
VI. CONCLUSION
Service engineering converges into a complex task of modeling business services involving various aspects to specify both execution and usage of services. Typically, the modeling process of business services results on several correlated models. On the one hand, there is the need to keep these models apart in order to attain the levels of abstraction to model different aspects of a service. This is mostly due to the fact that incorporating different models for different aspects of the service enables the separation of concerns during the design while attaining a staged development to evolve the service idea into an executable artifact. On the other hand, it is important to maintain the consistency and integrity of business processes models.
The ISE methodology for engineering services supports the modeling of business services based on perspectives and dimensions. This methodology directly meets the requirements of the independence of models, which however should be aligned and integrated. In this paper, we demonstrate the feasibility of our approach using only two models which takes parts on the tradition design of the user interface and business process of a service. To make the relationship between business process and UI models explicitly, we extend two well-accepted existing languages: the business process modeling language BPMN, and the user interface dialogue language DIAMODL with new event-based components, and implement model-driven transformation rules in order to provide support for business process integrated with user interface design. The extension of both models was necessary to focus on aspects of communication and synchronization.
Our future work aims to complete transformation rules of the ISE matrix and, thus, study in more detail the relationship between all models. Additionally, we plan the evaluation of ISE workbench for guiding independently improvements on the modeling languages at each perspective and dimension.
ACKNOWLEDGMENT
This research was funded by means of the German Federal Ministry of Economy and Technology under the promotional reference 01MQ07012. The authors take the responsibility for the contents.
REFERENCES
[1] J. Cardoso, K. Voigt, and M. Winkler, “Service engineering for the internet of services,” in Enterprise Information Systems X, 2008, pp. 17–25.
[2] S. White, “Introduction to BPMN,” IBM Cooperation, pp. 2008–029, 2004.
[3] H. Traetteberg, “A hybrid tool for user interface modelling and prototyping,” in CADUI’06. Bucharest, Romania: Springer, 2007.
[4] J. A. Zachman, “A framework for information systems architecture,” vol. 26, no. 3. NJ, USA: IBM Corp., 1987, pp. 276–292.
[5] T. Stahl and M. Völter, Model-Driven Software Development: Technology, Engineering, Management, 1st ed. Wiley&Sons, 2006.
[6] D. L. Parnas, “On the criteria to be used in decomposing systems into modules,” NY, USA: Springer, 2002, pp. 411–427.
[7] B. Baxley, “Universal model of a user interface,” in DUX ’03. NY, USA: ACM, 2003, pp. 1–14.
[8] J. J. Garret, The Elements of User Experience: User-Centered Design for the Web. CA, USA: New Riders Publishing, 2002.
[9] L. Constantine and L. Lockwood, Software for use: a practical guide to the models and methods of usage-centered design. NY, USA: ACM, 1999.
[10] A. P. Barros, G. Decker, and A. Grosskopf, “Complex events in business processes,” in BIS, ser. LNCS, vol. 4439. Springer, 2007, pp. 29–40.
[11] H. Traetteberg, “UI Design without a Task Modeling Language – Using BPMN and Diamodl for Task Modeling and Dialog Design,” in HCSE-TAMODIA ’08. Berlin: Springer, 2008, pp. 110–117.
[12] H. Traetteberg and J. Krogstie, “Enhancing the usability of bpm-solutions by combining process and user-interface modelling,” in PoEM, 2008, pp. 86–97.
[13] N. Sukaviriya, V. Sinha, T. Ramachandra, and S. Mani, “Model-driven approach for managing human interface design life cycle,” in MoDELS, 2007, pp. 226–240.
[14] N. Sukaviriya, V. Sinha, T. Ramachandra, S. Mani, and M. Stolze, “User-centered design and business process modeling: Cross road in rapid prototyping tools,” in INTERACT (1), 2007, pp. 165–178.
[15] K. S. Sousa, H. M. Filho, and J. Vanderdonckt, “Addressing the impact of business process changes on software user interfaces,” in BDIM, 2008, pp. 11–20.
[16] F. Paternò, C. Santoro, and L. D. Spano, “Designing usable applications based on web services,” in I-USED, ser. CEUR Workshop Proceedings. CEUR-WS.org, 2008. |
Optimizing Voice-over-IP Speech Quality Using Path Diversity
Mohamed Ghanassi and Peter Kabal
Electrical & Computer Engineering
McGill University
Montreal, Quebec H3A 2A7
Email: {mohamed.ghanassi, firstname.lastname@example.org
Abstract—In last few years, voice over Internet protocol (VoIP) has been gaining popularity as an alternative to traditional telephone by transmitting voice signals as packets over the Internet and private IP-based networks. However, voice packets experience loss, delay, and delay variation, which requires buffering, playout scheduling and loss concealment at the receiver. In this paper, we give an overview of a VoIP application and show how playout scheduling and loss concealment are jointly used to optimize perceived speech quality. We use this optimization criterion to design a histogram-based playout scheduling algorithm. Then, we identify the limitations of the VoIP application for this scheme and propose improvement using path diversity approach that can be implemented via a Service Overlay Network (SON). We present simulations results that show significant improvement of VoIP quality by using this approach.
1. INTRODUCTION
VoIP is a technology for transmitting voice over a data network, such as the Internet, using the Internet protocol (IP). Voice information is sent in digital form in discrete packets rather than in the traditional circuit-committed public switched telephone network (PSTN). Packets are moved along the most efficient path to their destination, where they get reassembled and delivered as voice. A major advantage of VoIP is cost saving by integrating data and real-time voice traffic on the same packet network.
However, packet switched networks were not originally designed for real-time transmission and VoIP technology does not always provide high-quality speech. This is mainly due to the fact that Internet protocol is a best-effort delivery service and Quality of Service (QoS) is not guaranteed. Voice packets experience varying delay (known as jitter) and in order to reconstruct a continuous signal, the received packets have to be buffered for sufficient time so that most of the packets are received before their scheduled playout time. With buffering and playout scheduling, there is a trade-off between delay and packet loss. If the playout delay is set long enough so that most of the packets will arrive before the scheduled playout time, the packet loss ratio (PLR) is very low. However, in this case, the end-to-end delay between the generation of speech and its playout at the receiver is large, which is not well tolerated in voice conversation. If the playout delay is set shorter, the PLR increases since some packets arrive after the scheduled playout time and are effectively lost.
Playout scheduling methods allow a certain amount of late packets. These packets, which are declared lost, as well as the packets dropped in the network, have to be replaced using Packet Loss Concealment (PLC) methods. PLC techniques that have been developed [1] are effective for loss of a single or a few packets. However, when packets are lost successively in a long burst, which occurs when the network delay suddenly increases for a short period of time, concealment of the lost packets is not very effective. Situations of burst loss have to be detected by the playout scheduling algorithm in order to adapt the playout delay accordingly.
Playout scheduling of packets is performed by searching for the best trade-off between delay and loss, corresponding to the optimum speech quality as perceived by service users. However, with severe network impairments, even the best trade-off achieved may translate to an unsatisfactory service level. One approach we can use to improve the VoIP service quality is to add redundancy to the voice signal. The information sent as redundant will be used by the receiver in order to recover the signal in case of packet loss. A commonly used redundancy scheme, known as Forward Error Correction (FEC), transmits information from a packet along with the original information in subsequent packets [2]. FEC methods are effective for isolated loss but in order to combat burst losses, they have to introduce additional delay in the speech signal. As an alternative to this technique, a path diversity based redundancy scheme was used to send redundant information on a second path and take advantage of uncorrelated loss and delay characteristics of the two paths [3].
Implementation of path diversity by explicitly sending packets over different paths can be achieved via IP source routing or Relay approaches [4]. With IP source routing, we can explicitly specify the intermediate nodes a packet should visit on route to its destination. However, this is currently difficult to implement in the Internet as not all ISPs support source routing. With the second approach, flexible routing is provided using relays placed at a number of nodes in the network. To send a packet from a point to another through an alternate path, the packet is simply sent to a relay where, upon reception, it is forwarded to its destination. Service Overlay Networks (SONs) have been developed as an effective means to provide flexibility in the routing process and have been proposed as a means to address the QoS issue for VoIP [5]. SONs are formed by inserting service gateways, which act as relay nodes
and perform service-dependent routing, at access boundaries in the Internet. Currently, SONs, combined with the peer-to-peer network model, are commercially implemented by VoIP service providers (see for example [6]).
In the work of Liang et al. [3], the advantage of using path diversity with multiple description coding (MDC) scheme relative to a single path scheme with FEC and with the same data rate was demonstrated. They used adaptive playout scheduling that sets a playout delay according to a specified loss rate. However, they did not evaluate the benefits of using path diversity when the trade-off between buffering delay and packet loss is optimum. In our work, we show how using a path diversity approach improves the optimum perceived speech quality in a VoIP conversation. Furthermore, we evaluate the improvement for a full redundancy scheme where all packets are sent as redundant and for a partial redundancy scheme where only a fraction of packets, judiciously selected, are sent as redundant information.
In this paper, we explore the approach of path diversity to improve VoIP performance. We design a playout scheduling algorithm based on the optimization of the perceived speech quality and that uses a histogram-based method for delay prediction. In Section 2, we present an overview of the existing playout scheduling algorithms and the design of our algorithm. In Section 3, we present the main ideas of improving VoIP service using path diversity and show the benefits of this approach.
II. DESIGN OF PLAYOUT SCHEDULING ALGORITHM
Fig. 1 shows a network trace of IP packet delays transmitted from UK to China [7]. Delay randomly fluctuates above a minimum value (corresponding to the propagation delay). Sometimes, it shows a spike by suddenly increasing and then decreasing to a level similar to that before the spike. A playout scheduling process is used to buffer packets before their playout. In a simple playout scheduling process, a fixed delay can be used but many packets could be lost in the case of a spike. Then, based on the fact that in a typical real-time voice conversation there are talkspurs (when the user is talking), and periods of silence (when the user is listening), per-talkspurt scheduling algorithms have been developed. In a per-talkspurt approach delays are adjusted at the beginning of each talkspurt and the silence periods are compressed or stretched. However with this approach, when a delay spike occurs during a talkspurt many packets can be lost. A recently introduced approach reduces the effect of spikes by adjusting the delay for packets during a talkspurt [8][9]. Packets for which delay has been adapted may have to be time-scale modified.
The starting point for designing an effective playout scheduling algorithm is to estimate the network delay for future packets. Methods commonly used are statistically-based, in the sense they use the statistics (either statistical parameters or distributions) of past delays to estimate the current playout delay. An early algorithm for playout delay was designed based on an auto-regressive estimation of mean and variation of the network delay [10]. Then, another family of scheduling algorithms that exploit delay statistics (such as pdf, histogram) have been developed [11][12][9]. A delay distribution is stored for the last $w$ packets and used to estimate the playout delay for the next packet. With this approach, we can specify the acceptable loss rate and the algorithm automatically adjusts the delay accordingly. Fig. 1 illustrates the concept of this approach. We generate a histogram of delays for the most recent $w$ packets (we used $w = 1000$), and the playout delay for the next packet is set such that the proportion of packets with an end-to-end delay larger than the end-to-end delay $d_s$ for that packet is equal to the specified target PLR.
Recently, scheduling algorithms have been developed based on the optimization of the perceived speech quality [13][14]. They use a model that translates network impairments (packet loss, packet delay) to perceived speech quality and estimate a playout delay that maximizes the speech quality. In these algorithms, the relation between playout delay and packet loss is based on the modelling of the tail distribution of the delays [13][14]. For designing our playout scheduling algorithm we will use this method of optimization of speech quality, but instead of modelling the packets delay with a specific distribution we use a more general technique that exploits the histogram of delays of received packets.
Spike delays may cause a degradation in speech quality if the playout delay algorithm does not react to their occurrence. A spike detection method has to be integrated in the VoIP application in order to distinguish a normal mode from a spike mode where the playout delay is set differently. A method commonly used sets the playout delay for future packets such that the end-to-end delay is equal to the measured delay for current packet. This technique may improve the performance of VoIP application but delay spikes remain very harmful and the only way to mitigate their effect is to reduce or avoid their occurrence. Using path diversity is a very efficient solution against delay spikes as we will show farther in this work.
Optimum perceived speech quality based algorithm
For designing a playout scheduling algorithm, we use speech quality criteria at the receiver. Fig. 2 shows an overview of a VoIP application. Impairments that affect a VoIP call
quality are network loss, network delay and delay jitter. Delay jitter is mitigated using a well designed playout scheduling algorithm. However, this algorithm induces additional delay and packet loss. It may also affect the speech quality if for example it uses time-scale modification of packets. A PLC technique is used to generate lost packets and the speech quality depends on its performance. This PLC technique may require an additional buffering of received packets before performing signal generation (for example if it uses future packets), which affects the total end-to-end (or mouth-to-ear) delay. The call quality depends on the speech quality achieved and on end-to-end delay.
Depending on the PLC technique used, the playout scheduling algorithm is designed based on the optimization of the call quality (or the perceived speech quality (PSQ)). Since the relation between network performance and PSQ is not straightforward, the E-model was created by the ITU to offer speech assessment based on network performance [15]. This model combines network impairments to provide a scalar quality rating value $R$, ranging from 0 to 100. An estimated Mean Opinion Score (MOS) for the call quality in the scale 1 to 5 can be obtained from the $R$-factor [15]. By considering only the impairments relevant to the design of a VoIP playout scheduling algorithm (end-to-end delay and loss), the $R$-factor can be written as [15]
$$R = 93.2 - I_d - I_e \quad (1)$$
where $I_d$ is the impairment factor caused by packet delay and $I_e$ is the impairment factor caused by packet loss. $I_d$ can be calculated using an analytical expression given in [15] and is shown in Fig. 3a. $I_e$ has been evaluated by subjective listening tests for different coders and for random loss [16]. For G.711, the coder we use in our analysis, results are shown in Fig. 3b.
Packet loss includes network loss and scheduling induced loss, which depends on the playout delay. If we can establish a relation $L(d)$ between the total loss and delay, the $R$-factor can be written in terms of delay as
$$R(d) = 93.2 - I_d(d) - I_e(L(d)) \quad (2)$$
Then, we can find the value of $d$ that maximizes $R(d)$ in (2) and set the playout delay accordingly. In order to find a relation between loss and delay, recent studies have modelled the delay distributions in the tail region as a Weibull [14] and a Pareto [13] distributions. However, we noticed that the playout delay may be very sensitive to the type of distribution used. In our design we use a more general method that makes no $a$ priori assumption on the delay distribution. We simply use the histogram of the most recent $w$ delays as shown in Fig. 1 to calculate the loss as a function of delay.
With our histogram-based method for $w = 1000$ and using measured packet delays of Fig. 1, we simulated the behaviour of our playout scheduling algorithm. In the case of a delay spike, we set the end-to-end delay for the next packet to the measured delay for current packet. Fig. 4 shows the end-to-end delay results. Variations in the end-to-end delay show the ability of the playout scheduling algorithm to track changing conditions in the network.
III. USING PATH DIVERSITY
The playout scheduling algorithm described in the previous section optimizes the PSQ by finding the best compromise between loss and delay in normal mode, and increases the playout delay in spike mode. Then, the PLC method uses the available speech signal to generate the missing packets. Hence, the performance of a VoIP application using this algorithm is limited by the best compromise between loss and delay achieved in normal mode, by the frequency of occurrence of delay spikes, and by the amount of speech available for the PLC method. In order to overcome these limitations and improve the performance of the VoIP application, we can send redundant packets through a different path independent of the main path. If the transmission bandwidth is available, we can improve the trade-off between loss and delay and reduce the frequency of occurrence of spikes simply by duplicating and sending all packets on the second path. If the transmission bandwidth is limited, we can use the second path to transmit only a part of the information. Previous studies showed that in a speech signal, certain packets are more important than others [17] in the sense that they are more difficult to regenerate well in case of loss. Then, by sending these packets as redundant on a second path, we can improve the speech quality at the receiver. We will analyze these path diversity schemes.
A. Full redundancy
If we are willing to significantly increase the bit-rate to get a better performance, we can duplicate 100% of packets and transmit them on a second path. The receiver selects the first arriving packet to reconstruct speech signal. However, in order to improve the VoIP performance by using this path diversity scheme, the second path has to be independent from the main path. Indeed, if the performances of the two paths are highly correlated, there is no significant gain in using the second path. Furthermore, the performances of the two paths in terms of minimum delay and jitter must not be very different. If for example packets from the second path arrive much later than those from the first path, they only can be used if the buffering delay is dramatically increased, which may compromise performance.
In addition to improving the best compromise loss-delay in normal mode, using this path diversity scheme reduces the effect of delay spikes on the speech quality. Indeed, when the two paths are independent, the probability of having both delays of the first and second paths in spike mode is small. And even if the delays of both paths are in spike mode, the scheduling algorithm sets the playout delay for the next packet according to the minimum of the current delays for the two paths.
The playout delay algorithm in the two-paths scheme is designed using the PSQ optimization method. The $R$-factor from (2) is used with a function $L(d)$ measured for the two-paths scheme. Assuming the two paths are independent, $L(d)$ is calculated by multiplying the loss functions corresponding to each path.
We simulated the behaviour of our playout scheduling algorithm for a two-paths scenario. Since we have no simultaneous delay measurements for two different paths with the same source and destination, we used for the first path, the measured delay trace of Fig. 1, and for the second path another portion of this trace. Results of end-to-end delay are shown in Fig. 4. Compared to the simulations results for a single path scheme, the end-to-end delay is significantly reduced in the periods where it was large for a single path scheme, and shows a lower variation with time. Furthermore, using path diversity, the effect of delay spikes is completely eliminated, which would highly improve the perceived speech quality.
Using the simulation results for single and two-paths cases and data of Fig. 3, we calculated the $R$-factor from (1). For each packet, the delay impairment factor is calculated for the simulated end-to-end delay and the loss impairment factor is calculated for a loss ratio measured for the most recent 1000 packets. Fig. 5 shows the improvement to the $R$-factor by using a two-paths scheme. The $R$-value shows a minimum of around 65 for the whole simulation period.
B. Partial redundancy
Duplicating 100% of packets in the second path is not efficient in terms of the bit-rate used. If we aim to improve the performance of VoIP application without significantly increasing the bit-rate, we can improve the speech quality at the receiver by sending only the most important packets (or the packets that are difficult to generate in case of loss) through a second path.
The importance of voiced frames preceded with unvoiced frames and transitional (unvoiced to voiced) frames for packet loss concealment and for speech quality has been confirmed in recent studies [17][18]. In a study of redundancy-based scheme for PLC of G.723.1 codec, it was shown that speech quality is significantly improved by using redundant information about important packets in the reconstruction process. With a simple method used for packet classification, it was determined that only 11% of packets are important [17]. Hence, by sending these packets as redundant information through a second path, the bit-rate is only slightly increased.
Design of a playout scheduling algorithm based on the optimization of PSQ requires a real-time identification of important packets and a characterization of the combined effect of packet loss for these packets on the second path and for all packets on the main path. Then, using the loss functions for the two paths $L_1(d)$ and $L_2(d)$, we can write the $R$-factor in terms of the end-to-end delay $d$ as
$$R(d) = 93.2 - I_d(d) - I_e(L_1(d), L_2(d)) \quad (3)$$
where $I_e$ is the loss impairment factor caused by the combined packet loss. The playout delay is set according to the value of $d$ that maximizes the $R$-factor.
In practice, a real-time selection process for important packets can be for example based on packet classification according to voiced/unvoiced criteria [17]. In our study, and for the purpose of showing the benefit and efficiency of this
redundancy scheme, we apply a non real-time method on a recorded speech signal. We identify the important packets by dropping from the speech signal one packet at a time and estimating the MOS after applying the built-in G.711 PLC method [19]. MOS estimation is performed using an objective method of Perceptual Evaluation of Speech Quality (PESQ) [20]. We select important packets as those corresponding to a MOS value less than $MOS_o - 2\sigma$, where $MOS_o$ is the MOS measured in a situation where no packet is dropped, and $\sigma$ is the variance of MOS for all packets. With this selection criteria, the proportion of important packets in the speech signal is 21%.
To characterize the impact of packet loss for the first and second paths on speech quality, we use the same speech signal and simulate different loss scenarios for these packets on the second path as well as for all packets on the first path. For fixed PLRs $PLR_1$ and $PLR_2$ for first and second paths respectively, we randomly select from the speech signal a group of packets corresponding to $PLR_1$ (they correspond to packets lost on path 1), then we select from the speech signal another group of important packets corresponding to $PLR_2$ (they correspond to important packets lost on path 2). Then, we identify packets that are in both groups, which are considered lost on both paths and are removed from the original speech signal. Then, we apply the PLC method to the resulting speech signal for which a MOS value is estimated using the PESQ tool. For each value of $PLR_1$ and $PLR_2$, this experiment is repeated 100 times and the resulting MOS values are averaged. Then, we convert MOS scores to $R$-values using a relation given by the E-model [15], and calculate a raw loss impairment factor as $I_r = 100 - R$. For a single path scheme, we map these $I_r$ values to the ITU’s values of Fig. 3b and calculate a mapping function that we use to convert $I_r$ to loss impairment factor $I_e$ for a two-paths scheme.
Fig. 6 shows the loss impairment factor estimated with the method described above, corresponding to different packet loss ratios for the first and second paths. We used the values of 0, 20, 50, 80, and 100% for $PLR_2$. For other values, the loss impairment factor can be interpolated. Using these data in (3) and delay measurements of Fig. 1, we calculated the $R$-factor corresponding to the two-paths scheme. From the histogram of packet delays we calculated the function $L_1(d)$ and we used the same function for $L_2(d)$ (assuming that the two paths have the same performance). Fig. 7 shows the $R$-factor at different end-to-end delay values for single path and two-paths schemes. The playout scheduling algorithm sets a playout delay that corresponds to the maximum value of the $R$-factor, which is, according to Fig. 7, improved by 15 in a two-paths scheme.
IV. CONCLUSION
In this paper, we have proposed two different path diversity schemes for improving VoIP performance. After we presented an overview of a VoIP application and showed how its components affect its performance, we first designed a new playout scheduling algorithm based on the optimization of perceived speech quality and using the histogram of packet delays in a single path scheme. Then, we identified the limitations of a VoIP application for this scheme and proposed improvement techniques using full and partial path diversity approaches. We presented simulations results that showed a significant improvement for the perceived speech quality of the VoIP application by using path diversity.
ACKNOWLEDGMENT
The authors would like to thank NSERC, Canada for supporting this research.
REFERENCES
[1] C. Perkins, O. Hodson, and V. Hardman, “A survey of packet loss recovery techniques for streaming audio,” *IEEE Network*, vol. 12, pp. 40–48, Sept-Oct 1998.
[2] V. Hardman, A. Sasse, and A. Watson, “Reliable audio for use over the Internet,” in *Proceedings of INET ’95*, June 1995, pp. 171–178.
[3] Y. J. Liang, E. G. Steinbach, and B. Girod, “Real-time Voice Communication over the Internet Using Packet Path Diversity,” *Proceedings of the ninth ACM international conference on Multimedia*, vol. 9, pp. 431–440, Sept. 2001.
[4] J. G. Apostolopoulos, “Reliable Video Communication over Lossy Packet Networks using Multiple State Encoding and Path Diversity,” *Visual communications and Image Processing (VCIP)*, pp. 392–409, Jan. 2001.
[5] Z. Duan, Z.-L. Zhang, and Y. T. Hou, “Service Overlay Networks: SLAs, QoS, and Bandwidth Provisioning,” *IEEE/ACM Transactions on Networking*, vol. 11, no. 6, pp. 870–883, Dec. 2003.
[6] S. Guha, N. Daswani, and R. Jain, “An Experimental Study of the Skype Peer-to-Peer VoIP System,” in *Proceedings of The 5th International Workshop on Peer-to-Peer Systems (IPTPS ’06)*, Santa Barbara, CA, Feb. 2006.
[7] L. Sun and E. Ifeachor, “Prediction of perceived conversational speech quality and effects of playout buffer algorithms,” in *Proc. IEE ICC*, Anchorage, USA, May 2003, pp. 1–6.
[8] F. Liu, J. Kim, and C.-C. J. Kuo, “Adaptive delay concealment for Internet voice applications with packet-based time-scale modification,” in *Proc. IEEE ICASSP*, vol. 3, May 2001, pp. 1461–1464.
[9] Y. J. Liang, N. Farber, and B. Girod, “Adaptive playout scheduling and loss concealment for voice communication over IP networks,” *IEEE Trans. Multimedia*, vol. 5, no. 4, pp. 532–543, Dec. 2003.
[10] R. Ramjee, J. Kurose, D. Towsley, and H. Schulzrinne, “Adaptive playout mechanisms for packetized audio applications in wide-area networks,” in *Proc. IEEE INFOCOM*, vol. 2, June 1994, pp. 680–688.
[11] S. B. Moon, J. Kurose, and D. Towsley, “Packet audio playout delay adjustment: Performance bounds and algorithms,” *Multimedia Systems*, vol. 6, pp. 17–28, Jan. 1998.
[12] K. Fujimoto, S. Ata, and M. Murata, “Statistical analysis of packet delays in the Internet and its application to playout control for streaming applications,” *IEICE Trans. Communications*, vol. E84-B, no. 6, pp. 1504–1512, June 2001.
[13] ——, “Adaptive playout buffer algorithms for enhancing perceived quality of streaming applications,” in *Proc. IEEE GLOBECOM*, vol. 21, no. 1, Nov 2002, pp. 2463–2469.
[14] L. Sun and E. Ifeachor, “New models for perceived voice quality prediction and their applications in playout buffer optimization for VoIP network,” in *Proc. IEEE ICC*, Paris, France, June 2004, pp. 1478–1483.
[15] I.R.G.107, “The E-Model, a computational model for use in transmission planning,” Mar. 2003.
[16] I.R.G.113, “Transmission impairments due to speech processing,” Feb. 2001.
[17] L. Tosun, “Dynamically adding redundancy for improved error concealment in packet voice coding,” MSc thesis, McGill University, Dec. 2004.
[18] L. Sun, G. Wade, B. Lines, and E. Ifeachor, “Impact of packet loss location on perceived speech quality,” in *Proc. of 2nd IP Telephony Workshop (IPTEL ’01)*, Columbia University, New York, Apr. 2001, pp. 114–122.
[19] ITU-T Recommendation G.711, “Appendix I: A high quality low-complexity algorithm for packet loss concealment with G.711,” Sept. 1999.
[20] ITU-T Recommendation P.862, “Perceptual evaluation of speech quality (PESQ): An objective method for end-to-end speech quality assessment of narrow-band telephone networks and speech codecs,” Feb. 2001. |
REGULATORY INFORMATION DISTRIBUTION SYSTEM (RIDS)
ACCESSION NBR: 8812160071 DOC. DATE: 88/11/30 NOTARIZED: NO DOCKET #
FACIL: 50-250 Turkey Point Plant, Unit 3, Florida Power and Light C 05000250
50-251 Turkey Point Plant, Unit 4, Florida Power and Light C 05000251
AUTH. NAME: CONWAY, W.F. AUTHOR AFFILIATION: Florida Power & Light Co.
RECIP. NAME: ERNST, M.L. RECIPIENT AFFILIATION: Region 2, Ofc of the Director
SUBJECT: Forwards mgt-on-shift weekly rept for wk starting 881118.
DISTRIBUTION CODE: D036D COPIES RECEIVED: LTR 1 ENCL 1 SIZE: 29
TITLE: Turkey Point Management Onshift Program
NOTES:
| RECIPIENT ID CODE/NAME | COPIES LTR ENCL | RECIPIENT ID CODE/NAME | COPIES LTR ENCL |
|------------------------|----------------|------------------------|----------------|
| DRP/ADR-2 | 1 1 | NRR/DRP-1/2 | 1 1 |
| PD2-2 PD | 1 1 | EDISON,G | 2 2 |
INTERNAL:
AEOD 1 1
NRR DEPY 1 1
NRR MORISSEAU,D 1 1
NRR/ADT 12-G-18 1 1
NRR/DRIS DIR 9A 1 1
OE LIEBERMAN,J 1 1
REG FILE 01 1 1
EXTERNAL:
LPDR 1 1
NSIC 1 1
NRC PDR 1 1
NOTE TO ALL "RIDS" RECIPIENTS:
PLEASE HELP US TO REDUCE WASTE! CONTACT THE DOCUMENT CONTROL DESK,
ROOM P1-37 (EXT. 20079) TO ELIMINATE YOUR NAME FROM DISTRIBUTION
LISTS FOR DOCUMENTS YOU DON'T NEED!
TOTAL NUMBER OF COPIES REQUIRED: LTR 22 ENCL 22
Mr. Malcolm L. Ernst
Acting Regional Administrator, Region II
U.S. Nuclear Regulatory Commission
101 Marietta Street, N.W., Suite 2900
Atlanta, Georgia 30323
Dear Mr. Ernst:
Re: Turkey Point Units 3 and 4
Docket Nos. 50-250 and 50-251
Management-on-Shift Weekly Report
Pursuant to the Nuclear Regulatory Commission Order dated October 19, 1987, the attached summary of Management-on-Shift (MOS) reports is submitted. The Plant Supervisor-Nuclear Shift Reports are also being submitted.
Should there be any questions on this information, please contact us.
Very truly yours,
W. F. Conway
Senior Vice President - Nuclear
WFC/RHF/gp
Attachment
cc: J. Lieberman, Director, Office of Enforcement, USNRC
Dr. G. E. Edison, Project Manager, NRR, USNRC
Senior Resident Inspector, USNRC, Turkey Point Plant
R. E. Tallon, President, FPL
The following is a list of the major research areas in which the Department of Mathematics has been active during the past few years:
1. **Algebraic Geometry**
- Study of algebraic varieties and their properties.
2. **Number Theory**
- Research into the properties of numbers, particularly integers.
3. **Differential Equations**
- Analysis of equations involving derivatives.
4. **Functional Analysis**
- Study of vector spaces with additional structure.
5. **Topology**
- Examination of properties preserved under continuous transformations.
6. **Probability Theory**
- Study of random phenomena and their mathematical models.
7. **Statistics**
- Collection, analysis, interpretation, and presentation of data.
8. **Numerical Analysis**
- Development of algorithms for solving mathematical problems.
9. **Mathematical Physics**
- Application of mathematical methods to physical problems.
10. **Applied Mathematics**
- Use of mathematical techniques to solve real-world problems.
11. **Mathematical Logic**
- Study of formal systems and their properties.
12. **Combinatorics**
- Study of discrete structures and their properties.
13. **Graph Theory**
- Study of graphs and networks.
14. **Cryptography**
- Development of secure communication methods.
15. **Optimization**
- Finding the best solution among many possible solutions.
16. **Game Theory**
- Study of strategic decision-making in competitive situations.
17. **Mathematical Finance**
- Application of mathematical methods to financial markets.
18. **Data Science**
- Analysis and interpretation of complex data sets.
19. **Machine Learning**
- Development of algorithms that enable computers to learn from data.
20. **Artificial Intelligence**
- Creation of intelligent machines capable of performing tasks that require human intelligence.
These areas represent the breadth and depth of research conducted within the department, reflecting its commitment to both theoretical and applied mathematics.
Four MOS Observers were on shift. M. D. Shepherd, Training Supervisor, St. Lucie Plant (11/18-20/88, days); W. Detwiler, Westinghouse Electric Corporation (11/21-25/88, days); J. W. Patterson, Westinghouse Electric Corporation (11/18-21/88, nights); and G. M. Smith, Westinghouse Electric Corporation (11/21-25/88, nights).
Unit 3 remained in mode 5 and Unit 4 remained defueled for the duration of the reporting period.
The independent observers did not note any questionable work practices. They did note one area for improvement about operators submitting design change recommendations.
The Plant Supervisors-Nuclear reported one questionable work practice on valves being disassembled without specific clearances requested for that work. They also noted one area for improvement regarding work controls on the connection of test leads.
ATTACHMENT: MOS DAILY REPORTS
The following is a list of the most common types of errors that can occur during the process of data entry:
1. **Typing Errors**: These occur when the data is entered incorrectly due to a mistake in typing, such as entering a number instead of a letter or vice versa.
2. **Omission Errors**: These happen when some information is not entered at all, either because it was not available or because the person entering the data forgot to include it.
3. **Duplication Errors**: These occur when the same piece of information is entered more than once, either by accident or intentionally.
4. **Inconsistency Errors**: These happen when the data entered does not match the format or structure expected by the system, such as entering a date in the wrong format.
5. **Validation Errors**: These occur when the data entered does not meet certain criteria set by the system, such as entering a negative value for a quantity that should be positive.
6. **Format Errors**: These happen when the data entered does not conform to the required format, such as entering a phone number without including the area code.
7. **Syntax Errors**: These occur when the data entered does not follow the rules of the programming language used to create the system, such as using a reserved keyword as a variable name.
8. **Logical Errors**: These happen when the data entered does not make sense in the context of the system, such as entering a date in the future for an event that has already occurred.
9. **Security Errors**: These occur when the data entered is not secure, such as entering sensitive information in an insecure environment.
10. **Performance Errors**: These happen when the data entered causes the system to run slowly or crash, such as entering a large amount of data into a database that is not optimized for handling large volumes of data.
To minimize these errors, it is important to have a thorough understanding of the data being entered and the system that will be used to store and process it. It is also important to have a clear set of guidelines and instructions for data entry, as well as regular training and support for those who will be entering the data. Finally, it is important to have a robust error detection and correction process in place to identify and fix any errors that do occur.
None
B. Areas for Improvement/Recommendations/Actions Taken
While performing 0-PMI-046.1, Boric Acid tank level indicator calibration, Unit 3 Residual Heat Removal (RHR) was isolated and a Significant Event made to the NRC due to inadvertant arcing when leads were being connected between a digital multimeter and terminals 13 and 14 in 3-QR-6, causing breaker 3P05-5 to trip, de-energizing among other equipment PC-3-403 that gave a false hi Reactor Coolant System (RCS) pressure signal causing closure of MOV-3-750 loop 3C RHR pump suction stop valve. Proper or improved work controls such as using shielded leads and OTSC to the procedure to allow using probes should be considered.
(88-3074)
C. Good Practices/Professionalism Observed
Good operator response to an off-normal situation that required immediate, positive and correct action, with prompt usage of 3-ONOP-050 for loss of RHR subsequent actions. The event from the beginning of closure of MOV-3-750 to the restart of the RHR pump lasted 5 minutes.
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
Reviewed By: [Signature]
William Detwiler
(Observer)
A. Plant Evolutions Observed
- Loss of Residual Heat Removal (RHR)
- Start of 3B1 main circulating water pump
- Instrument bus inverter autoswap to CVT
- Unit 4 containment tour
- Shift relief
- 0600 outage meeting
- Shift turnover meeting
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
None
E. Professionalism, Summary of Shift, Comments
Good job by the control room day shift concerning the loss of RHR. Quick action was taken to restore the RHR flow. The PSNs took personal responsibility to classify the incident, investigate it and report it to the NRC. An OTSC was quickly initiated to prevent recurrence.
Reviewed By: Operations Superintendent-Nuclear
Date: 11/22/88
Management Review By:
GM-N Date SVP Date VP Date
Gregg Smith
(Observer)
A. Plant Evolutions Observed
- Cold shutdown operations, units 3 and 4
- Shift briefing - nights
- Shift turnover - peaks to nights
- Plant tour of secondary systems - units 3 and 4
- Containment tour - unit 4
- Tour of Radiation Control Area (RCA)
- 3-OSP-062.4, Safety Injection System Full Flow pump and valve test (from control room)
B. Immediate Safety Problems
None observed
C. Questionable Work Practices
None observed
D. Areas for Improvement
None at this time
E. Professionalism, Summary of Shift, Comments
Routine operations - no unprofessional behavior observed.
Reviewed By: [Signature]
Operations Superintendent-Nuclear
Date: 11/22/88
Management Review By:
PM-N [Signature] Date: 11/22/88
SVP [Signature] Date: 11/22/88
VP [Signature] Date:
The following is a list of the major research areas in which the Department of Mathematics has been active during the past few years.
1. **Algebraic Geometry**
- Study of algebraic varieties and their properties, including moduli spaces, birational geometry, and arithmetic geometry.
2. **Analysis**
- Research in harmonic analysis, partial differential equations, and functional analysis.
3. **Combinatorics**
- Study of discrete structures and their properties, including graph theory, combinatorial optimization, and enumerative combinatorics.
4. **Differential Equations**
- Analysis of ordinary and partial differential equations, with applications to physics, engineering, and biology.
5. **Geometry**
- Study of geometric structures and their properties, including differential geometry, geometric topology, and geometric group theory.
6. **Number Theory**
- Study of the properties of integers and their relationships, including algebraic number theory, analytic number theory, and Diophantine equations.
7. **Probability and Statistics**
- Research in probability theory, stochastic processes, and statistical methods.
8. **Topology**
- Study of topological spaces and their properties, including algebraic topology, geometric topology, and low-dimensional topology.
9. **Applied Mathematics**
- Development of mathematical models and methods for solving real-world problems in science, engineering, and industry.
10. **Mathematical Physics**
- Application of mathematical techniques to problems in physics, including quantum mechanics, statistical mechanics, and condensed matter physics.
11. **Mathematical Biology**
- Use of mathematical models to study biological systems, including population dynamics, evolutionary biology, and neuroscience.
12. **Mathematical Finance**
- Application of mathematical techniques to financial markets, including risk management, portfolio optimization, and derivative pricing.
13. **Mathematical Education**
- Research on teaching and learning mathematics, including curriculum development, assessment, and teacher professional development.
14. **Mathematical Software**
- Development of software tools for mathematical computation and visualization, including symbolic computation, numerical analysis, and data analysis.
15. **Mathematical Logic**
- Study of formal systems and their properties, including model theory, proof theory, and computability theory.
These research areas are interconnected and often overlap, reflecting the interdisciplinary nature of modern mathematics. The department continues to foster a vibrant research environment, supporting both fundamental and applied research across these diverse fields.
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
Reviewed By: [Signature]
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operability
Reviewed By: L.W. Pierce Date: 11/23/88
William Detwiler
(Observer)
A. Plant Evolutions Observed
- 0600 outage meeting
- Shift turnover meeting
- Testing unit 3 A Diesel Generator
- Testing 286 valve in Chemical Volume Control System (CVCS)
- Preparation for 8 hour test run on unit 3 Diesel Generator
- Unit 3 containment tour
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
None
E. Professionalism, Summary of Shift, Comments
The day shift did a good job with the A Diesel Generator test preparations. Because the procedure was carefully reviewed a problem was found. An OTSC was immediately written to take care of it. The FSN carefully reviewed the OTSC and noticed a step missing. The OTSC was then corrected. The careful review of the procedure saved time and ensured the success of the test.
Reviewed By: [Signature]
Operations Superintendent - Nuclear
Date: 11/23/88
Management Review By: [Signature]
PM-N Date: 11/23/88
SVP Date: 11/23/88
Gregg Smith
(Observer)
A. Plant Evolutions Observed
- Pre-shift briefing (nights)
- Shift turnover - peaks to nights
- Unit 4 containment tour
- Radiation Control Area (RCA) tour
- Tour of Units 3 and 4 secondary plant
- "A" Emergency Diesel 8 hour run (terminated)
- Portions of maintenance performed on "reotemp" meter and level switches for skid tank for "A" EDG
B. Immediate Safety Problems
None observed
C. Questionable Work Practices
None observed
D. Areas for Improvement
None at this time
E. Professionalism, Summary of Shift, Comments
-Routine operations, (cold shutdown)
-No unprofessional behavior observed
-Peak shift PSN, AFSN, and lead engineer for EDG worked well together to ascertain the extent of problems with "A" EDG, determine 8 hour test run should be terminated, and initiate paperwork necessary to correct problems so the test run could be recommenced in an expeditious time frame.
Reviewed By: [Signature]
Operations Superintendent-Nuclear
Date: 11/23/88
Management Review By:
PM-N Date SVP Date VP Date
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
Reviewed By: [Signature]
William Detwiler
(Observer)
A. Plant Evolutions Observed
- 0600 outage meeting
- Shift turnover meeting
- Unit 4 containment tour
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
None
E. Professionalism, Summary of Shift, Comments
Quiet shift
Reviewed By: [Signature]
Operations Superintendent-Nuclear
Date: 11/25/88
Management Review By:
AC 1/25/88
PM-N Date
SVP 1/25/88
VP Date
Gregg Smith
(Observer)
Shift: Nights
A. Plant Evolutions Observed
- Pre-shift briefing (nights)
- Shift turnover - peaks to nights
- Tour of units 3 and 4 secondary plant
- Tour of Radiation Control Area (RCA) and unit 4 containment
- Commencement of "A" Emergency Diesel Generator (EDG) 8 hour test run
- 0-OP-003.3 120 V vital instrument AC system (swap from spare to normal)
B. Immediate Safety Problems
None observed
C. Questionable Work Practices
None observed
D. Areas for Improvement
None at this time
E. Professionalism, Summary of Shift, Comments
-Routine holiday operation
-No unprofessional behavior observed
-The RCO did a good job of determining that there was not an existing portion of 0-OP-003.3 "120 V Vital Instrument AC System" procedure which deals with transferring instrument bus back to its normal inverter after an auto transfer due to a momentary fault (restoring unit 3 Qualified Safety Parameter Display System (QSPDS) channel A). Section 7, which deals with transfer, has you verify the inverter is in a standby condition per section 5 (which assumes the inverter is de-energized initially). The RCO brought this to the attention of the FSN and APSN who properly modified the pre-requisites steps per O-ADM-201 (section 5.2.12). The on-shift PUP representative was then consulted and requested to investigate the procedure and hopefully develop a new section in the procedure 0-OP-003.3 which specifically provides guidance for the situation which occurred.
Reviewed By: Operations Superintendent-Nuclear
Date: 11/25/88
Management Review By:
PM-N Date
BVP Date
VP Date
The following is a list of the major research areas in which the Department of Mathematics has been active during the past few years.
1. **Algebraic Geometry**
- Study of algebraic varieties and their properties.
- Focus on moduli spaces, birational geometry, and arithmetic geometry.
2. **Analysis**
- Functional analysis, harmonic analysis, and partial differential equations.
- Research into spectral theory, operator algebras, and geometric measure theory.
3. **Combinatorics**
- Graph theory, combinatorial optimization, and discrete mathematics.
- Focus on extremal problems, Ramsey theory, and probabilistic methods.
4. **Differential Equations**
- Ordinary and partial differential equations, including nonlinear systems.
- Applications to physics, engineering, and biology.
5. **Geometry**
- Differential geometry, Riemannian manifolds, and geometric analysis.
- Study of curvature, geodesics, and minimal surfaces.
6. **Number Theory**
- Algebraic number theory, Diophantine equations, and analytic number theory.
- Research into prime numbers, modular forms, and elliptic curves.
7. **Probability and Statistics**
- Stochastic processes, statistical inference, and data analysis.
- Applications to finance, biology, and social sciences.
8. **Topology**
- General topology, algebraic topology, and geometric topology.
- Focus on homotopy theory, knot theory, and manifold theory.
9. **Applied Mathematics**
- Mathematical modeling, numerical analysis, and computational methods.
- Applications to engineering, economics, and environmental science.
10. **Mathematical Physics**
- Quantum mechanics, statistical mechanics, and condensed matter physics.
- Research into quantum field theory, string theory, and integrable systems.
These areas represent the breadth and depth of mathematical research conducted within the department, reflecting both theoretical and applied aspects of mathematics.
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
Reviewed By: [Signature]
| Date | Time | Location | Event Description |
|------------|--------|-----------|-------------------|
| 2023-10-01 | 14:00 | Room A | Workshop on Data Analysis |
| 2023-10-02 | 09:00 | Auditorium| Annual Meeting |
| 2023-10-03 | 16:30 | Lab B | Research Presentation |
| 2023-10-04 | 10:00 | Conference Center | Symposium on Sustainable Practices |
*Note: All events are held in the University of California, Berkeley.*
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
Reviewed By: [Signature]
W. L. Derwiler
(Observer)
A. Plant Evolutions Observed
- Load rejection test 4A Diesel Generator
- Shutdown of 4A Diesel Generator
- Plant tour
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
There are certain parts of systems that are either poorly designed, do not work as designed or are constantly malfunctioning due to poor design. They include blowdown flow control valves, blowdown flow indication, condenser makeup spray valves and steam dump control valves. Submitting a PWO will not fix the problem or will not fix them for more than a short time. The proper path would be to submit a request for Engineering assistance (REA). But after talking to several RCOs and PSNs, there is a reluctance to submit REAs which could address design problems. They feel that the REA will be ignored or rejected due to the cost of changing system designs. If Turkey Point is to be one of the best managed plants in the nation then these design problems need to be documented and addressed, otherwise improvements may never be made. Operators, especially RCOs and NOs should be encouraged to submit REAs since they are closest to the problems. Some sort of response with pertinent information (e.g., action to be taken) should be returned to the submitter within a reasonable time even if the problem cannot be corrected in the near term.
(88-3089)
E. Professionalism, Summary of Shift, Comments
Could not tour units 3 or 4 containment because holiday routine limited HP coverage of containment to critical path work only.
Good job by unit 3 RCO when he insisted that the instrumentation hook-ups and chart speed be verified before allowing the load reject test for 4A diesel generator to begin.
Reviewed By: Operations Superintendent-Nuclear
Management Review By:
Gregg Smith
(Observer)
Shift: Night
A. Plant Evolutions Observed
- 4-OSP-059.1 Source Range Nuclear Instrumentation Analog Channel operation test
- Shift turnover (peaks to nights)
- Pre-shift meeting (nights)
- Secondary plant tour (units 3 and 4)
- Commencement of "A" Emergency Diesel Generator 8 hour test run
B. Immediate Safety Problems
None observed
C. Questionable Work Practices
None observed
D. Areas for Improvement
None at this time
E. Professionalism, Summary of Shift, Comments
Note: Due to limited HP coverage (holiday routine) containments were closed, therefore no containment tour.
No unprofessional behavior observed
Routine holiday operations
Reviewed By: [Signature]
Operations Superintendent - Nuclear
Date: 11/25/88
Management Review By: [Signature]
PM-N Date: 11/25/88
[Signature]
Date: 11/25/88
VP Date:
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
Reviewed By: [Signature]
M. D. Shepherd
(Observer)
A. Plant Evolutions Observed
- 0600 outage meeting
- Pre-shift meeting
- Routine control room operations
- Plant tour
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
None
E. Professionalism, Summary of Shift, Comments
None
Reviewed By: [Signature]
Operations Superintendent-Nuclear
Date: 1/2/89
Management Review By:
PM-N Date SVP Date VP Date
John Patterson (Observer)
A. Plant Evolutions Observed
- Unit 3 in mode 5, unit 4, defueled
- Shift turnover/briefing
- Tour of intake structure
- Draindown of unit 4 intake cooling water (ICW) piping
- Tour of unit 3 containment
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
None
E. Professionalism, Summary of Shift, Comments
Routine shift operations
Reviewed By: [Signature]
Operations Superintendent-Nuclear
Date: 11/21/88
Management Review By:
PM-N Date SVP Date VP Date
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
Reviewed By: [Signature]
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
M. D. Sheehard (Observer)
A. Plant Evolutions Observed
- 0800 outage meeting
- Pre-shift briefings
- Licensed operator requal training - simulator
- Routine control room operations
- Plant tour
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
None
E. Professionalism, Summary of Shift, Comments
None
Reviewed By: Operations Superintendent-Nuclear Date: 1/25/89
Management Review By: PM-N Date SVP Date VP Date
John Patterson
(Observer)
Shift: Night
A. Plant Evolutions Observed
- Unit 3 in mode 5, unit 4 defueled
- Shift turnover/briefing
- Tour of intake structure
- Tour of unit 4 containment
- 3-OSP-62.4 Safety Injection system full-flow pump and valve test
- Unit 4 condenser waterbox dehumidification
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
Routine shift operations
E. Professionalism, Summary of Shift, Comments
Routine shift operations
Reviewed By: [Signature]
Operations Superintendent - Nuclear
The following is a list of the major research areas in which the Department of Mathematics at the University of Illinois at Chicago is active:
1. Algebra and Number Theory
2. Analysis
3. Applied Mathematics
4. Combinatorics
5. Geometry and Topology
6. Logic and Foundations
7. Probability and Statistics
8. Representation Theory
These areas encompass a wide range of mathematical disciplines, from pure mathematics to applied mathematics, and include both theoretical and computational aspects. The department is committed to advancing knowledge in these fields through research, teaching, and collaboration with other institutions and researchers around the world.
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
Reviewed By: [Signature]
None
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
Reviewed By: [Signature]
M. D. Shepherd
(Observer)
A. Plant Evolutions Observed
- 0800 outage meeting
- Pre-shift briefings
- Licensed operator requal training - simulator
- 1400 outage meeting
- Routine control room operations
- Plant tour
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
None
E. Professionalism, Summary of Shift, Comments
None
Reviewed By: [Signature]
Operations Superintendent-Nuclear
Date: 11/21/88
Management Review By:
PM-N Date SVP Date VP Date
The following is a list of the major topics covered in this thesis:
1. Introduction
2. Literature Review
3. Methodology
4. Results and Discussion
5. Conclusion
This thesis aims to provide a comprehensive analysis of the topic, with a focus on the methodology used to gather and analyze data. The results and discussion section will present the findings and their implications, while the conclusion will summarize the main points and suggest areas for future research.
References:
[1] Smith, J. (2010). *A Comprehensive Guide to Research Methods*. New York: Oxford University Press.
[2] Jones, M. (2015). *Methodological Approaches in Social Science Research*. London: Sage Publications.
[3] Brown, L. (2018). *Data Analysis Techniques for Social Scientists*. Washington D.C.: American Psychological Association.
John Patterson
(Observer)
Shift: Nights
A. Plant Evolutions Observed
- Unit 3 in mode 5, unit 4, defueled
- Shift turnover/briefing
- Tour of intake structure
- Tour of unit 4 containment
- OSP-200.1 main security entrance radio check
- 4-OP-079 section 7.2, filling "A" steam generator via unit 4 condensate transfer pump
B. Immediate Safety Problems
None
C. Questionable Work Practices
None
D. Areas for Improvement
None
E. Professionalism, Summary of Shift, Comments
The shift took prompt, decisive action to identify and mitigate the consequences of a feed system leak.
Reviewed By: [Signature]
Operations Superintendent - Nuclear
Date: 11/21/88
Management Review By:
PM-N Date SVP Date VP Date
Valves 4-20-121 and 4-20-221 were disassembled in accordance with two different PWOs. Credit was taken for two clearances already hanging. However, no clearance requests were made for these jobs, thus there was no review to assure these valves were properly isolated. Neither clearance listed on the PWOs properly isolated either valve. Consequently when we attempted to fill the "A" Steam Generator, we dumped 3,000 to 4,000 of demineralized water through these disassembled valves.
(88-3073)
B. Areas for Improvement/Recommendations/Actions Taken
None
C. Good Practices/Professionalism Observed
Routine operations
viewed By: [Signature] Date: 11/21/88 |
I AGREE for myself, and/or my child, my/our administrators and assigns, in consideration for my, and/or my child’s, participation in this United States Pony Clubs, Inc. (USPC) activity to the following:
I AGREE that I choose to participate voluntarily in an USPC activity, as a rider, driver, handler, lessee, owner, agent, spectator, volunteer, and/or trainer. I am fully aware and acknowledge that horse sports and USPC activities involve inherent dangerous risks of accident, loss, and serious bodily injury including, but not limited to, broken bones, head injuries, trauma, pain, suffering or death (“Harm”). I fully understand that this release covers, but is not limited to, inherent risks of an equine activity which mean a danger or condition that is an integral part of an equine activity, including but not limited to, any of the following:
- The propensity of an equine to behave in ways that may result in injury, death, or loss to persons on or around the equine;
- The unpredictability of an equine’s reaction to sounds, sudden movement, unfamiliar objects, persons, or other animals;
- Hazards, including, but not limited to, surface or subsurface conditions;
- A collision with another equine, another animal, a person, or an object;
- The potential of an equine activity participant to act in a negligent manner that may contribute to injury, death, or loss to the person of the participant or to other persons, including but not limited to, failing to maintain control over an equine or failing to act within the ability of the participant.
I AGREE to release the USPC, its successors or assigns, officials, officers, directors, employees, agents, personnel, volunteers and affiliated organizations from all claims including, but not limited to, claims for money or property, disability, covenants, actions, suits, causes or action, obligations, debts, costs, expenses, attorneys’ fees, judgments, orders and liabilities of whatsoever kind or nature in law, equity or otherwise, whether now known or unknown, suspected or unsuspected, and whether concealed or hidden, including but not limited to any state or federal statutory or common law claim or remedy of any kind whatsoever arising out of or in any way connected with any Harm to me or my horse and for any Harm caused by me or my horse to others, even if the Harm resulted, directly or indirectly, from the negligence of the USPC or the USPC activity.
I AGREE to expressly assume all risks of Harm to me or my horse, including Harm resulting from the negligence of the USPC or the USPC activity, and specifically agree to the applicable state statute/law regarding equine/farm animal activity liability and signed posting (if any), in any state in which I or my child participates in a USPC activity. A true copy for all state statutes in effect at the time of the execution of this agreement is attached hereto and incorporated herein by reference as if set out fully in the text of this document.
I AGREE to indemnify (that is, to pay any losses, damages, or costs incurred by) the USPC and the USPC activity and to hold them harmless with respect to claims for Harm to me or my horse, and for claims made by others for any Harm caused by me or my horse in the USPC activity.
I AGREE that neither I, nor any one claiming through me, will hereafter bring, commence, prosecute or maintain, or cause or permit to be brought, commenced, prosecuted or maintained, any suit or action, either at law or in equity, in any court in the United States or in any state thereof, or elsewhere, against the USPC, its successors or assigns, for, on account of, arising out of, or in any way connected with any Harm to me or my horse, and that neither I, nor any one claiming though me, will enforce, prosecute, or recover upon, or attempt to enforce, prosecute, or recover upon, any claim or right of action whatsoever, which I, or any one claiming though me, may now have or hereafter assert, in any way connected with claims for Harm to me or my horse, and for claims made by others for any Harms caused by me or my horse at the USPC activity.
I AGREE this Agreement is the entire agreement of the parties, and supersedes all prior oral and written understandings and agreements. This Agreement may be modified only by a written amendment signed by both parties.
I AGREE that if any provision of the Agreement is found to be invalid or illegal by a court of competent jurisdiction, the remaining provisions shall be construed as if the affected provision had not been included in order to effectuate the intent of the parties.
In the event this form is signed by the parent/guardian of a child, then all representations and acknowledgements herein are expressly made by, for, and on behalf of the parent/guardian and child.
By signing below, I AGREE to be bound by all applicable USPC rules and all terms and provisions of the USPC activity. I acknowledge that I enter into this release after having read the same, and place my signature hereto of my own free voluntary act and deed. By signing below, I represent to the USPC that I fully understand its contents, that I do not need any further explanation, and I waive any further explanation.
REQUIRED — all signatures must be originals, not photocopies.
_________________________ _________________________ and ____________________________ _________________________
Original Signature of Participant Date Original Signature of Parent(s) or Legal Guardian(s) Date
Unless participant is over the age of majority, i.e., dependent upon state law.
ALABAMA
WARNING: Under Alabama law, an equine activity sponsor or equine professional is not liable for an injury to or the death of a participant in equine activities resulting from the inherent risks of equine activities, pursuant to the Equine Activities Liability Protection Act.
ARIZONA
WARNING: Contestant is aware of the inherent risks associated with equine activities and is willing and able to accept full responsibilities for his or her own safety and welfare. Contestant releases the Event Sponsor from liability unless Event Sponsor is grossly negligent or commits wilful, wanton or intentional acts or omissions.
ARKANSAS
WARNING: Under Arkansas law, an equine activity sponsor is not liable for an injury to, or the death of, a participant in equine activities resulting from the inherent risk of equine activities.
COLORADO
WARNING: Under Colorado Law, an equine professional is not liable for an injury to or the death of a participant in equine activities resulting from the inherent risks of equine activities, pursuant to section 13-21-119, Colorado Revised Statutes.
DELAWARE
WARNING: Under Delaware Law, an equine professional is not liable for an injury to or the death of a participant in equine activities resulting from the inherent risks of equine activities, pursuant to 10 Delaware Code Section 8140.
FLORIDA
WARNING: Under Florida law, an equine sponsor or equine professional is not liable for an injury to, or the death of, a participant in equine activities resulting from the inherent risks of equine activities.
GEORGIA
WARNING: Under Georgia law, an equine activity sponsor or equine professional is not liable for an injury to or the death of a participant in equine activities resulting from the inherent risks of equine activities, pursuant to Chapter 12 of Title 4 of the Official Code of Georgia Annotated
ILLINOIS
WARNING: Under the Equine Activity Liability Act, each participant who engages in an equine activity expressly assumes the risks of engaging in and legal responsibility for injury, loss, or damage to person or property resulting from the risk of equine activities.
INDIANA
WARNING: Under Indiana law, an equine professional is not liable for an injury to, or the death of, a participant in equine activities resulting from the inherent risks of equine activities.
IOWA
WARNING: UNDER IOWA LAW, A DOMESTICATED ANIMAL PROFESSIONAL IS NOT LIABLE FOR DAMAGES SUFFERED BY, AN INJURY TO, OR THE DEATH OF A PARTICIPANT RESULTING FROM THE INHERENT RISKS OF DOMESTICATED ANIMAL ACTIVITIES, PURSUANT TO IOWA CODE CHAPTER 673. YOU ARE ASSUMING INHERENT RISKS OF PARTICIPATING IN THIS DOMESTICATED ANIMAL ACTIVITY
Iowa Statement of Inherent Risks
A number of inherent risks are associated with a domesticated animal activity. A domesticated animal may behave in a manner that results in damages to property or an injury or death to a person. Risks associated with the activity may include injuries caused by bucking, biting, stumbling, rearing, trampling, scratching, pecking, falling, or butting.
The domesticated animal may act unpredictably to conditions, including, but not limited to, a sudden movement, loud noise, an unfamiliar environment, or the introduction of unfamiliar persons, animals, or objects.
The domesticated animal may also react in a dangerous manner when a condition or treatment is considered hazardous to the welfare of the animal; a collision occurs with an object or animal; or a participant fails to exercise reasonable care, take adequate precautions, or use adequate control when engaging in a domesticated animal activity, including failing to maintain reasonable control of the animal or failing to act in a manner consistent with the person’s abilities.
KANSAS
WARNING: Under Kansas law, there is no liability for an injury to or the death of a participant in domestic animal activities resulting from the inherent risks of domestic animal activities, pursuant to sections 1 through 4. You are assuming the risk of participating in this domestic animal activity.
Kansas Statement of Inherent Risks:
Inherent risks of domestic animal activities include, but shall not be limited to:
(1) The propensity of a domestic animal to behave in ways i.e., running, bucking, biting, kicking, shying, stumbling, rearing, falling or stepping on, that may result in an injury, harm or death to persons on or around them;
(2) the unpredictability of a domestic animal’s reaction to such things as sounds, sudden movement and unfamiliar objects, persons, or other animals;
(3) certain hazards such as surface and subsurface conditions;
(4) collisions with other domestic animals or objects; and
(5) the potential of a participant to act in a negligent manner that may contribute to injury to the participant or others, such as failing to maintain control over the domestic animal or not acting within such participant’s ability.
KENTUCKY
WARNING: Under Kentucky law, a farm animal activity sponsor, farm animal professional, or other person does not have the duty to eliminate all risks of injury of participation in farm animal activities. There are inherent risks of injury that you voluntarily accept if you participate in farm animal activities.
LOUISIANA
WARNING: Under Louisiana law, an equine sponsor or equine professional is not liable for an injury to or the death of a participant in equine activities resulting from the inherent risks of equine activities, pursuant to R.S. 9:2795.3.
MAINE
WARNING: UNDER MAINE LAW, AN EQUINE PROFESSIONAL HAS LIMITED LIABILITY FOR AN INJURY OR DEATH RESULTING FROM THE INHERENT RISKS OF EQUINE ACTIVITIES.
MASSACHUSETTS
WARNING: Under Massachusetts law, an equine professional is not liable for an injury to, or the death of, a participant in equine activities resulting from the inherent risks of equine activities, pursuant to section 2D of chapter 128 of the General Laws.
MICHIGAN
WARNING: Under the Michigan equine activity liability act, an equine professional is not liable for an injury to or the death of a participant in an equine activity resulting from an inherent risk of the equine activity.
MISSOURI
WARNING: Under Missouri law, an equine professional is not liable for an injury to or the death of a participant in equine activities resulting from the inherent risks of equine activities pursuant to the Revised Statutes of Missouri.
MISSISSIPPI
WARNING: Under Mississippi law, an equine activity or equine sponsor is not liable for an injury to or the death of a participant in equine activities resulting from the inherent risks of equine activities, pursuant to this chapter.
NEBRASKA
WARNING: Under Nebraska law, an equine professional is not liable for an injury to or the death of a participant in equine activities resulting from the inherent risks of equine activities, pursuant to this act [Sections 25-21, 249 to 25-21, 253].
NORTH CAROLINA
WARNING: Under North Carolina law, an equine activity sponsor or equine professional is not liable for an injury to or the death of a participant in equine activities resulting exclusively from the inherent risks of equine activities. Chapter 99E of the North Carolina General Statutes.
OHIO
Ohio Statement of Inherent Risks:
"INHERENT RISK OF AN EQUINE ACTIVITY" MEANS A DANGER OR CONDITION THAT IS AN INTEGRAL PART OF AN EQUINE ACTIVITY, INCLUDING, BUT NOT LIMITED TO, ANY OF THE FOLLOWING:
(a) THE PROPENSITY OF AN EQUINE TO BEHAVE IN WAYS THAT MAY RESULT IN INJURY, DEATH, OR LOSS TO PERSONS ON OR AROUND THE EQUINE;
(b) THE UNPREDICTABILITY OF AN EQUINE’S REACTION TO SOUNDS, SUDDEN MOVEMENT, UNFAMILIAR OBJECTS, PERSONS, OR OTHER ANIMALS;
(c) HAZARDS, INCLUDING, BUT NOT LIMITED TO, SURFACE OR SUBSURFACE CONDITIONS;
(d) A COLLISION WITH ANOTHER EQUINE, ANOTHER ANIMAL, A PERSON, OR AN OBJECT;
(e) THE POTENTIAL OF AN EQUINE ACTIVITY PARTICIPANT TO ACT IN A NEGLIGENT MANNER THAT MAY CONTRIBUTE TO INJURY, DEATH, OR LOSS TO THE PERSON OF THE PARTICIPANT OR TO OTHER PERSONS, INCLUDING BUT NOT LIMITED TO, FAILING TO MAINTAIN CONTROL OVER AN EQUINE OR FAILING TO ACT WITHIN THE ABILITY OF THE PARTICIPANT.
OREGON
In accordance with Oregon Revised Statutes 30.691, Contestant (or persons or entities affiliated with Contestant), as a condition of participation in The Event, further waives the right to bring an action against the Event Sponsor (as defined in the Assumption of Risk, Waiver, or Release of Liability) for any injury or death arising out of riding, training, grooming or riding as a passenger upon the equine.
PENNSYLVANIA
Warning: Under Pennsylvania law an equine professional and equine activity sponsor is not liable for an injury to or death of a participant in equine activities resulting from the inherent risks of equine activities.
RHODE ISLAND
WARNING: Under Rhode Island Law, an equine professional, unless he or she can be shown to have failed to be in the exercise of due care, is not liable for an injury to, or the death of, a participant in equine activities resulting from the inherent risks of equine activities, pursuant to this chapter.
SOUTH CAROLINA
WARNING: Under South Carolina law, an equine activity sponsor or equine professional is not liable for an injury to or the death of a participant in an equine activity resulting from an inherent risk of equine activity, pursuant to Article 7, Chapter 9 of Title 47, Code of Laws of South Carolina, 1976.
SOUTH DAKOTA
WARNING: Under South Dakota law, an equine professional is not liable for an injury to or the death of a participant in equine activities resulting from the inherent risks of equine activities, pursuant to Section 42-11-2.
TENNESSEE
WARNING: Under Tennessee Law, an equine professional is not liable for an injury to or the death of a participant in equine activities from the inherent risks of equine activities, pursuant to Tennessee Code Annotated, title 44, chapter 20.
TEXAS
WARNING: UNDER TEXAS LAW (CHAPTER 87, CIVIL PRACTICE AND REMEDIES CODE), AN EQUINE PROFESSIONAL IS NOT LIABLE FOR AN INJURY TO OR THE DEATH OF A PARTICIPANT IN EQUINE ACTIVITIES RESULTING FROM THE INHERENT RISKS OF EQUINE ACTIVITIES.
UTAH
Pursuant to the Utah Equine Activity Liability Act, § 78-27b-101, et. seq., it shall be presumed that participants in equine or livestock activities are aware of and understand that there are inherent risks associated with these activities. An equine activity sponsor, equine professional, livestock activity sponsor, or livestock professional is not liable for an injury to or the death of a participant due to the inherent risks associated with these activities. "Inherent risk" with regard to equine or livestock activities means those dangers or conditions which are an integral part of equine or livestock activities, which may include:
(a) The propensity of the animal to behave in ways that may result in injury, harm, or death to persons on or around them;
(b) the unpredictability of the animal’s reaction to outside stimulation such as sounds, sudden movement, and unfamiliar objects, persons, or other animals;
(c) collisions with other animals or objects;
(d) the potential of a participant to act in a negligent manner that may contribute to injury to the participant or others, such as failing to maintain control over the animal or not acting within his or her ability.
VERMONT
WARNING: Under Vermont Law, an equine activity sponsor is not liable for an injury to, or the death of, a participant in equine activities resulting from the inherent risks of equine activities that are obvious and necessary, pursuant to 12 V.S.A § 1039.
WEST VIRGINIA
WEST VIRGINIA EQUESTRIAN ACTIVITIES RESPONSIBILITY ACT STATEMENT FOR HORSEMEN:
I ACKNOWLEDGE AND AGREE THAT I WILL PERFORM THE DUTIES REQUIRED OF ALL HORSEMAN UNDER THE WEST VIRGINIA EQUESTRIAN ACTIVITIES RESPONSIBILITY ACT, W.VA. CODE § 20-4-1, ET SEQ., WHICH INCLUDE THE FOLLOWING:
(1) Make reasonable and prudent efforts to determine the ability of a participant to safely engage in the equestrian activity, to determine the ability of the horse to behave safely with the participant, and to determine the ability of the participant to safely manage, care for and control the particular horse involved;
(2) Make known to any participant any dangerous traits or characteristics or any physical impairments or conditions related to a particular horse which is involved in the equestrian activity of which the horseman knows or through the exercise of due diligence could know;
(3) Make known to any participant any dangerous condition as to land or facilities under the lawful possession and control of the horseman of which the horseman knows or through the exercise of due diligence could know, by advising the participant in writing or by conspicuously posting warning signs upon the premises;
(4) In providing equipment or tack to a participant, make reasonable and prudent efforts to inspect such equipment or tack to assure that it is in proper working condition and safe for use in the equestrian activity;
(5) Prepare and present to each participant or prospective participant, for his or her inspection and signature, a statement which clearly and concisely explains the liability limitations, restrictions and responsibilities set forth under the West Virginia Equestrian Activities Responsibility Act.
WISCONSIN
NOTICE: A person who is engaged for compensation in the rental of equine or equine equipment or tack or in the instruction of a person in the riding or driving of an equine or in being a passenger upon an equine is not liable for the injury or death of a person involved in equine activities resulting from the inherent risks of equine activities, as defined in section 895.481(1)(c) of the Wisconsin Statutes. |
From
Chief Town Planner, Haryana-cum-Chairman,
Building Plan Approval Committee,
O/o Director, Town & Country Planning Department,
Haryana, Nagar Yojna Bhavan,
Madhya Marg, Sector 18, Chandigarh.
Tele-Fax: 0172-2548475; Tel.: 0172-2549851,
E-mail: firstname.lastname@example.org
Website www.tcpharyana.gov.in.
To
Pegasus Land & Housing Pvt. Ltd.,
Building No. 80, 15th Floor, Sector-44,
Gurugram-122003.
Memo No. ZP-1429/SD(DK)/2021/611 Dated: 12-01-2021
Subject: Approval of building plans of Affordable Group Housing Colony area measuring 5.03403 acres (Licence No. 19 of 2020 dated 01.08.2020), Sector-93, Gurugram being developed by Pegasus Land & Housing Pvt. Ltd.
Reference your application dated 08.10.2020 permission to re-erect the buildings in Affordable Group Housing Colony area measuring 5.03403 acres (Licence No. 19 of 2020 dated 01.08.2020), Sector-93, Gurugram Manesar Urban Complex in accordance with the plans submitted with it after receipt of ₹ 15,64,567/- towards Infrastructure Development Charges for 12% additional FAR being considered as an incentive for providing green buildings as per IGBC norms in view of provision of Code 6.5 of Haryana Building Code, 2017.
Permission is hereby granted for the aforesaid construction subject to the provisions of the Punjab Scheduled Roads & Controlled Areas Restriction of Unregulated Development Act, 1963 and Haryana Building Code-2017 subject to the following amendments, terms and conditions:
1. The plans are valid for a period of 2 years of the buildings less than 15.00 meters in height and 5 years for the multistoried buildings from the date of issuance of sanction, subject to validity of licenses granted for this scheme. However, you shall complete the project as per stipulation in Affordable Group Housing Policy issued vide Notification No. PF-27/48921 dated 19.08.2013.
2. The structural responsibility of the construction shall be entirely of the owner/supervising architect/Engineer of the scheme.
Further that:
a) The building shall be constructed in accordance to the Structure Design by Structure Engineer and certified by Proof Consultant on prescribed FORM BR-V (A2).
b) All material to be used for erection of building shall conform to I.S.I. and N.B.C. standards.
c) No walls/ceiling shall be constructed of easily inflammable material and staircases shall be built of the fire resisting material as per standard specification.
d) The roof slab of the basement external to the buildings if any shall be designed/constructed to take the load of fire tender up to 45 tones.
3. FIRE SAFETY:
(i) The colonizer and the Supervising Architect of the project shall be entirely responsible for making provisions of fire safety and fire-fighting measures and shall abide by all fire safety bye laws.
(ii) That you shall get approved the fire-fighting scheme in accordance with the section 15 of The Haryana Fire Safety Act 2009 and directions issued by the Director, Haryana Fire Services, Haryana, before starting the construction work at site.
4. No addition and alteration in the building plans/layout plan shall be made without the prior approval of DTCP. Further only figured dimensions shall be followed and in case of any variation in the plans, prior approval of DTCP shall be pre-requisite.
5. That you shall furnish the service plan/estimate of this scheme in accordance with approved building plans.
6. Based on the actual estimated cost of internal development of the commercial colony you shall furnish additional bank guarantee, if required.
7. The revenue Rasta if any passing through the site shall be kept unobstructed.
8. If any infringement of byelaws remains unnoticed, the Department reserves the right to amend the plan as and when any such infringement comes to its notice after giving an opportunity of being heard and the Department shall stand indemnified against any claim on this account.
9. The layout showing the electric installation shall have to be got approved from the competent authority before execution of work at site.
10. No person shall occupy or allow any other person to occupy any new building and before grant of occupation certificate, you shall apply for occupation certificate as per the provisions of Code 4.10 of the Haryana Building Code-2017 which shall be accompanied by certificates regarding completion of works described in the plans and it shall be accompanied by:
(i) Structural stability certificate duly signed by the recognized Architect & Structural Engineer.
(ii) A clearance from Fire Safety point of view from the competent authority.
11. The provision of letter boxes for each dwelling unit shall be made at the ground floor of each building.
12. The basements shall be used for parking and services as prescribed in the approved zoning plan and building plans. The parking lots proposed in the scheme shall be exclusively for the use of flat owners/residents of the group housing scheme. The parking lot shall not be leased out/transferred to any person who is not a flat owners/residents of the group housing complex.
13. That you shall comply with the conditions laid down in the Memo No. 189741 dated 01.12.2020 of SuperIntending Engineer (HQ), HSVP, Panchkula (Copy enclosed).
14. That you shall comply with the conditions laid down in the Memo No. 523 dated 03.12.2020 of Director General, Urban Local Bodies-cum-Fire Services, Haryana, Panchkula (Copy enclosed).
15. GENERAL:
(i) That you shall obtain the clearance/NOC as per the provisions of the Notification No. S.O. 1533 (E) Dated 14.9.2006 issued by Ministry of Environment and Forest, Government of India before starting the construction/execution of development works at site.
(ii) That you shall strictly comply with the directions of MoEF Guidelines, 2010 while raising construction and comply with the instructions of Director, Town and Country Planning, Haryana, Chandigarh issued vide orders dated 14.5.2015 which is also available on the departmental website www.tcp.haryana.gov.in.
(iii) That you shall submit the fire-fighting scheme duly approved in accordance with the section 15 of the Haryana Fire Safety Act 2009 and directions issued by Director, Urban Local Bodies Haryana before starting the construction work at site.
(iv) That you shall submit the approved Electrical Service Plan from competent authority and certificate to this effect that adequate arrangement has been put in place before grant of occupation certificate.
(v) That the rain water harvesting system shall be provided as per Central Ground Water Authority norms/Haryana Govt. notification as applicable.
(vi) That you shall use only Light-Emitting Diode lamps (LED) fitting for internal lighting as well as Campus lighting.
(vii) That you shall not construct the building beyond 30.0 mtrs. without getting the valid NOC from AAI.
(viii) That you shall ensure the installation of Solar Photovoltaic Power Plant as per the provisions of order No. 22/52/2005-5Power dated 21.03.2016 issued by Haryana Government Renewable Energy Department.
(ix) That you shall strictly comply with the directions issued vide Notification No. 19/6/2016-5P dated 31.03.2016 issued by Haryana Government Renewable Energy Department.
(x) That if any, site for Electric Sub Station is required, same will be provided by you in the colony.
(xi) That provision of parking shall be made within the area earmarked/designated for parking in the colony and no vehicle shall be allowed to park outside the premises.
(xii) That you shall follow provisions of section 46 of 'The Persons with Disabilities (Equal Opportunities, protection of Rights and full Participation) Act, 1995' which includes construction of Ramps in public buildings, adaption of toilets for wheel chair users, Braille symbols and auditory signals in elevators or lifts and other relevant measures for Hospitals, Primary Health Centre and other medical care and rehabilitation units.
(xiii) That you shall abide the terms and conditions of the Undertaking/Affidavit submitted in the office of Administrator, HSVP, Gurugram in compliance of Order dated 16.07.2012 of the Hon’ble High Court and shall not extract groundwater for construction purposes.
(xiv) That you shall abide by the terms and conditions of Affordable Group Housing Scheme issued vide Notification No. PF-27/48921 dated 19.08.2013.
(xv) That you shall give the advertisement as per the instructions issued vide memo No. 15772 dated 18.07.2014.
(xvi) That you shall deposit the balance amount of Labour Cess in future, time to time as per progress in construction at site.
(xvii) That you shall submit the scanned copy of the approved building plans in CD format within one week to this office from the issuance of this letter.
16. Environment: That you shall strictly comply with the directions of MOEF Guidelines, 2010 while raising construction. In addition, you shall comply with the instructions of Director, Town & Country Planning, Haryana, Chandigarh issued vide order dated 14.05.2015, available on the Departmental Website www.tcpharyana.gov.in at URL :https://tcpharyana.gov.in/Policy/Misc392%20OA%20No.%2021%20of%202014%20Vardhaman%20Kaushik%20Vs.%20UOI_ors.pdf in compliance of the orders dated 10.04.2015 passed by Hon’ble National Green Tribunal in OA No. 21 of 2014, which are as under:
(i) You shall put tarpaulin on scaffolding around the area of construction and the building. You are also directed that you shall not store any construction material particularly sand on any part of the street/roads.
(ii) The construction material of any kind that is stored in the site will be fully covered in all respects so that it does not disperse in the Air in any form.
(iii) All the construction material and debris shall be carried in the trucks or other vehicles which are fully covered and protected so as to ensure that the construction debris or the construction material does not get dispersed into the air or atmosphere, in any form whatsoever.
(iv) The dust emissions from the construction site should be completely controlled and all precautions taken in that behalf.
(v) The vehicles carrying construction material and construction debris of any kind should be cleaned before it is permitted to ply on the road after unloading of such material.
(vi) Every worker working on the construction site and involved in loading, unloading and carriage of construction material and construction debris shall be provided with mask to prevent inhalation of dust particles.
(vii) Every owner and or builder shall be under obligation to provide all medical help, investigation and treatment to the workers involved in the construction of building and carry of construction material and debris relatable to dust emission.
(viii) It shall be the responsibility of every owner/builder to transport construction material and debris waste to construction site, dumping site or any other place in accordance with rules and in terms of Hon’ble NGT order dated 10.04.2015 referred above.
(ix) All to take appropriate measures and to ensure that the terms and conditions of the Hon’ble NGT order dated 10.04.2015 referred above in OA No. 21 of
2014 and the earlier orders passed in said case should strictly comply with by fixing sprinklers, creations of green air barriers.
(x) Compulsory use of wet jet in grinding and stone cutting.
(xi) Wind breaking walls around construction site.
(xii) That you shall ensure that least dust has emitted into air/atmosphere and all steps are taken to prevent the same.
(xiii) That all the builders, who are building commercial, residential complexes which are covered under the EIA Notification of 2006, shall provide green belt around the building that they construct and compliance of the same shall be ensured prior to issuance of occupancy certificate. The width of green belt will be kept 1.5 meter along boundary wall within site, along periphery.
(xiv) If any person, owner and/or builder is found to be violating any of the conditions stated in this order and/or for their non-compliance such person, owner, builder shall be liable to pay compensation of ₹ 50,000/- per default in relation to construction activity at its site and ₹ 5,000/- for each violation during carriage and transportation of construction material, debris through trucks or other vehicles, in terms of Section 15 of the NGT Act on the principle of Polluter Pay. Such action would be in addition not in derogation to the other action that the Authority made take against such builder, owner, person and transporter under the laws in force.
(xv) All the owners/builders shall ensure that C&D waste is transported in terms of this order to the site in question only and due record in that behalf shall be maintained by the builders, transporters and NCR of Delhi.
(xvi) It is made clear that even if constructions have been started after seeking Environmental Clearance under the EIA notification 2006 and after taking other travel but is being carried out without taking the preventive and protective environmental steps as stated in above said order dated 10.04.2015 passed by NGT and MOEF guidelines, 2010, the State Government, SPCB and any officer of any Department as afore-stated shall be entitled to direct stoppage of work.
17. As per the condition of the licence, you were required to convey ultimate power load requirement to the Department. You have not intimated so far. The same may be supplied within a month from the issuance of this letter.
18. That you shall complete the project as per terms & conditions of Affordable Group Housing policy in a time bound manner from the date of earlier approved building plans/environmental clearance.
19. The zero period for interest over payment of fee and charges shall be subject to the condition that:
a) You shall pass on the corresponding benefits to the allottees to ensure that the hardships faced by the end-users also get mitigated to that extent.
b) You shall endeavor to complete the projects either within the original timeline as agreed upon in the sale-purchase agreement or as revised with prior consent of RERA.
20. On the basis of IGBC “Gold Rating” Certificate duly submitted by the colonizer, and as per provision of Code 6.5 of Haryana Building Code, 2017, the additional FAR of 12% is allowed. Since, the final rating will be issued by IGBC after completion of total building complex/project, therefore, final occupation for three times the area of additional FAR (which is sought to be availed, as incentive for green building), shall be withheld till the final rating from IGBC is obtained. However, if the colonizer fails to achieve the final rating, which is lesser than the provisional rating, the occupation certificate of all building complex shall be issued after compounding the additional FAR (i.e. difference of additional FAR from provisional rating & final rating) at the ten times of the rates of EDC applicable at the time of submission of occupation certificate.
This sanction will be void ab initio, if any of the conditions mentioned above are not complied with.
DA/As above & One set of Building Plans
(Hitesh Sharma)
Senior Town Planner (M)HQ
Member Secretary
For: Chief Town Planner, Haryana-cum-Chairman,
Building Plan Approval Committee,
Town & Country Planning Department,
Haryana, Chandigarh.
Endst. No. ZP-1429/SD(DK)/2021/______________ Dated: - ____________
A copy is forwarded to the following for information:-
1. Haryana State Pollution Control Board, Panchkula with the request that the compliance of the instructions issued by NGT shall be monitored and strict compliance to be ensured.
2. MD, HVPNLI, Planning Directorate, Shakti Bhawan, Sector-6, Panchkula with request to assess the power utility site requirement as per ultimate power load requirement.
3. Administrator, HSVP, Gurugram.
4. Senior Town Planner, Gurugram.
5. Superintending Engineer (HQ), HSVP, Panchkula.
6. District Town Planner, Gurugram along with one set of approved building plans.
7. District Town Planner (E), Gurugram.
8. Nodal Officer, website updation.
9. Assistant Divisional Fire Officer O/o Director, Urban Local Bodies, Haryana, Panchkula.
Encl: as above
(Hitesh Sharma)
Senior Town Planner (M)HQ
Member Secretary
For: Chief Town Planner, Haryana-cum-Chairman,
Building Plan Approval Committee,
Town & Country Planning Department,
Haryana, Chandigarh. |
The REGULAR meeting of the Board of Trustees of Columbus School District #6 was held on Tuesday, August 15, 2023, in the middle school commons area. Chair Morse called the meeting to order at 6:00 p.m. A quorum was established, and the meeting began with the Pledge of Allegiance to the Flag.
Those participating were Trustee Joe Morse, Trustee Ryan Meier, Trustee Sue Ivankovich, Trustee Stephanie Wright and Trustee Jay Baum. Superintendent Jeff Bermes, Middle School Principal Joseph Morse, Elementary Principal Tom Kurtz, High School Principal Jim Browning, BCEA Union President Vickie Gibson, Laurie Martinez and Lisa Westervelt were also in attendance.
A. Approve Board Meeting Minutes
B. Approve Warrants, Extracurricular Balances & Expenditures.
Trustee Baum made a motion to accept the agenda Trustee Ivankovich seconded. The motion was voted on and passed unanimously.
Trustee Baum made a motion to accept the consent agenda with addition to resignations of Christina Nelson and Cecelia Peterson. Trustee Meier seconded. The motion was voted on and passed unanimously.
The sub list will be available at the September meeting.
At 6:09 p.m. the board and those present toured the apartment complex purchased by the District.
PRESENTATIONS/PUBLIC COMMENT REGARDING AGENDA ITEMS
Laurie Martinez spoke on behalf of her concerns of the dress code policy in the High School Handbook.
OLD BUSINESS
A. Construction Update:
- Carpet is currently being installed in both the High School and Elementary buildings.
- The HVAC systems are almost completely installed.
- The track has been painted with exception of the inside portion because the track is not a 400 meter track. One spot that was painted on the track has been turned into warranty.
- Grandstand wall will be completed before the first home game.
- ESSER money has been spent out. Remainder of building projects will come out of the Elementary Bond and Metal Mines.
Elementary building underground sprinklers are almost completed.
Bids are coming in for the roofing on our new teacher housing.
B. **Coaches and Activity Handbook Adoptions for 2023-24**
It was discussed to change the meal allowances for our teams to the following: When the District provides meals, $12 Breakfast, $20 Lunch, $25 Dinner. $50 total (per diem per day at Tournaments that are overnight when breakfast is provided by the motel.) $30 (per diem per day at State Tournaments with no lodging.) $25 (per diem per day at District and Divisional Tournaments with no lodging.) Gratuities: when appropriate (sit down meals), the coach/sponsor will include, but not exceed 15% of the total meal bill. These amounts are MAXIMUM including beverage and/or dessert.
Trustee Wright made a motion to accept the 23-24 Coaches and Sponsors Activity Handbook with the creation of the meals. Trustee Meier seconded. The motion was voted on and passed unanimously.
C. **Review of the Property Management Contract**
All rental funds will be placed in Fund 20 up to $10,000 for each district with the remainder being placed in Fund 182. Trustee Baum made the motion to approve the 23-24 property management contract with Park Realty to manage the district owned rental property at 127 N 4th. Trustee Ivankovich seconded. The motion was voted on and passed unanimously.
**NEW BUSINESS**
A. **Handbook Adoptions for 2023-2024**
Elementary Handbook changes:
- Breakfast will be served before school from 7:55-8:15 a.m. in the cafeteria.
- Parents will be notified at 6, 12, and 18 absences. These are absences that are not counting medical absences. Those medical absences must have a doctor’s note to be excused.
- If your child is too sick to go outside for recess, they are probably too sick to be at school.
- As a general rule, gum is not permitted due to it getting on the carpets, under desks, and so on. But, some teachers do allow gum in the classroom. Please check with the classroom teacher whether or not gum is allowed in the classroom before sending gum with your child(ren).
Middle School Handbook changes:
- 8th Grade Recognition will be scheduled for the last Tuesday of the school year, at 5:00 p.m.
High School Handbook changes:
- On page 1, the start and end time of the school day has been updated. This year, the day will go from 8:10 a.m. – 3:35 p.m. with 37 minutes for lunch.
- On page 8, nicotine and drugs were added along with tobacco as items students are not allowed to have in their possession at any time in the school, on school grounds, or at activities where the student is competing or performing as a student.
- On page 10, we could add another item about marijuana specifically if that is something the board feels is necessary. Superintendent Bermes will reach out to Mr. Weldon, District Attorney on proper procedure to follow in regards to how to handle marijuana cases.
- It was brought to the boards attention that there are many concerns with the high school dress code policy and what is considered appropriate.
Trustee Baum motioned to approve the elementary, middle school and high school handbooks as presented, with proposed changes to come to the high school handbook. Trustee Ivankovich seconded it. The motion was voted on and passed unanimously.
B. **Transportation Handbooks 2023-2024**
There are no changes to the transportation handbook. Trustee Ivankovich made a motion to adopt the 2023-2024 Transportation Handbook. Trustee Baum seconded. The motion was voted on and passed unanimously.
C. **Adopt Elem & HS 2022-2023 Trustees Financial Summaries**
Superintendent Bermes and Business Manager Kari Econom reviewed the 22-23 Trustees Financial Summary with the board. Trustee Ivankovich made a motion to approve the 22-23 Trustee Financial Summary. Trustee Meier seconded. The motion was voted on and passed unanimously.
D. **Adopt Elem. & HS 2023-2024 Final Budgets**
Superintendent Bermes discussed the District’s final budgets for the 23-24 school year. Changes were made on the mills based off of the difference in the taxable values from 22-23 and 23-24. Trustee Baum made a motion to adopt the 23-24 budget. Trustee Ivankovich seconded. The motion was voted on and passed unanimously.
E. **Hiring of Extra-Curricular Positions and New District Hires**
Trustee Ivankovich made a motion to accept the following spring extra-curricular sports: Golf – Head Jeromey Burke, Asst. Bob Fitzgerald; HS Track – Head Jen Crago, Asst. Mark Albrecht; Asst. Thea Karamazadeh; MS Track – Jaime Holbert and Michael Robinson; Softball – Head Rachelle Gravance, Asst TBD; Baseball – Head Travis West, Asst. Cactus Anderson. Other extra-curricular – Paul Barta – 6th grade boys basketball. Trustee Meier seconded. The motion was voted on and passed unanimously.
F. **Joliet District’s Request to Enter District to Turn Around**
The Joliet School District is requesting permission to enter our district on Highway 421 to turn around at the designated turn around at the county line again this year. They will not be picking up any students within our district boundaries and are requesting this entrance into our district for safety purposes to turn around safely. Trustee Baum moved to allow the bus to turn around at the county line. Trustee Wright seconded. The motion was voted on and passed unanimously.
G. **Out-of-District Application**
It was discussed to reopen the out of district application reconsideration starting August 15, accepting applications through August 18th at 2 p.m. Trustee Wright made the motion to reopen the out of district application from tonight August 15th through 2 p.m. on August 18th to be reconsidered for enrollment. If the student meets the district requirements. Trustee Meier seconded. The motion was voted on and passed unanimously.
Trustee Baum made a motion to approve the out-of-district applications of the Smith, Brito and Snap Families. Trustee Ivankovich seconded it. The motion was voted on and passed unanimously.
**Administrative Reports**
A. **Elementary Report:**
- Current enrollment: 323 (Last years at this time was 327) 14 of the 323 are new registrations.
- New teacher orientation will be Friday August 18th.
- Interviewing for two PARA positions.
- Open house will be Monday, August 21 from 4:00-6:00 PM.
- Wednesday, August 23 is the first day of school.
- There will be two third grade classes and three fourth grade classes this year. Due to a smaller class size in the third grade and a larger class size in the fourth grade.
- Carpet is about finished in the K/1 classrooms.
- A fence was installed between the “big toy” and the faculty parking lot.
- Math workbooks will be a new edition called Eureka Squared.
- The Big Jay Show based out of Billings will be buying a new book every month for grades K-5. The community of Columbus stepped up and helped the Big Jay Show donate for this cause.
B. **Middle School Report**
• Current Enrollment: 182
• 6th grade and new student orientation night will be August 21 at 4 p.m.
• First day of school is August 23, start time 8:05 a.m.
• Yearbook Changes: Cadie Frickel-Secretary, TBD-Librarian, Ramon Ochoa-Band/Choir 6th, 7th and 8th. Honor Society-Beth Ann Steffan.
• Bell Schedule time changes.
• 8th grade recognition will be scheduled for the last Tuesday of the school year, at 5:00 p.m.
C. High School Report
• Current Enrollment: 250
• Senior class will have two foreign exchange students who will be leaving at the end of the first semester. In January, we will be getting a transfer student from Arkansas who is completely blind. We will need to prepare our teachers for this student starting in October.
• Handbook changes.
D. Superintendent Report
• Turf Tank is up and running. It took 36 minutes to complete the practice field.
• Beartooth Stock Growers would like to fundraise to give beef to school district.
• Any person interested in the items up for bid needs to email the superintendent.
• Current positions open with the district are: kitchen help, paraprofessionals and Librarian. Will possibly put out an ad for Librarian Aide.
• The partitions in the high school boy’s bathroom were delivered warped and will not connect. New ones were ordered and will be here before school starts.
• New employees will be required to come introduce themselves to the board at the September board meeting.
COMMITTEE REPORTS
Transportation:
Personnel Committee:
Buildings:
Finance: They met to approve and sign warrants.
Technology:
Policy:
ADJOURNMENT:
Trustee Baum made a motion to adjourn the meeting at 8:47 pm, Trustee Wright seconded it. The motion was voted on and passed unanimously.
Attest:
Karen Econom
Chairman |
Abstract
The ISO 17025 standard ("General Requirements for the Competence of Testing and Calibration Laboratories") requires that measurement uncertainty be taken into account when statements of compliance are made. In other words, when a calibration laboratory calibrates an instrument and produces a calibration certificate indicating that the calibration verification procedure "passed" or "failed", it is important that, for each test point, the measurement uncertainty be first calculated and then used in the determination of the test result. Guardbanding is a primary technique for assuring compliance with this 17025 requirement. Considerations of efficiency and productivity in calibration laboratories require full automation whenever possible. It is therefore desirable for automated calibration software to include guardbanding capability. This paper describes a flexible, configurable implementation of guardbanding in an automated calibration software system. A number of techniques are discussed, including both table-based and formula-based methods. Facilities for customization of the guardbanding algorithm are presented. A description of generated result data is included.
1. Introduction
The fundamental concept of guardbanding is to restrict the Pass/Fail limits applied to a calibration test based on a defined criterion.
The purpose of guardbanding is to control the risk of accepting an out-of-tolerance unit, or rejecting an in-tolerance unit.
Guardbanding affects the result of a calibration test. Without guardbanding the result of a test will be Pass or Fail. With guardbanding the result of a test will be Pass, Fail, or Indeterminate. Furthermore, a Pass or Fail test result without guardbanding may change to a result of Indeterminate with guardbanding.
Example
Suppose a calibration test with the following parameters is performed:
- Cardinal Point = 10 V DC
- Test Specification = ± 1%
Specification Limits = 9.9 V to 10.1 V
Measurement Uncertainty = 0.03 V
Measured Value (UUT Reading) = 9.92 V
Notice that in this example the measured value is within the specification limits. If a Pass/Fail determination were made without considering the measurement uncertainty, the result of the test would be Pass. However, when the measurement uncertainty is considered the true value becomes $9.92 \pm 0.03$ V. The true value may be as low as 9.89 V, which is below the lower specification limit.
When the measured value is close to a specification limit, the result is said to be Indeterminate. The precise definition of “close to” depends on the chosen guardbanding strategy.
2. Motivation
The primary motivation for guardbanding is derived from the standards document ISO 17025. In Section 184.108.40.206 the standard states:
*When statements of compliance are made, the uncertainty of the measurement shall be taken into account.*
A plausible interpretation of the 17025 statement is that when a calibration report is produced, and the report makes a specific statement as to the Pass/Fail result of the calibration, either for the overall calibration or for individual tests, the Pass/Fail determination must be made in a manner which considers the measurement uncertainty.
If a calibration report is produced which simply itemizes a series of tests, reporting for each test the reference value, the UUT value, the nominal test specification, and the measurement uncertainty, but does not report a Pass/Fail result, then it’s reasonable to interpret 17025 as not requiring that the measurement uncertainty be *taken into account* because, in this case, no statement of compliance is being made.
3. Strategies
Guardbanding strategies range from very conservative to no guardbanding.
The most conservative strategy is to tighten the specification limits by the full amount of the measurement uncertainty.
The least conservative method is to do nothing. The specification limits are used *as is*. This approach is traditional, but takes no account of the measurement uncertainty. MIL STD 45662, which requires a 4:1 Test Specification Ratio (TSR), uses this approach. This approach is relatively easy to implement, and the risk of passing an out-of-tolerance
UUT is usually considered to be acceptable if 4:1 TSRs can be maintained. However, in many areas of calibration it is difficult to maintain 4:1 TSRs.
Strategies which fall between the most and least conservative approaches involve dynamically adjusting the test limits based on various criteria.
Strategies may be broadly classified as follows:
§ Strategies which adjust the test limits as a function of the measurement uncertainty.
§ Strategies which adjust the test limits as a function of the test uncertainty ratio (TUR).
§ Strategies which adjust the test limits as a function of the test specification ratio (TSR).
§ Strategies that adjust the test limits based on an algorithm which directly attempts to maintain constant risk of accepting an out-of-tolerance UUT.
§ Strategies that statistically analyze historical calibration data for the UUT and/or references used in the calibration process.
§ Strategies that statistically analyze population data for references and UUTs.
Strategies that analyze historical calibration data require access to a database containing the required data. Such access may be impractical in some situations. For example, a 3rd-party calibration laboratory may not have the required historical calibration data. Similarly, strategies that require access to data for a population of instruments are typically impractical except for the instrument manufacturer. Although strategies based either on historical calibration data or instrument population data may have merit, they are beyond the scope of this paper.
4. Automation
In addition to categorizing guardbanding according to the particular strategy or algorithm used, it is also possible to categorize guardbanding based on the level of automation.
4.1 Manual Guardbanding
When guardbanding is performed manually, software to automate the calibration is not used. Guardbanded test limits must be calculated prior to the calibration. Software applications like Excel may be used to assist in the calculation of the guardbanded limits. When the calibration is performed it is the responsibility of the operator to determine the result of each test as *Pass*, *Fail*, or *Indeterminate* based on the calculated guardbanded limits. Results are then recorded, either on paper, or manually entered into a calibration data collection software system.
4.2 Semi-Automated Guardbanding
Semi-automated guardbanding involves using software designed for automated calibration, but without any facilities specifically designed to automated guardbanding. Test limits encoded in a calibration procedure may be adjusted according to a specified guardbanding criterion.
For example, suppose a particular calibration test requires a voltage source which generates 10 mV at 10 kHz. Suppose further that the UUT’s test specification at this test point is ± 0.01%. A calibration procedure language must provide a way for the procedure writer to encode the test specification, either in the procedure itself, or in an external data source (e.g., a database table). The procedure writer may decide, then, based on a particular guardbanding algorithm to tighten the test specification by 20%. To accomplish this it is necessary to change the test specification from ± 0.01% to ± 0.008%.
The virtue of this approach is that once a particular procedure has been updated to use guardbanded limits, the procedure may be re-used. Full automation with guardbanding is achieved this way. The disadvantages of this approach, however, are significant:
§ It requires a significant effort to update the calibration procedure.
§ The guardbanding may be static, rather than dynamic. For example, it may be difficult to vary the guardbanding factor based on the calculated measurement uncertainty.
§ The reporting system may not be able to easily distinguish between a calibration performed with guardbanding and one performed without guardbanding.
Some calibration software packages have sufficient flexibility that it is possible to overcome some of the disadvantages listed above. For example, the procedure writer may be able to directly take responsibility for calculating the guardbanded test limits in a dynamic way. The calibration procedure writer may also be able to generate result data containing information required to allow the reporting system to properly distinguish between guardbanded and non-guardbanded calibrations. However, even if a software package has the required, low-level flexibility to support guardbanding in this way, it is nevertheless a task which requires significant effort on the part of the procedure writer, and also complicates the calibration procedure, which may make maintenance more difficult. Furthermore, this approach entails hardwiring the guardbanding algorithm into the calibration procedure itself, which precludes allowing the end user to modify the guardbanding approach without changing the calibration procedure.
It is also worth noting in this context that some calibration software package implement facilities which, while not originally developed to support guardbanding, may be adaptable to a guardbanding purpose. For example, a calibration software package may include the ability to annotate a particular test as a marginal pass, or as significantly out of tolerance, or as an accredited calibration.
4.3 Automated Guardbanding
To improve upon the manual and semi-automated guardbanding methods discussed above, it is necessary to fully automate the guardbanding process. A satisfactory automated guardbanding solution should implement the following features:
§ The calculation of the guardbanded test limits should be automatic.
§ The procedure writer should be able to choose from a variety of built-in guardbanding algorithms.
§ It should be possible for the procedure writer to enable or disable guardbanding.
§ Result data sufficient to satisfy an auditor’s requirements should be saved in the calibration database.
§ It should be easy to generate a calibration report which distinguishes between guardbanded and non-guardbanded tests, and which shows both the guardbanded and nominal specification limits for each guardbanded test.
§ A procedure writer should be able to take an existing calibration procedure which was not written with guardbanding in mind and easily change it to a procedure for which guardbanding is enabled.
§ The software should be flexible and customizable. The guardbanding method and other guardbanding parameters should be configurable on a per-test, per-procedure, per-workstation, or per-site basis.
§ It should be possible to define guardbanding algorithms which differ from the built-in methods provided by the software application.
The following sections describe a number of built-in guardbanding methods.
4.3.1 TSR-Based Methods
*TSR*-based guardbanding methods are methods that use the *Test Specification Ratio (TSR)* as the independent parameter in the determination of the guardbanding factor.
The *TSR* is defined as:
\[ TSR = \frac{\text{(Test Specification)}}{\text{(Reference Specification)}} \]
Note that nomenclature has occasionally been inconsistent. *TSR* is sometimes referred to as *TAR* (Test Accuracy Ratio), and sometimes as *TUR* (Test Uncertainty Ratio). In this paper *TUR* refers to a ratio which involves the calculated measurement uncertainty, not simply the reference accuracy. *TSR*, as used here, is equivalent to *TAR*.
As a specific example, consider the *TSR-based* method defined by the following rules:
§ If the *TSR* is greater than 4.0, use 100% of the specification limits. (The guardbanding factor is 1.0).
§ If the *TSR* is between 1.5 and 4.0, use 80% of the specification limits. (The guardbanding factor is 0.8).
If the $TSR$ is less than 1.5, the validity of the test may be called into question because the reference accuracy and the UUT accuracy are too close. However, a particular organization may wish to define a guardbanding factor which applies when the $TSR$ is between 1.0 and 1.5.
The method described in the example above is easy to describe and relatively easy to implement. It does not require that the software have the ability to compute the measurement uncertainty, only that it have access to the reference accuracy and the nominal test specification of the UUT.
To make the method presented in the example above more flexible the automated calibration software should:
§ Allow the procedure writer or site administrator to configure the $TSR$ thresholds. In the example above the thresholds are 1.5 and 4.0. It should be possible to specify any reasonable $TSR$ value as a threshold.
§ Allow the guardbanding factors to be configured by the procedure writer or site administrator. In the example above the guardbanding factors are 0.8 and 1.0. The software should allow any reasonable values to be specified as guardbanding factors.
§ As a further generalization the software should allow a user-specified table of $TSR$ values with a corresponding guardbanding factor associated with each $TSR$ value. Any reasonable table size should be allowed. Allowing an arbitrary table in this fashion provides the user with a method to implement, in effect, a piece-wise linearization of any guardbanding function desired, as long as the function is defined as having one independent parameter (the $TSR$) and one dependent parameter (the guardbanding factor).
§ In addition, the software should also provide a configurable flag which allows the user to determine whether or not to linearly interpolate between points in the table or interpret the table as a step function, choosing the more conservative guardband factor for values which fall between points in the table.
4.3.2. TUR-Based Methods
A TUR-based guardbanding method is similar to a $TSR$-based method, but uses as the independent variable the Test Uncertainty Ratio (TUR) rather than the Test Specification Ratio.
The TUR is defined as:
$$\text{TUR} = \frac{\text{Test Specification}}{\text{Measurement Uncertainty}}$$
The Test Specification here is not normalized. The value stated by the manufacturer of the UUT is used directly. The Measurement Uncertainty here is the expanded measurement uncertainty, not the standard (1-sigma) uncertainty.
All of the configurable options described in the section on *TSR*-based methods (arbitrary table size, user-settable thresholds, user-settable guardband factors, and an interpolation flag) should be available with a *TUR*-based method as well.
Although a *TUR*-based method is essentially similar to a *TSR*-based method when viewed with respect to the guardbanding algorithm per se, *TUR*-based methods are nevertheless considerably more difficult to implement because the software must have the ability to calculate the measurement uncertainty for each calibration test.
A GUM-compliant measurement uncertainty calculation imposes significant additional requirements on the automated calibration software:
- The software must be able to automatically repeat each test a specified number of times, accumulating the measurements in order to calculate the standard deviation and include an uncertainty component based on the standard deviation.
- The software must be able to determine the resolution of the unit under test (UUT) in order to include an uncertainty component based on the UUT resolution.
- The software must allow the procedure writer or user to specify values of optional uncertainty components, and then calculate the standard uncertainty based on the reference accuracy (or accuracies if a test involves multiple references), the UUT resolution, the standard deviation of the measurements, and all specified optional uncertainty components.
- The software must implement the Welch-Satterthwaite approximation in order to determine the expanded measurement uncertainty. This in turn requires that the software allow the procedure writer or end user to specify various uncertainty-related parameters like degrees of freedom values, sensitivity coefficients, and confidence values.
If the calibration software has the capability to determine the measurement uncertainty as described above, implementing a *TUR*-based guardbanding method becomes simply a variation of implementing a *TSR*-based method.
### 4.3.3. Normalized TUR-Based Methods
A guardbanding method based on the normalized TUR is the same as the method based on the non-normalized TUR, as described in the preceding section, except that the independent variable is the normalized TUR.
The Normalized TUR is defined as:
\[
\text{NTUR} = \frac{\text{(Normalized Test Specification)}}{\text{(Standard Uncertainty)}}
\]
The standard uncertainty is available because the calculation of the expanded uncertainty involves first determining the standard uncertainty, then multiplying it by an appropriate coverage factor to determine the expanded uncertainty.
Determining the normalized test specification requires that the procedure writer be able to indicate the confidence associated with the UUT test specification, in order that the software has the information required to normalize the test specification to a 1-sigma value. In the absence of a direct specification of the confidence associated with the UUT specification the software application should use a default value of 2 sigma.
4.3.4. Measurement Uncertainty-Based Methods
A measurement uncertainty-based guardbanding method is a method that directly uses the expanded measurement uncertainty as the independent variable in the algorithm which determines the guardbanding factor.
Let:
\[
GUL = \text{guardbanded upper limit} \\
GLL = \text{guardbanded lower limit} \\
EU = \text{expanded measurement uncertainty} \\
GBF = \text{guardbanding factor}
\]
Then, using a measurement uncertainty-based guardbanding method:
\[
GUL = (\text{Upper Specification Limit}) - (GBF * EU) \\
GLL = (\text{Lower Specification Limit}) + (GBF * EU)
\]
The calibration software should allow the \( GBF \) (guardbanding factor) to be specified on a per-site, per-workstation, per-procedure, or per-test basis in a hierarchical fashion so that a more restrictive specification overrides a more general specification.
Setting the \( GBF \) to 1.0 effectively implements the conservative approach of tightening the nominal test specifications by the full amount of the calculated expanded measurement uncertainty. This approach was articulated in the Guide 25 Draft 5 recommendation, a precursor to the 17025 standard. The 17025 standard itself, however, does not directly state a requirement to implement the conservative approach of using a measurement uncertainty-based method with the \( GBF \) set to 1.0. Rather, the requirement is stated in a more flexible way, allowing the metrologist discretion in determining an appropriate guardbanding method. Using a measurement uncertainty-based guardbanding method with a \( GBF \) of less than 1.0 may require additional justification based on an analysis of the UUTs being calibrated and the references being used to calibrate them.
4.3.5. Result Data
Regardless of the guardbanding method selected, the software application should save result data in the calibration database sufficient to satisfy audit requirements and facilitate report generation.
Result data should include at least:
- Specification Limits
- Guardbanding Enable/Disable Flag
- Guardbanding Method
- Guardbanded Test Limits
- Guardband Factor
- Test Result (Pass, Fail, or Indeterminate)
- Overall Result of Calibration (Pass, Fail, or Indeterminate)
An ambiguity arises in connection with the determination of the overall result of the calibration procedure when one or more individual test results are *Indeterminate*.
Without guardbanding the overall procedure result is considered to be a *Pass* only if all individual tests within the procedure have a test result of *Pass*. However, with guardbanding the situation is less clear. Three approaches may be identified:
- In the first approach an individual test result of *Indeterminate* is regarded as equivalent to *Pass* in the determination of the overall test result.
- In the second approach an individual test result of *Indeterminate* is regarded as equivalent to *Fail* in the determination of the overall test result.
- In the third approach an individual test result of *Indeterminate* is regarded as equivalent to neither *Pass* nor *Fail*. This approach leads to the situation in which the overall result of the calibration procedure may be reported as *Indeterminate*.
Since user requirements may differ with respect to determining which of the three approaches to handling *Indeterminate* results is best at a particular site, the calibration software application should be configurable in this respect.
**Example**
Suppose a calibration verification procedure with 10 distinct tests is executed, and suppose the test results are 6 *Passes* and 4 *Indeterminates*.
With the first approach the overall result will be *Pass*, because the 4 *Indeterminate* results will be regarded as equivalent to *Pass*.
Using the second approach the overall result will be *Fail*, because the 4 *Indeterminate* results will be regarded as equivalent to *Fail*.
Finally, using the third approach the overall result will be *Indeterminate* because the 4 *Indeterminate* results are grouped with neither the *Pass* nor the *Fail* results.
### 4.3.6. Determination of the Test Result
As mentioned above, guardbanding affects the result of a calibration test. This section describes in some detail the typical method by which an automated calibration software application determines the test result, then shows how the method is modified to accommodate guardbanding.
The software first calculates the *Percent Error*, the error as a percentage of the test tolerance. The software then determines the test result by comparing *Percent Error* to 100%.
- If *Percent Error* is less than 100% the test result is *Pass*.
- If *Percent Error* is greater than 100% the test result is *Fail*.
- If *Percent Error* is exactly equal to 100% the test result is, by default, *Pass*, but the system should be configurable in this respect.
In order to determine *Percent Error* the software must first determine *Error* and *Test Tolerance*.
#### Determining the *Error*
Every calibration test compares two values, the *UUT Value* and the *Reference Value*.
*Error* is a signed value that represents the difference between the two values being compared. The sign is determined by convention.
*Error* is defined as:
\[
\text{Error} = \text{Measurement} - \text{Expected Value}
\]
Depending on the type of the test either:
**CASE 1:** *Measurement* is *UUT Value*, and *Expected Value* is *Reference Value*,
or
**CASE 2:** *Measurement* is *Reference Value*, and *Expected Value* is *UUT Value*.
In CASE 1 the *Error* is positive if the *UUT Value* (the *Measurement*) is greater than the *Reference Value* (the *Expected Value*), zero if the two values are the same, and negative otherwise.
In CASE 2 the *Error* is positive if the *Reference Value* (the *Measurement*) is greater than the *UUT Value* (the *Expected Value*), zero if the two values are the same, and negative otherwise.
**Determining the Test Tolerance**
The *Test Tolerance* of a test is specified in the calibration procedure. In order to calculate *Percent Error* the *Test Tolerance* must be converted to the same form as the *Error* (namely, absolute units). The procedure language implemented by the calibration software may allow the *Test Tolerance* to be represented in different ways. Examples are:
- Percentage of *Nominal*
- PPM of *Nominal*
- Percentage of *Range*
- PPM of *Range*
- Absolute Units (i.e., a floor specification)
Typically the *Test Tolerance* will be represented as an expression involving one or more terms, where each term is in one of the forms listed above. The software must convert each term to base units (for example, volts) and sum the terms to generate a single number which represents the *Test Tolerance*.
A minor complication is that the *Test Tolerance* may be asymmetrical. For example one might have a *Test Tolerance* of “+1 mV –2 mV”. When the *Test Tolerance* is asymmetrical the software must choose either the upper or the lower value when the test evaluation is performed. This choice is typically made based on a comparison of the *UUT Value* and the *Reference Value*. If the *UUT Value* is greater than or equal to the *Reference Value* the upper tolerance value is used as the *Test Tolerance*. Otherwise, the lower tolerance value is used as the *Test Tolerance*.
After *Error* and *Test Tolerance* have both been determined, *Percent Error* can be calculated as:
\[
\text{Percent Error} = \frac{|Error \times 100|}{TestTolerance}
\]
The sign of the *Error* does not matter in determining *Percent Error*.
When guardbanding is enabled, the test result determination is as described above, but with an intervening step which adjusts the *Test Tolerance* prior to the calculation of *Percent Error*.
5. Conclusion
Since there is no consensus at this point on precisely how to apply guardbanding in all situations, an effective automated calibration software solution must provide a flexible and customizable implementation. The software should:
- Allow the user to enable or disable guardbanding.
- Allow the user to specify the guardbanding method.
- Allow the user to specify a table of parameter thresholds with associated guardband factors.
- Allow the user to choose whether or not to interpolate between points in the table.
- Support guardbanding when the Test Tolerance is asymmetrical.
- Allow the user to specify how Indeterminate results are handled in the determination of the overall result of a calibration verification procedure.
- Provide mechanisms to allow the user (procedure writer) to invoke external programs or retrieve data from external files, where the external programs and data files are used to specify guardbanding factors and parameters.
- Implement configurability for all parameters in a hierarchical fashion where a per-test specification overrides a per-procedure specification, a per-procedure specification overrides a per-workstation specification, and a per-workstation specification overrides a per-site specification.
- Generate result data in a standard SQL database to allow an auditor to verify the guardbanding method, and to allow the user to generate calibration reports that show all necessary data related to guardbanding.
- Implement defaults such that the procedure writer may enable guardbanding without extensive modification of the calibration procedure. In the limiting case it should be possible to enable guardbanding without any modification of a calibration procedure.
The guardbanding approach described in this paper is intended to be flexible and customizable, and to allow most users to easily implement guardbanding in a way which satisfies their organizational requirements. However, as with most software projects, there are opportunities for improvement. Two particular opportunities for enhancement are:
- The table-based method described above, where the independent parameter is determined by the user, and an associated guardband factor is specified for each row in the table, could be made more general by allowing a mathematical expression to determine the guardband factor on each row of the table.
- More work should be done on guardbanding methods that directly maintain a constant risk of accepting an out-of-tolerance UUT.
6. References
1. ISO 17025, *General Requirements for the Competence of Testing and Calibration Laboratories*.
2. David Abell, Agilent Corp., *17025 Calibration Definition – Executive Summary, Use of Guard Bands*, 2002
3. ISO/DIS 14253.1 E 1995, *Geometrical Product Specification – Inspection by Measurement of Workpieces and Measuring Instruments*
4. David Deaver, Fluke Corp., *How to Maintain Your Confidence*, 1993
5. David Deaver, Fluke Corp., *Guardbanding with Confidence*, 1994
6. David Deaver, Fluke Corp., *Managing Calibration Confidence in the Real World*, 1996
7. David Deaver, Fluke Corp., *Guardbanding and the World of ISO Guide 25 – Is there Only One Way?*, 1999 |
Expression of bax and bcl2 Genes in MDMA-induced Hepatotoxicity on Rat Liver Using Quantitative Real-Time PCR Method through Triggering Programmed Cell Death
Mitra Behroozaghdam, Mehrdad Hashemi, Gholamreza Javadi, Reza Mahdian, and Mansoureh Soleimani
1Department of Genetics, Faculty of Sciences, Sciences and Research Branch, Islamic Azad University, Tehran, IR Iran
2Department of Genetics, Tehran Medical Sciences Branch, Islamic Azad University, Tehran, IR Iran
3Molecular Medicine Department, Biotechnology Research Center, Pasteur Institute of Iran, Tehran, IR Iran
4Cellular and Molecular Research Center, Iran University of Medical Sciences, Tehran, IR Iran
5Department of Anatomy, Iran University of Medical Sciences, Tehran, IR Iran
*Corresponding Author: Mehrdad Hashemi, Department of genetics, Tehran Medical Sciences Branch, Islamic Azad University, Tehran, IR Iran. Tel: +98-2122006664, Fax: +21-22080049, E-mail: firstname.lastname@example.org
Received 2014 October 15; Revised 2014 November 23; Accepted 2014 December 7.
Abstract
Background: 3-methylenedioxymethamphetamine (MDMA) is a synthetic and psychoactive drug, which is known popularly as Ecstasy and has toxic effects on human organs.
Objectives: Considering the potential toxic interaction, this study was performed to quantify the expression of bax and bcl2 genes in MDMA-induced hepatotoxicity on rat liver. Subsequently, we evaluated pentoxifylline as a possible protective drug on hepatotoxicity.
Materials and Methods: Adult male Wistar rats weighting 250 - 300 grams were used in the study. The rats were equally distributed into four experimental groups (5 rat/group). MDMA was dissolved in PBS and injected intraperitoneally (IP) including untreated control, MDMA (MDMA dissolved in PBS), treated-1 (MDMA followed by PTX) and treated-2 (PTX followed by MDMA). All animals given MDMA received 3 doses of 7.5mg/kg with two hours gap between doses. Liver tissue was removed after anesthetizing. Subsequently, RNA isolation, cDNA synthesis and Real-Time PCR were performed. Finally, data analyzed statistically to determine significantly differences between the groups (P value < 0.05).
Results: Using Real-Time quantitative PCR results, the gene expression ratio of bcl2 were calculated 93.80±20.64, 340.45 ± 36.60 and 47.13 ± 5.84 fold in MDMA, treated-1 and treated-2 groups, respectively. Furthermore, this ratio for bax gene obtained 2.13±0.33 fold in MDMA, 1.55 ± 0.26 fold in treated-1 and 10.44 ± 1.56 fold in treated-2 groups.
Conclusions: The present study focused on molecular mechanism of MDMA in programmed cell death using gene expression quantification of a pro-apoptotic and anti-apoptotic gene in MDMA-induced hepatotoxicity. The results showed that MDMA prompted apoptosis in liver and pentoxifylline protected against hepatotoxicity before and after taking MDMA.
Keywords: Pentoxifylline, Anti-Apoptotic Gene, MDMA, Pro-Apoptotic Gene, qReal-Time PCR
1. Background
3-methylenedioxymethamphetamine (MDMA) is a synthetic, psychoactive drug structurally similar to the stimulant methamphetamine and the hallucinogen mescaline (1). Although MDMA has been commonly considered as a safe drug, there are increasing evidences of its toxicity. MDMA is a neurotoxic substance for serotonergic and dopaminergic neurons and decreases 5-HT transporters (2). MDMA is known popularly as Ecstasy and affects human organs mainly brain, liver, kidney and heart. Liver is an organ of immense complexity that has fascinated mankind since antiquity (3). It comprises of multiple phenotypically distinct cell types. One type of polarized epithelial cell is the hepatocyte. Hepatocytes regulate intermediary metabolism and have an important role to detoxify endo- and xenobiotic, manufacture critical circulating proteins and generate bile acid dependent bile flow (3). According to the recent epidemiological data, the use of MDMA has increased almost in all parts of the world (4). Although many investigations reported MDMA-induced liver damage (5, 6), the fundamental mechanism accounting for hepatic toxicity is poorly understood.
Pentoxifylline (3, 7-Dihydro-3, 7-dimethyl-1-(5-oxohexyl)-1H-purine-2, 6-dione) has been classified as a haemorheological drug and a phosphodiesterase inhibitor. Pentoxifylline (PTX) is used for the treatment of intermittent claudication caused by peripheral artery disease. PTX significantly reduce blood viscosity in patients with...
peripheral arterial disorders and increase erythrocytes deformability in healthy subjects and patients with peripheral vascular disease (7).
Cell deaths have been classified by consensus agreement including apoptosis, necrosis, necroptosis, autophagy and cornification (8, 9). Apoptosis is a specific morphological aspect of cell death characterized by membrane blebbing, shrinkage of the cell, chromatin condensation and nuclear fragmentation (3). During apoptosis, pro- and anti-apoptotic genes interact with each other to control the integrity of the cell. The B-cell lymphoma-2 (Bcl-2) family is a group of proteins that regulate mitochondrial dysfunction (10). This family comprises anti-apoptotic (*bcl2*, *Bcl-xl*, *Bcl-w*, *Mcl-1* and *A1*), pro-apoptotic multidomain effector proteins (*bax*, *Bak*, and *Bok*) and BH3-only pro-apoptotic members such as *Bid*, *Bim*, *Bad*, *Bik*, *Bmf*, *Hrk*, *Noxa* and *Puma* (3).
2. Objectives
Considering the potential toxic interaction, this study was performed to evaluate hepatotoxic effects of MDMA on adult rat hepatocytes. Subsequently, we investigated the effect of pentoxifylline as possible protective drug on hepatotoxicity before and after taking MDMA.
3. Materials and Methods
3.1. Animal and Drug Administration
Sample size and power in each group were calculated with 95% confidence level, 5% confidence interval and 10 rats, respectively. In this experimental study, adult male Wistar rats weighting 250 - 300 grams (Cellular and molecular research center, Iran university of Medical Sciences, Tehran, Iran) were used. The animals were housed under reversed 12:12 hours light and dark cycles at constant room temperature (22 ± 2°C). All procedures were performed in accordance with ethics committee of cellular and molecular research center, Iran university of medical sciences and European communities council directive of November 1986 (86/609/EEC).
All chemicals were purchased from Sigma except PTX powder, which was gifted kindly by the Amin Pharmaceutical Co. (Esfahan-Iran). However, pure MDMA was gifted by Dr. Foroumadi, faculty of pharmacy and pharmaceutical sciences research center, Tehran university of medical sciences.
The rats were equally distributed into four experimental groups (5 rat/group) including untreated control, MDMA (MDMA dissolved in phosphate buffered saline), treated-1 (MDMA followed by PTX) and treated-2 (PTX followed by MDMA). MDMA was dissolved in PBS and injected intraperitoneally (IP). All animals given MDMA received 3 doses of 7.5 mg/kg with two hours gap between doses. In treated-1 group, pentoxifylline (100 mg/kg) was injected simultaneously with third dose of MDMA. In treated-2, MDMA was given one week after pentoxifylline.
3.2. Tissue Collecting
Two weeks after the first drug administration, all rats were anaesthetized by Ketamine and xylazine before killed. The liver was removed, dissected and frozen immediately in liquid nitrogen to store at -80°C.
3.3. Total RNA Isolation and cDNA Synthesis
Total RNA was isolated from 5 mg liver tissue using High Pure RNA isolation kit (Roche, Germany). According to the handbook kit, the samples were treated with DNase I enzyme supported in the kit to avoid DNA contamination. Finally, Optical Density ($A_{260}/A_{280}$ and $A_{260}/A_{230}$) and concentration were measured using Nanophotometer 2000c (Thermo Science, USA).
Reverse transcription reaction was performed using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Lithuania). Reaction volume was 20 µL and components were used including 1 µg of total RNA as template (11 µL), RevertAid RT 200 U/µL (1 µL), RiboLock RNase Inhibitor 20 U/µL (1 µL), Random Hexamer primer (1 µL), dNTP Mix 10 mM (2 µL) and Reaction Buffer 5X (4 µL). Then, samples were incubated for 10 minutes at 25°C, 60 minutes at 42°C and 5 minutes at 75°C.
3.4. Oligonucleotide Set Design
In the present study, *bcl2*, *bax* and Glyceraldehyde-3-phosphate dehydrogenase (*gapdh*) genes were selected as targets and internal reference gene, respectively. The sequences of interest genes were obtained from NCBI database and primer sets were designed via GeneRunner and Primer Express software v.3.0 (Applied Biosystems, Foster City, USA) and analyzed in Basic Local Alignment Search Tool to avoid secondary structure and homology with other genome region. Oligonucleotide sequences are shown in Table 1.
| Oligo Name | Sequence (5' → 3') | Amplicon, bp |
|------------|-------------------|--------------|
| *gapdh*-F | AAGTTCACGGCACAAGTTCAAGG | 22 |
| *gapdh*-R | CATACTCAGCACCACTCATCC | 22 |
| *bax*-F | AGGGTGCGCTGGGAAGGC | 17 |
| *bax*-R | TGAGCGAGGCGCGTGAGG | 17 |
| *bcl2*-F | ATCGCTCTGTGGATGACTGAGTAC | 24 |
| *bcl2*-R | AGAGACAGCCAGGAGAAATCAAAC | 24 |
3.5. qReal-Time PCR Data Analysis
Real-Time PCR is a high-technique monitoring amplification process step by step. In the assay, SYBR Green I was used as reporter dye. Real-Time PCR reactions were performed via Rotor-Gene Q apparatus (Qiagen, Germany). Total PCR reaction volume was 25 µL including 12 µL of SYBR Green I PCR Master Mix (TaKaRa, Japan), 1µL of forward and reverse oligonucleotide (400 nM), cDNA template (300 ng) and ddH$_2$O. Thermal cycling programme was performed for 30 seconds at 95°C (first denaturation), following 5 seconds at 95°C and 25 seconds at 60°C for 40 cycles and melting curve analysis ramping from 65°C to 95°C and rinsing 1°C each step.
Amplification efficiency for target and reference genes validated using 4-fold dilution series of control cDNA template as 2000, 200, 20 and 2 ng. Then, standard curve was drawn by plotting the logarithmic input cDNA concentration versus mean CT and the slope was determined. PCR reaction efficiency was calculated via the following formula;
$$E = \left(10^{\frac{-1}{\text{slope}}}\right) - 1$$
Expression level of target genes were calculated using comparative threshold cycle formula. The expression level of target genes to reference gene in treated samples compared to controls was calculated using the following formula:
$$\Delta \Delta CT = [(mCT_{\text{target}} - mCT_{\text{reference}})_{\text{test sample}} - (mCT_{\text{target}} - mCT_{\text{reference}})_{\text{control sample}}]$$
where $\Delta \Delta CT$ is calculated by the following equation (ii):
Equation 3.
$$\Delta \Delta CT = [(mCT_{\text{target}} - mCT_{\text{reference}})_{\text{test sample}} - (mCT_{\text{target}} - mCT_{\text{reference}})_{\text{control sample}}]$$
3.6. Statistical Analysis
In this study, all mathematical procedures were calculated using Statistical Package for the Social Sciences software (SPSS Inc v. 22). The statistical operations included mean ratio (M), Standard Deviation (SD), Determination coefficient (R²), Confidence Intervals (95% CI) and Standard Error of Mean (SEM). Furthermore, to determine differences between gene expression of interest groups, One-way ANOVA, non-parametric analysis (independent sample Kruskal-Wallis test) and then, Tukey (Post hoc test) were performed. P-value < 0.05 was considered statistically significant.
4. Results
4.1. Real-Time RT-PCR
The slope of standard curve for targets and reference genes were calculated from -3.316 to -3.344 (Figure 1). PCR efficiencies, additionally for all interest genes were obtained between 97.0 and 99.0%. the amplicons of $gapdh$, $bcl2$, and $bax$ were melted at 85.5°C, 86.2°C and 89.3°C, respectively (Figure 2). Melting curve analysis confirmed the specific amplification of fragments and no primer dimmers formation.
Figure 1. Standard Curves: Mean Threshold Cycle Versus Log cDNA Concentration
Y-intercept and Determination coefficient for interest genes; $bax$ ($y = -3.34x + 35.285$ and $R^2 = 0.99$), $bcl2$ ($y = -3.314x + 27.979$ and $R^2 = 1.00$) and $gapdh$ ($y = -3.32x + 39.509$ and $R^2 = 0.99$).
Figure 2. Melting Curve Analysis of the Target and Reference Genes for PCR Products. A: $gapdh$, B: $bcl2$, C: $bax$.
Real-Time quantitative PCR results revealed that gene expression ratio of anti-apoptetic gene, $bcl2$ were 93.80 ± 20.64, 340.45 ± 36.60 and 47.13 ± 5.84 fold in MDMA, treated-1 and treated-2 groups, respectively (Figure 3). Moreover, this ratio for $bax$, pro-apoptotic gene were 2.13 ± 0.33 fold in MDMA, 1.55 ± 0.26 fold in treated-1 and 10.44 ± 1.56 fold in treated-2 groups (Figure 3, Table 2).
Statistical analysis showed significant difference for \( bcl2 \) expression between experimental groups. The expression of this gene was increased in treated-1 compared to MDMA and treated-2 (P-value < 0.001) and between MDMA and treated-2 group (P-value < 0.01). Moreover, \( bax \) gene expression significantly increased in treated-2 compared to MDMA and treated-1 group (P-value < 0.001) and no significant difference between MDMA and treated-1 group (P-value > 0.05).
The ratio of \( bax \) to \( bcl2 \) (\( bax/bcl2 \)) for MDMA, treated-1 and treated-2 calculated as 0.02, 0.005 and 0.222, respectively.
5. Discussion
Although MDMA alone or in combination with other abused substances can damage various organs (such as liver, heart, brain and kidney), its use has been increased over the past decade. Many researches performed to figure out the mechanism of MDMA-mediated neurotoxicity (2, 12-14), but little research is performed about the mechanisms of MDMA-induced hepatotoxicity (2, 15, 16). As Liver is a target organ for MDMA toxicity, the present study focused on molecular mechanism of MDMA in apoptosis using gene expression quantification of a pro-apoptotic and an anti-apoptotic gene in hepatotoxicity. Data analysis showed that MDMA induced programmed cell death in liver switching pro-apoptotic gene on. This is demonstrated by upregulation of \( bax \) gene in treated-2 group approximately 7-fold than treated-1 (MDMA followed by PTX). The expression of \( bax \) increased in MDMA and treated-2 compared to normal controls (P value < 0.001). An investigation on apoptotic mechanisms triggered by MDMA (alone or combined) in immortalized human HepG2 cells showed down-regulation of BCL-2 and \( bcl_{-\chi} \) mRNA levels with concurrent upregulation of \( BAX \) and \( BAD \) genes (17).
Cristina Montiel-Duarte et al. reported that MDMA caused apoptosis in freshly isolated rat hepatocytes and cell line of hepatic stellate cells (HSC), as shown by chromatin condensation of the nuclei and accumulation of oligonucleosomal fragments in the cytoplasm. In both cell types, apoptosis correlated with decreased levels of \( BCL_{-\chi} \), release of cytochrome C from the mitochondria and activation of caspase-3 (2). These results suggested that liver cells apoptosis could be involved in hepatotoxicity of MDMA. Beitia et al. (18) demonstrated that MDMA significantly increased ALT, AST and ALP enzymes and reduced the number of hepatocytes in male Wistar rats. In another study, Kwan-Hoon Moon et al. showed that MDMA causes oxidative inactivation of key mitochondrial enzymes, which contributes to mitochondrial dysfunction and subsequent liver damage in MDMA-exposed animals (19).
Pentoxifylline (PTX) is a tri-substituted xanthine derivative designated chemically as 1-(5-oxohexyl)-3,7-dimethylxanthine and a hemorheologic agent, i.e. an agent that affects blood viscosity (20). However, the precise mechanism of Pentoxifylline and its effects in clinical improvement are still unknown. Recently, PTX with combined anti-inflammatory (TNF-α inhibition) and anti-fibrogenic properties has been found to be useful in patients with acute alcoholic hepatitis (21, 22). In the present study, PTX considered as a protective drug on MDMA-induced hepatotoxicity. The findings showed up-regulation of \( bcl2 \) gene in treated-1 compared to MDMA and treated-2 groups (\( P \) value < 0.001). Recent evidences showed that PTX act against MDMA through increasing expression of anti-apoptotic genes. In patients with severe alcoholic hepatitis, prednisolone and pentoxifylline are the first line therapy (23). Sharifi et al. reported that the number of apoptotic bodies significantly decreased in non-simultaneous injection of MDMA and PTX group than the other groups except control (15). In the parallel analysis, Movassaghi et al. performed chromosome staining in rat liver to investigate apoptotic bodies using TUNEL (Terminal deoxynucleotidyl transferase dUTP nickend labeling) assay (16). They showed that number of TUNEL-positive cells in rat liver was significantly decreased in MDMA followed PTX group compared to MDMA and vehicle groups. Besides, they published that apoptotic bodies were significantly decreased by PTX in the liver and there was no significant difference between the number of apoptotic bodies in MDMA followed PTX group and control groups (16).
The ratio of \( bax \) to \( bcl2 \) in interested groups showed pro-apoptotic gene to anti-apoptotic gene expression. In the present study, this ratio (\( bax/bcl2 \)) decreased in treated-1 group (finally received PTX) compared to other groups. Moreover, this ratio increased in treated-2 and MDMA compared to treated-1 (\( P \) value < 0.001).
**Acknowledgments**
Authors would like to thank Amin Pharmaceutical Co (Esfahan-Iran) for the gift of pentoxifylline, Dr. Foroumady for kindly gift of MDMA and research center of Islamic Azad university, Tehran medical sciences branch, Tehran, Iran.
**Footnotes**
**Authors’ Contribution:** Mitra Behroozaghdam: performed laboratory operation, manuscript preparation extensively. Mehrdad Hashemi: design of the study, provided technical support, data analysis and revised the paper. Reza Mahdian: analysis of data and provided technical support. Gholamreza Javadi: editing the manuscript and provided technical support. Mansoureh Soleimani: performed the statistical analysis and editing the manuscript.
**Funding/Support:** This work was supported by cellular and molecular research center, IRAN university of medical sciences. This study was not supported by official grant. However, it was performed by personnel budget.
**References**
1. Cagiano R, Bera I, Sabatini R, Flace P, Vermesan D, Vermesan H, et al. Effects on rat sexual behaviour of acute MDMA (ecstasy) alone or in combination with loud music. *Eur Rev Med Pharmacol Sci.* 2008;12(5):285–92. [PubMed: 1902421]
2. Montiel-Duarte C, Varela-Rey M, Oses-Prieto JA, Lopez-Zabalza MJ, Beitia G, Cenarruzabetia E, et al. 3,4-Methylenedioxymethamphetamine ("Ecstasy") induces apoptosis of cultured rat liver cells. *Biochim Biophys Acta.* 2002;1588(1):26–32. [PubMed: 12379310]
3. Malhi H, Guicciardi ME, Gores GJ. Hepatocyte death: a clear and present danger. *Physiol Rev* 2010;90(3):1165–94. doi: 10.1152/physrev.00061.2009. [PubMed: 20664081]
4. Cole JC. MDMA and the "ecstasy paradigm". *J Psychoactive Drugs.* 2014;46(1):44–56. doi: 10.1080/02791072.2014.878148. [PubMed: 24830185]
5. Downey C, Daly F, O'Boyle KM. An in vitro approach to assessing a potential drug interaction between MDMA (ecstasy) and caffeine. *Toxicol In Vitro.* 2014;28(2):231–9. doi: 10.1016/j.tiv.2013.10.021. [PubMed: 2424539]
6. Payance A, Scotto B, Perarnau JM, de Muret A, Bacq Y. Severe chronic hepatitis secondary to prolonged use of ecstasy and cocaine. *Clin Res Hepatol Gastroenterol.* 2013;37(5):e09–13. doi: 10.1016/j.clinre.2013.06.003. [PubMed: 23910059]
7. Movassaghi S, Nadia Sharifi Z, Mohammadzadeh F, Soleimani M. Pentoxifylline Protects the Rat Liver Against Fibrosis and Apoptosis Induced by Acute Administration of 3,4-Methylenedioxymethamphetamine (MDMA or Ecstasy). *Iran J Basic Med Sci.* 2013;16(8):922–7. [PubMed: 24106597]
8. Kroemer G, Galluzzi L, Vandenabeele P, Abrams J, Alnemri ES, Baehrecke EH, et al. Classification of cell death: recommendations of the Nomenclature Committee on Cell Death 2009. *Cell Death Differ.* 2009;16(1):3–11. doi: 10.1038/cdd.2008.150. [PubMed: 18846107]
9. Moher D, Pham B, Ausejo M, Saenz A, Hood S, Barber GG. Pharmacological management of intermittent claudication: a meta-analysis of randomised trials. *Drugs.* 2000;59(5):1057–70. [PubMed: 10852639]
10. Youle RJ, Strasser A. The BCL-2 protein family: opposing activities that mediate cell death. *Nat Rev Mol Cell Biol.* 2000;1(9):47–57. doi: 10.1038/108108. [PubMed: 18097445]
11. Kamalya AR, Shahrokhi F, Shamsian E, Hayat Nosalei M, Diba Najia P, Hashemi M, et al. Determination of sensitivity and specificity of a novel gene dosage assay for prenatal screening of trisomy 21 syndrome. *Clin Biochem.* 2012;45(5):267–71. doi: 10.1016/j.clinbiochem.2011.11.013. [PubMed: 22178104]
12. Bolton JL, Trush MA, Penning TM, Dryhurst G, Monks TJ. Role of quinones in toxicology. *Chem Res Toxicol.* 2000;13(3):355–60. [PubMed: 10725110]
13. Milhazes N, Silva-Oliveira T, Martins P, Garrido J, Oliveira C, Rego AC, et al. Synthesis and cytotoxic profile of 3,4-methylenedioxymethamphetamine ("ecstasy") and its metabolites on undifferentiated PC12 cells: A putative structure-toxicity relationship. *Chem Res Toxicol.* 2006;19(10):1294–304. doi: 10.1021/tx060123i. [PubMed: 17070098]
14. Roseman I, Leech K, Fielding A, Nutt DJ, Carhart-Harris RL. The effects of psilocybin and MDMA on between-network resting state functional connectivity in healthy volunteers. *Front Hum Neurosci.* 2014;8:204. doi: 10.3389/fnhum.2014.00204. [PubMed: 24904346]
15. Sharifi ZN, Movassaghi S, Foroumadi A, Hashemi M, Jafari Semnani S, Atashi M. The study of inhibitory effect of pentoxifylline on apoptosis of male Wistar rat hippocampus after long-term use of ecstasy. *J Dev Biol.* 2011;3:85–92.
16. Movassaghi S, Sharifi ZN, Mohammadzadeh F, Soleimani M. Pentoxifylline Protects the Rat Liver Against Fibrosis and Apop17. Dias da Silva D, Carmo H, Lynch A, Silva E. An insight into the hepatocellular death induced by amphetamines, individually and in combination: the involvement of necrosis and apoptosis. *Arch Toxicol*. 2013;87(12):2165–85. doi: 10.1007/s00204-013-1082-9. [PubMed: 23820845]
18. Beitia G, Cobreros A, Sainz I, Cenarruzabeitia E. Ecstasy-induced toxicity in rat liver. *Liver*. 2000;20(1):8–15. [PubMed: 10726956]
19. Moon KH, Upreti VV, Yu LR, Lee JJ, Ye X, Eddington ND, et al. Mechanism of 3,4-methylenedioxymethamphetamine (MDMA, ecstasy)-mediated mitochondrial dysfunction in rat liver. *Proteomics*. 2008;8(18):3906–18. doi: 10.1002/pmic.200800215. [PubMed: 18780394]
20. Kale R, Bajaj A, Mathew D. Development of matrix diffusion controlled drug delivery system of pentoxifylline. *Int J Pharm Pharm Sci*. 2010;2(1):22–30.
21. Akrivlidis E, Botla R, Briggs W, Han S, Reynolds T, Shakil O. Pentoxifylline improves short-term survival in severe acute alcoholic hepatitis: a double-blind, placebo-controlled trial. *Gastroenterology*. 2000;119(6):1637–48. [PubMed: 1113085]
22. Louvet A, Diaz E, Dharnany S, Coevoet H, Texier F, Thevenot T, et al. Early switch to pentoxifylline in patients with severe alcoholic hepatitis is inefficient in non-responders to corticosteroids. *J Hepatol*. 2008;48(3):465–70. doi: 10.1016/j.jhep.2007.10.010. [PubMed: 18164508]
23. Casanova J, Bataller R. Alcoholic hepatitis: Prognosis and treatment. *Gastroenterol Hepatol*. 2014;37(4):262–8. doi: 10.1016/j.gastrohep.2014.02.001. [PubMed: 24656653] |
EXHIBITS
Table of Contents
EXHIBIT 1: Ordinance No. CS-382
ORDINANCE NO. CS-382
AN ORDINANCE OF THE CITY COUNCIL OF THE CITY OF CARLSBAD, CALIFORNIA, ADOPTING A ZONE CODE AMENDMENT AND A LOCAL COASTAL PROGRAM AMENDMENT TO UPDATE THE CITY’S DENSITY BONUS REGULATIONS TO REFLECT CHANGES IN STATE LAW.
CASE NAME: DENSITY BONUS AMENDMENTS 2020
CASE NO: ZCA 2020-0001/LCPA 2020-0005
WHEREAS, Sections 65915 – 65918 of the California Government Code, known as State Density Bonus Law, requires a city or county to provide a developer that proposes a housing development within the jurisdictional boundaries of that city or county with a density bonus and other incentives or concessions for the production of lower income housing units, or for the donation of land within the development, if the developer agrees to construct a specified percentage of units for very low income, low-income, or moderate-income households or qualifying residents and meets other requirements.
WHEREAS, on Oct. 9, 2019, California Governor Gavin Newsom signed Assembly Bill 1763 (“AB 1763”) into law, which amended Section 65915 to further encourage and incentivize the application of State Density Bonus Law; and
WHEREAS, SB 1763 took effect Jan. 1, 2020, and existing provisions of the City of Carlsbad Municipal Code are inconsistent with the new law provisions; and
WHEREAS, California Government Code Section 65915(a) requires that all cities adopt an ordinance that specifies how compliance with State Density Bonus Law will be implemented; and
WHEREAS, staff has prepared a Zone Code Amendment ZCA 2020-0001 and Local Coastal Program Amendment LCPA2020-0005 pursuant to Chapter 21.52 of the Carlsbad Municipal Code, Section 30514 of the Public Resources Code, and Section 13551 of California Code of Regulations Title 14, Division 5.5; and,
WHEREAS, the Carlsbad Zone Code is the implementing ordinance of the Carlsbad Local Coastal Program, and therefore, an amendment to the Zone Code also constitutes an amendment to the Local Coastal Program; and
WHEREAS, pursuant to California Coastal Commission Regulations, a six-week public review period for the Local Coastal Program Amendment began on May 15, 2020 and ended on June 26, 2020; and
Sept. 15, 2020
EXHIBIT NO. 1
Ordinance
Carlsbad LCP-6-CAR-20-0078-2
California Coastal Commission
WHEREAS, on May 20, 2020, the Airport Land Use Commission reviewed and found the proposed Zone Code Amendment consistent with the adopted McClellan-Palomar Airport Land Use Compatibility Plan; and
WHEREAS, on June 17, 2020, the Planning Commission held a duly noticed public hearing as prescribed by law to consider ZCA 2020-0001/LCPA 2020-0005; and
WHEREAS, the Planning Commission adopted Planning Commission Resolution No. 7373 recommending to the City Council that ZCA 2020-0001/LCPA 2020-0005 be approved; and
WHEREAS, the City Council of the City of Carlsbad held a duly noticed public hearing as prescribed by law to consider ZCA 2020-0001/LCPA 2020-0005; and
WHEREAS, at said public hearing, upon hearing and considering all testimony and arguments, if any, of all persons desiring to be heard, the City Council considered all factors, including written public comments, if any, related to ZCA 2020-0001/LCPA 2020-0005; and
NOW THEREFORE, the City Council of the City of Carlsbad, California, ordains as follows:
1. The above recitations are true and correct.
2. The findings of the Planning Commission in Planning Commission Resolution No. 7373 shall also constitute the findings of the City Council.
3. Chapter 21.86 of the Carlsbad Municipal Code is hereby repealed and replaced to read as follows:
Chapter 21.86 DENSITY BONUS
21.86.010 Purpose.
The public good is served when there exists in a city, housing which is appropriate for the needs of and affordable to the public who reside within that city. There is in the City of Carlsbad a need for housing affordable to various groups, such as lower income, moderate income and senior citizen households. Therefore, it is in the public interest for the city to promote the construction of such additional housing through the exercise of its powers and utilization of its resources to facilitate the development of quality housing affordable for these types of households.
A. It is the purpose of this section to specify how compliance with Government Code Section 65915 et seq. ("State Density Bonus Law") will be implemented, as required by Government Code Section 65915, subdivision (a).
B. It is the purpose of this section to implement the goals, objectives and policies of the Housing Element of the city’s General Plan.
C. It is the purpose of this section to provide the implementing framework, as it relates to affordable housing density bonuses, and offer concessions and incentives for eligible housing developments which are consistent with the city's long-standing commitment to provide for affordable housing.
21.86.020 Definitions.
The definitions found in State Density Bonus Law shall apply to the terms contained in this section.
21.86.030 Applicability.
A housing development as defined in State Density Bonus Law shall be eligible for a density bonus and other regulatory incentives that are provided by State Density Bonus Law when the applicant seeks and agrees to provide very-low, low or moderate income housing units, or units intended to serve seniors, transitional foster youth, disabled veterans, homeless persons, and lower income students in the threshold amounts specified in State Density Bonus Law. A housing development includes only the residential component of a mixed-use project. A commercial development as defined in Section 21.86.110 shall be eligible for a commercial development bonus as provided in Section 21.86.110.
The granting of a density bonus, incentive or concession, pursuant to this section, shall not be interpreted, in and of itself, to require a general plan amendment, development code amendment, zone change, other discretionary approval, or the waiver of a city ordinance or provisions of a city ordinance unrelated to development standards.
21.86.040 Application Requirements.
A. Any applicant requesting a density bonus and any incentive(s), waiver(s), parking reductions, or commercial development bonus provided by State Density Bonus Law shall submit a density bonus report as described below concurrently with the filing of the planning application for the first discretionary permit required for the housing development, commercial development, or mixed-use development. The requests contained in the density bonus report shall be processed concurrently with the planning application. The applicant shall be informed whether the application is complete consistent with California Government Code Section 65943.
B. The density bonus report shall include the following minimum information:
1. Requested Density Bonus.
a. Summary table showing the maximum number of dwelling units permitted by the zoning and general plan excluding any density bonus units, proposed affordable units by income level, proposed bonus percentage, number of density bonus units proposed, total number of dwelling units proposed on the site, and resulting density in units per acre.
b. A tentative map and/or preliminary site plan, drawn to scale, showing the number and location of all proposed units, designating the location of proposed affordable units and density bonus units.
c. The zoning and general plan designations and assessor’s parcel number(s) of the housing development site.
d. A description of all dwelling units existing on the site in the five-year period preceding the date of submittal of the application and identification of any units rented in the five-year period. If dwelling units on the site are currently rented, income and household size of all residents of currently occupied units, if known. If any dwelling units on the site were rented in the five-year period but are not currently rented, the income and household size of residents occupying dwelling units when the site contained the maximum number of dwelling units, if known.
e. Description of any recorded covenant, ordinance, or law applicable to the site that restricted rents to levels affordable to very-low or lower income households in the five-year period preceding the date of submittal of the application.
f. If a density bonus is requested for a land donation, the location of the land to be dedicated, proof of site control, and reasonable documentation that each of the requirements included in California Government Code Section 65915, subdivision (g) can be met.
2. Requested Concession(s) or Incentive(s).
In the event an application proposes concessions or incentives for a housing development pursuant to State Density Bonus Law, the density bonus report shall include the following minimum information for each incentive requested, shown on a site plan if appropriate:
a. The City’s usual development standard and the requested development standard or regulatory incentive.
b. Except where mixed-use zoning is proposed as a concession or incentive, reasonable documentation to show that any requested incentive will result in identifiable and actual cost reductions to provide for affordable housing costs or rents.
c. If approval of mixed-use zoning is proposed, reasonable documentation that nonresidential land uses will reduce the cost of the housing development, that the nonresidential land uses are compatible with the housing development and the existing or planned development in the area where the proposed housing development will be located, and that mixed-use zoning will provide for affordable housing costs or rents.
3. Requested Waiver(s).
In the event an application proposes waivers of development standards for a housing development pursuant to State Density Bonus Law, the density bonus report shall include the following minimum information for each waiver requested on each lot, shown on a site plan if appropriate:
a. The City’s usual development standard and the requested development standard.
b. Reasonable documentation that the development standards for which a waiver is requested will have the effect of physically precluding the construction of a development at the densities or with the concessions or incentives permitted by California Government Code Section 65915.
4. Requested Parking Reduction.
In the event an application proposes a parking reduction for a housing development pursuant to California Government Code Section 65915, subdivision (p), a table showing parking required by the zoning regulations, parking proposed under Section 65915, subdivision (p), and reasonable documentation that the project is eligible for the requested parking reduction.
5. Child Care Facility.
If a density bonus or incentive is requested for a child care facility in a housing development, reasonable documentation that all of the requirements included in California Government Code Section 65915, subdivision (h) can be met.
6. Condominium Conversion.
If a density bonus or incentive is requested for a condominium conversion, reasonable documentation that all of the requirements included in California Government Code Section 65915.5 can be met.
7. Commercial Development Bonus.
If a commercial development bonus is requested for a commercial development, the application shall include the proposed partnered housing agreement and the proposed commercial development bonus, as defined in Section 21.86.110, and reasonable documentation that each of the standards included in Subsection 21.86.110(C) has been met.
8. Fee.
Payment of any fee in an amount set by resolution of the City Council for staff time necessary to determine compliance of the Density Bonus Plan with State Density Bonus Law.
21.86.050 Density Bonus.
All calculations are rounded up for any fractional numeric value in determining the total number of units to be granted, including base density and bonus density as well as the resulting number of affordable units needed for a given density bonus project.
A. If a housing development qualifies for a density bonus under more than one income category, or additionally as a senior citizen housing development as defined in Chapter 21.84 and State Density Bonus Law, or as housing intended to serve transitional foster youth, disabled veterans, homeless persons, or lower income students, the applicant shall identify the categories under which the density bonus would be associated and granted. Density bonuses from more than one category can be combined up to the maximum allowed under State Density Bonus law.
B. The density bonus units shall not be included in determining the number of affordable units required to qualify a housing development for a density bonus pursuant to State Density Bonus Law.
C. The applicant may elect to accept a lesser percentage of density bonus than the housing development is entitled to, or no density bonus, but no reduction will be permitted in the percentages of required affordable units contained in California Government Code Section 65915, subdivisions (b), (c), and (f). Regardless of the number of affordable units, no housing development shall be entitled to a density bonus of more than what is authorized under State Density Bonus Law.
21.86.060 Incentives.
A. Incentives include incentives and concessions as defined in State Density Bonus Law. The number of incentives that may be requested shall be based upon the number the applicant is entitled to pursuant to State Density Bonus Law.
B. Nothing in this section requires the provision of direct financial incentives for the housing development, including, but not limited to, the provision of financial subsidies, publicly owned land, fee waivers, or waiver of dedication requirements. The city, at its sole discretion, may choose to provide such direct financial incentives.
21.86.070 Local Coastal Program Consistency.
A. State Density Bonus Law provides that it shall not be construed to supersede or in any way alter or lessen the effect or application of the California Coastal Act of 1976 (Cal. Public Resources Code § 30000 et seq.), and further provides that the granting of a density bonus or an incentive shall not be interpreted, in and of itself, to require a local coastal plan amendment.
B. For development within the coastal zone, any requested density bonus, incentive(s), waiver(s), parking reduction(s), or commercial development bonus shall be consistent with all applicable requirements of the certified Carlsbad Local Coastal Program, with the exception of density.
21.86.080 Review Procedures.
All requests for density bonuses, incentives, parking reductions, waivers, or commercial development bonuses shall be considered and acted upon by the approval body with authority to approve the development within the timelines prescribed by California Government Code Section 65950 et seq., with right of appeal to the City Council.
A. Eligibility for Density Bonus, Incentive(s), Parking Reduction, and/or Waiver(s) for a Housing Development. To ensure that an application for a housing development conforms with the provisions of State Density Bonus Law and the Coastal Act, the staff report presented to the decision-making body shall state whether the application conforms to the following requirements of state law as applicable:
1. The housing development provides the affordable units or senior housing required by State Density Bonus Law to be eligible for the density bonus and any incentives, parking reduction, or waivers requested, including the replacement of units rented or formerly rented to very-low and low income households as required by California Government Code Section 65915, subdivision (c)(3).
2. Any requested incentive will result in identifiable and actual cost reductions to provide for affordable housing costs or rents; except that, if a mixed-use development is requested, the application must instead meet all of the requirements of California Government Code Section 65915, subdivision (k)(2).
3. The development standards for which a waiver is requested would have the effect of physically precluding the construction of a development at the densities or with the concessions or incentives permitted by California Government Code Section 65915.
4. The housing development is eligible for any requested parking reductions under California Government Code Section 65915, subdivision (p).
5. If the density bonus is based all or in part on donation of land, all of the requirements included in California Government Code Section 65915, subdivision (g) have been met.
6. If the density bonus or incentive is based all or in part on the inclusion of a child care facility, all of the requirements included in California Government Code Section 65915, subdivision (h) have been met.
7. If the density bonus or incentive is based all or in part on the inclusion of affordable units as part of a condominium conversion, all of the requirements included in California Government Code Section 65915.5 have been met.
8. If the housing development is in the coastal zone, the requested density bonus and any requested incentive(s), waiver(s), or parking reduction(s) are consistent with all applicable requirements of the certified Carlsbad Local Coastal Program, with the exception of density.
B. If a commercial development bonus is requested for a commercial development, the decision-making body shall make a finding that the development complies with all of the requirements of Subsection 21.86.110(C), that the city has approved the partnered housing agreement, and that the commercial development bonus has been mutually agreed upon by the city and the commercial developer. If the project is in the coastal zone, the decision-making body shall also find that the commercial development bonus is consistent with all applicable requirements of the certified Carlsbad Local Coastal Program, with the exception of density.
C. The decision-making body shall grant an incentive requested by the applicant unless it makes a written finding, based upon substantial evidence, of any of the following:
1. The proposed incentive does not result in identifiable and actual cost reductions to provide for affordable housing costs, as defined in California Health and Safety Code Section 50052.5, or for affordable rents, as defined in California Health and Safety Code Section 50053; or
2. The proposed incentive would be contrary to state or federal law; or
3. The proposed incentive would have a specific, adverse impact upon public health or safety or the physical environment or on any real property that is listed in the California Register of Historic Resources, and there is no feasible method to satisfactorily mitigate or avoid the specific adverse impact without rendering the housing development unaffordable to low and moderate income households. For the purpose of this subsection, specific adverse impact means a significant, quantifiable, direct, and unavoidable impact, based on objective, identified, written public health or safety standards, policies, or conditions as they existed on the date that the application for the housing development was deemed complete.
D. The decision-making body shall grant the waiver of development standards requested by the applicant unless it makes a written finding, based upon substantial evidence, of any of the following:
1. The proposed waiver would be contrary to state or federal law; or
2. The proposed waiver would have an adverse impact on any real property listed in the California Register of Historic Resources; or
3. The proposed waiver would have a specific, adverse impact upon public health or safety or the physical environment, and there is no feasible method to satisfactorily mitigate or avoid the specific adverse impact without rendering the housing development unaffordable to low and moderate income households. For the purpose of this subsection, specific adverse impact means a significant, quantifiable, direct, and unavoidable impact, based on objective, identified, written public health or safety standards, policies, or conditions as they existed on the date that the application for the housing development was deemed complete.
E. If any density bonus, incentive, parking reduction, waiver, or commercial development bonus is approved pursuant to this chapter, the applicant shall enter into an affordable housing agreement or senior housing agreement with the city pursuant to Section 21.86.090.
21.86.090 Affordable Housing Agreement and Senior Housing Agreement.
A. Affordable Housing Agreement. Except where a density bonus, incentive, waiver, parking reduction, or commercial development bonus is provided for a market-rate senior housing development, the applicant shall enter into an affordable housing agreement with the city, in a form approved by the city attorney, to be executed by the city manager, to ensure that the requirements of this section are satisfied. The affordable housing agreement shall guarantee the affordability of the affordable units for a minimum of 55 years or a longer period of time if required by the construction or mortgage financing assistance program, mortgage insurance program, or rental subsidy program; shall identify the type, size and location of each affordable unit; and shall specify phasing of the affordable units in relation to the market-rate units.
B. Senior Housing Agreement. Where a density bonus, waiver, or parking reduction is provided for a market-rate senior housing development, the applicant shall enter into a restrictive covenant with the city, running with the land, in a form approved by the city attorney, to be executed by the city manager, to require that the housing development be operated as “housing for older persons” consistent with state and federal fair housing laws.
C. The executed affordable housing agreement or senior housing agreement shall be recorded against the housing development prior to final or parcel map approval, or, where a map is not being processed, prior to issuance of building permits for the housing development. The affordable housing agreement or senior housing agreement shall be binding on all future owners and successors in interest.
D. The affordable housing agreement shall include, but not be limited to, the following:
1. The number of density bonus dwelling units granted;
2. The number and type of affordable dwelling units
3. The unit size(s) (square footage) of target dwelling units and the number of bedrooms per target dwelling unit;
4. The proposed location of the affordable dwelling units;
5. Schedule for production of affordable dwelling units;
6. Incentives or concessions or waivers provided by the city;
7. Where applicable, tenure and conditions governing the initial sale of the affordable units;
8. Where applicable, tenure and conditions establishing rules and procedures for qualifying tenants, setting rental rates, filling vacancies, and operating and maintaining units for affordable rental dwelling units
9. Marketing plan; publication and notification of availability of affordable units;
10. Compliance with federal and state laws;
11. Prohibition against discrimination;
12. Indemnification;
13. City’s right to inspect units and documents;
14. Remedies.
21.86.100 Design and Quality.
A. The city may not issue building permits for more than 50 percent of the market rate units until it has issued building permits for all of the affordable units, and the city may not approve any final inspections or certificates of occupancy for more than 50 percent of the market rate units until it has issued final inspections or certificates of occupancy for all of the affordable units.
B. Affordable units shall be comparable in exterior appearance and overall quality of construction to market rate units in the same housing development. Interior finishes and amenities may differ from those provided in the market rate units, but neither the workmanship nor the products may be of substandard or inferior quality as determined by the city.
C. The number of bedrooms of the affordable units shall at least equal the minimum number of bedrooms of the market rate units.
21.86.110 Commercial Density Bonus.
A. The following definitions shall apply to Commercial Density Bonus:
1. “Commercial development” means a development project for nonresidential uses.
2. “Commercial development bonus” means a modification of development standards mutually agreed upon by the city and a commercial developer and provided to a commercial development eligible for such a bonus under Subsection 21.86.110(C). Examples of a commercial development bonus include an increase in floor area ratio, increased building height, or reduced parking.
3. “Partnered housing agreement” means an agreement approved by the city between a commercial developer and a housing developer identifying how the commercial development will provide housing available at affordable ownership cost or affordable rent consistent with Subsection 21.86.110(C). A partnered housing agreement may consist of the formation of a partnership, limited liability company, corporation, or other entity recognized by the state in which the commercial developer and the housing developer are each partners, members, shareholders, or other participants, or a contract between the commercial developer and the housing developer for the development of both the commercial development and the housing development.
B. When an applicant proposes to construct a commercial development and has entered into a partnered housing agreement approved by the city, the city shall grant a commercial development bonus mutually agreed upon by the developer and the city. The commercial development bonus shall not include a reduction or waiver of fees imposed on the commercial development to provide for affordable housing.
C. The requirements for commercial development bonus are as follows, which also be described in the partnered housing agreement:
1. The housing development shall be located either: (A) on the site of the commercial development; or (B) on a site within the city that is within one-half mile of a major transit stop and is located in close proximity to public amenities, including schools and employment centers.
2. At least 30 percent of the total units in the housing development shall be made available at affordable ownership cost or affordable rent for low-income households, or at least 15 percent of the total units in the housing development shall be made available at affordable ownership cost or affordable rent for very low-income households.
3. The commercial developer must agree either to directly build the affordable units; donate a site consistent with subparagraph 1 above for the affordable units; or make a cash payment to the housing developer for the affordable units.
D. Any approved partnered housing agreement shall be described in the city’s Housing Element annual report as required by California Government Code Section 65915.7, subdivision (k).
21.86.120 Interpretation.
If any portion of this chapter conflicts with State Density Bonus Law or other applicable state law, state law shall supersede this chapter. Any ambiguities in this chapter shall be interpreted to be consistent with State Density Bonus Law.
21.86.130 Inclusionary housing.
All housing development projects are required to provide affordable housing units in accordance with Chapter 21.85, Inclusionary Housing, of this title. If an applicant seeks to construct affordable housing to qualify for a density bonus in accordance with the provisions of this chapter, those affordable dwelling units provided to meet the inclusionary requirement established pursuant to Chapter 21.85 of this title shall be counted toward satisfying the density bonus requirements of this chapter.
21.86.140 Severability.
If any provision of this chapter or its application to any person or circumstances is held invalid, the remainder of the chapter and the application of the provision to other persons not similarly situated or to other circumstances shall not be affected. (Ord. CS-102 § CXVI, 2010; Ord. NS-794 § 11, 2006)
EFFECTIVE DATE OF THIS ORDINANCE APPLICABLE TO PROPERTIES OUTSIDE THE COASTAL ZONE:
This ordinance shall be effective thirty days after its adoption; and the City Clerk shall certify the adoption of this ordinance and cause the full text of the ordinance or a summary of the ordinance prepared by the City Attorney to be published at least once in a newspaper of general circulation in the City of Carlsbad within fifteen days after its adoption.
EFFECTIVE DATE OF THIS ORDINANCE APPLICABLE TO PROPERTIES INSIDE THE COASTAL ZONE:
This ordinance shall be effective thirty days after its adoption or upon Coastal Commission approval of LCPA 2020-0005, whichever occurs later; and the City Clerk shall certify the adoption of this ordinance and cause the full text of the ordinance or a summary of the ordinance prepared by the City Attorney to be published at least once in a newspaper of general circulation in the City of Carlsbad within fifteen days after its adoption.
INTRODUCED AND FIRST READ at a Regular Meeting of the Carlsbad City Council on the 1st day of September, 2020, and thereafter
PASSED, APPROVED AND ADOPTED at a Regular Meeting of the City Council of the City of Carlsbad on the 15th day of September, 2020, by the following vote, to wit:
AYES: Hall, Blackburn, Bhat-Patel, Schumacher.
NAYS: None.
ABSENT: None.
APPROVED AS TO FORM AND LEGALITY:
CELIA A. BREWER, City Attorney
MATT HALL, Mayor
BARBARA ENGLESON, City Clerk (SEAL) |
Applying Design Patterns to Flexibly Configure Network Services in Distributed Systems
Douglas C. Schmidt
firstname.lastname@example.org
http://www.ece.uci.edu/~schmidt/
Department of Electrical & Computer Science
University of California, Irvine 92607*
This paper appeared as a chapter in the book *Design Patterns in Communications*, (Linda Rising, ed.), Cambridge University Press, 2000. An earlier version of this paper appeared in the International Conference on Configurable Distributed Systems, Annapolis, Maryland, May 6–8, 1996.
Abstract
This paper describes how design patterns help to enhance the flexibility and extensibility of communication software by permitting network services to evolve independently of the strategies used to passively initialize the services. The paper makes three contributions to the study and development of configurable distributed applications. First, it identifies five orthogonal dimensions of passive service initialization: service advertisement, endpoint listening, service handler creation, passive connection establishment, and service handler activation. Second, the paper illustrates how design patterns have been used to build a communication software framework that supports flexible configuration of different strategies for each of these five dimensions. Third, the paper demonstrates how design patterns and frameworks are being used successfully to develop highly configurable production distributed systems.
1 Introduction
Despite dramatic increases in network and host performance, developing extensible communication software for distributed systems remains hard. *Design patterns* [1] are a promising technique for capturing and articulating proven techniques for developing extensible distributed software. A design pattern captures the static and dynamic structures and collaborations of components in a software architecture. It also aids the development of extensible components and frameworks by expressing the structure and collaboration of participants in a software architecture at a level higher than (1) source code or (2) object-oriented design models that focus on individual objects and classes.
This paper examines design patterns that form the basis for flexibly configuring network services in applications built by the author and his colleagues for a number of production distributed systems. Due to stringent requirements for reliability and performance, these projects provided an excellent testbed for capturing and articulating the key structure, participants, and consequences of design patterns for building extensible distributed systems.
The primary focus of this paper is the *Acceptor* component in the *Acceptor-Connector* pattern [2]. This design pattern decouples connection establishment and service initialization from service processing in a networked system. The *Acceptor* component is a role in this pattern that enables the tasks performed by network services to evolve independently of the strategies used to initialize the services *passively*.
When instantiated and used in conjunction with other patterns, such as Reactor [2] and Strategy [1], the Acceptor-Connector pattern provides a reusable component in the ACE framework [3]. ACE provides a rich set of reusable object-oriented components that perform common communication software tasks. These tasks include event demultiplexing, event handler dispatching, connection establishment, routing, dynamic configuration of application services, and concurrency control.
This paper is organized as follows: Section 2 motivates the Acceptor-Connector pattern by illustrating how it has been applied in production application-level Gateways; Section 3 outlines the *Acceptor* component of the Acceptor-Connector pattern; Section 4 illustrates how to implement *Acceptor*'s flexibly and efficiently by applying the Wrapper Facade [2], Strategy, Bridge, Factory Method, and Abstract Factory design patterns [1]; Section 5 outlines how Acceptors have been used to implement application-level Gateways; Section 6 discusses related patterns; and Section 7 presents concluding remarks.
2 Background and Motivation
2.1 Separating Connection establishment and service initialization
Many network services, such as remote login, file transfer, and WWW HTML document transfer, use connection-oriented protocols, such as TCP, to deliver data reliably between two or more communication endpoints. Establishing connections between endpoints involves the following two roles:
1. The *passive* role, which initializes an endpoint of communication at a particular address and waits passively for the other endpoint(s) to connect with it.
2. The *active* role, which actively initiates a connection to one or more endpoints that are playing the passive role.
The intent of the Acceptor-Connector pattern described in this paper is to decouple passive initialization of a service from the tasks performed after the service is initialized. This pattern was discovered by generalizing from extensive experience building reusable communication frameworks for a range of distributed systems [3]. In all these systems, the tasks performed by a service are independent of the following:
- **Which endpoint initiated the connection:** Connection establishment is inherently asymmetrical since the passive endpoint *waits* whereas the active endpoint *initiates* the connection. After the connection is established, however, data may be transferred between endpoints in a manner that obeys a service’s communication protocol, which can be structured as peer-to-peer, request-response, oneway streaming, etc.
- **The network programming interfaces and underlying protocols used to establish the connection:** Different network programming interfaces, such as sockets or TLI, provide different routines to establish connections using various underlying transport protocols. After a connection is established, however, data may be transferred between endpoints using standard read/write system calls that communicate between separate endpoints in a distributed application.
- **The strategies used to initialize the service:** The processing tasks performed by a service are typically independent of the initialization strategies used to (1) advertise the service, (2) listen for connection requests from peers, (3) create a service handler to process those requests, (4) establish the connection with the peers, and (5) execute the service handler in one or more threads or processes. Explicitly decoupling these initialization strategies from the service behavior itself enhances the extensibility, reusability, and portability of the service.
2.2 Motivating Example
Figure 1 illustrates how the Acceptor-Connector pattern has been used to implement multi-service, application-level Gateways, which is described further in [4]. A Gateway
is a mediator [1] that routes data between services running on Peers located throughout a wide area and local area network. From the Gateway’s perspective, Peer services differ solely by their message framing formats and payload types. Several types of data, such as status information, bulk data, and commands, are exchanged by services running on the Gateway and the Peers. Peers are located throughout local area networks (LANs) and wide-area networks (WANs) and are used to monitor and control network resources, such as satellites, call centers, or remote branch offices.
The Gateway uses a connection-oriented interprocess communication (IPC) mechanism, such as TCP, to transmit data between its connected Peers. Connection-oriented protocols simplify application error handling and can enhance performance over long-delay WANs. Each communication service in the Peers sends and receives status information, bulk data, and commands to and from the Gateway using separate TCP connections. Each connection is bound to a unique address, e.g., an IP address and port number. For instance, bulk data sent from a ground station Peer through the Gateway is connected to a different port than status information sent by a tracking station peer through the Gateway to a ground station Peer. Separating connections in this manner allows more flexible routing strategies and more robust error handling if connections fail or become flow controlled.
One way to design the Peers and Gateway is to tightly couple the connection roles with the network services. For instance, the Gateway could be hard-coded to play the active connection role and initiate connections for all its services. To accomplish this, it could iterate through a list of Peers and synchronously connect with each of them. Likewise, Peers could be hard-coded to play the passive role and accept the connections and initialize their services. Moreover, the active
and passive connection code for the Gateway and Peers, respectively, could be implemented with conventional network programming interfaces like sockets or TLI. In this case, a Peer could call socket, bind, listen, and accept to initialize a passive-mode listener socket and the Gateway could call socket and connect to actively initiate a data-mode connection socket. After the connections were established, the Gateway could route data for each type of service it provided.
However, the tightly coupled design outlined above has the following drawbacks:
- **Limited extensibility and reuse of the Gateway and Peer software:** For example, the mechanisms used to establish connections and initialize services are independent of the type of routing service, e.g., status information, bulk data, or commands, performed by the Gateway. In general, these services tend to change more frequently than the connection and initialization mechanisms.
- **Inflexible connection roles:** There are circumstances where the Gateway must play the *passive* connection role and the Peers play the active role. Therefore, tightly coupling the software that implements connection establishment with the software that implements the service makes it hard to (1) reuse existing services, (2) extend the Gateway by adding new routing services and enhancing existing services, and (3) reconfigure the connection roles played by Peers and the Gateway.
- **Non-portable and error-prone interfaces:** Using low-level network programming, such as sockets or TLI, is non-portable and error-prone. These low-level interfaces do not provide adequate type-checking since they utilize low-level I/O handles. It is easy to accidentally misuse these interfaces in ways that cannot be detected until run-time.
Therefore, a more flexible and efficient way to design the Peers and Gateway is to use the Acceptor pattern.
### 3 The Acceptor-Connector Pattern
This section presents a brief overview of the Acceptor-Connector pattern. A comprehensive discussion is available in [2].
**Intent:** The intent of the Acceptor-Connector pattern is to decouple connection establishment and service initialization from service processing in a networked system.
**Forces:** The Acceptor-Connector pattern resolves the following forces for distributed applications that use connection-oriented communication protocols:
1. **The need to reuse connection establishment code for each new service.** Key characteristics of services, such as the communication protocol or the data format, should be able to evolve independently and transparently from the mechanisms used to establish the connections. Since service characteristics change more frequently than connection establishment mechanisms, separating these concerns helps to reduce software coupling and increase code reuse.
2. **The need to make the connection establishment code portable across platforms that contain different network programming interfaces.** Parameterizing the Acceptor-Connector’s mechanisms for accepting connections and performing services helps to improve portability by allowing the wholesale replacement of these mechanisms. This makes the connection establishment code portable across platforms that contain different network programming interfaces, such as sockets but not TLI, or vice versa.
3. **The need to enable flexible service concurrency policies.** After a connection is established, peer applications use the connection to exchange data to perform some type of service, such as remote login or HTML document transfer. A service can run in a single-thread, in multiple threads, or multiple processes, regardless regardless of how the connection was established or how the services were initialized.
4. **The need to ensure that a passive-mode I/O handle is not accidentally used to read or write data.** By strongly decoupling the connection establishment logic from the service processing logic, passive-mode socket endpoints cannot be used incorrectly, e.g., by trying to read or write data on a passive-mode listener socket used to accept connections. This eliminates an important class of network programming errors.
5. **The need to actively establish connections with large number of peers efficiently.** When an application must establish connections with a large number of peers efficiently over long-delay WANs it may be necessary to use asynchrony and initiate and complete multiple connections in non-blocking mode.
**Structure and participants:** The structure of the key participants in the Acceptor-Connector pattern is illustrated in Figure 2. The Acceptor and Connector components are factories that assemble the resources necessary to connect and activate Svc_Handlers. Svc_Handlers are components that exchange messages with connected Peers.
The participants in the Connection Layer of the Acceptor-Connector pattern can leverage off the Reactor pattern. For instance, the Connector’s asynchronous initialization strategy establishes a connection after a reactor notifies it that a previously initiated connection request to a Peer has completed. Using the Reactor pattern enables multiple Svc_Handlers
to be initialized asynchronously within a single thread of control.
To increase flexibility, **Acceptor** and **Connector** components can be parameterized by a particular type of IPC mechanism and **SVC_HANDLER**. The IPC mechanism supplies the underlying transport mechanism, such as C++ wrapper facades [2] for sockets or TLI, used to establish a connection. The **SVC_HANDLER** specifies an abstract interface for defining a service that communicates with a connected Peer. A **Svc_Handler** can be parameterized by a **PEER_STREAM** endpoint. The **Acceptor** and **Connector** components associate this endpoint to its **Peer** when a connection is established.
By inheriting from **Event_Handler**, a **Svc_Handler** can register with a **Reactor** and use the Reactor pattern to handle its I/O events within the same thread of control as the **Acceptor** or **Connector**. Conversely, a **Svc_Handler** can use the Active Object pattern and handle its I/O events in a separate thread. The tradeoffs between these two patterns is described in [4].
Figure 2 illustrates how parameterized types can be used to decouple the Acceptor-Connector pattern’s connection establishment strategy from the type of service and the type of connection mechanism. Application developers supply template arguments for these types to produce Application Layer **Acceptor** or **Connectors**. This design enables the wholesale replacement of the **SVC_HANDLER** and IPC mechanism, without affecting the Acceptor-Connector pattern’s service initialization strategy.
Note that a similar degree of decoupling could be achieved via inheritance and dynamic binding by using the Abstract Factory or Factory Method patterns described in [1]. Parameterized types were used to implement this pattern since they improve run-time efficiency. In general, templates trade compile- and link-time overhead and space overhead for improved run-time performance.
**Dynamics:** Figure 3 illustrates the dynamics among participants for the **Acceptor** component of the pattern. These dynamics are divided into the following three phases:
1. **Endpoint initialization phase**, which creates a passive-mode endpoint encapsulated by **PEER_ACCEPTOR** that is bound to a network address, such as an IP address and port number. The passive-mode endpoint listens for connection requests from **Peers**. This endpoint is registered with the **Reactor**, which drives the event loop that waits on the endpoint for connection requests to arrive from **Peers**.
2. **Service activation phase**. Since an **Acceptor** inherits from an **Event Handler** the **Reactor** can dispatch the **Acceptor**’s **handle_event** method when connection request events arrive. This method performs the **Acceptor**’s **Svc_Handler** initialization strategy, which (1) assembles the resources necessary to create a new **Concrete_Svc_Handler** object, (2) accepts the connection into this object, and (3) activates the **Svc_Handler** by calling its **open** hook method.
3. **Service processing phase**. After the **Svc_Handler** is activated, it processes incoming event messages arriving on the **PEER_STREAM**. A **Svc_Handler** can process incoming event messages in accordance with patterns, such as the **Reactor** or the Active Object [2].
The dynamics among participants in **Connector** component of the pattern can be divided into the following three phases:
1. **Connection initiation phase**, which actively connects one or more `Svc_Handler` with their peers. Connections can be initiated synchronously or asynchronously. The `Connector`'s `connect` method implements the strategy for establishing connections actively.
2. **Service initialization phase**, which activates a `Svc_Handler` by calling its `open` method when its connection completes successfully. The `open` method of the `Svc_Handler` then performs service-specific initialization.
3. **Service processing phase**, which performs the application-specific service processing using the data exchanged between the `Svc_Handler` and its connected `Peer`.
### 4 Applying Design Patterns to Develop Extensible Acceptors
This section describes how to implement a highly configurable instance of the `Acceptor` component from the Acceptor-Connector pattern by applying other design patterns, in particular Wrapper Facade [2], Strategy, Bridge, Factory Method, and Abstract Factory [1]. These patterns enable an `Acceptor` to flexibly configure alternative strategies for *service advertisement*, *endpoint listening*, *service handler creation*, *service connection acceptance*, and *service activation*. In this section, we focus on the `Svc_Handler` and the `Acceptor` components shown in Figure 2. The `Connector` component can be implemented in similarly.
#### 4.1 The `Svc_Handler` Class
This abstract C++ class provides a generic interface for processing services. Applications customize this class to perform a particular type of service. The C++ interface for the `Svc_Handler` is shown below:
```cpp
template <class PEER_STREAM>
// Type of IPC mechanism.
class Svc_Handler {
public:
// Pure virtual method (defined by subclass).
virtual int open (void) = 0;
protected:
// Instance of IPC mechanism.
PEER_STREAM stream_;
};
```
Each `Svc_Handler` contains a communication endpoint, called `peer_stream_`, of parameterized type `PEER_STREAM`. This endpoint is used to exchange data between the `Svc_Handler` and its connected peer. After a connection is successfully accepted, an `Acceptor` activates a `Svc_Handler` by calling its `open` method. This pure virtual method must be overridden by a concrete `Svc_Handler` subclass and performs service-specific initializations.
#### 4.2 The Acceptor Class
This abstract C++ class implements the generic strategy for passively initializing network services, which are implemented as concrete `Svc_Handler`s. An `Acceptor` instance coordinates the following five orthogonal dimensions of passive service initialization:
1. **Service advertisement**, which initializes the `peer_acceptor_` endpoint and announces the availability of the service to interested peers.
2. **Endpoint listening**, which waits passively for peers to actively initiate connections on the `peer_acceptor_` endpoint.
3. **Service handler creation**, which creates and initializes a concrete `Svc_Handler` that can communicate with the new peer.
4. **Passive connection establishment**, which uses the `peer_acceptor_` endpoint to accept a connection initiated actively by a peer.
5. **Service handler concurrency activation**, which determines the type of concurrency mechanism used to process data exchanged with peers.
The `Acceptor`’s `open` method is responsible for handling the first two dimensions. The `Acceptor`’s `accept` method is responsible for handling the remaining three dimensions.
The following interface illustrates the methods and data members in the `Acceptor` class:
```cpp
template <class SVC_HANDLER,
// Type of service handler.
class PEER_ACCEPTOR>
// Type of passive connection mechanism.
class Acceptor {
public:
// Defines the initialization strategies.
typedef Strategy_Factory<SVC_HANDLER,
PEER_ACCEPTOR>
STRATEGY_FACTORY;
// Initialize listener endpoint at <addr>
// according to specified <init_strategies>.
virtual void open
(const PEER_ACCEPTOR::PEER_ADDR &addr,
STRATEGY_FACTORY *init_strategies);
// Embodies the strategies for creating,
// connecting, and activating <SVC_HANDLER>’s.
virtual void accept (void);
protected:
// Defines strategy to advertise endpoint.
virtual void advertise_svc
(const PEER_ACCEPTOR::PEER_ADDR &);
```
The **Acceptor** is a C++ template that is parameterized by a particular type of **PEER_ACCEPTOR** and **SVC_HANDLER**. The **PEER_ACCEPTOR** is the type of transport mechanism used by the **Acceptor** to passively establish the connection. The **SVC_HANDLER** is the type of service that processes the data exchanged with its connected peer. Parameterized types are used to efficiently decouple the service initialization strategies from the type of **Svc_Handler**, network programming interface, and transport layer connection protocol. This design improves the extensibility of the **Acceptor** and **Svc_Handler** components by allowing the wholesale replacement of various strategies.
Figure 4 visually depicts the relationship between the classes that comprise the **Acceptor** implementation. The five strategies supported by the **Acceptor** to passively initialize **Svc_Handler**s are illustrated and described below.
1. **Service advertisement strategies**: The **Acceptor** uses its service advertisement strategy to initialize the **PEER_ACCEPTOR** endpoint and to announce the availability of the service to interested peers. Figure 5 illustrates the common strategies configured into the **Acceptor** to advertise services:
- **Well-known addresses**, such as Internet port numbers and host names;
- **Endpoint portmappers**, such as those used by Sun RPC and DCE;
- **X.500 directory service**, which is an ISO OSI standard for mapping names to values in a distributed system.
2. **Endpoint listener strategies**: The **Acceptor** uses its endpoint listening strategy to wait passively for peers to actively initiate a connection to the **PEER_ACCEPTOR** endpoint. Figure 6 illustrates the following common strategies configured into the **Acceptor** to wait for connections:
• **Reactive listeners**, which use an event-demultiplexer, such as a Reactor [2], to listen passively on a set of endpoints in a single thread of control;
• **Threaded listeners**, which use a separate thread of control for each listener.
3. **Service handler creation strategies:** The Acceptor uses its creation strategy to initialize a Svc_Handler that will communicate with the new peer. Figure 7 illustrates the following common strategies configured into the Acceptor to create Svc_Handlers:
- **Demand creation**, which allocates a new Svc_Handler for every new connection;
- **Singleton creation**, which only creates a single Svc_Handler that is recycled for every connection;
- **Dynamic creation**, which does not store the Svc_Handler object in the application process until it is required, at which point the object is dynamically linked into the process from a shared library.
4. **Passive connection establishment strategies:** The Acceptor uses its passive connection establishment strategy to accept a new connection initiated actively by a peer. Figure 8 illustrates the following common strategies configured into the Acceptor to accept connections from peers:
- **Connection-oriented (CONS) establishment**, which uses connection-oriented protocols, such as TCP, SPX, or TP4;
- **Connectionless (CLNS) establishment**, which uses the Adapter pattern [1] to utilize a uniform interface for connectionless protocols.
5. **Service handler concurrency activation strategies:** The Acceptor uses its activation strategy to determine the type of concurrency mechanism a Svc_Handler will use to process data exchanged with its peer. Figure 9 illustrates the following common strategies configured into the Acceptor to activate Svc_Handlers:
- **Reactive activation**, where all Svc_Handlers execute within a single thread of control by using the Reactor pattern [2];
- **Thread activation**, where each Svc_Handler executes within its own separate thread;
- **Thread pool activation**, where each Svc_Handler executes within a pool of threads to increase performance on multi-processors;
- **Process activation**, where each Svc_Handler executes within a separate process.
The next section illustrates how different Acceptors can be configured flexibly to support alternative strategies without requiring changes to its external interface design or internal implementation.
4.3 Using Design Patterns to Implement an Extensible Acceptor
The ACE implementation of the Acceptor-Connector pattern applies the Wrapper Facade [2], Factory Method, Strategy, Bridge, and Abstract Factory patterns described in [1]. These patterns facilitate the flexible and extensible configuration and use of the initialization strategies discussed above. Below, each pattern used in the ACE Acceptor is described, the design forces it resolves are outlined, and an example of how the pattern is used to implement the Acceptor is presented.
Using the Wrapper Facades Pattern: The Wrapper Facade [2] pattern encapsulates the functions and data provided by existing non-OO APIs within more concise, robust, portable, maintainable, and cohesive OO class interfaces. The ACE Acceptor uses the Wrapper Facade pattern to provide a uniform interface that encapsulates differences between non-uniform network programming mechanisms, such as sockets, TLI, named pipes, and STREAM pipes.
Figure 10 illustrates how the ACE Acceptor uses the Wrapper Facade patterns to enhance its portability across platforms that contain different network programming interfaces, such as sockets but not TLI, or vice versa. In this example, the PEER_STREAM template argument of the Svc_Handler class can be instantiated with either a SOCK_Stream or a TLI_Stream, depending on whether the platform supports sockets or TLI. The Wrapper Facade pattern ensures that these two classes can be used identically by different instantiations of the Svc_Handler class.
Using the Strategy Pattern: The Strategy pattern [1] defines a family of algorithms, encapsulates each one as an object, and makes them interchangeable. The ACE Acceptor uses this pattern to determine the passive initialization strategies used to create, accept, and execute a Svc_Handler. By using the Strategy pattern, an application can configure different initialization strategies without modifying the following algorithm used by accept, as follows:
```cpp
template <class SVC_HANDLER, class PEER_ACCEPTOR> void
Acceptor<SVC_HANDLER, PEER_ACCEPTOR>::accept (void)
{
// Create a new <SVC_HANDLER>.
SVC_HANDLER *svc_handler =
make_svc_handler ();
// Accept connection from the peer.
accept_svc_handler (svc_handler);
// Activate <SVC_HANDLER>.
activate_svc_handler (svc_handler);
}
```
Figure 11 illustrates how the Strategy pattern is used to implement the Acceptor’s concurrency strategy. When the Acceptor is initialized, its Strategy_Factory configures the designated concurrency strategy. As shown in Figure 9, there are a number of alternative strategies. The particularly strategy illustrated in Figure 11 activates each Svc_Handler to run in a separate thread of control. Since all concurrency algorithms are encapsulated in a uniform interface, however, it is easy to replace this strategy with an alternative one, such as running the Svc_Handler in a separate process.
Using the Bridge Pattern: The Bridge pattern [1] decouples an abstraction from its implementation so that the two can vary independently. The ACE Acceptor uses this pattern to provide a stable, uniform interface that is both open
(i.e., extensible) and closed (i.e., does not require direct code changes).
Figure 12 illustrates how the Bridge pattern is used to implement the Acceptor’s connection acceptance strategy.
**Figure 12: Using the Bridge Pattern**
(the Bridge pattern is used for all the other Acceptor strategies, as well). When a connection is established with a peer, the Acceptor’s accept method invokes the accept_svc_handler method. Instead of performing the passive connection acceptance strategy directly, however, this method forwards the method to the appropriate subclass of Accept_Strategy. In the example shown in Figure 12, this subclass establishes the connection using a connection-oriented protocol. Since the Bridge pattern is used, however, an application can change the Acceptor’s connection acceptance strategy to an alternative strategy. For example, it can change to the connectionless version shown in Figure 8 without requiring any changes to the code in accept.
Another advantage of using the Bridge pattern is that a subclass of the Acceptor can override its make_* methods to avoid the additional overhead of indirection through strategy objects on every call. In this case, the accept method uses the Template Method pattern [1]. In the Template Method version of accept the steps in the Acceptor’s passive initialization algorithm are fixed, but can be overridden by derived classes.
**Using the Factory Method Pattern:** The Factory Method pattern [1] defines a stable interface for initializing a component, but allows subclasses to specify the details of the initialization. The ACE Acceptor uses this pattern to allow each initialization strategy used by the Acceptor to be extended without modifying the Acceptor or Svc_Handler implementations.
Figure 13 illustrates how the Factory Method pattern is used to transparently extend the Acceptor’s creation strategy. The Creation_Strategy base class contains a factory method called make_svc_handler. This method is invoked by the make_svc_handler Bridge method in the Acceptor to create the appropriate type of concrete Svc_Handler, as follows:
```cpp
template <class SVC_HANDLER, class PEER_ACCEPTOR> void
Acceptor<SVC_HANDLER, PEER_ACCEPTOR>::accept (void)
{
creation_strategy_ ->make_svc_handler ();
}
```
An implementation of a creation strategy based on the demand strategy could be implemented as follows:
```cpp
template <class SVC_HANDLER> SVC_HANDLER *
Demand_Strategy<SVC_HANDLER>::make_svc_handler (void) {
// Implement the 'demand' creation
// strategy by allocating a new <SVC_HANDLER>.
return new SVC_HANDLER;
}
```
Note that it is the responsibility of the Acceptor’s Strategy_Factory to determine the type of subclass associated with the creation_strategy_.
**Using the Abstract Factory:** The Abstract Factory pattern [1] provides a single interface that creates families of related objects without requiring the specification of their concrete classes. The Acceptor uses this pattern to simplify its interface by localizing all five of its initialization strategies into a single class. The Abstract Factory pattern also ensures that all selected strategies can work together correctly.
Figure 14 illustrates how the Abstract Factory pattern is used to implement the Status_Acceptor taken from the Gateway example describe in Section 5. This example instantiates the following Strategy_Factory template:
```cpp
template <class SVC_HANDLER,
// Type of service handler.
class PEER_ACCEPTOR>
// Type of passive connection.
```
Figure 14 shows the creation and concurrency strategies—the other strategies are handled similarly. The Status_Strategies factory instructs the Status_Acceptor to dynamically create each Status_Router, which will execute in its own thread of control. This example illustrates the following points:
- The Abstract Factory pattern is often used in conjunction with the Factory Method pattern. For example, the Strategy_Factory abstract factory simplifies the interface to the Acceptor by consolidating all five initialization strategy factory methods in a single class.
- The Abstract Factory pattern ensures that various strategies can work together correctly. For instance, the Strategy_Factory can be subclassed and its various make_* Factory Methods can be overridden to create different types of initialization strategies.
- Subclasses of the Strategy_Factory abstract factory can be used to ensure that conflicting initialization strategies are not configured accidentally. For example, the singleton creation strategy may conflict with the thread concurrency strategy since multiple threads of control will attempt to access a single communication endpoint. A Strategy_Factory subclass can be defined to check for these conflicts and report an error at configuration time.
5 Example: Implementing Extensible Application-level Gateways Using the Acceptor
This section illustrates how the application-level Gateway described in Section 2 uses the pattern-based Acceptor component from Section 4 to simplify the task of passively initializing services whose connections are initiated actively by Peers. In this example the Peers play the active role in establishing connections with the Gateway.
Defining Svc_Handlers for routing peer messages: The three classes shown below, Status_Router, Bulk_Data_Router, and Command_Router, process routing messages received from Peers. These classes inherit from Svc_Handler, which allows them to be passively initialized by an Acceptor, as shown in Figure 15. Each class is instantiated with a specific type of C++ IPC wrapper facade that exchanges data with its connected peer. For
example, the classes below use a SOCK_Stream as the underlying data transport delivery mechanism. SOCK_Stream is an ACE C++ wrapper facade that encapsulates the data transfer functions in the socket interface. By virtue of the Strategy pattern, however, it easy to vary the data transfer mechanism by parameterizing the Svc_Handler with a different PEER_STREAM, such as a TLI_Stream.
The Status_Router class routes status data sent to and received from Peers.\footnote{To save space, these examples have been simplified by omitting most of the detailed protocol logic and error handling code.}
```cpp
class Status_Router :
public Svc_Handler<SOCK_Stream>
{
public:
// Performs router initialization.
virtual int open (void);
// Receive and route status data from/to peers.
virtual int handle_event (void);
// ...
};
```
The Bulk_Data_Router class routes bulk data sent to and received from Peers.
```cpp
class Bulk_Data_Router :
public Svc_Handler<SOCK_Stream>
{
public:
// Performs router initialization.
virtual int open (void);
// Receive and route bulk data from/to peers.
virtual int handle_event (void);
// ...
};
```
The Command_Router class routes bulk data sent to and received from Peers.
```cpp
class Command_Router :
public Svc_Handler<SOCK_Stream>
{
public:
// Performs router initialization.
virtual int open (void);
// Receive and route command data from/to peers.
virtual int handle_event (void);
// ...
};
```
Defining Acceptor factories to create Svc_Handlers: The three classes shown below are instantiations of the Acceptor template:
```cpp
// Typedefs that instantiate <Acceptor>s for
// different types of routers.
typedef Acceptor<Status_Router, SOCK_Acceptor> Status_Acceptor;
typedef Acceptor<Bulk_Data_Router, SOCK_Acceptor> Bulk_Data_Acceptor
typedef Acceptor<Command_Router, SOCK_Acceptor> Command_Acceptor;
```
These typedefs instantiate the Acceptor template with concrete parameterized type arguments for SVC_HANDLER and PEER_ACCEPTOR. A SOCK_Acceptor wrapper facade is used as the underlying PEER_ACCEPTOR in order to accept a connection passively. Parameterizing the Acceptor with a different PEER_ACCEPTOR, such as a TLI_Acceptor, is easy since the IPC mechanisms are encapsulated in C++ wrapper facade classes. The three objects shown below are instances of these classes that create and activate Status_Routers, Bulk_Data_Routers, and Command_Routers, respectively:
```cpp
// Accept connection requests from
// Gateway and activate Status_Router.
static Status_Acceptor status_acceptor;
// Accept connection requests from
// Gateway and activate Bulk_Data_Router.
static Bulk_Data_Acceptor bulk_data_acceptor;
// Accept connection requests from
// Gateway and activate Command_Router.
static Command_Acceptor command_acceptor;
```
Defining strategies to initialize Svc_Handlers: The three classes shown below are instantiations of the Strategy_Factory described in Section 4.2:
```cpp
// Typedefs that instantiate different types
// of <Strategy_Factory>.
typedef Strategy_Factory<Status_Router,
SOCK_Acceptor>
Status_Strategies
typedef Strategy_Factory<Bulk_Data_Router,
SOCK_Acceptor>
Bulk_Data_Strategies;
typedef Strategy_Factory<Command_Router,
SOCK_Acceptor>
Command_Strategies;
```
These typedefs instantiate the Strategy_Factory template with concrete parameterized type arguments for SVC_HANDLER and PEER_ACCEPTOR. The three objects shown below instantiate these classes to specify the initialization strategies for Status_Routers, Bulk_Data_Routers, and Command_Routers, respectively:
```cpp
// Creates a multi-threaded <Status_Router>.
Status_Strategies threaded
(new Well_Known_Addr,
new Reactive_Listener (Reactor::instance ()),
new Demand,
new CONS,
new Multi_Thread);
// Creates a multi-processed <Bulk_Data_Router>.
Bulk_Data_Strategies process
(new Well_Known_Addr,
new Reactive_Listener (Reactor::instance ()),
new Demand,
new CONS,
new Multi_Process);
```
// Creates a single-thread reactive <Command_Router>.
Command_Strategies reactive
(new Well_Known_Addr,
new Reactive_Listener (Reactor::instance ()),
new Demand,
new CONS,
new Reactive (Reactor::instance ()));
Each Strategy_Factory configuration shown above uses the well known address service advertisement strategy, the reactive listener strategy, the demand Svc_Handler creation strategy, and the connection-oriented acceptance strategy. To illustrate the flexibility of the Acceptor-Connector pattern, however, each Strategy_Factory implements a different concurrency strategy, as follows:
- When the Status_Acceptor is activated by Status_Router it runs in a separate thread.
- When activated by Bulk_Data_Acceptor, the Bulk_Data_Router runs as a separate process.
- When activated by Command_Acceptor, the Command_Router runs in the same thread as with the Reactor singleton [1], which is used to demultiplex connection requests for the three Acceptor factories.
Note how changing the concurrency strategy does not affect the Acceptor class. Thus, the Acceptor’s generic strategy for passively initializing services can be reused, while permitting specific details, such as the PEER_ACCEPTOR, SVC_HANDLER, and selected initialization strategies, to change flexibly.
**The main() gateway function:** The main gateway initializes the Acceptors with their well-known ports and initialization strategies, as follows:
```cpp
// Main program for the Gateway.
int main (void) {
// Initialize Acceptors with their well-known
// ports and their initialization strategies.
status_acceptor.open
(INET_Addr (STATUS_PORT), &threaded);
bulk_data_acceptor.open
(INET_Addr (BULK_DATA_PORT), &process);
command_acceptor.open
(INET_Addr (COMMAND_PORT), &reactive);
// Loop forever handling connection request
// events and processing data from peers.
for (;;)
Reactor::instance ()->handle_events ();
}
```
The listener strategy configured for each Acceptor is reactive, as shown in Section 5. Therefore, the program enters an event loop that uses the Reactor singleton to detect all connection requests from Peers within a single thread of control. When connections arrive, the Reactor singleton dispatches the associated Acceptor, which (1) creates an appropriate type of Svc_Handler on demand to perform the service, (2) accepts the connection into the handler, and (3) activates the handler. The concurrency strategy configured into each Acceptor dictates how every Svc_Handler it creates will processes events.
Figure 16 illustrates the relationship between Acceptor-Connector pattern components in the Gateway after four connections have been established. The various *Routers exchange data with their connected Peers using the type of concurrency strategy designated by their Strategy_Factories. Meanwhile, the *Acceptors continue to listen for new connections.
### 5.1 Known Uses
**UNIX network superservers:** Superserver implementations such as Inetd [5], Listen [6] and the Service Configurator [7] from the ACE framework use a master acceptor process that listens for connections on a set of communication ports. In Inetd, for example, each port is associated with a service, such as the standard Internet services FTP, TELNET, DAYTIME, and ECHO. The acceptor process decouples the functionality of the Inetd superserver into two separate parts: one for establishing connections and another for receiving and processing requests from peers. When a service request arrives on a port monitored by Inetd, it accepts the request and dispatches an appropriate pre-registered handler to perform the service.
**CORBA Object Request Brokers (ORBs):** The ORB Core layer in many implementations of CORBA [8] uses the Acceptor-Connector pattern to passively and actively initialize connection handlers when clients request ORB services. For example, [9] describes how the Acceptor-Connector pattern is used to implement the ORB Core portion in The ACE ORB (TAO), which is a high-performance and real-time implementation of CORBA.
Web Browsers: The HTML parsing components in Web browsers such as Netscape and Internet Explorer use the asynchronous version of the connector component to establish connections with servers associated with images embedded in HTML pages. This pattern allows multiple HTTP connections to be initiated asynchronously. This avoids the possibility of the browser’s main event loop blocking.
Ericsson EOS Call Center Management System: This system uses the Acceptor-Connector pattern to allow application-level Call Center Manager event servers [10] to establish connections actively with passive supervisors in a networked center management system.
Project Spectrum: The high-speed medical image transfer subsystem of project Spectrum [11] uses the Acceptor-Connector pattern to establish connections passively and initialize application services for storing large medical images. Once connections are established, applications send and receive multi-megabyte medical images to and from the image stores.
ACE: Implementations of the generic Svc_Handler, Connector, and Acceptor components described in the Implementation section are provided as reusable C++ classes in the ACE framework [3]. Java ACE [12] is a version of ACE implemented in Java that provides components corresponding to the participants the Acceptor-Connector pattern.
6 Related Patterns
[1, 13, 2] identify and catalog many architectural and design patterns. This section examines how the patterns described in this paper relate to other patterns in the literature.
The intent of the Acceptor-Connector pattern is similar to the Configuration pattern [14]. The Configuration pattern decouples structural issues related to configuring services in distributed applications from the execution of the services themselves. This pattern has been used in frameworks for configuring distributed systems, such as Regis [15], to support the construction of a distributed system from a set of components. In a similar way, the Acceptor-Connector pattern decouples service initialization from service processing. The primary difference is that the Configuration pattern focuses more on the active composition of a chain of related services, whereas the Acceptor-Connector pattern focuses on the passive initialization of a service handler at a particular endpoint. In addition, the Acceptor-Connector pattern also focuses on decoupling service behavior from the service’s concurrency strategies.
The intent of the Acceptor-Connector pattern is similar to that of the Client-Dispatcher-Server pattern [13] in that both are concerned with the separation of active connection establishment from subsequent service processing. The primary difference is that the Acceptor-Connector pattern addresses passive and active connection establishment and initialization of both synchronous and asynchronous connections. In contrast, the Client-Dispatcher-Server pattern focuses on synchronous connection establishment.
The service handlers that are created by acceptors and connectors can be coordinated using the Abstract Session pattern [16], which allows a server object to maintain state for many clients. Likewise, the Half Object plus Protocol pattern [17] can help decompose the responsibilities of an end-to-end service into service handler interfaces and the protocol used to collaborate between them.
The Acceptor-Connector pattern may be viewed as an object creational pattern [1]. A creational pattern assembles the resources necessary to create an object and decouples the creation and initialization of the object from subsequent use of the object. The Acceptor-Connector pattern is a factory that creates, passively connects, and initializes service handlers. Its accept method implements the algorithm that listens passively for connection requests, then creates, accepts, and activates a handler when the connection is established. The handler performs a service using data exchanged on the connection. Thus, the subsequent behavior of the service is decoupled from its initialization strategies.
7 Concluding Remarks
This paper describes the Acceptor-Connector pattern and illustrates how its Acceptor component has been implemented using other patterns to develop highly flexible communication software. In general, the Acceptor-Connector pattern is applicable whenever connection-oriented applications have the following characteristics:
- The behavior of a distributed service does not depend on the steps required to passively or actively connect and initialize a service.
- Connection requests from different peers may arrive concurrently, but blocking or continuous polling for incoming connections on any individual peer is inefficient.
The Acceptor-Connector pattern provides the following benefits for network applications and services:
It enhances the reusability, portability, and extensibility of connection-oriented software: The Acceptor-Connector pattern decouples mechanisms for connection establishment and service initialization, which are application-independent and thus reusable, from the services themselves, which
are application-specific. For example, the application-independent mechanisms in the `Acceptor` are reusable components that know how to establish a connection passively and to create and activate its associated `Svc_Handler`. In contrast, the `Svc_Handler` knows how to perform application-specific service processing.
This separation of concerns decouples connection establishment from service handling, thereby allowing each part to evolve independently. The strategy for establishing connections actively was written once, placed into the ACE framework, and reused via inheritance, object composition, and template instantiation. Thus, the same passive connection establishment code need not be rewritten for each application. In contrast, services may vary according to different application requirements. By parameterizing the `Acceptor` with a `Svc_Handler`, the impact of this variation is localized to a single point in the software.
**Improves application robustness:** By strongly decoupling the `Acceptor` from the `Svc_Handler` the passive-mode `PEER_ACCEPTOR` cannot accidentally be used to read or write data. This eliminates a class of subtle errors that can arise when programming with weakly typed network programming interfaces such as sockets or TLI.
However, the Acceptor-Connector pattern can also exhibit the following drawbacks:
**Additional indirection:** The Acceptor-Connector pattern can incur additional indirection compared to using the underlying network programming interfaces directly. However, languages that support parameterized types, such as C++, Ada or Eiffel, can implement these patterns with no significant overhead when compilers inline the method calls used to implement the patterns.
**Additional complexity:** The Acceptor-Connector pattern may add unnecessary complexity for simple client applications that connect with only one server and perform one service using a single network programming interface. However, the use of generic acceptor and connector wrapper facades may simplify even these use cases by shielding developers from tedious, error-prone and non-portable low-level network programming mechanisms.
Open-source implementations of the Acceptor-Connector and Reactor patterns are available at URL `www.cs.wustl.edu/~schmidt/ACE.html`. This URL contains complete source code, documentation, and example test drivers for the C++ components developed as part of the ACE framework [3] developed at the University of California, Irvine and Washington University, St. Louis.
**References**
[1] E. Gamma, R. Helm, R. Johnson, and J. Vlissides, *Design Patterns: Elements of Reusable Object-Oriented Software*. Reading, MA: Addison-Wesley, 1995.
[2] D. C. Schmidt, M. Stal, H. Rohnert, and F. Buschmann, *Pattern-Oriented Software Architecture: Patterns for Concurrency and Distributed Objects, Volume 2*. New York, NY: Wiley & Sons, 2000.
[3] D. C. Schmidt, “Applying Design Patterns and Frameworks to Develop Object-Oriented Communication Software,” in *Handbook of Programming Languages* (P. Salus, ed.), MacMillan Computer Publishing, 1997.
[4] D. C. Schmidt, “Applying a Pattern Language to Develop Application-level Gateways,” in *Design Patterns in Communications* (L. Rising, ed.), Cambridge University Press, 2000.
[5] W. R. Stevens, *UNIX Network Programming, First Edition*. Englewood Cliffs, NJ: Prentice Hall, 1990.
[6] S. Rago, *UNIX System V Network Programming*. Reading, MA: Addison-Wesley, 1993.
[7] P. Jain and D. C. Schmidt, “Service Configurator: A Pattern for Dynamic Configuration of Services,” in *Proceedings of the 3rd Conference on Object-Oriented Technologies and Systems*, USENIX, June 1997.
[8] Object Management Group, *The Common Object Request Broker: Architecture and Specification*, 2.3 ed., June 1999.
[9] D. C. Schmidt and C. Cleeland, “Applying a Pattern Language to Develop Extensible ORB Middleware,” in *Design Patterns in Communications* (L. Rising, ed.), Cambridge University Press, 2000.
[10] D. C. Schmidt and T. Suda, “An Object-Oriented Framework for Dynamically Configuring Extensible Distributed Communication Systems,” *IEEE/BCS Distributed Systems Engineering Journal (Special Issue on Configurable Distributed Systems)*, vol. 2, pp. 280–293, December 1994.
[11] G. Blaine, M. Boyd, and S. Crider, “Project Spectrum: Scalable Bandwidth for the BJC Health System;” *HIMSS, Health Care Communications*, pp. 71–81, 1994.
[12] P. Jain and D. Schmidt, “Experiences Converting a C++ Communication Software Framework to Java,” *C++ Report*, vol. 9, January 1997.
[13] F. Buschmann, R. Meunier, H. Rohnert, P. Sommerlad, and M. Stal, *Pattern-Oriented Software Architecture - A System of Patterns*. Wiley and Sons, 1996.
[14] S. Crane, J. Magee, and N. Pryce, “Design Patterns for Binding in Distributed Systems,” in *The OOPSLA ’95 Workshop on Design Patterns for Concurrent, Parallel, and Distributed Object-Oriented Systems*, (Austin, TX), ACM, Oct. 1995.
[15] J. Magee, N. Dulay, and J. Kramer, “A Constructive Development Environment for Parallel and Distributed Programs,” in *Proceedings of the 2nd International Workshop on Configurable Distributed Systems*, (Pittsburgh, PA), pp. 1–14, IEEE, Mar. 1994.
[16] N. Pryce, “Abstract Session,” in *Pattern Languages of Program Design* (B. Foote, N. Harrison, and H. Rohnert, eds.), Reading, MA: Addison-Wesley, 1999.
[17] G. Meszaros, “Half Object plus Protocol,” in *Pattern Languages of Program Design* (J. O. Coplien and D. C. Schmidt, eds.), Reading, MA: Addison-Wesley, 1995. |
TULLIP
A Projected Thesaurus Universalis Libri Lithographici Illustrati Persorum
Ulrich Marzolph
Göttingen
Considering the large amount of illustrative data available from the Islamic world, the need for adequate tools for comparative research appears to be compelling. In reality, however, the vast majority of studies in Islamic art deal with specific items or topics, such as particular artists, individual manuscripts, or specific scenes. The lack of a general and comprehensive survey of Islamic iconography\(^1\) is particularly significant for non-specialists dealing with Islamic art, such as historians of literature or folklorists. While these specialists by virtue of their professional interest demonstrate a distinct expertise for the textual backdrop of a given illustration,\(^2\) the illustration’s iconographical significance in a wider context would be beyond their apprehension. In consequence, specialists in literature often remain ignorant of significant iconographic aspects, while historians of Islamic art rarely display a profound knowledge of textual aspects relating to the illustrative material they deal with. The projected ‘Union Catalogue of Illustrated Persian Lithographed Books,’ *Thesaurus Universalis Libri Lithographici Illustrati Persorum* (TULLIP), aims to contribute to closing this gap between the disciplines.
At present, TULLIP is no more than a bulb waiting for adequate conditions to sprout and develop into a full-fledged project. While these conditions are being explored, I would like to profit from the opportunity to present TULLIP’s general agenda for further discussion. TULLIP constitutes a follow-up to my earlier study
---
1 Highly useful surveys of specific topics include M.V. Fontana, *Iconografia dell’Ahl al-Bayt: Immagini di arte persiana dal XII al XX secolo* (Naples 1994); J. Gierlisch, *Mittelalterliche Tierreliefs in Anatolien und Nordmesopotamien: Untersuchungen zur figürlichen Baudekoration der Seldschuken, Artuqiden und ihrer Nachfolger bis ins 15. Jahrhundert* (Tübingen 1996), and, most recently, E. Sims, B.I. Marshak and E.J. Grube, *Peerless Images: Persian Painting and Its Sources* (New Haven, London 2002). The latter work contains a detailed presentation of ‘the primary themes of Iranian imagery’ (91–330).
2 See, e.g., O. Grabar and C. Robinson (eds), *Islamic Art and Literature* (Princeton 2001).
Narrative Illustration in Persian Lithographed Books.\textsuperscript{3} While that study serves as an introduction to a previously unexplored field, it remains my conviction that the genre of illustrated Persian lithographed books is particularly suited for a comprehensive survey of iconographical features: the genre is fairly limited and, hence, surveyable; its imagery is not as complex as that of Persian miniature painting and, hence, more readily available for deciphering and categorizing; large amounts of data are more easily accessible than for any other category of Islamic art; and, finally, the natural way in which text and illustration are combined allows communication between the various disciplines concerned. The present sketch of TULLIP will outline the following points:
The subject: What exactly are illustrated Persian lithographed books?
The scope: Why a \textit{thesaurus universalis}, an international union catalogue of these items?
The bibliographical component: Where and how to acquire the relevant data?
The methodological component: What should TULLIP be expected to document?
The result: How should TULLIP be developed in order to offer optimum conditions for its users?
\section*{The Subject}
What exactly are illustrated Persian lithographed books?\textsuperscript{4} The history of the printed book in Iran is a comparatively recent one.\textsuperscript{5} Besides short-lived experiments in the pre-modern era, the art of printing as a continuous practice was introduced to Iran only in the Qajar period.
\begin{itemize}
\item The first ever book printed in Iran was produced in movable type. It is a collection of \textit{fatwas} relating to the Russian-Persian war, published in 1233/1817.
\item The first book printed in Iran by way of the lithographic technique\textsuperscript{6} is a copy of the Koran published in Tabriz, dated either 1248/1832, or 1250/1834.
\item The first illustrated lithographic book produced in Iran is a copy of Maktabi’s \textit{Leili va Majnun} published in 1259/1843. Illustrating lithographic books became a current practice in Iran as of 1263/1847.
\item Shortly after the year 1270/1854, printing in movable type ceased altogether. For about two decades, all books published in Iran were produced by way of lithographic printing.
\item Even though printing in movable type was taken up again in 1290/1874, lithographed books continued to be produced in large numbers. Gradually, printing in movable type won the upper hand, and lithographic production started to dwindle. The last items of the genre were published around the middle of the twentieth century.
\end{itemize}
TULLIP is to consider books in the Persian language printed in Iran by way of lithography roughly in the century between 1850 and 1950. One reason for the lasting prevalence of lithography in Iran, besides its aesthetic attraction, was its capacity to combine text and illustration in a comparatively inexpensive way. In consequence, hundreds of illustrated lithographed books were produced in Iran, above all in the areas of classical, religious, and narrative literature. At a later stage, and increasingly towards the latter half of the Qajar period, other genres of lithographed books also contain illustrations, such as travel literature, works of history, medicine, chemistry, astronomy, religious dogmatics or military education, besides translations of European novels and the emerging genre of the Persian schoolbook.\textsuperscript{7} Eventually, the illustrations in all of these books will have to be taken into account. As a starting
\begin{footnotes}
\item U. Marzolph, \textit{Narrative Illustration in Persian Lithographed Books} (Leiden 2001).
\item For a detailed discussion of the subject see U. Marzolph, ‘Zur frühen Druckgeschichte in Iran (1817–c.1900). 1: Gedruckte Handschrift/Early Printing History in Iran (1817– c. 1900). 1: Printed Manuscript.’ in E. Hanebutt-Benz, D. Glass, and G. Roper (eds), \textit{Sprachen des Nahen Ostens und die Druckrevolution: Eine interkulturelle Begegnung/Middle Eastern Languages and the Print Revolution: A Cross-Cultural Encounter} (Westhofen 2002), 249–68 (text), 271–2 (footnotes), 511–17 (description of exhibits), 538–9 (bibliography), and plates 112–21.
\item The most comprehensive survey of this field remains O. P. Scheglova, \textit{Iranskaya litografirovannaya kniga} (Moscow 1979); see also Sh. Bābāzādē, \textit{Tārikh-e chāp dar Irān} (Teheran 1378/1999); Marzolph, \textit{Narrative Illustration}, 12–21.
\item A. Senefelder, \textit{The Invention of Lithography}, translated from the original German by J.W. Muller (Pittsburgh 1998); W. Weber, \textit{A History of Lithography} (New York, Toronto, London 1966); D. Porzio (ed.), \textit{Lithography: 200 Years of Art, History & Technique}, translated from the Italian by G. Culverwell (Secaucus, NJ 1982).
\item For illustrated schoolbooks see particularly M.H. Mohammadi and Z. Qā’eni, \textit{Tārikh-e adabiyāt-e kudakān-e Irān}, vols. 1–4 (Teheran 1380/2001).
\end{footnotes}
point, research will be restricted to the archive of more than 10,000 individual illustrations that has served as the basis for my monograph study.\(^8\)
The Scope
The most important rationale for proposing a thesaurus universalis, an international union catalogue, derives from the lack of reliable bibliographical documentation.\(^9\) In addition, it relates to the fact that no single library or national conglomerate of libraries holds sufficient quantities of relevant books. Libraries in Iran at first sight would appear to contain promising holdings, yet at second sight prove to be highly problematic. Various reasons account for this evaluation. First, books in Iran were meant to be used. In consequence, many of the books preserved in Iranian libraries are torn and tattered, often missing the first and last pages, and often being in a moderate state of preservation. Second, public libraries are a comparatively recent phenomenon in Iran. Still today, some of the best stocked libraries are held by individuals who eventually might or might not donate their possession (as vaqf) to a public or semi-public institution. Hence, neither the Iranian National Library nor any other institution in Iran can be expected to hold but a fragment of the total number of books actually produced. Third, in Iran the awareness of the cultural importance of Qajar book production has only been growing over the past two decades. Up until the restrictions imposed after the revolution of 1978–9, Qajar books were sold and exported in considerable numbers. Meanwhile, most public libraries have stacked their holdings of historical book production in separate sections with limited access. A research trip in the summer of 2002 to important libraries in the five major Iranian cities of Teheran, Shiraz, Esfahan, Mashhad, and Tabriz showed that most libraries pragmatically regard the date 1320, corresponding more or less to 1900, as dividing historical from modern book production. Some of the larger institutions, such as the Teheran National Library and the Library of Congress or the Mashhad Central Library of the Sanctuary have made the bibliographical documentation of their complete holdings, including historical books, accessible by way of online databases in Persian.
In contrast, many of the provincial libraries with limited resources, while shelving their historical books separately, have not even been able to start proper cataloguing.
Outside Iran, a number of major libraries in European and North American cities are known to possess substantial collections of Persian lithographed books rivalling Iranian holdings in respect to both rarity and quality of preservation. No survey of ‘the Persian book’ can afford to disregard these holdings, above all those in Saint Petersburg,\(^{10}\) London,\(^{11}\) Cambridge, Paris, and Berlin. Accordingly, my previous research has taken into consideration the holdings of some 35 libraries in 13 countries, situated above all in Asia and Europe. Future research ought to take into account the holdings of North American libraries (Harvard, Princeton, New York Public Library, University of California at Los Angeles, and others), in addition to those in the Indian subcontinent\(^{12}\) as well as in the Caucasian and Central Asian republics, such as Azerbaijan, Georgia, and Tajikistan. As a point of specific interest, the available documentation shows that European scholars often regarded Persian historical book production as collectables. Hence, some of the best preserved items are contained in private collections. It might not come as a surprise that several smaller libraries, such as those in the German cities of Frankfurt, Cologne,\(^{13}\) Halle, or Gotha, whose holdings partly derive from private collections, variously hold extremely rare items that would escape the researcher’s attention if those libraries were disregarded. Given this situation, TULLIP will have to rely on a truly international cooperation in order to arrive at a next to complete assessment of Persian lithographic book production.
The Bibliographical Component
In order to acquire the relevant data, the proposed theoretical agenda necessitates human and technical resources in dimensions difficult to realize. Human resources include the preparation of surveys relating to holdings in Iran and its neighbouring Asian countries, Europe, and North America. Particular attention will have to be given
---
\(^8\) For the archive’s scope see Marzolph, *Narrative Illustration*, 230–69.
\(^9\) The data on lithographed books in the *Fehrest-e ketâbhâ-ye châpi-ye fârsî az âghâz tâ âkhar-e sâl-e 1345 [1966], bar-asâs-e fehrest-e Khân-Bâbâ Mosbâr va fâhâres-e Anjoman-e ketâb*, ed. E. Yâr-Shâter, vols 1–3 (Teheran 1352/1973), even though the work remains an admirable and generally useful bibliographical tool, are spurious and unreliable.
\(^10\) O.P. Sheheglova, *Katalog litografirovannykh knig na persidskom yazyke v sobranii Leningradskogo otdeleniya Instituta vostokovedeniya AN SSR*, vols 1–2 (Moscow 1975); ed., *Katalog litografirovannykh knig na persidskom yazyke v sobranii vostochnogo otdela nauchnoi biblioteki im. A. M. Gorkogo Leningradskogo gosudarstvennogo universiteta* (Moscow 1989).
\(^11\) E. Edwards, *A Catalogue of the Persian Printed Books in the British Museum* (London 1922).
\(^12\) ‘Ā. Noushâhi, *Fehrest-e ketâbhâ-ye fârsî-ye châpi-e sangi va kam-yâb-e Ketâbkhâne-ye Ganjbakhsh* (Islamabad 1350/1971).
\(^13\) K. Amir-Arjomand, *Katalog der Bibliothek des schiitischen Schrifttums im Orientalischen Seminar der Universität zu Köln*, ed. A. Falaturi, vols 1–6 (Munich, New York, London, Paris 1996).
to private holdings in Iran, some of which are as rich in terms of holdings as they are difficult to assess. Preparing such a survey will have to rely on a team of researchers from various countries, each contributing with their national expertise. Confining any such survey to the major international libraries containing Persian holdings would be a fatal decision, as any library, and particularly the smaller ones, can reasonably be argued to hold unique items. Technical documentation of the holdings would best be achieved by way of microfilm, as this technique is available in the majority of libraries worldwide. It should be noted, however, that even the Teheran National Library acquired its microfilm equipment as recent as less than ten years ago, and the first items I received from that library in the early 1990s were still prepared as photocopies. Some libraries, particularly the Caucasian and Middle Asian ones, might not command the necessary technical equipment, hence requiring the availability of mobile equipment. Once the relevant items have been documented, they should be made accessible in international data formats, preferably those facilitating their compatibility in various kinds of presentation.
The Methodological Component
In the long run, TULLIP is expected to document a large number of Persian lithographed books containing illustrations of any kind. The data are to be collected in three stages: First, work will be confined to the presently available set of about 10,000 illustrations. Second, new data from Iran should be collected; and third, data from other Asian, European, and North American libraries are to be added. Special care in documentation is to be given to the relation of text and illustration, a relationship that has previously often been neglected. Even though the printed texts of the lithographic period show a lesser degree of variation than manuscript texts, a certain amount of variation cannot be excluded. This evaluation also applies to the position of illustrations within a given text. As illustrations in some of the larger sized books, such as the *Shāhnāme*, fill a whole page from top to bottom, documenting the related text will mean to include at least the preceding and following page or pages. Moreover, the identification of specific scenes will often not be possible without a detailed knowledge of the full text of a given book. As the majority of texts printed in lithographed editions in the Qajar period is not available in modern editions, let alone critical ones, the documentation should not only comprise a given book’s illustrations but rather its complete text. In general terms, the systematic documentation and analysis of the illustrations will have to comprise both their iconographical assessment in terms of generic description and detailed catch-word analysis as well as their relation to a specific passage of an individual text.
The Result
In order to offer optimum conditions for its users, it goes without saying that TULLIP is to be developed as a multi-dimensional database, preferably available by way of internet. In addition to basic bibliographical documentation and identification of the illustrations, special care is to be devoted to supplying structured information that in the long run might serve as a basic constituent for a general assessment of Islamic iconography. Various major published systems employed in international iconographical research might serve as a basis for the development of an individual structure of iconographical description, while taking into account the particular requirements of the Islamic data.
A major source of inspiration for the creation of the required database will be the ICONCLASS system as developed by Dutch art historian Henri van de Waal.\(^{14}\) ICONCLASS is an iconographic classification system, essentially devised to supply ready-made descriptions for works of art, whether two-dimensional illustrations or emblematic imagery or three-dimensional sculptures. ICONCLASS works with a strict hierarchical order, starting with ten major categories labelled with the numbers 0 thru 9: (0) Abstract, Non-representational Art; (1) Religion and Magic; (2) Nature; (3) Human Being, Man in General; (4) Society, Civilization, Culture; (5) Abstract Ideas and Concepts; (6) History; (7) Bible; (8) Literature; (9) Classical Mythology and Ancient History. Already this outline shows the system’s obvious rootedness in Western culture. Data from the Persian and Islamic context will rarely fit into categories (0) thru (6), virtually none into category (7), and the overwhelming majority into categories (8) and (9). This situation suggests that in order to use ICONCLASS for the classification of illustrations in Persian lithographed books one would have to expand the main categories of literature and mythology whilst
\(^{14}\) Besides the information available by way of internet (http://iconclass.let.ruu.nl) see, e.g., R. van Straten, *Iconography, Indexing, Iconclass: A Handbook* (Leiden 1994).
diminishing the less important ones. In the context of Qajar lithographed books, literature implies the three categories of classical literature, particularly mystical, religious, pedagogical or otherwise instructive literature; religious literature, above all the works relating to the genre of *maqātil* or *rouze-khvānī*, i.e. works treating the martyrdom of Hosein b. ‘Ali at Karbala that is of pivotal importance for the definition of the Shiite creed; and works of a narrative, fictional, or otherwise primarily entertaining character.
The second published system of iconographical identification to serve as a source of inspiration for TULLIP is François Garnier’s *Thesaurus iconographique*. This system has originally been developed with the practical purpose of cataloguing works of art in French museums, and has since been applied to tens of thousands of items. Garnier’s system is much more detailed and, apparently, less predisposed than ICONCLASS. Moreover, it allows classification on various levels that are classed in the two large groups of *thèmes* (‘general subjects’) and *sujets* (‘specific subjects’). Levels range from (1) general specifics of representation, (2) nature, (3) the human body and natural environment to (16) geographical (17) biblical, (18) mythological, (19) historical subjects. As a particularly interesting point, Garnier’s system also takes into account implicit aspects of representation, such as (21) *courant de pensée* (‘intellectual background’) and (22) *périodisation* (‘imagery depicting certain periods’).
Other systems whose experience might prove useful for developing TULLIP’s system of iconographical classification include the Princeton *Index of Christian Art* or the Marburger *Inventarisations-, Dokumentations- und Administrations-System (MIDAS)*. At any rate, whatever the source of inspiration, TULLIP will have to develop its own system specifically adjusted to its purpose. Beyond the factual description of a given illustration, additional data will have to position the illustrations both diachronically and synchronically, the latter category supplying links to identical illustrations both in editions of the same work of literature as well as in other works. In addition, a detailed analysis of components such as background, gestures, and material culture will enable a multi-dimensional comparison of these aspects across different works as well as editions of the same work.
The only internet presentation at least partly aiming at a similar goal known to me to date is the *Shahnama Project*, parts of which are available on an internet site published by Princeton University, while a major project funded by the Arts and Humanities Research Board of the British Academy under the supervision of Charles Melville (Cambridge) and Robert Hillenbrand (Edinburgh) is ongoing. The Princeton website contains the documentation of all illustrations in five *Shāhnāme* manuscripts located in Princeton libraries, plus a survey of the dispersed leaves of the so-called Great Mongol (formerly ‘Demotte’) *Shāhnāme*. Viewing includes thumbnail surveys, identification of scenes, and related text passages from the English translation (Warner). The complete list of illustrated episodes in the various *Shāhnāmes* worldwide, an amended version of the list originally compiled by Edward Davis and Jill Norgren, provides an extremely useful starting point for the general iconographical identification of *Shāhnāme* episodes in lithographed editions. Similar listings have been published for two other major works of Persian literature, the poet Nezāmi’s *Khamse* and the corpus of fables commonly known as *Kalila va Dimna*. Similar comparative corpora have already been prepared by the present author for other works frequently published in lithographed editions, such as the Persian translation of the *Arabian Nights* (*Hezār va yek shab*), the seventeenth century collection of proverbs, Mohammad-‘Ali Hablerudi’s *Jāme’ al-tamsil*, or the most popular work of *rouze-khvānī*-literature, Jouhari’s *Tufān al-bokāʾ*.
By surveying a limited range of illustrated books, TULLIP will succeed in supplying basic data for a general survey of Islamic iconography. As the analyzed genre is both highly specific and chronologically late, these data will have to be
---
15 F. Garnier, *Thesaurus iconographique: système descriptif de représentations* (Paris 1984).
16 L. Heusinger, *Marburger Informations-, Dokumentations- und Administrations-System (MIDAS)*, 4th edn (Marburg 2000).
17 http://humanitas.princeton.edu/shahnama/
18 http://www.oriental.cam.ac.uk/shah
19 O. Grabar and Sh. Blair, *Epic Images and Contemporary History: The Illustrations of the Great Mongol Shahnama* (Chicago, London 1980).
20 E. Davis and J. Norgren, *A Preliminary Index of Shahnameh Illustrations* (Ann Arbor 1969).
21 See U. Marzolph, ‘Illustrated Persian Lithographed Editions of the Shāhnāme,’ in *Edebiyāt* (in print); id., *Images of Power: Illustrated Persian Lithographed Editions of the Shāhnāme* (Teheran in print).
22 L.N. Dodkhudeeva, *Poemy Nezami v srednevekovoi minyaturnoi zhivopisi* (Moscow 1985).
23 E.J. Grube, ‘Prolegomena for a corpus publication of illustrated Kalilah wa Dimna Manuscripts’, *Islamic Art* 4 (1990/91), 301–481; see also U. Marzolph, ‘The Lithographed Kalilah Illustrations to Tales from the Anvār-i Suhayli and Kalila va Dimnah Tradition in Lithographed Editions of the Qājār Period’, *Islamic Art* (in print).
developed and adjusted in order to fit the broader context. One of the numerous questions that TULLIP might help to answer is concerned with the relationship between lithographic illustration and popular iconography. When considering book production in the Near and Middle East, one has to keep in mind that books do not only contain printed words. In largely illiterate societies such as nineteenth century Iran, the illustrations included in lithographed books might even have conveyed stronger and more lasting impressions than the printed word, similar to the impact exercised by illustrations in other popular media, such as tile-work, stucco-work, glass painting or picture carpets.
Osman Zeki Bey
and
His Printing Office the *Matbaa-i Osmaniye*
Nedret Kuran-Burçoğlu
Yeditepe University, Istanbul
**Introduction**
Osman Zeki Bey was an Ottoman calligrapher who lived in the second half of the nineteenth century in Istanbul. He was the printer who was given the first legal permission to print the Qur’ān, by the Palace. This paper which is based on family documents and research done by the great granddaughter of the printer aims to present Osman Zeki Bey’s life story within the socio-cultural context of his time,¹ as well as giving information about his printing office the *Matbaa-i Osmaniye* and its publications. By doing this it also aims to highlight a particular phase of the Ottoman history of printing.
**Osman Bey, the Printer**
The calligrapher Osman Zeki Bey who got the first legal permission from the Ottoman Palace to print the Qur’ān, the holy book of Islam, was the first Chamberlain (*Başmabeyinci*) at the Palace during the reign of Abdülhamid II. His birth date is unknown. His burial date is given on his tombstone as 1888 (1301).
His father, Hakkakzāde Mustafa Hilmi Efendi was a well-known calligrapher and his name ‘Hakkakzāde’ indicates that Osman Zeki Bey’s grandfather had been a
---
¹ The author of this paper started her research on Osman Bey’s Printing Office and its publications in January 2000. The initial research that had been mainly based on the author’s family documents, focused on Osman Bey’s life and general information about his Printing Office. That part was completed the same year and the following two articles were published: Nedret Kuran-Burçoğlu, ‘Saray’dan İlk Defa Kur’an-ı Kerim Basma İzni Alan Hattat: Matbaacı Osman Bey’, *Tarih ve Toplum Dergisi*. 35, issue 209 (2001) 312–20, İstanbul: İletişim Yayınları; Nedret Kuran-Burçoğlu, ‘Matbaacı Osman Bey’, *Journal of Turkish Studies: Essays in Honour of Barbara Flemming* Edit. by Şinasi and Gönül Tekin; guest editor: Jan Schmidt, 26/II (2002), 97–112, Harvard University: NELC. The second article covers some additional information about family members that is reflected in the family tree. |
Terminal Schwann Cells Lead Synapse Remodelling following Injury
Zarin Zainul
DOI: http://dx.doi.org/10.1523/ENEURO.0028-14.2014
Oulu Centre for Cell-Matrix Research, Biocenter Oulu, Faculty of Biochemistry and Molecular Medicine, University of Oulu, FIN-90014 Oulu, Finland
Key words: astrocytes; CNS; PNS; remodelling; Schwann cells; synapse
Significance Statement
This commentary article describes the importance and significance of the article recently published by the Kang and colleagues in *The Journal of Neuroscience* in 2014. Kang and colleagues provided new features of injury induced synapse remodelling. This commentary article summarizes the important findings of Kang and colleagues with the appropriate commentary. In addition to the article by Kang and colleagues, many other very exciting and recent studies about synapse remodelling in peripheral and central nervous system have also been included.
Although the neuromuscular junctions (NMJs) of adult animals are stable in their structure, the structure of the synapse may alter under certain circumstances: on account of development, aging (Kang and Lichtman, 2013), disease, and reinnervation after injury (Rich and Lichtman, 1989). Axon regeneration in the peripheral nervous system (PNS) after injury is well known, and is greatly supported and served by Schwann cells (SCs), glial cells in the PNS. Terminal Schwann cells (TSCs) that cover motor nerve terminals are known to extend their processes and lead regenerating nerve terminals to reinnervate adjacent postsynaptic sites after injury. Once the axons have been restored to their synaptic sites, remodeling of the contacts takes place. Structural changes in the synaptic site after reinnervation are associated with the speed of reinnervation. Fast reinnervation leads to modest remodeling, whereas delayed reinnervation results in more structural changes at the synapse (Rich and Lichtman, 1989). Important questions in all these events are how TSCs take the lead to regenerate synapse and how they play a role in synapse remodeling?
Kang and colleagues (2014) sought to answer these questions by sequential vital imaging of mice expressing GFP in their SCs and CFP in their motor neurons. They performed lateral, medial, and double nerve crush to achieve variable periods of denervation. As a consequence of different denervation methods, differential axon terminal branching patterns and directions of the reinnervation were observed (Kang et al., 2014, their Table 2). To explore the role of TSCs in synapse remodeling, Kang and colleagues (2014) first studied the behavior of the TSCs and regenerating axons during synapse remodeling. On denervation, TSCs occupy the synaptic area, but gradually abandon a variable portion of the acetylcholine receptors (AChRs) depending on the length of the denervation period (Kang et al., 2014, their Fig. 2); however, what drives TSCs to abandon AChR sites remains unclear. Multiple reinnervation, which is known to occur during reinnervation after injury, was observed here (Kang et al.,
ACHRs that were isolated from the others and lacked SC coverage were susceptible to loss (Kang et al., 2014, their Figs. 4, 6), suggesting that AChR loss after reinnervation is associated with the abandonment by SCs (Kang et al., 2014 their Fig. 5). Here, it is clear that besides inducing and guiding regenerating axons, TSCs also affect postsynaptic receptor sites during regeneration and remodeling. Since postsynaptic receptor loss is the consequence of the remodeling process after injury, it is clear that there will be more receptor loss in the aged animals where junctions are already fragmented. Thus, it is interesting to know how TSCs lead synapse remodeling in older mice following injury. Nevertheless, this study raises the question of how the absence of SC coverage at denervated postsynaptic sites leads to receptor loss. Here, it seems that in reinnervated receptors, innervating axons cause elimination of unoccupied receptor sites by reorganizing the postsynaptic sites to coincide with the nerve terminal location, although the exact factors that contribute to the postsynaptic receptor loss as a consequence of synapse remodeling remains unclear. The major questions here are how TSCs sprouting is induced after injury, and what are the driving factors behind this sprouting? A possible explanation could be that the induced expression of ErbB2 in SCs enhances sprouting of TSCs in a manner reminiscent of the sprouting that occurs after denervation. This indicates that NRG/ErbB signaling may be involved in TSC sprouting, and it could be evaluated by inactivating ErbB receptors on TSCs after nerve injury (Hayworth et al., 2006).
During synapse remodeling, branches of axon terminals in regenerating axons initially followed the SC processes within the old synaptic sites (Kang et al., 2014, their Fig. 3), and formed escaped fibers (Kang et al., 2014, their Figs. 3A, 4, 8). In some cases, reinnervation stimulated sprouting of SCs that appeared to reinnervate previously abandoned receptors (Kang et al., 2014, compare their Figs. 3B, 4). A similar event can be seen in frogs where TSCs extend their processes on the arrival of regenerating axons. In addition, TSCs in frogs express neuronal agrin, which may induce AChRs clustering under TSCs sprouts.
Addition of new AChRs was observed during the reinnervation process after injury (Rich and Lichtman, 1989), and can also be seen in Kang et al. (2014, their Fig. 8). Regenerating escaped fibers contained synaptic vesicles and seemed to be capable of forming a functional synapse (Kang et al., 2014, their Fig. 9). In their study, Kang and colleagues (2014) focused entirely on the morphologic aspects of TSC-mediated synapse reinnervation. Physiology of the reinnervated synapses was not examined here, and it would be of great interest to examine it further. One important question: how do the SCs communicate to the axons during regeneration after denervation? Recently, it was proposed that dedifferentiated SCs secrete exosomes that are internalized by axons. Growth cone morphology is then shifted into the pro-regenerating phenotype by the exosomes (Lopez et al. 2013).
Findings from the TSCs are relevant to astrocytes, which are glial cells in the CNS and responsible for the portion of synapse remodeling and mediating synapse elimination in the developing and adult brain. Developing mice lacking Megf10 and Mertk phagocytic pathways failed to refine retinogeniculate connections and returned an excess number of synapses. Moreover, in adult mouse brain, astrocytes continued to engulf both excitatory and inhibitory synapses, suggesting that astrocytes are the critical players in the synapse remodeling (Chung et al. 2013). Since astrocytes have been shown to be involved in synapse remodeling, it is very interesting to know whether astrocytes also regulate injury-induced synapse remodeling in the CNS. Furthermore, like TSCs, astrocytes are also needed for the formation and maturation of synapses in the CNS. In cultured retinal ganglionic cells, removal of astrocyte feeding layers leads to decreased synapse numbers, suggesting that astrocytes are needed for synapse maintenance (Slezak and Priefer, 2003). Unlike PNS axon regeneration, capability of the CNS in axon regeneration is extremely poor due to diminished intrinsic regenerative capacities and lack of glial support (Lopez et al. 2013; Zainul, 2013). Despite positive roles of astrocytes in the CNS, astrocytes do not seem to be a favorable candidate in CNS axon regeneration due to the formation of glial scar, which is the physical barrier through which axons cannot further elongate. Astrocytes in glial scars become hypertrophic and release proteoglycans that contribute to the poor CNS axon regeneration. However, ablation of proliferating reactive astrocytes disrupted scar formation that consequently causes intensified inflammation, tissue damage, demyelination, and impaired functional recovery. It was reported in the early 1980s that the peripheral nerve environment, mainly composed of SCs, is more favorable for the CNS axon regeneration than CNS milieu. After such an excellent finding, SCs transplantation into the injured CNS is being extensively explored as an approach to overcome the problem of poor CNS regeneration (Deng et al., 2014). Moreover, transplantation of genetically modified SCs to secrete neurotrophin and chondroitinase better improved axon regeneration and functional recovery (Kanno et al., 2014). Interestingly, in a recent study, astrocytes revealed increased phosphorylation expression by Cdc2 and integrin activation, which are positively associated with the increased neurite outgrowth (Toy and Namgung, 2013). Recently, the FDA has approved clinical trials for SC transplantation in spinal cord injury. Further studies about the roles of astrocytes and comparing the common features shared with SCs may provide better understanding of the facts and factors that contribute to the poor regenerative response in the CNS.
In addition to synapse remodeling, TSCs also participate in the long-term maintenance of adult NMJs, although the maintenance mechanism as such is still not fully understood (Reddy et al., 2003). It is of great interest to know whether the TSCs provide mechanical support or trophic support (Feng and Ko, 2008). It has been shown that mature SCs occupy segregated territories, and such segregation is absent during development, lost after injury, and swiftly restored after reinnervation following injury. The territory of a single TSC is constrained by competition with its glial neighbors and axon terminal, but not by axonal activity, revealed by laser ablation of single SCs and axons (Brill et al., 2011). It has been observed that TSCs and nerve terminal sprouting is absent in muscle after injury when nitric oxide is blocked (Marques et al., 2006); however, the primary source of nitric oxide is not
known since it is expressed by TSCs, nerve, and muscle. Moreover, NRG-1 signaling is upregulated in SCs after derivation, suggesting that NRG-1 signaling may be responsible for inducing SCs sprouting (Carroll et al., 1997). Furthermore, TSCs seem to suppress nerve terminal plasticity when Semaphorin 3A is selectively expressed in TSCs only at fast twitch muscle fibres after nerve injury (De Winter et al., 2006). A plethora of studies demonstrate that denervated SCs express a number of genes, for example, DCC and Uncoordinated (Unc)5H2, which are receptors of netrin-1, osteopontin, and SC secretory apolipoprotein that positively regulate functional recovery after PNS injury (Zainul, 2014). To study the overall function of the SCs in mammalian NMJs, it would be interesting to perform loss-of-function studies with gene knock-outs that lead to the ablation of SCs, or else to employ a complement-mediated cell lysis approach (Reddy et al., 2003). This would enable us to better understand the functions of SCs in axon regeneration and synapse remodeling.
One problem, however, is that the mice lacking SCs die immediately after birth, which would mean resorting to genetically altered mice with the reduced expression of SCs. It is known that physical activity determines the functioning of the NMJs and, interestingly, it has been recently reported that a muscle-specific increase in the expression of peroxisome proliferator-activated receptor γ co-activator 1α (PGC-1α) significantly promoted NMJ remodeling even when physical activity was not induced. The plastic changes in the NMJs were not restricted to the postsynaptic sites, but also modulated presynaptic morphology and function, thus confirming the pivotal role of skeletal muscle in NMJ remodeling (Arnold et al., 2014). However, Arnold and colleagues (2014) did not examine injury-induced NMJ remodeling, and it is known that the incomplete recovery following peripheral nerve injury can result in changes in the distribution of muscle fiber-types (Mendler et al., 2008). Based on the findings by Arnold and colleagues (2014), it would be highly interesting to examine whether SCs are involved or whether PGC-1α has direct, SC-independent effects on reinnervation in adult mice with PGC-1α overexpression after denervation.
In summary, Kang and colleagues (2014) demonstrated the cause of synapse remodeling and provided evidence for new features of the synapse remodeling after denervation. What is abundantly clear from their elegant work is that TSCs actively play a pivotal role in synapse remodeling following nerve injury. Their findings nevertheless raise many questions for future consideration, such as why the SCs abandon their previous synaptic sites, and what is the mechanism that leads to the failure of reinnervation at certain synaptic sites. Molecular signaling pathways that are responsible for the interactions among SCs, axon terminals, and the postsynaptic muscle fibres during synapse remodeling also remain to be further elucidated. Unmasking the underlying molecular mechanism will add vital insights to our understanding of the biology of synapse remodeling. Moreover, a therapy could be developed for shortening the denervation period or for promoting TSCs maintenance without axonal trophic support, which would be essential for reducing growth of SCs away from the synapses and could consequently promote actual synapse recovery.
References
Arnold AS, Gill J, Christe M, Ruiz R, McGuirk S, St-Pierre J, Tabares L, Handschin C, (2014) Morphological and functional remodelling of the neuromuscular junction by skeletal muscle PGC-1alpha. Nat Commun 5:3569. CrossRef Medline
Brill MS, Lichtman JW, Thompson W, Zuo Y, Misgeld T (2011) Spatial constraints dictate glial territories at murine neuromuscular junctions. J Cell Biol 195:293–305. CrossRef Medline
Carroll SL, Miller ML, Frohnhert PW, Kim SS, Corbett JA (1997) Expression of neuregulins and their putative receptors, ErbB2 and ErbB3, is induced during wallerian degeneration. J Neurosci 17:1642–1659. Medline
Chung WS, Clarke LE, Wang GX, Stafford BK, Sher A, Chakraborty C, Joung J, Foo LC, Thompson A, Chen C, Smith SJ, Barres BA (2013) Astrocytes mediate synapse elimination through MEGF10 and MERTK pathways. Nature 504:394–400. CrossRef Medline
Deng LX, Walker C, Xu XM (2014) Schwann cell transplantation and descending propriospinal regeneration after spinal cord injury. Brain Res. Advance online publication. doi:10.1016/j.brainres.2014.09.038.
De Winter F, Vo T, Stam FJ, Wisman LA, Bar PR, Nicolou SP, van Muiswinkel FL, Verhaagen J (2006) The expression of the chemorepellent semaphorin 3A is selectively induced in terminal Schwann cells of a subset of neuromuscular synapses that display limited anatomical plasticity and enhanced vulnerability in motor neuron disease. Mol Cell Neurosci 32:102–117. CrossRef
Feng Z, Ko CP (2008) The role of glial cells in the formation and maintenance of the neuromuscular junction. Ann N Y Acad Sci 1132:19–28. CrossRef Medline
Hayworth CR, Moody SE, Chodosh LA, Krieg P, Rimer M, Thompson WJ (2006) Induction of neuregulin signaling in mouse schwann cells in vivo mimics responses to denervation. J Neurosci 26:6873–6884. CrossRef Medline
Kang H, Lichtman JW (2013) Motor axon regeneration and muscle reinnervation in young adult and aged animals. J Neurosci 33:19480–19491. CrossRef Medline
Kang H, Tian L, Mikesik M, Lichtman JW, Thompson WJ (2014) Terminal Schwann cells participate in neuromuscular synapse remodeling during reinnervation following nerve injury. J Neurosci 34:6323–6333. CrossRef Medline
Kanno H, Pressman Y, Moody A, Berg R, Muir EM, Rogers JH, Ozawa H, Itoi E, Pearse DD, Bunge MB (2014) Combination of engineered Schwann cell grafts to secrete neurotrophin and chondroitinase promotes axonal regeneration and locomotion after spinal cord injury. J Neurosci 34:1838–1855. CrossRef Medline
Lopez-Verrilli MA, Picou F, Court FA (2013) Schwann cell-derived exosomes enhance axonal regeneration in the peripheral nervous system. Glia 61:1795–1806. CrossRef Medline
Marques MJ, Pereira EC, Minatel E, Neto HS (2006) Nerve-terminal and Schwann-cell response after nerve injury in the absence of nitric oxide. Muscle Nerve 34:225–231. CrossRef Medline
Mendler L, Pinter S, Kiricsi M, Baka Z, Dux L (2008) Regeneration of reinnervated rat soleus muscle is accompanied by fiber transition toward a faster phenotype. J Histochem Cytochem 56:111–123. CrossRef Medline
Reddy LV, Koirala S, Sugiyara Y, Herrera AA, Ko CP (2003) Glial cells maintain synaptic structure and function and promote development of the neuromuscular junction in vivo. Neuron 40:563–580. Medline
Rich MM, Lichtman JW (1989) In vivo visualization of pre- and postsynaptic changes during synapse elimination in reinnervated mouse muscle. J Neurosci 9:1781–1805. Medline
Slezak M, Pfrieger FW (2003) New roles for astrocytes: regulation of CNS synaptogenesis. Trends Neurosci 26:531–535. CrossRef CrossRef
Toy D, Namgung U (2013) Role of glial cells in axonal regeneration. Exp Neurolobiol 22:68–76. CrossRef Medline
Zainul Z (2013) Delete genes to repair nerves: knockouts and axon regeneration in peripheral and central nervous system. Am J Neuroprotec Neuroregen 5:25–31. CrossRef
Zainul Z (2014) Positive regulatory genes in peripheral axon regeneration and functional recovery following injury. Am J Neuroprotec Neuroregen 6:4–9. CrossRef |
A Look In the Mirror: Attacks on Package Managers
Justin Cappos Justin Samuel Scott Baker John H. Hartman
Department of Computer Science, University of Arizona
Tucson, AZ 85721, U.S.A.
{justin, jsamuel, bakers, firstname.lastname@example.org
ABSTRACT
This work studies the security of ten popular package managers. These package managers use different security mechanisms that provide varying levels of usability and resilience to attack. We find that, despite their existing security mechanisms, all of these package managers have vulnerabilities that can be exploited by a man-in-the-middle or a malicious mirror. While all current package managers suffer from vulnerabilities, their security is also positively or negatively impacted by the distribution’s security practices. Weaknesses in package managers are more easily exploited when distributions use third-party mirrors as official mirrors. We were successful in using false credentials to obtain an official mirror on all five of the distributions we attempted. We also found that some security mechanisms that control where a client obtains metadata and packages from may actually decrease security. We analyze current package managers to show that by exploiting vulnerabilities, an attacker with a mirror can compromise or crash hundreds to thousands of clients weekly. The problems we disclose are now being corrected by many different package manager maintainers.
Categories and Subject Descriptors
K.6.5 [Security and Protection]: Invasive software; C.2.0 [Computer-Communication Networks]: General—Security and protection; K.4.1 [Social Issues]: Abuse and Crime Involving Computers
General Terms
Security
Keywords
Package Management, Mirrors, Replay Attack
1. INTRODUCTION
Package managers are a popular way to distribute software (bundled into archives called packages) for modern operating systems [1, 2, 3, 21, 22, 25, 26, 28, 31, 32]. Package managers provide a privileged, central mechanism for the management of software on a computer system. As packages are installed by the superuser (root), their security is essential to the overall security of the computer.
This paper evaluates the security of the eight most popular [9, 19] package managers in use on Linux: APT [1], APT-RPM [2], Pacman [3], PackageKit [4], pkgtool [5], urpmi [28], YaST [6], and YUM [32]. One example is the popular package manager for BSD systems called ports [22] and a popular package manager in the research community called Stork [26]. These package managers use one of four different security models: no security, cryptographic signatures embedded within packages, signatures on detached package metadata, or signatures on the root metadata (a file that contains the secure hashes of the package metadata).
This work demonstrates that there is an ordering to the amount of security provided by the different package manager security models. This order is preserved even as other security weaknesses in this package managers are corrected. Having no signatures allows the most egregious attacks, followed by having only package signatures, having signatures on detached package metadata, and finally having signatures on the root metadata, which provides the most security. However, there are usability concerns with different signature mechanisms, most notably the ability to verify a stand-alone package (a package obtained from a source other than the main repository). Signatures on root metadata do not provide a mechanism to verify stand-alone packages and so the user is likely to install such packages without using security checks. In contrast, package managers that use signatures on detached package metadata or signatures on packages can verify stand-alone packages.
Because of the usability strengths and weaknesses of different techniques for providing security, this work recommends a layered approach created by combining two techniques: signatures on the root metadata and either signatures on packages or package metadata. This technique provides the security strengths and the usability strengths of both types of signatures with an overhead of between 2-5%. The layered approach has been added to the Stork package manager and is now in use by thousands of clients around the world.
While we find that vulnerabilities in package managers do exist, vulnerabilities are not always exploitable in the real world. We examine this by looking at the security of popular distributions. We find that it is trivial for an attacker to control an official package mirror for a popular
distribution (such as Ubuntu, Debian, Fedora, CentOS, and openSUSE) and therefore to be in the position to launch attacks on clients. To mitigate this threat, many distributions use mechanisms to distribute requests to multiple mirrors or provide certain information from a trusted source. We show that some of these mechanisms actually decrease the security of users by making it easier to target attacks.
2. BACKGROUND
2.1 Package Formats
Packages consist of an archive containing files and, in most cases, additional embedded package metadata. For a given package, the embedded package metadata contains information about any other packages that must be installed in order for it to operate (the dependencies), functionality the package possesses (what the package provides), and various other information about the package itself. The most popular package format, RPM [23], has space for one signature. Other popular package formats have no standard field for signatures, although in some cases extensions exist to support signatures [8, 11].
2.2 Package Managers
Clients use a package manager to install packages on their system. A package manager gathers information about packages available on package repositories. Almost all package managers automatically download requested packages as well as any additional packages that are needed to correctly install the software. This process is called dependency resolution. For example, a requested package foo may depend on libc and bar. If libc is already installed, then libc is a dependency that has been resolved (no package must be installed to satisfy the dependency). If there is no installed package that provides bar, then bar is an unresolved dependency and a package that provides bar must be installed before foo may be installed. The package manager may be able to locate a package that provides bar on a repository.
The packages that are chosen to fulfill dependencies may have unresolved dependencies of their own. Packages are continually added to the list of packages to be installed until either the package manager cannot resolve a dependency (and produces an error) or all dependencies are resolved.
2.3 Repository
A package repository is usually an HTTP or FTP server from which clients can obtain packages and package metadata. The package metadata is usually just a copy of the embedded package metadata for all packages on the repository. Package managers download the package metadata from a repository so that they know which packages are available from that repository. This also provides the package manager with dependency information needed to perform dependency resolution. To facilitate convenient downloading of package metadata, most repositories store all of the package metadata in a small number of compressed files.
In addition to the package metadata, almost all repositories have a root metadata file. The name and location of the root metadata file varies by different repository formats, but the contents are similar. The root metadata provides the location and secure hashes of the files that contain the package metadata.
Figure 1: Repository Layout. The root metadata, package metadata, and packages may all optionally have signatures depending on the support of the package manager. Arrows point from a secure hash to the file it references.
Figure 1 shows the layout of a typical repository. A package manager downloads the root metadata and uses that to locate the files containing the package metadata. The package manager then downloads the package metadata. The package metadata is used to determine package availability as well as for dependency resolution. Packages are then downloaded and installed. The root metadata, package metadata, and packages may be signed depending on the security model of the package manager.
2.4 Mirror
It is common for a distribution to have more than one server from which users can download packages and package metadata. There is usually a main repository for a distribution whose contents are copied by many separate mirrors. A mirror typically contains exactly the same content as the main repository and is updated via rsync or a similar tool. A mirror differs from a main repository in that a mirror is not intended to have packages directly added to it or removed from it by its administrators. Packages are added or removed only on the main repository, and the mirrors later obtain the changes when copying the main repository.
A mirror can be public (available for anyone to use) or private (restricted to a specific organization). A mirror may also be endorsed by a distribution for public use, typically when that the distribution is in contact with the mirror maintainers. This type of mirror is called an official mirror (the terminology used outside of this document varies by the distribution). Official mirrors are by definition public because the distribution is endorsing their use to the public.
It should be noted that some distributions do not use official mirrors hosted by outside organizations. One example is tiny distributions that can support all of their clients by a small number of repositories that the distribution directly controls. Another example is a distribution that requires users to pay. These costs often are used to support a set of internally maintained mirrors for the distribution. Alternatively, a distribution may allow or require each organization using the distribution to set up their own private mirrors for the organization’s own use.
However, official mirrors hosted by outside organizations are the predominant mechanism for software distribution with all but two popular distributions [9, 19] relying on official mirrors. Official mirrors are essential for most distributions to reduce cost and management overhead.
3. THREAT MODEL
The threat model used in this paper involves an attacker that can respond to requests made by a package manager. This can be a man-in-the-middle, an attacker that has tricked a client into contacting the wrong server (e.g. through DNS cache poisoning), or an attacker who has gained control of an official mirror for a distribution. The threat model is as follows:
- The attacker can serve the client arbitrary files.
- The attacker does not know what package the client will request a priori.
- The attacker \textit{does not} have a key trusted to sign packages, package metadata, or the root metadata. — Note that mirrors do not usually possess the private key used to sign files, they only copy previously signed files from the main repository.
- The attacker has access to outdated packages, outdated package metadata, and outdated root metadata files. — There are many outdated mirrors on the Internet where an attacker can obtain these files.
- The attacker is aware of vulnerabilities in some outdated packages and is able to exploit those vulnerabilities. — This is possible by looking at change logs and updates to software source files or downloading an exploit toolkit [18].
- The attacker does not know of a vulnerability in the latest version of any package. — Zero-day vulnerabilities are unlikely to be known by many attackers.
- If a package manager supports signatures, signatures are useful. If a client or distribution chooses not to use signatures supported by their package manager, they are as vulnerable as if they used a package manager that does not support signatures.
- Expiration times in the root metadata are used, if supported, and current (un-expired) root metadata does not contain any vulnerable versions of packages — The root metadata is a single, small file so it is feasible for the main repository to sign it relatively frequently with short expiration times.
3.1 Attacks
Given this threat model, there are several attacks that may be used on a client. The impact of these attacks varies, but all allow the attacker to either crash or control the client’s computer (possibly via exploiting a package with a known vulnerability). Each of the following attacks is effective on at least some of the package managers we studied:
- \textit{Arbitrary Package} The attacker provides a package they created in place of a package the user wants to install.
- \textit{Replay Attack} An attacker replays older versions of correctly signed packages or metadata, causing clients to install an old package with security vulnerabilities the attacker can exploit. The attacker can then compromise the client by exploiting the vulnerable package. Note that most package managers will not downgrade an existing package, so a replay attack only works when the package manager installs a new package, not when it updates an existing package.
- \textit{Freeze Attack} Similar to a replay attack, a freeze attack works by providing metadata that is not current. However, in a freeze attack, the attacker freezes the information a client sees at the current point in time to prevent the client from seeing updates, rather than providing the client older versions than the client has already seen. As with replay attacks, the attacker’s goal is ultimately to compromise a client who has vulnerable versions of packages installed. A freeze attack may be used to prevent updates in addition to having an installed package be out of date.
- \textit{Extraneous Dependencies} The attacker rewrites the package metadata to have additional packages installed alongside a package the user intends to install. For example, the attacker provides metadata that incorrectly states that package \texttt{foo} depends on \texttt{bar}. This will cause \texttt{bar} to be installed when it is not desired or needed. If \texttt{bar} has a security vulnerability, this allows an attacker to compromise the client’s system.
- \textit{Endless Data} This attack is performed by returning an endless stream of data in response to any download request. This may cause the package manager to fill up the disk or memory on the client and crash the client’s system.
4. SECURITY OF PACKAGE MANAGERS
The security of a package manager varies depending on how signatures are used to protect data. This section explores the security strengths and weakness of signatures on different data along with implementation pitfalls observed in package managers (and how to fix them). This section then classifies the security of different signatures into a list ordered by increasing security.
The discussion groups package managers with similar security characteristics together. The first group of package managers do not use signatures (Section 4.1). The second group of package managers use signatures on packages but do not use signatures on package metadata or root metadata (Section 4.2). The third group of package managers use signatures on package metadata but not on the root metadata (Section 4.3). The final group of package managers use signatures on the root metadata (Section 4.4).
4.1 Package Managers Without Security
There are three popular package managers that do not provide security: Pacman, ports\footnote{A version of ports used by NetBSD did support package signatures at one time [7], but this has been obsoleted and is not maintained or used.}, and Slaktool. These package managers do not sign packages, package metadata, or the root metadata file. As a result, any attacker that controls a mirror can trivially launch an arbitrary package attack by responding to client requests with malicious software.
4.2 Package Signatures
YUM and urpmi rely solely on signatures embedded in packages to provide security. There is no protection of package metadata or the root metadata. As a result, an attacker can launch replay or freeze attacks and have clients install vulnerable packages. An attacker can choose to include any vulnerable versions of signed packages they wish on the repository.
However, neither YUM nor urpmi verify that the package metadata they initially received for dependency resolution matches the embedded package metadata of subsequently downloaded packages. This allows an attacker to forge package metadata to launch an extraneous dependencies attack. This means that if there exists a package that the attacker knows how to exploit and which the user does not have installed, the attacker can force it to be installed whenever any other package is installed by the user. The result is that an attacker can compromise essentially every client that installs or updates a package.
Both YUM and urpmi are also vulnerable to endless data attacks. For example, when YUM is given a repomd.xml file of unlimited size, it exits without printing an error after the filesystem is full — leaving the huge file on disk. Since no information is logged or printed about the error, this makes discovering the problem complicated (especially if YUM runs via auto-update).
Fixing the package managers Even without changing the signature methods, it is possible to modify these package managers to prevent the extraneous dependencies attack. Doing so requires verifying that downloaded package metadata is the same as the embedded metadata in any downloaded package. Furthermore, endless data attacks can be prevented by capping the size of downloaded data. However, these changes are not adequate to protect against replay or freeze attacks. Given the large number of packages that need to be checked to prevent replay or freeze attacks, we believe that adding root metadata signing is the most practical way to address this issue.
4.3 Package Metadata Signatures
The Portage and Stork package managers use signatures on package metadata; however, they do so in different ways. Each package in Portage has a separate, signed package metadata file for each version of the package. The package metadata contains the secure hash of the package (possibly along with hashes of related files such as patches). In contrast, Stork users create a single file that contains a timestamp and the secure hash of the package metadata for all of the packages that the user trusts. Users can also delegate trust to other users and all users typically delegate trust to a single “distribution” user. The analysis of Stork therefore focuses on the security of the packages trusted by the distribution user because the security of the distribution user affects all clients.
These package managers are not vulnerable to extraneous dependencies attacks because the signatures protect the package metadata. However, resistance to metadata tampering does not imply that the package manager is resistant to all attacks. Both package managers are vulnerable to endless data attacks.
In both Portage and Stork, an attacker can launch freeze attacks. In Portage, since each package has a different file for metadata signatures, an attacker can choose to have any combination of packages (such as those that include only older versions with known vulnerabilities) available on the mirror. Portage is vulnerable to replay attacks in the same way as freeze attacks.
In Stork, all of the package metadata hashes are in the same signed file. This prevents an attacker from choosing package metadata that existed on the repository at different times. Stork checks that timestamps on files are increasing to prevent replay attacks but does not prevent freeze attacks.
Fixing the package managers It is possible to add replay attack protection to Portage through timestamp checking, as well to add protection against endless data attacks to both Portage and Stork. However, both package managers will still be vulnerable to freeze attacks.
4.4 Root Metadata Signatures
The package managers APT, APT-RPM, and YaST use signatures on the root metadata. All three of these package managers optionally support package signatures as well, but this functionality is not widely used in practice.
Package metadata is stored in compressed files and the secure hashes of those files are stored in the root metadata file. As the root metadata is protected by a signature, the package metadata is protected from tampering which prevents extraneous dependencies attacks. In addition, the signature on the root metadata prevents a mirror from hosting versions of packages that were on the main repository during different time periods. The attacker must choose a time period of the main repository to copy and provide exactly those files when launching a replay attack. Unfortunately, none of these package managers check the order of timestamps to prevent replay attacks or have any mechanisms to prevent freeze attacks or endless data attacks.
Fixing the package managers It is easy for these package managers to protect against endless data attacks as well as to prevent replay attacks. Replay attacks can be prevented by adding a timestamp to the root metadata and checking that any newly downloaded root metadata is not older than the version the client last obtained. In fact, all three package managers have a timestamp available and merely need to add this check.
To mitigate the effectiveness of freeze attacks, package managers could add an expiration time to the root metadata. Clients would refuse to use a root metadata file if the current time is greater than the expiration time. Since the root metadata is a single, small file, it is feasible to re-sign this file often and require every mirror to be frequently updated (most distributions already require their public mirrors to update at least once a day).
4.5 Classification
The security mechanisms and vulnerabilities of the package managers are summarized in Figure 2. All of the package managers studied are vulnerable to endless data attacks as well as freeze attacks. Different package managers have varying resistance to other attacks such as replay attacks. Depending on the package manager’s security mechanisms, the result can be any of the following, where those listed first also imply those listed after: arbitrary packages created by the attacker are installed, any vulnerable package can be installed alongside non-vulnerable packages a client installs using an extraneous dependencies attack, mismatched outdated packages are installed (in that they existed on the
| Name | Signature | Package Installation | Metadata Abuse |
|----------|-----------|----------------------|----------------|
| Pacman | nothing | arbitrary | arbitrary |
| ports | nothing | arbitrary | arbitrary |
| Minnow | nothing | arbitrary | arbitrary |
| YUM | (1) | alongside | arbitrary |
| urpmi | (1) | alongside | arbitrary |
| Portage | (2) | mismatch | replay / freeze|
| RPM | (1), (2) | snapshot | replay / freeze|
| APT | (1)*, (3) | snapshot | replay / freeze|
| YaST | (1)*, (3) | snapshot | replay / freeze|
| APT-RPM | (1)*, (3)*| snapshot | replay / freeze|
**Figure 2:** Package managers, their protection mechanisms and vulnerabilities. The protection mechanisms are numbered (1) packages, (2) package metadata, (3) root metadata. ** indicates that support exists but is not in common use.
| Signatures Protecting | Best Case Package | Common Case Package |
|-----------------------|-------------------|---------------------|
| | Metadata Abuse | Metadata Abuse |
| No Security Package | arbitrary | arbitrary |
| | mismatch | replay / freeze |
| | | alongside |
| Package Metadata | mismatch / snapshot | freeze |
| | | mismatch |
| | | replay / freeze |
| Root Metadata | current | none |
| | | snapshot |
| | | replay / freeze |
**Figure 3:** Classification of package manager protection schemes. This demonstrates both the security that is possible to achieve using a scheme as well as what is commonly provided by existing implementations.
main repository at different times), or installed packages will come from a collection of outdated packages that all existed at the same time on the main repository (i.e. in the same snapshot).
Based on the observation and analysis of the security in existing package managers, it is possible to similarly classify the security mechanisms. As Figure 3 shows, one can obtain an ordering of the security of the mechanisms. Clearly, having no signatures allows the most attacks and is the most vulnerable. Signatures on packages provide a definite improvement over no signatures, but gives the attacker the ability to manipulate metadata arbitrarily and provides attackers the ability to populate a mirror with packages of mismatched versions, or, if package metadata isn’t verified using the signed packages, the ability to cause vulnerable packages to be installed alongside any non-vulnerable packages. Signatures on package metadata prevent the attacker from doing more than replaying or freezing the package metadata, but if these signatures are embedded in files the attacker can still mismatch versions of packages. By preventing replay and freeze attacks in package managers that sign the root metadata, a package manager will only install current packages and is immune to metadata tampering.
5. ADDITIONAL USABILITY NEEDS
This section focuses on additional usability requirements users have for package management. Most importantly, the use case where a user has an uninstalled package on their computer they want to verify.
The standard use case of the package managers and their security mechanisms is where a user needs to securely install software from a repository or mirror. However, it is not uncommon for a user to have a stand-alone package that was created by a party they trust and that they need to verify is free from tampering. Stand-alone packages are packages that are not obtained through the package manager’s normal channels at install time. Stand-alone packages may have been obtained manually from unofficial sources or may even be packages a user has created. Another source provides an extended discussion [29] describing why stand-alone package verification is an important and desirable feature.
The signing of only root metadata does not allow any practical means to verify stand-alone packages. Package managers that use signed root metadata could be modified to keep copies of all metadata obtained from the repository for future verification of stand-alone packages, but this only helps for packages a user manually downloads from the same repository that they access through their package manager. This also fails to satisfy one of the primary reasons given for being able to verify stand-alone package signatures: verifying signatures for files when they are only available for manual download and installation, not through a repository.
Package managers that sign package metadata tend to be more able to meet the needs of stand-alone package verification than the package managers who only sign root metadata. However, the way in which package metadata is stored has a major impact on usability in this case. Similarly with package managers that sign only root metadata, package managers would need to store old package metadata and this would only be of use for verifying stand-alone packages that came from a repository the user normally uses. In other cases, the user would need to be sure to always keep the signed package metadata with the package for verification purposes. This is far from an ideal solution.
Signatures embedded in packages are thus the most practical option and provide the greatest ease of use when stand-alone packages must be verified. All that a user needs in order to verify a package is the package itself. A drawback with having signatures in the package is that signatures are constrained by the limits of the package format so multiple signatures may not be supported.
Using signatures embedded in packages for stand-alone package verification does have complications, though. Notably, users must have the requisite public keys available in order to verify package signatures. They must also ensure on their own that packages they are installing are not outdated or have vulnerabilities. However, there are many scenarios where a user can use embedded package signatures in a way that increases security in their specific situation.
6. DEPLOYMENT EXPERIENCE
To gain more experience with what security mechanisms work well in practice, we modified the package manager Stork and added root metadata signing. We added an expiration time to the root metadata to prevent freeze attacks. Since Stork already supported both package signatures and package metadata signatures, this allowed us to experiment with all types of signatures in a single package manager.
The changes to Stork for root metadata signing were tested and beta-deployed and finally incorporated into the production release. Interestingly, Stork differed from all other secure package managers in that there was no key already trusted by clients to validate communication from the repository. As the only signed files in Stork were the package
metadata files signed by individual users, there had never been a need for the repository to have its own key that the clients trusted. This required distributing a repository key to clients in order for them to make use of the new root metadata signatures. The key was included with the initial release of the updated version of Stork. This initial key distribution was secure because the majority of users, through their trusted packages files, delegate trust to the Stork team to provide them updated Stork packages. Stork’s design meant that users would not be using this key for trusting packages but rather only verifying metadata files downloaded from the repository.
The resulting changes were transparent to the users both in terms of performance and usability. The overhead of using package metadata signing along with root metadata signing was measured and found to be negligible (between 2-5%). Ultimately, there were few comments about the addition of root metadata signatures since the existing security mechanism (package metadata signatures) was retained without modification. Though transparent to the users, they gained increased security through the addition of root metadata signatures.
We examined packages on the Stork repository to find that user-uploaded packages did not include package signatures, indicating that researchers were not using the optional package signature feature of Stork. While far from conclusive, this implies that package metadata signatures and package signatures are redundant. We reason that when multiple signature mechanisms are provided, either package metadata signatures or package signatures are sufficient for usability purposes and it is unnecessary to support both.
To conclude, we feel one should work hard from both a security and a usability standpoint. We propose that package managers should use multiple security mechanisms. It is clear that root metadata signatures should be included because of their security benefits. It also seems advantageous to have either package metadata signing or package signatures for usability. By combining root metadata signatures with either signed package metadata or signed packages, a package manager can obtain a high degree of security and excellent usability without significant performance impact.
7. PACKAGE MANAGERS IN PRACTICE
This section examines additional functionality that is provided by some of the most popular distributions (Ubuntu, Debian, Red Hat Enterprise Linux, Fedora, CentOS, openSUSE, and SUSE Enterprise Linux) that impacts the security of package managers in practice. It should be noted that we did a cursory examination of other popular distributions (Gentoo, Mandriva, LinuxMint, Sabayon Linux, Slackware, KNOPPIX, Arch Linux, and MEPIS Linux) and did not find any additional security practices that significantly altered the security over that provided by their package manager. We did not survey any smaller distributions, but we strongly suspect that the majority of them are at least as vulnerable as their package manager is by default.
In this section, we first examine the feasibility of an attacker obtaining an official mirror for a distribution. We examine this by setting up mirrors for five popular distributions that have official mirrors run by outside parties (Ubuntu, Debian, Fedora, CentOS, and openSUSE). We then examine the security pitfalls in using a security repository for package updates and a set of untrusted mirrors for the core distribution. Next, we examine how the mechanisms by which requests are distributed to different mirrors impacts security. We then examine the use of HTTPS to try and determine its effect and applicability. The section concludes with a look at the discussed distributions and a comparison of their security characteristics.
7.1 Obtaining a Mirror
To evaluate the feasibility of controlling mirrors of popular distributions, we set up public mirrors for the CentOS, Debian, Fedora, openSUSE, and Ubuntu distributions. A fictitious company (Lockdown Hosting) with its own domain, website, and fictitious administrator (Jeremy Martin) were used as our organization maintaining the mirrors. A server with a monthly bandwidth quota of 1500 GB was leased for $200 per month through The Planet (www.theplanet.com).
Setting up a public mirror for each distribution involved acquiring the packages and metadata from an existing mirror and then notifying the distribution maintainers that the mirror was online and available for public usage. The distributions varied in terms of the degree of automation in the public mirror application and approval process as well as whether newly listed mirrors have traffic immediately and automatically directed to them. In all cases, the distributions accepted our mirror and added it to the official mirror list for use by outside users. We saw traffic on our mirrors from a variety of clients, including military and government computers. More detail about our mirrors can be found in a tech report [5].
7.2 Security Repository
Debian and Ubuntu both use an official repository that serves security updates (packages that fix vulnerabilities). This prevents a mirror from launching a replay attack because the package manager will use the latest version of a package which will be available from the security repository. However, this protection does not extend to a man-in-the-middle attacker since the repositories do not support HTTPS.
Both Debian and Ubuntu use several mirrors beside the security repository. In the case where the security repository is down, an attacker can use a mirror to serve outdated content that was originally from the security repository to perform a replay attack. In addition, using multiple mirrors makes Ubuntu and Debian users much more vulnerable to endless data attacks.
7.3 Mirror Selection
In many distributions, not every source of data is created equal. For example, openSUSE distributes all of the metadata from a central source and only outsources package requests to mirrors. In this section we examine the impact that these practices have on the security of different distributions.
OpenSUSE uses a download redirector that sends some requests to mirrors. However, the download redirector serves package metadata and root metadata directly. This means that the client gets the metadata from a trusted source (not a mirror). While this doesn’t protect against all of the attacks a client may face (such as endless data attacks by mirrors or replay attacks by a man-in-the-middle attacker), it does make it much more difficult to launch an attack.
On CentOS, clients contact a central service which redirects the client’s requests to official mirrors. Requests for all types of content (packages, package metadata, and root metadata) are directed to mirrors that are not controlled by CentOS. The specific mirror a client is redirected to may be different for each file requested. The result is that an average CentOS client contacts many more mirrors than would normally be the case. This makes it much easier for an attacker to conduct an endless data attack on many clients but complicates replay or freeze attacks because the same repository may not be contacted for packages and metadata. However, an attacker can still effectively launch an extraneous dependency attack.
On Fedora, by default clients contact a central service called MirrorManager that is very similar to the CentOS service. However, MirrorManager allows a mirror administrators to specify that clients in an IP address range should use only their mirror. This allows easy targeting of attacks (to a specific country or organization) and reduces the number of other parties who will consume resources on the mirror. Note that the mirror need not have an IP address in the IP address range that it targets.
Online mirror redirectors handle failures differently than a security repository. If a mirror redirector fails, the clients will stop trying to communicate and fail (as opposed to “failing-open” by skipping the failed repository).
Many distributions including Gentoo, KNOPPIX, Arch Linux, Debian, and Ubuntu also have mirror selection tools that find a nearby mirror. These tools do not improve the security of users since they are only used to select a single mirror to download all content from. In fact, it should be noted that popular mirror selection tools for Debian and Ubuntu (such as netselect-apt or Software Sources) do not present the official security repository. This means that users who have used these tools to find faster or more reliable mirrors are unlikely to use the security repository for updates!
7.4 Verifying Mirrors
Many distributions use an automated mechanism to verify that mirrors are staying up-to-date. While we did not attempt to deviate from the correct, up-to-date status on our mirrors, we posit that an attacker can determine the IP address of the checking server and serve up-to-date content to the checking server while serving malicious content to users it is attacking. Note that this is trivial on distributions like Fedora that allow a mirror administrator to selectively target users, because the only references coming from outside the targeted range should be the checking server. It is publicly acknowledged by some distributions [13] that detecting and tricking these bots is trivial and so we do not feel this is a viable security mechanism.
7.5 HTTPS
Red Hat Enterprise Linux (RHEL) and SUSE Linux Enterprise are the only distributions that we found which widely support or use HTTPS. When investigating how YUM uses HTTPS, we discovered it does not validate SSL certificates. This means that while the communication uses HTTPS, there is no validation that the endpoint YUM communicates with is correct (allowing a man-in-the-middle to pose as the repository).
However, the Red Hat security team informed us that on RHEL, YUM communicates with the Red Hat Network servers using a special plug-in instead. During our discussion with the Red Hat security team, they realized there were issues in how SSL was used in the plug-in as well and began working on a fix for it.
7.6 Comparison
We examine the additional security mechanisms provided by different distributions and comment on their effectiveness against attack.
- SUSE Linux Enterprise doesn’t use mirrors hosted by outside parties and uses HTTPS so it is not vulnerable to the attacks described in this paper.
- OpenSUSE, by using a download redirector, provides significant protection from malicious mirrors. However, any other attacker, such as a man-in-the-middle, that can respond to client requests can perform replay or freeze attacks. If the download redirector fails or is unreachable (e.g. due to a denial-of-service attack), users cannot get updates but are not at risk of replay attacks. Despite the download redirector protecting users from replay or freeze attacks from malicious mirrors, users are still vulnerable to endless data attacks from mirrors.
- Ubuntu and Debian have similar security to openSUSE with the additional problem that if the security repository fails, they become vulnerable to replay or freeze attacks from mirrors.
- Red Hat Enterprise Linux is not at risk from malicious mirrors (since it uses no mirrors hosted by outside parties) but is vulnerable to man-in-the-middle attackers because of flaws in their HTTPS implementation. A malicious party who can act as a man-in-the-middle can launch any of the attacks that YUM is vulnerable to.
- With a man-in-the-middle attacker, CentOS is vulnerable to every attack (much like RHEL). In the case of a malicious mirror, their download redirector makes it more difficult to launch attacks that require the attacker to provide a snapshot of packages (like a replay or freeze attack), but the potential for abuse from endless data and extraneous dependencies attacks is still large.
- Fedora has the same vulnerabilities as CentOS but with the additional problem that an attacker can target attacks to a specific IP range. This also allows an attacker to launch attacks that require the attacker to provide a snapshot of the repository.
- Other distributions such as Gentoo, Mandriva, LinuxMint, Sabayon Linux, Slackware, KNOPPIX, and MEPIS Linux do not significantly alter the protection provided by their package manager.
8. RESULTS
To understand the impact of an attacker that controls a mirror, a trace of package requests was conducted on a CentOS mirror and used to estimate the number of clients that could be compromised or crashed by an attacker. For all of the popular distributions our mirror could have launched
Figure 4: CentOS Mirror Traffic. This figure presents the cumulative package downloads and requests from unique IP addresses for the CentOS mirror over a 13 day period.
Figure 5: Compromised Clients CDF. This figure presents the cumulative number of clients compromised over a 13 day period for an attacker with 25 vulnerable versions. This figure shows the effect of the security mechanism on the number of clients compromised.
an endless data attack to crash all of the clients that visited it. The number of unique IP addresses that contacted our mirror represents an upper bound on the number of clients that could be crashed each week.
8.1 Mirror Traces
The CentOS mirror was chosen for this analysis because it was the longest mirror experiment that was conducted, lasting 13 days. The package access trace gathered from the CentOS mirror is shown in Figure 4. The number of package requests and number of requests from unique IP addresses increase roughly linearly over this time period. Assuming the CentOS user base is not growing faster than our mirror serves clients, we would expect the number of unique IP addresses to flatten out over time, however our trace is not long enough to capture that effect. Since clients are counted as unique by IP address, multiple clients behind a NAT box or proxy are counted as a single client. There are many instances where a single IP address has several orders of magnitude more package requests than the median client, often with many requests for the same package. This implies that clients are using NAT boxes or proxies in practice.
The openSUSE and Fedora mirrors (not shown) had similar traffic effects as the CentOS mirror. The Debian and Ubuntu mirrors (not shown) were both up for only a short period, but did not demonstrate this effect. We suspect this is because they do not automatically distribute requests among the mirrors, instead requiring manual selection by a user or the use of a tool like netselect-apt. Since our mirrors were only listed a few days, they did not attract a large number of Debian or Ubuntu users.
8.2 Package Versions and Vulnerability
To perform our analysis, we needed to know the distribution of package versions over time on the mirror, as well as which of those versions are vulnerable to attack. Information about the 58165 versions of the 3020 RPM CentOS packages used in the last year was captured. The update times were captured and used in the data set to determine if different versions existed on the main repository at the same time. This information was used to estimate compromises for those package managers that require a snapshot (packages that were all on a repository during the same time period).
Determining which package versions are remotely vulnerable to attack proved to be difficult and we are unaware of a data set that provides a good model of this. To compensate, we randomly chose a set of vulnerable versions from all non-current versions of packages. In practice, we believe that an attacker would be more likely to work to discover vulnerabilities in old versions of popular packages as these would allow the attacker to compromise more clients. This work does not capture this effect.
A client that uses a mirror with a package manager that has missing or inadequate security (Pacman, Slaktool, ports, YUM, and urpmi) is considered compromised whenever they install a package from the mirror. A client that uses a package manager that allows an attacker to mismatch vulnerable package versions (Portage) is considered compromised when it installs any package with a known vulnerability. For clients with snapshot vulnerability to replay and freeze attacks (Stork, APT, YaST, APT-RPM), the attacker chooses a snapshot time and can compromise clients who install a package that the attacker has a vulnerable version that was current at that time. We defer details of how the attacker chooses the snapshot time to a tech report [5].
Replay attacks only impact clients that are installing new packages instead of those that are updating existing packages. From the mirror’s request log, it isn’t possible to determine whether a client is installing or updating a package. This means that results for Portage, Stork, APT, APT-RPM, and YaST represent an upper bound. The package managers without security (Slaktool, ports, Pacman) are not impacted because an attacker can create arbitrary packages. Similarly, the results for YUM and urpmi are accurate because as long as there exists a vulnerable package the client hasn’t updated to a newer version, an extraneous dependencies attack can compromise the client.
8.3 Number of Compromised Clients
Using the CentOS trace and version information, the number of clients compromised by a malicious mirror was estimated (Figure 5). Note that this estimation doesn’t take into account the security practices of the distribution. As described in Section 7.6, users from distributions with additional security practices may be safer than what is described here. We provide information about the security of the package manager because many of the smaller distributions reuse package managers for more popular distributions and use only the security mechanisms provided by their package manager.
As the true number of vulnerable packages is not known, we use a value of 25 vulnerable packages (the effect of varying the number of vulnerable packages is shown in Figure 6). These plots show that the security model of the package manager has a great impact on the number of clients that can be compromised. A client that restricts attackers to mismatched vulnerable package versions reduces the maximum estimated number of compromised clients by about a factor of 4 to around 900 over the 13 day period. Package managers whose security mechanisms require a snapshot reduce the maximum estimated number of clients compromised to under 500. A package manager with signatures on the root metadata and protection against replay and freeze attacks (for example, modified Störk) will not have any compromises from an attacker that controls a mirror.
The package managers that allow the attacker to mismatch packages or that require a snapshot vary based upon the number of vulnerable packages. The effect of varying the number of vulnerabilities is shown in Figure 6. The number of clients with vulnerabilities that the attacker can exploit is on the x-axis. The plots show that package managers that allow an attacker to choose different versions of packages that existed on the root repository at different times (ismatch) are more vulnerable than package managers that require a snapshot.
This figure clearly shows that a package manager that requires an attacker to present a consistent set of packages provides better security than one that allows an attacker to mismatch packages. Somewhat disturbing is the significant number of clients that can be compromised if there are only 5 vulnerable package versions.
Our leased server was bandwidth-limited and mirrored multiple distributions simultaneously for cost reasons. An attacker would likely expend more bandwidth or set up multiple mirrors to capture additional traffic, thus leading to more compromises.
9. DISCLOSURE AND DISCUSSION
After disclosing these vulnerabilities to package manager developers via Ryan Giobbi at CERT [12], we were contacted by proactive developers at Gentoo and openSUSE who confirmed our findings and began working on fixes [14]. We were also able to get in touch with developers at CentOS, Red Hat, Fedora, Ubuntu, and Debian and had confirmation and discussion about how we recommend they fix the issues we uncovered [10]. We believe that most of the other distributions will follow suit and either patch the vulnerabilities themselves or ask their package manager developer to do so. In the interim, we provide some suggestions about how distributions and users can increase their security:
- If the package manager supports HTTPS and it correctly checks certificates, the distribution can set up repositories or mirrors that support HTTPS transfers. This will not protect against a malicious mirror, but will prevent a man-in-the-middle attack from an outsider.
- The distribution can review their mirror policy carefully and validate administrator credentials before putting a mirror on the official mirror list. The distribution should also review their policies for deciding how mirrors are allocated traffic to prevent a mirror from being able to target attacks (perhaps indirectly by advertising high bandwidth).
- The distribution’s mirrors should use a secure connection (e.g. SSH) to synchronize with the main repository to prevent an attacker from impersonating the main repository.
- Users should check that the versions of the packages their package manager recommends for installation are recent through multiple sources.
10. RELATED WORK
Many package managers have GUI front-ends [15, 27]. These GUI-based tools are usually just a different interface to the functionality provided by a command line package manager and so are identical from a security standpoint.
There are many techniques that help to support software security such as systems that ensure the authenticity and integrity of software (including SFS-RO [16], SUNDR [17], Deployme [20], and Self-Signed Executables [30]), and code signing certificates [6]. These are complimentary to the solutions presented in this paper.
There are several systems that access multiple mirrors to improve download performance and avoid DoS attacks. Byers et al. [4] describe using Tornado codes to improve performance by downloading from several mirrors simultaneously. This has the side-effect of allowing the client to make progress even if one of the mirror misbehaves. Sharma et al. [24] describe having the client “hop” between mirrors.
while downloading a file. This prevents an attacker from launching a DoS attack on the client because the attacker does not know which mirror the client will use next. We hope that raised awareness will result in organizations trying these techniques in practice.
11. CONCLUSION
This work identifies security issues in ten popular package managers in use today. Furthermore, we demonstrate that while some security mechanisms used by distributions may help to prevent attacks, others may actually decrease security. We estimate that when ignoring any additional security mechanisms, an attacker with a mirror that costs $50 per week can compromise between 150 and 1500 clients each week. These security issues have been disclosed to distributions that are currently working to fix these problems.
12. ACKNOWLEDGMENTS
We would like to thank our shepherd Steven Murdoch and the anonymous reviewers for their insightful comments and feedback. We would like to thank Ryan Giobbi at the CERT for helping us to responsibly disclose these issues to the package manager maintainers. We would like to thank Jake Edge, Dag Wieers, Kees Cook, Jan iankko Lieskowsky, Robin Johnson, Ludwig Nussel, Peter Poeml, Marcus Meissner, Josh Bressers, Tomas Hoger, Jason Guntherope, Kyriacos Pavlou, Mike Piatek, Chris Gniady, Wenjun Hu, and Tadayoshi Kohno for their comments on our research. We would also like to thank the members of the Stork project for their help in creating Stork and their input on this research. This work was funded in part by the PlanetLab Consortium.
13. REFERENCES
[1] Debian APT tool ported to Red Hat Linux. http://www.apt-get.org/.
[2] APT-RPM. http://apt-rpm.org/.
[3] Arch Linux (Don’t Panic) Installation Guide. http://www.archlinux.org/static/docs/arch-install-guide.txt.
[4] J. Byers, M. Luby, and M. Mitzenmacher. Accessing multiple mirror sites in parallel: using Tornado codes to speed up downloads. INFOCOM’99, Eighteenth Annual Joint Conference of the IEEE Computer and Communications Societies. IEEE, 1, 1999.
[5] J. Cappos, J. Samuel, S. Baker, and J. Hartman. A Look In the Mirror: Attacks on Package Managers. Technical Report TR-08-06, Department of Computer Science, University of Arizona, Jul 2008.
[6] Introduction to Code Signing. http://msdn2.microsoft.com/en-us/library/ms37361.asp.
[7] A. Crooks. The netbsd update system. In ATEC ’04: Proceedings of the USENIX Annual Technical Conference, pages 17–17, Berkeley, CA, USA, 2004. USENIX Association.
[8] debigs. What is debigs. http://linux.arboretum.com/cn/linux101/g/debsigs.htm.
[9] DistroWatch.com. Editorial: How Popular is a Distribution? http://distrowatch.com/weekly.php?issue=20070827#features.
[10] M. Domsch. Re: YUM security issues... https://www.redhat.com/archives/fedora-infrastructure-list/2008-July/msg00114.html.
[11] man dpkg-sig. http://pvet.fr/man/linux/commandes/dpkg_sig.
[12] R. Giobbi. Vulnerability Analysis Blog: Safely Using Package Managers. http://www.cert.org/blogs/vuls/2008/07/using_package_managers.html.
[13] J. Hughes. HughesJR.com – Attacks on Package Managers - ummm... http://www.hughesjr.com/content/view/22/i/.
[14] R. H. Johnson. [gentoo] Index of /users/robbat2/tree-signing-gleps. http://viewcvs.gentoo.org/viewcvs.py/gentoo/users/robbat2/tree-signing-gleps/.
[15] The KPackage Handbook. http://docs.kde.org/development/en/kdeadmin/kpackage/
[16] D. Mazières, M. Kaminsky, M. F. Kaashoek, and E. Witchel. Separating key management from file system security. In Proc. 17th SOSP, pages 124–139, Kauai Island Resort, SC, Dec 1999.
[17] D. Mazières and D. Shasha. Building secure file systems out of Byzantine storage. In PODC ’02: Proceedings of the twenty-first annual symposium on Principles of distributed computing, pages 108–117, New York, NY, USA, 2002. ACM.
[18] milw0rm - exploits vulnerabilities : videos : papers : shellcode. http://www.milw0rm.com.
[19] Netcraft: Strong growth for Debian. http://news.netcraft.com/archives/2005/12/05/strong_growth_for_debian.html.
[20] K. Oppenheim and P. McCormick. Deployme: Tellme’s Package Management and Deployment System. In Proc. 14th Systems Administration Conference (LISA ’00), pages 187–196, New Orleans, LA, Dec 2000.
[21] Gentoo-Portage http://gentoo-portage.com/.
[22] Installing Applications: Packages and Ports. http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/ports.html.
[23] RPM Package Manager. http://www.rpm.org/.
[24] P. Sharma, P. Shah, and S. Bhattacharya. Mirror hopping approach for selective denial of service prevention. Object-Oriented Real-Time Dependable Systems, 2003.(WORDS 2003). Proceedings of the Eighth International Workshop on, pages 200–208, 2003.
[25] Slackware Package Management. http://www.slacksite.com/slackware/packages.html.
[26] Stork. http://www.cs.arizona.edu/stork.
[27] Synaptic Package Manager - Home. http://www.nongnu.org/synaptic/.
[28] URPMI. http://www.urpmi.org/.
[29] dkg-sig support wanted? http://nixforums.org/about101637-asc-15.html.
[30] G. Wurster and P. van Oorschot. Self-Signed Executables: Restricting Replacement of Program Binaries by Malware. In 2nd USENIX Workshop on Hot Topics in Security, Boston, MA, Aug 2007.
[31] YaST - openSuSE. http://en.opensuse.org/YaST.
[32] Yum: Yellow Dog Updater Modified. http://linux.duke.edu/projects/yum/. |
Randomized Algorithms for Throughput-Optimality and Fairness in Wireless Networks
Atilla Eryilmaz, Eytan Modiano and Asuman Ozdaglar
Laboratory for Information and Decision Systems
Massachusetts Institute of Technology
Cambridge, MA, 02139
Emails: {eryilmaz, modiano, firstname.lastname@example.org
Abstract—In this paper, we study multi-hop wireless networks with general interference models, and describe a practical randomized routing-scheduling-congestion-control mechanism that is guaranteed to fully utilize the network capacity, and achieve fair allocation of the resources. Using the framework described in this paper, low complexity distributed algorithms can be developed for a large class of interference models and fairness criteria (such as proportional and max-min fairness). Earlier distributed algorithms proposed in this context have been highly interference model dependent, and can guarantee at most 50% utilization of the achievable throughput. This is the first work that assures 100% utilization and also provides fair allocation in multi-hop wireless networks.
I. INTRODUCTION
The control of data networks to achieve high throughput and fair allocation of the resources among competing users (or flows) is clearly one of the most important problems in data communications. There has been considerable interest and progress in the development of algorithms that address the issues of efficiency and fairness for both wireline and wireless networks.
In the case of fixed arrival rates (i.e. inelastic traffic), Tassiulas and Ephremides provided in their seminal work [25] a joint routing-scheduling algorithm that can achieve the highest throughput without violating the stability of the network. Such throughput-optimal policies make dynamic routing and scheduling decisions to avoid highly congested nodes. For wireline networks, these policies can be implemented in a distributed fashion by using buffer occupancy information of only the neighboring nodes. In contrast, in wireless multi-hop networks, there is no known scheme for low complexity implementation of throughput-optimal policies. Many other throughput-optimal algorithms developed later share the same weakness [1, 21, 19, 9].
It is well-recognized that reducing the complexity of the centralized computation is an essential requirement for the development and implementation of throughput-optimal policies. For the case of switches, Tassiulas has shown how the use of randomized algorithms can reduce the complexity of the centralized computation [24]. One of the main contributions of our work is to extend the use of randomized algorithms for potential distributed implementation of routing and scheduling in the context of multi-hop wireless networks.
The issue of fair service of elastic traffic, where the rate of an elastic flow is assumed to be controllable, is first considered by Kelly et al. [11, 12] in the context of wireline networks. The authors developed de-centralized algorithms that have strong ties to notions in market economics. The main idea behind these algorithms was for each source to measure the congestion level it experiences and to adjust its flow rate accordingly with its utility function in order to fully utilize the resources. These ideas have been extended to different scenarios and algorithms (e.g. [15], see [22] for a review). However, all of these works were developed for wireline networks, assumed fixed routes, and ignored the stochastic nature of the traffic.
More recently, it has been realized that ideas of flow control can be successfully utilized together with the dynamic routing and scheduling algorithm described above [13, 23, 7, 18, 8]. It has been shown that mean rates arbitrarily close to the fair allocation can be achieved without violating stability. However, the routing and scheduling part of the algorithm inherited the centralized optimization problem of [25], which is impractical to solve for multi-hop wireless networks.
The design and analysis of distributed implementations of the above cross-layer approach attracted a lot of attention from the community. In particular, [14, 27] provided algorithms that guarantee 50% utilization of the stability region for node-exclusive-spectrum-sharing (NESS) interference model, where each feasible allocation forms a matching\(^1\) of the graph. It has been shown in [3] that 33% can be guaranteed even when the algorithm operates in a totally asynchronous manner (see [2] for a definition). Distributed implementations exist for this particular interference model if a significant portion of the capacity is sacrificed. For other general interference models, it has been shown that the amount of sacrifice is even greater. For example, [28] and [4] consider the two-hop interference model\(^2\) and show that the guaranteed fraction of the capacity region drops with the increasing number of neighbors. In particular, for a grid topology, only 12.5% of the capacity region is achieved, which is very discouraging.
\(^1\)In a matching, no two links that are incident on the same node can be active simultaneously.
\(^2\)In the two-hop interference model, a transmission over a link \((n, m)\) is successful if and only if all the neighbors of \(n\) and \(m\) are silent at the time.
Our main contributions in this work are listed next.
- By extending the approach in [24] to the multi-hop wireless network scenario, we prove that the capacity region of the network can be fully utilized with a practical scheduling-routing algorithm. Thus, no fraction of the throughput need to be sacrificed for a practical implementation.
- In the case of elastic traffic, we propose a cross-layer mechanism that is composed of a decentralized congestion controller operating in parallel with the practical scheduling-routing algorithm discussed in the previous item. We prove that this mechanism asymptotically achieves fair division of the resources across the flows.
- All of our analysis considers a generic interference model that can be applied to a large set of graph theoretic collision models considered in the literature.
It is important to note that the proposed mechanism can be implemented with low complexity distributed algorithms for specific interference models of practical interest (see [6]).
The rest of the paper is organized as follows. In Section II, we describe the system model. Then, in Section III, we describe the practical algorithm and show its throughput-optimality. Section IV introduces the congestion control mechanism for elastic traffic and establishes its fair characteristics. We provide our concluding remarks in Section V, and some of the proofs in the Appendix.
II. SYSTEM MODEL
Consider a fixed wireless network that is represented by a graph \( \mathcal{G} = (\mathcal{N}, \mathcal{L}) \), where \( \mathcal{N} \) denotes the set of fixed nodes and \( \mathcal{L} \) denotes the set of undirected links. We assume a time slotted system with synchronized nodes, where each slot is just long enough to accommodate a single packet transmission. Suppose there is a set \( \mathcal{F} \) of end-to-end flows traversing the network, where each flow \( f \in \mathcal{F} \) is given by the node pair \( (s(f), d(f)) \). Here, \( s(f) \) denotes the source node of flow \( f \) and \( d(f) \) denotes the destination node. We allow for multiple routes between each source-destination pair. We consider both the inelastic and elastic traffic cases.
Associated with each destination a buffer is maintained at each node. We let \( q_{n,d}[t] \) denote the length of the queue at node \( n \) keeping packets destined for node \( d \) at the beginning of slot \( t \).
**Definition 1 (Stability).** A given queue is called stable if \( \mathbb{E}[q_{n,d}[\infty]] < \infty \), where \( q_{n,d}[\infty] \) denotes the random variable with distribution given by the steady-state distribution of \( \{q_{n,d}[t]\} \). The network is stable if all queues are stable; and unstable otherwise.
We consider a general interference model formulation that contains all of the graph theoretic collision models considered in the context of scheduling (e.g. NESS [20, 14, 27, 3], or two-hop interference models [28, 4]). We say that two links interfere if their concurrent transmissions collide, and assume that if two interfering links are activated in a slot, both of the transmissions fail. We use \( \mathcal{I}(l) \subset \mathcal{L} \) to denote the set of links that interfere with link \( l \). Typically, this set will contain links from the local neighborhood of \( l \). As an example, the matching constraint of the NESS model implies that for link \( l = (n,m) \), \( \mathcal{I}(l) := \{ l' \in \mathcal{L} : l' \in i(n) \text{ or } l' \in i(m) \} \), where \( i(n) \) denotes the set of links that are incident to node \( n \). We will prove our results for the general model.
We use \( \pi = \{\pi_{(n,m)}\}_{(n,m) \in \mathcal{L}} \) to denote a link activation (or allocation) vector, and \( \Pi \) denote the feasible set of allocations that complies with the interference constraints. An allocation is feasible if and only if no two links in the set interfere with each other. As an example, for the NESS interference model, \( \Pi \) corresponds to the set of matchings of \( \mathcal{G} \). We introduce the notation \( \pi^d_{(n,m)} \) to distinguish packets destined for different nodes: at any given slot, \( \pi^d_{(n,m)}[t] \in \{0, 1\} \) is 1 if link \( (n,m) \) serves a packet destined for node \( d \) in that slot, and 0 otherwise. Notice that \( \pi_{(n,m)}[t] = \sum_{d \in \mathcal{N}} \pi^d_{(n,m)}[t] \), for all \( t \). Also note that it is sufficient to restrict our attention to policies that sets \( \pi^d_{(n,m)}[t] \) to zero whenever \( q_{n,d}[t] = 0 \), for all \( (n,m) \in \mathcal{L} \), because any other policy that sets \( \pi^d_{(n,m)}[t] = 1 \) when \( q_{n,d}[t] = 0 \) can be replaced by a policy with \( \pi^d_{(n,m)}[t] = 0 \) without affecting the evolution of the queue-lengths. Then, we can write the evolution of a particular queue, say \( q_{n,d} \), as
\[
q_{n,d}[t+1] = \\
(q_{n,d}[t] - \pi^d_{out(n)}[t] + \sum_{f \in S_{n,d}} x_f[t] + \pi^d_{into(n)}[t]),
\]
where \( x_f[t] \) is the number of exogenous flow-\( f \) arrivals to the network in slot \( t \), and \( S_{n,d} \triangleq \{ f \in \mathcal{F} : s(f) = n, d(f) = d \} \) denotes the set of flows that start at node \( n \) and are destined to node \( d \). Also, \( \pi^d_{into(n)}[t] \triangleq \sum_{k:(k,n) \in \mathcal{L}} \pi^d_{(k,n)}[t] \) is a shorthand for the number of packets entering node \( n \) that are destined for node \( d \). Similarly, \( \pi^d_{out(n)}[t] \) is the number of packets leaving node \( n \) and are destined for node \( d \). We set \( q_{d,d}[t] = 0 \ \forall t \).
**Definition 2 (Capacity (Stability) Region).** Let \( \mathcal{G} = (\mathcal{N}, \mathcal{L}) \) be a given network and \( \Pi \) be the set of feasible allocations. The capacity (or stability) region \( \Lambda \) of the network is given by the set of vectors \( r = (r_f)_{f \in \mathcal{F}} \) for which there exists \( \pi^{d(f)}_{(n,m)} \geq 0 \), for all \( (n,m) \in \mathcal{L} \) and \( f \in \mathcal{F} \), such that both the flow conservation constraints at the nodes and the feasibility constraints are satisfied, as given below:
(C1) For all \( n \in \mathcal{N} \) and \( f \in \mathcal{F} \), we have\(^3\)
\[
r_f 1_{s(f)=n} + \sum_{k:(k,n) \in \mathcal{L}} \pi^{d(f)}_{(k,n)} = \sum_{m:(n,m) \in \mathcal{L}} \pi^{d(f)}_{(n,m)},
\]
(C2) \( [\pi_{(n,m)}]_{(n,m) \in \mathcal{L}} \in \text{Conv}(\Pi) \).\(^4\)
It is shown in [25, 19] that \( \Lambda \) is the set of mean arrival rates for which there exists a policy that stabilizes the network.
III. STABILITY FOR INELASTIC TRAFFIC
The focus of this section is throughput-optimality under stability for the inelastic traffic scenario. We provide simple
\(^3\)We use \( 1_A \) as the indicator function of event \( A \).
\(^4\)Conv\((A)\) denotes the convex hull of set \( A \), which is the smallest convex set that includes \( A \). The convex hull is included due to the possibility of timesharing between feasible allocations.
routing-scheduling mechanisms that can support any mean arrival rate in the capacity region without violating stability (such mechanisms are said to be throughput-optimal [25, 21, 19, 7]). In earlier work [24], Tassiulas used randomized schemes to provide a low complexity stabilizing algorithm for switches using a centralized controller. In this section, we will extend the use of randomized throughput-optimal schemes for multi-hop networks with general interference models. In particular, we show that practical algorithms satisfying several simple conditions can be designed to achieve full utilization of the capacity of multi-hop wireless networks with general interference models.
To this end, we introduce the following notation for link weights: \( w_{(n,m)}[t] = w_{(m,n)}[t] \triangleq \max_d |q_{n,d}[t] - q_{m,d}[t]| \). This is also referred to as the differential backlog\(^5\) over link \((n,m)\) and can be interpreted as a measure of the importance of the link. Consider the following allocation vector
\[
\pi_w^*[t] \in \arg \max_{\pi \in \Pi} \sum_{l \in L} w_l[t] \pi_l \equiv \arg \max_{\pi \in \Pi} (\mathbf{w}[t] \cdot \pi).
\]
(2)
This allocation rule is called the back-pressure policy. Once the \(\pi_w^*\) is determined according to (2), only the commodity that maximizes the differential backlog over link \((n,m)\) is served over that link at the chosen rate. If the back-pressure policy is applied at every time slot, it is known to be throughput-optimal. However, it requires a centralized controller that knows \(\mathbf{w}[t]\) at every time-slot, and also communicates the allocation vector \(\pi_w^*[t]\) instantly to all the nodes of the network. These requirements make the implementation of this algorithm impractical for the multi-hop wireless network scenario.
The idea is to use a random algorithm, instead of the optimal one described in (2), which yields a feasible allocation \(\tilde{\pi}[t] \in \Pi\) at every time slot, which is not necessarily equal to \(\pi_w^*[t]\), but has a positive probability \(\delta > 0\) of being equal to \(\pi_w^*[t]\). Thus, we have
\[
P(\tilde{\pi}[t] = \pi_w^*[t]) \geq \delta, \quad \text{for all } \mathbf{w}[t] \text{ and } t.
\]
(3)
Once the allocation \(\tilde{\pi}[t]\) is chosen, the actual allocation \(\pi[t]\) is determined according to the following evolution.
\[
\pi[t+1] = \begin{cases}
\pi[t] & \text{if } \mathbf{w}[t] \cdot \pi[t] \geq \mathbf{w}[t] \cdot \tilde{\pi}[t] \\
\tilde{\pi}[t] & \text{otherwise}
\end{cases}
\]
(4)
The above randomized algorithm was introduced in [24] in the context of switches, where there exists a centralized scheduler. A similar approach has been used in developing a distributed implementation for networks restricted to single-hop communication with matching constraints and Bernoulli arrival processes [17]. As we will prove shortly, it turns out that the two conditions (3) and (4) are sufficient to achieve throughput-optimality in a more general setting.
The algorithm updates its allocation vector only if the proposed allocation of the randomized algorithm yields a better objective function. Such an algorithm can be divided into two parts: PICK and COMPARE, where PICK chooses a feasible schedule \(\tilde{\pi}[t]\) satisfying (3), and COMPARE communicates the relevant weight information to other nodes in the network to perform (4). Distributed implementations of both PICK and COMPARE can be developed for a given interference model (see [6] for an example).
**Theorem 1.** Assume that \(x_f[t]\) is independent and identically distributed (i.i.d.)\(^6\) for all \(t\) and \(f\) with \(\mathbb{E}[x_f^2[1]] \leq A < \infty\). Then, for any mean arrival vector \(\lambda := (\mathbb{E}[x_f[1]])_f \in \text{interior}(\Lambda)\), the above randomized policy is stabilizing.
**Proof:** We start by noting that \(\mathbf{y}[t] \triangleq (\mathbf{q}[t], \pi[t])\) forms a Markov Chain. Then, we study the mean drift of the following Lyapunov function of the state \(\mathbf{y} = (\mathbf{q}, \pi)\):
\[
V(\mathbf{y}) = \sum_{n \in N} \sum_{d \in N} q_{n,d}^2 + \left( \sum_{l \in L} w_l((\pi_w^*)_l - \pi_l) \right)^2 \\
= \underbrace{\mathbf{q} \cdot \mathbf{q}}_{\triangleq V_1(\mathbf{y})} + \underbrace{(\mathbf{w} \cdot (\pi_w^* - \pi))^2}_{\triangleq V_2(\mathbf{y})}
\]
(5)
The proof uses the following two key lemmas the proofs of which are moved to Appendix A and B.
**Lemma 1.** Let \(\Delta_1(\mathbf{y}[t]) \triangleq \mathbb{E}[V_1(\mathbf{y}[t+1]) - V_1(\mathbf{y}[t]) \mid \mathbf{y}[t]]\), then for some \(\gamma > 0\) and \(c_1 < \infty\), we have
\[
\Delta_1(\mathbf{y}[t]) \leq -\gamma \sqrt{V_1(\mathbf{y}[t])} + 2\sqrt{V_2(\mathbf{y}[t])} + c_1.
\]
**Lemma 2.** Let \(\Delta_2(\mathbf{y}[t]) \triangleq \mathbb{E}[V_2(\mathbf{y}[t+1]) - V_2(\mathbf{y}[t]) \mid \mathbf{y}[t]]\), then for some \(\delta > 0\) and \(c_2, c_3 < \infty\), we have
\[
\Delta_2(\mathbf{y}[t]) \leq -\delta V_2(\mathbf{y}[t]) + c_2 \sqrt{V_2(\mathbf{y}[t])} + c_3.
\]
**Proof of Theorem 1 (Continued):** Combining these lemmas, we upper-bound \(\Delta(\mathbf{y}[t]) \triangleq \mathbb{E}[V(\mathbf{y}[t+1]) - V(\mathbf{y}[t]) \mid \mathbf{y}[t]]\) by
\[
-\gamma \sqrt{V_1(\mathbf{y}[t])} - \delta V_2(\mathbf{y}[t]) + (2 + c_2) \sqrt{V_2(\mathbf{y}[t])} + c_1 + c_3.
\]
Thus, if we consider the scenario where \(V(\mathbf{y}[t]) \geq B\) for a sufficiently large \(B < \infty\), then we can guarantee that
\[
\Delta(\mathbf{y}[t]) \leq -\epsilon \sqrt{V_1(\mathbf{y}[t])} + \hat{B},
\]
(6)
for some \(\epsilon > 0\) and \(\hat{B} < \infty\). This is true because in this scenario, since we have \(V_1(\mathbf{y}[t]) \geq B - V_2(\mathbf{y}[t])\), we can write
\[
\mathbb{E}[V(\mathbf{y}[t+1]) \mid \mathbf{y}[t]] \leq -\frac{\gamma}{2} \sqrt{V_1(\mathbf{y}[t])} - \frac{\gamma}{2} \sqrt{B - V_2(\mathbf{y}[t]) - \delta V_2(\mathbf{y}[t])} + (2 + c_2) \sqrt{V_2(\mathbf{y}[t])} + c_1 + c_3.
\]
(7)
Notice that the sum of the expressions in (7) and (8) can be made negative by choosing \(B\) large enough. Hence (6) holds when we let \(\epsilon = \gamma/2\) and choose \(\hat{B}\) large enough.
The proof of positive recurrence of the chain follows from Foster’s Criteria (cf. [22] or [16]). Let us use \(\mathbf{y}[\infty]\) to denote the random variable to which \(\{\mathbf{y}[t]\}_t\) converges. Due to the \(f\)-ergodic theorem of [16], the drift condition of (6)
---
\(^5\)This definition of differential-backlog is slightly different from the ones in the literature, which is due to the assumption of undirected links here.
\(^6\)The assumption of i.i.d. arrivals is not critical to the analysis. The same results continue to hold for processes with mild ergodicity properties ([10]).
is equivalent to \( \mathbb{E}[\sqrt{V_1(y[\infty])}] < \infty \). Invoking the definition of \( V_1(\cdot) \) results in the intended stability result.
Theorem 1 shows that under very mild conditions, if a randomized scheduler can be found that satisfies (3), and the control in (4) can be performed, then the stability will be achieved for any stabilizable incoming traffic.
IV. FAIR SERVICE OF ELASTIC TRAFFIC
In the previous section, we proved the throughput optimality of a randomized scheme for inelastic traffic, provided that the exogenous arrival rates are stabilizable. Thus, with that scheme, even if stability is achieved for a given mean arrival vector, the system may be seriously underutilized. Ideally, it would be desired to know the stability region of the network and choose the arrival rates to fully utilize the network resources. However, as noted in [26], the task of determining whether a given set of arrival rates is in the characterized stability region becomes an intractable problem as the network size grows.
In this section, we extend the throughput-optimal algorithm so that the available capacity of the network is fully utilized and fairly shared among the flows in a completely decentralized and dynamic fashion without the need to explicitly characterize the stability region. To that end, we change the considered traffic model from inelastic to elastic, where in the latter case the mean arrival rates of the sources can be modified.
To define fairness, as is standard in the recent literature (e.g. [11, 12, 15, 22, 7, 18, 13]), we use a utility function, \( U_f(\cdot) \), of mean arrival rates that is a measure of flow \( f \)'s preferences. Throughout, we will assume that this function is concave and non-decreasing. We call the allocation, \( x^* \), that satisfies
\[
x^* \in \arg \max_{x \in \Lambda} \sum_{f \in \mathcal{F}} U_f(x_f)
\]
the fair allocation. Notice that this is the allocation that maximizes the aggregate utility of the network. It is known that by defining \( U_f(\cdot) \) appropriately different fairness criteria of interest, such as proportional or max-min fairness, can be achieved (see [22] for an extensive review).
Next, we describe the so called Dual Congestion Control mechanism that is implemented at the source of each flow in a completely decentralized fashion. Variations of this mechanism are studied recently in the literature [7, 18, 14, 23].
**DUAL CONGESTION CONTROL MECHANISM:** Assume that every flow has access to the its entry point queue-length information, i.e. flow \( f \) knows \( q_{b(f), e(f)}[t] \) for all \( t \). Then, at the beginning of each time slot \( t \), flow \( f \) generates \( x_f[t] \) packets satisfying
\[
x_f[t] = \min \left\{ U_f^{-1} \left( \frac{q_{b(f), e(f)}[t]}{K} \right), M \right\},
\]
where \( M \) and \( K \) are positive scalars.
The policy is easy to implement at each source because it only requires the queue length of the buffer at the source, and the individual utility function of the flow. Note that the only common information required at all the sources is \( K \). Once \( K \) is determined, the flow control mechanism can operate in a completely decentralized fashion in parallel with the randomized routing-scheduling algorithm described in Section III.
**Theorem 2.** For some finite constants \( C_1, C_2 \) we have
\[
\sum_{n \in \mathcal{N}} \sum_{d \in \mathcal{N}} \overline{q}_{n,d} \leq C_1 K
\]
\[
\sum_{f \in \mathcal{F}} U_f(\bar{x}_f) \geq \sum_{f \in \mathcal{F}} U_f(x^*_f) - \frac{C_2}{K}
\]
where \( \bar{x}_f \triangleq \lim_{T \to \infty} \frac{1}{T} \sum_{t=0}^{T-1} \mathbb{E}[x_f[t]] \), and similarly for \( \overline{q}_{n,d} \).
**Proof:** We start by introducing a relaxed version of the optimization problem (9): for any \( \varepsilon > 0 \), let
\[
x^*(\varepsilon) \in \arg \max_{x \in \Lambda(\varepsilon)} \sum_{f} U_f(x_f),
\]
where \( \Lambda(\varepsilon) = \{ x \in \Lambda : [x_f + \varepsilon]_f \in \Lambda \} \). \( \Lambda(\varepsilon) \) is a subset of the capacity region \( \Lambda \) with an \( \varepsilon \)-strip of its positive surface is deleted. It is not difficult to see that \( x^*(\varepsilon) \to x^* \) as \( \varepsilon \to 0 \).
We use the same Lyapunov functions \( V(y), V_1(y) \) and \( V_2(y) \) introduced in Section III (c.f. (5)) to prove the theorem. The following theorem studies the single-step mean drift of \( V_1(y) \). The proof of this lemma is provided in Appendix C.
**Lemma 3.** For some constants \( \epsilon > 0, c_1 < \infty \),
\[
\Delta_1(y[t]) \leq -\varepsilon \sum_{n,d} q_{n,d}[t] + 2K \sum_{f} \mathbb{E}[U_f(x_f[t]) | y[t]]
\]
\[
-2K \sum_{f} U_f(x^*_f(\varepsilon)) - \frac{\varepsilon}{|\mathcal{N}|^2} \sqrt{V_1(y[t])} + 2\sqrt{V_2(y[t])} + c_1
\]
**Proof of Theorem 2 (Continued):** In what follows, we will omit \([t]\) for convenience. By combining the result of Lemma 3 with that of Lemma 2, we can write: \( \Delta(y[t]) \)
\[
\leq -\varepsilon \sum_{n,d} q_{n,d} + 2K \sum_{f} U_f(x_f)
\]
\[
-2K \mathbb{E}[\sum_{f} U_f(x^*_f(\varepsilon)) | y] + c_1 + c_3
\]
\[
-\varepsilon \frac{\sqrt{V_1(y)}}{|\mathcal{N}|^2} - \delta V_2(y) + (2 + c_2) \sqrt{V_2(y)}
\]
\[
\leq c_4 - \varepsilon \sum_{n,d} q_{n,d} + 2K \{ \sum_{f} \mathbb{E}[U_f(x_f)|y] - \sum_{f} U_f(x^*_f(\varepsilon)) \}
\]
where the inequality \((a)\) follows from the fact that (13) is upper-bounded as argued in the proof of Theorem 1.
By taking expectations of both sides of the last inequality and summing over \( t = 0, 1, \cdots, T - 1 \), we get
\[
\mathbb{E}[V(y[T]) - V(y[0])] \leq Tc_4 - \varepsilon \sum_{t=0}^{T-1} \sum_{n,d} \mathbb{E}[q_{n,d}[t]]
\]
\[
+2K \sum_{t=0}^{T-1} \sum_{f} \mathbb{E}[U_f(x_f[t])] - 2KT \sum_{f} U_f(x^*_f(\varepsilon)).
\]
Since \( V(y) \geq 0 \) for all \( y \), we can re-arrange the terms in (14) to obtain the following inequality.
\[
\frac{1}{T} \sum_{t=0}^{T-1} \sum_{n,d} \mathbb{E}[q_{n,d}[t]] \leq \frac{V(y[0])}{\varepsilon T} + \frac{c_4 + 2M|F|K}{\varepsilon}.
\]
The proof of (11) is complete when we let \( T \to \infty \), and define \( C_1 \triangleq \frac{2M|F|}{\varepsilon} \).
If, on the other hand, we re-arrange the terms in (14) differently, we can get
\[
\frac{1}{T} \sum_{t=0}^{T-1} \sum_f \mathbb{E}[U_f(x_f[t])] \geq \sum_f U_f(x^*_f(\varepsilon)) - \frac{V(y[0])}{2KT} - \frac{c_4}{2K}.
\]
Hence, the proof of (12) is complete when we use Jensen’s inequality to write \( U_f(\bar{x}_f) \geq \frac{1}{T} \sum_{t=0}^{T-1} \mathbb{E}[U_f(x_f[t])] \), and define \( C_2 \triangleq \frac{c_4}{2} \).
\section{Conclusions}
In this work, we provided a framework for the design of practical cross-layer algorithms for multi-hop wireless networks that are throughput-optimal and fair. Previously proposed low complexity implementations for multi-hop wireless networks have been greedy in nature. Despite their ease of implementation, such policies have been shown in recent works to have poor throughput performance. The greatest weakness of greedy policies is their lack of memory. In this work, we considered the generalization of a practical randomized policy first introduced in [24] for switches to multi-hop networks. We proved that with the use of a small memory unit at the nodes, this policy achieves 100% of the available capacity of the network. We also considered the case of elastic flows and provided a decentralized congestion control algorithm that works in parallel with the randomized algorithm. We showed that the resulting cross-layer algorithm allocates resources fairly across users.
We note that the extra overhead that will appear in our randomized algorithm can be reduced at the expense of larger delay, but \textit{without sacrificing from throughput}. Also, our initial findings suggest that the additional complexity is comparable to the complexity of the greedy policies. In our companion paper [6], we show that the development and analysis of low complexity, distributed PICK and COMPARE algorithms (as described in Section III) for the two-hop interference model.
\section*{Appendix}
\subsection*{A. Proof of Lemma 1:}
\textit{Proof:} We start by arranging the terms of \( \Delta_1(y[t]) \).
\[
\Delta_1(y[t]) = \mathbb{E}[(q[t+1] - q[t]) \cdot (q[t+1] + q[t]) | y[t]]
\]
\[
\overset{(a)}{\leq} c_1 + 2 \sum_{n,d} q_{n,d}[t](\lambda_{n,d} + \pi^d_{into(n)}[t] - \pi^d_{out(n)}[t])
\]
\[
+ 2 \sum_{n,d} q_{n,d}[t](\pi^d_{into(n)}[t] - \pi^d_{out(n)}[t]
\]
\[
- \pi^d_{into(n)}[t] + \pi^d_{out(n)}[t]),
\]
where inequality \((a)\) is due to the fact that \( \mathbb{E}[x_f^2[t]] \leq A \), and that \( \pi^d_{into(n)}[t] \) and \( \pi^d_{out(n)}[t] \) are both upper-bounded by the maximum degree of \( G \). In (15) and (16), the optimal allocation \( \pi^*_w \) is added and subtracted. For notational convenience, we use \( \lambda_{n,d} \triangleq \sum_{f \in S_{n,d}} \lambda_f \). Also, \( \pi^d_{into(n)} \) and \( \pi^d_{out(n)} \) are defined similarly to \( \pi^d_{into(n)} \) and \( \pi^d_{out(n)} \).
Since the arrival rate vector \( \lambda \) is known to be within the stability region \( \Lambda \), we can upper-bound the second expression in (15) by \( -\gamma \sqrt{V_1(y[t])} \) (e.g. see [5]). Next, notice that
\[
(16) = 2w[t] \cdot (\pi^*_w[t] - \pi[t]) \leq 2\sqrt{V_2(y[t])}.
\]
Combination of these results yields the proof.
\subsection*{B. Proof of Lemma 2:}
\textit{Proof:} We start by analyzing the conditional expectation
\[
\mathbb{E}[V_2(y[t+1]) | y[t]] \leq (1 - \delta) \mathbb{E}[(w[t+1]
\]
\[
\cdot (\pi^*_w[t+1] - \pi[t+1]))^2 | y[t], \pi[t+1] \neq \pi^*_w[t+1]].
\]
Next, we write the evolution of the weight vector \( w[t] \) as \( w[t+1] = w[t] + r[t] - z[t] \), where \( r[t] \) denotes the vector of packets entering the queues and \( z[t] \) is the vector of packets leaving the queues. We can describe these vectors more precisely, but this is not necessary for the proof. The only important factor is the boundedness of them. Now, we can write \( \mathbb{E}[V_2(y[t+1]) | y[t]] \)
\[
\leq (1 - \delta) \mathbb{E}[(w[t] \cdot (\pi^*_w[t+1] - \pi[t+1])
\]
\[
+ (\pi^*_w[t+1] - \pi[t+1])
\]
\[
\cdot (r[t] - z[t]))^2 | y[t], \pi[t+1] \neq \pi^*_w[t+1]].
\]
It is not difficult to upper-bound the expression in (18) with a finite constant. To bound (17), we make two observations:
(i) \( w[t] \cdot \pi^*_w[t+1] \leq w[t] \cdot \pi^*_w[t] \) due to the definition of the optimal allocation vector \( \pi^*_w \) in (2).
(ii) \( w[t] \cdot \pi[t+1] \geq w[t] \cdot \pi[t] \) due to the update rule (3) of the randomized algorithm.
By combining these observations, we can find constants \( c_2, c_3 < \infty \) that satisfy
\[
\mathbb{E}[V_2(y[t+1]) | y[t]] \leq (1 - \delta)V_2(y[t]) + c_2 \sqrt{V_2(y[t])} + c_3,
\]
which completes the proof.
\subsection*{C. Proof of Lemma 3:}
\textit{Proof:} The proof starts with the following upper bound on \( \Delta_1(y[t]) \) (cf. (15) and (16)):
\[
\Delta_1(y[t]) \leq c_1 + 2\sqrt{V_2(y[t])} + 2\sum_{n,d} E[q_{n,d}[t] \sum_{f \in S_{n,d}} x_f[t] | y[t]] \\
- 2 \sum_{(n,m) \in L} \sum_d \pi^d_{(n,m)}[t](q_{n,d}[t] - q_{m,d}[t]) \\
= c_1 + 2\sqrt{V_2(y[t])} + 2\sum_f K E[U_f(x_f[t]) | y[t]] \\
- 2 \sum_f E[KU_f(x_f[t]) - q_{b(f),e(f)}x_f[t] | y[t]] \\
- 2 \sum_{(n,m) \in L} \sum_d \pi^d_{(n,m)}[t](q_{n,d}[t] - q_{m,d}[t]),
\]
where we get the last equality by adding and subtracting \(2K \sum_f E[U_f(x_f[t]) | y[t]]\), and by noting that
\[
\sum_{n,d} q_{n,d}[t] \left( \sum_{f \in S_{n,d}} x_f[t] \right) = \sum_f q_{b(f),e(f)}[t]x_f[t].
\]
Notice that the Dual Congestion Control mechanism is designed to optimize the expression in (19), and the Back-pressure Policy is designed to optimize (20). Bounds to these expressions are obtained in [18], which are reproduced below.
\[
\sum_f E[KU_f(x_f[t]) - q_{b(f),e(f)}x_f[t] | y[t]] \\
\geq K \sum_f U_f(x^*_f(\varepsilon)) - \sum_f q_{b(f),e(f)}[t]x^*_f(\varepsilon)
\]
\[
\sum_{(n,m) \in L} \sum_d \pi^d_{(n,m)}[t](q_{n,d}[t] - q_{m,d}[t]) \geq \sum_{n,d} q_{n,d}[t](x^*(\varepsilon) + \varepsilon).
\]
The first inequality is straight-forward since \(x^*(\varepsilon)\) is an element of the set of feasible rate vectors (i.e. vectors in \([0, M]^{|F|}\)) over which the Congestion Controller of (10) maximizes the the expression on the left. The second inequality follows from the fact that \(x^*(\varepsilon) = (x_f(\varepsilon))_f\) is at least \(\varepsilon\) away from the boundary of \(\Lambda\) for all \(f\).
By canceling common terms, we obtain
\[
\Delta_1(y[t]) \leq -2\varepsilon \sum_{n,d} q_{n,d}[t] + 2K \sum_f E[U_f(x_f[t]) | y[t]] \\
- 2K \sum_f U_f(x^*_f(\varepsilon)) + 2\sqrt{V_2(y[t])} + c_1.
\]
Finally, we use the following lower bound \(\sum_{n,d} q_{n,d}[t] \geq \frac{\sqrt{V_1(y[t])}}{|N|^2}\), which, when substituted into the previous expression, yields the statement of the lemma.
\[\blacksquare\]
**References**
[1] M. Andrews, K. Kumaran, K. Ramanan, A. Stolyar, R. Vijayakumar, and P. Whiting. Scheduling in a queueing system with asynchronously varying service rates, 2000. Bell Laboratories Technical Report.
[2] D. P. Bertsekas and J. N. Tsitsiklis. *Parallel and Distributed Computation: Numerical Methods*. Athena Scientific, Belmont, MA, 1997.
[3] L. Bui, A. Eryilmaz, R. Srikant, and X. Wu. Joint asynchronous congestion control and distributed scheduling for wireless networks. Proceedings of IEEE Infocom 2006.
[4] P. Chaporkar, K. Kar, and S. Sarkar. Throughput guarantees through maximal scheduling in wireless networks. In *Proceedings of the Allerton Conference on Control, Communications and Computing*, 2005.
[5] A. Eryilmaz. Efficient and fair scheduling for wireless networks. PhD thesis, University of Illinois at Urbana, Champaign, August 2005.
[6] A. Eryilmaz, A. Ozdaglar, and E. Modiano. Polynomial complexity algorithms for full utilization of multi-hop wireless networks, 2006. LIDS Report, submitted.
[7] A. Eryilmaz and R. Srikant. Fair resource allocation in wireless networks using queue-length based scheduling and congestion control. In *Proceedings of IEEE Infocom*, Miami, FL, March 2005.
[8] A. Eryilmaz and R. Srikant. Resource allocation of multi-hop wireless networks. In *Proceedings of International Zurich Seminar on Communications*, February 2006.
[9] A. Eryilmaz, R. Srikant, and J. Perkins. Stable scheduling policies for fading wireless channels. In *Proceedings of IEEE International Symposium on Information Theory*, 2003. A longer version appeared in the *IEEE/ACM Transactions on Networking*, 2005.
[10] A. Eryilmaz, R. Srikant, and J. R. Perkins. Stable scheduling policies for fading wireless channels. *IEEE/ACM Transactions on Networking*, 13:411–425, April 2005.
[11] F. P. Kelly. Charging and rate control for elastic traffic. *European Transactions on Telecommunications*, 8:33–37, 1997.
[12] F. P. Kelly, A. Maulloo, and D. Tan. Rate control in communication networks: Shadow prices, proportional fairness and stability. *Journal of the Operational Research Society*, 49:237–252, 1998.
[13] X. Lin and N. Shroff. Joint rate control and scheduling in multihop wireless networks. In *Proceedings of IEEE Conference on Decision and Control*, Paradise Island, Bahamas, December 2004.
[14] X. Lin and N. Shroff. The impact of imperfect scheduling on cross-layer rate control in multihop wireless networks. In *Proceedings of IEEE Infocom*, Miami, FL, March 2005.
[15] S. H. Low and D. E. Lapsley. Optimization flow control, I: Basic algorithm and convergence. *IEEE/ACM Transactions on Networking*, 7:861–875, December 1999.
[16] S. Meyn and R. Tweedie. *Markov Chains and Stochastic Stability*. Springer-Verlag, 1993.
[17] E. Modiano, D. Shah, and G. Zussman. Maximizing throughput in wireless networks via gossiping. In *ACM SIGMETRICS/FIP Performance*, 2006.
[18] M.J. Neely, E. Modiano, and C. Li. Fairness and optimal stochastic control for heterogeneous networks. In *Proceedings of IEEE Infocom*, pages 1723–1734, Miami, FL, March 2005.
[19] M.J. Neely, E. Modiano, and C.E. Rohrs. Dynamic power allocation and routing for time varying wireless networks. In *Proceedings of IEEE Infocom*, pages 745–755, April 2003.
[20] G. Sasaki and B. Hajek. Link scheduling in polynomial time. *IEEE Transactions on Information Theory*, 32:910–917, 1988.
[21] S. Shakkottai and A. Stolyar. Scheduling for multiple flows sharing a time-varying channel: The exponential rule. *Translations of the AMS, Series 2*, A volume in memory of F. Karpelevich, 207:185–202, 2002.
[22] R. Srikant. *The Mathematics of Internet Congestion Control*. Birkhäuser, Boston, MA, 2004.
[23] A. Stolyar. Maximizing queueing network utility subject to stability: Greedy primal-dual algorithm. *Queueing Systems*, 50(4):401–457, 2005.
[24] L. Tassiulas. Linear complexity algorithms for maximum throughput in radio networks and input queued switches. In *Proceedings of IEEE Infocom*, pages 533–539, 1998.
[25] L. Tassiulas and A. Ephremides. Stability properties of constrained queueing systems and scheduling policies for maximum throughput in multihop radio networks. *IEEE Transactions on Automatic Control*, 36:1936–1948, December 1992.
[26] L. Tassiulas and A. Ephremides. Dynamic server allocation to parallel queues with randomly varying connectivity. *IEEE Transactions on Information Theory*, 39:466–478, March 1993.
[27] X. Wu and R. Srikant. Regulated maximal matching: A distributed scheduling algorithm for multi-hop wireless networks with node-exclusive spectrum sharing, 2005. Submitted to *IEEE Conference on Decision and Control*.
[28] X. Wu and R. Srikant. Bounds on the capacity region of multi-hop wireless networks under distributed greedy scheduling. In *Proceedings of IEEE Infocom*, 2006. |
The people who walked in darkness have seen a great light; those who lived in a land of deep darkness—on them light has shined. 2 You have multiplied the nation, you have increased its joy; they rejoice before you as with joy at the harvest, as people exult when dividing plunder. 3 For the yoke of their burden, and the bar across their shoulders, the rod of their oppressor, you have broken as on the day of Midian. 4 For all the boots of the trampling warriors and all the garments rolled in blood shall be burned as fuel for the fire. 5 For a child has been born for us, a son given to us; authority rests upon his shoulders; and he is named Wonderful Counselor, Mighty God, Everlasting Father, Prince of Peace. 6 His authority shall grow continually, and there shall be endless peace for the throne of David and his kingdom. He will establish and uphold it with justice and with righteousness from this time onward and forevermore. The zeal of the LORD of hosts will do this. Isaiah 9:2-7
“Stir Us Up” is our theme for Advent this year at Augustana as we move into the month of December with great hope and expectation, thanksgiving and gratitude for all that God is doing in this congregation for the common good and in the world in which we live.
Our calling in the Christian Community is to proclaim the good news of God in Christ through word and deed as we seek to serve all people following the example of Jesus while striving for justice and peace in all the earth. As we call on our nation to welcome all refugees and to end the violence, and as we celebrate our diversity, worship is vital.
Gathering in Worship together is the first action we take each week to give thanks to God for the unconditional love we have experienced in Jesus Christ and to be stirred up to be about the ministry of our daily lives both individually and together. In worship we gather with God’s faithful people to continue in the covenant God made with us in Holy Baptism to embrace the life-giving gifts of God in the hearing of the word and the sharing of the Lord’s supper.
We can’t underscore the importance of worship enough as the Holy Spirit gathers us, enlightens us and then sends us out for our ministry of healing and hope, compassion and forgiveness, justice and peace, equity and inclusion, reconciliation and renewal. For those who may have been away from worship for a few weeks or even months due to life’s changes and challenges, the next Sunday on the calendar is always the right Sunday to come and celebrate that the Lord is good!
Advent, the beginning of the new church year, is an ideal time to invite others to join us in our life together at Augustana as we give ourselves away for the life of the world. It is a wonderful time to go deeper and wider in the sharing of our time, talents and treasures to the glory of God. My favorite mentors in giving in response to God’s love always said they never regretted giving generously to the work of God’s church for the sake of the world because God has been so generous to us. To give is to live with generosity which fills us with love and renews our spirits. It is simply good for us.
As we pray for God to “Stir Us Up” even more at Augustana we will gather around weekly themes; the 1st Sunday of Advent being as “People of the Promise,” including the Intergenerational Advent Festival at 10a and the Rose City Mixed Quartet at the 6p Jazz Service, and the blessing of Prayer Shawls at 8:30 and 11 a.m. On the 2nd Sunday of Advent our theme is “You Are My Messengers” as we encourage one another to reach out to others with the good news of God’s love in Jesus Christ inviting them into God’s community of hope. Extra copies of the enclosed Advent Schedule are available to give away.
On the 3rd Sunday of Advent we will ask the question “Have You Rejoiced or Exulted Lately” as we celebrate the goodness of God and hear the Christmas Story through our Annual Children’s Christmas Pageant at 10a and woven through the 6pm Jazz Service. On the 4th Sunday of Advent we’ll join Mary the mother of Jesus in the Magnificat “My Soul Magnifies the Lord” from the first chapter of Luke throughout our day of Worship.
Then On Christmas Eve we’ll have three incredible services at 7:00, 9:00 and 11:00pm and Christmas Day at 9:30am singing Glory to God in the Highest and Peace to All People on Earth. Our sermon Christmas Eve will be “Stir Us Up, O Prince of Peace.” We are expecting 1200 to 1400 worshipers at our Christmas Services and all will leave with hope and the renewed call to Go Tell It On The Mountain that God loves all. Yes “Stir Us Up” Lord Jesus to do your work for the life of the world.
Christ’s Peace and Joy,
Pastor Mark
Year-end Giving: This has been an incredible year of ministry at Augustana. Over the next four weeks we have a major goal to achieve to complete the year. Like many churches and non-profits, our December giving is what gets us to the finish with a large amount of our annual giving coming now. We enter December in the midst of our annual year-end push with $100,000 in offerings and special gifts to do all we set out to do and meet our obligations; we will keep you posted each week. It will take all of us together. We have shown in the past we can do this to answer God’s call. Now is always the time to place the Lord’s church first in our year-end giving. All of us are needed this year, members and friends of Augustana, locally and around the country and world. As always, we know that God will provide. Our giving is an act of worship that takes us deeper on our spiritual journey.
Pledge cards for 2019 will be dedicated at worship on Sunday, December 23. We encourage each household to make a pledge this year.
CARE OF CREATION HINTS FOR THE HOLIDAYS
When decorating, consider …
-- not purchasing a flocked tree (can’t be composted)
-- using items from nature
-- using LED lights and putting them on timers
When purchasing gifts, consider …
-- buying local, sustainable items
-- focusing on simple and meaningful (maybe handmade?)
-- doing a gift drawing within your family (each person buys a gift for one person)
When wrapping gifts, consider …
-- not buying foil or glossy paper (can’t be recycled)
-- getting creative with brown paper bags or newspaper
-- saving bows and gift bags for reuse
The 2018 Augustana Gala Auction was a tremendous success and set a number of records. The event sold out in early September (a record 172 tickets were sold). The ballroom was abuzz as guests sampled appetizers, sipped on beverages, chatted with one another, and bid on a record number of silent auction items. Oohing and aahing could be heard as guests checked out the fabulous dessert dash display. Fun was had by those who tried their hand at the wine ring toss; a record 126 bottles were donated. The total raised after auction operating expenses is $51,863.39, an increase of $10,350.21 over 2017. $11,250 was disbursed to Lutheran Community Services Northwest’s Intensive Case Management Program ($9,750 from the paddle raise and $1,500 from the Augustana Foundation). $11,408 was raised for a new Wolf range and ovens for the fellowship hall kitchen. The range has been purchased and should be installed in early December.
While auction night is just one evening out of the year, its impact extends well beyond that single night. It is seen throughout the year in the many important ministries and advocacies Augustana provides or is involved in. Our church community is broadened by the friends and family we invite to attend. And, the dinners/events which occur throughout the year promote and develop friendships, both old and new.
THANK YOU to everyone who donated, volunteered, attended the auction, and prayed for its success. -Janis Bell and Chris Grumm
We pray for:
Francisco Aguirre, Dora Reyna & family Ron Beck Chris Bergerson Beth Boyce Alison Embler-Brown Allen Brown, Jr. Rev. Tom Campbell-Schmitt Lou Carman Nick & Sara Drew Keith Erdmann Deaunka Geiger Norene Goplen Brian Gregory Shirley Greyfox George Hall
Kelsey Hansen Leah Higgins Al Hornish, back surgery Jenny Isabella Eddie Johnson April Lane Don Lombard Peter Minus Nicole Mongrain Ann Nelson Scott Niedermark D’Norgia Price Carol Reitz Andrea Robideau Gary Rosenthal Becky Roth Jenny Schmidt Jerrime Shelton
Erin Spadelli Mary Jo Swalwell (Kevin Smith’s mother) Fred Trujillo, Sr. Jim Turner, hip surgery Robert Wallace Luci Wells Miles Wilson
We pray for all members of the armed services during these tenuous times in the world.
We pray for restoration of all communities devastated by war, disease, weather and disaster; for all in need throughout the world that they may receive help and hope; for peace, tolerance, love and respect for all; for at-home members of our faith community: Signe Carlson
Please let the office and Pastor Knutson know if you are having surgery or are being hospitalized by calling the office, and pastoral care visits will be arranged. Please call any time about other pastoral care needs.
Prayer List: If you or someone you know wishes to be on the prayer list (with their permission) please let the office know. We list only those who request to be listed.
Worship each Sunday at 8:30 & 11 a.m., 6 p.m.
Christian Education each Sunday at 9:45 a.m.
Sunday, December 2 I Advent
Intergenerational Advent Festival, 9:50-10:50 a.m.
Rose City Mixed Quartet at Jazz Service, 6 p.m.
Jeremiah 33:14-16; Psalm 25:1-10; 1 Thessalonians 3:9-13; Luke 21:25-36
Sunday, December 9 II Advent
Malachi 3:1-4; Luke 1:68-79; Philippians 1:3-11; Luke 3:1-6
Sunday, December 16 III Advent
New Members Received
Christmas Pageant Dress Rehearsal 9:50 a.m.
Christmas Pageant with Jazz Service, 6 p.m.
Zephaniah 3:14-20; Isaiah 12:2-6; Philippians 4:4-7; Luke 3:7-18
Sunday, December 23 IV Advent
Pledges Dedicated
Micah 5:2-5a; Luke 1:46b-55; Hebrews 10:5-10; Luke 1:39-45
Sunday, December 30 I Christmas
1 Samuel 2:18-20, 26; Psalm 148; Colossians 3:12-17; Luke 2:41-52
Adult Forum
Advent Festival for all ages
Living Advent with Mary Anne Turner
Children’s Pageant dress rehearsal for all ages
No Forum
No Forum
Monday, December 24 CHRISTMAS EVE Candlelight Carol Services
“Stir Us Up, O Prince of Peace”
7 p.m. (Jazz/Gospel) 9 p.m. (Jazz/Gospel) 11 p.m. (Classical)
Tuesday, December 25: Christmas Day 9:30 a.m. Worship
8:30 Contemporary Choir rehearses
6:15-7:15 p.m. on Thursdays, directed by Andy Fort
December 13th 6:30 - 7:15 p.m. COME SING WITH US! The 8:30 Choir is having another special invitation choir performance on December 16th. If you’ve ever thought about singing in the choir, come join us! We will have one rehearsal on December 13th from 6:30 - 7:15 p.m. in the sanctuary, and then a performance the following Sunday, December 16th. We will use a special arrangement of People Look East, guaranteed to be both easy and enjoyable. See you there!
11:00 Sanctuary Choir New members always welcome! If you are not sure, try it anyway! We give free tune-ups! 7:20-9 p.m. Thursdays, directed by Dr. Walter Krueger
Sunday School Choir rehearses during Sunday School time, directed by Andy Fort
Handbell Choir rehearses Wednesdays, 7:15-8:45 p.m., directed by Joyce Hornish
Come and make a joyful noise!
Augustana, a Sanctuary Congregation since 1996: See more information about the cutting-edge work of the Evangelical Lutheran Church in America and the Lutheran Immigration and Refugee Service on immigration reform at www.elca.org. The Lutheran community has been a major mover on immigration for decades.
Augustana is a Sanctuary Congregation recognizing that we are on holy ground in the heart of the city. For more information on the Sanctuary Movement, see the Augustana website and the Interfaith Movement for Immigrant Justice website.
Weekly bulletin announcements received in the church office by noon Tuesday will be included for the following Sunday. Monthly newsletter announcements received by noon on the third Monday will be included for the following month. Email announcements directly to Joyce Kindschuh at email@example.com
St. Francis Dining Hall Thanksgiving Meal: A big thank-you to the 65 Augustana volunteers of all ages who worked from 9:30-11:30 a.m. on Thanksgiving Day to assist in the preparation of the meal for over 300 sisters and brothers in Portland who are homeless.
Thanks Augustana family and friends! 652 cans of food were donated for the NE Emergency Food Program, which more than qualifies for the Thrivent matching gift of $500 which will go to NEFP. Excellent work, everyone!
Augustana Giving Tree: We will again be helping clients of the HIV Day Center of Ecumenical Ministries of Oregon. Please take an “ornament” from the display in the narthex, and bring the requested item back no later than Dec. 16. Thank you so much for your generosity!
Advent Surprises
Our Christmas Trees are up and they look beautiful! A big thank you to Kathy and John Kirsch for getting them again this year, and to those who helped set them up and decorate.
The Sanctuary has been transformed once again to the wonderful blue of Advent. A big thank you to the Sanctuary Team.
Our new Wolf oven with six burners, a griddle and two ovens is now in, thanks to Marty Bell and his crew including George Hall, Ansel Johnson, Jerry and Colleen Benson and the employees of Rose’s Equipment and Supply. List price for this oven is $12,000 retail, $9,800 wholesale, and it was purchased for $7,250!! We offered a prayer for our old Wolf ovens of 52 years.
Marilyn Keller is again a key musician in the wonderful Gospel Christmas downtown, Dec. 7, 8, 9. She can get group ticket prices before it sells out.
Susan Johnson is in Inherit the Wind at the Lakewood Theater in Lake Oswego. Contact the theater for ticket info.
Next New Members Class: Sunday, December 9, 9:50 a.m. New members will be received on Sunday, December 16.
The memorial service for Kaydlynn Kleber was Saturday, Nov 10.
Church Council Christmas party is Tuesday, December 11; all are welcome.
Wildfires in California: Lutheran Disaster Relief is on the ground assisting. Special offerings may be placed in the offering plate with the note Wildfire disaster response.
Project Lemonade collects hats, gloves, and scarves for foster children in Oregon and SW Washington, and Thrivent will help. Please donate items in the marked box in the narthex! For further info, contact Elizabeth Morse (503) 706-1285.
Your 2019 Offering envelopes are available to pick up in the narthex. Please find your set and take home. It costs over $3 each to mail them. If you’d like envelopes and don’t have a set please stop by the church office and we’ll give you one.
Stewardship Pledge Cards will be available soon. Pledges will be dedicated on Sunday, December 23.
Nominations for Church Council and Foundation are being accepted through Friday, December 7. Forms are available in the office and the narthex. The Nominating Committee’s first meeting is Saturday, December 8. The Vision & Budget Team’s first meeting is Saturday, Dec. 8.
New in Augustana’s Library: The Qur’an and the Bible. Gabriel Said Reynolds, section 11. Miracles. Luke Timothy Johnson, section 4. Queer Theology. Lina Marie Lonstad, section 26. A Secular Age. Charles Taylor, section 24. And many more! Come check them out! (And don’t overlook the Fiction section, either.)
Lift Every Voice Oregon, which is officed at Augustana, has two pieces of legislation in Legislative Council for the 2019 season. The first is to ban the sale of assault weapons and the second is to ban the sale of large capacity magazines in Oregon. Much more to come in January. Volunteers are always welcome.
Friend in Need: Each year, Augustana gives thousands of dollars in direct aid including bus tickets and food to those who live on the streets or are on the edge. Gifts to the Augustana Friend in Need fund are always welcome.
Food is always needed for our food cupboard here as well as for NE Emergency Food Program. Bring donations to the narthex whenever you’re able.
Members of Augustana receive The Lutheran magazine as part of our congregation subscription at a yearly cost of $8 per household, paid by Augustana. Please give what you can to help offset the costs of our congregation subscription. Thanks!
Volunteer Here, We Need You!
For a fun, spirit-filled service opportunity, come help with parament changes and sanctuary preparation for Advent and Christmas. There will be a Sanctuary Work Party from 8:30-10a on Saturday Dec. 22. We’ll also be working on Dec. 24 (Christmas Eve) from 8:30-10am. No special skills needed, just willing hearts and hands.
The 11 am service needs greeters and adult acolytes. Please contact Sue Van Winkle if you are willing to serve in either of these important roles. Thank you. firstname.lastname@example.org
Moses Movers seeks two volunteers to help deliver food to the Northeast Food Pantry once a month on a Tuesday or Saturday morning from 9:00-10:30am. Call the church office 503-288-6174.
Saturday Breakfast Work Party: Join Ansel Johnson for breakfast on Saturdays from 8-9 a.m. and stay for about an hour to help with projects in and outside the church. Have a great breakfast, make new friends and help out Augustana. Many thanks to Marty Bell and all the property volunteers.
Advent Fair, December 2 at 9:50 a.m. in the Fellowship Hall
The 1st Sunday of Advent is December 2 and is our annual Advent Fair. This is a multi-generational, preparation time for all. Come and join the fun downstairs and make your Advent Log, Christmas presents for the birds, other crafts and enjoy delicious munchies. Decorate cookies, get your Advent calendar and browse Advent books that Beth has brought from the library. There is no better way to begin a meaningful season of Advent and Christmas than to join in the learning and fun. This is a great time to bring friends, family and neighbors to Augustana and get into the spirit.
Children’s Christmas Pageant: Sunday December 16th at 10 a.m. in the Sanctuary and again at the 6 p.m. Jazz Service.
PAGEANT REHEARSAL:
SATURDAY, DEC 15 AT 10 a.m.
Children will gather in the Fellowship Hall at 10 a.m. on Saturday, Dec. 15 to get into costume and rehearse the songs. Pageant rehearsal is in the Sanctuary until around noon, and then we go downstairs for the Birthday Party for Jesus with hot dogs, cake and drinks. We should be done by 1 p.m. Hosting this year’s Birthday party are Heather McClelland and Todd Gienapp. We customarily bring gifts to Jesus during the party and this year are collecting items for the Boise-Eliot Backpack Program. Please bring non-perishable food items on Saturday.
The performances of the pageant, directed again this year by Kerstin Cathcart will be on Sunday, December 16. All children come downstairs for costumes at 9:50 a.m. and the production should begin shortly after 10. Everyone is encouraged to come and see the pageant; for many, it is the highlight of the Christmas season. The old, old story, told in a fresh new way by the children never fails to delight. Feel free to bring family and friends to the pageant on Sunday morning, or at the Jazz service on Sunday night.
Confirmation students please plan to be at rehearsal on Saturday, getting the littles into costume and helping them move around the facility. Then again on Sunday, both at the 9:50 time and again at 6 p.m.
Any parents who want to help with rehearsal or performances, please let Rebecca or Kerstin know.
We will have regular Sunday School classes for all ages on December 9. There will not be class Dec 23, Dec. 30 or January 6.
Regular classes resume January 13.
High School News
Cheese, Chess and Checkers is the last Friday of the Month, December 28 from 7-9 PM in the basement. Intergenerational, fun & games for all ages.
High School Students are having great meetings and Bible studies every Sunday at 9:50 a.m. Attendance is in the 6-12 range and friends are always welcome.
A Pizza Party with College Students is being organized over the Christmas break to hear about their schools. Once the date is set, all High School youth will be notified.
Small Group Opportunities for Fellowship and Health
Join the Augustana Travelers:
Wed. Dec. 12: 11:30-11:45 AM arrive for 12 noon seating for Christmas Luncheon at Albertina Kerr, NE 22 & Glisan St.; $24 per person; 8 spots left. Wonderful Augustana tradition; join us for singing hymns & reading of Scriptures about the Christmas miracle. Delicious meal as well! Contact Kathy Jones w/any questions 503-781-1684
Knotty Needlers Prayer Shawl Ministry: Knitters and crocheters are invited Wednesday, December 5, 10a-12n. Beginners are welcome and yarn is provided.
Book Club meets Tuesday, December 11.
The Women’s Prayer and Support group meets each Monday morning at 9:30 in the Library. We welcome new women at any time. If you have questions, contact Rebecca at 971-400-9381 or email@example.com
Tai chi: Moving for Better Balance classes: Tuesdays 10 AM; Thursdays 1 PM; Saturdays 9 am, all in the Fellowship Hall.
Important Dates
Sunday, January 20: Annual Martin Luther King Jr. Jazz Gospel Service. This year will focus on our Lift Every Voice Oregon Campaign to ban the sale of assault weapons and large capacity magazines in Oregon. We have two bills submitted for the Oregon Legislature in 2019. This will be an all-star night from 6:00-8:00 p.m.
Sunday February 3 is our Annual Meeting and celebration of ministry with a 10:00 a.m. Combined Service, 11:30 Potluck and the meeting from 12-1:30 p.m. sharp!
## December Worship and Hospitality Schedule
| Time | December 2 | December 9 | December 16 | December 23 | December 30 |
|---------------|------------------|------------------|------------------|------------------|-----------------|
| **8:30 a.m.** | | | | | |
| Greeters | Loyce and Wayne Forsgren | Denise Gates Glen Jacobs | Debby and Don Willis | Joanne and Darrell Mecklem | Diane and Sil Davis |
| Readers | Carol Reitz Maria Grumm | Gwen Corum Cynthia Ingebretson | Diane and Sil Davis | Jennifer Fort Paul Fox | Kathy Allen Kirsch Elizabeth Morse |
| Assistant | Wayne Reitz | Karen Kinzey | Eli Kruger | Sue Best | Janey Paterno |
| Acolyte | Katelyn Haslebacher | Grant Haslebacher | Maya Kruger | Lola Conner | Audrey Stiefvater |
| Altar Guild | Michael Hayes | Marilyn Williams Larry Rank | Kelli and Ben Joy | Kathy Allen Kirsch John Kirsch | LeAnne Jabs |
| **11:00 a.m.** | | | | | |
| Greeters | Janis and Marty Bell | Kristi Davenport Arno Reifenberg | Dennis Brooks | Beverly and Richard North | Susan Van Winkle |
| Lector | Linda Knudson | Laurie Bregitzer | Wayne Westfall | Myrna Sheie | Joyce Olson |
| Assistant | Roger Olson | Anna Bandick | Alison Embler-Brown | Joyce Kindschuh | Roger Olson |
| Acolyte | Cheryl Davis | Susan Van Winkle | | | |
| Cookies | Susan Van Winkle | Kathy Jones | Lynn and Nancy Gray | Janis Bell | Pat Gebhardt |
| Altar Guild | Joyce Olson Susan Van Winkle | Joyce Olson Susan Van Winkle | Joyce Olson Susan Van Winkle | Joyce Olson Susan Van Winkle | Joyce Olson Susan Van Winkle |
New Acolytes are welcome at all services. Children may begin in 3rd grade and continue through High School.
### December Birthdays
God’s Blessings to our members celebrating birthdays! If your birthday is not listed, please call the office; we need the date!
Thanks.
| Name | Date |
|--------------------|--------|
| Charles Kohn | 12/02 |
| Kristen Haslebacher| 12/05 |
| Lottie Jensen | 12/05 |
| Josi-Lynn Dulaney | 12/07 |
| William Iboshi | 12/07 |
| Benjamin Iboshi | 12/07 |
| Lisette Saad | 12/08 |
| Cheri Beaulaurier | 12/11 |
| Engred Chai | 12/12 |
| Gino Pieretti | 12/12 |
| Sharon Eastvold | 12/13 |
| Jessica Borg | 12/16 |
| David Schmidt | 12/16 |
| Katie Witham | 12/16 |
| Heather Briggs | 12/17 |
| Alanna Deady | 12/17 |
| Robert Trimble | 12/20 |
| Roseta Akin | 12/22 |
| Courtney Beck | 12/22 |
| George Hall | 12/22 |
| Violet Whitcomb | 12/22 |
| Susan Hanset | 12/25 |
| D’Norgia Price | 12/25 |
| Claudia Scholtz | 12/25 |
| Laura Downs | 12/26 |
| Renee Ignacio | 12/26 |
| David Krogh | 12/26 |
| Linda Nelson | 12/26 |
| Marilyn Williams | 12/27 |
| Nancy Boros | 12/30 |
| Cole Delaney | 12/30 |
| Paul Cedfeldt | 12/31 |
| Krista Hornish | 12/31 |
| Karen Kinzey | 12/31 |
| Sunday | Monday | Tuesday | Wednesday | Thursday | Friday | Saturday |
|--------|--------|---------|-----------|----------|--------|----------|
| 2 Advent I | 3 | 4 | 5 | 6 | 7 | 8 |
| 9:30 a Contemporary Worship Svc | 9a Prayer/Study L | 10a Tai Chi FH | 11a-2p PreK FH | 11a-2p PreK FH | 1p Tai Chi FR | 9-3 Familias FR |
| 9:50a Intergenerational Advent Fair | 11a Classical Worship Service | 11a-1:30p Pack FH | 12a AMA L | 12a SA SS | 6:15p Contemporary Choir | 12n SA SS |
| 11 a Rose City Mixed Quartet with Jazz / Gospel Worship Service | 6p Book Club | 6:30p CAT FH | 6p MA SS | 7:25p Sanctuary Choir | 7p NA FR | 8:30 Breakfast work party |
| 7 p AA 27 | 7:15-8:15p Handbell rehearsal | 8:30-10a Sanctuary prep party | 8:30-10a Sanctuary prep party | 9a Tai Chi FH | 9a EcoFaith FR | 9a Tai Chi FH |
| 9 Advent II | 10 | 11 | 12 | 13 | 14 | 15 |
| 8:30 a Contemporary Worship Svc | 9a Prayer/Study L | 10a Tai Chi FH | 11a-2p PreK FH | 11a-2p PreK FH | 1p Tai Chi FR | 8:30 Breakfast work party |
| 9:50a Children’s Choir sings all ages | 11:12 PreK FH | 11a-1:30p Pack FH | 12a SA SS | 12a SA SS | 6:15p Contemporary Choir | 8:30-10a Sanctuary work party |
| New Members Received | 11 a New Member Class L | 11a-1:30p Pack FH | 6p Book Club Rm 27 | 7:25p Sanctuary Choir | 7:15-8:15p Handbell rehearsal | 9a Tai Chi FH |
| 6 p Jazz / Gospel Worship Service | 6p MA SS | 6p Council Christmas party | 4-8p Katch party FH | 8p AA FR | 9-12 Swish Sub class, 27 |
| 7 p AA 27 | | | | | | 10-1p Families Pageant Rehearsal & Birthday Party for Jesus |
| 16 Advent III | 17 | 18 | 19 | 20 | 21 | 22 |
| New Members Received | 9a Prayer/Study L | 10a Tai Chi FH | 11a-2p PreK FH | 11a-2p PreK FH | 1p Tai Chi FH | 8:30 Breakfast work party |
| 8:30 a Contemporary Worship Svc | 11:12 PreK FH | 11a-1:30p Pack FH | 12a SA SS | 12-5p Familias L | 6:15p Contemporary Choir | 8:30-10a Sanctuary work party |
| 9:50a Children’s Christmas Pageant | 12a SA SS | 6p MA SS | 5:30p CAT Rm 27 | 7:25p Sanctuary Choir | 7p NA FR | 9a Tai Chi FH |
| 11 a Classical Worship Service | | | | | | 8p AA FR |
| Children’s Christmas Pageant at Jazz / Gospel Worship Service | 7p AA 27 | | | | | |
| 23 Advent IV | 24 Christmas Eve | 25 Christmas Day | 26 | 27 | 28 | 29 |
| Dedication of Pledges | 8:30-10a Sanctuary prep party | 9:30a Christmas Day worship service | | | | |
| 8:30 a Contemporary Worship Svc | 9:30a Christmas Day worship service | | | | | |
| 11 a Classical Worship Service | 7p Jazz | | | | | |
| 6 p Jazz / Gospel Worship Service | 9p Jazz/Gospel | | | | | |
| Sanctuary prep party after jazz | 11p Classical | | | | | |
| 7 p AA 27 | | | | | | 8:30 Breakfast work party |
| 30 Christmas I | 31 | | | | | |
| 8:30 a Contemporary Worship Svc | 9a Prayer/Study L | | | | | |
| 11 a Classical Worship Service | | | | | | |
| 6 p Jazz / Gospel Worship Service | | | | | | |
| 7 p AA 27 | | | | | | |
The mission of Augustana Lutheran Church is to witness to the Good News of God in Jesus Christ as a multicultural and multinational congregation of justice-seekers and peacemakers who welcome and affirm every person.
### 2018 Augustana Church Council
| President | Chris Delaney |
|--------------------|---------------|
| Ayn Frazee, Secretary | Greg Halverson |
| Ken Petersen, Treasurer | LeAnne Jabs |
| Pastor Mark Knutson | Gabrielle Petersen |
| Francisco Aguirre | D’Norgia Price |
| Anna Bandick | Fiona and Duncan Taylor |
| Rose Christopherson | Robert Trimble |
| Jessi Khangura / Ejvin Berry |
|-----------------------------|
| Betsy, Mark, Rebekah Witham |
| Susan and Ramona Nieman |
| Todd Gienapp |
| Carol and Wayne Reitz |
| Lisa Wenzlick |
### Augustana Lutheran Church Staff
**Pastor:** Rev. Mark Knutson
**Admin. Asst.:** Joyce Kindschuh
**Gifts for Ministry:** Jenny Schmidt
**Education Coord:** Rebecca Dodson
### Musicians
**8:30 am Music Coordinators:** Maile Grumm, Jenny Schmidt, Joe Paterno, Andy Fort
**8:30 am Band:** Maile Grumm, Jenny Schmidt, Mark Foat, Eric North, Joe & Janey Paterno, Barbara Passolt, Noah Fort, Steve Aguila, Keri Haslebacher
**Organist & Sanctuary Choir Director:** Dr. Walter Krueger
**Pianist, Contemporary & Children’s Choir Director:** Andrew Fort
**Sunday School music:** Andy Fort
**Jazz Service:** Marilyn Keller, George Mitchell, Ron Steen, Kevin Deitz
### 2018 Foundation Board
| Chair | Linda Knudson |
|---------------|---------------|
| Allen Brown, Jr. | Julia Brown, Council President |
| Edward Running, Treasurer | Pastor Mark Knutson |
### Pastoral Associates
| Rev. Roger Olson | Rev. Sandy Walker | Deaconess Julia Brown |
|------------------|-------------------|-----------------------|
| Rev. Jake Jacobson | Rev. Tom Campbell-Schmitt | Deacon Sue Best |
| Rev. Susanne Olin | Rev. Patty Campbell-Schmitt |
| Rev. David Brownmiller | Rev. Don Voeks | |
The Cdc42 effectors Ste20, Cla4 and Skm1 down-regulate the expression of genes involved in sterol uptake by a MAPK-independent pathway
Meng Lin*, Heike Unden*, Nicolas Jacquier†, Roger Schneiter†, Ursula Just* and Thomas Höfken*
* Institute of Biochemistry, Christian Albrecht University, 24098 Kiel, Germany
† Division of Biochemistry, University of Fribourg, 1700 Fribourg, Switzerland
To whom correspondence should be addressed
Thomas Höfken, Institute of Biochemistry, Christian Albrecht University Kiel, Olshausenstrasse 40, 24098 Kiel, Germany. Tel.: +49 431 8801660; Fax: +49 431 8802609; E-mail: email@example.com
Running head: Regulation of sterol uptake by PAKs
List of abbreviations: ABC, ATP-binding cassette; BR, basic-rich; CRIB, Cdc42/Rac-interactive binding; PAK, p21-activated kinase; PH, pleckstrin homology; TLC, thin-layer chromatography
In *Saccharomyces cerevisiae*, the Rho-type GTPase Cdc42 regulates polarized growth through its effectors, including the p21-activated kinases (PAKs) Ste20, Cla4 and Skm1. Previously, we demonstrated that Ste20 interacts with several proteins involved in sterol synthesis that are crucial for cell polarization. Under anaerobic conditions, sterols cannot be synthesized and need to be imported into cells. Here, we show that Ste20, Cla4 and Skm1 form a complex with Sut1, a transcriptional regulator that promotes sterol uptake. All three PAKs can translocate into the nucleus and down-regulate the expression of genes involved in sterol uptake including the Sut1 targets *AUS1* and *DAN1* by a novel mechanism. Consistently, deletion of either *STE20*, *CLA4* or *SKM1* results in an increased sterol influx and PAK overexpression inhibits sterol uptake. For Ste20 we demonstrate that the down-regulation of gene expression requires nuclear localization and kinase activity of Ste20. Furthermore, the Ste20-mediated control of expression of sterol uptake genes depends on *SUT1* but is independent of a mitogen-activated protein kinase (MAPK) signalling cascade. Taken together, these observations suggest that PAKs translocate into the nucleus, where they modulates expression of sterol uptake genes via Sut1, thereby controlling sterol homeostasis.
Introduction
The small Rho GTPase Cdc42 plays a central role in the regulation of cellular polarity in eukaryotic cells (Etienne-Manneville, 2004; Jaffe and Hall, 2005). In the budding yeast *Saccharomyces cerevisiae*, Cdc42 regulates different types of polarized growth during various phases of its life cycle, including budding during vegetative growth, mating between haploid cells of opposite mating types, and filamentous growth upon nutrient limitation (Park and Bi, 2007). Cdc42 promotes polarized growth through multiple pathways including polarization of the actin cytoskeleton and directed vesicle trafficking. Among the Cdc42 effectors that trigger these pathways are Ste20, Cla4 and Skm1, members of the p21-activated kinase (PAK) family of serine/threonine protein kinases (Hofmann et al., 2004; Park and Bi, 2007).
Cdc42 recruits Ste20 and Cla4 from the cytoplasm and activates them at the plasma membrane at sites of polarized growth. Therefore, Ste20 and Cla4 are enriched at tips of buds and mating projections (Peter et al., 1996; Leberer et al., 1997; Holly and Blumer, 1999). The recruitment of Ste20 and Cla4 to these sites is governed by protein-protein as well as protein-lipid interactions. PAKs carry a conserved Cdc42/Rac-interactive binding (CRIB) domain that mediates binding to Cdc42 and regulates their activity (Cvrckova et al., 1995; Peter et al, 1996; Leberer et al., 1997; Lamson et al., 2002; Ash et al., 2003). Ste20 and Cla4 also bind to Bem1, a scaffold protein that brings Cdc42, its activator Cdc24 and either Ste20 or Cla4, into close proximity (Leeuw et al., 1995; Gulli et al., 2000; Bose et al., 2001; Winters and Pryciak, 2005; Yamaguchi et al., 2007). In addition, phosphoinositide-binding domains promote the association with membrane lipids. Cla4 carries a pleckstrin homology (PH) domain and Ste20 contains a short basic-rich (BR) region domain, both of them binding to membrane lipids.
Ste20 regulates multiple MAPK pathways that control mating, filamentous growth and osmotic stress response, and it is also involved in exit from mitosis and hydrogen peroxide-induced apoptosis (Leberer et al., 1992; Ramer and Davis, 1993; Liu et al., 1993; Roberts and Fink, 1994; O'Rourke and Herskovitz, 1998; Raitt et al., 2000; Höfken and Schiebel, 2002; Ahn et
Probably the best characterized function of Cla4 is the assembly of the septin ring, which plays a fundamental role in cytokinesis and cell compartmentalization (Weiss et al., 2000; Schmidt et al., 2003; Kadota et al., 2004; Versele and Thorner, 2004). In addition, Cla4 regulates mitotic entry and exit (Höfken and Schiebel, 2002; Seshan et al., 2002; Sakchaisri et al., 2004). Very little is known about Skm1 and no clear function has been attributed to this PAK (Martin et al., 1997). Interestingly, Ste20 and Cla4 interact with Erg4, Cbr1 and Ncp1, which are all involved in sterol biosynthesis, and the deletion of the corresponding genes results in various polarity defects, suggesting that sterol biosynthesis is crucial for cell polarization (Ni and Snyder, 2001; Tiedje et al., 2007).
Many aspects of sterol homeostasis are conserved between yeast and human, and ergosterol, the predominant sterol of yeast, is structurally and functionally related to sterols of higher eukaryotes (Sturley, 2000). Ergosterol biosynthesis can occur only when oxygen is available. Although sterol synthesis is an energy-consuming process, cells do not take up significant amounts of exogenous sterol under aerobiosis. The physiological significance of this phenomenon, termed aerobic sterol exclusion, is poorly understood (Lewis et al., 1985). It may be a way for cells to ensure that only the best fitting sterols accumulate in its membranes (Parks and Casey, 1995). In contrast, under anaerobic conditions, when sterol biosynthesis is compromised, cells become capable of importing sterols whose presence in the medium is then required for growth. Since completely anaerobic conditions are difficult to maintain, most studies employed sterol auxotrophs or mutants in heme synthesis. Heme acts as an intermediary in regulating the expression of oxygen-responsive genes. Therefore, deficiency in heme biosynthesis, e.g. in a hem1Δ background, mimics anaerobic conditions in the presence of oxygen and as a consequence these cells accumulate sterol from the medium (Gollub et al., 1977). The molecular mechanisms of sterol import are poorly understood. Sterol uptake is controlled by the transcriptional regulators Sut1, Upc2 and Ecm22, which are members of the Zn(II)2Cys6 family of DNA binding proteins (Schjerling and Holmberg, 1996). Under
anaerobic conditions Upc2 and Sut1 up-regulate the expression of the ATP-binding cassette (ABC) transporter *AUS1* and *PDR11*, and the putative cell wall mannoprotein *DAN1* (Figure 1) (Regnacq et al., 2001; Wilcox et al., 2002; Alimardani et al., 2004). These proteins then mediate sterol influx. Presumably, other proteins contribute to this process as well (Wilcox et al., 2002; Alimardani et al., 2004).
Here, we characterize the link between proteins regulating sterol uptake and the PAKs Ste20, Cla4 and Skm1. Our data suggest a model (Figure 1), according to which PAKs shuttle into the nucleus, where they interact with Sut1 and possibly other transcriptional regulators and down-regulate the expression of *AUS1*, *DAN1* and *PDR11*. As a consequence, sterol influx is also negatively regulated. Thus, PAKs seem to play an important role in the control of sterol uptake.
Materials and Methods
Yeast strains, plasmids and growth conditions
All yeast strains are listed in supplemental Table S1. The strains used in this study were in the YPH499 background with the exception of strains used for filamentous growth and sterol uptake. For filamentous growth the Σ1278b background was used (MLY48, MLY49, PPY966 and THY696). Cholesterol uptake experiments were performed with the BY4742 background (Winzeler et al., 1999) (MLY180, MLY181, MLY185, THY725, THY734, THY737, YRS1707 and YRS1962). Yeast strains were grown in yeast extract, peptone, dextrose (YPD) or synthetic complete (SC) medium. For induction of the GAL1 promoter, yeast cells were grown in yeast extract, peptone (YP) or SC medium with 3% raffinose instead of glucose. Galactose (final concentration 2%) was added to induce the GAL1 promoter. hem1Δ cells were grown in medium supplemented either with 40 µg/ml δ-aminolevulinic acid or with 80 µg/ml ergosterol solubilized in Tergitol NP40/ethanol (1:1) and 1% Tween 80. Yeast strains were constructed using PCR-amplified cassettes (Longtine et al., 1998; Knop et al., 1999; Janke et al., 2004).
All constructs used in this work are listed in supplemental Table S2.
Split-ubiquitin technique
The split-ubiquitin screen using STE20 as bait is described in Tiedje et al. (2007). For the interaction assays, $10^5$ ste20Δ cells carrying the split-ubiquitin plasmids were spotted on SC-His/Leu and SC-His/Leu/Ura plates and were grown for 2 days at 30°C.
Pull-down assays and antibodies
GST, GST-Sut1, His$_6$-Sec6 and His$_6$-Sut1 were expressed in *E. coli* BL21 (DE3) and purified using glutathione-sepharose (Amersham Biosciences) and Ni-NTA agarose (Quiagen), respectively. These immobilized recombinant proteins were presented to yeast lysates of 3HA-STE20, 3HA-CLA4 and 3HA-SKM1, respectively, for 90 min at 4°C in IP buffer (20 mM Tris pH 7.5, 100 mM NaCl, 10 mM EDTA, 1 mM EGTA, 5% glycerol, 1% NP-40, 1% BSA). After 5 washes with IP buffer, the associated proteins were eluted with sample buffer and analyzed by immunoblotting.
Monoclonal mouse anti-HA (12CA5) was obtained from Roche. Rabbit polyclonal anti-Cdc11 and goat polyclonal anti-GAPDH antibodies were from Santa Cruz Biotechnology, and rabbit polyclonal anti-GFP was from Fitzgerald Industries. Secondary antibodies were from Jackson Research Laboratories.
**Microscopy**
For fluorescence microscopy, cells were fixed with 4% formaldehyde for 10 min, washed twice with PBS and resuspended in water or in 0.25 µg/ml DAPI in 50% glycerol. Cells were examined with a Zeiss Axiovert 200M fluorescence microscope equipped with a 100 × Plan oil-immersion objective and images were captured using a Zeiss AxioCam MRm CCD camera.
**β-Galactosidase assays**
Densities of cell cultures were measured by optical density at 600 nm (OD$_{600}$). 1-10 ml of cells was harvested by centrifugation and resuspended in 1 ml Z buffer (100 mM sodium phosphate [pH 7.0], 10 mM KCl, 1 mM MgSO$_4$, 50 mM β-mercaptoethanol). Cells were permeabilized by addition of 20 µl chloroform and 20 µl 0.1% SDS. After 15 min incubation at 30°C the reaction was started by addition of 160 µl o-nitrophenyl-β-D-galactopyranoside (4 mg/ml in 100 mM sodium phosphate, pH 7.0) incubated at 30°C until the solution became yellow and
the reaction was stopped by addition of 400 µl 1 M Na₂CO₃. Samples were centrifuged and the OD_{420} and OD_{550} of the supernatant was determined. β-Galactosidase activity was calculated in Miller units as $1,000 \times [\text{OD}_{420} - (1.75 \times \text{OD}_{550})] / \text{reaction time (min)} \times \text{culture volume (ml)} \times \text{OD}_{600}$.
**Invasive growth and pheromone response assays**
For agar invasion assays, $10^5$ cells of an overnight culture were spotted on YPD and grown for 2 days at 30°C. Plates were photographed before and after being rinsed under a gentle stream of deionized water.
For examination of the formation of a mating projection, logarithmically growing cells were incubated with 1 µg/ml α-factor for 3 hours. These cells were fixed with formaldehyde for morphological examination.
**Sterol uptake**
Sterol uptake was essentially analyzed as described before (Reiner et al., 2006). Δhem1 mutant cells were cultured in δ-aminolevulinic acid-containing media, washed, and diluted in minimal media supplemented with Tween 80 (5 mg/ml), cholesterol (20 µg/ml), and 0.025 µCi/ml [¹⁴C]-cholesterol (American Radiolabeled Chemicals Inc, St. Louis, MO). After 2, 4, 8, and 24 h, equal OD units of cells were collected, washed with 0.5% Tergitol. [³H]-palmitic acid was added to the cell pellet as internal standard. Cells were disrupted with glass beads in the presence of chloroform/methanol (1:1) and lipids were extracted into the organic phase. Lipids were separated by thin-layer chromatography (TLC) (Merck, Darmstadt, Germany) with the solvent system petroleum ether/diethylether/acetic acid (70:30:2; per vol.) and free cholesterol and cholesteryl esters were quantified by scanning with a Berthold Tracemaster 40 Automatic TLC-Linear Analyzer (Berthold Technologies, Bad Wildbad, Germany). TLC plates were then
exposed to a phosphor imager screen and visualized using a phosphorimager (Bio-Rad Laboratories, Hercules, CA).
Results
PAKs interact with the transcriptional regulator Sut1
Previously we described a screen to identify novel interactors of Ste20 using the split-ubiquitin technique (Tiedje et al., 2007). This method is based on the ability of N- and C-terminal halves of ubiquitin to assemble into a quasi-native ubiquitin (Johnsson and Varshavsky, 1994; Wittke et al., 1999). If two proteins, which are attached to the N- and C-terminal halves, respectively, interact, the ubiquitin peptides may be forced into close proximity and a ubiquitin-like molecule is reconstituted (Figure 2A). Ubiquitin-specific proteases recognize this molecule, but not its halves, and cleave off the enzyme Ura3, that carries an additional arginine at the extreme N-terminus (RUra3). The freed RUra3 reporter is rapidly degraded by proteases of the N-end rule, resulting in uracil auxotrophy. Conversely, growth on 5-FOA indicates protein interaction as well, because 5-FOA is converted by Ura3 into 5-fluorouracil, which is toxic for the cell. Employing this technique, we identified Sut1 as a putative Ste20 interacting protein (Figure 2B). *SUT1* encodes a transcriptional regulator involved in sterol uptake under anaerobic conditions (Bourot and Karst, 1995; Ness et al., 2001). To confirm the interaction between Ste20 and Sut1 by an independent approach, recombinant GST and GST-Sut1 purified from *E. coli* were bound to glutathione-sepharose beads, which were then incubated with a yeast extract of *STE20-3HA* cells. Ste20-3HA interacted with GST-Sut1 but not with GST alone (Figure 2C). Next, it was tested whether Cla4 and Skm1, which are related to Ste20, also interact with Sut1 using a pull-down assay. 3HA-Cla4 and 3HA-Skm1, both expressed from yeast, bind specifically to recombinant His<sub>6</sub>-Sut1 but not to the unrelated His<sub>6</sub>-Sec6, which was used as a negative control (Figure 2D).
Nuclear localization of PAKs
Sut1 localizes to the nucleus (Ness et al., 2001), whereas Ste20 can be found in the cytosol and at the plasma membrane at sites of polarized growth (Peter et al., 1996; Leberer et al., 1997) (Figure 3B). Therefore, it is not clear where the interaction between Sut1 and Ste20 takes place. Analysis of the Ste20 sequence revealed a putative bipartite NLS between amino acids 272-288 with two clusters of positively charged amino acids separated by a spacer of 11 amino acids (Figure 3A). Notably, this putative NLS overlaps with the BR domain, which targets Ste20 to the plasma membrane (Takahashi and Pryciak, 2007). In fact, the first of three basic clusters within the BR domain (BR-1) is identical to the second cluster of the NLS (Figure 3A). To test whether the identified NLS is functional, it was fused with GFP and expressed in yeast cells. This GFP fusion also contained a homodimerizing GST moiety, which can help to reveal weak localization determinants by increasing binding avidity (Winters et al., 2005). However, the GFP-NLS fusion protein was found in the cytoplasm (data not shown). We reasoned that the NLS might not be recognized by the nuclear import machinery, possibly due to steric hindrance. Therefore, we added a short stretch of amino acids (SGAGAGAGAGAIL), which is used as a spacer for GFP fused to a protein of interest (Miller and Lindow, 1997) at the C-terminus of the NLS. This protein (termed GFP-Ste20\textsuperscript{272-288*}) exclusively localized to the nucleus as demonstrated by DAPI co-staining (Figure 3B). To rule out the possibility that the attached spacer sequence was responsible for the nuclear localization, another unrelated random sequence (LGYFLFFEGGPGTQFAL) was C-terminally fused to the NLS. This fusion protein (termed GFP-Ste20\textsuperscript{272-288**}) was restricted to the nucleus as well (Figure 3B). From these data we conclude that the NLS is functional, but requires a short non-specific stretch of amino acids at the C-terminus.
As mentioned above, the NLS overlaps with the BR domain (Figure 3A) (Takahashi and Pryciak, 2007). To find out whether the BR domain alone is sufficient for nuclear localization, we compared the localization of GFP fused with Ste20 residues 272 to 400 (comprising the entire NLS and BR domain plus some C-terminal amino acids) and Ste20 residues 285 to 400
(comprising only the second cluster of positively charged amino acids of the NLS, the complete BR domain and some C-terminal amino acids). Notably, GFP-Ste20\textsuperscript{272-400} was strongly enriched in the nucleus and present in the cytoplasm and at the plasma membrane, whereas GFP-Ste20\textsuperscript{285-400} was associated with the bud cortex and only a weak nuclear signal was observed (Figure 3B). Thus, the entire NLS is required for efficient nuclear targeting of Ste20, whereas the BR domain alone, which includes the second basic cluster of the NLS, plays only a minor role in this process. Importantly, in contrast to wild type Ste20 fused to GFP, GFP-Ste20 without the complete NLS (GFP-Ste20\textsuperscript{\Delta272-288}) was excluded from the nucleus (Figure 3, C and D). Similarly, much of GFP-Ste20 lacking the first basic cluster but with an intact BR domain (GFP-Ste20\textsuperscript{\Delta272-284}) was no longer present in the nucleus (Figure 3, C and D). However, the effect was less pronounced compared to GFP-Ste20\textsuperscript{\Delta272-288} (Figure 3D). Together, these data show that the BR-1 region contributes only to a minor extent to nuclear targeting of Ste20. Furthermore, our observations suggest that during normal vegetative growth at least some Ste20 localizes to the nucleus. Consistently, in few cells we observed nuclear localization of wild type Ste20 (Figure 3E). However, since Ste20 was only weakly enriched in the nucleus compared to the surrounding cytoplasm, the number of cells with a nuclear Ste20 signal was difficult to quantify.
Very little is known about Skm1, and its localization has not been reported (Martin et al., 1997). Therefore, we first analyzed the localization of full-length Skm1 fused to GFP. \textit{SKM1-GFP} under control of its endogenous promoter could be detected by immunoblotting, but was to weak to be visualized by fluorescence microscopy (data not shown). Therefore, we overexpressed \textit{GFP-SKM1} from the inducible \textit{GAL1} promoter. Unexpectedly, full-length Skm1 was strongly enriched in the nucleus with only a faint signal in the cytoplasm and at the plasma membrane (Figure 4A). Between amino acids 213 to 230 we identified a putative bipartite NLS (\textbf{KRTNSIKRSVSRTLRKKGK}; positively charged amino acids of the two basic clusters are underlined). Skm1 lacking this putative NLS was no longer enriched in the nucleus (Figure
We also made point mutations in either the first or the second cluster of positively charged amino acids. The mutant protein Skm1\textsuperscript{C1}, in which Lys213 and Arg214 of the first basic cluster were changed to Ala, mostly localized to the cytoplasm (Figure 4A). In a second mutant, Arg227, Lys228 and Lys230 of the second positively charged cluster were changed to Ala. The corresponding protein, termed Skm1\textsuperscript{C2}, was not or only faintly enriched in the nucleus (Figure 4A). By immunoblotting it was confirmed that Skm1 wild type and mutant proteins were expressed at comparable levels (Figure 4B). Thus, both clusters play an important role in nuclear targeting of Skm1. We also fused Skm1 residues 213 to 230 to GFP. This fusion protein was present in the cytoplasm (data not shown), but a slightly larger Skm1 fragment (residues 201-230) exclusively localized to the nucleus (Figure 4C). Taken together, it seems that, similar to Ste20, the predicted Skm1 NLS is functional but requires a few more amino acids to target GFP to the nucleus.
Like Ste20, Cla4 associates with the bud cortex (Holly and Blumer, 1999) (Figure 4E) and analysis of the Cla4 protein sequence did not reveal an NLS. To determine whether Cla4 could translocate into the nucleus, a series of Cla4 fragments fused to GFP were constructed (Figure 4D). In line with a previous report (Wild et al., 2004), Cla4 fragments lacking the N-terminal 180 residues, which contain a PH domain that binds to membrane lipids, did not associate with the plasma membrane (Figure 4E). A fragment containing residues 181-550 exclusively localized to the nucleus, whereas a smaller portion of Cla4 (residues 181-400) was cytoplasmic (Figure 4E). Therefore, we reasoned that a nuclear targeting sequence may be present between residues 400-550. Indeed, a fragment comprising this sequence and also a slightly smaller fragment (amino acids 450-550) were only found in the nucleus (Figure 4E). Further truncations resulted in nuclear enrichment but in these cells we also observed cytoplasmic staining (data not shown). Thus, residues 450-550 are sufficient for proper nuclear targeting of Cla4. Taken together, we demonstrated that all three PAKs can localize to the nucleus, where they could interact with the transcriptional regulator Sut1.
PAKs specifically down-regulate the expression of *AUS1*, *DAN1* and *PDR11*
Sut1 promotes sterol uptake under anaerobic conditions by enhancing the transcription of *AUS1* and *DAN1* (Regnacq et al., 2001; Alimardani et al., 2004). Therefore, we wanted to know whether the expression of these genes also depends on Ste20, Cla4 and Skm1. To this end, the promoter regions of *AUS1*, *DAN1* and *PDR11*, were fused to the reporter gene *lacZ*. *PDR11* was included because it also mediates sterol influx but is not regulated by Sut1 (Wilcox et al., 2002; Alimardani et al., 2004). Using these constructs, we confirmed that *AUS1*, *DAN1* and *PDR11* are not expressed in wild type cells under aerobic conditions (Figure 4, A-C). In contrast, in a *hem1Δ* background mimicking anaerobic conditions these genes were strongly induced (Figure 4, A-C). Furthermore, *SUT1* overexpression under aerobic conditions led to expression of *AUS1* and *DAN1* but not of *PDR11* (Figure S1). This is in line with previous reports (Regnacq et al., 2001; Alimardani et al., 2004) and indicates that *AUS1* and *DAN1* are both target genes of Sut1. Interestingly, *AUS1* expression levels were markedly reduced in *hem1Δ* cells overexpressing either *STE20*, *CLA4* or *SKM1* (Figure 5A), whereas high levels of either *GIC2*, which, like the PAKs, is a Cdc42 effector (Brown et al., 1997; Chen et al., 1997), or *YCK1*, a kinase that is involved in cell polarity (Robinson et al., 1997), did not affect *AUS1* expression (Figure 5A). Thus, the effects of PAKs are specific. Overexpression of *STE20* and *CLA4*, respectively, but not of *SKM1* lowered *DAN1* levels (Figure 5B). Finally, high levels of *CLA4* and *SKM1*, but not of *STE20* decreased *PDR11* expression (Figure 5C). In summary, all three PAKs down-regulate the expression of genes involved in sterol uptake, but with very different specificities. High levels of *GIC2* and *YCK1* did not affect *DAN1* and *PDR11* expression levels (Figure 5, B and C).
Next, it was tested whether deletion of PAKs has an effect on the expression of *AUS1*, *DAN1* and *PDR11*. We observed higher *AUS1* levels in cells lacking *STE20* (Figure 6A), and *SKM1*
deletion led to increased *PDR11* expression (Figure 6C). Notably, these results are consistent with the data obtained following PAK overexpression (Figure 5, A and C).
To test whether the observed effects of PAKs on gene expression are mediated by Sut1, expression studies were performed in cells lacking *SUT1* using *AUS1* as a reporter. The deletion of *SUT1* did not affect *AUS1* expression levels (Figure 7A), which is not surprising since *AUS1* expression is also under control of Upc2 and possibly other transcriptional regulators, which could compensate for the loss of *SUT1* (Wilcox et al., 2002). High levels of either *STE20*, *CLA4* or *SKM1* in the absence of *SUT1* led to no (Figure 7B) or only a minor reduction of *AUS1* levels (Figure 7, A and C). Thus, the PAK-mediated down-regulation of *AUS1* depends mostly on *SUT1*. Possibly, PAKs to some extent also regulate Upc2 and maybe other transcriptional regulators of *AUS1*. This would explain the slight decrease of *AUS1* levels observed following overexpression of *STE20* and *CLA4* in cells lacking *SUT1*.
The role of Sut1 was also examined in cells deleted for *STE20*. The additional deletion of *STE20* in the absence of *SUT1* had no effect on *AUS1* expression levels (Figure 7D). This is consistent with the notion that specific inhibition of Sut1 by Ste20 results in a down-regulation of *AUS1* expression. In *ste20Δ* cells, Sut1 seems to have a higher activity, which results in stronger *AUS1* expression, possibly due to the missing inhibition by Ste20. As mentioned above, the *sut1Δ* strain probably expresses *AUS1* at levels comparable to the wild type because of proteins such as Upc2, which are functionally redundant with Sut1. The fact that *AUS1* expression is unaltered in *sut1Δ ste20Δ* cells suggests that Ste20 negatively regulates for the most part Sut1 and not or only to a minor extent other transcriptional regulators like Upc2. Whereas, the experiment in Figure 7D suggests that Sut1 activates *AUS1* expression in a manner that is antagonized by Ste20, the experiments in Figure 7, A-C show that overexpression of PAKs inhibits *AUS1* expression in a manner that requires *SUT1*. It seems that Sut1 negatively regulates *AUS1* expression, a process that is promoted by high PAK levels. Importantly, in both models the effect on *AUS1* expression would be the same but it is not clear
how these two different modes of action can be reconciled. Unfortunately, it is not known how Sut1 controls the expression of genes such *AUS1*. Thus, it is even less clear how PAKs could regulate Sut1 activity.
Next, the mechanisms by which PAKs contributes to the regulation of gene expression were further characterized. For these experiments we focused on the regulation of *AUS1* by Ste20. Whereas overexpression of wild type *STE20* decreased *AUS1* levels, overexpression of a mutated *STE20* lacking a major part of the NLS but leaving the BR domain intact (*STE20*<sup>Δ72-284</sup>) did not affect *AUS1* (Figure 8A). Thus, the nuclear localization of Ste20 is essential for its role in the control of gene expression. High levels of *STE20*<sup>K649A</sup>, an allele that lacks kinase activity (Wu et al., 1995), resulted in a minimal decrease of *AUS1* expression, but this effect, which is possibly due to minute residual kinase activity, was much weaker than following overexpression of wild type *STE20* (Figure 8A). Thus, Ste20 kinase activity is required for the regulation of *AUS1* expression. Notably, wild type *STE20* and the described mutant alleles were expressed at comparable levels as confirmed by immunoblotting (Figure 8B). Therefore, the observed effects are not due to different amounts of Ste20 protein. We also examined phenotypes of *STE20* mutant alleles when expressed from the native *STE20* promoter. Deletion of *STE20* led to an increased *AUS1* expression in the *hem1Δ* background (Figure 6A). Therefore, we brought back wild type and mutated versions of *STE20* under control of its own promoter into the *ste20Δ hem1Δ* strain. As expected *STE20* reduced the expression of *AUS1* (Figure 8C). In contrast, the kinase dead derivative *STE20*<sup>K649A</sup> and the NLS mutant *STE20*<sup>Δ72-284</sup> had no effect on *AUS1* levels (Figure 8C), suggesting that Ste20 kinase activity and nuclear localization are required for the inhibition of *AUS1* expression.
In a similar experiment, we tested whether the NLS of Skm1 plays a role in the down-regulation of sterol uptake genes. As shown in Figure 6C, *SKM1* deletion results in an up-regulation of *PDR11*. Integrating wild type *SKM1* into the *skm1Δ hem1Δ* double mutant
reduced *PDR11* expression to normal levels, whereas *SKM1ANLS* had no effect (Figure 9). Thus, nuclear localization of Skm1 seems to be important for the down-regulation of *PDR11*.
**The regulation of *AUS1* expression is independent of Ste20 MAPK signalling**
Ste20 controls the expression of numerous genes by the activation of MAPK cascades regulating mating, filamentous growth and the hyperosmotic response (Roberts and Fink, 1994; O’Rourke and Herskowitz, 1998; Raitt et al., 2000). All three MAPK pathways share the MAPK kinase kinase Ste11, which is a target of Ste20 and essential for Ste20-dependent activation of MAPK cascades (Roberts and Fink, 1994; Wu et al., 1995). To examine whether the Ste20-mediated transcriptional changes observed by us depend on MAPK signaling, we determined *AUS1* expression in cells lacking *STE11*. The decrease of *AUS1* levels following *STE20* overexpression was independent of *STE11* (Figure 10A). Thus, Ste20 seems to regulate gene expression by two distinct mechanisms: the activation of MAPK cascades and a Sut1-dependent pathway. Since the down-regulation of *AUS1* required nuclear Ste20 (Figure 8, A and C), we tested whether the nuclear localization of Ste20 was also necessary for its role in filamentous growth and mating. As described previously, Ste20 is essential for haploid filamentous growth and the formation of a mating projection in response to pheromone (Leberer et al., 1992; Roberts and Fink, 1994) (Figure 10, B and C). Importantly, we observed normal filamentous growth for cells with Ste20 lacking the NLS (Figure 10B). These cells also responded normally to mating pheromone by arresting in G$_1$ phase and formation of a mating projection (Figure 10C).
Taken together, Ste20 controls gene expression by two different mechanisms. MAPK cascades depend on Ste11, but do not require nuclear Ste20. In contrast, the Sut1-mediated pathway relies on nuclear Ste20, but is independent of Ste11.
Ste20, Cla4 and Skm1 negatively regulate sterol uptake
Since Ste20, Cla4 and Skm1 down-regulate the expression of genes involved in sterol influx, it was tested whether individual deletion of either *STE20*, *CLA4* or *SKM1* affects sterol uptake. To this end heme-deficient wild type and deletion strains were grown in the presence of \[^{14}\text{C}]\text{cholesterol} and uptake and esterification were quantified by TLC analysis of radiolabeled lipids. Consistent with the gene expression data, increased levels of free and esterified sterol were observed for *ste20Δ, cla4Δ* and *skm1Δ* (Figure 11A). Sterol uptake was also analyzed in a strain that simultaneously overexpresses *STE20, CLA4* and *SKM1*. In these cells, sterol import was markedly reduced compared to the wild type (Figure 11B). Thus, PAKs have a negative effect on sterol uptake. Finally, it was tested whether the inhibition of sterol import by Ste20 depends on nuclear localization of this PAK. Sterol uptake of cells expressing *STE20^{NLS}* from its own promoter was indistinguishable from the wild type (Figure S2), suggesting that Ste20 may also have functions in sterol uptake that do not require nuclear translocation (see Discussion).
Discussion
Kinases of the PAK-family have a well established role in cell polarization (Hofmann et al., 2004). Budding yeast expresses three PAKs: Ste20, Cla4 and Skm1. To gain insight into molecular mechanisms of Ste20 action, we screened for interactors of Ste20 using the split-ubiquitin technique (Tiedje et al., 2007). Here, we show that Sut1, a transcriptional regulator that controls sterol uptake (Bourot and Karst, 1995; Ness et al., 2001), forms a complex with Ste20. By using pull-down assays we demonstrate that not only Ste20 but also Cla4 and Skm1 bind to Sut1. Sut1 localizes exclusively to the nucleus, even after strong overexpression (data not shown) (Ness et al., 2001). Thus far, Ste20 and Cla4 have been shown to localize mainly to the plasma membrane at sites of polarized growth and to the cytoplasm, whereas the localization of Skm1 has not been reported (Peter et al., 1996; Leberer et al., 1997; Holly and Blumer 1999). In this work, we show strong nuclear enrichment of Skm1 and we reveal an NLS. Further, we identify an NLS for Ste20. The fact that a mutated Ste20, which lacks the entire NLS or a major part of it, is excluded from the nucleus under normal growth conditions, suggests that at least a small amount of Ste20 is present in the nucleus at any time. Consistently, in few cells, we observed a nuclear enrichment of wild type Ste20. Cla4 does not contain an obvious NLS and the full-length protein is not enriched in the nucleus. Nevertheless, we observed that several Cla4 fragments that lacked the PH domain translocated into the nucleus. The PH domain is necessary for membrane association of Cla4 by binding to phosphoinositides (Wild et al., 2004). We suggest that similar to Ste20, Cla4 that is not associated with membranes can localize to the cytoplasm and the nucleus. Therefore, it seems likely, that the interaction between Sut1 and PAKs takes place in the nucleus. This notion is confirmed by our observation, that the down-regulation of the Sut1 target *AUS1* by Ste20 and the inhibition of *PDR11* expression by Skm1 depends on the nuclear localization of these PAKs.
The expression of *AUS1*, *DAN1*, *PDR11* and of other genes contributing to sterol uptake is repressed under aerobic conditions. In anaerobiosis, transcriptional regulators such as Sut1 and Upc2 allow expression of these genes, which is required for sterol uptake under these conditions (Regnacq et al., 2001; Wilcox et al., 2002; Alimardani et al., 2004). Here, we demonstrate that PAKs negatively regulate the expression of genes involved in sterol influx. Interestingly, we observed very different expression profiles for Ste20, Cla4 and Skm1. Ste20 regulates the expression of *AUS1* and *DAN1*, Skm1 controls *AUS1* and *PDR11*, whereas Cla4 inhibits the expression of *AUS1*, *DAN1* and *PDR11*. The reason for these different specificities is not clear.
Importantly, we demonstrate for all PAKs that the down-regulation of gene expression depends for the most part on *SUT1*. Since *AUS1* and *DAN1* are target genes of Sut1 and PAKs form a complex with Sut1, it seems very likely that PAKs control gene expression via Sut1. Nevertheless, it cannot be excluded that PAKs mediate their effects on transcription through other factors such as Upc2 or the less well characterized Ecm22 (Shianna et al., 2001) as well. *PDR11* is also regulated by Cla4 and Skm1 but its transcription is controlled by Upc2 and there is no evidence that *PDR11* is a target of Sut1. Further, we observed a minor reduction of *AUS1* levels in the absence of *SUT1* following overexpression of either *STE20* or *CLA4*. Regulation of Upc2 or other factors by Ste20 and Cla4 could account for these weak effects on *AUS1* expression. We observed that sterol uptake in contrast to the regulation of *AUS1* expression is independent of the Ste20 NLS. The reason for this discrepancy is not clear. However, sterol uptake is a more complex process than the transcription of an individual gene. Possibly, Ste20 contributes also to sterol uptake by other mechanisms. Ste20 could control the activity of transcriptional regulators such as Upc2 and Ecm22, which localize to the cytoplasm as well (Marie et al., 2008).
Thus far, the mechanisms of sterol uptake are only poorly understood. It seems that the transcriptional regulators involved in this process act redundantly, but very little is known
about Ecm22 and it is not clear how Sut1 promotes the expression of *AUS1* and *DANI*. Therefore, it is difficult to establish the mechanisms of inhibition of Sut1 and possibly other transcriptional regulators by PAKs. The down-regulation of *AUS1* expression requires Ste20 kinase activity. Nevertheless, we did not observe Sut1 phosphorylation by PAKs using in vivo and in vitro kinase assays (data not shown). Thus, it is conceivable that PAKs control transcriptional regulators via an unknown protein.
The control of gene expression by Ste20, Cla4 and Skm1 described here is clearly distinct from the regulation of transcription by Ste20 via MAPK modules as previously reported. During mating, filamentous growth and in response to hyperosmolarity Ste20 activates the MAPK kinase kinase Ste11 by phosphorylation, which eventually results in a change of the transcriptional pattern of numerous genes (Roberts and Fink, 1994; Wu et al., 1995; O'Rourke and Herskowitz, 1998; Raitt et al., 2000). As demonstrated for *AUS1*, the Sut1-mediated regulation of expression is independent of Ste11. Furthermore, it requires nuclear localization of Ste20. In contrast, the MAPK pathways promoting filamentation and the formation of a mating projection are independent of nuclear Ste20, but Ste11 is essential for these processes. Thus, the two mechanisms seem to be independent.
It has been reported that Ste20 translocates into the nucleus during hydrogen peroxide-induced cell death and directly phosphorylates histone H2B (Ahn et al., 2005). This presumably results in chromatin condensation. However, it has not been examined whether Ste20 alters gene expression under these conditions. Whether the inhibition of a transcriptional regulator involved in sterol uptake by Ste20 also involves chromatin modification as suggested for hydrogen peroxide-induced cell death remains to be tested.
Importantly, we demonstrated that the control of genes involved in sterol import mediated by Ste20, Cla4 and Skm1 affects sterol uptake. As expected from their role as negative regulators, the deletion of either *STE20*, *CLA4* or *SKM1* results in an increased sterol influx and
subsequent esterification. Furthermore, sterol uptake was markedly reduced in cells overexpressing all three PAKs.
Our data raise the question why Ste20, Cla4 and Skm1 regulate sterol uptake. The fact that all three PAKs are involved in this process, suggests that the control of sterol influx is crucial for the cell and possibly for cell polarization. Previously, we could show that Ste20 binds to Erg4, Cbr1 and Ncp1, which all catalyze important steps in sterol biosynthesis (Tiedje et al., 2007). Interestingly, these proteins are also involved in bud site selection, apical bud growth, mating, filamentous growth and exit from mitosis (Ni and Snyder, 2001; Keniry et al., 2004; Tiedje et al., 2007). These observations highlight the importance of sterol synthesis for cell polarization. Since Ste20 plays a crucial role in all these processes as well, it seems likely that Ste20 and sterol biosynthetic proteins act in the same pathway(s). In the absence of oxygen, sterols cannot be synthesized and cells completely depend on import of sterols from the extracellular medium. Under these conditions, control of sterol uptake may be as important for cell polarization as sterol biosynthesis in aerobiosis. Therefore, it is not surprising that PAKs regulate sterol uptake. Importantly, Cla4 not only has a negative effect on sterol uptake but also down-regulates sterol biosynthesis and storage (manuscript in preparation). Thus, it seems that Cla4 modulates sterol homeostasis under aerobic conditions and that all PAKs are involved in sterol import under anaerobiosis. In addition, homologues of oxysterol-binding proteins, a family of proteins, which regulate synthesis and transport of sterols, were found to participate in Cdc42-dependant cell polarity (Kozminski et al., 2006). How sterols contribute to cell polarization has been discussed controversially (Pichler and Riezman, 2004; Alvarez et al., 2007) and the elucidation of underlying molecular mechanisms will require further work.
In summary, we describe here a novel function for all three PAKs. They control transcription in the nucleus, which results in a change of sterol influx rate.
Acknowledgements
We are very grateful to Nils Johnsson for providing us with the split-ubiquitin library. We thank Thierry Berges for plasmids, Peter Pryciak for constructs and communicating unpublished data and Silke Horn for excellent technical support. The project was supported by the Deutsche Forschungsgemeinschaft (HO 2098/2 and HO 2098/3 to TH) and the Swiss National Science Foundation (PP00A-110450 and 3100-120650 to RS).
References
Ahn, S.H., Cheung, W.L., Hsu, J.Y., Diaz, R.L., Smith, M.M., and Allis, C.D. (2005). Sterile 20 kinase phosphorylates histone H2B at serine 10 during hydrogen peroxide-induced apoptosis in *S. cerevisiae*. Cell 120, 25-36.
Alimardani, P., Régnacq, M., Moreau-Vauzelle, C., Ferreira, T., Rossignol, T., Blondin, B., and Bergès, T. (2004). SUT1-promoted sterol uptake involves the ABC transporter Aus1 and the mannoprotein Dan1 whose synergistic action is sufficient for this process. Biochem. J. 381, 195-202.
Alvarez, F.J., Douglas, L.M., and Konopka, J.B. (2007). Sterol-rich plasma membrane domains in fungi. Eukaryot. Cell 6, 755-763.
Ash, J., Wu, C., Larocque, R., Jamal, M., Stevens, W., Osborne, M., Thomas, D.Y., and Whiteway, M. (2003). Genetic analysis of the interface between Cdc42p and the CRIB domain of Ste20p in *Saccharomyces cerevisiae*. Genetics 63, 9-20.
Bourot, S., and Karst, F. (1995). Isolation and characterization of the *Saccharomyces cerevisiae* SUT1 gene involved in sterol uptake. Gene 165, 97-102.
Brown, J.L., Jaquenoud, M., Gulli, M.P., Chant, J., and Peter, M. (1997). Novel Cdc42-binding proteins Gic1 and Gic2 control cell polarity in yeast. Genes Dev. 11, 2972-2982.
Chen, G.C., Kim, Y.J., and Chan, C.S. (1997). The Cdc42 GTPase-associated proteins Gic1 and Gic2 are required for polarized cell growth in *Saccharomyces cerevisiae*. Genes Dev. 11, 2958-2971.
Cvrckova, F., De Virgilio, C., Manser, E., Pringle, J.R., and Nasmyth, K. (1995). Ste20-like protein kinases are required for normal localization of cell growth and for cytokinesis in budding yeast. Genes Dev. 9, 1817-1830.
Etienne-Manneville, S. (2004). Cdc42-the centre of polarity. J. Cell Sci. 117, 1291-1300.
Gollub, E.G., Liu, K.P., Dayan, J., Adlersberg, M., and Sprinson, D.B. (1977). Yeast mutants deficient in heme biosynthesis and a heme mutant additionally blocked in cyclization of 2,3-oxidosqualene. J. Biol. Chem. 252, 2846-2854.
Höfken, T., and Schiebel, E. (2002). A role for cell polarity proteins in mitotic exit. EMBO J. 21, 4851-4862.
Hofmann, C., Shepelev, M., and Chernoff, J. (2004). The genetics of Pak. J. Cell Sci. 117, 4343-4354.
Holly, S.P., and Blumer, K.J. (1999). PAK-family kinases regulate cell and actin polarization throughout the cell cycle of Saccharomyces cerevisiae. J. Cell Biol. 147, 845-856.
Jaffe, A.B., and Hall, A. (2005). Rho GTPases: biochemistry and biology. Annu. Rev. Cell Dev. Biol. 21, 247-269.
Janke, C., Magiera, M.M., Rathfelder, N., Taxis, C., Reber, S., Maekawa, H., Moreno-Borchart, A., Doenges, G., Schwob, E., Schiebel, E., and Knop, M. (2004). A versatile toolbox
for PCR-based tagging of yeast genes: new fluorescent proteins, more markers and promoter substitution cassettes. Yeast 21, 947-629.
Johnsson, N., and Varshavsky, A. (1994). Split ubiquitin as a sensor of protein interactions in vivo. Proc. Natl. Acad. Sci. USA 91, 10340-10344.
Keniry, M.E., Kemp, H.A., Rivers, D.M., and Sprague, G.F. (2004). The identification of Pel1-interacting proteins that genetically interact with Cla4 may indicate a link between G1 progression and mitotic exit. Genetics 166, 1177-1186.
Knop, M., Siegers, K., Pereira, G., Zachariae, W., Winsor, B., Nasmyth, K., and Schiebel, E. (1999). Epitope tagging of yeast genes using a PCR-based strategy: more tags and improved practical routines. Yeast 15, 963-972.
Kozminski, K.G., Alfaro, G., Dighe, S., and Beh, C.T. (2006). Homologues of oxysterol-binding proteins affect Cdc42p- and Rho1p-mediated cell polarization in Saccharomyces cerevisiae. Traffic 7, 1224-1242.
Lamson, R.E., Winters, M.J., and Pryciak, P.M. (2002). Cdc42 regulation of kinase activity and signaling by the yeast p21-activated kinase Ste20. Mol. Cell. Biol. 22, 2939-2951.
Leberer, E., Dignard, D., Harcus, D., Thomas, D.Y., and Whiteway, M. (1992). The protein kinase homologue Ste20p is required to link the yeast pheromone response G-protein beta gamma subunits to downstream signalling components. EMBO J. 11, 4815-4824.
Leberer, E., Wu, C., Leeuw, T., Fourest-Lieuvin, A., Segall, J.E., and Thomas, D.Y. (1997). Functional characterization of the Cdc42p binding domain of yeast Ste20p protein kinase. EMBO J. 16, 83-97.
Leeuw, T., Wu, C., Schrag, J.D., Whiteway, M., Thomas, D.Y., and Leberer, E. (1998). Interaction of a G-protein beta-subunit with a conserved sequence in Ste20/PAK family protein kinases. Nature 391, 191-195.
Lewis, T.A., Taylor, F.R., and Parks, L.W. (1985). Involvement of heme biosynthesis in control of sterol uptake by Saccharomyces cerevisiae. J. Bacteriol. 163, 199-207.
Liu, H., Styles, C.A., and Fink, G.R. (1993). Elements of the yeast pheromone response pathway required for filamentous growth of diploids. Science 262, 1741-1744.
Longtine, M.S., McKenzie, A., Demarini, D.J., Shah, N.G., Wach, A., Brachat, A., Philippsen, P., and Pringle, J.R. (1998). Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae. Yeast 14, 953-961.
Marie, C., Leyde, S., and White, TC. (2008). Cytoplasmic localization of sterol transcription factors Upc2p and Ecm22p in S. cerevisiae. Fungal Genet. Biol. 45, 1430-1438.
Martín, H., Mendoza, A., Rodríguez-Pachón, J.M., Molina, M., and Nombela, C. (1997). Characterization of SKM1, a Saccharomyces cerevisiae gene encoding a novel Ste20/PAK-like protein kinase. Mol. Microbiol. 23, 431-444.
Miller, W.G., and S.E. Lindow. 1997. An improved GFP cloning cassette designed for prokaryotic transcriptional fusions. *Gene* 191, 149-153.
Myers, A.M., Tzagoloff, A., Kinney, D.M., and Lusty, C.J. (1986). Yeast shuttle and integrative vectors with multiple cloning sites suitable for construction of lacZ fusions. *Gene* 45, 299-310.
Ness, F., Bourot, S., Régnacq, M., Spagnoli, R., Bergès, T., and Karst, F. (2001). SUT1 is a putative Zn[II]2Cys6-transcription factor whose upregulation enhances both sterol uptake and synthesis in aerobically growing Saccharomyces cerevisiae cells. *Eur. J. Biochem.* 268, 1585-1595.
Ni, L., and Snyder, M. (2001). A genomic study of the bipolar bud site selection pattern in Saccharomyces cerevisiae. *Mol. Biol. Cell* 12, 2147-2170.
O'Rourke, S.M., and Herskowitz, I. (1998). The Hog1 MAPK prevents cross talk between the HOG and pheromone response MAPK pathways in Saccharomyces cerevisiae. *Genes Dev.* 12, 2874-2886.
Park, H.O., and Bi, E. (2007). Central roles of small GTPases in the development of cell polarity in yeast and beyond. *Microbiol. Mol. Biol. Rev.* 71, 48-96.
Parks, L.W., and Casey, W.M. (1995). Physiological implications of sterol biosynthesis in yeast. *Annu. Rev. Microbiol.* 49, 95-116.
Peter, M., Neiman, A.M., Park, H.O., van Lohuizen, M., and Herskowitz, I. (1996). Functional analysis of the interaction between the small GTP binding protein Cdc42 and the Ste20 protein kinase in yeast. EMBO J. 15, 7046-7059.
Pichler, H., and Riezman, H. (2004). Where sterols are required for endocytosis. Biochim. Biophys. Acta 1666, 51-61.
Raitt, D.C., Posas, F., and Saito, H. (2000). Yeast Cdc42 GTPase and Ste20 PAK-like kinase regulate Sho1-dependent activation of the Hog1 MAPK pathway. EMBO J. 19, 4623-4631.
Régnacq, M., Alimardani, P., El Moudni, B., and Bergès, T. (2001). SUT1p interaction with Cyc8p(Ssn6p) relieves hypoxic genes from Cyc8p-Tup1p repression in Saccharomyces cerevisiae. Mol. Microbiol. 40, 1085-1096.
Reiner, S., Micolod, D., Zellnig, G., and Schneiter, R. (2006). A genomewide screen reveals a role of mitochondria in anaerobic uptake of sterols in yeast. Mol. Biol. Cell 17, 90-103.
Roberts, R.L., and Fink, G.R. (1994). Elements of a single MAP kinase cascade in Saccharomyces cerevisiae mediate two developmental programs in the same cell type: mating and invasive growth. Genes Dev. 8, 2974-2985.
Robinson, L.C., Menold, M.M., Garrett, S., and Culbertson, M.R. (1993). Casein kinase I-like protein kinases encoded by YCK1 and YCK2 are required for yeast morphogenesis. Mol. Cell. Biol. 13, 2870-2881.
Sauer, N., and Stolz, J. (1994). SUC1 and SUC2: two sucrose transporters from Arabidopsis thaliana; expression and characterization in baker's yeast and identification of the histidine-tagged protein. Plant J. 6, 67-77.
Schjerling, P., and Holmberg, S. (1996). Comparative amino acid sequence analysis of the C6 zinc cluster family of transcriptional regulators. Nucleic Acids Res. 24, 4599-4607.
Shianna, K.V., Dotson, W.D., Tove, S., and Parks, L.W. (2001). Identification of a UPC2 homolog in Saccharomyces cerevisiae and its involvement in aerobic sterol uptake. J. Bacteriol. 183, 830-834.
Sikorski, R.S., and Hieter, P. (1989). A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122, 19-27.
Sturley, S.L. (2000). Conservation of eukaryotic sterol homeostasis: new insights from studies in budding yeast. Biochim. Biophys. Acta 1529, 155-163.
Takahashi, S., and Pryciak, P.M. (2007). Identification of Novel Membrane-binding Domains in Multiple Yeast Cdc42 Effectors. Mol. Biol. Cell 18, 4945-4956.
Tiedje, C., Holland, D.G., Just, U., and Höfken, T. (2007). Proteins involved in sterol synthesis interact with Ste20 and regulate cell polarity. J. Cell Sci. 120, 3613-3624.
Wilcox, L.J., Balderes, D.A., Wharton, B., Tinkelenberg, A.H., Rao, G., and Sturley, S.L. (2002). Transcriptional profiling identifies two members of the ATP-binding cassette transporter superfamily required for sterol uptake in yeast. J. Biol. Chem. 277, 32466-32472.
Wild, A.C., Yu, J.W., Lemmon, M.A., and Blumer, K.J. (2004). The p21-activated protein kinase-related kinase Cla4 is a coincidence detector of signaling by Cdc42 and phosphatidylinositol 4-phosphate. *J. Biol. Chem.* 279, 17101-17110.
Winters, M.J., Lamson, R.E., Nakanishi, H., Neiman, A.M., and Pryciak, P.M. (2005). A membrane binding domain in the ste5 scaffold synergizes with gbetagamma binding to control localization and signaling in pheromone response. *Mol. Cell* 20, 21-32.
Winters, M.J., and Pryciak, P.M. (2005). Interaction with the SH3 domain protein Bem1 regulates signaling by the Saccharomyces cerevisiae p21-activated kinase Ste20. *Mol. Cell. Biol.* 25, 2177-2190.
Winzeler, E.A., et al. (1999). Functional characterization of the S. cerevisiae genome by gene deletion and parallel analysis. *Science* 285, 901-906.
Wittke, S., Lewke, N., Müller, S., and Johnsson, N. (1999). Probing the molecular environment of membrane proteins in vivo. *Mol. Biol. Cell* 10, 2519-2530.
Wu, C., Whiteway, M. Thomas, D.Y., and Leberer, E. (1995). Molecular characterization of Ste20p, a potential mitogen-activated protein or extracellular signal-regulated kinase kinase (MEK) kinase kinase from Saccharomyces cerevisiae. *J. Biol. Chem.* 270, 15984-15992.
**Figure legends**
**Figure 1.** Transcriptional regulation of genes mediating sterol uptake.
In the presence of oxygen, yeast cells synthesize sterols and are unable to take up sterols from the extracellular medium. In contrast, under anaerobiosis cells rely on sterol uptake, because sterol synthesis requires oxygen. Under these conditions, the transcriptional regulators Sut1, Upc2 and Ecm22 promote the expression of *AUS1*, *DAN1* and *PDR11* and less well characterized genes (Y), which mediate sterol import. The underlying mechanisms of this transcriptional regulation are largely unknown. Here, we show that the PAKs Ste20, Cla4 and Skm1 can translocate into the nucleus, where they bind to the transcriptional regulator Sut1. The fact that the PAKs negatively regulate expression of *AUS1*, *DAN1* and *PDR11*, suggests that PAKs inhibit Sut1 and/or Upc2 and Ecm22 by a novel mechanism. As demonstrated for Ste20, the down-regulation of gene expression requires its kinase activity. However, since there is no evidence for Sut1 phosphorylation by PAKs, an unknown protein (X), phosphorylated by PAKs, may be involved in the control of transcriptional regulators. Since sterol synthesis is crucial for cell polarization under aerobic conditions, it seems likely that sterol uptake plays an equally important role in the absence of oxygen.
**Figure 2.** PAKs form a complex with Sut1.
(A) The split-ubiquitin technique. See text for details. USPs, ubiquitin-specific proteases.
(B) Ste20 interacts with Sut1 using the split-ubiquitin system. $10^5$ cells of the indicated plasmid combinations in a *ste20Δ* background were spotted onto media either lacking histidine and leucine to select for the plasmids or lacking histidine, leucine and uracil to monitor protein interactions. The unrelated proteins Rd1 and Ubc6 served as negative controls.
(C) Ste20 interacts with Sut1 in vitro. Purified GST and GST-Sut1 bound to glutathione sepharose beads were incubated with an extract from cells expressing *STE20-3HA* from the plasmid pKA86. Eluted proteins were analyzed by immunoblotting using anti-HA antibodies.
(D) Cla4 and Skm1 bind to Sut1. Recombinant His$_6$-Sut1 and His$_6$-Sec6 bound to Ni-NTA beads were incubated with lysates from cells expressing *3HA-CLA4* and *3HA-SKM1*, respectively. Eluted proteins were analyzed by immunoblotting using anti-HA antibodies.
**Figure 3.** Ste20 localizes to the nucleus.
(A) Sequence of the NLS and BR domain. The NLS is composed of two clusters of basic residues, whereas the BR domain consists of three clusters (BR-1, -2 and -3) (Takahashi and Pryciak, 2007). The second cluster of the NLS is identical to BR-1.
(B) The Ste20 NLS is functional. Exponentially growing wild type cells carrying the indicated Ste20 fragments fused to GST-GFP were incubated with galactose for 1 hour. Cells were subsequently fixed with formaldehyde and stained with DAPI. One asterisk denotes the fusion with the peptide SGAGAGAGAIL. Two asterisks indicate addition of the sequence LGYFLFFEGGPGTQFAL.
(C) Ste20 lacking the NLS is excluded from the nucleus. Logarithmically growing wild type strains carrying the indicated *GFP-STE20* alleles under control of a galactose-inducible promoter on a centromeric plasmid were cultivated in galactose-containing media for 1 hour. Arrowheads indicate the nucleus.
(D) Quantification of (C). n > 100
(E) In few cells, Ste20 was enriched in the nucleus. Exponentially growing wild type cells harboring GFP-Ste20 under control of a galactose-inducible promoter on a centromeric plasmid were incubated with galactose for 1 hour.
**Figure 4.** Nuclear localization of Skm1 and Cla4.
(A) Skm1 is strongly enriched in the nucleus. Exponentially growing cells carrying the indicated *GFP-SKM1* alleles were induced for 1 hour by the addition of galactose. Cells were fixed with formaldehyde and stained with DAPI.
(B) Wild type *SKM1* and its derivatives are expressed at comparable levels. The indicated strains were incubated with galactose for 1 hour. Cells were lysed and equal amounts of protein extract were analyzed by immunoblotting using antibodies against GFP and GAPDH (loading control).
(C) Residues 201-230 are sufficient for nuclear targeting of Skm1. The Skm1 fragment was expressed as galactose-inducible GST-GFP fusion.
(D) Fragments used to map the sequence that targets Cla4 to the nucleus. The kinase domain and regions that are required for membrane association are highlighted. The CRIB domain is the minimal Cdc42-binding motif (Cvrckova et al., 1995), a proline-rich region (Pro) mediates binding to the scaffold protein Bem1 (Winters and Pryciak, 2005), and the PH domain is required for the association with membrane phosphoinositides (Wild et al., 2004).
(E) Cla4 fragments localize to the nucleus. Cla4 fragments were expressed as GST-GFP fusions from a glactose-inducible promoter in wild type cells.
**Figure 5.** PAKs down-regulates the expression of *DAN1*, *AUS1* and *PDR11*.
(A) *AUS1* expression is regulated by Ste20, Cla4 and Skm1. The indicated strains harboring a plasmid on which the *lacZ* was fused to the promoter region of *AUS1* were grown in selective raffinose medium in the presence of ergosterol and Tween 80. The indicated genes were overexpressed for 18 hours by the addition of galactose. Shown is the mean of β-galactosidase activity of at least 6 independent experiments with SD.*, *P < 0.0001*, compared with *hem1Δ* cells without any overexpression.
(B) Ste20 and Cla4 control the expression of *DAN1*. The indicated strains carrying a *DAN1-lacZ* plasmid were treated as described in (A). Shown is the mean of β-galactosidase activity of
at least 6 independent experiments. *, $P < 0.0001$; **, $P < 0.001$, compared with hem1Δ cells without any overexpression.
(C) PDR11 expression is negatively regulated by Cla4 and Skm1. β-galactosidase activity of the indicated strains harboring a PDR11-lacZ plasmid was determined in at least 6 independent experiments. *, $P < 0.0001$; **, $P < 0.005$, compared with hem1Δ cells without any overexpression.
**Figure 6.** Expression of AUS1, DAN1 and PDR11 in PAK deletion strains.
(A) STE20 deletion results in higher AUS1 expression. The indicated strains carrying an AUS1-lacZ plasmid were grown in selective medium supplemented with ergosterol and Tween 80, and β-galactosidase activity was determined in at least 6 independent experiments. wt, wild type; *, $P < 0.001$, compared with hem1Δ cells.
(B) PAK deletion does not affect DAN1 levels. Shown is the mean β-galactosidase activity of at least 6 independent experiments with SD of the indicated strains harboring DAN1-lacZ on a plasmid.
(C) Cells lacking SKM1 express higher PDR11 levels. β-galactosidase activity of the indicated strains harboring a PDR11-lacZ plasmid was determined in at least 6 independent experiments. *, $P < 0.0005$, compared with hem1Δ cells.
**Figure 7.** Down-regulation of AUS1 expression by PAKs depends on SUT1.
(A) Effect of SUT1 deletion on AUS1 levels following STE20 overexpression. The β-galactosidase activity of the indicated strains was determined in 6 independent experiments as described in Figure 5A. *, $P < 0.0001$; **, $P < 0.05$, compared with hem1Δ cells without STE20 overexpression.
(B) Regulation of *AUS1* expression by Cla4 depends on *SUT1*. Shown is the mean β-galactosidase activity of 6 independent experiments with SD of the indicated strains. *, $P < 0.0001$; **, $P < 0.05$, compared with *hem1Δ* cells without *CLA4* overexpression.
(C) The down-regulation of *AUS1* expression by Skm1 requires *SUT1*. β-galactosidase activity of the indicated strains was determined in at least 6 independent experiments. *, $P < 0.0001$, compared with *hem1Δ* cells without *SKM1* overexpression.
(D) Effect of *SUT1* deletion on *AUS1* expression in cells lacking *STE20*. The β-galactosidase activity of the indicated strains was determined in at least 6 independent experiments as described in Figure 6A. *, $P < 0.0005$, compared with *hem1Δ* cells.
**Figure 8.** The regulation of *AUS1* expression requires kinase activity and nuclear localization of Ste20.
(A) Wild type *STE20* and mutant versions were overexpressed in the indicated strains as described in Figure 5A. Shown is the mean β-galactosidase activity of at least 6 independent experiments with SD. *, $P < 0.0001$; **, $P < 0.05$, compared with *hem1Δ* cells without *STE20* overexpression.
(B) Wild type *STE20* and its mutant alleles are expressed at comparable levels. The indicated strains were incubated with galactose for 18 hours. Cells were lysed and equal amounts of protein extract were analyzed by immunoblotting. An N-terminal 3HA tag allowed the detection of Ste20. Cdc11 was used as loading control.
(C) Wild type and mutant alleles of *STE20* expressed from the native *STE20* promoter were integrated into *ste20Δ* and *ste20Δ hem1Δ* cells. *AUS1* expression was determined in at least 6 independent experiments as described in Figure 6A. *, $P < 0.0005$, compared with *ste20Δ hem1Δ* cells.
**Figure 9.** The down-regulation of *PDR11* expression requires nuclear localization of Skm1.
*SKM1* and *SKM1ΔNLS*, respectively were integrated into *skm1Δ* and *skm1Δ hem1Δ* cells. *PDR11* expression was determined in at least 6 independent experiments as described in Figure 6C. *, $P < 0.0005$, compared with *skm1Δ hem1Δ* cells
**Figure 10.** Ste20 regulates gene expression by two distinct mechanisms.
(A) The down-regulation of *AUS1* expression is independent of the MAPK kinase kinase Ste11. The β-galactosidase activity was determined in 8 independent experiments as described in Figure 5A. *, $P < 0.0001$, compared with *hem1Δ* cells without *STE20* overexpression.
(B) Haploid invasive growth does not require nuclear localization of Ste20. $10^5$ cells of the indicated strains were spotted on a YPD plate and incubated for 2 days at 30°C. Pictures were taken before and after gentle rinsing with water.
(C) The NLS of Ste20 is not necessary for the formation of a mating projection. Exponentially growing cells of the indicated strains were incubated in selective medium with 1 μg/ml α-factor for 3 hours. Cells were then fixed with formaldehyde and the percentage of cells with a mating projection was determined. Given is the mean of three independent experiments with SD bars (n > 100 for each experiment).
**Figure 11.** Ste20, Cla4 and Skm1 negatively regulate sterol uptake.
(A) Ste20, Skm1 and Cla4 over-accumulate free sterols. Heme-deficient cells of the indicated genotype were cultured in media containing [14C]cholesterol for the indicated period of time. Cells were collected, lipids were isolated, separated by TLC and quantified by radio-scanning. Data are normalized to free sterols and steryl ester levels in wild type cells after 24 hours of uptake. Values are means and SD of two independent experiments. The *aus1Δ pdr11Δ* mutant was used as negative control.
(B) Overexpression of PAKs results in a decreased sterol import. For PAK overexpression, *SKM1* and *CLA4* were placed under control of the *GAL1* promoter by genomic integration. These cells also harbored *STE20* expressed from the *GAL1* promoter on a plasmid. Wild type cells carried the corresponding empty vector. Both strains were grown in SC-Ura medium supplemented with galactose and [¹⁴C]cholesterol. Sterols and steryl ester were analyzed as described in (A).
PAKs
transcriptional regulators
sterol uptake proteins
cell polarization
A
B
C
D
GST-Sut1
Ste20-3HA
His\textsubscript{6}-Sec6
His\textsubscript{6}-Sut1
3HA-Cla4
3HA-Skm1
A
BR
BR-1
BR-2
BR-3
272KKQTEKSYYSSSSKKRKSGNSGTLRMKDVFVQNIKR511
NLS
B
GFP DAPI P_GAL1-GST-GFP fusions
vector
STE20272-288*
STE20272-288**
STE20272-400
STE20285-400
C
GFP DAPI P_GAL1-GST-GFP fusions
STE20
STE20Δ272-288
STE20Δ272-284
D
Cells with Ste20 excluded from nucleus (%)
Ste20 40% 50%
E
GFP DAPI P_GAL1-GST-GFP-STE20
A
GFP DAPI
GFP-Skm1
GFP-Skm1Δ213-230
GFP-Skm1C1
GFP-Skm1C2
B
wild type GFP-Skm1 GFP-Skm1Δ213-230 GFP-Skm1C1 GFP-Skm1C2
α-GFP α-GAPDH
C
GFP DAPI
P_GAL1-GST-GFP-SKM1201-230
D
PH CRIB Pro kinase
181-400
181-550
400-550
450-550
E
GFP DAPI P_GAL1-GST-GFP
CLA4
CLA4181-400
CLA4181-550
CLA4400-550
CLA4450-550
A
β-galactosidase activity (U)
AUS1
HEMI hem1Δ
overexpression
B
β-galactosidase activity (U)
DAN1
HEMI hem1Δ
overexpression
C
β-galactosidase activity (mU)
PDR11
HEMI hem1Δ
overexpression
A
β-galactosidase activity (U)
AUS1
HEM1 hem1Δ
wt ste20Δ cla4Δ skm1Δ
B
β-galactosidase activity (U)
DANI
HEM1 hem1Δ
wt ste20Δ cla4Δ skm1Δ
C
β-galactosidase activity (mU)
PDR11
HEM1 hem1Δ
wt ste20Δ cla4Δ skm1Δ
*
A
β-galactosidase activity (U)
HEMI
hem1Δ
overexpression
B
HEMI
hem1Δ
overexpression
α-HA
α-Cdc11
C
β-galactosidase activity (U)
ste20Δ
ste20Δ hem1Δ
* * *
β-galactosidase activity (mU)
PDR11
skm1Δ
SKM1ΔNLS
skm1Δ hem1Δ
SKM1ΔNLS
*
A
β-galactosidase activity (U)
- + - + - +
wild type ste11Δ hem1Δ hem1Δ ste11Δ
STE20 overexpression
B
total growth filamentous growth
wild type ste20Δ
ste20Δ + STE20 ste20Δ + STE20ΔNLS
C
Cells with mating projection (%)
vector vector STE20 STE20ΔNLS
wild type ste20Δ
A
wild type
% wild type
Time (h)
aus1Δ pdr11Δ
% wild type
Time (h)
ste20Δ
% wild type
Time (h)
skm1Δ
% wild type
Time (h)
cla4Δ
% wild type
Time (h)
cholesterol
steryl ester
B
wild type
% wild type
Time (h)
PAK overexpression
% wild type
Time (h) |
The Las Vegas Stadium Authority Board met in regular session in full conformity with law and bylaws of said Board at the regular place of meeting in the Commission Chambers, Government Center, Las Vegas, Clark County, Nevada on Wednesday, the 23rd day of May, 2018 at the hour of 9:00 a.m. The meeting was called to order at the hour of 9:01 a.m. by Chairman Hill and on roll call, the following members were present, constituting a quorum of the members thereof:
ITEM NO. 1. Call to Order, Roll Call and Establish Quorum
Chairman and Members:
Steve Hill
Lawrence Epstein
Ken Evans
Bill Hornbuckle
Mike Newcomb
Jan Jones Blackhurst
J. Tito Tiberti
Tommy White
Absent:
Dallas Haun
Also present:
Laura Fitzpatrick (Ex-Officio Member)
Lori Sabella, Deputy Clerk
Beth Koch, Deputy Clerk
DISCUSSION: At this time, Chairman Hill called the roll and advised that a quorum was present.
Chairman Hill closed Agenda Item No. 1.
ITEM NO. 2 Public Comment
At this time, Chairman Hill asked if there were any persons present in the audience wishing to be heard on any items on the agenda as posted.
SPEAKER(S): Present
The Board was addressed by interested parties who expressed concerns regarding the participation of minority businesses on the Stadium project; Chronic Traumatic Encephalopathy (CTE), and concussions in football; the gubernatorial race; and room tax revenues.
There being no other persons present in the audience wishing to be heard on any items listed on the agenda as posted, Chairman Hill closed the public comments.
ITEM NO. 3 Approval of Agenda with the Inclusion of Any Emergency Items and Deletion of Any Items (For possible action)
FINAL ACTION: It was moved by Member Tommy White and seconded by Member Ken Evans and carried by unanimous vote of the members present that the agenda be approved.
Chairman Hill closed Agenda Item No. 3.
ITEM NO. 4 Chairman/Board Member Comments
DISCUSSION: Chairman Hill commented on the progress being made on the stadium, and added that other administrative issues would be addressed today.
Member Evans shared a brief review of the Community Benefits Plan Site Committee’s activities; and staff responded to questions posed by Member Hornbuckle regarding the decks being used in the stadium.
Chairman Hill closed Agenda Item No. 4.
ITEM NO. 5 Review and Potentially Approve Minutes of the Las Vegas Stadium Authority Board Meeting on March 22, 2018 (For possible action)
FINAL ACTION: It was moved by Member Tommy White, seconded by Member Ken Evans and carried by unanimous vote of the members present that the minutes of the Regular Meeting held on March 22, 2018 be approved.
Chairman Hill closed Agenda Item No. 5.
ITEM NO. 6 Review and Potentially Approve Minutes of the Las Vegas Stadium Authority Board Meeting on March 28, 2018 (For possible action)
FINAL ACTION: It was moved by Member Ken Evans, seconded by Member Tommy White, and carried by unanimous vote of the members present that the minutes of the Regular Meeting held on March 28, 2018 be approved.
Chairman Hill closed Agenda Item No. 6.
ITEM NO. 7 Review and Potentially Approve Minutes of the Las Vegas Stadium Authority Board Meeting on April 12, 2018 (For possible action)
FINAL ACTION: It was moved by Member Ken Evans, seconded by Member Tommy White, and carried by unanimous vote of the members present that the minutes of the Regular Meeting held on April 12, 2018 be approved.
Chairman Hill closed Agenda Item No. 7.
ITEM NO. 8 Receive a Status Update from Staff on Administrative Items Relating to the Operation of the Stadium Authority
DISCUSSION: A representative from Applied Analysis provided an update on administrative matters, including a review of funding notices; project status reports; budget processes; Stadium Authority bonds issuance; preparations for the audit due to the State in November; and room tax contributions.
Chairman Hill clarified that room tax revenue is generated approximately two months after earning, and is not accounted for on a full accrual basis due to governmental accounting methods; and staff then responded to Chairman Hill’s question regarding how room tax projections will work in the future.
Staff continued the presentation advising a Request for Quotation (RFQ) for an Authority Construction Representative had been issued, advising that eleven firms had submitted proposals; three finalists were selected and additional
proposals were received; interviews were conducted and staff has recommended that the contract be awarded to Grand Canyon Development Partners (GCDP).
Also discussed was the Community Benefits Plan Oversight Committee’s activities, advising a portal on the Las Vegas Stadium Authority’s website had been created which will be updated with information pertaining to the committee.
Upon questioning by Member Evans, a representative from Applied Analysis clarified the process on handling residual funds, advising that the topic will be discussed through an upcoming agenda item.
Chairman Hill closed Agenda Item No. 8.
ITEM NO. 9
Receive a Status Update from LV Stadium Events Company on the Progress of the NFL Stadium Project
SPEAKER(S): Present
DISCUSSION: A representative from the LV Stadium Company provided a report on the matter which included details relating to project completion status; substantial completion time frame; workforce diversity; box culvert relocation; and project costs through mid-May.
In response to questioning by Member Evans, a representative from the LV Stadium Company advised that a RFQ for a master concessionaire was drafted with the expectation of a vendor equity being required; and advised that a portal was being created to announce business opportunities which can be shared in an effort to engage small local businesses.
Member Hornbuckle requested an update on the Nevada Department of Transportation’s (NDOT) schedule for the roadway adjustments and improvements necessary to accommodate the anticipated increase in traffic. The representative from the LV Stadium Company discussed the changes that will be made including high occupancy carpool lanes on Interstate-15 with off-ramps at Harmon Avenue; and commented on major improvements to Tropicana Avenue and to the center median on Hacienda Avenue.
Concerns were expressed by the representative from the LV Stadium Company regarding NDOT meeting the completion dates prior to opening day, but assured the Board that the County and NDOT were working together to create a timely delivery of all facets of the roadway changes required, and
assured the Board that the situation was being closely monitored.
Chairman Hill closed Agenda Item No. 9.
ITEM NO. 10 Review and Potentially Approve a Professional Services Agreement with Grand Canyon Development Partners to Provide General Construction and Development Advisory Services as the Authority Construction Representative and, if Approved, Authorize the Chairman to Sign the Agreement (For possible action)
DISCUSSION: Chairman Hill introduced the item, followed by a brief update from a representative of Applied Analysis regarding documents that were submitted by Grand Canyon Development Partners.
FINAL ACTION: It was moved Member Bill Hornbuckle, seconded by Member Tommy White, and carried by unanimous vote of the members present that the recommendation be approved.
Chairman Hill closed Agenda Item No. 10.
ITEM NO. 11 Review and Potentially Approve and Authorize the Chairman to Sign the Release of Settlement Agreement between the Stadium Authority, Clark County, and the Nevada Department of Transportation (For possible action)
DISCUSSION: Following introduction of the item, a representative from Brownstein Hyatt Farber Schreck provided a report on the matter and verified that all issues relating to the 1994 settlement agreement had been resolved, and that the title to the land in question was clear.
FINAL ACTION: It was moved by Member Tommy White, seconded by Member Bill Hombuckle, and carried by unanimous vote of the members present that the recommendation be approved.
Chairman Hill closed Agenda Item No. 11.
ITEM NO. 12 Review and Potentially Approve and Authorize the Chairman to Sign the Personal Seat License Purchase and Sale Agreement Between the Clark County Stadium Authority, Financing Trust I, and Raiders Football Club, LLC (For possible action)
DISCUSSION: Following introduction of the item, a representative from Applied Analysis spoke on major changes to the Personal Seat License (PSL) agreement which included allowing PSL upsizing; the elimination of multiple banks for clearing and holding accounts; and administrative clean up corrections.
In response to a question from Member Tiberti, staff advised there were multiple authors to the agreement including DLA Piper, Morgan, Lewis & Bockius LLP, Brownstein Hyatt Farber Schreck; and also language utilized from other stadium project contracts.
FINAL ACTION: It was moved by Member Tito Tiberti, seconded by Member Ken Evans, and carried by unanimous vote of the members present that the recommendation be approved.
Chairman Hill closed Agenda Item No. 12.
ITEM NO. 13 Review and Potentially Approve and Authorize the Chairman to Sign the Authority PSL Account Agreement Between the Clark County Stadium Authority, Financing Trust I, Raiders Football Club, LLC, and Bank of America, N.A. (For possible action)
DISCUSSION: Following introduction of the item, a representative from Applied Analysis reviewed changes that were made to the PSL agreement and procedures which included the concept of the initial purchase; security language that confirmed there was a single bank for all transactions; and added language regarding future amendments.
FINAL ACTION: It was moved by Member Ken Evans, seconded by Member Tito Tiberti, and carried by unanimous vote of the members present that the recommendation be approved.
Chairman Hill closed Agenda Item No. 13.
ITEM NO. 14 Review and Potentially Approve and Authorize the Chairman to Sign the Calculation Agency Agreement Between Financing Trust I, the Clark County Stadium Authority, and Raiders Football Club, LLC (For possible action)
DISCUSSION: Following introduction of the item, a representative from Applied Analysis explained the purpose of the Calculation Agency Agreement, and reviewed the non-material changes that were made to the document which
included verbiage on the initial purchase; clarification of banks; and various administrative clean up items.
FINAL ACTION: It was moved by Member Tito Tiberti, seconded by Member Ken Evans, and carried by unanimous vote of the members present that the recommendation be approved.
Chairman Hill closed Agenda Item No. 14.
ITEM NO. 15 Review and Potentially Approve a Revised Project Budget for the Las Vegas Stadium Project (For possible action)
DISCUSSION: Following introduction of the item, a representative from Applied Analysis reviewed the Stadium Project budget report.
In response to a question from Member Hornbuckle, the representative from Applied Analysis advised that the site report will be issued prior to the funding budget; and that a simpler format of the report and budget will be created and provided to the Board.
FINAL ACTION: It was moved by Member Tito Tiberti, seconded by Member Bill Hornbuckle, and carried by unanimous vote of the members present that the recommendation be approved.
Chairman Hill closed Agenda Item No. 15.
ITEM NO. 16 Review and Potentially Approve, Adopt, and Authorize the Chairman to Sign a Resolution Establishing the Stadium Authority Waterfall Residual Fund (2965) (For possible action)
DISCUSSION: Following introduction of the item, a representative from Applied Analysis explained the various aspects of the Waterfall Residual Fund and the purpose behind the resolution to the document.
In response to Member Hill’s question, the representative from Applied Analysis advised that the $5 million in capital funds would not be included in the waterfall and would be kept separate.
FINAL ACTION: It was moved by Member Tommy White, seconded by Member Ken Evans, and carried by unanimous vote of the members present that the recommendation be approved.
ITEM NO. 17 Review and Potentially Approve Adopt, and Authorize the Chairman to Sign a Resolution to Augment the Fiscal Year 2018 Stadium Authority Fund (2960) Budget (For possible action)
DISCUSSION: Following introduction of the item, a representative from Applied Analysis presented details and clarifications relating to the augmentation of the FY 2018 Stadium Authority Fund budget.
The representative from Applied Analysis responded to Member Evans regarding the amount of funds going into the next fiscal year.
FINAL ACTION: It was moved by Member Tommy White, seconded by Member Ken Evans, and carried by unanimous vote of the members present that the recommendation be approved.
Chairman Hill closed Agenda Item No. 17.
ITEM NO. 18 Review and Potentially Approve the Revised Fiscal Year 2018 Transfer Reconciliation Schedule and Authorize the Listed Transfer Amounts for Various Stadium Authority Funds (For possible action)
DISCUSSION: Following introduction of the item, the representative from Applied Analysis discussed the transfer reconciliations process, and advised that regulation and recording is a requirement of the State of Nevada.
In response to Member Evans, the representative of Applied Analysis advised that funds cannot be accessed until after they have flowed through to the waterfall account.
FINAL ACTION: It was moved by Member Tommy White, seconded by Member Ken Evans, and carried by unanimous vote of the members present that the item be approved.
Chairman Hill closed Agenda Item No. 18.
ITEM NO. 19 Public Hearing – 9:30 a.m. Conduct a Public Hearing on the Fiscal Year 2019 Tentative Budget for the Clark County Stadium Authority (For possible action)
DISCUSSION: There being no objections, Item No. 19 was taken in conjunction with Item No. 20.
Following introduction of the item, Chairman Hill asked if there were any persons present in the audience wishing to be heard on the matters.
SPEAKER(S): None
There being no persons present in the audience wishing to be heard on the matters, Chairman Hill closed the public hearing.
A representative from Applied Analysis commented on various issues including matters relating to the tentative budget including a letter from the Department of Taxation that was issued to the Clark County Chief Financial Officer which indicated that the tentative budget was filed in accordance Nevada Revised Statute No. 354596; a review of taxes; interest earnings; expenditures; debt service principal; room tax revenues; excess revenues from bonds; a summary of accounts; debt schedules, and budget reports.
FINAL ACTION: It was moved by Member Tommy White, seconded by Member Ken Evans, and carried by unanimous vote of the members present that the recommendation be approved.
Chairman Hill closed Agenda Item No. 19.
ITEM NO. 20 Review and Potentially Approve the Final Budget for the Clark County Stadium Authority for Fiscal Year 2019; and Direct the County to Transmit the Approved Documents to the State of Nevada Department of Taxation, as Prescribed by Law (For possible action)
DISCUSSION: There being no objections, Item No. 20 was taken in conjunction with Item No. 19 (see Item No. 19 for discussion).
FINAL ACTION: It was moved by Member Tommy White, seconded by Member Ken Evans, and carried by unanimous vote of the members present that the recommendation be approved.
Chairman Hill closed Agenda Item No. 20
ITEM NO. 21 Public Comment
At this time, Chairman Hill asked if there were any persons present in the audience wishing to be heard on any items not listed on the posted agenda.
SPEAKER(S): Present
The Board was addressed by interested parties who spoke regarding several issues including CTE and concussions in football; the potential of a team strike once the stadium opens in 2020; and opposition to the Stadium project.
There being no other persons present in the audience wishing to be heard on any items listed on the agenda as posted, Chairman Hill closed the public comments.
ITEM NO. 22 Adjournment (For possible action)
The meeting was adjourned at the hour of 11:01 a.m. |
Power flow variation based on extreme learning machine algorithm in power system
Labed Imen, Labed Djamel
Laboratory of Electrical Engineering, Department Of Electrical Engineering, University of Constantine 1, Algeria
ABSTRACT
The main focus of this paper is a study that empowers us to understand how the temperature variation affects the transmission line resistance and as a result the power flow analysis with a specific end goal to assess losses in the electrical network. The paper is composed of two sections; the first part is a power flow study under normal conditions utilizing the neural network approach while the second investigated extreme learning machine algorithm efficiency and exactitude. Extreme learning machine algorithm has been used to settle several complications in power system: load forecasting, fault diagnosis, economic dispatch, security, transient stability; Thus, we proposed to study this technique to figure out this sort of complex issue. The study was conducted for IEEE 30 bus test system. The simulation results are exposed and analyzed in detail at the end of this paper.
Corresponding Author:
Labed Imen
Laboratory of Electrical Engineering,
Department of Electrical Engineering,
University of Constantine 1, Algeria
Email: email@example.com
1. INTRODUCTION
Electric power systems are real-time energy delivery systems. Real-time means that power is generated transported and supplied the moment we turn on the light switch [1]. Electric power systems are not storage systems like water and gas systems. Instead, generators produce the energy as the demand calls for it [2, 3].
The huge amount of power generated in a power station (hundreds of MW) is to be transported over a long distance (hundreds of kilometers) to load centers to cater power to consumers with the help of high-voltage transmission line in order to reduce transportation losses. Transmission and distribution losses reach a rage of 4% to 5% of the total consumption. The higher-voltage transmission lines require larger structures with longer insulator strings in order to have greater air gaps and needed insulation.
Conductor material (all wires), type, size, and current rating are key factors in determining the power handling capability of transmission lines. A conductor heats up when current flows through it due to its resistance, conductors tend to increase their resistance with an increase in temperature, rated by how much current causes them to heat up to a predetermined amount of degrees above ambient temperature. The amount of temperature rise above ambient (i.e., when no current flows) determines the current rating of a conductor. For example, when a conductor reaches 70°C above ambient, the conductor is said to be at full load rating [1].
Scientists used to mention the dynamic thermal rating in literature. Conductors transporting energy in low temperatures are more efficient. For an overhead transmission line, the maximum temperature it supports is 80C° [4-6]. The conductor transmitting 700 amperes at 40C; can transport 1200 amperes at 10C°, which implies, from summer season to winter, transmission line capacity can reach the double [4, 6, 7].
The conductor temperature varies in the range of 0° to 60°C or even larger, resulting changes of the line resistance [3, 8, 9].
Meanwhile, the resistance fluctuations are explicitly taken into account and coupled to power flow analysis. Power flow fluctuations can cause various phenomena ranging from frequency and voltage stability, thermal contingency, to critical scenarios such as components outages and black out [10]. For the purposes of load flow determination and system state estimation, a variety of methods has been used. The development of these methods is mainly led by the basic requirements of load flow calculation, which can be summed up as
a. the convergence properties
b. the computing efficiency and memory requirements
c. the convenience and flexibility of the implementation
Mathematically, the load flow problem is a problem of solving a system of nonlinear algebraic equations. With the scale of power system continually expanding, the dimension of load flow equations now becomes very high, thus research is still very active. Newton Raphson technique stands to be the most important approach to solve the load flow. Moreover, this accurate technique became unacceptable for real-time electric power system operation (dynamic), the problems result mainly from the speed of operation and the dependability (no missing operations), hence the introduction of the artificial intelligence or more precisising extreme learning machine due to its extremely fast training, good generalization, and universal approximation/classification capability [11]. It has been demonstrated to have excellent learning accuracy/speed in various applications, such as face classification, image segmentation, human action recognition, prediction, and functions approximation. ELM offers effective solutions for the electric power system as it is demonstrated; in the article [12] a novel version of ELM, which can be applied to train a single hidden layer recurrent neural network. ELM was presented and employed in order to achieve higher forecasting electricity load accuracy. In [13] authors introduced an ELM-based predictor, then it was developed for the power system. The predictor has been used for real-time to enhance the dynamic security of the power systems. Article [14] explores a fault diagnosis method for the key mechanical components of the wind turbine. The method based on improved extreme learning machine (IELM) combined the fault characteristics of gear and bearing of wind turbine drive system. A robust observer to detect and isolate sensors fault in a variable speed wind turbine is discussed in [15]. The designed approach was based on the ELM algorithm for the prediction of unknown model dynamics to design the robust non-linear observer. The article [16] envisages the application of craziness-based PSO to enhance the performance of ELM for feature classification of single and combined power quality disturbances. Thus we introduced this fast and accurate approach in this paper to examine the variation of the transported power, load flow and losses due to the influence of temperature on conductor resistance. To start with, we attempted to study the first proposition which is the neural network. Then for further optimization and in terms of precision, exactness, and proficiency, extreme learning machine algorithm was applied to address the above problem.
2. RESEARCH METHOD
2.1. Neural networks
The origin of artificial neural networks comes from the biological neuron modeling test by Warren McCulluch and Walter Pitts. Initially, Neural networks' objective was: patterns recognition, classification, then it becomes very interesting in all domains [17].
2.1.1. Neuron Model
Indeed, the simple architecture of a single neuron with a single layer is composed essentially:
First, the scalar input $p$ which is multiplied by the scalar weight $w$ to form the product $wp$. This operation takes place at the integrator which is one of the important parts of the neuron; the result is again a scalar added to a scalar bias $b$ to form the net input $n$.
The result $n$ is then transformed by a transfer function $f$ which produces the neuron output $a$. [18, 19], the model is shown in Figure 1.
2.1.2. Neural Network Learning
Learning is a process by which the free parameters from a neural network are adapted, through a simulation process by the environment in which the network is contained [17, 19, 20].
The supervised and unsupervised methods are essentially the two types of learning for a neural network. In our case, we are going to use the supervised learning where the training is controlled by an external agent who watches the answer that the network supposed to generate from the determined entrance. The supervisor compares the output of the network with the expected one and determines the amount of modifications to be made on the weights [18, 21, 22].
The supervised learning can be done through the three following Paradigms:
a. Error correction learning
b. Reinforcement learning
c. Stochastic learning.
During the artificial neural network training by error correction, the weights of the communication links are adjusted trying to minimize a function of cost depending on the difference between the desired values and the obtained one from the output network [19, 23].
It is required to determine the modifications on weights using the committed error, the correction error is calculated as the following expression:
\[ \Delta w_{ki} = \eta y_j (d_k - y_k) \]
(2)
Where:
- \( \Delta w_{ki} \) is the weight variation of the connection between the neurons \( j \) from the anterior layer and the output layer node \( k \).
- \( y_k \): the desired neuron output.
- \( y_j \), and \( y_k \) are the output values produced in the neuron \( i \) and \( k \) respectively.
- \( \eta \) a positive factor denominated the learning rate \( 0 < \eta < 1 \) [18, 24, 25].
We can say that the new weight is proportional to the committed error produced by the network and the answer of the node \( j \) in the previous layer as shown in the following formula:
\[ w_{kj}^{\text{actual}} = \Delta w_{kj} + w_{kj}^{\text{anterior}} \]
(3)
The total error is known as the mean squared error expressed in the equation (4) [26, 27]:
\[ \text{Error}_{\text{global}} = \frac{1}{2P} \sum_{p=1}^{P} \sum_{K=1}^{N} \left( y_K^{(p)} - d_K^{(p)} \right)^2 \]
(4)
Where:
- \( N \) is the number of neurons in the output layer and \( P \) is the number of examples in the training sample.
Normally to minimize this criterion function, is employed learning rule given by the descending gradient [18, 19].
\[ \Delta w_{kj} = - \eta \nabla_{w_{kj}} (\text{Error}_{\text{global}}) \]
(5)
If \( \eta \) takes a little value the learning process is made smoothly which gives; as a result, an increment in the time of convergence to a stable solution. In the other hand if \( \eta \) has an important value the speed of learning gets increased but there will be a divergence risk and this causes the instability of the system [25-27].
### 2.2. Extreme Learning Machine ELM
Huang et al. proposed an extreme learning machine algorithm, as a new learning scheme working for single hidden layer feed-forward networks (SLFNs) [28,29]. ELM is entirely different from traditional iterative learning algorithms as it randomly selects the input weights (\( W \)) and biases (\( b \)) for hidden nodes and analytically calculates the output weights (\( \beta \)) by finding the least square solution.
In doing so, it is proven that the training error can still be minimized with even better generalization performance [30]. Notably, the learning speed of ELM can be thousands of times faster than traditional feed-forward network learning algorithms like BP algorithm. Compared with conventional learning algorithms, ELM not only tends to reach the smallest training error but also to obtain the smallest norm of weights [29, 31].
For \( N \) arbitrary distinct samples \((x_i, t_i)\) where:
\[
x_i = (x_{i1}, x_{i2}, \ldots, x_{in})^T \in R^n
\]
\[
t_i = (t_{i1}, t_{i2}, \ldots, t_{im})^T \in R^m
\]
Standard SFLNs with \( N^\infty \) nodes and activation function \( g(x) \) are mathematically modeled as:
\[
\sum_{i=1}^{N^\infty} \beta_i g(w_i \cdot x_j + b_i) = o_j
\]
\( j = 1, \ldots, N \)
Where \( w_i = (w_{i1}, w_{i2}, \ldots, w_{in})^T \) is the weight vector connection \((ith)\) hidden node and input nodes.
\( \beta_i = (\beta_{i1}, \beta_{i2}, \ldots, \beta_{im})^T \) is the weight vector connecting the \((ith)\) hidden node and output nodes, and \( b_i \): is the threshold of the \((ith)\) hidden node. \( w_i \cdot x_j \) denotes the inner product of \( w_i \) and \( x_j \).
That standard SFLN with \( N^\infty \) hidden nodes and \( g(x) \) activation function can approximate these \( N \) samples with zero.
\[
\sum_{j=1}^{N^\infty} \|o_j - t_j\| = 0
\]
What means there exist \( \beta_i, w_i \) and \( b_i \) such that:
\[
\sum_{i=1}^{N^\infty} \beta_i g_i(x_j) = \sum_{i=1}^{N^\infty} \beta_i g(w_i \cdot x_j + b_i) = t_j
\]
Equation (10) can be written compactly as
\[
H\beta = T
\]
Where:
\[
H(w_1 \ldots w_N, x_1 \ldots x_N, b_1 \ldots b_{N^\infty}) = \begin{bmatrix}
g(w_1 \cdot x_1 + b_1) & \ldots & g(w_{N^\infty} \cdot x_1 + b_{N^\infty}) \\
\vdots & \ldots & \vdots \\
g(w_1 \cdot x_N + b_1) & \ldots & g(w_{N^\infty} \cdot x_N + b_{N^\infty})
\end{bmatrix}_{N^\infty \times N}
\]
\[
\beta = \begin{bmatrix}
\beta_1 \\
\vdots \\
\beta_{N^\infty}
\end{bmatrix}_T
\quad \text{and} \quad
T = \begin{bmatrix}
t_1 \\
\vdots \\
t_N
\end{bmatrix}_T
\]
Thus, the extreme learning machine algorithm can be summarized as follows:
Step1: Randomly assign input weight \( W_i \) and bias \( b_i \), \((i = 1, 2, \ldots, N)\).
Step2: Calculate the hidden layer output matrix \( H \).
Step3: Calculate the output weight \( \beta \) by \( \beta = H^T T \)
If hidden layer output matrix is non-square (that means \( N^\infty < N \)), the determinant of \( H \) \( \det(H) = 0 \) and (rank \( H < N^\infty \)), in linear algebra matrix, is assessed as a singular matrix that cannot be directly inverted [32-34].
\[
\beta = (H^T H)^{-1} H^T T
\]
The inverse of \( H \) cannot be determined if \( H \) is not a full-rank matrix. In this case, the generalized inverse matrix, known as pseudo inverse, is used. In order to calculate the pseudo inverse, various approaches
can be used. The most widely known type is the Moore–Penrose pseudo-inverse [35]. The pseudo-inverse of the matrix $H$, namely $H^\dagger$, can be computed via least square solution:
$$H^\dagger = (H^T H)^{-1} H \quad (14)$$
This solution is accurate as long as the square matrix $(H^T H)$ is invertible. However, it is singular in some applications. In SLFN, it is singular in most of the cases since there is a tendency to select $m << n$. In ELM, Huang et al., have solved this problem using Moore–Penrose pseudo inverse of H matrix to overcome this problem [32, 36, 37].
3. RESULTS AND ANALYSIS
In order to verify the proposed two approaches; ELM and neural networks, the paper study is conducted on IEEE 30 bus test system shown in Figure 2, it comprises of 6 generators, bus number one (1) is considered as a slack bus. Likewise, 24 loads powered over 41 transmission lines.
![Figure 2. Single Line Diagram of IEEE-30bus [38]](image)
3.1. Line Flow Solution Using Newton-Raphson Algorithm
Table 1 displayed the line flow results at 25°C°. It is the initial step to provide, active and reactive power and the total system lossesThe whole system is operating under normal conditions. These results are used to be the input matrix for the neural network training phase.
| i to j | P (MW) | Q (MVAR) | S (MVA) | Losses (MW) | Losses (Mvar) | i to j | P (MW) | Q (MVAR) | S (MVA) | Losses (MW) | Losses (Mvar) |
|--------|--------|----------|---------|-------------|---------------|--------|--------|----------|---------|-------------|---------------|
| 1-2 | 177,778| -22,148 | 179,152 | 5,464 | 10,524 | 12- 13 | 0,000 | -10,291 | 10,291 | 0,000 | 0,133 |
| 1-3 | 83,221 | 5,127 | 83,378 | 2,808 | 7,085 | 12- 14 | 7,856 | 2,442 | 8,227 | 0,075 | 0,155 |
| 2-4 | 45,712 | 2,705 | 45,792 | 1,106 | -0,517 | 12- 15 | 17,857 | 6,947 | 19,161 | 0,217 | 0,428 |
| 2-5 | 82,990 | 1,703 | 83,008 | 2,995 | 8,178 | 12- 16 | 7,208 | 3,363 | 7,954 | 0,053 | 0,112 |
| 2-6 | 61,912 | -0,958 | 61,920 | 2,048 | 2,264 | 14- 15 | 1,582 | 0,687 | 1,724 | 0,006 | 0,005 |
| 3-4 | 78,012 | -3,158 | 78,076 | 0,771 | 1,344 | 15- 18 | 6,014 | 1,744 | 6,262 | 0,039 | 0,080 |
| 4-6 | 70,126 | -17,526 | 72,282 | 0,604 | 1,179 | 15- 23 | 5,001 | 2,956 | 5,810 | 0,031 | 0,063 |
| 4-12 | 44,121 | 14,646 | 46,489 | 0,000 | 4,685 | 16- 17 | 3,654 | 1,451 | 3,932 | 0,012 | 0,027 |
### 3.2. The Neural Network
#### 3.2.1. The training phase using Levenberg–Marquard algorithm
a. The training algorithm is Levenberg-Marquard (trainlm), as it is shown in (Figure 3).
b. The performance is the mean squared error (mse) (Figure 4).
c. The convergence is satisfied at t=15sec (Figure 3).
d. The performance value reached $3.15e-19$ (Figure 4).
In tables 1 and 2 and also (Figure 5), it is evident that the neural network outputs and targets are completely identical. The main squared error reaches the value $3.15e-19$ (Figure 4). Bias and weights are adjusted.
Here below we can see clearly the neural network architecture and training phase exposed in Figures 3, the performance in Figure 4 and the regression in Figure 5 respectively.
#### Table 2. Line flow based on a neural network at 25°C
| i to j | P (MW) | Q (MVAR) | Losses (MW) | Losses (Mvar) | i to j | P (MW) | Q (MVAR) | Losses (MW) | Losses (Mvar) |
|--------|--------|----------|-------------|---------------|--------|--------|----------|-------------|---------------|
| 1-2 | 177,778| -22,148 | 5,464 | 10,524 | 12- 13 | 1,705 e-13 | -10,291 | 1,421e-14 | 0,133 |
| 1-3 | 83,221 | 5,127 | 2,808 | 7,085 | 12- 14 | 7,856 | 2,442 | 0,075 | 0,155 |
| 2-4 | 45,712 | 2,705 | 1,106 | -0,517 | 12- 15 | 17,857 | 6,947 | 0,217 | 0,428 |
| 2-5 | 82,990 | 1,703 | 2,995 | 8,178 | 12- 16 | 7,208 | 3,363 | 0,053 | 0,112 |
| 2-6 | 61,912 | -0,958 | 2,048 | 2,264 | 14- 15 | 1,582 | 0,687 | 0,006 | 0,005 |
| 3-4 | 78,012 | -3,158 | 0,771 | 1,344 | 15- 18 | 6,014 | 1,744 | 0,039 | 0,080 |
| 4-6 | 70,126 | -17,526 | 0,604 | 1,179 | 15- 23 | 5,001 | 2,956 | 0,031 | 0,063 |
| 4- 12 | 44,121 | 14,646 | 7,105e-15 | 4,685 | 16- 17 | 3,654 | 1,451 | 0,012 | 0,027 |
| 5-7 | -14,205| 10,500 | 0,151 | -1,687 | 18- 19 | 2,775 | 0,765 | 0,005 | 0,010 |
| 6-7 | 37,523 | -1,885 | 0,367 | -0,598 | 19- 20 | -6,730 | -2,645 | 0,017 | 0,034 |
| 6-8 | 29,528 | -3,754 | 0,103 | -0,558 | 21- 22 | -1,877 | -1,594 | 0,001 | 0,001 |
| 6-9 | 27,693 | -7,322 | -7,105e-14 | 1,594 | 22- 24 | 5,654 | 2,788 | 0,043 | 0,067 |
| 6-10 | 15,823 | 0,653 | -7,993e-15 | 1,278 | 23- 24 | 1,770 | 1,293 | 0,006 | 0,012 |
*Power flow variation based on extreme learning machine algorithm in power system (Labeled Imen)*
Table 4 illustrates ELM training phase parameters.
| Table 4. ELM training phase parameters |
|----------------------------------------|
| Extreme learning machine | Values |
| Training time(s) | 0.0469 |
| Testing time(s) | 0.0000 |
| Training performance | 3,666e-25 |
| Testing performance | 8.248e-27 |
ELM model is built and ready to be used in real time.
Table 5 shows the line flow results provided by the ELM trained model for the two resistance variation: $R_{T=30C^\circ}$ and $R_{T=05C^\circ}$
| i to j | R(pu) 30C° | P (MW) | Q (MVar) | Loss (MW) | Loss (MVar) | R(pu) 05C° | P (MW) | Q (MVar) | Loss (MW) | Loss (MVar) |
|--------|------------|--------|----------|-----------|-------------|------------|--------|----------|-----------|-------------|
| 1-2 | 0.0195 | 178,010| -23,312 | 5,593 | 10,5910 | 0.0177 | 176,840| -17,501 | 4,960 | 10,273 |
| 1-3 | 0.0460 | 83,370 | 4,788 | 2,871 | 7,1190 | 0.0416 | 82,639 | 6,479 | 2,561 | 6,961 |
| 2-4 | 0.0581 | 45,709 | 2,525 | 1,127 | -0,5200 | 0.0525 | 45,719 | 3,432 | 1,024 | -0,507 |
| 2-5 | 0.0481 | 83,105 | 1,320 | 3,061 | 8,2080 | 0.0435 | 82,543 | 3,232 | 2,738 | 8,069 |
| 2-6 | 0.0592 | 61,909 | -1,262 | 2,087 | 2,2630 | 0.0535 | 61,918 | 0,268 | 1,890 | 2,270 |
| 3-4 | 0.0134 | 78,099 | -3,531 | 0,788 | 1,3500 | 0.0121 | 77,678 | -1,682 | 0,703 | 1,320 |
| 4-6 | 0.0121 | 70,138 | -18,018 | 0,618 | 1,1880 | 0.0109 | 70,080 | -15,564 | 0,549 | 1,147 |
| 4-12 | 0.0000 | 44,155 | 14,581 | -1,774e-12| 4,6890 | 0,0000 | 43,989 | 14,901 | 1,885e-13 | 4,669 |
| 5-7 | 0.0468 | -14,155| 10,72 | 0,156 | -1,6820 | 0.0424 | -14,395| 9,613 | 0,133 | -1,704 |
| 6-7 | 0.0272 | 37,485 | -2,102 | 0,374 | -0,6000 | 0.0246 | 37,669 | -1,010 | 0,341 | -0,594 |
| 6-8 | 0.0122 | 29,551 | -4,254 | 0,106 | -0,5560 | 0.0110 | 29,520 | -1,763 | 0,094 | -0,563 |
| 6-9 | 0.0000 | 27,673 | -7,341 | -2,527e-13| 1,5950 | 0,0000 | 27,690 | -7,249 | -1,175e-13| 1,590 |
| 6-10 | 0.0000 | 15,832 | 6,631 | -4,927e-14| 1,1990 | 0,0000 | 15,823 | 6,638 | -1,779e-14| 1,177 |
| 6-28 | 0.0172 | 18,809 | -9,886 | 0,061 | -13,0820 | 0.0155 | 18,857 | -9,348 | 0,055 | -13,102 |
| 8-28 | 0.0648 | -0,575 | 2,286 | -3,533e-14| -4,3680 | 0.0586 | -0,574 | -2,702 | 2,68e-12 | -4,371 |
| 9-10 | 0.0000 | 27,693 | 6,772 | -2,629e-13| 8,1000 | 0,0000 | 27,690 | 6,617 | -1,848e-14| 0,8060 |
| 9-11 | 0.0000 | -3,388e-13| 15,708 | -3,024e-14| 0,4650 | 0,0000 | 2,698e-13| -15,456 | -2,733e-13| 0,449 |
| 10-17 | 0.0330 | 5,3610 | 4,431 | 0,015 | 0,0370 | 0.0862 | 5,415 | 4,347 | 0,013 | 0,037 |
| 10-20 | 0.0954 | 9,0270 | 3,5720 | 0,0820 | 0,1810 | 0.0298 | 9,029 | 3,509 | 0,074 | 0,180 |
| 10-21 | 0.0354 | 15,7410| 9,8430 | 0,1120 | 0,2370 | 0.0320 | 15,701 | 9,837 | 0,101 | 0,236 |
| 10-22 | 0.0741 | 7,5870 | 4,4880 | 0,0530 | 0,1070 | 0,0670 | 7,568 | 4,499 | 0,048 | 0,106 |
| 12-13 | 0.0000 | -2,176e-13| -10,3200| -1,758e-13| 0,1330 | 0,0000 | 1,435e-13| -10,174 | 1,691e-13| 0,130 |
| 12-14 | 0.1255 | 7,8580 | 2,433 | 0,0760 | 0,1550 | 0,1135 | 7,849 | 2,474 | 0,069 | 0,155 |
| 12-15 | 0.0674 | 17,8760| 6,9330 | 0,2220 | 0,4290 | 0,0610 | 17,783 | 7,002 | 0,199 | 0,426 |
| 12-16 | 0.0963 | 7,2200 | 3,3460 | 0,0550 | 0,1130 | 0,0871 | 7,158 | 3,430 | 0,049 | 0,112 |
| 14-15 | 0.2253 | 1,5820 | 0,6790 | 0,0060 | 0,0050 | 0,2037 | 1,580 | 0,719 | 0,006 | 0,006 |
| 15-18 | 0.1093 | 6,0180 | 1,7320 | 0,0400 | 0,0800 | 0,0908 | 6,001 | 1,794 | 0,036 | 0,079 |
| 15-23 | 0.1019 | 5,0130 | 2,9460 | 0,0320 | 0,0630 | 0,0922 | 4,957 | 2,996 | 0,029 | 0,063 |
| 16-17 | 0.0840 | 3,6660 | 1,4340 | 0,0120 | 0,0270 | 0,0759 | 3,609 | 1,518 | 0,011 | 0,027 |
| 18-19 | 0.0651 | 2,7780 | 0,7520 | 0,0050 | 0,0100 | 0,0589 | 2,765 | 0,815 | 0,005 | 0,010 |
| 19-20 | 0.0346 | -6,7270| -2,6580 | 0,0170 | 0,0340 | 0,0313 | -6,740 | -2,595 | 0,016 | 0,034 |
*Power flow variation based on extreme learning machine algorithm in power system (Labeled Imen)*
### 3.4. Discussion of results
#### 3.4.1. Neural networks
This case study demonstrates the application and practical benefits of the neural network with backpropagation, as it is introduced here to solve a nonlinear problem which is the variation of the conductor resistance due to the temperature fluctuations and its influence on the power flow and losses changes; attempting to optimize the time of simulation.
The Newton raphson method generated the dataset to be used for the training patterns process. The input matrix is composed of: line resistance (R), load demand \((P_D)\), \((Q_D)\)(real and reactive power), bus voltage magnitude \((V_i)\). The neural network model developed for IEEE 30 bus test was trained to provide power system values which are power generation (real and reactive power) \((P_g)\), \((Q_g)\) and system losses \((P_l)\), \((Q_l)\)(real and reactive).
The training phase took 15s to meet the convergence (Figure 3), the performance reached the best value \(3.15e-19\) at 15s (Figure 4). The network parameters are adjusted (bias and weights) and it is ready to be utilized in real time (for the two different values of temperature: T1=30C° and T2=05C°).
The conductor’s resistance increases as its temperature increases and so losses, due to; either the current load or the weather conditions. Results shown in (Table 3) are acquired in milliseconds (neural network results). The objective of work is to minimize time by utilizing the neural network for such changing of temperature. Also to assess the generated powers and losses during the 24 hours per day and through the entire year within milliseconds.
#### 3.4.2. Extreme learning machine
Most of our effort has focused on Extreme learning machine algorithm in this paper. Even for ELM the simulation process containing two steps; the training-validation and the test phase. The dataset is randomly partitioned into two non-overlapped sets, to form the input training -validation matrix and the test one. The input matrix is composed of line resistance R, load demand (real and reactive power) \((P_D)\), \((Q_D)\), bus voltage magnitude \((V_i)\). The target matrix is containing power generation (real and reactive power) \((P_g)\), \((Q_g)\) and system losses (real and reactive) \((P_l)\), \((Q_l)\). The algorithm seems to be very fast, it achieves the highest accuracy \(10^{-27}\) (Figure 13) and fastest training time: 0.0469(S) and also testing time 0(S) (figure 12).
In order to further verify the validity of the improved built model, and the generalization; two extreme values of the resistance 30C° and 05C are introduced, results are displayed in Table 4.
According to ELM based power flow, we estimated to match results offered by the neural network with further optimization in time and more exactitude (Table 5). It is evident that the regression accuracy of ELM algorithm is superior to the NN.
Losses at:
\[
T= 25C° \text{ are } 17,599 \text{ MW}, 22,244 \text{ Mvar.}
\]
\[
T= 30C° \text{ are } 17,985 \text{ MW}, 22,407 \text{ Mvar.}
\]
\[
T= 05C° \text{ are } 16,79 \text{ MW}, 21,641 \text{ Mvar.}
\]
Mathematically losses are covered by the plant connected to the slack bus (slack number one 1), but in the real interconnected power system; the slack bus generator is not always a single equivalent machine located in a control area, losses can be interpreted as exchanges and power transactions across multi-area systems. This may impose an additional operating cost. The price increases as well as the slack bus production increases.
ELM makes an improvement in the electrical system response; the grid parameters can be prevented instantly at any moment, and the plants generation can be rescheduled to meet load demand (in addition to losses), to maintain power system security, to avoid contingencies and load shedding (voltage drop,
frequency fluctuation), to prevent system violations and overloads, to predict the effect of outages and to minimize the fuel consumption.
Additionally, ELM could likewise be extended to make use of a large scale database, (Within [05 30] C°), cause it is already trained, and also for more advanced complications in power system.
Table 6 lists the mean comparison results of ANN and ELM. Extreme Learning Machines (ELMs) are very controversial and very fast machine learning models that perform very well. ELM results are better than ANN, due to the ability of the excellent generalization performance and the learning speed which can be thousands of times faster than traditional neural networks.
| Table 6. Comparison | Neural network | Extreme learning machine |
|---------------------|----------------|--------------------------|
| Training time(s) | 15 | 0.0469 |
| Testing time(s) | / | 0.0000 |
| Training performance| 3,15e-19 | 3,666e-25 |
| Testing performance | / | 8,248e-27 |
4. CONCLUSION
Power system security requires system monitoring in real time, where the database is continuously updated. This paper investigated the impact of the temperature variation on overhead power lines. This influence on the conductor's resistance cannot be neglected. We noticed different variations on the behavior of the electrical grid, notably: the system losses, the generated powers, and the line flow.
We introduced Extreme learning algorithm to solve this problem. The basic ANN modelization (feed forward back propagation) seems to be appropriate for the application IEEE 30 Bus system; but in terms of accuracy and exactitude of results offered by ELM (the timing minimization, performance), we can say economic objectives are realized (the optimization in time, fuel consumption and the reduction of the operation cost); ELM stands to be the most gifted version of the artificial intelligence techniques.
REFERENCES
[1] W. B. Steven and M. E. El-Hawary, “Electric Power System Basics”, Series Editor, IEEE PRESS, Wiley-Interscience a John Wiley & Sons, INC, Publication. pp: 3.
[2] Michel Janssen and Christoph Lehner, *the Cambridge Companion to Einstein*, Volume 1.
[3] M. Bockarjova and G. Andersson, "Transmission Line Conductor Temperature Impact on State Estimation Accuracy," *Power Tech, IEEE Lausanne*, pp. 701-706, 2007.
[4] P. Dey et al., "Influence of conductor designs and operating temperature on the economics of overhead lines," *PROC. IEE*, vol. 118(3/4), 1971.
[5] Youmin Li et al., "An OFF Based Approach For Static Voltage Stability Evaluation Considering Heat Balance Rule Of Overhead Line," *International Conference on Electric Utility Deregulation and Restructuring and Power Technologies* November 26-2, Changsha, China, 2015.
[6] I. Alibzu et al., "Tension -Temperature Behaviour of an Overhead Conductor in Operation," *Reliability of Transmission and Distribution Networks, IETConference*, 05 Apr 2012.
[7] L. Kotin., "A proposed algorithm for an overhead transmission line conductor temperature rise calculation," *International Transaction on Electrical Energy Systems Int*, vol. 24, pp. 578-59, 2014.
[8] A.Selakov et al., "Hybrid PSO-SVM method for short-term load forecasting during periods with significant temperature variations in city of Burbank", *Applied Soft Computing*, vol.16, pp. 80-88, 2014.
[9] M. M. Saied., "Assessing the dynamic rating of overhead transmission lines," *European Transaction on Electrical Power*, vol. 17, pp. 526-536, 2007.
[10] M. Schlapfer and P. Mancarella, "Probabilistic Modeling and Simulation of Transmission Line Temperatures under Fluctuating Power Flows," *IEEE Transactions on Power Delivery*, vol 26, no 4, pp. 2235 -2243.
[11] J. Tang et al., "Extreme Learning Machine for Multilayer Perceptron," *IEEE Transactions on Neural Networks and Learning Systems*, vol. 2 no. 4, pp. 809-821, 2016.
[12] Ö. F. Ertugrul., "Forecasting electricity load by a novel recurrent extreme learning machines approach," *International Journal of Electrical Power & Energy Systems*, vol. 78, pp. 429-435, 2016.
[13] Xu, Y et al., "Extreme learning machine-based predictor for real-time frequency stability assessment of electric power systems," *Neural Computing and Applications*, vol. 22, no 3-4, pp. 501-508, 2012.
[14] Wu, B et al., "Fault diagnosis for wind turbine based on improved extreme learning machine," *Journal of Shanghai Jiaotong University (Science)*, vol. 22, no. 4, pp. 466-473, 2017.
[15] A. El Bakri, et al., "Extreme learning machine-based non-linear observer for fault detection and isolation of wind turbine," *Australian Journal of Electrical and Electronics Engineering*, pp. 1–9, 2019.
[16] S. Vidiya and V. Kamaraj, "Particle swarm optimized extreme learning machine for feature classification in power quality data mining," *Automatika*, vol. 58, no. 4, pp. 487-494, 2017.
[17] V. L. Paukar and M. J. Rider, "Artificial neural networks for solving the power flow problem in electric power systems," *Electric Power Systems Research*, vol. 62, pp. 139-144, 2002.
[18] D. Abdellah and L. Djamel, "Power System Economic Dispatch Using Traditional and Neural Networks Programs," *International Journal of Scientific & Engineering Research*, vol. 3, no. 4, 2012.
[19] E. E. Elattar., "Prediction of wind power based on evolutionary optimized local general regression neural network," *IET Gener. Transm. Distrib.*, vol. 8, no 5, pp. 916-923, 2014.
[20] W. Yixing *et al.*, "Deep Learning Neural Network for Power System Fault Diagnosis," *IEEE Proceedings of the 35th Chinese Control Conference July 27-29, Chengdu, China*, 2016.
[21] R. Aggarwal and Yonghua Song, "Artificial neural networks in power systems. II. Types of artificial neural networks," *Power Engineering Journal*, vol. 12, no 1, pp. 41-47, 1998.
[22] M.Safari and M.Sarvi, "Optimal load sharing strategy for the wind, a diesel battery hybrid power system based on imperialist competitive neural network algorithm," *IET Renew. Power Gener.*, vol. 8, no. 8, pp. 937-946, 2014.
[23] L.G.Chen *et al.*, "Group-based chaos genetic algorithm and a nonlinear ensemble of neural networks for short - term load forecasting," *IET Gener. Transm. Distrib. & the Institution of Engineering and Technology*, pp. 1-8, 2016.
[24] N. Geetha, and P. Renuga, "Multilayer feed-forward neural network approach for optimal dispatch of UPFC embedded pool-bilateral electricity market, Ranking," *ETEP*, vol. 12, no 2, 2002.
[25] L.A. Laboissiere, et al., "Maximum and minimum stock price forecasting of Brazilian power distribution companies based on artificial neural networks," *Appl. Soft Comput. J. vol.* 35, pp. 66-74, 2015.
[26] S.N. Singh and K.N. Srivastava. "Evolving Artificial Neural Network Models for voltage Contingency," *International Transaction on Electrical Energy Systems*, vol. 12, pp. 127-133, 2014.
[27] H. Hadj *et al.*, "Feedforward neural network-based transient stability analysis of electric power systems," *European Transaction on Electrical Power*, vol. 16, pp. 577-590, 2006.
[28] G. Yang, *et al.*, "TOSELM: Timeliness Online Sequential Extreme Learning Machine," *Neurocomputing*, vol. 128, no 27, pp. 119-127, 2014.
[29] X. Zhixiang *et al.*, "modified extreme learning machine with sigmoidal activation functions," *Neural Comput & Applic*, vol.22, no. 3-4, pp. 541-550, 2013.
[30] Y.Xu *et al.*, "Real-time transient stability assessment model using extreme learning machine," *IET Generation, Transmission & Distribution*, vol. 5, no 3, pp. 314-322, 2011.
[31] K. Yakup *et al.*, "LU triangulization extreme learning machine in EEG cognitive task classification," *Neural Comput & Applic*, pp. 1-10, 2017.
[32] M.Tawfek *et al.*, "An Advanced Approach for Optimal Wind Power Generation Prediction Intervals by Using Self-Adaptive Evolutionary Extreme Learning Machine."
[33] A.Yuan *et al.*, "Generalization ability of extreme learning machine with uniformly ergodic Markov chains Peipei," *Neurocomputing*, vol. 167, pp. 528-534, 2015.
[34] G. B. Huang and C. K. Siew, "Extreme learning machine with randomly assigned RBF kernels," *Int. J. Inf. Technol*, vol. 11, no. 1, pp. 16-24, 2005.
[35] B. P. Jiangtao *et al.*, "Extreme learning machine for ranking: Generalization analysis and applications Hong Chena," *Neural Networks*, vol. 53, pp. 119-126, 2014.
[36] G. Yang, *et al.*, "TOSELM: Timeliness Online Sequential Extreme Learning Machine," *Neurocomputing*, vol. 128, no 27, pp. 119-127, 2014.
[37] I. Bahiuddin, et al," Comparing the linear and logarithm normalized extreme learning machine in flow curve modeling of magnetorheological fluid," *Indonesian Journal of Electrical Engineering and Computer Science*, vol. 13, no 3, pp. 1065-1072, 2019.
[38] Hadi Saadat., *Power System Analysis*, Mc Graw –Hill Series in Electrical and computer engineering 1999.
**BIOGRAPHIES OF AUTHORS**
**Imen LABED**: was born in (1980), she received the M.S. degrees from University of Constantine 1. She is currently a Ph.D. student at a laboratory of Electric Engineering, Department Of Electrical Engineering Constantine.University of Constantine 1(Algeria). Her research activities are currently interesting on the economic dispatch, forecasting, electrical load flow, neural networks, Extreme learning machine, and the renewable energy impact on the electrical power system. E-mail: firstname.lastname@example.org
**LABED Djamel** is a Head of department of electrical engineering of the University Mentouri Constantine, Director of electrical engineering laboratory Constantine LGEC, Director of Research. He received an ingenereate in Electrical from University of Annaba, a Master in Electrical Research Network from the Ecole Polytechnique of Montreal, Canada, and a Ph.D. in Electrical Engineering from the University Mentouri Constantine. His research activities focus on Dispersed Generation; Analysis of different types of problem quality of electric power caused by renewable energy sources; Various techniques for optimizing the optimal power flow; Stability of groups; Power flow method for DC networks integrated into AC. E-mail: email@example.com |
Unrecognized Potential:
The Relationship Between the Royal Navy and Vickers Armstrong, 1930-39
by
Yvonne Leslie Youngberg
A THESIS
SUBMITTED TO THE FACULTY OF GRADUATE STUDIES
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF ARTS
DEPARTMENT OF HISTORY
CALGARY, ALBERTA
AUGUST, 1994
©Yvonne Leslie Youngberg 1994
THE AUTHOR HAS GRANTED AN IRREVOCABLE NON-EXCLUSIVE LICENCE ALLOWING THE NATIONAL LIBRARY OF CANADA TO REPRODUCE, LOAN, DISTRIBUTE OR SELL COPIES OF HIS/HER THESIS BY ANY MEANS AND IN ANY FORM OR FORMAT, MAKING THIS THESIS AVAILABLE TO INTERESTED PERSONS.
L'AUTEUR A ACCORDE UNE LICENCE IRREVOCABLE ET NON EXCLUSIVE PERMETTANT A LA BIBLIOTHEQUE NATIONALE DU CANADA DE REPRODUIRE, PRETER, DISTRIBUER OU VENDRE DES COPIES DE SA THESE DE QUELQUE MANIERE ET SOUS QUELQUE FORME QUE CE SOIT POUR METTRE DES EXEMPLAIRES DE CETTE THESE A LA DISPOSITION DES PERSONNE INTERESSEES.
THE AUTHOR RETAINS OWNERSHIP OF THE COPYRIGHT IN HIS/HER THESIS. NEITHER THE THESIS NOR SUBSTANTIAL EXTRACTS FROM IT MAY BE PRINTED OR OTHERWISE REPRODUCED WITHOUT HIS/HER PERMISSION.
L'AUTEUR CONSERVE LA PROPRIETE DU DROIT D'AUTEUR QUI PROTEGE SA THESE. NI LA THESE NI DES EXTRAITS SUBSTANTIELS DE CELLE-CI NE DOIVENT ETRE IMPRIMES OU AUTREMENT REPRODUITS SANS SON AUTORISATION.
ISBN 0-315-99530-0
Name: Yvonne Leslie Youngberg
Dissertation Abstracts International is arranged by broad, general subject categories. Please select the one subject which most nearly describes the content of your dissertation. Enter the corresponding four-digit code in the spaces provided.
**History, European**
| Subject Categories | Code |
|--------------------|------|
| **COMMUNICATIONS AND THE ARTS** | |
| Architecture | 0729 |
| Art History | 0374 |
| Cinema | 0700 |
| Dance | 0378 |
| Fine Arts | 0357 |
| Information Science | 0223 |
| Journalism | 0391 |
| Library Science | 0349 |
| Mass Communications | 0708 |
| Music | 0413 |
| Speech Communication | 5457 |
| Theater | 0465 |
| **EDUCATION** | |
| General | 0515 |
| Administration | 0514 |
| Adult Continuing | 0516 |
| Agricultural | 0517 |
| Art | 0518 |
| Bilingual and Multicultural | 0292 |
| Business | 0688 |
| Community College | 0707 |
| Curriculum and Instruction | 0706 |
| Early Childhood | 0518 |
| Elementary | 0516 |
| Finance | 0277 |
| Guidance and Counseling | 0519 |
| Health | 0503 |
| Higher | 0745 |
| History of | 0520 |
| Home Economics | 0518 |
| Industrial | 0521 |
| Language and Literature | 0747 |
| Mathematics | 0280 |
| Music | 0522 |
| Philosophy of | 0978 |
| Physical | 0523 |
| Psychology | 0525 |
| Reading | 0533 |
| Religious | 0526 |
| Secondary | 0714 |
| Social Sciences | 0534 |
| Sociology of | 0540 |
| Special | 0529 |
| Staff Training | 0530 |
| Technology | 0710 |
| Tests and Measurements | 0288 |
| Vocational | 0747 |
| **LANGUAGE, LITERATURE AND LINGUISTICS** | |
| General | 0679 |
| Ancient | 0289 |
| Linguistics | 0291 |
| Modern | 0291 |
| General | 0401 |
| Classical | 0594 |
| Comparative | 0292 |
| Modern | 0297 |
| Old | 0516 |
| American | 0591 |
| Asian | 0305 |
| Canadian (English)| 0352 |
| Canadian (French) | 0355 |
| English | 0393 |
| German | 0311 |
| Latin American | 0312 |
| Middle Eastern | 0313 |
| Romance | 0313 |
| Slavic and East European | 0314 |
| **PHILOSOPHY, RELIGION AND THEOLOGY** | |
| General | 0422 |
| General | 0318 |
| Biblical Studies | 0321 |
| Clergy | 0322 |
| History of | 0320 |
| Philosophy of | 0322 |
| Theology | 0469 |
| **SOCIAL SCIENCES** | |
| Anthropology | 0323 |
| Archaeology | 0324 |
| Cultural | 0325 |
| Physical | 0327 |
| Business Administration | 0310 |
| General | 0454 |
| Accounting | 0272 |
| Banking | 0454 |
| Management | 0338 |
| Marketing | 0338 |
| Consumer Studies | 0385 |
| Economics | 050 |
| General | 050 |
| Agricultural | 0703 |
| Commerce-Business | 0505 |
| Finance | 0505 |
| Industry | 0509 |
| Labor | 0510 |
| Theory | 0511 |
| Folklore | 0338 |
| Geography | 0366 |
| Geology | 0351 |
| History | 0578 |
| General | 0485 |
| Agricultural | 0749 |
| Analytical | 0487 |
| Biochemistry | 0487 |
| Inorganic | 0488 |
| Nuclear | 0488 |
| Organic | 0490 |
| Pharmaceutical | 0490 |
| Physical | 0494 |
| Polymer | 0495 |
| Radiation | 0495 |
| Mathematical | 0405 |
| Physics | 0405 |
| General | 0405 |
| Acoustics | 0986 |
| Astronomy and | 0758 |
| Astrophysics | 0606 |
| Atmospheric Science | 0608 |
| Atomic | 0608 |
| Electronics and Electricity | 0607 |
| Elementary Particles and High Energy | 0758 |
| Fluid and Plasma | 0759 |
| Molecular | 0609 |
| Nuclear | 0710 |
| Experimental | 0621 |
| Optics | 0752 |
| Radiation | 0756 |
| Solid State | 0759 |
| Statistics | 0463 |
| **Applied Sciences** | |
| Applied Mechanics | 0346 |
| Computer Science | 0984 |
**THE SCIENCES AND ENGINEERING**
| Biological Sciences | Code |
|---------------------|------|
| Agriculture | 0473 |
| General | 0285 |
| Animal Culture and Nutrition | 0475 |
| Animal Pathology | 0476 |
| Food Science and Technology | 0359 |
| Forestry and Wildlife | 0478 |
| Plant Culture | 0479 |
| Plant Pathology | 0480 |
| Plant Physiology | 0817 |
| Range Management | 0777 |
| Wood Technology | 0746 |
| Biology | 0306 |
| General | 0287 |
| Biostatistics | 0308 |
| Botany | 0309 |
| Cell | 0329 |
| Ecology | 0329 |
| Embryology | 0333 |
| Genetics | 0349 |
| Herpetology | 0353 |
| Limnology | 0773 |
| Microbiology | 0410 |
| Molecular | 0307 |
| Neuroscience | 0317 |
| Oceanography | 0416 |
| Physiology | 0433 |
| Radiation | 0821 |
| Veterinary Science | 0778 |
| Zoology | 0472 |
| Biophysics | 0786 |
| Medical | 0760 |
| Earth Sciences | 0425 |
| Biogeochemistry | 0976 |
| Geochemistry | 0976 |
| Geodesy | 0370 |
| Geology | 0372 |
| Geophysics | 0373 |
| Hydrology | 0388 |
| Mineralogy | 0411 |
| Paleontology | 0415 |
| Paleozoology | 0426 |
| Palynology | 0418 |
| Parasitology | 0427 |
| Polynology | 0427 |
| Physical Geography | 0368 |
| Physical Oceanography | 0415 |
| Health and Environmental Sciences | 0768 |
| Health Sciences | 0566 |
| Audiology | 0300 |
| Chemotherapy | 0972 |
| Dermatology | 0325 |
| Education | 0350 |
| Hospital Management | 0769 |
| Human Biology | 0300 |
| Immunology | 0982 |
| Medical and Surgery | 0564 |
| Medical Physics | 0564 |
| Nursing | 0569 |
| Nutrition | 0416 |
| Obstetrics and Gynecology | 0380 |
| Occupational Health and Therapy | 0354 |
| Ophthalmology | 0381 |
| Pathology | 0978 |
| Pharmacology | 0419 |
| Pharmacy | 0572 |
| Physiology | 0382 |
| Public Health | 0564 |
| Radiology | 0574 |
| Recreation | 0575 |
| Speech Pathology | 0460 |
| Toxicology | 0383 |
| Home Economics | 0386 |
| Engineering | 0577 |
| General | 0335 |
| Agricultural | 0339 |
| Automotive | 0540 |
| Biomedical | 0541 |
| Chemical | 0542 |
| Civil | 0543 |
| Electrical and Electronic | 0544 |
| Heat and Thermodynamics | 0348 |
| Hydraulic | 0545 |
| Industrial | 0546 |
| Marine | 0547 |
| Materials Science | 0792 |
| Mechanical | 0548 |
| Metallurgy | 0743 |
| Mining | 0531 |
| Nuclear | 0552 |
| Packaging | 0549 |
| Petroleum | 0765 |
| Sanitary and Municipal | 0554 |
| System Science | 0790 |
| Geotechnical | 0492 |
| Operations Research | 0796 |
| Plastics Technology | 0795 |
| Transportation Technology | 0794 |
| Psychology | 0621 |
| Behavioral | 0384 |
| Clinical | 0522 |
| Developmental | 0620 |
| Experimental | 0623 |
| Intelligence | 0624 |
| Personality | 0625 |
| Physiological | 0989 |
| Psychiatric | 0549 |
| Psychometrics | 0532 |
| Social | 0451 |
Nom
Dissertation Abstracts International est organisé en catégories de sujets. Veuillez s.v.p. choisir le sujet qui décrit le mieux votre thèse et inscrivez le code numérique approprié dans l'espace réservé ci-dessous.
Sujet
CODE DE SUJET
Catégories par sujets
HUMANITÉS ET SCIENCES SOCIALES
COMMUNICATIONS ET LES ARTS
Architecture ........................................... 0739
Beaux-arts ............................................. 0357
Bibliothéconomie ....................................... 0359
Cinéma .................................................. 0428
Communication verbale .................................. 0459
Communications ......................................... 0398
Danse .................................................... 0376
Histoire de l'art ....................................... 0377
Journalisme .............................................. 0413
Musique .................................................. 0371
Sciences de l'information ............................... 0723
Théâtre .................................................. 0465
EDUCATION
Cybernétique ............................................ 515
Administration .......................................... 0514
Art ....................................................... 0479
Collèges communautaires ................................. 0275
Commerce ................................................ 0688
Économie domestique ..................................... 0300
Éducation ................................................ 0516
Éducation précoce ....................................... 0518
Éducation sanitaire ...................................... 0300
Enseignement agricole ................................... 0517
Enseignement bilingue et multiculturel .................. 0382
Enseignement industriel ................................ 0521
Enseignement primaire .................................. 0524
Enseignement professionnel .............................. 0527
Enseignement religieux .................................. 0527
Enseignement secondaire ................................ 0529
Enseignement spécial .................................... 0529
Enseignement supérieur .................................. 0745
Evaluation ............................................... 0312
Finance .................................................. 0277
Formation des enseignants ............................... 0530
Histoire de l'éducation ................................ 0279
Langues et littérature .................................. 0279
LANGUE, LITTÉRATURE ET LINGUISTIQUE
Littérature .............................................. 0376
Linguistique ............................................. 0290
Médievale ................................................ 0297
Modernes ................................................ 0297
Anglaise .................................................. 0591
Américaine .............................................. 0591
Anglaise .................................................. 0592
Allemande ................................................ 0305
Canadienne (Anglaise) ................................... 0352
Canadienne (Française) .................................. 0352
Allemande ................................................ 0311
Germancie .............................................. 0311
Latino-américaine ....................................... 0312
Moyen orientale ......................................... 0312
Romane ................................................... 0313
Slave et est-européenne ................................. 0314
SCIENCES ET INGÉNIERIE
SCIENCES BIOLOGIQUES
Agriculture .............................................. 0473
Alimentation et technologie alimentaire ................. 0359
Culture ................................................... 0379
Élevage et clivestration ................................ 0475
Exploitation des pêcheries ............................... 0777
Pathologie animale ...................................... 0476
Pathologie végétale ..................................... 0480
Physiologie végétale .................................... 0817
Syndicalisme ............................................. 0778
Technologie du bois ..................................... 0746
Biologie
Généralités .............................................. 0306
Anatomie ................................................ 0287
Biologie (Statistiques) .................................. 0307
Biologie moléculaire .................................... 0307
Botanique ................................................ 0309
Cellule ................................................... 0309
Ecologie .................................................. 0329
Entomologie .............................................. 0333
Génétique ................................................ 0359
Limnologie .............................................. 0793
Microbiologie ............................................ 0410
Neurologie .............................................. 0417
Océanographie ........................................... 0416
Pharmacologie ........................................... 0333
Radiation ................................................ 0821
Science vétérinaire ..................................... 0778
Zoologie .................................................. 0472
Biophysique .............................................. 0368
Généralités .............................................. 0786
Medicale .................................................. 0760
SCIENCES DE LA TERRE
Biogéochimie ............................................. 0425
Géochimie ................................................ 0956
Géodésie .................................................. 0370
Géographie physique ..................................... 0370
Lecture ................................................... 0525
Mathématiques ............................................ 0380
Musique ................................................... 0522
Philosophie .............................................. 0422
Religion ................................................... 0422
Généralités .............................................. 0318
Crééologie .............................................. 0319
Éthique catholique ...................................... 0319
Histoire des religions ................................... 0320
Philosophie de la religion ............................... 0322
Théologie ................................................ 0469
SCIENCES SOCIALES
Anthropologie ............................................ 0324
Archéologie .............................................. 0324
Culturale ................................................ 0325
Physique .................................................. 0325
Droit ...................................................... 0398
Économie ................................................ 0398
Généralités .............................................. 0501
Commerce-Afrique ........................................ 0505
Économie sociale ........................................ 0505
Économie du travail ..................................... 0510
Finances .................................................. 0508
Histoire .................................................. 0511
Théorie ................................................... 0511
Études américaines ...................................... 0345
Études africaines ....................................... 0345
Études féministes ....................................... 0453
Folklore .................................................. 0345
Géographie .............................................. 0366
Gérontologie ............................................. 0351
Gestion des édifices ..................................... 0310
Généralités .............................................. 0310
Administration ........................................... 0270
Banques ................................................... 0270
Comptabilité ............................................. 0272
Marketing .................................................. 0338
Histoire .................................................. 0578
Histoire générale ....................................... 0578
SCIENCES PHYSIQUES
Bionomédicale ............................................ 0541
Chaleur et froid ........................................ 0348
Condensation ............................................. 0348
(Emballeage) ............................................. 0549
Gène aérospatial ........................................ 0538
Gène physique ........................................... 0542
Gène civil ................................................ 0543
Gène électronique et mécanique .......................... 0544
Gène industriel .......................................... 0546
Gène mécanique .......................................... 0546
Gène physique ........................................... 0542
Ingenierie des systèmes .................................. 0790
Mécanique navale ........................................ 0547
Métallurgie .............................................. 0743
Science des matériaux ................................... 0794
Technologie industrielle ................................. 0745
Technique minière ....................................... 0351
Techniques sanitaires et environnementales .............. 0554
Technologie hydraulique ................................. 0545
Mécanique appliquée ..................................... 0346
Généralités .............................................. 0349
Matières plastiques ..................................... 0795
Technologie ............................................. 0795
Recherche scientifique .................................. 0795
Textiles et tissus (Technologie) ........................ 0794
PSYCHOLOGIE
Généralités .............................................. 0521
Personnalité .............................................. 0525
Psychologie .............................................. 0349
Psychologie clinique ..................................... 0521
Psychologie du comportement ............................. 0521
Psychologie du développement ........................... 0620
Psychologie expérimentale ............................... 0622
Psychologie infantile ................................... 0524
Psychologie physiologique ............................... 0589
Psychologie sociale ..................................... 0451
Psychiatrie .............................................. 0532
THE UNIVERSITY OF CALGARY
FACULTY OF GRADUATE STUDIES
The undersigned certify that they have read, and recommend to the Faculty of Graduate Studies for acceptance, a thesis entitled "Unrecognized Potential: The Relationship Between the Royal Navy and Vickers Armstrong, 1930-39" submitted by Yvonne Leslie Youngberg in partial fulfillment of the requirements for the degree of Master of Arts.
[Signature]
Supervisor, Dr. J.R. Ferris, Department of History
[Signature]
Dr. H.H. Herwig, Department of History
[Signature]
Dr. D.M. Peers, Department of History
[Signature]
Dr. Terry Terriff, Strategic Studies Program
12 Aug 94
Date
Abstract
With the huge increase in naval armaments orders between 1935 and 1939, the relationship between Vickers Armstrong and the Royal Navy should have been without conflict. However, the events of 1929-1935 ensured that this was not the case. After a decade of stability, the dual blow of the Depression and the London Treaty in 1930 made the Admiralty’s formerly stable policy chaotic. The capital ship replacement programme scheduled to begin in 1931 was cancelled, and the Admiralty’s efforts to preserve capacity were thwarted. By 1935 British shipbuilding capacity was cut in half.
The problems of 1929-1935 were exacerbated by the Admiralty’s actions in 1935-1939. Vickers’ experiences over the period illustrate that the Admiralty failed to use existing capacity efficiently. Its attempts to force the firm to increase gun mounting production only caused friction and made the firm less willing to turn production completely over to rearmament.
Acknowledgements
I would like to thank all the people who made writing this thesis a much less painful process than it might have been. The staff at the MacKimmie Library were most helpful, and special thanks must go to the Interlibrary Loan office. Lorraine Baker and Magdalena Gisler were especially kind in dealing with a very frazzled graduate student.
My supervisor, Dr. John Ferris, deserves special thanks for his help and the late nights he spent editing my work. Without his assistance this thesis would not have been possible. I would also like to thank Dr. H. Herwig, Dr. D. Peers, and Dr. T. Terriff for serving on my defence committee.
Many friends lent their support during the months I spent working on this. I would like to thank Leslie Schultz, Shannon Chace-Hall, Corinne Mahaffey, James Sterrett and Sylvia Price for keeping me sane and focused.
Finally, I must thank my family, whose guidance, support and love have been constant. Special thanks go to my mother, Ann Youngberg, who has always been there for me, and who encouraged me to follow my dreams even when they took me half a country away.
To my grandmother, Annabell (Anna) Stuart, 1922-1987.
# Table of Contents
Approval Page .......................................................... ii
Abstract ................................................................ iii
Acknowledgements ......................................................... iv
Dedication ................................................................ v
Table of Contents ......................................................... vi
List of Tables ................................................................ vii
Introduction ................................................................ 1
Notes ........................................................................... 7
Chapter One: The Development of Naval Policy ............... 8
Post-War Industrial Attrition and the London Naval Conference. 10
The Effects of the London Naval Agreement .................... 24
Notes ........................................................................ 41
Chapter Two: Rearmament ............................................. 46
The Defence Requirements Committee ............................ 47
The Defence Policy and Requirements Committee (1935) .... 56
Rations, Feast, and Famine ........................................ 63
New Standards of Naval Power .................................... 66
Notes ........................................................................ 77
Chapter Three: Vickers Armstrong and Rearmament .......... 81
Capacity to Spare ..................................................... 83
Capital Plant Extensions ........................................... 93
Business As Usual .................................................... 98
Notes ........................................................................ 108
Conclusion ................................................................ 112
Bibliography ................................................................ 118
List of Tables
Table I: Proposed British Cruiser Tonnage ........................................... 18
Table II: Proposed American Cruiser Tonnage ....................................... 19
Table III: Age of British Battleships by 31 December 1936 .................. 31
Table IV: Comparison of Effective Estimates and Vote 8 (Shipbuilding, Repairs & Maintenance), 1930 to 1935 ............................................. 36
Table V: Vote 8 by Section, 1930 to 1935 ............................................ 37
Table VI: Complete Ships (as of 31 March 1939) .................................. 60
Table VII: Forecasted Annual Estimates, 1937 to 1946, DRC Fleet vs. Fleet A . 71
Table VIII: Naval Estimates and Appropriations, 1937 to 1940 .............. 73
Table IX: Average Capacity Used, Barrow Works, 1933 to 1935 .............. 85
Table X: Average Capacity Used, Elswick Works, 1933 to 1935 .............. 85
Table XI: British & Foreign Naval Armaments Sales, 1933 to 1935 ............ 86
Table XII: Summary of Naval Armaments and Shipbuilding Sales, 1936 to 1939 . 88
Introduction
This thesis has three purposes: to explain why British naval rearmament during the 1930s was so difficult, to explore the forces which shaped rearmament policy, and to illustrate these issues by examining the relationship between the Royal Navy and the premier armaments firm of Great Britain, Vickers Armstrong. This topic falls on the margins between several branches of history – policy, businesses, and technology. Perhaps for this reason, the area has rarely been treated as a single and coherent topic in its own right. In fact, however, the history of naval rearmament requires the examination of two distinct levels of decision making: one being the Cabinet, the Admiralty, and other departments in Whitehall, the other including the Admiralty, armaments firms, and their labour force. Each of the levels has had its historians, but no one has examined them in detail. The connection between government policy during rearmament and its execution by individual armaments firms is not an area well documented by historians. Rearmament policy has been dealt with extensively, while some work has been done on the procurement policies of the Admiralty. The connection between the Admiralty and specific firms is not well documented, yet it is vital to understanding rearmament. Britain had the most extensive naval armaments capacity in the world and the Admiralty received huge estimates for construction throughout the 1920s and in late the 1930s, yet the Navy was unprepared for war in 1939. The performance of the Admiralty and armaments firms in naval construction between 1933 and 1939 was not good, yet precisely the same organization did far
better at the task during World War II. To understand why this occurred it is necessary to study naval construction at two levels simultaneously; the naval policy in Whitehall and the firms’ perspective - the experience of those who tried to meet the Navy’s needs.
Some aspects of this topic have been well documented. This is particularly true of the development of rearmament policy. Stephen Roskill’s extensive work on interwar naval policy remains the best overall account of naval developments during the rearmament period. He argues that the desire for disarmament and faith in the League of Nations among both the elites and public slowed Britain’s response to the aggressive posture of Germany and Japan. The faith in a diplomatic solution explains why Britain was so determined to pursue armaments limitation both in 1930, with the London Naval Treaty, and even as late as 1936 with the signing of the second London Naval Treaty. In particular, Roskill argues, it was the decision in 1930 to hold another disarmament conference in 1935 that slowed Britain’s move towards rearmament, though other factors, especially the Treasury’s inability to understand the Navy’s needs, were also significant.\(^1\) Robert Shay, in *British Rearmament in the Thirties*, also blames the events of 1929 – 1935 for Britain’s poor response to the need to rearm, but for another reason. Shay focuses on the actions of decision makers in government. He is especially critical of the National Government’s role, noting that circumstances prevented any real opposition to its policies and that what little opposition occurred was soothed by token measures such as the appointment of Thomas Inskip or the dismissal of Viscount Swinton.\(^2\) He is also critical of the
services’ lack of unity, which gave the Treasury the opportunity to play the services off against each other and thereby to retain some control over expenditure.\(^3\)
Conversely, G.C. Peden, in *British Rearmament and the Treasury, 1932-1939*, argues that while the Treasury must share some of the blame for the late start of rearmament, it played a valuable role during rearmament itself. Since the services regarded rearmament as a chance to feast after years of fiscal restraint, someone had to try to keep them under control. In doing so, the Treasury managed to impose a set of priorities on rearmament that took into account the relationship between expenditure and economic capacity.\(^4\) G.A.H. Gordon, like Roskill, Shay, and Peden, also attributes the difficulties of rearmament to the actions or lack of them in the early 1930’s. His *British Seapower and Procurement between the Wars*, however, is most critical of the Treasury and its near fanatic belief in the deterrent value of a strong economy. While Gordon feels that its concern was warranted by the experience of World War I, he nevertheless argues that the Treasury’s response was too dogmatic. A more balanced concept involving economic and military considerations was required.\(^5\) These representative works illustrate the view of the majority of historians who argue that, for various reasons, the events of 1929-1935 are to blame for Britain’s delays in rearming; that rearmament in practice began to take off around 1935, only to proceed slowly for various reasons, mostly stemming from financial limits and the inability of British industry to meet the orders for new construction.
Most of the arguments regarding the relationship between rearmament and armaments capacity, however, are brief and do not include much primary research.
The specialist work which has been done suffers from two faults: it either does not go into enough detail, or else it only deals with the issue from the perspective of the government. The three volumes in the series on the official economic history of World War II which relate to the topic suffer from both of these problems. *Factories and Plant*, by William Hornby, is essentially a compilation of facts and figures, tonnages produced, and the like. The problems that hampered rearmament are not analyzed, nor are the experiences of specific firms considered.\(^6\) *The Administration of War Production* by J.D. Scott and Richard Hughes devotes only eighteen pages to pre-war naval production and concentrates on the organization of the Admiralty while ignoring how the Admiralty related to armaments firms.\(^7\) Finally, *British War Production* by M.M. Postan, still the most thorough examination of the economics of rearmament and wartime construction, goes into greater detail about the capacity available during the inter-war years, but only from the government's perspective.\(^8\)
The arguments about rearmament put forward in this thesis reinforce most of the conventional views in the field. As many have argued, I believe that the problems of rearmament were largely due to the events of 1929-35. The dual blow of the London Treaty of 1930 and the Depression crippled naval policy and the shipbuilding industry. Had either one of these two concurrent events not occurred, the problems would still have existed but in a diminished form. The general decline of the armed forces and the late start on the rearmament programme would still have created problems, but Britain would have been better prepared in 1939. Again, the scale and the heat of the interdepartmental rivalries of the era cannot be ignored. In particular,
throughout the interwar years, the Treasury and the Admiralty were locked in a fierce struggle to dominate strategic policy; each struck the hardest blows it could. Thus, the Treasury often tried to drive the Navy to the level of being only the third largest in the world. For its part, the Admiralty always tried to get everything it possibly could from the government while it was in a position to do so. It caused dramatic reversals in naval policy on several occasions. Chapters One and Two focus on the course of these issues during the interwar years, with particular emphasis on some issues which have been overlooked - the actual fate of the Admiralty’s procurement of warships and armaments.
This thesis, differs from existing scholarship in two main ways, which will be found in Chapter Three. First, it links naval policy to its practical implementation in the form of procurement and the success and efficiency of this process. It shows that the conventional view, with its focus on the formulation of policy in Whitehall rather than its execution on the slips, is adequate. If one only considers the political sphere, for example, the Admiralty was increasingly successful from 1934-35 onwards and able to get its way on matters of policy. Of equal importance, however, is the fact that the Admiralty failed to implement this policy for reasons which can only be understood by its relations with industry, from the firm’s perspective. For that purpose, this thesis will focus on one firm in particular, Vickers Armstrong. Secondly, this thesis attempts to fill a gap in the history of Vickers and, more broadly, the naval arms industry. The only other scholarly work dealing with the firm in this period, *Vickers: A History* by J.D. Scott, deals specifically with the firm and
its corporate history. It is an excellent account of the firm’s business relations, including its work for the Admiralty, but it does not place the firm in a broader context.\(^9\) Until that is done, our understanding of British naval policy between the wars will be incomplete.
Notes
1. Stephen Roskill, *Naval Policy Between the Wars* v. I (London: Collins, 1968.), p. 67.
2. Robert Paul Shay, *British Rearmament in the Thirties* (Princeton: Princeton U.P., 1977.), p. 281.
3. Shay, p. 282.
4. G.C. Peden, *British Rearmament and the Treasury, 1932 - 1939* (Edinburgh: Scottish Academic Press, 1979), pp. 179-81.
5. G.A.H. Gordon, *British Seapower and Procurement during the Wars* (Maryland: Naval Institute Press, 1988.), p. 287.
6. William Hornby, *Factories and Plant* (London: HMSO, 1959), *passim*.
7. J.D. Scott and Richard Hughes, *The Administration of War Production* (London: HMSO, 1955.), *passim*.
8. M.M. Postan, *British War Production* (London: HMSO, 1952), *passim*.
9. J.D. Scott, *Vickers: A History* (London: Weidenfeld & Nicolson, 1962.), *passim*.
Chapter One: The Development of Naval Policy
In the decade following World War I, Britain was in a secure position at sea. While the terms of the Washington Naval Treaty of 1922 forced the Admiralty to scrap twenty-four completed capital ships and to abandon the plans to build four new super-Hoods, the Treaty actually increased the pace of naval construction in Britain. Modifications to the ships retained under the terms of the treaty such as the design and construction of 6,000 ton cruisers as well as 8-inch gun, 10,000 ton cruisers, and of two 35,000 ton battleships (the Nelson and the Rodney) provided more than enough new work to keep the British shipbuilding industry occupied between 1922 and 1926. The ten-year holiday in the construction of new and replacement capital ships (save the Nelson and the Rodney) threatened the survival of specialized capacity. However, the Admiralty chose to preserve that capacity by providing subsidies to certain firms and maintained capacity in general through a slow and steady replacement program of the vessels not restricted under the treaty. All of this hinged on the assumption that the ten-year naval holiday would end in 1931 and that a sustained and major rearmament program would start in 1931-32. The Admiralty pursued a policy of large and continual construction with the aim of maintaining the largest fleet in the world, of keeping it modern, of retaining the extra capacity necessary to construct a huge navy for emergencies, and to sustain an increased level of construction in the 1930s. This policy was successful until the late 1920s. The Admiralty preserved both the strength of the Royal Navy and the shipbuilding capacity required to maintain it.
This position declined between 1929 and 1935 as a result of two events. First, the London Naval Treaty wrecked the Admiralty’s plans for ship replacement and postponed the start of the planned reconstruction of capital ships. This in turn devastated the Admiralty’s procurement policy - for its arrangements with the naval armament industry had assumed that a major reconstruction programme would begin in 1931-32. The Admiralty’s efforts to preserve shipbuilding capacity through the lean years of the 1920s was for naught; whereas capacity was adequate in 1930, it was cut in half by 1935. These direct consequences of the London Naval Treaty crippled Britain’s ability to rearm its navy later in the 1930s. Secondly, the global Depression which began in 1929 placed the shipbuilding industry in a double bind. Precisely as the scale of naval shipbuilding orders declined, the Depression caused a significant decline in orders for merchant tonnage. Whereas in 1929 over 1.5 million tons of new merchants vessels were built, only 133,000 tons of merchant ships were launched in 1933.\(^2\) The simultaneous collapse of naval and mercantile orders forced many firms to cease armaments production or to go out of business altogether. The result of the London Naval Treaty and the financial restraint of the period 1929 to 1935 was an aging navy incapable of maintaining its strategic requirements and a decimated shipbuilding industry unable to meet the needs of a crash rearmament programme. Ironically, Britain was far better prepared in 1928 than in 1934 for the naval circumstances of the rearmament period.
Post-War Industrial Attrition and the London Naval Conference.
With the end of World War I, the British shipbuilding industry experienced an inevitable period of contraction as the surplus capacity necessary for wartime production was eliminated. The experiences of the industry during the 1920s can be broken down into three distinct periods: the early 1920s, in which there was a reduction in capacity as firms returned to peacetime production; the mid 1920s, in which the situation stabilized with assistance from the Admiralty; and the period from 1929 to 1935 when naval policy and naval armament capacity was crippled by the London Naval Treaty. These were years of both massive disarmament and rearmament. As a result of the Washington Conference, orders for battleships were cancelled in 1922; between 1920 and 1929 over forty obsolete battleships and battle-cruisers were sold or scrapped.\(^3\) Simultaneously, between 1920 and 1923, the Admiralty completed or converted many warships left on the stocks from 1918, especially cruisers and aircraft carriers; between 1922 and 1926 it laid down far more warship tonnage than any other naval power; and between 1927 and 1929 it provided enough orders to keep the industry alive until 1931.\(^4\)
Despite the orderly reduction of capacity during the 1920s, the Admiralty continually recognized the need to maintain various types of plant for emergency expansion and the major replacement programme which was scheduled to begin in 1931. In 1920 the Admiralty Board agreed that a minimum level of naval capacity should be preserved to sustain a construction programme which would allow four capital ships, eight light cruisers, sixteen destroyers, and miscellaneous smaller craft
to be completed within $2\frac{1}{2}$ years. In order to preserve this capacity, the Board agreed that if there were not sufficient orders subsidies would have to be used.\(^5\) By 1922, however, the Washington Treaty raised a new problem. The ten-year building holiday for capital ships greatly reduced the market in Britain and the world for some specialized capacity, specifically regarding firms producing armour plate and heavy shell. Thus, the Board had to decide whether it was best to keep the relevant British firms in operation or else to let them close down and revive them when they were required. If the firms were allowed to close, it would be very difficult to bring them back into operation as the skilled labour they required would have dispersed to other areas. While it would be possible to restart operations at such plants by using a nucleus of trained labour from other firms, this could not be guaranteed. At the same time the Board recognized that if the plants were closed down temporarily, firms would demand compensation in order to keep the capacity in being, and they would require further compensation if, after a few years, it was decided to allow the firms to close. Neither the firms nor the Admiralty had a clear picture of where policy was heading and consequently no decision could be made on the matter at that point.\(^6\)
The Admiralty’s assessment of this issue illustrates clearly its entire approach to the problem of maintenance of capacity. In March 1922 the Sea Lords considered the amount of plant required to complete cruiser construction programs between 1922 and 1931, the capital ship replacement program which was scheduled to begin in 1931, and the question of how easily the retained plant could be expanded to meet the extra construction requirements of wartime expansion. The issue of the production of
heavy shell was easily settled - the two British firms, Hadfields and Firths, agreed to keep their plant in operation without subsidies so long as the Admiralty divided its orders evenly between them. The issue of armour plate was more complex. The Sea Lords concluded that the five main armour producers (Armstrong, Beardmores, John Brown, Cammell Laird, and Vickers) had a total average output of 44,000 tons per year and a maximum output of 60,000 tons. Annual requirements for armour until 1931 were estimated at 3,000 tons, which, during the resumption of capital ship construction between 1931 and 1941, would rise to 18,000 tons. By subsidizing the production of the three biggest firms between 1922 and 1931, the Admiralty could maintain the capacity for an average output of approximately 26,000 tons of armour, with a maximum of 40,000 tons. The preservation of such an output would permit the Admiralty to construct its full capital ship replacement program, with the surplus capacity to produce armour for one extra capital ship per year in case of emergency.\(^7\) This the Admiralty proceeded to do - that is, through various means it offered hidden subsidies to some firms to maintain their plant while allowing others to leave the market. Thus the Admiralty preserved capacity, but not all of it; as the case of armour production illustrates, the Admiralty knew what its requirements were and would be in the future, added a generous margin for emergencies, and then took various steps to ensure that its needs could be met. This example also illustrates how dependent both the Admiralty and the armaments firms were on the proposed capital ship construction programme of 1931-41.
In 1925 the Board of Admiralty reconsidered all these issues. At this time, the
Sea Lords reconsidered the minimum power of expansion which should be maintained, and increased it. It was estimated that in the first year of a Far Eastern war, Britain would have to lay down seven capital ships, two aircraft carriers, thirty cruisers, forty destroyers, and thirty-two submarines. It also agreed to continue subsidies to the two major shell manufacturers, Hadfields and Firths, so long as they offered to continue their policy of dividing all orders in that sector. It agreed that the capacity available for the manufacture of naval guns at Vickers Armstrong had to be preserved, but preferably without subsidies. As long as it was economically viable, orders were to be given to preserve the gun-making capacity at Beardmores, but no subsidy could be given.\(^8\) These decisions show other methods which the Admiralty employed to preserve capacity. In some cases even if it did not place any orders with a firm, it subsidized the latter to keep the plant in being; in others, it concentrated orders with certain firms to keep their plants operating; it convinced still other firms to support currently unprofitable plant or even to expand plant for which no market would exist before 1931, either through the promises of subsidies or assurances of orders in the future. Work was divided as much as possible and new construction was concentrated in private dockyards to keep as much capacity as possible alive. However, such efforts were complicated by the need to preserve specialized capacity which only the government could use; generally, only large firms such as Vickers-Armstrong could preserve capacity in submarine and heavy armament construction which might not make profits for years on end.\(^9\)
While the government was willing to support the preservation of specialized
capacity, its interests differed from those of British shipbuilders. The Admiralty’s main concern was to maintain the capacity required for strategic purposes, to which end it tried to maintain as many firms as possible and expected them to operate at a loss for several years. Conversely, the firms necessarily thought in terms of profit. Thus, in the years following World War I, they took measures of their own to ensure the orderly reduction of capacity. The National Shipbuilders Society (NSS) was formed in 1929 to purchase redundant capacity and sell the capital equipment. Over the next ten years, the NSS eliminated 1.25 million tons of capacity while compensating the affected firms; between 1920 and 1929, about one million tons of capacity were reduced without the benefit of compensation to the firms involved.\(^{10}\)
This consolidation of capacity shaped both the industry and its skilled labour. Some firms, such as the Coventry Ordinance Works and the Darlington Forge Company, went out of business, while others, like Birmingham Small Arms, stopped producing armaments entirely. Major firms such as Cammell Laird, Beardmores, and Firths continued to produce armaments but only on the most limited scale. By 1933 only Vickers was able to produce armaments on a large scale\(^{11}\), and only because of its merger with Armstrong Whitworths in 1927.\(^{12}\) As a result of this attrition, unemployment in the shipbuilding industry, which had been 5% in 1920, reached 64% in 1932. Roughly 320,000 men were employed in the industry in 1920, which declined to 66,000 in 1932.\(^{13}\) Emigration and a shift to other industries resulted in a drain of skilled labour, the impact of which was felt dearly as the shipbuilding industry began to expand again in the mid- to late- thirties. In 1935 the total insured
labour force was about one half of what it had been in 1930, or 157,000 men.\textsuperscript{14}
A difficult adjustment from war to peacetime production was inevitable and necessary for the shipbuilding industry in the 1920s, and it was done with success. In 1929 the industry was still vibrant and able to meet the Admiralty’s requirements; however, between 1930 and 1935, the number of berths suitable for the construction of naval vessels declined by half.\textsuperscript{15} It was the policies laid down at the London Naval Conference of 1930 that caused this dramatic decline. These events stemmed from a combination of factors - a decline in the Admiralty’s political power in Whitehall and a rise in Anglo-American rivalry over naval strength. In the late 1920s, Anglo-American relations had been hampered by the Anglo-French disarmament agreement and the failure of the Geneva Disarmament Convention in 1927. In mid 1928, it was announced that Britain would accept the French position on land armaments while the French would do the same for the British proposals regarding cruisers. Fearing the impact this would have on any upcoming negotiations, President Coolidge retaliated in November by calling for American naval supremacy. This was followed by a bill providing funds for the construction of fifteen cruisers, which was passed by Congress in February, 1929.\textsuperscript{16} After the failed conference at Geneva in 1927, there was some public sentiment for real progress towards disarmament and Ramsay MacDonald’s new Labour government was determined to provide it. It believed that the key reason for the failure of the Geneva conference had been the interference of the naval “experts” of the countries involved; by taking a personal lead in the negotiations, MacDonald hoped to prevent this from happening again. After conversations with American
officials, MacDonald envisioned a conference aimed at the reduction of naval arms, attended by the political representatives of Great Britain, the United States, Japan, Italy, and France, with a three power conference [Great Britain, the United States, and Japan] dealing with arms limitation, to be held should Italy and France refuse to cooperate.\textsuperscript{17} In either case, the naval experts were to remain behind the scenes, giving advice during the conference as required and working out the strategic details of any agreement reached by the negotiators.\textsuperscript{18} As a result, the London Naval Conference of 1930 can be characterized as a politicians' conference, motivated by political considerations as opposed to strategic concerns.
With the failure of the Geneva Conference, both Britain and the United States were determined that the next conference would be a success. To ensure this, extensive negotiations were carried out before the conference took place to define the conference's objectives. It was announced publicly that the object of the negotiations was naval arms reduction, with the United States and Great Britain agreeing to naval parity. Parity would be based on equal strength, not numerical equality, and was to be measured using a "yardstick" devised by American and British naval officers.\textsuperscript{19} The key elements of the yardstick were the displacement, guns, and age of a ship; it would cover all combatant ships, with separate categories for capital ships, carriers, cruisers, destroyers, and submarines.\textsuperscript{20}
The concept of the yardstick, a formula which could evaluate the characteristics of a naval vessel and assign it a particular military value, was first proposed by Hugh Gibson, the American representative to the League Preparatory
Committee on Disarmament at Geneva in April 1929.\textsuperscript{21} Despite its flaws, British negotiators liked the concept of a yardstick because it could potentially help solve the difficulty of reaching an agreement on Anglo-American cruiser parity. Britain had more small [6" gun] cruisers than the United States, most of which were old and obsolete, as were Britain’s 7.5" gun cruisers. The United States had more 8" gun cruisers under construction or projected than Britain, although the latter had far more actually built. Thus, although Britain had a greater number of cruisers, the Admiralty feared that the United States might have newer cruisers of superior design in the near future; to decide what constituted parity on these terms was difficult, as negotiators at Geneva had discovered in 1927.\textsuperscript{22} However, it must be emphasized that the United States was really a paper tiger because its potential naval strength was not translated into actual ships. In 1930, Britain had 323,680 tons of cruisers built or building, which rose to 411,550 tons in 1936. The United States had only 215,050 tons of cruisers built or building in 1930, and only 304,450 tons in 1936. While the Americans were a potential threat to Britain, the threat was never truly realized.\textsuperscript{23}
It is ironic that after so much debate, the concept of the yardstick only served to get the powers to attend the conference; it was of little help in achieving an agreement because of its inherent flaws. British delegates hoped to press for a reduction in the size of capital ships from 35,000 to 25,000 tons displacement, with a maximum total tonnage of 375,000 tons for the United States and the British Empire, respectively. It was also hoped that the other powers would agree to reduce the maximum calibre of guns to 12" from 16", and to extend the lifespan of a capital ship
to twenty-six years from twenty. The British anticipated some opposition to their capital ship proposals, especially from the United States, which wanted to suspend building until 31 December 1936. It was believed, however, that the United States would accept the British proposal should the building holiday prove unacceptable, as would Japan. Britain also hoped to reduce the size of aircraft carriers from 27,000 to 25,000 tons displacement, with a total tonnage of 125,000 and a lifespan of twenty-six years.\textsuperscript{24}
Central to this conference was an agreement on cruiser strength and particularly the definition of parity between the United States and Great Britain. In preliminary negotiations, Britain announced that it was prepared to accept a minimum cruiser tonnage of 339,000 tons by 31 December 1936, if other powers would also substantially reduce their strength in these warships. The 339,000 tons included:
\textit{Table I: Proposed British Cruiser Tonnage}
| British Empire | Tons |
|---------------------------------------|----------|
| Fifteen 8-inch ships | 146,800 |
| Fourteen new 6-inch ships | 90,720 |
| Twenty-one older 6-inch ships | 101,480 |
In light of the British figures, the United States considered that parity with the Royal Navy would require a total of 315,000 tons. The table below illustrates the tonnage breakdown.
Table II: Proposed American Cruiser Tonnage
| United States | Tons |
|-------------------------------|----------|
| Twenty-one 8-inch ships | 210,000 |
| Ten Omahas | 70,500 |
| Five 6,900 ton 6-inch ships | 34,500 |
These figures would give Britain only 24,000 tons advantage in the cruiser class despite the greater age of British cruisers. The British contended that this could not constitute parity. The United States would retain a fighting superiority, while Japan would surely push for an increase in its total cruiser tonnage to meet its demand for 70% of the United States’ 8" cruiser tonnage. Thus Britain would retain an ageing cruiser fleet while the United States would develop a technologically superior fleet and Japan would build at least four more 8" cruisers. To avoid these problems, the British hoped to convince the United States to accept only eighteen 8" cruisers and a total tonnage of 285,000 tons. Meanwhile, Britain would reduce her 6" cruiser tonnage by 14,000 tons.\(^{25}\)
These proposals rested on the Admiralty’s position that a minimum of fifty cruisers was required to meet Britain’s strategic needs. With the reduced cruiser tonnage of 325,000 tons, this figure could be met only by reducing the number of replacement cruisers and retaining older and smaller vessels in service, or by redesigning the replacement vessels and reducing their displacement. In order to do either of these things, the British hoped to change the "yardstick" in their favour, with
one ton of 8" tonnage being equal to 2.2 tons of 6" tonnage. In other words, one 10,000 ton 8" cruiser would be the approximate equivalent of three 7,000 ton 6" cruisers. Despite the major concession by the Americans on this point, which would require the United States to accept fewer cruisers than the British, the Admiralty still remained sceptical. In order to meet the maximum tonnage restriction, the British would have to scrap twenty-two cruisers. Although the Royal Navy had planned to scrap fifteen of the vessels due to wartime service, it had also hoped to replace them. The other seven cruisers could be made to last until 1937 but, in turn, this would force a much larger construction program in 1937 and 1938. By extending the lifespan of these ships and limiting the rate of replacement under the terms of the treaty, six cruisers would have to be replaced in 1937 and six more in the year following.\textsuperscript{26}
Throughout the preliminary negotiations, Ramsay MacDonald’s government acted in full cooperation with the Admiralty. A weak Board of Admiralty, aware of its own political isolation and that MacDonald’s policies were supported by the strongest departments of state, the Foreign Office and the Treasury, abandoned the long term policies which the Royal Navy had maintained for a decade. It was forced to accept a proposal that the capital ship building holiday be extended from 1931 to 1936.\textsuperscript{27} After much hesitancy, the Admiralty grudgingly accepted the figure of fifty cruisers as the minimum required to meet British needs. Admiralty officials also agreed to tolerate an American preponderance of heavy 8" cruisers, as offset by the British superiority in the light 6" class. This position was much lower than the
Admiralty had demanded in 1927 for the Geneva conference but, at various stages in the 1920s the Admiralty had accepted that figure as meeting British needs. In 1929 it was considered acceptable so long as Europe remained stable and peaceful for the next decade. In a note from Lord Chatfield to N.H. Gibbs, author of the volume on rearmament policy in the official British history of World War II, Chatfield stated that relations between the Admiralty and the Cabinet did, however, reach a breaking point late in, or just after, the preliminary negotiations. The First Lord then overruled the Sea Lords and a political decision was made by Ramsay MacDonald to accept the demands of the United States and Japan. The Prime Minister ignored the advice of the Sea Lords, and they were not invited to the cabinet meeting at which the decision was made.\textsuperscript{28}
Despite the initial cooperation between MacDonald and the Admiralty, relations soon turned sour, largely because the Admiralty and the Treasury were divided over the building programmes which were to be carried out under the proposed London agreement. Based on the initial discussions about the treaty, the Admiralty proposed a negotiating position for the conference. Its proposals would produce a steady replacement program which would result in naval estimates of approximately £56 million per year, a program which was in fact more expensive than that carried out under Stanley Baldwin’s Conservatives.\textsuperscript{29} The Treasury was angered at the idea that a disarmament treaty could actually increase defence spending. In a pointed memo on naval spending, the Treasury noted that the figure of £56 million did not even include spending on capital ships and it rejected the Admiralty’s programme.
The Treasury asked that the Navy Estimates be reduced by at least £10 million from their current level.\textsuperscript{30} In order to achieve this aim, the Treasury proposed a complete revision of naval policy. Using the Navy Estimates of 1929 as a working example, it broke naval expenditure down into four key areas: pensions, maintenance, Fleet Air Arm, and new construction. Of these, substantial reductions could be achieved only in maintenance and new construction, especially in cruisers. The Treasury put forth three cases. Case A, the Admiralty proposals, included a fleet of fifty cruisers, 144 destroyers and fifty-five submarines, with lifespans of twenty, sixteen, and twelve years, respectively. In this case the annual cost for new construction would be £8,430,000. Case B involved the same number of ships, but extended their lifespans to twenty-four years for cruisers, twenty years for destroyers, and fifteen years for submarines. This would reduce the new construction needs to £6,843,000 per year. Case C involved a fleet of forty cruisers, 100 destroyers, and forty submarines, with the same lifespans as in Case B. New construction could be reduced to £4,946,000 per year.\textsuperscript{31}
The Treasury advocated Case C to reduce not only the cost of new construction but also of maintenance, the only other area in which real savings could be achieved. Case C would also affect a permanent reduction in the size of the fleet, below the size advocated by the Admiralty. The Treasury advocated this position for two reasons: it would reduce rather than just limit the level of armaments and it would cut spending to a point which would assist the financial well-being of the country. Rather than the Admiralty’s conception of naval program, based on the principle of parity with the
United States and actually resulting in a higher level of expenditure, the Treasury advocated that Britain "do all that we can to encourage an all-round reduction of armaments which save for the acceptance of the principle of parity with the U.S.A. will preserve present relativities generally, but on a level which we can afford." Thus, Britain would be perceived as leading the move towards disarmament while also saving a substantial amount of money. The Treasury hoped to avoid an overly strict interpretation of the doctrine of parity, and thus to avert the rivalry which occurred after the Washington Treaty as powers built vessels up to the maximum levels permitted. It claimed that the government faced two tough alternatives: either take risks and demonstrate a commitment to disarmament, or accept that even with the treaty no real reduction in spending or naval armament would be achieved.
The Admiralty countered these drastic proposals in a memorandum to Cabinet four days later. This emphasized that the Treasury's proposals completely overturned the government's policy, the very policy that had guided the preliminary negotiations which had made the upcoming conference possible. These proposals, moreover, ignored the minimum level of naval strength required by the Empire, which the Dominions had been assured would be maintained even after the conference. While the Treasury, of course, was free to examine the estimates and propose economies, the Admiralty's duty was to meet the Empire's defensive requirements. Ideally, the Admiralty required seventy cruisers to defend the Empire's sea communications, provided the United States was not considered a possible enemy. Twenty-five cruisers were required with the battlefleet; another twenty-five cruisers to guard British
interests in the Indian Ocean, Persian Gulf, and the Western Pacific; ten more to protect the Home and Mediterranean waters; and ten cruisers again for work in the South Atlantic, Eastern Pacific, Australia, South Africa, New Zealand, and Canada. Should war necessitate the organization of convoys in the Atlantic, even seventy cruisers would be woefully inadequate. Only on the assumption of extraordinary stable international conditions had the Admiralty agreed to accept a total of fifty cruisers.\(^{34}\)
Even the proposals based on the principle of parity, the Admiralty emphasized, did not leave Britain favourably placed against other powers. The Americans would have more 8” cruisers, the Japanese would have a ratio of greater than 5:3 in the same vessels, the French, Japanese, and Italians had more flotilla leaders, and Japan and France had more submarines. Of all these powers, Britain was most reliant on naval power for defence, and its maritime position would be endangered if the Treasury’s arguments were heeded. While the latter’s arguments were attractive, they were also misleading; while money would be saved in the short term as replacement construction was deferred, but long term costs would increase. Eventually, the Admiralty maintained, the ships would have to be built and they would be more expensive when the costs of restarting unused facilities were factored in.\(^{35}\)
**The Effects of the London Naval Agreement**
The London Naval Agreement of 1930 did not have nearly so drastic an effect as on the Royal Navy as the Treasury had advocated, and it certainly met the goals defined by the government before the conference. The capital ship building holiday
was extended until 1936 as MacDonald, though not the Admiralty, had hoped. Britain was permitted a total cruiser tonnage of 339,000 tons, and the United States 323,500 tons. Numerically, the Japanese challenge was prevented as Britain was allowed fifteen heavy cruisers, the Americans eighteen, and the Japanese twelve.\textsuperscript{36} However, the London Naval Agreement affected the Royal Navy, naval construction, and the shipbuilding industry in two ways. First, through its decisions on cruisers and by extending the building holiday for the replacement of capital ships, the London Treaty left Britain with, on average, the oldest heavy surface warships of the naval powers. Hence, in effect, if hostilities were to break out in the middle 1930s, virtually the whole fleet would need immediate replacement. Secondly, the treaty massively reduced the size of shipbuilding orders between 1931 and 1936, as compared to the period 1922 to 1929. All this, in turn, following on the short rations of the period 1927 to 1929 and the simultaneous impact of the Depression, gutted the naval armaments industry.
Of the 339,000 tonnage maximum for British cruisers, nearly 249,000 was in the form of existing vessels. The remaining 90,000 tons was to be used for the construction of ten 7,000 ton and four 5,000 ton 6" cruisers. These vessels were to be constructed over a three year period; replacement construction would be carried out in the years following. Although Britain was permitted to construct two destroyer flotillas per year under the terms of the treaty, the Admiralty thought it best to propose one leader and one flotilla per year.\textsuperscript{37} By forcing the Admiralty to cut its planned new construction, and limiting the extent of its replacement construction, the
London Agreement virtually guaranteed that later rearmament would be incredibly difficult because it forced a huge reduction in shipbuilding capacity. The shipbuilding industry had lost over one million tons of excess capacity in the decade following World War I, and between 1929 and 1934 further reduced its capacity by over one million tons more. The age of the fleet meant that, if a threat to British security arose in the 1930s, it would have to be almost entirely reconstructed; the lack of naval orders which resulted from the London Agreement meant that the capacity simply wouldn’t exist to accomplish this.
Under the circumstances, it was paramount for the Admiralty to preserve as much capacity as possible through a planned program of replacement. The Admiralty was less than successful in achieving this goal, in part for reasons beyond its control. The shackles of the London Treaty and of finances combined with the pressure of operational requirements complicated even routine planning in an extraordinary fashion. In 1932, for example, the Admiralty considered a report outlining a plan for carrier replacement. This report assumed that the fleet would be entirely reconstructed to meet the Washington and London Treaty standards by 1950. Such a fleet would need 405 aircraft, of which 250 would be based on aircraft carriers. The wartime strength of such a fleet was set at five large carriers with fifty planes per carrier.\(^{38}\) Based on these assumptions, Admiralty staff recommended that at the disarmament conference scheduled for London in 1935 Britain should pursue a minimum carrier tonnage of 110,000 tons, and try to set the maximum size of individual carriers at 22,000 tons. The report noted that the three oldest carriers of
the Royal Navy, the *Eagle*, *Hermes*, and *Argus*, were too slow to keep up with the fleet, unable to launch aircraft in all weather, and likely not able to survive in wartime.\(^{39}\) Hence, the report recommended that a new 22,000 ton carrier be laid down in 1933 and, on its completion in 1936, the *Hermes* and *Argus* should be scrapped or put in reserve.\(^{40}\)
The comments made on this report by senior Admiralty officials demonstrate the difficulty in devising long term plans in this uncertain era. They could not be sure of the plans of other powers, especially the United States, of what they planned to build and what designs they were developing. They emphasized the need to maintain Britain’s position for the disarmament conference scheduled for 1935. Scrapping the *Hermes* and *Argus* would put Britain well below the tonnage limit of the time and therefore in a weak bargaining position. Thus, it was suggested that the new carrier which was to be included in the 1933 programme should be deferred until 1936, and that the decision about when to scrap the *Hermes* and *Argus* be delayed until 1936 or 1937. That is, the operational requirements of the fleet should be subordinated to politics, something which was usually anathema to the Royal Navy.
Other officers emphasized that any decision should be delayed until better launch/arrester equipment was available or the needs of the Royal Navy were made clear.\(^{41}\) This example shows that the Admiralty could not decide on a design for the new carriers, partly due to its own faults - a lack of knowledge about the direction of other nations’ intentions or technological difficulties - but also due to factors relating to the London Naval Treaty. The first problem was the fault of the Admiralty, while
the second was due to factors which were beyond its control.
Nowhere did these difficulties produce greater problems than with regard to the establishment of coherent programmes of capital ship refit and replacement. Admiralty officials analyzed this issue at length in 1933. With a twenty-six year lifespan, a capital ship required a refit every eight to nine years in order to replace electrical wiring, repair or replace boiler tubes, to conduct any other necessary repairs, and to modernise the vessel. The program of large repairs was based on the ideal case that no more than three capital ships and one aircraft carrier would be in the dockyards at any one time. Of course, should the forecasted repairs to even one ship take longer than expected, this figure would increase. The refit schedule, moreover, assumed that no ship would develop serious unexpected defects; this was not always the case. The *Royal Sovereign*, the *Resolution* and the *Repulse* all had serious problems with their inner hulls and bulges which might have shortened the time between refits, lengthened the time required for repair, and limited their lifespan. Fortunately, these defects were corrected at an early refit and did not reappear.\(^{42}\)
Refits, moreover, were not simply a technical problem, but a strategic one. When deciding which battleships could be taken out of service for a refit, the Admiralty had to consider the strengths of other navies. Once a ship was taken out of service and the refit begun, it could not be made ready for service on short notice. Ships undergoing refit were put on twelve months' notice, that is, they could not rejoin the fleet for at least a year. Only with ships near the end of a refit could there be some flexibility and the process hastened.\(^{43}\) Between 1935 and 1938 it estimated
that, at any given time, the United States would have five capital ships in need of repair, France could have five or possibly six ships ready at short notice, all four Italian ships would be under repair, and Japan would have one or possibly two under repair at any time. Even considering the United States and Italy as neutral, Great Britain would have twelve capital ships, Japan eight, and France five or possibly six. Notably, the Admiralty was uncertain of the total capital ship strength of the United States and Italy and refused even to consider the possibility of either power becoming an enemy.\textsuperscript{44} Understanding the plans of the other naval powers was vital, lest Britain find herself at an unexpected disadvantage with regard to capital ship strength, and such understanding was recognizably hard to achieve. The key here is that the Admiralty was already taking a serious look at the naval standard of the time and attempting to show its deficiencies. This memo was drafted in 1932 and the naval standard at that time required that the Royal Navy be able to send a deterrent force against Japan to the Far East while simultaneously maintaining a deterrent force in British Home Waters. With the expansion of the Japanese navy, and the growth of the French and Italian navies, the Admiralty did not feel it could meet this standard and make provisions for the adequate maintenance of the fleet at the same time.
These problems regarding refits meant that the proposed twenty-six year lifespan of individual warships and the desired strength of the standing fleet strength often could not be achieved. Faced with a need to maintain a higher capital ship strength, the Admiralty either had to extend the lifespan of a ship or decrease the duration of refits. Thus, the \textit{Repulse} and \textit{Renown} were given forecasted lifespans of
thirty-five and thirty-four years, respectively. As the Director of Naval Construction, A.W. Johns, noted, this assumption was optimistic even with effective repairs, modernisation, and regular maintenance - that is, the warships might break down unexpectedly.\textsuperscript{45} The possibility of extending ships’ lives was limited because, as a result of the tonnage limitations imposed by the various disarmament conferences, ship designers were forced to sacrifice durability in order to save weight. Another key factor was cost, for as a ship became older the cost of refits became prohibitive. Major defects which appeared due to age could not be repaired easily or cheaply.\textsuperscript{46} The other option open to the Admiralty, planning refits of shorter duration, required that a lower standard of completion be accepted on non-essential systems (for example, ventilation), in order to speed up the process. This possibility, again, was limited. Older vessels required that substantial time be spent to repair defects; to shorten refits and cut expenses less time was spent on modernizing these ships, which were soon to be decommissioned.\textsuperscript{47}
In any case, shortened refits were possible in theory, but not practice. In 1933, in theory, a capital ship could be refitted in fifteen months if deck protection over the engine rooms and magazines was not completed, or in twenty-four months if such protection was installed. In practice, such time estimates could only be achieved if the yearly estimates met the financial requirements of the Royal Dockyards, if there was no unforeseen interference in the refit programme which diverted men to other tasks, if the dockyards could hire additional temporary employees as required, and if the Royal Dockyards built little new construction, with most and preferably all of that
contracted out to private firms.\textsuperscript{48} In short, all ideal conditions had to be met to complete refits in the Royal Dockyards on time; these facts coupled with the lengthy time between refits meant that there was little chance of shortening their duration. In short, the Admiralty was caught in an impossible situation. Financial constraints and treaty restrictions prevented the construction of new ships; the need to keep a fleet in being adequate to meet the threat of other navies made it necessary to lengthen the lifespan of British ships; yet the refits which were required to extend these lifespans could rarely be completed in time to keep the fleet at its required strength. Unless conditions were perfect, the Royal Navy was forced to operate at a strength below that which was necessary to meet its strategic requirements.
All this was complicated by a further problem - the issue of the replacement of the Royal Navy’s fifteen battleships, which was scheduled to begin in 1936. The Admiralty forecast that by 31 December 1936, considering existing refit/replacement schedules and treaty obligations, the age of those vessels would be as follows.
\textbf{Table III: Age of British Battleships by 31 December 1936}
| Number of Ships | Age (as of 31/12/36) |
|-----------------|----------------------|
| 3 | 21 years |
| 8 | 20 years |
| 1 | 19 years |
| 1 | 16 years |
| 2 | 9 years |
Thus, by 1943, seven years after the expiry of the London Naval Treaty, eleven capital ships would have reached their age limit. Based on a rate of replacement of
one capital ship per year, some of those vessels would have to be kept in service long after they had passed the twenty-six year limit. In particular, the *Repulse* and the *Renown* would have to remain in service until they were well over thirty years old.\(^{49}\)
As a result of the London Treaty, Britain would have the world’s oldest fleet until almost 1950.
Faced with such a situation, the Admiralty realized that a faster programme of replacement construction could meet minimum defensive requirements. As well as convincing the government that it was necessary to accept the extra expense of such a construction programme, the Admiralty also had to get other naval powers to agree on an age limit of twenty-six years for capital ships. In doing so, the Admiralty hoped that all the powers would agree that any power had the right to advance its battleship construction programme so that none of its ships exceeded the twenty-six year limit.\(^{50}\) Obtaining such an agreement was also viewed as a way to counter the perceived threat from Japan. The Admiralty felt a stronger position was required in the Far East not only to deter Japanese aggression, but also to protect British interests should Japan go to war with the United States or the Soviet Union while Britain remained neutral. As it was believed that Japan was likely to break the capital ship ratio laid down at the Washington Conference in 1921, the Admiralty felt it best to speed up replacement construction in order to free capacity for the new construction which might be necessary to counter Japanese naval expansion in the future.\(^{51}\) The advantage of getting other powers to agree to an age limit for capital ships was threefold - Britain could build new ships without appearing to violate the spirit of existing disarmament
agreements; it could maintain existing shipbuilding capacity; and it could free up capacity for new construction which might be necessary later on. Moreover, as the Admiralty also recognized at the time, all of these arguments were futile. It could make plans and propose solutions, but until the government's policy changed, no action could be taken. This issue, however, does indicate the Admiralty's views regarding armament capacity in the years immediately after the London treaty.
The Admiralty viewed increasing the pace of replacement construction as necessary not only to meet Britain's naval commitments but also to save the shipbuilding industry. The Admiralty wanted to give as many contracts as possible to as many British shipbuilding firms as possible in order to keep a minimal level of capacity in being. It recognized the need to preserve the plant and the concentration of skilled labour in the industry. Without a compensating increase in orders for mercantile shipping to replace lost naval business, the industry was woefully weak and areas specific to naval construction were being hardest hit. In order to remedy this, in August 1933 the Director of Naval Contracts proposed a slow increase in the number of contracts given to private firms, while hastening the rate of laying down capital ships. This would allow firms to gradually increase levels of men and material in the yards without compromising the quality of the initial orders. While the first ship ordered from a private yard might take four years to complete, subsequent orders might easily be completed in less than three. Once plant and labour were organized, further orders could be completed with less disruption. Hence, the Admiralty estimated that a replacement program based on one capital ship per year was worse
than no order at all. Some ten firms would be involved in major naval contracts, and on a cycle of one capital ship laid down per year, each firm would have at least six years of idle time between contracts.\textsuperscript{52} If a firm received one contract every ten years, and the contract took four years for the firm to complete, the result was a wait of six years before the next order was placed. This would hurt the Royal Navy as contracts would be delayed, and it would also disrupt the firms’ entire operation. Money would be lost as plant was activated and deactivated and labour would simultaneously be disrupted. The already highly mobile workforce was likely to be diluted as skilled workmen travelled from yard to yard following the contracts. The diminished pool of skilled workmen would decline still further as men left shipbuilding in favour of more stable work.
To avoid these problems, the Admiralty hoped to order more than one capital ship per year while limiting the number of firms involved in the scheme so that each firm received a second order before the first was completed. The target was a second order in under three years, a pace which the armour manufacturers could match.\textsuperscript{53} Thus firms could retain skilled tradesmen and simply move them to the next vessel as they completed their jobs with the first. Nor was the Admiralty alone in proposing this option; the desperation felt by labour and management in the shipbuilding trades was reflected in a deputation from the firms to the Admiralty in 1930. Unemployment was so severe that the firms begged for an accelerated construction program; otherwise, unemployment in the industry was expected to jump from 23% to over 50%. As the President of the Shipbuilding Employers’ Federation pointed out,
there could be no better time for an accelerated program; prices were the lowest they had ever been and the government was simply avoiding the inevitable. Money would have to be spent on construction eventually and there could be no better time than the present.\textsuperscript{54}
Despite the advocacy of the Admiralty and the pleas from the shipbuilding industry, plans for accelerated construction were for naught. They were blocked by a combination of international negotiations and fiscal restraint. Rather than begin a slow and steady replacement program, the government considered a complete cancellation of new naval construction in 1931. Of course, the Admiralty opposed any such move and maintained that it could only increase the uncertainty in the global naval situation; the government relented.\textsuperscript{55} A steady programme of construction was predictable, whereas any building holiday would eventually end, and as it drew to a close, tension would result as the naval powers watched each other’s actions and feared for the worst. The Admiralty also stressed that such a holiday would cripple the industry. Skilled tradesmen would be lost to more secure employment, most notably the aircraft industry, and new workers would not be apprenticed. A lack of contracts from the Royal Navy, moreover, would also reduce foreign orders; the best advertisement for British shipbuilding was the Royal Navy and if it stopped building, other countries might well lose respect for British industry. Firms would have either to eliminate capacity or continue operating at a loss; as few firms were willing to lose money the Admiralty would have either to subsidize them or else expand the capacity in the Royal Dockyards. In either case, the result would be greater expense in the
long term.\textsuperscript{56}
Despite the flurry of proposals which the Admiralty presented to the government, very little was accomplished and the problems they were intended to solve continued unabated. Many of the steps which the Admiralty proposed were adopted once rearmament began in 1935, but not before. They were blocked by the terms of the London Naval Agreement and by the stringent financial conditions of the period 1929 to 1935.
Naval estimates were erratic in the early 1930s, as is illustrated in Table IV below. The size of Effective estimates fell marginally between 1930 and 1933 and then sharply increased in the years 1934 and 1935. The flux in Vote 8 was particularly dramatic. Although relatively stable between 1930 and 1932, it rose sharply in 1933 and 1934, and increased by just under £5,000,000, or 25%, in 1935.
\textit{Table IV: Comparison of Effective Estimates and Vote 8 (Shipbuilding, Repairs & Maintenance), 1930 to 1935.}\textsuperscript{57}
| Year | Effective Estimate | Vote 8 |
|------|--------------------|--------------|
| 1930 | £43,618,686 | £15,983,643 |
| 1931 | £42,806,926 | £15,566,205 |
| 1932 | £41,255,258 | £15,887,597 |
| 1933 | £44,465,550 | £18,524,658 |
| 1934 | £48,603,114 | £20,994,482 |
| 1935 | £51,595,895 | £25,897,792 |
Table V: Vote 8 by Section, 1930 to 1935.\(^{58}\)
| Year | (I) Personnel | (II) Material | (III) Contract Work |
|------|---------------|---------------|---------------------|
| 1930 | £6,235,810 | £4,198,553 | £5,549,280 |
| 1931 | £6,331,401 | £4,517,053 | £4,717,751 |
| 1932 | £6,267,764 | £4,190,785 | £5,429,048 |
| 1933 | £6,192,587 | £4,420,159 | £7,911,812 |
| 1934 | £6,531,500 | £4,808,656 | £9,654,326 |
| 1935 | £7,173,391 | £6,562,803 | £12,161,598 |
Thus, although some of the increase in Vote 8 after 1932 was caused by an rise in the costs of personnel and material in the Royal Dockyards, it stemmed mostly from contract work to private industry. Between the years 1930 and 1935 contract expenses more than doubled, with the greatest part of the increase coming in the years 1933 to 1935. The sudden increase in the Effective estimates, and particularly in Vote 8, as a result of rearmament was precisely the situation which the Admiralty had hoped to avoid since 1919, and about which it had warned the Treasury and the Cabinet. It created problems when firms were suddenly given a number of orders immediately after a number of years in which they had been starved of orders, with a resulting loss of capacity.
The low expenditure between 1930 and 1933 and the later rise in the naval estimates reflects the ongoing battle between the Admiralty and the Treasury. The Treasury succeeded in cutting the 1931 estimates and perhaps with some reason. In a memo dated 2 February 1931 the First Lord of the Admiralty advocated a programme over three years of nine large cruisers, three small cruisers, four leaders, thirty-two
destroyers, twelve submarines, and miscellaneous smaller vessels for a total cost of £36,777,400.\textsuperscript{59} The usual arguments were made about the need for a steady construction programme and the benefits to labour and industry, while the First Lord stressed that only £52,600 would be spent in 1931 on this programme.\textsuperscript{60} The Admiralty’s arguments were overturned by the Chancellor of the Exchequer, who pointed out that while costs were low in the first year of the construction programme, spending would rise sharply in subsequent years - to £2 million in 1932 and £3 million 1933. In fact, the Admiralty’s programme would hamper the industry, as spending would drop in 1931 and then increase sharply. This went completely against the Admiralty’s own belief that steady expenditure on consistent programmes was the only way to preserve armaments capacity. Instead, the Treasury proposed a smaller programme by eliminating one large cruiser, two destroyers, and two sloops, saving a total of £2,525,000 over three years. The Treasury’s amended programme would stabilize spending over the three year period and thereby cause less disruption to labour and industry.\textsuperscript{61} In the end, its arguments for a reduced programme were successful and not entirely unsound.
The Admiralty’s estimates of the period 1932 to 1935, however, reflect the failure of the London Naval Treaty to limit naval armaments and the slow rise of a new naval menace. In 1933 the Admiralty had to revise its original 1933 program to account for the larger size of the 6" gun cruisers planned by other powers, particularly the American plan to build 10,000 ton cruisers armed with fifteen 6" guns. Rather than one \textit{Leander} class cruiser (7,250 tons) and three \textit{Arethusa} class cruisers (5,400
tons), the Admiralty had to request either three cruisers of a new design (8,900 tons, later known as the *Minotaur* class) or two new design cruisers and one *Arethusa* class cruisers. The latter proposals was adopted after discussions with the Treasury. Under the revised program, the cost of cruiser construction could be reduced by £200,000 to £5,600,000.\(^{62}\)
In 1934, the Admiralty attempted to overcome the challenge presented by the Japanese construction of six large 6" gun cruisers, and four even larger American cruisers. The Admiralty wanted to counter with six *Minotaur* class cruisers, though it also recognized that a new treaty was required to limit construction of these new, larger 6" cruisers. The Admiralty also requested that the orders be placed with contractors at the beginning of the financial year to allow firms to take advantage of the summer weather.\(^{63}\) While the 1935 construction programme was initially considered a normal year, American actions again necessitated a change in Admiralty plans. Under the London Treaty (1930), up to 16% of a nation’s destroyer tonnage could be built in the form of flotilla leaders. In 1934, the United States announced its intention to build up to this maximum, a policy followed by Japan since 1930. As a result, the Admiralty requested that the destroyer flotilla normally constructed every year be replaced by seven destroyers of 1,830 tons. After overage vessels were scrapped, this would leave the Royal Navy with heavy destroyers totalling 22,000 tons, or 14.5% of her total destroyer tonnage.\(^{64}\) The Chancellor of the Exchequer approved this change to the 1935 programme on 1 March 1935 as well as the decision to lay down a second destroyer flotilla brought forward from the 1936 programme,
thereby increasing the 1935 programme to two destroyer flotillas.\textsuperscript{55} Thus, in 1934 and 1935, we the Admiralty began to overcome the restrictions of the Treasury. This reflects a resurgence in the Navy’s growing political influence. The increased threat from Japan and the growth of German power had forced the government to reconsider its commitment to disarmament, a change reflected by the appointment of the Defence Requirements Committee in 1933. However, the half decade of neglect of armaments industries coupled with the government’s reluctance to take action on the committee’s proposals meant that the problem of rearmament appeared insurmountable.
Notes
1. Roskill, *Naval Policy*, v. I, p. 331-32.
2. William Hornby, *Factories and Plant* (London: HMSO, 1958.), p. 37.
3. Philip Pugh, *The Cost of Seapower* (London: Conway, 1986.), p. 157, Fig. 6/4
4. J.R. Ferris, *Men, Money and Diplomacy* (Ithaca: Cornell U.P., 1989), p. 135-37.
5. "Maintenance of Power of Expansion in the Manufacture of Armament Material" Board of Admiralty Minutes, 28 January, 1920. ADM 167/60.
6. "Maintenance of Power of Expansion" Board of Admiralty Minutes, 21 March, 1922. ADM 167/66.
7. Report on meeting of the Sea Lords to consider retention of certain plant to meet future naval construction requirements. Board of Admiralty Minutes, 21 March, 1922. ADM 167/66.
8. Board Minutes, 25 June 1925, ADM 167/71.
9. Gordon, p. 77.
10. ibid.
11. 'Shipbuilding and Armaments Industry', 2 March 1933, CAB 21/371.
12. Lesley Richmond & Bridget Stackford, *Company Archives* (Aldershot: Gower, 1986.)
13. Hornby, p. 39.
14. M.M. Postan, *British War Production* (London: HMSO, 1952) p. 4
15. ibid.
16. B.J.C. McKercher. "'Our Most Dangerous Enemy': Great Britain Pre-eminent in the 1930's" *International History Review* (XIII, 4, November, 1991, pp. 751-783), p. 764.
17. David Carlton, *MacDonald vs. Henderson* (London: Macmillan, 1970), p. 106-7.
18. R. MacDonald, regarding conversation with Dawes & Gibson, 25 June, 1929. *Documents on British Foreign Policy 1919-1939*. Series II, V. I (London: HMSO, 1947), pp. 10-11.
19. MacDonald to Dawes, 8 July, 1929. *Documents on British Foreign Policy 1919-1939*, Series II, V. I, pp. 19-20.
20. Dawes to Macdonald, 12 July, 1929. *Documents on British Foreign Policy 1919-1939*. Series II, V. I., pp. 22-23.
21. Roskill, *Naval Policy*, v.II, p.37.
22. Raymond G. O’Connor. "The "Yardstick" and Naval Disarmament in the 1920's" *Mississippi Valley Historical Review* Vol XLV, No. 3, (December 1958.), p. 455.
23. B.J.C. McKercher. "'Our Most Dangerous Enemy': Great Britain Pre-eminent in the 1930's." *International History Review*. XIII, 4 (November, 1991, pp. 751-783), p. 765.
24. C.P. 5(30) "London Naval Conference, 1930" CAB 24/209
25. C.P. 5(30), "London Naval Conference, 1930." CAB 24/209.
26. C.P. 5(30) "London Naval Conference, 1930." CAB 24/209.
27. Carlton, p. 121-23.
28. N.H. Gibbs, *Grand Strategy: Rearmament Policy* V. I (London: HMSO, 1976.), p. 29.
29. Gibbs, p. 29.
30. C.P. 12(30) "Memorandum by the Treasury on Financial Aspects of the Naval Conference." 16 December, 1929. CAB 24/209.
31. C.P. 12(30) "Memorandum by the Treasury on Financial Aspects of the Naval Conference." 16 December, 1929. CAB 24/209.
32. C.P. 12(30) "Memorandum by the Treasury on Financial Aspects of the Naval Conference." 16 December, 1929. CAB 24/209.
33. C.P. 12(30) "Memorandum by the Treasury on Financial Aspects of the Naval Conference." 16 December, 1929. CAB 24/209.
34. C.P. 13(30) "Financial Aspects of the Naval Conference" 20 December, 1929. CAB 24/209.
35. C.P. 13(30) "Financial Aspects of the Naval Conference" 20 December, 1929. CAB 24/209.
36. C.P. 128(30) "The London Naval Treaty, 1930." 22 April, 1930. CAB 24/211.
37. Roskill, *Naval Policy*, v.II, p. 73.
38. 'Carrier Replacement Program', 1932, ADM 1/9342, p. 2.
39. The Eagle had a top speed of 24 knots, the Hermes 25 knots and the Argus 20 knots. The Hermes and the Argus were the smallest carriers, with tonnages of 10,850 and 14,450 tonnes respectively.
40. 'Carrier Replacement Program,' 1932, ADM 1/9342, p. 3.
41. ibid, p. 4-8.
42. 'Battle Fleet - Large Repairs to 1940', 1934, ADM 1/8774/107/M.01784/33.
43. 'Battle Fleet - Large Repairs to 1940; Director of Plans - Memo', 1934, ADM 1/8774/107/M.01784/33.
44. 'Battle Fleet - Large Repairs to 1940; Director of Plans - Memo', 1934, ADM 1/8774/107/M.01784/33
45. 'Battle Fleet - Large Repairs to 1940; Director of Naval Contracts - Memo', 1934, ADM 1/8774/107/M.01784/33.
46. 'Battle Fleet - Large Repairs to 1940; Director of Plans - Memo'. 1934, ADM 1/8774/107/M.01784/33.
47. 'Battle Fleet - Large Repairs to 1940; Director of Naval Contracts - Memo'. 1934, ADM 1/8774/107/M.01784/33.
48. 'Battle Fleet - Large Repairs to 1940; Director of Dockyards - Memo', 1934, ADM 1/8774/107/M.01784/33.
49. 'Battle Fleet - Large Repairs to 1940; Director of Plans - Memo'. 1934, ADM 1/8774/107/M.01784/33.
50. 'Battle Fleet - Large Repairs to 1940'. 1934, ADM 1/8774/107/M.01784/33.
51. ‘Battle Fleet - Large Repairs to 1940; Director of Plans - Memo 1934, ADM 1/8774/107/M.01784/33.
52. ‘Battle Fleet - Large Repairs to 1940; Director of Naval Contracts - Memo’. 1934, ADM 1/8774/107/M.01784/33.
53. ‘Battle Fleet - Large Repairs to 1940; Director of Naval Contracts - Memo’. 1934, ADM 1/8774/107/M.01784/33.
54. Cabinet Paper 424(30), Deputation for Shipbuilding Employers on the Subject of Naval Work, 15/12/30, CAB 24/217.
55. Naval Building Holiday, 1931, ADM 1/8748/138, p. 5.
56. ibid, pp. 5-6.
57. Appropriation Account, Navy Services, 1930, British Sessional Papers 1931-32, Appropriation Account, Navy Services, 1931, British Sessional Papers 1932-33, Appropriation Account, Navy Services, 1932, British Sessional Papers 1933-34, Appropriation Account, Navy Services, 1933, British Sessional Papers 1934-35, Appropriation Account, Navy Services, 1934, T 161/710/S34880/34, Appropriation Account, Navy Services, 1935, T 161/S34880/35.
58. Appropriation Account, Navy Services, 1930. British Sessional Papers, 1931-32, Appropriation Account, Navy Services, 1931. British Sessional Papers 1932-33, Appropriation Account, Navy Services, 1932. British Sessional Papers, 1933-34, Appropriation Account, Navy Services, 1933. British Sessional Papers, 1934-35, Appropriations Account, Navy Services, 1934. T 161/710/S34880/34, Appropriation Account, Navy Services, 1935. T 161/S3480/35.
59. Naval Construction Programme, 1931, Cabinet Paper (29)42, CAB 24/219, Appendix A.
60. Naval Construction Programme, 1931, Cabinet Paper (29)41, CAB 24/219, pp. 1-2.
61. Cabinet Paper (29)42, Naval Construction Program CAB 24/219, p. 3.
62. C.P. 233(33) "Programme of New Construction, 1933. Proposed Alterations" 24 October, 1933. CAB 24/243.
63. C.P. 311(33) "Program of New Construction for 1934." 21 December, 1933. CAB 24/245.
64. C.P. 42 (35) "Programme of New Construction for 1935." 25 February, 1935. CAB 24/253.
65. C.P. 198(35) "Programme of New Construction for 1935." 17 October, 1935. CAB 24/257.
Chapter Two: Rearmament
There were two distinct phases in British naval rearmament in the 1930s. The first corresponded with the formation of the Defence Requirements Committee in 1933, which was created to investigate the material deficiencies of the three services and report on how to remedy them. These recommendations were not fully honoured. Had they been acted upon, Britain would have been in a much better position when rearmament began in earnest in 1936. Secondly, as the political situation worsened in 1935, the DRC was once again asked to report on what could and should be done to prepare the services for the growing threat from Germany and Japan. The changes in policy which it advocated illustrate a fundamental shift in defence preparations - from the move overcoming deficiencies to preparing for war. Unfortunately, this decision came too late; since little had been done to make good the deficiencies outlined in 1934, the services had much work to do before rearmament could really begin. Their attempts to do so were hampered by limits to Britain’s financial power. Instead, by failing to eliminate military deficiencies early in the decade, later plans for rearmament were doomed to fall short because too much had to be done too quickly from the same limited financial and industrial base. No service was more significantly hampered by these problems than the Admiralty, which until 1936 was forced to operate within the limits of the London naval treaty of 1930 and to account for the intended naval conference of 1935.
The problems caused by the rush to rearm in the late thirties were exacerbated by the Admiralty’s pressures to adopt a new, and much more expensive, standard of
naval strength. Throughout this period the Admiralty engaged the Treasury in a bitter struggle over naval policy and it finally achieved a remarkable victory. By 1936-37, the Admiralty was free to pursue virtually any naval policy it liked, and it did.
Unfortunately, this political victory came too late for the policies to have effect before the outbreak of World War II. The result of these pressures, some outside of the Admiralty’s control and others not, led the Admiralty to pursue an extraordinary policy of feast and famine with the British armament industry.
**The Defence Requirements Committee**
By 1933 the British government recognized the need to bolster its sagging defence forces. The new National Government (a coalition of Conservatives, National Labour and Liberals, with Ramsay MacDonald as Prime Minister) was far less optimistic about British security than its Labour predecessor had been. Two events were central in forcing this reevaluation: the Japanese invasion of Manchuria in September 1931 and its siege of Shanghai in 1932, and the rise to power of Adolf Hitler in January 1933. Of these two events, Japan’s aggressive behaviour in the east caused the greatest immediate concern and was directly responsible for the cancellation of the Ten Year Rule. Under the provisions of the rule, the services were to assume that no major war would break out within the coming decade and plan accordingly. The period was adjusted annually depending on the existing international situation. While the Ten Year Rule was neither the sole basis of defence policy in the 1920s nor the sole cause of its flaws,\(^1\) its cancellation indicated that Whitehall no longer
believed that the world would be stable and peaceful. Meanwhile, the Navy concluded that Japan would not adhere to the spirit of the London Naval Treaty, and the fear of a Japanese menace again became central in its eyes. The rise of a potential enemy in the east was coupled with the growing possibility of German hostility. It was Adolf Hitler’s new government which jolted the British government into finally evaluating the country’s military readiness.
The potential menace of Japan and its implications for defence planning was considered by the Chiefs of Staff Sub-Committee (COS) in a report to the Committee of Imperial Defence (CID) on 17 March 1932. The COS concluded that Britain lacked the naval forces to protect British possessions in the Far East, or the bases necessary to maintain the fleet in the area. Thus, they felt, the entire region was open to Japanese attack. The Army and the RAF were similarly weakened and unable to protect British interests. The only solution was the abolition of the Ten Year Rule: in effect, the service argued that Japan was a potential threat and that Britain should consider rearmament to match it.\(^2\) The Treasury retorted that Britain could not afford to abolish the Ten Year Rule as a basis for military planning. If Britain could not defend its Eastern interests militarily, neither could it afford to do so. The purpose of the Ten Year Rule, the Treasury argued, was two-fold: to prevent the military from planning for events which the Government believed to be only remotely possible, and to avoid preparations which were "beyond the financial capacity of the country to provide against."\(^3\) Thus, while the Treasury accepted that Britain did not have the military capability to defend its Far Eastern interests, it denied that it was financially
feasible to remedy the situation. Until the country was in a better financial position, Britain would have to accept the possibility that its Far Eastern possessions might be lost. In any case, throughout the 1920s and 1930s, the Treasury denied that this danger existed. It held that no clash of interests existed between Britain and Japan, or that a war was likely - the problem was one to be solved by diplomacy, not force.
On 22 March the CID considered both statements and gave the military a partial victory. It agreed that the Ten Year Rule should be abandoned as a guide for defence planning, but emphasized that planning for the possibility of war was not to require an increase in defence expenditure. In the words of Robert Shay, "[t]he services had won the right to plan, but not to spend." In fact, abolishing the Ten Year Rule in 1932 had no more effect than creating it had done in 1919.
Hitler's assumption of power in Germany gave rise to a great deal of pessimism in Britain and France. The Foreign Office warned the government that the situation might produce an arms race between Germany and Britain and France and lead to an escalation of tension and perhaps a European war by 1937. The departure of Germany from the Geneva Disarmament Conference in October 1932 made this worst-case scenario seem all the more likely. The COS in turn warned that Germany could present a direct threat to Britain within three to five years. This would put Britain in great danger as the Army was simply incapable of fighting a war, even a limited one, on the Continent. The COS also raised another crucial point. The deficiencies were not easily remedied because the Army's requirements went beyond a need for men and armaments; the country did not have the capacity to produce the
armaments which the Army required in adequate quantities.\textsuperscript{5} Nor, the COS warned, were the other services in better shape for a major war.
Accordingly, the Defence Requirements Committee (DRC), a sub-committee of the Committee for Imperial Defence, was established to investigate this situation and made its report in February 1934.\textsuperscript{6} The DRC was asked to consider how the past years of economy and disarmament had affected the services and to recommend a program to meet their worst deficiencies. It was not asked to recommend a rearmament programme, and it was limited to proposing responses to threats presented by Japan and Germany in the Far East, Europe, and India. Japan was considered to be the more immediate concern, while it was recognized that the greatest long term threat came from the reviving Germany.\textsuperscript{7} The DRC was composed of very senior figures; its members included the Chiefs of Staff, Sir Warren Fisher, Permanent Secretary of the Treasury, Sir Robert Vansittart, Permanent Secretary of the Foreign Office and the committee chair, Maurice Hankey.\textsuperscript{8} The committee was well balanced in the areas of defence, finance, and foreign policy. Its civilian members recognized the danger at hand and the need to strengthen British defences, but they held different views on how to do so.
This fact was of fundamental significance to the Navy. By and large, the three service chiefs favoured a balanced programme, one which would divide new expenditures roughly evenly between them, and they actually were less enthusiastic about expanding their size than were their civilian counterparts. Fisher’s chief concern was to bolster British security while keeping expenditure within the limits of a
balanced budget. To accomplish this, and to concentrate resources on the greater threat he perceived from Germany, Fisher advocated a diplomatic solution with Japan. Britain could not afford the navy required to dominate Pacific and European waters, but the Navy alone could not counter the German menace. Rather than heavy naval expenditure, Fisher felt that the RAF could best counter the German threat. Fisher was supported on this specific point by the CIGS, General Archibald Montgomery-Massingberd, and the CAS, Air Marshal Edward Ellington, as well as Vansittart.\(^9\)
Conversely, it was Admiral Ernle Chatfield who, with the support of Maurice Hankey, advocated a strong British presence in the Far East, and sought to focus resources on the Navy.\(^{10}\)
The DRC’s report was delivered on 28 February 1934 with only slight dissension; Vansittart, the member most outspoken in his support of the RAF, even more so than Air Marshal Ellington, felt that the report did not adequately address the needs of RAF rearmament.\(^{11}\) The report recommended a five-year program, and a balanced one, to meet the what it perceived to be the worst deficiencies among the services; the total cost of the program was forecast as £71,322,000. The Home Defence Air Force was to be raised to fifty-two squadrons, while the Army was to be increased so as to be able to put four infantry divisions, one tank brigade, and one cavalry division into action on the Continent.\(^{12}\) Since the DRC’s authority was limited to making good deficiencies, its recommendations and the estimated cost of its programme did not include new naval construction. That issue, in any case, would turn on the planned naval conference in 1935. However, the DRC did recommend
that the Royal Navy modernize its capital ships as other naval powers were doing. Essential stores were to be built up and naval bases were to be improved; the Singapore naval base was to be made operational by 1938.\textsuperscript{13} The provisions included £29,750,000 total expenditure, as follows:
| Item | Cost |
|-------------------------------------------|----------|
| Fleet Air Arm | £1,900,000 |
| Capital Ship Modernization | £1,900,000 |
| Personnel | £8,000,000 |
| Anti-submarine equipment | £250,000 |
| Local Sea Defences | £600,000 |
| Fuel Reserves | £6,700,000 |
| Ammunition, torpedoes, mines, depth charges | £1,700,000 |
| Stores | £4,300,000 |
| Singapore base | £4,400,000 |
While the Admiralty was happy with the findings of the DRC, its optimism was short-lived. After the Cabinet considered the report, it was sent to the Ministerial Committee on Disarmament [the DC(M)] for debate. The DC(M) included the political heads of the services, the Treasury, and the Foreign Office; it was chaired by the Prime Minister, Ramsay MacDonald, or, when he was unavailable, the Lord President of the Council, Stanley Baldwin. The DC(M) considered the DRC’s report over a period of three months, discussing the matter in thirteen sessions. As a
deliberate choice, one initiated by Warren Fisher, the DRC’s report had been formulated with little concern for the financial implications of the program. That is, it defined what the services required to make good their most serious deficiencies, with little concern regarding how to finance the programmes or to assign priorities to the needs of the various services. When the DC(M) considered the report, however, it had to take account of precisely these issues of costs and priorities. Here no service was threatened more than the Navy.\textsuperscript{14}
After the London Naval Treaty, British naval planning was based on the need to protect British home waters against a European-based threat while simultaneously maintaining a small naval force in the Far East to deter the Japanese. Given the expansion and modernization of the Japanese navy, coupled with the possible resurgence of Germany as a naval power, the RN would require greater force to meet these commitments - unless the latter were changed. This is precisely what some members of the DC(M) advocated; rather than increase expenditure to meet these commitments, they argued, Britain’s Far Eastern commitment should be abandoned in favour of dealing with its greater needs in Europe. Treasury officials led the way in advocating a formal change of policy which subordinated needs in the Far East in favour of building a force to deter Germany. To do so, a massive increase in the size of the RAF was proposed, which would be offset by a corresponding limitation on naval spending, while diplomatic measures would deal with the problem posed by Japan.\textsuperscript{15} In a paper presented to the DC(M), the Chancellor of the Exchequer, Neville Chamberlain proposed that the bulk of monies for deficiencies be given to the RAF,
while the Army be limited to £19,100,000 over five years [as opposed to £40,000,000 as provided by the DRC recommendations]. Most shocking of all, he proposed that all naval building be cancelled, a move which would save £8,000,000. In doing so, the amount spent would be reduced by nearly 30%, to £69,300,000.\textsuperscript{16} To many members of the DC(M), this represented too radical a change. In the words of the First Lord of the Admiralty, Bolton Eyres-Monsell, it meant "altering the whole basis of imperial policy and letting go the principal link which holds the Empire together."\textsuperscript{17} Fortunately for the Admiralty, the DC(M) did not endorse the new policy proposed by the Treasury. However, the cost of the deficiencies programme outlined in the DRC’s report was reduced from a total of more than £70 million to £50 million, not including the cost of new naval construction, with the Admiralty’s share reduced by almost 60% to £13 million.\textsuperscript{18} Thus, the navy was forced to pay a disproportionate share of the cost.
The effectiveness of the DRC in addressing the serious deficiencies in the services was limited by the DC(M)’s attempt to cut costs; its effectiveness was also limited due to the length of time which passed before the committee’s recommendations were finally acted upon. The DRC report was presented to Cabinet on 28 February 1934 and sent to the DC(M) in early May. The DC(M) debated the report for over three months before issuing its revised deficiencies programme. Once this revised programme was issued, the services were permitted to negotiate the approved expenditures with the Treasury as part of the formulation of the annual estimates for 1935. The new policy, and the deficiency programme, was not made
public until the government issued its White Paper concerning defence on 4 March 1935.\textsuperscript{19} Thus, Britain did not react to the situation of 1933 before 1935.
In the year between the DRC report and the White Paper, the global situation had become even worse, particularly with regard to the naval situation. By 1935, Britain faced the ugly prospect of three potentially hostile naval powers. The Japanese announced their intention not to renew the London Naval Treaty in 1936, while Italy’s designs on Abyssinia became more apparent, as did the threat to British interests in the Mediterranean.\textsuperscript{20} Germany launched its third "pocket" battleship in 1934 and, with the delivery of the British White Paper, Hitler abrogated the terms of the Versailles Treaty. No longer fettered by the terms of Versailles, Germany presented a growing threat in European waters.
While external threats grew, the lack of governmental response to the DRC’s recommendations meant Britain did nothing to strengthen its depleted forces. The impact of the DC(M)’s alterations to the deficiencies programme laid out in the DRC’s report was not as drastic as it could have been, but it did have a dramatic effect on British naval strength. In effect, Britain chose to sacrifice defence preparations against Japan, and to subordinate maritime to air requirements. While it remained official British policy to defend its commitments in the Far East, the resources to do so simply did not exist. While the British dallied, the Japanese modernized; the gap between what the British intended to do and what they were capable of doing was great in 1932. It became even wider by 1935. The DRC was created to eliminate deficiencies and ensure that, even if Britain did not rearm, her
defence forces would not fall any further behind those of other powers. However, the Cabinet produced that very situation by failing to act on the committee’s recommendations quickly and completely. Yet, in hindsight, 1933 to 1935 was the period when large scale rearmament had to begin in order to meet the dangers at hand. If immediate and major action was not taken, Britain would be condemned a few years later to pursue even larger and more expensive programmes which it could not satisfactorily complete. In 1934 the DRC put forth a programme which involved only one tenth the expenditure that was authorized three years later. As Stephen Roskill pointed out, the expenditure of 1937 proved to be insufficient; had the government accepted the recommendations of the DRC, perhaps the later rearmament efforts would have been more successful.\(^{21}\)
**The Defence Policy and Requirements Committee (1935)**
While the German naval threat to Britain was potentially serious, it was the expansion of its land and air forces that proved to be of immediate concern as the year progressed. In March 1935 Germany announced the beginning of conscription while Hitler claimed to have achieved air parity with the west. The situation was considered severe enough to require another reevaluation of defence policy. With disarmament no longer a real possibility, the DC(M) was replaced by a new body, the Defence Policy and Requirements Committee (DPRC). The DPRC revived the DRC in early July to consider how to revise defence programs to meet the growing German threat. The DRC delivered an interim report on 24 July 1935 which contained only two
recommendations. The first was that, due to a lack of armaments capacity, Britain could not reach a state of military preparedness sufficient to counter the German and Japanese threat until 1 January 1939. The second simply stated that, since a parliamentary recess was upcoming, the committee needed the authority to reach a decision on programmes so that they could be undertaken immediately. The priority was to be on achieving security as quickly as possible, with the financial considerations to be of secondary importance.\textsuperscript{22} The DRC’s interim report marked the turning point in the 1930s between eliminating military deficiencies and rearmament. Treasury demands for a balanced budget, fiscal responsibility, and the primacy of the economic health of the country were slowly giving way to the need to address the serious weaknesses of the three services.
On 21 November 1935 the DRC presented its third report which addressed the needs of the services in light of the new guidelines outlined in its interim report. For the Royal Navy, it recommended a fundamental change in policy. Since 1932, the accepted standard of naval power had been to ensure that the Royal Navy could maintain a deterrent force in the Far East capable of protecting British interests until the rest of the fleet could arrive, while simultaneously maintaining a deterrent force in British home waters. In its report the DRC recommended that the government accept the Admiralty’s proposal that there be a revival of a two-power standard; that is, the Royal Navy be capable of defending British interests in the Far East from Japanese aggression (as opposed to merely deterring it), while simultaneously keeping a force able to counter Germany in European waters.\textsuperscript{23} The new standard was based on the
actual defence of British interests, as opposed to merely deterring attack. While not as dramatic as the historical definition of the Two-Power Standard, the DRC’s recommendation reflected the acceptance of the global nature of British naval commitments and the imperative need to meet them, something which the first DRC report had not done.
Despite its acceptance of the new naval standard, the DRC’s first goal was to ensure that the Royal Navy could meet the requirements of the existing one. In other words, the construction programme laid down in the report was intended to eliminate deficiencies rather than to meet the demands of a two-power standard. As Roskill has argued, since the first DRC programme had failed to eliminate deficiencies, the second DRC programme had first to do so before it could work towards its true goal of expansion. In particular the condition of the armaments industry hampered the expansion of the Royal Navy. The report noted that, due to the lack of foreign orders and the small size of recent British contracts, the naval armaments industry was in a weakened state. Men and material had to be gathered before any large scale construction could begin. As a result, during the period 1936 to 1939, the Admiralty could simply plan for the replacement of over-age ships; only when industrial capacity had expanded could construction begin on the new vessels which were required to increase the overall strength of the navy.\(^{24}\) The only possible way to speed construction under existing conditions was through a change in the Admiralty’s ordering practices. In particular, by ordering vessels earlier in the fiscal year work could begin over the summer months which would cause fewer construction delays.\(^{25}\)
The DRC recommended that seven capital ships be laid down over the next three years; two in 1937, three in 1938, and two in 1939. This would give the navy a new ship strength equal to that of Germany and Japan combined, provided that neither power increased its rate of construction. The need for such a high rate of construction stemmed from the capital ship building holiday imposed by the Washington Agreement in 1921 and extended by the London Naval Treaty; after fifteen years of virtually no replacement there were several over-age vessels which had to be replaced in a short space of time. In the DRC’s view, construction after 1939 would be dictated by the actions of other powers. If Germany and Japan did not lay down more capital ships, Britain would have matched their combined strength by ordering just two more capital ships in 1940. After 1940, it might at best be possible to resume a regular replacement programme of one ship per year but, with no treaties regulating capital ship construction, this situation could become far worse.\textsuperscript{26}
In order to meet the deficiencies in the 1932 naval standard, the DRC accepted the Admiralty’s proposal for a replacement program of five cruisers per year. If the new standard was accepted, the committee noted, more new cruisers would have to be constructed and more over-age vessels retained in service. The DRC accepted the existing Admiralty plan to build four aircraft carriers over the period 1936 to 1942, but recommended an increase in the construction of destroyer flotillas and flotilla leaders. Should the total destroyer strength laid down under the London Naval Treaty prove sufficient, one destroyer flotilla would be laid down every second year between 1938 and 1942. If the new naval standard was adopted, however, more construction
would be needed.\textsuperscript{27}
As a result of the DRC's construction programme, by 1939 the British navy was expected to reach numerical equality on paper with the German and Japanese navies.
\textit{Table VI: Completed Ships (as of 31 March 1939)}\textsuperscript{28}
| Class of Ship | British Commonwealth | Japan | Germany |
|---------------------|----------------------|---------|------------------|
| Capital Ships | 3 modern | 2 modern| 2 new |
| | 3 modernized | 7 modernized| 3 \textit{Deutschlands} |
| | 9 non-modernized | 9 total | 5 total |
| | 15 total | | |
| Aircraft Carriers | 6 | 5 | 1 |
| 8" Cruisers | 15 | 12 | 3 |
| Large 6" Cruisers | 12 | 6 | 0 |
| Small 6" Cruisers | 28 | 18 | 6 |
| Destroyers | 144 | 93 | 38 |
| Submarines | 45 | 89 | 38 |
The DRC's plan for new construction was expected to cost £225,125,000 over the period 1936 to 1942.\textsuperscript{29} The cost of new construction would level off after 1939 if some form of naval limitation remained in effect, but it would rise should this not be the case. The cost of new construction would also be substantially higher should the government accept the DRC's recommendation of the new naval standard. While the cost of the program seemed high, the DRC was careful to point out that, in fact, the figure did not really represent a net increase in the cost of new construction. In an
average year, the Admiralty expected to spend approximately £16,500,000 on new construction; that figure had been lower since 1929 because of the holiday in capital ship construction, and because of those actions these warships now had to be replaced at a greater than usual rate.\textsuperscript{30}
While the greatest part of the expenditure recommended by the DRC was committed to new construction, the report also detailed a further expenditure of £29,000,000 to overcome other deficiencies. This included the modernization of capital ships, anti-submarine equipment, and reserves of ammunition. It was proposed that three capital ships be completely modernized and four others be partially modernized; only the three post Jutland warships \textit{Nelson}, \textit{Rodney}, and \textit{Hood} did not need such attention. If, the DRC noted, no agreements limited the age and number of capital ships, and if other powers retained their overage vessels, then other British capital ships would have to be modernized in order to maintain its naval strength.\textsuperscript{31} In any case, due to the increased number of ships, personnel levels would have to be increased. The DRC recommended an increase in naval personnel of 35,600 men, or approximately 30%, over the period 1936 to 1942; the committee members felt that even more personnel would be necessary, but the Admiralty refused to accept a larger increase lest it affect efficiency.\textsuperscript{32}
Thus, the DRC’s recommendations for naval expenditure were generous and it largely accepted the Admiralty’s proposals on major policy issues. In particular, it wholeheartedly approved the new standard of naval strength, which represents a marked departure from the DRC’s first report. When the report was presented to the
DPRC, most of the recommendations for the Royal Navy were accepted without question; its only concern was the new standard. The DPRC feared the lack of skilled labour and material would make it difficult to meet the requirements of the existing naval standard, while the requirements of the new standard had not been worked out in detail. The ministers refused to accept a new standard without knowing the sort of fleet it would entail.\textsuperscript{33}
While considering the DRC’s report, moreover, the DPRC also looked closely at the difficulties of meeting the requirements of the three services. Specifically, it focused on the rationalization of industrial production so as to ensure that the needs of the services could be met. The DPRC knew that Britain could not overcome the result of fifteen years of neglect in three years. Thus, its goal was to produce as much as possible in the short time available, without disrupting the British economy. Two tasks had to be achieved in order to do this. First, capacity had to be expanded to meet the greatly increased demand; secondly, reserve capacity had to be developed for use after the outbreak of war. As British armaments firms simply could not meet the demands of the services in wartime, this meant that shadow capacity would have to be developed in other sectors of the economy. While the Principle Supply Officers Committee had done some work in this area, more remained, and the increased demands of rearmament meant that part of the shadow capacity would have to be utilized in peace.\textsuperscript{34} Shadow capacity, in which non-armaments firms were provided with the plant and the training to produce military goods, was conceived of as a way to expand the country’s production in a time of war without diverting capacity from
civilian production peace. That the DPRC considered using shadow capacity in peacetime, despite the government’s policy of ‘business as usual’, illustrates the lack of British armament capacity.
The third DRC report and the DPRC recommendations which arose from it reflect the conflict between intent and effect. The DRC had been formed in 1933 to address deficiencies, but the government had failed to heed its recommendations. In 1935 the DRC’s intent changed from deficiencies to rearmament, but rearmament could not proceed until the services and their allied industry first overcame the deficiencies which the government had failed to address in 1933. In the case of the Royal Navy, the third report of the DRC was dominated by measures which would just barely allow it to match the existing naval standard. While it was recognized that this existing standard was inadequate and approved the idea of a new standard, it realized the achievement of a new standard was impossible in the limited time available. Shipbuilding capacity in 1935 was half that of 1930, and firms could barely handle the measures outlined in the DRC’s third report. Despite this, the Admiralty chose to capitalize on the turn of the political tide in order to pursue even more ambitious programmes. It would succeed in doing so, but its political and policy success would actually hamper the process of naval rearmament.
**Rations, Feast, and Famine**
By 1935-36, Treasury control over the services’ estimates and policies was rapidly waning. The Treasury was particularly concerned with the Admiralty’s
estimates, which had risen by over 70% between 1935 and 1938 in support of a policy to which the Treasury objected.\textsuperscript{35} By 1937, Britain could no longer meet the growing expense of rearmament out of revenue, and the Defence Loans Bill was passed in March 1937. The bill provided for the establishment of a consolidated fund of £400,000,000 limited to five years, ending 31 March 1942.\textsuperscript{36} While the services regarded this move as a positive one, the Treasury was determined to minimise the amount of money borrowed for defence expenditure; it feared the long-term costs of borrowing such large sums of money. Given the fragile state of the British economy at the time, there was much to be said for this position. The Treasury was also concerned that the services were trying to do too much in a limited span of time, and thus were hurting themselves in the long run, and again its views had some force.\textsuperscript{37} In a memo to the Board of Admiralty, the Treasury pointed out that a huge and sudden increase in demand was bound to cause a corresponding rise in prices, and it might push the naval armaments industry back into the same difficult experience after World War I, when a massive slump had followed a building boom.\textsuperscript{38} In effect the Treasury used the Admiralty’s own argument of consistent construction programmes against it. Above all, the Treasury feared that rearmament might leave Britain with armed forces that it could not afford to maintain.\textsuperscript{39}
Almost immediately after the Defence Loan was announced, the Treasury attempted to deal with these problems by rationing defence expenditure. It argued that, since it was impossible to completely meet Britain’s defensive needs in short term, defence spending had to be prioritized. Since Germany presented the greatest
direct threat to Great Britain, building a deterrent force against German aggression was given top priority, well above Japan; as the RAF was the only force believed capable of "taking the fight to the enemy", it was given first claim on spending among the services. The Treasury wished to relegate the Navy to second and the Army to third priority, to share whatever monies were left after the RAF had received what it needed. The Treasury was not isolated in its opinion, for it reflected the conventional thinking among ministers. In a memo to Cabinet in October 1937 even Hankey, a friend of the Navy, ranked the needs of the services in the following order: the RAF and defence from air attack, the RN and the defence of naval bases, and, finally, the Army.\textsuperscript{80} In order to meet defence needs, the Treasury wished to set a limit of £1,500,000,000 to service spending over the following five years. Under the scheme, the programs already authorized by the Treasury would continue, but the services would provide the Treasury with detailed sketches of the rest of their programs, their estimated cost of construction, and the cost of maintenance. The Treasury itself would decide how much the country could afford, and the DPRC would decide which programmes would have to be cancelled.\textsuperscript{81}
The Admiralty openly opposed this attempted reassertion of the Treasury's control over spending, but it was not greatly affected by the new scheme. While the RAF and the Treasury had powerful political support, so did the Navy. Rationing and the increased concerns over the total cost of rearmament certainly prevented the Cabinet from authorizing the Admiralty's new Naval Standard and the new standard fleet, but this did not prevent the Admiralty from pursuing it. The result of the
Treasury’s counter-attack of 1936 was identical to those of the mid 1920s. Service spending was cut, but far less than the Treasury demanded, and the Army bore the brunt of the economies.\textsuperscript{42}
\textit{New Standards of Naval Power}
The rearmament period was one of remarkable political success for the Admiralty. The first DRC report supported the navy’s needs, although its recommendations were largely ignored by the Cabinet. With the third DRC report, and the favourable response from the DPRC in 1935, the Admiralty’s fortunes had definitely changed for the better. While the DRC essentially provided for the reconstruction of fleet only up to the 1932 naval standard, it also proposed that Cabinet consider the Admiralty’s new standard. Encouraged, the Admiralty drew up plans for the fleet necessary to meet the new standard, which it revised to include the protection of trade. Despite some misgivings by the government and the Treasury, the Admiralty received massive estimates and directed its policy and estimates to meet an ambitious and unauthorized standard.
This process began in 1932 when the Ten Year Rule was abolished, and the plans division of the Royal Navy began to investigate its readiness for war and to review its future policy. Before the DRC was established, the First Sea Lord, Admiral Frederick Field, ordered the Naval Staff to examine the cost of meeting existing deficiencies and the program of new construction over the period 1932 to 1942. Although the inquiry was not approved by the Admiralty board, it became the
basis of the Admiralty’s discussions with the DRC the following year.
It was with the discussion by the DPRC in 1935 that the new direction of naval policy really began to take shape. The obvious weakness of the naval standard of 1932 was that Britain could in effect handle only one enemy at a time. Thus, if the British were forced to meet a Japanese threat in the East, home waters would be left only weakly defended and vice versa. The expansion of the French and Italian navies, as well as the potential threat posed by Germany, made this an unattractive scenario. The growth of other naval powers around the globe and the London treaty had eroded the comfortable British superiority which had existed after the Washington treaty. As a result, for planning purposes the Admiralty considered a “Two-Power Standard”, which would enable Britain to simultaneously send a fleet to the Far East to counter Japan while keeping a force in Home Waters large enough to match any European enemy (specifically Germany). It was this standard that the Admiralty proposed to the DRC in 1935.
It must be noted that the concept of a naval standard was more a political tool than a concrete strategic concept. It is less an indication of British naval strength than how Britain perceived its enemies. A change from a One-Power Standard to a Two-Power Standard does not indicate a change in British strength, but a change in the scale of maritime dangers and the definition of which states Britain felt most threatened by. The One-Power Standard which was the basis of naval planning between 1919 and 1932 had different meanings attached to it at different times. For example, in 1919-20 the One-Power Standard was applied against the United States.
While the British were concerned about the growth of the Imperial Japanese Navy, the American desire to achieve naval parity by 1925 was considered the greater threat. However, to directly challenge the United States on the issue could either cause a war or give the Americans reason to use political means, such as recalling Britain’s outstanding war debts, to solve the matter. Despite their claims, the Americans would not build a navy equal to Britain’s even if they wanted the right to, so long as the British did not force the United States’ hand. Simply by not considering the Americans a threat the British could maintain real naval security while abandoning the symbol of naval supremacy.\textsuperscript{45} By late 1920 the practical implications were even less clear than the standard itself. Was the One-Power Standard, as interpreted against the United States, to require a fleet larger or smaller than the American fleet? Equal to the American navy only in United States’ waters or the entire American navy plus any other two navies, as the Admiralty wanted?\textsuperscript{46}
While the DRC was more kindly disposed towards the Royal Navy in 1935 than in 1934, the ministerial committee which reviewed its report, the DPRC, was not fully convinced of the need for a new standard of naval power. It recognised that only rearmament could meet the threat from Germany and Japan, but the committee also recognized the limits of British industry. Rearming to a new standard by 1939 was beyond the capability of industry and therefore not of immediate concern; moreover, the Admiralty had not precisely defined what the new standard would involve. The DPRC was not willing to accept the new standard without knowing the cost and size of the fleet which it would produce, and with good reason. In the loose
form proposed by the Admiralty, the standard could be interpreted many ways, from merely meeting the 1932 standard on paper, to a standard which would require a fleet able to meet the Japanese and German fleets at a time and place of those powers' choosing. The fleet to fulfill such a standard would have to be significantly larger than the combined strengths of the German and Japanese navies. However, the DPRC and especially the DRC did encourage the Admiralty to present a more precise version of this concept for future consideration. Asking the Admiralty to define a programme for expansion was guaranteed to create a split between the Admiralty, the Treasury and the government, which was exactly what happened as each side developed its own idea of where naval policy should head. Nor was this the only source of confusion. For the next four years the Admiralty maintained two separate forecasts of estimates and ship construction.
In its proposals to the DRC, the Admiralty had outlined a two power standard in which the navy could send a fleet to the Far East to counter a Japanese threat while simultaneously keeping a force in home waters capable of fighting a war with Germany. Essentially, this was a similar standard to the One-Power Standard of the mid 1920s. In its report to the Cabinet, the DPRC recommended that the Admiralty be encouraged to work out the fleet requirements of such a standard so that the matter could be considered more carefully.\textsuperscript{47} Rather than merely defining the requirements of the proposed standard, the Admiralty first raised the scale of the standard. To the already ambitious requirements of meeting a simultaneous threat from Japan and Germany, the Admiralty added another stipulation - that the Royal Navy be able to
protect British maritime trade at all times, anywhere in the world, from a threat by anyone.\textsuperscript{48} The Admiralty advocated this new "New Standard" in order to turn the proposed Two-Power Standard into a Three-Power Standard. The French and the Italians had not been party to the London Naval Agreement (1930) and their fleets presented a possible threat to British naval power.\textsuperscript{49} By 1935-36, much to its dismay, the Admiralty was forced to treat Italy as a potential enemy. By incorporating the protection of trade within the new naval standard, the Admiralty could respond to the growing threat from Italy by claiming that it endangered British maritime trade in the Mediterranean Sea. Additionally, the Admiralty could better justify its requests for a larger cruiser fleet, the central point of contention in naval policy for a decade, by referring to the protection of trade.
The naval construction set out in the DRC's third report and accepted by the DPRC came to be known in Admiralty documents as the DRC Fleet. At the same time, the Admiralty devised plans for a fleet based on the 1937 Naval Standard, which it called Fleet A, or the New Standard Fleet. The key features of this fleet included:\textsuperscript{50}
\begin{tabular}{ll}
Capital Ships & 20 \\
Aircraft Carriers & 15 (including three on long notice reserve) \\
Cruisers & 100 \\
Destroyer Flotillas & 22 \\
Submarines & 73 \\
\end{tabular}
Fleet A represented a major increase in the size of the Royal Navy, with five more capital ships, nine more aircraft carriers, and forty-five more cruisers than the DRC fleet, with a corresponding increase in cost. For example, between 1939 and 1941 it would require the laying down of six capital ships, four aircraft carriers, seventeen
cruisers, three destroyer flotillas, and twenty submarines.\textsuperscript{51} The difference in cost, if the new naval standard was approved, is illustrated in the table below.
\textit{Table VII: Forecasted Annual Estimates, 1937 to 1946, DRC Fleet vs. Fleet A}\textsuperscript{52}
| Year | DRC Fleet | Fleet A |
|------|---------------|---------------|
| 1937 | £105,065,000 | £105,065,000 |
| 1938 | £122,000,000 | £124,500,000 |
| 1939 | £123,000,000 | £131,000,000 |
| 1940 | £103,500,000 | £123,500,000 |
| 1941 | £81,000,000 | £115,500,000 |
| 1942 | £77,500,000 | £111,500,000 |
| 1943 | £79,500,000 | £111,000,000 |
| 1944 | £88,000,000 | £112,000,000 |
| 1945 | £91,000,000 | £107,500,000 |
| 1946 | £86,000,000 | £98,000,000 |
While expenses were comparable in the first few years, expenditure on the DRC fleet stabilised in 1940; unstabilised expenditure, that is, the yearly expenditure including the cost of new construction on Fleet A would continue until at least 1944. Further, once construction costs were complete, the stabilised cost of maintenance was considerably higher for Fleet A: £103,500,000 as compared to £88,000,000.\textsuperscript{53} There were ways to reduce this expenditure, however; the Admiralty proposed extending the lifespan of certain classes of ships, cutting expenditure on maintenance of Fleet Air Arm aircraft, and increasing reliance on reserve manpower.\textsuperscript{54} Despite these measures, there was no avoiding the substantial increase in expenditure which would result if Fleet A was approved.
By 1 February 1938, the Admiralty estimated that the aggregate cost of the annual estimates on the Fleet A basis for the years 1939 to 1941 had risen from £349,000,000 to £403,000,000 due to increases in the cost of labour and materials and the addition of major new items.\textsuperscript{55} By August 1938, the figure had risen to £443,013,000. This increase in cost of 25% in twelve months finally caused the Cabinet to flinch. Though the government was still willing to discuss the new naval standard and Fleet A, the Admiralty was told that naval estimates could not exceed £410,000,000 over the years 1939 to 1941. This figure would be increased or decreased depending on the over- or under-expenditure in 1938 Estimates.\textsuperscript{56} In a sense, this was a tacit refusal by the government to accept the Admiralty’s new fleet; if the moderate increase in expenditure over the DRC fleet during the first few years of the proposed program was unacceptable, it was unlikely the government would be willing to absorb the high costs of Fleet A which would continue after the DRC program was scheduled to end. In response to the Cabinet’s concerns, the Admiralty reduced the cost of Fleet A from £443,013,000 to £417,757,000 by revising the cost of new construction. It noted that further reductions were possible, but not advisable; these measures included reducing the number of ships involved in the anti-aircraft rearmament programme, eliminating the plan to prepare underground fuel storage facilities, and not constructing a second cordite factory. However, if these reductions were made, the figure could be reduced to £410,287,000.\textsuperscript{57} The rapid growth of the proposed cost of Fleet A would likely have doomed it to failure had the war not broken out, but in the interim it gave the Admiralty a valuable bargaining tool.
Whenever the Treasury demanded that estimates be cut the Admiralty used Fleet A as a yardstick; as it was considerably more costly than the DRC fleet, the latter seemed an equitable compromise. While the new naval standard was not approved, it was used by the Admiralty to justify its estimates with success.
The appropriations accounts, which indicate what the Admiralty actually spent in a given year, show how successful the Admiralty was in achieving its demands for increased expenditure in the late 1930's.
**Table VIII: Naval Estimates and Appropriations, 1937 to 1940**
| Year | Gross Estimate | Gross Appropriation | Effective Estimate | Effective Appropriation |
|------|----------------|---------------------|--------------------|------------------------|
| 1937 | 108,016,387 | 105,581,046 | 98,026,658 | 95,680,561 |
| 1938 | 131,946,619 | 131,473,378 | 121,947,868 | 121,519,501 |
| 1939 | 153,666,781 | 181,770,565 | 143,524,989 | 172,263,428 |
| 1940 | —— | 384,162,378 | ——— | 375,206,130 |
In 1937, the Admiralty received approximately the same amount it required to meet either the DRC or Fleet A proposals. However, in 1938 and 1939, the Admiralty received considerably more money than even proposed in the Fleet A scheme; in 1938 it was granted £6,973,378 more than it requested under the Fleet A scheme, and in 1939, the difference was more than £60,000,000. The size of the 1939 increase was largely due to the worsening political situation and the eventual outbreak of the war which prevented the rationing of expenditure. Nevertheless, it is obvious that the Admiralty was still successful in getting the monies it wanted. The Treasury’s efforts at rationing the services did not hamper the Royal Navy.
Throughout the 1930s, the more the Admiralty obtained what it wanted, the more it demanded. Once the government accepted the provisions of the DRC’s third report, rearmament to meet the 1932 naval standard, the Admiralty immediately devised another standard with greater scope which required a correspondingly larger fleet. When asked to tell the Cabinet what was entailed in a two-power standard, the Admiralty responded by turning it into a three-power standard. In its proposals to the first and second DRC inquiries, the Admiralty outlined a fleet which, it argued, could meet the Japanese and German navies. By 1938, the Admiralty maintained that the DRC fleet would defend vital British interests only should relations with the Japanese and Germans improve. It justified the new standard fleet to the Treasury by explaining that better relations with Japan would only slightly reduce the cost of the fleet because, for reasons of security, it was better to decrease readiness by keeping a portion of the Reserve Fleet on extended notice.\(^{59}\) In other words, the standing size of the fleet would be reduced but, by keeping more ships in reserve rather than scrapping them, very little money would be saved. In concrete terms, the fleet, should relations improve with Japan, would require:\(^{60}\)
| Type | Number |
|-----------------------|--------|
| Capital Ships | 20 |
| Aircraft Carriers | 13 |
| Cruisers | 83 |
| Destroyer Flotillas | 16 |
| Submarines | 73 |
Improved relations with Germany would have a greater effect on fleet requirements. In that case, the DRC fleet would roughly suffice; this fleet would include:\(^{61}\)
Capital Ships 20 (some maintained by the Dominions)
Thus, by 1938, the Admiralty had completely revised its requirements and claimed that only the New Standard Fleet could achieve the security the DRC fleet was supposed to provide. While this changed position can be explained in part by the growth of a number of possibly hostile naval powers, it also reflects the Admiralty’s attempts to push through its demands at a time when it had little opposition from other governmental departments. It intended to establish under the present favourable circumstances a policy which would strengthen the Royal Navy for decades to come. It was not thinking or preparing simply for a war against Germany around 1939-42, but for the defence of British interests at sea through the 1950s.
The Admiralty was remarkably successful in achieving its goals during this period, but its victories were in the sphere of politics rather than policy. The DRC and the DPRC were open to the Admiralty’s suggestions, and they supported its requests with regard to rearmament. Not only did the Admiralty get its entire rearmament programme approved, it also gained considerable support for its new standard of naval strength and the fleet that was needed to achieve it. While the new standard and its fleet were never approved, the Admiralty was successful in getting very large estimates accepted, and the Treasury’s attempts at rationing were not particularly painful. In some ways, the Admiralty’s political success was counterproductive. It masked a pattern of famine followed by feast in its dealings with arms.
firms. Shipbuilding capacity could barely handle the requirements of the DRC fleet; those of Fleet A were beyond it. The Admiralty’s pursuit of two separate standards undoubtedly helped to increase the size of naval estimates, but it also made coherent planning difficult. However, getting programmes approved was not the same as getting them constructed. Despite having the largest naval shipbuilding capacity in the world at its disposal, the Admiralty did not use it efficiently. The Admiralty’s successes were of greater significance in Whitehall than at sea.
Notes
1. J.R. Ferris, *Men, Money and Diplomacy* (Ithaca: Cornell U.P., 1989.), p.29.
2. Robert Shay, *British Rearmament in the Thirties* (Princeton: Princeton U.P.), pp. 22-23.
3. Shay, p. 23.
4. Shay, p. 24.
5. Shay, p.27-28.
6. N.H. Gibbs, *Rearmament Policy* (London: HMSO, 1976) p. 94.
7. Gibbs, p. 94.
8. Roskill, *Naval Policy* v.II, p. 168.
9. Roskill, *Naval Policy*, v. II, 168.
10. Roskill, *Naval Policy* v. II, pp. 168-69.
11. Roskill, *Naval Policy Between the Wars*, v. II, (Annapolis: Naval Institute Press, 1976.), p. 169.
12. Shay, p. 33.
13. Gibbs, pp. 96-7.
14. G.A.H Gordon, *British Seapower and Procurement between the Wars* (Annapolis: Naval Institute Press), p. 124.
15. Gordon, p. 124.
16. Shay, p. 39.
17. Gordon, p. 125.
18. Gordon, p. 125.
19. Gordon, pp. 124-25.
20. Gordon, p. 126.
21. Roskill, *Naval Policy*, v.II, p. 172.
22. Stephen Roskill, *Hankey, v. III* (London: Collins, 1974), p. 173.
23. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259
24. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259.
25. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259.
26. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259
27. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259
28. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259
29. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259
30. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259
31. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259
32. D.R.C. 37. Committee of Imperial Defence. Defence Requirements Subcommittee. Third Report, 21 November, 1935. CAB 24/259
33. C.P. 26 (36) "Programmes of the Defence Services" 12 February, 1936. CAB 24/259.
34. C.P. 26(36) Programmes of the Defence Services. 12 February, 1936. CAB 24/259.
35. Gordon, p.176.
36. Defence Loans Bill. T 175/96.
37. Gordon, p. 176.
38. Board Minutes, 1937. Friday, 12 February, 1937. ADM 167/96.
39. Shay, p. 162.
40. Roskill, *Hankey*, v.III pp. 290-91.
41. Shay, pp. 162-63.
42. Gordon, p. 177.
43. Gibbs, pp. 117-18.
44. J.R. Ferris, "The Symbol and Substance of Seapower: Great Britain, the United States and the One-Power Standard, 1919-1921" in *Anglo-American Relations in the 1920s* B.J.C. McKercher, ed. (Edmonton: U of A Press, 1990, pp. 55-80.), pp. 57-58.
45. Ferris, *Men, Money and Diplomacy*, pp. 54-57.
46. ibid, p. 79.
47. C.P. 26 (36) Programmes of the Defence Services. 2 February, 1936. CAB 24/259.
48. Forecast of Navy Estimates, DRC Fleet and Fleet A: Draft Memo on a New Standard of Naval Strength. ADM 116/4433, v. 1.
49. Forecast of Navy Estimates, DRC Fleet and Fleet A: Draft Memo on a New Standard of Naval Strength. ADM 116/4433, v. 1.
50. Forecast of Navy Estimates, DRC Fleet and Fleet A: Draft Memo on a New Standard of Naval Strength. ADM 116/4433, v. 1.
51. Revised Forecast of Navy Estimates, 1939-41. Forecast of Naval Expenditure. Naval Defence Estimates and Expenditure. ADM 116/4434, v. 2.
52. Navy Defence Estimates & Expenditure. Analysis of Requirements, DRC vs. Fleet A. ADM 116/4433.
53. Navy Defence Estimates & Expenditure. Analysis of Requirements, DRC vs. Fleet A. ADM 116/4433.
54. Navy Defence Estimates & Expenditure. Defence Expenditure. ADM 116/4434 v.2.
55. Navy Defence Estimates & Expenditure. Forecast of Naval Expenditure in 1939, 1940, and 1941. ADM 116/4434, v. 2.
56. Navy Defence Estimates & Expenditure. Naval Expenditure 1931-1941. August, 1938. ADM 116/4434, v. 2.
57. Navy Defence Estimates & Expenditure. Naval Expenditure 1931-1941. August, 1938. ADM 116/4434, v. 2.
58. Appropriation Account, Navy Services, 1937. T 161/904/S34880/37, Appropriation Account, Navy Services, 1937. T 161/904/S34880/37, Appropriation Account, Navy Services, 1939. T 161/1627/S34880/39, Appropriation Account, Navy Services, 1940. Great Britain. Parliamentary Papers, 1940-41.
59. New Standard of Naval Strength. Board Memo on a New Standard of Naval Strength. T 161/780/S.42000.
60. Forecast of Navy Estimates, DRC Fleet and Fleet A: Draft Memo on a New Standard of Naval Strength. ADM116/4433, v. 1.
61. New Standard of Naval Strength. Board Memo on a New Standard of Naval Strength. T 161/780/S.42000.
Chapter Three: Vickers Armstrong and Rearmament
One might have expected the Admiralty’s political gains during the late 1930’s to have been a blessing for arms firms and a means to restore the relationship between capacity, construction, and security which had existed in the 1920s. This was not the case. The Admiralty’s political victories were not translated into concrete achievements regarding rearmament. While the Admiralty succeeded in getting bigger naval estimates, its ordering practices meant the money was not used effectively. An examination of Britain’s premier naval armaments firm, Vickers Armstrong [Vickers], illustrates this. Indeed, the firm’s perspective is essential for a full understanding of the failure of British rearmament. After suffering through years of a weak naval armaments market, the rapid increase in naval construction should have been a great boost to Vickers. In anticipation of precisely these circumstances, the firm had spent a great deal of money maintaining its naval yards and thought itself ready for rearmament. At Barrow alone Vickers spent nearly £100,000 per year to keep its plant and equipment up to date.\(^1\) However, the relationship between the Admiralty and the firm were not smooth during this period. Until 1935, the firm’s naval armaments capacity was barely utilized at all due to a lack of orders, as one might expect; after 1935, much of the firm’s capacity still remained idle due to a lack of orders. There were two principal areas of friction between the Admiralty and Vickers during the period: the inefficient use of Vickers’ capacity and the Admiralty’s attempts by to manipulate the firm into focusing its capacity on the production of one
of the most specialized forms of armaments, gun mountings. The attempts by the government to limit Vickers’ foreign sales was also a cause of frustration. Vickers did its best to ensure that the boom years would not compromise its long-term security. It believed that the Admiralty’s policies went against the best interests of the firm.
To understand the relationship between Vickers and the Admiralty, it is necessary to consider the Admiralty’s procurement policy. As only a small part of new construction was built in the Royal Dockyards, which concentrated on repair and maintenance, most new warships were built by contract in private shipyards. The Admiralty was responsible for design, ordering, and inspection of completed work. Firms were responsible for placing sub-contracts, purchasing materials, and managing labour. Where possible, competitive tender was used. However, the Admiralty took some care to ensure orders were placed to benefit the firms and the Navy by preserving capacity.\(^2\) This was particularly true with shipbuilding, for "the placing of orders for warships could not without courting disaster be made blindly subject to competitive tender, the good of the industry as a whole had to be considered."\(^3\) It appears that the different works at Vickers submitted tenders on the same items, so Barrow and the Naval Yard might compete for shipbuilding orders, while Elswick and Barrow might compete for the same gun mounting orders; the Vickers works competed with each other as well as with other firms for contracts. It was the decision of the Admiralty as to which firm or plant would get the contract, based on factors such as price, length of time to completion, preservation of capacity and the
Capacity to Spare
Vickers was a large firm with a number of different works and subsidiaries. However, its naval armaments production was concentrated at three sites: Barrow, in northwestern England, Elswick and the Naval Yard, both near Newcastle. Barrow was the largest of the three, with capacity for shipbuilding and other naval armaments. Elswick and the Naval Yard were smaller, specialized operations; Elswick focused on gun mountings and gun production and the Naval Yard concentrated on naval shipbuilding. In analyzing the capacity used and available in the three key Vickers naval works, two distinct periods are apparent. The first, ending in December 1935, is characterized by consistent under-use, frequently well below 50% of capacity. In particular, a complete lack of orders resulted in the closure of the Naval Yard until the second quarter of 1936. This period witnessed a gradual increase in the percentage of capacity used, but figures remained well below full production. The second period begins in the first quarter of 1936 and is characterized by a sudden rise to almost full use of the capacity in all departments in all three yards. However, this sudden rise in aggregate productivity masks two more serious trends: the problems caused by the inefficient order procedures of the Admiralty and its attempts to force Vickers to specialize in certain areas of construction, specifically the production of gun mountings. While virtually every department was working at full capacity, this happened only because under-used departments took on part of the workload of other
departments, which in turn received more orders than they could handle. This problem was most apparent at the Barrow yard, probably due to its full range of armament production.
Between 1933 and 1935 the feast and famine ordering cycle was already apparent and its impact is apparent when one studies the firms Quarterly Reports over the period. These reports provide an in-depth analysis of the operations of each of the Vickers’ yards; they were prepared by the yard administrators for the firm’s head office and shareholders. In 1933 the Barrow yard received orders for one submarine and two destroyers, while work proceeded on a cruiser and a submarine ordered by the Admiralty the previous year.\(^4\) The Admiralty orders to Barrow in 1934 included two destroyers, one minelaying sub,\(^5\) and sub-contracts for machinery, boilers, and gun mountings.\(^6\) With the increase in work from the Admiralty, the Naval Yard was reopened; in the second quarter of 1934 it received an order for a cruiser hull\(^7\) and later that same year another British cruiser was laid down while preliminary work began on orders for two destroyers and one cruiser.\(^8\) Orders for 1935 at Barrow included a minelaying sub in the first quarter,\(^9\) a provisional order for two destroyers in the final quarter,\(^10\) and a large quantity of assorted gun mountings. On the other hand, while many orders were placed with the Naval Yard in 1934, none were given in 1935.\(^11\) The inconsistent order patterns are reflected in Vickers’ sales figures. In 1933 Vickers’ total shipbuilding sales in all yards, excluding merchant vessels, were £837,637; naval armaments sales totalled £2,093,142.\(^12\) Vickers’ total shipbuilding sales for 1934 were £1,599,030, while total naval armaments sales were £2,909,769.\(^13\)
Vickers’ total shipbuilding sales for 1935 were £2,011,897, a figure inflated by a huge jump in foreign orders, and naval armaments sales totalled £2,046,560.\(^{14}\) The clump of shipbuilding orders in 1934, and the lack of orders in 1935, caused disruption in the supply of men and material. After expanding rapidly and experiencing inefficiencies in the process, Vickers was forced to lay off employees as work fell off. The tables below illustrate the utilization of capacity at the Barrow and Elswick works, as well as the impact of Vickers’ foreign armaments sales.
**Table IX: Average Capacity Used, Barrow Works, 1933 to 1935\(^{15}\)**
| Department | 1933 | 1934 | 1935 |
|------------------|------|------|------|
| Shipyard | 40% | 64% | 67% |
| Gun Mountings | 72% | 69% | 83% |
| Marine Engines | 44% | 63% | 73% |
| Sub Engines | 56% | 71% | 45% |
| Foundries | ---- | 92% | 99% |
**Table X: Average Capacity Used, Elswick Works, 1933 to 1935\(^{16}\)**
| Department | 1933 | 1934 | 1935 |
|------------|------|------|------|
| Gun Mountings | 47% | 56% | 66% |
| Guns | 29% | 28% | 34% |
Table XI: British & Foreign Naval Armaments Sales, 1933 to 1935
| Quarter Ending | British Orders | Foreign Orders |
|----------------------|------------------|----------------|
| 30 June 1933 | £685,066 | £59,793 |
| 30 September 1933 | £274,405 | £9,507 |
| 31 December 1933 | £1,223,528 | £15,931 |
| 31 March 1934 | £1,011,091 | £880,738 |
| 30 June 1934 | £623,899 | £29,836 |
| 31 December 1934 | £3,187,074 | £465,423 |
| 30 June 1935 | £1,503,960 | £68,206 |
| 31 December 1935 | £2,417,458 | £1,574,741 |
These tabulations reveal four interesting features of Vickers’ production during the period 1933 to 1935. The first is a notable and continual rise from a very low base of shipbuilding sales caused by a sharp rise in British naval orders. Secondly, the sharp variation quarter-by-quarter in Vickers’ sales to foreign powers, which sometimes were insignificant, and on other occasions, critical. Thirdly, although the percentage of capacity used at Barrow and Elswick gradually increased, it was consistently under-used. Only the foundries at Barrow regularly operated at nearly full capacity. Finally, the Admiralty was increasingly funnelling its orders for gun mountings to Vickers, for completion at the Barrow or Elswick works.
The situation at Elswick is particularly interesting – here the percentage in capacity used for gun mountings increased nearly four times over that for gun production. Vickers received these orders because it was the only firm which could
produce gun mountings for the heaviest naval guns, although Beardmores had retained some heavy gun mounting capacity\textsuperscript{18}. While this increase in orders was logical, it also caused some concern among company officials. The production of gun mountings was a very specialised area, requiring plant which was not easily converted to other types of production. This gave Vickers increased business during rearmament and almost a monopoly in the area, but at the risk of having to support capacity which was of little use during periods of low naval construction.
Finally, the case of Vickers’ Naval Yard illustrated the erratic nature of Admiralty ordering practices. The yard was closed until 1934 due to a lack of orders. It was only reopened when a substantial order was placed by the Admiralty, and then the following year it did not receive a single new order while many contracts were lost to other firms.
The Admiralty’s “feast or famine” approach to naval construction caused difficulties for all firms; rather than being able to keep a stable number of employees continually employed in all trades, firms were forced to hire and then release the members of various trades as work progressed on all orders simultaneously. This forced workmen to travel to various yards as they were required which, coupled with the migration of skilled tradesmen to other industries, produced a serious shortage of labour by late 1935.\textsuperscript{19} While the practice might have made sense from the Admiralty’s perspective, because it kept the maximum number of firms operating in the armaments industry, it caused the firms much difficulty.
The sudden and ever rising increase in orders that occurred from 1936 onward
had a disruptive effect on Vickers’ naval yards. During this period, the difficulties caused by the Admiralty’s ordering practices are most evident; production was slow due to a lack of skilled labour and a steady supply of material. Moreover, the Admiralty’s concentration of orders for mountings with Vickers is apparent when one considers the sales figures over the period, as is the decision not to increase orders for warships.
**Table XII: Summary of Naval Armaments and Shipbuilding Sales, 1936 - 1939**
| Year | Naval Shipbuilding | Naval Armaments |
|------|--------------------|-----------------|
| 1936 | £2,490,202 | £2,927,922 |
| 1937 | £1,991,680 | £3,822,129 |
| 1938 | £2,726,767 | £4,131,632 |
| 1939 | £3,237,033 | £7,138,751 |
The halting process of these orders is clearly exemplified in the case of the main Vickers shipyard, Barrow, and the Naval Yard. After a decline in shipbuilding orders after 1933-34, the Barrow works received a large number of contracts in 1936. In the first quarter of the year, it was given contracts for two destroyers and one submarine as well as a large order for mountings and guns. The Naval Yard, however, received no orders at all; Vickers itself decided to transfer the destroyers from Barrow to the Naval Yard. In the next quarter Barrow received an additional order for mountings and guns, while the Naval Yard received one for two destroyers, its first in over a year. In the third quarter of 1936, Barrow received an order for a battleship, and Vickers chose to build the hull at the Naval Yard to keep the yard
running.\textsuperscript{23} Finally, at the end of the year, Barrow received orders for an aircraft carrier, four submarines, and a large number of mountings and guns. Once again, the firm transferred part of the work to the Naval Yard.\textsuperscript{24} During 1936, most departments at Barrow were working at 100% of capacity, but only because some capacity was converted to mounting and gun production. Despite the increase in shipbuilding orders, the Barrow shipbuilding department used an average of 76% capacity.\textsuperscript{25}
Essentially, Vickers was forced to rationalize the Admiralty’s orders on its own and, by doing so, unwillingly followed the Admiralty’s policy of concentrating gun mounting orders at Barrow, while placing contracts with another Vickers plant which the Admiralty chose to ignore. The Naval Yard, a small yard specializing in naval shipbuilding, was reopened in mid 1934 when Admiralty orders increased. Rather than permit it to close again once orders dropped off, and thus face the extra costs required should further orders force the yard to reopen, Vickers transferred some of the Admiralty’s shipbuilding orders to the Naval Yard.
After a heavy year in 1936, the contracts given to Vickers declined in 1937. Despite having received orders for only two destroyers in two years, the Naval Yard did not receive a single contract in 1937. Barrow was given an order for mountings in the second quarter of 1937\textsuperscript{26}, for an aircraft carrier and a submarine in the third quarter\textsuperscript{27}, and for another submarine at the end of the year.\textsuperscript{28} Once again, Barrow’s departments worked at 100% capacity to fill mountings orders, while the shipbuilding department worked at only 88% of capacity.\textsuperscript{29} 1938 was another lean year for the Vickers naval yards. Barrow received only one order for mountings in the third
quarter of the year, although existing orders kept most departments fully occupied. The Naval Yard, after three years with almost no direct contracts, had very little work; only one order, for a cruiser, was received in 1938.
After these two lean years, the Admiralty suddenly placed a large number of orders with both Barrow and the Naval Yard in 1939. In the first quarter of 1939 Barrow received orders for battleship machinery, while the Naval Yard was given the hull. Both also received an order for a cruiser, and Barrow an extensive order for mountings and guns, while the Naval Yard accepted an order for two cruisers. In the second quarter, Barrow received another large mountings order, which was followed by orders for a cruiser, twelve submarines, machinery for three destroyers, and mountings. Despite the number of shipbuilding orders received, the Barrow shipbuilding department worked at an average of only 59% capacity.
This fact points to some of the hidden structural problems of rearmament. Partly due to factors beyond its control and partly due to its own choice, in the 1930s the Admiralty followed a pattern of large armaments orders to specific plants followed by lean years. This, in turn, produced a vicious cycle of under-used capacity followed by delays in production due to shortages of labour and materials. By September 1936, the Barrow yards was forced to limit its repair work as the amount of new construction on hand had depleted the already limited pool of skilled labour. By the end of the year, the entire shipbuilding industry was faced with a shortage of joiners, plumbers, sheet iron workers, fitters, painters, shipwrights, and riveters. While the number of apprentices had risen, the Barrow yard alone required at least a
hundred more. Unskilled labour was also scarce.\textsuperscript{37}
In this period, labour for shipbuilding was organized in a complex way, one centred on skilled craftsmen rather than mass production lines. Ideally, a shipyard extended back from a river or shore with the berths located near the material stockyard, and the work and shop processes arranged in sequence between. The building process began with the creation of the working drawings and the ordering of material, which was assembled and prepared for construction. The sections (structural bars) were then bent, machined, and marked; at the same time, in an adjacent part of the yard, the plates were marked, cut, machined, and shaped. Once these two processes were completed, the sections and plates were assembled into units, which were then taken to the berths for assembly. At the berths, large cranes were used to install the ship’s heavy machinery.\textsuperscript{38} Naval construction, of course, also required special production such as the fabrication of guns and gun mountings. These were constructed while work on the sections and plates was underway and were installed after the units were assembled. Once the basic structure was completed, the finishing trades, that is, plumbers, carpenters, electricians, painters, and riggers were brought in to complete and fit out the ship. For this reason, a steady stream of orders was required because entirely different tradesmen were required to handle different points in the construction process. If a large order was received and then followed by a lull, the shipyard would be forced to hire large numbers of tradesmen for the initial stage of the process, only to release them and then hire the tradesmen required for the next stage of production. To meet its labour needs, Vickers imported skilled tradesmen
from other regions, but this was only a temporary solution. However, once these men were able to find work in their trades closer to home they would return, and the problem would start over again.
These problems were exacerbated by the inconsistency of the Admiralty’s orders. The aircraft carrier ordered in 1937, for example, did not require the same sort of heavy construction needed for other ships, so more orders were required to keep the men employed and prevent them from leaving for jobs in other yards, but it did require the work of trades which were not involved in the construction of surface warships. Shortages of rivets and armour were beginning to cause delays in construction. At the same time, the volume of orders for mountings had outstripped the capacity of Vickers’ yards and delays were becoming a problem.\(^{39}\) By September 1937, steel shortages had forced the yards to dismiss many platers, riveters, caulkers, and drillers; meanwhile, the lack of shipbuilding orders meant that the rest would soon have to be let go as well.\(^{40}\)
By March 1938, precisely when naval rearmament was surging forward, the situation in the Vickers’ yards was bleak. The firm had been unable to secure a steady supply of steel plates for ship construction\(^{41}\) and, as a result, many of the workers had been laid off. Most of the shipbuilding orders were nearly completed, and thus the workers in the finishing trades would soon have to be laid off; given the lack of new contracts and orders in progress, it would be years before these tradesmen would be needed.\(^{42}\) Over the course of the next twelve months, steel production continued to decline and just the key workers were retained in the yards. The
situation only began to improve with the outbreak of war, with increased steel production and the receipt of more orders.\textsuperscript{43}
Difficulties in obtaining a steady flow of supplies was at the heart of Vickers’ difficulties during the period 1936 to 1939. The sudden increase in demand by the entire armaments industry meant suppliers could not keep up. There were shortages of raw materials, such as steel, and plant, such as machine tools. This disrupted work on contracts and, as orders were delayed, the Admiralty chose to place orders with other firms. So intricate was the relationship between time, the cycle of production, labour, and material that the smallest delay or shortage in virtually any area of the endeavour could throw an entire order or yard wildly off schedule. The result was a cycle in which a lack of material or labour delayed work, which caused a drop in orders, which produced layoffs due to a lack of orders, and more delays due to shortages of labour and supplies. While the Admiralty was not entirely to blame, its actions certainly made the problem worse.
\textit{Capital Plant Extensions}
While the Admiralty was not particularly successful in utilizing Vickers’ capacity during peacetime, it was more successful in expanding the latter for use in wartime. Prior to the Prime Minister’s Statement on Defence in 1935, Vickers had thoroughly re-equipped its works and thus felt the firm was ready for the increased demands of rearmament. However, it soon became obvious that these preparations would be insufficient and that further measures were necessary, particularly with
regard to the development of extra gun mounting capacity.\textsuperscript{44} To better meet the demand for armaments two steps were taken to increase Vickers’ capacity. Under the Government Defence Programme, the government agreed to provide grants to allow firms, including Vickers, to expand their plant. This programme began in 1936 and continued until the outbreak of the war. Additionally, in a deal negotiated directly with Vickers in 1937, the British government purchased the Scotswood works from Armstrong Whitworth & Co., and then handed them over to be operated by Vickers.
Under the terms of the Scotswood agreement, the government paid for the installation of new tools and leased the plant to Vickers for a twenty-eight year term. Under this lease, Vickers had the right to terminate its occupancy at the end of seven, fourteen, or twenty-one years, and also had the right to first refusal should the government choose to sell the works. Vickers agreed to take over the plant on 2 August 1937 along with the existing stock, stores for orders in progress, and general stores. Vickers assumed all contracts in progress at the time of transfer and continued to employ most of the personnel at the plant.\textsuperscript{45} The terms of the agreement provided Vickers with the opportunity to expand its operations at little risk. Start-up costs were minimal, a favourable rent had been negotiated, a labour force was already on site, and there was no risk of being caught with costly excess capacity should the armaments market diminish. At the same time, should it prove advantageous, Vickers could acquire the plant outright at a later date. Since the Admiralty wanted Vickers to specialize in the production of gun mountings, the Scotswood agreement allowed the firm to do so without tying itself to highly specialized plant which would be nearly
impossible to convert for other purposes. The generosity of this agreement shows that just as the Admiralty held the upper hand over Vickers in the allocation of contracts, it desperately needed Vickers’ help in expanding specialized capacity.
The acquisition of Scotswood had an important effect on Vickers’ armament capacity. By 1938, Vickers estimated its probable annual output of naval armaments in time of war as follows:
| Type | Barrow | Elswick |
|-----------------------------|--------|---------|
| 16" Mountings | 4.5 | 4.5 |
| 6" Triple Mountings | 18 | 12 |
| or 9.2" Coast Mountings | 15 | 9 |
| 5.25" Mountings | 20 | 16 |
| | | 18 |
| 4.7" Mountings (a) Mark XIX Twin | 40 | 24 |
| or (b) Mark XX Twin | 26 | 16 |
| or (c) Mark XVIII Single | 120 | 72 |
| 4.5" Twin Mountings | 16 | 25 |
| Pom-Pom Mountings | 36 | 50 |
| 4" Twin Mountings | 50 | |
The above figures were based on the assumption that labour shortages would not
create complications and that the firm could continue its practice of using subcontractors for some of the work. The only cost foreseen in maintaining such an extensive program was the purchase of some heavy machinery at an approximate cost of £120,000. But, there existed one drawback. Vickers would have difficulty producing enough 16" guns and mountings for the planned three capital ships per year; it could do so only by funding plant extensions at Sheffield and Elswick (at an estimated cost of £430,000). If the extension was completed, the Sheffield works could produce twenty-five sets of 16" forgings and two relines per year, while Elswick could produce an additional seven 16" guns (increased from 4.5) and two 15" or 16" relines per year. If the extensions were not completed, Sheffield could only produce eighteen 16" guns.\textsuperscript{46}
While the extra production at Scotswood appears at first to have had only a limited impact on Vickers' total production, that is not the case. The acquisition of Scotswood had three positive effects for the firm and the country. First, it increased the country's total armaments production as the plant had previously produced mostly civilian goods. Secondly, it gave Vickers an extra and sorely needed source of skilled labour, which could be easily retrained.\textsuperscript{47} Finally, it increased Vickers' plant for armament production without endangering the survival of the firm. Without the generous agreement of 1937, Vickers would not have taken this step.
Vickers gained not only from the Scotswood agreement, but also from other programs directed at the armaments industry as a whole. In 1936 the British government approved funding for various plant extensions to Vickers works at
Barrow, Elswick, Crayford, and Dartford under the Government Defence Programme. The funds approved at that time totalled £807,000, though this figure increased as the program continued; Barrow and Elswick alone received over £900,000. At the same time the English Steel Corporation, a Vickers subsidiary, was granted £1,000,000 to build an additional armour plate plant.\textsuperscript{48} Barrow, being one of the most modern of the Vickers works, received only three grants, given to the firm in the first quarter of 1937. These were - Grant #71, for the purchase of machine tools for gun mounting, £409,116; Grant #72, for a new heavy gun mounting Shop, £240,500; and Grant #77, for a new pattern store, £14,750.\textsuperscript{49} On the other hand, Elswick received eleven grants, totalling £270,500:
\begin{tabular}{ll}
Grant: & Amount: \\
Special Capital #3 - Guns & £35,500\textsuperscript{50} \\
Special Capital #4 - Breech Mechanisms & £15,500\textsuperscript{51} \\
Special Capital #7 - Cartridge Cases & £10,000\textsuperscript{52} \\
Special Capital #9 - Extension to Heavy Gun Mounting Shop & £45,000\textsuperscript{53} \\
Special Capital #10 - Fuzes & £5,000\textsuperscript{54} \\
Special Capital #11 - Gun Mountings & £20,000\textsuperscript{55} \\
Special Capital #12 - General Purposes & £4,000\textsuperscript{56} \\
Special Capital #13 - New Gun Pit, 24 Shop & £5,500\textsuperscript{57} \\
\end{tabular}
Special Capital #16 - Guns £21,500\textsuperscript{58}
Special Capital #17 - Breech Mechanisms £8,500\textsuperscript{59}
Special Capital #19 - Increase capacity for 4.5" guns and liners £100,000\textsuperscript{60}
Of all the Vickers’ works producing naval goods, Elswick was the most heavily involved in the production of guns and mountings, and the grants illustrate the Admiralty’s determination to have Vickers specialize in the production of these goods. Vickers was willing to do so if the Admiralty would subsidize the process and underwrite the risks. This, in turn, illustrates another fundamental factor in the industrial side of rearmament. Vickers was determined to avoid being caught up in the rush to rearm. It was unwilling to follow Admiralty policy blindly, and to ensure its own survival, the firm was determined to continue its civilian production and foreign armaments sales even during the peak of peacetime rearmament.
\textit{Business As Usual}
Despite the extensive requirements of the British armed forces as they began to rearm, Vickers sought to keep British rearmament production from completely dominating its productive capacity. Based on its experiences after 1918, Vickers believed that the armaments orders must ultimately slow down and that if the firm was not careful, it would be caught with a huge workforce and plant geared for the production of warplanes, guns, and warships as opposed to huge profits and the
ability to manufacture locomotives, automobiles, tractors, and merchant shipping. In 1922 and 1923 the situation at Barrow had been so grim that birds were seen nesting in the unused cranes, and the men still working were met at the yard's gates every afternoon by children begging for food.\(^{61}\) While, under such circumstances, some of the capacity could be converted, the process could be costly and the dislocation to the firm immense. Fortunately for Vickers, the three services recognized the validity of its position and did nothing to prevent the firm from continuing to fill other orders. Aside from armament production destined for the British forces, Vickers continued to fill foreign orders as well as to produce non-military goods. Vickers was not alone in its determination to carry on its normal business, for it was the explicitly stated policy of the government throughout the rearmament period. While some politicians demanded that civilian production be curtailed in favour of increasing armaments production, the government disagreed. A reduction of civilian production would lead to either a decrease in the standard of living or an increase in imports and a resulting adverse balance of trade. While the former was viewed as political suicide, the latter presented the greater danger of economic dislocation. Thus, Treasury officials were able to convince the government that "business as usual" was the safest option.\(^{62}\) It is characteristic of the complexities of rearmament that, on the one hand, Britain had too little naval armaments, while on the other hand it encouraged its leading arms manufacturers to use much of their capacity to arm other powers.
The War Office and the Admiralty supported Vickers' international trade, as they recognized that this business was central to the preservation of capacity.\(^{63}\) So too
did the company’s corporate officials, who included a number of retired senior military officials, with access to the corridors of power. The company’s CEO during the rearmament era, Sir Charles Craven, was an ex-naval officer with a flair for the dramatic. One story tells of him spiriting a Japanese naval delegation into seclusion away from London for a week, then returning to Barrow with a substantial contract for submarines. Craven was no less determined to drum up commercial orders; he was said to have boarded Lord Inchcape’s yacht on the Riviera and remained there “genially piratical” until Inchcape agreed to give Vickers an order from his New Zealand Shipping Company.\(^{64}\) While Craven could use his extensive connections at the Admiralty, General Sir Noel Birch, the former Master General of the Ordnance, who was appointed to the Board of Directors in 1927, provided a vital link with the War Office. The inventor of the Birch self-propelled gun, he was vital in assisting the firm’s design teams. As the firm could only sell abroad weapons which it had designed itself (that is, without War Office collaboration), Birch’s knowledge of tactics and weaponry was invaluable.\(^{65}\)
While the Admiralty in particular had always been willing to do whatever was possible,\(^*\) or at least whatever the Treasury would permit, to encourage firms to preserve capacity for armaments production, it was fiscally advantageous to have other countries provide a steady stream of orders to keep skilled tradesmen employed and materials in ready supply. Moreover, foreign orders helped to bolster Britain’s image as a global power, with the technology to maintain her position as such, and it gave
\(^*\) see Chapter One, pp. 11.
Vickers’ order books a certain degree of stability. The War Office and the Royal Air Force had lengthy "wish lists", but these needs varied from year to year as defence plans changed and estimates were devised. Only the Admiralty could be reasonably definite about what it would be ordering in the coming years and even that did not seem secure at the time, as the events of 1929-30 had overturned its policy of a decade. Again, in March 1936 one of Vickers’ directors, Sir Noel Birch, noted that there was a debate about the abolition of battleships within the Admiralty, which would not be settled until a new minister was appointed. Moreover, the political situation threatened Vickers as well. If international tension could be eased, the firm could well experience another round of defence cuts. Rearmament could not last forever. In Birch’s opinion it was better to accept both foreign and British orders and worry about a lack of capacity when and if the problem arose. To turn away foreign orders would be to cripple Vickers’ international trade for years to come. Conversely, to accept such orders would bolster Britain’s military position. In Birch’s words, "if we have to defend ourselves sooner than we expect, the mere fact that we have all these perfectly excellent 75 AA batteries in manufacture, will be of the greatest use to the army." The preservation of foreign trade was a necessity to maintain British foreign markets and supplement British military orders.
Birch’s opinion was echoed by the firm’s managing director, Sir Charles Craven, although in a somewhat less bombastic fashion. In a visit to Admiral Chatfield at the Admiralty, and to the War Office later that same month, Craven emphasized the need to maintain foreign business, though he also privately assured
officials that British needs could still be met. So long as this could be done and foreign clients were willing to pay, Craven welcomed the extra business. However, despite Craven’s promise that British orders would be given priority, he recognized the dilemma in which too many orders would place the firm. Vickers’ northern works had yet to begin work on contracts totalling £11,000,000, with further orders from the Admiralty and War Office still to be placed. Late deliveries of vital equipment, especially machine tools, had slowed work on existing orders, while Craven estimated that the firm’s workforce was 25% below that required to keep pace with current orders. To postpone the production of orders which were soon to be placed by the British services in order to meet foreign obligations would be disastrous, as the services would simply take their business elsewhere; yet to abandon foreign orders to meet the services’ needs would threaten the survival of the firm.\(^{67}\)
At the same time, foreign orders represented a sizeable part of Vickers’ normal business. As an example, in the last quarter of 1934, Vickers had a Belgian order for tanks forthcoming and had secured armaments orders totalling £332,000 from Belgium, Greece, and Turkey.\(^{68}\) Nor was this an isolated event - between 1930 and 1934, in the midst of a global economic crisis, Vickers’ foreign armaments sales from the Elswick works alone were £1,831,084.\(^{69}\) A similar situation existed at the Crayford works. Until the outbreak of war, foreign orders comprised over half of the work completed in the yard.\(^{70}\) In the long run, Craven noted, to limit Vickers’ foreign trade would reduce its ability to meet British armaments needs, unless the government would subsidize it to keep the extra capacity in being. Craven
emphasized that Vickers was willing to give priority to British orders and accept price controls or any other conditions the government chose to apply in order to retain the right to continue direct negotiations in certain cases.\textsuperscript{71} If the government chose to take control over negotiations for armaments sales to foreign powers or, worse yet, ban sales entirely, all the efforts Vickers’ had made to increase its foreign trade would be wasted and the firm might wither as a result.
Despite Vickers’ arguments, in July 1938 the government took over the negotiation of armament sales to foreign powers. Craven, fearing the worst, quickly sent a letter to Admiral Chatfield asking him to use his influence to have the rules relaxed.\textsuperscript{72} Admiral Chatfield’s reply illustrates the confusion which existed during this period. He noted that the three services interpreted government decisions in different ways, and held that the Air Ministry and the War Offices had interpreted the government’s measures of July 1938 far too strictly. The aim of the policy was to coordinate industrial output so as to maximize the benefits to Britain and its allies. Chatfield interpreted these new rules to mean that foreign powers would make general inquiries about arms sales to the British government, which would decide whether to supply the foreign power and, if so, whether to do so directly or to refer the order to a commercial interest. In most cases, the orders would be passed on to armaments firms, which would then deal with the foreign power directly. Chatfield also felt that the new rules did not apply to negotiations already in progress, a key concern to Craven. Chatfield also promised to make inquiries and attempt to get the rules relaxed.\textsuperscript{73} Chatfield’s cooperation and willingness to get the new regulations relaxed
show that Craven, a former Royal Naval officer, had chosen the right service to complain to. Such lobbying efforts were a key feature of Vickers business methods during this period, and attempts to pit the services against each other were frequently successful.
The commitment before September 1939 of both the government and the firm to "business as usual" shaped the actual conditions in the Vickers shipbuilding yards. In a series of interviews which took place in 1960, Vickers' employees discussed their memories of the interwar period. With the onset of rearmament, a number of individuals recalled the difficulties of resuming a full production level. Mr. D. Rodger, manager of the Vickers Marine Department in Newcastle, stated that very few men had left the region during the Depression. When the Naval Yard reopened in 1934 most of the jobs were filled by men who had been previously employed by Vickers, and on most mornings men waited outside the gates looking for work. Despite the ready supply of labour, the yard still experienced problems when it was reopened and a lot of money had to be spent to get the plant operating again. Once production resumed, the previous surplus of labour disappeared and, by 1937, the yard experienced serious shortages of skilled labour.\(^{74}\)
While activity in the various Vickers naval yards had greatly increased after 1935, most of those interviewed in 1960 remembered the pace as busy but not frantic. Only with the outbreak of war were the works truly swamped; an employee at Elswick stated that he worked seven days a week between 1939 and 1942. At that time, in order to speed construction, the Admiralty stationed a captain and staff at Elswick
with authority to approve ships’ drawings. These drawings had previously had to be sent to Bath for approval.\textsuperscript{75} Despite the increased work which stemmed from naval rearmament, the yards continued to produce civilian products, including cement plant, condensing and feed heating plant, pumping plant, and winding machinery until the outbreak of war. These goods used exactly the same type of capacity which produced marine engines and guns, but Vickers chose not to use that capacity for armaments production. This decision was supported by many in the works, who feared for the future once the rearmament program ended and orders slowed. This response was understandable, considering the events of 1922-1934.\textsuperscript{76} Given the experience of the rapid reductions following World War I, Vickers and its employees regarded the naval building boom as a temporary phenomenon, a common reaction among firms and unions during rearmament. Perhaps due to the specialized nature of his department, one senior official interviewed stated that there was no advantage in converting capacity used for civilian products to armaments production. Mr. J.B. Brown, Manager of the Gun Mounting Department, was adamant that civilian production did not interfere with armaments production and that to turn men from the production of cement machinery to gun mountings would have only slowed his department down. He also emphasized that the Admiralty had full knowledge of the extent of civilian goods being produced in the yards during this period; with the large number of Admiralty officials in the yards the work would have been impossible to hide.\textsuperscript{77}
After the weak armaments market of 1929-1934, the push for rearmament should have been a blessing to Vickers and its capacity should have been fundamental
to rearmament. However, for a number of reasons, this was not the case. Despite the Admiralty’s large new construction programmes, it failed to use Vickers’ capacity efficiently. Its practice of dumping a large volume of orders on a firm and then providing only a few more contracts in the following years hampered the supply of labour and materials. This produced delays in the completion of orders, which made the Admiralty withhold orders, which made the situation even worse. The Admiralty’s inefficient use of capacity created one set of problems for Vickers; its attempt to force the firm into increasing its capacity for gun mountings created another. While the firm was willing to increase its production in this area, it was determined to avoid being left with a large amount of unusable capacity once the naval orders declined. In this respect, the firm was remarkably successful, for it was able to get subsidies and work out special arrangements like the Scotswood purchase, which increased Vickers’ gun mounting capacity while minimizing its liabilities.
Vickers was also successful in ensuring that the British rearmament programme did not completely overtake its other areas of production. The firm’s determination to continue producing civilian goods and its demands that the government allow it to continue its foreign arms sales was vital in ensuring its long-term corporate stability. Thus, neither the Admiralty nor Vickers itself even tried to make maximum use of the firm’s capacity – what emerged was a compromise between the Admiralty’s desire for full production and the firm’s desire to preserve alternate markets.
Vickers’ experiences during the rearmament period illustrate that the process must be studied from the firm’s perspective as well as the government’s. While clear
trends about rearmament emerge at the policy level, they are contradictory at the firm level. In one area of production, Vickers had more orders than capacity, while in another, not nearly enough. While the Admiralty required more warships it could not provide Vickers with enough orders for rational production, so Vickers used the excess capacity to fill orders for foreign navies. At the crux of the matter was the conflicting interests of the two parties involved. The Admiralty had to consider the security of the nation and its ability to fulfill its responsibilities. Vickers, on the other hand, had to consider the profit involved in rearmament. While these goals are different, and often contradictory, each was valid. Thus, the relationship between Vickers and the Admiralty shows that the Admiralty's political gains of the early 1930s were not translated into a great success for the armaments industry. Even with careful planning, Vickers had difficulty meeting the Admiralty's demands.
Notes
1. Scott, p. 218.
2. J.D. Scott and Richard Hughes, *The Administration of War Production* (London: HMSO, 1955.), pp. 7-8.
3. Scott & Hughes, p. 8.
4. Quarterly Report - quarter ending 31 March 1933. Vickers 168
5. Quarterly Report - quarter ending 31 March 1934. Vickers 172.
6. Quarterly Report - quarter ending 31 December 1934. Vickers 175.
7. Quarterly Report - quarter ending 30 June 1934. Vickers 173.
8. Quarterly Report - quarter ending 31 December 1934. Vickers 175.
9. Quarterly Report - quarter ending 31 March 1935. Vickers 176.
10. Quarterly Report - quarter ending 31 December 1935. Vickers 179.
11. Compiled from Quarterly Reports, 1935. Vickers 176, 177, 178, 179.
12. Summary of Sales, 1931 - 1939. Vickers 722.
13. Summary of Sales 1931 - 1939. Vickers 722.
14. Summary of Sales, 1931 - 1939. Vickers 722.
15. Compiled from 1934 Quarterly Reports. Vickers 176, 177, 178, 179.
16. Compiled from 1934 Quarterly Reports. Vickers 176, 177, 178, 179.
17. Q.R. ending 30 June 1933, Vickers 169; Q.R. ending 31 December 1933, Vickers 171; Q.R. ending 31 March 1934, Vickers 172; Q.R. ending 30 June 1934, Vickers 173; Q.R. ending 31 December 1934, Vickers 175; Q.R. ending 30 June 1935, Vickers 177; Q.R. ending 31 December 1935, Vickers 179.
18. Hornby, p. 58.
19. Quarterly Report - quarter ending 31 December 1935. Vickers 179.
20. Taken from Summary of Sales, 1931 - 1939. Vickers 722.
21. Quarterly Report - quarter ending 31 March 1936. Vickers 180.
22. Quarterly Report - quarter ending 30 June 1936. Vickers 181.
23. Quarterly Report - quarter ending 30 September 1936. Vickers 182.
24. Quarterly Report - quarter ending 31 December 1936. Vickers 183.
25. Compiled from 1936 Quarterly Reports. Vickers 180, 181, 182, 183.
26. Quarterly Report - quarter ending 30 June 1937. Vickers 185.
27. Quarterly Report - quarter ending 30 September 1937. Vickers 186.
28. Quarterly Report - quarter ending 31 December 1937. Vickers 187.
29. Compiled from 1937 Quarterly Reports. Vickers 184, 185, 186, 187.
30. Quarterly Report - quarter ending 30 September 1938. Vickers 190 (i).
31. Quarterly Report - quarter ending 31 March 1938. Vickers 188.
32. Quarterly Report - quarter ending 31 March 1939. Vickers 192.
33. Quarterly Report - quarter ending 30 June 1939. Vickers 193.
34. Quarterly Report - quarter ending 30 September 1939. Vickers 194.
35. Compiled from 1939 Quarterly Reports. Vickers 192, 193, 194.
36. Quarterly Report - quarter ending 30 September 1936.
37. Quarterly Report - quarter ending 31 December 1936.
38. D.J. Eyres, *Ship Construction* (London: Heinemann, 1972), pp. 103-104.
39. Quarterly Report - quarter ending 30 June 1937. Vickers 185.
40. Quarterly Report - quarter ending 30 September 1937. Vickers 186.
41. The quarterly report indicates that steel production had dropped to 25% of normal output. It is not clear whether this refers to the production of the English Steel Corporation, Vickers subsidiary, or to national production.
42. Quarterly Report - quarter ending 31 March 1938. Vickers 188.
43. Quarterly Report - quarter ending 30 September 1939. Vickers 194.
44. Letter from Sir Charles Craven to Admiral Backhouse, 7 December 1938. Vickers 675.
45. Vickers Limited Board Minutes, 14 July 1937. Vickers 1225.
46. Letter from Sir Charles Craven to Admiral Backhouse, 7 December 1938. Vickers 675.
47. Scott, p. 225.
48. Vickers Limited Board Minutes, 21 May 1936. Vickers 1371.
49. Quarterly Report - quarter ending 31 March 1937. Vickers 184.
50. Quarterly Report - quarter ending 30 June 1936. Vickers 181.
51. ibid.
52. ibid.
53. ibid.
54. ibid.
55. ibid.
56. ibid.
57. ibid.
58. Quarterly Report - quarter ending 30 September 1936. Vickers 182.
59. ibid.
60. Quarterly Report - quarter ending 31 December 1936. Vickers 183.
61. Scott, pp. 144-45.
62. R.A.C. Parker. "British Rearmament 1936-9: Treasury, trade unions and skilled labour." English Historical Review. XCVI, 379 (April, 1981, pp. 306-343), p. 309.
63. Letter from Sir Noel Birch to Sir Charles Craven, 12 March 1936. Vickers 603
64. Scott, p. 160.
65. Scott, p. 188.
66. Letter from Sir Noel Birch to Sir Charles Craven, 12 March 1936. Vickers 603
67. Telegraph from Sir Charles Craven to Sir Noel Birch, 13 March, 1936. Vickers K613.
68. Memo- Military and Air Armaments, 3 December 1934. Vickers 532.
69. Summary - 1930-34 - SALES. Reply to Royal Commission on International Arms Trade Questionnaire - Vickers Armstrong - Elswick. Vickers 79.
70. Interview with Mr. Lane, 14 January 1960. Vickers 674.
71. Letter from Sir Charles Craven to Admiral Chatfield, 1 August 1938. Vickers 532
72. Letter - Craven to Chatfield, 1 August 1938. Vickers 532.
73. Letter from Admiral Chatfield to Sir Charles Craven, 8 August 1938. Vickers 532.
74. Interview with Mr. D. Rodger, March 1960. Vickers 722
75. Interview with Mr. Hall, 26 September 1960. Vickers 674.
76. Interview with Mr. E.P. Beach, March 1960. Vickers 722.
77. Interview with Mr. J.B. Brown, March 1960. Vickers 722.
Conclusion
The Royal Navy was in a secure position following World War I; indeed, even after the Washington Naval Treaty of 1922. While capital ship replacement was postponed until 1931, the construction of various categories of warships gave shipbuilding firms enough work to survive. During this period, the Admiralty also formulated long-term plans which hinged on the replacement of the capital ship fleet over the decade following 1931. During this same period, the British shipbuilding industry went through a rough time of adjustment from wartime to peacetime production. After almost a half-century of growth, the end of World War I left the industry with too much capacity for the immediate needs of armaments production. Even the Admiralty recognized this fact and allowed the inevitable readjustment to take place; it only interfered to ensure the survival of the specialized capacity required for the capital ship replacement program. In fact, the rationalization of the industry combined with judicious actions by the Admiralty sufficed to keep the industry in an acceptable state until 1929. Where an overabundance of capacity existed, the Admiralty chose to funnel its orders so to keep a select few firms operational, while other firms were allowed either to cease armaments production, go out of business, or survive on their own strength. In some cases, such as the production of shells, the Admiralty was forced to subsidize firms to keep the capacity in being. While orders were often low during the period, the Admiralty recognized the need to preserve the capacity for emergency needs, and did so.
This period of security ended in 1929, for two reasons: the London Naval
Treaty of 1930 and the government’s fiscal restraint following the global depression of 1929. The London Treaty devastated British naval policy for it extended the holiday on capital ship construction and restricted construction on other classes of warships. Throughout the 1920s the Admiralty had used the construction of cruisers, destroyers, and other vessels to keep the shipbuilding industry working until capital ship replacement could commence. This option was lost after the treaty was signed as was the incentive which the Admiralty had used to encourage firms to keep currently unprofitable plant in being, that is, the promise of large and continued capital ship orders after 1931. All this crippled the British shipbuilding industry, particularly because the Depression limited the orders of merchant vessels which might otherwise have compensated for the drop in naval orders. The treaty simultaneously reduced the navy’s ability to defend British interests. By severely limiting the size of the Royal Navy, it made the simultaneous defence of the Far East and European waters nearly impossible. At the same time, the global depression heightened the government’s emphasis on fiscal restraint, which further hampered the Navy’s ability to perform its assigned role. The period 1929-34 marked the peak of the Treasury’s control over the Naval Estimates, which only declined with the rise of the Japanese and German threat.
In 1933, with the resurgence of Germany and the growing menace of Japan, the dire position of the three British services was finally recognized. The Defence Requirements Committee was appointed to investigate the worst deficiencies and to eliminate them. It recommended a large, balanced deficiencies program designed to deal both with the immediate threat of Japan and the long term danger of Germany.
As a result, the Royal Navy’s needs were well-addressed by the program, for it was vital in countering both possible threats. Unfortunately, the DRC’s report was virtually ignored. No decisions were made on the matter until early 1935 and the delay cost Britain dearly. By failing to act immediately and to begin countering the deficiencies which had developed in all the services since 1919, the government made the task of rearmament essentially impossible. In fact, the Royal Navy was less capable of protecting British interests after the DRC’s report than before.
The increase in international tension forced the DRC to reconvene in 1935, with a mandate for rearmament, not deficiencies. The DRC accepted the Admiralty’s demands for a large new construction program which would bring the Navy up to the standard it was formally intended to meet. While the DRC also accepted the Navy’s request for a new Two-Power Standard, the DPRC merely recommended to the Cabinet that the Navy be asked to study the new standard, and further define what fleet it would require.
The rearmament period was marked by the growing political influence of the Admiralty. Its political position had waned since 1925 and reached its nadir during the years following 1929. While the Treasury was successful in controlling naval expenditure during the period 1929 to 1933, its ability to do so decreased in the years following. The Admiralty’s success in dealing with the DRC spilled over into its dealings with the Treasury, and naval estimates. At the same time, the more the Admiralty gained, the more it demanded. When the first DRC report accepted the Admiralty’s demands, it returned with a new naval standard in 1935. When the DRC
accepted that new standard, and the DPRC encouraged the Admiralty to develop the idea, the latter then increased the two-power standard to a three-power standard and proposed a new and much larger fleet to enforce it. The Admiralty then maintained that the only the new standard fleet could meet the Japanese and German threat, the task which the DRC fleet was originally supposed to accomplish.
These were sizeable victories, but they were not translated into concrete progress in rearmament. The Admiralty was able to deal more efficiently with politicians than British industry. As its dealings with Vickers illustrate, serious problems with the Admiralty’s procedures for procurement hampered the progress of the rearmament program. Inconsistent ordering prevented the efficient use of Vickers’ shipbuilding capacity. By ordering heavily from Vickers one year and then offering very few contracts in the next two years, the delivery of vessels was delayed. As Vickers’ dealt with sudden orders, it was forced to hire large numbers of tradesmen, only to release them soon after when the work was completed and no further orders were in sight. This process continued with different groups of tradesmen throughout the shipbuilding process. The Admiralty’s response to the delays which stemmed from shortages of labour and inconsistent supplies of materials, problems which it created, was to cut further orders and to thereby make the problem worse. This vicious circle continued until the outbreak of war and only under the circumstances of a total war, with all the capacity used to the limit all the time, were these problems overcome. Once this happened, British production was organized very well. Until then, however, different firms performed at different levels of efficiency with chaotic
While orders for ships were inconsistent, orders for other naval armaments were not. The Admiralty wished to make Vickers its sole supplier of gun mounts, while Vickers feared that as a result it would be driven into a dangerously overspecialized situation. Not wishing to take such a risk, the firm responded in two ways: it demanded that the Admiralty subsidize the expansion of the firm’s gun mounting capacity and it doggedly followed the government’s adage of “business as usual” by continuing to produce goods for civilians at home and for the military up until the beginning of the war. Both of these steps were taken, and the end result was beneficial to both British armaments capacity and to Vickers. The Admiralty’s efforts to force Vickers to adopt what the firm feared would be an unprofitable course of action actually made the firm even more determined not to become the prisoner of rearmament. Therefore, Vickers was all the more determined to continue business which it believed would bring long term profit instead of relying on the short term profits from rearmament.
This thesis illustrates the difference between formulating policy and executing it, and three conclusions are evident. First, the Admiralty’s attempts at long range planning were thwarted by interdepartmental rivalries, both the competition between the services for limited defence budgets and the greater rivalry between the Treasury and the Admiralty. When the Admiralty had powerful political supporters, and the support of other departments like the Foreign Office, it was successful. When this was not the case, it was crushed. The dramatic swings in the balance of power over
naval decision making crippled long term procurement and the Navy, armaments firms like Vickers, and ultimately Britain itself paid the price. Secondly, British rearmament was a flawed process. It began too late, and from too low a level, and the problem was not merely in actions taken in Whitehall but in the industrial implementation of policy. Finally, all this shows that rearmament cannot be fully understood without assessing the interest, the capabilities, and the experiences of the armament firms. The Admiralty’s dealings with Vickers show that political victories do not necessarily ensure success procurement. The Admiralty won many political fights from 1934 to 1939, but it could not translate this success into concrete progress on construction. In fact, its political success probably made the situation worse by causing confusion as the Admiralty tried to follow both DRC rearmament policy and its own version of rearmament. The DRC rearmament plan was supposed to be completed in 1939, while the Admiralty’s rearmament scheme continued past 1942; in trying to fulfill the requirements of both schemes, the Admiralty entered World War II less prepared than it should have been, and could have been.
Bibliography
Primary Sources
OFFICIAL RECORDS
*Public Record Office, Kew*
ADM 1 Admiralty and Secretariat Papers
ADM 116 Admiralty and Secretariat Cases
ADM 167 Board of Admiralty Minutes and Memoranda
CAB 21 Registered Files
CAB 24 Cabinet Memoranda
T161 Supply Files
T175 Hopkins Papers
*Published Documents*
Woodward, E.L. and Rohan Butler. *Documents on British Foreign Policy 1919-1934*
Series II, Vol. I. London: HMSO, 1947.
*Sessional Papers*
Appropriation Account, Navy Services, 1930, British Sessional Papers 1931-32.
Appropriation Account, Navy Services, 1931, British Sessional Papers 1932-33.
Appropriation Account, Navy Services, 1932, British Sessional Papers 1933-34.
Appropriation Account, Navy Services, 1933, British Sessional Papers 1934-35.
CORPORATE ARCHIVES
Vickers Archives, Cambridge University Library.
Secondary Sources
BOOKS
Aldcroft, Derek. *The Interwar Economy: Britain 1919-1939*. New York: Columbia University Press, 1970.
Ashworth, William. *Contracts and Finance*. London: HMSO, 1953.
Barnett, Correlli. *The Audit of War*. London: Macmillan, 1986.
Bond, Brian. *British Military Policy Between the Two World Wars*. Oxford, 1980.
Carlton, David. *MacDonald versus Henderson*. London: Macmillan, 1970.
Cooper, Duff. *Old Men Forget*. London: Rupert Hart-Davis, 1953.
Eccles, H.E. *Logistics in the National Defence*. 1959.
Eyres, D.J. *Ship Construction*. London: Heinemann, 1972.
Ferris, J.R. *Men, Money and Diplomacy*. Ithaca, NY: Cornell University Press, 1989.
Gibbs, N.H. *Rearmament Policy*. London: HMSO, 1976.
Gilbert, Martin. *The Roots of Appeasement*. London, 1966.
Gordon, G.A.H. *British Seapower and Procurement between the Wars*. Annapolis: Naval Institute Press, 1988.
Graham, Gerald. *The Politics of Naval Supremacy*. Cambridge: Cambridge University Press, 1965.
Grant, Sir Allan. *Steel and Ships*. 1950.
Hancock, W.K. *British War Economy*. London: HMSO, 1975.
Hicks, Ursela. *The Finance of British Government*. Oxford: Clarendon, 1970.
Hornby, William. *Factories and Plant*. London: HMSO, 1958.
Howard, Michael. *Defence Policy in the Era of the two World Wars*. London, 1972.
Inman, P. *Labour in the Munitions Industry*. 1975.
*Jane’s Fighting Ships of World War II*. London: Bracken, 1989.
Johnson, F.R. *Defence by Committee*. Oxford: 1960.
Keegan, John. *The Price of Admiralty*. London: Hutchison, 1988.
Kennedy, Gavin. *The Economics of Defence*. London: Faber & Faber, 1975.
Kennedy, Paul M. *The Rise and Fall of British Naval Mastery*. New York: Chas. Scribner's Sons, 1976.
Kennedy, Paul M. *The Realities Behind Diplomacy*. London, 1981.
Marder, Arthur. *The Anatomy of British Sea Power*. Hamdon, Connecticut: Archon, 1964.
McKercher, B.J.C., ed. *Anglo-American Relations in the 1920s*. Edmonton: University of Alberta Press, 1990.
Middlemas, K. *The Diplomacy of Illusion*. 1972.
Neidpath, James. *The Singapore Naval Base and the Defence of Britain’s Eastern Empire, 1919-1941*. London, 1981.
d’Ombrain, N. *War Machinery and High Policy*. 1973.
Peden, G.C. *British Rearmament and the Treasury: 1932-1939*. Edinburgh: Scottish Academic Press, 1979.
Pimlott, Ben. *Hugh Dalton*. London: Jonathan Cape, 1985.
Pollard, Sidney. *The British Shipbuilding Industry, 1870-1914*. Harvard University Press, 1979.
Postan, M.M. *British Production*. London: HMSO, 1952.
Pugh, Philip. *The Cost of Seapower*. London: Conway Maritime Press, 1986.
Ranft, Brian, ed. *Technical Change and British Naval Policy, 1860-1939*. London: Hodden & Stoughton, 1977.
Richardson, Dick. *The Evolution of British Disarmament Policy in the 1920’s*. London: Pinter, 1989.
Richardson, H.W. *Economic Recovery in Britain, 1932-9*. London: Weidenfeld and Nicolson, 1967.
Richmond, Admiral Sir Herbert. *Statesmen and Sea Power*. Cambridge, 1946.
Richmond, Admiral Sir Herbert. *National Policy and Naval Strength*. 1928.
Richmond, Lesley and Bridget Stackford. *Company Archives*. Aldershot: Gower,
Rodger, N.A.M. *The Admiralty*. Great Britain: Terence Dalton, 1979.
Roskill, Stephen. *Hankey, Man of Secrets Vol. II 1919-1931*. London: Collins, 1972.
Roskill, Stephen. *Hankey, Man of Secrets Vol. III 1931-1963*. London: Collins, 1974.
Roskill, Stephen. *Naval Policy Between the Wars Vol. I The Period of Anglo-American Antagonism 1919-1929*. London: Collins, 1968.
Roskill, Stephen. *Naval Policy Between the Wars Vol. II The Period of Reluctant Rearmament 1930-1939*. Annapolis: Naval Institute Press, 1976.
Scott, J.D. and Richard Hughes. *The Administration of War Production*. London: HMSO, 1955.
Scott, J.D. *Vickers: A History*. London: Weidenfeld and Nicolson, 1962.
Shaw, G.C. *Supply in Modern War*. 1938.
Shay, Robert. *British Rearmament in the Thirties*. Princeton: Princeton University Press, 1977.
Sumida, Jon. *In Defence of Naval Supremacy*. Boston: Unwin Hyman, 1989.
Taggert, Robert, ed. *Ship Design and Construction*. New York: Society of Naval Architects and Marine Engineers, 1980.
**JOURNAL ARTICLES**
Bywater, Hector. "The Coming Struggle for Sea Power" *Current History*. v. 41, October 1934, pp. 9-16.
Gueritz, Rear Admiral E.F. "Nelson’s Blood: Attitudes and Actions of the Royal Navy 1934-1945." *Journal of Contemporary History*. 16 (1981), pp. 487-99.
McKercher, B.J.C. "'Our Most Dangerous Enemy:' Great Britain Pre-eminent in the 1930's." *International History Review*. XIII, 4, November, 1991, pp. 751-83.
O’Connor, Raymond G. "The "Yardstick" and Naval Disarmament in the 1920's"
Parker, R.A.C. "British Rearmament 1936-9: Treasury, trade unions and skilled labour." *English Historical Review*. XCVI, 379, April 1981, pp. 306-43.
Sumida, Jon. "'The Best Laid Plans'": The Development of British Battle-Fleet Tactics, 1919-1942." *The International History Review*, XIV, 4, November, 1992, pp. 661-700.
**THESES**
Alch, Mark, Lee. "Germany's Naval Resurgence, British Appeasement and the Anglo-German Naval Agreement of 1935." PhD Dissertation, UCLA, 1977.
Bright, Charles C. "Britain's Search for Security 1930-36: The Diplomacy of Naval Disarmament and Imperial Defence." PhD Dissertation, Yale, 1970. |
SOPHOMORES VICTORS IN ANNUAL TRACK MEET
Defeat Freshmen by Score of 40 to 32—Curtis Stars For Sophomores.
Saturday afternoon, in the Gym, the Sophomore track team defeated the Freshman aggregation by the clean score of 40 to 32. The first-year men did excellently in the sprints and distance runs, but were easily outclassed in the field events, however. In the events, Adams, C. T. Guehing, Benson and Loomee for the Freshmen, did very well, while among the Sophomores, Curtis, O'Brien and Jones were the stars. Both teams secured four first places, but the clean sweeps which the Freshmen had very recently made in the field events, together with two unexpected second places in the distance runs, turned the tide in favor of the Sophomores.
The number of entries from the Freshman class was more than double the number of the other entrants. Indeed, in the distance runs, the field was made up almost entirely of Freshmen, and the Sophomores had only one or two runners to oppose the many Freshmen. However, these men were good, and won several places. Coach Kanally was greatly disappointed at the performance of the Freshman team in the field events, and is in hopes that more men will come out and try the shot-put, high jump and pole vault. The latter (Continued to Page 3.)
SOUTHERN CLUB DANCE.
To Be Held at Hotel Tuileries On March 15.
The Technology Southern Club will hold one of its popular dances at the Hotel Tuileries on Saturday evening, March 15, from 8 to 12 o'clock. This dance is not to be limited to the members, as was the case with the last one. Students wishing to procure invitations should see some member of the committee. The admission price is two dollars and a half per couple, including refreshments.
The dances given by the Southern Club at the Hotel Tuileries have been among the most enjoyable functions at the Institute. This one bids fair to come up to all expectations. The matrons will be Mrs. George H. Ramsay, Mrs. Alfred E. Burton, Mrs. Charles M. Sipforder and Mrs. D. Gay.
FRESHMAN MEETING.
Class Dinner to Be Discussed in Huntington Hall.
The Freshman Class will hold a meeting in Huntington Hall on Tuesday, March 11, at 1 o'clock. At this meeting the subject of the advisability of holding a class dinner or smacker will be considered, and the question of the unpaid class dues will also be brought up. The Governing Board voted to pay the class dinner for the bulletin board which is to be bought in the Union. This board is to contain a list of all the members of the different classes, so that when there are besides the names of those who have paid their class dues. About three-quarters of the class have paid their dues, so the others had better hurry up and pay up or their names will appear unchecked.
HARVARD CHESS TEAM DEFEATS TECH CLUB
Technology Team Unable to Win From Crimson's Best Material.
Last Saturday evening the Tech Chess Club met the Harvard team in a game. This match was the second of a series of games being played by the two teams. The first match was played only recently, and resulted in a Tech victory for the red team over Harvard, apparently determined to prevent the repetition of an unsuccessful match on their part, such a game which represented not only the Harvard team, but also some of Harvard's best material in chess.
The members of the Technology team found no easy opponent in the Crimson team, and the playing was very close. The men of both teams displayed excellent form. The Tech men had evidently found their equals. They made a hard struggle to complete the good work with which they opened, and chances of continuing the struggle as they might luck was gaining them, and the best they succeeded in doing was to obtain one draw. The luckiest man, C. E. Nordwood, '14, whom his opponent, E. F. Lange, was unable to defeat.
The tactics of the better members of the team were defended, but in nearly every played matches. The following men represented Harvard: C. E. Nordwood, C. F. Parker, H. B. Valkar, D. M. Beers, F. M. Currier and E. R. V. Hurley; and defeated Re (Continued to Page 2.)
SPRINGFIELD TO GET N. E. I. C. A. A. GAMES
On Condition That Rent of Field Is Lowered—May Be Held at Tech Field.
At a meeting of the Executive Committee of the N. E. I. C. A. A., which was held Saturday evening in the Copley-Place Hotel, it was decided that the New England Inter-collegiate track championships will be held in Springfield, Mass., in April. However, this decision is neutral, and if the expenses of holding the meet at Springfield cannot be considerably reduced, the meet may be held at Tech Field. The meets have been held at Springfield the past two years, and the expenses have been nearly equal to the receipts in each case. Therefore, the committee decided that the championships will be awarded to Springfield only on the condition that the rent of Pratt Field, where the games are held, is lowered. The Tech authorities have been making efforts to secure the games for Tech Field, and have been letting the members of the Association requesting their support in the matter. However, a majority seems to favor holding the meet at Springfield. Tech will only secure it in case of a refusal to lower the rent of Pratt Field. The next meeting will be held on Sunday and Saturday, May 23 and 24. This date will be very suitable to the Tech team, since it is almost a week previous to the time set for the final examinations.
Since Harry Adams, who had been referee for many years, has not declared himself for or against the new date, it was decided to appoint Frank H. Bigelow, the former Harvard captain, as referee. The Tech team will play the meets at Worcester for many years.
The summary:
| Date | Teams |
|------------|------------------------|
| March 1 | Milton H. S. |
| March 1 | O'Hearn, Scully, r.f...1.g. Stanley (Adams) Fox, c. Leary McBride, l.s...1.g. Kelly (Cushing) Dutton, r.g...1.f. Kelly (Cushing) |
| The score: | Tech, 15; 21; Milton H. S. 19. Goals from floor: Carder, 2; Leary, 3; O'Hearn, 2; Smith, 3; McBride, 2; Dutton. Goals from fouls: Kelly, 2; Morse, 3; Leary. Referee, Duggan. Timer, P. Mulvey. Time, 26-15 minute periods and 6 minutes overtime period. |
E. E. SOCIETY TRIP
Two trips to the South Boston power stations of the Boston Elevated will be taken by the Electrical Engineering Society on Tuesday and Wednesday, March 11 and 12 respectively. This plan is similar to the one carried out last year to enable the men to get a better idea of the work done by having their men go on each trip. This is also of advantage to the men who cannot go on Tuesday and can spare the time on Wednesday. The exact time and place will be announced later in tomorrow's issue of THE TECH.
MEMORIAL MEETING IN COPELY HALL TONIGHT
In Memory of Prof. Despradelle, Architect of Genius and Renown.
Tonight, in Copley Hall, there is to be held a memorial meeting in memory of the late Professor Désiré Despradelle. The meeting will be under the auspices of the Boston Society of Architects, the Massachusetts Institute of Technology, and the Boston Architectural Club. At the meeting an exhibition of the works of Prof. Despradelle will be opened in Copley and Alston Halls. This exhibition will remain open to the public until Saturday evening, March 15. Because of the limited space the admission will be by ticket, but the exhibition will be open to the public.
The memorial meeting will be under the care of R. Clifton Sturgis, president of the Boston Society of Architects, and President Maurin of Technology will be the first speaker. Two clergymen, Rev. W. B. Kling and Rev. J. H. B. Smith, warm friends of Professor Despradelle, will speak of his personal qualities, while two brothers architects from New York will give him a professional standpoint. These are Whitney Warren, of Warren & Wetmore, architects of the New Grand Central Station, and Thomas Higginson, of Carrère & Hastings, who planned the New York Public Library.
It is interesting that the associations and Technical Club have an exhibition of the drawings of this remarkable man are largely educational. Aside (Continued to Page 2.)
1915 BASKETBALL.
Sophomores Defeat Milton High In Fast Overtime Game.
Saturday afternoon, at Milton, the Sophomore basketball team defeated the Milton High School team by the score of 31 to 19. As indicated by the score, the game was an exceedingly close one, and an average period of six minutes was required to break the tie. Carder and Leary did well for the Tech, while McBride and Scully starred for the Sophomores.
The summary:
| Date | Teams |
|------------|------------------------|
| March 1 | Milton H. S. |
| March 1 | O'Hearn, Scully, r.f...1.g. Stanley (Adams) Fox, c. Leary McBride, l.s...1.g. Kelly (Cushing) Dutton, r.g...1.f. Kelly (Cushing) |
| The score: | Tech, 15; 21; Milton H. S. 19. Goals from floor: Carder, 2; Leary, 3; O'Hearn, 2; Smith, 3; McBride, 2; Dutton. Goals from fouls: Kelly, 2; Morse, 3; Leary. Referee, Duggan. Timer, P. Mulvey. Time, 26-15 minute periods and 6 minutes overtime period. |
1915 TECHNIQUE.
Meeting Tomorrow at 5 O'clock in 21 Rogers.
McC. Werlich, chairman, ex-officio, of the 1915 Technique Electoral Committee, announces again that the first meeting of the new elected body will be held at 5 o'clock, Tuesday, March 11, in 21 Rogers. All twenty-seven men who make up the committee should be present at this meeting, as the aims, objects and idea of the Technique will be outlined by Lawrence C. H. H. Selby, '15, of 1913 Technique. Also a committee to draw up a constitution will be appointed. Another meeting will probably be held Monday to consider and discuss the constitution.
WEATHER.
Forecast for Boston and vicinity: Monday probably fair and warmer; moderate to brisk southwest winds.
CALENDAR.
Monday, March 10, 1913.
4.15—Tech Shop Meeting—Union
Tuesday, March 11, 1913.
1.00—1916 Class Meeting—Huntington Hall.
5.00—1915 Technique Meeting—21 Rogers
Electrical Engineering Society Trip.
Wednesday, March 12, 1913.
5.00—Meeting of Cercle Francais—29 Lowell
5.00—Catholic Club Meeting—Union.
7.00—1915 Track vs. Boston College High—Gym.
Electrical Engineering Trip.
Do you patronize our advertisers?
SHOW REHEARSALS TO COMMENCE TODAY
Coach Sanger to Start Active Work—First Cut in Cast Tonight.
Saturday afternoon at 2 o'clock, a tailor from C. W. Ware & Co., of 115 Broad street, started measuring all the men in the cast and chorus for their costumes. The dressmaker has awarded the contract because of the fact that their prices are lower than any other firm.
The first lot of costumes will be ready on or about Friday of this week, as soon as pictures will be taken soon after that date as possible. The remainder, however, will be ready later and will not hinder the progress of the photographs.
Mr. Ernest Sanger will be present at the cast rehearsal this afternoon, at 4:15 o'clock, to take charge from that time on. The men will have many lines at all will not be allowed by Coach Sanger to read their parts. The coach will make his first cut this afternoon, and most of them will not be called as principals will be used in the chorus as understudies.
Mr. McCrady will have charge of the chorus rehearsal to be held also at 4:15. All men in the chorus must bring their own, as no practice will be given without this. Stage Manager Flase has not as yet arranged a schedule for this week's rehearsal, but this program will be announced later.
CHESS MATCH
(Continued from Page 1.)
respectively, W. Grant, W. E. Merrill, '13; H. T. Chandler, '14; L. B. Hoyt, '13; J. P. Pratt, '14; and R. D. Brown, '14. F. Fanning of Harvard, tied with C. E. Norwood, '14, of the Tech Chess-Club.
However, by their defeat last Saturday, the Tech Club will play the third game of the series against the Harvard team. All the sports of the men's team will be bent toward endeavoring to inflict a defeat upon the Harvard team, and the next match promises to be a very lively and interesting one.
COMMUNICATION.
To the Editor of THE TECH:
Dear Sir:—Since writing the communication which you so kindly published on March 7, I have a letter from Professor Weston, the Secretary of the Intercollegiate Peace Association. He says that "it will not be necessary to hold the convention before April 18, or perhaps, in case of need, a few days later. The local convention of the club may be held at the end of the first or the beginning of the second week in April. As the organization wishes to make a start at once, should you reply to Professor Weston till Thursday, March 13, and, thus, if there are no applications, to indicate that the institute will not be represented.
Yours very truly,
A. T. Robinson.
MEMORIAL MEETING.
(Continued from Page 1.)
from the fact that the showing will be important in that the sketches and plans will reach the press. The space of the two walls, the architects will show to thinking people how the plans of their buildings are evolved. It is popularly supposed that the architect jumps at once to the drawing of his personal or his exterior, but he truth is that many smaller and many smaller ideas are worked out independently to be brought together into the completed whole.
The principle which underlies the teaching methods of a successful school; it is known to be the basis of instruction at the Ecole des Beaux Arts and how completely it was followed at Tech-
MILLION DOLLAR SHOW.
A "million dollar" art engine exposition has just closed at the University of Wisconsin. The show was divided into four sections, the farm power engine, the automobile, the shop, the motorcar race division and the motorboat section. Large crowds of people viewed the exhibit and the numerous women visitors seemed as much interested as the men.
The Freshman meet with the Boston College High marks the close of the athletic season, and the men hope to begin work at the field in less than a week.
Company B is said to have been told by their captain that it was a matter of luck to have been excused from drill Wednesday, but if so the next month will tell.
The American Brass Company
ANSONIA BRASS & COPPER BRANCH
Ansonia, Connecticut
Manufacturers of
Bare and Insulated Copper Wire and Cable
Drawn Copper Rods, Bars and Strips.
Brass Sheets, Rods, Wire and Tubes.
Sole Manufacturers of
TOBIN BRONZE"
(Trade Mark Registered)
If you will give us your order early, Why a discount FOR THIS REASON? To distribute our work over a greater period. It is impossible for us to take care of our trade when confined to a few weeks known as the "rush season," hence this offer.
Our Spring and Summer Wooolsens are now on our counters, and exceed in quality, beauty and variety of design anything we have ever shown.
BURKE & CO. Incorp. Tailors
18 SCHOOL STREET and 843 WASHINGTON ST., BOSTON
7 MAIN ST., ANDOVER, MASS. HANOVER, N. H.
McMORROW
College Shoes for College Men:
2385 Washington Street, Boston, Mass.
OPPOSITE YOUNG'S HOTEL.
Architectes Français: 1884, First Rougave, Deschaumes and Edmond Rougeve, 1886, Architect, diploma from the Government School, first Second Grand Prix de Rome, which was the highest rank; 1889, Lauréate of the Institute de France; and the Grand Prix International des Buildings and National Palaces; 1900, was made officer of the French Academy; 1901, member of the Salon, Paris, making him hors concours.
He came to the Massachusetts Institute of Technology in 1890, and became Roth Professor of Architecture, and within a few months of his death was named director of the department. In 1910-11, special lecturer at Harvard University on Architectural Design.
Are you a regular subscriber to THE TECH?
IMPORTANT MEETING OF "CERCLE FRANCAIS"
New Constitution To Be Acted Upon—French Play Will Be Started.
Wednesday afternoon, at 5 o'clock, the Cercle Francais will hold the second of its regular meetings in Room 102, on the third floor. As originally planned, the club will start a regular program, and on account of some minor business which is to be discussed at that time, the members are urged to be present.
The special committee which has been drawn up for the club will present it for the final ratification of all the members. After voting on one member and attempting to settle other business, the reading of a French play will be begun. Palliseron's La Grande ou le St'Enfant, which is considered to be one of the best of the present day French short dramas, as been selected to start on. Parts will be assigned to different members, readings will be made, and general discussion of the play will follow.
It was decided upon at the first meeting, the meeting being conducted in French, and it is desired that all members and those who have signed the invitation to joining attend its meeting. The treasurer desires all the men to come prepared to pay their dues, which are $1.00 per year. The next meeting of the society will be on Wednesday, March 19.
INTERCLASS MEET.
(Continued from Page 1.)
ment was almost as much of a farce as in the open interclass meet, and many of the men showed versatility. The men who have any desire to help their class and the Institute should remember this and at least make a try for the team in this event.
The fact that no admission was charged had some effect, as about fifty of the fellows attended the meet. The first event was the 35-yard dash, which was won by Adams, '16, easily winning the first heat. Smith, '15, snatching second heat from Ross, '16. Thomas, '15, was away right with the gun in the third heat, and beat his teammate, in the fourth heat, and fastest man of any of the meets, 4.35 seconds, omis and Comiskey took first and second, respectively, in the third heat. The 100-yard dash was a very close semi-final heat, each being closely pressed by their opponents. Curtis and Smith won the semi-finals, the Sophomores, Loonis, '16, had a bad start, but became strong at the finish and managed to beat Smith himself. In the final, the men went away well together, and sped to the door in a line, Adams having the advantage if such there was, a scant six inches. At the finish two Freshmen entered forges their ahead, and Adams, '16, went to a fine place. However, he threw himself for the tape a few feet too soon and was nosed out by Loonlis, '16, who won for third. O'Neill, '16, took the lead at the start of the mile, and held it until the last lap. Then Bennett, '16, who had already crept up from seventh place, took the lead, and set such a fast pace thereafter, he was never in danger. The fight for the place of honor was very close. Zeppler, '16, took second on the sixth lap, and had the lead on the seventh lap, but was overtaken by Bennett, '16, who won the race. The Freshman team was second, only passed by White at the fourteenth lap, only to be passed by White at the next lap. Thereafter, Bennett ran the last few yards of Zeppler, and it seemed though the latter was doomed, but released a splendid sprint to the two laps, and although Hastie had upon him, he was unable to beat him to the line.
Johnny O'Brien led for about ten yards in the quarter after which he was passed by Curtis, who retained the lead throughout the remainder of the race. Adams was second, but was surpassed by Reed, who forced Curtis hard but was unable to pass him. Adams easily won third from O'Brien, and the other heat of the Freshmen men distinguished themselves by knocking over all three hurdlers. Adams, '16, went in and into the air, Curtis easily won first, then, Teson beating Adams out for second. The Freshmen took the lead at the start in the half, and a fast pace. His position was never endangered. Dean, Hill, and Rooper fought hard for second place, and when Rooper obtained second place soon after the start and could not be ousted. Dean was third, closely followed by Rooper, '16.
Teson and Fribus, both Sophomores, tied for first in the high jump, but were going to the second heat. Clarke, '16, who was looked upon to take second, was unable to place. Reason, '16, won the shot-put for the Freshmen, Fribus, '16, and O'Brien, '15, winning second and third, respectively. The efforts of the rest of the entire team were brilliant. The Sophomores had an easy time in the shot-put. O'Brien won the teammate, Clarke, second. Walk obtained third place for the Freshmen. As in the pole vault, many of the men who were shot-putters were unable to move native at the same time, and the Freshman team had no one who was capable of forcing the second try. It was at the fourteenth lap, only to be passed by White at the next lap. Thereafter, Bennett ran the last few yards of Zeppler, and it seemed though the latter was doomed, but released a splendid sprint to the two laps, and although Hastie had upon him, he was unable to beat him to the line.
DOCTOR GIFFORD TO GIVE T.C.A. TALK
"God's Building" Subject For Popular Speaker Next Thursday.
The T.C.A. has been fortunate in securing Dr. O. P. Gifford for its regular Thursday talk this week. As many of the fellows know, Dr. Gifford is one of the most prominent ministers in Massachusetts. He is pastor of the Brookline Baptist Church, and is particularly noted for his opposition to the appropriation of State money for the support of children's schools.
Gifford's first talk to Tech students was given a year ago last Fall, on the subject, "A City Without a Church." Such a timely subject was enjoyed by all who heard it. Many of the fellows enjoyed it so much that they asked to hear the Doctor again. Since then, he has given two more talks, each as successful as the first. The T.C.A. has every reason to expect that this Thursday speech will be as well received as his previous talks. Those who have attended Dr. Gifford's previous talks should listen for this one, and the others should think twice before they let slip an opportunity of hearing this popular speaker. The subject this week is "God's Building," but experience has shown that the Doctor will have some interesting stories to tell, for his words seldom give away the character of his talks. But as he is famous for his epigrams we can certainly predict a large supply of good sayings.
CHEMICAL SOCIETY.
Professor Passano will speak before the Chemical Society Friday evening, March 13, on the subject of "Scientific Management." Those who have been in Professor Passano's classes are confident that he will have some interesting points upon his subject. The meeting will begin at 7:45 P.M.
PRESS & HERBERT TAILORS
Announces the arrival of a complete Spring and Summer Stock of High Grade Fabrics.
We are making a specialty of full dress suits. A suggestion to you. Come up and see about a new dress suit, in time for the Junior Week Parties. We can handle you right for both quality and price. Tel. B.B. 2087.
486 Boylston Street
KEEP YOUR PRELIMINARY THESIS NOTES NEAT AND READABLE while fresh in your mind. It is hard to read some of those scratchy notes after the ink grows cold. What's the solution?
RENT A TYPEWRITER. What Typewriter?
THE OLIVER;
which has been in use exclusively in THE TECH offices for the past two years and has satisfactorily fulfilled every condition that has been imposed upon it.
You can see Estabrook, the Oliver man, any time you like in THE TECH office. Make a date with him.
See the "Styles" at 146 Congress Street
THE OLIVER TYPEWRITER COMPANY
HOTEL BRUNSWICK
EUROPEAN AND AMERICAN PLANS
HERBERT H. BARNES, Proprietor
collegiates this Spring, and in the next few years, it is necessary that there be more men trying for the field events. Couch Kanally will be glad to see any man who has any ability or desire to make good in any of these events.
The summary was as follows:
Shot-put—Won by C. Loonis, '16; second, W. Adams, '16; third, E. Smith, '16. Time, 44.5 seconds.
Pole vault—Won by J. Zeppler, '16; second, J. Hastie, '16. Time, 5 minutes, 4 seconds.
High jump—Won by J. O'Brien, '15; second, R. Ober, '16. Height, 5 feet 6 inches.
High jump—Won by E. Teson, '15; second, J. O'Brien, '15; third, W. Hill, '16. Distance, 31 feet.
Pole vault—Won by L. Laurason, '16; second, F. Fribus, '15; third, J. O'Brien, '15. Height, 9 feet 6 inches.
The Technique posters around upon the various bulletin boards are among the most prominent. They are all very fine pieces of work.
The Freshmen will now prepare to get revenge at the annual Spring track meet.
BOSTON OPERA HOUSE
MON., 8 to 11. AIDA, Melis, Guy, Zemaitello, Rossi, Cond., Moranzoni.
WED., 8 to 11. SAMSON ET DALILA, Guy, Zemaitello, Riddle, Marfoinesi, Cond., Andre-Caplet.
FRI., 7:30 to 11. THE BLUE FOREST, Melis, Fisher, Swartz, Amenden, De Potter, Riddle. Followed by THE SECRET OF SUZANNE, Nielsen, Scotti, Devaux, Cond., Andre-Caplet.
SAT., 2 to 5. FAUST, Nielsen, Swartz, Amenden, Martin, Segurcho, Conda Strong.
SAT., 8 to 11:20. ONLY APPEARANCE OF LINA CAVALIERI. Also, LUCIEN MOUTON in CARMEN, Melis, Fisher, Muratore, Marfoinesi, Cond., Andre-Caplet.
SUNDAY ORCHESTRAL CONCERTS
8:15 to 5:30.
March 16, RUBINSTEIN, Ganz, Soloist. Downtown Office, Steinert's, 162 Boylston St.
Mason and Hamlin Pianos Used. Address Mail Orders to Box Office.
TUESDAY EVENING, MARCH 18. ONLY APPEARANCE OF ENRICO CARUSO. I. PAGLIACCI. Tickets now on sale.
The London Shop
MEN'S FURNISHINGS
478 Boylston Street
Our Spring Suits and Overcoats will be ready early in February also Spring Hats, Haberdashery, Shoes and Leather Goods at our Boston Sales Offices:
LAWRENCE BUILDING
TREMONT AND WEST STREETS
Send for Illustrated Catalogue
Ask for WARD'S STATIONERY
Wardwove Writing Paper and Envelopes. Fraternity Stationery. Engraved Invitations, Reception and Visiting Cards. Banquet Menus. Coats of Arms, Crests and Monograms, Students Supplies.
WARD'S
57-63 Franklin St.
BOSTON.
MARCEAU
Photographer
160 Tremont St.
Special rates to Students
Young Men's Hats for Every Occasion
COLLINS & FAIRBANKS COMPANY
383 WASHINGTON STREET - BOSTON
CLASSIFIED ADVERTISEMENTS AND NOTICES
Faculty Notices free.
Activities, 4c per line.
Individuals, 10c per line, 6 words to a line
LOST—February 28, in Rogers, or Eng. A, a Waterman Safety Pen. Return to "Cage." Reward. C. A. Higdon.
(Pr-14)
LOST—In Union, 1913 Class Pipe. Please return to "Cage."
(Pr-14)
LOST—Wednesday, in the Gym. Gold ring with green stone. Return to Cage and receive reward.
WILL the man who took the wrong raincoat in the Union Dining Room on Thursday, February 27, leave a note at the Cage for M. J. Sayward, 1914?
THIRD YEAR.
Section 2 of Courses I and XI will meet Thursday afternoons, from 2 to 4, in Mr. Room 50, Engineering Building B, for exercises in Structural and Field Geology.
WALTER HUMPHREYS,
Registrar.
To M. I. T. Men
who do not eat at the
TECH UNION DINING ROOM
We want your opinion.
Let us know what you think of our food and service. Then tell us what YOU want for
"THE PLACE ACROSS THE RIVER."
Richard's School of Dancing
Class for beginners Monday, Advanced pupils Saturday. So-called "easy" classes, also lessons by appointment daily. Learn the latest dance, the only dance that is really popular for the Promenade. 30 Huntington Ave. :::: Boston
Hinds Hand Laundry Co.
3 APPLETON STREET
Tol. Tremont 850
6-7 Branch Office
Copley, Plaza Haberdasher
16 HUNTINGTON AVE.
ONE DAY SERVICE
Oak Grove Creamery Co.
431-437 Boylston Street
Most Homelike Restaurant in Boston
Open from 6:30 A.M. to 8:00 P.M.
LUNCH ROOM UP ONE FLIGHT. ELEVATOR
Quickest possible service from 11:30 to 3
The Original
TECH BARBER SHOP
Increased Business, Increased Help. Everything Sterilized
Copley Square
Second Floor
Students' used clothing and other personal effects bought by
KEEZER
360 COLUMBUS AVENUE
(Near Dartmouth St.)
Highest prices paid for same.
Toll. 9146. Phone. Write or Call
Open Evenings to 9 o'clock
M. LEVENGSTON
Dress
Prince Albert
and Tuxedo Suits
To Let
12 School Street, Boston
Room 1
Telephone, 3315-J Main
HERRICK, COPLEY SQUARE
Choice Seats for all Theatres
Phone B. B. 2325
Key Number Connecting Five Phones
Charles Wesley Hearn
Artist Photographer
729 Boylston St.
Official Photographer for PORTFOLIO 1912 & 1913
Most reasonable rates for excellent work.
Have your picture taken now and exchange with your friends in the Senior Class who are soon to leave you.
E. F. MAHADY CO.
Prescription Opticians
SHELL SPECTACLES
AND EYEGLASSES
671 Boylston St.
Near Public Library
Kodaks, Supplies, Developing and Printing.
All Goods Required by Students at
MacLachlan's
502 Boylston Street
Drawing Instruments and Materials, Fountain Pens Text-Books
THE ADVANTAGES WE OFFER
PROGRESSIVE methods, large resources and two conveniently located, thoroughly modern bank buildings, combine to make the Old Colony Trust Company the most desirable depository in New England.
Interest-bearing accounts subject to check; may be opened either at Court Street or Temple Place, and the two offices used interchangeably for the transaction of your banking business.
You are cordially invited to inspect the modern Safety Deposit Vaults at either office.
OLD COLONY TRUST CO.
Main Office
Court Street
Branch Office
Temple Place |
COSNet: An R Package for Predicting Binary Labels in Partially-Labeled Graphs
Marco Frasca and Giorgio Valentini
April 30, 2024
Dipartimento di Informatica
Università degli Studi di Milano
Via Comelico 39/41, 20135 Milano, Italy
email@example.com
Scope and Purpose of this Document
This document is a user manual for COSNet, the software implementing the model developed by Bertoni et al. (2011), Frasca et al. (2013). It provides an introduction into how to use COSNet. Not all features of the R package are described in full detail. Such details can be obtained from the documentation enclosed in the R package.
Contents
1 Introduction 2
2 COSNet 2
3 An Example of the Usage of COSNet for the Functional Classification of Yeast Genes with the Functional Catalogue 3
3.1 Data Loading ........................................... 4
3.2 Predicting Labels for Unlabeled Nodes with COSNet .... 5
3.3 Processing the Results .................................. 6
3.4 Assessing COSNet Performance .......................... 6
4 An Example of the Usage of COSNet for Inferring Gene Ontology Labels for Fly Genes 7
4.1 Data Loading ........................................... 7
4.2 Predicting GO Labels with COSNet ...................... 8
4.3 Result Evaluation ....................................... 8
1 Introduction
This package implements the algorithm COSNet (Bertoni et al. 2011, Frasca et al. 2013), which has been proposed for predicting node labels in partially labeled graphs, especially when labelings are highly unbalanced. In this context, nodes represent the instances of the problem, whose labels can be positive (+) or negative (-). Unbalanced labeling means that one class (usually the negative class) considerably outnumbers the other one. Many real world problems are characterized by few positives and much more negatives, such as the gene function prediction, where genes having the most specific biomolecular functions are very few (Ashburner et al. 2000), or in medical diagnosis of cancer, where patients having a certain cancer (positive class) are the large minority. The instances are only partially labeled, and the aim is to extend the labeling to all the instances. In this context, imbalance-unaware algorithms may suffer high decay in performance when classifying new instances (Ling and Sheng 2007). COSNet automatically learns from the input data the model parameters able in dealing with the label imbalance, and efficiently infers binary labels for the unlabeled instances in the graph.
Formally, the input of COSNet is represented by a weighted graph $G = (V, W)$, where $V$ is the set of nodes and $W = (w_{ij})_{i,j=1}^n$ is the symmetric weight matrix with null diagonal: the weight $w_{ij} \in [0, 1]$ denotes a similarity index of node $i$ with respect to node $j$. The labeling of $V$ in positive $V_+$ and negative $V_-$ nodes is known only for a subset $S \subset V$, while is unknown for $U = V \setminus S$. The aim is to extend the labeling to nodes in $U$, that is inferring a bipartition of $U$ in positive $U_+$ and negative $U_-$ instances. The labeling imbalance can be represented through a coefficient $\epsilon = |S_+|/|S_-|$, where $S_+$ and $S_-$ are the sets of positive and negative examples, respectively. The labeling is considered highly unbalanced when $\epsilon << 1$.
2 COSNet
COSNet is a binary classifier based on a parametric Hopfield network which embeds the partial labeling and node similarities $W$ and predict the binary labels for the unlabeled nodes through an asynchronous dynamics, which
updates only the unlabeled neurons. In order to deal with label imbalance, the model introduces two parameters, $\alpha \in [0, \frac{\pi}{2}]$ and $\gamma \in \mathbb{R}$, determining respectively the neuron activation values $(\sin \alpha, -\cos \alpha)$ and the neuron activation thresholds. The parameters are automatically learned by an efficient supervised procedure on the basis of the input data. The initial state for neuron $v \in V$ is set as follows:
$$x_v(0) = \begin{cases}
\sin \alpha & \text{if } v \text{ is positive labeled} \\
-\cos \alpha & \text{if } v \text{ is negative labeled} \\
0 & \text{if } v \text{ is unlabeled}
\end{cases}$$
For each unlabeled node $i$, the initial state $x_i = 0$ is changed to $-\cos \alpha$ or to $\sin \alpha$ according to the following asynchronous dynamics:
$$x_i(t) = \begin{cases}
\sin \alpha & \text{if } \sum_{j=1}^{i-1} W_{ij} x_j(t) + \sum_{k=i+1}^{n} W_{ik} x_k(t-1) - \gamma > 0 \\
-\cos \alpha & \text{if } \sum_{j=1}^{i-1} W_{ij} x_j(t) + \sum_{k=i+1}^{n} W_{ik} x_k(t-1) - \gamma \leq 0
\end{cases}$$
where $n = |V|$ and $t$ is the current time. At each time $t$, the state of the network is $\boldsymbol{x}(t) = (x_1(t), x_2(t), \ldots, x_n(t))$, and a Lyapunov state function named energy function is associated to the network:
$$E(\boldsymbol{x}) = -\frac{1}{2} \sum_{i,j=1 \\ j \neq i}^{n} W_{ij} x_i x_j + \sum_{i=1}^{n} x_i \gamma$$
The dynamics converges to an equilibrium state $\hat{\boldsymbol{x}}$ corresponding to a minimum of $E$ (Frasca et.al 2013), which is used to infer the bipartiton $(U_+, U_-)$ of $U$: $U_+ = \{i \in U, \hat{x}_i = \sin \alpha\}$ and $U_- = \{i \in U, \hat{x}_i = -\cos \alpha\}$. Furthermore, COSNet can been adopted as ranker by assigning to each neuron a score related to its internal energy at equilibrium. More precisely, the score assigned to neuron $i \in U$ is the following:
$$s(i) = \sum_{j \neq i} (W_{ij} \hat{x}_j - \gamma) \quad (1)$$
### 3 An Example of the Usage of COSNet for the Functional Classification of Yeast Genes with the Functional Catalogue
In this section we apply COSNet to predict the functions of yeast proteins. We adopt a binary protein-protein interactions data set of 2338 yeast proteins from the STRING data base (von Mering et al. 2002), contained in the R
package *bionetdata* and the corresponding Functional Catalogue (FunCat) annotations.
### 3.1 Data Loading
First, let us load the library and check the data:
```r
> library(COSNet);
COSNet: Cost-Sensitive algorithm for binary classification in graphs.
> library(bionetdata)
> data(Yeast.STRING.data)
> dim(Yeast.STRING.data)
[1] 2338 2338
> rownames(Yeast.STRING.data)[1:10]
[1] "YJR121W" "YAL009W" "YGR165W" "YLR298C" "YOL052C" "YOL051W" "YPR083W"
[8] "YOR110W" "YGR032W" "YML015C"
The named squared matrix *Yeast.STRING.data* contains 1 in position \((i, j)\) if the corresponding proteins interact, 0 otherwise. In the same package, in order to define the binary labels, we load the annotations of 176 FunCat classes for the proteins included in *Yeast.STRING.data*. Annotations refer the funcat-2.1 scheme, and funcat-2.1 data 20070316 data, available from the MIPS web site.
```r
> data(Yeast.STRING.FunCat)
> dim(Yeast.STRING.FunCat)
[1] 2338 177
> rownames(Yeast.STRING.FunCat)[1:10]
[1] "YJR121W" "YAL009W" "YGR165W" "YLR298C" "YOL052C" "YOL051W" "YPR083W"
[8] "YOR110W" "YGR032W" "YML015C"
> colnames(Yeast.STRING.FunCat)[1:10]
[1] "00" "01" "01..01" "01..01..03" "01..01..06"
[6] "01..01..06..05" "01..01..09" "01..02" "01..03" "01..03..01"
```
---
\(^1\) The package can be downloaded at http://cran.r-project.org/web/packages/bionetdata/index.html.
\(^2\) http://mips.gsf.de/projects/funcat
The number of columns is 177 because the authors added a dummy class "00". Even in this case, Yeast.STRING.FunCat is a binary matrix whose $i,j$-th component is 1 if protein $i$ is annotated with class $j$, 0 otherwise. Note that the row names of both Yeast.STRING.FunCat and Yeast.STRING.data are identical. We first exclude the dummy class, that is useful only when hierarchical computations are performed, and then we select some classes and, since COSNet needs $\{1,-1\}$-labels, we change each 0-component of Yeast.STRING.FunCat to -1.
```r
> ## excluding the dummy "00" root
> to.be.excl <- which(colnames(Yeast.STRING.FunCat) == "00")
> Yeast.STRING.FunCat <- Yeast.STRING.FunCat[, -to.be.excl]
> ## choosing the first 35 classes
> labeling <- Yeast.STRING.FunCat[, 1:35]
> ## number of positive labels
> colSums(labeling)
01 01.01 01.01.03 01.01.06 01.01.06.05 01.01.09
859 170 32 45 27 56
01.02 01.03 01.03.01 01.03.01.03 01.03.04 01.03.16
49 161 50 26 38 57
01.03.16.01 01.04 01.05 01.05.02 01.05.02.04 01.05.02.07
36 239 277 56 26 54
01.05.03 01.05.25 01.06 01.06.02 01.06.02.01 01.06.06
35 78 122 33 25 21
01.07 01.07.01 01.20 02 02.01 02.07
119 87 39 196 27 23
02.10 02.11 02.13 02.13.03 02.19
24 24 65 34 35
```
```r
> Yeast.STRING.FunCat[Yeast.STRING.FunCat == 0] <- -1
```
### 3.2 Predicting Labels for Unlabeled Nodes with COSNet
Now we predict the node labels through a 5-fold cross validation procedure implemented by the function cosnet.cross.validation provided by the COSNet package. This procedure at each time hides the labels in a fold and predicts them with COSNet.
```r
> out <- cosnet.cross.validation(labeling, Yeast.STRING.data, 5, cost=0)
```
Note that the cost parameter of COSNet is set to 0: this means that the unregularized version is adopted. Now we test the regularized version of COSNet on the same data:
```r
> out.r <- cosnet.cross.validation(labeling, Yeast.STRING.data, 5, cost=0.0001)
```
3.3 Processing the Results
The output of the `cosnet.cross.validation` function is a list whose fields are: "labels", named matrix containing the input labels; "predictions", named matrix containing the binary predictions; "scores", named matrix containing the predicted scores according to Eq. (1).
```r
> predictions <- out$predictions
> scores <- out$scores;
> labels <- out$labels;
> predictions.r <- out.r$predictions
> scores.r <- out.r$scores;
> labels.r <- out.r$labels;
```
3.4 Assessing COSNet Performance
We now evaluate the performance of COSNet in terms of F-score, Area under the ROC Curve (AUC) and in terms of Precision at $x$ Recall level (PxR). We use the R package `PerfMeas`, which provides functions to compute the performance measures we need:
```r
> library(PerfMeas);
> ## computing F-score
> Fs <- F.measure.single.over.classes(labels, predictions);
> ## Average F-score
> Fs$average[4]
F
0.3426235
> Fs.r <- F.measure.single.over.classes(labels.r, predictions.r);
> # Average F-score for the regularized version of COSNet
> Fs.r$average[4]
F
0.373728
> ## Computing AUC
> labels[labels <= 0] <- 0;
> labels.r[labels.r <= 0] <- 0;
> auc <- AUC.single.over.classes(labels, scores);
> ## AUC averaged across classes
> auc$average
[1] 0.7150491
```
> auc.r <- AUC.single.over.classes(labels.r, scores.r);
> ## AUC averaged across classes for the regularized version of COSNet
> auc.r$average
[1] 0.7187172
> ## Computing precision at different recall levels
> PXR <- precision.at.multiple.recall.level.over.classes(labels,
+ scores, seq(from=0.1, to=1, by=0.1));
> ## average PxR
> PXR$avgPXR
0.1 0.2 0.3 0.4 0.5 0.6 0.7
0.55186943 0.48994551 0.39963257 0.28602119 0.18971808 0.10466840 0.07222500
0.8 0.9 1
0.06004248 0.05153318 0.04181462
> PXR.r <- precision.at.multiple.recall.level.over.classes(labels.r,
+ scores.r, seq(from=0.1, to=1, by=0.1));
> ## average PxR for the regularized version of COSNet
> PXR.r$avgPXR
0.1 0.2 0.3 0.4 0.5 0.6 0.7
0.61817169 0.52044172 0.43354676 0.31842160 0.20956440 0.11586193 0.07029459
0.8 0.9 1
0.05933335 0.05205932 0.04182582
4 An Example of the Usage of COSNet for Inferring Gene Ontology Labels for Fly Genes
In this section we show an application of COSNet in predicting the Gene Ontology (GO) (Ashburner et al. 2000) labels for 9361 Drosophila melanogaster genes. The input similarity matrix is obtained by integrating several types of data, including co-expression, genetic interactions, protein ontologies and physical interactions. The Gene Ontology annotations (release 15-5-13) with 3-300 annotated genes have been considered. Both data and annotations can be downloaded at http://frasca.di.unimi.it/cosnetdata/.
4.1 Data Loading
> ## reading similarity network W
> W <-
+ as.matrix(read.table(file=paste(sep="", "http://frasca.di.unimi.it/",
+ "cosnetdata/u.sum.fly.txt"), sep=" "))
> ## reading GO annotations
> GO.ann.sel <-
+ as.matrix(read.table(file=paste(sep="", "http://frasca.di.unimi.it/",
+ "cosnetdata/GO.ann.fly.220.127.116.11_300.txt"), sep = " ",))
> GO.classes <- colnames(GO.ann.sel)
> ## changing "." to ":"
> GO.classes <- unlist(lapply(GO.classes, function(x){
+ substr(x, 3, 3) <- ":"; return(x)}))
> colnames(GO.ann.sel) <- GO.classes;
4.2 Predicting GO Labels with COSNet
Now we determine a random partition in 3 folds of the input data, hide the labels of the genes in one of these folds (test set), and use the labels in the other 2 folds as training set for COSNet.
> n<-nrow(W);
> ## selecting some classes to be predicted
> classes <- c("GO:0009605", "GO:0022414", "GO:0032504",
+ "GO:0002376", "GO:0009888", "GO:0065003");
> labels <- GO.ann.sel[, classes]
> ## for COSNet negative labels must be -1
> labels[labels <= 0] <- -1;
> ## Determining a random partition for the class GO:0009605 in 3 folds
> ## ensuring that each fold has a similar proportion of positives
> folds <- find.division.strat(labels[, 1], 1:n, 3)
> ## hiding the labels of the test set (the fold of index 1)
> labels[folds[[1]], ] <- 0;
> ## predicting the hidden labels for each class with COSNet
> res <- apply(labels, 2, function(x, W, cost){
+ return(COSNet(W, x, cost))},
+ W = W, cost = 0.0001);
4.3 Result Evaluation
The function COSNet returns a list with five members, including binary predictions, ranking scores, and learned parameters. We now show how to compute, for instance, the AUC and the P10R achieved for the first GO term. Moreover, we show the value learned for the model parameters.
> library(PerfMeas);
> ## last predicted term
> term.ind <- 6;
> scores <- res[[term.ind]]$scores;
> test.genes <- names(scores);
> test.labels <- as.vector(GO.ann.sel[test.genes, term.ind]);
> pos.labels <- sum(test.labels > 0)
> pos.labels
[1] 91
> alpha <- res[[term.ind]]$alpha
> gamma <- res[[term.ind]]$c
> alpha
[1] 1.459915
> gamma
[1] 4.914017e-17
> AUC <- AUC.single(scores, test.labels)
> AUC
[1] 0.7851084
> P10R <- precision.at.recall.level(scores, test.labels,
+ rec.level = 0.1)
> P10R
[1] 0.4545455
5 Usage of COSNet for Predicting Therapeutical Categories of Drugs
This section shows an application of COSNet in predicting drugs categories from DrugBank for 1253 DrugBank drugs. The drug similarity matrix and the corresponding labels are available in the R package bionetdata, and contain respectively the Tanimoto chemical structure similarity scores among the considered drugs and the 0/1 labels for 45 drug categories, where in position $i,j$-th we have the value 1 whether the drug $i$ is associated with the drug category $j$, 0 otherwise.
5.1 Loading the data
> library(bionetdata);
> ## similarity matrix DD.chem.data
> data(DD.chem.data);
> ## label matrix DrugBank.Cat
> data(DrugBank.Cat);
5.2 Learning the Model Parameters
We show now how to learn the COSNet parameters using the package function optimizep. Fixed one drug category, first of all, we determine a random stratified partition of the data using the function find.division.strat provided by the COSNet package, and then hide the labels of one fold and use the other folds as training set.
```r
> n <- nrow(DD.chem.data);
> drugs <- rownames(DD.chem.data);
> drug.category <- c("Cephalosporins");
> labels <- as.vector(DrugBank.Cat[, drug.category]);
> names(labels) <- rownames(DrugBank.Cat);
> ## Determining a random partition in 5 folds ensuring that each
> ## fold has a similar proportion of positives
> folds <- find.division.strat(labels, 1:n, 5)
> labels[labels <= 0] <- -1;
> ## hiding the test labels (the fold of index 1)
> test.drugs <- folds[[1]];
> training.drugs <- setdiff(1:n, test.drugs);
> labels[test.drugs] <- 0;
```
Now we need to project the training nodes into a bidimensional space, step necessary for the learning procedure (Frasca et al. 2013), and we can do it using the package function generate_points, which returns a list with two fields: pos_vect, the named vector of abscissae of projected points and neg_vect, the named vector of ordinates of projected points. Then we can call the function optimizep to learn the parameters of the model.
```r
> points <- generate_points(DD.chem.data, test.drugs, labels);
> str(points)
List of 2
$ pos_vect: num [1:1253, 1] 6.43 7.12 3.75 5.94 1.33 ...
..- attr(*, "dimnames")=List of 2
... ..$ : chr [1:1253] "DB00115" "DB00116" "DB00117" "DB00118" ...
... ..$ : NULL
$ neg_vect: num [1:1253, 1] 211 247 191 188 106 ...
..- attr(*, "dimnames")=List of 2
... ..$ : chr [1:1253] "DB00115" "DB00116" "DB00117" "DB00118" ...
... ..$ : NULL
> opt_parameters <- optimizep(points$pos_vect[training.drugs],
+ points$neg_vect[training.drugs], labels[training.drugs]);
```
5.3 Running the Sub-network of Unlabeled Nodes to Infer Labels
We now extend the learned parameters to the sub-network of unlabeled nodes and run it with the package function `runSubnet`.
```r
> ## alpha parameter
> alpha <- opt_parameters$alpha;
> ## gamma parameter
> gamma <- opt_parameters$c;
> ## optimal F-score achieved during learning phase
> ## procedure (see Frasca et al. 2013)
> Fscore <- opt_parameters$Fscore;
> res <- runSubnet(DD.chem.data, labels, alpha, gamma, cost=0.035);
```
5.4 Extracting Predictions
The output of the `runSubnet` function is a list with three fields: "state" which is a named vector containing the binary predictions; "scores", named vector containing the scores described in Eq. (1); "iter", integer representing the network iterations needed to reach the fixed state. By means of this vectors, we can now compute the prediction performance.
```r
> library(PerfMeas)
> str(res)
List of 3
$ state : Named num [1:251] -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 ...
..- attr(*, "names")= chr [1:251] "DB00123" "DB00133" "DB00138" "DB00140" ...
$ scores: Named num [1:251] -10 -8.02 -6.35 -7.44 -9.17 ...
..- attr(*, "names")= chr [1:251] "DB00123" "DB00133" "DB00138" "DB00140" ...
$ iter : num 4
> res$iter
[1] 4
> labels <- as.vector(DrugBank.Cat[, drug.category]);
> names(labels) <- rownames(DrugBank.Cat);
> test.names <- names(res$scores);
> AUC <- AUC.single(res$scores, labels[test.names]);
> AUC;
[1] 1
> P10R <- precision.at.recall.level(res$scores,
+ labels[test.names], rec.level=0.1);
> P10R;
```
5.5 Cross Validating COSNet in Predicting Drug Therapeutical Categories
> library(bionetdata);
> data(DD.chem.data);
> data(DrugBank.Cat);
> labels <- DrugBank.Cat;
> labels[labels <= 0] <- -1;
> out <- cosnet.cross.validation(labels, DD.chem.data,
+ 5, cost=0.035);
> Fs <- F.measure.single.over.classes(labels, out$predictions);
> Fs$average[4];
F
0.3824548
> labels[labels <= 0] <- 0;
> auc <- AUC.single.over.classes(labels, out$scores);
> auc$average
[1] 0.8180897
> PXR <- precision.at.multiple.recall.level.over.classes(labels,
+ out$scores, seq(from=0.1, to=1, by=0.1));
> PXR$avgPXR
0.1 0.2 0.3 0.4 0.5 0.6 0.7
0.62530488 0.58520807 0.50996769 0.44960582 0.37873653 0.26178586 0.19865210
0.8 0.9 1
0.15501125 0.11415291 0.06054957
References
Bertoni, A., Frasca, M., Valentini, G. (2011). COSNet: A Cost Sensitive Neural Network for Semi-supervised Learning in Graphs. *ECML/PKDD* (1) **6911**, 219-234.
Frasca, M., Bertoni, A., and Valentini, G. (2013). A neural network algorithm for semi-supervised node label learning from unbalanced data. *Neural Networks*, **43**(0), 84–98.
Von Mering, C., Krause, R., Snel, B., Cornell, M., Oliver, S., Fields, S., and Bork, P. (2002). Comparative assessment of large-scale data sets of protein-protein interactions. *Nature* **417** 399–403.
Ashburner, M. et al. (2000), Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. *Nature genetics* **25** (1) 25–29.
Ling, C. X. and Sheng, V. S. (2007). Cost-sensitive Learning and the Class Imbalanced Problem. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.