The Changing Character of the National Library of Medicine, 1965
Page  1

THE CHANGING CHARACTER OF THE NATIONAL LIBRARY OF MEDICINE Martin M. Cummings, M.D.* * Presented at the Wellcome Historical Medical Library, London, England, July 6, 1965. [Mr. Poynter - distinguished guests:] It is a great pleasure for me to be here today in this center of learning with which we have had so many pleasant affiliations over the years. I plan to talk about some of the changes taking place at the National Library of Medicine with particular reference to the support of medical historical scholarship. However, before I do, I should be more than remiss if I failed to acknowledge the great debt that all of us interested in studies in this field owe to Dr. Poynter and his staff at the Wellcome Historical Medical Library. The Wellcome catalogs of manuscripts of incunabula, and of printed books through 1640 are models for all to emulate. Current Work in the History of Medicine has been a boon to medical historians throughout the world. Dr. Poynter's role in organizing the annual conference on the history of medicine in Britain has been a great stimulus to medical historical activity in [unreadable] country. I understand that he has even lured away a number of American investigators supported by the National Institutes of Health who found this library irresistible in its appeal to the scholar. In many ways, Dr. Poynter, what we are doing in Bethesda is little more than following the path you have already marked out. However, we are taking a few plunges through new territory which may be of interest to you. I am sure you will all agree that one of a library's most serious responsibilities to the scholarly community is to inform it of the

Page  2 full potential of its holdings. The time-honored card catalog, whatever its inconveniences, has on the whole served the community well. But a card catalog serves only that fortunate portion of the community which can make personal visits to important collections. Unfortunately, the publication of book catalogs serving a far wider audience has been prohibitively expensive for the majority of libraries. Hence, the remarkable and distressing fact, that of the major general collections in the world today, only the British Museum, the Bibliotheque Nationale, and the Library of Congress have published adequate records of their holdings. An encouraging sign of late has been the recent publication of volumes 1-2 of the cooperative catalog of several major Italian libraries, and I am not forgetting the more modest published bibliographic contributions of many specialized libraries. But it remains true, that to examine the resources of most libraries in Europe as well as in the United States, it is usually necessary to make a personal visit at great expense both of time and money. A certain amount can be done by correspondence, and interlibrary loan inquiry. But haphazard inquiries such as these are likely to be protracted. This is a lamentable but understandable state of affairs. Fortunately, the needs of medical historians are to some extent met by the published catalogs of our respective libraries. At the moment, our History of Medicine Division is preparing two catalogs dealing respectively with the 16th and 17th century. The first, which lists over 4,800 16th century items, is now ready for the printer, while

Page  3 the second is making steady progress. In view of the fact that our eighteenth century collections number some 40,000 items, full cataloging of these will be deferred. At the moment, we are trying to establish main entries for all eighteenth century items with a view to interfiling our checklist cards into one unified sequence. It will be several years before we publish an eighteenth century catalog, and when we do, the descriptions published will probably be considerably curtailed. Needless to say, we look forward to the publication of the continuation of the Wellcome catalog for the period 1641 to 1850. A word about our History of Medicine Division's experience with "rare book cataloging." Discussion of NLM cataloging problems goes back to World War II when for security reasons the Division was moved to Cleveland. Originally, the problem was one of simply providing a shelf-list for the books. Once this original record was completed (admittedly inadequate, as items bound with others were initially ignored), the subsequent process of adding to the catalog record became known as "check listing." The theory and methods behind this varied from time to time, notably with each change of personnel. Colonel Frank Bradway Rogers estimated that the Division had a minimum of 20,000 books published up to 1801. The time taken to catalog a rare title during 1952 was on the order of 6 hours per title, or roughly, with a staff of five catalogers, 1500 titles a year. Of these, about 500 would be newly acquired titles, so the backlog would be reduced at the rate of only 1,000 titles a year. At that rate,

Page  4 HMD's collections would, according to Rogers' estimate, have taken twenty years to catalog. Actually, the picture was blacker than he knew: the task might well have taken 60 years, for if we count in all the pamphlets and theses, which, as you well know, are often difficult to catalog anyway, our [historical] collection now totals some 60,000 items. Clearly in a situation such as this, something had to give. And the answer seemed to be: diminish excessive bibliographical detail. The scholar must be presented with a minimum of necessary information. This would comprise an accurate entry for the author, translator, editor, and contributors to a work; an accurate title page transcription; full imprint data; collation by leaves or pages (or in the case of unfoliated or un-paginated books, a collation by signatures); and the occasional provision of notes describing physical imperfections in the copy cataloged, or citing pertinent bibliographies to which the scholar could be referred for further detail. We believe this represents a rational approach to our cataloging problems. Where reference works exist, it seems pointless to duplicate their information: where the cataloger can himself make a contribution, discovering something previously unknown, then he should by all means take the time to record it, providing always that the discovery he makes is really significant. A new cataloging manual has been written, the Senior Cataloger has himself processed 4,800 items from the sixteenth century, and two of his colleagues who joined the staff more recently have processed some 3,500 seventeenth century items.

Page  5 So much for statistics, which in themselves are tedious, I grant, but I mention them for a good reason. Our first duty is to the historian interested in the text. For this reason, while we have curtailed certain types of descriptive notes and bibliographic history in order to simplify and speed the program, in selected instances the catalogers are now analyzing composite works more extensively than before, in order to bring out texts which quite often are not otherwise available. Few of us with the collecting instinct are content simply to make known what already exists in our libraries. Stagnation is bound to result if we make no effort to improve our resources for scholars. Acquisition therefore counts for much of our staff time, as it doubtless does for yours. With our modest budget and existing strength, we are able to add only a few hundred early imprints a year. Scholarship needs more than early imprints, however, and our present planning is being directed primarily toward other areas: Early manuscripts on microfilm, modern manuscripts of recent and contemporary scientists, and what has come to be called in America "oral history." Our History of Medicine Division once only considered materials prior to 1801. Its cataloging and curatorial responsibilities for printed books in general still do. But I am glad to say that it now considers the 19th and even the 20th century--which is nearly two-thirds gone--valid subjects for historical study. [There are a million items in this portion of our collection.] To me personally this is the most interesting part of medical history, and I look

Page  6 forward with some excitement to developments in this area. Traditionally, medical historians, except those working in the medieval period, have relied very heavily on printed works. This is, I think, closely related to the attitude that grants priority in scientific discovery to the first published paper, and clearly it is the published papers that establish and represent the public body of scientific and medical knowledge. Much remains to be done in the history of medicine in the use of printed material, of course. I think not merely of the great monuments of medicine which are the pride of a "rare book" collection, but even more of the many lesser works-or even of the lowly thesis--that throw light on the characteristic medicine of a time. To my mind, such large collections as NLM's 17,000 theses of the 18th century and earlier have never been properly exploited for the light that they might throw on medical education in those years. There is much to be learned in medical history, however, that cannot be found in print. We all know with what avidity every scrap of paper bearing on the career of Vesalius or Harvey is sought after and studied. I am not suggesting that in this day of massive documentation all scraps even of our most distinguished scientists will ever take on comparable value. But no one can hope to understand the growth of a scientist's thought, the inter-relations of members of the scientific community, or the development of policies governing medical institutions in our present age without unpublished records.

Page  7 Therefore we have begun to seek out and preserve the correspondence, notebooks, diaries, and other unpublished materials of distinguished and representative biomedical scientists and physicians for our own library, and to help stimulate a more widespread interest elsewhere. We have made a modest beginning in this field and are starting to organize for public use materials already acquired. Since our primary goal is preservation of the material in the most appropriate repository, rather than just in our own, Dr. Blake the Chief of our History of Medicine Division has participated in such groups as the Conference on Science Manuscripts, and has urged the importance of preserving medical records to archivists and state and local historians. The fact that we may be interested in the papers of some professor may often, I have no doubt, induce the authorities of his own institution to look with a more interested view on such materials themselves. A few of our medical institutions do have active and alert archival programs. If more do so, it will be all to the common good. In this day of the telephone, airplane, and roundtable conferences, however, many of the most important communications between individuals may never be recorded on paper. Of course, scholars and scientists, no less than politicians, have always discussed their problems face to face without recorders, and lack of documentation for significant events is not simply a phenomenon of present-day technology. Indeed, it is quite likely that our own age is far better documented than any previous one. An inordinate amount of material, however, seems to be

Page  8 documentation for the public record, or documentation for obfuscation. We may need approaches other than those of the traditional historian to the mass of paper, especially in government and institutional archives, to tell us--or to tell the future historian for whom we are now preserving this material--what is or what was to the contemporary eye significant. We also may need other approaches to the time to tell us what was never put down on paper. This is the function of oral history. There are in the United States a number of centers actively engaged in the preparation of oral history records. Columbia University, where the program was originated by Professor Allan Nevins more than 15 years ago, has the longest record of experience in this field. The program has included some medical figures, but has concentrated on persons of political importance, to the neglect of many other human activities. We believe, therefore, that the field of medicine is open for a genuine contribution. Oral history, as we look upon it, is more than the recording of casual conversations with the notable figures of medicine, so that future generations may be inspired by voices from the past. Modern technology-- [especially] the tape recorder--makes oral history possible, but the most important ultimate product is the transcript of what may be an extended series of interviews. The experience of a number of practitioners of the oral history art suggests that it is essential to have trained interviewers well versed in the subject who have carefully planned their line of questioning by preliminary research in the relevant documents. The result should be a thoughtful discussion of certain

Page  9 problems or areas in which the person interviewed has been active, or even in some cases an autobiographical memoir. Oral history will provide no magic answer to the historian's questions. The interview record is subject to sources of error just as other records are, and must be used by the historian just as critically as his other forms of evidence. But it will, we hope, add a new dimension of understanding to the record of history being preserved today for the benefit of historians tomorrow. Oral history depends on one device of modern technology: the tape recorder. The National Library of Medicine is probably better known, however, for its applications of another product of the electronic age, the modern computer. I am sure that those of you who have worked on the preparation of Current Work in the History of Medicine appreciate the amount of labor required to keep abreast of current literature in the history of medicine. The National Library of Medicine too has been aware of this problem and decided sometime ago to study how it might help to meet this need without duplicating existing efforts. Quite clearly the one thing that was not needed was a new "current awareness" tool, since you have been taking care of this so well. For this reason, we offered to help your effort and I trust that the rough monthly printout from MEDLARS that we have started sending to you is proving useful. Current Work does become a bit tedious to use for retrospective searching, now that it has reached over 40 issues. You have solved this problem very handily here, of course, with your cumulative card index,

Page  10 of which our own staff, I might say, is exceedingly jealous. It has seemed to us, therefore, that the most useful next step in medical-historical bibliography, and one that is particularly appropriate for machine processing, is the preparation of cumulative bibliographies. As a first step in this direction, we made a study of the historical citations stored in MEDLARS, and found that of the approximately 2500 medical journals regularly indexed for Index Medicus, over 900 contained historical articles in one or more issues, for a total of some 2000 articles per year. We recognize that there are in addition about 100 medical-historical journals not included in Index Medicus, some thousands of medical journals not in the MEDLARS store which may contain occasional historical articles, and, beyond that, there are still several thousand more non-medical journals ranging from Archeology and Classical studies to local history, which are not indexed in Index Medicus. All told, we expect the total to run, after eliminating trivia, to at least 2,500 citations per year. Sorting, duplicating, and cumulating this mass of citations, and providing all the biographical, geographical, and chronological as well as topical approaches demanded by historians seems to us a task admirably suited for machine manipulation, [and we are beginning to use our computer for this purpose. We are most hopeful of cooperation between your library and ours in what I should like to see advance as a joint effort. We would like to do more, however, than simply make the material available. For one thing we share with you, I am sure, a conviction that the study of medical history has something of value to offer to

Page  11 contemporary medicine--not to be sure in the way of new drugs to be discovered in old pharmacopoeias, but for perspective, insight, and perhaps even a little more humility than is ordinarily met with in an age of alleged "miracle" drugs. Besides, it is pointless to spend a great deal of money on the preservation and processing of such material, and then allow it to remain on the shelves unread. Ideally, one would have a large circle of visiting readers, but in fact, in Bethesda as in other historical medical collections, the number of steady visitors [to the history division] is small. Those who do visit personally are avid readers, and use the collections in depth, particularly the academicians who descend on us and enliven things in the summer. A wider audience is reached via interlibrary loan. But our figures for last year show that only 2.7% of the entire historical medical collections were actually utilized, one way or another. There are many reasons for this. First, those interested in the history of the health sciences will always be a select minority, and of the interested group many will find the language barrier a problem. I need not remind you of the high percentage of medical and scientific works published in Latin before 1801, and of the sad number of medical men [myself included] with "small Latin and less Greek." Second, even physicians with a classical background may well feel diffident about approaching some of this early material, particularly the manuscripts and books produced in the first 150 years of printing.

Page  12 Some acquaintance with paleography is essential for study of the earlier periods. With the increasing use of the vernacular language in scientific communication during the later centuries, both in original works and in translations, epitomes, and adaptations, there is ample scope [for study by] a physician or scientist with the basic minimum of French and German. Of the major scientific works, too, many have been fortunately translated into English at one period or another. What then should the Library do, to encourage and promote research in the history of medicine? First, through our Extramural Program, which in this respect works closely with the History of Medicine Division, we support, through grants, research in the history of medicine by University scholars. [Also,] we are providing financial support for the training of medical historians at two leading centers of instruction in the United States (Yale and Johns Hopkins). We are planning to offer special fellowships for research in our own Library. Second, I believe, the History of Medicine Division should set an example by participating directly in scholarly projects of one kind or another. We believe our Division members have a duty to study the material in their charge, and in time, to publish the results of their study. [This year] I have increased the size of the staff to permit more time for such efforts.

Page  13 Finally, I should say a few words about other changes taking place at NLM. We, [by harnessing the power of the computer for bibliographic services] have been making library historyover the past few years. Our MEDLARS program now has more than 300,000 citations stored on magnetic tape. Of particular significance is the practical experience gained in the first year of MEDLARS' operation. Despite imperfections in our subject heading list, improvements needed in indexing practices and techniques, and some retrieval difficulties imposed by these deficiencies, the results [of reference retreival] have been most heartening. Experience gained from more than 2,3000 [computer] searches performed for physicians, scientists, teachers, librarians, and others, has established the practicability of MEDLARS beyond question. Appreciating that there is as yet no satisfactory method of evaluating effectiveness of information storage and retrieval systems, the Library has relied heavily on consumer reaction and appraisal. Based on evaluation of critical reports to the Library, the percentage of missed entries is minimal; furthermore, the relevance ratio: as determined by the individual requesters' evaluation of demand bibliographies now ranges between 60 and 70 percent [and recall of documents in the system exceeds 90%. Studies of more refined evaluative techniques will be undertaken cooperatively with Cloverdon (is this correct?) of this country.]

Page  14 For the Library to index all of the world's currently published substantive biomedical literature in the depth required by specialized information centers seems impracticable. The Library intends to avoid the waste of unnecessary duplication by supplying specialized information centers with relevant bibliographies on magnetic tapes, with the hope that these centers will then analyze and refine the material to meet their particular requirements. Already the Library and several universities have made cooperative arrangements to test this concept. Because MEDLARS was designed to serve three purposes, i.e., demand bibliographies, recurring bibliographies, and high speed printing, it is likely to prove more economical and more utilitarian than a system designed to serve a single function. Also, its location in a research library makes it possible to provide the original text or photocopies

Page  15 of material required by users and to supplement the computer system with traditional reference services. Research and Development The Library realizes that MEDLARS is only an initial response to the need for improving biomedical communication and modernizing library techniques. It is in this spirit that we plan to carry out the following activities. First, in accordance with its legal responsibility for wide dissemination of scientific and technical information in medicine, the Library initiated in 1964 a three-phase program to decentralize MEDLARS. This program calls for production of compatible, duplicate tapes to be made available to medical schools, research institutions, governmental organizations, and industry for use in their own computers. These groups would then have the same search and retrieval capabilities as MEDLARS. In the first phase of decentralization, the Library has started two pilot studies: one at the University of California in Los Angeles, using computer equipment which is not compatible with the Library's, and the second at the University of Colorado using compatible equipment. The University of California is reprogramming the Library's Honeywell tapes for use on IBM computers. [These tests will be completed next month.] It will test the feasibility of supplying regional services and also by using the tapes to support the activities of a specialized brain research center. [The University of] Colorado [is using] MEDLARS tapes for experimental purposes such as selective dissemination of information. [Our experience here indicates that one may search a months citations in a given field at a cost of 10 dollars per search - using less than two minutes (?) of computer time.]

Page  16 The second phase will be the formation of a network of additional search centers in areas with large concentrations of scientific manpower. Interest in this part of the program is so high that requests for sharing in the MEDLARS' searching capability have already been received from more than 40 university medical centers, private corporations, and government agencies. To aid it in its selection of additional search centers, the Library has developed criteria based on the service potential of the institution; its computer resources, and its interest in conducting further research and development based on use of MEDLARS tapes. The third phase of the decentralization program will be the provision of both data and program tapes at cost to interested institutions within the United States and abroad, and experimentation with the linkage of established centers in the United States through use of data transmission equipment. As many of you know, one such [Medlars] center is now being organized here under the auspices of the Ministry of Education and Science. A second activity planned at the library is the automation of our aquisitions and cataloging systems. These library operations are obvious candidates for automation. The Library has performed a systems analysis of these functions and has designed a system to improve these capacities through use of available computer equipment. Computer programming is now under way to implement the first step of this new system. Mechanization of the cataloging operation will enable the Library to institute a rapid cataloging service that

Page  17 other libraries can use for their own acquisitions and cataloging activities. The System will produce our printed book catalog and a five-year cumulation. It will also enter citations to selected monographs into MEDLARS. [This system will be operational Jan 1, 1966.] Finally, plans are now under way for the development of a [new] graphic image storage and retrieval system that will permit rapid photocopy retrieval of the full text of documents in the Library's collection. What the Library envisions is a system that will tie the bibliographic capabilities of MEDLARS to the graphic capabilities of this new photocopy system. [This work is being ___ out by NLM staff with assistance from our National Bureau of Standards.] By its development of improved plans for dissemination of information, the National Library of Medicine is fostering a greater awareness and a better understanding of research and development efforts in behalf of public health and clinical medicine, and a more rapid translation of research into clinical application. The Library's transformation from a passive repository of information to an active ally of the researcher, teacher, and clinician has stimulated increased use of [our] medical library facilit[y] and has converted other medical libraries to dynamic energizers of thought through better communication of public information. [Since 1962 the use of our library has increased at a rate of 20% per year. This last year we provided 150,000 inter-library loans, 55,000 ___ service and 30,000 reference services in addition to the published bibliographic services such as I. M.] [Mention ____ Library Construction Act of 1965 5 year program (125 million dollars) for library construction, ___, research, ____ + _____ + development of reference medical library networks.] While the Library works for the future, it has not forgotten its historic heritage. The National Library of Medicine dates its founding from 1836, but it really began to grow just a hundred years ago, in 1865, when John Shaw Billings was placed in charge of what was then the Library of the Surgeon General's Office, U. S. Army.