University of Virginia Library

Search this document 


expand section 
expand section 
expand section 
expand section 
collapse section 
Some Bibliographical Irregularities in the Shakespeare Fourth Folio by Giles E. Dawson
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 

expand section 


Page 93

Some Bibliographical Irregularities in the Shakespeare Fourth Folio
Giles E. Dawson

I. Collation and Pagination

THE FOURTH FOLIO OF SHAKESPEARE (1685) was printed in three sections by three different printers, the first of whom Professor Bowers has recently shown to have been Robert Roberts.[1] The collation of F4 is i 2 iA4 A-Y6 Z4, 2B-2Z6 *3A-*; 3D6 *3E8, 3A-4B6 4C2. Roberts' section, which, as Bowers demonstrates, includes the preliminary sheets, contains the Comedies and ends with Z4; the second section contains the Histories and Troilus, Coriolanus, Titus, and Romeo and ends with *3E8; the third contains the rest of the Tragedies and the apocryphal plays. In the second section errors in signatures are uncommonly numerous, and some of these are not without interest. The most serious of the errors consist of the misprinting of sigs. 2B1 through 2C3 as B, B2, B3; C, C2, C3. 2D3 is printed D3. 2E3 is not printed. 2O1 is printed Pp. *3E1 and *3E2 are printed without the asterisks; but *Eee3 and *Eee4 are correctly so printed. Being important in the gathering and folding of sheets, signatures are in most books printed much more carefully than pagination; we do not often meet with such gross carelessness as these sheets exhibit. Nor was it here allowed to pass unnoticed. On the contrary, once these errors were discovered the printer took drastic steps to correct at least those most likely to make trouble. The discovery appears to have occurred while 2O1 was in the press and near the


Page 94
end of its run, for in only about fifteen per cent of the thirty-seven copies which I have examined at the Folger has a press correction been made. In all the rest of the thirty-seven the misprinted signature has been corrected with a pen. No other press corrections were, so far as I have found, made, but in every copy 2B1 and 2C1 have been amended by pen, likewise 2C3 (but not 2B3 or 2D3). In many copies 2C2 has been amended by pen, and in most 2E3, omitted in the printing, has been written in. A typical specimen copy may be seen in the Methuen facsimile, where all the corrections noted above except 2C2 will be found. The hand—most readily recognizable in the added 2E3—is the same in all copies which I have seen. The press correction of one, and only one, of these errors—and that the last significant one made—together with the pen correction of the misprinted copies of this sheet, strongly suggests that the earlier errors were not discovered until this point was reached and that all the other pen corrections were ordered forthwith and effected in this unknown printer's shop. Sections 1 and 3 contain no significant errors in signing.

The pagination of the volume is A1-H6, 1-96; I1-O1, 99-160; O2-X5, 163-254; X6-Z3, 253('243')-272; 2B1-*3E8, 1-328, 3A1-4C2, 1-302. In addition to the errors indicated in this formula which affect the total count of Section 1, the following simple errors occur in all copies (ten in number) in which I have examined the pagination: 109 is printed for 107, 111 for 109, 186 for 190, 187 for 191, 221 for 219, 234 for 246, 243 for 253 bis. In nine of the ten copies 33 is misprinted 23, and in three 164 is misprinted 160, and 169-171. This carelessness in the printing of page numbers is not uncommon either in kind or, I believe, in degree. Far more unusual is the excellent pagination in Section 2, where I have found no errors of any kind. Section 3 falls but little short of this ideal, having, so far as I have found, only one simple misprint: 176 for 167. All the errors in Section 1 noted above (as well as that in Section 3) may be seen in the Methuen facsimile, with the exception of pp. 164 and 169.

II. Reprinted Sheets

Sixteen and a half sheets of the middle section occur in two printings. Most copies, including that reproduced in the Methuen facsimile, contain the original printing, the reprinted sheets having


Page 95
been found in only six copies—four in the Folger Library, two in the New York Public Library. The table below shows the distribution of the reprinted sheets in the six copies.

Discovery of the reprinted sheets (which for the sake of brevity I will hereafter call F5, since they constitute a fifth folio printing) would presumably have been deferred even longer were it not for an obvious clue. For some obscure reason the side rules and the foot rules were in the printing of the F5 sheets omitted.[4] Each


Page 96
normal page is boxed with rules, side, head, and foot, with an additional horizontal rule beneath the headline and a vertical rule between the two columns of text. The reprinted leaves uniformly lack the side rules and the horizontal foot rule,[5] though all others are as in the original F4 printing.

The absence of the rules is the conspicuous difference between the two states of the sheets—the flag which gives notice of their presence. A further comparison shows that one state is a reprint of the other. Except for the rules they do not differ notably in general appearance, but upon close inspection innumerable small differences become apparent throughout. Collation reveals just such variants as we might except.[6] There can be no doubt as to which kind of sheets are the reprinted ones—the rules tell us that much. Collation of the original F4 sheets with the corresponding sheets of F3 (1663-4) and F5 give the expected results: F4 was printed from F3, F5 from F4. In F5 obvious F4 misprints are in the main corrected. Its text shows a marked tendency towards modernization of spelling and towards a regular use of the apostrophe in possessives. New misprints are about as numerous as corrections. None of the alterations point to the hand of anyone but the compositor.

The explanation for the reprinting is not far to seek. The distribution of F5 sheets in the six copies affected—with fifteen in Folger 28 and only one in NYPL Lenox—suggests that the printer of this middle section of F4 either carelessly or fraudulently under-printed these sheets. When the gathering of all sheets into volumes was nearly finished and the stacks were all getting low, a few stacks were exhausted or it was seen that they soon would be, leaving a stock of something more than two hundred different good sheets—how many of each I will explore presently. At what time in relation to the original printing the discovery of the shortages and the subsequent reprinting are likely to have taken place can be determined


Page 97
only conjecturally. Several avenues of approach to this problem must be investigated.

First, what significance can be attached to the omission of side and foot rules? The complete boxing in of type pages with rules was first commonly practised toward the end of the sixteenth century and became increasingly popular in the early decades of the seventeenth, until, by 1630, it was almost universal—though never quite so—in books printed in England. Soon after 1640 we begin to find unruled pages with increasing frequency. Charles Edmonds' Commentaries of Julius Cœsar, 1655, though a folio with double columns is entirely without rules. Burton's Anatomy, 1676, a folio printed in double columns, is ruled exactly as are the pages of F5—horizontal above and beneath the headlines, vertical between columns. The Works of the Learned Sir Thomas Brown, Kt., 1686, a folio in single columns, is printed with double head-rules only, except in the index, the pages of which are ruled exactly as those of F5 are. In fact I have found folios so ruled in every decade from 1600 to 1720. Clearly then the rules of the F5 pages cannot be regarded as direct evidence as to the date of printing. It does seem safe to say, however, that they do indicate either a date substantially later than the original printing in 1685 or a different printing house. For if Herringman had discovered the short count of seventeen or more sheets soon after their delivery to him he would presumably have demanded that the printer of the central section make the deficiency good, and it is scarcely conceivable that the same printer would then have produced reprints so strikingly different from the originals in appearance as these are. We may well ask why any printer at any probable time would have done so, and this is a question to which I can furnish no very satisfactory answer. That the printer did not have enough rules is out of the question. It must have been simply a mistake of some sort. Perhaps the workman actually responsible for the make-up of the formes, not fully aware of the nature of the job in hand, only followed the custom of the house for folio printing in two columns prevailing at that period. Had the printing been done in the shop that had produced the original F4 sheets, and with no considerable lapse of time, such an error would not be likely.

Evidence derived from watermarks leads to much the same


Page 98
conclusion. Virtually all of the normal sheets of the volume—not only in the middle section but in all three—show some variety of a group of watermarks bearing the name DVAVLEGARD beneath a large shield surrounded by a sort of ruffled or fluted collar. I have found here only one completely different and unrelated mark—a small, plain, crowned shield bearing a simple fleur-de-lis and the initials RO, beneath which is the name N. Porte or O. N. Porte.[7] No watermark which I have seen in F4 occurs in any F5 sheet. There, instead, are five different marks—three forming a related family, the other two another family.[8] Since I have not succeeded in finding any of these elsewhere they yield no direct evidence as to the date at which the F5 sheets were printed. But they again argue either for a date later than 1685 or for another printer.

Finally there is the evidence of spelling and the regular use of the apostrophe in possessives. The spelling shows a good deal of what might be called modernization: will for wil, Doll. for Dol., Country for Countrey, warlike for warlick, Lion for Lyon, and the like. The last decades of the seventeenth century and the early years of the eighteenth were a period of marked standardization of spelling—probably not to be matched in any like period before or since. This tendency did not begin in 1685; indeed it had no beginning, but was in my opinion accelerating more or less steadily in the second half of the century. Such changes as I have listed, with the possible exception of the use of apostrophes, would be expected, or at least would not be surprising, at any date after 1650. In fact the same sorts of modernization abound in F4, the actual count of spelling changes (from the F3 text) per


Page 99
page being there higher than in F5. As we should expect, the F4 compositor, while conforming in general to the standardizing trend of the day was not consistent and left many relics of the F3 spelling. The F5 compositor or compositors carried on in the same direction. Much the same may be said too of the increased use of the apostrophe in possessives: the F4 compositor introduced a few of these; the F5 compositor was more consistent and carried this modernization further.

It can hardly be said then that we have any clear direct evidence of a substantially later date for F5 than for F4. The most that we can say is that the absence of side and foot rules, the completely different watermarks, and a strong tendency towards modernization of spelling, taken all together, make it appear highly probable that F5 was printed either in a different shop or at a date a good deal later than the printing of F4. But the employment of another printer would in itself be strong evidence of the passage of time between the two printings. The two come to much the same thing.

And if I reconstruct the probable history of the sheets of the volume correctly it seems unlikely on the face of it that the need for reprinting would have been observed at once. The volume was issued with a variety of imprints—three in all:

  • (1) LONDON, | Printed for H. Herringman, E. Brewfter, and R. Bentley, at the Anchor in the | New Exchange, the Crane in St. Pauls Church-Yard, and in | Russel-Street Covent-Garden. 1685.
  • (2) LONDON, | Printed for H. Herringman, E. Brewfter, R. Chiswell, and R. Bentley, at the Anchor | in the New Exchange; and at the Crane, and Rose and Crown in St. Pauls | Church-Yard, and in Russell-Street Covent-Garden. 1685.
  • (3) LONDON, | Printed for H. Herringman, and are to be fold by Joseph Knight | and Francis Saunders, at the Anchor in the Lower Walk | of the New Exchange. 1685.
Henry Herringman owned in 1685 at least a half interest in the copyrights of most of Shakespeare's plays, and the fact that his name, and only his, appears in all the imprints, always in the first position, makes it appear that he was the principal if not the sole capitalist in the publication.[9] The other booksellers named were probably small investors, each receiving a stipulated number of


Page 100
copies. It is significant that all the copies containing F5 sheets possess, or probably did possess, title-pages with the first and commonest of the three imprints.[10] This fact does not in itself, of course, prove that the shortages were not soon discovered, for it is natural to assume that the stipulated numbers of sheets to go to the booksellers named in the imprints would be counted out for them at once upon publication. Herringman, primarily a wholesale dealer, would retain the greater part of the stock, and except for a number of sheets gathered into volumes for immediate sale I suspect that the rest would have gone into his warehouse in bundles of ungathered sheets. Nothing is known about the date at which the edition was exhausted, but we may safely assume that a work as big and as dear as this remained in print for fifteen or twenty years; had it been otherwise, presumably, another edition would have been called for before 1709, the year in which Tonson brought out Rowe's edition. By 1700, fifteen years after publication, the smaller investors would long since have disposed of their stocks, the principal publisher would have got his money back, and sales, each one of which would be cash in his pocket, would have slowed to a trickle. Herringman's successor,[11] whoever he was, would have in his warehouse 229 bundles or stacks of ungathered sheets, from which, as need arose, a dozen or so copies would be made up. It seems to me a reasonable if untested hypothesis that at some such time the shortage of some sheets was discovered. Sheet 2O3:4 may well have been the first exhausted. A count then revealed that others were near exhaustion, and a calculation of


Page 101
costs and of the probability of future sales showed that it would be profitable to go to the expense of reprinting the seventeen or more sheets required to make the remaining stock good.

The only questions yet to be considered are how many copies of the good sheets of F4 remained and how many different sheets were underprinted. No precise answers can of course be given, but some calculations can be made on the basis of the table above.

While not attempting a census of extant copies, I have myself examined fifty-seven and have received reliable reports on nine others.[12] Six of these, or approximately ten per cent, contain one or more F5 sheets (if we count Folger 13, which, though its single leaf is not indigenous, represents a copy with a whole sheet). Sixty-six is a large enough number to be considered representative, and I therefore assume that something like ten per cent of all copies now or at any time existing would contain one or more of the F5 sheets. This means that the most deficient of the sheets of the middle section of F4—probably 2O3:4—was approximately ten per cent short. If this is substantially correct, then, supposing that the publisher's contract with each of the three printers called for two thousand copies, printer number two delivered only 1800 copies of 2O3:4. Even if the impression was only half this size there remained when 2O3:4 was exhausted one hundred copies of each of the good sheets. There could have been no question then of the profitableness of reprinting the short sheets, assuming that sales were still at an even moderately good level.

At least seventeen different sheets required reprinting. Were there probably more? The table gives us every reason for believing that there were and that others still exist and will be found.[13] Had Folger 28 perished or gone to another collection not seen by me I


Page 102
should have missed three of the seventeen sheets. Folger 33 contains another apparently unique example. The mathematical probability that others would be found if sixty-five further copies were examined is so high as to amount almost to certainty. Since two out of the six copies shown in the table contain four apparently unique examples I am inclined to believe that the number of sheets reprinted in 1700 (or whenever it was) must have been somewhere in the neighborhood of twenty-five.

III. Sig. Li, pp. 123-4, Section I

Though many persons who have used the Fourth Folio must have observed that pp. 123-124 in Love's Labour's Lost are set in 8-point type instead of the 12-point used elsewhere throughout the volume, no one has been sufficiently curious about it to work out the explanation, or at any rate, thought it worthwhile to print it.[14] No variant is involved here, for all copies, including the Methuen facsimile, are in this respect alike.

Throughout F4 normal columns contain 74 lines of type. The four columns of pp. 123-124 contain 91, 90, 91 and 90 lines. All other pages of the L gathering, including those of L6, the leaf conjugate with L1, are printed in the usual type and with the usual number of lines.

The explanation is not difficult. Printing began with the inner sheet, L3:4, probably with its inner forme. L2:5 was printed and perfected next. Then the working off of either the inner or the outer forme of L1:6 was begun. At this point the proofreader discovered that a block of text had been omitted and that some resetting would be required. If possible the resetting must be confined to L1, since L2r:5v was already printed off or so far along that any resetting here would involve the discarding of much valuable paper as well as presswork. And, as we can see, it proved to be possible. Had the printing of L2r:5v not been finished or well along, the resetting could have been, and would have been, spread out so as to avoid such unsightly pages as 123-124 are.

For their unsightliness is not caused only by the smaller type. Additional compression has been accomplished by the elimination


Page 103
of all of the usual white spaces at stage directions, by the printing of two very short speeches in one line (in seventeen places), and (in one place) by forcing three lines of verse into two type lines. A further small gain resulted from the fact that much of the first column of p. 123 is prose, which allows more words to the line in the smaller type. Thus no true measure of the total compression obtained is to be arrived at by the mere counting of lines and the discovery that the four reset columns contain 362 lines instead of the usual 296.

But in another way we can calculate the amount of compression almost precisely. The text which now fills pp. 123-124 (L.L.L. V.i.13 to ii.254, in the Globe edition) occupies 425 lines in F3, counting all white spaces at stage directions. If we assume that this was originally set up in the same way in F4—that is with the same spacing—296 lines would have been accommodated, leaving a balance of 129 lines. Since F3 is set 66 lines to the column, 129 lines is almost exactly one page. We can safely assume then that what the proofreader found missing from the F4 had been one page of the F3 text. This means that the omission was due to the compositor, for only he could have been concerned with F3 pages. Nothing about the F3 text of the passage involved suggests any reason for the omission or any clue as to which page was omitted. We have no way of knowing certainly that the whole first forme, presumably the inner, of L1:6 was not actually wrought off before the detection of the omission. But it appears probable that had this been the case L1 would have been cancelled so that L6 could be salvaged, reducing the loss of paper by fifty per cent. That no such cancellation was made suggests that the omission was discovered in the first forme in the press.



M. W. Black and M. A. Shaaber call attention to the three separately-printed sections in Shakespeare's Seventeenth-Century Editors, 1632-1685 (1937). Bowers' identification is in "Robert Roberts: A Printer of Shakespeare's Fourth Folio," Shakespeare Quarterly, II (1951), 240-246.


The strikingly different appearance of these partially unruled pages is noticed in Contributions to a Catalogue of the Lenox Library, No. V. Works of Shakespeare, Etc. (New York, 1880), p. 41, where the lack of side rules in 2O3:4 is commented on. A previous owner of Folger 7 also observed at least some of the pages without side rules and pencilled inside the front cover a note attempting to explain the conditions: 'Some lls appear to be in proof state before lined borders were printed in.' It was this note which first drew my attention to the reprinted leaves.


At least I have found no exception. It is conceivable that some sheets were reprinted with all rules, but if they were they have escaped me, for they could be discovered only by chance or by an exhaustive (and exhausting) collation, which I have not attempted; I have only borne the possibility in mind and been on the lookout for such sheets.


I have not collated the whole of the 70 pages, but instead made a spot check, collating some fifteen or twenty scattered columns. It then began to appear doubtful that further hours of collating would reveal more significant facts than I had already found.


I cannot claim to have looked at every sheet or nearly every one, but I have seen a great many of them. Edward Heawood, Watermarks (Hilversum, 1950), reproduces (nos. 654, 671) two very similar but not identical DVAVLEGARD marks. Others may be seen in the second folio Beaumont and Fletcher, Fifty Comedies and Tragedies, 1679. The bearings on the shields are various and always complex and unheraldic in appearance. In all I have distinguished four markedly different DVAVLEGARD marks in F4.


Ia a large crowned shield bearing a post horn, with initials WR; Ib a large crowned shield bearing a fleur-de-lis, with initials WR over IG; Ic the same but without IG; IIa grapes with a crooked stem at the top and initials MLP (or M P?); IIb grapes with small fleur-de-lis at top and initials IP. In every sheet except two of the F5 sheets the same mark appears throughout all copies. Nor is that all, for they tend to occur in groups. All the sheets (as shown in the table) from 2I3:4 through 2O3:4 contain IIb; from 2T3:4 through 2Y1:6 contain Ia; from 2Z2:5 through 3B2:5 contain IIa.


G. E. Dawson "The Copyright of Shakespeare's Dramatic Works," Studies in Honor of A.H.R.Fairchild (University of Missouri Studies, 1946), pp. 11-35.


I say 'probably did possess' because one, the NYPL Lenox copy, now has the Knight-Saunders imprint; but Mr. P. N. Rice, chief of the reference department (to whom and to whose assistants I am much indebted for skilful, prompt, and patient help in examining the NYPL copies and answering my troublesome questions) writes that this copy, which has been rebound by Bedford, is sophisticated and that the title and frontispiece 'might well have been inserted in the place of others.' Four of the other five copies have the first title. Of the Folger copies no. 7 appears to be in the original binding and perfect throughout; 28 (rebound) has had at least one interior leaf added (see note 2 above), but all the preliminaries appear to be indigenous; 33 is rebound, but handwriting on the title and on many other leaves proves its title to be the original; for 13, which has no title, see note 3 above. Of the twenty-nine Folger copies of F4 having titles, twenty-three have the first imprint, two the second, and four the third.


Herringman did not die until 1704, but he had turned his retail trade over to Knight and Saunders as early as 1684, as C. W. Miller has shown in "Henry Herringman, Restoration Bookseller-Publisher," Papers of the Bibl. Soc. of Amer., XLII (1948), 292-306. If any stock of the Fourth Folio remained as late as 1700 it is unknown who may have held it.


The fifty-seven which I have seen are: Folger 37, B.M. 4, Trinity College Cambridge 4, Trinity College Dublin 4, Huntington 4, Bodleian 1, Cambridge University 1, Library of Congress 1, Rugby School 1. The nine which I have not seen are: NYPL 6, Harvard 1, Birmingham Public Library (England) 1, University of Virginia 1.


The validity of my statistical approach to this question would be affected adversely if it were to be assumed that H. C. Folger was aware of the existence in some copies of the reprinted sheets and made special efforts to acquire copies possessing them, but all indications are that he was not aware of them and that he acquired the four copies just as he did the other thirty-three. The fact that four out of the thirty-seven copies contain F5 sheets, then, is the result of pure chance which might equally well recur in any other thirty-seven copies. I hope that owners or custodians of Fourth Folios not mentioned in note 12 who find F5 sheets will communicate their findings to me or to Professor Bowers or otherwise make them known.


The use of 8-point type was first pointed out to me in April 1950 by Professor D. S. Robertson, Vice-Master of Trinity College, Cambridge, who had observed it in a fragment of F4 which he owns.


Page 104