College and Research Libraries • MATHILDE V. ROVELSTAD Open Shelves/Closed Shelves in Research Libraries This article traces the traditional method of making books accessible through their topical arrangement on library shelves, outlines the rea- sons for the abandonment of this method in European research li- braries, explores its modified reappearance in postwar academic li- braries, and cites modern arguments against the practice of making an entire research collection available on open shelves. AMERICAN LmRARIES OF ALL KINDS en- deavor to make books easily accessible, and as a result most libraries open their shelves to their readers. This practice does not prevail, however, in all li- braries throughout the world. Large aca- demic research libraries abroad usually have closed stacks, and scholars gain ac- cess to these collections through cata- logs, bibliographies, and indexes. The maintenance of open stacks with materials accessible through a classified arrangement is very costly. In view of modern research needs and budget pres- sures experienced by libraries, the high costs involved in this practice become more and more questionable and re- quire justification. To explore this is- sue, this paper will trace the tradition of open access; outline reasons for its abandonment in European research li- braries; demonstrate its reappearance in modified form since World War II, par- ticularly in German libraries; and cite modem arguments which challenge the . validity of classifying the holdings of a research library on the shelves for the benefit of those who wish to browse. If one defines a library as a collection Mathilde V. Rovelstad is professor, De- partment of Library Science, Catholic Uni- versity of America, Washington, D .C. organized for use, access to it is essen- tial. The question then becomes wheth- er this access is more successfully accom- plished through the use of traditional bibliographical sources or through the bibliothecal approach which allows browsing in the shelves. Most foreign academic and research libraries arrange their collections in a chronological se- quence according to the order of acqui- sition by the library. Browsing in such a situation is meaningless since books of different subject fields are placed next to each other; and without special nota- tional provisions individual volumes of a work, or monographs in a series, are not kept together. This practice is di- rectly opposed to that used in American libraries where books are classified and available on open shelves. THE TRADITION OF OPEN SHELVES Historically, open access to shelves·, where books are placed together by sub- ject, is the oldest and simplest way of providing for the use of a collection. It is a method successfully demonstrat- ed in most private libraries, and is a very economical retrieval device for any small collection. Even though little is known about early practices in library organization, it is obvious that the rna- /457 458 I College & Research Libraries • September 1976 terial assembled in these libraries must have been arranged in some orderly fashion if it were to be retrieved easily. In -fact, catalogs that have survived from antiquity indicate that some . rough subject order already had been maintained for the . tablets and rolls in earthen jars, wooden chests, wall niches, ·and later in the armaria of Roman times. 1 ·The monastery libraries which were dominant during the Middle Ages had very small collections compared with those of antiquity. They served commu- nities of learned monks and lay persons with scholarly interests and religious backgrounds. For ease of use, the ver- nacular books, the "lay library," were usually separated from those used for study. Since they generally had to be used on location, a rough subject order was the typical form of organization. 2 For a long time the arrangement itself had to serve as a finding device since there were no catalogs. This was possi- ble bec~use the collections were very· small; . even medium-size libraries did not hav.e more than a few hundred. codices. In larger collections there were headings to guide the user to the loca- tion of the literature. Pictures of au- thors on cupboards or walls would sug- gest the presence of related books, such as in the library of Isidore of Seville ( 56~636) , or the furniture units them- selves were designated, such as an arma- ria imperiali in the library of Charle- magne (742-814). The monasteries of St. Gall and the Reichenau provided captions for sub-units.3 . A unique medieval custom required a periodic inspection of library volumes and was still mentioned in Richard de Bury's Philobiblon.4 This practice was probably the reason why lists of the items , contained in a furniture unit were compiled. The absence as well as the presence ()f a work could thus be checked. With groWing collections these lists also facilitated the ~etrieval of the works in the library. Actually, they were early catalogs with a double function: as shelflists they provided inventory control and indicated location; and be- cause related items were kept together, they also served as subject catalogs. It is evident that ease of access to the assem- bled books was of concern since a typi- ·cal list displayed the titles in homoge- . neous groupings. For example, it would begin with the Bible, then list the Church Fathers, theologians, and an- tique authors, and end with the artes liberales.5 After the twelfth century the period of the great monastery libraries came to an end, and universities emerged as the carriers of scholarly teaching and learning. Since their book collections served a more varied group than those of the monasteries, they included the profane disciplines taught at the col- leges. The library at the college en- dowed by Robert de Sorbonne ( 1250) in Paris set the pattern for others. As in the monastery collections, books were arranged to suit the needs of their users. They were grouped according to the four faculties of the college: theology, medicine, law, and philosophy. Within each group there was a rough arrange- ment according to the first letter of the author's name. But even this order was not strictly maintained since works were comparatively easy to find in the small collections. In 1289 the Sorbonne library was divided into a noncirculating refer- ence library (libra ria magna) with 330 volumes, of which the heavily used books were chained to twenty-six desks, and a small circulating library (libraria parva) with 1,290 books consisting of duplicates and less valuable items.6 The Renaissance and Baroque age could boast beautiful -library rooms- architectural gems-in royal and prince- ly residences. Books were stored in al- coves on series of tiers surrounding 4 l • great halls. The owners generally al- lowed a congenial public the use of their collections in these halls, but with the spreading of education there were already demands from a larger public for access to the knowledge stored in li- braries. Gabriel N au de ( 1600-1653), in his famous treatise on library practice, suggested that libraries should be open for public use and available to .. the humblest of those who may reap any benefit thereby."7 The circle of people admitted became larger when some li- braries opened their doors to the "pub- lic," which meant at that time learned individuals who were not members of the immediate academic community. The Ambrosiana in Milan ( 1603) was the first large library adopting this lib- eral policy, and it was soon imitated by others, such as the Bodleian at Oxford ( 1612) and the Mazarine in Paris ( 1643). Cardinal Richelieu ( 1585-1642) wanted the Sorbonne also to be a .. pub- lic" library where scholars would be ad- mitted at certain hours daily, as well as ccmessieurs les curieux et les etrangers," that is, outsiders not known to the aca- demic community.8 This public still rep- resented a very small and select group of individuals and certainly did not in- clude the average man on the street. As books in these libraries did not circulate, they had to be used in the li- brary rooms themselves. Catalogs, when they existed, were still unsophisticated; consequently, it was the orderly arrange- ment that guided the reader to the lit- erature. At the end of the seventeenth century, detailed classified arrange- ments, already popular in the smaller libraries, spread to the larger ones. With the eighteenth century they became more systematic, a reasonable corollary to the Age of the Enlightenment's pref- ·erence for encyclopedic thought. A logical and systematic order was consid- ered an indispensable aid for the schol- ar to guide him to what he needed. Open Shelves/Closed Shelves I 459 There were sizable collections, such as the library of the Elector of Saxony in Dresden, whose 17 4,000 volumes until 1796 were accessible solely because of their detailed systematic arrangement. 9 Many theoretical treatises su.pported such detailed orders. Particularly well known was Konrad Gesner's Pandectae sive Partitiones universales (Zurich, 1548), a subject guide to his Bibliotheca universalis (Zurich, 1545). It followed the existing university faculties and be- came a model for the arrangement of books on the shelves. to ABANDONMENT OF OPEN SHELVES Toward the end of the eighteenth century two practices emerged which created barriers in the traditional reader/book relationship and whiCh greatly influenced modem library prac- tice and theory. One was the fact that until the middle of the nineteenth cen- tury, modem conveniences, such as heat and lighting, were slow in finding their way into libraries. To make studying more comfortable, small rooms were made available for reading purposes in the larger libraries. These rooms could be easily heated and, furthermore, al- lowed observation of readers.ll Now the user was no longer surrounded by the collection; the needed volumes had to be delivered to the reading room for use there. It was only a small step to close off the book rooms themselves and to ad- mit the users to the reading rooms only. The other factor contributing to the closing of stack areas was the steady growth in the amount of scholarly lit- erature. The most efficient way to shelve incoming material seemed to be in large multitier stacks with parallel ranges close together to increase the storage capacity. Thus stacks were created as separate areas from the rooms where the books were used. Increasingly, the reading rooms became the only areas open to readers, while the stacks were 460 I College & Research ·Libraries • September 1976 closed or, in rare circumstances, open only to privileged users. Closing of stack areas also helped to prevent losses. Other developments during the nine- teenth century also had an influence in this change. Many new universities had been established in Western Europe dur- ing the .second half of the century. New disciplines were created in. which · re- search was encouraged. Academic com- munities more than doubled in some .co~ntries and depended on the univer- sity libraries . to support their scholarly pursuits. However; as library appropria- tions did not increase proportionately, large libraries found themselves tied to administrative practices which were in- adeq:uate to cope with the many new books to be processed and with greater demands for services. A bottleneck was the. shelflist, which had to serve both for location control and as the library's subjec::t catalog. The addition of new volumes necessitated .a constant shifting of the .books on the shelves, changes of alreaqy cumbersome notations, and ad- jusvnents in the catalog. As a conse- quence of such complicated operations, backlogs of unprocessed· books devel- oped .. The . only . solution to this problem seemed to be the abandonment of the prevailing subject arrangement of the literature on the shelves and the employ- ment of the most economical storage possible. Thus the systematic library catalog could be relieved of its shelflist function, and new accessions could be integrated much more speedily. The re- sults of this decision were far-reaching. The abandonment of any systematic or- der on the shelves made it unprofitable for a reader to go to the .stacks directly, since similar material would no longer be together. The closing of ·shelves to readers was a logical consequence of a nonsystematic arrangement. Librarians ever since have rationalized this deci- sion. The new method, born of necessity, employed a shelf-finding device based on sequential numbering. This method was not actually a new one, as evidenced in the catalogs of Durham· (twelfth century), or Canterbury (around 1300), where new acquisitions were added in numerical order after an initial subject grouping.12 A famous example of this arrangement, heavily criticized by Naude, was the arrangement of the col- -lection of the Ambrosiana in Milan ( 1602).18 The nineteenth century pacesetter for shelving order by accession number was the Bibliotheque Nationale in Paris, and from there it spread rapidly to other western European countries. As the Bibliotheque N ationale had been un- able to process the overwhelming flood of incoming works, a decision was made to close the old stacks where books had been grouped by subject and to start a "classement mechanique," which, due to official instructimis, was now de rigueur for the large libraries under the central control of the French government.14 When Leopold Delisle, who was then al- so director of the Bibliotheque Nation- ale, suggested in his influential hand- book on library practices a numerical arrangement of new works, the French municipal libraries voluntarily followed these recommendations.15 Strict numeri- cal order was found to be particularly economical when the material was first grouped according to size. As recently as the 1937 edition of Crozet's standard handbook, grouping into nine sizes for scholarly libraries and into six for pub- lic libraries was proposed.16 The French pattern of strict numeri- cal arrangement within a size group did not find general acceptance in German academic libraries even though they had experienced the same financial problems and were faced with the same backlogs. There the emerging · library profession had developed the "subject specialist- <(• t • • librarian," ·who was responsible for the building and organization of the collec- tion in . his field of specialization and in related subject areas. A strict numerical order would scatter the works in his field, separate him from the collection, and prevent him from evaluating its strength and weaknesses. Since collection building was the librarian's prime re- sponsibility, such an arrangement was not considered suitable.17 Instead, the arrangement adopted by German ·li- braries was a modification of the French practice, in which the material was first divided into subject' groups and then subdivided according to book size and accession number. Numerical arrange- ment in these two forms spread rapidly among the scholarly libraries in Eu- rope.18 · As a result, with the exception of reference collections in reading rooms, : scholarly libraries had closed stacks. A major debate on the advantages of numerical vs. systematic arrangement and the issues closely connected with it -open shelves · ·vs. closed shelves-then took place. Administrative and econom- ic considerations had the upper ·hand, and Georg Leyh, the foremost defender of numerical arrangement, pronounced the "dogma of systematic arrangement" as anathema, and created a new dogma of the numerus currens. 19 According to Leyh, Fritz Milkau, then director of the Pnissian State Library, had reasoned that ·· 90 percent of all reader requests were for material of the last decade. Leyh concluded; therefore, that an ar- tahgement according to acquisition by the library would be most practical. As new books would be shelved together, the library attendant could find the re- quested volumes within a small area and could deliver them quickly to the circu- lation desk: Leyh argued further that because · of the · continuing lack of funds; space, and personnel, the ar- rangement by classification created large Open Shelves/ Closed Shelves I 461 backlogs and that new accessions ·could not easily be accommodated by the fre- quently complicated notation systems. He also expounded th~ obsolescence of existing classification schemes and the great efforts involved in keeping pace with the development in all branches of knowledge required to update the sys- tems. He reasoned that it was impossible to assemble all the literature on a sub- jeCt in one place, since classification schemes separated works and placed them in many locations in the system ac- cording to a variety of disciplines. Con- sequently, researchers would have to gather material from ;many locati~ns. They would do much better to find the needed ·works through the traditional bibliographical sources and thus avoid the wasted time hunting . through the .stacks. 20 · As Leyh had indica,ted, a sequential order keeps new material generally to- gether in annual layers, and whole sec- tions c~m thus be easily removed and placed elsewhere. Following Leyh's thinking, some large libraries have creat- ed separate chronological sections, 'such as the Vittorio Emanuele in Rome which has ten chronological groups, ~nd the Ziirich Zentralbibliothek where there are . four such sections, two already transferred to a storage facility.21 The adoption of a mechanical ar- rangement and the resulting exclusion of students from the stacks isolated the European university library and fos- tered, especially in Germany, the devel- opment of institute libraries. These lat- ter are characterized by open shelves and quick acquisition of special literature. They receive strong financial support from the university administration be- cause the faculty, recognizing the value of open shelves which bring the ·stu- dents into immediate contact with the .literature of their field, .present strong cases for them. 462 f College & Research Libraries • September 1976 OPEN SHELVES IN AMERICAN LmRARIES While most prominent European uni- versity libraries excluded their students from direct access to their collections, the public library movement in the United States prospered and had already begun to influence public library theor~ in Scandinavian countries. The Amen- can brand of democracy determined the development of a philosophy of public library work which was different fr?m that in central European countnes. Within this movement .. open access" to the collection was the most radical and · tradition-breaking innovatipn. It emerged as a library technique whi<:h again had the needs of the reader m mind and thus signified a return to the direct book/ reader relationship of old. Many sociopolitical developments con- tributed to its general acceptance, but the influence of a few factors was par- ticularly strong. Some college libraries which allowed their students free access to their collections may have served as examples for public libraries. As these libraries and the student bodies served were both small, it was natural to use the available books directly in the li- brary rooms. At Brown University this practice had begun in 1848; Cornell was another early example; and at Mt. Holyoke Seminary, teachers and stu- dents could study in comfortable al- coves in which the books were ar- ranged.22 The democratic concept of equal opportunity for all citizens de- manded the acceptance of this policy in public libraries. In addition to teach- ers and clergymen who already had ac- cess to the shelves of public libraries, 23 extensions of this privilege to others now became reasonable.24 Not all librarians, however, sub- scribed to this belief. The idea of ex- tending library services to an "unknown public" generated arguments, which can be followed for several decades in the editorials and articles of the Library Journal and in the reports and debates of the ALA conferences. Many pessi- mistic prophecies were made for li- braries if book stacks were opened. It was feared that 'theft, misplacements, and mutilations would be rampant. However, there seemed never to be any doubt that open shelves would contrib- ute significantly to the self-education of interested citizens. Librarians were well aware that open shelves were not a panacea for everyone. The general read- er would be at a loss in a large collec- tion and would not know where to tum, whereas the serious student, knowing what he wanted, could find it quickly. 25 At the conference of librarians held in London in 1877, the majority spoke against open access; even forward-look- ing Melvil Dewey was one of them.26 The subject was again debated at the ALA conferences at San Francisco in 1891 and at Lake Placid in 1894. How- ever, at Atlanta in 1899 and at Montreal in 1900, members were now overwhelm- ingly in favor of open access.27 The tide had turned. Many foreign visitors to American and British public libraries, where sim- ilar practices had emerged, were im- pressed with open shelves, "le grand avenir des bibliotheques;, and described them enthusiastically.28 But, except for the Scandinavian countries, the impact upon European continental librarian- ship was not great. Open access libraries were the exception in France and Ger- many, not to mention southern Euro- pean countries. A British writer, Derrick J. Bott, called the German practice- which he considered general throughout Europe-the most highly developed closed system.29 Its purpose was to edu- cate, and this was thought to be best .ac- complished when the staff had close con- tact with readers. Closed stack collec- tions were characteristic of all types of libraries in continental Europe. As t .. • j shown earlier, in university libraries this practice was the result of administrative considerations . . THE IMPACT oF WoRLD W .An II World War II had a great impact up- on - continental librarianship. It marked a new beginning and brought about drastic changes in library philosophy and services. Many buildings and cata- logs were destroyed and valuable collec- tions and treasures dispersed or lost. Much work and thought was necessary to rebuild and reorganize libraries. The library situation in Germany presented a particularly interesting case, because destruction from the war and the influ- ence of the political ideologies of the occupying powers gave a . new orienta- tion to the reconstruction efforts. After the war, German academic li- braries were much concerned with mak- ing the remaining collections available as quickly as possible. To expedite this process, prewar administrative methods were chosen again because they had al- ready proved to be efficient and econom- ical. Books were arranged in most li- braries by size, with a numerical suborder. 30 Philosophical considerations were postponed for later exploration. However, public libraries in Germany were in a different situation. They had been of little importance and influence, but now a new era had begun. As a part of the democratization process in- troduced by the occupying powers, a new comprehensive school system was crea,ted with equal educational oppor- tunities for all. Information . programs were set up and _cultural exchanges es- tablished to demonstrate in a most prac- tical way accomplishments from abroad. As a result, different attitudes devel- oped, and they created new demands for reading and information. In turn, they also had a strong impact on the de- velopment of public libraries. Well- publi<;ized ex~mples of American li- Open Shelves/ Cl~sed Shelves I 463 brarianship in major German cities were the Amerika -Hauser with their open stack libraries. The impressive Amerika Ge- denkbibliothek in Berlin, established with United States funds, was a highly visible and advertised German library which demonstrated American practices soon imitated by a generation of new public libraries. The open shelf policy is a natural practice in a public library where the collections are current and small enough to be meaningful to a browser. Even though German academic libraries could no longer remain insensitive to the open shelf policy demonstrated so ef- fectively by the American example and by the new emerging public libraries, the academic libraries had large, spe- cialized collections which were not clas- sified and which were frequently housed in buildings unsuited to change. Their strongly defended tradition gave priori- ty to collection building rather than to service. German visitors to the United States observed, however, that the openness of the American research library had made it more and more a center of study. The open shelves and . study cubicles had brought scholars and books together rather than separating them.31 It seemed . obvious that the American library was oriented towards its clientele and that its services were determined by the needs of its users. Several papers read at the 1956 conference of academic librarians in Berlin surveyed the situation and ex- plored the function of the academic li- brary in postwar Germany. It had be- come evident that the general university library had isolated itself because of its closed stack policy and its restricted ~~les of access, and so institute libraries grew . rapidly. These latter, because of their -open shelves .and ease of use, had developed at such an astounding rate that in 1966, according . to Pflug, 80 per- cent of the acquisitions budgets of two 464 I College & Research Ltbraries • September 1976 North German universities had been al- lotted to them rather than to the general university library.32 It was argued that, if the university library again wanted to assume its role as a center of re- search, it must give priority to new re- search requirements and to needs of its clientele. Open shelves, the historically natural relationship with library users, appeared now as the pivotal point which could again regain for the library its central role within the organization it was sup- posed to serve. 83 It was also recognized that, because German traditions and conditions were different from those in the United States, foreign practices could not be adopted without scrutiniz- ing their suitability and adaptability. 34 The old arguments concerning closed shelves and open shelves were once more the center of debate. It was agreed that administrative considerations could no longer dictate library practice and that new approaches had to be explored. In order to bring the readers back into fo- cus, they had to be identified. The Ger- man academic library served two dis- tinct groups: the research worker and the student. The researcher still sought access to a highly specialized literature primarily through traditional catalogs, bibliographies, and indexes; the student, par-ticularly the beginner, was part of a rather homogeneous group that relied heavily on general literature to satisfy class requirements. A solution had to be found that would meet the needs of both of these groups. THE COMPROMISE Open access as demonstrated in Anglo-American countries did not seem to be an ideal solution for German re- search libraries. For some time it had been evident that there is a maximum size for an open shelf collection beyond which open shelves become a liability and a luxury, and even a disservice . to readers and staff. 35 The conre,pt of free access, which was originally conceived as an aid to ·the user, had in the transi- tion from the small library to the large brought with it many of the difficulties which it had proposed to eliminate. A compromise has evolved in the practice of German academic libraries. In order to serve both groups of read- ers, closed stacks are used to house the specialized research material, and divi- sional reading rooms have been devel- oped for students in their first years of study, as well as for general readers. Closed stack areas are dictated by the architectural plan of the older library buildings, by the sequential arrange- ment of the collections which would be very costly to classify, and mostly by the belief that browsing is not a profitable activity for the research worker. The reading rooms, which originally housed primarily the noncirculating reference collections are now enlarged to include a scaled-down version of the entire li- brary collection. This practice has be- come the subject of much discussion: How large should the open collection be? What should it contain? How should it be arranged? How should the reader be instructed in its use? The character and size of the library itself must determine the number and size of the divisional reading rooms.36 The purpose of the open collection is to put students into imrr1ediate and di- rect contact with a well-selected basic body of literature in their fields of study. The collection could be rather large, with 100,000 volumes still consid- ered acceptable. 37 The literature assem- bled must always represent an integral unit and must not be a collection of disjointed small subject groups. Users must be able to perceive their subject fields in their entirety, both in their or- ganization and in connection with relat- ed areas. 38 This is true whether there is a large reading room or several smaller reading rooms. In the university library at Constance, for example, there is no t :J ~- . general library collection but a number of special reading rooms. 39 Kluth suggested that the open collec- tion should contain a good representa- tion of the literature most frequently used, works of established value on a given subject, and important new rele- vant titles, but excluding historical works. The currency and relevance of the assembled literature must be scruti- nized daily; requests received at the ref- erence desk and titles popular in circula- tion should guide in this task. 40 This maintenance of the collection is an im- portant responsibility for librarians, providing another dimension to their work. In order to keep like material to- gether, the reading room collections should not be divided into several groups according to size. Because open shelf collections are designed with the general reader in mind, teaching is an important activity for the librarian in- volved. As in all defenses of open ac- cess, the fact that it also supports aca- demic teaching has not been over- looked.41 Just as German librarians in their consideration of the concept of open access have evolved the above compro- mise, ·librarians in the United States and England, traditionally the strongest de- fenders of open shelves, are beginning to reexamine their ideas. In many cases open shelves have become an anachro- nism no longer serving the ·needs of the modem research worker. Critics ·· are aware, Jor example, that the large num- b er of books .assembled in · a university _library on a giveri subject defeats brows- ing. 42 Furthermore, in his· recent study George Pitemick concluded that "it is no longer realistic to hope to provide the same degree of accessibilitY to every item in [the university's] ever growing stocks."43 · Several developments indicate the de- sire . to restore to· the us~r a manageable body ·of literatt~re : This is evidenced in the · transfer of little-used material Open Shelves/Closed Shelves I 465 from open access into storage libraries, in the formation of departmental and divisional libraries, and in the growing number of undergraduate libraries on large university campuses. Leyh' s argu- ment of sixty years ago, pointing to the impossibility of keeping all of the li- brary's holdings on a subject together, has been emphasized today by the tre- mendous overlapping in modem disci- plines. The prevailing classification sys- tems scatter subject matter widely. Jour- nals are at the heart of research in many disciplines, and their contents are retrievable only through bibliographical sources, not through browsing. A fur- ther fragmentation has arisen because of information available in new for- mats that cannot be easily interfiled in a regular shelf sequence. This is the case not only for audiovisual materials, such as tapes and cassettes, but also for the increasing number of publications in microform. Another modem development which makes browsing for the researcher im- possible is the fact that no longer can the large library present the universitas litterarum. As these libraries are forced into greater specialization, the universal- ity of knowledge must be reconstructed cooperatively through coordinated ac- quisition plans, reference networks, and costly interlibrary loan activities. . The need to control the growing amount of data in the sciences,. which could .no longer be accessed through tra- ditional classification devices, has result- ed in externally produced data bases in machine-readable form. Access to such data bases is provided through terminals located in libraries. Accord~ng to .a re- cent survey of ·the Association of Re- search Libraries, computer-based biblio- graphic searching ·has . already become a · viable and effective research tool in university libraries.44 In this process, ma- terial of high in_terest potential can be retrieved in great · specificity and in many combinations and much more ·466 1 College & Research Libraries· September 1976 effectively than through traditional , browsing in the stacks. As a result, the arguments· for brows- ing on open shelves for the sake of serendipity, are no longer convincing. Through browsing, the researcher can access only a small portion of the li- brary's potential information resources with a great expenditure of his time. In terms of modern efficiency, the high costs of -classifying an entire research collection and of housing it in · a syste- matic manner are, therefore, seemingly out of proportion to the · alleged bene- fits a relatively small group of users may derive. · CoNCLUSION The solution worked out by German research libraries is interesting . and · may be adopted by other large research li- braries: closed stacks, employing com- pact shelving, for research · literature and modern reading rooms tailored to the requirements of general readers. The traditional retrieval method via catalogs and indexes for the library- owned specialized literature could be supplemented by computer technology that allows access to an enormous amount of data in machine-readable form at remote locations. The librarian will work closely with the researcher in the formulation of the search proce- dure and exploration of ·informational sources available. Consequently, there would be a .. far more flexible interac- tion than would result by wandering through the ·stacks, as if one were shop- ping in a supermarket."45 With specialized collections made available through catalogs, bibliogra- phies, indexes, and terminals, the gen- eral collection housed on open shelves could be arranged according to tradi- tional classification schemes or perhaps by more functional methods to meet user needs. As some modern studies have indicated, these differ widely in various areas.46 · The above considerations indicate that the practice of making the entire holdings of a research library available on open shelves may no longer be justi- fiable in terms of economic efficiency and reader needs. In times of budgetary stringencies it is always beneficial to look to history, to inquire into practices used elsewhere, and to reexamine cur- rent methods. The provision of a com- bined closed and open shelf arrange- ment can reconcile a library philosophy of service with the requirements of modern administrative principles. ' REFERENCES 1. Carl Wendel, "Das griechisch-romische Al- tertum," in Georg Leyh, ed., Handbu.ch der Bibliothekswisseruchaft, 2. AuH. ( Wies- baden: Harrassowitz, 1955) 3, pt. 1:140. 2. Georg Leyh, "Aufstellurig und Signaturen," in Handbuch der Bibliothekswissenschaft (Wiesbaden: Harrassowitz, 1961) 2:688. 3. Wendel, "Das griechisch-romische Alter- tum," p.140-41; Leyh, "Aufstellung," p.685. 4. Richard Aungerville, The Love of Books: Being the ·Phiwbiblon' of Richard de Bury (London: Burns & Oates, .1905), p.l16-17. 5. Carl Christ, ••Das Mittelalter," in Hand- bu.ch der Bibliothekswissenschaft 3, pt. 1 : 271. 6. Ibid., p.430. 7. Gabriel Naude, Advice on Establishing a Library (Berkeley: Univ. of California Pr., 1950), p.74. 8. Ludwig Klaiber, "Die franzosischen Biblio- theken seit der Renaissance," in Handbuch der Bibliothekswissenschaft 3, pt. 1:704-5. 9. Georg Leyh, "Die deutschen Bibliotheken von der AufkUi.rung his zur Gegenwart," in H andbuch der Bibliothekswissenschaft 3, pt. 2:121. 10. Ibid., p.129. 11. Gustav Abb, ••Die Bibliotheksbenutzung," in H andbuch der Bibliotheksv;issenschaft 2:417 . . 12. Leyh, "Aufstellung," p.698. 13. Naude, Advice, p.68. 14. "Circulaire a propos de !'instruction gen- erale relative au service des bibliotheques · ~ t 1 universitaires, 1878," cited by Hugo Alker, "Aufstellung und Signaturen der Univer- sitatsbibliothek Wien, ein Beitrag zur Geschichte des numerus currens," Zentral- blatt fur Bibliothekswesen 68:242 (1954). . 15. Leopold Delisle, Instructions llementaires et techniques pour la mise et le maintien en ordre des livres d'une bibliotheque, 4. ed. (Paris, 1910). 16., Leo Crozet, Manuel pratique du bibliothe·- caire, nouvelle ed. (Paris: Librairie Emile Nourry, 1937), p.78, 81. 17. Leyh, "Die deutschen Bibliotheken," p.378. 18. Alker, "Aufstellung und Signaturen," p.243. 19. Zentralblatt fur Bibliothekswesen 29:241 (1912); 30:97 (1913); 31:398 (1914). 20. Zentralblatt fur Bibliothekswesen 30:131 (1913). 21. Walter Bauhuis, "Katalogprobleme und Losungen," Zentralblatt fur Bibliothekswes- en 61:112 (1947). 22. U.S. Office of Education, Public Libraries in the United States of America, Their His- tory, Condition and Management (Wash- ington, D.C.: U.S. Govt. Print. Off., 1876), p.90, 108. 23. Herbert Putnam, "Access to the Shelves, a Possible Function of Branch Libraries;.' Library ]ournal16:65 (1891 ). 24. Frederick J. Teggart, "Access to the Shelves," Library ] ournal 24:667 ( 1899); Thomas Wentworth Higginson, "Access to Shelves," in Arthur Bostwick, ed., The Li- brary Without Walls (New York: Wilson, 1927). p.155. 25. Teggart, "Access to the Shelves," p.671. 26. Library Journal2:272 (1877). 27. "Dana at the Montreal Conference," Li- brary ]ournal25:153 ( 1900). 28. Eugene Morel, Bibliotheques (Paris: Mer- cure de France, 1908) 2:181. 29. Derrick J. Bott, "Fifty Years of Open Ac- cess," Library Association Record 46:212 (1944). . .30. Aranka Racz, "Szechenyi-Bibliothek und deutsche Staatsbibliothek, Zentralblatt filr Bibliothekswesen 71:426-45 ( 1957). 31. Hermann Tiemann, "Neue Leses.aalaufga- Open Shelves/Closed Shelves I 467 ben in den wissenschaftlichen Bibliothe- ken," Zeitschrift fii,r Bibliothekswesen und Bibliographie 3:175 (1956). 32. J oachiin Stolzenburg, "Bibliothekssystem und systematische Aufstellung," Zeitschrift fur Bibliothekswesen und Bibliographie 14: 305 ( 1967 ). 33. Rolf Kluth, "Die Freihandbibliothek," Zeitschrift fur Bibliothekswesen und Bib- liographie 7: 109 ( 1960). 34. Tiemann, "Neue Lesesaalaufgaben," p.172. 35. F. W. Ratcliffe, "Problems of Open Access in Large Academic Libraries," Libri 18:109 (1968). 36. Kluth, "Die Freihandbibliothek, p.106. 37. "Freihandbibliothek," Lexikon des Biblio- thekswesen ( Leipzig, 1969 ) , p.265. 38. Tiemann, "Neue Lesesaalaufgaben," p.174. 39. Stolzenburg, "Bibliothekssystem," p.308. 40. Kluth, "Die Freihandbibliothek," p.99--100. 41. Peter Moosdorf, "Zur Problematik der Freihandausleihe in Hochschulbibliothe- ken, Thesen und Vorschlage," Zentralblatt fur Bibliothekswesen 86: 152--64 ( 1972). 42. Gilbert W. King, Automation and the Li- brary of Congress, a Survey Sponsored by the Council on Library Resources (Wash- ington, D.C.: Library of Congress, 1963), p.23. 43. George Piternick, "Book Storage in Aca- demic Libraries, a Report Submitted to the Council on Library Resources" ( V ancou- ver, 1974). 44. Jeffrey J. Gardner, David Wax, and R. D. Morrison, Jr., "The Delivery of Computer- Based Bibliographic Search Services by Academic and Research Libraries," ARL Management Supplement 2 (Sept. 197 4). 45. King, Automation and the Library of Con- gress, p.23. 46. F. Celoria, "The Archaeology of Serendip," Library Journal 94:1846-48 (May 1, 1969); S. M. Apted, "General Purpose Browsing," Library Association Record 13: 228-30 ( 1971 ) ; Philip M. Morse, "Search Theory and Browsing," Library Quarterly 40:391-408 ( 1970). '------ -----_j · · List of new and forthcoming titles - lstoricheskie Zapiski. Index Great Britain. Cabinet ·Office • . · Vols. 1 - ~0 (1937- 1972). Cabinet. History Seri8$, · Compile