College and Research Libraries By ROBERT B. DOWNS / American Library Cooperation in Review FOR TWO or three generations American cooperation, and it is not difficult to perceive . library leaders have been urging various the reasons why they have been prolific for types of joint effort. The potentialities of the past seventy-five years. Such lists do library cooperation .are widely recognized not call for indefinite commitments on the and, for certain fields, accomplishments part of the individual library, they do not hav~ been ·notable. In bibliographical ! require the library. to give up anything, most enterprises, for example, no profession can of them are of immediate practical use, and point to finer cooperative achievements than it is a matter of considerable pride for the Poole~s Index~ the Union List of Serials~ institution's holdings to be well represented,. the National Union Catalog in the Library These factors have aided in the creation of of Congress, and similar undertakings monumental works like the Union List of successfully carried through by American Serials in Libraries of the United States librarians. and Canada~ List of the Serial Publications On the other hand, certain phases of of . Foreign Governments~ International library cooperation have made slight head- Congresses and Conferences_, I840-1937.~ way, though vigorously and convincingly American Newspapers~., 1821-1936., and _ promoted. Perhaps the time has c~me, Brigham's Bibliography of American N ew1- therefore, for realistic stocktaking to see if papers_, 1690-1820., not to mention hundreds we are proceeding on the right track and of similar lists restricted to smaller regions to determine as objectively as possible in and special subjects or types of material. wHat dir~ctions future efforts are likely Union lists of books which have thus far to produce the most fruitful results. As a appeared have been mainly concerned with preliminary, an attempt will be made here ,large sets, early imprints, and specialized to summarize the various kinds of coopera- topics. Among recent examples are Still- tive arrangements developed to date, with well's Incunabula in American Libraries., some indication of their present status.1 Bishop's Checklist of America'!- Copies of Union Lists uShort-Title Cataloguen Books., Thomson's "Monographic Holdings of American Li- braries in the Medieval and Renaissance , Union lists hav~ _m_e_t )Vith a more cordial Fields," 2 Historical Records Survey's reception ·than any other sort of library 1 Several previous · studies proved helpful in the preparation of this review, especially: Bishop, W'illiam \Varner. "Resources of American Libraries." Li· brary Quarterly 8:445-79, October 1938; Wilson, Louis Round. "Resources of Research Libraries: A Review." College ana R-esearch Libraries 5:259-66, June· 1944; and a more general report, "Cooperation and Coordination in Higher Education." American Council on Education Studies, series I, vol. :z, no. 5, April 1938. SEPTEMBER., 1945 numerous publrcations in the American Im- prints Invent ory series, Emeneau's Union List of Prin,ted Indic Texts and Transla- tions in American . Libraries., Gardner's Union List of Selected Western Books on 2 Progress of Medieval and Renaissance Studies, Bulletin 18::z8-52, June 1944. 407 I ChinaJ and Karpinski's Bibliography of Math ematical Works Print ed in America Through 1850. The manuscript field has not. been neglected : extensive lists were issued by the Historical Records Survey, and there are more limited works, as Pole- man's Census of Indic Manuscripts in the United States and Canada and De Ricci and Wilson's Census of M edieval and Renaissance Manuscripts in the United States and Canada. Obviously, the number of union lists pertaining to ·specialized sub- jects and categories of material can be, and is likely to be, expanded indefinitely. In fact, many are in preparation; an important instance is the checklist. of . about twelve thousand Russian titles in American li- bra:ries, now being compiled in the Library of Congress. The chief drawb acks to union lists of the character mentioned are the considerable expense to libraries in checking their holdings for inclusion, the fact that the lists soon go out of date and reqlJire revision, the small editions in which issued, and the danger of ,excessive multiplication of lists dealing with minute segments of knowledge. Union Catalogs Berthold 's "Directory of Union Catalogs in the United States" 3 records a total of I I 7 catalogs now functioning in this coun- try, divided among several · principal types: . national, regional, local, subject, exchange, and Library of Congress depository catalogs. The oldest of these, th~ National ·Union Catalog in Washington, dates from the beginning of the present century. Only one regional catalog, that of the Califor.nia State Library, antedates 1~0. The vast federal relief program of t·h~ 1930's fur- nished the stimulus for dozens of new union catalog~-city, county, state, regional, ex- change, and subject-widely distributed 3 Dow n s, Rob ert B., ed. Union C(l.tafogs _ i n t he U nit ed S t ates. Chi cago, Am~ican Libra ry Associa- tion, 19 42 , p. 351-91. over the nation. For several years a mass . of free labor from federal government relief agencies was available for compiling cata- logs. The gradual, and finally complete, withdrawal of this aid has placed some of the catalogs in . a difficult financial position, forcing them to carry on with restricted budgets and staffs. It would be a fair statement to say that most union catalog sponsors have not been particularly concerned with fitting their catalogs into any kind of national plan, and, consequently, some duplication of effort, questionable regional divisions, and other lack of integration are evident. , A comprehensive survey of the union catalog - situation was made in 1940-41 by a group of investigators under the sponsorship of the A.L.A. Board on Resources of American Libraries and was subsequently. published.4 Recommendations were made therein for the future coordination and development of union catalogs, to insure thorough coverage of every portion of the country without needless overlapping and with due considera- tion to fiscal support. · Standing at the summit of our system of union catalogs is the great National Union Catalog in the Library of Congress. It is generally agreed that the _maximum development and expansion of this catalog should be the primary objective of any union catalog program for the country. A committee of the Association of Re- search Libraries .was appointed in 1942 to study ways and means of obtaining a full record for the National Union Catalog of all · titles held by libraries in the United States. 5 The committee came to the conclusion that publication in book form of the · Catalog ot Books Represented by Library of Congress Printed Cards offered a possible solution to the problem; approximately one hundrtp selected libr~ries 4 Op. cit . 6 Downs, Robert B. "Expanding the National Union Catalog." A .L.A. Bulletin 37:432-34, November i 943. 408 COLLEGE AND RESEARCH LIBRA R IES I • were asked to check volumes of the ~ub­ lished depo.sitory catalog and to report any titles not recorded to the National Union Catalog. A favorable response to the proposal was received from a majority of libraries approached. By reason of dif- ficult wartime condition~, especially labor shortages, the process of checking has not ptoceeded as rapidly as originally hoped for, although substantial contributions have been forwarded to Washington. A further step in the growth of the National Union Cata- log is the incorporation of entries from the leading regional union catalogs. An ap- p~opriation from Congress in 1943 is making. possible the checking of the Cleveland and Philadelphia catalogs for this purpose, add- ing the holdings of hundreds of individual libraries. That we are far from the goal of a complete record in the National Union Catalog of all books in American libraries is strikingly demonstrated in LeRoy Merritt's study contained in Union Catalogs in the United States which, on the basis of ~xtensive sampling, shows about four mil- lion titles thus far lacking. Bibliographical Centers Closely related to but broader in scope1 than the union catalog is the bibliographical center. The pioneer organization of this type is the Bibliographical Center for Re~ search, Rocky Mount'ain Region, at Denver, which grew out of various experiments in library cooperation. Creation of the Den- ver center was begun in 1935, and since that time the development has ·been rapid. Thirteen institutions in Colorado, Wy- oming, Utah, and New Mexico are now contributing financial support. The center's most basic tool is a union catalog, compris- ing a Library of Congress depository cata- log, printed cards issued by John Crerar, Folger Library, and Princeton University, and records of holdings of some thirty libraries in the Rocky Mountain states-a SEPTEMBER, 1945 total of 3,400,000 cards. In addition, there is a collection of ten thousand .volumes of national, trade, and subject bibliog- raphies, library and sales catalogs, and several special card files. Present activities of the Denver bibliographical center fall into these main categories: location of books and other printed material, serving as a clearinghouse for interlibrary loa'ns in the region, supplying cataloging data, and pro- moting coope.rative book acquisition. Rec- ords show the center is being used con~ stantly by a variety of persons, among them college faculty members and other teachers, students, businessmen, government officials, authors, and adult study groups. Ov_er thirty thousand inquiries were received in 1943, according to a recent report.6 · Patterned aft~~ the Denver center are the Pacific Northwest Bibliographic Center at Seattle and the Philadelphia Bibliographical Center. The first is sponsored by the University of Washington and the second by the University of Pennsylvania. The resources, types of service, and general objectives are similar in all three centers~ There is an unmistakable trend on the part of other union catalogs to expand and take on the functions of a bibliographical center, for they find, sooner or later, that their card files are insufficient to furnish the wide range of information for which they are called upon and they must develop other facilities in or.der to proyide satisfacto~y services. Because of time and transportation factors and relatively inadequate collections for research in the Pacific Northwest and · Rocky Mountain areas, t'he bibliographical centers in those regions are particular! Y vital, giving them resources far beyond the capacity of any single institution. Descriptions of Resources Another device for providing information e The Bibliographical Center for Research, Rocky Mountain Region. (City Club Pamphlet, No. 27) Denver, 1944, p. 42. 409 about library resources is descriptions of holdings. The first attempts of this nature were limited to lists of special collections, perhaps with brief notes, s~ch as those by Johnston, Mudge, and Richardson. More recent examples have generally been planned on a broader scale, to cover resources / as a whole, npt simply special collections. They vary in comprehensiveness from 'reports on single libraries to surveys national in scope. Among the best of the guides to individual \institutions are those for the New York Public Library, 7 Harvard, 8 University of Pennsylvania, 9 and American Antiquarian Society. 10 The same procedure has been applied to cities11 and to regions. 12 In the city and regional studies all types of libraries and all important subj~ct divisions repre- sented were described. Several country- wide investigations covering a variety of fields have also been completed or are in process. These include the Joint Committee on Library Research Facilities for National Emergency's survey of materials in science and technology potentially useful for war purposes, 13· and the Special Libraries As- sociation's projected four-volume work, of 7 New York (City) Public Library. Guide to the Reference Collections of the Library. New York City, New York Public Library, 1941. 416p, 8 Harvard University Libr a ry. Library o{ Ha1'vard University. (Special Publications, 6) Cambndge, Har- vard University Press, 1934. r86p. 9 Bibliographical Planning Committee of Philadelphia. Faculty Survey of the University of Pennsylvania Li- braries. (Philadelphia Library Resources, No. I) Philadelphia, University of Pennsylvania Press, 1940. 202p. 1o American 1Antiquarian Society. Guide to the Resources of ·the Society. Worcester, The Society, 1937· 98p. 11 Downs, Robert B., ed. Resources of New York City Libraries. Chicago, American Library Associa- tion, 1943; Library ,.'and Reference Facilities in the Area of the Distric t of Columbia . Washington, Li- brary of Congress, 1944. 95P·; Bibliographical Plan- ning Committee of Philadelphia. Philadelphia Li- braries and Their Holdings. Philadelphia, University of Pennsylvania Press, 1941. 46p. 12 Downs, Robert B., ed. Resources of Southern Libraries. Chicago, American Library Association, 1938. 370p.; Van Male, John. Resources of Pacific Northwest Libraries. Seattle, Pacific Northwest Li- brary Association, 1943. 404p.; Carlson, William H. "Library Resources in the Land of Little Water." A.L.A. BulLetin 34:617-27, October 1940. 13 Joint Committee on Library Research Facilities for National Emergency. Guide to Library Facilities for National Defense, rev. ed., ed. by Carl L. Cannon. Chicago, American Library Association, 1941. 235p. which the first has been issued. 14 An annual report on notable materials added to Ameri- can libraries has appeared in the Library Quarterly~ beginning with IQ40.15 An attempt was also made simply to list, on the basis of opinions by specialists, the principal collections in libraries of the United States. 16 From the point of view of the subject ex- pert, surveys concentrating on limited 'fields are perhaps of greatest value, and the num- ber of such studies is steadily increasing.17 The techniques followed in these various surveys --of --resources - have--differed widely. · Some have depended entirely on question- naires and correspondence, others on personal visits by outside investigators or local library staffs; most have followed careful plans, a few have developed hap- hazardly. Their usefulness to the research worker naturally varies in accordance with their completeness, the importance of li- braries included, the amount of specific detail, t.he convenience of arrangement, and similar factors. The library resources survey should be regarded as coxnplementary to the union catalog, neither taking the place of the other. Library Specialization or Division of Fields We come next to a type of library co- u. Special Libraries Association. Special Library Resources, ed. by Rose ·L. Vormelker. New York City, Special Libraries Association, 1941. vol. I, 764p. . lilLibrary Quarterly 10:157·91, April 1940; 11:257- 301, July 1941; 12:175-22o, April 1942; 14:Z32-58, April 1944. 1G Downs, Robert B. "Leading American Library Collections." Library- Quarterly 12:457-73, July 1942. 17 Examples: Wilson, Louis R., and Downs, Robert B. "Special Collections for the Study of History and Literature in the Southeast." Papers of the Biblio- graphical Society· of America 28:97·131, 1934; Gilder, Rosamond1 and Freedley, George. Theatr~r Collections . in Libranes and Museums. New York City, Theatre Arts, 1936. 182p.; Powell, L~wrence C. "Resources of Western Libraries for Research in History." Pacific Historical Review 11 :263-80, September 1942; Carlson, William H. "Scandinavian Collections in the Libraries of the United States." Scandinavian Studies and Notes 15:217-38, August 1939; r6:2'91·303, November 1941; Poleman, Horace I. "Facilities for Indic Studies in America.'' American Council of Learned Societies Bulletin, No. 28:27-107, :May 1939; Hilton, Ronald, ed. Handbook of Hispanic Source Materials and Research Organizations in the United States. Toronto, University of Toronto Press, 1942. 441p. 410 COLLEGE AND RESEARCH LIBRARIES operation on which there is less general agreement than for any · of the activities outlined above. From a practical point of view it is impossible for even the largest libraries, as they are now organized, to hold more than a fr action of the world's literature. Therefore, acquisition agree- ments among libraries would appear, theoretically at least, to be the logical and ·sensible solution. Nevertheless, a consider- able amount of skepti~ism about the practic- ability of agreements for dividing collecting interests among libraries still prevails. Possibly Fremont Rider 18 has hit upon a fundamental weakness when he points out that the scholar is never reconciled to having his research materi'als in some other library, but insist~ on havin g them immediately at hand. Even Mr. Rider's microprint pro- posals, however, are based upon' a high degree of specialization among libraries. Handicaps to library specialization were described further by Taube. 1 9 The per- sistence of interest in the problem, however, is demonstra~ed by two special conferences, with published proceedings. 20 The fir st of these, sponsored by the A.L.A. Board on Resources of American Libraries discussed ' at length difficulties in t4e way of agree- ments -but voted. ~nanimous approval of a resolution on their desirability. The latest contribution to the subject of specialization is the proposal of a committee of the Liprary of Congress Librarian's Council, composed of Keyes D. Metcalf, Archibald MacLeish, and Julian P. Boyd, 21 · 11 Ride r , · Fremont. Th•! Scho lar an d th e Fut ure of the Research Library. New York City, H adham Press, I 944, p. 79-84 . • 19 • Ta~be,, Mor.timer. " 'fhe R ea lities of Library Spe- ctahzatton. L•brar'll Quarterly 12: 246-56, April 1942 . . ~ A.L.A. B«?ar.d on Resources of American Libraries. ar.'V' Specv:,z ation; Proceedings of an Info rmal fer nee. .~hicago,. American Library A ssociat io n, . 148p.; Proceedings of the Conference on Li- brary Specializa tion; " "An Agreement for Regional Libr ary Specialization in the Pacific Nor thwest;" "Ap- pend ix." PNLA Quarterly 8:5 2-59, J a nuary 1944. 21 Me tcalf, Keyes D. 1 and Williams, Edwin E. "Pro~osal f_or ~ ~ivis't on of .~e~ponsibility a mong Ame.nc•n Ltbrar~es ~? the Acqutsttton and Recording of Ltbrary Matenals. College and R es earch Librari es s:xos-os, March 19 44. . SEP1EMBER~ 1945 for a thorough coverage of the world's literature by American libraries. . This would be accomplished by having cooperat- ing institutions, each of which would have agreed to specialize in one or more given divisions of knowledge, acquire at least one copy of every book of potential research · interest published anywhere in the world. The Library of Congress classification would be used as a basis for subject divisions. A second important step in the plan· is to have every book thus obtained promptly listed in the National Union Catalog at the Library of Congress. Its sponsors realize their program cannot be placed in full effect until the war's end but are pro- : ceeding with preliminary details. ' The library profession has a basis of experience extending over a considerable ~ period of time for specialization agreements, though not on such an. ambitious scale as the Metcalf-MacLeish-Boyd proposal. In New York City, fo'r example, an under- standing between Columbia University and the New York Public Library dates back to I 8g6; certain . fields are definitely al- located to one library or the other, and they consult with each other in the twilight zone not covered by this understanding. Working arrangements also prevail with more specialized institutions in the city. An extraordinarily effective program has also long existed among a g roup of. Chicago libraries: John Crerar, Newberry, U niver- sity of Chicago, and Chicago Public Library. The original plan· became effective about 1895; at that time the Newberry Library assumed responsibility for literature, history, and the arts, and John Crerar agreed to cover the natural, physical, and social sciences. Other assignments subsequently were made to the University of Chicago, Chicago Public Library, and other Chicago libraries. In the South one of the most noteworthy examples of library cooperation is that existing between Duke University 411 I· and the University of North Carolina ;22 included is a division of collecting interests for a number of subject fields, newspapers, and public documents. Another fnstance is in Nashville, Tenn., where Vanderbilt U ni- versity, George Peabody College, and Scarritt College have worked out compre- hensive plans for developing joint research collecti~ms as part of a far-reaching program of library c~operation. 23 Numerous other illustrations could be cited of ~uccessful local agreements for sharing acquisition responsibilities. Of special interest are the following: the arrangement, begun about 1927, among the universities of Michigan and Minnesota, the John Crerar and Newberry libraries for purchasing cooperatively certain publications of European local academi s and societies; an understanding among the University of Minnesota, Minneapolis Public Library, Minnesota Historical Society, and Minne- apolis Athenaeum in the fields of genealogy, local history, fine arts, music, and Scan- dinaviana; a similar agreement among Brown University Library and other li- braries in Providence, for collecting local history, art, law~ mathematics, Latin American literature, American poetry, Lincolniana, and state publications; a divi- sion in Cleveland among Western Reserve University, Cleveland Public Library, Western . Reserve Historical Society, the Museum of Art, and the Natural History Museum; the centralization of order work for the six libraries of the Oregon State System of Higher Education ;24 a division, of the Latin American field among Duke 22 Downs, Robert B., and Branscomb, Harvie. "A Venture in University Library Cooperation." Library Journal 6o:877-79, Nov. xs, 1935; Pratt, E. Carl. "Library Cooperation at Duke and North Carolina Uni- versities." College and Research Libraries 2:142-45, March 1941. 23 Kuhlman, A. F., ed. Development of University Centers i,n the South. Nashville, Joint University Libraries, 1942, p. 53-128. 2• Described in detail, along with various other types of library cooperation, in Mrs. Mildred H. Lowell's College and University Library Consolidations. Eu- gene, Oregon State System, 1942. 136p. University, University of North Carolina, and Tulane University; and an extensive cooperative program for preserving state and local newspapers undertaken by the University of Virginia, Virginia State Li- brary, and some two dozen other Virginia libraries. Of a very specialized nature is the agreement among about thirty-:four New England and New York libraries for the preservation of advertising sections of periodicals. Photographic Repr~duction The widespread use of microfilm, begin- ning about a decade ago, has led to co- operative efforts of several kinds. It was realized early that establishment of a laboratory for . film in every library would be uneconomical ; hence, experts in the field . have urged that a few fully-equipped labora- tories be set up to serve a much larger number of libraries. The city-wide service provided by the University of Chicago Libraries' Department of Photographic Reproduction and by the American Docu- mentation Institute's Bibliofilm Service, in Washington, are examples. On . a com- mercial basis, University Microfilms of Ann Arbor, Mich., is performing in a similar manner. A coordination of · microphoto- graphic laboratories is under consideration for Philadelphia. 25 Another step toward microfilm cooperation is directed toward making more generally available work already completed; this was one' of the pur- poses of a recent union list.26 The possibilities of cooperation in the reproduction of large, expensive works are virtually limitless. Among notable projects of that nature .are th~ filming, for some sixteen libraries, of English book_s before I 5 50, as listed in the Short 211 Heilemann, J. J. "Coordination of Microphoto- graphic Labor_atories." Philadelphia Bibliographical Center. Documentation on a Regional· Basis. Phila- delphia, 1944, p. 25-26. 26 Philadelphia Bibliographical Center, Union List of Microfilms. Philadelphia, The Center. 1942. Sup- plement, 1943. 412 COLLEGE AND RESEARCH LIBR~RIES / Catalogue/ the microfilming of all extant magazines published in the continental United States before I 8oo; and the microprinting, in process, of the British Sessional Papers for the nineteenth century. Another is the reproduction by photo-offset of the Catalog of Books Represented by Library of Congress Printed Cards. It was recently reported that files of approximately I 50 major A~rican newspapers are avail- able in film fon~nd in a substantial num- ber of cases the cost of filming was shared by two or more libraries. Before our entrance into the present war a project sponsored by Harvard University received a representa- tive group of about fifty newspapers from . the principal countries of the world and microfilmed them for distribution to Ameri- ca.n research libraries; war conditions have forced temporary suspension of the enter- prise. Also dealing with a special type of material, the Library of Congress and the University of North Carolina Library jointly sent an operator with a microfilm camera to visit libraries throughout the country for the purpose of assembling a complete record of the proceedings of legis- lative bodies of the American colonies, ter- ritories, and states. Going far beyond these relatively limited undertakings is Fremont Rider's proposal to have each im- portant research library become a publisher of microcards in fields assigned to it, the cards to be offered for sale to other libraries. 27 Cooperative Cataloging and Processing A cooperative cataloging project, under the sponsorship of the Library of Congress and the A.L.A. Cooperative Cataloging Committee, was initiated in I932, with the collaboration of libraries in the :United States and Canada. During the .ten-year period, I933-4j, the Library of Congress 21 Rider, l?P· cit., p. I 76-209. SEPTEMBER, 1945 Aa ~ ... received from cooperating libraries n.tf'" edited card copy for about sixty thousand titles, chiefly new foreign books and' mono- graphs in scholarly series. It was recently reported 28 that 365 libraries have cooperated in the contribution of titles for catalog entries which have been printed by the Library of Congress. In the same direction the Library of Congress has made arrange- ments with certain cooperating libraries by which each library is to supply catalog copy for the current official publications of its home state. The plan is to be gradually ~ extended to cover all states. In addition, libraries in several of the larger cities have agreed to catalog the official publications of their respective cities and to supply copy to J · the Library of Congress for printing. U ni- versity libraries are being asked to furr?-ish catalog copy for publications issued by their ' institutions, including doctoral dissertations and the products ot university presses. Another problem of increasing concern to our large research libraries is the huge accumulation of uncataloged materials, ar- rears from preceding years. A series of · recommendations to meet the problem on a national scale was offered by Kellar. 29 A plan for centralizing the cataloging of ar- rears has been suggested for the Philadelphia area,.3° For the past three years a group of Colorado librarians has had under con- sideration an even more inclusive program, looking toward the possibility of centraliz- ing technical processes for all types of libraries in the region. 31 28 Library of Congress. Descriptive Cat a loging Divi- sion. Cooperative Cataloging Manual for Use of Con- tributing Lib1·aries. W as hington, U.S. Government Printing Office, 194 4, p. so- 6o. 29 Kellar, Herb ert A. Memoranda on Library Co· op eration. Washington, 194 1, p. 18-2'9. ao Linderoth, ~rs. uth With erbee. "Centralized Cata loging Projet Phila d elphia Bi bliographical Center. Doc ation on a Regional Basis. Phila- delphia, 1944, p. -9. a1 Colorado College and Head Librarians Conference. "First Report of the Special Committe e, Centralized Technical Processes and Bookbuying." August 194 2 ; "Second Report." February 1943; "Planning Studies on Centralization." October 1942 (Mimeographed). 413 ·- ' _Duplicate Exchanges The system of exchanges for duplicate publications which the Medical Library As- sociation has had in operation' since 1899 is generally acknowledged to be the most effective scheme thus far tried for disposing of such material. . These are the essentials of the plan: lists of duplicates from all co- operating libraries are consolidated in a central office, distributed to-association mem- bers, and records of wants returned to the central office, with the larger libraries in the organization receiving priority for items available. . In 1940 a Periodical Exchange Union, limited to the problem of duplicate periodi- cals, was set up under the sponsorship of the Association of College and Reference Librari~. 32 · The procedure differs from the M.L.A. plan in several respects, chief of which is that lists of duplicates are circu- lated among the participating libraries in the order of the size of their annual budget for . periodicals; libraries making the largest ex- penditures for periodicals have priority over others, on the theory that the material needed by them will be more rare and difficult to ·obtain. An ingenious proposal for disposing of duplicate publications has been offered by Phineas Lawrence Windsor, of the U niver- sity of Illinois, and others. The plan, in brief, is to box up duplicates, unlisted, and ship them to the nearest large library; the receiving library would select anything wan ted, pass on the remainder to another library, and so on until everything was distributed or any residue could be dis- carded. Presumably, the duplicates would need to be roughly classified by subject · or type under this scheme. - 12 Van Deusen, Neil C. "Periodical Exchange Union." College and Research Libraries 2: 288, June 1941. Name changed to Duplicate Exchange Union in 1944 and functions expanded. See Thompson Donald E. "Duplicate Exchange Union." College 'and Re- search Libraries 6:xs8-6o, March 1945. Central Storage Ware houses For the past forty-two years, since Presi- de~t Eliot of Harvard first advanced it, the idea of inexpensive centralized storage for little-used books has been discussed. Up to now, however, so far as the writer is aware, only one such cooperative storehouse has been constructed, namely the New England Deposit Library in Boston, serving Har- vard U ~iversity, Boston Public Library, . Massachusetts Institute of Technology, and other institutions in the area.33 Economy of storage, elimination of duplication, and divi- sion of fields among libraries are primary objectives . A central storage warehouse for the Mid- west, with Chicago as a center, has been advocated for ,several years by President Hutchins, of the University of Chicago, and colleges of the Connecticut Valley have discussed central storage of large reference collections for the college libraries i~ the district. No concrete development has yet come from either proposal. Regional Library Development In several sections of the United States broad programs of regional library coopera- tion, embodying two or more of the specific types of cooperation outlined above, are under way or under consideration. The bibliographical centers of Denver and Seattle of cou~se hav~ aspects touching on many phases of regional cooperation. In the Atlanta-Athens area of . Georgia, Emory University, GeGrgia School of Tech .. nology, University of Georgia, and other libraries have joined in a serie~ of measures to improve and coordinate their resources. Recommendations for combining a group of North Texas institutions into a regional system for cooperative purposes were made by A. F. Kuhlman, on the basis of his as Metcalf, Keyes D. "The New England Deposit Library." Library Quarterly 12:622-28, July 1942. 414 . COLLEGE AND RESEARCH LIBRARIES survey of the situation there. 34 Long-range plans for the development of university library centers in the South were discussed at a Nashville conference in I 944· 35 Conclusion It is inevitable that lack of space a~d lack of information prevent mention of addi~ tional examples of, or proposals for, library cooperation which should be included in a complete record. A number of interesting plans are taking shape in the field of index- ing and abstracting. The present status of the state document center program, inaugu- rated in 1930, would be worth investigation for its bearing on questions of library in- tegration. The problem of eliminating un- wise competition among American libraries in postwar foreign book buying is a matter of concern now receiving particular at- tention from the Association of Research Libraries. Plans being formulated for ex- tensive reprinting of publications originating 1M Kuhlman, A. F. The North Texas Regional Li- braries; An Inquiry into the Feasibility and D esira- bility of Developing Them as a Cooperativ.e Enterprise. Nashville, Tenn., George Peabody College Press, I943· ssp: 35 Conference of Graduate Deans and Librarians, Nashville, Tenn. Development of Library ResoU1' ces and Grad11ate Work in the Cooperative University Centeru of the South, ed. by Philip G. David so n and A. F. Kuhlman. Nashville, Joint University Li- braries, I 944· 8 I p. in Axis-controlled countries will also involve the cooperation of research libraries. "The Checklist of Certain Periodicals" in scien- tific and technical fields published in enemy territory since 1939, a union list of Ameri- can library holdings now being compiled in the Library of Congress, will provide a useful foundation for this purpose. From a close study of accomplishments to date, there would appear to be certain im- portant principles which have influenced the success or failure of various kinds -of library cooperation. First, distance is a handicap, and it is easier for libraries not too far removed from each other to work together. Second, regional library cooperation has its greatest opportunities in those areas with inadequate book resources. Third, libraries should not be asked to give up anything but rather to assume positive responsibilities and , receive direct benefits. Four-th, agreements must be flexible enough to provide for ex- pansion and adjustment. Fifth, complete elimination of duplication between libraries is not possible or desirable. Finally, only a comparatively limited number of libraries are at present equipped to make any sub- stantial or effective contribution to a general prograrp of cooperation on the research level. * * * Comment by CHARLES H. BROWN Union Catalogs and Problems df~ Collecting It is not easy to add to Mr. Downs's com- prehensive inventory of cooperative projects in the field of librarianship nor to comment on most of them. One or two points, how- ever, forced themselves upon the attention of at least one librarian, somewhat like the proverbial sore thumb. We librarians have not made much progress in reaching understandings on acquisition poli- cies. The reasons are obvious. Without af- fecting university policies, librarians <;.an agree on union lists of serials, union catalogs, co- opeiativ~ cataloging, and many other coopera- SEPTEMBER~ 1945 tive enterprises. When acqmsttwn policies are concerned, our graduate colleges are im- mediately involved. So long as university administrators, deans of graduate colleges, and the faculties themselves do not realize . the necessity for some agreement on the vari- ous fields of specialization in research which their universities should undertake, then the librarians can do little. Fortunately, there are indications that this need is beginning to receive more attention in university circles. It certainly is more of a credit to an institu- tion to have a few outstanding departments 415 than to have inany fields in which instruction is given on a mediocre scale. Many universi- ties giving graduate instruction for the doc- tor's degree in the sciences do not have library facilities to support work in those disciplines. In a study now under way in the Iowa State College Library, the one hundred periodicals most frequently cited in representative jour- nals in chemistry and botany were checked in the Union List of Serials. In chemistry only twenty to twenty-two institutions ·in the United States possessed three-fourths of these most-cited sets, although the study by the American' Council on Education1 in 1934 would indicate that thirty-seven were quali- fied for instruction for the doctorate in chem- istry. A.t present over sixty institutions are actually giving the doctor's degree in this field. In botany the study by the American Council would indicate that thirty institutions are considered qualified to give graduate work. Only about fifteen of these possess three- fourths of the one hundred most-cited sets of periodicals in this field. Again, many other institutions not listed in the study by the American Council are attempting to give in- struction for the doctorate in this field. In- terlibrary loans, photostats, and films prove of some assistance, but ·no scientist wants to depend to a very considerable extent on inter- library loans, much less on films and micro- print. Indeed, for much scientific research, research publications must be available on the campus for frequent and extensive examina- tion. In the humanities conditions may be even worse. A .university appoints a professor of history whose research is in the field of Ameri- can colonial history. The librarian builds up a collection. In a few years the professor moves elsewhere, and the librarian is required to build up a research collection for a man specializing in the Civil War period. As yet, librarians are unwilling to adopt a proposal for the lending en bloc of publications more valuable elsewhere. Yet some of us are be- ginning to 'realize that the cost of storing large collections of little-used material ·is becoming an increasing burden and eventually will tax the resources of even our wealthiest institu- tions. Libraries in. a given region will be 1 American Council on Education. Report of Com- mittee on Graduate Instruction. Washington, D.C., April 193 4 . compelled eventually to agree on certain areas in which they will collect exhaustively and other areas in which their acquisitions will be limited. It will be far better for us to attempt to work out some priQ.ciples of co- operative acquisition for publications than to wait until some such policy is forced upon us by our administrators. Under the present system many libraries have attempted to cover so many fields that their collections have become inadequate for research in any field . Furthermore, this tendency to include some- thing of everything, without any complete coverage in any field, has resulted in wide- spread duplication of publications rather easy to obtain and in a corresponding . lack by all libraries in the United States of any copy of many publications which have been or may be urgently needed for research. Publications Not Available Elsewhere In view of the skepticism of librarians in regard to cooperative acquisition of publica- tions, the Metcalf-MacLeish-Boyd committee wisely emphasized the need of acquiring publi- cations which apparently are not available anywhere in the United States. As noted by Mr. Downs, much progress has been achieved in certain fields of the humanities. No men- tion is made, however, of the need in certain fields of science. Some of us have found dur- ing the last few years that no copies of cer- tain publications urgently needed for the war effort could be located in the United States. If these publications had appeared several hundred years ago, as in the case of the hu- manities, the lack would QOt create amaze- ment, but no copy of many scientific publications published during the last forty years could be found anywhere in the United States. If the Metcalf-MacLeish-Boyd com- mittee can provide for some checking of all publications of foreign countries to make sure that at least one copy is available in the ·United States, the committee will be render- ing a service beyond measure to science and industry. If the committe.e had exist~d and been successful in its work in the prewar days, certain requests for material by our armed forces during the last · few years would not have proved so embarrassing. Liprarians generally have developed their acquisitive tendencies to the extreme. The 416 COLLEGE AND RESEARCH LIBRARIES sole rule for accepting a book as a gift seems to be that it is one not already in the collec- tion. It seems to make little difference whether some other library might be able to use the book to better advantage · or not. It may be embarrassing to persuade a potential donor that books he wishes to give to one library might better be deposited in another. The policy of accepting publications obtain- able as gifts and buying without too much system has resulted in some situations which seem almost ludicrous. Many libraries have collections of German doctoral dissertations, all relatively easy to obtain, but when· it comes to dissertations from universities in minor European countries, not one copy may be found anywhere in 'the United States. We have had no system for cooperative acquisition. Without doubt, certain cities in this coun- try . are generously provided with research material. The various libraries in many cities have duplicate copies of publications which are little used, while in other sections of the · country no copy can be found. Certainly research would be greatly stimulated by a · redistribution of research publications on the basis of need, especially in view of the short- age of such publications, which is certain to become worse. The proposal to transfer cer- tain collections which are duplicated in one city to other cities, or even to other countries, has more merit than appeared at first. Pos- sibly microprint will eventually solve all of our problems, but the day for that does not appear to be at hand. A millennium has the habit of remaining at a distance when we attempt to approach it. Union Catalogs One more bit of heresy. Except for local use, the writer cannot see the reason for union catalogs, outside of the Library of Congress, in the regions east of the Missis- sippi. An air mail letter can reach the Li- brary of Congress within twenty-four hours. We c~n obtain information as to the location of a book more satisfactorily from the N a- tiona! Union Catalog of the Library of Congress than from any other source. Why write a center in Chicago, Cleveland, Cincin- nati, or even Philadelphia, when more infor- mation can be obtained from the Library o'f Congress than from any of these cities? For the benefit of libraries in the immediate neigh- borhood of Chicago, a union catalog at Chi- cago might be of assistance, although it would be expensive. For most librarians, the union catalog of the Library of Congress will be the final recourse no matter what other umon catalogs may be set up. * * * Comment by KEYES D. METCALF Division of Fields of Collecting Mr. Downs's article· is admirable in every particular-as a statement of fact and for its critical judgment about the adequacy and the practicability of what has been accomplished or proposed in the way of cooperation in the past. It is difficult to say more or to supple- ment what has already been said. Three minor comments are submitted on the present situation, however. First, the Harvard foreign newspaper microfilm project is still going strong, although it has not been possible to keep up the microfilming of papers from parts of the world that were occupied by the totalitarian powers. · Second, as a librarian in the Boston area, I am glad to go on record to the fact that the New England Deposit Library has been running along smoothly, financially and otherwise, in spite of SEPTEMBER~ 1945 the_ fact that lack of manual labor has made it ·impossible for libraries to send as much material to the deposit library as would have been the case in other times. It is expected that a second unit will be needed soon after the close of the war. And, third, it seems worth while to suggest that the critical point in cooperative cataloging lies in whether or not the Library of Congress or any other agency that might take its place can ever bring itself to accept cooperatively prepared copy without full revisio~. So far, altera- tions have cost more than they · are worth. They have tended to hold back the whole cooperative cataloging program, which . in the postwar period should be ready to expand greatly. In addition to these comments, further con- 417 sideration of division of fields seems called for. The research libraries of the country, as Mr. Rider explains so well in his book The Scholar and the Future of the Research Library, have been doubling on the average every sixteen to twenty years for generations. I am sure that Mr. Rider will agree that, after a library reaches a certain size and age, the rate of growth must stop. I am not ready to name the age, or to say whether the size is one, five, or ten million volumes. I might have said three million, since the New York Public Library, Harvard, and Yale have slowed up since reaching that figure, if I had not j&st finished reading the 1944 Annual Report of the Librarian of Congr ess telling of a net increase of 481,733 volumes in that library in one year. It seems evident, how- ever, that the growth of libraries, like the rate of growth of anything else, whether it be that of an individual, the population of a country, the number of students in a univer- sity, or the size of ships, cannot increase in- ... definitely. The growth of libraries must slow down partly because of the lack of material to collect-it is impossible to believe that the number of books published will .double every sixteen years indefinitely-but it must also slow down because the cost of building con- struction, of acquisition, of cataloging, and of service will at some point become so great that they will take more money than is avail- able. The library in a university, for instance, cannot continue to increase its expenses more rapidly than other parts of the university without taking a larger and larger percentage of the total resources, and there i~ a limit as to how far a library can go in that way with- out becoming more of a nuisance than a blessing. When the time comes that the rate of growth must decrease and the library finds that it cannot continue to collect as exten- sively in all fields as in the past, it is sug- gested that there are at least four different courses of action that may be taken, as fol- lows: 1. A library can definitely adopt the policy of becoming more selective in all fields. This is perfectly possible; but am I mistaken in believing that a library that is selective in all fields, and not really outstanding in any, may be a very good library but can never become a great library-a library with a country-wide reputation to which visiting scholars will come in large numbers, a library where productive research can be carried out on a large scale, a library that can be called truly outstanding? 2. A library may c~ntinue to try to do every- thing that it has done in the past but do it less and less well. It will then become overextended, the quality of its direct service to the public and of its cataloging and its collections will all decline, and it will become a Grade C or D library compared to others. It would not be difficult for any of us to think of a number of libraries that have become overextended, have tried to do more than they could with the funds that were made available to them, and have fallen down on the job. I am sure none of us want-s to slide into that group deliberately if we can help it. 3· A library, when it finds that it cannot continue to keep up with its previous rate of growth, instead of trying to cover all its present fields might, for part of its work, fall back on interlibrary loan; on sending many of its ad- vanced students to other libraries to find their material; on microfilm reproducti~ns for par- ticular items that are wanted; or, if Mr. Rider's dream comes true, on microcards. All of these throw the burden on someone else, while pro- viding little or nothing in retu'rn, and sooner or later would result in an unbearable situa- tion and a loss of reputation. 4· A library might finally go along with a division of fields, as proposed by the Metcalf- MacLeish-Boyd committee and outlined in Col- lege and R esearch Libraries for March 1944.1 By this plan a library would become more selective in mo st fields but mo r e inclusive than before in certain limited fields for which it agrees to assume re sponsibility and for which it will freely furnish book's to others by inter- library loan, by photographic reproduction, or by caring for visiting scholars, thus retaining its self-re spect when it calls on other libraries for help. Having lived with this plan for some months now, it is easy to see the objections to it. It is not going to be .easy to organize on even a small scale. It is going to be very difficult to persuade libraries to reach the necessary agreements and, having reached them, to continue with them. There will be frequent complications when a university pro- fessor who has built up a strong collection dies or transfers to another university, with the result ·that his collection proves to be in 1 M etcalf, Keyes D., and vVilli a ms , Edwin E . "Propo sal f o r a Divi sion of Responsibility amon g Ameri can Libraries in the Acquisition and Recording of Library Materials." College and Research Libraries 5 :ro5-o9, March 1944. 418 COLLEGE AND RESEARCH LIBRARIES the wrong place for future use. It should be admitted that the plan is not a cure-all, but it will help w hen it comes to the specialized collections. It is these special collections that are most expensive to acquire, catalog, and maintain, but it is of first importance to have them somewhere in the country, and a co- operative program in connection with them seems desirable. Note may properly be ma,de here of prog- ress in the above-mentioned committee's plans. The surveys of Belgian and Mexican publica- tions have been completed; those for Sweden, Spain, and Peru are well under way; others will come along in due course. The results so far are enlightening. For Belgium it was found that '79 per cent of the titles selected from the 1937 lists as being of possible interest to research libraries cannot be found in any one of the fifty-five libraries that reported. • These include practically all of the large gen- eral research libraries in the country. The total cost of all the priced items published in 1937 would have been only $1160. That is, if one university library had been willing in the year 1937 to acquire all books of research importance published in the regular trade in Belgium in that year, it would have cost the institution only $u6o, and the other libraries of the country would hav_e felt secure in their knowledge that they could be selective as far as Belgium was concerned, because all the material could be readily found elsewhere in the United States. * * * Comment by RAYNARD C. SwANK Cooperative Subject Bibliography The librarian who reads Mr. Downs's article ''American Library Cooperation in Re- view" may feel proud of the cooperative achievements of his profession; and, if he is a cataloger or a bibliographer, he may feel, as does this author, especially pleased that no field of library activity is more notably repre- sented than the bibliographical. To union lists, union catalogs, bibliographical centers, descriptions of resources, and cooperative cataloging, more than half of Downs's re- view is ... devoted. Yet these achievements represent for the most part but half the field of enumerative bibliography-that half which concerns the description and location of specified books or collections. The . other half, which concerns the listing of books pertaining fo specified sub- jects, is but meagerly represented. This omission is not an oversight. Indeed, it ac- curately reflects the present stage in a normal development of bibliographical enterprise. The foundation upon which any system of subject bibliography must rest is patently the finding list. Unless books can be located, there is no point in seeking references to them in subject lists. That this foundation is al- ready being well laid at the interlibrary level is evidenced by an impressive array of such cooperative works as the Union List of S e- rials~ American Newspapers~ 182 1-1936~ and SEPTEMBER~ 1945 the National Union Catalog in the Library of Congress. Yet it is equally patent that the . finding list can never achieve its greatest usefulness without subject catalogs or bib- liographies to supply references to books which subsequently need to be located. Provision for an adequate subject approach to library materials, als~ at the interlibrary level, is the indispensable next step in the development of a complete bibliographical system. The librarian thus far has not altogether neglected this other half of the general biblio- graphical problem. The cooperative catalog- ing project, although primarily concerned with descriptive cataloging, aids in the assignment of subject headings for books entered in the card catalogs of individual libraries. But the subject catalog of the individual library com- plements the author catalQg of that library only, not the union auth~r catalog or the union list. As long as a person selects books from the subject catalog of one library, he will have no use for a finding list of books in other libraries. At the interlibrary P,lane a partial subject lead is offered by descriptions of the resources of various groups of libraries; but, valuable as these general descriptions are, they do not actually supply references to the materials on any subject. For lists of actual references to subject matter not contained in particular libraries, one must still depend 419 wholly on pure subject bibliography-that heterogeneous mass of apparatus which has . been created primarily by the working scholar and which has not yet received the concen- trated attention of the library profession. The need of a more adequate subject ap- proach to books has been sharply felt by many librarians in recent years. Proposals for the compilation of new union subject cata- logs, in ·both card and book forms, have roused occasional discussion. This author, among others, has suggested greater emphasis in libraries on the compilation and exploitation of pure subject bibliographies. Striking out in another direction, Mr. Rider, who had previously advocated a plan for the publica- tion of union dictionary catalogs, has now startled the library world with the possibili- ties of a microcard catalog, a major purpose of which would be the provision of a more effective subject approach to research mate- rials. The exact nature of the bibliographical sys- tem which may eventually best satisfy the subject ne'eds of the reading public is anything but clear at this time. Indeed, the exact nature of the need itself is but vaguely known. There is still much, very much, to be learned about bo'Oks and, especially, about readers be- fore any comprehensive reorganization can be safely undertaken. Nevertheless, several basic propositions ha·ve already emerged in forms sufficiently clear to command the at- tention of every thoughtful librarian. If they _ are valid, as this author believes they are, the general character of future subject bibliog- raphy is indicated. Provision of Subject Approach I. The provision of an adequate subject approach to books is a general bibliographical problem, not merely a problem of library cataloging. The library cataloger is and should be vitally interested in the matter, but he cannot readily solve it without reference to the experience and achievements of others who have long grappled with the same problem. Indeed, any proposal must be viewed with suspicion which does not account for the large amount of bibliographical activity apparent now, as always, among working scholars and professional bibliographers, as well as many ' librarians. No analysis can be sound which does not seek out the forces which reared such monumental works as the Cambridge Bibliography of English Literature, or which compel scholar after scholar to compile supplem~nts, however clumsy, to such tools as Hammond's Chaucet: manual, or which oblige readers advisers the country · over to develop special lists and files of references found useful for different types of readers. On the other hand, any program which does not take into consideration the real values of library cataloging must also be viewed askance. This is not a matter for either the cataloger or the bibliographer alone; it is their common problem. And not until each gains an enlarged conception of the whole, a firmer grasp of the purposes and accomplishments of the other, and, above all, a more realistic awareness of actual needs-not until then can a satisfactory subject scheme be developed. 2. An adequate subject approach can be provided only by ,a system of special tools aimed directly at special needs·, not by general tools aimed broadly at g eneral needs. One of the most disturbing elements · in traditional cataloging thought has been the conviction that efficient service requires the general use of single, universal, subject tools for all readers. This shotgun method assumes a mass of average readers with unoriginal problems about stereotyped subjects. It as- sumes 1urther a general unity of knowledge and a need in every re.ader to have all parts of that knowledge simultaneously at his fin- gertips. A more sensitive, discriminating method singles out comparatively small groups of readers, with fairly homogeneous needs, and then aims directly at them. For purely descriptive and locational purposes, the com- prehensive, standardized, author tool may be altogether desirable, since books and their locations do not vary with readers' interests; but, for the selection of books to fit individual needs-and all needs are individual-the spe- cial approach is likely to be more rewarding than the general. Based on Value to Reader 3· The selection of materials to be listed under su·bjects should be based on value to the reader, not on ihe incidence of materials in particular libraries or on the circumstances of publication. The criteria which determine what is listed 420 COLLEGE AND RESEARCH LIBRARIES or not listed in many tools are manifold and confusing. In a large number of subject catalogs and indexes, as well as many "bib- liographies," the material must first belong to the library, or libraries, or to the type of material being indexed. Second, it must to a large extent possess bibliographical . inde- pendence-that is, it must consist of a separate work with its own title. Third, in many instances, it must appear in a prescribed physi- cal form-a volume of so many pages, with or without boards, not on microfilm, or the like. Finally, then, if it still qualifies, it may be examined for relevance to subjects or readers' interest. But, meanwhile, a multi- tude of impertinent factors have so condi- tioned the results that no one can say precisely why this or that title appears or does not appear or precisely what need the list is capable of filling. From the point of view of the person who seeks references to guide his reading, such tools are low in biblio- graphical significance. Significance, in this sense, is the fulness of meaning which a list holds for the reader, and the most significant list is the one about which it is possible to say, "Here are refer- ences to the materials-all the materials and nothing but the materials-relating to· this subject and of value to this reader." The tools which describe the subject con- tent of libraries may obviously achieve high significance when tjhat content itself is sig- nificant. There are special collections of which the catalogs approximate exhaustive bibliographies. There are also rigorously weeded collections of which the catalogs are valuable selective bibliographies. But most book collections are ei~her too small to be exhaustive or too great to be selective, and no collection can be both. But, meanwhile, the limitations of no single collection, or group of collections, need be imposed upon subject lists and subsequently upon the reader. On the con- trary, the subject list is the one device which can and should be used to help the reader, as well as the librarian, to surmount precisely those limitations. Subjects for Study 4· The headings in subject lists should re- fer to subjects for study, not merely subjects of books. In most existing subject tools, including library catalogs and many bibliographies, the SEPTEA-!BER_, 1945 headings point to the subjects of the books listed under diem. Under "U.S.-History- Colonial period," for example, are listed books about that period of American history. In othe.r tools the headings represent, in a broader sense, subjects which the reader may wish to study. Under those headings are then listed not only materials which are, strictly speaking, about those subjects but also others which are necessary to the study of the sub- ject. For example, under "U.S.-History- Colonial period" might be listed, in addition to the histories about the period (i.e., the secondary sources), significant social and political works of the period (i.e., the primary sources). Or, . for a student of The Dunciad, a catalog or bibliography might list under that heading the texts of the work itself and primary sources relating to its composition and publication-stationer's records, letters, etc.-as well as the history and criticism of the · poem. Books, in other words, may bear different relations to subjects, especially to literary, historical, and social subjects, and the . list which places under a heading only the mate- rials which are specificaliy about that subject may fail to guide the reader to other useful, if not essential, sources. uSubject-to-Book" Method 5· Most effeciive in the compilation of sub- ject lists is the usubject-to-book" process, not the "book-to-subject" proces-s. There are at least two different ways of compiling subject lists, and these are anti- thetic. The book-to-subject process is char- acteristic of library cataloging and indexing, wherein the compiler takes a book in hand and lists it under subjects where it may be found useful. As book after book is processed, the number of entries in all lists, or under all subjects, grows by accretion. The subject- to-book process is characteristic of "pure" subject bibliography, wherein the compiler takes a subject in hand and lists under it all books which may be found useful. As subject after subject is processed, the number of lists, instead of entries, grows by accretion. Both methods have advantages, but those of the subject-to-book process seem the more important. First, the subject-to-book process offers a way of meeting important needs first, by dwelling on crucial subjects, letting others wait for time and circumstance. In the 421 book-to-subject process, the controls relate primarily to forms of materials and the flow of new acquisitions into the library, often with the resuh that urgency goes begging. Second, the subject-to-book process brings the compiler into closer rapport with the library user. The compiler, like the reader, begins realistically with a specific problem and looks for books to solve it. His objective-to as- semble useful books about a subject-is always clearly before him. In the book-to- subject process, the objective is too often simply to subject-head the book or to classify the library; and rarely does the compiler, who works at a distance from the reader, stop dispersing books among subjects long enough to see how meaningfully they are assembling on the other side. And, third, it is easier to circumscribe the books relating to a subject than the subjects to which a book may relate. Books are tangible objects which can be seized upon and analyzed for relevance to a definite need, but the needs themselves are as different as people and as intangible as the imagination. Many people know the literature of some subject, but few know all the needs which any book will satisfy. It must be acknowledged, however, that the crude, preliminary sorting of books by subjects, as they first come from the press or the antiquarian's satchel, can take place only by the book-to-subject process. Hut this orig- inal subject indexing, whether done in library catalogs, in national or trade bibliographies, or elsewhere, is preparation only for the more intensive and critical work of the subjec:;t bibliographer. Character of System These propositions, if indeed they are valid, suggest the general character of the subject system which should be developed to comple- ment the union author catalog or finding list. On the one hand, they militate against any_ effort to construct union subject catalogs or union dictionary catalogs analogous to the present catalogs of individual libraries. Such catalogs, as commonly conceived, would be general purpose tools based on the content of libraries and compiled by the book-to-subject process. On the other hand, they clearly support the possibility of enlarging and sys- tematizing, in cooperation with scholars and bibliographers, the already indispensable re- sources of special subject bibliography. The success of the subject portion of Mr. Rider's proposed microcard catalog would depend largely upon the selection of materials which can be published in that form. If, as he hopes, complete coUections of the useful materials on special subjects can be ~issued, bibliographical significance would simultan- eously be achieved. This might well come to pass if the process were begun with the com- pilation of adequate lists of materials on selected special subjects and then with the publication of microc~rds for all those mate- rials. - The development of an adequate system of subject bibliography, whatever the form it takes, will require years of patient study and experiment. There are no short-cuts, no panaceas. First must come studies, not merely of books but of readers, and not merely of readers but of significant types of readers whose particular interests can be segregated and defined. These interests must then be translated into subjects, groups of subjects, and types of reading on those sub- jects. Then must come the search for books to fill those special needs and the publication of lists which can be recommended wherever those needs exist. For scholars in different fields there must be · comprehensive bibliog- raphies bearing directly on crucial problems and reflecting current theories and methods of research in those fields. For the farmer, the clubwoman, the ·industrial worker, the adolescent, or whoever else may display sig- nificantly different interests, there must be other lists, many of which may depart wholly from the academic subject categories of the scholar. Such an objective, to say the least, is ambitious and one not soon to be realized. But in the long view the results would be richly; rewarding. And it is an objective which, if it is ever to be achieved, will exact the utmost of cooperative eff~rt from librar- ians, scholars, and bibliographers alike. 422 COLLEGE AND RESEARCH LIBRARIES