In January 1675 Humphrey Prideaux wrote to his friend Henry Ellis telling him about Fell’s discovery of the after hours activities of his copper-plate printers at the Sheldonian Theatre in Oxford:

‘The presse hath often furnished me with something to tell you. You little thinke it hath been imployed about printeing Aretins postures. I assure you we were like to have had an edition of them from thence were it not that last night the whole work was mard. The gentlemen of All Souls had got them engraved, and had imployed our presse to print them off. The time that was chosen for the worke was the eveneing after 4, Mr. Dean [that is Fell himself] after that time never using to come to the theator; but last night, being imployed the other part of the day, he went not thither till the work was begun. How he tooke to find his presse workeing at such an imployment I leave to you to imagin. The prints and plates he hath seased, and threatens the owners of them with expulsion; and I thinke they would deserve it were they of any other college than All Souls, but there I will allow them to be vertuous that are bawdy only in pictures.’

It’s a good story and shows that these 17th century Oxford dons were familiar with Aretino’s Postures, whose suppression by this time was clearly not as complete as is sometimes supposed – they did not, as Lynn Lawner wrote in 1988 ‘disappear from circulation for exactly 400 years’. Aretino’s Postures, or I Modi were the subject of Robert Hellenga’s novel The sixteen pleasures, the story is of a young book-restorer from the Newberry Library in Chicago who goes to Florence after the 1966 flood to see what she can do to help, and while working in a convent library discovers the only surviving copy of the 1525 edition of I Modi, bound up with a prayer book in a later embroidered binding. I Modi began life as a suite of erotic engravings after drawings by Raphael’s pupil, the mannerist artist Giulio Romano, done in about 1524. The engraver was Marcantonio Raimondi, one of the best Renaissance engravers and one of the first to specialise in interpretative engravings of other artists works. Soon after publication of the suite of engravings Pope Clement VII ordered the prints to be burned, and Marcantonio was imprisoned. But they were none-the-less re-issued in 1525 with a printed text, a series of sexually explicit sonnets, specially written by Pietro Aretino to accompany them. There were presumably many piracies, but the only one to have survived is a single copy of an edition of 1527 with rather poor reduced woodcut copies of the illustrations, which came to light in 1928.

But in 1675, 150 years after the original publication, an edition was in circulation in Oxford and available for the gentlemen of All Souls to have copied, presumably they were working from an edition which are now quite lost. If it were only the history of engraved erotica that was a shadowy area of bibliography, we would perhaps not be surprised. What is surprising is that the whole technical history of engraved illustrations in books is almost unknown -- suppressed by a generation of literary scholars who wanted us to believe that pictures pollute the purity of the text. This afternoon I’m going to be talking about the secret history of the production of books with engravings in them.

The copper-plate printers at the Oxford Press who were attempting to print pornography after hours were almost certainly different workmen from those employed in letterpress printing, a totally different craft requiring different skills. The Oxford Press was in fact highly unusual in directly employing copper-plate printers, and this is an essential point which I want to emphasise. Printing the text and printing the pictures, if they were engraved, were separate operations, for the most part carried out in different shops. Even at the great Plantin press in Antwerp, where punch-cutting and typefounding were carried out under the same roof as letterpress printing, copper-plate printing was contracted out during the whole of the seventeenth century. It was only at the beginning of the eighteenth century, in 1703 when for some reason there was a serious decline in the standards of the copper-plate printers of Antwerp, that Balthasar Moretus ordered a copper plate printing press and began to hire his own workmen to use it. The Cambridge University Press did own a copper plate press in the late seventeenth century, but interestingly did not directly employ copper-plate printers, hiring itinerant printers to work it. These are the exceptions that prove the rule: copper plate printing for book work was for the most part carried out in separate workshops, by firms independent of the printing houses, and this was the situation for the whole of the hand-press period.

Just as the gentlemen of All Souls had got their plates engraved and had taken them to Bishop Fell’s printers to run off copies, it was normally the publisher who commission the engravings for a book and then contracted out the printing of them. Notice tht the gentlemen of All Souls are described as the owners of the Aretino plates. Sometimes authors were responsible for commissioning engraving, as Hook was for the Micrographia of 1665, and he performed the same service for Robert Boyle for at least one of his books, the New experiments and observations touching cold printed in the same year. I suspect that authors may often have retained ownership of portrait plates, which perhaps explains the curious case of the portrait of James Howell by Claude Melan and Abraham Bosse which first made its appearance in a French edition of Howell’s Dodona’s grove printed in Paris, and turns up, re-worked, in Howell’s Londinopolis, printed in London in 1657. But on the whole publishers commissioned the engraved plates, retained ownership of them, and contracted out their printing.

Engravers themselves must have owned or had access to rolling-presses for proofing, and it is interesting that it was the engraver David Loggan who supplied one of the rolling-presses to Bishop Fell’s press in Oxford in 1669. Engravers could and no doubt did undertake edition work for booksellers in the seventeenth century, but it does seem there were also specialist copper-plate printing firms. Bosse himself says in his 1645 manual that printing was not his craft, that is, he was an engraver, not a printer. In England copper-plate printers were not controlled by any guild and did not have to be members of the Stationer’s company, and very little is known of their organisation.

There is a famous invitation card of about 1685

Loving Society of Rolling Press Printers

in the Pepys Library for a dinner organised by the ‘Loving Society of Rolling-Press-Printers’, but if Pepys, who took a keen interest in engraving, had not pasted this invitation into one of his scrap books we would know nothing of this society. The invitation is to a William Hawkins who is unknown, and of the two stewards who signed the card, one, Joseph Nutting, is recorded as an architectural engraver, but the other, John Brodstock, is unknown. We do know the names of one or two other copper-plate printers of the seventeenth and eighteenth-centuries from a few trade cards and surviving records of printers and publishers, but before Bewick, in the early nineteenth century, there is no archive of a copper-plate printer, and there must have been dozens or even hundreds of firms of whom we know absolutely nothing. This is in very marked contrast with the letterpress printers who’s public and private lives have been documented in the minutest detail.

Miel Copper Plate Printer [Spectaculum naturae & artium, in vier Sprachen, Deutsch, Lateinisch, Franzoisisch und Italianisch. Erste [ – Zwente] Lieferung. Berlin, bey George Ludewig Winter, 1761–1765. John Wilhelm Meil (1733–1805) succeeded Chodowiecki as director of the Academy at Berlin. ‘Il est encore aujourd hui très estimé’ (Benezit).

A huge range of crafts is involved in bookproduction, far greater than for any other manufactured product of the pre-industrial age. The basic ones of punch-cutting, typefounding, paper-making, type-setting, printing and binding each has its range of ancillary crafts producing equipment and supplies. So perhaps we should not be surprised that even putting ink to paper might have to be done in two different workshops. The reason, of course, is that taking an impression from the raised surfaces of type and woodcut requires a different press from the one needed to pick up the ink from the lines engraved in a flat plate. Mechanically this is not perhaps as self-evident as it may seem, so it is worth quickly going through the reasons why the two kinds of press have to be different before describing their working.

In letterpress or relief printing, the printing surface is level and the non-printing areas are cut away so they do not touch the paper. In copper-plate or intaglio printing, the lines which are to print are cut into the surface of the plate to varying depths, and the white areas are produced by the polished surface of the plate. In printing a copper-plate the paper is forced into the lines to pick up the ink, and the surface of the plate is in full contact with the paper. With this in mind, we can appreciate that much more pressure will be needed to print from an engraved copper plate than from a raised surface. In printing a copper plate the load is spread over the whole area of the plate and enormous pressure is needed to mould the paper into the engraved lines; but in printing from a raised surface it is only the inked surface, a fraction of the area of a page of type, that takes the load of the press. The big difference then between printing from raised type and from flat copper plates is the very much greater pressure needed to print from copper plates. The only way to obtain sufficient pressure for copper plate printing is by the use of rollers, because with rollers the pressure generated by the press is concentrated in a line that moves across the plate, rather than being spread over the whole surface. Pressure is weight divided by area, so if the area of contact is reduced to a narrow band and the weight remains the same, the pressure is much greater.

In the printing house

Miel Letterpress

the press is operated by a team of two men, taking turns at inking -- the beater -- and laying on the paper and taking the impression -- the puller. The forme containing type and woodcut illustrations is inked by the beater using a pair of ink balls, the other pressman, the puller, then lays the printing paper on the tympan, then in one smooth movement folds down the frisket onto the tympan and the two together onto the forme; turning the rounce handle he runs the carriage under the platten; pulls the bar which turns the screw and forces the platten down onto the forme. In wooden presses not enough force could be generated to have a platten large enough to cover a large forme, so that two pulls were needed, the puller moving the carriage in between. Having pulled the bar -- there is enough elasticity in a wooden press for it to spring back of it own accord -- the puller runs the carriage out; lifts tympan and frisket; and takes off the printed sheet. Meanwhile the beater has been taking a dab of fresh ink and re-distributing the ink on the ink balls, glancing at the last printed sheet to check his work, and is now ready to ink the forme again. This is a fairly complicated process: 2 distinct movements for the beater and 9 for the pressman (I hope you’ve been keeping up, there will be a test later). Yet astonishingly speeds of 250 impressions an hour, that is one impression every 14.5 seconds, were quite normal. Working a ten hour day or more output figures at the Cambridge University Press at the end of seventeenth century were regularly between about 1500 and 3000 impressions a day, or between 750 and 1500 perfected sheets, that is sheets printed on both sides.

Almost everything is different about copper-plate printing.

Bosse 1542

There are about the same number of movements, but the whole process is much much slower, mostly because the inking process is so slow. The only early output figure I have found for wooden rolling presses is 750 sheets a day at the Cambridge University Press at the end of the seventeenth century. This is however in line with figures given much later in Berthier and Boitard’s 1836 manual, where it is stated that 400 impressions a day are standard for a three man crew, which could be raised to 700 impressions for a four man crew. Thus the Cambridge University Press figures suggest that a press crew or four men was being employed. When D.F. McKenzie analysed these figures he assumed that the payment of 5s a day was for a single workman, so he was led to the conclusion that copper-plate printing was more remunerative than letterpress printing at 1 and 6 to 3 shillings a day. But if my suggestion is right then the 5 shillings had to be divided amongst four workmen, making them worse paid than the letter press printers. In several other ways the size of the press crew is I believe of great significance to bibliography, and I don’t think it has been considered before. I will return to it later.

There wont be time this afternoon to go into the economics of copper plate printing (even if I fully understood them, which I don’t), but notice that these figures suggest that it cost about the same to machine one plate as three letterpress sheets for an edition of 500 copies. That is one plate per 24 pages of a quarto, 48 pages of an octavo. Assuming that paper accounts for half the production costs, and leaving aside composition and engraving, we can see that adding a plate to a quarto of 50 sheets (400 pages) in an edition of 500 copies is like adding 3 sheets, which adds 6% to machining costs or 3% to overall production costs. I don’t want to suggest that these are correct figures yet, but this calculation does I hope show what might be done with more data.

The CUP figures suggest a four man crew, Bosse shows three in the 1642 print, though in his 1645 manual he says nothing about the number of workmen. Let us see how the process worked with three men as shown in the print. The plate is first warmed on the brazier, which softens the rather stiff ink and makes it easier to force it into the engraved lines cut into the plate. The workman at the back of the illustration is inking the plate with a rubber, a roll of linen cloth, working the ink into the engraved lines. Superfluous ink is then wiped from the surface in two stages, first with rags and then with the palm of the hand. The first part of the wiping process is not shown; in the second stage of wiping, as we can see, the workman is using the palm of his hand using whiting -- ground chalk -- as a cleaning agent. The whiting is probably contained in the bowl on the left. (I’m told that traditionally copper-plate printers, because of their inky palms, would greet visitors with a sort of masonic hand-shake using only their thumb and first two fingers.) Wiping is the slowest process and is shared between two men in a four man press crew. It requires skill not to take the ink out of the lines, and at the same time to leave the surface of the plate perfectly clean. Since the surface of the plate produces the white areas of the image, the slightest smudge, or a fine scratch that holds ink, will show up. The plate is now nearly ready to be placed on the plank of the press for the pressman to take off the impression, but the edges still need a final wipe, and this is done by the pressman in a three man crew (since the single wiper has his work cut out to keep up with the other men) or in a four man crew by the pressman.

The simple mechanism of the rolling press can easily be understood from the engraving. The star wheel is attached to the top roller, the bottom roller is free-running, so that turning the spokes draws the plank through the press, sliding over the supports at either end and the bottom roller rotating in the opposite direction. The press is entirely constructed of wood, except for bearings for the rollers which are lined with iron plates. Above and below these bearings can be seen the packing of card to adjust the pressure, interleaved with felts to give the press some elasticity.

On the rolling press, in contrast to letterpress printing, neither the plate nor the paper is fixed to the press in any way, their positions depending on the pressman’s care in placing both for each impression. The procedure which Bosse describes in 1645 is as follows. With the blankets firmly nipped by the roller and thrown back over it, a sheet of dry paper of the same size as the printing paper is placed on the plank of the press; the inked plate is then positioned on this sheet to give the desired margins. The dampened printing paper is then laid over the plate, lining it up with the undersheet, followed by another sheet of dampened paper, the maculature, and the blankets. The whole assembly is then drawn through the rollers with a firm even pressure on wheel. The pressman then has to move to the other end of the press. The blankets are lifted and placed over the roller again, the maculature removed and the print carefully peeled off the plate with both hands. For the next impression, the press is worked in the opposite direction. The dampened printing paper is placed on a board on top of the press and the pressman is provided with a table at each end of the press to receive the printed sheets as impressions are taken first in one direction and then the other. Only one of these tables is shown in the 1642 print. At the end of the day, or the next day, they can be hung up to dry.

These details are, as I said, from the copper-plate printing manual published by Abraham Bosse in 1645. It is the first published manual, and the only original one for nearly 200 years. It first appeared as an appendix to his manual of etching and engraving

Traicte de mainières de graver en taille douce sur l’airin. Par le moyen des eaux fortes, & les vernix durs et mols. Ensembel de la façon d’en imprimer les planches & d’en construire la presse, & autre choses concernans lesdits arts. Par A. Bosse, graveur en taille douce. (Treatise on the manner of engraving on copper. By means of acids and soft and hard varnishes. Together with the method of printing the plates and of constructing the press, and other things concerning the said arts. By A. Bosse, copper-plate engraver.

Abraham Bosse (1602-1676), the pupil of Callot, was one of the best etcher-engravers of his time, and though as I mentioned earlier he modestly states that the printing of copperplates is not his own craft, he must have had considerable first-hand experience of the rolling press, and had access to the workmen who built and operated such presses. Most of the book is taken up with the techniques of etching and engraving, the part on the construction and operation of the rolling press taking up less than 20 pages, but illustrated with 6 plates. The first four plates are measured drawings of the rolling press and its parts [SLIDE 5.1-4], the fourth a view from one end of the press showing a plate laid on the plank [SLIDE 5.5], and the sixth an illustration of the pressman [SLIDE 5.6], clearly derived from the large separate print published in 1642 which we have just been looking at. John Evelyn translated Bosses’s treatise into English at the request of the Royal Society and intended to publish it as an appendix to his Sculptura of 1662, but hearing that the engraver William Faithorne was bringing out a translation he put it aside. In fact Faithorne only published a translation of the first part of Bosse’s manual, on etching and engraving, leaving out the printing part. Evelyn seems to have known this, so why he did not still publish his translation of the printing part I don’t understand. No English version was available until Evelyn’s translation was published from his manuscript as an appendix to A.H. Church’s 1906 edition of Sculptura. Evelyn’s translation is interesting because he was an amateur etcher and was in contact with printer’s in London and gives English names for the technical terms – for example he writes that the star wheel is ‘by our workemen called the Wheele’ …and the plank is ‘the Table (as our men call it)’; the plate is cleaned with ‘the Tampon or (as we in England call) Rubber’. And he says that the frame is made of oak and the rollers of walnut, details which are not mentioned by Bosse.

Had Evelyn published his translation in 1662 as he planned, we would have had a manual for copper-plate printing in English 21 years before

Moxon’s manual for letterpress printing of 1683. There were however numerous adaptations of Bosse’s treatise and translations into languages other than English. But apart from a paragraph in C.F. Partington’s The engraver’s complete guide probably published in 1825, no other original copper-plate printer’s manual was published before Berthier and Boitard’s 1836 manual. All the earlier published accounts, incuding those in the Encyclopédie and the Encyclopédie méthodique derive from Bosse. There is also the tantalising references to An essay on engraving and copper-plate printing, 1732, by J.H. Hauckwitz, a copper-plate printer, but no copies of this book are now known to exist.

Anthony Dyson has illustrated 3 surviving seventeenth and eighteenth century wooden rolling presses

Plantin

French Atelier

Science museum

perhaps someone here knows if there are others?

Wooden presses were still in use in 1836 and are described in detail by

Berthier and Boitard

Wooden press from Bertier and Boitard.

I want now to turn from what can be learnt from illustrations, manuals and surviving presses to the books themselves, and the evidence of their own production history that they contain. What I think we should be looking at is 1. the paper on which the engravings are printed; 2 how the engravings are positioned on the sheet; 3. where plate leaves are inserted in the books; 4 whether they are printed several at a time on a single sheet, or printed individually on cut leaves; and 5 how engravings are combined with text on the same sheet.

3.1 the paper on which the engravings are printed

The paper on which inserted plates are printed is almost invariably different from the paper stock used to print the text of a book and is generally of a better quality. However, ordinary book paper can perfectly well be used to print engravings as can be seen in the many books in which engravings are printed in the text, and these books are printed on paper no different from any other book. The real reason that plates are printed on different paper is not that better paper is strictly necessary for everyday bookwork, but has to do with the fact that the copper-plate printers were independent of the letterpress printers, and their paper was bought in a different way. In book printing, the paper was ordered and paid for by the publisher, or the printer who financed the publication. He paid the printer to print on the paper he supplied, and the printer did not hold his own stock of paper. By contrast I am convinced that the copper-plate printer supplied his own paper. A knowledge of paper, and the best kind to choose for a particular job, has always been part of the copper-plate printer’s expertise. Even at Oxford, a book like Walter Charleton’s Exercitationes de differentiis & nominibus animalium of 1677, which has engravings printed in the text as well as on inserted leaves, inserted plates printed on a different paper stock from the text. I think that it was normal for the copper-plate printers to purchase their own paper.

3.2 how the engravings are positioned on the sheet

There are at least three parameters for the placing of the copper plate on a piece of printing paper. Which side of the paper it is printed on, its orientation relative to the sheet, and the position relative to the edges of the paper. The first two of these parameters I have tested in only very small samples so I don’t want to make too much of this, but it does look as if there is a preference for printing on the mould side of the paper, which is not surprising as it is smoother, and for placing the plate with the chain lines parallel to the short side of the image, that is the chain lines are horizontal on an upright plate, as in a quarto book, but this is not linked to the format of the book in which the plate is inserted. I think that this orientation relative to chain lines is probably related to workshop practice, but I would not like to speculate on how this works at the moment. I only report these preliminary findings to suggest an area of investigation which might be fruitful.

The third parameter, the position of the plate relative to the edge of the sheet is certainly of significance because it tells us something about the planning of a book. When engraved plates are to be inserted in books, a sufficient margin is usually left on one side to provide a stub to be folded round the gathered sections so that the sewing goes through a fold between the plate leaf and its stub. The stub will almost invariably be seen coming out a few pages back, or further on in the book, looking as if a leaf has been cut out (alarming for the novice book-collector).

A familiar problem of using books with plates is referring to the plate while reading the text. The solution to this problem common in the seventeenth and eighteenth centuries was to bind the plates as throwouts,

Newton Arithmetica universalis, second English 1728

Same

Throwout plates are folded into the book in such a way that they can be opened out with the image thrown clear of the text block, so that any page can be read while looking at the illustration. The image itself may be no larger than the page of text, or it can be larger and folded several times, and it can be constructed by the binder or planned by the printer. If it is constructed by the binder, this is done by binding in a blank leaf, called a full apron, and hinging the plate to the outer edge. A variant is to hinge the plate to the foredge of a text leaf, but this is quite rare. In looking at English science books I have been struck by how often throwouts were planned by the printer. The copper-plate is printed to one side of the sheet leaving a blank space slightly larger than the eventual trim-size of the book so that the throwout and apron are one. There is sufficient consistency from copy to copy to confirm that this was planned at an early stage, so that the copper-plate printer must have been told where to position the plate on the sheet.

3.3 Where plate leaves are inserted in books

In the hand-press period the position and folding pattern of plates can vary from copy to copy because these features of the books construction are to some extent determined by the wishes of the owner, or the traditions of the binder. However this does not mean that that these details were not planned by the publisher or master printer. Except for books with large sections of consecutive plates, the plates were generally inserted in the quired sheets by the master printer before they were delivered to the bookseller or binder. He therefore had control over their position, and the binder’s easiest course would be to leave the plates where he found them. In fine work the binders manuals say that plates are to be removed before the book is beaten; this suggests that in everyday work they are not removed, and even in fine work the obvious thing would be to put them back where they came from, though there is more likelihood that they would be moved. This seems to be borne out by the fact that copies of seventeenth century books in contemporary workaday bookseller’s bindings, tend to have the plates in the same places and similarly folded, but they are more variable in bespoke bindings.

Of course plates often have an engraved page number to show where they are to be bound, or they are referred to in the text as ‘see plate facing page such and such’, and there are sometimes printed directions to the binder at the beginning or end of the book. Occasionally, especially in the eighteenth-century, these include detailed folding instructions. But whether there are obvious clues like these or not, I think that how plates are bound into books is evidence of design on the part of the publisher, modified by the bookbinder.

3.4 whether printed several at a time on a single sheet, or printed individually on cut leaves

Having looked at the paper, comparing watermarks and deckle edges in different copies of a number of seventeenth-century English science books with inserted engraved plates, I have come to the tentative conclusion that it was quite common to print plates singly on cut sheets, that is on pieces of paper not much bigger than the eventual size of the plate-leaf in the finished book; but there is also evidence of printing several plates at once on larger sheets which were then cut up by the binder and inserted in the proper places before or as the book was bound. This is of course in contrast with the letterpress parts of a book which are folded and sewn without any cutting, the leaves only becoming separate once the book is trimmed by the binder, or opened by the reader with a paper knife if issued uncut. I don’t know why engravings were sometimes printed singly on cut leaves, and sometimes several at once on a larger sheet, but I wonder if it could have had something to do with the size of the press crew. With three or four men inking and wiping plates, a pressman might be kept busy pulling impressions as the plates were presented to him one at a time. On the other hand one man working alone would perhaps find it more economical to ink and wipe several plates, then lay out and print them on a single sheet of paper all at once with only one pass through the press.

There is something else very intriguing about the three or four man press crew. With 3 men, four plates are being printed concurrently, since one plate is warming on the brazier while the other three are being inked, wiped and printed. With a four man team 5 plates are in concurrent production. So if a book had only one or two plates, could this mean they would have been printed concurrently with the plates for another book? This is where identification of the paper might really come into its own, perhaps making it possible to identify a particular copper-plate printer working on the plates for two or more books, perhaps even doing work for two different publishers at the same time. I don’t know how practical this really is, paper identification is a nightmare, and often a mixture of paper stocks used in one book. This happens in the letterpress parts of books too, where printers are using up remnants of paper stocks left over from other jobs, but the use of mixed paper stocks is more pronounced in the plates.

3.5 how engravings are combined with text on the same sheet.

12.2 12.3 Leonardo Trattato della pittura, Paris 1651

Printing several copperplates on one sheet at one pass through the press is almost certainly the way engravings were printed in letterpress sheets if more than on engraving appears on one side of the sheet, since it seems almost inconceivable that they were printed one by one. The type has to be set leaving blank spaces for the illustrations, the letterpress is run off and the sheets delivered to the copper-plate printer to add the engraved illustrations and decorations. There is documentary evidence that the letterpress was normally printed first, and there are good practical reasons for doing this. The most compelling is that if it were the other way round, it would be very difficult for the printer to tell the copper-plate printer exactly where to place the impressions, unless he had already set the type; and if he has done that he is not going to keep the type standing while the copper-plate printer puts the prints on otherwise blank sheets.

It is commonly observed that copperplates in text are not always very well aligned, even in a prestigious production like this Leonardo. One problem was that the dampened printing paper might expand differentially depending on how wet it was, and the sheets would be further deformed by printing the letterpress, and this might vary from sheet to sheet. There seems to have been no secure way of achieving perfect registration, and Fertel in his printer’s manual of 1723 cautions that enough space must be left to make it easy for the copper-plate printer. My guess is that placing the plates was done by an extension of the method described by Bosse for positioning a single plate. As Bosse describes it, a sheet of the printing paper is laid on the plank of the press, and the plate laid on this in the position that the impression is to go, as Bosse says, ‘to give the correct margins’. The printing sheet is then lined up on this undersheet and the impression will then be in the same place on the printed sheet that the copper physically occupied on the undersheet. Now if the undersheet were a printed sheet, with spaces left for the engravings, the coppers could be put in position on it, probably with the aid of lines drawn on it. Lining up the printing sheet on the undersheet should then get the illustrations in the right places. However lining up deckle edges cant be very precise, and an obvious refinement would be to line up the point holes in the undersheet and the printed sheet. I don’t know if this was done. (I hav’nt explained about press points and point holes, but those who are familiar with letterpress printing will know that their position is constant relative to the type).

4 DESCRIPTIVE BIBLIOGRAPHY

How do illustrations work, and how is this affected by the technology of print? In Edward Hodnet’s analysis illustrations in imaginative literature do three things: they represent a scene in the text; they interpret the text; or they are decorative. Representation; Interpretation; Decoration. Some illustrations do two or more of these things. If an illustration represents something specific in the text, then we prefer to have it placed close to the relevant scene; but interpretation may be more general, like a frontispiece that can be placed outside the narrative, before the book or chapter begins. The illustrations in technical and scientific books have a more concrete connection with their texts, but they too can be seen as fitting into the same scheme. One might describe a plate as representational if it is used where it is hard to convey something in words alone. As Henry Wooton wrote in 1694 ‘Engraving upon Wood, or Copper, is of great Use in all those Parts of Knowledge where the Imagination must be assisted by sensible Images’. More general interpretation is offered by a plate like the engraving of Boyle’s air pump in his New experiments of 1660 which explains the working of the pump, a general understanding of which is necessary to follow the descriptions of experiments that follow. Decorative material is often related to the subject of the book, but is not linked directly to the text.

In other words different kinds of illustration have different relationships with the text. This relationship may be revealed by the internal reference system used, or the lack of one. There are three principal ways in which text can refer to images: first the illustration is placed adjacent to the text which refers to it; second a reference system is used in the text, such as plate and figure number; third an explanatory caption, not part of the linear narrative of the text is attached to the plate. If the image is not keyed to the text, this tells us that it is probably not necessary for a literal reading of the text, but has some other function. Thus portraits and allegorical frontispieces are richly informative, but are perhaps not to be ‘read’ as part of the text. Frontispieces may also function as advertisements, along with pictorial titlepages, giving a quick visual summary of the contents of the book. Other un-keyed illustrations, including headpieces, tailpieces and historiated initials as well as inserted plates, may also function allegorically, or as commentary, or to indicate the content of a chapter visually, as visual place markers, or as embellishments to enhance the sale of the book. Not that we should see decoration as purely a matter of prettyficating: typography and decoration have always served to alert the reader to the kind of content that might be expected, whether academic or popular, serious or lightweight, and therefore mediate their reading of the text.

My point is that the physical relationship of text and images, including position and folding, as well the internal reference system, is crucial for an understanding of how the graphic elements of a book work.

Some of what I have said is speculative, but I hope that at least I have drawn attention to how much needs to be done. The technical history of the printing of engravings in books is waiting to be written. There are no bibliographic wooden rolling presses: no one has yet built a replica wooden rolling press to find out at first hand what the problems of operating it are. The only experience anyone has is of operating nineteenth century iron rolling presses, of the kind in use in many artists studios.

Man Ray photo of Meret Oppenheim, 1933.

The basic principle of the rolling press is so simple, and it remains the same in the iron rolling press, except for the addition of gearing and a flywheel, that it has been assumed that printing on a wooden rolling press would not be any different. I doubt if this is the case.

Side by side with historical research and experiment, bibliographical methods need to be developed for the graphic elements in books for two reasons: first, because books carry their own printing history within them; secondly because describing the way illustrations are physically related to the text will lead to a better understanding of what the visual material in a book is for, and how it works. I would like to propose a new science of non-typographic bibliography. Like traditional bibliography it will work at three levels, as I have tried to show this afternoon: the first level is printing history from external sources; the second is analytical, that is printing history determined from the evidence in the books; and the third level is descriptive, recording the content and function of graphic elements in books.

I have purposely said nothing about what is engraved on the plate, not because this is not a legitimate concern of non-typographic bibliography, far from it, but because what is engraved on the plate, or rather the image of it in the ink transferred to paper has so far engaged all the attention of bibliographers and historians, and I have wanted to suggest ways to redress the balance. To pursue the history of the engraved elements of books, we need to understand the relationships and techniques of the author, publisher, artist, engraver, letterpress printer, copper-plate printer and binder. All I am saying is that if historians are to deal properly with such topics as the ‘Semantics of Graphics’, or ‘Rhetoric and Graphics in Micrographia’ (to use two real examples of the titles of papers on scientific illustration), then they have as much need of a knowledge of printing history as the traditional textual scholar.

But more than this, as Thomas Tanselle has argued, in textual studies the physical form of the text is important irrespective of literary theory. Whether you are interested in the author’s intention, or the publisher’s intention, or the reader’s reaction, the study of texts must be based on the study of books. In this view, since graphics are part of the physical form of the book, they cannot be ignored.

‘History of the Book,’ as a discipline, also demands the study of the whole book, including its reception.

To pursue the history of illustration, the study of texts, and ‘History of the Book’, it is high time that non-typographic bibliography be brought up to the standards of historical accuracy and logical reasoning that letterpress bibliography has enjoyed for decades. When these convergence criteria are met, the two can become a common currency.