College and Research Libraries Automating Bibliographic Research: Identifying American Fiction, 1901-1925 Geoffrey D. Smith, Diane Vizine-Goetz and Edward T. O'Neill The American Fiction Project of the Ohio State University Libraries was the focus of a coopera- tive research effort between Ohio State and the OCLC Office of Research. The purpose of the American Fiction Project was to improve access to Ohio State's early twentieth-century Amer- ican fiction collection and to identify and acquire new titles for the collection. For its part, OCLC identified catalog records in the OCLC database-the Online Union Catalog-for items suspected to be within the project's scope (American fiction published from 1901 through 1925). The results and analysis of the sample records suggest that the methodology employed can be of great assistance for the compilation of comprehensive bibliographies and for collection development. ~~ merican fiction has long been a ~ ~ focus of collection development M_~ at the Ohio State University Li- braries. Ohio State's William Charvat Collection of American Fiction (fiction from 1787 through 1900) compares favorably with similar collections at the American Antiquarian Society, Yale Uni- versity, the Library of Congress, Harvard University, and the Huntington Library. Ohio State's holdings are particularly strong in the 1876-1900 period. Addition- ally, Ohio State possesses a generous number of American fiction titles (includ- ing many first editions) from 1901 through 1925 in its general collection. In the late 1950s, this collection was further strength- ened by the purchase of a large collection of American literature from the Library Company of Philadelphia. Many titles of American fiction, especially publications from the first quarter of the twentieth cen- tury, were included in this acquisition. Recognizing the significance of its early twentieth-century American fiction collec- tion, Ohio State began a project to estab- lish a comprehensive bibliographic file of fiction for the period and to improve bib- liographic access to the collection. The project is referred to as the American Fic- tion Project (AFP). 1 Its specific purpose is to identify, acquire, and catalog, in full ac- cord with national standards, the corpus of American fiction published from 1901 through 1925. 2 At present, Ohio State's holdings in this subject area-more than 10,000 titles-are rivaled only by the hold- ings of the Library of Congress. Providing ready access to the materials will be of sig- nificant value to scholars of literary his- tory, publishing and printing history, and book illustration and design. Moreover, since popular fiction expresses social, po- litical, economic, and religious attitUdes of the times, the American fiction collection will also be useful for research in other hu- Geoffrey D. Smith is Curator of The William Charvat Collection of American Fiction at Ohio State University Libraries, Columbus, Ohio 43210. Diane Vizine-Goetz is Research Scientist and Edward T. O'Neill is Senior Research Scientist at OCLC Online Computer Library Center, Dublin, Ohio 43017-0702 . 252 manistic disciplines, such as history, soci- ology, anthropology, and folklore. The Office of Research of the OCLC On- line Computer Library Center, in a collab- orative research effort with the Division of Special Collections at Ohio State, identi- fied catalog records in the OCLC database-the Online Union Catalog-for items suspected to be within the project's scope. As a result of these efforts, many ti- tles of American fiction that have been overlooked in previous scholarly research will become part of the canon of American literary and cultural history. The cooperative effort of OCLC and Ohio State proved beneficial to both par- ties: OCLC was able to compare its biblio- graphic database against an established bibliographic file; Ohio State was able to use the OCLC database to refine and fur- ther enrich its bibliographic file of Ameri- can fiction for the period. THE AMERICAN FICTION PROJECT BIBLIOGRAPHY As a consequence of the effort to orga- nize its current holdings and acquire new titles from the period, a comprehensive bibliography of American fiction is being created by Ohio State. The criteria for se- lection of titles in the bibliography are modeled after Lyle H. Wright's biblio- graphic work in earlier American fiction, i.e., first American book appearance of adult fiction by American authors during the period. Included are "novels, short stories, tall tales, allegories, tract-like tales, and fictitious biographies and tra- vels."3 AFP' s extensive working bibliography was compiled from a number of sources, including R. Glenn Wright's Chronological Bibliography of English Fiction in the Library of Congress through 1950 and the PZ 1 shelf- list (short fiction) of the Library of Con- gress. 4 This file was enriched through comparison with the Huntington Li- brary's file of American fiction, 1901-1930 (on loan to Ohio State through the cour- tesy of the Huntington Library), the New York Public Library's Dictionary Catalog of the Research Libraries, copyright records, and listings in Publishers Weekly. More than 350 additional bibliographic Automating Bibliographic Research 253 sources have been examined for confirma- tion of existing AFP records and discovery of new records. These sources include standard guides to American studies such as Bibliographical Guide to the Study of the Literature of the U.S.A. 5 and Tanselle' s Guide to the Study ofUnite_d States Imprints; 6 regional bibliographies such as An Anno- tated Bibliography of California Fiction, 1664-1970 7 and Ohio Authors and Their Books; 8 genre bibliographies such as The Checklist of Science-Fiction and Supernatural Fiction9 and The Novels of World War I: An Annotated Bibliography; 10 and author bibli- ographies including A Golden Anniversary Bibliography of Edgar Rice Burroughs11 and A Bibliography of Christopher Morley. u Nu- merous cultural and historical studies, critical studies, biographies, and book dealer catalogs were also consulted. The working American fiction biblio- graphic file is a manual file that comprises thirty boxes of index cards. The file con- tains catalog records of appropriate items and records of authors and titles deemed inappropriate for inclusion in the bibliog- raphy. (Inappropriate records are desig- nated 11 omits.'') Thus, the working biblio- graphic file contains more records than the approximately 131 000 titles confirmed to be first American printings by Ameri- can fictionists from 1901 through 1925. The majority of II omitted'' records are for foreign authors, collections of anecdotes, juvenile works, jest books, folklore, es- says, periodical fiction, and reprints of earlier American fiction. The omissions follow the criteria established by Lyle Wright in his work on pre-1901 American fiction. The omitted titles and authors are important to the file, for substantial re- search is often required to determine whether or not an author or title is appro- priate for inclusion. By retaining "omits" in the working file, redundant research ef- forts can be prevented when the omitted authors or titles appear in other biblio- graphic sources. THE OCLC DATABASE AS AN AID TO LITERARY RESEARCH AND BIBLIOGRAPHY Catalog records in the OCLC database that met the general requirements of AFP 254 College & Research Libraries were identified by OCLC. The offline search was limited to OCLC machine- readable cataloging (MARC) records in the books format entered into the data- base prior to July 1985. Some records, however, containing cataloging that was judged to be incomplete or unreliable were excluded from consideration. Exam- ples of such records are (1) minimal-level records, (2) cataloging-in-publication rec- ords, (3) order-level records, and (4) error- and-warning-level records. A search was performed using a record- identification algorithm based on the proj- ect's title selection criteria: the first Ameri- can book appearance of adult fiction published between 1901 and 1925. The al- gorithm used the data elements classifica- tion number, fiction code, publication date, and juvenile work code in a series of tests to identify candidate records with the following characteristics: (1) fiction, (2) published during the specified period, · and (3) not a juvenile work. The algorithm was applied as follows: 1. Database searches were performed to find all records classed as American fic- tion according to the Library of Congress Classification (LCC) or the Dewey Deci- mal Classification (DDC). The class num- bers used are given in tables 1 and 2. When a record contained both an LCC number and a DDC number, the LCC number was preferred. Records that did not match on classification number were rejected except for unclassed records. Rec- ords that contained no classification num- ber were retained because many libraries, especially public libraries, do not classify fiction. 2. Unclassed records and records that TABLE 1 LffiRARY OF CONGRESS CLASSES FOR AMERICAN FICTION Oass Number/Range PS 991-3369 PS 3500-3549 PZl PZ3 OassNumber Description American authors, 19th century American authors, 1900-1960 Fiction in English, Collections Fiction in English, Individual authors 1750-1950 May1987 TABLE2 DEWEY DECIMAL CLASSES FOR AMERICAN FICTION Oass Number/Range 813 817 818 OassNumber Description Fiction Satire and humor Miscellaneous writings matched on classification number were then checked to determine if the item was published between 1901 and 1925, inclu- sive. Records that passed the publication date test were then checked to determine if the record was for a juvenile work. Rec- ords for juvenile works were rejected. Records with the specified classification numbers were saved for listing. The re- mainder of the algorithm was applied to the unclassed records. 3. The fiction code was checked to de- termine if the record was for a work of fic- tion. Records identified as fiction were saved for listing. , 4. An unclassed record not identified as fiction was listed only when it met all of the remaining tests: a. language is English; b. published in the United States; c. not a government document; d. not a conference publication. Records for reprints were intluded in the listings even though AFP' s biblio- graphic file is limited to first editions. There are several reasons for this decision. A copy of the original edition may not be held by an OCLC member institution. It is even possible that no copy of the original edition exists, and the reprint may be the closest derivative of the original edition. More likely, however, the work is held by a member institution but was cataloged long before OCLC was formed. Therefore, no bibliographic record for the item would be found in the OCLC database unless the record had been entered retrospectively. In any case, the information obtained · from the reprint is often sufficient to iden- tify the original edition using other biblio- graphic tools. Even when the original edition is avail- able, information about reprints estab- lishes a book's popularity and forms the basis for determining the complete biblio- graphic transmission of a text, an impor- tant aspect of a complete, ideal bibliogra- phy. Nevertheless, a reprint edition does not always indicate wide sales or circula- tion of the original edition. For instance, limited editions of organizational or church materials might warrant reprint- ing, or privately published novels may eventually be released by a sympathetic publisher as reprints (though actually the first trade edition). Records identified by the database search were printed alphabetically by au- thor and title, with some variation due to machine alphabetization capabilities. For example, diacritical marks divided an au- thor's surname, resulting in alphabetical arrangement by the syllables preceding the diacritical mark (e.g., Benet becomes Ben et, arranged alphabetically following Below and preceding Benchley). Also, the grammatical articles (a, an, and the) were included as first words in titles. Records were confined to one line with the follow- ing information in sequence: OCLC num- ber, date of publication, author's sur- name, and title (see figure 1). Computer listings were further divided by classifica- tion: LCC PS and PZ: DDC 813 and 817-818; unclassified with fiction indica- tor; and unclassified without fiction indi- cator. The output format was suitable for com- parison to AFP' s bibliographic file, which is also arranged alphabetically by author and title. Since the printed output, even in abbreviated form, numbered more than 900 pages, listings of full OCLC record printouts were impractical in terms of con- venience, cost, and time. For rapid com- parison of OCLC records with AFP rec- ords, the abbreviated records provided sufficient data in the vast majority of cases (although, in retrospect, full authors' names would have been desirable) and al- lowed for easy record keeping. Finally, the full OCLC record could be retrieved by OCLC number when additional informa- tion was needed. EVALUATION OF THE OCLC SEARCH RESULTS The results of the search of the OCLC database for potential catalog records of Automating Bibliographic Research 255 American fiction published between 1901 and 1925 were analyzed by the Ohio State University Libraries. To ensure a thor- ough analysis, a sampling of approxi- mately 10% of the AFP' s card file (three of thirty boxes, arranged alphabetically by author and extending from A through Brou) was compared with the equivalent sampling of the OCLC records (A through Brou, 4,798 author and title entries of 49,408 total entries). The remaining 90% of the records will be checked in a more cur- sory manner, i.e., looking for new titles only. What might appear as an onerous analysis (4,798 title entries) was mitigated by experience. The AFP bibliographer, fa- miliar with the fiction of the period and American literature generally, was able to eliminate many titles on sight inspection of the OCLC listings. Reprints of works by popular authors such as Louisa May Al- cott or John Kendrick Bangs were recog- nized as inappropriate to project scope. In many cases, nonfiction titles (poetry, drama, essays, biographies) were either known to be nonfiction or described as such in the title or subtitle (e.g., Selected Poems of Craven Langstroth Betts or Jacob Leisler: A Play of Old New York). Juvenile works, such as L. Frank Baum' s Rinkitink in Oz, were also easily identified. Since fic- tion can be assigned any title by its author, however, no record was omitted on sight inspection unless the bibliographer was certain of its inappropriateness for the AFP file. The 10% AFP sample file yielded 1,247 titles appropriate for inclusion in a bibliog- raphy of American fiction within the es- tablished criteria. The corresponding OCLC sample contained 910 of the 1,247 titles, a 72.9% hit rate. Since the AFP file has been developed over a five-year pe- riod, a 72.9% hit rate of appropriate titles is significant. Further analysis of the sam- ple file indicated that 1,033 titles were de- rived from R. Glenn Wright's Chronologi- cal Bibliography of English Fiction in the Library of Congress through 1950 and 214 ad- ditional titles from research of other printed sources by AFP staff. A comparison of the OCLC catalog rec- ords with the bibliographic records de- rived from R. Glenn Wright's bibliogra- OCLCNo. 2645001 6121457 6362903 4444725 9220720 6320521 1492062 4555113 273503 3732966 8933691 5298416 4558222 780059 3690902 2335416 981133 908609 4885659 728704 *Indicates reprint Date 1921 1909 1908 1910 1920 1912 1924 1901 1920 1913 1910 1904 1919 1912 1920 1912 1907 1901 1905* 1923 Author Angell Angert Annesley Annesley Annie Laurie Ansell Anthony Anthony Anthony Appleton Arctander Armstrong Arnold Arnold Arnold Arundel Atherton Atherton Atherton Atherton Title Smiles: An Original Comedy Drama in Four Acts Is Mark Twain Dead? The Door of Darkness Wind Along the Waste Roses and Rain Happy Houses Razzberry A Victim of Circumstance: A Novel Margaret Fuller The Moving Picture Boys in the West Guilty? Incense of Sandalwood Miss Emeline's Kith and Kin The Hermit of Lover's Lane The School of Sympathy MotorBoat Boys on the Great Lakes Ancestors: A Novel The Aristocrats The Bell in the Fog, and Other Stories Black Oxen FIGURE 1 AFP Notations Omission in AFP file (Drama) Omission in AFP file (Essay) Omission in AFP file (British author) Omission in AFP file (British author) New omission (Inspirational literature) New omission (British author) Title in AFP file Title in AFP file Omission in AFP file (Biography) Omission in AFP file Guvenile) Title in AFP file New omission (British author) Title in AFP file New title Omission in AFP file (Nonfiction) New omission Guvenile) Title in AFP file Title in AFP file Title in AFP file (Reprint counted as omission) Title in AFP file Sample printout of records identified by OCLC database search phy, however, indicates further the efficacy of bibliographic utility database searching as an initial step in the creation of broad-ranging bibliographic files, such as a period or genre bibliography. As men- tioned above, the majority (approximately 83%) of records in the AFP file was derived from Wright's bibliography. The Wright bibliography is a reproduction of the Li- brary of Congress' shelflist, completed in 1973, for fiction of the English-speaking world. Researchers examined each record in Wright to determine its potential for inclu- sion in a bibliography of American fiction published between 1901 and 1925; made copies of possible inclusions, which were mounted on index cards and filed by au- thor name; researched pertinent biblio- graphic sources to ascertain eligibility for the file; checked Ohio State's online cata- log for institutional holdings; and, finally, coded the cards with additional, relevant information discovered during the re- search process. The OCLC search results contained 970 of the 1,033 Wright titles in the 10% sample, or 93.9% of the original American fiction card file. Compilers of a comparable bibliographic file (e.g., Amer- ican fiction, 1926-1950), then, could ex- pect, prior to consulting printed sources, a base file of OCLC records to include a high percentage of appropriate titles. These statistics indicate two major points: traditional research of printed bib- liographies are essential for establishing a thorough bibliography of American fic- tion; but, online databases, such as the OCLC database, can aid in establishing a significant bibliographic file (72. 9% of the titles discovered over a five-year period) in about one quarter the time. DISCOVERY OF NEW TITLES For AFP, the discovery of new titles was preeminent during the analysis of the OCLC search results. Thirty new titles were discovered-seven by examination of complete OCLC catalog records and re- search of in-house reference tools (e.g., National Union Catalog, Publishers Weekly) and twenty-three by inspection of texts or photocopies obtained through interlibrary loan. To date, an additional eleven titles Automating Bibliographic Research 257 could not be confirmed as either fiction or nonfiction after in-house research. The unconfirmed titles are being requested through interlibrary loan for inspection. The results of the analysis are shown in ta- ble3. TABLE 3 ANALYSIS OF OCLC SAMPLE Titles in AFP file 910 Omissions in AFP file 3,152 Duplicate editions 91 Authors in AFP file under variant name 21 New omissions 583 New titles 30 Unconfirmed titles 11 Total 4,798 The confirmation of thirty titles, previ- ously unrecorded after years of intensive research, is noteworthy. Twenty-six were by authors not previously in the American fiction file. The titles were either not listed as fiction in the many traditional printed sources or, in one case, not listed in any of the sources consulted, including the Na- tional Union Catalog. For instance, neither National Union Catalog nor National Union Catalog Supplements showed a record for Mary E. Bell's Mammy's Story and Verses. 13 In short, the thirty titles might well have remained unidentified as American fiction if not for the OCLC database and this proj- ect. Projecting on the 10% sample of a total of thirty new titles, AFP can anticipate un- covering approximately 300 new titles through OCLC research efforts. Of the thirty new titles identified to date, twenty were from LCC PS, six from LCC PZ, two from unclassified records with fiction indi- cator, one from unclassified records with- out fiction indicator, and one from DDC. Since the AFP bibliographic file was compiled principally from the LC PZ shelflist, a greater number of new titles from LCC PS, as compared to PZ, would be expected, though six titles from PZ is not insignificant. The new titles from PS can be attributed, in part, to classifying principles that require a work to be classed with the predominant subject or form. Classifying a work as poetry, essays, mis.:. cellanies, etc., often inhibits the identifica- tion of fiction included in the text. In some 258 College & Research Libraries cases, fictional pieces may be distinct works with no relation to the overall struc- ture of the text; in other cases, the fiction may be an integral and unifying element of the work's total concept. One method for identifying LCC PS for fiction, outside of OCLC' s search of its database, would be to screen the LC shelflist, as R. Glenn Wright did for LCC PZ (ten years of re- search according to Wright). 14 Two new ti- tles from the unclassified catalog records , justify searching unclassified materials. The one new title in the DDC listings is ac- counted for, principally, by the order of the OCLC search that used LCC classes when both LCC and DDC numbers were present, although the DDC records may yet yield more results in the complete analysis. SIGNIFICANCE OF NEW TITLES Equally important as the number of new titles located is the importance of the texts and themes themselves. Several of the new titles address pertinent literary, his- torical, and cultural themes of the period that are of special interest to scholars. Adam Abet's Social Conscience15 addresses ethnic intolerance and nationalism on the American home front during World War I. In a more positive vein, Samuel Hopkins Adams, in The Beggar's Purse, 16 provides a humorous account of the need for fiscal re- straint and sacrifice at home while the doughboys are fighting the war to end all wars. More than fifteen years prior to the Prohibition Act, Kate Anderson, in A Fight Against Odds, 17 dramatized the prevalent view toward the destructive influences of alcohol. Occasionally, an author presents a sympathetic, though often patronizing, view of black America's strides toward equality, as Isabella Andrews does in The Failure of Cunningham. 18 More often, how- ever, writers of the period perpetuated the plantation myth and stereotypes of super- stitious, indigent blacks, as does Julia Keirn Baker in The Wandering Joy19 and C. H. Beazley in Crackerlings and Caramels. 20 In The Inscrutable Woman, 21 Edward David Baron depicts the ethnic prejudice against numerous immigrant groups. Though most of these works are examples of nei- ther gre~t literary craft nor social enlight- May 1987 enment, they are reflective of American sentiment of the period and may actually represent the majority view from the com- mon man's experience. Many of the newly identified titles ema- nate from special interest or regional and local presses. Other presses, in addition to those cited for the works above, include Billington Press of Ossining, New York; L. Graham Co. of New Orleans; Hunter and Co., Richmond, Virginia; Stockton Press, Baltimore, Maryland; J. P. Bell Co., , Lynchburg, Virginia; and Broadman Press, Nashville, Tennessee. Identifica- tion of little-known presses may initiate investigation into regional publishing ac- tivities as distinct cultural statements about local perceptions and attitudes. In the case of both subject matter and re- gional expression, these new titles will be all the more valuable to scholars because of their omission from traditional printed sources. In addition, G. Thomas Tanselle has noted publishing history's "undeni- ~ able relevance . . . for meaningful literary history''; he notes further ''that we do not have a complete inventory of regional im- prints, nor histories of all significant pub- lishers, small and large, nor analyses of all book-publishing centers, nor edited ver- sions of all important records of publish- ing firms." 22 Archives of the major publishing houses ;- of the period are generally available. Houghton Mifflin, for instance, maintains its own library of company archives and the National Union Catalogue of Manuscript Collections lists numerous additional mate- rials about Houghton Mifflin, including papers of company officers and corre- spondence of authors. The archives of Harper and Brothers, located at Columbia University, are now available on micro- - film for the years 1817-1914.23 It is likely, however, that few if any records exist for small presses from three-quarters of a cen- tury ago. The products of those presses, ... the books themselves, may be the sole, primary sources of diverse regional and sectarian publishing activities. In addi- tion, since these small presses, by finan- cial restrictions alone (not overlooking their specialized market) seldom, if ever, advertised or announced titles in the trade journals (principally Publishers Weekly and the Publishers' Trade List Annual), the OCLC search results will be a substantial beginning for investigation of regional publishers of fiction. For instance, the in- ordinate number of local southern pub- lishers may suggest an effort by those businessmen to promote a different view of American society than the one the mainstream, eastern publishing centers were providing. The addition of these ti- tles to a bibliographic file of American fic- tion, then, should aid scholarly research into local and regional historical culture; for, addressed to limited, local audiences, and hence concerned with local issues, re- gional fiction is less inhibited by social and political constrictions that affect the sub- ject matter and presentation of nationally distributed fiction of the period. THE AMERICAN FICTION PROJECT BIBLIOGRAPHIC DATABASE Titles appropriate to project scope are cataloged (6,000 titles to date) on the OCLC system and entered onto a discrete database of machine-readable biblio- graphic records maintained by Ohio State. Cataloging for the project goes beyond traditional cataloging information to in- clude "rare book" transcription of the Automating Bibliographic Research 259 pagination; provide access points to names of publishers, printers, illustrators, and binding designers (when applicable); and denote established sources of biblio- graphic reference. The American Fiction Project, an ongo- ing project, will add the new titles to its database, which, when completed, will be available to institutions and individuals on computer output microfiche as well as magnetic tape. SUMMARY From American Fiction Project's per- spective, OCLC' s search results proved useful for enrichment and refinement of AFP' s bibliographic file. The results and analysis suggest that the methodology employed can be of great assistance to bib- liographic scholars. Furthermore, OCLC and AFP' s cooperative research and anal- ysis will have very real benefits for the li- brary and scholarly community in the form of the American fiction database. The joint endeavor will be a distinct contri- bution to literary research in terms not only of increasing access to materials, but also in developing innovative research methods that will influence the direction of future scholarship. REFERENCES AND NOTES 1. The American Fiction Project's Final Technical Report to the U.S . Department of Education is available on ERIC, ED 260 730. 2. From April1983 through March 1985 the American Fiction Project operated under two one-year grants from the U.S. Department of Education's Strengthening Research Library Resources program (Title 11-C of the Higher Education Act) in order to purchase and begin full cataloging of American fiction from the first quarter of the twentieth century. 3. American Fiction, 1876-1900: A Contribution toward a Bibliography (San Marino, Calif.: Huntington Library, 1966), p.ix. 4. R. Glenn Wright, Chronological Bibliography of English Fiction in the Library of Congress through 1950 (Boston: G. K. Hall, 1974). 5. Clarence Gohdes and Sanford E. Marovitz, Bibliographic Guide to the Study of the Literature of the U.S.A., 5th ed. (Durham, N.C.: Duke University Press, 1984). 6. G. Thomas Tanselle, Guide to the Study of United States Imprints, 2v. (Cambridge: Belknap Press, 1971). 7. Newton D. Baird and Robert Greenwood, An Annotated Bibliography of California Fiction, 1664-1970 (Georgetown, Calif.: Talisman Literary Research, 1971). 8. William Coyle, ed., Ohio Authors and Their Books (Cleveland, Ohio: World Publishing, 1962). 9. E. F. Bleiler, The Checklist of Science-Fiction and Supernatural Fiction (Glen Rock, N.J .: Fire bell Books, 1978). 10. Philip E. Hager and Desmond Taylor, The Novels of World War I: An Annotated Bibliography (New York: Garland, 1981). 260 College & Research Libraries May 1987 11. H. H. Heins, A Golden Anniversary Bibliography of Edgar Rice Burroughs (West Kingston, R.I.: Donald M. Grant, 1964). k, 12. P. Alfred Lee, A Bibliography of Christopher Morley (New York: Doubleday, Doran, 1935). 13. Mary E. Bell, Mammy's Story and Verses (Baltimore, Md.: Stockton Press, n.d.). 14. R. Glenn Wright, V.1, p.iii. 15. Adam Abet, Social Conscience (Bridgeport, Conn.: Co-operative Publishing Co., 1920). 16. Samuel Hopkins Adams, The Beggar's Purse (Boston: Smith & Porter, 1918). 17. Kate Anderson, A Fight against Odds (Chicago: Woman's Temperance Publishing Assn., 1903). 18. Isabella Andrews, The Failure of Cunningham (Richmond, Va.: Hampton Institute, 1908). 19. Julia Keirn Baker, The Wandering Joy (New York: Broadway Publishing, 1910). 20. C. H. Beazley, Crackerlings and Caramels (New York: Broadway Publishing, 1910). 21. Edward David Baron, The Inscrutable Woman (New York: Broadway Publishing Company, 1910). 22. G. Thomas Tanselle, ''The Historiography of American Literary Publishing,'' Studies in Bibliogra- phy 18:3-5 (1965). 23. The Archives of Harper and Brothers, 1817-1914 (Cambridge and Teaneck, N.J.: Chadwyck- Healey Microfilm Edition, 1980). ACQUISITION PERSPECTIVES 1. ANY BOOK IN PRINT . . . . means delivery to your library of all available books from any publisher or distributor in the U.S. or Canada. There is no list of publishers you must check ... WE DELIVER THEM ALL ... including trade, scientific/techni- cal, text, university presses-, paper- backs, associations, small presses, Canadian, and software. Send us your direct orders and ex- perience the nadded value" received when using Book House. CALL TOLL-FREE TODAY 1-800-248-1146 In Canada & Michigan CALL COLLECT (517) 849-2117 OCLC Vendor No. 17397 SAN 169-3859 the BSOK H IJSE JOBBERS SERVING UBRARIES WITH ANY BOOK IN PRINT SINCE 1112 208 WEST CHICAGO STREET JONESVILLE, MICHIGAN 49250 WILSONDISC OFFERS • CD·ROM Access Search each database on a separate compact disc. And, with the touch of a button, you can update your search online-at no additional charge! • Online Access For access to the most current data, subscribers to each WILSONDISC database re- ceive unlimited online search time in that database. You pay only the nominal hourly telecommunications rate-no additional charge! Four Search Modes for Users at All Levels 1 BROWSE Mode The same access points as the familiar Wilson printed indexes. 2 WILSEARCH ® Mode Easy-to-follow menu options simplify search formulation. 3 WILSONLINE® Mode The same commands available with the WILSON LINE online retrieval system. 4 EXPERT Mode Expanded screen handling, windowing functions, and more. WILSONDISC Databases Each database is available as a separate compact disc updated and cumulated quarterly. • Applied Science & Technology Index • Art Index • Biography Index • Busmess Periodicals Index • Cumulative Book Index • Education Index • General Science Index • Humanities Index • Index to Legal Periodicals • Library Literature • Readers' Guide to Periodical Literature • Social Sciences Index Unlimited Online Searching with Your WILSONDISC Subscription Subscribers to a WILSONDISC database can also take advantage of unlimited online searching m that database for the entire term of their subscription. The only additional fee is the nominal WILSONDISC telecommunica- tions charge. SPECIAL OFFER! Demonstration Disc Available Thespecially-pricedWILSONDISC Demonstration Disc contains six months of data from 16 Wilson databases, and allows all of the CD-ROM search capabilities available with WILSON DISC. The Demonstration Disc is available at $99, deductible from your first invoice for a regular WI LSONDISC database subscription. To Order Call Toll-Free: 1·800·367·6770 D In New York State call 1-800-462-6060 ; in Canada , call collect 212-588-8400. THE H.W. WILSON COMPANY