University of Virginia Library

II

It is obvious that many more books are catalogued with fairly brief entries by librarians than are accorded detailed descriptions by bibliographers and that there are many more library cataloguers at work than there are descriptive bibliographers. It is not surprising, therefore, that more attention has been given over the years to the principles and practices of library cataloguing than to those of descriptive bibliography.[22] The present Anglo-American Cataloging Rules (AACR) have developed from a tradition that can be traced back to Panizzi's British Museum rules of 1841 and includes Charles A. Cutter's Rules for a Printed Dictionary Catalogue (1876), the American and British Library Associations' Catalog Rules of 1908, the "Preliminary American Second Edition" of those rules in 1941, the Rules for Descriptive Cataloging in the Library of Congress (1947, 1949), and the A. L. A. Cataloging Rules for Author and Title Entries (1949).[23]

At many stages along the way there have been formal discussions, committee meetings, investigative reports, and institutes to plan revisions and new developments in the code. In the fifteen years before the publication of AACR in 1967, Seymour Lubetzky prepared proposals for new rules (1953) and drafts of rules reflecting the work of the Code Revision Committee (1958, 1960); official institutes on revision of the code were held at Stanford (1958) and McGill (1960); other conferences on cataloguing took place at the University of Chicago (1956) and at St. Andrews, N. B. (1961); and an International Conference on Cataloguing Principles was held in Paris in 1961, with 53 countries and twelve international organizations represented.[24] Since then, discussion has continued, as at the conferences on AACR at the Universities of Toronto and British Columbia in 1967 and at the University of Nottingham in 1968,[25] and some further changes have been made in AACR (in a ten-page supplement added to the 1970 impression and in the version of Chapter 6 published separately in 1974).[26] As the product of so much deliberation, the AACR must be of interest to the descriptive bibliographer, both because any well-considered approach to the recording of books is relevant to his concerns and because this one in particular reflects the cumulative experience of several generations of cataloguers and will exert great influence. Michael Gorman has been quoted as saying, "This is not only the best cataloguing code we have, it is also the best we are likely to have for a very long time."[27] It is not unfair, therefore, to expect AACR to be based on a clear understanding of the kinds of relationships among bibliographical activities which were outlined above. And it is legitimate to scrutinize the extent to which the recording of physical details as directed by the rules is useful to descriptive bibliographers.

Chapter 6 of AACR, on the "descriptive cataloging" of separately published monographs, is naturally the focus of attention for the descriptive bibliographer. Perhaps the best way to begin an examination of its approach is to look at Rule 141, on "collation."[28] Traditionally what is called the "collation" in a library catalogue card or entry consists of three parts: pagination, illustrations, and size. The "preliminary note" to this rule emphasizes its concern with physical details: the "collation" is called "the cataloger's description of the

physical work" (p. 47), and one of the aims is "to present a picture of the physical characteristics of the work to the reader" (p. 48). But one should notice that "work" rather than "book" is the word chosen and that the physical characteristics are recorded in order to help the reader "both in identifying the work and telling him something of its nature." Furthermore, another aim is "to ensure that all those parts of the work are described which would be retained in the binding or rebinding of the work," implying that the emphasis is on the substance of the work and that other integral leaves, such as those containing advertisements, are unimportant. It is clear, even from this introductory statement, that the "collation" is to be principally concerned with an indication of the extent of the contents of the book and not with the book itself. There is no reason to object to this emphasis, except that the "collation" has been defined as the "description of the physical work." If "work" here means "book," the usage is imprecise and the statement untrue. If it is being used carefully, in distinction to "book," the inclusion of the word "physical" still causes a problem: since the work exists physically only as embodied in the book, the physical description can only be based on the book, which may contain elements (such as advertisements) which are not part of the work.

The rules for recording pagination (or foliation) reflect the same ambiguity. First one is told (141B1a) that the "extent of a work" is to be indicated in terms of pages, leaves, or columns, depending on the method followed in the book being catalogued. The implication is that the cataloguer is concerned with the characteristics of the physical book, since the method of numbering employed in a given book is not related to the extent of the work; if the sole interest were in indicating extent, all figures for all books could be converted to a single unit, such as pages. The same impression is conveyed by the further rule (141B1b) that arabic or roman numbers or letters are to be used, following the practice of the book. But this rule ends with the statement that "Pages or leaves numbered in words, or in characters other than Arabic or Roman, are designated in the collation in Arabic figures." Thus the emphasis has shifted to an indication of the extent of the work, eliminating a report of the actual system of numbering used. What is the rationale, one may ask, for allowing the nature of the characters employed in numbering the pages or leaves of a given book to determine whether the cataloguer reports in his entry a characteristic of the book (the actual system used) or a characteristic of the work (its extent, measured in convenient terms)?

If the numeration in a book is divided into two or more series, the

North American AACR—following a long-standing tradition in library cataloguing—requires the recording of the "last numbered page or leaf of each numbered section" (141B1c). In many books, of course, the last page of text is not numbered, and this rule clearly places the emphasis on recording a physical detail (which pages are in fact numbered) rather than on specifying with precision the extent of the work.[29] Yet the emphasis shifts the other way in the determination of what constitutes a section: "either a separately numbered group of pages, or leaves, or an unnumbered group which, because of its length (one fifth or more of the entire work), or its importance, should be mentioned."[30] When bulk or importance becomes a criterion for the inclusion of information, certainly no attempt is being made to provide an accurate accounting of the physical structure of the book. But, then, if a small or "unimportant" section of text can be omitted, the representation of the extent of the work is not entirely accurate, either. Indeed, the aim, as it emerges two paragraphs later (141B1e), is only to provide an approximate idea of the bulk of the work: one is told that a correction may be required if "the last numbered page or leaf does not represent the total number, or approximately the total number, of pages or leaves in the work or in the section."[31] The same mixture of aims appears in the instructions for recording the pagination: the figure representing a group of unnumbered pages is to be enclosed in brackets (141B1c), thus emphasizing a physical detail; but where the numbering changes from roman to arabic within a sequence (e.g., i-viii, 9-176), the whole sequence is to be represented by the arabic total (141B1e), thus emphasizing the extent of the section rather than the physical details of the numbering. Similarly, advertisements which constitute separate groups of pages (whether numbered or unnumbered) are to be disregarded (141B1c), placing the emphasis on the work, not the book; but if the advertising pages continue the page numbering of the text, the last page number in the sequence is to be given, with a parenthetical indication of which pages the advertisements occupy (141B11), thus making the physical detail of pagination dominant over the content of the pages.[32] One of the awkward situations produced by these rules is illustrated at the end of rule 141B1c itself. Since the rules require that a note be provided to call attention to the presence and extent of a "bibliography" (that is, a reference list) in a book, and since such a "bibliography" might well occur on a final unnumbered page, provision must be made for referring to such a page in certain instances. The solution offered in this rule is illustrated by the pagination record "86, [1] p." and the note "Bibliography: p. [87]." Aside from the awkwardness of referring to the 87th page in two different ways, the basic difficulty is that the use of brackets implies a concern with the actual pagination, while the necessity for adding the "[1]" arises solely from the nature of the material printed on that page. There could be still more unnumbered pages, which would not be recorded because their content did not demand reporting. The principal interest, clearly, is in the content, and pagination references derived from physical description do not always serve that purpose efficiently; but, used in this way, they do not serve the purposes of the descriptive bibliographer either, because they do not necessarily form a complete record.

The handling of various special problems connected with pagination further reveals this awkward mixing of approaches. When there is no numbering in a book at all, the printed pages are counted and the number placed in brackets—or, if the figure is over 100, the number may be estimated (141B2). And when there are several (more than three) "numbered main sections," the numbers on the last numbered page of each section are added together and presented in the form "968 p. in various pagings" (141B3b). Both these rules obviously emphasize the work, not the book. Why, then, is the numbering of the individual main sections to be reported when there are no more than three of them, with other lesser sections recorded in the form of a total, as in "xiv, 226, [44] p." (141B3a)?[33] The fact that

there are fewer main sections does not alter the cataloguer's aim; and the resulting series of figures represents more than one system, since the bracketed figure here is a total of two or more sequences (it could also, in another situation, refer to a single unnumbered section). Furthermore, the bracketed figure itself could result from more than one system, if some of the sections it refers to are unnumbered and some numbered, since all printed pages are counted in unnumbered sections and only the last numbered page in numbered ones. If the primary interest is, as it would seem to be, in recording the extent of the work, what is the point of introducing an element of physical description which complicates that record and yet does not, because of its ambiguity, furnish an offsetting benefit to a descriptive bibliographer?

Two other rules about pagination deserve to be commented on. One describes the treatment of works in more than one volume (141C): when the pagination of the volumes is separate, only the number of volumes is to be recorded; but when the pagination is continuous, it is to be added in parentheses, as "2 v. (xxxi, 999 p.)." This rule is doubly peculiar. In the first place, it is difficult to understand why the physical division into two or more volumes renders a reference to pagination unnecessary, when pagination—rather than "1 v."—is considered the appropriate way to indicate the extent of a work in one volume. After all, some two-volume works are shorter than some one-volume works.[34] Second, it is not clear why the continuity of pagination is a reason for recording the paging; the pagination is either worth listing or not worth listing, but the fact that it starts over in the second volume does not make it irrelevant. The logic is even further confused in the statement that "Separately paged preliminary matter in volumes after the first is ignored unless it is important; if it is important, the work is not considered as being paged continuously"—in which case the pagination is not noted at all. One is left with the anomaly that the presence of "important" matter in a separately paged preliminary section in the second volume of a two-volume work is a reason for eliminating the record of pagination entirely.[35] Surely this is a prime example of the situation in which

a physical detail of bookmaking is allowed to interfere with the effective indication of the extent of the work.[36]

The other pagination rule which requires particular comment deals with incomplete copies (141B12): "If the last part of a work is wanting, and the paging of a complete copy cannot be ascertained, paging is given in the form 179+ p., with note of the imperfection." Aside from the illogical reference to "the last part of a work," when a book can be defective in other places as well, the problem with this rule is its conception of the function of a catalogue listing. Whereas the rules previously discussed have shown some confusion about the distinction between books and works, this rule reveals some indecision about whether the undertaking is a catalogue or a bibliography. The implication here is that the pagination of a complete copy, when known, is recorded in the collation line (presumably with a note somewhere pointing out the defect in the copy under examination). But if these are catalogue rules—not rules for bibliographies, which refer to ideal copies—the basis for each entry must be the book present in the collection being catalogued.[37] The emphasis may be on the content of the book rather than on its physical features, but any physical features mentioned must conform to the characteristics of the specific copy at hand. Details about the characteristics of a complete copy may be useful, but they are strictly supplementary. Perhaps the role of the Library of Congress in supplying printed catalogue cards to other

libraries has helped to weaken the concept of a catalogue entry as an accounting of a specific copy; in any event, a code of cataloguing rules should not contribute to the confusion by implying that a catalogue card or entry refers to an ideal rather than an actual copy.[38]

The second element in the collation, following the designation of the pagination or foliation or number of volumes, is a brief reference to any illustrative matter in the book. It consists of nothing more than the abbreviation "ill.", "unless particular types [of illustrations] in the work are considered important enough to be specifically designated"; when that occurs there are several specific terms, like "diagrams," "maps," "music," or "portraits," to choose from (141D1a). As with pagination, the intent is obviously to suggest something about the content of the work, not to record the precise physical structure of the book; but the emphasis here is on the nature of the illustrations, not their extent. A later rule (141D4) does permit specifying the number of illustrations, but only if they are numbered or "if the number can readily be ascertained"; and any numbers given are to be arabic and are not to appear in brackets even if the illustrations themselves are unnumbered. This rule, unlike the rules for pagination, reveals no indecision regarding aims, for the focus is entirely on content: such physical details as the manner of numbering the illustrations are not allowed to intrude into a statement about the illustrations. A problem arises, however, from the fact that the previous part of the collation, the pagination statement, may also refer to plates (141B1d) and to music (141B10) when they occur in separately paged or unpaged sections or on pages not otherwise covered by the

notation of pagination. Examples given are "xvi, 246 p., 24 leaves of plates" and "74 p., 15 p. of music." Apparently the rationale is that this part of the collation line indicates the "extent" of the work and would be misleading without the mention of these major elements; the second part then takes up the nature of the illustrative matter as a whole, whether it occurs on separate pages or on pages which are included in the numbering of major sequences.[39] This illustration statement thus becomes a commentary on one aspect of the content of the pages recorded in the pagination statement. Two questions immediately come to mind. First, if the extent of a work in numerical terms is to be supplemented by some comment on the manner of presentation of the material, why are illustrations singled out for comment? And why are illustrations defined to include genealogical tables and graphs (141D1a) and to exclude tables in general (141D1b)? Second, if other groups of pages need not be labeled in the pagination statement, why should those containing plates and music be named? Plates may be scattered through a volume, but as far as the measurement of the extent of the work is concerned they would seem to be no different from the "lesser variously numbered or unnumbered sections" (141B3a) for which a single unlabeled total is to be provided. The treatment of illustrations thus raises another question about the purposes of the pagination statement. As for the illustration statement itself, the problem is less one of aims than of consistency in carrying them out. One wonders whether the expression "ill." (or even one of the more precise terms) is informative enough to bother including; but the question clearly has to do with reference bibliography, not descriptive bibliography, for physical description is not intended.

The third part of the collation is an indication of size, consisting of the measurement in centimeters (rounded off to the next higher full centimeter) of the height of the binding (141E1). This measurement is of course a physical detail, but it is only one of several measurements

which would be of interest to the descriptive bibliographer. The purpose of including this one measurement in a catalogue entry generally stressing the content of works rather than the form of books is puzzling; and the four reasons furnished in a "Preliminary note" to this rule do nothing to suggest an answer.[40] First we are told, "The size of the work is included in the catalog entry as an aid in finding the work on the shelves." Of course "book" is meant instead of "work" in both instances, and this imprecision reflects a basic confusion as to the purpose of this information. The interest is unquestionably in the work, as stated, but the reason provided can only refer to the book; and locating a book by its size—even as a device to supplement other techniques—is certainly a primitive method of information retrieval.[41] The second stated reason for recording the height of a book is "as an aid to the user of the catalog in selecting a desirable edition." This preposterous point scarcely requires comment, for the correlation between the height of a binding and the desirability of the edition it covers would be relevant (if at all) only to the choice of books for reading in bed or for packing in luggage. The other two reasons are that the height "serves the reader who wishes to borrow the work through interlibrary loan or who wishes to order a photocopy of the work or a part of it." Again, "book" is meant; and the person who would be influenced by the height of the book in his request for a loan or a photocopy cannot be very seriously interested in the work it contains. The only justifiable reason for including the height of a binding in an entry oriented toward the content of the book is one that is not mentioned: the height could be regarded as a supplement to the pagination details, further indicating the extent of the work by suggesting the size of the pages. But this function—indicating "the space occupied by the work"—would be served still better by the specification of two or three dimensions, as required for broadsides (141E3) and "unusual formats" such as "boxes or cans" (141E6). The discussion of "size" (that is, height), as it stands, is not well thought through and provides no sensible reason for the inclusion of that detail; if no better reasons are to be offered, the requirement of specifying height is a flagrant example of the insistence on a physical detail which is unnecessary in relation to the emphasis of the entry and inadequate to serve as an aid to the physical bibliographer.[42]

The rules for description in the North American Text of AACR developed from—and remained close to—those in the 1949 Rules for Descriptive Cataloging in the Library of Congress. And both these codes represent a considerable simplification of what had been proposed in 1941 in the "Preliminary American Second Edition" of the A. L. A. Catalog Rules. The pagination rules set forth there result in such illustrations as "xii p., 5 l., [3], 219 p." and "v, 365, [3] p., 2 l."— which suggest careful attention to physical details but are nevertheless intended to indicate the extent of the work, not of the book. One rule, for instance, states that "Blank leaves at the beginning of a book are not counted even if they have apparently been included in the paging"; and another requires that intermediate unpaged matter be reported as leaves "when some or all of the leaves are blank on one side, except that unpaged matter continuing the text from a preceding numbered page is given as a page, even if printed on a leaf one side of which is blank" (rules 271-272). Despite the elaboration of rules such as these,[43] the system does not manage unambiguously to convey just which pages contain printed matter (the number of blank pages in the groups designated as leaves in the illustrations cited above is not determinable from the formulas)[44] —and it certainly does not provide a register of all the pages in a book. Dissatisfaction with these proposed rules was fortunately widespread and began even before their publication, for a note facing the title page of the 1941 volume

acknowledges that there had been "considerable disagreement as between some catalogers and some administrators." But the complaints and the ensuing discussions too often resulted merely in requests for simplification, without a reexamination of the underlying function of the catalogue entry—without, that is, exploring why the elaboration of detail did not further the aims of the entry.[45] The rules had indeed become too complex, but not in any absolute sense. They had become inappropriately complex because the complexity arose from the notation of physical details, when the function of the pagination record was to suggest the extent of the work and was not primarily concerned with the physical book. The resulting formula was bound to be an inefficient and finally unsuccessful instrument for conveying information about either the work or the book. It was more dramatically unsuccessful than the present rules; but they still suffer from the same confusion. Even the British Text of AACR, which is more logical in its presentation of rules for description and its requirements than the North American,[46] falters from indecision regarding the purpose of including physical details. The AACR treatment of pagination may look good in comparison with that in the complex 1941 rules; but the act of simplifying the rules has not altered the underlying problem which made those earlier rules unsatisfactory.

Cataloguers and librarians themselves have been uneasily aware that the collation statement is a trouble spot, the treatment of which has never been satisfying. Herman H. Henkle, in the Studies of Descriptive Cataloging (1946) which formed part of the deliberation leading up to the 1949 Library of Congress Rules, summed up the problem:

This passage is instructive: Henkle's opening statement accurately sets forth the issue; but his summary of the arguments on both sides shows how the discussion generally focuses on the amount of detail involved rather than on the alternative functions of the collation as expressed in his earlier comment. (Certainly the arguments of those favoring "brief collation," as recorded here, are extremely weak; but that does not mean that theirs is necessarily the weaker position, for their arguments simply do not touch the basic question.)[47] Some years after the Library of Congress Rules appeared, Leonard Jolley described Library of Congress cataloguing as "still avowedly bibliographical"[48] and questioned the value of including the collation at all, since without an identification of type sizes and layout the number of pages does not very accurately denote the size of the work and since the details provided "do not produce a statement of pagination upon which a bibliographer can rely in all cases" (p. 132). Like Henkle, he saw the central issue, and he stated it even more trenchantly: With this kind of statement before them, the planners of AACR should have been able to confront the real problem and produce a set of rules for collation firmly based on a well-defined view of its purpose. Instead, it was decided that the discussions preceding the 1949 Rules constituted a largely sufficient basic reconsideration of the rules for description and that the rules for entry and heading were the ones which now demanded full-scale rethinking;[49] as a result the AACR rules for description are disappointingly similar, in their confused underlying principles, to what had existed before. This fact has not gone unobserved. Andrew Osborn has said, "I am much concerned because in the AA code the rules for descriptive detail are not in the same class as the rules for entry and heading."[50] And R. O. Linden has pointed out in AACR "a confusion in general between the bibliographical, and what might be termed the evaluative function of collation."[51] In his discussion of the rule regarding the date of a volume, he makes a comment about the meaning of "edition" which again would apply to other rules for description: "Two approaches appear possible—one, a definition that gives emphasis to the bibliographical character of the work, the second, a definition that is based on the intellectual content. Two values appear to be confused here" (p. 50). It is clear that this confusion has been recognized not only by descriptive bibliographers but by those within the library profession as well.

Of course, as these comments suggest, it is not merely the collation line which reveals a confusion between books and works. I have concentrated on that part of the entry, particularly the pagination statement, as a telling illustration of the problem; but the problem is not confined to that element. For instance, the treatment of the title pages of books bears some awkward traces of a concern with physical detail. The basic rule for the "transcription" of title makes clear that a literal transcription is not intended, for exactness is required only "as to order, wording, spelling, accentuation, and other diacritical marks" but "not necessarily as to punctuation and capitalization." Furthermore, if "diacritical marks are omitted from the title page, they are added in conformity with the usage in the text" (134B1). The emphasis is clearly on the content of the title, not on its formal presentation or typographic layout. Yet when long titles are abridged (as they are "if this can be done without loss of essential information"), three dots are required to mark the ellipsis (134B2). This requirement is understandable when part of the title quoted follows the omission, for not to indicate the omission in such instances would simply be irresponsible quotation; but when the omission occurs at the end of the quoted part of the title, one could argue, as with ordinary quotations within a text, that the ellipsis dots are unnecessary. The recording of the title is admittedly a special type of quotation, since punctuation and capitalization need not be followed; but it nevertheless is a quotation (concerned with words and the accompanying marks conventional to the language), not quasi-facsimile transcription (concerned with the typography and layout in which those words and marks are presented).[52] Other recorded details, aside from titles (main titles, subtitles, series titles, and so on), need not be regarded as quotations, however, but as reports of information. Therefore, when the author's name is provided as the heading for the entry, it seems unnecessary to repeat the name following the title, as the basic rule requires (134D1).[53] The concern of this rule is obviously not with the physical form of the title page, because it recognizes that the

author's name may have to be taken from a different position on the title page or even from somewhere else in the volume; but there is a lingering sense that the exact form in which an author's name appears in a book must be recorded, even when fuller information about him (his complete name and perhaps his dates) is already provided in the heading. In regard both to the ellipsis dots and to the repetition of the author's name, one could argue that in some cases their presence might suggest or convey important information (as when the form of the author's name on the title page is considerably different from his established name cited in the heading), and in these cases their inclusion would be justified, since the goal is to be informative regarding substantive, not formal, matters. But the criterion for inclusion, given the emphasis of the entry as a whole, must turn on the relevance of the detail as information about the work or author, not on an assumption that the mere physical presence of the detail in a particular form is relevant in itself.

The treatment of some parts of the title page reflects this principle more firmly than that of other parts. If a subtitle, for example, is printed at the head of the title page, above the title, it is silently transposed to a position following the title in the entry (134C4b).[54] And the imprint is regularized into the order place-publisher-year, regardless of the order on the title page, and neither this rearrangement nor omissions of words need be specified (136A, 136C1, 138A).[55]

The inconsistency in the handling of different parts of the title page is strikingly shown by the fact that data for the imprint statement can be taken from elsewhere in the book and recorded without brackets,[56] whereas the author's name must appear in brackets if it is taken from somewhere in the book other than the title page (132B, 134D1).[57] Few people, I think, would question the propriety of the rearrangement of the material so as to produce relatively uniform catalogue entries, and the fact that this approach is so widely accepted suggests a broad understanding—whether consciously expressed or not—that library catalogue entries serve largely a reference function. Even the descriptive bibliographer generally assigns a standardized title to each of his descriptions to aid the reader in locating them; his focus is of course on physical description, but that aspect of his work which involves reference bibliography entails standardization for efficiency of reference. The library cataloguer, unlike the descriptive bibliographer, is primarily concerned with reference bibliography, and thus the body of each of his entries can be expected to be a standardized presentation of facts, not a transcription of forms. In the AACR treatment of title-page information, those few rules which imply some obligation to offer physical description stand out, against this background, as incongruous and, indeed, confused.[58]

It should come as no surprise that the emphasis of library cataloguing, as reflected in AACR, is on what may be called reference bibliography, where the primary concern is the intellectual content of books. Of the two conventional divisions of library cataloguing, subject cataloguing—or classification—obviously deals with content; what may be less clear at first is that the other division, so-called descriptive cataloguing, does so as well.[59] Because descriptive bibliography treats of books as physical objects, some confusion may be caused by the use of the term "descriptive cataloguing" to denote an activity which does not. The difficulty, however, is not entirely one of terminology. The present cataloguing code, AACR, in all its versions, states that "The collation is the cataloger's description of the physical work and is limited to standard bibliographical terminology" (132A; 1967 texts, 131).[60] Yet, as this examination of the rules for collation indicates,

attention is not given to physical details for their own sake but as clues suggestive of the extent or nature of the work contained in the book being catalogued. The descriptive cataloguer's job, as set forth in these rules, has a basically different aim from that of the descriptive bibliographer, and the "standard bibliographical terminology" employed is not that which is standard in the field of descriptive bibliography.[61] What the library cataloguer normally means by "descriptive" is "annotated with certain largely physical details which help to characterize the content of a book." The objection to the cataloguer's practice, as codified in AACR, is not that he gives too much attention to the work and neglects the book: it is entirely proper that he should emphasize the work. The flaw in the AACR is that some of its recommendations for handling physical details reflect a failure to keep this goal firmly in mind and to recognize the relationships between reference and physical bibliography. The result is a lack of decisiveness and singleness of purpose in a number of rules, producing in turn certain data in a form not entirely appropriate to either interest. Descriptive bibliographers should have no quarrel with reference bibliography; but their respect for it is not likely to increase so long as it can appear at times as merely a less precise form of descriptive bibliography. Part II of the Anglo-American Cataloging Rules, as a product of great deliberation and a document destined to have wide influence, is disappointing in that it is not able clearly to place its subject in relation to descriptive and reference bibliography and thus to offer rules informed by a well-defined point of view.