College and Research Libraries Two ARL Approaches to Counting Holdings of Research Libraries A T THE TWENTY-FIRST MEETING o f t h e Association of Research Libraries ( A R L ) held in N e w York City, March 2, 1944, R o b e r t B. Downs brought u p the need f o r more u n i f o r m standards of sta- tistics of library holdings. H e believed a committee should be a p p o i n t e d to at- tempt to standardize such statistics. In the discussion that followed, D o n - ald Coney expressed the view that titles were m o r e important than volumes in determining holdings. Errett W . McDiar- m i d proposed that processed volumes be counted, since many volumes are neither cataloged nor accessioned. Downs replied that the number of volumes organized and ready for use should be the criterion. Henry B. Van Hoesen questioned the need for u n i f o r m statistics o r the real value of comparing them. A m o t i o n was then passed to a p p o i n t a committee to at- tempt to develop standard practices in statistics.1 A Committee on Statistics of Library Holdings was appointed in January 1945, with Downs as chairman. It made its first report at the twenty-third meeting of A R L on June 22, 1945 in N e w York City, r e c o m m e n d i n g that library holdings and annual additions be reported in terms of bibliographical items. A f t e r some discus- sion, Downs m o v e d the a d o p t i o n of the report. His m o t i o n was seconded and it passed. W h i l e the report was adopted, the A R L Minutes d o not indicate f o r what purpose it was adopted but it was sug- gested that the report be sent to member 1 A R L Minutes 21:16. This article is based primarily upon the action of the Association of Research Libraries as reported in the Minutes of its meetings. A positive microfilm of the Minutes for meetings 1 through 42, 1932-54, is avail- able from the Microreproduction Laboratory of the Massachusetts Insitute of Technology for $10. By A . F. K U H L M A N Dr. Kuhlman is Director, Joint University Libraries, Nashville, Tennessee. institutions of A R L f o r their study and advice. It was realized that there were real difficulties in changing statistics, both because it w o u l d upset comparative sta- tistics of past years and because it w o u l d be expensive to recount. T h e executive secretary, Paul N o r t h Rice, was directed to send at least ten copies of the report to every member institution of A R L . 2 T h e result of this action was that many libraries sent in criticisms and sugges- tions and the Downs committee revised its report and presented it at the twenty- fourth meeting in Chicago, December 29- 30, 1945. Downs admitted that virtually all libraries c o m m e n t i n g on the recom- mendations indicated that they w o u l d not make the system retroactive but c o u l d put it into effect f o r current acquisitions. H e also stated that "consid- erably m o r e than a majority of the As- sociation's members answered the Com- mittee's questions, some of them sending detailed comments and criticisms deal- ing with the preliminary recommenda- tions. W i t h this additional background, the Committee believes the subject has been adequately explored, and that the proposals it is n o w prepared to offer are practicable, reasonable and will b e gen- erally adopted, if approved b y the As- sociation." 3 T h e report was then accepted with the understanding that the chairman of the committee consult with the Library Serv- ice Division and with the A L A Commit? tee on Statistics and that he then have 2 A R L Minutes, 23:7, 19-20. 3 A R L Minutes 24:8, 16-18. M A Y 1 9 6 0 207 the report published in a library periodi- cal. T h i s was d o n e in an article in the Library Quarterly, January 1946. In his article D o w n s went b e y o n d the report of his A R L committee. H e de- veloped a helpful statement on difficul- ties involved in counting library hold- ings. H e acknowledged that the most widely used system of counting holdings was by physical volumes and that to change established routines and apply new rules retroactively w o u l d be an un- dertaking of great magnitude, particu- larly for large libraries. H e discussed the merits and limitations of three methods of measuring library holdings: the physi- cal v o l u m e count, the bibliographical unit count, and measuring linear feet of materials o n shelves. In discussing the unit for counting he used the definitions of a v o l u m e adopted by A L A and the United States Office of Education and a m o r e specific definition by R a n d o l p h G. Adams, stressing that a volume is any bibliographical item with a title or title page which is fully pre- pared for use. Accessibility was stressed as a criterion in the definitions, and by Downs, as the prime factor in counting volumes. C o u n t i n g should be confined to materials intended to f o r m part of the permanent research collection. Downs favored counting by biblio- graphical items. H e r e c o m m e n d e d that in counting multiple items b o u n d be- tween two covers o n e should record as a v o l u m e any item which has a title o r title page of its own and which w o u l d be counted as a v o l u m e if b o u n d separately. But he was aware of the danger of " p a d - d i n g . " T o avoid it, a n u m b e r of items b o u n d between two covers p r o b a b l y should not b e regarded as separate bib- liographical units if they constitute a connected series. T h u s , " t o count every d o c u m e n t in the collected edition of a government's publications as a biblio- graphical unit w o u l d swell total figures for library holdings to almost astronomi- cal proportions." Downs also included a discussion of •some of the factors responsible for lack of uniformity in statistics of holdings of different libraries. A separate count of important non-volume material by type was recommended, such as: manuscripts, microproductions, sound-recordings, mu- sic scores, maps, and prints. T h e committe's recommendations for counting were summarized at the end of Downs' article. At the twenty-fifth meeting of A R L , June 19, 1946 in Buffalo, N e w York, Downs reported that he had met the di- rective of the Association of contacting the Library Service Division and A L A Committee on Statistics, that he assumed his report was n o w officially adopted by the A R L and he expressed the h o p e that members of the Association w o u l d put the committee's recommendations into practice, insofar as feasible.4 At the twenty-sixth meeting of A R L , December 29, 1946 in Chicago, its execu- tive secretary, Paul North Rice, reported he had received inquiries about h o w many A R L members had put into oper- ation the m e t h o d of keeping statistics r e c o m m e n d e d by the A R L committee. T h e chairman then asked the g r o u p h o w many had adopted the new plan. Repre- sentatives of three libraries—Illinois, In- diana, and the Library of Congress—in- dicated their libraries had d o n e so. T h e r e u p o n , D o w n s was requested to make a survey of A R L members to de- termine h o w many had adopted the scheme.5 At the twenty-seventh meeting of A R L in Washington in March 1947, Downs presented by title only his report of the A R L Committee on Statistics of library holdings and it was reproduced as an ap- p e n d i x to the minutes of the meeting. T h e inquiry made by the committee was answered b y thirty libraries—two- thirds of the A R L membership. A b o u t one-half of those replying appeared to 4 A R L Minutes 2 5 : 6 . 5 A R L Minutes 26.6. 208 C O L L E G E A N D R E S E A R C H L I B R A R I E S be following all, or a substantial part of the committee's recommendations. T h i s group included most of the largest re- search libraries in the country. Most of the other libraries replying stated that their statistics were based upon accession records, i.e., the physical volume count. T w o difficulties were reported to a gen- eral adoption of a count of holdings by bibliographical items: (1) Libraries would have to make a retroactive count of their entire collection to be consistent, and they could not afford it; and (2) the committee's recommendations would cur- rently add to the cost of compiling sta- tistics because they called for more com- plete records. A majority of those replying, said Downs, regarded uniformity in statistics of library holdings as desirable, but many doubted its feasibility. Downs concluded: "Obviously, statistics will mean little unless agreements can be reached on some common rules. As time goes on, li- braries following different practices will go farther apart rather than closer, and will n o longer be comparable." 6 At the thirtieth meeting of A R L in Chicago, January 30, 1948, the execu- tive secretary, Charles W . David, urged the reopening of the question of how to count library holdings since it had not been settled to the satisfaction of all con- cerned. Downs stated he was not adverse to having the matter reopened and he moved that the A R L Committee on Sta- tistics be reconstituted with a new mem- bership. T h e motion was seconded and carried unanimously.7 At the thirty-first meeting of A R L , June 11, 1948 in Philadelphia the new committee, under the chairmanship of Guy R . Lyle, was authorized to devise and recommend some simple method of recounting book stocks and to report back to a later meeting of the Associa- tion.8 Lyle's committee on Counting Library 0 A R L Minutes 27:13,26. 7 A R L Minutes 30:9. 8 A R L Minutes 31:18. Holdings presented its report, dated January 13, 1949, at the thirty-second meeting of A R L on January 20, 1949 in Chicago. H e asked that his committee be discharged, but it was pointed out that so important a report deserved far more careful consideration than could be given to it at that meeting. H e was urged, and he agreed, to permit the com- mittee to remain in being at least until the next meeting when it was hoped the report could be given careful considera- tion and important decisions could be made.9 T h e committee had recommended the "physical count," in preference to the "bibliographical unit," and had worked out rules for the former method. O n March 3, 1949, the executive secretary, Charles W . David, distributed the report of the Committee on Counting Library Holdings to the membership and trans- mitted Lyle's request that at the Cam- bridge meeting of A R L on March 31, 1949 members should express a prefer- ence for one of the two methods. Mem- bers unable to attend that meeting were to send their vote by mail to the execu- tive secretary in advance of the meeting. T h e report of the Lyle Committee on Counting Library Holdings described briefly why it had decided to recommend the physical volume count. In May of 1948, the committee inquired of seventy- five libraries, including all A R L mem- bers, as to which of three methods of counting holdings they preferred. Re- plies from fifty-nine libraries were re- ceived—twenty-three favored a count by bibliographical unit, thirty-two a count by physical volume and four a count by piece. Only twenty-eight A R L members replied. Of these, eleven favored a count by bibliographical unit and seventeen a count by physical volume. In its preliminary deliberations the Lyle Committee was struck by two things: (1) N o one, apparently, had ever bothered to establish clearly the rides " A R L Minutes 32:32, 55-59. M A Y 1 9 6 0 2 0 9 for counting by physical volume such as the Downs committee had done for counting by bibliographical unit. (2) T h e committee doubted seriously whether any change would make for greater uniform- ity in counting for libraries beyond a certain size. " I n view of the variety and complexity of materials received by a large research library, the committee felt that no concept of uniformity in count- ing could be more than ideal." Since the relative merits of the two systems could not be determined until the rules for counting by physical vol- umes were formulated, the committee drew u p a plan for counting by physical volume. After studying both methods of count- ing the committee concluded that count- ing by physical volume is preferable to counting by bibliographical unit be- cause: (1) Most libraries are now com- mitted to a count by physical volume. A change to a retroactive count of biblio- graphical units would be burdensome and expensive. (2) T h e count by physical volume can incorporate many of the good features of the bibliographical unit method of counting without exaggerat- ing or inflating statistics. T h e committee in a sub-appendix showed the difference in count of eight titles selected at ran- dom. These, counted bibliographically, totaled 141 units but counted by physi- cal volumes they represented only 19 vol- umes. (3) T h e physical volume count is easier than counting by bibliographical unit. (4) Counting by physical volume is simple and inexpensive to administer. A t the thirty-third meeting of A R L , at Cambridge on March 31, 1949,10 the executive secretary, Charles W . David, opened the discussion of the report of Lyle's Committee on Statistics by recall- ing that in March 1947 the A R L had voted its approval of a method of count- ing library holdings by bilbiographical units rather than by physical volumes. Many libraries, however, had not ac- 1 0 A R L Minutes 33:11-14. cepted this decision, and there had been numerous protests. T h e r e u p o n a new committee had been appointed under the chairmanship of Mr. Lyle and its report had been distributed to A R L members at the preceding meeting in January. Lyle had urged that a formal vote of the Association be taken as to method of counting library holdings. In the discussion that followed, Downs, chairman of the earlier committee which had recommended counting by biblio- graphical unit, said "that he had once thought uniformity possible but that he had become disillusioned on this subject and believed that no action taken here would have much effect." He thought one more expression of preference would be futile and therefore moved that the re- port of the committee be accepted and that the committee be discharged with thanks. His motion was voted d o w n — 16 to 9. T h e r e u p o n letters were presented that had been received by Lyle. G. Flint Purdy, who had been chairman of the committee of A C R L which had annually compiled statistics for college and univer- sity libraries, said that he thought the Lyle committee had done an extraordi- narily good j o b and that the method of counting by physical volumes (rather than by bibliographical unit) as recom- mended by Lyle's committee seemed to him " t o be about as far as we can go at the moment in establishing a standard and practicable means of measuring the contents of libraries." H e and Ralph M. Dunbar (letter to Lyle) of the Library Service Division suggested methods for refining the committee's recommenda- tions of counting by physical volumes. A vote taken on the two methods of counting showed that A R L members stood as follows: T w e l v e favored count- ing by bibliographical unit and twenty- nine favored counting by physical vol- ume. It was suggested that A R L members, in reporting for the annual Princeton 210 C O L L E G E A N D R E S E A R C H L I B R A R I E S statistical compilation, hereafter indicate which method of counting they have used. T h i s recommendation was unani- mously approved, but unfortunately has not been observed by all libraries follow- ing the bibliographical count. T h e re- port of the Lyle Committee was pub- lished in the January 1950 issue of CRL, pp. 69-72. T h u s the Association of Research Li- braries has wrestled with the problem of counting holdings of research libraries in the work of two able committees: the Downs committee, favoring counting of volume material in terms of bibliograph- ical items, plus separate counts for vari- ous types of non-volume material; and Lyle's committee, favoring the physical volume count. Each committee has sup- plied definitions and rules for the method of counting it favored. With both plans before it, the Association voted 29 to 12 favoring the physical volume count. Eleven years have passed since this vote was taken and they seem to have proved that Downs was right when he said at Cambridge in March 1949 that he be- lieved any action taken by A R L would have little effect in producing uniform- ity in counting holdings. Now, in 1960, we seem to be reaching a situation that Downs warned against at the A R L meeting in Washington in March 1947 when he said, "statistics will mean little unless agreements can be reached on some common rules [for counting holdings]. As time goes on, li- braries following different practices will go farther apart rather than closer, and will n o longer be comparable." W e may well have reached that state already. If the university libraries that have reported their holdings in terms of bib- liographical items had only earmarked them as such, that would have helped somewhat. But it still would not have told by what percentage the number of physical volumes in a given library had been inflated, whether by 20 per cent, 30 per cent, 40 per cent, or what. Whether we like it or not the size of university libraries has become a factor in institutional rivalry in attracting top- flight faculty members and graduate stu- dents. T h a t is one good reason why sta- tistics should be made as comparable as possible. T w o sources seem primarily responsi- ble for an inflationary count when the bibliographical unit is used: 1. In the rules for counting by biblio- graphical unit the Downs committee recommended at the December A R L meeting in Chicago in 1945 that: " I n counting multiple items b o u n d between two covers, record as a volume any item having a title or title page of its own, and which would be counted as a volume if b o u n d separately, i.e., base statistics on bibliographical units."1 1 T o follow this rule in counting monographic ma- terial and a great mass of official govern- mental publications would result in seri- ous padding. T h u s for instance, at the Joint University Libraries the Hearings of the 85 th Congress have been assem- bled, bound and counted in 185 physical volumes. But to apply the bibliographi- cal measuring rod would swell the count to at least 1,085. 2. Equally serious inflation of count- ing holdings arises from counting micro- prints and microcards each as a volume, for in many cases it requires many micro- cards to reproduce a single physical vol- ume. It is n o wonder that at the recent A R L meeting in Chicago on January 27, 1960 the urgent need for uniform policies in counting library holdings was stressed by Jens Nyholm. N o action was taken by the A R L group because it was thought ac- tion should be deferred until the A L A Statistics Coordinating Committee makes its report. It is to be hoped that that Committee will produce standards that will provide greater uniformity in count- ing library holdings. 11 A R L Minutes 24:16-17. M A Y 1 9 6 0 211