1994

In presenting this year's annual report, I shall first comment on some of the events that marked the year, paying particular attention to developments that are likely to interest potential grantees and other readers. Immediately following this discussion is a special section describing the current status of the Foundation's electronic Journal Storage Project (JSTOR), which is something of a departure from the Foundation's normal mode of operation. Then, consistent with past practice, I yield the balance of the pages to a colleague who will write about a particular program or topic. This year, I have asked Stephanie Bell-Rose to discuss the Foundation's activities in the field of field of immigrant policy studies. Other program discussions from previous years include minority fellowships, eastern europe, culture and the arts, population policy, and literacy. 

Although the topics emphasized in this report are varied, they may suggest, collectively, more of a focus on public and social policy than one usually associates with this Foundation. That would be a misreading. All told, we made 351 grants in 1994, as shown in the detailed listing at the back of the Foundation's printed annual report. The distribution of appropriations by field and type of activity is consistent with long-established patterns at the Foundation. A

These emphases are appropriate, in our view, because of the intrinsic value of these fields and because they appear to be of less interest to a number of other sources of philanthropy. The Foundation is fortunate in that it is able to sustain these historic commitments while simultaneously supporting other activities and projects (sometimes for more limited periods of time), which it believes also have high potential.

The private grantmaking foundations, perhaps more than any other type of organization, nonprofit or for-profit, depend heavily on the wisdom and guidance of their boards of trustees. The lack of market sanctions and the relative absence of active and often critical constituencies (such as faculties and student newspapers, in the case of colleges and universities) combine to vest trustees of foundations with great responsibility, as well as considerable freedom of action. A book, Inside the Board Room, published this year contains a fullerexplanation of the reasoning behind this conclusion.

One test of any board is its ability to achieve smooth transitions in its membership. In 1994, the Trustees of The Andrew W. Mellon Foundation faced the task of having to replace John R. Stevenson, an attorney with a distinguished career at Sullivan and Cromwell and a former Legal Advisor to the Secretary of State, who retired as Trustee after having served the Foundation faithfully since 1975. An exceptionally conscientious Trustee, Mr. Stevenson combined his interest in law (especially the law of the sea), and in art and art museums, with special knowledge of Latin America.

At their June meeting, the Trustees elected W. Taylor Reveley, III, of Richmond, Virginia, to succeed Mr. Stevenson. Mr. Reveley, also a distinguished attorney, is with the firm of Hunton & Williams and was a law clerk to Justice Brennan. An expert on environmental law, he is also a devotee of liberal arts colleges, the author of a well-known book on the War Powers Act, and an experienced trustee, having served on the boards of numerous nonprofit organizations in the fields of art and culture, religion, and higher education.

This Foundation has been unusually fortunate in the close working relationships that have existed between the Trustees and its presidents. At their December 1994 meeting, the Trustees voted to recognize the debt which the Foundation owes to two past presidents, John E. Sawyerand Nathan M. Pusey. The Trustees decided to pay tribute to Mr. Sawyer's leadership by naming the Foundation's new program of seminars on comparative historical and cultural studies, and the graduate and postdoctoral fellows supported through them, for him: they will be known henceforth as the "Sawyer Seminars" and the "Sawyer Fellows." This action seemed particularly appropriate because the new seminars program (described later in the report) is in many ways an intellectual offspring of earlier programs at the Foundation initiated by Mr. Sawyer, including especially the "Fresh Combinations" program. The 1987 Annual Report (pdf) contained a review of the "Sawyer Years" at the Foundation. At a still earlier point in the evolution of the Foundation, Mr. Pusey played a critically important role in pursuing the development of programs appropriate to a truly national foundation. In recognition of his contributions, the Trustees have named the Foundation's handsome new library for him -- an action which also seemed especially appropriate in light of Mr. Pusey's leadership in establishing the Foundation's first library and appointing its current director, Kamla Motihar.

As a direct beneficiary of the contributions made by Presidents Pusey and Sawyer in defining the style of the Foundation, I take special pleasure in recording these well-deserved recognitions in this report. It is also my sad duty to add a postscript: On February 7, 1995, Mr.Sawyer died at his home in Woods Hole, Massachusetts. He will be missed sorely by his legions of admirers at Williams College, at the Woods Hole Oceanographic Institute, at this Foundation, and at the innumerable institutions, worldwide, which his philanthropic instincts and personal dedication nurtured and strengthened.

In 1994, the Trustees also took advantage of a singular opportunity to honor Paul Mellon's lifelong support of the National Gallery of Art by making the largest appropriation in this Foundation's history: an endowment grant of $15 million. The National Gallery has a unique place in the cultural life of the nation and a unique relationship to the Foundation, which has been a staunch supporter for many years. The purpose of this latest grant was to strengthen the capacity of the Gallery to serve its basic scholarly and curatorial objectives, including the conservation of art, to which the Foundation has been a primary contributor. A further purpose of this leadership gift, served in part through its timing, was to encourage contributions from others toward the successful completion of the Gallery's New Century Fund campaign.

Because of the unique circumstances which led to this grant, the Trustees decided that it should not displace other grants which the Foundation would make through its regular programs. Thus, this appropriation was provided over and above the regular grantmaking budget -- which explains why total appropriations in 1994 rose to nearly $120 million, compared with the lower-than-planned total of $90 million in 1993. In 1995, the staff members of the Foundation are working within a grantmaking budget of $105 million.

I have chosen to concentrate the remainder of my comments in this report on: (a)programs in which the Foundation has largely completed its main work, at least for the time being; and (b) areas of new or increased emphasis. Many of the Foundation's most important ongoing programs cannot be placed within either of these domains (the Mellon Minority Undergraduate Fellowship Program is one example, support of the performing arts is a second, and basic research in ecology is a third). A number of these other "main line" topics have been discussed in recent reports, and in future years I expect to continue the practice of focusing each year's report on a limited set of subjects.

Some of the Foundation's programs are designed to last for limited periods, and disciplined efforts are made to conclude such initiatives in a timely way so that the Foundation has the flexibility to increase its support in other areas, as new opportunities arise. Three areas of declining programmatic emphasis are described below: (1) literacy; (2) Eastern Europe; and (3)Foundation research on nonprofit institutions and the NTEE codes.

Literacy

In 1994, the Foundation made its last group of large grants (totaling over $8 million) for research into fundamental problems of cognition and learning as they apply to the acquisition of literacy. The Russell Sage Foundation, and its president, Eric Wanner, have worked in partnership with the Foundation in selecting and reviewing promising initiatives in this broad field, and we believe that some pathbreaking work is being done under the sponsorship of the two foundations.

From its beginnings in 1988, the Foundation's program has had objectives far more ambitious than those of most national efforts to improve literacy. The goal has not been simply to improve children's mastery of the three Rs, but to help children become actively literate across a broad array of intellectual competencies. Active literacy is the ability to acquire knowledge on one's own and to use it in solving new problems; it entails a capacity to think critically about new information and to learn independently after the scaffolding of the classroom has been removed. In pursuit of this goal, the Foundation's program has attempted to capitalize on the findings of basic research about the learning process that have emerged in the cognitive sciences over the past 20 years, especially as they might be applied to the educational needs of disadvantaged students. Over the past six years, the Foundation has supported seven major research projects and programs.

Experience has taught us to respect the difficulties of bridging the gap between basic cognitive research and educational practice. Cognitive scientists often have difficulty breaking away from abstract studies of learning in laboratories in order to make full contact with the inherently messier problems of achieving educational change in classrooms. Of the seven projects the Foundation has supported, two have taken that difficult step and attempted to design and sustain functioning learning environments. One is the approach devised by Ann Brown and Joseph Campione of the University of California at Berkeley which, while focusing on environmental science, encourages students and teachers to become a community of learners. It is now being used in Oakland elementary schools. The other is an innovative computer-based after-school program (the "5th Dimension"), designed by Michael Cole of the University of California at San Diego, which is now being studied in six cities across the country.

Both of these projects demonstrate, albeit in quite different ways, the potential benefits of applying cognitive science to the fostering of active literacy. But just as these projects convince us of their educational promise, they also remind us of their fragility. To sustain and spread an educational innovation, it is necessary not only to define a curriculum, but also to teach teachers how to teach it, to create a battery of supporting materials and exercises (often computerized), to develop a strategy for implementing the program in new settings, and to devise methods of evaluation that can be used to detect whether the program is achieving the desired results -- a particularly difficult problem when the goal is more a set of mental processes than mastery of a specific subject matter. We are persuaded that educational innovations conceived by cognitive scientists will not have discernible impact on American education until and unless they address all of these problems, which may appear mundane from the point of view of the laboratory but are, in fact, challenging puzzles in their own right.

Over the past two years, we have worked to develop a new model for organizing research efforts of this more applied kind. The essential idea which underlies this model is that learning how to sustain and spread educational innovations is best achieved by research consortia or partnerships, each one organized around a particularly promising curricular innovation. It is now clear that excellent people in cognitive science and education are willing to commit themselves to working on applied projects in this partnership mode. Accordingly, as its final large investment in the field of literacy, the Foundation made two sets of grants in 1994 to collaborative projects intended to advance the work begun at Berkeley and at San Diego.

Eastern Europe

The Foundation is also proceeding with its previously announced plan to reduce its level of new commitments in Eastern Europe. An additional $8 million was appropriated for projects in Eastern Europe in 1994, beyond the total of approximately $29 million appropriated over the previous five years. The purposes and characteristics of the Foundation's activities in Eastern Europe were described in detail by the Foundation's Senior Advisor in this area, Richard E.Quandt, in the 1992 annual report (pdf).

The Foundation has worked especially hard to promote business training and library collaborations and computerization, and we believe that, as a result, a very much stronger infrastructure, promising long-term support of these activities, has been put in place. Progress continues to be made in institutionalizing and preserving the most successful initiatives undertaken with Foundation support. In 1995, we anticipate investing an additional $5 million to complete the most important of these projects. In 1996, we anticipate that further support for projects in Eastern Europe may be in the $2 million range.

Foundation Research on Nonprofits and the NTEE Codes

As a final example of an area in which work was largely concluded in 1994, it is gratifying to report that the National Center for Charitable Statistics received a final grant from the Foundation to complete its massive recoding of the National Taxonomy of Exempt Entities (NTEE), which is used to classify nonprofit organizations by field of activity. Even more encouraging is the specific purpose of the final grant: to institutionalize the new method of classification within the Internal Revenue Service, thereby ensuring the regular updating of the files and the proper classification of new entities. As a result, it should now be possible for students of the nonprofit sector, and other users of the Business Master File of the IRS, to examine trends within defined fields of activity and to analyze the tax returns (Form 990s) filed by all 501(c)(3) organizations within specified fields.

In 1994, staff of the Foundation also completed research of their own on the nonprofitsector. First, staff members published a broad-gauged study of the"institutional demographics" of the nonprofit sector. This collaborative work is intended to introduce readers to the wealth of data contained in the Business Master File, to "frame" the nonprofit universe by describing patterns of institutional births and deaths by field, and to test propositions such as the assertion that earned income is becoming a significantly larger share of total revenues for charitable nonprofits (by no means always true, we found). We hope that this publication proves useful in stimulating additional research that will assist the work of nonprofits, inform grantmaking activities of foundations, and contribute some basis of factual knowledge to public policy debates.

Research by Foundation staff on two companion studies was also largely completed in 1994. Jed Bergman finished his study of the histories of five independent research libraries, in which he describes their transformation from "income spenders" (able to finance their activities almost exclusively from existing endowments) to "fund seekers" (obligated to seek large amounts of contributed income on a continuing basis if they are to balance ever-increasing budgets). Kevin Guthrie also finished almost all of his research on the evolution of The New-York Historical Society -- a saga which, in its most recent stages, has attracted much attention in the press. The Society's very survival has been threatened, and some of its actions have been highly controversial (including a deaccessioning program that led to the auctioning ofEuropean paintings). The tortuous path traversed by the Society, from the early 19th century to the present day, yields lessons for other entities facing similar challenges and poses large questions of public policy. We expect both of these studies to be published by Jossey-Bass, in either late 1995 or early 1996. Our own staff members are now working primarily on projects in higher education, and for the foreseeable future, the Foundation's support of research on the nonprofit sector will be limited to studies carried out by others.

As always, the Foundation attempts to identify and address new opportunities, and in the following pages I discuss three areas in which we have either started new work or increased emphasis in 1994: (1) comparative studies of the historical and cultural origins of contemporary issues; (2) the impact of electronic technologies on scholarly communication and libraries; and (3)research on higher education.

Comparative Historical and Cultural Studies: The Sawyer Seminars

In last year's report, I described in some detail the Foundation's thinking about scholarly work classified under the area studies rubric, and the reasons why it has seemed less likely than it once did to provide an optimal framework for many kinds of training and research. Subsequently, 17 universities judged among the strongest in the humanities, social sciences, and studies of foreign cultures were invited to design new seminars intended to serve the following purposes: to provide an opportunity for comparative study of the historical and cultural origins of contemporary social, political, and economic developments; to engage university faculty and advanced graduate students in research on topics that transcend regions and time periods; and to allow such work to proceed outside permanent organizational forms which might impose ongoing financial burdens on universities.

The initial responses from universities to the letter announcing the program were enthusiastic. It was evident that there was widespread interest in regularizing collaborations between humanists and social scientists, allowing greater graduate student traffic across departmental boundaries and, very important for some universities with largely autonomous area studies centers, in beginning to bridge the gulf separating the disciplines and centers.

A distinguished advisory panel was appointed by Harriet Zuckerman, the Foundation's staff member with primary responsibility for this program, to help in making selections among the 59 different seminars proposed by the 16 universities that ultimately made submissions. In the first round of what we hope will be a continuing competition, 20 seminars were chosen forfunding. Perhaps the best way to illustrate the faculty responses to this initiative is by describing a few of the first group of Sawyer Seminars.

• One seminar to be offered at the University of Chicago is "resolutely historical" and focuses primarily on Europe. It examines the historical roots of modern ideologies oftolerance, the development of and opposition to ideals of tolerance in philosophical, religious, scientific, political, and artistic contexts in the 16th through the 18th centuries. The seminar will attempt to determine how the idea developed that diversity and contention are themselves valuable and to differentiate tolerance in this sense from tolerance of diversity as a necessary evil. It will trace the emergence of practices of tolerance and repression, especially but not only by drawing on the history of the theater and music for case materials.

• A seminar to be offered at Duke is titled Public Culture and Transnationalism. It begins with the premise that contemporary transnational flows of people, goods, and ideas in postcolonial South Asia and Africa, Latin America, and East Asia are arresting but far from new. The seminar will compare these contemporary developments with earlier instances of migration, cultural diffusion, and economic exchange, will examine the connections between these flows and economic development, and will take up the effects of transnational flows on public culture.

• Members of the departments of comparative literature, English, and romance languages at Harvard proposed a seminar that will involve, among others, lawyers, political scientists, historians, and a neurologist-linguist from MIT. The seminar will study language and nationhood in North America--why, for example, Americans cherish the notion that the nation was originally monolingual when bilingualism has always been common, how constitutional bilingualism has worked, and the influence on national identities of "cultural contact zones," borderlands where several languages are spoken and nationality is blurred. Puerto Rico, Quebec, and the United States are the principal cases to be examined; sessions will also take up the German experiment in Pennsylvania, the demise of Native American languages, enforcedspeaking of English or Spanish by slaves, the contributions of newspapers and musical lyrics in foreign languages to the shaping of collective identities and, finally, multilingual punning.

• A seminar to be offered at Northwestern, on Islamic fundamentalism, begins with the medieval antecedents of the seeming paradox that Muslims revere Islamic principles of governance but, nonetheless, are willing to accept political systems they regard as at odds with those principles. The tension between the ideal and the real in politics in Islam will be traced historically and examined in the contemporary period in the heartland of Islam and in Muslim populations in Western Europe and West Africa. The seminar is to involve Northwestern faculty and students who work in history, anthropology, comparative religion, and Asian and African languages.

If, as we expect, these seminars prove successful as settings for advanced training as well as for stimulating scholarship and research, the Foundation will invite additional proposals.

Electronic Technologies, Scholarly Communication, and Libraries

Over the next three to five years, the Foundation expects to invest heavily in the broad field of scholarly communication. We intend to give special attention to the implications of electronic technologies for publishing scholarly materials, conserving them, and making them more readily available to users, especially faculty members and students working in college and university libraries. The rationale for making such investments, especially at this time, has been described elsewhere, most particularly in a Foundation publication distributed widely by the Association of Research Libraries and in last year's annual report. The Foundation is employing two distinct approaches in choosing among the myriad opportunities for useful work in this rapidly developing field.

First, the Foundation itself, working with key partners (particularly faculty at the University of Michigan and an able advisory committee), has undertaken to develop an electronic journal storage prototype called "JSTOR." In brief, JSTOR is intended to provide a highly convenient and cost-effective mode of access to the back issues of core journals in the arts and sciences. Because of the combination of legal, technical, managerial, and economic issues that had to be resolved in order to launch even a pilot version of JSTOR, it seemed better to oversee the work directly than to ask an external entity to coordinate all aspects of such a complex process. Fortunately, the Foundation has been able to engage Ira H. Fuchs, Vice President for Computing and Information Technology at Princeton University, as Senior Advisor and technical director of JSTOR. In our view, JSTOR has great promise, and its current status is described in some detail in a separate section at the end of this part of the report.

Second and more predictably, the Foundation is using its standard grantmaking apparatus to support an array of projects designed by leading researchers and practitioners at a wide variety of universities, university presses, libraries, and professional associations. Richard H. Ekman and Richard E. Quandt are directing this program, and I describe below the thinking that has guided its development, the criteria employed in recommending grants, and, to illustrate current priorities, some specific grants made in 1994. In designing the Foundation's grantmaking program, Messrs. Ekman and Quandt decided to place considerable emphasis at the outset on "electronic publishing," a term which covers a variety of technologies and approaches. The most widely encountered forms of electronic publishing at present include: (1) network-based distribution of computer files containing the text of preprints; (2) the CD-ROM or network-based distribution of abstracts or bibliographies; (3) the network-based distribution of journals or individual journal articles; and (4) the CD-ROM or network-based distribution of multimedia products, which allow the user to make easy transitions among text, visual images, and sound images. Some of these applications are created with relatively little effort, but are also relatively crude in appearance and function, while others have required much greater investments, but provide the user with hypertext links and powerful search engines.

It is generally the case that these electronic products have been created with more enthusiasm for the promises of the new technologies than careful calculation of the costs of alternative approaches. In particular, the creators of these electronic publications have characteristically volunteered their time for organizing and running electronic publishing ventures and have used computer time that otherwise would have been unclaimed on existing computers. There is no doubt that the various electronic publications have enhanced the ability of scholars and students to gain access to literature by increasing the scope of the materials that can be accessed and reducing the time necessary to do so. Indeed, some have estimated that it costs 10-30 percent less to produce an all-electronic journal than a comparable hard-copy one. Yet, the economics of electronic publication remain a largely unexplored subject.

In making initial grants in this somewhat amorphous field, the Foundation has sought tosupport a variety of "natural experiments" selected on the basis of their promise and their consistency with two basic grantmaking criteria. We decided, first, to concentrate our support on projects that imaginatively combine existing technologies, since there seems to be enough hardware and software available to permit the testing of a wide variety of ideas. Second, we decided to require that close attention be paid to economic realities and to the cost-effectiveness of various approaches. Thus, we have asked all of those who submit proposals to incorporate procedures for measuring the capital and current costs of the proposed new developments and for comparing them to the costs of alternative (usually traditional) approaches. Also, grantees are expected to monitor closely the acceptance and use of these products by students, faculty, and other users.

The nine grants made in 1994 under this specific program are listed in the printed annual report of The Foundation (as are the grants made in support of JSTOR). Let me highlight four of them.

• An appropriation of $150,000 was made to the Massachusetts Institute of Technology to establish an exclusively electronic journal to be called The Chicago Journal of Theoretical Computer Science. Publication delays in computer science journals are typically one to two years and subscription prices are at the high end of the price range for scholarly periodicals. MIT Press anticipates that a new electronic journal, with editorial leadership based at the University of Chicago, will provide useful price competition for existing journals. The new journal will maintain the customary professional standards used by scholarly journals but guarantee electronic publication of an article within two weeks of acceptance. Subscriptions will be sold to individuals (at $30/year) and libraries (at $125/year) with essentially no restrictions onhow the published articles may be used.

• An appropriation of $400,000 was made to Johns Hopkins University, for use by the Johns Hopkins University Press in support of Project Muse, an effort to make available online all 40 of the scholarly journals the press currently publishes. A prototype of the project was recently completed, based on current issues of three journals, Configurations, Modern Language Notes, and English Literary History , and these journals are now available on the Internet with fully formatted text; subject, title, and author indexes; hypertext links to tables of contents, endnotes, and illustrations; full searchability of both text and tables of contents; and improved illustrations. A key feature of Project Muse is that it involves established and distinguished journals, including core titles in several disciplines, rather than new, untested journals. One of the innovations of Project Muse has been to develop software, freely available to anyone who wishes to use it, for converting the compositor's PostScript files, used in the production of printed journals, to Hypertext Mark-Up Language (HTML) files, which are used by Mosaic clients. This process enables Hopkins to prepare at low cost substantial amounts of text for online publication.

• An appropriation of $700,000 was made to Columbia University to develop a comprehensive evaluation methodology for its ongoing activities in creating a digital library. The project's objective is to make available within the Columbia community in online, networked electronic form (1) Columbia University Press reference works, such as the Concise Columbia Encyclopedia, Granger's Index to Poetry, the Columbia Dictionary of Quotations, the Columbia Guide to Standard American English , and other works;(2) some 100 Oxford University Press monographs in selected fields, starting with neuroscience;(3) selected high-use titles from Simon and Schuster; and (4) various texts for the study of the humanities, such as Greek and Latin texts and primary works in philosophy, religion, and history. Readers will be offered the opportunity to read online, request prints of sections or whole books, and place online orders for paper copies of books. All costs will be tracked, but, at least at this stage, no per-use charges will be levied. This project also intends to examine alternative models for acquiring and sharing intellectual property.

• An appropriation of $750,000 was made to Cornell University to coverpart of the costs of the "Making of America" digital library project. Cornell University, in cooperation with the University of Michigan, sought support for the pilot phase of a multi-institutional partnership designed to preserve and make accessible through digital technology a significant volume of primary source material on the history of the United States. During the first phase, 10,000 monographs documenting the period from 1850 to 1877 will be selected, scanned, and made available online across institutional boundaries. 

Research in Higher Education

We have concluded that it would be timely for the Foundation, over the next few years, to encourage and sponsor a series of studies of issues important to higher education -- many, but not all, to be conducted by outside investigators. In developing a current research agenda, our intention is to build on past work. Over the years, the Foundation has sponsored a number of studies of higher education, with two main objectives in mind. The first has been to understand better, and to help others understand better, factors affecting the well-being of higher education as a whole, including particular groups of colleges and universities. The intended audience has been policy-makers, leaders of colleges and universities, and students of higher education. A second objective has been to inform the grantmaking activities of the Foundation in the broad field of higher education and scholarship -- the program area in which the Foundation has invested most heavily from its earliest days.

In the normal course of events, the Foundation no doubt would have continued to support research in higher education -- but in an ad hoc manner. The case for proceeding more systematically rests, in part, on our sense that this is a time of considerable uncertainty about the role of higher education in this country and abroad and much debate (often unfocused and sometimes confused) over key issues. An advantage of sponsoring a wide-ranging series of studies at more or less the same time is that, in some instances, participating scholars may benefit from one another's work. In addition, we thought that some leading scholars might be more inclined to work in this area (which has not always enjoyed high prestige) if they knew that other able people were involved. Finally, the appearance of a series of studies within a limited period might call more attention to the importance of the issues under discussion than the sporadic publication of individual reports and books.

Our current research agenda contains approximately 20 projects. Some are historical or comparative in character. For example, the Centre for the Study of Education and Training at Lancaster University received a grant in 1994 to carry out a comparative study of the academic professions in England, Germany, the Netherlands, and Sweden. This research will build on surveys in 15 countries coordinated by the Carnegie Foundation for the Advancement of Teaching. The Lancaster study aims first to create a comparative profile of the academic profession -- its composition and activities -- in these four countries. Then, the analysis will focus on topics of importance in rapidly expanding systems of higher education (in Europe, a process termed "massification") such as planned and unplanned diversification of institutions, the situation of women, the roles of junior staff, and the effects of government efforts to manage university affairs.

Other research projects are concerned with the changing characteristics of various fields of study. We are particularly excited about a projected volume of essays on changes in the humanities over the last 25 years. The proposed topics include: "Patronage, Who Pays? The Marketplace and the Humanities;" "The Humanities as Social Conscience;" "What is Taught in the Humanities Classroom and How," and "Print, Books, and Libraries in the Electronic Age." This project is being directed by Alvin Kernan, Senior Advisor to the Foundation for the humanities. Some of the faculty members asked to contribute essays will also be invited to spend time at the National Humanities Center working on their topics and meeting with other scholars in residence.

Still other projects are focused quite directly on contemporary policy issues. Michael McPherson of Williams College and Morton Schapiro of the University of Southern California area to work on a new study of how families pay for college, which will update and extend the results reported in their earlier book, Keeping College Affordable. New data of high quality are available from panel surveys, and McPherson and Schapiro are optimistic that these resources will permit even sharper conclusions to be drawn concerning the effects of various financial aid policies on access and choice.

In conducting this new research, McPherson and Schapiro also intend to take advantage of a parallel study being carried out by two Foundation staff members, Elizabeth Duffy and Idana Goldberg, who are conducting an in-depth analysis of trends since the mid-1950s in applications, offers of admissions, financial aid outlays of various kinds (including the provision of "merit aid"), and enrollments at a wide array of private and public institutions in Ohio and Massachusetts. One objective of their research is to understand more fully than anyone does at present what has happened to the pools of candidates for liberal arts colleges and to the nature of competition for prospective students. The budgets of most of the private schools are "tuition driven" (that is, are heavily dependent on tuition revenue), and thus their futures depend in no small part on projected trends in enrollment and associated expenditures on financial aid. At the same time, these institutions are committed to maintaining (or improving) their quality and the diversity of the student populations which they serve.

A study of a very different kind is being conducted by Charles Clotfelter of Duke University, under the aegis of the National Bureau of Economic Research. Clotfelter is examining trends in institutional costs at three research universities (Chicago, Duke, and Harvard) and at one liberal arts college (Carleton). A distinctive characteristic of Clotfelter's work is that it is based on data collected at the departmental as well as the institutional level. He expects to be able to learn, for instance, to what extent unit costs have risen more rapidly in a science department than in a department in the humanities or social sciences -- and, if the costs of doing science have risen especially rapidly, to indicate why this has been the case.

The Foundation is also sponsoring a major study of faculty retirement. Orley Ashenfelter and David Card of Princeton University are working with data provided by TIAA-CREF to analyze retirement patterns and the reasons for them. Now that mandatory retirement has been ended, it is important to know how decisions to retire, at various ages, are related to variables such as: gender, marital status, and family obligations; fields of study; salary histories and salary levels; personal wealth, as reflected partly in TIAA-CREF accumulations; the structure of pension plans (defined contributions versus defined benefits); early retirement incentives of various kinds; pre- and post-retirement benefits (especially health benefits); working conditions (teaching loads, access to libraries and laboratories, quality of students and colleagues); and the individual's professional accomplishments and standing as a teacher and scholar (will the abler faculty retire earlier?). Ashenfelter and Card propose to analyze retirement decisions both retrospectively and prospectively, and their research should be of considerable value to institutions concerned with retirement issues and prospective faculty staffing, as well as of general interest to labor economists.

To provide a fuller sense of the range of research being carried out under the Foundation's higher education research agenda, I should note, first, that Foundation staff are continuing to study the effects of current grantmaking programs such as those intended to improve doctoral education while reducing attrition and time-to-degree and those designed to increase the number of minority students pursuing PhDs in the arts and sciences. Also, a wide variety of other studies are underway, including historical investigations of the interplay between religion and science in 19th century American universities; the alternative ways in which colleges and universities have sought to "manage pluralism" and to benefit from more diverse student bodies; and the evolving role of athletics within academically selective colleges and universities. In time, we hope to sponsor more work on the governance of higher education and to encourage broad-gauged studies of the historical evolution of systems of higher education in Europe and other parts of the world.

The original impetus for JSTOR came when I was informed, in my capacity as a trustee of a liberal arts college (Denison University, in Granville, Ohio), that the shelf space in the collegelibrary was filled to overflowing and that the college had to contemplate spending another $5million to relieve the crowding and permit new acquisitions to be shelved. Inspection revealed that journals published prior to 1990, combined with government publications, occupied more than a quarter of all shelf space. I suspected that this was a common problem for college and university libraries, likely to become even more serious over time, and it seemed to me that electronic technologies ought to be helpful in finding new solutions. As is so often the case, the objectives of the project became more ambitious as it evolved. While JSTOR is still intended to save space and capital costs, we are now giving more attention to its immediate utility to users. This shift in emphasis is due in large part to what we have learned about both the abundant technical possibilities for creating a highly accessible database and the associated economics, which are very encouraging.

But I am getting ahead of the story. Let me go back and restate the objectives of the project as they are now understood:

1. To create faithful electronic replications of back issues (defined as those publishedprior to 1990) of core journals that will be of archival quality -- in order to address issues of conservation and preservation such as broken runs (missing issues), mutilated pages, and long-term deterioration of paper copy;

2. To improve dramatically access to journal literature for faculty, students, andother scholars by linking bitmapped (fax-like) images of the pages of journals to one or morepowerful search engines -- and, in many instances, by providing access to much more complete sets of journals than now exist on the library shelves of particular campuses;

3. To study the effects of providing high-quality electronic access on the usage of the back issues of journals in a variety of academic settings; and

4. To address some of the vexing economic problems of libraries by easing storage problems, saving the prospective capital costs involved in building more shelf space, and reducing operating costs associated with retrieving back issues from the stacks and reshelving them.

A frequently asked question should be answered directly: Why does JSTOR cover only issues of journals published before 1990? Why are current issues excluded? There are two reasons. First, current issues of journals are generally more readily accessible than back issues. Second, JSTOR seeks to avoid creating economic difficulties for publishers who generate theoverwhelming share of their revenue by selling subscriptions to current issues. We want to work with the publishers of journals, and to give them something of value, not to compete with them.

While JSTOR itself will not include current issues, it is possible that some publishers will be interested in taking the electronic platform that the Mellon Foundation proposes to give them (the bitmapped images of their back issues) and connect to it electronic access to current issues. This could prove to be a highly attractive, relatively inexpensive option for publishers -- and of real benefit to users, who would then have a single means of searching all issues of a journal, from the oldest to the most recent, for a subject or author of interest.

In the Pilot Phase of the project, we decided to include

In moving from concept toward a real prototype, the first hurdle to be cleared was obtaining copyright permissions. Following extended discussions, we have now obtained permission to bitmap back issues from the copyright holders of all ten journals, and to incorporate these images in an electronic file that will be made available online, on a royalty-free basis and in perpetuity, at a limited number of sites. The principles and forms of agreement that have resulted from these discussions may themselves be of value in suggesting new kinds of relationships between journal publishers and users, at a time when there is much debate over the application of copyright laws to electronic texts. Our experience suggests that there is much common ground. We are encouraged by the enthusiasm of publishers for the project and by their desire to cooperate in every way.

At the same time that we were seeking copyright permissions, we were also exploring the best way of creating the electronic database of back issues that is central to the project. The Foundation has made two grants to the University of Michigan to oversee this process and produce an integrated prototype that will be highly user-friendly. Fortunately, JSTOR complements nicely the work that was already underway at Michigan on Project TULIP (sponsored by Elsevier) and, more generally, on digital libraries (supported in large part by NSF). Thus, the Michigan group has already encountered -- and largely solved -- many, though certainly not all, of the problems involved in making journal literature available online.

The basic electronic file is to be produced by using a scanning technique which makes bitmapped images that capture the full features of each page. These electronic "pictures" allowthe user to see on a screen the page of the journal exactly as it appears in hard copy: fonts, formats, equations, figures, graphs, and illustrations, as well as text, are replicated precisely. But bitmapping also has two disadvantages. Since the bitmapped images are stored as a series of dots, rather than as text, they cannot be searched for key words or names; also, they require far more computer storage space than does text alone.

For these reasons, we decided to add two other features to the electronic database: (1) a searchable index which allows users to obtain bibliographic references to all articles in the database containing a key word or words in their title (such as "Federal Reserve"), or written by a particular author, or published in a certain journal in a defined set of years; and (2) an Optical Character Recognition (OCR) capacity that will enable users to search the actual text pages (not just the table of contents) of journals included in the electronic file. Thus, it will be possible for a user to search for all references to "Federal Reserve" wherever the words appear, in a title, in text,or in a bibliographic reference -- and then to view on screen the articles which contain this phrase. Both of these capabilities are made possible by processing each of the bitmapped images with OCR software that scans the images and converts any characters it can "read" into text which can be stored compactly and searched easily.

The final component of JSTOR, as now envisioned, is software which will link the bitmapped images of the journals to the index and to the OCR-version of the text, thereby allowing the user to call up on the screen faithful images of the pages of the journal article that have been found by searching either the index or the text itself. The software being developed at Michigan will allow readers who have viewed articles on their screens to print them (in whole orin part) if they conclude that it would be useful to have a single hard copy for further reference.

In sum, strictly from the users' perspective, what is the appeal of JSTOR? The "selling" points are the speed, ease, and efficiency offered by electronic access to perfect replications of the full pages of back issues of key journals. From a linked desktop computer, a reader will be able to page through journal indexes and actual articles with much more speed and convenience than if standing in the stacks of even an extremely fine library and then make copies of material that is of particular interest. Printing out parts of articles at a work station would be a welcome substitute for the more laborious forms of note-taking or copying that are now needed.

While the objectives were clear enough, we did not know, until early in 1995, whether we could find a vendor who could do the bitmapping and OCR-scanning in a cost-effective way. We have taken to heart our own admonitions to others about the need to pay close attention to the economics of projects, and we were unsure how much JSTOR would cost to produce and use. Careful investigation of the capacities of a number of alternative providers of the services inquestion yielded extremely encouraging results. A contract has now been signed with DigitalImaging, Inc., of Anaheim, California, a vendor which will bitmap the pages at a very highstandard of resolution (600 dots per inch) and provide an OCR-scan that is accurate at the 99.95 percent level. Inspection of sample work confirms that such demanding specifications can be satisfied -- and at a cost-per-page that is extremely attractive.

One interesting discovery that we made in the process of obtaining bids is that working from paper copies of back issues of journals, rather than from microfilm, produces higher quality results and is -- to our surprise -- considerably cheaper. This conclusion has important implications beyond JSTOR.

The work of bitmapping, OCR-scanning, and software development is proceeding well, and we hope to have a pilot version of JSTOR, or at least of significant parts of it, up and running at five test sites in the spring of 1995: the Bryn Mawr/Haverford/Swarthmore library network, Denison University, Williams College, the University of Michigan, and (in a slightly different mode) Harvard University. Needless to say, we are eager to learn how JSTOR, for all its seeming appeal, will in fact be accepted by actual users. Efforts are underway to benchmark the current usage of these journals at the test sites so that it will be possible to gauge the impact of electronic access on patterns of use. Some have speculated that "no one is interested any more in backissues of anything." We shall see. In any event, even if usage is modest, JSTOR should have major implications for library storage.

If the pilot succeeds (which will be determined by both the application of technical standards and the responses of users), we would plan to move at once into a more comprehensive testing stage that, at a minimum, would encompass a larger number of educational institutions of widely varying kinds, including perhaps some that are overseas. Current agreements with publishers limit us, however, to 50 sites, and we are committed to a general stock-taking if and when the project expands to that stage.

Even now, we are thinking about a series of questions concerning the future of JSTOR,and a meeting of the project's advisory committee will be held in the spring of 1995 to address many of them, albeit in a preliminary way. They include:

• New features. Recognizing the dangers of "creeping featurism," and the necessity of not allowing costs to rise above acceptable levels, are there one or two new features that we should consider adding to the electronic database? We are already exploring the desirability and feasibility of adding SGML "tags" to the OCR-version of the database. Tagging would enable users to distinguish separate elements of the text, such as headings, footnotes, bibliographic references, and so on, which in turn could facilitate even more sophisticated searching, including the use of hypertext. Adding this feature might also give users the option of retrieving a highly usable OCR-form of the database that would make fewer demands on network capacity than the transmission of bitmapped images.

• More sites and fields. What priority should be assigned to including various types of institutions within the first 50 sites? How rapidly should we seek to include more academic fields? Which ones?

• Copyright agreements. If we are to move beyond the initial 50-site limitation, what kind of copyright agreement can be negotiated that will be faithful to the objectives of JSTOR and agreeable to the publishers?

• Linkages with current issues of journals. Should we explore arrangements with publishers interested in linking electronic publication of their current issues to the JSTOR-provided platform of back issues? If so, what pricing mechanism would satisfy the legitimate needs of the publishers and yet not undercut the objectives of JSTOR, which include avoiding any charge-by-use arrangement and making access to back issues available at extremely modest cost?

• Investment and cost-sharing questions. More generally, how should the future costs of JSTOR be apportioned? The Foundation has made all of the investments needed to start the project, and expects to cover all of the "up-front" costs. If the project expands significantly, however, it will be necessary for others to contribute. The potential economies of scale are truly extraordinary, since once the start-up investments have been made in developing the software and creating the electronic files, the marginal costs of adding sites are very low indeed. Adding fields and journals will, on the other hand, continue to require significant outlays, and there may also be questions of storage of the files, and transmission, to be worked out.

• Organizational issues. Any consideration of expansion of the project also raises questions of management. It is evident that the Foundation itself cannot manage the project indefinitely, and careful thought must be given to the identification of suitable institutional partners. Are there existing organizations that should take over the project or should some new entity, perhaps a consortium, be created for this purpose?

In any case, one can envision having JSTOR, or a JSTOR-like product, available to essentially all appropriate nonprofit users, in the US and outside it -- in effect, to scholars and students worldwide. Educational institutions with limited holdings of journals could be given, more or less instantaneously and at very low cost, access to journal literature that they probably never anticipated having at their disposal.

It is because of what we perceive as the great promise of JSTOR that I have written at some length about a project still very much "in process." There are many judgments yet to bemade, and we would welcome comments and suggestions so that the best ideas will emerge. Working with our collaborators at Michigan, our advisors, and other interested parties, the Foundation is committed to do all that it can to see that JSTOR achieves its full potential. 

William G. Bowen
President
The Andrew W. Mellon Foundation
140 East 62nd Street
New York, NY 10021
(212) 838-8400