here are more questions than than answers about the production of books in Victorian England. Thackeray dealt with some of the most prominent publishers in the business, about whom books have been written, and yet not enough is known about the detailed work methods or the economic network of book production.1 The problem stems in part from the fact that histories of publishing have focused on the glamorous workings of the shop front - the relations between publishers and authors and booksellers, many of whom were publishers themselves. Frank Mumby and Ian Norrie's Publishing and Bookselling recounts the origins and highlights of publishing and the progression of publishing firms through inheritance and mergers, but just what happened to a manuscript when it came into a publishing house and how it was transformed into a book is a story never told straight through. Instead it tells how anonymous manuscripts turned out to be by so-and-so, how such and such an author felt about the copyediting, how another was interested in the choice of type fonts or binding, how decisions were made about another title page, and so on. Such anecdotes are valuable in suggesting the processes involved, and they suggest questions about who designed the title pages, who did the copyediting, what training and education were common among compositors and pressmen, and what specific production tasks, if any, were undertaken by the business partners.
In addition, histories of printing have concentrated on machinery, guild rules, trade unions, and papermaking and have recounted these matters in terms of trade and economic forces separate from the roles of individual workers in the production of individual works. The generalizations are very useful for an overall picture but frustratingly short on [146/147] specific information for specific printers in specific years. The idea that not much changed in the practice of printing from the seventeenth through the nineteenth centuries in spite of the advent of machine presses and steam power allowed Marjorie Plant in The English Book Trade: An Economic History of the Making and Sale of Books and Philip Gaskell in A New Introduction to Bibliography to omit detailed analysis of the passage of a manuscript through production in a Victorian publishing and printing house. Indeed, such specific information seems not to exist anywhere to be recounted. The works of these students of printing history do, however, indicate the questions to be asked.
The first major question has to do with division of labor, What jobs were related to the publishing, printing, and bookselling businesses, and who did them? These three divisions were just emerging as separate businesses in the nineteenth century, and they merged almost as often as they separated. The firm of Bradbury and Evans, for example, began as a printery and developed publishing as a later part of its growing business. Smith, Elder began as a trading company, supplying books as well as other materials to the colonies, then developed publishing, and fifteen years later acquired a printery.
The importance of other related businesses becomes apparent from the question of division of labor. At the two ends of book production there must, obviously, be writers and book purchasers (one supposes the latter to be readers, too, but for trade purposes it does not matter what is done with the book once it is purchased). Between these two stands an industry requiring financial backing and a multitude of skills.
How publishers dealt with authors and, to a degree, with booksellers is fairly well documented.2 How they dealt with the manufacturers of paper and ink and with typefoundries, who did the purchasing, how much [147/148] shopping around was involved, how many employees had regular work at the printshop and bindery is less well known.3 It is difficult to know also whether traveling salesmen or agents worked for the publishers, for the book retailers, or for themselves. And yet all these questions concern the conditions of book writing, publishing, and reading and have a bearing on why books were written and published as they were and what markets were reached effectively.
From the anecdotes about publishing and book production and from the histories of publishing houses and of printing, bookbinding, and bookselling a list of jobs related to production can be made. What cannot be known is whether a different person performed each of these or whether one worker was employed for a combination of them.
Once a decision to publish was made, the manuscript was given to a compositor for typesetting. Surviving manuscripts for Thackeray's works, obviously used as setting copy, do not have copy editor's marks or book designer's specifications. Presumably, then, the publisher or the shop foreman or whoever acted as the book's designer talked it over with the compositors, who then implemented the plan as they set type. Each compositor imposed his own version of the house rules, if there were any, onto the portion of the manuscript assigned to him as he set it.4 Compositors could set type directly into pages with running heads and page numbers or as galley slips of pure text. And then proofs would be pulled. We do not know, for Bradbury and Evans or Smith, Elder, precisely who read the proofs, but it seems clear that the compositor, who was paid by the amount of type he set, not by the number of hours he worked, was charged for any mistakes he had made. Since making corrections was not a profitable way to spend time, the compositor had to balance his desire for speed with the demand for accuracy and a need to anticipate corrections in the manuscript called for by the author or other proofreaders. Thus, a [148/149] good compositor had to be a good copy editor. Any eccentric authorial usage was likely, therefore, to be smoothed over by the compositors' editing instincts under conditions that required speed, not deliberation over textual nuances and innovations
Proofreading, according to descriptions of some printeries, was often done by boys reading against copy. It is not known if that ever occurred with a Thackeray work. Though, with the exception of a few fragments of Vanity Fair, proofs for his works before 1857 do not survive, Thackeray did get proofs; he may have seen second corrected and even third corrected proofs before printing began. It is clear that he had frequent access to the premises of his major publishers, Chapman and Hall, Bradbury and Evans, and Smith, Elder, so it is possible that in the time-honored tradition recounted in Henry Esmond when Dr. Swift encounters Henry at the printer's house presumably on a similar errand, Thackeray often read proof in the shop.
The compositor would correct the type and turn it over to the pressmen for printing. Most of Thackeray's work up through 1857 was printed by Bradbury and Evans - even those works published by Chapman and Hall and Smith, Elder. We do not know how many presses the firm had or what kind they were. How fast, for example, could a number of Vanity Fair be produced? Markings on the manuscripts show that one installment would be divided up among as many as four or five compositors who could finish the work easily in a day. Proof corrections probably took longer to point out than to implement. If final corrections came into the printing house as late as the twenty-ninth of the month for a part that had to be available for sale on the first day of the next month, there were at most two days for the operation. Occasionally the ledgers show charges for night work. Pressing and binding and distributing were all that remained to be done. Two presses, each mounted with both inner and outer forms for one sheet, could print 2,500 copies on one side; the sheets were then turned end on end and perfected on the other side. Each printed sheet would then have two identical copies of a gathering for the book. Cut in half, 5,000 copies would result. If handpresses were used producing, say, 250 copies per hour, the work would take twenty-two hours. If some of the sheets were perfected before the full 2,500 copies were printed on one side, there would be copies available for binding well before the printing was completed. Presumably the wrappers, steel-plate illustrations, and advertising materials could have been prepared in advance or printed at other presses. Bradbury and Evans was obviously a substantial printing establishment, though perhaps not as big as Clowes's printing factory that [149/150] in 1866 employed 568 men and had over fifty operational presses [Plant, pp. 357-58]. Many other firms had eight to ten presses, but Bradbury and Evans must have had more because it was publishing and printing Punch weekly, as well as serials by Dickens and Surtees at the same time it was bringing out Vanity Fair, and the company continued to be printers for other London publishers.
And so the printed sheets went to the bindery. It is very likely that there was a bindery on the premises at Bradbury and Evans. The practice up until the 1830s had been for publishers, whether or not they were also the printers, to announce to the trade that a work would be available in quires (unbound) on such and such a day. Wholesalers and agents would crowd in, buy, and take off to be bound the copies they hoped to sell to libraries and retail booksellers (Mumby, p. 19). But by the 1840s most publishers were supplying books in their own cloth bindings. The ledgers show that to be the case with Thackeray's works published by Bradbury and Evans.
This summary would not be complete without commentary on stereotyping, for that process produced a number of bibliographical complexities of real interest. Eight years after the publication of Pendennis, Henry Bradbury, son of Thackeray's publisher, explained the increased use of stereotyping: "The demand for literature had so materially increased, that to print off large editions on speculation of immediate or eventual sale, or to keep whole works standing in type for indefinite periods to meet demands for reprints, was equally hazardous: in one case there was an uncalled-for outlay of capital invested in paper and print; in the other, there was an unnecessary accumulation of works standing in type, and the consequent loss of the repeated use of these types — a principle quite in opposition to the true spirit of the first invention of moveable types. " Stereotyping, Bradbury contended, avoided both hazards. Earlier, however, T. C. Hansard had warned in a treatise on printing that an "extravagant notion prevails ... of the exceeding economy of stereotyping; it would therefore be of advantage to give a fair view of the case. On an average, the cost of stereotyping may be taken as the same as that of composition, or even higher. It is quite clear, therefore, that if the first edition of a work is stereotyped, the speculator at once incurs the expense of printing two editions, minus the press-work." [Bradbury, pp. 32-33; Hansard, p. 130] In opposition to [150/151] Hansard's estimate, number 1 of Pendennis cost £4.15 to compose, £3.17.6 to correct, but only £4 to stereotype. Furthermore, the phrase "minus the press-work" includes the savings in paper. The first pressrun of a work for which stereotypes have been cast can be much smaller, for additional printings can be produced at no extra cost in case of a sales hit. On the other hand, most books never reach a second printing, so Hansard was right to suggest that in many cases the stereotypes would prove to be an added and unnecessary expense.
While Bradbury seemed not to agree with Hansard, the actual practice in the printing of Vanity Fair and Pendennis was a compromise between the two views. Perhaps it was with an eye to economy that the printer prepared copies of the parts of these serials from type in insufficient numbers to see how sales went and if it would be worth plating the novels, but it is as likely that the printers, who had had years of experience with Thackeray, simply were not given enough time between the submission of copy and publication deadlines to allow for the plating process before the first printing. The results of the two methods of printing are visible in the quality of the product.
When a book is printed from standing type, particularly if the print run is relatively large, the type is liable to shift in the forms. The evidence of this shifting is easily revealed by machine collations done on the Hinman Collator. The fact that type was jostled and respaced during or between printings, as well as that type damage and corrections occurred, shows up as flickering images in an otherwise stable page. Some jostle of type is visible to the naked eye. Some examples, signifying nothing more than that Pendennis was printed from type, occur in volume 1 at page 97, lines 17, 18, 19, and 21, where the first two letters of each of these lines have dropped down in some copies, and at the page number for page 259, where the 9 is raised in some copies and lowered in others. Also at page 149.6 the line is loose and extends into the right margin; the letters move from side to side, leaving, in some copies, spaces with words; those noted are w ith and th e.
Stereotyping freezes such movement. Thus damage, corrections, and type shifting that show up in late printings from type often carry over into the plates. Also new changes and some restorations appear in the plates, for if the type has been used for printing, it must be cleaned and can be corrected or realigned before the casting molds are made. One surprising and at first baffling fact about Vanity Fair and Pendennis was that the type page of some copies were slightly smaller than others. This turned out to be due primarily to the difference between copies printed front type and copies printed from stereotyped plates. In Pendennis, for example, the type [151/152] page (i.e., the size of the page measured from the top of the running head to the bottom of the last line and from margin to margin horizontally) was larger in some copies than in others (by 1 ½ mm horizontally and 2 to 4 min vertically. In 1858 Henry Bradbury wrote, "The method of stereotype still in use consists in making plaster-of-Paris moulds of pages of types, and casting plates in type-metal-" The process of casting plates was given in greater detail by T. C. Hansard. A fine-grain plaster-of-paris is brushed on the types, and coarser plaster poured over that. When sufficiently set, the plaster mold is removed from the type and placed in an oven to dry and to raise its temperature to equal that of the molten lead used for the plate. In lectures on the subject, Thomas Bolas wrote, "Although plaster expands at the moment of setting, it afterwards contracts in drying, and a plaster stereotype is therefore smaller than the original by about 1-80th linear." Elsewhere J. Southward added: "The contraction in plates taken by the plaster process is much greater than by the papier-mâché. This is owing to the shrinkage of the mould in the baking, and in the dipping-pan. In the paper process, the baking is performed while the mould is on the type; but the plaster is put into the oven by itself, and the evaporation of the moisture causes the contraction. It may be safely stated that a page of crown quarto [about 19 by 25 cm] will shrink about a nonpareil in length [about 2 mm], and a thick-lead in width [about 1 mm]." Hence, a plate made from the contracted mold would print a smaller type page than would the standing type [Bradbury, p.33; Hansard, pp. 125-130; Bolas, p.24; Southward, p.78].
Since books printed during this period — whether with type or with stereotyped plates — were printed on dampened paper, paper shrinkage during drying introduced dimensional changes that complicate simple measurement as a way to detect the use of stereotypes. In Pendennis horizontal type-page measurements vary from 9.45 to 9.8 cm, Copies printed front stereotypes range in width from 9.45 to 9.6 cm, and those printed from type measure from 9.6 to 9.8 cm in width, so that a page measuring 9.6 cm might result from either process. This problem is resolved by dealing with the evidence for one sheet or gathering at a time: in no case did a given leaf produce the 9.6 cm measurement in both its type-printed and plate-printed form. For a plate-printed sheet that varied from 9.45 to 9.55 cm, the corresponding type-printed copies usually varied from 9.6 to 9.7 cm or wider. I concluded that paper shrinkage [152/153] variation amounted to a smaller difference than did the shrinkage caused by the stereotyping process [Shillingsburg, Detecting Stereotype, pp. 2-3].
In Vanity Fair and Pendennis copies with the smaller type pages consistently show more type damage than the larger ones. The cases in which the smaller copies have superior readings show definite evidence of conscious rearrangement of type such as respacing. In identifying impressions, therefore, smaller pages (printed from plates) can be assumed to be later than the larger pages (printed from type). The undoubtedly linear progression from type to plates indicated by damage noted in copies of the books obviates the possibility of simultaneous printing from type and plates as T. C. Hansard indicated was the practice even before 1841 in the printing of the Penny Magazine [Hansard, pp. 133-34].
Whether it is worth the effort necessary to learn so little is — as Kathleen Tillotson says at the end of "Oliver Twist in three Volumes" - — a matter of taste [Tillotson, p. 132]. And as Mr. Foker says in Pendennis, there's no accounting for tastes.
Last modified 30 November 2021