University of Virginia Library

The Descriptive Bibliography of American Authors
by
G. Thomas Tanselle

In 1942 Jacob Blanck remarked that "one who attempts a bibliography of almost any nineteenth-century American author has embarked on a task that may prove endless."[1] All bibliographies are endless tasks, but this statement calls attention to the paucity of bibliographical work upon which the student of an American author can draw. One may feel that the situation has now improved, as a result not only of the bibliographies published in the intervening twenty-five years and the increased attention to American literature in bibliographical journals but also of the monumental Bibliography of American Literature (1955- ), effectively conducted by Blanck himself. This work is currently providing the first bibliographies of many authors and the first dependable ones of many others. But the BAL is not — nor is it intended to be — a substitute for full-length detailed bibliographies of individual American authors. Unhappily, then, it is still possible to say that we do not have an entirely satisfactory bibliography for a single major American writer.

That this judgment is not unduly harsh can be demonstrated by a brief survey of past accomplishments — or, indeed, by momentary reflection. For, aside from BAL, what is there for Emerson and Hawthorne? Where does one go for Melville or Whitman or Thoreau, for Hemingway or Faulkner? The bibliographies of Poe and Mark

Twain, of James and Eliot, have their strong points, but even here much remains to be done. It has become a cliché to say that no bibliography can please all its users, because they come to it with such various purposes; and some kind of mortal antipathy between scholars and collectors is usually cited. The foolishness of this point of view is, fortunately, becoming more recognized year by year, as scholars begin to see the relevance of physical details (the collector's "points") to the establishment of a sound text and as collectors begin to understand that their duty, as preservers of the records of the past, extends as much to late reprints as to first "editions." Scholars should be grateful that the interest in first "editions" provided the incentive for the earliest bibliographical work on American authors; but the day has passed when a bibliographer can announce without any sense of shirking his responsibility — as H. W. Cathcart did in his 1905 bibliography of Hawthorne — that he has made no attempt to enumerate later editions "as they could be of no use to the collector of first editions." Today both collectors and scholars — if they are serious (and there are perhaps more dedicated men in the first group than the second) — demand the same things of a bibliography: that it provide a complete account (as complete as surviving material permits) of the publication of a given author's writings, that it be a history of the forms in which those words have been presented to the public, from their initial appearance to the present day (or some other specified point in time). To criticize presently available bibliographies for not being "entirely satisfactory" is not, therefore, to ask naively for an impossibility; admittedly bibliographies will always be riddled with errors and will always be unfinished, but they can at least be satisfactory — from the point of view of all their users — in their approach to the material.

Too often the question of approach is narrowed down to one of technique or form, and many futile misunderstandings have arisen because one man did not approve of the collation formula used by another or did not want to make the effort required to comprehend it. Prefaces to bibliographies abound with deprecatory allusions to bibliographical debates, which the author of the bibliography has avoided by his wholesome "common-sense" (and often quite unscholarly) method. Such a passage as the following, from W. B. S. Clymer and C. R. Green's 1937 bibliography of Robert Frost, belies certain very real merits which the work possesses:

Surely this is a curious kind of logic to find underlying a scholarly investigation. Standardization of the form for recording bibliographical details is obviously a desideratum, and the choice of an appropriate form in any given instance has a direct bearing on the resulting quality and usefulness of the "content" — in other words, any argument about form, separated from a context, is misguided. One would expect bibliographies published since 1949 — the year of Bowers's Principles of Bibliographical Description — generally to exhibit a superior deployment of material. But to say this is not to imply that the chief criterion of a bibliography is necessarily its adherence to the formal recommendations set forth there; to do so would be grossly to misunderstand the purpose of that work. A more meaningful statement would be that neglect of (or hostility to) the Principles in a bibliography is often symptomatic of an omission of important information — if not of an unscholarly approach to the material.

One of the most important services which the BAL is performing may well turn out to be the widespread attention it is drawing to the fact that scholars and collectors require the same kind of bibliography. It gives collations of gatherings (not simply pages) for nineteenth-and twentieth-century books; it is more precise than previous bibliographies in describing publisher's cloth, with all references keyed to photographs; and, most important of all, it records later impressions and editions, not just the first or second. The writers of author-bibliographies must recognize that their research provides information not merely about one author but about the publishing and printing practice of his time; if a bibliography is not conceived of as a record of the forms of publication of certain written works (rather than the minimum amount of information deemed necessary for identification of certain editions or impressions), it is not contributing to the store of bibliographical knowledge that will enable the next author-bibliographer to reach conclusions otherwise impossible.[2] One can never know which facts, considered unimportant at present, may take on significance in the future through the fuller accumulation and analysis of data. The record should therefore be as full as possible, and it is at this point that formal considerations enter. Clearly, as Madan pointed

out long ago,[3] there must be a degressive principle by which certain materials are treated less fully than others — but not of the sort which allows a bibliographer to append to a relatively elaborate description of a first impression a brief list of later imprints (not even distinguishing between impressions and new editions). If one major fault can be singled out in the American author-bibliographies of the past, it is their almost total failure to provide genealogies of editions — to trace the course of a given text through various sets of plates (perhaps issued under different imprints but from the same setting) or various later editions; and this essential information can be presented only if the bibliographer is willing to examine late printings with the care he bestows on early ones and to admit that he cannot ignore textual matters.

American literary scholarship at the present time offers a unique opportunity for the production of this kind of author-bibliography on a wide scale. The intensive textual work now in progress on many major nineteenth-century authors under the general leadership of the Modern Language Association's Center for Editions of American Authors forms the perfect base for descriptive bibliographical scholarship, for most of these projects are forming excellent collections of books (both early and late printings) and are inevitably in the process discovering many new textual changes, binding variants, impressions, and issues. For example, the discovery of duplicate typesetting in the first edition of The Scarlet Letter and of altered standing type in the second, as announced in the Centenary Hawthorne, provides a striking demonstration of the interdependence of detailed physical description and textual decisions. It would be a sad loss if such information were not organized into bibliographies, and most of these projects have happily made arrangements for the preparation of a bibliography in connection with their new edition — among those so far announced are bibliographies of Irving, Hawthorne, Melville, Whitman, and Mark Twain. On the eve of so much activity, it seems appropriate to review the past course of American author-bibliographies. The story is not only a chapter in the history of American scholarship; it is an object lesson as well.

* * * * * *

Interest in American bibliography was at first directed toward the broad field of "Americana" in general, and it would be possible to see

its origins in such enumerations as Bishop White Kennett's Bibliothecae Americanae Primordia (1713), Obadiah Rich's Catalogue of Books, Relating Principally to America (1832), or Henri Ternaux's Bibliothèque Américaine (1837). The only description contained in the early nineteenth-century lists, however, was an occasional indication of format (a simple "4to" or "8vo"), as in D. B. Warden's Bibliotheca Americo-Septentrionalis (1820), Nicolas Trübner's Bibliographical Guide to American Literature (1855), and certain sale catalogues (like John Russell Smith's Bibliotheca Americana, 1865); but Harrisse's Bibliotheca Americana Vetustissima (1872 additions) included quasi-facsimile transcriptions of title pages. Henry Stevens, that illustrious American in London, issued in 1854 his short-lived periodical Stevens's American Bibliographer, for the publication of trial versions of descriptions to be included in his larger work; some of them indicated line-endings in title-page transcriptions. There followed a stream of works from his hand, and he opened the first volume of his Historical Nuggets (1862) with a passage which has been re-echoed by many later bibliographers: Joseph Sabin, whose publication in 1868 of the first part of his great Bibliotheca Americana marked an important step forward, was also carrying on a campaign for higher bibliographical standards in the columns of his house organ, the American Bibliopolist, during the early 1870's (9 vols., 1869-77). Its pages furnished a place for the exchange of bibliographical notes and queries, and he issued with it "A Handy Book About Books," with useful information for bibliographers. By 1870 he was using vertical lines to indicate line-endings in title-page transcriptions (II, 180-187), and in an 1873 article entitled "What Is Bibliography?" he announced, "We define material bibliography to consist of a careful copy of the title, place, date, size, &c., of a book" (V, 84-86). He went on to explain the meaning of format and even proposed a professorship of bibliography: "Then perhaps we may find the meaning of the word will be more widely known" (V, 87).

In the same year that Stevens was remarking in his Nuggets, "We have never seen a perfect Catalogue" (III, vi), the New York firm of

Leon & Brother issued its famous catalogue, First Editions of American Authors (1885), generally considered the first checklist to reflect the growing interest in collecting American literature as opposed to general Americana.[4] It was not of course descriptive, but it listed for sale the complete works of an amazing number of authors and, perhaps even more surprisingly, stressed in its preface the importance of first editions from a textual point of view: "they are of the highest bibliographical and literary interest. . . . they often contain passages or poems which are omitted from later editions. . . . by collecting the first editions of an Author we have the benefit of a pure and unchanged text." This catalogue was followed in 1893 by Herbert Stuart Stone's First Editions of American Authors (with an introduction on contemporary collecting and collectors by Eugene Field) and in 1897 by Patrick K. Foley's American Authors, 1795-1895. These works had a long period of usefulness (and are still referred to occasionally) because of the large number of authors included. But the checklist which has dominated the field is Merle Johnson's American First Editions, first published in 1929. Johnson did not describe the books listed but attempted to mention points which would identify their true first printings. His pioneer use of type-batter evidence (here and in his Mark Twain bibliography) was often misunderstood and ridiculed, and some collectors preferred the easier path offered them by Foley — an attitude satirized in a jingle by Oscar Wegelin: "Some letter dropt, a broken type, / How very, very drolly; / Such things as these, were ne'er thought of / By dear friend P. K. Foley."[5] Johnson revised his work in 1932, and Jacob Blanck brought out further revised editions of it in 1936 and 1942;[6] its continued popularity is indicated by the fact that a fifth printing of this fourth edition was issued as late as 1965, and some people now are as hesitant to relinquish it for BAL as they were to give up Foley three decades before. Other works appeared in the early thirties, offering somewhat more detailed examinations of a smaller number of books, such as Richard Curle's Collecting American First Editions (1930), Jacob Schwartz's 1100 Obscure Points (1931), or I. R. Brussel's Anglo-American First Editions: West to East (1936). In 1950 Carroll A. Wilson's Thirteen Author Collections of the Nineteenth Century provided better descriptions than had previously been available for most of the eleven American authors included. And the BAL in 1955 was the culmination of this tradition of multiple-author bibliographies, far removed from the checklists which were its ancestors.

The growth of interest in collecting and in bibliography, of which these works give evidence, forms the background against which the development of the book-length single-author bibliography should be viewed. One does not begin to look for this sort of publication before the last years of the nineteenth century, the period when many scholarly organizations were founded on both sides of the Atlantic. In the United States attention was first given to the statesmen, with the Bibliotheca Hamiltoniana (by Paul Leicester Ford, 1886), Bibliotheca Jeffersoniana (by Hamilton B. Tompkins, 1887), and Bibliotheca Washingtoniana (by William S. Baker, 1889); and Wilberforce Eames in 1890 and 1891 investigated the bibliography of Eliot's Indian works. By the turn of the century the only literary figures who had been treated were Franklin (by Paul Leicester Ford, 1889), Thoreau (by Samuel Arthur Jones, 1894), and Thomas Bailey Aldrich (by Francis Bartlett, 1898). These contained little description; but the indication of line-endings, format, and pagination in Jones's Thoreau, for example, provided at least as much information as was available in the English bibliographies then published — James Gibson's Burns (1881), Wise and Smart's Ruskin (1893), Temple Scott's or Buxton Forman's Morris (both 1897), or Wise's Browning (1897).

The first decade of the twentieth century saw a flurry of American author-bibliographies, largely created by the Houghton Mifflin Company's consciousness of its heritage, as the successor to the great firm that had published so much of the important nineteenth-century literature. In 1905 it began the first American series of author-bibliographies, with Nina E. Browne's A Bibliography of Nathaniel Hawthorne. The works in this series were not really descriptive at all (since they gave a short — or sometimes full — title for each book, followed by pagination and a symbol to represent size), but — since the publisher's files were at hand — they furnished more information about publication and number of impressions than one might have expected, particularly the full listing of later editions (with no attempt, however, to show relationships of plates). The bulk of each volume consisted of an alphabetical list of individual works, with an indication of the first periodical and first book appearance of each. Hawthorne was followed in 1906 by George W. Cooke's Lowell, in 1907 by George B. Ives's Holmes, and in 1908 by Francis H. Allen's Thoreau and Cooke's Emerson. All five volumes were arranged on the same plan, but some

(like Holmes and Thoreau) indicated line-endings in full titles.

Earlier in the decade Oscar L. Triggs had prepared a Whitman bibliography (in Complete Writings, X [1902], 139-233), but it was inconsistent in providing title-page transcriptions and was not intended to be descriptive. Two library lists of this period were Hamilton B. Tompkins's George Henry Calvert (1900) and Winifred Mather and Eva G. Moore's Thomas Wentworth Higginson (1906). Victor H. Paltsits in 1903 made a step forward in A Bibliography of the Separate and Collected Works of Philip Freneau. Not only did he give line-endings and an indication of size (in format terms); he also recorded signatures (as "[A]-C in fours"), reproduced many title pages, and listed locations of copies. In 1905, besides Browne's Hawthorne, two other bibliographies of Hawthorne appeared. One of them, by Wallace Hugh Cathcart for the Rowfant Club, was clearly aimed at collectors and thus complemented Browne's: while it took up only first "editions," it indicated line-endings and bindings, even offering some discussion of states and binding variants. The other was Jacob Chester Chamberlain's description of his own collection, which in a sense began another series, for Luther S. Livingston's Longfellow in 1908 was presented as one of "The Chamberlain Bibliographies" (Livingston noted line-endings and bindings but gave collations of gatherings only if signatures were present). One other bibliography of this decade was LeRoy Phillips's A Bibliography of the Writings of Henry James (1906); though it indicated line-endings, it was very meager in its offering, allowing "Cloth," for example, to suffice for a binding description and declining to investigate the complicated matter of James's transatlantic publication — nor were its faults removed in the "enlarged edition" of 1930.[7]

Near the end of the decade the theory of bibliographical description was enormously advanced by the remarks of Pollard and Greg,[8] but the result was not to stimulate author-bibliographies on this side of the Atlantic, for only two are noteworthy in the entire second decade: Livingston's Lowell in 1914 (parallel to his Longfellow) and Charles F. Heartman's Hugh Henry Brackenridge in 1917 (not the first and by no means the last of his bibliographical publications). These were

also the years of F. F. Sherman's checklists (Bliss Carman, 1915; Thomas Bailey Aldrich, 1921). The decade of the 1920's, however, is a different matter. The general prosperity made book-collecting popular, and the interest in "modern firsts" as well as the rediscovery of earlier authors (notably Melville) prompted bibliographies of those writers. Frank Shay's Walt Whitman in 1920 provided scant information, but Meade Minnigerode's Some Personal Letters of Herman Melville and a Bibliography (1922) attempted much more, giving line-endings, pagination and format symbol, unusually full contents list, binding description, and price. For some books he furnished a long essay of commentary, including an account of textual variants and quotations from reviewers, and he tried to list later printings.[9] This book was published by Edmond Byrne Hackett under his imprint of The Brick Row Book Shop, which was to reappear in 1927 on Lawson M. Melish's bibliography of Edith Wharton; like Minnigerode, Melish gave special attention to his lists of contents, noting divisional titles and plates. Although his formula for pagination, such as "pp. (viii) + 254," was inadequate, as usual at the time, he did give exact measurements of pages and discussed bindings.

The other bookshop which issued bibliographies in the twenties — even more actively — was the Centaur Book Shop of Philadelphia. Each volume in "The Centaur Bibliographies of Modern American Authors" was a small, attractively produced limited edition, printed at the Torch Press, generally carrying a prefatory essay by the author involved; since many of the authors were only in mid-career, blank pages were included at the end labeled "Future Collations." The first in the series, in 1922, was H. L. R. Swire's A Bibliography of the Works of Joseph Hergesheimer. For each first edition it gave a title-page transcription with line-endings marked (not quasi-facsimile); then a paragraph called "Collation," which began in the pattern of "pp. (vi) + 327, consisting of . . ." and proceeded with the contents; then a paragraph indicating size (as "Crown 8vo, 7¼ x 4⅞"), binding material, lettering on the binding (transcribed), end papers, and edges; and finally a note on variant bindings, special issues, or later editions. It represented what was generally expected of a collector's bibliography at the time, and the other volumes followed in an unusually consistent fashion: Vincent Starrett's Stephen Crane in 1923; Guy Holt's James Branch Cabell, Carroll Frey's H. L. Mencken, and Scott Cunningham's

Carl Van Vechten, all in 1924; Edward H. McDonald's Theodore Dreiser in 1928 (which included reference to English editions, but only from the English Catalogue); Vincent Starrett's Ambrose Bierce in 1929; and I. R. Brussel's Cabell (a thorough revision of Holt) in 1932.

The other bibliographies of the twenties were less satisfactory. Carolyn Wells and Alfred F. Goldsmith's A Concise Bibliography of the Works of Walt Whitman (1922) was generally vague, though it gave brief binding descriptions.[10] Henry Cole Quinby's Richard Harding Davis (1924) used the form "12mo, 19 cm." for size, followed by contents, total pagination, binding, publication information, and very discursive notes. Only occasionally would he mention gatherings, in footnotes which illustrate the unnecessary complications created by the use of format designations for size: one footnote on "12mo" reads, "So according to the size rules of the American Library Association and in the publishers' advertisements; but really and truly, it is composed of eight-leaf signatures" (p. 18). Joseph Lawren's James Gibbons Huneker (1925) gave nothing more than title-page transcriptions and lists of contents; Oscar Wegelin's John Esten Cooke (1925) was another in Heartman's Americana series. In 1929 Vrest Orton's Dreiseriana (labeled as one of "The Chocorua Bibliographies") presented details about the publication of Dreiser's work (including some later editions) in essay form, with occasional interspersed collations; it is best described by its subtitle, "A Book About His Books." To find the outstanding author-bibliography of the decade, one must turn from these separately published works to one which was part of a larger undertaking, Wilberforce Eames's bibliography of John Smith for the twentieth volume (1927) of Sabin (pp. 218-265). His title-page transcriptions (with line-endings) and formula for pagination, as "4to, pp. (8), 1-19, 24-42," were not above the general standard, but they were followed by an enormous quantity of annotation in small type, giving the contents and facts of publication and revealing the painstaking work which lay behind the bibliography. Thomas H. Johnson has with justice called it a "model of bibliographical scholarship in dealing with an extraordinarily complex subject";[11] if Eames's formulas have been surpassed in precision, his careful notes are still impressive.

Despite the Depression, a large number of bibliographies were published during the thirties. If the collecting fever had called attention to the need for bibliographical studies of American authors, the wholesale revaluation of the American past in the twenties intensified

the demand; the result was not only more bibliographies, but more scholarly ones. The decade is framed by one of the monuments of American bibliography: in 1930 Thomas J. Holmes issued Increase Mather and His Works, a short-title list serving as a preview to the two-volume Increase Mather which appeared the following year; and in 1940 he published the three-volume Cotton Mather as well as a volume on The Minor Mathers. At the time of its inception, this work represented the most serious bibliographical attention which had yet been paid to an American author (or group of authors) and the first extensive use of title-page reproductions in an American bibliography. It employed format terms accurately, reported gatherings in concise fashion, located copies, and noted later printings. One objection which may be raised is that its alphabetical arrangement of material is not as meaningful as a chronological one would have been; also, the generous inclusion of title-page facsimiles for all major titles does not obviate the bibliographer's responsibility for complete quasi-facsimile transcriptions. Even with its faults, the Increase Mather demonstrated more clearly than any previous American work that descriptive bibliography was a scholarly discipline, and I think it may fairly be said that it is the earliest American author-bibliography which does not need to be done over.[12]

The general output of these years was characterized by continued attention to contemporary or recent writers, as in the twenties, with a simultaneous effort to treat the standard nineteenth-century authors in more detail. As for the moderns, Bradford F. Swan's Henry Blake Fuller came in 1930; and 1931 alone saw Lucius Beebe and Robert J. Bulkley's Edwin Arlington Robinson, Ralph Sanborn and Barrett Clark's Eugene O'Neill (which included English editions), Cecil Johnson's George Sterling, Edwin O. Grover's Percy MacKaye (a "biographical bibliography"), and Louis H. Cohn's Ernest Hemingway. These were not distinctive, but they generally provided the expected notes on pagination, contents, and binding (collation of gatherings was not expected) and sometimes offered title-page reproductions. The Hemingway, though it gave no publication data, did recognize textual matters, to the extent of saying, "There are slight textual differences between the American and English editions of Men Without Women"

(p. 46); but its imprecision in the use of bibliographical terms is shown by another comment on the same book: "The first edition of this book may be determined by its weight only, as the second printing of the book took place without any typographical changes or corrections" (p. 29). In 1932 Barton Currie's Booth Tarkington gave prices and publication facts, along with the usual listing of total pagination (called "collation"), contents, and binding; but Lucile Gulliver's Louisa May Alcott was hardly descriptive (with "16mo, cloth"), yet it did allude to binding variants, sometimes gave the size of editions, and listed foreign editions. By this time title-page transcriptions indicating line-endings — but not in quasi-facsimile style — were a standard element of an author-bibliography, but the imprecise account of pagination became, if anything, more cumbersome, as "1 blank leaf + pp. xxii + pp. 204 + 2 leaves," in Lavinia R. Davis's bibliography of Edith Wharton (1933).[13] A number of other bibliographies in these years followed the same general pattern and require little comment: Sydney S. Alberts's Robinson Jeffers and Harvey Taylor's Sinclair Lewis in 1933, P. D. and Ione Perkins's Lafcadio Hearn in 1934 (the descriptions are proudly announced as made "from actual inspection"),[14] Alfred P. Lee's Christopher Morley in 1935 (significant for its attention to dust jackets and its measurement of the type-page), Alexander C. Wirth's Lizette Woodworth Reese in 1937, W. B. S. Clymer and C. R. Green's Robert Frost in 1937 also (with thorough notes on binding variants, and format terms used for "foliation rather than size"),[15] Paul S. Clarkson's William Sydney Porter in 1938 (the size of gatherings is indicated — as "in eights" — without saying how many), Arthur Mizener's Archibald MacLeish in 1938 also, and A. J. Hanna's Irving Bacheller in 1939 (with very little description but data on number of copies in each printing).

Against this background six bibliographies of the thirties, on six major authors, stand out. First, William R. Langfeld and Philip C. Blackburn's Washington Irving (1933) — which originally appeared in installments in the Bulletin of the New York Public Library in 1932 — provided very careful discussions of each work, including comments on textual variations among the earliest printings or editions; its weaknesses were a failure adequately to define terms like "issue" or

"state," a neglect of signature collation, and the omission of information on later editions. This last fault was recognized by Stanley Williams, who called it a "collector's bibliography" which laid "foundations for studies of later editions";[16] he — with Mary E. Edge — proceeded to furnish such a study in 1936 (A Bibliography of the Writings of Washington Irving), attempting to list all editions and translations, but with no discussion of textual matters and no bibliographical description. In 1934 two major authors were treated. One was John W. Robertson's Edgar A. Poe, in two beautifully printed volumes from the Grabhorn Press, the second of which was a "commentary" on the bibliography in the first. The work is useful for the numerous title-page reproductions but is unsuccessful in its attempt to show spacing in transcriptions and provides no formulary collations. The other, A Descriptive Bibliography of the Writings of James Fenimore Cooper by Robert E. Spiller and the Irving bibliographer Philip C. Blackburn, revealed a more thoughtful concern with bibliographical principles. Defining an author-bibliography as "the record of his literary life" (thus including more than just first "editions"), they recognized "the immaturity of the science" of bibliography for modern books in working out their own form: The serious consideration of such matters is commendable, even though the passage shows some confusion about the meaning of "format" and the purpose of signature collations. Also noteworthy is their degressive principle, with four lengths of entry, despite some further confusion about "firsts": "American firsts, or firsts issued in other countries in cases where no American first exists, are accorded full entry . . . a shorter form of collation has been used for all authorized English and Continental firsts, whether they preceded the American firsts or not." Finally, one respects their effort to define terms, even when the definitions seem vague today: "The use of the word 'state' instead of 'issue' to describe reprintings with only minor variations is helpful and has been adopted here" (p. 12). The Cooper is important for its attention to the rationale of bibliographical description, even if a number of its specific decisions have been outmoded by later discussion.

Next, in 1935, appeared Merle Johnson's Mark Twain (a revision of a 1910 work, but essentially a new bibliography) — not an unusual production except in one respect, its detailed examination of various states, especially those produced by type batter (as in the six-page discussion of Huckleberry Finn). Individual items are difficult to locate (since there are no item numbers or running titles) except through the extremely full index, which is an integral part of the work, providing the only consecutive listing of the reappearance of specific titles. (In its index and style of description, the Mark Twain was parallel to Karl Yost's 1937 Edna St. Vincent Millay, issued by Harpers in an identical binding, to form a brief "series" of bibliographies of Harper authors; Yost was careful in his comments on bindings and on publication and copyright data but was interested only in firsts and confused in his use of "edition.") The year following the Mark Twain, Charles Beecher Hogan published A Bibliography of Edwin Arlington Robinson (Yale, 1936) — supplementing it in PBSA, XXXV (1941), 115-144, with "New Bibliographical Notes," also issued as a separate. The result of prolonged collecting and research, it may be criticized more for its omissions than for its treatment of the details chosen for inclusion; in an article on his experiences as a bibliographer Hogan asserted, "my book was written with the intention of aiding collectors rather than graduate students, and if one is comparing my description in the presence of the book itself, a listing of the book's contents seemed to me to be quite supererogatory"[17] — a point of view which ignores the fact that more information may make possible the detection of previously unsuspected variants and distorts the whole purpose of descriptive bibliography. The sixth of these major bibliographies has received more praise than any of the others: Thomas F. Currier's A Bibliography of John Greenleaf Whittier (1937), which Jacob Blanck in 1942 called "probably the best single-author bibliography ever done in the United States."[18] The work is particularly full in its listing of reprints and later editions, and it belongs in that very small group of bibliographies of this time which paid any attention at all to gatherings (in

the form "Sig. [1]-7 in 8's; 8 in 2"). Its weakest feature is the transcription of title pages — one is surprised to see so understanding a bibliographer remark that "there is but one instance . . . where the lining has the slightest significance" (p. 4). The book remains, nevertheless, an example of extremely careful work.

In the early 1940's only three bibliographies were in any way distinctive. Despite the fact that Greg's "A Formulary of Collation"[19] had been available long enough to be assimilated, most bibliographers still ignored the basic physical structure of the book and were content with the old formula of size-pagination-contents-binding, or less — as Julian Sawyer's Gertrude Stein (1941), and the even less descriptive Yale catalogue of the same year; Oscar Wegelin's revised 1941 edition of his William Gilmore Simms (giving only transcriptions, pagination, and locations of copies), and A. S. Salley's complementary work (1943) listing the contents; Malcolm Young's "bibliography" of Paul Elmer More (1941), only a checklist; Robert W. Daniel's brief catalogue of Faulkner (1942), which did use quasi-facsimile transcriptions and listed reprints; and George R. Preston's Thomas Wolfe (1943), again with reprints. Of the three major bibliographies, one was another Poe (1940; revised, 1943)[20] — this time by Charles F. Heartman and James R. Canny as part of "Heartman's Historical Series." Although at the opposite pole from Robertson's in point of physical appearance, it was more dependable, revealed thorough checking, and provided a good census of copies (but unfortunately did not make clear which ones had been examined). As its title indicated (A Bibliography of First Printings), it did not include overseas editions, a particularly fascinating study in the case of Poe. There were no collations of gatherings, but the authors at least saw that a well-executed bibliography could serve the needs of all, since they specifically aimed their book at "the Poe scholar, the research worker, and to a certain extent the collector and dealer" (p. 11). Also in 1940 Thomas H. Johnson published The Printed Writings of Jonathan Edwards. Its descriptions were brief (the

contents paragraph quite sketchily treated), but its approach was significant in several respects: it used "format" correctly and reported format only "when evidence is present" (p. x); it employed Greg's formulary, at least in part (a collation line would read, "Sm 8vo; 18.1 x 11 cm.; π3, B-D4, [2], ii, 25, [1] p."); it located copies; and it seriously investigated later editions, such as those of the Tract Societies.[21] The following year Jacob Blanck published his Harry Castlemon, Boys' Own Author. If the writer treated was minor, the bibliographical research was not. Blanck's collations were often in words rather than a formula, but the important fact is that gatherings did get mentioned (as "[1]-13 in 8's"), and the rest of the description was equally thorough, with elaborate notes including copyright information and publication prices.

This bibliography was dedicated to J. K. Lilly, whose interest in book-collecting was to become increasingly prominent in succeeding years, for it was his generosity that made possible the next landmark in American author-bibliography — the series of four volumes on Indiana writers published by the Indiana Historical Society between 1944 and 1952: Anthony J. and Dorothy R. Russo's James Whitcomb Riley (1944); Dorothy R. Russo's George Ade (1947); Dorothy R. Russo and Thelma L. Sullivan's Booth Tarkington (1949) and Bibliographical Studies of Seven Authors of Crawfordsville, Indiana (1952).[22] All four were handsomely produced by the Lakeside Press (with many collotype illustrations), and all followed the same pattern. For each book there was a title-page transcription (in quasi-facsimile) and paragraphs on collation (formulary collation of gatherings, with measurement of leaves and description of paper), pagination (contents), illustrations, binding, publication date and price, and notes. When the contents of a book had previously appeared in periodicals, full details of original publication were also given (the periodical items were again listed in a later section, without mention of the book, but all the references were brought together in an excellent index). Unfortunately the location of copies examined was not indicated (except in a few unusual instances), and only first "editions" were described (though later ones were listed, and even states of Grosset and Dunlap reprints were sometimes differentiated), with a resulting lack of attention to

textual matters. Aside from these reservations, one can agree with F. G. Melcher that "No useful feature of bibliography-making seems to have been omitted."[23] The four volumes form a sustained achievement and without question constitute the finest "series" of author-bibliographies produced in the United States.

During these years there were a few more bibliographies of importance, set off against an increasing number of author-checklists.[24] In 1948 the New York Public Library published another in its outstanding series, William M. Gibson and George Arms's William Dean Howells (serialized in the Bulletin in 1946-47). The great amount of material unhappily necessitated not merely an unattractive double-column layout but a severely condensed form of description: the transcription of the title was followed in one paragraph by pagination and size, in the pattern "p. ii + x + 134, 7¼ x 4½ inches" (with no collation of gatherings), contents, end papers, and binding. Later editions were briefly mentioned, but the numbering system — with each book lettered consecutively under its year, as 70A, 70B, etc., for 1870 — was not designed to show relationships between various printings and editions of a single work.[25] Also in 1948 appeared a new version of Vincent Starrett's 1923 Stephen Crane, by Ames W. Williams and Starrett. The thorough rewriting brought the book into line with modern bibliographical practice in many respects (formulary collation, careful listing of contents, use of data from Publishers' Weekly), so that it is above the level of the average bibliography. The New York Public Library in 1949 published its next bibliography, at the opposite extreme from the Howells in size — Winthrop Wetherbee's Donn Byrne. Its strong points were an equal attention to American and English first printings (with reprints listed) and detailed notes with publication (but not copyright) information; its collations, however, were overly condensed,

in the form "12mo (19 x 13 cm.). v, 282 p.," followed by the enumeration of contents (with some transcription) and description of binding. Jack Potter's John Dos Passos (1950), with good formulary collations and occasional discussions of dust-jackets, was vague in noting copies examined ("several dozen copies" of each Doran title) and reporting later printings ("all books went into later printings unless otherwise indicated"). In the same year as the last of the Indiana volumes came Donald C. Gallup's T. S. Eliot (1952), the outgrowth of many years' interest in Eliot as shown by his 1937 Yale catalogue and 1947 checklist. An admirable job, it nevertheless did not discuss textual relationships nor later printings ("Later impressions and editions are generally ignored except where some particular significance attaches to them," p. ix). It has perhaps received most attention for its pagination formula, based on the system of the American Library Association (used on Library of Congress cards), in which printed preliminary pages adding up to the first numbered page need not be specified, nor do final blank versos. Like any carefully defined system, it has the merit of being unambiguous; but it can be cumbersome to use and more difficult to follow than a straightforward record of all pages, with unnumbered ones italicized: compare "xix pp., 1 leaf, 297, [1] pp." with "i-v vi-xix xx-xxii, I 2-297 298." (To be sure, the latter method does not show which unnumbered pages are blank, but the contents paragraph serves this function.) In any case, the choice of this method for a scholarly bibliography of an important author has furnished, as I. R. Willison says, "the occasion for a major debate on the essentials of modern bibliography."[26]

In the remainder of the 1950's there were four bibliographies of special significance, beginning with the posthumous publication of Thomas F. Currier's Oliver Wendell Holmes (1953), edited by Eleanor M. Tilton for the Bibliographical Society of America. This huge work surpassed the high standard of Currier's Whittier and was especially noteworthy for its use of publishers' files and account books. Its plan was similar to the Whittier, with books and pamphlets (English as well as American firsts) described chronologically and individual poems arranged alphabetically. Gatherings were noted, in words rather than formulas, and paragraphs covered binding, inserted advertisements, impressions, contract, and contents; later English and American (other than Houghton Mifflin) editions were briefly recorded in an appendix. The most noticeable defect (as in the Whittier) was the inadequate (and inconsistent) transcription of title pages. One may

concur with Oscar Cargill in calling the work a "superlative accomplishment" and a "touchstone for the profession" without agreeing that it should be the model in form for future bibliographies of our major authors.[27] Four years later, in 1957, came a two-volume bibliography of James Branch Cabell from the University of Virginia. The first volume, by Frances Joan Brewer, was not distinguished, without quasi-facsimile transcription or any collation except total number of pages, though it was somewhat more detailed on bindings. It is the second volume, by Matthew J. Bruccoli, which makes the bibliography important; entitled Notes on the Cabell Collections at the University of Virginia, it represents the first large-scale application of certain modern bibliographical techniques to the problem of distinguishing impressions in machine-printed books. Each setting of each Cabell book is divided into its constituent impressions, determined by means of collation on the Hinman machine (assuming that textual variants are evidence of a new impression), measurement of gutters, and analysis of type damage. For ease of reference, it is unfortunate that the two volumes of the work could not have been consolidated (particularly in view of the discrepancies that occur, as in the reporting of pagination), but the value of the second volume as a pioneer effort is obvious. In the same year Leon Edel and Dan H. Laurence's bibliography of Henry James was published as the eighth volume in Rupert Hart-Davis's series of "The Soho Bibliographies" (and a revised edition appeared in 1961). The work set a very high standard (with formulary collations and some discussion of the complicated textual problems), and the preface showed that many of the serious difficulties in the description of modern books had been thoughtfully confronted; but one might wish for quasi-facsimile transcriptions (instead of all capitals), fuller treatment of later editions, and a less casual attitude toward the designation of colors.[28] The following year the New York Public Library issued the most elegantly produced of all its distinguished bibliographies, Emily E. F. Skeel's Noah Webster (edited by Edwin H. Carpenter, Jr.), printed in a limited edition by The Thistle Press. The research was done with great care and all later editions duly recorded, but the published form of the descriptions was an unhappy choice: title-page transcription (without line-endings or quasi-facsimile style); pagination — as "119, (1) pp." — measurement, and contents (without signature collation, except where there was no pagination, as on p. 233); location of copies; and notes.[29]

Between the time of these bibliographies and the next Soho on an American figure, in 1963, the numerous brief "bibliographies" can simply be listed. When no comment is made, one may assume that, despite the excellent examples of the late 1940's and 1950's, the standard is about the same as in the 1920's, with no formulary collations and minimal notes on contents and bindings: Samuel French Morse's Wallace Stevens (1954; revised and enlarged in 1963 as one of the checklists published by Alan Swallow of Denver); H. D. Rowe's Hart Crane (1955), also published by Swallow; Klaus W. Jonas's Carl Van Vechten (1955), not really descriptive; Frank dell'Isola's Thomas Merton (1956); C. Harvey Gardiner's William Hickling Prescott (1958); Martha Cox's Maxwell Anderson (1958), not intended to be descriptive; Elmer D. Johnson's Thomas Wolfe (1959); W. H. Hutchinson's Eugene Manlove Rhodes (1959); Edgar Branch's James T. Farrell (1959), a checklist only; William White's John Ciardi (1959), W. D. Snodgrass (1960), and Karl Shapiro (1960); Dan H. Laurence's Robert Nathan (1960), which follows Gallup's Eliot in pagination formula;[30] George J. Firmage's E. E. Cummings (1960); Betty Adler and Jane Wilhelm's H. L. Mencken (1961), a thorough checklist; Henry Hardy Heins's Edgar Rice Burroughs (1962; revised, 1964); David Kherdian's William Saroyan (1965), which includes descriptions and (even two illustrations) of dust jackets; Hensley C. Woodbridge, John London, and George H. Tweney's Jack London (1966), the largest of the Talisman Press bibliographies; and the series compiled by Eugene P. Sheehy and Kenneth A. Lohf, including Marianne Moore (1958), Yvor Winters (1959), Frank Norris (1959), and Sherwood Anderson (1960).[31]

Three bibliographies stand out from this group because they employ a formulary collation of gatherings: Charles M. Adams's Randall Jarrell (1958); J. M. Edelstein's Thornton Wilder (1959); and Cecil K. Byrd's "Check List of the Melcher Lindsay Collection" (Indiana University Bookman, December 1960), a praiseworthy job. James B. Meriwether's The Literary Career of William Faulkner (1961), though not strictly a bibliography, is the indispensable bibliographical tool for Faulkner and takes account of the English editions. The eighteenth volume of the Soho series was Donald Gallup's second bibliography of a major modern writer, Ezra Pound (1963), which maintained the high standard he had set in the Eliot; it employed his special system for recording pagination but did not treat later editions and impressions. In 1964 the University Press of Virginia published another bibliography which used the same analytical techniques as Bruccoli's Cabell: William W. Kelly's Ellen Glasgow, edited by Oliver Steele. All the impressions of each edition were noted, with the evidence for the classification; but the book is not a bibliography in the usual sense because only the distinguishing points are given: "the notes are sufficient to identify any copy of an Ellen Glasgow book printed to date." The aim was to provide a guide for identification rather than a record of the forms of an author's publications. In the editor's words, "although the descriptions are not fully analytical, the notes are the result of a thorough bibliographical analysis." At the end of our survey, then, we have returned to a work superficially like the earlier collector's handbooks, but scholarly in method and intent. It is an analytical study, not a descriptive bibliography; but it has demonstrated some of the tools which must be a part of any future bibliographer's equipment.

* * * * * *

A glance at the record reveals how much remains to be done before even the principal authors are adequately covered. For the colonial and federal periods, only the Mathers, Edwards, and Webster have been carefully treated, though C. William Miller's work on Franklin is nearing completion. Of the early nineteenth-century writers, Cooper, Irving, and Poe have been provided with usable bibliographies; but among the major mid-century authors — Emerson, Thoreau, Hawthorne, Melville, Whitman — not a single one has been given a satisfactory account. If Whittier and Holmes have been dealt with handsomely, there is nothing for Bryant, Longfellow, and Lowell. Mark Twain, Howells, Henry James, and Crane have been provided for, but in varying degrees. And in the twentieth century, the vogue of certain writers has produced erratic patterns, whereby there are three bibliographies

of Cabell, two of Van Vechten, and even a large volume on Edgar Rice Burroughs; it is perhaps surprising that at least Robinson, Eliot, and Pound have been given something of their due. Minor writers often command devoted attention, and the bibliographies of Riley, Ade, Tarkington, and "Harry Castlemon" have set higher standards than those reached in most works on major figures.

No doubt the public reaction to author-bibliographies partly accounts for this state of affairs. They are notoriously bad risks for a publisher, and there is constant pressure to prune each work and reduce the printing costs. Only Houghton Mifflin among major American publishers has ever shown any extended interest in author-bibliographies, although Harpers did begin an informal series with Mark Twain and Millay, and other commercial publishers have occasionally issued bibliographies of their own authors' works. For the most part, however, American author-bibliographies have appeared under the imprints of bookshops (Brick Row, Centaur), libraries (New York Public, Yale), or scholarly (nonprofit) organizations (Indiana Historical Society). The phenomenon of Rupert Hart-Davis is unusual in England, and he has no American counterpart. But the publishers by no means should bear the principal blame for the poor quality of most author-bibliographies: the profession itself is at fault. Bibliographers in university positions continually find that their colleagues have little idea what they are up to; professors of literature concede the necessity for good bibliographies but often give little evidence of knowing how to judge bibliographical work. If their reviews would not call a sketchy bibliography "detailed" or refer condescendingly to a study of "misprints" but instead would enumerate errors and demand a scholarly approach, the general level of bibliographies might gradually rise. But an understanding of the purposes of bibliography cannot be expected in an atmosphere which supports such statements as the following, about Hogan's Robinson: "frankly for collectors rather than scholars, with emphasis upon exact descriptions of editions, bindings, and text, and a minimum of biographical material."[32] There have been, of course, some responsible reviews in the past — one thinks of David Randall's long series for Publishers' Weekly [33] — but there have not

been enough. A bibliographer deserves to be faced with the kind of scrutiny offered by James M. Osborn in his well-known review of Hugh Macdonald's Dryden [34] or by James B. Meriwether in his recent review of Frederick Woods's Churchill (which could be taken as a model of the careful checking required).[35]

On the other hand, it is easy to abuse bibliographies unfairly, and there are few bibliographies so bad that they are worse than none at all. As the Times comments, "A reviewer . . . is always conscious that the sheer ballast of fact in an author bibliography, however inadequate, will always give it useful substance of a kind that unoriginal history, biography or criticism can never have."[36] And if, as Lawrence Wroth once said of bibliographical analysis, "There are few tasks more exacting, more demanding of sustained reflection and of the exercise of the constructive imagination,"[37] then a relatively small number of good bibliographies can be expected. Nevertheless, no valid excuse can be found for haphazard and sloppy work. To quote the Times once more, in its general survey of the Soho bibliographies, the amount and kind of detail in a bibliography may reflect the state of the source material but always reveal "the degree of enthusiasm and understanding which the individual has brought to his task."[38] Though the gradual advance of bibliographical knowledge may alter the particular demands made upon an author-bibliographer, his "understanding," or approach, is the fundamental test.

The history of the descriptive bibliography of American authors bears out this point. But it also demonstrates that the less narrow the approach, the more likely that future specialized demands will be met. If past experience is to be instructive, the bibliographer must realize that he cannot ignore anything as irrelevant. Information which seems to him useless can furnish someone else with essential data on book-trade or publishing practice that will, in turn, provide a meaningful context for the original facts.[39] More specifically, this inclusiveness

must encompass all later editions and impressions[40] — both details from publishers' records and descriptions of examined copies (often more difficult to locate than "firsts"). An account of textual relationships among these various editions also forms a natural part of the story (twenty-five years ago Osborn was saying in his Dryden review, "The time is not far distant when such a statement will be considered a fundamental responsibility of the scholar-bibliographer"). Quasi-facsimile transcriptions and formulary collations of gatherings are central elements (even for machine-printed books), along with a contents note offering transcriptions of other preliminary pages. And, to mention only one more debatable issue, dust jackets cannot be ignored on the grounds that they are not an integral part of a book; their importance obviously warrants treatment, and they can at least be described, in their various states, as separate entities. It is to be hoped that the bibliographies of the major American authors now being edited[41] can take proper advantage of an unusual opportunity to combine new bibliographical techniques with the accumulated experience of a seventy-five-year-old tradition. In 1934 Percy Muir could say, "The state of modern bibliography is worse than deplorable. It is almost hopeless."[42] One is perhaps not being too optimistic to feel that the present outlook is somewhat brighter.

Notes