University of Virginia Library


Page 145

Mark III: New Light on the Proof-Reading for the First Folio of Shakespeare

IN 1932 DR. E. E. WILLOUGHBY PUBLISHED, IN A now classic monograph, a reproduction and a brief discussion of the sole example of corrected proof for the First Folio of Shakespeare then known.[1] The present writer described a second example in an article published in 1942.[2] It is now possible to present a third example;[3] and an analysis of this, particularly in the light of evidence furnished by the other two, raises a number of questions about stop-press correction that may be of interest—and that seem to point to the soundness of certain conjectures that could heretofore be made only with great diffidence. Not that complete confidence is now possible: the new evidence is slight and rather slippery—but all the more interesting, perhaps, for this very reason.

The first of our three concrete examples of First-Folio proof-reading occurs in an isolated leaf upon which are printed two pages of Anthony and Cleopatra. One side of this leaf shows the uncorrected state of page 352 (xx6v) of the First Folio text. This page is clearly marked with a proof-reader's corrections, and these corrections, and these only, are the ones embraced in the text


Page 146
presented by most of the surviving copies of the book. Quite probably, as is strongly suggested by both our other examples, this single leaf was once a part of a complete First Folio. Otherwise its chances of survival would have been extremely poor; and it is known to have been discovered in the 19th century "in a parcel of fragments of the First Folio"[4] (which seems more likely to mean in a group of leaves from a First Folio than, for instance, in a bundle of printer's waste from Jaggard's shop). In any event the surviving leaf is certainly only a part of what once was a full sheet. But what the other leaf of the sheet looked like—whether or not page 341 (XXIr), the conjugate forme-page of page 352 (XX6v), also contained a proof-reader's corrections—this, unfortunately, we cannot say. The question, however, is not so idle as it may at first seem, for it is basically a question about proofing methods, methods of which we would gladly know much more than we do, and about which our other examples appear to have something to teach us.

The proof-sheet first fully described in 1942 is to be found in the Jonas copy of the First Folio, No. 47 in the Folger collection. Forme VV3r:4v (pages 333 and 336, in Othello) is in the uncorrected state. Page 333 shows, in what is apparently the same hand and the same ink as the marks in the Anthony and Cleopatra page, the proof-reader's marks that resulted in the corrected state of this page. The conjugate forme-page 336, however, is unmarked.

Another First Folio in the Folger collection (No. 48, once the W. T. Spencer copy) provides our new evidence, this time in a section of the book involving two further tragedies, Hamlet and King Lear. Page 281 (qq1r, in Hamlet) presents a special problem, to be considered later in this paper. Let us first look at the conjugate forme-page 292 (qq6v, in Lear). Here, in what seem again the same hand and ink as before,[5] a third set[6] of proof-reader's corrections appears. (See Plate I.) The changes called for are not numerous—only five in all. These changes were all made, and must have been made, indeed, before any large number of uncorrected


Page 147
states had been printed off: 76 of the 82 copies examined (or about 92.7%) show the corrected state of this page (as in Plate II); only 6, including the copy with the proof-reader's marks, (or about 7.3%) were printed before correction was effected. The variant pairs are given below. (Positions are identified as to column within the page and line number within the column. The additional references in parentheses are to act, scene, and line in the Globe edition of King Lear. The uncorrected readings, which can easily be examined further in the Lee facsimile, where no proof-reader's cancellations obscure them, are given first.)

  • 1. a 21 (II, ii, 52-3)
    liues, he dies that▊ſtrikes
    liues, he dies that ſtrikes
  • 2. a 46 (II, ii, 80)
    holly cords atwaine
    holy cords▎ atwaine
  • 3. b 6 (II, ii, 104)
  • 4. b 31 (II, ii, 129)
  • 5. b 46 (II, ii, 141)
    ſi ttill
    ſit till

It may be noted of these variants that the earlier state is in every case wrong, but that correction in no case implies reference to copy. The errors are all patently compositorial, with the possible but improbable exception of "holly" for "holy" in a 46. Neither word appears in the quarto text and one or the other must rest on the authority of some manuscript.[7] But even if the earlier Folio reading accurately reproduces the copy for the Folio print, both the error and the correct reading are sufficiently obvious to be recognized by the proof-reader independently of any other "authority." We may also observe of this variant that the correction was not faultlessly effected, since the space-quad added after "cords" to compensate for the lost "l" was not properly pushed down and so took ink. There is, perhaps, nothing very surprising in the fact that the correction of a 21 not merely removed the faulty setting after "that", the only thing in the line that was marked by the reader, but also resulted in improved spacing before


Page 148
"he": the wider spacing here is an evident consequence of reducing the space between "that" and "strikes". Yet it is clear that the correction called for could more easily have been made otherwise, without disturbing so much type, and that the compositor who effected the required change was interested in producing generally well-spaced lines when possible; that, in short, some attention was paid to typographical refinement. And this is also shown by 2 (despite the inking space-quad) and by 5, though in this instance it is the proof-reader rather than the compositor who is responsible for a minor improvement—an improvement remarkable for its very triviality. Between this kind of thing and a serious and sustained concern for textual accuracy, however, there remains a very wide gulf.[8]

Page 281 (qq1r, the conjugate forme-page of 292) presents a special problem because it obliges us to define variant somewhat more narrowly than heretofore. A collation of 80 copies of this page reveals no variants in the sense either of verbal differences or of differences in spelling, punctuation, capitalization, or the like. Yet the page seems nevertheless to have been proof-read; for a minor typographical imperfection, although of a kind to which we ordinarily have paid very scant heed, appears to have been corrected in it, and corrected at the same time as the five small errors in page 292 listed above. The facts are these. In b 33 on page 281 (corresponding to V, ii, 358, in the Globe Hamlet) the last two letters of "thee" for some reason stood so high in the original setting that, whereas these two letters seem very heavily inked in the printed page, the "h" seems lightly inked and the right-hand part of it is barely perceptible (see Plate III—or the Lee facsimile). The word so appears in all 6 of the copies which show the uncorrected state of page 292. In all but one of the 74 others which show corrected page 292 all four letters of the "thee" in page 281 are uniformly inked (see Plate IV for a random example). The one exception is provided by Folger copy No. 40, which shows the corrected state of page 292 but the ill-printed "thee" in page 281. This, however, is surely the exception that proves the rule. For close examination demonstrates that the two leaves in question are not in this copy conjugate, as is perhaps most readily shown by


Page 149
the fact that both qq1 and qq6 here bear a watermark: Folger No. 40 is clearly a "made-up" First Folio.

The absolute constancy of the relationship between the two settings of "thee" in page 281 and the two patently variant states of the text in page 292 clearly indicates that the changes in both pages came about at the same time, after the single unlocking for correction to which we know the forme was subject; and if all of these changes were intentional, it follows inevitably that page 281, no less than page 292, presents us with truly variant states that are the result of proof-reading. For we must certainly accept the principle that whenever a text is deliberately changed in the course of its transmission, genuine, if not always "substantive," variants are produced. "But how," it may be objected, "can we know that the change in page 281 was intentional? Obviously enough the crucial issue is not the relative triviality of the difference between two supposed 'states' but whether human volition can properly be adduced to account for this difference; and what is there to make us suppose direct human agency here responsible?"

Unhappily, page 281 does not show any sign of the proof-reader's mark that would resolve this difficulty. The suggestion presents itself, moreover, that the two insufficiently pushed-down types in page 281 b 33 might just possibly have assumed their proper places, as it were by accident, when the forme was loosened by the unlocking required for the correction of the errors in page 292. Such an explanation, however, will hardly do. It is initially unsatisfactory, especially in the light of the extreme unevenness of the "thee" in the 6 copies here considered uncorrected. Are such completely even impressions as we find in the 73 corrected copies likely to be the accidental result of work on the other page of the forme? How are we then to explain, if even impressions are thus easily and unintentionally produced, the unevenness in the stage direction "Enter Osricke." a few lines farther along in page 281? Here the "c" was somewhat high in the original setting, so that the preceding "i" inked poorly; and approximately the same unevenness appears to be characteristic of all copies. And the same thing is true of the "cracke a Noble heart" about ten lines below, where the consecutive letters "e a N" were a little high and therefore seem overinked, though in slightly different degree—in all copies. Only in b 33, then, does unevenness give way to perfect


Page 150
regularity of impression. The difference is fairly extreme and the change occurred, not gradually, but at once; and it occurred at a time when five other small changes were introduced elsewhere in the forme, demonstrably as a result of proof-reading. The change in question is not likely to have been accidental, as other evidence from nearby lines indicates. Deliberate alteration as a result of proof-reading seems the only reasonable explanation.[9]

If then we are obliged to assume that page 281 was proof-read, why does it not show, like its conjugate forme-page, the proof-reader's marks? Probably for the very simple reason that the proof-reader, unlike the pressmen, considered the single page, rather than the two-page forme, his basic working unit. The peculiarities of another pair of conjugate forme-pages, 333 and 336, in Othello, will perhaps now be recalled. Both pages exist in two clearly distinguished states, although only a single correction was made in page 336; and the corrected and uncorrected states of the one page are invariably found paired with the corresponding states of the other. In the Jonas copy, moreover, page 333 carries the proof-reader's corrections; but page 336 is unmarked. "The most satisfactory explanation of these facts," it was suggested in 1942, "seems to be that the proof-reader handed his corrections to the compositor as soon as he had read one page of an impression taken from the forme for proofing purposes, and then used another impression as proof for the second page. Such a procedure is entirely plausible, since it would obviously save considerable time. (It is also possible, of course, that an impression from the uncorrected forme was given to each of two proof-readers and that one read p. 333 while the other was reading p. 336. This method would also save valuable time; but there is no need to postulate an additional workman.) We are doubly unfortunate, therefore, in that the


Page 151
Anthony proof-sheet is fragmentary. Since it represents only half of a two-page forme . . . it can provide us with no further basis for conjecture as to how commonly the method that seems to have been used in the proofing of VV4v:3r was employed elsewhere during the printing of the Folio."[10]

Now, however, we have new evidence. For in pages 281 and 292 of the Spencer copy we appear to find precisely the same peculiarities as in pages 333 and 336 of the Jonas copy. And hence we may conclude that, at least some of the time and when dealing with that portion of the First Folio that includes Hamlet, Lear, Othello, and (probably) Anthony and Cleopatra, the proof-reader returned a sheet after correcting only one of its pages. Presumably —for only thus would such a practice effect real gains—he then proceeded to correct the other page in another copy of the full sheet. The required alterations could by this means be begun sooner than otherwise; and by the time these changes had been made in the first type-page the reader would be able to return his second proof-page, suitably marked for immediate correction.

Whether or not this practice was the usual one throughout the printing of the First Folio, and perhaps of a great many other books of the period as well, we still cannot say with assurance; but certain suggestions present themselves. Probably it would not be an efficient method whenever large numbers of errors were likely to be encountered (so that the second proof-page might not be marked and ready as soon as required); nor is it a method that would ever be employed when the reader contemplated a careful comparison of proof-page and copy, since really conscientious proofing of this kind would certainly require much more time than a rapid reading of which the primary intent was merely the elimination of obvious typographical error.[11]

The implications of this last consideration are obviously of some moment; and although the subject cannot here be farther


Page 152
pursued, at least one of these implications may be mentioned. If the stop-press variants in the First Folio, which we now have good reason to believe very abundant, were ordinarily the product of a method which itself precluded correction by reference to copy, it follows that we should be particularly interested in the readings of the various formes that make up the volume before these were subjected to a process of arbitrary sophistication. Many of the First Folio's variants will no doubt be found to involve only the "accidentals" of which Dr. Greg speaks elsewhere in the present volume and to which less interest usually attaches than to "substantives." But substantives too there will unquestionably be; and a careful attention to the uncorrected members of variant substantive pairs may sometimes enable us to get closer to what Shakespeare himself wrote than the corrected readings that are but the swift guesses of a printing-house reader who seems to have been more concerned with the appearance than with the accuracy of the Folio text. It is still, of course, imprudent to generalize too widely about the press-correction method, or methods, employed during the printing of the First Folio; but all of the evidence so far available certainly indicates the customary use of the none-too-rigorous method described above.

The examination of forme qq1r:6v throughout the entire Folger collection of First Folios brings to light one further and rather surprising fact that may also, finally, be recorded here. In addition to Folger No. 48, the copy which carries the proof-reader's mark on page 292, the copies showing the uncorrected state of the forme are Folger Nos. 15, 31, 47, and 69; and the Lee facsimile reveals that the Chatsworth-Devonshire copy is also uncorrected in this forme. (Folger No. 40, clearly a "made up" copy in which the two leaves in question are not conjugate, is uncorrected in page 292 only.) And now let us once more recall forme VV3r:4v in Othello. The uncorrected state of this forme is found in Folger Nos. 15, 31, 47, and 69; and the Lee facsimile also shows the uncorrected state. In short, except that qq1r:6v is also uncorrected in No. 48 and in one page of made-up No. 40, the very same copies and no others are uncorrected in two different formes—and in two formes, one involving Hamlet and Lear, the other Othello, that cannot have been dealt with consecutively in Jaggard's printing-house. Clearly

Plate I

Page Plate I

Plate II

Page Plate II

Plate III

Page Plate III

Plate IV

Page Plate IV


Page 153
there must have been a more systematic method of separating the sheep from the goats, and of keeping the two breeds apart, than we have sometimes supposed.[12]



E. E. Willoughby, The Printing of the First Folio of Shakespeare (1932), Frontispiece and pp. 62-64.


Charlton Hinman, "A Proof-Sheet in the First Folio of Shakespeare," The Library, 4th ser., XXIII (1942), 101-7. Photographs of the pages and states discussed accompany this article.


All three are among the collections in the Folger Shakespeare Library in Washington. I am deeply indebted to the director and staff of this great library for innumerable kindnesses, including permission to publish the reproductions that accompany this paper.


Willoughby, op. cit., p. 63.


As Drs. Willoughby, McManaway, and Dawson all agree after independent examinations of the pages in question; but the evidence, as these experts also agree, remains too scanty for really positive identification.


Third, of course, only in order of discovery. The three sets came into being in just the reverse order: the present set (page 292), the Othello set (page 333), then the Anthony set (page 352).


See "The Copy for F," pp. 7-18 in George Ian Duthie, Shakespeare's "King Lear" (1949), esp. p. 17.


With one notable exception in the Othello proof-page (333), there are no corrections that imply reference to copy in any of the three proof-pages now known.


To readers who are not satisfied by this argument I can only offer an appeal to the evidence, as yet unpublished, that gives me my own greatest confidence in the conclusion expressed above. During the summer of 1949 I collated pages 305-40 (18 formes that include the whole of Othello and parts of Lear and of Anthony and Cleopatra) throughout the entire Folger collection of 79 First Folios. The work was done with mechanical aid made possible by a generous grant from the Old Dominion Foundation. Among the numerous variants that my "machine" turned up was abundant evidence that the proof-reader was constantly at pains to correct minor typographical infelicities—including space-quads taking ink (however slightly) and several uneven impressions. On more than one occasion the press was stopped for the sole purpose of correcting a single badly inking type.


Hinman, op. cit., pp. 103-4.


Of some interest in this connection are the statistics presented above: approximately 7.3% uncorrected as against 92.7% corrected states in a total of 82 copies examined. These percentages suggest, for an edition of 1250 copies, a total of about 90 uncorrected states, or probably rather less than half an hour's work by an uninterrupted press. To remove a forme from the press, to unlock it and correct one of its type-pages, would presumably take only about the same time, at most. But even if it took slightly longer, it would surely not take so much time as would be required for a proof-reader to make a thoroughgoing collation of a proof-page with its copy.


It will be interesting in this connection to learn (I do not myself yet know) which copies contain the uncorrected state of XXIr:6v in Anthony and Cleopatra. Familiarity with other variant formes among the Folger First Folios makes me bold to declare, however, that we cannot ordinarily expect any such strict correspondence as the one just noted.


Page 154