College and Research Libraries Coordinating Collection Development: The RLG Conspectus Nancy E. Gwinn and Paul H. Mosher Collection development officers in libraries whose parent institutions are members of the Re- search Libraries Group, Inc., have a new collection evaluation tool, the RLG Conspectus. The Conspectus is an overview, or summary, arranged by subject, of existing collection strengths and future collecting intensities of RLG members. It serves as a location device or collections considered as national resources and as a basis for assignment of primary collecting responsibilities. The authors trace the antecedents of the Conspectus to work of ALA commit- tees, the initialRLG Collection Development Committee, and a group known as GNOMES + 2. Descriptions of the data-gathering process and of the online version of the database precede an outline of the benefits, realized and anticipated, to individual institutions, as well as the partnership. The Association of Research Libraries is conducting an experiment to see if the methodology can be extended to non-RLG association members. hen Columbia, Yale, the New York Public Library, and Har- vard formed the Research Li- braries Group, Inc . (RLG) in 1974, they created a rare opportunity for cooperation within the world of research librarianship. RLG was more than just an- other library consortium. This small com- munity of universities, their libraries, and an independent research library formed a partnership to achieve planned, coordi- nated interdependence in response to the threat posed by a climate of increasing ec- onomic restraint and financial uncer- tainty. RLG was a group of homogeneous, geographically proximate institutions, similar in goal and function, and with a history of earlier cooperative endeavor. Their commitment to active, mutual sup- port was strong. In its first years, RLG experienced the withdrawal of Harvard, the addition of Stanford University, and the adoption of the computer-based bibliographic pro- cessing system (BALLOTS) developed at Stanford. By the close of 1982, there were twenty-six full, affiliate, and associate members and sixteen special members of the partnership. BALLOTS, RLG's tech- nical processing system, was overhauled, with many of the traumas that accompany large-scale technological innovation, into RUN (the Research Libraries Information Network), a bibliographic utility support- ing many of the functions of its parent consortium. Following the acquisition of RUN and the expansion of membership, RLG's other principal programs-collection management and development, shared resources, and preservation-were recon- stituted in somewhat different form . RUN, coupled with central staff access to other resources of the Stanford computer facility, had a substantial impact on the development and operational nature of these programs, and each deserves its own study. For the Collection Manage- ment and Development Program, how- ever, the availability of computer re- Nancy E. Gwinn is associate director of program coordination, Research Libraries Group, Inc., Stanford, Cali- fornia, and Paul H. Mosher is director for collection development, Stanford University Libraries, California. 128 Coordinating Collection Development 129 sources made possible the construction of the RLG Conspectus, a · collection evalua- tion instrument to facilitate coordinated collecting activity. This article describes the history, operation, and future of the Conspectus. The Conspectus is an overview, or sum- mary, of existing collection strengths and future collecting intensities of RLG mem- bers. Arranged by subject, class, or a com- bination of these, its divisions contain standardized codes that describe collec- tion/ collecting levels on a scale of 0 to 5 (with 5 as "comprehensive"). But the Conspectus planners also harbored a larger vision. Assuming its successful develop- ment, they hoped the Conspectus would become the cornerstone of a larger na- tional cooperative effort (one now being studied by the Association of Research Li- braries) among all the principal research li- braries of the nation, for the eventual ben- efit of generations of scholars. The invention of the RLG Conspectus de- rived from the fortuitous conjunction of individuals sharing common interests and goals, the expansion of RLG membership, and agr~ement that something like a na- tional collection development policy would be necessary to protect the research capacity of the nation's universities from the impact of repeated and unfavorable economic cycles. Using this tool, research libraries could focus collective resources on appropriate distributed but coordi- nated effort, thus ensuring availability of unique or rare titles to the nation's scholars. To this was added the availabil- ity of staff and computer support from RLG. The ideas that eventually were forged into the Conspectus can be traced to three sources: GNOMES + 2, groups within the American Library Association, and the initial work of the first RLG Col- lection Development Committee. GNOMES+ 2 In 1978 at the ALA Annual Conference, a group of chief collection development officers heard John Finzi (now director of the Collections Development Office of the Library of Congress) present a position paper on a "new Farmington plan." Rec- ognizing that no single library, including the Library of Congress, had or could ac- quire the entirety of world book produc- tion, Finzi called for distribution of collec- tion responsibilities both for ''exotic'' regions or areas and for certain classes of material in other subjects or disciplines. Calling itself GNOMES + 2, this group consisted of the collection development officers of the Seven University Group li- braries: Stanford, Cornell, Harvard, Yale, Chicago, Columbia, and Princeton, plus the New York Public Library and the Li- brary of Congress.* A shared concern for the future of research library collections , was evident in the room. The group's ma- jor worry was how to rationalize the col- lecting powers of the nation's major re- search libraries to achieve adequate support for both foreign area studies and subjects or disciplines, for which materials are fugitive or under poor bibliographic control or distribution. In the course of the discussion, GNOMES + 2 members com- mitted themselves to devising an appro- priate scheme-:-analogous to the Far- mington plan, but different in _scope-and proceeded to carry the idea to other fo- rums. RLGANDALA · The chief collection development offi- cers of the four RLG members formed the first Collection Development Committee in 1974. By 1978 their work was providing important practical and theoretical con- cepts to help shape the growing idea of a nationwide plan. This group forged anini- tial program that included two key com- ponents: 1. The analysis of collection develop- ment needs and programs at each institu- tion, the preparation of collection devel- opment policy statements, and development of the means to coordinate and rationalize information from the four policies. *The reference is to the nickname of the Seven University Group who humorously call themselves "gnomes" because they meet in dark and smoke-filled rooms underground. 130 College & Research Libraries 2. The allocation of "primary collecting responsibilities,_, or subjects, geographi- cal areas, and forms of material by which a h r un er o ar ·ership, responsi- bility for collecting in certain areas. Primary collecting responsibilities (PCRs) were assigned for publications from countries in Africa, Central America, and Eastern Europe, for intergovern- mental agencies and international organi- zations, and within the fields of architec- ture and journalism. These PCRs were distributed among members when fields were identified as "non-conflict"-that is, there were no competing or overlapping programs of importance among the mem- ber institutions. The third source shaping the concept of the RLG Conspectus derived from discus- sions and work within various units of the American Library Association. Members of GNOMES + 2 and RLG were active in ALA as well and carried their ideas and commitment to meetings of the Chief Col- lection Development Officers of Large Re- search Libraries Discussion Group and of subject specialist sections within the Asso- ciation of College and Research Libraries, where complementary work was already under way. The RTSD Collection Devel- opment Committee, which had begun to prepare a series of guidelines to foster and facilitate management of library collec- tions, included a guideline for the prepa- ration of collection development policy statements with definitions of collection levels adapted from those articulated by RLG (see appendix A}. 1 A set of language identifiers modified from RLG was also in- cluded. THERLG COMMITIEE REGROUPS By 1979, following the move of RLG's central staff to Stanford, representatives to the revitalized, newly named Collection Management and Development Commit- tee brought to their first meeting a com- munity of shared activities, interests, and goals reflecting both these earlier discus- sions and developments in their own in- stitutions. Together with strong leader- ship from David Starn and Paul Mosher, chair and vice-chair, and John Haeger of March 1983 the RLG central staff, the ideas, activities, goals, and shared products made up the stuff and substance from which the new committee formed its programs. The back- bone became the RLG Conspectus. In January 1980 the committee received and endorsed a subcommittee recommen- dation that "the committee develop an RLG collection policy statement . . . to serve as a vehicle for cooperation with the Library of Congress and other major re- search libraries in developing an eventual national research resource collection of materials held severally by RLG and other major research libraries, with primary col- lecting responsibilities distributed among those libraries and LC, and with LC acting as a kind of' system equalizer' to minimize the impact of local program change on na- tional research library resources." 2 Mem- bers further agreed that the coordinated policy statement should carry information on existing collection strengths as well as current collecting intensities. At the same meeting, the committee brainstormed ideas on the nature of coop- erative collection development and agreed to several objectives, the achieve- ment of which would require support by a cooperative RLG collecting policy. They included: 1. The need to identify collection strengths nationally. 2. Mutual reliance and interdepen- dence in providing research materials. 3. Establishment of a tool to identify collecting levels at other institutions, to al- low for changes, and to assess their signif- icance. 4. Capacity to control better the physi- cal growth of library collections and oper- ating costs, and to distribute both collect- ing responsibilities and savings that might result. 5. Development of a mechanism to lo- cate needed research materials more ade- quately. 6. Rationalization and standardization of format and terminology of local collec- tion development policies to enable li- braries to achieve the above goals. 7. Development of a mechanism whereby an institution may store or dis- pose of locally unneeded materials with Coordinating Collection Development 131 the knowledge and assurance that materi- als will be available elsewhere. 8. Establishment of a means for relating collection policy to preservation policy, both institutionally and cooperatively. 9. Development of a mechanism for re- lating collection policy and responsibility to cataloging priorities and for establish- ing centers of cataloging. Paul Moshe·.c agreed to develop a pro- posal for constructing an RLG policy. ~~coNSPECTUS" COINED Three months later, the committee ap- proved the Mosher proposal on the format and content for the RLG collection devel- opment policy, to be known as the RLG Conspectus. The word "conspectus" was defined as a ''breakdown of subject fields in such a way as to allow distributed col- lection responsibilities for as many fields as possible. " 3 This term has received gen- eral acceptance as a way of differentiating the national descriptive policy statement now in formation from those of individual institutions. Moreover, the term reflects the distribution of collection strengths and collecting intensities in a way that can fa- cilitate planning, but without the prescrip- tive implications of a "policy." It was rec- ognized that each participating institution must be left free to create its own policies as local programs and funding permitted; the Conspectus was viewed as a means to encourage coordination of those individ- ual institutional efforts for the greater ben- efit of libraries and their users across the nation, without giving up local autonomy. The committee also adopted a number of general principles, summarized as fol- lows: 1. The Conspectus must be easy to use, flexible enough to meet changing needs, and capable of elaboration in order to treat adequately each field and subfield. 2. The need for specificity in collection strength and collecting intensity would normally be most useful and necessary for fields in which many or most libraries col- lect at a fairly high level and which involve a wide range and large number of materi- als, such as medieval history, German his- tory, French literature, or sociology. 3. The Library of Congress classifica- tion should form a general framework for the Conspectus through the use of its vari- ous schedules, but other subject descrip- tors, outlines, or breakdowns could be used for academic fields not adequately covered by LC. 4. Recognizing that such a complex project as the Conspectus might never be fully completed, or might be completed hastily in only a marginally useful man- ner, it should be phased in at logical stages according to an established timetable. In determining where to start within the LC classification schedule, the committee used the National Shelflist Measurement Project data as a guide. By starting with fields that reflected the largest acquisition and cataloging efforts of member libraries, the eventual collecting assignments, even partially worked out, would represent a significant achievement. Since linguistics, languages, and literature (class P), com- bined with history (classes C, D, E, and F) collectively represented 39 percent of all ti- tles held by research libraries, these fields were chosen to start. Subsequent work would proceed through the classification roughly based, in descending order, on the number of volumes represented in re- search library collections. At the same time, the committee recog- nized that work on Conspectus segments for area studies programs, such as East Asia, or subjects representing special RLG interests, such as art and architecture, might be undertaken in a parallel time- table if it seemed desirable to do so. Committee members and RLG · central staff worked hard throughout the balance of the year to revise definitions of collect- ing levels and language codes and to con- struct work sheets for data collection and a format for data presentation. Ultimately, the committee envisioned an interactive, online format that would allow access to the database by subject, institution, LC class, geographical area, or other useful descriptors, since, in hard copy, the entire document would be several hundred pages long and cumbersome to use. EAST ASIA EXPERIMENT While the collection development offi- cers were making their own plans, RLG 132 College & Research Libraries established an East Asian program and be- gan to mount a major development effort to build the capacity for handling East Asian scripts in RLIN. Since East Asian vernacular collections a generally housed separately from general research collections yet cover all subject fields, it appeared that a methodology for gather- ing data about them could be effectively employed. Such a strategy would satisfy twin objectives. As a microcosm of the general collections, problems encoun- tered could be resolved and the methodol- ogy revised before a great commitment of time had been made. Secondly, since East Asian materials are often expensive, any data that would help rationalize and dis- tribute the cost seemed worthy of collec- tion. RLG central staff prepared a broad sub- ject outline based on the Library of Con- gress classification scheme and asked East Asian librarians to provide separate val- ues on a scale of 0 to 5 to describe existing collection strength and current collecting intensity for materials in Chinese, Japa- nese, and Korean languages. The experi- ment was a success. When the values were spread out on a grid, patterns began to take shape that reflected what most cu- rators thought to be true about the relative strengths of their collections. Fear that in- stitutional pride might result in a gross overrating of collection strength proved unfounded. When inconsistencies were spotted, they were resolved through dis- cussion. Heartened by this experience, RLG cen- tral staff, in close consultation with spe- cialized subject bibliographers, began to prepare work sheets to cover other subject areas, and the RLG Conspectus became a reality. Immediately, of course, the prob- lem of using the LC scheme to describe to- day's interdisciplinary research collec- tions became apparent. For example, LC classes C, D, E, and F describe only one type of history; today's historians regu- larly use materials in B (philosophy and religion), H (social sciences), J (political science), L (education), and other classes in writing social and intellectual history. The same holds true for other disciplines. Studies have shown that some 55 percent March 1983 of the titles used by sociologists nation- wide are given numbers outside of the LC class for sociology. 4 Initial work sheets therefore attempted to draw from all classes any field that sup- ported research effort in one. The history division, for example, contained a number of lines for reporting from other classes, such as the H class for economic history, J for constitutional history, etc. Later, as work on the online file progressed, it be- came clear that this was less important and possibly misleading, since the system could draw together data from through-. out the schedule, and the methodology was changed. By the fall of 1980, data collection was under way. RLG has now collected data for subjects that account for more than 76 percent of RLG libraries' collections. In addition to language, literature, and his- tory, these include: art and architecture, philosophy and religion, most physical sciences, music, economics, political sci- ence, sociology, law, and East Asian and South Asian studies. These fields have been collectively divided into more than 2,700 subjects and geographical subdivi- sions. Provision has been made for brief, informative notes to clarify or enhance specific data by highlighting subject col- lections of particular strength or problems of assessment. On the drawing board are work sheets for government documents, life sciences, geography and earth sci- ences, technology, medicine, and Latin American studies. Others planned for the near future are anthropology, psychol- ogy, and education. From the foregoing, it is clear that a "two-track" system of reporting has emerged. In addition to following an LC class arrangement, Conspectus segments are being completed for major area studies collections that cut across all classes of knowledge. In part this reflects acquisi- tions practices, which are often geograph- ically based. But it also recognizes the self- imposed obligation of American research libraries to provide adequate bibliographic coverage for all areas and peoples of the world, a national need that became once again acute following the death of the Far- mington plan. Coordinating Collection Development 133 RLG CONSPECTUS ONLINE Recognizing the complexity and length of the entire document, and the problems that would arise from trying to manipulate its bulk to arrive at needed information, RLG central staff has mounted all of the existing data online, so it can now be readily retrieved. This new interactive database, called the RLG Conspectus On- line, can be searched by subject, class, col- lection level, and institution, among other values. The system's flexibility and ease of use greatly enhance the utility of the data. Development work on the database re- ceived funding support from the New York Public Library, an RLG member. If a bibliographer in a northeastern re- search library, for example, wishes to make a decision as to whether or not to purchase an expensive new multivolume set in the field of European demography, he or she could first search the RLIN bib- liographic files to see if any RLG library had ordered it. If no record were found, the bibliographer could then switch to the Conspectus Online to look for collection lev- els in other libraries. Using subject words such as "economic" and "demography," the bibliographer finds a record contain- ing collection data for that field and dis- covers that Princeton has both a compre- hensive (levelS) historical collection and a commitment to continue to collect at that level in a wide variety of foreign languages (5/SW). (See figure 1.) Moreover, as a backup the New York Public Library has both an existing collection and ongoing collecting policy at level 4 (a strong re- search collection) and a note that indicates an emphasis on English, French, German, Italian, Portuguese, and Spanish lan- guages. This information has increased the bibli- ographer's options. It is now possible to decide to depend on Princeton (or NYPL as a last resort) for access to the set under the liberal lending policies of the RLG Shared Resources Program and to use the money that would have been spent for CEC029) ECONOMICS AND SOCIOLOGY - ECONOMICS · Demography PCR:None HB879-3700 COSG 3/3F CSUG 4/4F CTYG 4/4F CUBG 4/4W CUDG 4/4E DCLC 4/4W ILNG 4/4E MAAR 4/4E MDJG 2/3F MIUG 3/3W NHDG 3/3F NJPG 5/SW NJRG 3/3E NYCG 3/3W NYCX 4/3F CCI for foreign materials is 3. NYPG 4/4W For materials in languages other than English, French, Garman, Italian, Portuguese and Spanish, collecting is limited to bibliographies and indexes, dictionaries, government directories, demography, statistical results of censuses end other statistical reports, end official government gazettes. HYUG 3/3E PASG 4/4F PATG 3/3E PAUG 4/4E RIBG 4/4F Demography collection supplemented by departmental library •. FIGURE 1 Display from RLG Conspectus Online Database 134 College & Research Libraries purchases in areas more closely tied to lo- cal academic programs.* Wider-ranging collection decisions could also be influenced by use of the Con- spectus database. An RLG institution that did not have a high-level academic pro- gram in economic demography, for exam- ple, could decide to reduce its collecting effort from a present level4 (advanced re- search) to level 3 (advanced study), with attendant cost benefit, relying on Prince- ton's level 5 collection (along with the level 4 collections of other institutions) to answer the infrequent call for such materi- als from its own faculty and students. Clearly, one of the advantages of this data-gathering effort is to increase options for local decision making. Within the RLG partnership, the work has led to other benefits, some obvious, but others unan- ticipated. Forming Collection Development Policy Statements Each library contributing data to the Conspectus inevitably will construct a com- plete, useful local collection development policy, representing its future intentions as well as its historical collection strengths. Collection development plan- ning, trade-off decisions in terms of col- lecting strengths or intensities, and changes in collecting levels suggested by changes in academic programs (or by fi- nancial exigency) can much more easily be made, and their impacts understood. Once the local changes are made, RLG central staff can quickly revise the data, and the change is rapidly communicated to all member libraries. For institutions that already had devised collection policy statements, the conver- sion to the Conspectus was relatively sim- ple. Even so, the process itself has proved to be an effective training aid for staff members, who may greatly increase their knowledge about collections. March 1983 Rationalizing Collecting Responsibilities As each portion of the Conspectus is com- pleted, a subcommittee of the Collection Management and Development Commit- tee reviews it in detail. For the present, this initial review focuses on subject areas where there are no research level (level 4 or 5) collections within RLG or where there are only one or two. If there are at least three collections at level 4 or 5 (re- search or comprehensive) among RLG members, the subject is considered ade- quately covered. If two or fewer member libraries have research-level collections, it is possible that this is an "endangered" field and that a member should be sought to accept a primary collecting responsibil- ity (PCR). To make such a recommenda- tion about an uncovered or poorly covered field, the subcommittee considers first if a level 4 collection is necessary within RLG or if a combination of level 3 collections, overlapped, may contain a sufficient num- ber of unique titles to equal a level 4 or 5 collection and thus adequately cover the subject area. The subcommittee then reports its rec- ommendations to the full Collection Man- agement and Development Committee. The library for which an assignment is suggested has an opportunity to explain whether or not it would be likely to accept responsibility for an area. RLG recognizes that local academic programs dictate a li- brary's collection practices to a great ex- tent, and local constraints as well as na- . tional needs are considered when a library is asked to maintain a level4 or 5 collection (or to upgrade from a lower level to a higher one) in the interest of the partner- ship. Any library may opt not to do so and may then notify the committee that local program or financial problems make it necessary to reduce a collecting level, even if the institution has accepted a pri- . mary collecting responsibility. So far the *As part of the Shared Resources Program, RLG members have committed themselves to give prior- ity to interlibrary loan requests from RLG members and to respond within three days to any request. Members also agree to use the United Parcel Service for shipping and to loan material that has been received but not yet cataloged. No fees are charged among members. Coordinating Collection Development 135 equalizing factors of the other member li- braries plus the Library of Congress have been found more than adequate to handle the few cases that have occurred. Assignment of a primary collecting re- sponsibility to one institution may affect collecting policies elsewhere-or it may not. Faculty and library commitment to support of local academic programs may remain high, even though another institu- tion may possess a larger collection or budget. The level of local collecting activ- ity remains a local decision. The Conspec- tus is intended only to increase local op- tions. It does not dictate local policy. As part of its analysis, the subcommittee also takes into account the existence of other strong, accessible research collec- tions outside the RLG partnership. While there is consensus that RLG libraries rep- resent a high proportion of available scholarly publications among their 75 mil- lion volumes, even their combined hold- ings plus the Library of Congress do not represent the entire universe, particularly in certain specialties. RLG's present plan is to recognize strengths outside the part- nership and to look for eventual collabora- tion on a larger scale so that these may be taken into account in an expanded na- tional research collecting scheme. Nine portions of the Conspectus have · been analyzed and more than 150 PCR as- signments made in subjects as disparate as Russian architecture, Chicano litera- ture, Chinese medicine, pastoral theol- ogy, and Finnish history. This represents about 20 percent of the fields identified for possible assignment. Assignment of the others was considered unnecessary or postponed for a variety of reasons: (1) the volume of publishing or scholarly interest in the field was minimal (e.g., Manx lan- guage); (2) the field was well covered out- side of the partnership (e.g., band music); (3) the number of collections at level 3 within the partnership was sufficient (the hypothesis, again, that multiple holdings at level 3 collectively equal a strong level4 collection); 5 (4) the assignment was post- poned pending receipt of data from a member thought to have the strongest col- lection; or (5) the assignment was post- poned awaiting additional information from an area studies segment. Accepting primary collecting responsibility obliges the member institution to continue collect- ing and processing materials at the level at which the assignment was accepted, to maintain the materials in good condition, and to make them available to other RLG partner libraries within the scope of the RLG Shared Resources Program, unless relieved of the commitment by RLG. Within this framework, the Library of Congress serves as a kind of equalizer for the system. Should none of the RLG li- braries have strong academic programs in the study and history of the Irish or Gaelic languages, for example, and should no member wish to upgrade its collecting to the level considered desirable by mem- bers, the Library of Congress has agreed to consider accepting a primary collecting responsibility for that field, possibly even upgrading its collecting activity, in order to provide support for the system. In re- turn, LC looks forward to being able to de- pend on the collecting responsibilities of other major research libraries within the country and to take those opportunities, when desirable, to cut back on its own col- lecting in certain areas. Resource for Reference and Interlibrary Loan The RLG Conspectus Online database is likely to be a rich resource for the interli- brary loan librarian who needs to find and borrow older titles not represented in RUN or other finding tools. If a location cannot be determined from other sources, the librarian may search the Conspectus for a strong collection in the subject and direct the request first to that library. Since the RLG interlibrary loan subsystem allows the requester to define a hierarchy of po- tential sources for a title, the Conspectus can help establish a logical order. If a nega- tive response is received from the first in- stitution queried, the request automati- cally is sent to the second, and so on. The reference librarian is likely to find the Conspectus of equal utility in directing faculty and graduate students to strong re- search holdings in subjects out of scope for the home institution. Used this way, the Conspectus becomes a practical, gen- 136 College & Research Libraries eral guide to a greater world of scholarly resources. Regional and Specialized National Planning There are signs that the RLG Conspectus methodology is being used, in part or as a whole, as a basis for other efforts at re- gional or nationwide cooperative plan- ning. Members of the Colorado Organiza- tion for Library Acquisitions (COLA), a subgroup of the Colorado Alliance of Re- search Libraries, use a modified version of the RLG definitions for collection levels as a basis for assigning points to reach a ranked order of expensive items sug- gested by members for cooperative pur- chase. (Colorado State University is an RLG member.) Likewise, work sheets for the Japanese portion of the East Asian conspectus division were used by all insti- tutions with significant Japanese collec- tions in the western United States to col- lect data in preparation for a Western Regional Japanese Library Conference held in January 1982 at Stanford Univer- sity. Five of the thirteen libraries repre- sented were RLG members. A group of South Asian bibliographers, meeting in conjunction with the annual conferences of the Association of Asian Studies, pre- pared work sheets for South Asian area studies and are in the process of data revi- sion. The universities of Chicago, Wash- ington, and Wisconsin, all non-RLG members, are contributing to that project. Other initiatives based on Conspectus-like activities are developing within SALALM (Seminar on Acquisition of Latin Ameri- can Library Materials) and among univer- sity libraries in Indiana. In addition to the benefits outlined above, others can be anticipated. As the work progresses, the data may be used to reach decisions for RLG' s other coopera- tive programs, such as shared cataloging and preservation. At least one RLG mem- ber hopes to use the data as an aid to fund- raising efforts, that is, identifying for po- tential donors opportunities for gifts and endowments to build or maintain collec- tions of special strength or value as dem- onstrated in the Conspectus. Likewise, the data may help demonstrate to deans and March 1983 provosts the requirement for funding sup- port to build collections when new faculty appointments are made. DATA VERIFICATION As the work on the Conspectus has pro- ceeded, the complexity of its compilation and its pioneering nature have, not sur- prisingly, brought some problems to light. Since there are no adequate quantitative measures of collection utility or excel- lence, it was recognized from the begin- ning that judgment and discrimination would be involved in determining collec- tion levels. Comparative shelflist mea- surement data exists for some libraries ac- cording to a more simple Library of Congress scheme, and this data can be of help. However, not every RLG member li- brary contributed to that project, nor does data exist for every subject. Furthermore, the collections of many libraries have been built up over time from bulk gifts, pur- chases, or exchanges, which may tend to inflate the library's title count, without adding measurably to the quality or signif- icance of the collection. / To help ensure the veracity of the data, therefore, RLG has initiated a program of "verification" and "overlap" studies, which may be used in conjunction with shelflist measurements to guide libraries in assigning comparable collection levels. These studies are designed also to repre- sent the distribution of unique titles and the pattern of duplication among member libraries, which can provide interesting data for further planning of the RLG col- laborative effort. The value of these stud- ies was confirmed by a pilot collection evaluation project sponsored by RLG in the summer of 1981. A sample of 1,000 monograph and serial titles in the field of English literature was drawn from the Cambridge Bibliography of English Literature and the Modern Language Association bibliographies. These were checked against the holdings of four RLG members-Columbia, Yale, Stanford, and California-Berkeley-each of which had reported a level 4 (research) collection, and therefore could be expected to own the major published source materials re- quired for dissertations and independent Coordinating Collection Development 137 research. If one were to distribute the five levels of the scale evenly by percentage of holdings, an institution reporting a level4 collection should be expected to hold be- tween 70 and 85 percent of the titles con- sidered important. The results of this study showed holdings that ranged from 76 to 86 percent of the titles searched, illus- trating not only the accuracy of the re- P