University of Virginia Library

Search this document 


expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
collapse section 
Cases and Compositors in the Shakespeare First Folio Comedies by Paul Werstine
expand section2. 
expand section3. 
expand section4. 
expand section5. 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 

expand section 

Cases and Compositors in the Shakespeare First Folio Comedies
Paul Werstine

Much bibliographical research in the twentieth century has been devoted to identification of the type-setters who transmitted Shakespeare's plays from manuscript to print. Recent qualitative study of the compositors' work has justified the labour expended in compositor identification, since accuracy in transmission has been shown to vary widely from one workman to another.[1] Before Shakespeare's modern editors can further explore these qualitative differences, they must be convinced that identification of the compositors' stints in the plays is sound. Secure compositor identification,


Page 207
grounded on close analysis of a wide range of evidence, must logically precede qualitative analysis of the compositors' accuracy.

Compositor identification in the Shakespeare First Folio has occupied scholars since 1920 when Thomas Satchell discovered two compositors at work on the Folio Macbeth. Yet Charlton Hinman's Printing and Proof-Reading of the First Folio of Shakespeare of 1963 stands as the major contribution to the study. Hinman revolutionized compositor study when he plotted the recurrence, throughout the Folio, of individually distinctive types to identify the cases of type Jaggard's compositors used.[2] Often Hinman was able to associate individual compositors with specific cases of type: Compositor D usually set type from case z, Compositor B generally worked at case y, and Compositor A at case x. Founded on case evidence as well as spelling evidence, most of Hinman's compositor attributions have resisted convincing challenge for over fifteen years.

Only in quires D-H of the Comedies was Hinman unsuccessful in identifying cases. Quires D-F, he was forced to conclude, were apparently printed concurrently with non-Folio matter, very probably Wilson's Christian Dictionary, according to type-batter evidence visible in both works (Hinman, II, 370-371). Types from wrought-off Folio pages of quires D-F were often distributed, according to Hinman, for use in Wilson's work, which was set from the same cases employed in Folio production. Pages of the Dictionary in turn furnished types to Folio pages. Such irregular distribution practices deprived the analytical bibliographer of intelligible patterns of type recurrence in the Folio and thus prevented case identification. Irregular distribution persisted during work on quires G and H, but here Hinman found no evidence of concurrent printing. Instead, in these two quires, according to Hinman, different compositors frequently distributed parts of the same pages or columns (II, 386, 390). Such distribution yielded patterns of type recurrence—especially in quire G—which were at odds with those Hinman had plotted throughout most of the Folio where only two cases were used. When type-recurrence evidence in quires G and H failed to meet the norms for such evidence Hinman had developed, he relinquished the problem of case identification in G and H to later investigators. However, before he abandoned the problem, he left a wealth of evidence and valuable direction to researchers. Without intensive study of Wilson's Christian Dictionary, not attempted here, no one can now expect to solve all the problems of Folio quires D-H, but Hinman's suggestions regarding distribution practices in quires G and H offer a basis for re-examining his type-recurrence evidence in these quires. Such a re-examination may also illuminate problems in surrounding quires.

While scholars have yet to pursue Hinman's suggestions, bibliographical


Page 208
work on the Folio Comedies has proceeded in other directions. Although Andrew Cairncross's reassessment of Hinman's compositor identifications in the Folio has not won assent,[3] later investigations by T. H. Howard-Hill and John O'Connor have proved valuable.[4] Both these scholars were forced to rely only on spacing and spelling evidence to identify the compositors of quires D-H, but their skillful and conscientious analyses of this evidence strengthen the basis for study provided by Hinman. Together these scholars have confirmed Hinman's association of individual compositors with specific cases: between them, they have now identified all known z-case Folio pages as Compositor D's work. However, since spelling evidence in the problem quires of the Comedies is tricky, Howard-Hill and O'Connor still disagree on compositor attribution for some pages and columns. Re-examination of Hinman's typographical evidence made possible in part by their studies may resolve their disagreements.

Quire H

Compositors and Cases

Quire H provides the point of departure for this study because here for the first time in five quires Hinman discerned a largely intelligible pattern of type recurrence, and here later scholars have achieved substantial, if not total, agreement regarding compositor identification. Hinman recognized two compositors, B and D, at work, as well as many indications "that all the non-B pages were set from one case . . . and that H4v, H5, H6, and H6v [Compositor B's pages] were set from a different case" (II, 390). Howard-Hill later detected Compositor C's work on H1, 1va, 2b, 2v, 3a, and 4a; O'Connor agreed to these re-attributions but also gave H3v and the last twenty lines of H5v to Compositor C. Howard-Hill's discovery that three different compositors, B, C, and D, all worked on quire H introduces the possibility that the three may have worked from three different cases here, just as they did in quires N-Q (Hinman, II, 414-438). Without the knowledge that three compositors were at work, Hinman overlooked the possibility of three cases, but in his differentiation of Compositor B's case from the case used to set the non-B pages, Hinman provided the key to separating the cases used by B's fellow compositors. Hinman observed that in quire H Compositors B and D "commonly distributed type from the same wrought-off pages. As a rule they concerned themselves with different whole columns of those pages. As a rule but not always. Types from upper G6a reappear in H5 [B's page], a type from lower G6a goes to the forme-mate page [H2v, now known to be the work of Compositor C]. . . . Distribution practices that are not ordinarily in


Page 209
evidence in most of the Folio, in short, are certainly in evidence here" (II, 390). In evaluating the possibility that each of the three quire-H compositors used a different case, we must entertain the second possibility that, not just Compositors B and D, but all three compositors may have shared distribution of the same pages and columns.

Table I, a revised form of Hinman's graph for quire H, demonstrates that the quire was indeed set from three cases.[5] Since Hinman has already differentiated Compositor B's case from the others (II, 390-391), I shall concentrate only upon distinguishing from one another the cases used by Compositors C and D. For example,

i) Column G5b provides types to Compositor D's columns H3b and H4b as well as to the disputed page H3v (see lines 2, 3, 7.1, 8.2, 9.3, and 11.2 of Table I), but none to any page or column accepted by all scholars as set by Compositor C. Lower column G5a provides a type to each of Compositor C's page H2v and column H4a (see lines 5 and 17 of Table I), but none to any page or column set by Compositor D and none to H3v. However, upper column G5a does provide types to page H3v (see lines 1.1 and 1.2 of Table I). Had all of page G5 been distributed into a single case from which both Compositors C and D worked, we might expect that at least one or two of the nine identifiable types in quire H drawn from column G5b and upper column G5a would recur in pages or columns set by Compositor C. None does recur. Yet the presence of types from lower G5a in Compositor C's work confirms and extends Hinman's observation that the compositors of quire H sometimes shared distribution of the same columns from wrought-off pages. The observation is extended further in many of the following examples.

ii) Column G4b provides a type to each of Compositor D's columns H3b and H5va (see lines 8.2 and 23.2 of Table I), but to no page or column set by Compositor C; column G4a provides a type to each of Compositor C's column H3a and page H2v (see lines 11.1 and 18 of Table I), but not to any column or page set by D and not to the disputed page H3v. This evidence suggests that Compositor C distributed column G4a into one case, and Compositor D distributed column G4b into another.

iii) Lower column G6b also furnishes types to Compositor D's H5va as well as to his column H2a (see lines 22 and 23.2 of Table I), but to no work set by Compositor C, although upper G6b furnishes a type to the disputed lower column H5vb (see line 23.1) and lower G6a furnishes a type to Compositor C's page H2v (see line 18).

iv) Lower column H3a furnishes types to Compositor D's upper column H5vb and column H1vb (see lines 8.1 and 9.2), whereas only the middle of H3a furnishes a type to a column set by Compositor C, column H1va (line 9.1).

v) Column G2b furnishes types to Compositor C's H2v, H3a, and H4a


Page 210


Page 211
(lines 5, 8.1, 9.1, 9.2, 10 and 18), but to no non-C page, nor to the disputed page H3v. Column G2a furnishes types to H3b and H4b (lines 4 and 11.2), that is, to two columns set by D but to none by C.

vi) Column G1vb supplies types to Compositor C's H1, H3a, and H4a (lines 5, 7.2, 11.1, and 26), but to no work set by D, nor to H3v.

vii) Column H4b supplies types to Compositor C's H1 and H1va (lines 4 and 6), but not to Compositor D's H1vb and not to H3v.

viii) Column H4va supplies types to Compositor C's H1va and H2b (lines 12 and 13), but none to Compositor D's columns on the same pages, H1vb and H2a, and none to H3v.

Table I also shows that the upper parts of G2vb (l. 10), G3vb (l. 26), G5va (ll. 5 and 18), and H3vb (l. 1.1), as well as all of G1b and H5b (ll. 18 and 19), supplied types only to Compositor C's undisputed pages and columns. Lower column G5va (l. 6) and all of G6v, H3b, and H3va (ll. 1.2, 8.2, 9.3 and 23.2) supplied types only to Compositor D's undisputed pages and columns; the middle of G4vb provided one type to Compositor D's H2a and another to disputed page H3v (ll. 3 and 22); and the middle of G2vb and lower G3a furnished types to H3v alone. Thus type-recurrence evidence indicates that while Compositor C worked from one case, Compositor D worked from another. It further indicates that the disputed page H3v was set from the case at which Compositor D set his portion of quire H. In order to accept such overwhelming type-recurrence evidence, we need admit only the assumption that the three compositors distributed different columns of the same pages and different parts of the same columns, the very assumption Hinman required in order to distinguish Compositor B's case from the single other case Hinman believed was in use during work on quire H. On the basis of type-recurrence evidence in quire H alone, it is not yet possible to identify any of the three cases as the familiar x, y, and z cases. Thus I have used the case designators r, s, and t in Table I.

Case differentiation can be put to immediate use in resolving the dispute over compositor attributions for page H3v and lower column H5vb, assigned to Compositor D by Howard-Hill, to Compositor C by O'Connor. On H3v, as O'Connor noted, "the traditional evidence of doe, goe, here/heere is inconclusive . . . because of the number of long lines requiring justification" ("Compositors D and F," p. 91). As a result, scholars have been forced to rely on ambiguous spelling and spacing evidence. In assigning H3v to Compositor C, O'Connor could depend only on a single spaced terminal comma, a single indented two-line stage direction, three instances of you'll, and the contractions ye , and yu , and yt . These features are all characteristic of Compositor C but are also found in pages set by Compositor D or by Compositor F whose work is often difficult to distinguish from Compositor D's. For example, Compositor D's pages, like Compositor C's, sometimes contain spaced terminal commas: Q6, set from case z, has four. Compositor F set yu (D3) and you'll (B4v, B5, E3) and indented a two-line stage direction (A5). In denying H3v to Compositor D, O'Connor remarked on the absence of that


Page 212
compositor's preferential terminal -ie spellings, although in D's pages the ratio of-ie/-y spellings varies widely even for selected words: on N6v the ratio is 3:1, but on L5v it is 1:3 (O'Connor, 96). As O'Connor correctly acknowledged, "no piece of [the evidence for Compositor C] is conclusive by itself" (p. 91). With the demonstration above that H3v was set from the case at which Compositor D stood during composition of quire H, there is no longer any need to deny him this page. Indeed the assignment of H3v to Compositor D is confirmed by the page's thirty-six indented flow-overs cited by Howard-Hill in support of Hinman's original attribution of this page to Compositor D.[6]

Case identification provides less decisive evidence for lower column H5vb (Err., 1235-54). Whereas everyone has agreed to assign the rest of H5v to Compositor D, O'Connor attributed the last twenty lines of H5vb to Compositor C. These lines contain the spellings wee'll and gon, both strongly characteristic of Compositor C, according to O'Connor. The lines also contain one distinctive type, d24, last observed by Hinman in upper column G6b (see Table I, l. 23.1). Although four other distinctive types from G6b recur in Compositor D's pages of quire H (see Table I, ll. 22 and 23.2), all four come from lower G6b. Thus it is entirely possible that Compositor D distributed only the lower part of column G6b, while Compositor C distributed the upper part. If such is the case, Compositor C might well have set lower column H5vb. On the basis of type-recurrence evidence alone, it is equally possible, however, that Compositor D distributed all of column G6b and set all of page H5v. Only a combination of type-recurrence evidence with spelling evidence—a combination essential to compositor identification—makes it probable that Compositor C set lower column H5vb.

Distribution and Order

It has already been observed that Compositors B, C, and D frequently distributed type from the same wrought-off pages and columns. Such distribution practices are "abnormal in terms of later practice" (Hinman, II, 391) in the Folio Histories and Tragedies. They are not abnormal, however, in the context of other quires in the Comedies, such as quires N and Q, where again three compositors worked from three cases. Type-recurrence evidence in quire H raises the same question about the relative order in which individual pages were set that Hinman addressed in his discussion of quire N: to what extent did different compositors work simultaneously on pages and formes? (II, 417)

No forme in quire H was composed entirely by the same workman. Thus each forme provided an opportunity for simultaneous composition by at least two workmen and sometimes by all three. Probably the two pages of forme H2v:5 were composed simultaneously, since each page was set by a


Page 213

Summary for Quire H[7]

D distributed  -----  and set H3v(CI, free) 
C"  {G2b (BI)
{G5v upper a (FI)
{G1vb (EI) 
" H4a 
D "  {G2a(-)
{G5v lower a(-) 
and set H4b(BI) 
C "  -----  " H3a 
D "  -----  " H3b(DI) 
B "  {G5vb(-)
" H4v(EI) 
C "  {G6 lower a and upper b(AI)
" H2v(AI) 
B "  {G6 upper a[]
" H5(GI) 
D "  G6 lower b(-)  " H2a 
C "  H4va(EI)  "H2b(FI) 
D "  {G6v[]
{H3 lower a(-) 
" most of H5v(CI) 
C "  -----  " H5v lower b 
C "  {H3 middle a(-)
{H3v upper b(CI) 
" H1v
D "  H3va(-)  " H1vb(EI) 
B "  {H3 upper a(-)
{H3v lower b(-) 
" H6(BI) 
C "  H5b(GI)  " H1[] 
B "  H5a(-)
H2v upper b[] 
" H6v(DI) 
different compositor from a different case according to normal Folio practice. The same is true of H1:6v. The two columns of each of pages H4, 3, 2, and 1v were also set from different cases by different compositors and thus may also have been composed simultaneously. Furthermore, page H3 may well have been set by Compositors C and D from cases r and s at the same time Compositor B set the forme-mate page H4v from case t. The same may be true of forme H1v:6, where Compositors C and D shared composition of page H1v and Compositor B set page H6 alone. Yet the two pages of each of formes H3v:4 and H2:5v cannot have been composed at the same time, since, in each of these formes, Compositor D alone set a single page (or all but twenty lines of a single page in the case of H5v) as well as a column in the forme-mate.

There is also strong evidence in quire H of simultaneous composition of different formes. For example, in the Cr column, H1va, is a type from the middle of H3a, and types from upper H3a reappear in Bt's page H6; yet a


Page 214
type from lower H3a is to be found in Ds's share of H5vb. There can be but one adequate explanation: Compositors B and C must have begun distribution in preparing to set the fifth forme, H1v:6, before Compositor D had completed distribution in preparation for setting H5v in the fourth forme. Here is an instance of three compositors at work simultaneously on two successive formes. There are others. In the Bt page, H4v, there first reappear types from G1va and G5vb, but types from lower G5va are found in Ds's H4b and types from both upper G5va and G1vb recur in Cr's H4a. Such evidence suggests that the three compositors were sharing distribution of G1v and 5v while Compositors C and D were preparing to set H4, but Compositor B was preparing to set a page in the next forme, H4v. Centre-rule evidence also indicates simultaneous composition of successive formes. Centre rule DI was freed through distribution of page G1 in preparation for setting forme H2v:5. Yet centre rule DI was used to impose a page in the preceding forme, H3. Preparation for composition of forme H2v:5 therefore seems to have anticipated imposition of the earlier forme, H3:4v. Thus type-recurrence and centre-rule evidence indicates three instances where the compositors worked on different formes at the same time—rare instances discovered later only in quires N-Q.

Conflicting evidence is apparent in connection with the distribution of column G6b. Types from lower G6b first appear in Ds's columns H2a and H5va, but the single distinctive type from upper G6b does not reappear until lower H5vb, the work of Cr. Thus it would seem that Compositor C distributed upper G6b in preparation for completing Ds's work on H5v before Compositor D had begun work on H5v—clearly an impossibility. The simplest explanation is that Compositor C distributed upper G6b at about the same time as he distributed lower G6a, that is, in preparing to set H2v. Buried under types distributed from columns G1b and H4va and, as demonstrated above, the middle of column H3a, the single distinctive type from upper column G6b did not re-appear until Compositor C turned to finish Compositor D's page H5v.

While the production of quire H was abnormal by Folio standards, it need not be regarded as either haphazard or inefficient. This method of production must have proved useful, for it was repeated, in large part, in quire N where again there were the same three compositors at the same three cases available for work on the Folio. Another comparable example is to be found in quire G.

Anomalous Types

It is remarkable that three compositors often distributing type from the same pages and columns at the same time failed to produce in quire H even one anomalous appearance of a distinctive type. Instead it now becomes possible to explain two anomalous appearances of distinctive types recorded by Hinman. Upon discovering two distinctive types from upper H3a in Compositor B's H6 (p 27 and N22 in Table I, l. 10), Hinman concluded that "B's


Page 215
labours for H6 can hardly have been simultaneous with D's for H5v and we cannot well suppose that Compositor D began his distributing of H3 at some point near the middle of H3a, leaving the first twenty-five lines or so standing for B to distribute later. How we can best account for the presence of types p27 and N22 in H6 is hard to say, but they evidently got into B's case in some irregular manner" (II, 392). With the discovery that quire H was set by three compositors working from three cases, often simultaneously on different formes, the appearance of these types need no longer be regarded as anomalous. Compositor B's "labours for H6" probably were simultaneous with D's for H5v, just as, somewhat earlier, Compositor B's work on H4v probably coincided with the work of Compositors C and D on H4.

Quire G

Compositors and Cases

Hinman's analysis of the type-recurrence evidence in quire G failed to yield any case identifications. Perhaps as a result, there has been considerable disagreement among scholars about compositor attributions in quire G. The range of opinion is charted below:

Hinman  Howard-Hill  O'Connor  Werstine 
G1  B/D  B/C  B/C/B 
G1v   D? 
G2  D/C  F/C  F/C 
G4  D?  C/F  C/F  C/F 
G4v   D?  F/C  F/C  F/C 
G6  F/C  F/C  F/C 
G6v   A?  F/C 
As the chart shows, Hinman and O'Connor found three different compositors in quire G, while Howard-Hill discovered four. This multiplicity of compositors again introduces the possibility that at least three cases may have been in use here as they were in work on quire H. The possibility is strengthened by Hinman's significant observation that in quire G, as in quire H, the same pages "were partly distributed into one case and partly into another" (II, 387). Table II, a revised form of Hinman's graph for quire G, presents evidence to confirm the possibility that quire G was set from three cases.[8]


Page 216

Compositor B

Hinman convincingly attributed pages G2v, 3, and 3v to Compositor B on the basis of the high frequency of do, go, heere, and Isa. spellings in these pages (I, 407-408). Confirming Hinman's attributions, Howard-Hill pointed to evidence in these pages of Compositor B's use of unitalicized forms of Duke, Prouost, and Friar, the speech prefix Duke. and spaced medial commas. This broader range of evidence also permitted both Howard-Hill and O'Connor to assign page G1v and column G1a to Compositor B, but not column G1b. Although column G1b contains one do spelling, eight unitalicized forms of Duke, and a preponderance of Duke. speech headings, it also has two goe spellings in short lines, a doe spelling in a long line that perhaps required spelling adjustment for justification, and four Frier spellings, three of them italicized. This conflicting evidence led Howard-Hill to assign the column to Compositor D, while O'Connor preferred Compositor C. Further analysis of the spelling patterns in the column resolves the disagreement. All the spelling evidence on G1b that conflicts with Compositor B's habits occurs in the first twenty-two lines of the column: doe (l. 17j), goe twice (l. 5), Frier italicized (ll. 2-3, 6, 18) and Frier unitalicized (l. 22). The last forty-four lines contain such abundant evidence of Compositor B's hand that they must be assigned to him: the speech prefix Duke. eight times, the unitalicized form of Duke six times, and the spelling do in a short line.

Attribution of lower column G1b to Compositor B is confirmed by type-recurrence evidence provided by Hinman. This evidence shows that all the pages and columns attributed to Compositor B on the grounds of spelling were set from the same case—all, that is, except G3v for which no reliable type-recurrence evidence is available. For example,

i) Lower column F6a furnishes a distinctive type to each of Compositor B's pages G2v and G3 (see Table II, ll. 6 and 12), but to no other page or column, although upper column F6a furnishes a distinctive type to each of columns G4va and G2a (see Table II, ll. 9.1 and 15.1).

ii) Column F6b supplies types to Compositor B's page G2v and to his column G1a, as well as to lower column G1b, just assigned to Compositor B (see Table II, ll. 11 and 18), but to no pages or columns not set by B.

iii) Column G3va also provides types to Compositor B's G1a and lower column G1b and to his page G1v (see Table II, ll. 1 and 2), but provides none to pages and columns set by other compositors. Compositor B must also have distributed column F5va (see Table II, ll. 6 and 7.1), lower F5vb (l. 8), F6va (ll. 11 and 12), middle G3vb (l. 1), middle G3a and all G3b (ll. 6 and 7.1), G2va and lower G2vb (l. 11), for types from these columns also appear only on pages G3, 2v, 1v, column G1a and lower column G1b, with the single anomalous exception noted below.

At this point in our analysis of type recurrences, there is but the slightest evidence that Compositor B stood at the same case for quire G that he used for quire H—that is, case t. Column G3b, which supplied distinctive types to Compositor B's page G1v and column G1a, also supplied a distinctive type


Page 217


Page 218
to his page H6v (see Table I, l. 27). Yet the identification of case t in quire G will be confirmed repeatedly by recurrences found in the later quires I, M and N of types observed last in pages distributed by Compositor B to set quire G.

Compositor C

Hinman attributed to Compositors A and D the pages in quire G not set by Compositor B, but he acknowledged that the "spelling peculiarities in this quire are often far from conclusive" (II, 386). Howard-Hill and O'Connor have since agreed to assign G2b, 4a, 4vb, 5, 6b, and 6v to Compositor C. Evident in these pages and columns are C's characteristic spellings doe and goe, the speech prefixes Isab. and Duke., unitalicized forms of Prouost, Duke and Frier, and spaced medial and terminal commas. Based on this evidence (charted below), these attributions to Compositor C cannot be challenged except in the case of G6v, a part-page which contains the last twenty-two lines of the text of MM and "The names of all the Actors." There is, of course, some evidence of Compositor C's work on this part-page: two Duke. speech prefixes in the first column and, especially, the contraction wee'll in the second last line of the text in the second column. Although it is unlikely that any compositor but Compositor C would have set wee'll, it is equally unlikely that Compositor C set all of G6v, for the text in the first column contains the spelling here (in herein) and the second line of the second column includes an italicized form of Prouost, a word Compositor C never set in italics in quire G. It would appear then that Compositor C merely finished composition of G6v, setting the last few lines of the second column of text and, perhaps, "The names of all the Actors" as well. Part-page G6v belongs to the same forme as page G1, the second column of which, as already noticed, has been assigned to Compositor C by O'Connor. Although, as I have shown, the lower part of the column was Compositor B's, there is still much evidence to support attribution of upper column G1b to Compositor C: the spellings doe, in a long line, and goe, twice in short lines, the speech prefix Duke. in short lines, as well as two unitalicized forms of Duke/s and one of Frier. Here is enough to outweigh the three italicized forms of Frier in upper G1b, forms uncharacteristic of Compositor C who italicized Frier but one other time in quire G. It is difficult to say, however, precisely where Compositor C left off composition of G1b to allow Compositor B to finish the column. The last spelling in the column peculiar to Compositor C is Frier in l. 22, but the first distinctive type indicative of Compositor B's case is not found until l. 40. Thus Compositor C may have set as few as twenty-two lines or as many as thirty-nine lines of column G1b.

Type-recurrence evidence confirms that all the pages and columns now attributed to Compositor C were set from the same case, except for G6b, upper G1b and lower G6vb. With one anomalous exception, the distinctive types in G6b all come from column G4vb, which furnishes distinctive types to no other Folio page or column. The single distinctive type in upper G1b was last seen by Hinman in upper column G3a, which also provides no distinctive


Page 219


Page 220
types to other Folio pages. Lower G6vb contains no distinctive types. Attribution of G6b, upper G1b, and lower G6vb to Compositor C must rest on spelling evidence alone. Type-recurrence evidence demonstrates the integrity of the case from which Compositor C set the remainder of his output in quire G:

i) Lower column F2a supplies distinctive types to Compositor C's page G5 and his columns G4a and G4vb (see Table II, ll. 4.1, 5.1, 9.2, 10, and 14), but to no other pages or columns, although the upper four lines of F2a supply one distinctive type to G4va (see Table II, l. 9.1).

ii) Column F1vb also supplies a distintive type to each of Compositor C's column G4vb and page G5, as well as another to his column G2b (see Table II, ll. 9.2, 14, and 15.2), but to no other columns or pages.

Compositor C also distributed the middle of column F5vb (see Table II, ll. 4.1 and 14), column F6vb (l. 14) and column G4a (l. 4.1), since distinctive types from these columns also recur on page G5 and columns G4vb and G2b. Like Compositor B, Compositor C set his stint on quire G from the same case he used for his pages of quire H: G4a, which supplies distinctive types to Compositor C's column G2b, also supplies a distinctive type to his page H2v and another to his column H3a (see Table I, ll. 11.1 and 18). Additional type-recurrence evidence from quire H has already indicated that Compositor C distributed the upper parts of columns G2vb and G3vb. Distinctive types from these partial columns re-appear in Compositor C's column H3a and his page H1 (see Table I, ll. 10 and 26), although no distinctive type from upper G2vb or upper G3vb recurs in pages of quire G.

Compositor F

The remaining pages and columns of quire G were assigned by Hinman to Compositors A and D, but have since been re-assigned by O'Connor and Howard-Hill to Compositor F alone. The single exception is column G2a, assigned by Hinman and O'Connor to Compositor A(F) and by Howard-Hill to Compositor D. While this disagreement over column G2a is indicative of the difficulty encountered by scholars in distinguishing between Compositors D and F, this problem need not delay us here. It is sufficient to note that the highly consistent spelling pattern in the remaining pages of quire G (charted below) indicates the presence of just one more compositor. Following O'Connor and Howard-Hill, I have called this workman Compositor F. These scholars have shown that, unlike Compositor B, Compositor F of quire G used the spellings doe, goe and here, italicized forms of Duke, Prouost, and Frier, the speech prefixes Duk. and Isab. and non-spaced medial commas. The italicized forms, the speech prefix Duk. and the infrequency of spaced terminal commas on Compositor F's pages are useful in separating him from Compositor C. Only the division of part-page G6v requires additional comment. Recurrence, in column G6va, of two distinctive types last seen in column G5b (distributed into Compositor F's case s), the spelling here (in herein) also in G6va, and the italicized form of Prouost in the second line of


Page 221
column G6vb all demonstrate that Compositor F set all of G6va and began G6vb. Exactly where he left off composition of G6vb so that Compositor C could finish the column is difficult to say. The last distinctive indication of Compositor F's hand occurs in the second of the eleven lines in column G6vb, the first distinctive indication of Compositor C's hand in the second last line. Thus Compositor F may have set as many as twenty lines of text on G6v or as few as thirteen lines.

Type-recurrence evidence shows that all pages and columns attributed to Compositor F were set from a single case, with the exception of column G6a for which no evidence is available. Column G6a takes its three distinctive types from column G4va, a column which provides identifiable types to no other Folio page or column. Attribution of column G6a to Compositor F thus must depend on spelling and spacing evidence alone. Such evidence has proved convincing to Howard-Hill and O'Connor who have agreed to assign the column to Compositor F. The integrity of the case from which Compositor F set the rest of his stint is readily demonstrable:

i) Column F1va supplies a distinctive type to each of Compositor F's column G4va and page G5v (see Table II, ll. 9.1 and 16), but none to any other column or page.

ii) Upper column F6a provides a distinctive type again to Compositor F's column G4va, as well as another to the same compositor's column G2a (see Table II, ll. 9.1 and 15.1), but to no other pages or columns, although lower column F6a does provide distinctive types to Compositor B's pages G2v and G3, as already noted.

iii) Column G4b also provides two distinctive types to Compositor F's column G2a (see Table II, l. 4.2), but none to any other page or column in quire G. However, as observed earlier, column G4b also furnishes identifiable types to two columns set by Compositor D in quire H (see Table I, ll. 8.2 and 23.2). Therefore Compositor F's columns G2a and G4va and his page G5v must have been set from the same case Compositor D used in his work on quire H, that is, case s.

iv) Alone of the columns in quire G, Compositor F's column G4b takes a distinctive type from upper column F5vb (see Table II, l. 5.2), which also supplies a distinctive type (Hinman's h4o) to Compositor D's column H4b set from case s. Thus column G4b too must have been set from case s.

v) Finally, Compositor F's column G6va takes both its identifiable types from column G5b (see Table II, l. 13), a column which provides distinctive types to no other column or page in quire G, but which supplies eight identifiable types to columns and pages in quire H set from case s by Compositor D (see Table I, ll. 2, 3, 7.1, 8.2, 9.3, and 11.2). Therefore column G6va, like Compositor F's G4b, G4va, G2a, and G5v, was also set from case s.

Into case s were also distributed page F1 (see Table II, l. 16), upper column F2a (l. 9.1), and column F2b (ll. 4.2 and 5.2). According to type-recurrence evidence observed in quire H, distribution of lower column G3a, upper column G4vb, and the middle of column G2vb also fell to Compositor


Page 222


Page 223
F at case s, for distinctive types from these three partial columns are to be found again in the pages of quire H set from case s (see Table I, ll. 3 and 22). To account for all the distribution into case s during work on quire G, we must finally look as far ahead as quire I, where, in I3v, there reappears an identifiable type from lower G3vb. When I3v is also shown to have been set from case s, we may then confirm that lower G3vb, too, was distributed into that case as Compositor F worked on quire G.

Distribution and Order

Summary for Quire G

B distributed  -----  and set G3v(EI, free) 
C "  {most of F2a (AI)
{F5v middle b (FI) 
" G4a 
F "  {F2b and upper a (-)
{F5v upper b (-) 
" G4b(FI) 
B "  {F5va and lower b (-)
{F6 lower a and b (BI) 
" G3(AI) 
F "  {F6 upper a (-)
{F1va (DI) 
" G4v
C "  F1vb (-)  " G4vb(BI) 
B "  F6va (CI)
" G2v(CI) 
C "  F6vb (-)  " G5(DI) 
F "  F1 []  " G5v(FI) 
F "  G4b (FI)  " G2a 
C "  G4a (-)  " G2b (BI) 
B "  {G3va and middle b(EI)
{G3b and middle a (AI) 
" G1v (EI) 
F "  {G3 lower a (-)
{G4va and middle b (BI)
{G3v lower b (-) 
" G6a 
C "  {G4v upper and lower b (-)
{G3 upper a (-)
{G3v upper b (-) 
" G6b(AI) 
B "  G2va and lower b (CI)  " most of G1 (DI) 
C "  -----  " G1 upper b 
F "  G5b (DI)  " G6va, upper b 
C "  -----  " G6v lower b and
"The names of all the Actors" 

In quire G, Compositors B, C, and F co-operated in the distribution of the same wrought-off pages and columns, much as the three compositors of quire H did. Such abnormal distribution practices again raise questions about the relative order in which pages and formes were set and the extent to which different compositors worked simultaneously on different columns, pages and formes. Since none of the formes of quire G was the work of a single compositor, there was much opportunity for simultaneous composition sometimes by two workmen, sometimes by three. The two pages of forme G2v:5 may have been composed simultaneously, for each page was set by a different


Page 224
compositor from a different case, according to normal Folio practice. The two compositors who set G2v:5 shared distribution of page F6v and therefore were likely at work at the same time. Simultaneous composition of the pages of G1:6v is also probable for the same reasons, but with the minor variation that for a short time Compositor C aided Compositors B and F in setting both of the pages. However, the pages of forme G2:5v cannot have been set simultaneously. A single compositor, Compositor F, set both G2a and G5v from the same case. The columns of each of pages G4, 4v, 2, and 6 may have been composed at the same time: one column of each page was set by Compositor C from case r, the other by Compositor F from case s. Evidence that Compositors C and F shared distribution of the same wrought-off pages and columns in preparation for setting each of these four pages suggests they did indeed work on the pages simultaneously. Shared distribution of columns and pages also points to simultaneous composition of forme G1v:6 by three compositors.

Simultaneous composition of successive formes is a possibility in quire G, but not so strong a possibility as it was in quire H. Compositors B, C, and F each had a hand in the distribution of column F5vb. However, Compositors C and F were preparing to set page G4 of the first forme by distributing upper and middle F5vb, while Compositor B was readying his case to set page G3 of the second forme by distributing lower F5vb. Thus composition of the first two formes may have proceeded simultaneously, but such a conclusion is not inescapable. Perhaps Compositors C and F merely left lower F5vb for later distribution by Compositor B.

Of the method of production of quires F and G, Howard-Hill wrote: "The picture that this division of work presents of activities in Jaggard's printing-house is most unusual. It is scarcely to be conceived that the Folio would ever have been published had this been the ordinary method of working. . . . The irregular distribution of these quires arose, I suggest, not because Compositor D's practice was irregular (as Hinman thought) but because the division of setting amongst four compositors working from two cases led to unusual practices" (p. 80). The foregoing analysis has indicated that composition of quire G, at least, was divided amongst three different compositors, each at a separate case. Therefore activities in Jaggard's printing-house appear much less chaotic than Howard-Hill asserted. Instead, comparison of the methods used to produce quire G with those used to produce quire H suggests that a routine had been developed for setting by three compositors. In both quires, a single compositor set a whole page of each of the first two formes, while two compositors co-operated to set the forme-mates. Then Compositors B and C each set a full page in the third forme. Three columns of the fourth forme (almost three in quire H) were set by Compositor D in quire H and by Compositor F in quire G; the fourth column was, in each case, the work of Compositor C. The fifth forme in each quire was composed in exactly the same manner as the second forme, with Compositor B setting a full page of each forme and his fellow compositors sharing the forme-mate. Variations in the setting of the sixth formes of quires G and H


Page 225
may well have arisen because page G6v, unlike page H6v, is but a part-page. Thus the three compositors of quire G divided their labours in very nearly the same manner as the three compositors of quire H—not according to the "normal routine" that prevailed in the later two-case setting of the Histories and Tragedies, but in an orderly manner nevertheless.

Anomalous Types

Of anomalous appearances of distinctive types in the pages of quire G, Hinman wrote: "The distribution of most of the pages of quire F having been irregular, it is impossible to say which quire-F types, if any, appear anomalously in quire G. None of the types that are found in any two quire-G pages appears anomalously in the later of those pages" (II, 387). Re-examination of Hinman's type-recurrence evidence (exclusive of the first three formes of quire F distributed during work on quire F) reveals just one anomalous appearance of a distinctive type from quire F in quire G. Distinctive type p39 (see Table II, l. 17) is common to both F5va63 (distributed into Bt) and G6b53 (set by Cr). This single anomalous appearance of a distinctive type, of course, has no significance for the argument that Compositors C and F used different cases to set their portions of quire G, nor is one anomalous type sufficient evidence to question the use of separate cases by Compositors B and C.

Quire I

Compositors and Cases

Hinman identified two compositors at work on quire I, Compositor A, perhaps, or C at case x, and Compositor B at case y. Howard-Hill then assigned the six and a half pages set, according to Hinman, from case x (I3v-5a, 5v-6v) to Compositor C, but Howard-Hill's re-assignment was modified by O'Connor who offered a strong argument that page I3v was Compositor D's. Again Compositor B presents the fewest problems. Hinman demonstrated the integrity of case y, the case used by Compositor B to set pages I1-3 and column I5b (II, 396). It is sufficient here to indicate evidence presented in Hinman's graph for quire I that shows Compositor B must have stood at the same case, case y, when he set his stints on quires G and H. Such evidence consists of the recurrence (in quire-I pages set by By) of distinctive types Compositor B distributed in preparation for setting his portions of quires G and H. The By page I3 contains one distinctive type from each of columns G3b and G3va (see Hinman, II, 394, l. 8), both distributed by Compositor B into case t during his work on quire G. The By pages I3 and I1v and the By column I5b reveal distinctive types last seen in upper column G6a, lower column H3vb, and columns H2vb, H4a, and H4vb (see Hinman, II, 394, ll. 8, 12, 17 and 23), all distributed by Compositor B into case t during his stint on quire H. These seven type recurrences demonstrate the simple equation that case t, Compositor B's case for quires G and H, is the familiar case y


Page 226
at which Compositor B generally worked throughout the setting of the Folio.

The same kind of evidence indicates that case r, used by Compositor C to set his pages of quires G and H, was the familiar case x occupied by Compositor C for quire I. The Cx pages I4v and I6v contain single distinctive types last observed in lower column F2a and column F1vb (see Hinman II, 394, ll. 10 and 21), both distributed by Compositor C into case r during composition of quire G. The Cx pages I4, I4v, I5v, and I6 take distinctive types from columns G1b and H5b and the middle of column H3a (see Hinman, II, 394, ll. 4, 10, 15, 16, and 18), all distributed by Compositor C into case r during the setting of quire H. Re-examination of Hinman's type-recurrence evidence thus not only demonstrates that Compositor C's case r is indeed case x, but also confirms attribution of pages I4, 4v, and 5v-6v to Compositor C by identifying these pages as set from the case long occupied by that workman.

No type recurrences can establish that the disputed page I3v was also set from case x. This page contains ten identifiable types. Three of these were last seen in columns distributed by Compositor D into case s during the setting of quire H: sh26 in H3a55, u22 in H3b3, and sh23 in H3va32. Five more are taken from hitherto undistributed column H2va;[9] yet this column provides no distinctive types to Compositor C's page I4, forme-mate to I3v, and none to any other pages of quire I set by Compositor C. Thus type recurrence provides strong evidence that page I3v was not set from case x, but from case s, the case used by Compositor D for quire H. O'Connor's attribution of page I3v to Compositor D is thus confirmed.

Anomalous Types

Four anomalous appearances of distinctive types must be recorded. Distinctive type y28 is common to I5va5 (set by Cx) and H3b56 (distributed by Ds during composition of quire H); distinctive type h24 is common to I4va21 (set by Cx) and H3a62 (distributed by Ds again during work on quire H). In the wrought-off material, neither type occupied a peripheral position, nor was either one adjacent to a centre rule. Thus neither one was likely vulnerable to a stripping accident.[10] Even so, these two anomalous appearances can scarcely outweigh the seventeen regular appearances of distinctive types already noted in the Ds and Cx pages of quire I. Were the three-case hypothesis for quires G-I seriously in error, we might expect to find a much lower proportion of type recurrences explained by it.

A third apparently anomalous appearance of a distinctive type in fact provides further evidence for the three-case hypothesis. The Ds page I3v contains distinctive type D23, last seen in column G3vb, a column which supplied types to the Cx page H1 and the By page G1v (see Table II, l. 1 and Table I, l. 26). However, the distinctive type in Cx's page H1 is from


Page 227
G3vb6 in the upper part of the column, the distinctive types in By's page G1v are from G3vb23-39 in the middle part of the column, and the distinctive type in Ds's page I3v is from G3vb52 in the lower part of the column. Therefore three compositors in all likelihood shared the distribution of column G3vb, as they shared distribution of other columns of quires F, G, and H.

The fourth anomalous appearance was noted by Hinman: "The type represented in line 19 of the graph is G34. It is common to H5a22 (distributed into case y) and I5vb5 (set by Ax [that is, Cx]). It is immediately adjacent to the left-hand rule in page H5, and hence in a peripheral position there" (II, 397).

Quires K-Q

Compositors and Cases

In these seven quires, Hinman discovered four different compositors occupying three different cases. Compositor B alone used case y; Compositor C used case x, although, according to Hinman, so did Compositor A in quire K; and Compositor D alone used case z, except during composition of quires O and P, where Hinman found Compositor A standing at case z. Subsequent research has established a higher correlation between individual compositors and cases. Howard-Hill re-assigned Compositor A's quire-K pages to Compositor C so that Compositor C is now recognized as the only workman to use case x during composition of these seven quires, with the single exception of column K5b, set by Compositor B from case x, according to Howard-Hill. Howard-Hill also re-assigned Compositor A's quire-O pages, set from case z, to Compositor D, thereby strengthening the association between Compositor D and case z. At the same time, however, Howard-Hill weakened this association when he attributed the z-case pages of quire P to Compositor F and gave Compositor D column P1va, set from case x. Most of these anomalies in Howard-Hill's analysis were resolved by O'Connor, who assigned the disputed z-case pages of quire P to Compositor D and demonstrated that only lower column P1va was Compositor D's work and was set from case z. After the cumulative efforts of Hinman, Howard-Hill, and O'Connor, we now know that, with the exception of a single column (K5b), Compositor B alone used case y, Compositor C case x, and Compositor D case z for quires K-Q.[11]


Page 228
There is now general agreement that Compositor A was not represented in these quires.

In these seven quires there recur a significant number of distinctive types last seen in quires F-H. These recurrences (charted below) confirm that the same pattern of three distinct cases of type associated with three individual compositors is also evident in the composition of quires G, H and I.[12] Twelve

Hinman's distinctive type  last seen  then distributed into case  location of recurrence in quires K-Q  set from case 
1. y24  H5va22  Cx  K3b59  Cx 
2. h45  H2a65  Cx  K4va32  Cx 
3. o26  H2a57  Cx  K4va33  Cx 
4. S24  H5b8  Cx  K4vb48  Cx 
5. p25  H4va8  Cx  L4b49  Cx 
6. B28  H5b6  Cx  L3b54  Cx 
7. W23  H2b46  Cx  L2vb29  Cx 
8. st24  H3vb41  Cx  L1b62  Cx 
9. W30  H4b6  Cx  M3vb57  Cx 
10. N21  H4va30  Cx  N3b49  Cx 
11. P25  G2vb21  Cx  O2a8  Cx 
12. w34  G2a57  Ds  L1b65  Cx 
13. W45  H4vb23  By  K4a16  By 
14. W44  H3vb65  By  M5a46  By 
15. H33  G2vb55  By  M4va55  By 
16. F25  G5vb21  By  N2vb34  By 
17. P21  G2vb27  Ds  K5vb58  Dz 
18. ss24  H2va2  Ds  L4vb19  Dz 
19. )21  F1va58  Fs  L4va16  Dz 
20. B26  G5b7  Fs  L5vb59  Dz 
21. Y21  H2vb36  By  K5vb28  Dz 
22. st23  H2a59  Cx  K2vb15  Dz 
of these distinctive types (no.'s 1-12) recur in the pages of quires K-Q set by Compositor C from case x. Since all but one of these (no. 12) were last seen in material already identified as distributed by Compositor C, these type recurrences support the conclusion that Compositor C used case x for his stints on quires G and H and forme I3v:4. Compositor B's use of case y during the setting of these earlier quires is also confirmed again by the recurrence of the next four charted types (no.'s 13-16) in his pages of quires K-Q, all four from wrought-off columns he distributed to set pages of quires G and H. The remaining six types charted (no.'s 17-22) re-appear in pages set from case z by Compositor D. Hinman was forced to regard all six recurrences as anomalous because he argued that Compositor D's case z was not set up until composition of quire K had begun (II, 400). Yet four of these distinctive types were last seen in pages, columns, and partial columns already


Page 229
identified as distributed by Compositors D or F into case s. The reappearance of these four types in quire-K and quire-L pages set from case z establishes that case s was the case Hinman designated case z. Therefore case z was not set up de novo during work on quire K, but was instead the case from which Compositors D and F set pages of quires G, H and I.

Anomalous Types

Three anomalous appearances are listed among the types charted (no.'s 12, 21 and 22). I can discover no explanation for the anomalous appearance of numbers 12 and 22, but the third, no. 21, can be readily explained. Distinctive type Y21 (no. 21) is found in the lower part of column H2vb at l. 36 and found again in the Dz page K5v. Since column H2vb supplied three distinctive types to By pages, distribution of column H2vb is assigned to Compositor B in the chart. The three distinctive types mentioned are h44, common to H2vb32 and H6va44; e27, common to H2vb29 and I5b11; and o38, common to H2vb66 and I2b11. Only the last of these three distinctive types is evidence that Compositor B distributed the lower part of column H2vb, but it is not good evidence because it was located in the last line of the wrought-off column H2vb where it would have been especially vulnerable to a stripping accident. More probably, therefore, Compositor B distributed only the upper part of column H2vb, including ll. 29 and 32, while Compositor D distributed the lower part of the column, including l. 36 where Y21 was last seen before it recurred in the Dz page K5v. The recurrence of Y21 is probably not anomalous but instead becomes further evidence of Compositor D's use of case z during the setting of quire H. In summary, the three-case hypothesis for quires G-I explains all but two of the distinctive type recurrences charted (no.'s 12 and 22), whereas Hinman's two-case theory for quires G-I failed to explain six of these recurrences (no.'s 17-22).

Quires A-F

Thus far this article has concentrated on quires G-I; evidence has been drawn from quires K-Q only to confirm the three-case hypothesis as it applies to quires G-I. There is no need to search beyond quire Q for evidence of the three-case pattern, for during composition of quire R cases x and z were liquidated (Hinman, II, 441). However, study of the composition of quires G-I has implications for an understanding of work on the first six quires of the Folio, quires A-F. Since case z was not set up de novo during the setting of quire K, as Hinman believed, but was already in use for quire G, there remains a question about when case z was set up. In dealing with this question, we cannot rely as much on the kind of type-recurrence evidence used heretofore in case and compositor identification. Almost midway through the first six quires of the Comedies section—just after composition of page D4v, to be precise—type recurrence no longer yields patterns useful for case identification. Such patterns do not emerge again until quire G.


Page 230

In spite of this breakdown in type-recurrence evidence for quires D through F, one fact is evident: case z was not set up de novo for composition of quire G, but was already in use before work on this quire began. The pages and columns of quires G through Q set from case z contain no fewer than nine distinctive types last seen in pages of quires C, E, and F which were distributed before Jaggard's compositors turned to quire G.[13] If it is granted that so many appearances of distinctive types cannot all be dismissed as anomalous, we must search for evidence of the use of case z in Folio quires set before quire G. The search is frustrated to some extent by the hiatus in intelligible type-recurrence evidence that extends from quire D to quire F. Since study of the type recurrences in these quires has yielded no case identifications, there must remain a theoretical possibility that case z may have been set up de novo at any point during composition of these Folio quires or during composition of non-Folio material that was being printed concurrently with these Folio quires—probably Wilson's Christian Dictionary, according to type-batter evidence Hinman discovered in both works. Although such uncertainty cannot be entirely dismissed, there are, however, intelligible patterns of type-recurrence evidence in quires A-D4v that bear on the question about when case z was set up.

According to Hinman, quires A-D4v (in printing order) were set from only two cases. Compositor B set pages of the first four formes of quire A from case y and then, according to Hinman, yielded the case to Compositor C who worked on pages of the last two formes of quire A, as well as pages of quires B, C, and D. Both compositors were aided by Compositor A (now called F) who stood at a second case designated case x by Hinman.[14] Hinman's argument for this interpretation of the available evidence can be divided into two stages, the first successful, the second unsuccessful. In the first stage he used type-recurrence evidence to differentiate the case used by Compositor A(F) in quires A-D4v from the single other case Hinman thought was used first by Compositor B and then by Compositor C. Hinman found a pattern of 116 distinctive-type appearances in quire A, none of which is rendered anomalous by his two-case hypothesis for the quire (Hinman, II, 351). Quire B offers another 162 appearances, only two of which Hinman was forced to dismiss as anomalous. The first distinctive type to recur anomalously is u23, common to A4b48 (distributed by By) and to B5vb61 (set by A[F]); the second is distinctive type B30, common to A1va10 (distributed by By) and to B4vb42 (set by A[F]), but this distinctive type appears, in the earlier of the two pages, immediately adjacent to the centre rule where it would be vulnerable to a stripping accident (II, 358). Since these anomalous appearances scarcely constitute


Page 231
grounds for challenging Hinman's two-case hypothesis for quire B and since Hinman drew equally impressive evidence from quire C (II, 360- 364) and the first two formes of quire D (II, 366-371), there can be little doubt that at least two cases, Compositor A(F)'s case x and Compositor B and Compositor C's case y, were employed to set Folio quires A-D4v.

Hinman was, in my view, much less successful in the second stage of his argument when he attempted to demonstrate that only two cases, and not three, were used in quires A-D4v, that is, when he argued that the case used by Compositor C to finish quire A and to compose pages of quires B-D4v was the same case used by Compositor B to set pages of the first four formes of quire A. The only support for this stage of Hinman's argument can be found in the recurrence of just two distinctive types common to quire-A pages distributed by Compositor B into case y and to pages set by Compositor C in quires A and B.[15] Distinctive type h30 is common to A6va53 (distributed by By) and to B2va33 (set by C); distinctive type W22 is common to A4b6 (distributed by By) and to A2b45 (set by C). Since W22 is, in the earlier of the two pages, immediately adjacent to the centre rule where it would be especially vulnerable to a stripping accident, Hinman's identification of Compositor C's case for quires A-D4v as case y depends entirely on the recurrence of a single distinctive type distributed from a non-peripheral position. It will be recalled that Hinman dismissed as anomalous the recurrence, in a quire-B page set by Compositor A(F), of a single distinctive type (u23) which had been distributed from a non-peripheral position into case y by Compositor B during composition of quire A. By analogy, it would also be necessary to dismiss as anomalous the recurrence, in a quire-B page set by Compositor C, of this single distinctive type (h30), which had been distributed in exactly the same way by Compositor B.[16] Dismissing the recurrence of h30 as anomalous, in turn, entails rejecting Hinman's argument that Compositor C succeeded Compositor B at case y. It would then follow that Compositor C stood at a third case to set his stints on quires A-D4v. Thus from the beginning of Folio production with quire A, three cases—not two—were used by the Folio compositors. Since type-recurrence evidence breaks down in quire D and does not again become useful until quire G, it is impossible to be certain whether the three cases used to set quire A (two of which, as Hinman was able to demonstrate in the first stage of his argument, remained in use for quires B-D4v) were the same three cases used for quires G-Q. I thus merely follow Hinman's example in using the same case designators for quires A-D4v as for quires G-Q: Compositor B at case y worked on seven of the first


Page 232
eight pages of quire A; Compositor C at case x (now differentiated from case y) set two of the last four quire-A pages and pages of quires B-D4v; and Compositor A(F) at case z (which Hinman called case x) aided both the other compositors in succession.

Thus far differentiation of Compositor C's case, now called case x, from Compositor B's case y has proceeded in entirely negative terms; Hinman's failure to demonstrate that Compositor C used case y indicates that Compositor C used a separate case. Yet it is also possible to differentiate the two cases, x and y, in positive terms on the basis of type-recurrence evidence. Setting aside the two anomalous type recurrences (h30 and W22) already discussed, we can observe that pages A1, 6v, 1v, 6, and 4 supply some thirty-six distinctive types to quire-A pages set by Compositor B (see Hinman, II, 344, ll. 1-5, 7-11, 15, 16), but none to any page set by Compositor C; pages A3, 3v, and 2v and column A2b supply some thirty-seven distinctive types to pages of quires A, B and C set by Compositor C (see Hinman, II, 344, ll. 1, 7, 11, 13, 17; 352, ll. 7, 8, 17, 20-22; 360, ll. 1, 3, 5, 29, 31; and I,480), but none to quire-A pages set by Compositor B. Therefore seventy-three type recurrences demonstrate that while Compositor B worked from case y for quire A, Compositor C worked from his own case, case x.

Type-recurrence evidence also indicates when, after quire A, case y next supplied types for Folio composition. In the five pages Compositor B distributed during his labours in quire A, there are forty-four appearances of distinctive types. As already noted, in thirty-six of these possible forty-four instances, the distinctive types distributed into case y re-appear in the y-case pages of quire A set by Compositor B; four more identifiable types, as we have seen, recur anomalously in pages of quires A and B set from the two other cases, x and z, by Compositors C and A(F); and one (A31) is never found again in the Folio. The three remaining distinctive types do recur in the Folio, but are not to be seen until quire F, two in column F4b, one in column F5a (see Hinman, II, 376, ll. 6 and 19).[17] The extended delay in the recurrence of these three distinctive types deserves emphasis. In contrast, recurrences of the seventy-two distinctive types distributed into cases x and z from wrought-off pages of quire A by Compositors C and A(F) can all be located in quires A-D, with the exception of four distinctive types which do not recur at all


Page 233
in the Folio. (There are twenty-four recurrences in quire A, thirty-three in quire B, nine in quire C, and just two in quire D.) There can, then, be just one explanation for the long delay in the recurrence of the three distinctive types mentioned: case y, unlike cases x and z, was not used again for Folio composition until long after Compositor B had completed his work on quire A. Therefore Compositor C could not have used case y for his work in quires A-D4v, but instead must have used a separate case, case x. At the same time, Compositor A(F) used a third case, case z, to set his stints for quires A-D4v. The third case, case z, was therefore set up de novo at no point in Folio production, but was in use from the beginning.

To summarize: from the outset Jaggard's printing of the First Folio seems to have demanded the labours of three compositors, each standing at a separate case. Compositor B occupied case y, Compositor C case x, and Compositor F case z. Yet all three men never worked simultaneously on the Folio during composition of the first five quires (A-E). Sometimes a single compositor worked alone or sometimes a pair of compositors shared a forme of the Folio, probably because, as Hinman observed, Jaggard's shop still had composition of Wilson's Christian Dictionary in hand. This pattern of composition remained unchanged even after Compositor D had succeeded Compositor F with the beginning of work on quire F (Howard-Hill, pp. 78-82; O'Connor, pp. 101-110). By the middle of Folio quire F, however, Wilson's volume apparently no longer drew upon the time of Jaggard's compositors or the stock of types in their cases, and three workmen—Compositors B, C, and D—began to devote their time exclusively to the Folio. This new pattern of composition by three type-setters was to continue until work on quire I began, even though Compositor F took over from Compositor D again for quire G, only to be displaced again immediately after this quire by Compositor D who then occupied case z until it was liquidated for composition of quire R. When three compositors—B, C, and F—distributed types from the last half of quire F in order to set G, their work produced clear patterns of type recurrence so that their cases can be identified with some confidence for the first time since early in the composition of Folio quire D. These type-recurrence patterns indicate that complex distribution practices persisted throughout composition of quires G and H, with different compositors often co-operating to distribute types from the same wrought-off pages and columns. Yet there is nothing to indicate that the availability of three compositors and cases sowed confusion in the production of the Folio. Instead, the compositors followed essentially the same routine in dividing the labour of composition for both quires G and H. A similar routine can be observed in the method used to compose parts of quires N and Q, when again different compositors sometimes worked on successive formes at the same time. Not until quire R do we begin to find evidence of the two-compositor, two-case pattern of composition that was sustained throughout most of the last two sections of the Folio, the Histories and the Tragedies.


Page 234

This purely bibliographical investigation has at least two major implications for editors and textual critics of the early Folio Comedies. The first deals with cast-off copy for quires G and H. Ordinarily, as Hinman observed, when a quire was set in normal Folio printing order, there was need to cast off copy only for the first half of the quire, since the pages of the second half were set in their reading sequence: 4, 4v, 5, 5v, 6, 6v (II, 505-506). However, simultaneous composition of successive formes of a quire—possible for quire G, probable for quire H—would have necessitated casting off copy for both halves of the quire so that compositors could work concurrently from separate blocks of copy to set pages of the second half of the quire. Thus editors must be wary of compositorial attempts to stretch or squeeze the texts of MM and Err. into the space available. One such attempt is evident in the last half of quire H at the bottom of column H5vb, set by Compositor C, and the top of H6a, set simultaneously by Compositor B, where verse is compressed into prose. Second, identification of cases x, y, and z in quires G-I confirms attribution of pages in these quires to Compositors B, C, D, and F. Editors of the Folio Comedies may then have greater confidence in using the available studies of compositorial accuracy in arriving at judgments about when and how to emend.



John S. O'Connor, "A Qualitative Analysis of Compositors C and D in the Shakespeare First Folio," Studies in Bibliography, 30 (1977), 57-74.


Throughout this article "case" will refer to "all of the components of a single complex of cases" (C.J.K. Hinman, The Printing and Proof-Reading of the First Folio of Shakespeare [1963], I, 108.)


"Compositors C and D of the Shakespeare First Folio," PBSA, 65 (1971), 41-52; "Compositors E and F of the Shakespeare First Folio," PBSA, 66 (1972), 369-406.


T.H. Howard-Hill, "The Compositors of Shakespeare's Folio Comedies," SB, 26 (1973), 61-106; John S. O'Connor, "Compositors D and F of the Shakespeare First Folio," SB, 28 (1975), 81-117.


Table I retains the line numbering of Hinman's graph, while expanding it where necessary (i.e., l. 1 becomes ll. 1.1 and 1.2), so that the types referred to in the graph may be recovered from Hinman's tables (I, 425-490).


An "indented flow-over" is produced when a compositor carries over a line too long for his measure to the next line and indents it two or three ems from the left margin (O'Connor, p. 89).


Centre-rule designators (i.e., BI) are included from Hinman's summary for quire H on II, 389.


Excluded from Hinman's graph in its revised form are recurrences in quire G of distinctive types from the first three formes of quire F, which were distributed during composition of the last half of the quire. Hinman expressed no confidence in the evidence of these recurrences when he wrote of quire F: "Probably because irregular distribution practices were followed, we cannot say which types do and which do not appear anomalously in this quire, or at any rate in the last two formes of it" (II, 383). After examining type-recurrence patterns in quire F, I can only concur with Hinman's judgment, and thus I have limited attention to distribution carried on during composition of quire G.


The ninth distinctive type in page I3v is G24 common to I3vb36 and E3va10, and the tenth is discussed below under "Anomalous Types."


On stripping accidents, see Hinman, I, 128-130.


The exceptional column K5b may not be exceptional at all. A paucity of distinctive types in this column forbids conclusive identification of the case used to set it. There are only three distinctive types in column K5b, two from page I2v which provided identifiable types to no other Folio page, and one from I2a37. Since the only other distinctive type from column I2a is common to I2a43 and K5a51 (demonstrably set from case x by Compositor C), we might conclude that Compositor B set K5b from the same case, case x. Yet so often in the Comedies did compositors share distribution of the same columns when they also shared composition of a single page that it is equally possible that Compositor B distributed type from the upper part of column I2a, including I2a37, into case y, while Compositor C distributed lower I2a, including I2a43, into case x. If so, then Compositor B set column K5b from case y, and the association between individual compositors and cases is absolute for quires K-Q.


We are concerned here, of course, only with distinctive types distributed from pages F2-H5v in the order of printing, that is, the material distributed to prepare for composition of quires G and H and forme I3v:4.


The nine types are: W29 C4vb37 L2a59; ss24 E1va18 G5va26; M22 E2a17 G6a15; G24 E3va10 I3vb36; h37 F4vb49 G2a57; w34 F3a20 G2a57; f25 F2vb49 G4b57; ffi21 F3vb46 G5vb43; and y23 F3vb48 G5vb23.


As Hinman acknowledged, his use of the case designators x and y was arbitrary, since the hiatus in intelligible type-recurrence patterns already noted in quires D-F prevents identification of the cases used for quires A-D as the same cases used for quires G-Q (II, 348).


Hinman's footnote (II, 345, n.1) indicates his search for more reliable evidence and his inability to find it: "Forme A2v:5 seems to have no ordinary letterpress in common with either A1v:6 or A3v:4. . . . None, that is, of which we can be absolutely certain— though it may be that the same damaged type is indeed represented both by the defective 'n' in 'vnworthinesse' in A6b5 and by the very similar 'n' in the word 'In' at the beginning of A2vb65."


As Hinman wrote elsewhere, "Of course the testimony of a single already distributed type . . . is by no means conclusive" (II, 28).


Although columns F4b and F5a were not the work of Compositor B, but instead that of Compositor C, these three type recurrences can hardly be used as evidence that Compositor C's case for quires A-F was indeed case y, not case x. It is not likely that three types distributed during composition of quire A would have remained unused in the case from which Compositor C set pages of three quires (B-D), only to re-emerge together in two columns of a fourth quire (F). The concentrated recurrence of the three distinctive types in two quire-F columns suggests, instead, that the three had just lately been supplied to whatever as yet unidentified case Compositor C used for quire F. As Hinman was able to demonstrate, the compositor who set F4 distributed non-Folio material (likely Wilson's Christian Dictionary) just before he set the page, and thus the three distinctive types probably recurred in the Folio at this point because they were distributed into Compositor C's case from wrought-off pages of non-Folio material set from a different case (Hinman, II, 381-382).