The 39th Rare Books and Manuscripts Preconference (Washington, June 23-26, 1998).

Getting Ready for the Nineteenth Century:
Strategies and Solutions for the Acquisition, Cataloging, Preservation, and Research Use
of Published and Unpublished Materials from the Nineteenth Century.

The Case of Lucile

If the theme of this conference were the Fifteenth Century, I expect we would find ourselves in reasonable consensus that questions of acquisition, cataloging, preservation, and research use have largely been resolved. In each case, the possibility of surprise is there, but the main issues seem well blocked. This is true as well, though increasingly less firmly, for the Sixteenth and Seventeenth, even the Eighteenth Centuries. Which is, no doubt, one of the reasons that the organizers of this conference chose to focus on the Nineteenth Century, for which all these questions remain very lively indeed.

This afternoon I want to comment primarily on two of these issues: acquisition and cataloging. The other issues come in by implication, but for nineteenth century materials we have approached acquisition and cataloging in largely traditional ways, and the case of Lucile demonstrates that these ways will not be adequate. You cannot preserve or use what has not been acquired; and it is difficult to acquire material when you can't adequately describe what you do already have. It is equally difficult for readers to access materials which are inadequately described for their purposes.

Let me start with acquisition.

I recall seeing, some years ago, in the Library of Congress Rare Book Collection, a couple of shelves of variant paperback editions of Joseph Heller's Catch-22. Terry Belanger, as most of you will know, collects multiple copies of certain 19th and 20th century titles. Students in the University of Virginia Rare Book School classes use them to learn a variety of things about the history of the book. But it was Robert Rosenthal who predisposed me to start collecting Luciles thirteen years ago. Twenty-five years ago, as Curator of Special Collections at the University of Chicago, Bob was taking in every Lalla Rookh he found on the grounds that this one title, first published in 1817 and frequently reprinted over the next century, formed a capsule history of 19th century book arts. When I last saw the run in the lower levels of Regenstein Library, 60 or 70 copies had accumulated. [2021 Note: With the ability and interest of libraries/librarians to correctly identify "editions" of 19th-century books under active discussion, I want to add this note. Between 1980 and 2010, I took advantage of the occasional opportunity to acquire a Lalla Rookh of interest for binding or condition to acquire a small number of of copies, the most interesting of them an 1829 London edition in a clearly American binding of deeply blocked full calf. In 1910, I offered them to the Rare Book Librarian at Chicago and fairly quickly got back a note from an assistant saying, "Oh we have all of these." I was skeptical and asked if they had checked apparent duplicates for condition and binding. A bit less quickly this was answered with a slightly sheepish acknowledgement that, well, in fact only one was an "actual" duplicate and they would be happy to take the lot as a gift in Rosenthal's memory.]

So there is the occasional exception. But surely most of us agree that the central chore of collection development is that of providing readers the greatest number of titles with the least possible duplication. A new edition, particularly one with scholarly notes, may merit acquisition, but we add yet another copy of the same text only when there is reasonably clear evidence that multiple copies are required to meet the needs of readers. And even these duplicates are often weeded once the need has passed.

Indeed, as a kind of paradigm, one college library I contacted in an early survey of Luciles found that they had seventeen copies on their shelves, all accessioned as gifts before 1925, none of which had ever circulated. They kept two copies for their unusual bindings and, to my gratitude, allowed me to acquire the other fifteen.

This seems to me responsible. This library very likely did not "need" these copies, and I certainly intend to do my best to see that they are well preserved. Less responsibly, I think, another library answered my questions about their one rather uncommon edition, and thanking me for bringing it to their attention, mentioned they intended to discard it as no longer in scope of their collecting policy. They did not respond to my offers to buy it, and so, I suppose, by causing it to be looked at, I destroyed it.

We certainly face massive challenges in maintaining what our predecessors did acquire -- in maintaining, in William Gass's phrase, that "shelved up present which passes for time in a library." What our predecessors did not collect, however, suggests still other dimensions of the challenge we now face in reconstructing the history of publishing during the last half of the 19th century. Individual titles may or may not have been missed, but clearly there are whole classes of books which libraries missed, or avoided, between 1860, when Lucile was first published in the United States as a Ticknor & Fields "Blue and Gold", and 1938, when the last titles in A.L. Burt's Home Library, including Lucile, were remaindered out by Blue Ribbon Books.

The records of Ticknor & Fields were rather well preserved, and they have been stored at Harvard for some time. They were preserved in part because the firm was recognized at quite an early date as a publisher of American writers of lasting importance. The records were also preserved because, after several reorganizations, separations, and mergers, the final parts of Ticknor & Fields were absorbed by Houghton, Mifflin & Co. in 1889, and Houghton Mifflin Co. itself has had a long and fairly stable history down to today.

From the records of Ticknor & Fields and its successor firms, we know that just under 100,000 copies of Lucile were printed between 1860 and 1878 -- six different editions, two of them illustrated, in a total of 83 printings. Since five of the six editions were offered in two to four bindings -- cloth, half calf, tree calf, antique morocco, and the cloth was often of two or three colors with green, plum, and terra cotta favored -- there may be as many as 500 identifiably different [platings], printings, states, issues, and binding variants. The variation was probably not quite this great: a total of 250 to 300 variants seems more likely.

Of the 98,382 copies thus printed and bound, I know the location of 154, which collectively represent 49 variants. The Book Arts Press collection holds 15; I have 14; the other 125 are scattered among the 225 libraries which have reported holdings to me. Some variants are relatively well-represented -- 11 of the 2,550 copies in the first, 1860, printing are known -- but many are known in only one or two copies. For example, there were three printings of the second illustrated edition, a total of 4,050 copies, issued in paper wrappers. Just three copies are known, two with an 1871 imprint, one with an 1872 imprint.

To put this another way, as a library community, we can offer access to 15% of the copies Ticknor & Fields are known to have printed; and to between 15 and 20% of the variants. There are, doubtless, copies "out there" among booksellers still to be collected; and in this case, the characteristics of many of the variants not reported are predictable: they were printed from electrotyped plates, so the text blocks differ mainly in titlepage date; and the same binding plates and stamps were almost certainly used and re-used. Nonetheless, locating even single copies of the "missing" 80 to 85% is a daunting assignment. You can't preserve what you can't first acquire.

Until the late 1870s, Ticknor & Fields and its successors enjoyed something of a monopoly on Lucile. The 1860 edition and the first illustrated edition of 1868 were brought out in cooperation with the author's London publisher, Chapman & Hall. Ticknor & Fields enjoyed a reputation of making payments to foreign authors, something many, even most, American publishers of English writers did not do at this time, so the author, Owen Meredith, may have gotten a chunk of change from his American sales. [2021 note: He did: a penciled note in the T&F ledgers indiates a payment of $100.]

Foreign titles had no American copyright protection in this period, but American publishers did, by and large, respect a trade practice of not poaching on each other's titles. This gentleman's agreement began to break down in the 1870s, and it was gone, gone, by the end of the 1880s. The Copyright Act of 1891 for the first time extended copyright protection to writers in other countries if those countries, in turn, extended it to American writers, but it offered no protection to titles by foreign writers published before its enactment. A title with a juicy sales record, like Lucile, was fair game for anyone who took the effort to produce an edition. Some 75 American publishers did make that effort, each bringing out one, a handful, a dozen or two, or a couple of hundred editions between 1880 and 1920.

Before I proceed with that part of the story, however, some of you may be saying, Lucile? The book is a novel-length narrative poem composed of iambic pentameter lines [See footnote 1], about 6,000 of them, more or less rhymed, couplet by couplet, relentlessly through description, letters, dialogue, and even footnotes. The plot, based on a novelette by George Sand, concerns a wealthy young widow, Lucile, Comtesse de Nevers, who ten years before the story begins had courted with the English Lord Alfred Vargrave, who is about to marry another woman, Miss Matilda Darcy -- though mutual feelings remain. Now living on the Continent, Lucile is herself being aggressively courted by the French Duke Eugene de Luvois, and these people, with a small supporting cast, trek round Europe, from resort to spa to hotel, city to city, missing, meeting, and corresponding with one another. The treks offer Meredith much room in which to paint poetically the splendor of the Pyrenees and the Alps, storms and other scenes and events, all the while inserting philosophic asides about life's many excitements and disappointments.

After quite a lot of this, Lucile rejects the Duke and becomes a nursing nun, Soeur Seraphine. Some years later, in the Crimea, she finds an English soldier recovering too slowly from his wounds because the guardian of his French sweetheart opposes their marriage. The soldier is of course Vargrave's son, the French girl du Luvois's ward, and Lucile accomplishes reconciliation all round.

Owen Meredith, the name under which Lucile was published, was the son of novelist Bulwer-Lytton. Discouraged by his father from writing, Lytton entered the diplomatic service for a long and distinguished career. Among other postings, he was Viceroy of India in 1877 and supervised the proclamation of Victoria as Empress. In addition to Lucile, he cranked out several other, equally long, narrative poems and much miscellaneous verse, none of which found an audience like the one Lucile found in the United States.

I still have too much data to analyze to give you with full confidence an estimate of the total number of identifiably different printings and bindings in which this one book appeared between 1860 and 1920. But as a working estimate, take the 250 to 300 we can attribute to Ticknor & Fields before 1878. Add another fifty produced by the successor companies, including Houghton Mifflin (which kept at least one edition in print until 1925).

Other prolific producers included Henry Altemus of Philadelphia, A.L. Burt and H.M. Caldwell of New York, and the cluster of companies owned by John W. Lovell. These firms add no fewer than 200 editions to the total. Some of the remaining 65 firms produced a dozen or two editions; many produced only a few; some only one. Say they average 5. That adds 325. The most prolific producer was Thomas Y. Crowell of New York. As you see from my handout, Crowell described 120 octavo and quarto editions and 62 Handy Volume editions between 1879 and 1931. There is doubtless a bit more variation in illustrations, frontispieces, and titlepages than was accounted for in these 182 descriptions. Round to 200. That brings the total to a staggering -- I think staggering -- 1000 to 1100 variants. [2021 Note: I'm now confident the number exceeds 2000; my collection has just passed 1400, with very few "duplicates".]

One goal of what I've come to call The Lucile Project is to collect or record at least one copy of each variant. Again I have too much data still to analyze to say precisely what we lack, but it is a lot. I have 210 unique Luciles, the Book Arts Press has about 300, including about 75 I do not; and I have some the Book Arts Press lacks. The several hundred records from the 225 reporting libraries include more than a hundred variants which neither I nor the Book Arts Press hold. So, roughly, we can account for 400 to 450 variants, something less than half the total I estimate were produced.

A very substantial number of variants are known in only one copy, that one copy often in a library's circulating collection. Apart from a few variants in clearly splendid bindings, most Luciles are nice examples of turn-of-the-century book craft but not obvious candidates for rare book collections. That is troubling because nearly a third of the 250 copies reported to the National Union Catalog of Pre-1956 Imprints had disappeared or been rebound when I surveyed them in 1990. And it is perhaps important to note that 100 libraries reported just under 250 copies to NUC Pre-56. With duplicates eliminated, this is a miniscule portion of the 1000-1100 varients produced.

To repeat, 80 publishers overall have been identified as producing one or more editions of Lucile. For thirteen of them, 16%, no copies whatever are known. In some cases, this may be because one publisher took over another publisher's stock and did not bother to change titlepages. The copy of a Winn & Hammond edition, uniquely reported to NUC Pre-56 by the Detroit Public Library, has been lost. I cannot yet quantify my impression, but I think the smaller format, less expensive editions produced by publishers like Altemus, Caldwell, and Conkey will prove the most elusive.


Lucile is, of course, only one of thousands of titles published between 1860 and 1920, and it is highly unusual in that for 40 years it was a standard "gift" book, one of a fairly small number of titles which, presented at Christmas, New Year's, or most any occasion, were certain to be welcomed. Since the text was fixed, gift books competed on price and on embellishment. Prices ranged from ten cents for Handy Volume editions in decorated cloth to twelve dollars for carefully printed, well illustrated volumes bound in full leather. Embellishment was a matter of balancing cost with splendor. Editions which offered more or better illustrations, color frontispieces, decorated pages, and showy or novel bindings competed well against more dowdy presentations, particularly if they were less, or at least not more, expensive. Hence they rolled out, new and novel each year (or so the publishers promised), in cloth stamped with gold and inks in every style designers could imagine, in padded leathers stripped from a great assortment of animals, in limp ooze, in circuit calf and tree calf, in levant and turkey morocco.

I had thought initially to illustrate this talk with slides, but on reflection it occurred to me that not only would it take time but it would be redundant: recall any turn-of-the-century book you've handled. At least one Lucile would look pretty much like it.

If price and embellishment were central concerns, dating this stuff was not, and hence the cataloging problem. A few publishers specialized in elaborate editions and dated them, notably Frederick A. Stokes who brought out a series of increasingly ornate editions during the 1890s, the last designed by Will Bradley. But after 1880, most copies of Lucile are not dated. An imprint date, in fact, was probably a disadvantage as it would be harder to sell "old" stock in the face of new. The way in which editions are added to and removed from the Crowell list, for example, suggests that a particular version was kept in stock until sold out, which might take a year, or several years.

Sorting out the immense, mostly undated, production of Luciles is another of the goals of The Lucile Project. The fundamental difficulty is that the differences between undated editions is more often than not graphic, rather than textual, and no number of words is sufficient to describe what you have in hand with confidence that there is not another copy to which your description would equally well apply but which is, in fact, quite different. This part of the project, therefore, is based not only on surviving copies but on the publishers' catalogs collected in each annual volume of PTLA, the Publishers' Trade List Annual.

PTLA was created in 1873 by Frederick Leypoldt, who the previous year had founded the trade magazine, Publishers' Weekly. Leypoldt's intent was to establish a base on which a full catalog of United States imprints might eventually be founded, and his idea was straight-forward. Each year ask every American publisher to submit by a specified date a specified number of copies of a catalog of their publications. Collate the resulting catalogs, provide indexes and other preliminary matter, and bind the result into a volume which could be distributed back to booksellers, libraries, and other potential book buyers and distributors. The volumes are relatively uniform, page size about 8 by 11 inches, but grow thicker with each passing year as both individual catalogs and the number of publishers submitting them grow larger. By 1900, the volumes are nearly a foot thick and unwieldy in the extreme.

Content varies. Some catalogs were little more than lists with price and order information. Others offer substantial descriptions of individual books and series. Relatively few are illustrated in the 1870s; by the late 1880s, however, more and more are illustrated with line cuts and half-tones, most often of bindings, later dustjackets, but also of authors and manufacturing facilities. It is this fact which is making it possible to identify and date a substantial percentage of the many editions of Lucile which were published between 1885 and 1920.

To look in on this work, check the Lucile Project website. Here I am pulling together, publisher by publisher, information from PTLA and surviving copies of Lucile, linking descriptions and scans of PTLA illustrations to those surviving copies. Lucile was rarely the single book in an edition or series. Typically it was one of a dozen or so titles, but series sometimes ran to hundreds of titles, usually bound in uniform style. Consequently, as the project files become more complete, it will be possible not only to place Luciles accurately but also to identify other titles in the series. The Altemus and Burt files are as complete as I can now make them, and you might want to check undated imprints you hold against them. I'd be delighted to hear from anyone who finds this useful.

When the files are more complete, we hope to invite the cooperation of the booktrade, asking booksellers to check their stock against the publishers' files and offering me or the Book Arts Press any copies not yet in hand. These copies will come to Tulsa where I can scan and add them to the Project website. Meantime, if you run across Luciles, remember that they will find a welcome home with me or in the Rotunda at the University of Virginia.


Getting Ready for the Nineteenth Century: the case of Lucile, generalized to hundreds of similar titles, suggests that we must prepare now to add several miles of shelving in our stacks, adjust our acquisitions budgets, and invent efficient and effective ways to include images in our cataloging. The opportunities for cooperation are obvious.

One more word on the Publishers' Trade List Annual. Since relatively few publisher's catalogs survive, and those which do are widely scattered, PTLA is an invaluable resource for research of late 19th century imprints. During the 1970s, the Meckler Corporation published a microfiche edition of PTLA, 1903-1981. In 1996, the Guide to Microforms in Print priced this at $14,700. I have not been able to discover why the 1873-1902 volumes, highly valuable for all sorts of studies, were not selected for filming, and I do not know whose run was sacrificed for filming. The microfiche set lacks at least some pages. Meckler has transformed to Mecklermedia and moved on to electronic products, and I do not know the current location of the master microfilm.

An informal survey in 1995-1997 conducted via the Exlibris listserv suggested that relatively complete runs for the pre-1920 volumes are far from common. Full or nearly full runs of PTLA have so far been identified at the American Antiquarian Society; The British Library; Brown University; University of California, Los Angeles; Cornell University; University of Illinois; University of Indiana; University of Michigan; University of Minnesota; the Newberry Library; The New York State Library, and Rutgers. The University of California, Berkeley and Stanford, I'm told, divide a fairly long run. Libraries as large as the University of Texas, Austin, have only scattered volumes before 1900. [2021 Note: In the 1960s, the University of Iowa sent its run to the Center for Research Libraries, as did, one guesses, a number of other libraries. For a great more information on the history of PTLA], including an index of contributing publishers 1873-1947, see the PTLA tab in banner of this page (or click here).

The paper on which the catalogs submitted to PTLA were printed varies greatly in quality, and most of these large volumes are quite fragile. If someone needs an enormous, but enormously useful, scanning project, PTLA 1873-1902 awaits.


[2005 note: in 1999, I understood the verse scheme to be iambic pentameter couplets. This is not the case. An iamb is a metrical foot consisting of a one short followed by one long syllable, or of an unaccented followed by an accented syllable. In Lucile, each line has 12 syllables in a pattern of two unaccented followed by an accented syllable -- an anapestic meter. The verse can also be described as "duodecasyllabic rhyming couplets” (or “rhymed anapestic couplets” or even “anapestic couplets” or anapaestic tetrameter). Meredith takes liberties with both meter and rhyme: some lines strain the regularity of the scheme!]

Last revised: 14 February 2021