In 1534, the marital plans of Henry VIII led him to break with Rome, to deny the authority of the pope who would not grant him a divorce, and to declare himself the Supreme Head of the Church in England. But although he thereby ceased to be a Roman Catholic, the comparatively few reforms he allowed in the conduct of religion were more political than doctrinal. Those such as Thomas Cromwell (Henry’s vicar general and vicegerent in spirituals) and Thomas Cranmer (archbishop of Canterbury) who wanted to steer the church in directions inspired by European reformers had to move slowly and warily, because by no stretch of the imagination could Henry himself be described as a Protestant.On 3 March 1542, the Convocation of Canterbury decreed (probably at the suggestion of the Supreme Head) that from henceforth all church services in England should be ordered according to ‘Sarum use’:Footnote 1 that is, the version of the liturgy and rites associated since Norman times with the diocese of Salisbury, and already far more widespread throughout the kingdom than the uses of York, Hereford, and a few less influential sees. To ensure uniformity (and to hasten the demise of rival uses), the following January the king granted a joint patent to the former partners Richard Grafton and Edward Whitchurch, giving them a lifetime monopoly of printing all Sarum liturgies, namely
The masse booke/ the Graill, the Antyphoner, The Himptnall, The portaus, and the prymer bothe in Latyn and in Englishe of Sarum vse for the province of Canterbury … And … that they and their assignes oonly and none other person nor persons … haue libertie to printe the bookes abouesaid.Footnote 2
Grafton and Whitchurch, who would become the most important printers of Edwardian prayer books and would also contribute to those of 1559, had begun their careers as merchants with no intention of becoming printers.Footnote 3 Grafton served as an apprentice in the Grocers’ Company, and was made a freeman in December 1534 while in his early twenties. Whitchurch, probably a year or two Grafton’s junior, was freed from his apprenticehip in the Haberdashers’ Company in June 1536. After their first mercantile venture with another young Haberdasher ended in litigation, the partners’ shared zeal for religious reform led them to finance the printing in Antwerp of the translation that became known as the ‘Matthew’ Bible of 1537.Footnote 4 They followed that success with an even more ambitious project: a substantial revision by Miles Coverdale of that translation, to be printed in an unusually large format and destined to become known as the ‘Great Bible’. No press of the required size had ever been constructed or used in England, so the job was given to François Regnault in Paris.
About three-fifths of the printed sheets had already been shipped to England, and Regnault had almost finished the remainder, when the Inquisition summoned him and seized all the sheets still in his hands. But although the French authorities never released the confiscated sheets and eventually burned them, they not only allowed the publishers to acquire one of Regnault’s presses and some of his types, ornaments, and even employees, but were apparently the first to suggest that solution. And so Grafton and Whitchurch set up a printing house in the Greyfriars’ former precinct in London, and rapidly learned how to run it. By November 1539 they had replaced all the confiscated sheets and begun selling the Great Bible, and because Henry’s injunctions of 1538 had required every parish church in the land to acquire a copy, they proceeded to reprint the whole book six times with extensive financial assistance from another Haberdasher.Footnote 5
By the time they had finished supplying the nation’s churches with bibles, their printing house had become the largest and most productive yet seen in England. There is no obvious sign of dissension between them, and they continued to collaborate for many years, but sometime around the turn of 1542–43 Whitchurch took a share of the materials and set up a printing house of his own. Whether he did so before or after the two were jointly granted the patent for Sarum liturgies is uncertain. In late 1544 Grafton (alone) was appointed printer to the young Prince Edward, and the following May he and Whitchurch were given another joint patent, this time for a royally approved primer.Footnote 6 Meanwhile, whichever of them had custody of the actual patent for liturgies had apparently lost it, and in January 1546 they paid for an inspeximus exemplification of it.Footnote 7 Had they known that Henry had only a year left to live, and that Catholic service books would not be needed during the next reign, they could have avoided that expense.
The Book of the Common Prayer, 1549The accession of Edward VI in January 1547 brought promotion for Grafton, who replaced Thomas Berthelet as King’s Printer in April.Footnote 8 On the same day he and Whitchurch received a new patent for any and all
bookes concerning dyvyne seruice or conteyning any kinde of sermons or exhortacions that shalbe vsed suffred or Aucthorised in our Churches of Englande and Irelande … being in the Englysshe or Lattyn tongue.Footnote 9
Archbishop Cranmer had been working towards a vernacular form of the liturgy since the 1530s, and as early as 1544 had persuaded Henry to allow the publication of an English litany that could be included in the Sarum service. Under Edward he began anew, and (with the aid of other like-minded divines) by late 1548 he and his collaborators had prepared what would be published in March 1549 as The Book of the Common Prayer.Footnote 10 The Act of Uniformity that both authorized and imposed it (2 & 3 Edw. VI, c. 1) was introduced in the Commons by a bill that was read on 19 December 1548, but was redelivered to Secretary Smith rather than proceeding. It reappeared and was read in the Lords on 7 January, where after two more readings it passed on the 15th – but while both archbishops supported it, eight of the eighteen bishops present dissented. After three readings in the Commons it passed without division on the 21st.Footnote 11
Both Whitchurch and Grafton must have begun printing the book as soon as (or even before) the Lords had voted, because although the Act would not become law until it received the royal assent on 14 March, their colophons are dated the 7th and 8th of that month respectively.Footnote 12 The new form of service was to become compulsory at Pentecost next (9 June), so the printers still had three more months to make progress towards the ideal of providing at least one copy for every parish. How close they came to that goal, however, is unknown. Each produced at least four more editions dated 1549 (a date that probably became a mere formula before they had finished), but the exact order of those reprints is not completely clear. Whitchurch’s STC 16270 and 16270a are both dated May on the title-page and 4 May in the colophon, while his 16272 and 16273 are dated June on the title-page and 16 June in the colophon. In each case one of those colophon dates is merely reprinted from its copy, and it is quite likely that at least one edition was really finished as late as 1550. The precise dates of the Grafton editions are equally uncertain, because all have title-pages that claim the month as March 1549. The two that are probably latest (16274–5) have colophons dated June, but it is unclear which is the earlier. Meanwhile John Oswen of Worcester, who had been granted a patent to print books for church use in Wales and the Marches,Footnote 13 printed the only known edition in quarto (16271, dated 24 May) and a folio dated 30 July (16276). Although Humphrey Powell had no comparable patent, he printed a folio edition in Dublin in 1551 (16277),Footnote 14 apparently unaware that it was shortly to be replaced by a substantially revised version.
In Chapters 4 and 5 I shall discuss the physical structure of the first two folio editions of 1559 in some detail. In each case that structure was inherited from an edition of 1552, and how those earlier editions evolved will be examined in Chapter 2. Those editions in turn were shaped by lessons the printers learned while mass-producing the first Edwardian version in 1549.Footnote 15 But when I looked closely at the 1549 editions I realized that despite the historical importance of their contents they had never been studied really carefully as physical objects. Even the basic question of which edition was the first had never been properly investigated, and will here be answered for the first time.
In order to explain how I reached some of the conclusions here offered – how one can tell which parts of a book were printed first, or who printed them – it will be necessary to spell out in some detail how books were printed and, perhaps even more significantly, how they were not. The next section is therefore of particular importance, for unless it is read attentively some of the deductions offered later may not be properly understood. Sixteenth-century printed books (and this may be the single most noteworthy fact to learn and remember about them) were not manufactured one at a time, and did not emerge from the press one after another as if on a conveyor belt.
How the Books were Printed
Until Chapter 9, almost all the books mentioned in this one will be folio editions. All that means is that each pair of leaves (each bifolium) is a single sheet folded in half, so each leaf is half the size of a sheet of the paper used (which in the case of the prayer books measured approximately 38 × 28 cm, or 15 × 11 inches). A folio book consists of a series of quires (or gatherings), which occasionally consist of only a single folded sheet but are usually made up of multiple sheets (though seldom more than six) folded together, and therefore contain between four and twelve leaves (between eight and twenty-four pages). Each quire is eventually sewn through the fold to a series of cords that lie across the spine and will secure the boards to the finished book.
Each folio sheet has two pages printed on each side, and the pair of pages for one side of each sheet is called a forme.Footnote 16 Because very few printers had really large supplies of type, folios were usually printed by formes. If we use a quire in sixes as an example (three sheets, six leaves, twelve pages), the usual method of printing it was to cast off the text for the first five of those pages: to mark up the copy and indicate where each page should begin, and to begin setting with pages six and seven (the innermost forme of the quire). While that forme was being printed the compositors (type-setters) would set pages five and eight to be printed on the other side of all those sheets (to perfect them) – and after distributing the type from pages six and seven back into the type-cases, would set pages four and nine to print on the first side of the next heap of sheets.
Few printers could afford to keep supplies of type large enough to print more than a few folio pages before an earlier forme of type had to be scrubbed, rinsed, dried, and distributed back into the cases. If a thousand copies of a book were being printed, what existed halfway through the process was not five hundred copies of the book but a thousand copies of half of it, and no copy could be completed and sold until the very last forme was being printed.
To inform the eventual binder of the order in which the quires should be bound, each quire was identified by a signature (a letter or other character) printed below the text on its first page. To explain the order of the other sheets in the quire, below the text on the first page of each (a right-hand page, or recto) the relevant number would follow the signature. In a three-sheet quire designated D, therefore, below the text on the first three rectos would appear D (or D1), D2, and D3 respectively (or D.i., D.ii., and D.iii.).
Because the pages were seldom set or printed in text order, page-numbers were very easy to get wrong and comparatively seldom used. Numbering leaves (foliation) was more common (although also prone to error), but frequently done without. In a bibliographical study such as this one, leaves or pages are usually cited by the more reliable signatures (although they too can be misprinted), and referred to as (for example) leaf A5, page E3r (for recto, or front), page G1v (for verso, or back), and so on.Footnote 17
An unsigned quire in a book’s preliminaries (quires of prefatory material, dedications, contents lists, or anything else preceding the main text) is conventionally identified by bibliographers as π (Greek p for preliminary), and a second such quire would be called ππ or 2π. Sometimes preliminary quires are lettered, and if the letter is one also used in the signatures of the main text it is cited with a superior π prefixed (πA). But a letter not used to sign the main text (as lower-case a in the preliminaries of a book otherwise signed only in capitals) does not need a prefix.Footnote 18
A book’s signatures can also be used to describe its overall structure (or collation) by means of a formula. At this point I should perhaps reassure any reader whose reactions to the word formula include an aversion to anything suggesting either calculation or any branch of mathematics or science. A collational formula is no more than a compact, step-by-step description of how the book is made up: how many quires, how many leaves in each, and how they are signed. And while the complete formula for any of the earliest editions may look a little intimidating at first sight, to begin what I think of as the ‘archaeological’ approach, I intend to divide each formula up into sections: the preliminaries; what I shall call parts 1 and 2 of the main text (each part subdivided); and two belated afterthoughts.
The PreliminariesEach of the earliest editions of 1549 has a colophon (a statement of who printed it, where, and when, but not on the title-page where that would be called an imprint) dated in early March. Edward Whitchurch’s colophon (at the end of the book) is dated 7 March and Richard Grafton’s (at the end of the Communion service) a day later, so the common (and careless) assumption has long been that Whitchurch’s edition beat Grafton’s into the shops by a day and is therefore the editio princeps. I shall reexamine that conclusion later in the chapter, but since each title-page is dated simply ‘Mense Martij’ with no day specified (Figure 2), and since their preliminaries are identical in structure, I need record the formula for them only once:
The ‘2o’ (which I shall not repeat for the other sections) simply indicates that the book is a folio, in which each sheet contains two leaves (a quarto would be ‘4o’); the comma after πA8 merely separates the preliminaries from the main text. The first ‘quire’ is a single sheet: a bifolium whose first leaf has the title-page on the recto and a list of contents on the verso. Title-pages are almost never signed, but leaf ❧2 is signed with an ‘Aldine leaf’ rather than a letter. At its right extremity the tip of Whitchurch’s leaf bends downwards and Grafton’s upwards, but that is not important. What is significant is that both printers misprint the leaf number as ‘i.’ instead of ‘ii.’.
It could hardly be clearer that one of these sheets was printed from the other, rather than each independently from manuscript copy. It is unsurprising that the wording of the title is identical, and the minor differences of line-division are easily accounted for by the differences in size and proportion of the central spaces in the two woodcut compartments. More dramatic is the resemblance between the two lists of contents, in which although the spelling of individual words differs quite freely, each line-division in the entries that exceed one line is in exactly the same place (including the redundant double hyphen in ‘Communion of the=|same’ in item ix). The first paragraph of the Preface on ❧2r is necessarily divided differently because Whitchurch had to fit the text around a larger ornamental initial, but in the second paragraph and the whole of the second page all lines divide at the same point in both. It is also reasonably clear that whichever printer was the first to set and print this, it was the last sheet of his edition to be printed, as is often true of title-sheets.Footnote 19
The second preliminary quire (πA8, prefixed by π to distinguish it from the first quire of the main text) was also used as copy by whichever printer was the second to print it. But whoever printed it first may have done so at almost any time during the proceedings. The first page (πA1r) has only a section title that introduces the next three pages; they contain an explanation of the order in which the Psalms are to be read throughout each month (πA1v), a table illustrating that order (πA2r), and an explanation of the order in which the rest of the Bible is to be read (πA2v). The remaining twelve pages of the quire contain a liturgical calendar, with each month filling a page.
The calendar quire presented special challenges, and in each printing house would have been assigned to experienced workmen with specific skills. Thirteen of the sixteen pages needed to be set by compositors capable of handling tabular material: the table of Psalms (πA2r) and the more difficult nine-column calendar pages (similar to those of 1561 reproduced in Plates 6 and 7). One of the difficulties is that in such tables the vertical and horizontal rules, which are printed from thin strips of brass, cannot cross each other. In those tables, therefore, most of the vertical lines are really made up of short, line-high rules, each set in approximately the right place according to a mark scratched on the setting rule on which the compositor assembled the type.Footnote 20
Moreover, each forme of the calendar quire is printed in two colours, and not all pressmen had the necessary skills or experience for that. When set, the whole forme was first printed in red on a sheet of parchment. The words to be printed in red were then carefully cut out so that when the cut parchment was placed (as a frisket) between the type and a clean sheet of paper, only the selected words actually touched the paper. When all the sheets had been printed with those words in red the frisket was removed, and the forme was cleaned. The red words were taken out and replaced by spaces and quads so that the rest of the text and the rules could be printed in black.Footnote 21 Because of this extra difficulty, the Grafton calendar quire was printed in sufficiently large numbers to supply at least two editions (perhaps as many as 3,000 copies; perhaps more), and the preliminaries in at least his first two editions are essentially identical. Whitchurch, however, apparently printed only enough for a single edition, and arranged the preliminaries slightly differently for his next edition.
The Main Text, Part 1
At this point we meet parts of two collational formulae that are much simpler than they may seem at first sight. Here I have followed tradition by listing Whitchurch first, so let us begin by walking through his formula. His main text begins with an eight-leaf quire A, but continues with a series of eight six-leaf quires signed B–I. Another eight-leaf quire signed K is followed by eleven more six-leaf quires (L–X), and the section finishes with a third eight-leaf quire (Y). At first sight Grafton’s formula may look very different, but in fact it contains exactly the same number of leaves (138). The ‘major’ differences are that while both printers have eight leaves in quires K and Y, Whitchurch also does so in quire A but Grafton in quire V.
What I have here called part 1 of the prayer book deals with the ‘usual’ services for the whole year. Section 1a presents the order for Matins and Evensong, which are essentially the same for every day of the year (although a few special variants are indicated in those sections). Section 1b then prints all the special Introits, Collects, Epistles, and Gospels prescribed for use during the Communion services held on ninety special days thoughout the year, while Section 1c presents the ‘basic’ Communion service itself. If we divide the collational formulae by those sections the two editions appear even less different. In this case I have placed Grafton first, because there is clear evidence that he was the first to print part 1.
Grafton began, in fact, with section 1b (the Introits etc.), whose running titles throughout are ‘At the Communion’. He correctly predicted that section 1a would need no more than a single quire, and so began setting the texts for the first Sunday in Advent on the first page of a quire he signed B (numbering the recto page ‘Fol. vii.’ on the correct assumption that quire A would probably contain only six leaves). The work apparently proceeded regularly as a series of six-leaf quires until at least part of the way through quire L (whose leaves are numbered in roman numerals Lxi–Lxvi) and perhaps beyond, until a problem arose. Either part of the manuscript copy had been misplaced or the authorities decided only belatedly to add special texts for Easter Monday, but one way or another it became necessary to scrap K6 and to replace it with a quire with two more leaves (K8, each of whose last three leaves is therefore numbered ‘Fol.Lx.’). The work then proceeded without visible problems until quire T, whose completion left too much text remaining to fit into twelve pages. The copy for section 1c may not yet have been available, so rather than leave section 1b incomplete Grafton chose to use another eight-leaf quire (V8) whose last leaf was left blank.
It was probably at this point that he went back to the beginning of part 1 and began to work on quire A, containing the orders for Matins and Evensong. The compositor cast off the first five pages so he knew where to start setting the sixth page (A3v), readied his galley by heading it with a running title from a recently distributed verso page, and began to set. When he finished that page he began the next (A4r), using a running title from a recent recto but remembering to change the folio number in the top right corner to ‘iiij’. Once that forme was imposed and ready for the press he moved on to the next (A3r:4v), and then another. Since he had started from the middle of the quire one might expect inner forme A2v:5r to be next – and it may indeed have been the next to be set, although outer forme A2r:5v was the first to be printed. But it was not until the press had started printing inner forme A2v:5r that someone realized that although the folio numbers had been corrected, the actual running titles still read ‘At the Communion’ as throughout section 1b.Footnote 22 The press was stopped and the headlines were belatedly corrected to ‘Matins’ and ‘Euensong’ before the remaining copies were perfected, although sheets A3:4 and A2:5 were then apparently set aside, reprinted with the correct running titles, and not included in the first edition. But someone subsequently printed the words ‘Mattins’ and ‘Euensong’ many times on one or more sheets of paper, covered up the erroneous headlines with paste-on cancels (blank, in the case of A4r), and added the modified sheets to the heaps from which one or more later editions were gathered.Footnote 23 Finally, after finishing section 1a, Grafton apparently returned to the other end of part 1 and printed the Communion service on quires X6 and Y8, leaving the final page Y8v blank.
I have as yet offered no evidence that Grafton printed part 1 before Whitchurch did – but it is plentiful. Section 1a differs quite substantially between the two versions, but in quires B–M of section 1b the two settings agree page for page and indeed usually line for line, and any occasional departure is brought back into agreement within a very few lines. While that close agreement (especially as regards the anomalous quire K8) makes it clear that one printer was using the other’s sheets as copy, it does not show which was which, but after quire M that becomes obvious. In quire N the two versions begin to diverge, with each of Whitchurch’s quires containing a little more text than Grafton’s. When Grafton finished the last page of quire T he still had enough text left to fill fourteen pages: too much for a six-leaf quire but not quite enough to fill the sixteen pages of V8. Whitchurch’s carefully calculated casting off left his section 1b finishing neatly on V6r, leaving only a single final blank to focus attention on X1r and the beginning of the Communion service. So not only can we deduce that Whitchurch was using Grafton’s sheets as copy: we can also be reasonably certain that Grafton had finished quire V some time before Whitchurch reached quire N. The casting off required the whole contents of Grafton’s N–V (twenty-five printed sheets) to be available for marking up. There is no sign here of two printers racing to finish only a day apart: the distance between them would take a single press more than three weeks to close.
There is, moreover, one important difference between the two printers’ versions of section 1a (Matins and Evensong). Grafton was quite right when he assumed that a single six-leaf quire A6 would be all he needed to allow before beginning section 1b on B1r. Indeed, when he finally printed it the text filled only ten of the twelve available pages, so A6r was used for a completely unnecessary explicit (‘THVS EN=| deth the order of Matyns and | Euensong, through | the hole yere.’) and A6v remained blank. Whitchurch, however, used an eight-leaf quire A8, filled thirteen of its pages with text, reprinted the explicit on A7v, and left the whole leaf A8 blank.Footnote 24
In Matins on A2r, both printers begin the page with the same three rubrics, set as ten lines by Grafton but as twelve by Whitchurch. The third rubric calls for the recitation of the canticle Te deum laudamus in English, except in Lent when it is to be replaced by Benedicite omnia opera, also in English. Either Grafton or whoever wrote out his manuscript copy apparently assumed that Te deum in English was too familiar to need repeating, and so supplied only Benedicite in full,Footnote 25 but Whitchurch printed the whole text of both canticles. On A3r (Grafton) or A3v (Whitchurch) another canticle is called for, namely Benedictus – and once again Grafton omits it but Whitchurch includes it. Likewise in Evensong, rubrics call for the English recitation of both Magnificat and Nunc dimittis, both of which Grafton omits but Whitchurch includes. And that is why Whitchurch’s section 1a fills five more pages than Grafton’s.
Before I turn to the shorter part 2 it is necessary to add a rider to my statements about Whitchurch and his printing of section 1b, because as the section stands in all the extant copies I have seen he did not print quite all of it. Paging through quire X one might notice that while four of the six folio numbers are in lower case with no internal punctuation (Fol.cxxj.), leaves X3 and X4 both use a capital C followed by a period (Fol.C.xxiii.). And although few readers would be in a position to recognize it as not belonging to Whitchurch, the 28-mm capital D on page X3r is part of an alphabet of brass initials I have written about elsewhere, and which in 1549 belonged to the printer Nicholas Hill (originally vanden Berghe).Footnote 26 Hill had come to England from the Low Countries in about 1519, had taken out letters of denization in July 1544, and in 1548–49 he collaborated with Whitchurch on several books, including one reprint of the Great Bible and five editions of Erasmus’s New Testament Paraphrases.Footnote 27 In this case his printing of a single sheet hardly rises to the level of ‘sharing’, and may indicate that Whitchurch discovered something seriously wrong with his own sheet X3:4 but was too busy to print the necessary cancel (replacement) himself. If there was indeed a problem, its nature is beyond discovery, and Hill’s replacement for the most part follows Grafton line for line.Footnote 28 But because shared printing would play an increasingly important role in the Tudor prayer books, now is a good time to outline its history.
What could perhaps be considered the first English example was the three-volume Graunde Abbregement de le Ley of 1516 (STC 10954), of which John Rastell printed the first volume and Wynkyn de Worde the second and third. But while that project was shared, each volume is the work of one printer throughout. The beginnings of ‘real’ shared printing in England appear in the 1520s, when de Worde occasionally shared a book with one of his ex-apprentices, either to find him a small task to do (in 1521 John Skot printed the first quire of 10631.5 for him, but it is de Worde whose name appears in the colophon) or perhaps to help him meet a deadline (in 1524 de Worde contributed sheets D3:4 and E2:3 to 15050, otherwise printed by Robert Copland). Later in the 1520s John Rastell in London rather surprisingly shared at least four books with Peter Treveris across the river in Southwark, and in 1532 Treveris shared another with Robert Wyer in London.Footnote 29 From the mid-’30s to the mid-’40s about a dozen examples are known (not counting the Great Bible, mostly printed by François Regnault in Paris but completed by Grafton and Whitchurch in London after Regnault’s arrest). But during the Edwardian Reformation the book trade went into overdrive, and I know of more than thirty examples from those years. (Mary’s less bookish reign would apparently produce only half a dozen.) Moreover, while each of the earlier examples was the work of two printers only, three of the Edwardian examples each involved three printers. Motives for the practice could vary, but in one way or another the question of speed usually came into it, because two printers could always produce completed sheets more quickly than one. So the printers who were tasked with producing the books on which the Edwardian Reformation relied – bibles, the Paraphrases of Erasmus, and the successive English prayer books – often needed to enlist others to assist them.
Whitchurch (STC 16267) a–e6 f8 Grafton (STC 16268)Footnote 30 Aa–Ee6 f8
Here it seems unnecessary to add a separate presentation formally divided into the two sections of the text, because section 2b occupies only the last eight-leaf quire in each edition. Section 2a contains what one could describe as the ‘occasional’ services whose celebration is completely independent of the calendar: Baptism, Confirmation, Marriage, Visitation of the Sick, Burial, and the Purification (‘Churching’) of Women after childbirth.
Section 2b, however, begins with a ceremony that does depend on the Easter calendar, and is headed on its first page as ‘The firste daie of lente commonly called Ashe-wednesday’ (Whitchurch, f1r). In the contents list that is elaborated to ‘A declaration of scripture, with certein prayers to bee vsed the firste daye of Lent, commonlye called Ashwednesdaie’ (❧1v), and in 1552 and all subsequent versions of the prayer book that ceremony is known as the Commination. It finishes on f4v, and is followed on f5r by a four-page essay ‘Of Ceremonies: why some be abolished and some retayned’. That in turn is followed on f7r by ‘Certayne notes’ (five in number), ‘Finis.’, and Whitchurch’s colophon of 7 March (Figure 1a).
In the formula above and the foregoing paragraph it will be noticed that this time I have given Whitchurch’s edition priority – because here the evidence points in that direction almost as clearly as in part 1 it pointed towards Grafton. Whitchurch’s use of the capitulum (¶) in his signatures suggests that when he began to print he was unsure exactly what sequence of signatures would precede this section,Footnote 31 whereas Grafton’s two-letter signatures clearly follow his almost-completed single-letter alphabet.Footnote 32 And while Whitchurch foliated part 2 separately from i to xxxvii, when Grafton reprinted it he was able to continue the foliation from the end of his completed part 1 (C.xxxv–C.lxx, leaving the page of ‘Certain notes’ unnumbered).
At first sight there seem to be more differences than similarities between the two versions, although in most of the services Grafton’s first page ends with the same catchword as Whitchurch’s. But the drop capitals with which Grafton begins most prose passages of any length are seldom the same size as Whitchurch’s, so his line-breaks do not commonly match Whitchurch’s. Nor, indeed, do the page divisions, although each service occupies exactly the same number of pages in each edition. But Grafton was able to spread the text more evenly through the pages than had Whitchurch, whose layouts sometimes seem quite eccentric.Footnote 33 Before his compositors started work Grafton must have done some very careful casting off, but his dependence on the Whitchurch setting is perhaps best indicated by the folio number on Aa6r. That page’s ‘Fol. v.’ in fact misprints Whitchurch’s ‘Fol. vi.’ – but in Grafton’s numbering it should have read ‘Fol. C.xl.’
Three Remarkable Volumes
It is now useful to examine the seemingly unrelated question of how many copies of the earliest editions of 1549 survive. When searched in March 2020 the relevant ESTC entries listed seventeen copies of Whitchurch’s STC 16267 (ESTC S109513), but only seven copies of Grafton’s 16268 (ESTC S93744). Given that Whitchurch’s colophon is dated a day before Grafton’s, its status as the apparent first edition might perhaps be enough to have skewed the figures – but those numbers are not the end of the story. Of the seven Grafton ‘copies’, one is an isolated specimen of his colophon leaf, bound into a privately owned copy of a later edition. And the one listed as at Keble College, Oxford, is really a copy of 16269 wanting everything before B1, misreported some years ago to SOLO (Search Oxford Libraries Online) as 16268. Which apparently leaves the score at Whitchurch 17, Grafton 5.
But those numbers are not final either, because three extant volumes appear in both lists.Footnote 34 Each of those three has part 1 of Grafton’s edition (❧2 πA8, A–I6 K8 L–T6 V8 X6 Y8) followed by part 2 of Whitchurch’s (a–e6 f8). Each library has evidently treated its volume as two distinct items bound together, and in two of them the Grafton part has ‘(1)’ after the call number while the Whitchurch part has ‘(2)’. Each has clearly been believed to be a ‘made-up’ copy, cobbled together by a comparatively modern bookseller or collector from parts of two defective books. Moreover, because none of the three includes either Grafton’s Litany (in the final quire of his STC 16268, presumed lost with his part 2) or Whitchurch’s (the quire immediately preceding part 2 in his STC 16267, presumed lost with his part 1), each has been considered not only ‘made-up’ but also textually incomplete.
To accept that, however, we would have to imagine three copies of Grafton’s edition, each perfect at the beginning but seriously defective after part 1, each coming into the hands of a bookseller or collector who owned a more seriously defective (though more desirable) Whitchurch first edition whose part 2 was the only salvageable fragment. Some such explanation could have seemed plausible if the British Library copy (acquired in 1859) had been the only exemplar.Footnote 35 But the volume containing the Brasenose copy was catalogued there in the 1660s, and at least two of the early annotators of the New College copy practised their handwriting in both printers’ shares. All three are in good condition, and they outnumber the only two recorded copies of Grafton’s 16268, neither of which is quite perfect.Footnote 36
When examining part 1 I pointed out that for Whitchurch to have been able to cast off and re-divide the text of Grafton’s printed sheets N–T6 and V8 he would have needed access to at least twenty-five of the sheets that Grafton had already printed (and if Grafton had already finished section 1c, at least thirty-two sheets). But Grafton’s colophon dated 8 March is on Y8r, so unless that date is a fiction Whitchurch’s part 1 could not have been finished much before the end of the month (if indeed in March at all). And since his reprint of the Grafton preliminaries is slavishly precise even down to the misprinted signature ‘❧.i.’, his repetition of Grafton’s ‘Mense Martij’ on the title-page is hardly trustworthy. But in much the same way, Grafton’s edition of part 2 is manifestly later than his own part 1 (whose foliation it continues), and cannot have been finished before the date of either colophon.Rather than made-up copies constructed decades or centuries later, therefore, the three hybrid volumes are evidently copies of the prayer book as it was originally issued in early March 1549. They are complete copies of a book that was shared by the two printers: part 1 by Grafton with his dated colophon on his final page (Y8r) and part 2 by Whitchurch with his dated colophon on his final page (f7r). Neither of the pioneers of the formulary of collation had any idea how commonplace a practice shared printing was in the period on which their attention was focused, so neither Greg nor Bowers imagined that a convention was needed to indicate its presence. So let me propose that in comparatively simple cases such as this, a semicolon might usefully mark a division between printing houses.Footnote 37 This, then, is the collation of the true first edition of the 1549 prayer book:
2o: ❧2 πA8, A–I6 K8 L–T6 V8 X6 Y8; a–e6 f8.
But the Litany is not the only thing ‘missing’ from the three supposedly made-up copies, none of which has the note on retail prices found on Whitchurch’s f7v in all copies of STC 16267 that I have seen. This is an official order from King Edward (ostensibly ‘by the aduyse of’, but really by, Protector Somerset and the Privy Council), limiting the maximum retail price for an unbound copy to 2s. (24 pence), or for a copy ‘bounde in paste or in boordes’, 3s. 4d. (40 pence). Exactly when this was calculated is unrecorded (the extant registers of the Edwardian Privy Council seldom mention matters of this kind), but it may have been based on the printers’ original estimate, which would not have included the unexpected extra sheet in quire K8 or any other belated additions.Footnote 38 The British Library and New College copies each contain 186 leaves, or ninety-three printed sheets, so for 2s. unbound a retail purchaser would obtain 3.875 sheets for each penny.
There is, however, yet another layer to the story, revealed by a unique bifolium now bound between the two parts in the Brasenose copy. Signed with a Maltese cross (✠2), it contains the four canticles that Grafton had not included in section 1a but which Whitchurch inserted when he printed his enlarged version of quire A. As presently bound it follows immediately after the leaf with Grafton’s original colophon on its recto, although that is not where it was intended to be placed. It ends on its own final verso with a second Grafton colophon, printed from exactly the same lines of italic type used for his first, but with the date altered from ‘the.viij daye of Marche’ to ‘the.xvi.daye of Marche’ (Figures 1b and 1c). And between the end of the canticles and that colophon appears Grafton’s setting of what I shall now call the original note of those same prices: 2s. unbound or 3s. 4d. in paste or boards.
Whitchurch was evidently planning to reissue his part 2 with a reprint of Grafton’s part 1, and I suspect that he had already reprinted a substantial part of it, including the six-leaf quire A that lacked the canticles. His extant leaf A1 reprints Grafton’s more or less line for line and with the same catchwords on each page. In Grafton’s quire the conjugate leaf A6 contains nothing of significance: only the unnecessary explicit on A6r that Whitchurch might well have decided to do without. I therefore believe that at some date after 16 March Whitchurch cancelled his own A2–5, reprinted those eight pages as eleven new pages that included the canticles, and reinstated the explicit on his A7v only to avoid a completely blank opening. What had been blank A6 thus became the present blank A8.Footnote 39
The Grafton sheet ✠2 that Whitchurch probably used as copy for the added canticles also included the note about retail prices that clearly belongs at the end of the book. So Whitchurch evidently put his remaining copies of sheet f2:7 through the press again to add the note to page f7v. But he should have waited a little longer before doing so, because the four canticles were not the last of the belated additions.
At this point I need to confess that what STC classifies as each printer’s first edition of the book contains a two-sheet quire that I have not included in the partial collations I have presented above. In Whitchurch’s edition it is signed with a capitulum (4) and placed between parts 1 and 2; in Grafton’s it is signed with a Maltese cross and placed at the very end (✠4). Each was added sometime after 16 March, when the compilers realized that although the first of the rubrics following the Communion service (on Y7r in each printer’s edition) prescribes when the Litany should be said or sung, the Litany itself had been omitted from the book. In Whitchurch’s quire the Litany fills leaves 1–3 (the blank fourth leaf is usually missing). But Grafton reprints the canticles first, ignoring the verse structure that had filled three and a quarter pages in bifolium ✠2 and cramming them as prose into two pages and twelve lines. The Litany, too cramped to be allowed a heading but implicitly identified by the running titles, follows immediately after Nunc dimittis, and ends on ✠4v leaving just enough space for the note about retail prices and yet another Grafton colophon (this time undated).
On at least three detectable occasions the printers had either to add new sheets or replace ones already printed: Grafton had to replace his original quire K6 with K8, to add ✠2, and later to replace it with ✠4; Whitchurch had to replace A2–5 with A2–7 and insert quire 4. I suspect that Grafton may have reminded either Cranmer or some other Privy Councillor of these incidents – because on ✠4v the prices had been revised upwards: from 2s. to 2s. 2d. unbound, and from 3s. 4d. to 3s. 8d. when bound in paste or boards. The retail purchaser of an unbound copy would now get only 3.655 sheets for each penny (although Whitchurch had to wait until his next edition before he too could print the revised prices).Footnote 40To summarize the early history in print of the 1549 Booke of the Common Praier, then, let me resort to collational formulae, with the quires of the first edition underlined:
2o: ❧2 πA8, A–I6 K8 L–T6 V8 X6 Y8; a–e6 f8.
2o: ❧2 πA8, A–I6 K8 L–T6 V8 X6 Y8; a–e6 f8; ✠2.
For the third issue of the first edition (STC 16268), now known only from slightly defective copies at Durham Cathedral (F.IV.56) and Christ Church, Oxford (Gibbs.1), Grafton reprinted Whitchurch’s part 2 and for the first time added the Litany (following the canticles in a new quire that replaced ✠2):Footnote 41
Meanwhile Whitchurch had been busy reprinting Grafton’s preliminaries and part 1 (with a little assistance from Nicholas Hill), inserting the omitted canticles where they belonged in part 1a and adding the Litany in a new quire 4 at the end of his reprinted section:
Because 80 per cent of this fourth version (STC 16267, the editio formerly known as princeps) is newly printed, it is most realistically defined as the second edition, with part 2 reissued from the first.
But although 16267 can no longer claim priority, it is arguably the best of the early editions. It is the first to incorporate each of the four omitted canticles where it belongs in the daily services, and while the Litany more obviously ‘ought’ to be between sections 1a and 1b (where it would appear in and after 1552), Whitchurch’s position for it between parts 1 and 2 (not followed by Grafton until his fourth edition) is certainly better than at the very end as the second of two afterthoughts.
Corroboration from the Past
I had already worked out most of the above after first seeing the contents of bifolium ✠2, when cataloguer Sophie Floate sent me a photograph of the Brasenose title-page. And once I read the anonymous annotations below the woodcut border I realized that the essential facts had already been documented some four centuries ago. Sometime in or soon after the late sixteenth century, the book’s then owner reported as follows (as shown in Figure 2).
This former Service book of ^K Edwd, was printed 4 times in one yeare 1549: the first inpression was this by Grafton & Whitchurch; in wch, the Litanie wth the folowing praiers, was omitted, & Te Deũ, Benedictus, Magnificat, and nũc Dimittis were added to the later end.
The two most important details here are the writer’s unconcerned acceptance of the fact that the volume was the work of both printers whose colophons appear in it, and the observation that the canticles bifolium was at that time the last quire in the book, where Grafton would later place its successor, quire ✠4. The writer then tells us about Grafton’s next editions:
The 2 [impression] was in the same moneth of March, and had ye Litanie, Te Deũ &c/ in end of all.
The 3d & 4th [impressions] wasere in Iune folowing, and had Te Deũ &c/ in their proper places, after the 1 & 2 Lessons [in Matins and Evensong], and the Litanie at the end of the Comunion.
I suspect that those two ‘impressions’ were one copy of either 16272 or 16273 (Whitchurch) and a second of either 16274 or 16275 (Grafton). All four have June colophons, the canticles embedded where they belong, and the Litany between parts 1 and 2, and the writer would have had to examine them very closely to distinguish the members of either pair from each other.It was probably a subsequent owner of the Brasenose copy who moved bifolium ✠2 to its present position between Grafton’s Y8 and Whit-church’s a1: probably the same owner who appended the volume to the 1637 first edition of the prayer book ‘for the use of the Church of Scotland’ (STC 16606) and had them bound together. The writer of the title-page notes would have known better than to assume that the canticles without the Litany belonged after the Communion, despite the similarity of the two Grafton colophons. The only clue to the date of the notes is the final sentence, in which it is observed that the 1549 book
was translated into Latin, & sent to Martin Bucer, and on ^itwch hee wrote his Censure, wch is extant in his Opera Anglicana.
Exactly how the events of 1549 played out is beyond recovery, but what I believe to be a fairly plausible reconstruction can be built around some completely imaginary numbers. Let us therefore hypothesize that soon after the passage of the Act on 21 January, Grafton (the King’s Printer) and Whitchurch (his partner in the patent for ‘bookes concerning dyvyne seruice’) are told that the immediate need is for a few hundred advance copies so that the clergy in the parishes of the capital (and a few other important cities and dioceses) can all prepare to introduce the new services simultaneously (while copies will necessarily reach some parts of the country rather later). Grafton decides on a speedy initial edition of 600 (except of the preliminaries, of which he will print many hundreds more); Whitchurch calculates that he can print 1,200 copies of the substantially shorter part 2 (which contains only nineteen sheets beside Grafton’s seventy-four) by the same deadline. When they finish their shares in early March, the authorities decide to begin by distributing 300 of those copies, which are duly assembled and bound.Footnote 42
With 300 copies of part 1 still in his warehouse, Grafton is advised that the canticles should have been included in part 1. He therefore prints enough copies of bifolium ✠2 to update both those copies and any of the first 300 that have not yet been bound. Whitchurch withholds his 900 copies of part 2 so he can add the note of prices to f7v; Grafton begins to reprint 300 copies of part 2, and when the omitted Litany is supplied he reprints it with the canticles as quire ✠4 of the first all-Grafton edition (with a revised list of prices). Whitchurch meanwhile prints 900 copies of part 1 (with the canticles included where they belong) and appends the Litany to complete the first all-Whitchurch edition. If documentary evidence for the numbers and dates were suddenly to be discovered, nobody would be more surprised than I if they were really close to my guesses. But the observed facts do appear to indicate something on those lines.
The Reprints of 1549
We have already seen Whitchurch taking advantage of setting from printed copy, and rearranging part of the text for no apparent reason beyond saving a sheet by reducing Grafton’s quire V8 to his own V6. The capped retail price gave him an even better motive. He probably sold comparatively few copies to retail customers: a printing house was a factory rather than a bookshop, and while his premises probably included a shop, most of what he printed was sold at wholesale prices to retail booksellers. But whatever he demanded as a fair price from a shopkeeper who could not resell above the official price limit, reducing the number of sheets in the book would increase his profit margin. So it need come as no surprise that his next edition and its successors packed a little more text into every page, reduced the number of blank pages from eight to three, and (with a collation of πA10, A–P8 Q4 R–T8 V10) reduced the number of sheets in each copy from ninety-five to eighty-four. While the price for an unbound copy thereafter remained constant, the maximum price of a bound copy was raised again after two more editions, from ‘bounde in past or in bordes’ not above 3s. 8d. to ‘bounde in paste or in boordes couered with calues leather’ not above 4s.Footnote 43
Grafton’s reprinting strategy was rather different, and his next two editions were essentially page-for-page reprints, with the same collational formulae and the same number of leaves as his first. But while that may seem to waste an obvious opportunity, there are some potential advantages in reprinting page for page, and Grafton was evidently aware of them. For example, however many copies of a given sheet are wanted, in practice some sheets might be damaged or spoiled. The number of usable sheets therefore sometimes falls short of the number actually printed, and pressmen were anyway quite capable of miscounting when assembling the heap of clean sheets to be dampened and readied for the press run. The maximum number of complete copies that could be gathered from the ninety-five heaps of printed sheets was the number present in the smallest heap, after which all the sheets left in all ninety-four of the other heaps were unusable. But if the book were to be reprinted page for page, those leftover sheets could be used in the next edition, and commensurately fewer new copies of each relevant sheet printed.
It is, moreover, easier (and consequently more efficient) for compositors to reprint text line for line and page for page than to rearrange it. So while Whitchurch decided to incorporate the ‘missing’ canticles in his quire A8, Grafton twice chose to reprint his original quire A6 without them, leaving them cramped in quire ✠4.
Grafton, indeed, sometimes seems content to mingle the sheets from different print-runs without much regard for the order in which they were printed. No matter how small his original edition of part 1, he seems to have printed far more copies of his first setting of the preliminaries, which reappear in most copies of STC 16269 and some copies of 16269.5.Footnote 44 Not until his fourth edition did he follow Whitchurch’s lead and print them as a single five-sheet quire, thus allowing himself to use some red ink on the title-page. Most copies of what STC distinguishes as 16269 and 16269.5 appear to include at least a few sheets that may really ‘belong’ to the other, and as I have already noted, his original sheets A3:4 and A2:5 (with the incorrect running titles corrected by slip cancels) are found only in copies of one or other of those later editions.
Quire ✠4 is another puzzle. The only extant copy of STC 16268 with what I presume to be the original setting is the copy at Durham Cathedral, with ✠1r signed ‘✠.j.’ and line ✠4v1 reading ‘here’ in the outer forme of the outer sheet. The same setting of both formes is found in the copy of 16269.5 that EEBO mistakenly attributes to Cambridge University Library (really British Library C.25.l.2). But while the Huntington Library copy of 16269 apparently has the same setting of the inner forme, its outer forme has ‘✠.i.’ and ‘heare’ respectively. Sheet ✠2:3, however, has the same inner forme in all three copies, but parts of the outer forme differ – most notably where the shoulder-note identifying Nunc dimittis, correctly beside the first line of that canticle in 16268 and 16269, is several lines lower beside the beginning of the Litany in 16269.5.Footnote 45 Meanwhile the other copy of 16268 (Christ Church, Oxford, Gibbs.1), which lacked quire ✠4 when Kenneth Gibbs acquired it, was later supplied by Quaritch with a copy from an otherwise unknown edition printed after the next change in the note about prices.Footnote 46
Grafton did eventually take advantage of the opportunity to reduce the size (and consequently the cost) of reprints, and having taken rather more time over it than Whitchurch had, he did a better job. By eliminating all blank pages, and by being prepared to begin a new section on a verso (or even mid-page) rather than necessarily on a new recto, Grafton was able to print each of his last two editions (collating ❧10, A–R8 S–T6) in only seventy-nine sheets, instead of Whitchurch’s eighty-four.
There had been, however, a different advantage to be gained from Grafton’s earlier preference for repeating an existing structure. It is easier for a compositor to reset a forme page for page and line for line than to start at a mark somewhere in one page of copy (even printed copy) and finish precisely at a similar mark on another page. That advantage applies especially when the printer of the earlier edition asks a second printer to help out with the reprint, because he can simply hand over copies of whatever sheets are needed and request exact copies of them. Both Whitchurch and Grafton did indeed seek such help. Nicholas Hill, who had provided Whitchurch with the presumed cancel X3:4 in the first edition, was also asked to contribute twelve sheets to STC 16270 (K8; R8; N4:5; S1–3:6–8), twenty to 16270a (A1,2:7,8; C4:5; G8; I8; N8; R1,2:7,8; S1–3:6–8), and fifteen to 16273 (G1–3:6–8; I8; N8; R1,2:7,8; S2,3:6,7). Less expected is the presence in three of Grafton’s editions of quires printed by Robert Wyer (a freeman of the Salters’ Company who specialized in small pamphlets: often illustrated, rarely larger than octavo, and averaging fewer than four sheets each). Despite his usual habits, though, he was evidently asked by Grafton to contribute six folio sheets to STC 16269.5 (L6; Cc6), four to 16274 (C8), and twelve to 16275 (C8; H8; N8).
The End of Demand
Ideally, of course, the combined efforts of all the printers actively involved should have produced enough copies to supply at least one to each of England’s parishes, between 8,500 and 9,000 in number. Many parishes, of course, would need more than one copy, and parish clergy were by no means the only parties interested in both reading the book at home (a practice actively encouraged during the Reformation) and following the services while in church. We do not know the size of any of the extant editions, and it is quite possible that one or more small-format editions have completely perished.Meanwhile, however, there was no shortage of conservative critics for whom the 1549 book had gone far too far – or on the opposite side, of champions of reform who felt that the new liturgy retained too many traces of popery. In particular, while the Communion service could hardly be said either to preach transubstantiation or to require belief in it, those most convinced that the ceremony merely commemorates the Last Supper found the rubric and the accompanying words much too easy to interpret as implying a corporal Real Presence.
The body of our Lorde Iesus Christe whiche was geuen for thee, preserue thy body and soule vnto euerlastyng lyfe.Footnote 47
In his Edwardian writings on the Eucharist both before and after 1549 Cranmer repeatedly makes it clear that while Christ is really present during Communion, his presence is spiritual as distinct from corporal, and that the bread and wine remain unchanged in substance both during and after the event. As expressed, however, the 1549 words of administration allowed even diehard conservatives to accept the new ritual by interpreting it (or as its author would have insisted, misinterpreting it) in their own way. This was by no means the only part of the first Edwardian prayer book that failed to satisfy the most ardent reformers, so by the end of 1551 the authorities had decided that it should be replaced by a substantially revised version.