University of Virginia Library

Search this document 


expand section 
expand section 
expand section 
expand section 
collapse section 
[section 1]
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 

expand section 

From time to time bookseller's catalogues offer Vanity Fair in parts with more or less extensive descriptive notes and references to Van Duzer's 1919 description of his Thackeray library.[1] Occasionally the descriptions refer to other known copies of the novel for comparison. One such catalogue refers to "the Austin copy" as "the finest copy of 'Vanity Fair' ever offered at auction in America" and ends its headnote on the current offering as "One of the Two Finest Copies Ever Offered at Auction in This Country."[2] Who owns the finest copy of Vanity Fair is a question likely to interest book collectors more than literary critics. And as long as fineness is determined by references to the condition of the wrappers and the priority of advertisements unconnected with the novel's text, it will remain a question of little importance to the critic. But it is an important question to both collectors and critics, and this for reasons hitherto unknown.

The basics are, of course, well-known: the first edition[3] consists of 20 numbers in 19 separate installment parts, each with 32 pages of text, additional leaves of advertisements, and printed yellow paper wrappers. The last installment contains numbers 19 and 20. The first number was published on 1 January 1847 and the final double number was published on 1 July 1848. Following completion of the serial issue, left-over sheets from that issue along with newly reprinted sheets from the same typesetting were used for publication of the novel in one volume. The publisher's accounts record multiple printings of this edition and continuous


Page 119
availability of the novel in parts and book form throughout the 1850's and 60's.

This summary of the production of Vanity Fair gives a false impression of simplicity. And whether because as scholars and book collectors we prefer simplicity to facts or because the facts have been too difficult to ascertain, both scholars and book collectors have operated under certain naive assumptions which need dispelling—to wit: the idea that any given physical copy of Vanity Fair whether in parts or volume form might as a whole belong to a single printing (the production process of mid-nineteenth century serialized novels makes this as likely to be untrue as to be true); or the idea that the "printing," "issue," or "state" of the text of a serialized novel can be ascertained by reference to points in the wrappers. Regardless of the literary significance of the variants in Vanity Fair, their existence and distribution have an important bearing on the question of "fineness" or "priority" which till now has been inaccessible to book collectors. But the variants do, also, have literary significance and must attract the critics' notice. There are 210 variant readings between the printings within the first edition of which 150 are substantive, 17 being the addition, deletion, or substitution of passages ranging in length from three to seventy-five words. It would seem of some small importance to the literary critic to know that originally Mr. Jos Sedley in Vauxhall Gardens is referred to as a "fat bacchanalian" but that second thoughts delete that description,[4] or that Miss Crawley had originally signified her intention of dividing her "fortune equally" between Sir Pitt's second son and the family at the rectory, but that later the fortune is described merely as an "inheritance" to be divided without reference to proportion (77.25). And it must be of some interest to the followers of Becky's fortunes that in the original text the reason Colonel Crawley, though governor of Coventry Island, could settle on her only 300 pounds a year (out of an annual income of 3000) was that his revenues went to the "payment of certain debts and the insurance of his life," but that later the reference to life insurance was deleted (579.7-8). Indeed, both the critic and the printing historian stand to benefit as much as the book collector from the effort to determine who owns the finest copy of the first edition of Vanity Fair; for, though there is probably no satisfactory answer, the attempt at one requires that we focus attention not only on book-collecting values, but on the complex printing


Page 120
history of the book, and the artistic effects of the author's continued interest in the composition of the work during the printing processes. We can determine when those 210 variants first appeared in the text, who was responsible for them, and finally what significance they have to our understanding and assessment of the novel.

As David Randall pointed out in his 1948 notes toward a collation of Vanity Fair, there are five areas of concern for parts-issued books: front wrappers, back wrappers, inserted advertisements, plates, and text.[5] For Vanity Fair, by far the greatest interest to date has been lavished on the first three—the parts the author had the least do do with. Booksellers continue to identify their copies of Vanity Fair by reference to readings on the wrappers and by one or two famous "points" in the text which are not in fact points, one of the most often used "points" remaining unchanged throughout all printings of the first edition.[6] But these five concerns are merely the discrete parts of the product. Taken together they are the corporate results of the activities of publishing, printing, binding, and marketing, each of which affects the five-part product. In addition to the books themselves, we can bring to bear information from the publisher's records, the author's letters relative to the production, and our knowledge of how books were made in the midnineteenth century as well. Synthesizing all the evidence makes it possible to provide a satisfactory guide to the effects production history had on the novel, the specific result of production represented by a given copy of the book, and the significance the vicissitudes of production have for the student of the text and for the book collector.

Most traditional collectors of Vanity Fair probably think much of this investigation unnecessary. One recently assured me that a useful bibliography is one that identifies the printings in a simple straightforward way for book collectors. But identifying the printing or state represented by any given copy of Vanity Fair is seldom possible. It is often possible to identify the printing or state of each sheet or gathering within a given copy of the book, but most copies of Vanity Fair as wholes represent hybrids or mixtures of sheets from different printings—as a natural result of the methods of manufacturing and marketing the book.[7] When each monthly part was published, a certain number of


Page 121
copies of the part were printed from type. As each stack of printed sheets dwindled in the bindery, it was replenished by reprinting the depleted numbers. The result was that some parts were reprinted more often than others (see Chart I). The book was available in parts for years after it had also become available in book-form, and the book-form itself was bound in small lots ordered as needed for distribution; hence sheets from one printing can appear indiscriminately mixed with sheets from other printings in any given book—regardless of its present form, whether parts or volume. If a particular copy of the book has an early title-page and prefatory material the rest of the book may yet consist of sheets printed late or a mixture of early and late. Even if parts-issue had ceased once the book as a whole became available, one need only be reminded that eighteen months elapsed between the printing of part one and the printing of the title-page, and that, according to the publisher's records, during those eighteen months parts 1 and 7-13 were each reprinted once.[8]

Whereas I have been able to identify the order in which readings were introduced to the text and have given the readings in that order in the appended table, they may appear in confusing disorder in any given copy of Vanity Fair. Thus, in using the variants lists as a means of identifying a given book, each gathering must be identified separately. Furthermore, the Vanity Fair owner must be warned that the difference between parts and book-form issues was created by the bindery, not the press. In other words, the mere fact that a copy was issued in parts is no guarantee that the text is an early printing, nor does the fact that a copy is in book-form mean that it is composed of later printed sheets.

But enough of complexities. Following the completion of the parts issue on 1 July 1848 and during the succeeding fifteen years, Bradbury and Evans issued the first edition of Vanity Fair continuously in parts and volume forms. Each monthly part was printed at least six times, twelve parts were printed seven times, and two parts were printed eight times—prepared according to the schedule in Chart I. The novel was stitched into parts and bound in book-form for distribution according to the schedule in Chart II.

I. The first printing of each number was run off from standing (moveable) type. Positive identification of printings from type can be


Page 122
made by measuring the length of any printed line extending from margin to margin. If the measurement equals or exceeds 3 and frac1316ths inches (9.7 cm), the page was printed from standing type; copies printed from stereotyped plates measure 3 and ¾ths inches (9.6 cm) or less.[9]

States of the first printing were created when certain changes were effected in the standing type before stereotyped plates were cast. Though it is not clear whether these changes are stop-press corrections producing two states of the first printing or if they represent second printings from type, the former seems to me the more likely in view of certain evidence to follow concerning stereotyping. The changes in the standing type are listed in columns one through two of Chart III. In one instance, that of signature X in part number 10, three "states" of the "first printing" are distinguishable (see Chart III, 309.3 and 309.11).

In the publisher's account books, charges for stereotyping were entered at the same time as those for initial composition and corrections, and it is conceivable, perhaps even probable, that the stereotypes were cast immediately after the first printing from standing type. If that were the case, one would expect to find relatively commonly copies of Vanity Fair in parts with numbers 1 and 7-13 printed from stereotypes since these numbers were reprinted months before re-issue in book-form (see Chart I). In fact, however, both the normal complexities of original production and the rapacity of unscrupulous bookmen have produced copies of Vanity Fair in parts with sheets printed from stereotypes. The worst instance I know of the latter is the Heineman copy in the Pierpont Morgan Library which is expertly repaired so that only careful examination reveals that seven of the gatherings were made up by combining pages from at least two different copies—one printed from type and one from stereotyped plates—and that three additional whole gatherings printed from stereotypes were probably supplied surreptitiously: a classic case of collecting wrappers and advertisements rather than texts. However, not all "mixed copies" are aberrations; the publisher's records show that Vanity Fair was always available in parts so that as the years went by new-sold parts would contain late-printed sheets. Similarly, copies in volume form sometimes contain early, maybe first-printing, sheets either as a normal result of binding schedules or because some purchaser of early parts had his copy rebound in volume form. Nevertheless, certain copies (not all in parts) composed mostly of sheets printed from type have been noted with numbers 1 and 7-11 both printed from


Page 123
stereotyped plates and sharing all the readings characteristic of the last "state" of the printing from type.[10] Not only does this pattern correspond with the publisher's records, but it demonstrates that the major alterations in Vanity Fair, for these numbers at least, were effected after stereotyping. It is likely that similar copies of the book in parts also exist with numbers 12-13 printed from stereotypes but with readings corresponding to the final state of the first printing.

It is misleading to speak of a second printing of Vanity Fair, for not all parts reached a second printing at the same time. There was, nevertheless, a short time in which the publisher was producing books with a characteristic combination of first-printing sheets and second-printing sheets which it is tempting to refer to as a second stage of production.[11]

II. A. The second printing of numbers 1 and 7-13, apparently issued in combination with first-printing sheets of the rest of the book, is printed from stereotyped plates but shares the readings of the final state of the printing from type (see column three in Chart III). Numbers printed from stereotypes but with readings agreeing with the printing from type can be distinguished from type-printed numbers by the measurement described above (i.e. type-printed pages measure 3 and frac1316ths inches or more horizontally from margin to margin while stereotype-printed pages measure 3 and ¾ths inches or less). This second printing of numbers 1 and 7-13 can sometimes be distinguished from other, later stereotype printings by the readings in column three of Chart III.

Though the publisher's records give no clear indication of when the majority of substantive variants were introduced,[12] the existence of numbers 1 and 7-11 printed from stereotypes with unaltered readings proves conclusively that changes in those numbers post-date the stereotyping and suggests a similar pattern for the rest of the book. In addition, the evidence of machine collation, though inconclusive, seems to suggest


Page 124
that in numbers 2, 3, and 5 a few changes not actually noted in any copy of the book printed from type may in fact pre-date stereotyping.[13] It is possible that these changes were made in standing type which was not then used again before stereotypes were cast and that the further changes made after stereotyping were effected before the plates were used the first time. Hence, there may never have been copies of the book with the earlier changes only. The importance of the machine-collation evidence is that it suggests a priority of changes and supports the notion that interest in changing the text was continuous with Thackeray, not merely a one-time or haphazard concern. The changes that appear to pre-date stereotyping are all corrections or alterations of some magnitude and are probably authorial. (See asterisked entries in Chart III.)

II. B. The second printing of parts 2-6 and 14-20 was made from stereotyped plates incorporating considerable alterations and seems to have occurred at about the time of the third printing of parts 1 and 7-13. See Chart I for the rapidity of reprinting during the eight months following the serial's conclusion (July 1848 through February 1849). This second printing of parts 2-6 and 14-20 can be distinguished from the first printing by the measurement indicated above (i.e. the first printing is from type, the second from stereotypes) and by the readings listed in Chart III, column three.

III. The third printing of parts 1 and 7-13 was made for the most part from corrected stereotyped plates. However, the "correction" of signature Y entailed the removal of a woodcut dropped into the text. The result was that the text from that point to the end of the chapter had to be moved up 9.7 centimeters. Rather than trying to cut and move stereotyped plates, the printers reset the whole of pages 366-340. Thus,


Page 125
the first pages of signature Z are affected in the same way. At least one printing, probably the third, was run off from corrected plates except for pages 336-340 for which type was imposed along with plates. The readings characteristic of this printing are recorded in Chart III, column four.

IV. The fourth and later printings by Bradbury and Evans are not distinguishable from one another and are only occasionally identifiable at all. The publisher's records show at least six printings for each part, seven printings for twelve parts and eight printings for two parts. Signatures Y and Z are again significant, for the pages printed from type in the third printing were then stereotyped for subsequent printings and though there are no distinguishing readings, the type-page measurements indicate stereotyping.

The title-page, printed with parts 19-20, is the only part that can be distinguished in more than four printings. According to the publisher's records, the final part (containing numbers 19-20, the preface, and title-pages) was printed seven times by Bradbury and Evans before Smith, Elder, and Company acquired the stereotypes and back stock in 1865. The Bradbury and Evans printings occurred three times in 1848, once each in 1849 and 1855, and twice in 1864. Unlike the text, the title-page seems to have been reset for each printing, and six different settings have been identified. Though it is impossible to determine the precise order in which the title pages were prepared, it is beyond question that the first one given in Chart VI is the first printing. It seems likely that the two printings dated 1849 precede some if not all the others dated 1848 because the books they belong to have readings pre-dating readings in copies with the 1848 date. However, as I have taken pains to show, the state of one sheet cannot be used as evidence for the state of other sheets in the same book because mixtures of early and late sheets were a normal product of the bindery. Since each of the title pages was entirely reset, the apparent carry-over of a characteristic from one to the next would be fortuitous, not indicative of order.

When Smith, Elder, and Company acquired the copyrights and the back stock of stereotypes and printed sheets from Thackeray's other publishers in July, 1865, they acquired from Bradbury and Evans 5,001 copies of "various numbers" of Vanity Fair in parts and two copies bound up. Over the next eight months they printed enough "various numbers" to bring the total to 17,392 which were made into 865 copies of the book bound in cloth with 92 numbers left over. Though I have not yet encountered a copy dated 1865, a charge for new titles was entered


Page 126
in the account books for November of that year. A copy dated 1866 is at the Simon Fraser University Library, Burnaby, B.C., Canada.[14] The records also show an August 1868 printing of 250 copies, again with a new title page dated 1868; there is a copy (not personally examined) in the British Library. It is curious that Smith, Elder reprinted the first edition of Vanity Fair in 1868, for that is the year they brought out an entirely new edition of the book as part of a collected edition of Thackeray's works.

What does it all mean? For the book-collector, I hope the historical record reveals the true nature of the copies of Vanity Fair he owns, and that it establishes a range of representative copies he may seek to add to his collection. For the historian of printing, I hope the record with its combination of evidence from publisher's records and the books themselves confirms his sense of the complexity of the economic, technological, and artistic confluence that commercial book publishing was. For the textual editor the implications are patently clear. For the literary critic and student of Thackeray's works, I hope the record of variants provides a basis for further understanding Thackeray's attitude towards his text and supplies evidence leading to a clearer understanding of his concerns in the novel. Some changes seem to have relatively obvious motives: the elimination of Dobbin's lisp (all but one instance) at 50.18 and 105.18-19 is clearly intended to improve the image of the book's only gentleman. And the deleted reference to life insurance at 579.7-8 is the mere correction of an error, since at 501.3-4 Rawdon was unable to qualify for life insurance at all. But other changes suggest more subtle motives. Why, for example, did Thackeray find it inappropriate, at 50.11up, to call Jos a "fat bacchanalian"? And why, at 75.29, does he change Lady Muttondown to Lady Southdown and then decide, at 500.35, that it is also inappropriate to refer to Lady Southdown as Lady Macbeth? Why, at 40.10-11, is the sentence "And what can Alderman Dobbin have amongst fourteen?" (so reminiscent of St. John's account of the feeding of the five thousand) deleted? And why is Mrs. Blenkinsop, at 55.14-15, no longer allowed to opine to Pinner, apropos of Becky, that governesses are "neither one thing nor t'other." If nothing else, the alterations focus attention on passages that were, somehow, not right and for which the new readings are, somehow, better. But this article is not the place for critical speculation or evaluation; I only hope


Page 127
that this presentation of the evidence of revision will stimulate critical assessments providing a clearer understanding of Thackeray's achievement in Vanity Fair.[15]