College and Research Libraries 204 I College & Research Libraries • May 1970 zens will benefit enormously. Every city can and should have such a library as Dr. Martin envisions for Chicago, a "nerve center ... for contemporary information, in substance functioning as th~ fact bank, information switchboard, and special li- brary for the general populace."-Edwin Castagna, Enoch Pratt Free Library. Library Automation; a State of the Art Review. Ed. by Stephen R. Salmon. Chi- cago: American Library Association, 1969. 175p. $7.50. (73-77283). The papers presented at the ALA Pre- conference Institute on Library Automa- tion at San Francisco in June 1967 consti- tute this volume. The purpose of the insti- tute was to inform ALA members of the state of the art of library automation. It achieved its purpose, and with the prin- cipal exception of on-line applications de- scribed since 1967, it still constitutes an in- formative review for librarians not directly involved in research and development. Separate sections of the report are de- voted to acquisitions, cataloging, serials, and circulation, but the publication lacks an adequate review of information retriev- al. Necessarily lacking are descriptions of on-line systems in acquisitions, serials, and circulation that have been activated since 1967. Other sections discuss the MARC Project at the Library of Congress, networks, sys- tem analysis and design, and buildings. The MARC Project has had major devel- opments since 1967, which of course are not in Library Automation. On the other hand, system analysis and design is a time- less topic. One of the most interesting sec- tions is that by Robert H. Rohlf entitled "Building-Planning Implications of Auto- mation." This section does not give cook- book answers to those who wish detailed replies to the question "How will library automation affect the building I am plan- ning?" but it does give a valuable basis from which effective planning can proceed. Library Automation will be a useful and informative publication for some years to come.-Frederick G. Kilgour, The Ohio College Library Center. Cataloging U.S.A. By Paul S. Dunkin. Chi- cago: American Library Association, 1969. 159p. $5.00. ( 69-17830). Paul Dunkin has given us a brief survey of cataloging theory in the United States. He prefaces his book with an annotated list of the most influential writings on cata- loging; after which he summarizes the cat- aloging codes from Cutter's on. Then, un- der each problem area-entry, description, subject, classification, the catalog-he dis- cusses the major points of view and their theoretical bases. His expressed intention is to show why we catalog as we do. The categories, assumptions, and objec- tives of the transcendent theories are pre- sented with clarity. We see how we arrived at our current practices, that they do not form a coherent whole and reflect histori- cal not logical development. They are largely "the accumulation of what has been done in LC" ( p. 143) , a compromise of conflicting bibliographical objectives, par- ticularly of conflicting theories on "the pub- lic's needs and/ or wants. (They are not necessarily the same.)" We index the book collection both to lo- cate a work and to relate it to other works. That is our first principle. Cataloging at- tempts to do this systematically, and parts of Cutter's coherent but expensive system still stand. Parts have fallen under attack. But no matter how cogent or inviting later theories have been, the system has re- mained closed to any but peripheral and compromised changes, adopted usually for economic reasons and tending to make the system a less coherent whole. Mr. Dunkin shows us why we have arrived at our cur- rent practices. We all know what they are and what problems they raise in applica- tion and comprehension. Thus we enter works on "principles of authorship," not according to the title page statement the author and publisher have agreed on. Our forms of entry reflect wave after wave of opinion. We relate some types of material by added entries, others by uniform titles, and still others by form headings. Our sub- ject headings reflect a number of views on the uses of language, and a continual re- duction of attempts to apply them system- atically or to relate them fully. MARC finds it necessary to bolster our descrip- tions with explicit statements on such points as language of text, country of ori- gin, and index. Even the paging state- ment, shown to be most important in es..: tablishing editions, has gone wild with the acceptance of Title II descriptions. The catalog gets larger and more confusing. The attempt to tie cataloging at least physically to books was dismissed ten years ago in the Library of Congress' The Cata- loging-in-Source Experiment. This report, called by Dunkin "an amazing document," is one still deeply resented by catalogers outside the Library of Congress, who did not feel the experiment's pressures. Noth- ing since has promised immediate practical relief. Attempts to tie cataloging more log- ically or even more simply to books have added to the cost or to the confusion or to both. Mr. Dunkin has tried to limit himself to descriptive rather than critical analysis. The reader will be grateful to have the his- tory laid out concisely. This is an impor- tant book, intelligently done; if it emerges as a kind of epitaph to cataloging theory as we have known it, perhaps machines will someday release us and give us a chance at theories again.-Lois Hacker, Cornell University Libraries. Prolegomena to Library Classification. 3d ed. By S. R. Ranganathan, assisted by M. A. Gopinath. New York: Asia Pub- lishing House, 1967. 640p. (73-427373). It is with deep gratitude that I remem- ber my first encounter with the Prolego- mena. It (then in its second edition) opened my eyes with its clear statements of the problems of classification, as well as with its amazing revelation that anyone had gone so far toward their solution. This third edition is not a revision in the usual sense, but r ather a development of those parts of the second edition of the greatest generality, excluding much of the histori- cal, speculative, and practical discussions which (the author informs us) are being developed in two other books: Classifica- tion: Retrospective and Prospective, and D epth Classification and Its D esign. Thus the new Prolegomena consists, in a way, of three separate titles. Libraries ·With the Recent Publications j 205 second edition should not retire it to inac- tive storage unless they acquire all three new titles. If there is a work in which is concen- trated (and the word must be taken in a very strong sense) all that is most germinal in the theory of classification, it is the Prolegomena. Nothing else can rank with it except the 1876 Dewey and Cutter works, and perhaps the Gardin team's L'Automa- tisation des Recherches Documentaires. In this new incarnation it has become more than ever nothing but what-must-be-con- sidered-before ... , less a survey that in- cludes prolegomena! matter. No one (ex- cept the beginning student, who would in all but a very few cases be quite put off by the unaccustomed rigor of the mode of exposition and who would be in principle unaware of the aporia in the praxis that have led to this theoria) who is serious about understanding, constructing, apply- ing, or using any classification or system of indexing can afford to be uninformed about what Ranganathan works through here. The new edition would better have been (like the second) printed in En- gland; there are misprints in abundance, though most are not too serious-just ir- ritating. But there are a few weaknesses of a more serious sort. Interpolation (internal hospitality) in chain ( §LG) is not really explained, though Ranganathan along with everyone else assumes that Dewey's radix- fractional principle makes it possible. But it may instead be that only a faceted nota- tion does-and then only in a somewhat weak sense. Dichotomy is discussed in the proper pejorative light (§PC) , but its real function (positive/negative = enumera- tion/" others") is not mentioned. Figure 16 ( p. 367) is intended to show the complexi- ty of "the tree of knowledge"; it is so com- plex as to confuse, and the lack of explana- tory text makes it not a help but a hin- drance to the reader. UDC is made to seem to have Anteriorising Common Iso- lates (p. 448-449), which would assured- ly surprise most of its adherents; the lack of phase-relational flexibility in UDC (p. 462) is largely true, but the pioneer efforts of Kervegant have led at least to an offi- cial t est of a relator-schema of my own