University of Virginia Library

Much of the research for this essay was completed while on a Folger-NEH fellowship at the Folger Shakespeare Library, Washington, D.C., while the author was working on an edition of Coriolanus for the New Cambridge Shakespeare.

The Riverside Shakespeare, gen. ed. G. Blakemore Evans (Boston: Houghton Mifflin, 1974), p. 1437a. In 1985, in The Craft of Printing and the Publication of Shakespeare's Works (Washington D.C.: Folger, 1985), George Walton Williams still found the scholarly consensus to be Shakespeare's own `fair copy' of his `foul' or working papers (p. 84), and in `Today We Have Parting of Names: A Preliminary Inquiry into Some Speech-(Be)Headings in Coriolanus', Thomas Clayton agrees about Shakespearean holograph, but thinks the manuscript might have been used as a promptbook (MS copy, courtesy of the author, p. 24; forthcoming in Shakespeare's Speech-Headings and the Bibliographer, the Editor, and the Critic, ed. George Walton Williams, AMS Studies in the Renaissance 25 [New York: AMS Press]).

Albert Gilman, `Textual and Critical Problems in Shakespeare's Coriolanus', 1954 Ph.D. diss., University of Michigan, p. 147.

William Shakespeare, Coriolanus, ed. Philip Brockbank, The Arden Shakespeare (London: Methuen, 1976), p. 7.

In William Shakespeare: A Textual Companion (Oxford: Clarendon Press, 1987), gen. eds. Stanley Wells and Gary Taylor, Coriolanus appears in the summary chart of copy texts on p. 147. John Jowett's later discussion of the nature of the copy is more certain of a promptbook original, although the F compositors' copy may have come to them at one remove: `The inaccuracies in speech-prefixes might alternatively indicate that the copy for F was not the prompt-book itself but a transcript of it — in which case the prompt-book may or may not have been based on authorial fair copy' (p. 594).

R. B. Parker, ed., Oxford Shakespeare Coriolanus (Oxford: Oxford Univ. Press, 1994), p. 143. Parker thinks that the author's fair copy probably became the company's playbook and that the F copytext manuscript was probably the `penultimate draft in which he was still to some extent "thinking through" the play' (p. 147).

Charlton Hinman, The Printing and Proof-Reading of the First Folio of Shakespeare, 2 vols. (Oxford: Clarendon Press, 1963), 2: 506. In `Line Division in Shakespeare's Dramatic Verse: An Editorial Problem', Analytical and Enumerative Bibliography 8 (1984), Paul Werstine emends Hinman by noting that since four formes of Coriolanus were set before the first of Titus was composed, `it was not strictly necessary that copy for the whole of Cor. be cast off until after the last of these formes (2v:5) was in type. Perhaps then three pages of Cor. were not set from cast-off copy (aa4, 4v, 5v)' (p. 117 n. 30).

Hinman, Printing and Proofreading of the First Folio, 2: 163-64. Hinman's temporal calculation is based on the fact that `between cc3:4v and the postulated cancels in groups 7 and 8 come two full quires of the regular series and five intercalary formes'.

The suggestion, first made by Harley Granville-Barker, Prefaces to Shakespeare, 2 vols. (Princeton: Princeton Univ. Press, 1947), 2: 294, was developed by John Dover Wilson in his New Shakespeare edition of Coriolanus (Cambridge: Cambridge University Press, 1960): the elaborate SDs for this play, as for The Tempest, were `perhaps necessitated ... by the author's absence in Stratford at the time the play was being rehearsed' (p. 131); see also Martin R. Holmes, Shakespeare and Burbage (Chichester, Sussex: Phillmore & Co., 1978), p. 196.

In the New Variorum Shakespeare Antony and Cleopatra (MLA, 1990), Marvin Spevack questions a number of Wilson's list of characteristic Shakespearean spellings (New Shakespeare Coriolanus, pp. 133-34), but he offers support for the unusual use of sc for initial s, unmentioned by Wilson, in Scicinius for Sicinius (p. 374). Brockbank, in his Arden edition of Coriolanus (pp. 3-4), does discuss this example, and he adds several others `which may indicate Shakespearean idiosyncrasies'—such as shoot for shout, strooke for struck, god for good, too for to, especially in the form too't; see also Parker's Oxford Shakespeare edition, pp. 138-39.

On the pages of Addition II of More, see Shakespeare's Hand in `The Play of Sir Thomas More', ed. A. W. Pollard and J. Dover Wilson (Cambridge: Cambridge Univ. Press, 1923); for the other instances of sc for s, see the Oxford Textual Companion, p. 593, and Spevack, New Variorum Antony and Cleopatra, p. 374.

Alice Walker's conclusion in `The Folio Text of 1 Henry IV', SB 6 (1954) that Compositor B exhibited `habitual carelessness' and `was unusually prone to take liberties with his copy' (p. 55) has, on a more extensive survey of B's Folio work, been corrected by Paul Werstine in `Compositor B of the Shakespeare First Folio', Analytical and Enumerative Bibliography 2 (1978), 241-63, and `Line Division', cited in n. 6.

In his editior's preface to the most recent book of essays on the More manuscript, Shakespeare and 'Sir Thomas More': Essays on the play and its Shakespearean interest (Cambridge: Cambridge Univ. Press, 1989), T. H. Howard-Hill notes that his contributors at the very least agree that the case for Shakespeare as author of Addition II is as strong as any so far made to deny it or identify another playwright as Hand D (p. 2). Howard-Hill himself concludes that together the essays tip 'the balance further in favour of Shakespeare as the author and hand' of Addition II (p. 8).

In 'Spelling and the Bibliographer', The Library, fifth ser. 18 (1963), T. H. Hill concludes that it 'has not yet been settled whether a preference for one form of spelling predisposes an author or copyist towards another morphologically similar spelling' (p. 11). His own study of Ralph Crane shows that Crane, at least, was not so predisposed.

The probabiliity that an idiosyncratic spelling would be perpetuated through the labours of two regularising copyists is lower than for one, of course (Q Justice Scilens does not survive in F 2 Henry IV); but the fact that Scicinius is an uncommon Roman name, rather than a common English noun for which a copyist often had his own preferred spelling, might increase its chances of being carefully reproduced by both scribe and Compositor A. It should be noted, however, that A. C. Partridge cautions against depending on the initial sc as being spectacularly rare; in Orthography in Shakespeare and Elizabethan Drama (Lincoln: Univ. of Nebraska Press, 1964), he cites other contemporary examples (p. 62), and I have found additional ones, including settings in John Marston's Sophonisba (Q 1606) of Scyphax for Syphax, scilent for silent, and scilence for silence (twice).

Shoot appears on both Wilson's and Brockbank's lists (see n. 9); arrant is suggested by David Bevington in his New Cambridge Shakespeare edition of Antony and Cleopatra (Cambridge: Cambridge Univ. Press, 1990), p. 261.

Act, scene, and line numbers are cited from The Riverside Shakespeare (see n. 1) and, following the slash, TLN (Through-Line-Numbers) keyed to Charlton Hinman's Norton Facsimile of The First Folio of Shakespeare (New York: W.W. Norton, 1968). These will sometimes appear not to match, where the modern edition has relineated the F line(s). Because they are often pertinent to my argument, the F spelling (with only u/ v and long s normalised) and line breaks have been retained. F's terminal periods for SDs and unabbreviated SHs have been omitted as too distracting to appear repeatedly in mid-sentence. When medial or terminal punctuation, or its absence, is significant, it will be noted in the discussion.

In The Lives of the Noble Grecians and Romanes, transl. Thomas North (London, 1595): `So the Coriolans ... made a salye out upon them, in the which at the first the Coriolans had the better, and drave the Romaines backe againe into the trenches of their campe' (p. 238); later, Coriolanus `disguised himselfe in such array and attire, as he thought no man could ever have knowen him for the person he was, seeing him in that apparell he had upon his backe ... his face all muffled over' (p. 247).

See the Oxford Textual Companion introduction to All is True, p. 618, for a more detailed discussion of evidence for Shakespeare's apparent preference for has, and for the fact that the only other plays in which Compositor B, at least haphazardly following his copy, set a high proportion of ha's are Macbeth and F Hamlet, both generally agreed to have been set from scribal copy; for do's and does, see the Textual Companion introduction to F Lear, pp. 530-31. Less certain as indication of a scribal intermediary is Coriolanus' pattern of 40 Oh against 12 O, although O `predominates over Oh in nearly all the good quartos of Shakespeare's plays', according to MacD. P. Jackson, Studies in Attribution: Middleton and Shakespeare (Salzburg, 1979), p. 215; see also Gary Taylor and John Jowett, Shakespeare Reshaped, 1606-1623 (Oxford: Clarendon Press, 1993), Appendix II.

This and the preceding examples and figures are taken from the Oxford Textual Companion, p. 593, although I have corrected the number of a'th from the Textual Companion's inaccurate 29; the figures given include variant spellings that lack the final mark of elision, a'th and o'th. (The fact that in both All's Well and Coriolanus most a'th' were set by Compositor B suggests that he was more tolerant of this form, less prone to regularise to o'th', than Compositor A, who in Coriolanus set only 1 a'th' to 6 o'th'). On a'th' and o'th', John Jowett, the Oxford editor of Coriolanus, finds that `to some extent' these two forms are found in `clusters', and that `in particular the copy for 4.5.21-4.6.131/2285-2631 might have been in a second hand' (p. 593). Inspection of the passage in question, however, does not seem to substantiate this last point.

For instance, o'th' is common in both Coriolanus and Antony and Cleopatra; Coriolanus' 27 instances of a'th', however, distinguish it from Ant., which has only 1. In `The Use of Linguistic Criteria in Determining the Copy and Date for Shakespeare's Plays', Pacific Coast Studies in Shakespeare, ed. Waldo F. McNeir and Thelma Greenfield (Eugene: Univ. of Oregon Books, 1966), Frederick O. Waller finds that from about Twelfth Night on, there is a clear pattern of Shakespeare's `moving toward more informal, colloquial usages', with an increasing use of has and does, alongside the early forms hath and doth, and of the contracted forms of the (p. 7). However, the figures for Coriolanus, and the number with added apostrophes, are still anomalously high.

Brockbank suggests as a Shakespearean preference too for to, though he connects it with the form too't (see n. 9).

In his note to Hamlet 1.5.35 (Textual Companion, p. 404), Gary Taylor observes that the `Shakespeare canon contains 1,526 occurrences of tis to only 35 it's'. Only 3 it's appear in good quartos; most occur in plays generally agreed to have been set from scribal copy, and `another 7 occur in Folio texts of doubtful copy', Coriolanus (4) and Antony and Cleopatra (3). I think Coriolanus was set from a scribal transcript, and this is probably the case for Ant. as well.

It is of course possible that both instances result from the wrong type being in the compositor's `I' box, or from Compositor B's mistake as he repeated the word to himself to hold it in mind while he set his composing stick. In his Oxford Shakespeare edition of Coriolanus, Parker disagrees: he finds that Hand D's capital I `resembled an A or arabic 4' and was responsible for these errors (p. 139). Inspection of the plates in Sir Edward Maunde Thomspon's Shakespeare's Handwriting and Pollard and Wilson's Shakespeare's Hand in `The Play of Sir Thomas More' indicates that Hand D's I could be mistaken for 4 but not for A.

Partridge, Orthography, p. 58. Regardless of the identity of Hand D, Shakespeare's own punctuation was likely to have been light. In 'The Noisy Comma', Antony Hammond notes that most dramatic manuscripts 'before about 1620 are very sparsely punctuated, and what punctuation there is ... is based upon conventions different from those which the printers were in the process of establishing' (Crisis in Editing: Texts of the English Renaissance, ed. Randall McLeod [New York: AMS Press, 1994], p. 214).

Absence of final punctuation and Shakespeare's tendency to write initial c as a capital apparently created the new sentence at 4.1.5/2441 (bb4v), although the break violates the original sentence's sense: `You were us'd | To say, Extreamities was the trier of spirits, | That common chances. Common men could bear,...'. It perhaps accounts for the random capitalisation in mid-sentence elsewhere of verbal forms beginning with c. Another whole passage of mispointed verse comes at 3.3.67-71/2348-52 (bb4).

Jaggard's house style for F was to use an apostrophe to indicate an unvoiced ending to the past participle in verse, e.g. `answer'd'; instances of elision indicated by spelling alone in Hand D are `lyvd' (= lived) and `sylenct'. Initial elision in the More pages can be seen in `thippe' (= th'hip) and `thoffendor'.

Partridge, Orthography, pp. 102-03. For evidence from Jaggard's 1599 printing of Shakespeare's sonnet 144 in the quarto of The Passionate Pilgrim, see pp. 133-34.

In `Author, Compositor, and Meter: Copy-Spellings in Titus Andronicus', Papers of the Bibliographical Society of America 53 (1959), Hereward T. Price thinks it clear that Shakespeare `closely adapted spelling to meter, showing elision, where that was necessary, or printing the full form, just as the meter demanded' (p. 160). In his study of Spenser, Massinger, and Harrington, he finds that in most cases, when setting verse, the compositor would follow his copy (p. 185). If this is true, then the scribe would be the one to have expanded most of the elisions. Yet we cannot be certain that Shakespeare was consistent, at least for medial elision, or that there were as many instances as I have supposed. It is also possible that compositors had certain preferences of their own or for some words set colloquial pronunciation, since marv'llous (2682) occurs in a short line of prose.

In `Folio Compositors and Folio Copy: King Lear and Its Context', Papers of the Bibliographic Society of America 79 (1985), Gary Taylor finds that Compositor B would add anomalous apostrophes where the sense does not require them, so some beyond those imposed by the scribe might have crept in in the printing house (pp. 35, 49). An instance in Coriolanus appears at 4.1.37/2476: start's.

Ibid., p. 65.

Taylor and Jowett, Shakespeare Reshaped, Appendix I, p. 246.

Ibid.

For tables showing the proportion of exclamation marks taken from printed copy to those added by Compositor B, see Taylor, `Folio Compositors and Folio Copy', p. 66.

Werstine, `Line Division in Shakespeare's Dramatic Verse', p. 76. As will be apparent, my discussion of the lineation problems in Coriolanus is deeply indebted to this essay. On the mislineation in Hand D, see also the Oxford Textual Companion, p. 637.

Werstine, `Line Division', p. 97. Werstine's general assertions about Compositors A and B are based on his study of those F plays set from earlier printed copy and of the F plays set by these two compositors. See also Brockbank's discussion in his Arden ed., pp. 12-16.

Werstine, `Line Division', p. 104.

Ibid., p. 96. Werstine notes that only 12 of A's 15 combined verses and 19 of B's 29 occur on pages where they were `demonstrably short of space' (p. 95), and also that if the compositors' copy contained combined verses,`we should expect to find repeatedly in the printed text not only two lines of verse set on a single line of type, but also two lines of verse set as two lines of prose, as we do in both Antony and Coriolanus' (p. 121 n.40).

Ibid., p. 97. Of Compositor B, Werstine notes that the example of Antony `indicates that he reproduced irregularities [in line division] in his copy when he found them' (p. 101).

The Lives of the Noble Grecians and Romanes, p. 238. The 1595 edition corresponds in spelling with Shakespeare's text in conduites (compare Coriolanus, 2.3.242/1645), as well as in providing both the family names for Titus. For futher arguments on this edition as Shakespeare's source, see Brockbank's Arden edition, pp. 27-29.

The play's variant spellings of Corioles — Carioles (5 instances), Coriolus (1), Corialus (2) — may derive from misreadings by scribe or compositor ( Carioles appears in Compositor A's work as well as B's, although the latter two variants are set only by B); but all, or some, may originate in a Shakespearean casualness in reproducing North's Plutarch's Corioles. If the scribe and/or Compositor A were as careful here as with reproducing Scicinius, then at least Carioles as well as Corioles stood in the original manuscript text, or were misread as standing there.

In The Stability of Shakespeare's Text (Lincoln: Univ. of Nebraska Press, 1965), E. A. J. Honigmann offers a theory of `irregular composition' that would explain the Lartius/Latius variations differently. In his view, Shakespeare did not begin composition with 1.1 but rather wrote 2.1.161ff., 3.1, and perhaps parts of 1.1 `before spotting the misprint in his Plutarch'; the argument is clinched for Honigmann by Shakespeare's highly unusual repetition of an identifying `label' — repetition but `not in its original form' — for Cominius in the SD at 2.1.161/1060 (Cominius the Generall) and at 2.2.36/1241 (Cominius the Consul), the latter a scene not, in Honigmann's view, written consecutively after 2.1 (p. 147). I do not find this argument convincing, in part because, as I have indicated, there are other possible explanations for the Lartius/ Latius variations, and in part because the Cominius labels can be seen as consistent with continuous composition, since they designate distinct functions for this character in the two scenes: returning triumphant military general who then, in 2.2, reverts to his peacetime office of consul. The labels may also be meant to indicate costume changes.

Capell first added with others to the exit direction for Coriolanus and Cominius at 3.1.252/1980. The only Coriolanus supporters who have speaking parts after this exit are Menenius, one or two patricians (two Patri. SHs), and one or two senators (I Sen., Sena.). In his Oxford Shakespeare edition, however, Parker has Titus exit earlier, at 3.1.20/1698.

Evidence from Hand D's scene in Sir Thomas More is perhaps most easily studied in Thomas Clayton's transcription in The `Shakespearean' Addition in `The Booke of Sir Thomas Moore': Some Aids to Scholarly and Critical Shakespearean Studies, Center for Shakespeare Studies Monograph No. 1 (Dubuque: Wm. C. Brown, 1969); see also the Oxford Textual Companion, pp. 461-67. In `Today We Have Parting of Names', Clayton suggests that the kind of `decremental abbreviations' found in the More manuscript may lie behind some SH mistakes in Coriolanus (p. 13).

What we cannot tell is whether the Latius spelling represents Shakespeare's original or a later misreading of a hastily written Lartius whose minim strokes fell short of the full complement for r and were glossed over for an immediate t.

Indeed, the first instance on aa4 is Titus Lartius, even though only this character's first name had been given in the immediately preceding SD, Enter Titus with his Power, from the Pursuit. Either Compositor A reproduced an unusually full copy form, or he dropped Lartius from the SD to make it centre more attractively on the page and then added it to the SH, since he was only setting a part line there and had plenty of room for it. It should be noted that he set one Mar. SH on this page, but he was forced into this expedient because the verse line at 1.9.41/797 is a long one.

Compositor A's preferred spellings, established by comparing his relevant F work with its printed quarto copy, are part of the evidence detailed by Gary Taylor in `The Shrinking Compositor A of the Shakespeare First Folio', SB 34 (1981), 96-117.

Other marks of non-authorial copy suggested by Compositor A's first page are 1 ha's, 1 do's, 2 Oh, and perhaps the round brackets. Compositor A generally resists the do's spelling, and this is his only instance in Coriolanus; elsewhere he sets 2 does and 3 doth.

See above and n. 7.

Clayton, `Today We Have Parting of Names', pp. 12-14, 24. The sequence in question reads: Coriol., Coriol., Corio., Com. (mis-set for Cor.?), Cor., Cor., Cor.

If we consider the whole page, 4 brackets, 6 Oh, 3 ha's; if we consider only the second column, where the Coriol.-Cor. series appears, 2 pair of brackets, 4 Oh.

There are also two occurrences in Q1 Pericles (1609), but this is a corrupt text and not thought to have been set from Shakespearean autograph.

Facing a similar anomaly with the 6 Omnes (though only 8 All, all set by B) in Ant., Marvin Spevack sees the probable implication to be that `the first instance of Omnes,...set by Compositor E, was the copy's form. The other instances, set by B, are typical of his vacillation between what may be copy and the more economical All' (New Variorum Antony and Cleopatra, p. 368).

Possibly with (or wth) o, where the o had been smudged out of existence. In `A Prompt Copy of A Looking Glass for London and England', MP 30 (1932-33), Charles Reed Baskervill notes inconsistencies in the book- keeper's correction of SHs in this black-letter quarto that has been marked up for a prompt: in some cases the printed copy's mistake is heavily scored through, in others only lightly, and in 5 instances the `incorrect printed SH has been allowed to stand along with the MS entry', written to its left (p. 38). The same casualness might affect annotated SDs.

In his Arden edition of Hamlet (London: Methuen, 1982), Harold Jenkins includes among what he takes to be the book-keeper's annotations to Q2's foul-paper copy the duplicate sound cues and Cum Alijs (p. 42); in a footnote he suggests that this book- keeper was the same man who prepared the fair copy behind F Hamlet.

W. W. Greg, in The Shakespeare First Folio (Oxford: Clarendon Press, 1955), thought that because Coriolanus was intended to head the `Tragedies', act division was introduced at the time of printing as a mark of sophistication (p. 407). T. W. Baldwin thinks that, had this been the case, `one would have expected full dress of acts and scenes, not acts merely'; he suggests the `probability, not certainty, is that Coriolanus, Titus, and Caesar had recently been in production, so that availability may have been a weighty reason for using stage copies for them' (On Act and Scene Division in the Shakespeare First Folio [Carbondale: Southern Illinois Univ. Press, 1965], p. 105).

Wilfred T. Jewkes, Act Division in Elizabethan and Jacobean Plays, 1583- 1616 (Hamden, Conn.: Shoestring Press, 1958), pp. 96-101. Jewkes thinks act division in Coriolanus was probably not entered at the time of composition; to him it seems more likely that it was introduced at the same time as the additional SDs, and both `in connection with performance' (pp. 188-89).

Taylor, `The Structure of Performance', Shakespeare Reshaped, pp. 18, 30-31. J. Dover Wilson had suggested this adoption of act intervals in `Act and Scene Division in the Plays of Shakespeare', RES, 3 (1927), p. 394; see also Jewkes, Act Division in Elizabethan and Jacobean Plays, pp. 91-92, and Richard Hosley, `The playhouses and the stage', in Kenneth Muir and S. Schoenbaum, eds., A New Companion to Shakespeare Studies (Cambridge: Cambridge Univ. Press, 1971), p. 33.

Taylor has proposed a similar reading in `The Structure of Performance', p. 40. In his Oxford Shakespeare edition, however, Parker argues that the F division between Acts 3 and 4 `seems a mistake' and is perhaps therefore unShakespearean; yet since in his designation of three main movements in the play's action the first also overlaps an act division (between 1 and 2), he concludes that `the asymmetry may be intentional' and leaves unaltered F's notation of Acts 3 and 4 in his own text (p. 28).

Paul Werstine, `Narratives About Printed Shakespeare Texts: "Foul Papers" and "Bad Quartos"', SQ 41 (1990), p. 69. He urges that we abandon the desire to `fix the origins of the early printed versions upon single agents, when these texts were open to penetration and alteration not only by Shakespeare himself and by his fellow actors but also by multiple theatrical and extra-theatrical scriveners, by theatrical annotators, adapters and revisers (who might cut or add), by censors, and by compositors and proofreaders' (p. 86). See also Gary Taylor, `Post-script', Shakespeare Reshaped, p. 243.

Philip Williams, `New Approaches to Textual Problems in Shakespeare', SB 8 (1956), p. 6. Williams, however, believed this copyist to be the same one who transcribed Timon of Athens.

Brockbank, Arden ed. Coriolanus, p. 24.

OED Both a B 1b (`Extended to more than two objects.'). One of the OED examples is from 1 Henry VI: ` Margaret shall now be Queene, and rule the King: | But I will rule both her, the King, and Realme' (5.5.107-08).

Editors have generally assigned all these unnumbered speeches to `First Senator' or `First Soldier', but this procedure gives perhaps unintended prominence to single individuals (though those actors might be grateful). The desired degree of verbal involvement in the action by those on stage might, for a specific production, have been left to the rehearsal stage. In `Editorial Treatment of Foul-Paper Texts: Much Ado About Nothing as Test Case', RES n.s. 31 (1980), Stanley Wells argues against an editorial judgement to assign the 5 unnumbered Watch speeches in 3.3 of that play to `2 Watch', as is usually done: `To give him all the lines is to build up his part beyond justification, and to make mutes of all the other watchmen; it goes beyond the textual evidence' (p. 11); the point is just as well taken if the copy is a faithful transcript of an author's undesignated original. Editorial assignment of such speeches may do a disservice to a script intentionally left open.

It should be noted, however, that in some scenes with entry directions of indefinite number, speeches are all assigned: for instance, of the seven or eight Citizens who enter at the beginning of 2.3, only `1', `2', and `3' speak, and in 5.6 there are speeches for `1', `2', and `3' of the 3 or 4 Conspirators of Auffidius Faction.

There are also unnumbered speeches for the Queen's ladies and for the tribunes in Cymbeline, 1.5 and 3.7; 5 speeches for Watch in 3.3 Much Ado About Nothing, although 7 others are specified `I.' or `2'(see n. 63); and in Timon of Athens successive SHs Some speake and Some other at 3.5.86-87, but as these F texts are not thought to have been set from prompt, we cannot know whether a playhouse annotator would have particularised the SHs. Julius Caesar, which is thought to have been set from prompt copy or a transcript of it, has two unnumbered speeches for Sold. in 5.4, but since the next (and only other) two soldier speeches are numbered, the missing specificity for the first two may result from compositorial oversight.

In `"Foul Papers" and "Prompt Books": Printer's Copy for Shakespeare's Comedy of Errors', SB 49 (1988), Werstine observes that SHs for the 2 committeemen who enter at l. 142 in The Launching of the Mary are sometimes `Com: I' but also sometimes `Com:' or `Committ:'; in The Captives appears a SH `woman' when three possible female speakers are on stage (p. 242). On variable SHs in manuscript playbooks, see also Werstine's `McKerrow's "Suggestion" and Twentieth-Century Shakespeare Criticism', RenD, n.s. 19 (1988), pp. 152-53.

The revising Hand C in Sir Thomas More, who seems to have been preparing the manuscript for theatrical production, altered only 1 of 10 All SHs in Hand D's 3 pages. F Julius Caesar, set from prompt or a transcript of it, uses All in 3.2 to designate speech for the crowd of plebeians, though in some instances the words clearly should be further distributed among several distinct voices.

E. A. J. Honigmann, `Re-Enter the Stage Direction: Shakespeare and Some Contemporaries', ShS 29 (1976), p. 122. Thomas Clayton notes that some All speeches, such as `Speake, speake', call for `not necessarily only two but for n repetitions' (`Today We Have Parting of Names', p. 7).

If the scribe that I am hypothesizing made his transcript late, perhaps with the F publication in mind, then he too could have contributed mistakes and misreadings.

Clayton, `Today We Have Parting of Names', p. 8. Clayton notes that in this last case, the two uses of All would still require disambiguation; for the possibility of duplicate Alls as successive SHs, he points to Giorgio Melchiori's persuasive argument that the two instances of all in l. 38 of More mean some and others (`Hand D in "Sir Thomas More": An essay in Misinterpretation', SB 38 [1985], p. 101). Editors have chosen among these possibilities in their handling of this passage: The Riverside Shakespeare, Brockbank's Arden ed., and Parker's Oxford ed. assume one misplaced SH and so move the existing All up from TLN 1896 to 1894, thus giving all 3 lines to All; others move All up to 1894 and give this SH 2 lines, but then variously reassign 1896 (Malone to Cit.; Kittredge to Patricians; Hibbard to Menenius, thus merging 1896 with Menenius' subsequent lines).

Such is the reasoning, I assume, on which the Oxford Textual Companion conjectures AEDILES. I find this hypothesis attractive in terms of context and, more guardedly, on bibliographical grounds. Unfortunately, Coriolanus offers no control for this misreading: the only other relevant SH abbreviation is a single instance in 3.3, a scene which uses the spelling `Edile' (Edi.). We do not know what 'Aed[.]' (or 'Æd[.]'?) in Shakespeare's hand would look like. In Hand D's pages of the More manuscript, SHs are entered in secretary hand (with the exception of one Italian S), and in this script `All' and `Aed' could be easily mistaken. Indeed, `Aedile' is quite possibily itself a scribal sophistication; if so, and if the miniscule initial letters in the `all' SHs in the More manuscript are any pointers, Shakespeare may have left `edi' without a distinct dot for the final letter, with the result that the abbreviation was mistaken for the commoner `all'.

Clayton points out that, on the pattern of Hand D in More, r would more likely be mistaken for m than vice versa, and both the examples here discussed misread in that direction. He also cites a personal letter from George Walton Williams relevant to even shorter abbreviations for these two characters: in Romeo and Juliet `there are many instances in which it is pretty clear that Shakespeare forgot to add the tilde in various contractions' (`Today We Have Parting of Names', p. 25).

Brockbank, Arden ed. Coriolanus, pp. 161-62.

In his New Shakespeare ed. of Coriolanus, Wilson notes, besides the easy graphical confusion, that `Brut.'s [sic] reply is clearly addressed to Cor.; and Cor. replies in turn' (p. 196).

Clayton, `Today We Have Parting of Names', p. 13. Clayton reminds us that Hinman established a considerable break in the printing of quire bb (see p. 000 and n. 7 above), and the bb2v:5 forme was among those that had to be reset; he also notes that the cruxes on bb2va occur in a `span of F lines (TLN 1915-65) that may all have been on the same page' of the manuscript copy (p. 13).

Warburton reassigned the first quoted line to Coriolanus; as early as F2 (1632) the second was given a SH for Cominius.

Brockbank, Arden ed. Coriolanus, p. 210 n.

Clayton offers a version of this argument in `Today We Have Parting of Names', p. 18. In New Readings in Shakespeare, 2 vols. (Cambridge: Cambridge Univ. Press, 1956), C. J. Sisson suggested that TLN 1962-64 were a marginal addition, `prefixed by the compositor in error to the existing speech of Menenius' (2: 127), but Clayton, rightly I think, finds this argument unlikely on grounds of both meter and sense (p. 17).

The actor who plays `I. Cit.' would, of course, appreciate the emendation, and his part can be — and has been — strengthened even further through non-verbal stage business in Acts 2 and 3. The role of `I. Cit.' can thus become an attractively substantial minor part for an up-and-coming actor; in some modern productions it has been played by the understudy for Coriolanus, as it was in Peter Hall's 1959 production in which Albert Finney played First Citizen and understudy for Laurence Olivier.

Clayton, `Today We Have Parting of Names', p. 12.

Wilbur Sanders mounts a persuasive defense of retaining F in Wilbur Sanders and Howard Jacobson, Shakespeare's Magnanimity: Four Tragic Heroes, Their Friends and Families (New York: Oxford Univ. Press, 1978), p. 140. See also Michael Warren, `Textual Problems, Editorial Assertions in Editions of Shakespeare', in Textual Criticism and Literary Interpretation, ed. Jerome McGann (Chicago: Univ. of Chicago Press, 1985), pp. 31-32, and Richard Proudfoot, `Textual Studies', ShS 30 (1977), p. 204.

H. Eardley-Wilmot, Times Literary Supplement, 13 October 1950, p. 645. In his New Shakespeare ed. of Coriolanus, Wilson further elaborates on the appropriateness to Sicinius' argument of specifying the disease as gangrene, whereas this would be a tactical error for Menenius (p. 204 n). Brockbank suggests a simple misremembering on the part of Compositor B (Arden ed. Coriolanus, p. 214 n); see also Clayton, `Today We have Parting of Names', pp. 19-20.

Theobald conjectured `?', which Steevens adopted in the 1778 Variorum; Malone followed Rowe (1714) in providing a dash to indicate interrupted speech and speculated that had he been allowed to continue, Menenius would have added, `Is this just?'. Replying to Eardley-Wilmot in the TLS and assuming the F punctuation to be correct, A. P. Rossiter defends F's assignment to Menenius as being in character with his `personal style' of `dry irony' ( Times Literary Supplement, 20 October 1950, p. 661. Rossiter's argument does not, however, answer the objections to F made by either Wilson or Clayton (see preceding note).

Clayton, `Today We Have Parting of Names', p. 20.

Brockbank, who gives the line to Virgilia, finds it only `just conceivable that Coriolanus cries out in apprehension or distress, or even intends a dismissive blasphemy, but either would be inconsistent with the assumed role of "lonely dragon"' (Arden ed. Coriolanus, p. 240 n).

In Dramatic Documents from the Elizabethan Playhouses, 2 vols. (Oxford: Clarendon Press, 1931), W. W. Greg answered the question of what treatment the book-keeper would acccord the author's SDs: `as a rule he left them alone. So long as they were intelligible it mattered little to him the form in which they were couched' (1: 213)

Although he later altered his opinion on this matter, in The Editorial Problem in Shakespeare (Oxford: Clarendon Press, 1942), W. W. Greg thought a book-keeper might `not trouble to make specific such a direction as "with two or three Lords attendant"' (p. 38). In several articles detailing his study of the surviving manuscript playbooks, William B. Long has presented strong evidence that Greg's early statements, here and in Dramatic Documents (quoted in n. 86) were correct. In `"A bed | for woodstock": A Warning for the Unwary', MARDE 2, ed. J. Leeds Barroll and Paul Werstine (New York: AMS Press, 1985), Long observes that modern bibliographers wrongly demand that the `ideal' late 16th- and 17th-century `"promptbook" should be neat and orderly, containing complete and regular speech-heads, entrances and exits — with all manner of vagueness and ambiguities resolved. Unfortunately, the surviving ... playbooks exhibit no such features, even on an occasional basis' (p. 93). In `Beggars Bush: A Reconstructed Prompt-Book and Its Copy', SB 27 (1974), Fredson Bowers concurs: `In the surviving playbooks, bookkeepers have not changed playwrights' entrance directions that leave the number of extras indefinite' (p. 94; see also p. 131). In Coriolanus there are 4 lightly permissive entry directions, 10 SDs for a collective body (e.g. the Army of the Volces at 1.4.22/514, or a rabble of Plebeians at 3.1.179/1886), and 9 unspecified with others or with Attendants.

In Casting Shakespeare's Plays: London Actors and Their Roles, 1590-1642 (Cambridge: Cambridge Univ. Press, 1992), T. J. King suggests that the number provided for with Lords/Soldiers/Attendants was usually 3, although it might vary according to the number of mutes required for the play as a whole (p. 34). For Coriolanus, King estimates that 11 men could handle the 16 principal male roles (25 lines or more); along with 3 boys, they would be responsible for 93% of the play's lines. An additional 14 men would be needed for 34 small speaking parts, as well as 4 boys and 30 mutes. According to King's tables, of the F plays only Antony and Cleopatra and Henry VIII require larger casts (see Chap. 4).

Long, `"A bed | for woodstock"', pp. 106, 110. In `Modern Textual Theories and the Editing of Plays', The Library, sixth ser. 11 (1989), T. H. Howard-Hill notes that, unlike modern promptbooks, Renaissance playbooks `do not usually record information contained in other necessary playhouse documents', such as actors' parts, casting tables, or the `plot' (p. 111).

In the printed, black-letter quarto of A Looking Glass for London and England that has been marked as prompt, `flourish' appears either in the left margin opposite the space for its occurrence or above the printed entry SD (Baskervill, `A Prompt Copy of A Looking Glass for London and England', p. 37). A hasty annotator might cause ambiguity with either placement.

Peter M. Wright argues that these are two distinct SDs, meant to cover movement during a scene change: the Flourish accompanies the departure of the victorious Romans at the end of 1.9, but the Cornets are for Aufidius' entry, or what editors usually mark as 1.10 (`Stage Directions in Two Seventeenth-Century Folios: Ben Jonson's Folio of 1616 and the Shakespeare Folio of 1623', Ph.D. diss., UCLA, 1991, pp. 144-45,n. 14). A flourish, presumably of trumpets, followed by a separate flourish of cornets, while possible, seems unlikely; as John H. Long observes in Shakespeare's Use of Music: The Histories and Tragedies (Gainesville: Univ. of Florida Press, 1971), `the fourish concludes the council just ended by the Romans. There is no reason for Aufidius, who enters defeated..., to be greeted with a fanfare' (pp. 221- 22).

In Bibliographical Studies in the Beaumont & Fletcher Folio of 1647 (Oxford: Oxford Univ. Press, 1938 [for 1937]), R. C. Bald notes this as a typical change made in turning the transcript of a play into the playbook for performance: `careful attention to musical effects and noises off stage in a text is an indication that it has received the producer's attention, and is probably derived from a prompt-copy' (p. 79). The Oxford Textual Companion also notes that the music cues, as well as the detailing of the names of characters in entry directions and the misplaced Exeunt at TLN 1948, suggest a playhouse annotator (p. 593). Brockbank, on the other hand, while seeing `some evidence for the intervention of the book-keeper', finds `it is not decisive' (Arden ed. Coriolanus, p. 5).

W. J. Lawrence in Shakespeare's Workshop (rpt.; New York: Haskell House, 1966) notes that at the beginning of the 17th c. cornets are associated with the indoor private theatres, where their softer tone could be used more effectively than in the public amphitheatres (pp. 52-53). The King's Men acquired Blackfriars in August 1608, although all playing was prohibited due to plague until at least November 1609; Lawrence takes the call for cornets as a pointer to dating Coriolanus, `which cannot have been produced before the end of November 1609, and seems more likely to have seen the light a month or two later. In either case it would have been a Blackfriars production' (p. 63); see also Long, Shakespeare's Use of Music, pp. 222, 228-29, and Oxford Textual Companion, p. 594. Though less likely, Coriolanus could have been composed early in 1608 and opened at the Globe in the spring, before plague closed the theatres, then been reannotated for a subsequent Blackfriars performance. Such appears to have been the case with All's Well and Merchant of Venice, which also mix trumpets and cornets in flourishes; All's Well exists only in its F form; Q1 Merchant (1600) lacks any reference to cornets. See n. 55 for T. W. Baldwin's suggestion that Coriolanus had received a production not long before work on F began.

The first doubling at 2.1.0.2 may have been authorial (Enter Menenius with the two Tribunes of the | people, Sicinius & Brutus), since `commonly Shakespeare identified a character at his first entrance with a label stating his rank, profession, or relationship to another character' (Honigmann, The Stability of Shakespeare's Text, p. 147); these labels are only rarely repeated, however, so the latter instances are likely to be those of an annotator making sure that entry directions included the names of all the scene's important characters. Even at 2.1.0.2, the order suggests that the tribunes' proper names may have been tacked onto an existing direction (in contrast, 1.3.0 begins Enter Volumnia and Virgilia, mother and wife to Martius). See also the Oxford Textual Companion, p. 593, on this point.

A SD Retire or Aside for First and Second Servingman at 4.5.52/2707 would be helpful, but only Capell, followed by Keightley, has been seriously confused by its absence. For another possible example in 1.4, Enter Martius Cursing, see p. 000 above and n. 98; Troilus and Cressida, 4.5.158, offers another instance where Enter applies to characters who have previously only gone `aside'.

Bowers notes that the `common failure to mark ordinary and expectable exits wanting in Massinger's papers shows that scrupulous attention to exit directions is no absolute criterion for prompt-copy' (` Beggars Bush: A Reconstructed Prompt-Book and Its Copy', p. 132); see also Gary Taylor, `"Praestat difficilior lectio": All's Well that Ends Well', Renaissance Studies 2 (1988), p. 40. Some or all may also have existed in the manuscript but been omitted by oversight or because of space constraints (the probable cause of the missing exit SD for 1.8 and the shorter Exit for Exeunt at the end of 1.7); they may also have been clear in the actors' parts, though unnoted in the playbook.

1.4 presents Coriolanus's major staging challenge, though the SDs may have been less confusing in the manuscript copy, before Compositor B got through with it. Still, on the basis of a non-Shakespearean playbook, Fredson Bowers concludes that `If Knight's transcript of this prompt- book from Massinger's papers is at all typical, it is evident that unresolved tangles in the action and even in the casting could be transferred to print from a prompt-book or transcript' (` Beggars Bush: A Reconstructed Prompt-Book and Its Copy', p. 135); see also Baskervill, `A Prompt Copy of A Looking Glass for London and England', pp. 50-51, and Oxford, Textual Companion, p. 387.

J. W. Saunders, `Vaulting the Rails', ShS 7 (1954), p. 77. In his view, no exit is required because neither Marius nor the Roman soldiers really leave (or leave our sight). The SD Enter Martius Cursing is `similar to those illogically entered in other plays, when a player has not left the stage but has moved between two different visible areas'. Modern productions often employ a similar solution, with the Romans retreating to the verge of the stage or even the centre steps and centre aisle; Titus exits fighting and re-enters from the same direction to comment on Martius' bravery (so staged in the 1984 Peter Hall production at the National Theatre). Martius may exit fighting and almost immediately re-enter cursing, or he may be visible driving off the Volscians.

The centred SD Enter the Gati also suggests Compositor B might have been led by an ambiguous terminal symbol to see Gati instead of Gates and to assume another army is indicated. Just above, at TLN 514, he had centred Enter the Army of the Volces; here, at TLN 543, he centres what was presumably a marginal action-direction for what is in effect an exit.

The fact that aa3v is B's first page of Coriolanus may mean that, as yet unfamiliar with the play and especially its proper names, he was trying to follow his copy closely. Had he been willing at 1.7.0 to alter his spelling, use an & and jettison Caius, as he may have done in the first column and again at 1.8.0, he could either have saved a line for white space (1.7.0.4 is only one word long, Scout), or achieved his preferred direction format. With the exception of this instance and the short entry direction on aa1v, he elsewhere sets simply Martius in his Act 1 Sds.

The Oxford Textual Companion finds Enters an additional argument for scribal copy, since the indicative rather than imperative form `does not elsewhere appear in texts thought to be set from Shakespeare's papers' (p. 593). Yet this SD seems one of the most `authorial' in the play, containing information that neither paid scribe nor compositor would be likely to know, and in fact the indicative form does appear elsewhere in a text believed to have been set from holograph: Q1 Richard II, on B2 and I1 (at 1.3.6 and 5.2.84 in the Riverside edition). Alternatively, Enters could be a scribal addition, but it might also derive from a playhouse annotator or from a compositor's correction for having failed to head the SD with Enter. Its capitalisation suggests it may have been written in the margin of Compositor B's copy, perhaps even as Enter, meant to head the SD but miscued (or its cue misread) as to placement.

In his New Shakespeare ed. of Coriolanus, Wilson finds it `obvious that Auf.'s [sic] address to the Volscians precedes their being "driven in breathless"' (p. 167). Many editors still retain F, however, and simply add Exit for Aufidius after his final two lines.

On bb5, for instance, Compositor B has managed to get the second servingman's speech at 4.5.3-4/2657 onto one line by abbreviating `master' as `M.', sacrificing two of the three spaces after punctuation marks, and leaving off the period after Exit. Mislineation at TLN 2659-60 is probably intentional, so that the first half-line would leave a bit of white space under Enter Coriolanus, and running lines 2664 and 2665 together as prose allows room in the right margin for the off-center Enter second Servant.

Discussing the combining of two verse lines into one as a compositorial space-saving technique, Werstine notes that on aa3 `the combined verses occur near the bottom of the second column where there is no white space around two entrance directions' (`Line Division', p. 120 n. 38).

In setting the two penultimate pages of F 1 Henry 4 from printed copy, Compositor B compressed his copy by shortening SDs to hold them to one line and by omitting one altogether on f5v, at TLN 3066 (cutting from K3 of the quarto He spieth Falstaffe on the ground); see Eleanor Prosser, Shakespeare's Anonymous Editors (Stanford: Stanford Univ. Press, 1981), p. 202 n. 21. In The Printing and Proof-Reading of the First Folio of Shakespeare, 2: 508, Hinman notes B's probable omission of a SD in Antony and Cleopatra, due to poorly cast-off manuscript copy. On B's willingness to tamper with his copy's SDs, see also Werstine, `Compositor B of the Shakespeare First Folio', pp. 252, 257-59, and Williams, `New Approaches to Textual Problems in Shakespeare', p. 8.

There are other mis-settings on this page (bb1v), such as mislineation and Latius in the entry SD and SHs in 3.1. The mislineation is quite likely at least in part Compositor A's but, as argued above, the Latius misnaming probably stood in the copy. The unnumbered senatorial SH at 3.1.31/1712 may be related to an economising impulse, since the line is justified only by eliminating the usual space after the SH period, but senators are either unnumbered (two more on this page alone) or erratically numbered in the rest of this scene, in both compositors' stints and where justification is not an issue.

Honigmann, `Re-Enter the Stage Direction', p. 119.

In `New Approaches to Textual Problems in Shakespeare', Philip Williams observes that Compositor A tended to reproduce the copy position of SDs; Compositor B was more likely to normalise, and he `was reluctant to interrupt a speech to insert even a needed stage-direction' (p. 7). Williams does not, however, offer his evidence for this assertion; see below, n. 110.

In `Justification and Spelling in Jaggard's Compositor B', SB 27 (1974), S. W. Reid provides evidence that `when justifying a line B generally altered types toward the end of it before changing those toward the beginning' (p. 110). TLN 3538 is a case in point, for a shorter spelling of `vntill' would have allowed Compositor B to set the sentence's terminal period.

Five lines earlier in the same speech B does split a verse line in order to set Drum a farre off in the right margin, rather than set the SD on its own separate line, but farther down the same column a sound and action SD breaks up Coriolanus' tirade to the Roman soldiers. If Compositor B had a preference for uninterrupted speeches, it was not a particularly strong one. As noted earlier, in the second column of aa3 one sound cue has probably been left out because there was no room in the margin next to `Hark', and by this time B could not spare a separate line.