Gary Wiggins


The academic science librarian of today faces many challenges as we move toward a new century. Not the least of these is the transition from a collection which is primarily print- based to information sources which will exist substantially in electronic formats. This paper examines some of the concepts and dilemmas which an academic science library must consider in order to align its collection development activities with the changing environment of scientific and technical librarianship. In particular, the problem of serials access versus ownership is considered.

Of prime importance to collection development in large academic science libraries are the faculty, for it is their research programs and endeavors which determine the collection's scope and depth and the services offered by the library. Science faculty at academic institutions are generally engaged in 3 activities: research, teaching, and service. The larger the academic institution, the more the faculty are likely to concentrate their efforts on research. It is that activity which garners the research grants to fund the personnel and equipment needed to explore matters of interest in the laboratory. Ultimately, the results of the experiments will be published and contribute to the archival record of science which is vital for its survival.

Tuck et al. have identified four categories of research activity:

- collecting and analyzing published information

- collecting and analyzing primary data (including

experiments and field work)

- writing (reports, theses, etc.)

- communicating results of research through seminars,

conference papers, journal articles, and books.1

Experimental results are typically presented at symposia or conferences before appearing in published primary sources of information. In a sense, the oral presentations are part of the second activity: teaching. Teaching requires organizational skills and the ability to present concepts in a way that will be easily understood by the audience. Good teachers also have the capacity to test the target audience in a meaningful way to see if (and to what degree) they have grasped the material which has been selected for presentation.

Finally, academic faculty are engaged in service. The service may take the form of a post within the department (for example, chairing the laboratory safety committee) or within the larger academic community at the home institution (such as service on the faculty senate), or, more likely nowadays, service to the profession through a role in one of the professional societies. The last activity may involve arranging a symposium at a national meeting or serving on any of the many committees which most professional organizations generate. An academic science library will typically be heavily involved in the faculty's research and teaching enterprises, but much less so in the activities which center around service.

There are some services which academic libraries traditionally have provided free, and which will undoubtedly continue to be free. Among them are ready reference, circulation, and open access to materials. However, both librarians and users must recognize that information is not a free commodity. Fees for certain services which have traditionally been free should be considered. In-depth reference service, mediated online searching, and on-campus delivery of materials are examples of services where the labor costs have often been absorbed by the academic institution. Yet it is the labor costs which form a significant barrier to the provision of wider access services. Therefore, it is essential to define the core services and collections which ought to be offered at no charge to the user. In particular, much attention needs to be devoted to collection development and access services, for it is the access to published material which is still of utmost concern to the faculty at the beginning of this last decade of the twentieth century.

What used to be called the "acquisition" function in libraries has now been designated "collection management and development." This reflects the realities imposed by economics over the last two decades. Science librarians today do not simply acquire materials suggested by the faculty and other users. They actively evaluate the body of available sources, comparing them to the research and teaching interests of their primary clientele and selecting the material which will satisfy as much as possible of the current demand from the on-site collection. Even the mass of material from previous, more affluent decades of "acquisition" work does not escape their scrutiny, for collection management also encompasses weeding. The desirability of keeping a collection of little-used older material on-site must be measured against a rapidly dwindling amount of empty shelf space for newer material. Off-site storage facilities may be utilized, but increasingly, withdrawal (weeding) is the choice made by the collection manager.

Another buzzword which emerged in the 1980s is "access," a term generally applied to information sources which are not held at the geographic site of the academic institution. Calls for more use of access services are increasingly heard from library administrators, especially that new breed of assistant or associate deans of large academic libraries, the "chief collection development officers." Part of their job is to control expenditures and find space for all of the materials collected. One need only read the titles of a few "Specflyers" (publications of the Association of Research Libraries Office of Management Studies Systems and Procedures Exchange Center) to gauge the direction of modern acquisitions work: "Cooperative Collection Development" (no 111, February 1985) and "Serials Control and Deselection Projects" (no 147, September 1988). The recent report of the ARL Task Force on a National Program for Scientific and Technical Information lists five priorities, among them, "accelerating research libraries' efforts to coordinate programs in acquiring, maintaining, and preserving serials literature."2

The management of science serials collections has often been approached in the last two decades by assigning each serial subscription to a subject fund (chemistry, physics, biology, medical sciences, etc.). The responsible librarians were then given the charge to "manage" the fund, that is, to make the necessary cancellations when the allocated budget was exceeded, which frequently happened in the last 20 years. The difficulty with this approach is that it tends to ignore the interdisciplinary nature of science, an aspect which is becoming increasingly important in many areas of research. Too often the fund managers (librarians) were forced to cancel serial titles which were of less immediate need by their own faculty, but were critical to an emerging new area of study in another field. An often-used criterion is the "cost/benefit" of a particular serial title.3 Elaborate usage studies have been devised to prove that such venerable reference titles as the Beilstein Handbook of Organic Chemistry or translation journals simply were not contributing to the immediate mission of a given fund. Unfortunately, little weight is given to the effort and expense required to produce such works and the cost of obtaining the desired information via alternate paths when such sources are not available.

Collection development in academic science libraries has bordered on the chaotic in recent decades. This has resulted in part from a failure to adequately define "cooperative collection development" and "access" and in part from the lack of a coherent plan for integrating machine-readable sources into our collecting profiles. There are some ground rules which must be accepted before order can be accessed in collection development.

Rule number 1: Buy no material for any program/department until its priority in the institution's plan for the 1990s is clearly defined. This must be clearly and openly stated by the highest levels of the academic administration.

Rule number 2: Invest the resources available for collection development (including access) in a manner which reflects the long-term priorities of the academic institution. This rule must be enforced by the library administration.

It seems sensible to define the access/ownership expenditure ratio on the basis of actual university priorities. There should be at least a two-tiered division of expenditures for acquisition, that is, for actual material held (or leased) on-site, regardless of format. This could be defined in terms of the percentage of times that a researcher should expect to find a needed item at the home institution, perhaps 90% of the time for the top level of support, and 75% of the time for those supported at a lesser level.

Many academic institutions subsidize the costs of access to materials or electronic forms of information which are not locally owned (or leased). That may take the form of a flat fee for a photocopy (or telefacsimile) of journal articles, a subsidy for costs associated with searching an online database, etc. It could be argued that academic institutions should provide a uniform level of subsidy for access according to membership in user groups, regardless of the priority which the university places on the programs or departments from which the users come. There are significant annoyances in using many of the access services, despite the great strides in the technology which serves these areas. Not the least of those is the delay in obtaining the material. It is likely that for a good while to come, the time necessary to acquire information via information products accessible through a library will exceed the time it takes to find the same information in a library which actually owns the information source. Thus, an academic institution must reserve a significant portion of what was once called the acquisitions budget for access to alternate sources of information, both print and electronic. Costs which might be subsidized from this source of funds include interlibrary loan fees, search costs for remote online databases, and purchase, license, or royalty fees for electronic forms of information provided on-site in lieu of paper counterparts.

The basic problem with cooperative collection development programs in the sciences, if not in all disciplines, is that there is a core of material and information sources which all significant research libraries want to own or have access to. This is nowhere better illustrated than in the Research Libraries Group "4+ Project" for Chemistry. Starting with a list of journals which contribute the most references each year to Chemical Abstracts, the participants in the project supplemented the titles with rankings from Science Citation Index and other sources. The list was then pared down to 541 titles, and an agreement was reached that for each title on the list, n subscriptions would be maintained, where n varies from 1 to the maximum number of participants.4 The participating universities were seeking a long-term commitment for the maintenance of the titles, at least 5 years. Needless to say, there was a large group of titles for which all libraries chose to continue to maintain subscriptions.

Ralli has stated that "There is no substitute for direct access by scholars to the materials they require."5 But direct access is no longer limited to a user who physically retrieves books and journals sitting on library shelves. Information is becoming fully transportable to a user's workstation. In the future, it should be easier to transport information in computer- readable form from a library or other source than it is to make a photocopy today. Electronic formats of information will have the greatest and most far-reaching impact on academic research over the next decade. Libraries should, therefore, embrace a broader definition of access to include all document delivery techniques which use appropriate technology to easily deliver information from the location in which it is held to the user. As telefacsimile services become more widely available, the options for enhanced document delivery services increase dramatically.

Some recent research projects and new commercial services provide useful insights into the ingredients necessary for the successful transition to an electronic library. These are Cornell's CORE project, the British Library's Project Quartet, and the Colorado Alliance of Research Libraries UnCover services.

The Chemistry Online Retrieval Experiment (CORE) at Cornell University is an extensive investigation of the impact of an electronic chemistry library.6 Equipped with a database of ten years of online journals from the American Chemical Society and the indexing of those articles from Chemical Abstracts Service, the project has several ambitious goals. One of them is to determine the users' perception of the effectiveness of the electronic form of the primary journal as compared to paper. Another is to identify the most desirable traits of the user interface to an electronic system of journals. The CORE project also attempts to address the most severe limitation to online use of chemical journals--the lack of pictorial information in online files of full-text science journals. Drawings of apparatus, spectograms, chemical structures, reaction diagrams--all standard features of printed chemical journals--would be lost if the only access in the CORE project were to the ASCII text. In order to include such graphical matter, the pages with images were scanned from the microfilm versions of the printed ACS journals.

Among the potential benefits cited for an electronic chemistry library are superior distribution methods, cheaper, less- cumbersome methods of storing and archiving the material, and more effective access to library resources by end users.7 A key technology for the CORE project is the computer network, since the delivery of graphic images to a workstation requires considerably more network capacity than does textual data.

Project Quartet, so named for the four British academic participants (University of Loughborough, University of Birmingham, Hatfield Polytechnic, and University College, London), also included as participants the British Library Document Supply Centre and the ADONIS consortium of publishers.8 The project ran for three years, 1986-89, and had objectives similar to those of the CORE project, among them, to determine just how far the newer electronic media could replace traditional print sources and where the management of print-on-paper systems might be enhanced by such media.

Project Quartet focused on 4 distinct areas: online information systems, computer-based message systems, computer conferencing systems, and document delivery systems. 220 high-use biomedical journals on CD-ROM (the ADONIS collection) formed the major database used in the experiment. User studies conducted on the ADONIS portion of the project showed that there was no overwhelming demand for the service. A possible explanation offered by the researchers is "...that the service provided was not really sufficiently better than that obtainable using conventional methods for the user to be willing to make the switch."9 To be successful, a new service must provide sufficient incentive to the user to make its use routine. The Project Quartet researchers suggest that the document delivery function needs to be integrated with related services--end-user searching, current awareness bulletins via e-mail, one-stop ordering and delivery.10 Only then do they feel that the users will be persuaded to change their information search and request habits.

The Colorado Alliance of Research Libraries (CARL) has developed such an integrated system with its UnCover services. With CARL, the users can perform searches over the Internet in a bibliographic database of journal articles. Either word or author searches are possible. They can also scan contents pages of recent journals, select articles, and have them sent to a telefacsimile machine. The FAX document delivery service is generally provided within 24 hours, but the remarkable aspect of the service is that CARL retains the FAX image, thereby providing the next user who selects that article with a virtually instantaneous copy. It remains to be seen whether masses of end-users with credit cards in hand will avail themselves of CARL's UnCover services. Competition from Faxon Finder and Faxon Xpress, similar services from Faxon Research Services, Inc., will soon be available.11 Such integrated services point the way to the future of science librarianship. It is the implementation of services based on increased access which will define the science library of tomorrow.

In the mid-1970s, Eugene Garfield noted that a collection of 100,000-200,000 books and articles can form the active core of a library able to provide copies of 90 percent of all future citations.12 This contention was even more dramatically illustrated by Garfield in discussing the 1988 Science Citation Index. In that year, only 900 journals received 83 percent of the 8,000,000 citations processed for SCI's Journal Citation Reports.13 In the earlier article, subtitled "Pulling Weeds with ISI's Journal Citation Reports," Garfield proposed that only reprints of highly cited articles be retained by libraries. Although Garfield's company, the Institute for Scientific Information (ISI), had plans at that time to identify, collect, publish, and market a core collection of heavily-cited articles, such a product never materialized.

One can conjecture that the idea of a library of repackaged paper articles was ultimately deemed by ISI to be too radical a departure from the normal acquisitions patterns of the time to be marketable. However, times have changed. There exists today the technology to create a library of electronic articles tailored to the needs of a particular clientele. What is missing from the current scene is the motivation to do so. As long as libraries continue to measure quality in terms of complete physical volumes held, the idea will not take hold. However, a few bold success stories could change the value systems of both academic librarians and users. A library equipped to store and retrieve images received from a document supply source such as CARL or Faxon could begin to build a database of document images with a modest investment in equipment.14 A LAN-based FAX product could serve as the waystation for incoming images. An archival copy could be created before forwarding to the end-user, and eventually a collection of documents of particular relevance to the primary clientele would be amassed. Of course, a fee would have to be paid to the publisher for the retention of the image, but appropriate accounting software could keep track of the costs and also determine which records had failed to be productive in a given time span (perhaps 10 years).

The obvious benefit of such a scenario from the librarian's viewpoint is that the acquisitions budget could be used to acquire and retain both complete volumes of the core journals and additional individual units (articles) which are truly needed by the clientele. There would be an incentive for the publishers to weigh more carefully the items accepted for publication. Articles which no one cites or wants to read could not be easily concealed in a journal volume sold only as a complete product. In terms appropriate to the automobile industry, a library would have the option of purchasing the "foreign" (off-site) economy model or the "domestic" (locally-owned) fully-equipped model.

The science library community is poised to make a fundamental change in collecting policies in the 1990s--to move from the exclusive collection of journal volumes to a more flexible collecting profile that allows the permanent addition to their collections of individual articles from journals not locally held. What is needed to facilitate the collection of individual articles is a unique item identifier for each journal article, a code which is universally understood by librarians and users. With the development of the Serials Item and Contribution Identifier (SICI), we now have such a tool.15 SICI is the ANSI/NISO standard Z39.56 (1991) which defines a string of numbers and/or letters to uniquely identify a particular item (issue) and/or contribution (article) of a serial. The SICI code provides the capability to construct a database of locally held journal articles instead of (or in addition to) journal issues and volumes. The SICI standard will enable the precise indexing and searching of a file of article images. Some of the major scientific publishers, including Elsevier, Pergamon, and Wiley have pledged to print the barcoded version of the SICI item/issue identifier on the cover of journal issues starting in 1992. In effect, the SICI has the potential to revolutionize serials transactions in much the same way that Chemical Abstracts Service Registry Numbers for chemical compounds have revolutionized the searching of chemical databases--by providing a uniform access point to all references to a unique item. The SICI will surely become a standard data item in the records of the serials abstracting and indexing services. Then it should be straightforward to link the references retrieved from such searches to the file of original documents, whether in a remote or local database.

What is being proposed differs substantially from the system of document delivery and document use which is now in place. How can we enlist the support of the major players (publishers, document suppliers, librarians, and users) in order to transform the current system?

The two most difficult groups to convince will be the publishers and the users. Publishers will probably raise the threat of copyright infringement in order to preserve the journal volume instead of the journal article as their smallest marketable unit. The legal aspects of the reuse and repackaging of electronic data are murky at best. Libraries and publishing companies must confront and solve the copyright questions before significant progress toward the electronic library can occur. The owner of a nightclub who permits musicians to perform copyrighted music without paying royalties to the composer of the work is liable for damages. So too might librarians be held liable if we create databases which allow our patrons to illegally use the data without proper compensation to the copyright holder. Nevertheless, librarians must move the concept of cooperative collection development to its logical conclusion: that individual libraries will no longer be the customers of journal publishers. Instead, the publishers must deal with library consortia (alliances) whose members will collectively decide what titles to buy and in what numbers within the consortium. In that regard, the Research Libraries Group 4+ Project provides a good model for defining the core collections in various disciplines.

Another approach, involving Elsevier Science Publishers, is TULIP, The University Licensing Program.16 The 3-year program, scheduled for 1992-94, is described as the first step by Elsevier and 12-16 participating U.S. university libraries to develop a "viable" means for the electronic distribution of scientific information, defined as "economically and functionally acceptable to all parties in the distribution value chain."17 Limited to 35 materials science and engineering journals, the TULIP collection of Elsevier titles is said to encompass approximately half of the core journals in this area. Tapes will be shipped to one or more of the participating institutions who will make the data available to the others over the Internet. Interestingly, usage statistics in future reports about the use of the journals distributed electronically through the TULIP program must not identify specific journals and articles within journals. An open enrollment option will be available from January-June 1992 only to members of the Coalition for Networked Information who subscribe to at least half of the titles in the program. These two facets of the program perhaps reveal the publisher's concern about maintaining journal subscriptions. Participants in TULIP will make a single annual payment to cover single campus uses of the titles, with no further per-use payments required. However, there is an option to sell individual articles on demand to non- campus customers.

The TULIP program is significant in that a consortium of academic libraries with common interests has engaged in a research project with a major publisher. Publishers must find ways to work with library consortia to define the "most marketable unit" of a journal title, the collection of articles which will have the greatest chance of being purchased by all members of a consortium of large academic libraries. Just as scientific authors seem to have embraced the concept of "least publishable unit" for maximum exposure of their research results, so too must the publishers develop the concept of the "most marketable unit" for the continued survival of the scientific journal as we know it today. The "least marketable unit," the journal article, will be defined by the libraries who import the images of individual journal articles into their own local databases. Publishers must decide how and to whom the least marketable unit will be sold. Will it be available from document suppliers or libraries or simply be deposited in an archive at the publisher's location to be ordered at will by whoever needs it?

The users of journal articles must next be convinced that it is to their advantage to accept the concept of access. In order to succeed, librarians must define access in a much broader sense than is currently employed in the library world (access to materials not held locally). We must extend the concept of access to our own local collections and provide better (electronic) delivery even of the material we own. It is a fact that the single most valued service in an academic science library today is the provision of well-functioning photocopy machines. Science faculty expend enormous amounts of effort, directly or indirectly, to copy articles and to enter the bibliographic records for them into their personal information systems. The paper copies themselves generally are added to the considerable store of filing cabinets maintained in most faculty offices to house their "personal libraries."

It is this cycle of information gathering that science librarians must tap into in order to win faculty support. We must turn our attention to building local electronic repositories of document images and the attendant bibliographic records which will serve our primary clienteles better than is possible with the current system. More importantly, we must enlist the support of our users in the design of local multi-purpose document archives and indexing systems. The design must incorporate incentives to channel the energy and money currently being expended on photocopying into the production of the local database.

The source of the document images received by the user could be the library's local collection, as well as the collections of consortium partners, document suppliers, or publishers. The database would be searchable over a local- or wide-area network, with images delivered simultaneously to the user's workstation and to the locally maintained library database. Eventually, initiation of document delivery orders for items not locally held could be a shared responsibility with the users. A copy of the image would go into the local database as well as being shipped to the user. With an appropriate scanning device to substitute for the photocopy machine and a network link to their own private files, the end-user might even be persuaded to forego making paper copies of older articles in favor of the document images. Once the images enter the database, other legitimate users should have access to them with no more authorization than the barcodes commonly affixed to user IDs for circulation purposes nowadays. To increase the attractiveness to the users, the database must have value added by providing pipelines to the wordprocessing packages employed by them. The users must be able to easily query the database, to call forth the retrieved images, and to capture bibliographic information that meshes with their own method of manuscript production. Thus, a program for fitting the bibliographic references into manuscripts at appropriate points and in appropriate formats is essential.

One might ask why all of this activity should be taking place in each local science library. Why not assign some central agency, such as CARL, Faxon, or the Center for Research Libraries to perform the work once for all of us? In addition to the obvious need to have multiple collections of significant scientific work for archival purposes, there is the economic aspect to consider. At this point in time, the scientific publishing enterprise cannot survive if we tell the publishers to sell one copy to a document delivery counterpart of OCLC and let us affix our "access symbol" to the record as needed. It is unlikely that there will be a sudden shift from commercial publishers to university presses or to free or low-cost electronic formats for scholarly journals in this decade. For the foreseeable future, there must continue to be a diverse group of subscribers to share the publication costs of scholarly journals. Sanford G. Thatcher, Director of the Penn State Press and long-time Chair of the Copyright Committee of the Association of University Presses, has spoken to this issue.18 He argues that scientific journal publishing requires substantial resources in staff expertise and capital. While anyone with the ability to start a forum on a listserver can conceivably become an electronic journal "publisher," this is not devoid of costs, both human and real. The value added by publishers cannot be denied. The scientific community needs the scholarly review process which is coordinated by the publishers. Likewise, publicity and the maintenance of a subscription service are not without cost.

Be that as it may, it must be admitted that a portion of the cost of the current scientific publishing enterprise is due to financial incentives given by the publishers to the organizers of professional society symposia and conferences. Sometimes the expense of presenting a symposium is more than offset by the royalties paid by the publisher in exchange for the right to publish the papers.

There is a pressing need for a reassessment of the values and costs in scholarly publishing. The process has begun, and it will not be a painless one. The document access system proposed in this paper is transitional. It seeks to shore up and redefine the library customer base of the scientific publishers by promoting the concept of the article as a marketable unit. Librarians, publishers, and users need to work together to define acceptable subscription prices for journals, fees for the retention and reuse of document images or abstracting/indexing data, and fees for the value-added services provided by librarians. Then we will move toward a truly electronic library.


1. Tuck, Bill [and others]. Project Quartet. (British Library. Library and Information Research Report; 76); Cambridge: Cambridge University Press; 1990. (p.249)

2. Report of the ARL Task Force on a National Program for Scientific and Technical Information. Washington, DC: Association of Research Libraries; 1991 May. (p.6)

3. See for example:

Blick, Dawn; Sinha, Reeta. Maintaining a high-quality, cost- effective journal collection. College & Research Libraries News 51(8): 485-490; 1991 September.

Chrzastowski, Tina E. Journal collection cost-effectiveness in an academic chemistry library: results of a cost/use survey at the University of Illinois at Urbana-Champaign. Collection Management 14(1/2): 85-98; 1991.

4. The RLG lists in chemistry, business, and mathematics are available from: Distribution Services Coordinator, The Research Libraries Group, Inc., 1200 Villa Street, Mountain View, CA 94041-1100.

5. Ralli, Tony. Performance measures for academic libraries. Australian Academic & Research Libraries 18(1):1-9; 1987 March. (p.4)

6. Egan, Dennis E. [and others]. Creating and using an electronic library: Progress report on the CORE project. In Proceedings of EDD '91; Belcore/BCC Conference on Electronic Document Delivery; 1991 March 25-28; East Brunswick, NJ. (pp.238-251)

7. Ibid. (p.239)

8. Tuck. Op. cit. (p.3)

9. Ibid. (p.244)

10. Ibid. (p.248)

11. For more information on CARL's UnCover service, contact:

CARL Systems, Inc., 3801 E. Florida Avenue, Bldg. D, Suite 300, Denver, CO 80222 Phone: 303-758-3030 FAX: 303-758-0606.

The Minutes of the September 24-25, 1991 OCLC Research Libraries Advisory Committee indicate that OCLC is jointly developing a serials table of contents and document delivery project with Faxon. The database will contain bibliographic records from about 10,000 journal titles and will be linked to library holdings information. Members of the committee noted that "Faculty will not want to pay individually for such documents and will want the institution to cover charges." (p.6) They also note that "Un-mediated document supply requests directly from the user are inevitable." (p.6) For further information, contact:

Faxon Research Services, Inc., 14 Southwest Park, Westwood, MA 02090 Phone: 617-329-3350 ext. 407; FAX: 617-329-6291.

12. Garfield, Eugene. No-growth libraries and citation analysis; or, pulling weeds with ISI's Journal Citation Reports. In Essays of an Information Scientist; v. 2: 1974-1976; Philadelphia, PA: ISI Press; 1977: pp. 300-303. (p.301)

13. Garfield, Eugene. How ISI selects journals for coverage: quantitative and qualitative considerations. Current Contents 33(22): 5-13; 1990 May 28. (p.6)

14. The equipment and software necessary to establish a document archive database are available commercially. For example, Compulink's LaserFiche LAN can store document images and be integrated into a Novell network. The system handles computer files, paper documents, and FAX transmissions. The so-called "hydra" machines, multi-function devices for printing, faxing, scanning, and copying, could also be adapted for such an application. See Byte 16(4): 217; 1991 April for a list of companies that provide document imaging solutions.

15. Feick, Tina. SISAC news: ANSI/NISO Z39.56 1991 has passed! and SISAC update. Newsletter on Serials Pricing Issues NS 13; 1991 November 13.

16. Elsevier Science Publishers B. V. TULIP, The University Licensing Program. Working Plan 1; 1991 October 10.

17. Ibid. (p.2)

18. Thatcher, Sanford G. Letter, 1991 December 3, to Ann Okerson, Office of Scientific and Academic Publishing, Association of Research Libraries.