College and Research Libraries THOMAS J. WALDHART and LEROY G. ZWEIFEL Organizational Patterns Of Scientific and Technical Libraries: An Exainination Of Three Issues Three aspects of tlte problen~ of library reorganization are consid- ered with specific reference tq scientific and technological libraries: the· politics of centralization; the concept of accessibility; and the in- teraction of science and technology. Although these aspects are cen- tral to library reorganization, their impact on the structure of univer- sity scientific and technical libraries has yet to be understood in order to provide a rational basis of decision making by university adminis- trators, librarians, and library users. INTRODUCTION THE DIVERGENT PATTERNS OF ORGANIZA- TION which university libraries have as- sumed over the years, ostensibly to meet the information needs of their users in the most efficient and effective manner possible within resource limitations, rep- resent .a topic of continuing interest to university administrators, librarians, and library users. The Association of Research Libraries ( ARL) sponsored study, Problems in University Library Management, summarized some of the basic issues related to the organizational structure of university libraries as fol- lows: Librarians are caught between con- flicting pressures for centralization and decentralization of collections and, Thomas ]. W aldhart is assistant profes- sor, College of Library Science, University of Kentucky, Lexington; LeRoy G. Zweifel is professor, Engineering Sciences Library, University of Wisconsin, Madison. 426 I consequently, facilities. University ad- ministrators desire to hold duplication of collections and dispersal of services to a minimum. Faculty and graduate students press for decentralized de- partmental libraries. There appears to be a growing trend toward the estab- lishment of large, decentralized librar- ies covering several major academic fields, such as the life sciences. The more established trend of constructing separate facilities for undergraduate and graduate students continues. The issue of centralized versus de- centralized facilities poses major man- agement problems for university librar- ians. In planning new construction and considering changes in existing space utilization, the library must de- cide whether it is more efficient and effective to decentralize or to central- ize operations. Librarians indicate that little data are available to assist them in making such decisions. 1 Whether librarians, university admin- istrators, or library users are "caught be- tween conflicting pressures for centrali- zation and decentralization" depends upon their individual perspectives. Al- though these groups should participate in decisions relating to the organization- al structure of university library sys- tems, they are not equally involved in the making of decisions. In most cases, as university administrators control the resources necessary to accomplish major modifications in library systems, they play the major role in decision making. McAnally and Downs noted the appar- ently declining influence of university librarians and libraries in terms of their .ability to participate in high level de- cisions.2 In actuality, the extent to which university librarians can influence major library or university decisions de- pends largely on the historical back- ground and present conditions at a given university. If the university li- brarian has the respect and confidence of the administration and faculty he can be expected to have considerable in- fluence and may even provide leader- ship. If he lacks this respect and con- fidence or if major decisions related to the library system have traditionally been made by the faculty and/ or the ad- ministration, authority and responsibil- ity for decision making may shift from the library environment, with the uni- versity librarian primarily engaging in matters of implementation . . Although not explicitly stated in the ARL study, much pressure for decen- tralization of university library systems originates with the faculties and stu- dents of scientific and technical fields; thus, the organizational structure of sci- entific and technical libraries provides the focus for the following discussion. The commonly assumed advantages and disadvantages of centralized or de- centralized scientific and technical li- braries in university environments have been repeatedly discussed in the litera- ture of librarianship. 3 Suffice it to say that most university library administra- tors, and students of librarianship in general, oppose forms of subject decen- Organizational Patterns I 427 tralization which extend to the depart- mental level. They argue that such an organizational pattern: ( 1) generates numerous problems of communication, control, and coordination; ( 2) results in rising cost through the duplication of library services, materials, personnel, and records; and ( 3) encourages, wher- ever materials are not duplicated, the fragmentation of knowledge and the isolation of scientific and technical fields. The trend at many universities to organize scientific and technical libraries around a number of closely related aca- demic fields (e.g., life sciences, physical sciences, engineering sciences, etc.), may result from: ( 1) an attempt by univer- sity administrators and librarians to re- alize economies of scale (i.e., cost sav- ings in physical plant and operating ex- penses due to increased size); ( 2) recog- nition of the potential benefits of in- creased accessibility to users; ( 3) a com- promise between factions favoring greater centralization (e.g., a single sci- entific and technical library) and those favoring greater decentralization (e.g., departmental libraries ) ; ( 4 ) a desire to serve more adequately the information needs of interdisciplinary instructional programs and research; or ( 5) a com- bination of the above. Faculty members of academic depart- ments, in contrast to library administra- tors, often argue for subject decentral- ization to the departmental level. They contend that their work (primarily their research activities) requires unlimited access to library materials over time, and in addition that library materials should be physically situated in close proximity to the greatest number of potential users. Particularly with respect to scien- tific and technical libraries, they main- tain these factors should supersede what they view to be the essentially economic arguments of library administrators. Caught between the persuasive argu- ments of library administrators on the one hand and those of faculty on the 428 I College & Research Libraries • November 1973 other are university administrators, who must allocate university resources among competing demands. As a reorganization of a university library system represents a long-term commitment, decisions made today may severely limit future options. Centralization of library sys- tems almost certainly entails the initial expenditure of substantial sums of money in developing the physical plant to house the consolidated collection and the staff to provide services. Change in general is costly, and in a time of short money lack of adequate financial re- sources has a strong tendency to rein- force the status quo. Thus, unless uni- versity administrators are provided evi- dence of substantial savings in operat- ing expense or improvement in service, it is unlikely that proposals for major reorganization will receive enthusiastic administrative support. Although any long-range planning must be partially based on "soft" infor- mation (e.g., subjective judgments, pro- jections of student enrollment and uni- versity growth, changeable building pri- orities ) decision makers would like to have as much accurate or "hard" infor- . mation as possible regarding the future consequences of alternative courses of action. Unfortunately, most arguments are based exclusively upon the subjec- tive judgment of individuals possessing a strong bias either for or against cen- tralization. The profession's inability to accumu- late this objective information results fro:m a combination of factors. First, any attempt to evaluate the efficiency, effectiveness, and/ or benefits derived from a particular pattern of library or- ganization is at best a costly, time-con- suming, and in the end, somewhat sub- jective process. For instance, Evans and Borko identified a range of criteria which seem to hold some potential for accessing the performance of libraries; included are library coverage, response time, accessibility, cost, use, user satis- faction, and a number of combinations of each. 4 The selection of evaluative criteria represents a complex process. Al- though some are relatively easy to use (e.g., library coverage or response time), others are extremely difficult to opera- tionalize (e.g., cost-benefits or user satis- faction). Reducing the performance of an entire library system to a limited number of criteria runs the additional danger of underestimating the complex- ity of most library systems. The selec- tion of cost, for example, as the major performance criterion would probably be unacceptable to the majority of li- brary users, particularly if the study emphasized system cost over user cost or benefits. 5 Second, although levels of perform- ance could conceivably be determined for a given pattern of organization, it is unlikely that most library systems could examine the performance of al- ternative patterns of organization. Such an approach would require either the restructuring of the library system for experimental purposes, or perhaps the development of a complex model for simulating the alternative systems. In either case, excessive cost tends to reduce the viability of such an approach. Finally, where it is possible to gener- ate reliable information which may sup- port decision making, wide variations in local circumstances restrict its use to a given library system at a .Particular point in time. Some dangers inherent in evaluating the performance of varying patterns of library organization by com- paring different library systems, for in- stance a centralized system at university X with a dec~ntralized system at univer- sity Y, have already been discussed in de- tail by Richmond. 6 For the immediate future there ap- pears to be little likelihood that univer- sity administrators, librarians, or library users will possess the ability to predict accurately the level of benefits to be de- rived from given patterns of library or- ganization; too little is known about the relationship between library organiza- tion and system performance. Yet with- out this knowledge individuals are re- duced to relying exclusively on intuition or an appeal to authority to serve as a basis for decision making. Because of their potential impact, three aspects of the problem of library reorganization which have received lit- tle attention will be considered: the pol- itics of centralization; the concept of accessibility; and the interaction of sci- ence and technology. These topics will be examined to illustrate their impor- tance to the subject at hand and to indi- cate specific aspects of the problem which seem to deserve additional study. PoLITics OF CENTRALIZATION It has been observed that «university professors know more about everything else than they know about themselves and their habitat."7 Although knowl- edge of the academic system has in- creased appreciably over the last five years, information related to the role of the university library in that system -particularly the role of the university librarian in the social structure of the university-remains fragmentary. Two recent exceptions to this generalization are a paper by McAnally and Downs, and an interview with Robert A. Miller, reported by Lyle. 8• 9 Both provide in- sight into the complex social system of which the university library is an inte- gral part and illustrate the importance of the political realities which directors of university libraries must deal with on a continuing basis. Although the eco- nomic aspects of centralization receive emphasis in the literature of librarian- ship, the political aspects of centraliza- tion may really determine negotiations for changing library systems. In universities with traditional de- partmental decentralization of scientific and technical libraries, faculties have tended to be relu~tant to give up c'their" Organizational Patterns I 429 libraries. They cite the need for imme- diate physical accessibility to library ma- terials. Yet, a more compelling, although unstated, reason may be human resist- ance to ,change, and the possible conse- quent loss of administrative control. Where departmental libraries exist, faculties tend to retain considerable control over matters of library policy and procedures: library hours, avail- ability of library keys, duplication and selection policies, selection of the de- partmental librarian, establishment of library services, etc. As expected, prefer- ential treatment is often accorded fac- ulty members and graduate students of the department; library users from oth- er departments on campus frequently find these libraries, and their services, considerably less accessible. With cen- tralization, departmental faculties might lose this control over the operation of the library as well as their preferential treat- ment. Centralization confers most responsi- bility and authority for decision making to the office of the director of the cen- tralized facility, rather than upon the faculties of the individual departments. The librarian, rather than focusing on one department, must balance the needs of several departments; he must balance the needs of undergraduates against those of graduate students and faculty; the needs of research against those of instruction. Specifically, the librarian of a centralized facility must consider the scientific and technical information needs of the total academic community subject to limited resources. Although general improvement in ser- vice to the overall academic community might reduce service to specific depart- ments, centralization has considerable potential for improving library service to individuals throughout the academic community. The availability of in- creased resources ( physical, financial, and human) to centralized systems pro- vides an opportunity to develop library 430 j College & Research Libraries • November 1973 services which are generally not avail- able in highly decentralized systems-se- lective dissemination of information, systematic collection development by subject specialists, increased access to subjects of peripheral interest, im- proved facilities for study and research, exploitation of nonprint media, and document delivery services. In addition, it provides improved access to library materials for the rapidly growing num- ber of educational and research pro- grams which are inter-, multi-, or trans- disciplinary in nature. Too often, however, promises of im- proved library service made by librari- ans are not matched by performance, and most scientists and technologists in university environments are aware of this pahtful fact. Many faculty mem- bers have had an opportunity to make use of the university's central library (which represents, in their minds, an ex- ample of centralization). They have ob- served that the library was often over- crowded, extremely large, difficult to get to and once there, difficult to use, im- personal, possessed of loan policies which were restrictive, and stated by in- dividuals having little appreciation of their particular informational needs. The difference between promised and observed performance sometimes gen- erates a credibility gap of monumental proportions. The ability of the library administra- tor, usually the university librarian, to bridge this gap will mainly determine the extent of departmental faculty sup- port, which seems necessary to be effec- tive. Nicholson observed: Centralization will be successful . . . only if complete agreement is reached by faculty, university and library ad- ministration that it is the best way in which the greatest number can be ef- fectively served ... under reasonable financial expenditures.lO Most scientists and technologists are less concerned with how the "greatest number can be effectively served" than they are with how they can be more ef- fectively served. If a library adminis- trator cannot secure the active support of the faculty in such a venture, he must do everything possible to reduce active opposition. One way of measur- ing faculty reaction to centralization is through faculty response to questions of the following nature: 1. Does the library administrator have an understanding and appreciation of the information needs of scien- tists and technologists? Has this ap- preciation been demonstrated by his past interest in, and support of, the development of the scientific and technical libraries on campus? 2. What is his record with regard to library matters of primary concern to the scientific and technical com- munity within the university? 3. Will centralization mean a reduc- tion in library service to me? Can the library administrator be trusted to provide the kinds of services that he promises? Does he have the "power" within the university to deliver on such promises? 4. As the system grows in size and complexity can assurances be given that we will continue to be treated as individuals with individual in- formation needs? Without a perceived record of sensitiv- ity to departmental needs, the university library administrator will have to pro- vide sufficient guarantees that centraliza- tion will, in fact, result in improved li- brary services for scientists and technol- ogists. Little objective data exists on the role of the university librarian in the social structure of the university. Available in- formation primarily considers the rela- tionship of the director of libraries to the library committee (which may or may not be representative of the facul- ty of the university as a whole), library staff, students, and sometimes the uni- versity administration. Little is known · ~ about the attitudes of departmental faculties, particularly in the sciences and technology, toward the university librarian. In negotiations on the restructuring of the university library system, the de- gree of empathy displayed by the par- ticipants can be critical in determining , a program's success or failure. A brief survey of the backgrounds of the direc- tors of the twenty-five largest university libraries in the United States (and thus those thought most likely to be facing the question of centralization or decen- ~~alization) revealed that none possessed a Ph.D. degree in either the physical sci- \ ences, engineering, or the life sciences. In addition, twenty-two of the twenty- , five directors possessed backgrounds in the humanities or social sciences. 11 This considerable divergence in backgrounds might have a substantial impact on these negotiations. Sound sociological research is essential before an adequate understanding of the role of the university library in the academic community can be realized. This understanding is not only impor- tant for decision making, but would also contribute greatly to the establish- ment of realistic library goals, and ex- pectations of library service, in the aca- demic community. CoNCEPT OF AccESSIBILITY Library administrators tend to be less than sympathetic toward demands of science and technology graduate stu- dents and faculty for maximum acces- sibility (both physical and intellectual), because to increase the user's accessibil- ity almost certainly will result in an in- crease of operating costs. Library ad- ministrators do not have to defend user costs (i.e., the physical, psychological, and economic cost which users incur in interaction with the system) at budget- Organizational Patterns I 431 ary meetings, but they do have to de- fend library operating costs. An inter- esting example of the conflict between user cost and system cost is described by Dougherty in an evaluation of a docu- ment delivery system at the University of Colorado.l 2 Until university and li- brary administrators recognize that re- ducing library costs often increases user costs (and indirectly university costs), most university library systems will con- tinue to shift the cost burden to the user. Nevertheless, university library ad- ministrators have the obligation to pro- vide the most effective and efficient li- brary service possible within existing budgetary constraints. No matter how efficient a centralized library may be on the basis of library costs, it cannot be effective unless it is used by those for whom the system was designed. The ex- tent to which a library, or any informa- tion system is used, depends in large part upon the extent to which potential users perceive the system as being acces- sible. Scientists and technologists frequent- ly have emphasized the importance of having documents immediately acces- sible throughout the course of their teaching and research. Nevertheless, lit- tle is known when a library or informa- tion service becomes so inaccessible as to be not used. Defining accessibility in terms of the economic, psychological, and physical cost associated with using a given com- munication channel, Allen found that, for the groups of engineers studied, ac- cessibility was a critical variable related to the extent to which a communication channel was used. 13 Further, he conclud- ed that: Improving the quality of performance of a particular information service (or system) will not in and of itself, lead to increased use of the service. Before the improved information service can lead to increased performance, it must 432/ College & Research Libraries • November 1973 j be used: And the only way to increase of familiarity with the system is un- 1 use is through increased accessibility.l4 known. Obviously, the extent to which At least three dimensions of accessi- time, distance, and familiarity affect an bility are clearly identifiable: distance, individual's perception of accessibility time, and familiarity. With distance, depend on various conditions: ease of for instance, is a library in a building use; motivation of the user; his particu- next door inaccessible? Is one down the lar information needs; climatic condi- block inaccessible? How about one tions; the physical condition of the across campus? Although Raffel and user; and many other factors which are Shishko provide a methodology, based difficult or impossible to anticipate. on location theory, which may provide Although the concept of accessibility library administrators with some guid- is central to most arguments for the de- ' ance in the location of centralized or centralization of university scientific decentralized library systems, the meth- and technical libraries, and is recognized od provides little hope of resolving the by library administrators as an essential essentially psychological problem relat- to effective library service, it remains a ed to the point at which distance makes poorly understood aspect of the commu- a library inaccessible to the point of nication process. As such, it should be nonuse.15 used with caution as a measure of sys- Similarly, the amount of time re- tern performance for basing decisions quired to retrieve a document may in- on library organization until a syste- fluence a user's perception of accessibil- matic examination can be made. ity. Are documents located in storage, and which require two days to retrieve, considered inaccessible? Are materials acquired through interlibrary loan, and requiring at least two weeks to retrieve, inaccessible in the eyes of the user? Are materials housed in a centralized li- brary, which require twenty minutes walking time from an individual's of- fice inaccessible? Allen found that the engineer's per- ception of accessibility was influenced by past experience. The more familiar an engineer was with a given communi- cation channel the more accessible he perceived it to be. 16 Thus, if an indi- vidual makes use of the university's cen- tral library once or twice a year he may perceive that library to be relatively in- accessible as compared with, for in- stance, a departmental library which he uses twice a day. Many individuals who have used centralized scientific and tech- nical libraries have reported that their experiences have been somewhat unsatis- factory.17 Whether this lack of satisfac- tion is a result of low performance on the part of the library or the user's lack INTERACITON OF SciENCE AND TECHNOLOGY The extent to which departmental de- centralization represents an effective and efficient form of university library or- ganization partially depends on the amount of crossover (i.e., "the degree to which those at one market or point of origin [e.g., academic department] use libraries at more than one loca- tion") which exists within the library system. 18 If crossover is negligible, a strong argument can be made for situ- ating the libraries close to the primary market (e.g., the academic department); if crossover is substantial, libraries should be placed so that they are most convenient to the total user population. The extent of users with multidiscipli- 1 nary interests, therefore, is important in the determination of organizational structure. In terms of centralizing scientific and technical libraries, the extent to which these disciplines interact, particularly through the literature, should be exam- ined. Existing knowledge relating to the interaction of science and technology is equivocal. Storer observed that "it has been increasingly the case that scientific advances are directly responsible for technological advances (and) it is only because of the increasing dependence of technology on scientific progress and the closer ties between science and the common universe of discourse, that we now speak of the 'practical' importance of science."19 In contrast, Price, follow- ing an analysis of citation patterns in scientific and technical literature, ob- served that the "interaction between sci- ence and technology seems to proceed only slightly and with great difficulty through the literature."2o In an analysis of the communication patterns of engineers in industrial en- vironments, Allen presents an appeal- ing, and apparently valid, explanation of why the average engineer makes lit- tie use of research-oriented literature. He stated that ''most of the profession- al engineering literature is too mathe- matically sophisticated for the average engineer to comprehend. It is therefore inaccessible to him."21 If the average engineer lacks the mathematical sophis- tication to interpret the professional en- gineering literature, in all probability he likewise lacks the ability to interpret the research-oriented literature of sci- ence. If this argument were valid for ·engineers in .academic as well as indus- trial environments, establishing a cen- tralized library would not seem justified. To study the relationship between the citation of scientific literature and the institutional affiliation of engineers, Waldhart performed a citation analysis of a selected sample of articles written by engineers. 22 Results of this analysis clearly indicated that, contrary to popu- lar opinion, engineers who publish tend to make extensive use of scientific litera- ture.23 The major exception to this gen- eralization was those source articles pub- lished in trade journals. These source articles generally lacked references, and Organizational Patterns I 433 were authored by individuals who lacked the Ph.D. degree and possessed industrial affiliations. Trade journals were not generally employed by engi- neers with academic affiliations either as a citation source or as a publication outlet. In addition, engineers with academic affiliations were found to be more sci- entific in orientation (they cited a sig- nificantly higher proportion of refer- ences from science), and more litera- ture conscious or dependent (they cited a significantly higher number of refer- ences per source article ) , than engineers with nonacademic .affiliations.24 Although W aldhart studied only the engineer's use of scientific literature, the extent to which scientists cross over and make use of technological literature is generally felt to be considerably less, although no recent study of this subject exists.25 More likely, crossover by scien- tists occurs primarily within the struc- ture of science itself, rather than be- tween science and technology. Some changes in the scientist's use of techno- logical literature can be expected with the recent modification of funding pol- icies of the federal governm·ent which tend to emphasize applied as contrasted with basic research, and the growth of multidisciplinary research institutes de- voted to the study of environmental problems. 26 Thus from the point of view of the engineer in academic environments, it would appear that engineers could bene- fit from the improved accessibility to scientific literature which would result from a centralization of scientific and technical libraries. However, it should be emphasized that the reorganization of university library systems should not be bas·ed solely upon such evidence. It represents only a "bit" of information which may be useful for understanding the complexity of any decision to cen- tralize or decentralize scientific and tech- nical libraries. 434 I College & Research Libraries • November 1973 CoNCLUSIONs The Association of Research Librar- ies stated that "librarians indicate that little data are available to assist them in making such d ecisions (centralization or decentralization) ."27 In this light, two basic conclusions seem inescapable: first, because of the uniqueness of local cir- cumstances it is unlikely that a "general theory" of library organization, which can guide the decision-making process, will b e formulated in the near future; second, if librarians need data to sup- port the decision-making process, it will fall to them to generate such data. D ecisions relative to the organization- al patterns which university library sys- tems will assume in the future can be made on a "crisis" basis, where decisions are forced , often prematurely, by im- mediate problems which require solu- tion; or, they may be made on the basis of careful, long-range planning which attempts to deal "systematically with fu- ture opportunities, problems and alter- native courses of action."28 It seems patently apparent that the second course of action is preferable. Implicit in this course is the need for more intensive and extensive research related to there- lationship between organizational pat- terns of libraries and their perform- ance. This paper identified three areas felt to be particularly deserving of at- tention-many more could be detailed. The library community, both practition- ers and those primarily concerned with research, must take the initiative in con- ducting such studies. Failure to accept this challenge can only lead to a fur- ther degrading of the role of librarians as active participants in university d e- cision making. REFERENCES l. Booz, Allen & Hamilton, Inc., Problems in University L ibrary Management (Washing- ton, D.C.: Association of Research Librar- ies, 1970), p.35. 2. Arthur M. McAnally and Robert B. Downs, "The Ch anging Role of Directors of Uni- versity Libraries," CRL 34:103- 25 (March 1973 ). 3. Mari anne Cooper, " Organizational Patterns of Academic Science Libraries," CRL 29: 357- 63 (S ept. 1968) ; Lawrence Thompson, " The Historical Background of Departmen- tal Collegiate Libraries," Library Quarter- ly 12:49-74 (Jan. 1942); D. A. Wells, " In- dividual D ep artmental Libraries vs. Con- solidated Science Libraries," Physics Today 14:40-41 (May 1961); J. H. Shera, "How Much Is a Physicist's Inertia Worth?" Physics Today 14:42---43 (Aug. 1961) ; Har- vey Marron, "Science Libraries: Consoli- dated or Departmental?" Physics Today 16:34-39 (July 1963); Jean Legg, "Death of the D epartmental Library," Library Re- sources and Technical Services 9:351-55 (Summer 1965); Robert R. Walsh, "Branch Library Planning in Universities," Library Trends 18:210-22 (Oct. 1969); Michael J. Bruno, "Decentralization in Aca- d emic Libraries," Library Trends 19:311- 17 (Jan. 1971) . 4. G. Edward Evans and Harold Borko, Ef- fectiveness Criteria for Medical Libraries: Final Report (Institute of Library Re- search, University of California, Los An- geles, April 1970, ED 057 813). 5. F. W. Lancaster, "The Cost-Effectiveness Analysis of Information Retrieval and Dis- semination Systems," American Society fo r Information Scienc e Journal 22:12--27 (Jan.-Feb. 1971). 6. Phyllis Richmond, "Systems Evaluation by Comparison Testing," CRL 27:23-30 (Jan. 1966 ). 7. Nicholas J. Demerath, et al., Power, Presi- dents, and Professors (New York: Basic Books, 1967 ) . 8. McAnally, "Changing Role of Directors." 9. Guy R. Lyle, The Librarian Speaking: In- t erviews with University Librarians ( Ath- ens, Ga.: University of Georgia Press, 1970), p.106-13. 10. Rutgers University, Graduate School of Li- brary Service, Studies in Library Service, Studies in Library Administrative Problems (New Brunswick, N.J.: Rutgers University, 1960), p.133-56. .. 11. The twenty-five largest university libraries were identified through The Bowker Annu- al of Library and Book Trade Information (New York: Bowker, 1970), p.16-17. Di- rectors of the libraries were identified through The American Library Directory, 1970- 1971, 27th ed. (New York: Bowker, 1971 ). Biographical information for the di- rectors was identified through A Biograph- ical Directory of Librarians in the United States and Canada, 5th ed. (Chicago: American Library Association, 1970) . 12. Richard M. Dougherty, "The Evaluation of Campus Library Document Delivery Service," CRL 34:29--39 (Jan. 1973). 13. Thomas J. Allen and Peter G. Gerstberger, Criteria for Selection of an Information Source (Alfred P. Sloan School of Manage- ment, M.I.T., Working Paper #284-67, Sept. 1967, PB 176 899), p.23. 14. Ibid., .19-=20. 15. J effre Raffel nd Robert Shishko, "Cen- tralization vs. Decentralization: A Loca- tional Analysis Approach for Librarians," Special Libraries 63: 13!>-43 (March 1972). 16. Allen, Criteria for Selection, p.23. 17. Wells, "Individual Departmental Librar- ies," p.40-4l. 18. Raffel, "Centralization vs. Decentraliza- tion," p.142. 19. Norman W. Storer, The Social System of Science (New York: Holt, 1966), p.2. Organizational Patterns I 435 20. Derek J. de Solla Price, "Is Technology Historically Independent of Science? A Study in Statistical Historiography," Tech- nology and Culture 6:553-68 (Falll965). 21. Allen, Criteria for Selection, p.20-2l. 22. Thomas J. Waldhart, "The Relationship Be- tween the Citation of Scientific Literature and the Institutional Affiliation of Engi- neers" (Unpublished Ph.D. dissertation, In- diana University, Jan. 1973). 23. Janice M. Ladendorf, "Information Flow in Science, Technology, and Commerce; A Review of the Concepts of the Sixties," Special Libraries 61:215-22 (May-June 1970); Donald G. Marquis and Thomas J. Allen, "Communication Patterns in Applied Technology," American PsychOlogist 21: 1052-60 (Nov. 1966); Price, "Is Technolo- gy Historically Independent of Science?" p.553-68. 24. Waldhart, "The Relationship," p.l74. 25. Herman H. Fussier, "Characteristics of the Research Literature Used by Chemists and Physicists in the United States," Library Quarterly 19: 19-35 (Jan. 1949 ) . 26. Mina Rees, "The Saga of American Univer- sities: The Role of Science," Science 179: 19-23 ( 5 Jan. 1973). 27. Booz, Problems, p.35. 28. George A. Steiner, "The Critical Role of Top Management in Long-range Planning," Arizona Review 14:5-13 (Aprill966).