University of Virginia Library

The Printing of John Webster's Plays (III): The Duchess of Malfi
by
John Russell Brown

The examples that have been presented in this paper have demonstrated the utility of this particular sort of investigation in shedding light on mechanical and on compositorial practices that offer problems to the bibliographer'—so wrote George Walton Williams in his 'Setting by Formes in Quarto Printing' in Studies in Bibliography, XI (1958). His thesis was forcefully presented, and other bibliographers who study quarto editions of dramatic texts of the late sixteenth and early seventeenth centuries have followed his lead in observing type-shortages in relation to compositors' habits. And now, in proper sequence, a responsible editor of such texts must discover how his task can be aided by this knowledge. He has a new duty: he must determine the order in which the pages of his copy-text were set and then enquire whether this order has influenced each individual compositor's presentation of his copy. In this way he may be able to discover means for presenting a text closer to his author's intentions or the printer's copy.

He may be able to do this. So far few have tried, and, until those few have shown their gains clearly, other editors with pressing commitments may be unable to decide whether they should spare the time for this further research. It is in this situation that some processes used for editing John Webster's Duchess of Malfi in the light of knowledge concerning its original printing by Nicholas Okes in 1623 are set out in considerable detail in the following note.

From papers which have been published already many relevant facts are known about the kind of copy used by Okes for this play, about the two compositors who set it, and about other quarto plays from Okes' shop within a few years of 1623.[1] We know that The

Duchess was set from a scribal transcript, probably in the hand of Ralph Crane, and that the work was divided between two compositors in this manner:    

Compositor	A	A1-4v	B3-C2v	D3-E2v	F3-G2v	H3-I2v	K3-L2v	M3-N2v
Compositor	B	B1-2v	C3-D2v	E3-F2v	G3-H2v	I3-K2v	L3-M2v	N3-4.

We also know that some sheets of the second quarto of The Insatiate Countess, printed by Nicholas Okes in 1616, were divided between two compositors in the same way, and that those sheets were set by formes, the inner first, with the compositors working simultaneously; two of the four sheets were set in this order:    

Compositor	X	1v	2	1	2v
Compositor	Y	3v	4	3	4v

If the copy for The Duchess of Malfi was written as clearly and regularly as other manuscripts known to have been the work of Ralph Crane, this play might have been 'cast off' for setting by formes almost as easily as the printed copy for The Insatiate Countess. The same procedure of composition may have been used.

New evidence suggests that this guess is correct. In setting The Duchess there was a shortage of roman, capital T, which was made good by the use of italic; a shortage of italic, capital B, made good by roman; and a shortage of roman, capital W, made good by using two Vs; and these shortages occur in ways that suggest, though not with equal clarity, the same order of composition. The shortage of T is most useful. Compositor A first encountered the lack while setting his half of Sheet H. The following table shows the number of Ts and Ts on each page in the order in which they were used; X stands for a T used incorrectly for a T, and the superscript numeral for the number of times the particular letter occurs consecutively.

             

	1v or 3v	2 or 4	1 or 3	2v or 4v
H3-4v	T9	T6,X2	T7,X3	T,X3
I1-2v	T3	T10	T5	T4,X3
K3-4v	T11	T8	T2,X,T2,X, T,X,T	T3,X9
L1-2v	T9	T5,X	X,T,X,T7	X,T,X5
M3-4v	T6	T5	T2,X8	T7
N1-2v	T11	T,X5	T3,X3	T6,X,T

The shortages of T in H3-4v, K3-4v L1-2v and N1-2v all suggest that the inner-forme pages were set first; and those on L1-2v, in particular, that page 1 (or 3) was set before 2v (or 4v). It appears that each of these half-sheets was started with some supply of Ts, and that the shortage once felt usually continued with only slight alleviation until the end of the half-sheet. M4v may have been set before M3, but occurrences of Ts in N1-2v suggest that when a shortage was felt on N2 it continued with only slight relief on N1 and a considerable new supply was not obtained until the commencement of N2v; a new supply after M3 and before M4v would also account for M3-4v.

Shortages of B (chiefly used in prefixes for Bosola's speeches) again suggest that the inner-forme pages were set first. The incorrect Bs first appear on A's pages of Sheet E, and again on his halves of G and K; on B's pages they occur in sheets H,K,L and M. The following table shows incorrect Bs as X; it is arranged as before.

               

	1v or 3v	2 or 4	1 or 3	2v or 4v
E1-2v	B 9,X	X5	B 2	--
G1-2v	B 4	X,B	--	B 2
H1-2v	B,X	B	B 2	B
K1-2v	B,X3,B,X3	X4	B,X,B 5	B 3
K3-4v	B 5	B 6	B 8	B,X
L3-4v	B,X	X3,B 3	B	X7
M1-2v	--	B3	X,B 5	B 4,X,B 4

Here the evidence is clearest on K1-2v, where a shortage was felt almost at once but was relieved after beginning 1; on K3-4v, where the shortage is apparent only on the last page (Ts were lacking for the last two pages); and on L3-4v, where the shortage began to be felt on the first page and lasted half-way through the second where it was relieved with four Bs which were sufficient through the third page but not for the last.

The use of VV for W gives only the slightest evidence. But for C3-4v and E3-4v it again suggests that setting started with a supply and that the inner-forme pages came first, and for F1-2v it suggests that 1 preceded 2v, and that some replenishment was made before completing F1. X stands for VV in the table.

         

	1v or 3v	2 or 4	1 or 3	2v or 4v
C3-4v	W6	W4	X3,W3,X	W,X,W
E3-4v	W3	W4	W3,X,W	W5,X
F1-2v	W	W4	X3,W	W2
M1-2v	W5	W,X	W5	W3,X

These tables provide no evidence about some half-sheets in the book. But the completely regular alternation of the compositors' 'takes' encourages the belief that a regular sequence of composition was followed: that is, the copy cast off in advance and inner forme pages of each half-sheet set first. And this is further supported by editorial investigations.

A recurring problem for an editor of Webster is how to arrange the text as verse. In Okes' quarto of The Duchess numerous lines are obviously bungled, but more are possibly so. For Webster's versification is not always the kind that can be verified by a simple count: it ranges from the easily formal, to the slack, strained, subtle, or hesitant. Moreover he used incomplete verse-lines to point dramatic pauses between speeches and within them, and introduced very short passages of prose in scenes otherwise in verse.[2] Editors often differ about the correct presentation, and their disagreement can be of moment. Literary critics are particularly apt to find importance or beauty in Webster's pauses or metrical subtleties. Una Ellis-Fermor may speak for many:

She has noted lines which are 'impetuous, half-hushed now in horror' or 'slow, almost inaudible beginnings', and a 'rhythm' which 'gives no choice' of interpretation of character and meaning. George Rylands, Clifford Leech, Muriel Bradbrook are among other critics who have valued a subtle 'naturalism', a special emphasis, a dramatic tension, many fine and strong effects in the metrical arrangements of Webster's lines. An editor of The Duchess of Malfi must be especially scrupulous in these matters, and here he may find that enquiries into the order of composition of the quarto can help him.

If an editor believes that The Duchess was originally set in the order in which it is read, he will probably judge that its obvious mislineation occurs fortuitously. But once he supposes that the copy was cast off before composition and the two inner-forme pages of each half-sheet set first, he will suspect that mislineation is often due to inaccurate casting off and the compositors' attempts to keep in step. In each sheet Signatures 1v-2 and 3v-4 would be set consecutively, so that the first page of each of these pairs would be unlikely to have this kind of mislineation; any error could be carried over to its fellow, with the chance that it might be cancelled out by the opposite kind of error in casting off.

With added notation on the copy, it might be possible to carry over accumulated error from Signature 2 to 2v and 4 to 4v, and, if the two compositors worked perfectly in time with each other and cast off sheet by sheet as they worked through the copy, then some accumulated error from 4v might be accommodated on the following sheet. But miscalculation on Signatures 1, 2v and 3 would always have to be remedied on those pages; and, once one compositor lagged behind, miscalculation on Signature 4v would have to be dealt with on that page. With such considerations in mind, the places where the most responsible editors of The Duchess, F. L. Lucas (1927) and A. K. McIlwraith (1953), saw fit to change the lining of the quarto in ways that altered the number of lines of type to be set are shown in a table. An underlined numeral indicates how often, on the page of the quarto indicated, both editors reduced the number of lines of type—that is, when we may suppose that the compositor added a line to the number in his copy; ordinary type numerals show how often we may suppose that he saved a line. Numerals in brackets show, in the same manner, the occasions when only one of the two editors changed the lineation of the quarto.                          

	1v	2	1	2v	3v	4	3	4v
B	--	--	--	2	--	--	--	--
C	--	--	--	1	--	--	--	--
D	--	--	--	--	--	--	--	--
E	--	2	--	1	--	--	--	1
F	--	--	--	--	(1)	1(1)	--	--
G	--	--	--	--	--	--	1	(1)
H	--	--	(4,1)	--	--	1(1)	1	--
I	--	--	--	1	--	--	--	--
K	--	--	--	--	--	--	--	1
L	--	--	--	1	--	--	--	1(2)
M	(2)	1	1	1	--	--	(1)	--
N	--	2	--	(1,1)	--	--	1

This table at once confirms the accuracy of the hypothesis concerning the order of composition and helps the editor when he is uncertain how to present Webster's verse-lines.

The following facts confirm the order of composition.

21 times both editors modified the number of lines of type.

• 7 are on Signature 2v
• 7 . . . . Signatures 2 or 4
• 4 . . . . . . 1 or 3
• 3 . . . . Signature 4v
• 0 . . . . Signatures 1v or 3v

3 pages have 2 obvious mislineations; on each page both imply that the compositor added lines of type, or both that he saved.

Compositor B was responsible for fewer obvious mislineations than A (9 to 12), but he more frequently varied the number of lines on a page: E3v, I3v, K1 and K1v have 38 instead of 37; F2v, G3v and N3v have 36. It is noteworthy that none of these pages contains certain mislineation. Compositor A only varied from the standard on four occasions, setting 38 lines on D4v and G1v (where he made no mislineation), and on H3 (where he did save one line), and setting only 36 on N2v (where he made no certain mislineation). These variations demonstrate the results of inaccurate casting-off where the compositors were less careful about the precise balance of the printed pages.

The first way that the editor is helped by this new knowledge of the order of composition is in assessing the authority of the printer's copy: he may be virtually certain that this kind of mislineation is due to the printing of the play, not to its transcription by the scribe. Understanding where and how mislineation is likely to occur, he will be more reluctant to alter the line-arrangement of his copy-text elsewhere. Secondly the editor is helped in particular readings where it had previously been difficult to decide on the correct line-arrangement. Both the changes suggested, without support, for N2v can hardly be correct, for they cancel each other out in altering the number of lines of type; the editor should choose one or the other, or neither. And since A set a line short on that page against his usual custom, the suggested mislineation which would imply that he had added a line to those in his copy is the more likely to be correct. The change on H1 which implies that Compositor B added a line is unlikely to be right, if any of the other suggested changes are: and before accepting the large number of these, the editor would wish to find some special reason for the unprecedented inaccuracy. He will view the unconfirmed changes on F4, H4 and L4v particularly favourably, because they would alter the number of lines of type in accordance with the necessary re-arrangements.

Occasionally the editor will be able to propose a change where neither of his predecessors altered their copy-text. An example of this is on B4v, where the quarto has:

as one line, and 'must I cut?' is slightly out of alignment with the rest of the type and outside the width of the composing stick used for this page (marked clearly by the turnover of its second line). The typesetting suggests a special effort to accommodate a long line, and it occurs only three lines from the foot of the last page of the outer forme in a half-sheet. The editor would have bibliographical justification for re-arranging and for removing two elisions, printing the passage: The emphasis is different and a pause has been added; in performance, the two characters will seem the more deliberately to take stock of each other. Or again, on I2v (the page most prone to mislineation in any sheet, and here one on which both earlier editors have re-arranged the verse-lining once) another change may be introduced. It involves the same kind of alteration to the number of lines of type as the change generally accepted. The quarto has: as one line, using a turnover for the last word. Lucas and McIlwraith both retain this, the latter showing it in context as: But the editor who knows the likelihood of the compositor trying to save a line on this, the last page of the outer forme in the first half of a sheet, will be encouraged to print: In performance, there will be an impressive pause on 'chaos'; it will seem the end of the duchess' curse. And perhaps 'Remember' will sound more deliberate in its new position: certainly, the colon after Bosola's 'still', followed by the duchess completing his half-line, and the greater opportunity for deliberation in the metrical arrangement of the reply, will give a dramatically exciting impression of the duchess' composure even in the middle of her curses; by such changes of rhythm and 'transitions' of mood, the impression of great resources in a character's mind and spirit is often given in the theatre. By common consent we know that Webster delighted to use these devices, and so it behoves an editor of his plays to study every clue that will help to restore the original line-arrangement. Here a study of composition is well repaid.

Another reason for an editor undertaking this research is the likelihood, attested in many kinds of investigation, that the solution of one bibliographical problem will set others in a new and clearer perspective. The editing of The Duchess of Malfi does benefit in this way.

One of its problems is the authority of its stage-directions. There are comparatively few of them, as in some other plays with massed entries at the head of each scene. But one, 'A Coffin, Cords, and a Bell' (K1v; Lucas, IV.ii.164-166) might derive from a prompt-book, while two others, 'Ferdinand giues her a ponyard' (F4; III.ii.79-80) and 'Enter Bosola with a Guard' (H4; III.v.110) were added during presscorrection, after some sheets had been printed, and might have been supplied by the author himself together with the corrections to dialogue and a marginal note which more surely indicate his responsibility.[4] Here the editor's new belief in the accuracy of the verse-lining of the printer's copy and his knowledge of the sequence of composition will encourage a reappraisal. He will notice that while both compositors were prepared to alter the verse-lining of their copy to accommodate the text cast off for a particular page, they set most of the book, like dramatic manuscripts in Ralph Crane's hand, with each new speech beginning a new line of type. Many times they could have countered inaccurate casting-off by starting a new speech in the same line as the conclusion of the previous one. But this they did not do until Sheet I. Clearly a firm printing-house decision was reversed at this point. The following table shows how both compositors availed themselves of this means of adjustment on pages after Ilv. The numerals stand for the number of times, on the page indicated, a new speech begins in the same line of type as the last line, or part-line, of the previous speech.

         

	1v	2	1	2v	3v	4	3	4v
I	2	--	--	--	6	--	1	--
K	--	--	1	4	4	4	3	--
L	4	2	4	5	--	2	1	3
M	5	4	1	2	--	--	--	--

Once introduced, this kind of space-saving was used repeatedly. For a few sheets casting-off was probably modified in order to take advantage of it; but when, half-way through sheet M, it became obvious that the text could be accommodated comfortably within one more sheet, the early rule of starting each speech with a new line seems to have been re-inforced. If continuous printing occurred only in the last five or six formes, it would seem that the procedure was modified in order to avoid using a further sheet or half-sheet; but the new arrangement was started uncertainly, and was used consistently in only three or four half-sheets immediately following the tentative change.

These facts are relevant to a re-consideration of the authority of the stage-directions because the first four formes with continuous printing contain three directions, each beginning (as no others in the play) with the word 'Here . . .'. The setting of these directions suggest that they were added, like those found among the variants due to presscorrection, after the type had been set up. On I1v the direction is printed:

If this were added after composition of the text of the dialogue, two lines would have to be saved to accommodate it: and on this page two speeches are printed continuously (probably for the first time), saving two lines. On K1 the direction is printed: Again if this were a late addition two lines would have to be saved: and on this page one speech is printed continuously, and there are 38 instead of 37 lines of type. (The last part-line, the signature and the catchword are all, quite exceptionally, set in one line of type.) On I4, the direction: is centered, with a line's space each side, and is followed by the Madman's song, also in italic and running to within a one-line space of the full text-space for 37 lines. If the direction were added after the page was set up and if the compositor thought he should space it out to avoid confusion with the italicized song, four lines would have to be saved. In fact six lines are saved by continuous printing on the preceding page of 38 lines which, as we have seen, would have been set with I4.[5] It seems probable that the directions on I1v and K1 were added after the text had been set up from the printer's copy, and possibly that on I4. There are no features in the setting of I3 or K2v-4 which can suggest a reason for the tentative change from the procedure fixed from the start of the work.

Besides massed entries at the head of scenes, a dumb-show and simple marks for exits, there are only 16 stage-directions in the quarto and, as we have seen, two of these are among the press-variants for which Webster was probably responsible, and three appear to be late additions. The other eleven may now be scrutinized afresh for signs that they too were additions to the first type-setting. The directions are:

• i. he kneeles (C3v; I.i.477)
• ii. Enter Antonio with a Pistoll. (G1v; III.ii.166)
• iii. she shewes the poniard. (G1v; III.ii.177-178)
• iv. giues her a dead mans hand. (I1v; IV.i.50-54)
• v. A Coffin, Cords, and a Bell. (K1v; IV.ii.164-166)
• vi. They strangle her. (K2v; IV.ii.243-244)
• vii. Shewes the children strangled. (K3; IV.ii.272-273)
• viii. she dies, (K4v; IV.ii.381)
• ix. He kills the Seruant. (N2v; V.v.46)
• x. He wounds the Cardinall, and (in the scuffle) giues Bosola his death wound. (N3; V.v.70-72)
• xi. He kills Ferdinand. (N3; V.v.83)

None of these directions is in the formes (F inner, G outer and H inner) which have the 'Websterian' press-variants; but all except i are subsequent to F inner where Webster's attendance at the press has first been suspected. The one nearest to F inner is also remarkable for being the only one set on a line of type by itself, interrupting the dialogue. And it is on one of the three pages where Compositor A set 38 instead of 37 lines of type: like the three directions beginning 'Here . . .', this one could have been added after the page was set up but before any sheets were printed, being accommodated by using more than the usual text-space. Direction iv is placed, like others, to the right of the dialogue, but unlike them (and like the small type of the author's note added during press-correction on H2) it is not aligned with the rest of the type on the page: this direction could not have been set at the same time as the dialogue on which it comments; perhaps it also was added after the rest of the type had been set up, at the same time as the direction 'Here is discouer'd. . . .' which is on the same page. There is another kind of evidence for thinking that direction v was set in the same way. It occurs on a page where italic, capital Bs were in short supply. One was available for the first prefix 'Bos.', but by the 19th line, to the right of which the direction begins, three roman Bs had already been used and no more italic; after line 19, three more italic Bs were called for by the copy but the compositor had to be content with roman, and on the next page set the roman exclusively. Yet in 'Bell' of this stage-direction an italic B was used. It may have been set after the shortage of Bs had been relieved, that is after the two pages of the inner forme in its half-sheet had been set up.

Before investigations into type-shortages and the order of composition were undertaken, it was known that two stage-directions had been added after composition: now evidence has been found for believing that six more were similarly added, but before any surviving sheets had been printed off. These eight occur in the work of both compositors, so some source other than their copy must surely be posited. They might have been added because some printing-house corrector reversed a decision to omit some of the stage-directions found in the copy; but it is hard to believe that such crossed purposes would remain so long without being disentangled. Because we already believe that Webster visited the press to correct the printing of some formes, it is tempting to think that he was personally responsible for adding these stage-directions.

There is nothing in the position or type-setting of the remaining eight directions which argues against the same origin and procedures for them, save only that 'he kneeles' on C3v would be a very isolated

addition. If all sixteen were added in the same way, from the same source, only 9 formes would be involved out of 24.

One further aid to the editor that can come from a close study of the type-setting of The Duchess of Malfi concerns punctuation. We know that, like Crane manuscripts and unlike other plays from Okes' printing-house at the same time, the quarto has more than usual numbers of colons and semi-colons, brackets and hyphens. A table showing the page-by-page fluctuations in the number of colons set,[6] especially if the pages of each half-sheet are arranged by formes, suggests that on certain pages the punctuation was influenced by type-shortages, and in some cases these apparent shortages coincide with those already noticed from less equivocal evidence. On L1-2v, for example, the appearance of italic T on L2 and L1 suggests that a distribution of type was made on L1 after L1v and L2 had been set, and the number of colons on individual pages varies in accordance with such a procedure. Compositor A usually set more than a dozen on each page, but here the figures are:

   

L1v	L2	L1	L2v
16	9	5	21

All but one of the colons on L1 occur after the second of the two incorrect Ts. Sometimes an apparent shortage of colons is found on pages which also have fewer brackets or hyphens than most. L1 and 1v, for example, have no hyphens at all, and L2 has 3; L2v has 4. On M2v both Bs and Ws were in short supply, and on this page there are only two pairs of brackets (there are none on the preceding M1) and two hyphens, and no colons at all — it is the only page in the whole book without one. Type-shortages are indicated on K4v by one roman B in a prefix and nine italic Ts, and this suggests that the presence of only two pairs of brackets within the first 8 lines (there are none on the preceding K3) and no hyphen (there are only 2 at the top K3) are also due to type-shortage and not the printer's copy. There are 7 consecutive roman Bs in prefixes on L4v and on this page there is a drop in the number of colons and no brackets (there are only one and a half pairs on L3 and none on L4).

Such matters must be judged after long familiarity with the text, for the frequency of any mark of punctuation will vary according to the syntactical, rhetorical, metrical and dramatic qualities of the dialogue and to unknown variations in the scribe's, compositors' and, perhaps, author's techniques. But the editor of The Duchess of Malfi, bearing in

mind the copy's heavy demand on type for certain marks of punctuation, may believe, on some exactly limited occasions, that if he is to punctuate his text responsibly he must consider the influence of typeshortage. A study of the printing of the play can influence his work in such minute and finely-judged details as these.

notes