1927 reference in the new millennium: where is the Automat?

Abstract

James Ballard, director at the Boston Medical Library, tracked questions he received at the reference desk in 1927 to recognize the trend of queries and to record the information for future use. He presented a paper on reference services that listed sixty of his reference questions at the Thirtieth Annual Meeting of the Medical Library Association (MLA) in 1927. During a two-month period in 2001, the authors examined Ballard's questions by attempting to answer them with print sources from the 1920s and with the Internet. The searchers answered 85% of the questions with the Internet and 80% with 1920s reference sources. The authors compared Internet and 1920s print resources for practical use. When answering the questions with 1920s resources, the searchers rediscovered a time in health sciences libraries when there was no Ulrich's Periodicals Directory, no standardized subject headings, and no comprehensive listings of available books. Yet, the authors found many of the 1920s reference materials to be quite useful and often multifunctional. The authors recorded observations regarding the impact of automation on answering reference questions. Even though the Internet has changed the outward appearance of reference services, many things remain the same.

BACKGROUND

In the early part of the last century, the Automat was a popular eatery without wait staff. Diners served themselves by inserting nickels into machines after examining pies and savory dishes in lighted cases. The Automat, America's first fast-food restaurant, brought high-tech, inexpensive dining to a low-tech era [1]. A common library textbook published in 1927 suggested that someday an automat might develop in libraries to deliver information, thus reducing the need for the human component in reference services. James Wyer, director of the New York State Library School at Albany, compared possible future library automation with the Automat. He stated that the organization of a million books could someday be possible so that a “desired volume, like a bit of chocolate or gum, shall drop out of a chute after the proper manipulation of buttons, levers, slots and keys” [2]. Wyer reflected that reference work existed, because it was not yet possible to organize materials mechanically well enough to dispense them without personal service.

Today, the Internet, an automat-like system, is highly integrated into reference services. An automated system such as the Internet probably would have impressed Wyer and his colleagues, at least initially. Library patrons now typically have direct access to many electronic resources and often perform their own searches. A patron touching a few buttons or keys is now often instantly overloaded with information. Yet, the automated library of today is not self-service, like Wyer speculated. Librarians still actively serve and train patrons in libraries now heavily populated with computers.

Librarians provided personalized, yet manual, reference services in the 1920s. James Ballard, who from the age of fourteen served for sixty-three years at the Boston Medical Library [3], claimed that “Hour after hour, day after day, the telephone, the mail, the visitor in person, bring a constant flow of inquiries of all kinds, no two alike, simple and difficult, ordinary and bizarre” to the reference desk [4]. Ballard suggested keeping the reference area in a separate alcove, so reference services would not interrupt students. In this room, he kept 3,500 reference books. A second set of books, including the American Medical Directory, was stored in the cataloging room where the central telephone receiving station was located. Librarians quickly answered ready reference questions there without burdening the reference desk. Librarians answered price and purchasing questions with collection development resources kept in another office [5].

Ballard encouraged tracking reference questions over a period of months or years to evaluate the trends of queries and to record the information for future use [6]. Several period library textbooks recommended recording commonly asked questions on index cards for repeat reference [7, 8]. In his description of reference services, Ballard mentioned sixty of the reference questions that he received at the Boston Medical Library in 1927. He divided his questions into two categories: “Information” and “Reference and Research” (Table 1 and Table 2). Ballard did not provide answers to any of the questions.

Table 1 1927 “Information” questions

Table 2 1927 “refernce and research” questions

The first category of questions, which Ballard termed “Information,” consisted of questions that could be answered quickly by consulting specific reference sources (Table 1) [9]. Examples of queries included:

Requests for addresses and full names of physicians (local, national, and European), information concerning hospitals, addresses of drug houses, the addresses and prices of medical periodicals, the publishers and prices of medical books, lists of local specialists or of those in other states and foreign countries and of special hospitals. [10]

The most frequently asked question at the Boston Medical Library was, “Does Dr. —— belong to the Massachusetts Medical Society” [11]?

“Reference and Research” was the other category that Ballard used to classify patrons' questions (Table 2). Ballard defined “Reference” as queries that did not require extensive research. A specific book or article usually satisfied these requests [12]. Some of the questions in this category included Prohibition in relation to medicine (R-13) and the medicinal value of the vanilla bean (R-15). Ballard declared that research was closely allied with reference but involved more time. He warned that any reference question could quickly develop into a research question [13].

Because of the time involved in conducting library research in the 1920s, larger health sciences libraries often had separate fee-based bibliographic service departments. These departments created bibliographies, verified citations, and provided translation services. Metta Loomis, librarian at the University of Illinois, estimated that a complete ten-year bibliography on a single topic could cost $250 of a librarian's time in 1926. Most librarians would have to “slight many in order to give one two hundred and fifty dollars' worth of reference service” [14]. The Boston Medical Library planned to open this type of bibliographic service department in the fall of 1927 [15]. Janet Doe's new service at the New York Academy of Medicine in 1927 charged fifty cents per hour to fellows and $1 per hour to the public for bibliographic services [16]. In the first three months, this new service filled sixty-eight requests including bibliographies on topics such as color schemes for hospital rooms and the effects of telephones on hearing [17].

The outward appearance of reference services has changed during the past seventy-five years. New reference innovations have been developed. Old familiar print resources are now often available on the Internet. However, with the proliferation of Internet resources, it is unclear if answers to reference questions are easier to access. Ballard's questions were examined by answering them with 1920s print sources and with the Internet. This study contrasts Internet resources with historical print sources and compares their practical use. The study chronicles some of the developments in health sciences reference services before automation.

METHODS

At the University of Minnesota Bio-Medical Library, the authors examined Ballard's questions to explore how automation affected modern reference services and to develop an historical perspective. They attempted to answer Ballard's questions with 1927 print sources and with the Internet during a two-month period in the summer of 2001. A librarian and a reference assistant numbered and evenly divided up Ballard's sixty reference questions. The authors attempted to answer each question within a reasonable amount of time, although official limits were not established.

The authors developed criteria for answering questions using 1927 print sources and the Internet. The searchers first answered questions from print sources published before 1928. The University Libraries' electronic catalog (MNCAT), helped identify 1920s resources, because no card catalog remained [18]. If a title existed before 1928, but only a later edition was available, the resource was allowed provided the edition's publication date was no later than the 1930s. The searchers also answered Ballard's questions using the Internet. Internet resources included online indexes, directories, and services purchased by the University of Minnesota Libraries as well as any other Internet resource available to the public. The searchers treated the 1927 questions as current information requests when searching the Internet unless a question was date specific.

The searchers assembled the recommended books from the 1920s in an office at the Bio-Medical Library. Reference titles included those mentioned in Ballard's article [19]. Isadore Mudge's list and Evan Jay Crane's Guide to the Literature of Chemistry listed additional titles [20, 21]. The searchers consulted MNCAT and browsed the book stacks for other possible titles. They also found titles in the Owen H. Wangensteen Historical Library of Biology and Medicine and from storage at the Minnesota Library Access Center. Table 3 lists some of the health sciences reference books from the 1920s recommended by period books and journal articles.

Table 3 Selected reference works from the 1920s

The searchers informally recorded their answers and comments as they worked through Ballard's questions, using Blogger, a push-button Web-publishing utility [22]. By using Blogger, the searchers instantly recorded their thoughts and answers as a single, dated log on the Internet. The authors noted observations about which type of reference questions tended to be easy and difficult to answer using the Internet. They recorded similarities and differences about reference materials available in 1927 and in 2001.

RESULTS

The searchers successfully answered most of Ballard's sixty questions with 1920s reference sources and with the Internet (Table 4). Approximately 90% of Ballard's “Reference and Research” questions were answered with the Internet and with 1927 print sources. Three of the four “Reference and Research” questions that remained unanswered by the Internet were not answered with 1927 print sources either. Slight differences emerged in the searchers' ability to answer Ballard's “Information” questions. While the authors answered 81.5% of the questions with the Internet, they answered 66.7% of the questions with period reference sources. Overall, the searchers found answers for 85% of the questions on the Internet and 80% with 1920s reference materials. Table 1 and Table 2 summarize which questions were successfully answered.

Table 4 Percentage of 1927 questions answered with 1920s print resources and the Internet

DISCUSSION

Availability of materials.

Reference sources from the 1920s were not readily available to the searchers, nor were they necessarily familiar with them. The authors often had to identify and retrieve sources from storage or view them in special collections. On several occasions, only scattered editions of standard reference books remained from the 1920s at the university. For example, no editions of Stedman's Practical Medical Dictionary between the years of 1918 and 1930 survived. Other resources were ephemeral and were lost altogether. For example, no vertical files remained. Library of Congress Cards were also not available for use in this study.

Reference interview.

Many of Ballard's questions were vague and could have benefited from a live reference interview. The intent of the questions was often unclear, leaving the meaning open to interpretation. For example, the authors first considered sex and common sense by a woman (I-19) to mean how common sense related to gender and later to mean how the act of sex related to a woman's common sense. This query turned out to be an exact title of a 1922 book, Sex and Common Sense, which was authored by a woman [23].

Modern electronic reference via email, chat, and the Web lends itself to less personal interaction with patrons, leaving many queries subject to interpretation. Electronic reference requests are often vague, but people in person also often ask ambiguous questions. Loomis relayed this same experience in her paper about reference services in 1926. She remarked that the public's requests were often so misleading that it required “all of the librarian's intuition and experience to interpret” [24]. She claimed that efficient reference work required not only service but also “psychology” [25]. Reference services via semi-automated systems as well as in person often require follow-up questions for clearer interpretation.

Reference skills.

Doe mentioned in an interview with Estelle Brodman in 1977 that the modern day librarian required a “vastly different body of knowledge and collection of abilities” [26]. Doe stated that, after the automation age came, physicians had more information available directly to them. Consequently, she stated that librarians were now responsible for learning how to use library equipment and for teaching their patrons about them [27]. She stated that administration is “enormously magnified in the requirements of the librarian today” [28]. In contrast, she said, in the early days, “one concentrated on books and indexes and the reader who came in and wanted help personally” [29].

Constructing bibliographies in the time before the information age required different search skills. Lucretia W. McClure, formerly of the University of Rochester, though not a librarian of the 1920s, witnessed the emergence of automation. She felt that younger librarians had lost the art of reading the literature. A manual search, she stated, required librarians to use their natural brain ability and skills to “ferret out the answers” [30]. She said that before automation:

You could not put it in a machine and get a printout; you had to do it yourself. This was the really grand experience of reference. [31]

Automation brought full-text, keyword searching to the reference desk. As a result, less in-depth analysis in conducting manual searches is typically required at reference desks today.

Reference librarians also needed foreign language skills to be proficient in their jobs at the beginning of the twentieth century. Before World War I, health sciences libraries relied heavily on foreign reference sources. The Army Medical Library† depended on military attaches in other countries to purchase foreign materials [32]. The 1920s brought an influx of new American bibliographic aids developed to replace foreign ones lost during the war [33]. Even so, several of the recommended reference books of the time, such as Landolt-Börnstein Physikalischchemische Tabellen or Handwörterbuch der Sozialen Hygiene, were in German or other foreign languages [34, 35]. Doe commented that a person could not “get anywhere in medical librarianship” without some knowledge of French and German and a “little smattering” of Spanish and Italian [36].

Characteristics of the 1927 health sciences literature.

The body of health sciences literature was comparatively small in 1927 and began to grow significantly in the 1940s. A study appearing in the New England Journal of Medicine tracked the yearly weight of the Index Medicus over time. In 1927, the Index Medicus weighed approximately three kilograms, but by 1969, the weight of the Index Medicus had increased tenfold [37]. A follow-up study reported a slowing of the growth of the Index Medicus, but the weight still increased to thirty-seven kilograms in 1985 [38]. The size of the health sciences literature in 1927 was more manageable in a paper format than it is today.

It was less challenging to answer some of the questions with 1927 print sources than originally anticipated because of the smaller size of the health sciences literature. For example, one all-encompassing reference source, the Reference Handbook for the Medical Sciences, was extremely useful and easy to use. The series claimed to embrace “the entire range of scientific and practical medicine and allied science” [39]. The searchers easily answered several questions such as those about nail biting (R-30), volume of blood in the body (R-25), fear of death (R-6), and instructions for Caesarean section (R-5). In contrast, eight easy-to-use volumes could not adequately begin to cover the basics of today's health sciences literature.

Many current standard reference books existed in 1927, but their scope was broader and some were multifunctional. The Quarterly Cumulative Index Medicus listed new books and publisher information in addition to journal articles. The American Medical Directory, now published as the Directory of Physicians in the United States, was a single volume in 1927 but served as much more than a physician directory. This book also functioned as a directory of serials, providing information about circulation and costs, and contained addresses of libraries and medical schools [40]. The American Medical Directory answered several of Ballard's questions, ranging from the name of the librarian at the Army Medical Library (I-2), to the address of Lea and Febiger (I-8), to the most widely read medical journal (I-24). The searchers consulted this resource for several other “Information” questions (I-4, I-7, I-10, I-21, I-25).

Subject searches.

Subject questions were relatively easy to answer whether searched for on the Internet or in the 1927 Index Medicus, even though the searchers lacked extensive experience using the print resource. Many of the subject questions—ranging from those about left-handedness (R-16), to community nursing (R-3), to the diuretic properties of ammonium chloride and calcium chloride (R-18)—were easily answered via MEDLINE as well as the 1927 Index Medicus. C. C. McCulloch, Jr., librarian at the Surgeon General's Library‡, praised the Index Medicus in 1916 by stating that

With comparatively little trouble and with the expenditure of a few minutes time we have limited our search through the world's medical literature to a couple of hundred titles which we can look over and investigate at our leisure. [41]

The searchers also found the 1927 Index Medicus practical and easy to use, even though they were accustomed to searching MEDLINE electronically.

In 1927, the Index Medicus did not contain a fully developed standardized list of subject headings. Ballard stated that there was an acute need for a standardized list of subject headings for health sciences librarians at that time [42]. The first set of standardized subject headings intended primarily for the indexing staff at the Army Medical Library appeared in 1931. The American Medical Association printed 100 copies of this list, but the supply was quickly exhausted [43]. However, the 1927 Index Medicus contained several nonstandardized subject headings, which the searchers found extremely useful. For example, “Prohibition” listed several journal articles regarding Prohibition in relation to medicine (R-13).

“Reference and Research” questions that were difficult to answer with 1927 reference sources were often challenging to answer using the Internet. Questions involving historical change, such as when Osler changed his statement regarding pneumonia as a self-limited disease (R-22), were not found without expending excessive time. The searchers never found the reference suspected to be in The Lancet about the woman who lost her lover when she was twenty-five years old, developed an idea that he would return, and retained her youth and beauty for forty years (R-28). In print, this would have required long hours of thumbing through volumes of The Lancet. The Internet might have been useful for this question, but was of no help because this question was so dated and did not contain helpful key words.

Serials.

Locating and identifying serials was a challenge for librarians in the early part of the century, because there was not a comprehensive union list of serials. It was nearly impossible to know what journals other libraries had in their collections. M. Myrtle Tye, of Emory University, introduced and encouraged discussion of the need for a union list of medical serials at MLA's 1925 Annual Meeting [44]. The long-awaited first edition of the Union List of Serials in Libraries of the United States and Canada finally appeared in 1927 as a single volume. It contained 75,000 entries from 225 different libraries [45]. Loomis predicted that the impact of the publication of the Union List of Serials would be “beyond prophecy” [46]. Another author described this work as “epoch-making” [47].

Ballard's questions about classified serials were difficult to answer with 1927 print resources but were easily answered with the Internet. The searchers answered nearly all serials questions electronically by using ulrichsweb.com or Publist.com [48, 49]. Creating a list of medico-legal (R-19) or bio-chemical (R-20) periodicals was difficult with 1927 print sources. Both the 1927 Index Medicus and the American Medical Directory contained lists of serials and publishers, but neither classified them by subject. A work called Classification of Periodicals: Army Medical Library was finally identified [50]. This list, completed in 1927, was a simple list of classified serials with no introduction or explanation. It is unknown if this resource was widely distributed in the 1920s, because only two libraries report owning it today [51].

The first edition of Ulrich's Periodicals Directory, which classifies periodicals by subject, was not published until 1932. At that time, the Periodicals Directory covered popular culture topics such as “leather,” “radio,” “beauty culture,” and “canning and preserving” but did not cover the disciplines of medicine, law, or chemistry. The Periodicals Directory did cover the discipline of pharmacy under the heading of “drug-trade” and public health under “hygiene & public health” [52]. This book referred readers to other unclassified sources in the bibliography section at the end of the single, thin volume.

Books.

While the searchers easily answered most questions about books using the Internet, they experienced problems locating book information with 1927 print sources. By using OCLC WorldCat, they found it easy to locate almost any book whether or not the University of Minnesota owned it [53]. For example, it was easy to locate the book Sex and Common Sense (I-19), because the university owned two copies of it, and the searchers assumed it would have been included in the 1927 card catalog. However, the university did not own the book by L. Hill (I-5), Sunshine and Open Air or the Diary of a Physician (I-3), and finding information about them was challenging using 1927 print sources. The Index Catalogue did not list these titles, and their year of publication was not revealed in Ballard's questions.

The Army Medical Library card sets were not readily available to health sciences libraries around the country, until the information appeared in the Index Catalogue. Ballard wrote, “If the Army Medical Library could type and sell duplicates of its cards as they are reported it would be a great service to the medical libraries of the country” [54]. The Index Catalogue was produced directly from these cards, but there was up to a fifteen-year time lag depending on what volume was being produced in a particular year [55]. The third series of the Index Catalogue began in 1918 and the fourth series in 1936. This publication delay contributed to the difficulty in locating information about health sciences books with 1927 print sources.

Several types of cataloging cards were in print in the 1920s to help librarians locate book titles outside of their own collections. The Library of Congress printed complete sets of cataloging cards, and some libraries served as depositories [56]. In 1916, the Library of Congress actively printed cards for the Department of Agriculture and the Bureau of Education but not for the Surgeon General's Office [57]. The U.S. Public Health Service and the U.S. Hygienic Laboratory printed their own cards independently [58]. The publisher, Wiley & Sons, started issuing cards for new technical books instead of distributing flyers and leaflets [59]. Some of these card sets remain, but none were available to the searchers in this study.

The Internet greatly enhanced searching for book values. The searchers had no problem finding prices for used books on the Internet using sites such as abebooks.com, which lists independent bookseller's rare books [60]. To find book values in 1927, the searchers pored through old volumes of American Book Prices Current and Book Auction Records and were only able to find a price for the value of Biographica Medica, but not for Medical Botany (I-18) [61, 62]. Finding a newer book of that time, such as Canano, 1925 (1–27), was difficult, because publishers were slow to provide book information to Publishers Weekly and the Cumulative Book Index [63, 64]. The Index Medicus inadequately covered new books [65].

Vertical or card files.

Reference librarians in the 1920s often maintained vertical files of pamphlets, clippings, and miscellany. This practice helped them track social trends and current happenings. Mary Hazeltine, preceptor of the University of the Wisconsin Library School, emphasized the importance of using a vertical file for current awareness in 1922. She claimed that “The up-to-date librarian will study the community, its business, social and other organizations and will keep in touch with all movements, both local and world wide” [66]. Wyer's textbook described such a file as a favorite container for “an increasingly prolific yet indispensable flood of lowly, heterogeneous, ephemeral print” including “rivulets” of pamphlets, clippings, maps, pictures, bibliographies, and other unbound items [67].

Librarians have developed new ways of tracking ephemeral tidbits of information in the information age. The Internet, perhaps, is the current replacement of the once prominent vertical and card files. For example, the Bio-Medical Library tracks health-related newspaper stories, linking them to journal citations on the Internet via Health and Medicine in the News instead of filing newspaper clippings [68]. The library also keeps a list of some of the difficult reference questions and answers on the Internet, where it is accessible to anybody at the reference desk. Coincidently, the question about the number of quarts of blood in the human body (R-25) appears in Ballard's list as well as among the questions at the Bio-Medical Library. Contemporary librarians use the Internet much as the librarians of 1927 used a vertical or card file to track ephemeral information.

Some popular culture issues remain of interest over time, making information easily accessible on the Internet and in 1927 print sources. For example, there still is much interest in anti-vivisection literature (R-9). The question regarding the use of disease germs as an offensive in the next war (R-32) is of unusually high concern at this time. The answers to these questions were easily answered both via the Internet and from 1920s print sources, probably indicating a consistent level of public interest over time.

The Internet, much like a vertical file, is an excellent source of information about current popular culture, trends, and fads. However, information on the Internet may be transitory. Some of Ballard's questions were so dated that they were difficult or impossible to answer by using any Internet resource. Finding a person using the MacDonald's cure for high blood pressure (I-20) certainly was not looking for somebody who frequents the Golden Arches. The searchers did not find the address of the Defensive Diet League (I-12) or the components of a Ham's plaster (R-31) on the Internet or in 1927 print sources. The Boston Medical Library possibly provided answers to these questions from a vertical or card file. The Internet is still a comparatively young medium. The future will reveal how much previously posted material is preserved to reveal information about the popular culture of this era.

CONCLUSION

Loomis mentioned that reference librarians sometimes sidetrack themselves on difficult, but interesting, questions and neglect other patrons needing assistance with more mundane matters. Loomis stated that many times:

Such a reference quest has all the fascination of a hunt and would be thoroughly enjoyed by almost any librarian, but ninety per cent would steal time from other customers and other work in order to indulge in such sport. [69]

However, William Warner Bishop, director of the University of the Michigan Department of Library Science, emphasized the importance of having a historical perspective of one's work. He wrote, “[W]ithout a sense of the historic setting of his work, a man is almost as hopeless as is the man who lacks a sense of humor” [70]. The authors gained a historical perspective of health sciences reference during this study but did not “steal time” from other patrons while exploring Ballard's questions.

Reference services have changed during the last seventy-five years. In the 1920s, print sources served reference librarians well because of the size of the health sciences literature. Reference services in the 1920s relied heavily on individual librarian's skills and intuition with well-used print resources. The Internet, an automat-like system, infiltrates reference services in the new millennium, a type of automation librarians only speculated about in the 1920s. The Internet has now brought many reference sources directly to patrons. Searching for book and journal information is simplified with the Internet. Internet-based indexes allow full text as well as subject heading searching. The Internet often substitutes for vertical and card files, once so prevalent in libraries.

Even with the changes in reference resources and services over the years, some things remain the same. Several library innovations that were in development in the 1920s, such as the Union List of Serials and the Periodicals Directory, are still available today in electronic or paper format. The Index Medicus, now typically accessed via the Internet, is still heavily relied upon to answer “Reference and Research” questions. Overall, the searchers in this study had similar success rates in answering Ballard's questions with 1927 print sources and with the Internet. Some questions are difficult to answer no matter in which era they are asked. In this study, most of Ballard's “Reference and Research” questions that remained unanswered with 1927 resources remained unanswered by the Internet.

The fabulous food-dispensing contraption, the Automat, as it turned out, was more a fantasy of automation than a reality. Behind the curtain, many men and women were hard at work restocking the chambers as soon as food was removed. Fast-food restaurants replaced the Automat where people, not machines, again provided service [71]. An automat for information, as speculated in 1927 by Wyer, has not replaced personal reference service. Humans still work on the reference desk and behind the scenes developing new electronic resources. Machines do not understand the psychology required in the reference interview. It remains imperative to keep librarians as intermediaries between a reader and the right resource. Wyer concluded that the utmost use of great libraries could never be attained by mechanics, for interpretation is a “far more delicate and difficult matter than ever can be achieved by mechanical marvels” [72].