University of Virginia Library

The Historiography of American Literary Publishing
by
G. Thomas Tanselle

It has now been fifteen years since William Charvat made his important comments on "Literary Economics and Literary History" before the English Institute.[1] Those remarks suggested a great many lines of inquiry into publishing history which were then virtually unexplored and indicated their undeniable relevance— indeed, their crucial significance—for meaningful literary history. However difficult the facts may be to come by and however frequent the crossing of disciplinary boundaries that may be required, one should not have to insist on the essential importance of technological, geographical, and economic factors, or of publishers' policies and copyright laws, or of the facilities for mass education and for book distribution— the importance of such areas as these for an informed study of the course of literary history.

But if the relevance of these matters is obvious, what, one may ask, has been done to pursue the various lines mentioned by Professor Charvat? Or, to put the question more bluntly, how does it happen that, with so many graduate students searching for seminar and thesis topics, so few are directed toward work in a field where the labors of many men for many years to come will make only a beginning? It may be that the current fashions in literary scholarship are such that the instructor who must get out another article will choose to explicate a

given poem for the fiftieth time rather than survey the output of a given publisher for the first time. Or it may be that fewer members of English departments are interested in publishing history than in explication, and that fewer members of history departments are drawn to the history of publishing than to the history of labor organizations or steel companies. But surely one cannot argue that publishing history is less "literary" than the biography of an author or less "historical" than the analysis of a statesman's oratorical style. And in any field the accomplishment of certain basic tasks is a prerequisite to progress in more inclusive tasks—which is not to say that publishing history cannot be considered a legitimate subject of endeavor in its own right, but simply that literary scholars may think of it at times only as the groundwork for other investigations. At any rate, the fact remains that the history of American literary publishing can still be called, in relative terms, a neglected field.

To be sure, certain scholars, notably Professor Charvat himself, have drawn their graduate students into this area. In addition, American bibliographical journals have been hospitable to such research.[2] It is true also that, if the English had Cyprian Blagden to examine Elizabethan publishing, we have Rollo Silver and Lawrence Wroth to furnish us with details of the colonial publishers; if for nineteenth-century British publishing there are Edmund Blunden and Royal Gettmann, for American there are David Kaser, Warren Tryon, and Madeleine Stern; if the general course of British book production and distribution has been surveyed by Frank Mumby, the American has been treated by Hellmut Lehmann-Haupt.[3] We have our listings of

regional imprints, our index to Evans, our histories of individual houses, our registers of printers and booksellers in a given book-trade center, our editions of cost-books. But we do not have a complete inventory of regional imprints, nor histories of all significant publishers, small and large, nor analyses of all book-publishing centers, nor edited versions of all important surviving records of publishing firms.

We have, in other words, made an encouraging beginning, but we must recognize that the work of a handful of scholars, however dedicated and productive, can be no more than a beginning on a task of such magnitude. Comprehensive histories of American book production have been attempted,[4] and no one would deny that works like the Lehmann-Haupt volume have their usefulness—but it seems to me equally evident that further general histories would, at this stage, be pointless. Before anything approaching a full-scale history of American publishing can be embarked on, we must have the complete story of the individual publishers that make it up. It is helpful, as I say, to have had some of the broad outlines sketched in for us first; but the next step must inevitably be a return, for more intensive study, to the details on which those outlines were based. The writing of history always moves in such cycles, but in the writing of American publishing history we have arrived only at the beginning of the second stage, looking forward to the next level of synthesis and generalization. How are we to talk with authority about the characteristics of a particular publisher until we have listed every title he issued and discovered the circumstances of the publication of each title? How are we to comment, in any but pat generalizations, on the publishing of a particular period until we have made such a study of every publisher operative during that time? And how are we to trace with insight the course of American publishing from one period to another until we have examined each period in this way? It is my contention, therefore, that the small areas must be staked out next, before we can proceed to survey the wider, if only slightly greener, fields.

I

What I am saying has, of course, been often recognized—as one can see by glancing at past achievements. Until recently the principal activity, understandably enough, was concerned with colonial printing and bookselling. For the early period, the study of literary publishing per se has been, because of the nature of colonial book production and its utilitarian purpose, subordinate to a study of printers and their types, though one should not, for that reason, fail to examine the booksellers (or the bookselling function of the printers), for their techniques of promotion and distribution lie more nearly at the center of "publishing" (as opposed to "printing") history. It is natural that the presses of Stephen Daye, Benjamin Franklin, William Bradford, and William Parks would be among those which have attracted the most attention,[5] and natural, too, that the first printers in various colonies have frequently been the subjects of research.[6] As comprehensive guides to the period we do have the work of Lawrence Wroth, Douglas McMurtrie, and Rollo G. Silver,[7] which have supplanted the

earlier more casual commentaries;[8] and we have, in varying degrees of thoroughness, the inventories of imprints for particular areas[9] —but these can be construed as histories of literary publishing only in the most undigested form. We still need to have detailed accounts, with full descriptive bibliographies, of the work of other printers and booksellers—such as Robert Bell, Thomas Fleet, Samuel Kneeland, Benjamin Edes, Timothy Green, David Hall, Hugh Gaine, Benjamin Eliot, Daniel Henchman, and John Allen (of Boston), to name only a few out of hundreds.[10]

If less has been done generally with the literary publishers of the turn of the nineteenth century, when Philadelphia was replacing Boston as the center of publishing and when the functions of printer, publisher, and bookseller were vaguely beginning to take their present shape, at least the two leading figures, Isaiah Thomas and Mathew Carey, have received their due in the researches of Clifford K. Shipton and E. L. Bradsher. The study of Carey has proceeded, in the hands of Lawrence Wroth, to one of Carey's traveling salesmen, the illustrious Mason Weems, and has moved on, led by David Kaser, to the years when the firm was known as Carey & Lea.[11] But in neither of these

cases do we have attempted a complete catalogue of the imprint. Nor do we have studies of other interesting firms like Isaac Riley, Joseph Nancrede, Robert Aitken, West & Greenleaf, Samuel Hall, or James & Thomas Swords.

The nineteenth century has been the field of greatest activity in recent years. William Charvat's Rosenbach lectures of 1957-58 survey brilliantly the whole field of literary publishing for the first half of the century, and Donald Sheehan does the same thing for the Gilded Age. Madeleine B. Stern covers the entire century with a series of short well-documented accounts of seventeen important publishers,[12] including James P. Walker and Horace B. Fuller (publishers to the transcendentalists), John Russell of Charleston, Robert Fergus of Chicago, A. K. Loring, Frank Leslie, G. W. Carleton, and Dick & Fitzgerald. The house that assembled the greatest galaxy of literary figures was, of course, Ticknor & Fields, which has, accordingly, been studied most intensively—by Warren Tryon, William Charvat, and James Austin— while James Osgood, one of the later partners of the firm that descended from Ticknor & Fields, has been treated by Carl J. Weber.[13] The later history of that company, as it transformed itself into Houghton Mifflin, has not been told in detail, nor have the literary activities of the other important Boston house, founded in 1837 by Charles C. Little and James Brown. In fact, one may say that the literary aspects of the large general publishers have been dealt with much less fully than have those of certain smaller, experimental (and perhaps more exclusively literary) firms. Our principal source of information about the great general houses of the nineteenth century, such as John Wiley (1807), Wiley & Putnam (1838), James and John Harper (1817),

Daniel Appleton (1825), J. B. Lippincott (1836), and Charles Scribner (Baker & Scribner, 1846), is still only the commemorative or anniversary volumes which the firms themselves have issued from time to time.[14] Scholarly researches, including accurate catalogues of publications, have scarcely touched these firms (to say nothing of Jewett & Co., Derby & Jackson, and so on), nor the other large literary firms prominent later in the century, like E. P. Dutton (Ide & Dutton, 1852), Henry Holt (Leypoldt & Holt, 1866), Dodd, Mead (Taylor & Dodd, 1839), Thomas Y. Crowell (1870), Century (1881), and Frederick A. Stokes (1881).

Certain features of the literary book trade during this time have not gone unnoticed, however. Cincinnati, as the leading Western publishing center for much of the century (partly because of McGuffey's Readers), has been investigated by Walter Sutton,[15] and there have been briefer studies of other regional publishing activity, as well as regional book-trade directories.[16] But the phenomenon of the Bobbs Merrill Co. (Hood & Merrill, 1838) at Indianapolis, or of David McKay (1882) —known best perhaps to Whitman collectors—and John C. Winston (1884) at Philadelphia, has not been studied, nor have the numerous Chicago firms of this time, notably the forerunners of A. C. McClurg (W. W. Barlow, 1844). The whole problem of the geographical shifts in leadership, with New York assuming first place after midcentury, and with various outlying centers flourishing and declining, requires much further investigation, as does the popularity of reprints of British literature and the repercussions of edition binding. The relationships between authors and their publishers are important for understanding literary, as well as publishing, history, and they are beginning to be studied for the nineteenth century—in greatest detail, thus far, for Prescott, though that story involves English publishers as much as American.[17] This was also the age of dozens of companies

producing cheap reprints on the worst of papers but in the most flamboyant of bindings. That general movement has been surveyed,[18] but it needs more detailed study—though for the Beadle dime novels we can hardly expect anything more detailed than Albert Johannsen's three-volume work.[19] Raymond Kilgour is making important contributions to our knowledge of individual publishers of this time with his work on Roberts Brothers and on Estes & Lauriat, and we have some material on Mark Twain's publishing ventures, though not enough.[20] But the literary taste of the Gilded Age can never be adequately analyzed until we have thorough accounts of such publishers of subliterary works as Lee & Shepard, D. Lothhrop & Co., Porter & Coates, and Henry Altemus.

American literary publishers of the 1890's and the turn of the century were a colorful lot, but, despite their appeal (the piracies of Mosher and the Toulouse-Lautrec posters for Stone & Kimball were among the more glamorous, but not uncharacteristic, episodes), they have not been treated so fully as they deserve. If one thinks first of Stone & Kimball, it is partly because the history of that firm has been so thoroughly detailed by Sidney Kramer, in a book that includes bibliographical descriptions of all 309 titles issued and illustrates the use to which trade catalogues, advertisements, and book reviews may be put.[21] Hardly less influential as publisher to the American fin-desiècle

(and the British, too, since it was the American publisher of the Yellow Book) was the Boston firm of Copeland & Day, whose history and output have been studied by Joe Walker Kraus in a Master's thesis (modeled on Kramer's work) at the University of Illinois in 1941.[22] Mention of Illinois brings to mind how indebted we are to the Library School there for much of the knowledge we do have of this period. Particularly under the supervision of P. L. Windsor and Anne M. Boyd, Master's candidates and students in Library 102 prepared a series of pioneering theses and term papers in the 1930's (available at present in the stacks of the University Library).[23] We owe also to an Illinois doctoral dissertation our most thorough study of Thomas B. Mosher, the Portland publisher of limited editions, whose firm, though extending beyond the nineties, was in its heyday at the turn of the century.[24] English literary publishers had begun to establish American branches by this time (the earliest was Macmillan in 1869, followed in 1875 by Longmans Green and in 1896 by the Oxford University Press); but we have no detailed treatment of these branches, and the American branch of the principal British avant-garde publisher of the nineties, Elkin Mathews & John Lane, has been treated only peripherally in J. Lewis May's book.[25] Beyond these, what do we have for the nineties? Where can we go for information about Way & Williams, or Lamson, Wolffe, or R. H. Russell? Where do we learn about the important general publishers established during the nineties, such as L. C. Page (1892) and Frank N. Doubleday & S. S. McClure (1897)?[26] The story of that period still has to be written, and it will depend for its information on careful studies of all these publishers.

For the twentieth century, the scholarship is even more meager. The works that exist deal not so much with particular firms as with special subjects, such as private presses and fine printing,[27] best sellers,[28] university presses,[29] book clubs,[30] and even (though hardly literary) textbook and reference work publishers.[31] There is also no end to the general surveys of the state of the book industry—largely economic studies, not literary, though they will be useful to future historians of the literary economics of the mid-twentieth century.[32] But what solid work has been done on the individual publishers of literature in this century? Despite the logic of attempting to establish the record of these firms while it is still possible to consult the files or interview the persons involved, almost nothing has been achieved. Seemingly, no one has yet been attracted to those large publishers particularly associated with the first three decades of the century, like Frank N. Doubleday, George H. Doran, Stanley M. Rinehart, John Farrar, Alfred Harcourt,

Donald Brace, Harold Guinzburg, W. W. Norton, Richard L. Simon, and M. Lincoln Schuster. But, perhaps more surprisingly, the smaller, more experimental, more militantly literary or artistic or socialistic firms—the real representatives of the personal publisher in twentieth-century America—have not been studied either: such men as Mitchell Kennerley, B. W. Huebsch, Alfred A. Knopf, Albert and Charles Boni, Horace Liveright, Thomas Seltzer, Lincoln MacVeagh, Robert McBride, and Pascal Covici. There can be no question that the story of a great deal of the most important twentieth-century American literature is bound up with these men, and with other even smaller publishers, like Lieber & Lewis, Frank Shay, Egmont Arens, and the Sunwise Turn.[33]

The facts which author-bibliographies have turned up about some of these publishers cannot substitute for full-scale treatments of them as important men in their own right (not simply as the publishers of James Joyce, or Sherwood Anderson, or Edna Millay, and so on). The task is an essential one, and the materials for performing it are becoming more difficult of access every day, as anyone who has tried to compile a catalogue of a twentieth-century firm will testify. In absolute terms, of course, the materials are far more plentiful than for earlier periods, but that is to be expected; what is decidedly discouraging, however, is that scholars are not currently utilizing the masses of papers which, in the minds of too many publishers, are not worthy of preservation. We do at present have the means, if we have the will, to write the history of the publication of modern literature in America— and to write it fully, not as loose surveys of miscellaneous facts or as fanciful embroiderings of conjectured details, which some of our past histories of earlier publishers have had to be (or have been). Even a brief sketch of what has been accomplished in the historical study of American literary publishing reveals, besides the industry of a few dedicated scholars, the huge gaps, in all periods, which remain to be filled. And even a brief perusal of the prefaces to the studies mentioned here should provide a large enough record of difficulties, not to drive away future students, but to attract more of the same caliber.

II

The differences which one discerns between the kinds of research into publishing history performed for different periods—the emphasis at one time on typography, at another on sales and distribution, or at

another on biography—constitute an index to the changing categories of materials available for various periods, as well as to the changing nature of the publisher's calling. There are basically five kinds of materials for the study of American literary publishing: 1) records, papers, and other files of the firms themselves; 2) the published trade catalogues of the firms; 3) contemporary national or regional bibliographies, or retrospective lists; 4) reminiscences and memoirs of publishers and others associated with the book-trade; 5) contemporary references, ranging from the publishers' own advertisements or posters to book reviews, columns of literary gossip, and lists of "books received." On the basis of the relative availability or scarcity of the materials in each of these categories, we can divide the study of American publishing into certain periods, in terms of methodology. It should go without saying that the books themselves, which the publishers produced, form the central material of study, toward which these five categories are various routes; but the scarcity of the books to be studied (or, for modern periods, their great proliferation) remains another factor, beyond but not unrelated to these five, which affects the methodology and the reliability of publishing historiography.

Because the third category listed above, consisting of the national bibliographic registries, is generally at the center of all attempts to establish the output of a given publishing firm, it is well at the outset to be aware of the peculiarities of the records available for different periods. For American bibliography, the chief retrospective lists are Evans, Shaw-Shoemaker, and Sabin, and one must, of course, rely on retrospective (rather than current) lists for the early period. The original twelve volumes of Charles Evans' American Bibliography, published between 1903 and 1934, cover the years 1639 through part of 1799, chronologically by year and alphabetically by author within the year; each volume has two indexes, one of authors and one of subjects, and a list (but not index) of printers. Clifford K. Shipton published a thirteenth volume in 1955, bringing the record through 1800, and Roger P. Bristol has now provided an author-title index (1959) and an index of printers (1961) for the entire set. Though Evans (with 39,162 titles) supersedes the only previous record, the famous "Haven list" (a 350-page supplement prepared by Samuel Foster Haven for the 1874 edition of Isaiah Thomas' History of Printing), there are still a great many omitted items—so that, until these are all brought together, a number of other lists and articles, like the Waters list, must be consulted.[34] The years 1801 through 1819

(amounting to 50,192 items) are now covered, in nineteen yearly volumes, by the most recent of these retrospective researches (1958-1963), the work of Ralph R. Shaw and Richard H. Shoemaker; these volumes are not as yet indexed. The third of these great catalogues, Joseph Sabin's Bibliotheca Americana: A Dictionary of Books Relating to America . . . , with 106,413 entries in 29 volumes (published between 1868 and 1936, and completed by such eminent bibliographers as Wilberforce Eames and R. W. G. Vail), is not, technically, a national bibliography, since it includes works relating to America published in other countries, but it is useful both because of its large scope and because of the additional bibliographical references it gives for many items. In using Sabin one should keep in mind the reductions in its scope made in 1929, 1932, and 1933. Before volume XXI (1929) the scope was Sabin's original one—to attempt the listing of everything connected with the Americas (North and South) down to the date of the current volume itself; beginning with volume XXI, only works to 1876 were listed. In 1932 (with part 130, beginning with "South Carolina") the date was pushed back to 1860, and in 1933 (with part 141, beginning "Stuart") to 1840. For certain categories, the cut-off date was considerably earlier in each case, and literary scholars should remember that, from 1932 on, most poetry, drama, and fiction published after 1800 was omitted.[35]

For most of the nineteenth century, one may use contemporary bibilographies, chiefly Roorbach, Kelly, and the American Catalogue. Orville Roorbach's original publication was partially retrospective, since he first issued his Bibliotheca Americana in 1849, covering the years from 1820. He brought out a current supplement in 1850, followed by a new consolidated edition in 1852; thereafter, three supplements (in 1855, 1858, and 1861) brought the record to 1861. His work consists of an author-title alphabet (no subject index), and one peculiarity (aside from the errors and omissions) is that the various elements of each entry are arranged in columns, so that, if one has the patience, it is possible to search for a particular publisher's books by running one's eye down the publisher-column of each page. James Kelly's American Catalogue (2 vols., 1866, 1871) is the companion work, using the same arrangement, for the years 1861-1870.

National bibliography in the last three decades of the century is bound up with the official trade journal, Publishers' Weekly, founded in 1872 by Frederick Leypoldt—but actually a continuation of various earlier journals like Charles B. Norton's Literary Advertiser, Literary Gazette (1851-55), Literary Register (1852-56), and Literary Letter (1857-62), the American Literary Gazette and Publishers' Circular (1855-72), and the Trade Circular and Publishers' Bulletin (1869-71).[36] Leypoldt's first American Catalogue (2 vols., 1880) lists, in two alphabets (author-title and subject) the books in print on 1 July 1876; one feature of interest to students of publishing is that there is an appendix of publishers' lists of their books published since 1876. Under the editorship of R. R. Bowker, the office of Publishers' Weekly put out seven permanent supplements to the American Catalogue (covering 1876-84, 1884-90, 1890-95, 1895-1900, 1900-05, 1905-07, 1908-10). It should be understood that the American Catalogue is based on the weekly record of new publications as listed in Publishers' Weekly; and it should be remembered that there are two alphabets in each volume before 1900, one after. Publishers' Weekly, of course, continues to the present, but the American Catalogue was replaced at the turn of the century by two new publications of the now well-known H. W. Wilson Company, the Cumulative Book Index and the United States Catalog. The first of these was published monthly, beginning in 1898, and then cumulated annually into one author-title-subject alphabet. The difference between this publication and the United States Catalog is that the latter does not record all books produced in a given period but only those still in print at a particular time. The first volume (1900) lists the books in print at the beginning of 1899 (in two alphabets, one for author, one for title); the second edition (1903) does the same for 1902 (now in one author-title-and-subject index), the third (1912) for 1912, and the fourth (1928) for 1928 (there was also a supplement to the second edition in 1906 covering the books published 1902-05). After 1928 there is only the Cumulative Book Index, brought together in huge four-year cumulations (beginning with 1957-58, two-year cumulations);

at that time its scope was increased to include all books published in the English language.[37] The student of publishing should not fail to notice the value of the comprehensive lists of publishers and their addresses in each of these volumes.

The historian of twentieth-century American publishing has one further national bibliography at his disposal, one that is far more important for his purposes than the Cumulative Book Index—the Catalogue of Copyright Entries issued by the Copyright Office of the Library of Congress. Though this record is limited to works copy-righted in the United States and is therefore not a complete record of publication, it is of obvious usefulness for establishing the core of a publisher's output and dating the titles, since it has an index that lists, in addition to authors, the copyright holders—who are, frequently, the publishers. Thus, while the Cumulative Book Index can be used only to check on titles after they are known (since it has no publisher index), the index to the Catalogue of Copyright Entries can at times be used to establish a basic, though almost certainly partial, chronological catalogue of a given imprint (basic in the sense that these listings constitute proof of the existence of the books, recorded as they are from the actual deposit copies).

Theoretically, copyright records should not be limited in their usefulness to the twentieth century, because the first American copyright law was passed in 1790. But the story of the repeated efforts of dedicated men to establish a national copyright catalogue is a very discouraging one.[38] Our principal copyright laws are those of 1790, 1870, 1891, and 1909. The 1790 law did provide for the deposit and registry of copies, but this procedure was to be handled by the various district courts, resulting in inefficiencies and inaccuracies even for those books which the authors or publishers decided to register. Our fragmentary published record of early nineteenth-century copyrights consists, then,

of only two lists: 1) the 1822, 1823, 1824, and 1825 supplements to S. A. Elliot's A List of Patents . . . (1820), containing about 950 titles deposited between 1796 and May 1825; 2) Charles Coffin Jewett's "Copy-right Publications," in the form of two appendixes to the 1850 report of the Smithsonian Institution, listing 2,057 titles deposited between 1846 and 1850.[39] The only other related published record for these years is the Library of Congress catalogues of holdings, then issued complete at ten-year intervals with annual supplements. The copyright-deposit books are incorporated into the alphabetical listing in the complete catalogues, but some of the annual volumes (like the one in 1846) have copyright entries listed separately; for the first part of the alphabet, the Alphabetical Catalogue of the Library of Congress: Authors (1864-80) is a place to check for the period after mid-century. In using these catalogues or the Smithsonian's records, however, one must keep in mind the shifting status of these places as depositories. Deposit copies went (from the district courts) to the Secretary of State until August 1846; to the Smithsonian, the Library of Congress, and the Secretary of State from 1846 until early 1859; to the Department of the Interior and nowhere else from 1859 to 1865; and to the Library of Congress as well as to the Department of the Interior from 1865 until the act of 8 July 1870, when the Library of Congress was made the sole depository. The importance of the 1870 act lies in the fact that it consolidated the entire copyright administration, including the function of registry formerly assigned to the district courts, into one central office.

It was not, however, until the celebrated "international" copyright law of 3 March 1891 that official provision was made for a published catalogue of copyright entries. Beginning 1 July 1891—and continuing to the present—these records have been published, but their early years were so experimental in form that one is justified in thinking of the copyright Catalogue as a twentieth-century reference tool. From the beginning it has recorded (like the two earlier brief catalogues) books deposited and therefore physically existing (in contrast to the district court records, which listed registrations of forthcoming books that may never have appeared), and the Catalogue was declared by law in 1909 to constitute prima facie evidence of the facts it contained. But this weekly Catalogue of Title-Entries, as it was first called, did not include the dates of receipt of deposit copies until during its

seventh year (beginning 13 September 1897), did not give the official copyright dates until 1898, and did not provide cumulated indexes until 1901. As a matter of fact, it did not become really satisfactory for reference until July of 1906, when the responsibility for publication of the Catalogue was shifted from the Treasury Department to the Copyright Office itself; at that time Thorvald Solberg, to emphasize the improved form which the Catalogue was to take on in his office, began a new series in the volume-numbering and changed the title to the Catalogue of Copyright Entries.[40] From that time on, a search of the copyright records is relatively easy and can be accomplished by recourse simply to the printed Catalogue; but information about earlier copyrights remains very difficult of access, though it is all available in the Copyright Office, since the record books of the clerks of the district courts for the period 1790-1870 (transferred to the Copyright Office in 1870)[41] and the records of the Office itself for the period 1870-1897 are imperfectly indexed.[42]

The Catalogue from 1906, then, until 1940 (when the War made necessary, unfortunately, a very abbreviated form of entry and index)[43] is a most remarkable work of reference. The entries, printed from the linotype slugs used for the Library of Congress cards, contain all the information (as to date, pagination, title-page transcription, and so on) available on those cards, plus certain other facts, usually three: 1) the copyright date; 2) the date of deposit of two copies in the Copyright Office; 3) the date of receipt of an affidavit of American manufacture.[44] It should be understood that, before the Act of 1909, the official date of copyright was the date on which the title was entered for registry, not the date on which it was published or received; since that time the date of copyright has been construed as the official date of publication reported by the publishers. Clearly, the appearance of any book may have been delayed beyond the predicted publication date, and the actual publication date may differ from the copyright date listed; but the fact remains that, since 1909, the copyright records do give the intended (and often actual) publication date, while before that time they provide only a terminus a quo. The date of deposit of copies is usually later than the copyright date, but with different significance for the pre- and post-1909 years. Before 1909 it was required that copies be deposited not later than the publication date— a fact which pushes forward the terminus a quo; after 1909, since the copyright date is identical with the publication date, the copies, obviously, were generally deposited after publication. The Catalogue lists, for published books, only those actually received and not simply all the titles registered in the Office records.[45] The third item, the affidavit

of American manufacture, was required only after 1909 and often differs (if at all) from the deposit date by a matter of days.[46]

With this information about national bibliographies as a background, it is possible to divide American publishing roughly into three periods, in terms of the methodology required for historical research: 1) the period before the copyright law of 1790, in which only retrospective lists are available; 2) the years between 1790 and the early 1870's, in which fragmentary copyright records may sometimes be used to supplement the somewhat unreliable Roorbach and Kelly lists; and 3) the time since the early 1870's, which saw the establishment, not only of the Copyright Office as the consolidated central agency, but of the Publishers' Weekly, the Publishers' Trade List Annual, and the American Catalogue. One should remember, too, for each of these periods, to refer to whatever national bibliographies of other countries (particularly England) are available; for one's task in the study of an American publisher must include, for each title, information as to whether it originated with him, was reprinted from an earlier British edition, or consisted simply of the British sheets with an American (or joint) title page and casing. The English Catalogue is especially helpful since it provides, from 1800 on, the month, as well as the year, of publication for each book and since it has generally been cumulated quinquennially.[47] The whole area of the relation between English and American publication dates (for both English and American works) needs to be investigated, as David Randall has tantalizingly implied by entering American publication dates in his catalogue of the Lilly Library exhibition (1963) of books on the Grolier List.

For the first of these periods, since the function of printer and publisher was not always distinct, publishing history is almost identical with typographical history. Papers and records of colonial printers do not, as a rule, survive (though there are occasional documents like the Franklin & Hall account book at the New York Public), and printers' catalogues of publications for this period are scarce (though Franklin, Hugh Gaine, Samuel Hall, and others did issue them). The index to Evans can be used to establish a basic list for a given printer, one that

may then be supplemented by reference not only to other lists of early imprints but to contemporary newspapers. The printer who published a newspaper was likely to advertise his own books prominently, and booksellers' notices also appeared regularly. Though the latter will probably not contain names of printers, they are useful for approximating publication dates of particular titles and sometimes for ascertaining prices. For example, John Mein inserted extensive lists of books for sale in his shop in the Boston Chronicle, which was published by Mein & Fleeming.[48] Kneeland & Green's announcements in their own paper, the New England Weekly Journal, are another case in point—as when they say, on 4 April 1738, that "In a few Days will be Published" an Essay Concerning Silver and Paper Currency and then report, on 11 April, that the Essay "This Day is Published." One does not, however, expect great accuracy in these notices: the same paper announced Eliphalet Adams' Sermon . . . on . . . the Execution of Katherine Garret as "Just Published" on 4 July and again on 18 July. Nevertheless, when the Massachusetts Gazette, on 24 November 1786, lists a new edition of Webster's Spelling Book (for sale by Peter Edes and John Boyle) as "this day published," one does at least have a rough indication of the time of publication. But more advertisements, unfortunately, will be for runaway slaves or tea than for books. Reviews may occasionally furnish needed information about a work, but much of the time the reviewing was not extensive enough to be a substantial guide nor prompt enough to be useful for dating. Reminiscences are scanty, too, but there are the journals of Hugh Gaine or Franklin's Autobiography to provide some insight into the workings of the book trade and printing establishments at the time.[49]

The next period, 1790-1870, offers more materials for investigation. If the papers of publishers are not exactly plentiful for these years, they are much fuller than for the preceding period—the Lea & Febiger Papers in the Pennsylvania Historical Society, further Carey papers in the American Antiquarian Society along with the Isaiah Thomas material, the Fields Papers at Huntington, the Dix & Edwards Papers and

the Ticknor Papers at Harvard, the large collection of papers concerning the publishing industry in the Harvard Baker Library[50] —and our two published cost books, of Carey & Lea and Ticknor & Fields, pertain to this time. The practice of binding catalogues or announcements into the backs of published volumes became widespread (thus preserving many which would otherwise be lost), and their very usefulness, occasionally, in distinguishing between printings of a book is also indicative of their value as a guide to the chronology of an imprint. Advertisements are still a help, but once the boundary of 1790 is crossed the copyright ledger books of the district courts offer a possible source of corroboration. The fact that Royall Tyler's The Contrast, for example, was advertised as "just published" in the Massachusetts Centinel for 22 May 1790 brings one at least to the vicinity of the publication date —supported by the copyright entry in the Pennsylvania district court record book on 15 June 1790. As the century moved on, the proliferation of magazines and newspapers which reviewed books, like the Knickerbocker, Democratic Review, Burton's, Graham's, Atlantic Monthly, Scribner's, and Broadway Journal, provides further possibilities for checking dates. The first publication of "The Raven" in book form can thus be pushed back as far as April 1845, since Vandenhoff's Plain System of Elocution, in which it appeared, was reviewed in the New York Weekly News on April 19.[51] Of course, the whole perplexed copyright situation at this time must be taken into account in any explanation of a publisher's decisions.[52] A good source of technical information for mid-century is Jacob Abbott's The Harper Establishment (1855). And the reminiscences of publishers or their friends are always worth consulting—especially such storehouses of information as Joseph T. Buckingham's Personal Memoirs and Recollections of an Editorial Life (1852), J. C. Derby's Fifty Years Among Authors, Books, and Publishers (1884), John W. Moore's Historical, Biographical, and Miscellaneous Gatherings . . . Relative to Printers, Printing, Publishing and Editing . . . (1886), Samuel G. Goodrich's Recollections of a Lifetime (1857), and James T. Fields' Yesterdays with Authors (1871) and Biographical Notes and Personal Sketches (1881). Douglas McMurtrie has reprinted (1929) Henry Lemoine's Present State of Printing and Bookselling in America (1796), and Mathew Carey's Autobiography (1829) has been reissued (1942). Less well-known memoirs can be turned up, too, like Walter Carter's Autobiography and Reminiscence (1907), Uriel Crocker's Memorial (1891), John C. Holbrook's Recollections of a Nonagenarian (1897), or Hubert H. Bancroft's Literary Industries (1891), and an important series of 41 reminiscences of American publishers appeared in the New York Evening Post between November 1874 and September 1875. To the extent that booksellers in the earlier years more nearly performed the promotional function of the publisher than the printer did, one may find much relevant information in book-trade directories of a given area, especially the retrospective ones issued by the New York Public Library.[53]

The really distinguishing feature, however, about this period is the great number of retrospective lists of regional imprints that are continually being made available. To be sure, for the Atlantic coast the checklists of early imprints take one back into the preceding period and for some western states take one up as far as 1890, but in general one can say that the bulk of this work has dealt with the first three quarters of the nineteenth century. The undisputed dean of regional imprint recording was Douglas C. McMurtrie, who had issued checklists of imprints for Chicago (1836-50), Colorado (1859-76), Louisiana (1764-1810), Maine (1792-1820), Michigan (1796-1850), South Carolina (1731-40), Tennessee (1793-1830), Utah (1849-60), and Wisconsin (1833-50), in addition to numerous briefer studies of imprints, between 1927 and 1935, before the work of the American Imprints

Inventory began. He was the natural choice as director of that phase of the WPA's Historical Records Survey. Between 1936 and 1942, under his leadership, with a central office in Chicago, the Imprints Inventory brought out 37 checklists which cover early printing in Alabama (1807-40), Arizona (1860-90), Arkansas (1821-76), California (1833-55), Chicago (1851-71), Florida (1784-1860), Idaho (1839-90), Iowa (1838-60), Kansas (1854-76), Kentucky (1787-1820), Massachusetts (1801-02), Michigan (1796-1850), Minnesota (1849-65), Missouri (1808-50), Nebraska (1847-76), Nevada (1859-90), New Jersey (1784-1800), New Mexico (1784-1876), Ohio (1796-1820), Tennessee (1793-1850), Washington (1853-76), West Virginia (1791-1830), Wisconsin (1833-63), and Wyoming (1866-90).[54] These lists represent the work of 2000 staff members who amassed fifteen million slips by searching 5000 libraries; the original intention was to record all pioneer imprints to 1876 (when the American Catalogue began)—or to 1890 in eight western states—but the project was unfinished at the time the WPA terminated. Four later volumes growing out of the Inventory have appeared under the auspices of the Bibliographical Society of America: the Dakotas (1858-89) and Arkansas (1821-76), edited by Albert H. Allen in 1947; Rhode Island (1727-1800) by John E. Alden in 1949; and Oklahoma (1835-90) by Lester Hargrett in 1951. McMurtrie himself also continued to issue volumes separately from the WPA series, such as those for Chicago (1835-50), Indiana (1804-49), Mississippi (1798-1830), Montana (1864-80), North Carolina (1749-1800), North Dakota (1874-90), Oregon (1847-70), and a number of New York towns. In addition, he had a few distinguished forerunners, such as Charles R. Hildeburn on Pennsylvania, 1685-1784 (1885), Stephen B. Weeks on North Carolina (1890), J. H. Trumbull on Connecticut, 1709-1800 (1904), and Lawrence C. Wroth on Maryland, 1686-1776 (1922); and, since 1935, several other states have found able recorders, notably Alabama (1807-70), California (1833-62), Maine (1785-1820), Maryland again (1777-1810), South Carolina, Tennessee (1861-66), and Texas (1795-1860), along with work on Vermont (to 1820) and West Virginia (to 1863).[55] Since 1950, students in the Department of Library Science at the Catholic University of America have turned out over forty regional imprint lists as theses, covering Connecticut (1801-10), Delaware (1801-62), Maine (1821-35), Maryland (1811-76), Michigan (1851-55), Mississippi (1831-40), North Carolina (1801-20, 1866-76), Rhode Island (1821-50), and a number of cities and other areas.[56]

Though far from complete, and published in a maddeningly inconvenient way, these imprint records furnish our most reliable guides to the publications of this period and should always be consulted in addition to the considerably less trustworthy Roorbach and Kelly. When indexed by imprint, they can be used (like Evans) to establish a basic list for a given firm. The essential difference, however, between the recording of imprints and research into the history of publishing should be borne in mind. The student of imprints is interested in the development and spread of printing; the student of publishing history is concerned partly with printing, but also with the economic factors entering into relations with authors, public taste, book distribution, and the like. The distinction between printing and publishing, important for all periods, becomes so pronounced during this period that the recording of imprints by locality can be no substitute for the further analysis of imprints by publisher. Imprint recording has been pursued industriously and has provided raw material for the historian of publishing, but it in no way reduces his obligation to rearrange that material in other patterns which reveal other kinds of significance.

III

Since my own recent work has been concerned with publishing in the third of these periods, the years since 1870, I should like to comment on the available materials in somewhat greater detail and

illustrate some common problems in reconstructing a publisher's list with examples drawn from this period. Further, the publishing history of these years is in greatest need of scholarship—though much remains to be done in the earlier periods as well—and perhaps a fuller description of possible procedures and approaches is therefore in order. If the period is less challenging to the mind which enjoys a maximum exploitation of a minimum number of facts, it should be all the more attractive to the temperament which delights in finding order in an overwhelming mass of information. The huge increase in the volume of books produced makes this the most complex of the periods and necessitates an even greater amount of study in individual firms before a meaningful pattern of trends and movements can be discerned.

Of the three chief tools for the period, two began in the early 1870's (both originating in the office of Frederick Leypoldt), and it is their indispensability which largely justifies dating this third methodological period from their inception. One is the Publishers' Trade List Annual, a device for making permanently accessible the yearly catalogues of nearly all publishers. Its basic purpose is obviously to make current catalogues conveniently available to booksellers, but the accumulation of non-current volumes on library shelves is a great boon to the student of publishing, who would otherwise have to contact the individual publishers he was concerned with, only to find, perhaps, that the ephemeral catalogues for various years had disappeared from the files. Most publishers of any significance have inserted their annual catalogue of publications in PTLA, and those who issued only a few titles in a given year often have a portion of a page in a special indexed supplementary section. In the early years there was an index and then the Publishers' Weekly annual list as well, but by the late 1880's the index was abandoned, to be revived only in 1902-04 before the inauguration in 1948 of Books in Print; in the 1890's the PW list was also discontinued as duplicative, so that the Annual by the turn of the century had evolved into its present form, consisting only of publishers' catalogues. The practice certain publishers have had of listing even their out-of-print books, as well as their forthcoming publications, makes this work the obvious and basic tool for research on a firm's output; the difficulties encountered in establishing the list for a publisher who did not submit his catalogues (such as Samuel French in the 1870's or Mitchell Kennerley for most of the period of his firm's existence) is testimony of the degree to which the information in this work is both essential and taken for granted.

Since a publisher's intentions, as reflected in his catalogues, do not always coincide with the facts as they turn out, one must therefore have recourse to the second of these tools, Publishers' Weekly. Established in January 1872 as the Publishers' and Stationers' Weekly Trade Circular, and changing its name to Publishers' Weekly on 2 January 1873, it is the chief source of book-trade news as well as of weekly lists of new publications. At first called "Order List of Books Just Published or Forthcoming" (arranged by publisher), this weekly list became, by mid-February 1872, the "Alphabetical List of Books Just Published" (by author), followed by an "Order List" (by publisher), and, by 18 August 1877, simply the "Weekly Record." From the beginning, the first issue of each month carried an "Alphabetical Reference List" of all books listed during the preceding month (with reference to the week of full listing), and sometimes these lists were placed together as an annual record, though the usual annual record through 1885 was an author and title index to publishers' summaries of their output (and each publisher was urged to make his first advertisement each year a complete list of the preceding year's publications). In 1887, however, the last January issue contained a full annual list (for 1886) with reference to the full weekly listing of each title, a feature that continued through January 1911 (for 1910). The "Order List," a rearrangement of each week's listing by publisher (and of obvious value to the student of publishing) was discontinued with the issue of 1 February 1908. Though not handled consistently, the years 1900 through 1911 generally had quarterly cumulations as well as monthly and annual ones—that is, the April issue would list, not simply books of March, but those of the first three months of the year; July would contain a six-month cumulation; and October would put together the publications of July, August, and September.[57] But after October 1911 there were no more quarterly cumulations, only monthly; and the monthly lists ended with December 1918 (in January 1919). These cumulations—whether annual (through 1910), quarterly (through

October 1911), or monthly (through December 1918)—are easier to consult (as indexes to the weekly lists) than fifty-two weekly ones (and better, too, since they contain titles not present in the weekly record), but after 1918 (until 1935) one has no choice but to search each weekly list (not facilitated by the weekly title index instituted 1 July 1922). In July of 1935 a monthly title index began, serving the function of the old monthly cumulation, but it has now been replaced (as of January 1960) with the very efficient, separately published American Book Publishing Record. For the years in which there is no other form of cumulation, the indexes to announcements of fall and spring books can be helpful (in more recent years containing predicted publication dates).

The third main tool, the Catalogue of Copyright Entries, has already been described. But one should remember the significance of entries both in it and in Publishers' Weekly—particularly since some scholars have used one or the other as direct and reliable evidence of publication dates. A book may not actually appear until long after the publication date given in the copyright record, or it may be out (in some cases) long before its entry in Publishers' Weekly. Both records are quite accurate, as much as it is in their power to be, but everything depends on the habits of the publishers. One publisher may never meet his announced publication dates; another may always meet them. One may not get his books delivered to the office of Publishers' Weekly promptly (and then deliver a large batch at one time); another may have them sent immediately upon publication. Disagreements among scholars as to the relative reliability of these records have usually arisen from the relative promptness and efficiency of the publishers on whom they were working.[58]

It is possible to construct a fairly efficient procedure, based on the nature of these tools, for establishing a chronological catalogue of a given firm's output. Such a procedure would consist, basically, of four steps. First, one should check the catalogues of the firm (for every year of its existence) in the Publishers' Trade List Annual, making a card for every title listed and keeping a record in one corner of each card

indicating the years in which that title was listed. It is essential that every annual catalogue be examined, even for those firms which attempt to list all their publications, past as well as present, for certain titles inevitably are overlooked or dropped from year to year (or transferred to another publisher—these are, after all, order catalogues). Thus if one looked only at the last catalogue or two of B. W. Huebsch (which do in fact list most of his publications), one would miss Carl D. Thompson's Municipal Ownership, Melville Anderson's Happy Teacher, and Robert Aitken's Beyond the Skyline. Record should also be made on these cards of any other information supplied in the catalogue, such as descriptive remarks about new books or the fact that a book is "forthcoming" or "Ready in November." The completed set of cards will, in this way, be a presumably full listing of a firm's output, as announced by the firm itself, with an indication of the length of time each book remained in print. Since these catalogues more often than not are arranged alphabetically by author, it is an easy matter to finish this first stage of work with the cards also in alphabetical order. Before proceeding to the second step, one should take the time to make a typed list of these authors and titles: this list will serve as a brief reference guide for the imprint, since the cards will not be in alphabetical order again but will be transformed through succeeding stages into a chronological arrangement; it will also serve as the basis for the index to the imprint catalogue and will obviate a final re-alphabetizing of the cards.

This much accomplished, one's second stop should be at the English Catalogue shelf. The arrangement there is alphabetical by author in five-year cumulations. One can tell from the first date recorded on each card the probable publication year, and one simply starts through the cards checking each title that falls within the span of the English Catalogue volume for the earliest years of the firm's existence. For each of these titles published in England, one should record the publisher and month and year of publication (and perhaps the number of pages, a rough clue—later—to whether or not the English and American sheets are identical). Any good catalogue of an American imprint must contain information about English publication and its relation to the American publication by the firm involved, and this information might as well be entered on the cards while they are in alphabetical order. If the cards falling within each volume of the English Catalogue are separated from the others after being checked, one not only will be sure of having checked all the cards but will have them arranged alphabetically within five-year groupings. It is true, of

course, that certain titles coming near the beginning or end of each five-year span, if not located in the first volume consulted, should be checked in the preceding or succeeding volume, since they may have been published in England several months before or after their American publication. Common-sense shortcuts, too, suggest themselves: if one knows a particular author to be British, for example, one expects to find that his books have been published earlier in England than in America.

The third stage of this process, for publishers after 1891 (and especially after 1906), is a search of the Catalogue of Copyright Entries; for publishers in the two preceding decades this step must of necessity be omitted, but in those cases some attempt should be made to secure copyright information directly from the Copyright Office. One starts through the cards again, this time checking the titles for any one year in the index to the appropriate annual volume and entering the index numbers (in use through 1945, as described above) on the cards. Any copyrighted work will be listed under its author, but any work copyrighted in the name of the publisher will be entered also (through 1938) under his name; it is therefore simpler to check the authors' names first, so long as one does not fail to glance at the publisher's listing before leaving the index, to make sure that no books are recorded here which were not announced in the publisher's catalogues and for which one thus has no card. The cards are then arranged by index-numbers, the index-numbers looked up, and further data placed on the cards: accurate title-page information, number of pages, size, price (sometimes), official date of publication, date of deposit of copies, date of affidavit, and name of copyright holder. Naturally one checks the unlocated items, as before, in the next succeeding volume. At this stage one's cards are roughly in chronological order according to copyright (publication) date.

The fourth round takes one to Publishers' Weekly, where one's task is lighter or heavier depending on the kind of cumulations made in the years involved (as explained above). In any case one's job is to locate the "Weekly Record" in which each title is listed. If the year is one with an annual cumulation, the work is relatively painless; if it is one without even monthly cumulations, it is indeed exasperating, but the chronological arrangement of the cards offers some guidance. For the years before 1908, with the weekly "Order List," one can search each issue without great effort and should certainly do it, as a further check on the possibility of previously unnoticed titles. The date of the

"Weekly Record" will be placed on each card, with any other new information found in the listing (generally, the price and a brief description of the contents of the volume). Before each volume of Publishers' Weekly is closed, there is a general, as well as an advertisers', index to be consulted. Each of the firm's advertisements should be studied and the date of each advertisement for any title entered on its card (along with such phrases as "Ready Soon" or "Published last Tuesday").[59] It is only after one has accumulated all these dates—copyright date, deposit date, date of Publishers' Weekly listing, and advertising dates—that one can make an informed judgment about the actual time of the appearance of the volume; and it is likely that very little further shuffling of the cards will be necessary to place them in the most exact chronological order. The final task is simply to number the cards consecutively and enter the numbers on the alphabetical index sheets already prepared.

Unless one is particularly fortunate, however, the process does not operate quite this smoothly, and one will usually end up with a number of problem cards—titles found in the publisher's catalogues or advertisements and nowhere else, or titles first encountered in the copyright record or in Publishers' Weekly but seemingly never announced by the publisher. In the first of these cases, one may suspect that the books were announced by the publisher but never actually issued; still, one cannot be sure, since the works may not have been copyrighted and may not have been sent to the Publishers' Weekly office for record. So one begins searching for evidence of their existence—in the Library of Congress Catalogue, American Catalogue, United States Catalog, or CBI. If the LC catalogue has the book with the desired imprint, the question is answered; if not, but one of the other three has it, one must be wary. Unlike the copyright record, entries in those catalogues are not proof of a book's existence. The H. W. Wilson Company in the early years made a careful attempt to examine every book listed (a more attainable goal then than later), but in some cases it had to rely on information requested from a publisher (though never simply from his catalogue);[60] the publisher could report a book as "almost ready"

which never in fact appeared. For example, Thomas Seltzer's 1924 catalogue lists Babette Deutsch's Honey out of the Rock; his 1925 does not. The copyright entry is for 4 September 1925 in the name of the Appleton company; the LC copy has the Appleton imprint, as have the copies in the New York Public and Yale—yet the United States Catalog for 1924 lists the book as Seltzer's. A similar difficulty arises when a publisher who is currently issuing a given author's works decides to reprint all his earlier works. In such a case the publisher may announce for publication many titles which he never gets around to issuing—yet it is difficult to tell just which ones appeared (in the absence of conclusive information in Publishers' Weekly or the copyright records), since libraries are more likely to have the original editions. So the problem of tracking down the publication dates of some of the Huebsch reprints of Veblen's work is a taxing one. In addition to checking the card catalogues in as many libraries as possible, one should turn to any other reference works that list large numbers of books—such as dealers' or auction catalogues, or a work like Book Review Digest. (It may be that one will want the dates of some early reviews on each card, anyway, as a further aid to an accurate chronology—in which case one should go to this work between steps two and three above, when the cards will be alphabetical in five-year groupings; they will emerge in one-year groupings, equally convenient for the ensuing work with the copyright catalogue.)

The second possibility—that a certain title is found in Publishers' Weekly or the copyright record but not in the publisher's announcements—is less serious, for if it is in the copyright record, it definitely once existed, even if no large library seems to have it today. If the source of the title is only the Publishers' Weekly, one must be more cautious. Entries in the "Weekly Record" are generally made from the books themselves, but sometimes they are not (in these cases there is fortunately an indication that the book has not been seen). An illustration of the problem this situation can create involves Jean Parke's Psalms of the Heart Restored. Because the book was listed in Publishers' Weekly for 8 November 1919 as published by Scott & Seltzer, I had assumed it to be the first book bearing the imprint of this short-lived partnership, though I did wonder why it was not listed in the Scott & Seltzer catalogue. The book is not in the Library of Congress, but I came across it in a bookshop and discovered that it bears the imprint of Temple Scott alone. While Publishers' Weekly transcribes from title pages if a book is at hand, the listing does serve the purpose of telling book dealers where a book can be ordered, and a firm's new

name may therefore be used when the old one still appears in the imprint. (A useful feature of the PW record is the "c." which indicates that a book is copyrighted; if one has missed it in his previous search of the copyright entries, he can return.) A similar confusion in imprints may occur when, instead of inserting a cancel title page with the British sheets, an American publisher simply pastes his own label over the original imprint, as Huebsch did with some of E. D. Morel's works. Even an actual joint imprint may be recorded in America only under the name of the American publisher—which emphasizes, all the more, the necessity of personally examining every book. A final disturbing point about this second category of discrepancies is that it is conducive to nightmares in which one catches vague glimpses of still further entries, for books unannounced by the publisher himself, in some of the more inaccessible reaches of a file of Publishers' Weekly.

To compile a catalogue of an imprint, as this procedure illustrates, one must extract from the basic tools a category of information which they are not designed to yield up easily. The whole process, of course, has assumed that the publisher's own records are not available; but, even if they are, the same steps should be followed, both as a check on errors or omissions in them and as a source for information which they will usually not contain. The principal facts which only the publisher's files can supply are the dates of later printings (unless one can secure a late copy with a printing record on the verso of the title page) and the size of each printing. For the post-1870 period the papers of a great many publishers are still in their own offices. The appearance of the useful American Literary Manuscripts in 1960 revealed that at least Bobbs-Merrill, Longmans Green, McGraw-Hill, Norton, Prentice-Hall, and Viking have important holdings for the study of relations between literary authors and their publishers, but not all firms are desirous of being turned into research libraries (or even of having their papers housed in one), as the recent dispersal of material by Appleton-Century-Crofts demonstrates. If the records of some of the smaller firms of this period have perished through neglect, there is an increasing desire on the part of the libraries, vying in the manuscripts race, to acquire what remaining documents there are, while it is still possible. The Bentley Papers at Illinois, though not of an American house, are a case in point, as are the Henry Holt Papers at Princeton.[61] The New York Public Library has a number of important manuscript collections for this period, especially of the Century

Company (1880-1914), Knopf (since 1930), Crowell-Collier (since 1935), the publishers of many New York newspapers, and the National Association of Book Publishers; the Library of Congress has some Harper material, both Huntington and Houghton some Thomas B. Mosher, the American Antiquarian Society some Copeland & Day and some Lee & Shepard, and Columbia some Putnam. One may also, most of the time, track down some important manuscript material about a publishing firm by looking into the papers of its leading authors (which are more likely to be housed adequately). Thus the Sherwood Anderson Papers at the Newberry Library (like the Colonel House Papers at Yale) contain material relating to B. W. Huebsch, including a 1924 letter of Huebsch's when Anderson was considering moving to Liveright, that seems to me one of the significant documents in twentieth-century literary publishing.

As with the other categories of materials, there are far more (and far more useful) volumes of reminiscences for this period than for any previous one. The huge, basic books are Joseph Henry Harper's The House of Harper (1912) and I Remember (1934), George Haven Putnam's Memories of a Publisher (1915), George Palmer Putnam: A Memoir (1912), and Authors and Publishers (1897), George P. Putnam's Wide Margins (1942), George H. Doran's Chronicles of Barabbas (1935),[62] Henry Holt's Garrulities of an Octogenarian Editor (1923), and Walter Hines Page's A Publisher's Confession (1905) and Life and Letters (1925). Publishers of magazines have frequently furnished some of the best views of the publishing world, as L. Frank Tooker in The Joys and Tribulations of an Editor (1924), S. S. McClure in My Autobiography (1914), John Adams Thayer in Astir: A Publisher's Life Story (1910), and Edward Bok in The Americanization of Edward Bok (1920). But there are many other important firsthand accounts, among which the most revealing are Ferris Greenslet's Under the Bridge (1943), William W. Ellsworth's A Golden Age of Authors (1919), George Terry Dunlap's The Fleeting Years (1937), Robert Barrie's My Log (1917), John Barnes Pratt's Personal Recollections (1942), Theodore Bliss' Publisher and Bookseller (1941),

Ralph Fletcher Seymour's Some Went This Way (1945), Alfred Harcourt's Some Experiences (1951), E. Haldeman-Julius' The First Hundred Million (1928), Donald Friede's The Glittering Angel (1948), and Edward Uhlan's The Rogue of Publisher's Row (1956). Frederick A. Stokes, B. W. Huebsch, and Alfred Harcourt have delivered reminiscent lectures in the Bowker series,[63] and other publishers have at times written articles or sketches in a nostalgic vein.[64] At least two publishers, Alfred Knopf and B. W. Huebsch, have responded to interviews for the Oral History Research Office at Columbia, under Louis M. Starr, and have in this way provided voluminous material not otherwise available.[65]

One other related category consists of the commemorative volumes issued by various firms to celebrate their anniversaries. Some of these are quite distinguished, like the 1899 and the 1905-06 Houghton Mifflin volumes, Appleton-Century-Crofts' Fruit Among the Leaves (1950), and Knopf's Borzoi 1920 and Borzoi 1925, while others are nothing more than pamphlets presenting a brief sketch of a firm's history; but whether they represent staff hackwork or considerable research by a signed author (such as Roger Burlingame's histories of Scribner's and McGraw-Hill, Grant Overton's of Appleton, Charles Morgan's of Macmillan, or Quentin Reynolds' of Street & Smith),[66] there is an enormous array of them,[67] and they often can provide a starting point for further work. At any rate, the problem for the post-1870

period is evaluating so much evidence and cutting through so many documents, in order to build up, slowly and carefully, a history of each firm and its influence book by book.

If the emphasis in this discussion has been on publishing historiography as the construction of documented publishers' catalogues, I have not meant to imply that there are no other important areas of investigation. Mention of publishers' reminiscences and of correspondence with authors suggests in itself some of the relations between the history of publishing and the cultural, political, and economic conditions prevailing at any particular time. Moreover, a thorough picture of publishing cannot fail to notice publishers of periodicals as well as of books. Publishers and editors of avant-garde magazines, in Greenwich Village, say, in the 1910's, offered important outlets for writers, existing side by side with experimental publishers of books and occupying an important place in the publishing scene; similarly, the full story of American publishing will deal with the pioneer newspaper publishers and editors of each locality, who often served a large group of people as its sole publisher and symbol of the publishing world. Some book publishers, like Fields or Harper or Huebsch, established important periodicals, but other magazines of great significance for literary and intellectual history, like the Port Folio or McClure's or the Little Review, are important in their own right as forces in publishing and deserve to have their stories told as part of publishing history.[68] As for newspapers, certainly there is no dearth of memoirs by pioneering editors, waiting to be utilized and the data incorporated into larger patterns of frontier publishing.[69] Another area which constitutes a

whole field in itself, but with which the student of literary publishing should be familiar, is that of bibliopoly—for the autobiographies of book dealers, as well as their catalogues and those of auction houses, often prove to be convenient storehouses of information.[70]

The emphasis here, as I say, has been on the detailed history of individual imprints. It should be a truism, but apparently is not, that a knowledge of the lists of individual publishers underlies all broader investigations into publishing history. Generalizations about the out-put of any publisher, about the characteristics of publishers in any period, or about the trends in American publishing from one period to another, must begin there. When George Parker Winship, forty years ago, was surveying the scholarship of American printing history, he felt obliged to defend the calling of bibliography:

Bibliography has come a long way since then, and it is surely no longer necessary to shun the "sweeping together of scattered fragments" simply in order to demonstrate that bibliography's "own special field" consists of more than that. The fragments must in any case be swept before the "intelligible narrative" can come, and we are now in the fortunate position of being able to do the sweeping without being demoted to scullery-maids. It may even turn out that the man who has no compulsion to hide his broom will be the very man who can use it with imagination.

Notes