University of Virginia Library

Signing by the Page
by
B. J. McMullin

It is easy to assume that in the hand-press period (and beyond) the invariable way of indicating—for members of the printing house as well as for members of the binding shop—the order in which the pages were to fall within a gathering, sheet or volume was to sign a certain number of leaves in each gathering; even when pagination had become all but universal the practice of signing at least the first leaf of the gathering was retained. Confirmation that the trade did use the leaf as the basis for reference lies not merely in the system itself (which may be taken to equate with foliation) but also in the practice of sometimes signing a cancellans according to the same system (i.e. by leaf) when the cancellandum was not signed—for example, actually signing cancellans $7 as '$7' in a volume signed in its original state perhaps only $1-4. Bibliographers have not only regarded leaf signatures as the trade's mechanism for keeping track of type-pages (before and after printing) but also adopted the system as the means of referring to leaves (and pages), since not all publications are foliated, paginated or columnated; that is, reference by leaf, based on the structure of the gathering, has been regarded by modern bibliographers as the only acceptable form of reference. Certainly, some bibliographers of modern books have adopted pagination as the basic system of reference, since in most books of the last two centuries pagination is correct and therefore useable. Nonetheless, most bibliographers still prefer reference by leaf—or they use a dual system—in that, for purposes of bibliographical analysis, reference by leaf relates more clearly to the structure of the volume, sheet or gathering and also copes more readily with sequences of unnumbered leaves or pages.

Concurrently, however, there has also existed within the trade a system of reference using as its base not the leaf but the page, a system which may in

fact be coexistent with printing from moveable type itself. Surviving manuscripts (and printed texts) which have served as printer's copy often bear marginal references inserted in the printing house by the compositor to indicate page breaks in the typesetting.[1] The references usually take the form of a page number and a signature, the latter denoting the gathering plus the page within the gathering—thus '35/C3' indicates that the following page is page number 35 in the continuous numbering and the third page in gathering C (i.e. C2r). Such a system is intelligible and would appear to serve the needs of compositors, although—unlike leaf signatures—having no obvious application in the typesetting itself or in the subsequent printing and binding of the sheets. Nonetheless, in certain publications of the late seventeenth century and the eighteenth signed by the leaf in the normal way, there may be an occasional leaf 'signed' according to the page which its recto occupies within the gathering, in the same manner that the beginnings of pages are marked in printer's copy. This is the practice which I call 'signing by the page', the resultant signature a 'page signature'. Thus instead of being unsigned, page $8r may in certain circumstances be signed '$15', indicating, that is, that the recto of that particular leaf occupies the 15th page within the gathering. (Sometimes, too, 'p.' precedes the number, as '$ p.15'.)

The equation of number and page is obvious enough, but the rationale for the employment of such a system in the printed work is not so obvious; indeed the possible implications for the study of the trade in these two centuries—to be considered below—are quite surprising. Given that a system of reference to the page within the gathering did exist in the printing house, perhaps its employment in printed sheets needs to be explained simply by reference to a supposed change of procedure or to some kind of 'contamination'. On the other hand—given their apparent concentration in particular texts and particular printing houses—it is much more likely that the appearance of page signatures within a printed volume is to be explained by reference to a difference in purpose. If so, the purpose is not obvious to me. The following notes are offered, therefore, not so much as a vehicle for conveying information as a request for assistance in establishing whether the apparent implications are capable of substantiation from other sources. With one exception, my observations are limited to publications of the Oxford and Cambridge University Presses and the King's Printer; that I have seen only one example of signing by the page in the work of other printers should not at this stage be regarded as significant, though these three printing houses were the major producers of volumes which, on present evidence, were the most likely to contain such signatures. In the first place—again with the one exception—all the volumes containing leaves signed by the page are without pagination. Secondly, the vast bulk of the leaves so signed are cancellantia.

Although signing by the page is overwhelmingly to be associated with cancellantia, there are two other usages which I wish to illustrate before proceeding further, particularly as they highlight the major problem associated with the practice—i.e. establishing for whom they were intended, members of the booktrade or members of the public.

The first usage involves not actual signing but the employment of gathering-plus-page number as a means of internal reference, a practice patently directed to readers rather than to members of the trade. In its unambiguousness the usage differs from those which are the main concern of this essay; but that unambiguousness may serve to throw light on those more common instances in suggesting that they too may have been addressed to the reader. In the various volumes of congratulatory verse written by members of Cambridge University in the first half of the eighteenth century—many, if not all, without pagination—there are often lists of errata; thus in Epicedium Cantabrigiense in serenissimum Daniœ principem Georgium (1708), 2°, the 'ERRATA CORRIGENDA' begins: 'IN Scheda [i.e. sheet] notata I. pag.2 [i.e. I1v] lin.33. lege perenni.'; I1 is signed. Elsewhere—as in Academiœ Cantabrigiensis carmina, quibus decedenti augustissimo regi Wilhelmo III. parentat; et succedenti optimis auspiciis serenissimœ reginœ Annœ gratulatur [1702], 2°—'pag' may be omitted, thus: 'L4. lin.2. lege profudit', where 'L4' represents L2v; L2 is not signed.

The various collections of Cambridge congratulatory verse represented by these two examples reflect the general difficulty of referring readers to a particular point in the text when the volume itself lacks foliation, pagination or columnation and has no internal reference system (as in works such as the Bible, dictionaries and most classical texts). The Cambridge volumes are also unusual in employing this particular form of reference: I know of no other instances. Indeed it might be claimed that 'normally' the lack of a numbering system would have made the inclusion of an errata list impractical. For example, in his Lytle treatise . . . against the protestacion of Robert Barnes at the tyme of his death (London: Robert Redman, 1540; STC23209), which is a particularly late publication to be unnumbered, John Standish could only note that the courteous reader would find printer's errors on practically every leaf of his little book (like par for per), which errors should be blamed on the carelessness of the printer, not the laziness of the author—quicquid est vicii non meæ inertiæ, sed impressoris incuriæ imputes oro. Once books were folliated it was possible to be precise in the identification of printer's errors; the standard form of reference in errata lists in foliated volumes was 'Folio 37, page 1' for rectos and 'Folio 37, page 2' for versos—the term 'verso' itself was not available, being not found, according to the OED, until 1839. The two Cambridge volumes instanced here are outside the tradition—presumably dead by the beginning of the eighteenth century—in that they use signatures (not leaf or page numbers) as the basis of their references. The form of reference in the 1708 volume illustrates the general problem of referring to a page within a volume without a numbering system, that in the 1702 the

additional problem of how to infer the signing of an unsigned leaf. That both references were directed to the reader seems clear.

The second usage does involve signing by the page, but, in contrast with the usage of the Cambridge volumes of congratulatory verse, it is patently directed to a member of the trade; on the face of it, however, there is no connexion with the compositorial practice of marking copy. I know of only two instances, both of them bibles, the first a London-printed Welsh bible of 1718/17, the second an Oxford-printed English bible of 1726/25.

Eighteenth-century bibles not only normally lack pagination but can also be found bound up in two (or more) volumes. Sometimes the division into two during binding is quite arbitrary, seemingly being determined by the convenience of the binder; at other times the best possible division is made, given that few books begin at the head of a recto, let alone at the head of $1r. The two instances under consideration, however, are ones in which the point of division was determined in the printing house, not the binding shop.

In the first instance the two volumes were designed to contain (I) the Old Testament, (II) the New and (presumably) a selection from among a variety of other pieces, such as a metrical version of the Psalms and possibly the Book of Common Prayer, the Apocrypha, a concordance, an index, the Companion to the alter and other devotional pieces; the break between Old Testament and New does not coincide with the end of one gathering and the beginning of another. In the Welsh bible, Y Bibl Cyssegr-lan, London, by John Baskett, 1718/17, 8° (collating A-2P8 2Q6 [NT:] x2Q2 2R-3F8; in some exemplars the Testaments are separated by the Apocrypha, collating [a]-[i]8 [k]4), the title page to the New Testament, x2Q2r, is signed "Qq p.13'. The chainlines in the paper suggest that the bifolium x2Q1.2 was imposed as 2Q4.5 in the last gathering of the Old Testament; unlikely though it may seem, the signing appears to have been inserted as a guide to the binder which leaves were to be detached from 2Q or, to put it another way, which were to immediately precede 2R1. The aberrant signing can have been of use only to a member of the trade, but why this particular form of signature? The answer may be connected with the fact that 'normal' practice would have been to impose the first two leaves of the New Testament as 2Q7,8, leaving the binder to cope with the four disjunct leaves which would result from a decision to bind the bible in two volumes. (Later Cambridge practice would have been to impose the two leaves as in the 1718/17 Welsh bible but to sign the first one '**Qq**'.) Even if one discounts the claim that the aberrant signing in the 1718/17 Welsh bible was determined by the possibility that it would be bound in two volumes (all the dozen or so exemplars seen, however, are actually bound in one) the fact remains that the method of signing was dictated by the need to extract from the octavo sheet the bifolium which was to be bound as a separate gathering.

In the second instance the bible was designed to be divided into two volumes of relatively equal bulk, with additions perhaps to be distributed between the two—e.g. with the prayer book preceding the Old Testament and the metrical psalms following the New. The English bible printed in

Oxford, also by John Baskett, 1726/25, 4° in eights, was designed for division after Ezra, the second volume then beginning with Nehemiah. Had the bible not been designed so as to be capable of being bound in two volumes it would have collated A-2T8 2U4, [Apocrypha:] A-K8, [NT:] A-N8 O4; but Nehemiah begins on what would have been 2A7r, and the two sheets constituting 2A were therefore imposed in such a way as to produce two gatherings, 2A6 and x2A2, the first leaf of x2A being signed '[Aa 13]'. Here the equation between signing and division into two volumes is manifest, though the reason for employing that particular form of signing remains unclear. (Again it might be noted that normal eighteenth-century Cambridge practice would have been to sign the first leaf of the bifolium '**Aa**'.)

As already indicated, however, the vast majority of leaves signed by the page are cancellantia. And of those cancellantia the vast majority are to be found in exemplars of the Book of Common Prayer. Cancellation was required to bring the prayer book up to date when certain events took place within the royal family, notably the death of one monarch and the accession of another, since members of the royal family are mentioned in it by name. The earliest instances I have seen are those where prayer books of the reign of Charles II are brought into the reign of James II through the insertion of cancellantia—i.e. they date from 1685+. The case of the Book of Common Prayer I have dealt with elsewhere;[2] suffice it to repeat here that when cancellanda were themselves unsigned the corresponding cancellantia were sometimes—though by no means invariably—signed by the page. Thus in gathering B in the 1727 Oxford 12° (signed $1-6) in the exemplar in the State Library of New South Wales (Richardson 91), leaves 5, 8 and 10 are cancelled, the cancellantia being signed 'B5', 'B15' and 'B19'. (Note that the first cancellans is signed in the same way as its corresponding cancellandum—conventionally, by leaf; in other words, whoever was to insert the cancellantia in this edition had to be able to work within two systems of notation.) The incidence of cancellation did not diminish after 1727, but thereafter the practice of signing cancellantia by the page is uncommon—the last instance that I can report is in the Oxford 8° prayer book of 1762, where in three of the four exemplars seen (Bodleian, C.P. 1762 e.1(1); British Library, C.128.c.8 and 3407.b.17(1)—the fourth exemplar is National Library of Wales, OA 201) there are cancellantia at A8 and B6, signed respectively 'A15' and 'B11'.

The updated prayer books are particularly puzzling. The purpose of signing by the page is clear: it enables a cancellans to be correctly placed when the cancellandum is unsigned and the volume lacks pagination (or foliation or columnation). The first question raised by the practice is why such a system should have been employed when the trade was presumably accustomed to inserting cancellantia in volumes without pagination, as the more usual practice of not distinguishing cancellantia in any way, or of

marking them with nothing more than a typographic symbol, would suggest. The second question relates to the time at which the cancellation was effected, for there is an additional complication with the prayer books: whereas cancellation in general is to be seen as part of the process of creating 'ideal copy' prior to issue, in the case of the prayer books it appears to be part of a process carried out after issue. At first sight it might be assumed that cancellantia datable possibly to several years after the date in the imprint are evidence that stocks of particular editions of the prayer book remained unsold and therefore needed to be brought up to date before they could be sold. But the situation is more complex—at least to the extent that in some instances further editions in the same format had been issued by the same publisher in the intervening years. Perhaps the 'problem' is no more than a reflexion of the state of the publisher's warehouse. On the other hand— since signing cancellantia (by the leaf) was apparently never necessary as a guide to members of the trade and since the observed instances of this unusual form of signing cancellantia cluster in the prayer book, a work with a potentially volatile text—the system of signing by the page may be taken to suggest that it was employed as a guide to people outside the trade. In other words, the system of signing cancellantia by the page may be taken to suggest that it was designed as a guide for owners of prayer books who were faced with inserting cancellantia in exemplars in their possession. (I am attracted to the notion that at least slips were inserted by owners by the fact that they are sometimes pasted over red rules, which, I assume, were drawn after the sheets had been sold, though obviously before they were bound; alternatively one would have to suppose that booksellers or binders were unconcerned to have the ruling re-done at those points.) The question which this suggestion raises is: how then were cancellantia (including slips) made available to owners? Is there any evidence indicating (a) that cancellantia were indeed made available to owners of prayer books (and other publications?), and, if so, (b) what the mechanism for making them available was?

At this point it is proper to concede that there are various counter-arguments to the suggestion that signing cancellantia by the page was a system devised to assist owners of books (specifically prayer books) rather than members of the trade. In the first place cancellantia in prayer books normally appear to have been present when the volume was ploughed, and sometimes there are no obvious stubs to which an owner could have attached them. The case for cancellantia being made available to owners of prayer books is perhaps further weakened by the fact that there was no obligation on owners to bring their exemplars up to date other than by making manuscript emendations. The Act for the Uniformity of Publick Prayers (XIV. Carol. II), prefaced to the Restoration prayer book, provides that 'in all those Prayers, Litanies, and Collects, which do any way relate to the King, Queen, or Royal Progeny, the Names be altered and changed from time to time, and fitted to the present occasion, according to the direction of lawful Authority' (1662 folio, first edition, b3r). Any alterations in wording were in the event authorised by an order in council (or 'warrant'), which customarily begins:

'Whereas by the late Act of Uniformity, which establisheth the Liturgy, and enacts, that no form or order of common prayer be openly used, other than what is prescribed and appointed to be used in and by the said Book; it is notwithstanding provided, that in all those prayers, litanies and collects, which do any ways relate to the King, Queen or royal progeny, the names be altered and changed from time to time, and fitted to the present occasion, according to the direction of lawful authority: it is thereupon this day ordered . . . that the following alterations be made, viz.' After the changes are specified the warrant customarily concludes: 'And it is further ordered, that no edition of the Book of Common Prayer be from henceforth printed, but with the aforesaid amendments; and that in the mean time, till copies of such editions may be had, all parsons, vicars and curates within the realm, do (for the preventing of mistakes) with the pen correct and amend all such prayers in their Church Books, according to the foregoing directions. . . .'[3] The lack of an explicit requirement that cancellantia be secured and inserted does not prove incontrovertibly that no such mechanism existed. On the other hand it can no doubt be assumed that booksellers or printers would regard the injunction that no edition be printed without the specified changes as requiring them to effect those same changes in unsold stocks by a process of cancellation which might include any one or a combination of slips, single-leaf and bifoliar cancellantia, or even the replacement of whole gatherings (particularly those containing the state prayers).

Furthermore, there is an advertisement in the London Gazette no. 2012 (26 February-2 March 1684/5) which bears on the question. It reads:[4] 'This is to give notice, that the Common-Prayer-Book in all volumes is now to be had at the King's Printing-Office, with the alterations in the forms of prayer for the King and the Royal Family, according to His Majesty's direction and command.' Charles II had died 6 February, and the separately published warrant (British Library, C.21.f.2(14)) is dated 16 February, so that the prescribed form of words was known no more than two weeks before the date of the advertisement. The advertisement may be taken to imply that new editions of the prayer book in all formats ('volumes') had been set and run off in the intervening period. Equally, however, 'alterations' may be taken to suggest that the required changes had been made to stocks of bound sheets by means of cancellation of one kind or another and to stocks of unbound sheets by supplying the binder with the necessary cancellantia. The latter interpretation could also be taken to suggest that whatever exemplars were on hand—regardless of date of original publication—were altered, thus accounting for the phenomenon noted earlier: cancellantia appearing in editions of particular formats which were not the current ones. On the other

hand prayer books in a particular format do not always follow one another page for page, so that economies in setting cancellantia would not necessarily be available.

The speed with which the altered versions were made available suggests that they were not newly set editions but exemplars which were on hand and into which cancellantia were inserted prior to sale—i.e. whatever form of signing was employed for the cancellantia was intended as a guide to a member of the trade. But the updated prayer books seem never to be provided with new title leaves confirming their currency, an omission which is difficult to understand if the internal changes were made while the sheets were still in the printing house or the bookseller's shop. The cancellantia, then, may well have been united with the sheets after they had left the bookseller, and the form of signing them have been supplied as a guide to the owner.

There remain two further instances of signing cancellantia by the page which must be noted, since neither is a prayer book and one is apparently the work of a London trade printer, unfortunately anonymous. Though the evidence is not unambiguous, signing by the page in both instances appears on the face of it to be designed for the use of members of the trade (as part of the process of creating ideal copy), not members of the public (as a device to allow them to effect cancellation in exemplars in their possession).

The first of these additional two instances is the 1743 Oxford quarto bible, the first edition of the Bible printed by Mark and Robert Baskett since succeeding their father John as Printers to the University. Cancellation for the moment disregarded, it collates A2 A-3P8 3Q4 3R2, [NT:] x3Q4 x3R8 3S-4H8 4I2 (the first half of each gathering is signed). Of the four exemplars seen, three (British Library, 3050.ee.8(2); National Library of Wales, BS185. d43 4to; Durham University, Bamburgh Castle L.iii.7-10) contain 62 obvious cancellantia, concentrated in the first two alphabets, all signed by the page: A5 (signed 'A9'), A6 ('A11'), A7 ('A13'), B5 ('B9'), B7 ('B13'), B8 ('B15'), D6 ('D11'), D8 ('D15'), etc. In addition, in the BL exemplar I8 ('I15') is a cancel, though the leaf is integral in the other two; the texts are the same,[5] implying that a correction was made at press. The presence of the 63 cancellantia—all replacing unsigned cancellanda—is revealed by their signatures, but there is no reason to suppose that the cancellation in this volume is confined to the second halves of gatherings. Indeed, one would expect that there would be about the same number of cancellantia in the first halves of gatherings, signed in the conventional manner—i.e. as their corresponding cancellanda. Such, in fact, is probably the case: disturbances to the patterns of watermarks, chainlines and tranchefiles serve to identify a further 51 presumed cancellantia in the British Library exemplar. The fourth exemplar of the 1743 Oxford quarto (Bodleian, Bib.Eng. 1743.d.1) is quite unlike the other three, in that it is totally devoid of cancels—at least to the extent that routine

inspection has revealed none of the accepted stigmata of cancellation—so that a textual collation of it and one of the other three would serve to identify cancellantia not otherwise identified. In passing, it might be noted that such wholesale cancellation was effected for what seem—on the evidence of the lamentable textual state of contemporary bibles—quite trivial reasons; for example, the only textual difference in G8 ('G15') is that the cancellans corrects the erroneous 'bread' (vere 'beard') in Leviticus xxi.5, '. . . neither shall they shave off the corner of their bread. . . .' One might readily suppose that the Baskett brothers were anxious to establish for themselves a higher reputation for accuracy than that enjoyed by their father.

Had the Bodleian exemplar not survived, there would have seemed little doubt that signing by the page in the 1743 quarto bible was a device employed for the benefit of members of the trade, that the wholesale cancellation was part of the process of creating ideal copy. The Bodleian exemplar, however, may be taken to suggest that the cancellation in the other three exemplars was not after all occasioned by the process of creating ideal copy prior to issue; indeed the existence of two 'pure' forms (without and with the cancels) could be taken to imply the reverse: that, as in the prayer books, the cancellation took place after the issue of at least one exemplar, that the two varieties are separated temporally.

The sole instance known to me of signing by the page in a volume produced elsewhere than in the printing house of one of the three privileged printers (though the identity of the printer is not known) is Peter Heylyn's A Help to English history, 'continued to the first day of November, 1773' by Paul Wright (London, printed for the Editor: and sold by [nine booksellers in London, one in Cambridge, one in Oxford, two in Chelmsford]; by most of the booksellers in the kingdom; and by the Editor, at Oakley, near Quendon, Essex. 1773.), 8°. Following the seven leaves of preliminaries, A8 (-A8), comes (in all exemplars known to me) what in its original state was a half-sheet, signed *A3.*A4.*A5.[*A6] and consisting of a List of Subscribers, without pagination and with the final page blank. (The signing of the half-sheet suggests that ideally it should have been bound after or before A3, just as the dedicatory leaf '[*D3]' is bound before D3; A3r comprises a further dedication, to George III.)[6] In all but one exemplar[7] * A6 has been replaced by a leaf signed '*a7'; this form of signature can, I think, be explained only in terms of the recto of *A6 being the seventh page in the four-leaf gathering —i.e. I take it that this form of signing serves the same purpose as it does in the other volumes discussed. The purpose of the cancellation was presumably to create ideal copy by bringing the list of subscribers up to date: the effect is to add 28 names, four being incorporated in the reset T and W sequences on the recto, the remaining 24 constituting a separate sequence on the verso.

This last sequence is followed—in at least the Monash University exemplar (*942 H617H.w)—by an Addenda, dated 20 April 1774, which adds the names of two further subscribers and notes that two of the original ones have been succeeded by their sons. The existence of addenda on the cancellans is puzzling and suggests that there may be an intermediate stage, without the addenda, which would have been added by putting the page to press again.[8]

Signing cancellantia by the leaf clearly reflects what is assumed to be the standard form of reference within the trade. Is there, however, a date before which signing cancellantia by the leaf is unknown? And can a reason be advanced why such a form of designating cancellantia should be introduced —admittedly not universally—when the evidence of other volumes containing cancellantia indicates that the practice was not even necessary for keeping track of them? One has to assume, I think, that signing cancellantia was a means of assisting binders in inserting them correctly, but if binders were accustomed to working within a system which pre-supposed the capacity to infer the signing of an unsigned leaf, can binders at large be assumed to have at some stage lost that capacity? Whatever the explanation, it remains that binders would still have had to be able to locate the (unsigned) cancellandum: where cancellans and cancellandum are both found bound up in a volume it is by no means always the case that the latter has been slashed to indicate its status. Incidentally, how widespread was slashing in any case? And how did the trade keep track of an undesignated cancellans when the corresponding cancellandum was not signed and the volume itself lacked pagination?

The 1743 Oxford quarto bible may well be unique: cancellation in eighteenth-century bibles is most unusual, and the system of signing cancellantia in this one may well have been taken over from the prayer books produced in the same printing house. Likewise the edition of Heylyn may be a sport. On the other hand these two instances may suggest that a detailed examination of a wider range of eighteenth-century publications would show the practice of signing by the page to be more widespread than it appears to be from the evidence now available. But whatever the extent, the question remains: why did some printers, working over a period of nearly a century and located in London, Oxford and Cambridge, sign some leaves—specifically cancellantia—by the page? If signing in this way was designed to assist members of the trade, what were the circumstances which dictated its use and why was it not universally employed (even if only in volumes lacking pagination when replacing unsigned cancellanda)? If signing in this way was designed to assist members of the public, by allowing them to correctly place cancellantia in volumes which they already owned, what was the mechanism by which they were enabled to secure such cancellantia?

Notes