University of Virginia Library

The Printing by the Cambridge Press of A Platform of Church Discipline, 1649 [*]
by
Lawrence G. Starkey

A PLATFORM OF CHURCH DISCIPLINE, A quarto printed by the Cambridge, Massachusetts, Press in 1649, is important to religious historians as the foundation of New England Congregationalism. As such this book has been widely reprinted, both here and in England: Holmes describes twenty-six editions and three reissues from 1649 to 1893.[1] To bibliographers, the first edition of the Platform is important as the first extant work of Samuel Green, third Cambridge printer,[2] who operated the Press from

1649 until it ceased to print in 1692. The Platform also presents a number of interesting bibliographical problems, which have long been unsolved. The purpose of this paper is to clear up several such points about the printing of this book.

There has been no agreement about the dating of the Platform. Roden wrote that it was issued "in the late summer of 1649,"[3] but Winship believes that it was printed after 19 October 1649,[4] at which date the following entry was made in the minutes of the Massachusetts General Court:

I feel, however, that the book was printed before the autumn meeting of the General Court and that the wording of the entry in the Court records was copied from the printed title-page. Furthermore, on 17 December 1649 a London printer entered the title in the Stationers' Register; thus either a book or a manuscript had been dispatched to England before the General Court met.[6] Since crossings from New to Old England

in less than two months were unknown at that time, a copy of the Platform printed after the General Court meeting could not have reached England by 17 December. The only reason for delaying the printing of the book until the Court met would have been to make certain of the recommendation of the General Court; if an actual approval had been necessary, however, it is unlikely that an unapproved manuscript would have been sent to England. The evidence, I believe, while it is admittedly inconclusive, points to the Platform having been in print before 19 October 1649. The Court, as a consequence, was not ordering the manuscript to be printed, but instead simply recommending a book already in existence. The language of the Court entry seems to bear out this hypothesis.

Further information about the printing of the first edition of the Cambridge Platform is found in what may be called the first bibliographical document in the history of printing in English North America: a list compiled in 1656 by Stephen Daye and Samuel Green of printing done at Cambridge until approximately 1654, with some sketchy data about receipts from sales, printing costs, and quantities of paper used.[7] This document has been well-studied by Winship.[8] As we might expect, the data about the Platform was the first entry made by Green:

• Sinod booke. he [Dunster] had of Bro: Green 12:00:00
• finding papr. for ye impression
• abate for paper. 6 Rheame ¼ 02:05:00
• ------------------- --------
• Rest. -09.15.00 09:15:00

At the end of Green's part of the list, the £9 15s was added in as Dunster's profit. For the impression the latter supplied paper which Green appraised as worth £2 5s. Apparently Green did the printing, sold the copies to one or more booksellers,[9] took out enough money to pay himself for his labor,[10] then delivered the rest, £12, to Dunster. This procedure differed from that in effect both before and after the printing of the Platform; for every other book listed by Daye and Green, the printer was credited with a specified sum as his payment.

Since the Platform is a book of five and one-half sheets, the six and one-quarter reams of paper would have been sufficient for an edition of 568, which probably may be reduced to about 550 copies because of waste and imperfect sheets.[11]

The title-page of the Platform, which exists in two states, is overcrowded, as in most books printed by the Cambridge Press. It contains twenty-two lines of type (compared with thirty-eight in the text) and gives not only the title and imprint but also the circumstances of the book's preparation and three scriptural quotations. The following transcript is made of the title in its corrected state II.

The Platform collates as follows: 4°, πA6 A-D4, 22 leaves, pp. [2]π I 2-10, 1-29 30-32 (πseries in sq. bkts. immediately following hdl.); $4 (+πA5,6) signed (multiple letters indicate leaf no., as 'Aa' for A2); πA4-5 missigned 'Aaa', 'Aaaa'; πA2-6 in italic.

The title-page is πA1 (verso blank). A preface occupies πA2-6v. The seventeen chapters of the text begin on A1 and end on D3. D3v is blank. On D4 is a table of contents and a list of errata. D4v is blank.

The preface has a running-title, 'The Preface.' (πA2v-6v); there are no uniform running-titles for the text, but abbreviated chapter-titles are used as hdls., A1v-D3. When one chapter ends and another begins on the same page, the two chapter-titles are abbreviated and combined as the hdl. for that page.

The type is predominantly roman, with some italic; the text has side-notes in roman referring to the Bible by book, chapter, and verse; (D1) 38 ll. 155(164) x 98(114) mm., 82R. The same type had been first used by the Cambridge Press in 1645 for printing John Winthrop's A Declaration of Former Passages and Proceedings Betwixt the English and the Narrow-gansets.

Catchwords were used, but in a manner which emphasized the inexperience of the printer. They were seldom set over to the page margin, and there was little or no effort to make catchword capitalization agree with the first word on the succeeding page. Whenever syllables were used as catchwords, the compositor omitted the usual hyphen.

I have examined seven of the nine extant copies, as follows: University of Virginia, 175 x 130; New York Public Library (lacks D4), 181 x 132; Congregational Library, 171 x 127; American Antiquarian Society, 180 x 138; Huntington (microfilm); William Clements (microfilm); John Carter Brown (title has been cut out close to frame and mounted for binding), 177 x 131. In addition, I have received a detailed report on an eighth copy: Thomas W. Streeter, 178 x 132. There is a ninth copy in a private library which I have not seen.[12]

Of the eight copies upon which this study is based, two have an imprint which reads:
[I]

In the other six copies, the first line of the imprint has been altered:[13] [II]

There is good evidence for reversing the order of imprints favored by Winship, who believes that imprint II was first through the press.[14] Associated with imprint I is an error in a scriptural quotation: in the quotation from the 84th Psalm, 'How amiable are thy Tabernacles O Lord of Hosts?', the final 's' in 'Tabernacles' is lacking in the Brown and Streeter copies, which have imprint I. This quotation is found in a correct state in all copies which have imprint II, thus suggesting that the correction and the change of imprint were made at the same time. An examination of the seventh word in the third line of πA6v, which is in the same forme as the title-page, provides additional evidence. In the Brown and Streeter copies, with imprint I, the word is printed incorrectly as 'im', but in all copies with imprint II the word has been press-corrected to 'in'. From this previously unrecorded evidence, I conclude that the letter was changed when the forme was unlocked to alter imprint I to imprint II. Although I believe that these two instances are sufficient to prove the priority of imprint I, I feel bound to mention that the two states of the title-page are characterized by three other differences. With imprint I are found 'Eighth moneth' instead of 'Eight moneth', '84.' instead of '84' (no period), and a swash 'G' in 'GATHERED' instead of a plain italic 'G'. The presence or absence of the swash 'G' is not substantively significant, but the other two differences are manifestly more correct in state I of the titlepage than in state II. Since there is no question of textual alteration connected with them, I would contend that the 'h' in 'Eighth' and the period after '84.' were pulled out when the

loosened forme was inked after it had been unlocked to change imprint I to imprint II and to correct the misprint in the scriptural quotation.[14a] On the contrary, there is no space for an 's' in 'Tabernacle' when the letter is missing, which suggests that the letter was not pulled out in inking but was instead added as a result of press-correction.

There is some additional, though inconclusive, evidence that imprint I was actually first through the press. The separation of town and region in I is awkward. Winship maintains that an imprint would usually begin with the place, which is one of his reasons for regarding imprint I as a revised version.[15] If this reason were valid, one would expect to find Green henceforth setting his imprints with the place first. For the next seven years, however, he invariably set his name first and the place second. Only when, eleven years later, he became associated with Marmaduke Johnson, an experienced London printer, did Green habitually adopt the more conventional sequence in his imprints.[16]

Winship's other argument for the priority of imprint II is that it is found in the same copies as a misprint on iA6 recto, whereas imprint I appears in a copy where the misprint has been corrected. He points out that the leaf with the misprint is "the leaf that is joined to that of the title whether the half-sheet of this fold [i.e., gathering] was folded outside or inside the other four leaves."[17] Actually, this argument would have no bibliographical validity whatsoever for the second of Winship's postulates: if the half-sheet in the quarto (6's) gathering is the inmost fold and the first, second, fifth, and sixth leaves constitute the full sheet, the title on iA1r cannot be in the same forme as the misprint on iA6r, and thus no connection can exist between them. On the other hand, if it is possible to demonstrate what Winship felt was 'futile to guess,' that is, that the half-sheet is the outermost fold iA1.6, then since all four type-pages of the two leaves could have been imposed in the same forme—provided the fold were printed by half-sheet imposition, —any argument based on a relation between the misprint and the title must be scrutinized carefully. The problem, therefore, must be attacked from two points of enquiry: (1) which fold in the six-leaf preliminary gathering was printed as a half-sheet; and (2) if this fold was iA1.6, was it printed by half-sheet imposition (the only method which could bring the type-pages for iA1r and iA6r together in the same forme) or in some other manner which would separate them by formes.

(1) Following the title-leaf, the first gathering of the Platform continues with a ten-page preface, the whole quarto gathering being composed of six leaves and thus necessitating the first use of a half-sheet by a Cambridge printer. This poses to the bibliographer the nice problem whether Green quired the half-sheet within the folded full sheet as would have been

normal printing practice, or whether he printed the title on it and wrapped it around the full sheet. Although the problem has been thought insoluble,[18] watermarks conveniently provide the answer. By good fortune in the copy held by the American Antiquarian Society (the other copies are ambiguous[18a]), the watermarks link iA1 with iA6, and iA2 with iA5. Thus iA3.4, the inner fold, cannot be the half-sheet, for that would mean that each of the quarto leaves of a full sheet iA1.2.5.6 would have a watermark:an impossibility. The only conclusion is that iA1.6, the fold containing the title-page and the end of the preface, must be the half-sheet.

(2) Having established that the outer fold is the half-sheet, we may now turn to the question of its printing, for a bibliographical connection can exist between iA1r and iA6r only if the fold were printed by half-sheet imposition, that is, by placing all type-pages in one forme, with printing and perfecting of a full sheet being made from this forme and the halves of the full sheet subsequently being cut apart to furnish two identical copies of the half-sheet. First, however, it is necessary to examine what are the actual facts of coincidence between this iA6r misprint and its correction[19] in relation to the two states of the title. Winship's facts are in error here, for the misprint on iA6r is not, as he states, corrected in copies with imprint I though uncorrected in all copies with imprint II. Instead, this misprint appears in one of the preserved copies with imprint I (John Carter Brown) and is also found in two copies with imprint II (University of Virginia and Huntington) although

corrected in the other four copies with imprint II which I have examined and also in the Streeter copy with imprint I. This evidence puts a quite different complexion on the problem, for it demonstrates (a) the misprint was not corrected at the same time as the alteration in the title; (b) imprint I must have been first through the press. Determination of the precise method of printing thus becomes doubly necessary if we are to untangle the proper explanation for these facts.

Had iA1.6 been printed by half-sheet imposition, the type-pages must necessarily have been imposed in a single forme in the following relation to each other:

If we begin normal printing from this forme and lay each successive piece of paper, printed on one side only, on a pile, we should start with imprint I of the title, the misprint on iA6r, and the misprint on iA6v. The series of sheets printed with this state of the type we may call series X. As the second step in the printing, the press is stopped, the title is altered to imprint II, and coincidentally the misprint on iA6v is corrected. A second series of sheets, series Y, is thereupon printed on one side only with these characteristics and laid on top of series X in the gradually mounting heap of wrought-off sheets. Somewhat later the misprint on iA6r is detected, and the press is stopped to make this correction.[19a] The remaining sheets, series Z, are thereupon printed and laid on the pile in order.

To complete the process this whole pile is turned over so that series X is on top, and perfecting is executed, all three series being perfected by the forme in state Z. When this operation is followed, and the full sheets cut in half to give us the iA1.6 folds, we observe that we have secured a proportion of states which closely approximates those in the extant copies. The largest number of half-sheets contains imprint II and the corrected readings on iA6r and iA6v. A smaller number gives us the state of the Virginia and Huntington copies, with imprint II, iA6v corrected, but iA6r uncorrected. Finally, we have the smallest group, containing imprint I, iA6v uncorrected, but iA6r corrected—that is, the Streeter copy. It is clear, therefore, that if printing proceeded by half-sheet imposition as outlined above, no copy could be produced which would correspond with the John Carter Brown copy with imprint I, although in all other respects we have variants corresponding with the other known copies and in approximately the correct proportions.

There is, however, another alternative.[20] If the correction of the misprint in iA6r did not take place during the printing of the white-paper but instead was performed during the operation of perfecting, then if all of series X and a certain number of series Y (there would be no series Z of white-paper) had been perfected with the forme in the Y state and the press were stopped to correct iA6r (constituting state Z) during the early perfecting of the Y sheets, we should indeed have copies produced which agree exclusively with the John Carter Brown exemplum and the two known states of imprint II, but none at all of the state represented by the Streeter copy with imprint I.

For these reasons, it is necessary to enquire whether another

method of printing might not have been adopted which would produce copies in the states observed and in proportion to their preservation. This method is to be found in printing by cut sheets. According to this rather elementary procedure, familiar in the earliest days of printing, the full sheets were cut in half before any were printed, and thereupon each half-sheet was treated as a separate sheet, being printed from one forme and perfected from a different forme (inner and outer).

If this method were employed for the half-sheet in the Platform, the inner and outer formes would each have been made up from only two type-pages, as follows:

Whether the inner or the outer forme was first through the press is undeterminable and here of no consequence, since under any circumstances uncorrected printed white-paper is perfected by an uncorrected forme at the start and the overlap—as represented by the Virginia and Huntington copies—occurs according to the unequal proportion of each machined as a separate operation.

Although printing by cut sheets is a primitive method as compared with half-sheet imposition, there is every indication from his work that Green was not a sophisticated workman and that he may well have prided himself on successfully solving the problem he faced, especially if—as likely—he had never been instructed in the technique of half-sheet imposition. If we believe that iA1.6 was indeed printed by cut sheets, we are enabled to explain without difficulty the particular proportion of extant copies in each state, a matter impossible to explain by any theory of half-sheet imposition. Moreover, the difference

in time between the correction of misprints on iA6r and iA6v is more readily accounted for if they are in different formes than if we must assume in half-sheet imposition that two separate correction operations were made in the same forme. By cut-sheet printing only a relatively few copies of the inner forme need have been printed before the correction on iA6r was effected, and, indeed, this is the direct import of the evidence of the Streeter copy.

The central bibliographical fact with which we are concerned is clear. Printing was by cut sheets; moreover, imprint I must have been first through the press, and the alteration of this state to the form of imprint II has no causal connection with the correction of the misprint on iA6r as has been asserted. The two type-pages were in different formes, and hence alterations to these pages are from a bibliographical point of view completely independent even though the leaves are conjoined.

Commencing with iA2, and continuing with the subsequent gatherings in 4's, the recto of every leaf in the Platform bears a signature, instead of the two, or at most three, leaves customarily signed in a quarto gathering, all that are necessary for a binder. Green seems to have devised his own method of signing, using letters exclusively. The rectos of each gathering were signed with a combination of capital and lower case letters, for example: B, Bb, Bbb, Bbbb. The letter of the alphabet denoted the gathering, the number of letters, including the capital, the leaf. In the four gatherings of the text of the Platform, this system is worked out perfectly; but in the preface, printed last of all, there is some confusion, doubtless because of the quiring of the iA gathering as a quarto in 6's. The rectos of the preliminary gathering are signed in succession after the title page: Aa, Aaa, Aaa, Aaaa, Aa5. The last leaf, signed Aa5, was, of course, iA6; apparently Aa5 was intended as an abbreviation for Aaaaaa, i.e. capital A and five lower-case a's. The fourth and

fifth leaves of the gathering were incorrectly signed Aaa and Aaaa. The error was natural, for Green had only to forget for a moment that he was later to have a half-sheet; the two leaves in question would be correctly signed as the last two leaves of a quarto gathering like the others in the same book. Green abandoned this system of signing after printing the Platform, and similar signatures are never again found in the books printed by the Cambridge Press.

The list of nine errata found on D4r (of the inner forme) is unable, because of its position, to correct any errors on D1v, D2r (D3v is blank), or in the table of contents on the errata leaf itself, where two chapters are listed as beginning on the wrong pages.[21] That an erratum is listed for D1r proves that the outer forme of the D gathering was first through the press. Since the list of errata corrects none of the several errors in the preface, either, we may readily assume that the iA6 was printed last, as would be expected. The designation of the errata as 'faults escaped in some of the books thus amended' seems to indicate that corrections had been made in the text by stopping the press, and that one might expect to find the correct readings in some copies. A collation of eight of the nine extant copies, however, discloses that none of the errata was corrected. Because we have less than two per cent of the edition, it is impossible to tell whether corrections were actually made or whether Green was trying consciously to give the impression that he was a much more careful printer than actually he was. We know that later, while printing the preface, he did stop his press to correct errors in the half-sheet. He seems, however, to have paid no attention to several errors in the full sheet (iA2-5) of the preface, including one that was particularly noticeable: the misspelling of 'Preface' as 'Prefae' in the headline on iA4r.

Notes