Picture of Gabriel Egan G a b r i e l   E g a n  .  com

Year's Work in English Studies 2001: Shakespeare: Editions and Textual Studies

The current crises in theories of editing Shakespeare pivot on a single question: can we determine with tolerable certainty the kind of manuscript used as printer's copy for each of the early printings? Editors who think that we can tend to use this 'knowledge' to discriminate between multiple early printings to find the one they want to base their modern text upon and they conjecturally emend it by reference to their theories of how its errors came about, while editors who think that we cannot so discriminate tend to be more cautious, stressing the arbitrariness of their choices about base text and emendation. This year appeared two major critical editions of the same play, 3 Henry 6, one for The Oxford Shakespeare edited by Randall Martin, and one for the Arden Shakespeare edited by John D. Cox and Eric Rasmussen. The differences between them usefully illustrate the consequences of differing answers to the central question.

    Randall Martin's introduction to the Oxford Shakespeare 3 Henry 6 runs to 132 pages and is organized under 8 heads that move from 'Rediscovery and reception' through analyses of particular characters (Richard of Gloucester, Edward 4, Queen Margaret, but not Henry 6 himself), to Martin's view of the origins of, and relationships between, the early printings. The first edition was an octavo of 1595 (O) called The True Tragedy of Richard Duke of York, which Martin abbreviates to True Tragedy (the 1986 Oxford Complete Works chose Richard Duke of York), followed by quartos in 1600 (Q2) and 1619 (Q3), and the Folio text of 1623 bearing the familiar name of The Third Part of King Henry VI. Martin takes the now common view that True Tragedy was the second part of a two-part play the beginning of which is represented in the 1594 quarto called The First Part of the Contention of the Two Houses of York and Lancaster and that the play we know as 1 Henry 6 was a prequel written later to tell the pre-history to the two-parter. It is worth distinguishing True Tragedy as represented by O and the quartos from 3 Henry 6 as represented by the Folio because Martin thinks that substantial authorial revision separates them; they are not merely different names for the same thing. Changing titles are revealing, and True Tragedy's gives attention to York even though he dies in act 1 while Henry lasts to almost the end. Martin thinks that True Tragedy was Shakespeare's first version of the play, written in 1591, and that the 1595 octavo is an imperfect report of it, and that Shakespeare's longer version of the play was written 1594-6 and this is essentially what got into the Folio as 3 Henry 6. Martin's edition is based on the Folio.

    The Folio title is probably not authorial and it gives priority to Henry 6 without, however, mentioning his life or death, and moreover it "avoids the contemptus mundi associations hinted at by York's 'true tragedy'" (p. 20). 3 Henry 6 does indeed deepen the character of Henry. The play would have reminded people of the danger of two monarchs claiming one kingdom when Mary Queen of Scots arrived in England in 1568, which only ended with her execution in 1587. Sackville and Norton's Gorboduc, written early in Elizabeth's reign, was printed in 1590, the year before Shakespeare began on True Tragedy, and its representation of civil strife in a divided kingdom is alluded to in Shakespeare's play. An early performance of Gorboduc appears to have used a real company of soldiers in a formalized battle scene, showing the stage/reality crossover of "drill as theatrical rehearsal and combat as performance" (p. 23). Neoclassicism demanded that violence be reported, not shown, and Gorboduc breaks this rule and had to be excused for it in Sidney's Defence of Poetry. Sidney and Jonson tried to distinguish 'low' from 'high' dramatic art, a distinction that rather misrepresents Elizabethan drama's mingling of 'official' and 'popular' culture. Popular civic dramas such as those performed at Coventry would have large, well choreographed battles including female warrior-characters (but not female actors), and Martin notes that victory over the Armada in 1588 did not bring an end to military preparations, just the opposite: there was increasingly conscription by the government as well as the older kind of feudal conscription by lords raising troops from amongst their tenants. Hence the play's son (conscripted in London) who kills his father (conscripted in Warwickshire), and the wider theme of the "broken connection with local history" and the "uncertain embrace of metropolitan culture" (p. 32). The first London playhouse, the Theatre in Shoreditch, was near the muster ground of Finsbury Fields and the audiences may be expected to have appreciated (indeed, have experience of) drills being done well. On the stage weapons were "not simulated period props but actual contemporary equipment" (p. 33). Martin offers no evidence for this last alarming claim, and one assumes that weapons were blunted to prevent accidental slaughters in performance. As usual with the Oxford Shakespeare, footnotes are used to reference supporting materials, which makes for convenience of use at the cost of limiting space. Martin elects to give what is known as a 'deep' link to an article in the online journal Early Modern Literary Studies (p. 33n1)--meaning that the Internet address is specified right down to the particular document to be accessed--which is a practice to be deprecated because it wastes space and because the smallest error makes the link unusable. In this case there are several small errors and the link as printed does not work; writers should cite only the homepage of the journal (here, http://purl.oclc.org/emls) and provide sufficient additional information (title, volume, year) to allow the reader to navigate her own way down to the specific article.

    Continuing his exploration of contemporary contexts, Martin argues that history plays could be defended as tools for teaching military strategy and that where True Tragedy is overt and showy in its militarism, 3 Henry 6 is somewhat restrained, even pacifist. The opening stage direction of True Tragedy calls for the men to be wearing white roses, while 3 Henry 6 does not, and the latter increases the sense of confusion (who is who?) and makes the story more easily applicable to other conflicts (p. 26). Of course, Martin accepts that some differences between the True Tragedy and 3 Henry 6 might be due to the latter deriving from an authorial manuscript (so, Shakespeare had not yet thought to add the detail about the roses) or indeed to editing in the printing house. But Martin detects in other differences, such as Margaret and Prince Edward being captured separately in 3 Henry 6, the signs of subsequent authorial revision "toning down the sound and fury" and making the play's attitude towards war "more rueful". In short, 3 Henry 6 has been distanced from "official Elizabethan wartime and political contexts" (p. 37). Likewise True Tragedy lacks 3 Henry 6's imagery of war being like a sea (the water caught between the forces of the moon and the wind) that makes conflict seem like a natural condition rather than human sin.

    Martin ties his introductory sections on particular characters to the stage history, and under the Brechtian heading "The (resistible?) rise of Richard of Gloucester" he observes that one reason for the relative neglect of 3 Henry 6 was the success of Colley Cibber's adaptation of Richard 3 that held the stage between 1700 and 1821 and contained large sections of 3 Henry 6 (p. 46). Also, it is common to tack the beginning of Richard 3 onto the end of 3 Henry 6 by having Richard give his "Now is the winter of our discontent" speech, which gratifies audiences by linking the obscure play to the well-known one and gives a teleological reading: it was all leading up to the evil reign of Richard and then the good reign of Henry 7. But Martin thinks that 3 Henry 6 actually has a weak sense of historical causality and there is little justification of present actions by past ones; indeed, Shakespeare is less teleological than his sources (p. 49-50). Political motives for action quickly give way to blood-feud and competitive savagery, and then even familial ties cease to be a motive to action and all becomes "expedient violence" for "seizure of power" (p. 54). Under "'Edward IV'" Martin notes that Henry 6's entailing of the crown removed the "transgenerational continuity" that makes it "an abiding authoritative symbol", instead it is just a property; hence Edward 4 is not even given a coronation: "3 Henry VI's relatively subdued use of ritual denies Edward the myth-making spectacles of kingship" (p. 71). Here Martin gives the substance of his illuminating note reviewed last year (2000) on the regional and topical allusions of 5.1, which features places near to Shakespeare's home town and a 'John Somerville' whose namesake was probably his disgraced relative. Indicating just how far Shakespeare's play has influenced modern attitudes, Martin notes that Shakespeare downplays to the point of extinction the historical Warwick's reputation as one of the old class of martial aristocrats, loved for his courtesy and hospitality, and makes him more a self-serving setter-up and plucker-down of others; we think of him as the 'kingmaker' largely because of this play (p. 79). In telling "Margaret's story: a 'new' play" (pp. 82-96), Martin records that for most of the stage history of the play--until the mid-twentieth century, in fact--Margaret lost out in cutting and adaptation, yet now the part is frequently compared to that of Lear. There was a distinct trend to liken Margaret to Britain's Prime Minister Thatcher in productions by the English Shakespeare Company and the Royal Shakespeare company in the 1980s, combined with a noticeable anti-feminist backsliding from the progressive mid-century work; this tended to "rehabilitate patriarchal biases against an outspoken non-domestic woman" (p. 94).

    Martin's section on "The Original Texts: their history and relationship" is of greatest concern to this review. Martin claims that the copy for F is "generally agreed to be Shakespeare's manuscript", although there is debate about "its state" and whether other hands annotated it "in anticipation of use in the playhouse as a prompt book or script" (p. 96). It is hard to see what Martin means by "or" in that last phrase: 'script' is certainly a less contentious term than the wildly anachronistic 'promptbook' favoured by New Bibliography, but naming both does not make the claim any more tentative. Martin gives the standard New Bibliographical reason for thinking that F is based on authorial papers: some of its stage directions are "indefinite or vague", which suggests "'pre-performance'" status, and others are authorially descriptive rather than practically prescriptive (p. 97). As if to soften his line, Martin footnotes the work of Paul Werstine and William B. Long that showed that vague and indefinite stage directions were not necessarily absent from "playhouse copy", conjoining it to his own assertions with "however . . .". But if one accepts the validity of Werstine and Long's scholarship, one simply cannot use the evidence of 'permissive' stage directions to determine printer's copy; it will not do to simply name-check them and move on without stating where one stands on the matter. The problem recurs with other evidence for authorial copy, "Changes in speech prefixes [that] seem also to reveal subtle shifts in a character's function or status . . ." (the indicator first seized on by R. B. McKerrow (1935), and again Werstine's demonstration that these can be found in theatrical (as opposed to authorial) manuscripts is acknowledged but not refuted. On the matter of actors' names ("Gabriel", "Sinklo", and "Humfrey") occurring in the Folio text, Martin tangles with W. W. Greg much as Cox and Rasmussen do in the Arden version, as we shall see. Martin thinks that these names show that Shakespeare had specific people in mind for certain parts as he wrote, and in a footnote writes that "The names are unlikely to derive from a prompter annotating the play, since this kind of annotation typically takes the form of extra information or duplicate directions in extant playhouse manuscripts of the period. See Greg, Folio, pp. 114-15" (p. 98n5). It is worth remembering that Greg's view was that where an actor's name glosses a character's name, we are seeing signs of a prompter reminding himself who was playing a minor part. Where, as here, we find the actor's name instead of the character, Greg thought that this was typically authorial but should only occur where it would matter to the dramatist who played the part, since minor parts that anybody could take would not concern the dramatist during composition (Greg 1955, 117, 142). The problem here is that the actors' names are instead of character names (so, consistent with authorial copy) but the roles are minor ones that anyone could play (something about which the author should not care), and Martin cites Greg's firm view that in the present case the names could not have come from Shakespeare's pen (p. 99). In this Martin stands against Greg, whom he sees contradicted by the fact that actors names in "extant dramatic manuscripts" and early printings of Shakespeare "are overwhelmingly hired men rather than sharers", in support of which assertion he cites a page of John Dover Wilson's 1952 Cambridge edition of the play. Wilson does indeed discuss the matter of actors' names, and disagrees with Greg about what they tell us, but he makes no such claim about "extant dramatic manuscripts" generally, confining himself to Shakespeare alone.

    The New Bibliographical 'mast' is broad enough to accommodate disagreements between Wilson and Greg, yet having pinned his colours to it Martin remains tentative: "If these traces and anomalies point to F being Shakespeare's working papers" then we should consider whether they were annotated in the theatre. Departing from Greg and previous editors of this play in thinking that they were not, Martin follows William B. Long's lead (1985) in deciding that unannotated papers could have been used in the playhouse "as an acceptable, 'finished' script" (p. 100). One is entitled to ask if Martin means by "script" what he earlier meant by "prompt book"--he uses the word "prompter" on the previous page--and one detects here a trend. As reviewed here last year, Gordon McMullan's Arden Shakespeare edition of Henry 8 used the expression "a score for a stage play" to avoid the problematic word 'promptbook'. Martin notes that F has none of the features that McKerrow claimed book-keepers added to their manuscripts--anticipatory calls for actors and properties, stage directions naming properties needed later in a scene, names assigned to character roles, and anticipatory entry stage direction (McKerrow 1931, 270-72) but in the same footnote (p. 100n2) Martin admits that Werstine (1990) has "questioned" the application of these criteria and advised taking each document individually. This is hardly an adequate description of Werstine's critique of New Bibliography: if Werstine is even just mostly right, we have precisely nothing to tell us what copy underlay a given printing assessed solely on internal evidence. For Martin the most economical explanation is that F represents "Shakespeare's draft papers" and that a fair copy of these were sent off to get the Master of the Revels's licence and become the "official promptbook" . Little tweaks that seem unShakespearian (such as "Speaking to Bona", "Speaks to Warwick" in 3.3) might have been added by the Folio editors for readerly clarity (p. 101).

    In his narrative of True Tragedy's text, Martin records that O was reissued in 1600 to make Q2 and then in 1602 Thomas Millington (publisher of The Contention of York and Lancaster and True Tragedy) transferred his rights to Thomas Pavier, who in 1619 had William Jaggard (later publisher of F) print both plays together as The Whole Contention betweene the two famous houses, Lancaster and Yorke (Q3). The copy for our play in this composite Q3 was "an edited copy of O" (about this Cox and Rasmussen disagree, as we shall see) with just one passage possibly altered with authority and the others changes occurring in the printing house (p. 104). O is about 1000 lines shorter than F, needs the same number of actors, has some dramatic alternatives that are arguably preferable to F, and some verbal "anomalies" that are hard to explain by revision or printing error. Edmond Malone thought O an earlier, non-Shakespearian, version of the play but in the twentieth century it was mostly held a report of the play better represented by F. The latter view has recently come in for criticism, and Martin thinks there is compelling evidence to support revision and reporting: ". . . True Tragedy is a memorially reported early version of the play that Shakespeare substantially revised as 3 Henry VI" (p. 105). What has traditionally been thought to clinch the argument for memorial reconstruction being the source for O is Peter Alexander's argument that it has a corrupted version of the row between King Edward and his brothers Gloucester and Clarence about the daughter-heiresses of lords Hungerford, Scales, and Bonville being married (with King Edward's consent) to Hastings, the Queen's brother, and the Queen's son respectively, rather than to Gloucester or Clarence, who, as the king's brothers, should come first. 3 Henry 6 gets it right, and followed the sources, while True Tragedy omits the important fact that it is the queen's relatives being preferred that irks Gloucester and Clarence, and True Tragedy names Scales (rather than his daughter) and has him married to the daughter of "Lord Bonfield". This looks like the kind of error someone might make in dim recollection, although Steven Urkowitz claimed that True Tragedy makes good enough sense on its own;  that the details are not historically correct does not make it a bad text. For Martin the important point is this name "Lord Bonfield", which appears in no sources but does appear in Robert Greene's George a Greene which was published in 1599 with a title-page claiming that it was performed by Sussex's men. We know from its 1594 title-page that Titus Andronicus  was owned by Derby's (also known as Strange's), Pembroke's, and Sussex's men, so this is a link to True Tragedy since the 1595 title-page claims it was performed by Pembroke's. Thus "Bonfield" appears in two plays with no historical connections but both performed "by companies [Sussex's and Pembroke's] who shared scripts and personnel", something we know from the evidence of the Titus Andronicus title-page and other company history. Hence the name Bonfield "is a non-authorial interpolation by players", which supports Alexander's theory of memorial reconstruction, although there is nothing necessarily surreptitious about this (pp. 108-9). Also, True Tragedy has "Edward, rhou [sic] shalt to Edmund Brooke Lord Cobham" (A7r) where F has "You Edward shall vnto my Lord Cobham" (TLN 353, 1.2.40). The sources do not give this man a personal name, only a title, and it is hard to imagine that F represents something removed from the play, for if the motive was to not offend the Cobhams the name could have been taken out altogether. Moreover, True Tragedy is incorrect: the man's name was Edmund not Edward Brooke, so the likeliest explanation (as Hattaway suggested in the New Cambridge Shakespeare edition) is that the personal name was added by an actor, perhaps to allude to the Lord Cobham of Shakespeare's time, William Brooke (p. 110).

    Further, albeit weaker, evidence for memorial reconstruction is the phenomenon of characters betraying knowledge they could not yet have at a given point in the play. An example happens near the end of 2.5 in True Tragedy (C3v) when Exeter enters in the middle of a battle and says "Awaie my Lord for vengance comes along with him" (my emphasis), which word "him" has no antecedent. In F, however, this line appears slightly differently ("Away: for vengeance comes along with them" my emphasis, TLN 1275) and continues an ongoing onstage conversation. Martin's explanation of what happened is unfortunately foggy: "O's entry was apparently changed so that Prince Edward preceded Exeter on stage" (p. 111). This is a badly-worded ambiguity and might mean that the printed text O was changed in some way or that something was changed to make the printed text O; John Jowett's notational shorthand (MSO meaning 'the manuscript underlying O', MSF meaning 'the manuscript underlying F') is ideal for dispelling such confusion. Martin's suggestion seems to be that the reporters making O failed to have the Queen, Prince Edward, and Exeter enter as a group and instead had them enter successively, and then in response to this change the reporter(s) adjusted "them" to "him" because Exeter's line now responds only to what the Prince has just said, which would be about just one man, Warwick ("him"), whereas previously Exeter was responding to what the Prince had just said about Warwick and what the Queen said about Edward and Richard (thus "them"). The reporters then omitted Prince Edward's comment on Warwick so that Exeter's comment has no antecedent. I find this inherently implausible: the reporters change a line of dialogue ("them" to "him") to suit a change in stage direction, which is fairly fussy of them, and then they fail to notice that they have produced nonsense because another line of dialogue has been omitted. Alternatively, suggests Martin, F simply revises O, and then we have still the problem of Exeter's gibberish in O. Of the same kind of weak evidence for memorial reconstruction is the moment in O's 5.1 when Richard of Gloucester advises against entering the gates of Coventry in pursuit of Oxford's troops ("Weele staie till all be entered", E1v) in language that suggests that he knows that more (making up "all") are coming, which is in fact foreknowledge of the ensuing actions of Montague and Somerset. In F, by contrast, Edward simply cautions against going in because "other foes" (TLN 2741) may turn up, which phrasing Martin calls "strategically hypothetical" (p. 112). Finally, Martin observes that True Tragedy lacks the classical allusions that Shakespeare put in his other early plays, although of course one might argue that they were simply added to F as part of a process of authorial revision.

    The relationship between O and F is so complex, Martin argues, that it cannot be explained solely by memorial reconstruction or revision, rather both must be operating. Twentieth-century scholars who went beyond the theory of simple piracy as the reason for memorial reconstruction argued that O represents an abridgement for touring with fewer players, but they never quite agreed on how to do the calculations of doubling. Martin's calculation of the doubling shows that O and F need the same personnel: 13 men and 4 boys, plus a couple of non-speaking walk-ons. Thus a rationale for abridgement (to save parts) falls although one might still argue that O represents abridgement for shortened playing time. But O does not do its cutting simply, rather it is full of "complex rearrangement of scenes and lines" that seems oddly roundabout if the desire was just to save time. In some respects O actually expands on F (including having stage directions derived from the sources), so we cannot just say that O represents a badly-remembered F, nor that O represents a heavily censored text since it retains surely the most censorable event, the disinheriting of Henry's son. Martin supports Malone's conjecture that Shakespeare went back to the play that we know from True Tragedy and amplified it to make the play we know as 3 Henry 6 (p. 115). In this he follows other twentieth-century critics, but where they merely applied subjective criteria--True Tragedy being good enough to stand on its own and not merely a bad report--or were simply expressing poststructuralist dissatisfaction with New Bibliography, Martin thinks he has something more tangible to base his argument upon. By comparing O and F's dependence on Holinshed and Hall he attempts to show a pattern of authorial revision (p. 117). For this he uses three examples and promises more in an essay called "Reconsidering the texts of The True Tragedy of Richard Duke of York and 3 Henry VI" forthcoming in Review of English Studies. Since publication of Martin's edition the essay has appeared, although under the more definite title of "The True Tragedy of Richard Duke of York and 3 Henry VI: Report and revision" and it will be reviewed here next year. Martin's first example is in 2.1 where, between the towns of Wakefield and Towton, Warwick sizes the opposing Yorkist and Lancastrian forces: in O it is 50,000 Lancastrians versus 48,000 Yorkists, while in F it is 30,000 Lancastrians versus 25,000 Yorkists. Martin compares these numbers to those given in the sources. For the battle of Towton, Holinshed and Hall agree on 60,000 Lancastrians versus 48,660 Yorkists while for the second battle of St Albans Holinshed says 20,000 Lancastrians versus 23,000 Yorkists, while Hall mentions only the 23,000 Yorkists. Thus for scene 2.1, O seems to be getting its numbers from the battle of Towton while F gets its from the second battle of St Albans. So much seems clear, but Martin goes on "Thus it seems that O, with its figures linked to Towton, followed Hall, whereas F followed Holinshed . . ." (p. 118), which is a claim I cannot fathom since both sources report on both battles. In any case, the numerical correspondences do not see close enough to posit definite use of the sources and since an educated person would know something of the scale of the Wars of the Roses anyone might pick appropriate numbers unaided. Unfortunately, Martin's explanation of all this is tortuous.

    More clearly, Martin's second illustration from source use is that for the battle of Tewkesbury, O follows Hall in having Margaret and Prince Edward captured together, while F follows Holinshed in having Prince Edward captured separately. There are also in F a couple of pious lines from Margaret that might reflect Holinshed's unique report that she fled to a religious house. During the revision of the O version to make the F version, Shakespeare apparently turned from Hall to Holinshed, as he generally did with his history plays. Martin's final example is Clarence's return to the Yorkist side after supporting the Lancastrians for a while. Hall and Holinshed report that this was motivated by a "damsel, belonging to the Duchess of Clarence" persuading him of the unnaturalness of his actions, while Hall alone also reports Richard's agency in bringing his brother back over to the Yorkists' side, with whispered words, but reminds the reader of the damsel's prior work. O dramatizes Richard's agency (E1v-E2r) and gives him alone the credit for bringing Clarence back whereas Holinshed stresses instead Clarence's internal turmoil and his pretence to Warwick that he is still on the Lancastrian side, which is what F dramatizes. In F Clarence apologizes to his singular "brother" (that is, Edward ) for his betrayal whereas in O he refers to his "brothers" (Edward and Richard), thereby again stressing Richard's role as Hall does. In all, O seems influenced by reading Hall and F by reading Holinshed (p. 119-21).

    At this point Martin summarizes where we are ("Having established that O is a memorial report of an earlier version of the play which Shakespeare revised as F") and turns to F's use of O (or its derivatives) as printer's copy. McKerrow thought that the opening stage direction and first 18 lines of 4.2 in F were set directly from O or Q3, and Martin rightly comments that "none of the variants McKerrow cites is indubitably an error" (p. 122), although of course they are not "variants" but invariants, places where F follows what McKerrow thought was an error in O, for that is how you prove the dependence of one text upon another. For the Arden 2 edition, Andrew Cairncross went further and claimed that much of F was set from O. The 1986 Oxford Complete Works editors were sceptical of the McKerrow/Cairncross view but accepted that the F compositor might intermittently have glanced at Q3 and perhaps took a whole passage from it if his copy was not good. Martin, on the other hand, finds no evidence of Q3 being used in the printing of F and argues that one can explain the agreement of Q3/F against O/Q2 by "acceptable metrical variation, different chronicle details, and rewriting". Martin gives the example of George of Clarence saying in the Folio that Henry has passed a law "To blot out me, and put his owne Sonne in" to which Clifford replies "And reason too, / Who should succeede the Father, but the Sonne" (TLN 967-9). George's speech should, of course, be given to Edward (the son who has been blotted out) and Cairncross thought that it was the compositor following O (at least for the speech prefix) that caused the problem, for O has a different speech that does suit George of Clarence ("blot our brother out", B7v). In support of this view one can observe that at this point Q3 and F agree on some incidentals against O: the spelling "Parliament" (Q2/Q3/F) against "Parlement" (O) and the dividing of Clifford's next line ("And reason . . . the son") into two verse lines (Q3/F) rather than being a single long overlapped line (O/Q2). But Martin observes that F has a pleasing literary opposition of sons/fathers that O/Q3 spoils with its "blot our brother out" and that O/Q2/Q3 have Clifford say "And reason George" where F has "And reason too". If the F compositor was following Q3 at this point (rather than his manuscript copy), why did he changed "George" to "too" if his Q3 copy showed that George had indeed just spoken and hence "And reason George" (the Q3 reading) would be correct? No, Martin concludes, more likely F is a revised version of O and O's problems are those of "faulty reporting" (pp. 122-3).

    In seeking the dates of original composition and staging, of the reporting to make True Tragedy, and of Shakespeare's revision to make what got into F, Martin notes some fixed points (pp. 123-5). The play's composition cannot precede publication of Holinshed's Chronicles in 1587 nor be later than Robert Greene's death on 2 September 1592, since Greene famously alluded to a line from True Tragedy ("Tiger's heart wrapped in a player's hide"). The significances of three other dates are debatable: Henslowe's record of "harey the vj" being "ne[w]" on 3 March 1592, the plague closure starting 23 June 1592 (and lasting until 1594) and Thomas Nashe's allusion in Pierce Penniless to a performance of 1 Henry 6 by August 1592 ("brave Talbot . . . fresh bleeding"). We cannot be sure that Henslowe's "harey the vj" is 1 Henry 6 (as opposed to parts 2 or 3), but Roslyn Knutson showed that for multi-part plays, Henslowe consistently recorded if something were 'part two', so probably "harey the vj" is part one; Henslowe often neglected to state the first part number for a multi-part play. If True Tragedy was written after 1 Henry 6, then Contention of York and Lancaster and True Tragedy must have been written and performed between March 1592 ("harey the vj" being "ne[w]) and June 1592 (the theatre closure), which does not seem enough time. Just possibly, Greene got the "tiger's heart" line from a manuscript of the play, not from performance, so the plague closure is not relevant. Most likely is E. K. Chambers's explanation that Contention of York and Lancaster and True Tragedy formed a two-parter written before 1 Henry 6 (the prequel), but then the ownership of the different parts gets tricky. 1 Henry 6 we know was performed by Strange's men led by Edward Alleyn at the Rose (as Henslowe's Diary indicates) and True Tragedy's title-page says it was performed by Pembroke's men. Martin claims that Contention of York and Lancaster was definitely a Pembroke's men's play too even though its title-page is silent on the matter, and does not indicate why he thinks so (p. 126). Andrew Gurr showed that Pembroke's men was created to fill the Theatre when Edward Alleyn rowed with James Burbage in May 1591 and took his company of Strange's men away from the Theatre. Initially Pembroke's were successful, playing at Court during Christmas 1592 and 1593, but they failed in their provincial tour of summer 1593 and pawned their apparel and playbooks. Shakespeare seems to have retained control of his plays, since they ended up with the Chamberlain's men, as did he. So, who reconstructed True Tragedy from memory? It could have been Pembroke's men, if Shakespeare had the play in his possession and did not go on their provincial tour of 1592-3. The Titus Andronicus title-page suggests a traffic in playbooks and personnel from Strange's to Pembroke's to Sussex's, so alternatively those of Pembroke's company who did not move on to another one after its collapse in August 1593 might have tried touring, perhaps joining up with "a downsized Strange's" and/or Sussex's. Martin claims that the fact of the True Tragedy title-page mentioning only Pembroke's (not Strange's or Sussex's) suggests that a hard-up regrouping of Pembroke's "made the report sometime after August 1593, which they subsequently published in early 1595" (p. 127). Again, Martin's logic defeats me: why does mentioning only Pembroke's on the True Tragedy title-page suggest this? If they were poor in August 1593 and recollected True Tragedy to make some money from a printer, why wait until 1595 to get it printed? The argument here is too compressed even for a specialist to follow.

    In all events, True Tragedy was written before 1 Henry 6 opened in March 1592 and if written before May 1591 (the creation date of Pembroke's men) then it was most likely written for Strange's, with Edward Alleyn as "bigboond" Warwick (O, E3r), or if after then for Pembroke's. That True Tragedy was in performance by 1591 is suggested by its being echoed in The Troublesome Reign (published 1591) and by its echoing of Spenser's The Faerie Queene (published 1590). But what of the objection that 1 Henry 6 just feels like its early, less accomplished, work, too 'rough' to be a later-written prequel to Contention of York and Lancaster and True Tragedy? We can get around that by saying that 1 Henry 6 is not all by Shakespeare, and multiple authorship would also explain its link with Strange's (the "harey the vj" is definitely for Strange's at the Rose) but not Pembroke's; Martin gives the analogue of the multi-authored Sir Thomas More which belonged to Strange's (pp. 128-9). The final remaining question is 'when did Shakespeare revise the play to make the F text?' (pp. 130-2). The names of actors in the Folio texts of 2 Henry 6 and 3 Henry 6 help: John Holland and George Bevis appear in the former--although as reviewed last year, Roger Warren believes that Holland is a name from the sources, not an actor, and Bevis is the mythical figure (2000)--and "more certainly" (presumably a nervous glance at Warren's view, although his article is not cited) there are Gabriel Spencer, John Sincklo, and Humphrey Jeffes in 3 Henry 6 (p. 130). Spencer's death on 22 September 1598 gives us a terminus ad quem for the manuscript underlying Folio 3 Henry 6. From the minor parts played by Spencer and especially Jeffes (who we know of as an Admiral's men sharer later) we may guess that the manuscript is relatively early, else they would have bigger parts. Holland and Sincklo are in the plot of 2 Seven Deadly Sins "which was performed by Strange's Men at the Curtain around 1590, and certainly before 1592", for which claim Martin cites work by Greg, Gurr, and Scott McMillin and Sally-Beth MacLean. Actually, this is not certain and McMillin seriously entertained the possibility that 2 Seven Deadly Sins might be as late as 1594 (McMillin 1988) and in a forthcoming paper David Kathman dates it to 1597-8 on the basis of biographical knowledge about the actors named in the plot. Spencer, Sinclo, and Jeffes came together in the Chamberlain's men in 1594, so that is the earliest date of the revision that made the version of the play we know from the Folio, and the terminus ad quem is provided by the uncensored reference to Lord Cobham (discussed above), which must precede the controversy over 1 Henry 4  in 1596. There is also some evidence in the expansion of Margaret's oration in 5.4 that Shakespeare was reading Arthur Brooke's The Tragical History of Romeus and Juliet as he revised the play, which would make it roughly contemporary with Romeo and Juliet written in 1595 and a hint of Richard 2 confirms 1595 as the likeliest year.

    Having described precisely what he thinks of the materials he is working from, Martin is able is give a pleasingly crisp description of his editorial procedures: his edition is based on F, the expansion and revision of the play reported in O. F's variant passages are followed except in a few cases of "error, omission, or indispensable clarification", and where's O's stage directions are simply significantly different but not essential, they are merely recorded in the collation (p. 133). (Being post-theatrical, O's stage directions perhaps offer insights about how particular matters were settled in the theatre, but the revision to make F diminishes this value.) Martin's edition uses the Oxford Shakespeare's broken brackets for "plausible but debatable or ambiguous" stage directions and follows Stanley Wells's well-known rules on modernizing spelling. Because Martin thinks the Folio text substantially different from the O text, he "reluctantly" uses the Folio's title. I defer an examination of Martin's choices regarding particular textual cruxes until the introduction to Arden Shakespeare edition has been described so that the differing choices of the two editions can be compared directly. To conclude on Martin's Oxford edition it remains only to note his appendices. Appendix A is a "Commentary on Historical Sources" (pp. 327-56), a study of how the play relates to what is described in Hall and in Holinshed keyed to the line-numbers in Martin's text, so it is a set of commentary notes, more full than could be got onto the pages of the main text. Martin's comments are about the differences in the narratives as well as the literary qualities of what Shakespeare does with his source material. Appendix B "Montague" (pp. 357-60) is about how this character relates to two historical figures, Warwick's father and Warwick's brother, and to the character of Salisbury in 2 Henry 6. In Appendix C "Casting Analysis of 'True Tragedy' and '3 Henry VI'" (pp. 361-78) Martin builds on David Bradley's From Text to Performance (1992) and T. J. King Casting Shakespeare's Plays (1992), but he disagrees with their view that older boys could not double in non-speaking roles such as drummers, flag-carriers, and soldiers. Martin's view that they did perform such "hack work" (as Greg called it) comes from the plots of The Battle of Alcazar and Orlando Furioso. Martin reckons that O needs 15 men and 4 boys, while F needs 13 men and 4 boys, but in fact this is effectively the same thing because the two extras in O could be just walk-ons. In Appendix D "Queen Margaret's Tewkesbury Oration" (p. 379) Martin reprints the bit of Brooke's Romeus and Juliet from which Shakespeare took this speech, and Appendix E "Alterations to Lineation" (pp. 380-2) needs no explanation.

    The division of labour in John D. Cox's and Eric Rasmussen's Arden Shakespeare edition of 3 Henry 6 is made explicit: Cox thanks Rasmussen for editing the text and writing the textual introduction and textual notes (p. xv) and Rasmussen thanks Cox for "overturning centuries of editorial tradition by pointing to overlooked analogues that render emendation unnecessary and the Folio eminently defensible" (p. xvii). The style of Cox and Rasmussen's long introduction (176 pages) is quite unlike Martin's for the Oxford Shakespeare and unlike other Arden Shakespeare editions, for they set out to tell the story of "written engagement with the play (at least in English) from the earliest comment to the latest" (p. 4), surveying the reception rather than giving a reading of the play. Theirs is a huge undertaking, and some aspects of the reception (such as feminist criticism) are only sketched in. On the matters of Henslowe's receiving £3 16s 8d for "harey the vi" which was "ne[w]" on 3 March 1592 and Nashe's Piers Penniless being entered in the Stationers' Register on 8 August 1592, with its reference to "brave Talbot" bleeding again the English stage, and on Greene's Groatsworth being entered on 22 September 1592 with its allusion to True Tragedy, Cox and Rasmussen are on familiar ground surveyed above (pp. 5-6). O's stage directions are fuller than F's, although it is a third shorter overall, and call for use of the stage balcony ("on the walles" E1r). Cox and Rasmussen are unaccountably confident that the play was first performed at Henslowe's Rose and for details of its design they rely on Christine Eccles's flawed book 1990. Cox and Rasmussen reproduce a picture (p. 8) of Jon Greenfield's model of the first Rose (1587), which shows the theatre having no stage cover, but as discussed in last year's review, an erosion line one foot in front of the foundations of the Rose's stage (uncovered in 1989) clearly indicates water running off a roof over the stage. Like Randall Martin for the Oxford Shakespeare, Cox and Rasmussen think that Jonson's mockery of "York and Lancaster's long jars" being staged with "three rusty swords" is, like Sidney's criticism, a misapplication of Italian Neoclassicism to the English stage (pp. 9-10), but oddly they cite Jonson in original spelling ("iarres") although they modernized Henslowe's "harey the vj" to "harey the vi" on page 5.

    Cox and Rasmussen chart the stage history of the play from its first performances to the present, in particular via John Crowne's Restoration adaptation The Miseries of Civil War (pp. 12-14) and then nineteenth- and twentieth-century revivals that returned more to less to Shakespeare. Like Martin for the Oxford Shakespeare, Cox and Rasmussen reproduce photographs from notable twentieth-century productions, but they also devote nearly half a page to the picture of a horribly injured skull recovered from the site of the battle of Towton (p. 24), rather tenuously linked to the Clifford's death from an arrow in his neck. Cox and Rasmussen report on the 1999-2001 Royal Shakespeare Company production of 3 Henry 6 in its This England cycle but neglect to use that label, so adding to the difficulties of future theatre historians. This production ended with the Yorkists in the final scene walking over a stage covered with Henry 6's blood and Cox and Rasmussen comment that "Shakespeare's occasional pun on 'guilt' and 'gilt' has never been rendered more graphically" (p. 32). Without further explanation this remark is cryptic: what is the link between blood and gold-plating? Over-egging their critical pudding, Cox and Rasmussen claim that an amateur production of 3 Henry 6 by slave descendents on the Honduran island of Roatan in 1950, reported by Louise Wright George, "undoubtedly staged the most radical version" of the play, and in blissful ignorance of Bertolt Brecht and Jann Kott (p. 40).

    Cox and Rasmussen are careful to separate the question of whether Shakespeare alone wrote 3 Henry 6 from the question of O representing a memorially reconstructed version. Edmond Malone took the view that Greene's charge against Shakespeare ("beautified with our feathers") was one of plagiarism, and hence that Shakespeare had rewritten an existing play by George Peele, and that this is why the first printings of Contention of York and Lancaster (1594) and True Tragedy (1595) name Pembroke's men rather than the Shakespearian company of Chamberlain's/King's men. The memorial reconstruction hypothesis of Peter Alexander and Madeleine Doran provided a different way to explain True Tragedy's inferiority to 3 Henry 6, but it does not directly bear on the matter of authorship. Cox and Rasmussen think that the authorship question might be insoluble since even style detection by computer analysis is thrown off by "variations in orthography and typography and poor proofreading of early printed texts" (p. 47); they might have added also the problem of one writer imitating another's style. On the matter of computerized stylometry, Cox and Rasmussen report as though factual Don Foster's objections to the work of the Shakespeare Authorship Clinic at Claremont McKenna College, specifically the failure to "commonize" (regularize in matters of incidentals) the electronics texts used, and they give an over-generalized explanation of this procedure: "As a basis for accurate computer analysis, texts need to be 'commonized', i.e. rendered identical in textual accidentals such as spelling, punctuation and word breaks . . ." (p. 47n2). This may be true for some of the linguistic tests one may want to apply, but clearly not for tests that rely on idiosyncrasies of spelling, punctuation and word breaks.

    As part of their survey of the play's reception, Cox and Rasmussen offer potted histories of a number of 'criticisms': "Moral" (pp. 49-64), "Character" (pp. 64-81), "Historical" (pp. 81-113), "Psychoanalytic" (pp. 113-7), "New" (pp. 117-35), "Performance" (pp. 135-40), and "Feminist" (pp. 140-8). Several of these sections are too brief to be of use, but in the first the editors offer something of their own reading: sidestepping Tillyardism, they claim that 3 Henry 6 exhibits "'magical' thinking", a belief in the power of "spells, incantations, curses and blessings", in "prophecies, omens, 'prodigies',  oaths and swearing". About this the play is deeply ambivalent, and although such "'oppositional thinking'" (God/Devil, Good/Evil) was not done away with until the Enlightenment it could in Shakespeare's time be challenged by scepticism. This challenge was always ultimately futile since there was nothing to replace 'magical' thinking (pp. 57-9). Binary oppositions fused magical and moral thinking and were in the service of monarchial dynasties, but more than anything else Protestantism undermined 'magical' thinking from within by its "miracles are ceased" principle manifested in rejection of transubstantiation and exorcism (p. 60) This is a kind of deconstructivist reading, although the editors do not openly identify the self-destructing binary opposition in structural terms even after using the expression "complementary oppositions" (p. 59). The history plays articulate the crisis in 'magical' thinking, Cox and Rasmussen claim, and although there is some providentialism in the Henry 6 plays, much is not providential but man-made. The section on "Historical Criticism" opens with the surprising claim that "A 'turn to history' marked criticism and critical theory since the 1980s, as a reaction against the 'linguistic turn' of deconstruction" (p. 81). Deconstruction is as much a philosophical as a linguistic practice and one more properly seen in alliance with New Historicist and Cultural Materialist thinking than against it. Cox and Rasmussen use the notion of intertextuality to argue that establishing Biblical allusions and sources for Shakespeare is a fraught business since his culture was "saturated with the Bible" and we might easily mistake him getting something directly from there that actually came from another area of contemporary culture such as other plays or prose writings (pp. 88-90). Thus they do not emend the Folio's "Let me embrace the sower Aduersaries, / For Wise men say, it is the wisest course" (TLN 1422-3) because there is a Biblical analogue for it as it stands: "Agre with thine aduersarie quickely" (Matthew 5.25) which appears in the Geneva Bible and the Book of Common Prayer; Henry we know is carrying a prayerbook (p. 91). It is for this sort of supporting evidence enabling retention of Folio readings that Rasmussen thanks Cox for (p. xvii), although one must observe that the claimed analogy is not close.

    Establishing the play within the context of Shakespeare's early career, Cox and Rasmussen point out that it is "second only to Titus Andronicus in the number of words with the root 'venge'", which is "probably" a sign of Seneca's influence (p. 96). Such claims should always be accompanied by a statement of which texts were used to do the word-counting (or which concordance, if that is the source), and moreover rank order is not always as revealing as the raw data it conceals. Using the electronic edition of the Oxford Complete Works, I count Titus Andronicus having, at 43, nearly twice as many words based on 'venge' as 3 Henry 6, which has 23. The third place goes to Richard 3 at 20, as one might expect, but in fourth place is Cymbeline at 19, ahead of fifth-placed Hamlet at 18. The link between frequency of 'venge' words and Senecan influence does not seem quite as clear in the light of this evidence. Stylistically the joining of the separate labours of Cox and Rasmussen is largely seamless, but because their introduction is written to be readable as discrete sections there is necessarily repetition between them, and a point about critical prejudice against Tudor morality plays is made several times. Indeed, an entire inset quotation from Philip Brockbank's seminal essay "The Frame of Disorder" is produced on pages 64 and 124. This militates against a 'through-line' of argument and it is disconcerting to be told that Richard of Gloucester "is based on the morality play figure called the Vice" on page 106 of an introduction that has been referring to the Vice figure since page 78. Congruent with the editors' slightly shaky comments on recent literary and philosophical theory is their misuse of the word 'over-determined' to mean 'trying too hard' or 'forced' (". . . [Richard] Simpson's reading [that 3 Henry 6 is about 1580s politics] seems arbitrary and over-determined", p. 110) rather than in its proper sense of 'having more determining factors than the minimum necessary'. Especially weak is the section on "Psychoanalytic Criticism", which claims that Freudianism "has strong affinities with the inclination to see the human psyche as transcendent and homogeneous across cultures" (p. 113). This does Freud an injustice, since he was much concerned with how specific cultural forms make us unwell and his theory of the conscious/unconscious split is precisely the opposite of a homogeneous human psyche. Much better is the section on "New Criticism" that Cox and Rasmussen convincingly claim incorporates the 'metatheatrical' criticism of Anne Righter, James Calderwood, and John Blanpied in which the 'governing theme' (what New Critics always look for) is always the same: artistic creation itself. They approvingly cite, with a few reservations, Richard Levin's critique of this approach which pointed out that if every play is about artistic creation, we might as well all pack up and go home for the critics' work is done (pp. 130-4).

    Cox and Rasmussen begin their section on "The Texts of The True Tragedy and 3 Henry VI" (pp. 148-77) with a couple of useful summaries: a list of all the places where in editing F they have made "judicious use" of O, and that, although it is not conclusive, they intend to present evidence against the view that O is based on a memorial reconstruction and that F was printed from authorial papers (pp. 148-9). As is usual with this third Arden series, the text not used as the basis for the edition is quite superbly reproduced in facsimile at the back of edition. Cox and Rasmussen's departures from the editorial tradition begin with their assertion that Q2 (1600) is not an exact reprint of O: dozens of irregularly divided verse lines in O are relined, properly put back into verse in Q2 (p. 151). Likewise, Q3 (1619, the "Whole Contention" edition of Contention of York and Lancaster and True Tragedy) was identified as a reprint of O by Greg but Cox and Rasmussen have found 32 places where Q3 follows Q2's lineation rather than O's. This could happen by independent relineation--after all, the verse was there to be recovered--but "the simpler explanation" is that Q3 was reprinted from Q2 (p. 153). As the Q3 title-page claims, it is indeed "newly corrected" (there are nearly 300 substantive variants from O/Q2) and "enlarged" (Contention of York and Lancaster gains 11 new lines, True Tragedy gains one), although the authority for these changes and additions is disputable. Cox and Rasmussen dispute a claim about space-wasting in Q3 made by the editors of the Oxford Complete Works of 1986 (Wells et al. 1987, 205): signature Q3v does have a couple of extra lines of dialogue, but Cox and Rasmussen wonder why, if these were compositorial padding (as the Oxford editors have it), the man did not just wait until finishing the next page (Q4r), which completed the inner forme, and then see what needed to be done (p. 155). One answer might be that he feared cumulative error making the situation even worse by then. More clear is the case of putative expansion on Q4r that Cox and Rasmussen rightly observe makes no new lines so "can have nothing to do with problems of casting off" (p. 156). The agency and authority of Q3's variants are important in connection with the link between Q3 and F, and Cox and Rasmussen give the Hinman/Blayney compositor attributions for F, divided between compositor A and B for whom they conveniently list the respective Folio signatures, Folio page numbers, and corresponding act, scene, and line numbers in their edition. F and Q3 were printed in the shop of William Jaggard and Folio compositor B worked on both. As Cairncross pointed out, there are frequent F/Q3 agreements against O/Q2, one of which (O: "Henry and his sonne are gone, thou Clarence next" sig. E7r; Q3: "King Henry, and the Prince his sonne are gone" sig. Q3v; F1: "King Henry, and the Prince his Son are gone" TLN 3165) shows such strong F1/Q3 linkage that we should think that "Q3 may have been used in some way by the Folio compositors" (p. 157). To be convincing, a claim of dependence should ideally rest on agreement in error and since all three versions of the line are acceptable it is possible that each printing represents the reading of its underlying manuscript, which manuscripts differed for some reason.

    Cox and Rasmussen agree with the editors of the Oxford Complete Works that there is evidence for no more than the occasional consultation of Q3 in the setting of F, yet they offer a most surprising summary of the bibliographical stemma of textual descent: ". . . the first edition of 1595 (O) was reprinted with some minor changes in 1600; the second edition (Q2) was then reprinted with further revisions in 1619; this third edition (Q3) was then reprinted in a substantially revised form in 1623 (F1)" (p. 158). This stemma ignores the manuscript(s) entirely and describes only how the printed texts are related, and having just announced that ". . . Q3 was probably consulted only occasionally by the F compositors" it is bizarre to then use the phrase "was reprinted . . . in" to describe the Q3/F relationship. The hypothesis that O was printed from a memorial reconstruction of the play is based "rather precariously" on a single variant passage (4.1.47-57) about the matching of the heiresses of Hungerford, Scales, and Bonville to Hastings, the Queen's brother and the Queen's son. This is mangled in O but accurately follows Hall's Chronicle in F, which fact is the "linch-pin" of Alexander's argument that Cox and Rasmussen attempt to remove (pp. 161-3). There are, in fact, "significant orthographic correspondences" between True Tragedy and Hall's Chronicle in the spellings of Penbrooke, Norffolke, Fawconbridge, and Excester, whereas F has Pembrooke,  Norfolke, Falconbridge, and Exeter, and other O/Hall agreements against F include Warwick saying his men number "48. thousand" (O), "48,600" (Hall), but "fiue and twenty thousand" (F TLN 839, 2.1.180), Henry's having reigned for 38 years (O and Hall) rather than 36 years (F TLN 1831, 3.3.96), York retreating to Sandall Castle in Wakefield (O and Hall) which is not unspecified in F (TLN 235, 1.1.206), and his recollection of battles in Normandy (O and Hall) but in France (F TLN 395, 1.2.72). Less convincingly, there is a link between the phrasing of Hall's account of Warwick's landing in 1470 ("crying 'King Henry! King Henry! A Warwick! A Warwick!") and O ("All. A Warwicke, a Warwicke") against F ("They all cry, Henry" TLN 2216, 4.2.27) and O's order of scenes 4.4 and 4.5 is historically correct while F has the chronology reversed. In short, there is much in O that "could not have been derived from the version of the play preserved in the Folio text" (p. 163), and hence the hypothesis of memorial reconstruction receives another blow; Cox and Rasmussen do not entertain the possibility that simple revision might also separate O and F to account for these differences.

    The idea that memorial reconstruction might be done because the authorized book was left behind when a company toured was "effectively undermined" (p. 163) by Werstine's pointing out the entry in the Hall Book of Leicester dated 3 March 1583/4 which reads "No Play is to bee played, but such as is allowed by the sayd Edmund [Tilney], & his hand at the latter end of the said booke they doe play". One might respond that this evidence is open to the usual law of ambiguity in historiography: does a prohibition show that a thing never happened, because it was not allowed, or that it did happen, else why would anyone ban it? Formerly 'bad' quartos are indeed being critically rehabilitated as authorial first drafts, although the 1608 quarto of King Lear is a surprising example for Cox and Rasmussen to mention (p. 164) since no-one has ever claimed it is 'bad'. The editors acknowledge Steven Urkowitz's argument that O is an original authorial version that was later revised to make the play represented by Q, but "are cautious about advancing this conclusion since other explanations are certainly possible" and they include amongst the other explanations the use of Hall by those making the memorial reconstruction. Laurie Maguire has "puckishly" pointed out that the repetition of essentially the same lines in O ("For strokes receiude . . . I rest my selfe" C1r, "For manie wounds receiu'd . . . I yeeld to death" E2v) is traditionally attributed to memorial reconstruction (the reporter unintentionally anticipating himself) while the same phenomenon in Q2 Romeo and Juliet ("O true Appothecary . . . with a kisse I die" L3r) is attributed to authorial false start, and Cox and Rasmussen agree that neither 'proves' anything (p. 165). Actually, these are not the same thing at all since the repetition in O occurs across a gap of about 1800 lines while in Q2 Romeo and Juliet the repetition is on the same page. The former could have got into print without anyone failing to delete anything, since the 'error' is not easily noticed, while the latter must involve someone's failure to delete the repetition since no-one produces such a thing intentionally. To put it another way, it is exceedingly difficult to make a single hypothesis that covers both cases: a false start followed by failed deletion works for instance in Romeo and Juliet but not True Tragedy, anticipation by a reporter is good for True Tragedy but not Romeo and Juliet. Contrary to Maguire and Cox and Rasmussen, Greg did not claim that these things "'prove'" (Maguire's scare quotes) what kind of copy was used in the printings, only that they were "characteristic" of different kinds of copy. Cox and Rasmussen admit the evidence of memory in O's apparent aural garbling of words that F seems to have got right: "Wrath makes him death" (O B1r), "Wrath makes him deafe" (F TLN 513, 1.4.53); "his adopted aire" (O B2r), "his adopted Heire" (F TLN 561, 1.4.98); "Tygers of Arcadia" (O B3r), "Tygers of Hyrcania" (F TLN 622, 1.4.155); "Sore spent with toile as runners" (O C1r), "Fore-spent with Toile, as Runners" (F TLN 1057, 2.3.1); "the litnes of this railer" (O E5r), "the likeness of this Rayler" (F TLN 3013, 5.5.38); "Cyssels" (O E7v), "Sicils" (F TLN 3210, 5.7.39) (p. 165). Hence Cox and Rasmussen, although "dubious about the theory of memorial reconstruction by touring actors" do not think anyone can explain these homonymic errors saying that O was printed from authorial copy. Instead, they find "more plausible" the view put forward by Blayney with support from Humphrey Moseley's preface to the Beaumont and Fletcher Folio of 1647 that memorial reconstruction was done by actors to give private transcripts to their friends (p. 166).

    Like Martin editing the Oxford Shakespeare edition of 3 Henry 6, Cox and Rasmussen think they have caught Greg in a contradiction in The Shakespeare First Folio: "[occasionally] the substitution of the name of an actor, when the part is written with a particular performer in view" shows that the copy was foul papers, and yet of the names of Gabriel Spencer, John Sincler, and Humphrey Jeffes appearing in F 3 Henry 6 Greg writes: "[In no case is it of the least consequence who took these minor parts and their assignment] cannot possibly be attributed to the author" (p. 167). This does indeed look like an "internal contradiction" in Greg's writing if one omits, as Cox and Rasmussen do, the words I have placed in square brackets. By selective quotation (dropping the qualification) Cox and Rasmussen make Greg definite where he was tentative ("occasionally") and they ignore his earlier explicit claim that there "are two ways in which actors' names may find their way into dramatic manuscripts" (1955, 120 my emphasis), from the author's pen and from the prompter's. What matters, Greg claims, is how important it is that a particular man plays the part and how big the part is, for an author will care (and write about it) if it does matter dramatically, while a prompter will want to know who is doing it either way, especially if it is a minor part he might otherwise forget the casting of. Here it does not seem to matter if the particular named men are used (Sincler's famous thinness, for example, is not exploited in the scene) so it would seem to be a matter for the prompter not the author. After tracking the genesis of explanations about actors' names in printed plays for several pages (pp. 167-71) Cox and Rasmussen declare that the whole question has "little relevance to the editing process" (p. 172) and speculate whether there are not actors' names after all. Gabriel would be a good name for the divine messenger of 1.2 and Humphrey too might be fictional name, although they concede that "Sinklo" is a harder case to argue. (An obvious retort to this speculation is that even if these are fictional names, no-one utters them in performance so any aptness would be lost on an audience.) Cox and Rasmussen support the recent rejection of McKerrow's 'suggestion' that variant names in speech prefixes and stage directions show authorial copy, citing the usual work by Werstine and Long and also Richard Kennedy (Kennedy 1998). Generally, modernized editions regularize speech prefixes but there are special cases--aristocratic and monarchial titles get lost and won in the plays--and Cox and Rasmussen think that Lady Grey's speech prefix "Wid[ow]" in 3.2 is such a case: she is known to the audience only as a widow at this point and her widowhood is what makes Edward interested in her. Thus they break "with three centuries of editorial tradition by retaining the WIDOW SPs in 3.2" (p. 175). This smacks of caprice and taken further the same logic could change dozens of speech prefixes in modern editions bringing no great advantage and much confusion; a Hamlet without Claudius (just "KING") would be the next step, one supposes. Cox and Rasmussen conclude with their view that the compositional priority of True Tragedy and 3 Henry 6 is unknown, as are the natures of the underlying manuscripts (pp. 175-6), but one is then left wanting to know how they came to the decide on the latter as their base text, even if it was only the tossing of a coin. After their text, Cox and Rasmussen provide four appendices. The first is a facsimile of the 1595 octavo keyed to the line-numbers of their modernized text and to the Through Line Numbering of Charlton Hinman's facsimile of the Folio. Appendix Two is a doubling chart in which Cox and Rasmussen reckon that the play's 67 roles require 21 men and 4 boys, which is 8 more than Martin calculated in the Oxford Shakespeare. One reason for the difference is that Martin allows boys to double as soldiers and watchmen while Cox and Rasmussen have 8 men do nothing but silent soldiering with drums, trumpet, or colours. The third appendix lists the names of the battles depicted in the play, together with their dates, their outcomes, and where they appear in the play, and the fourth gives genealogical tables for the houses of Lancaster, York, and Mortimer.

    Regarding particular editorial choices, Randall Martin's Oxford Shakespeare 3 Henry 6 and John D. Cox and Eric Rasmussen's Arden Shakespeare 3 Henry 6 are here considered together. The play begins with a stage direction that brings on the Yorkist party, and for the Arden Cox and Rasmussen import to F the additional information provided by O that they have "white roses in their hats", and there is a similar detail for the entrance of the Lancastrians at Despite a long description of their views about the early texts, Cox and Rasmussen provide no rationale for borrowing stage directions from O and since F makes sense on its own, it is hard to see this other than as old-fashioned textual conflation; this borrowing recurs throughout the text and only interesting cases will be noted. Martin, on the other hand, declares his intention to ignore O and use only F except where his hand is forced by "error, omission, or indispensable clarification" (p. 133), so quite understandably he leaves O's detail about roses out of the opening stage direction and the subsequent one for the Lancastrians. At 1.1.19 Norfolk, observing the lopped-off head of Somerset, says "Such hope have all the line of John of Gaunt" and both editions take this from F without worrying as their predecessors have done that Shakespeare might have meant "Such hap" but "hope" got picked up from Richard of Gloucester's next line, "Thus do I hope to shake King Henry's head". Richard of Gloucester's encouragement to his brother Edward to live up to the family tradition of slaughtering for power is given in F as "For Chaire and Dukedome, Throne and Kingdome say" (TLN 749, 2.1.93) which makes reasonable sense (with "say" meaning 'declare yourself ambitious') and thus Martin uses it, whereas Cox and Rasmussen have the benefit of Richard Proudfoot's new and ingenious emendation to "'ssay" meaning 'make a successful attempt to gain'. Both editions use O's "idle thresher" where F has Warwick say "Or like a lazie Thresher with a Flaile" (TLN 789, 2.1.130), although only Cox and Rasmussen explain that "lazie" appears in the previous line too and must have been repeated by compositorial accident. Arden notes discussing "editorial amendations or variant readings" are supposed to be "preceded by *" (General Editors' Preface, p. xiii) but this one and many more in Cox and Rasmussen's editions lack the asterisk; I can detect no pattern in the few that do receive the mark. F's repetition of "lazy" makes sense and is defensible as the kind of imperfect language that suits the moment and the speaker (as Frank Kermode explored in Shakespeare's Language), so it is surprising that this should be thought a clear error needing emendation.

    Alone in the midst of battle, King Henry reflects on the quiet life: "So Minutes, Houres, Dayes, Monthes, and Yeares / Past ouer to the end they were created" (TLN 1172-3, 2.5.37-8). Cox and Rasmussen follow Rowe in inserting "weeks" between "days" and "months" because the preceding speech considers in turn the passing of minutes, hours, days, weeks, and years and because it fills out the metre, whereas Martin passes over the matter in silence. At 2.5.119 both editions use Dyce's "E'en for the loss of thee" instead of F's obviously faulty "Men for the losse of thee" (TLN 1257) and observe that the box holding types of the letter M lay directly below the one for letter E in the upper case used by compositors, so either the printer's hand went to M box by mistake or someone accidentally put an M in the E box. (Were the M box above the E box we could free the printers of all suspicion and blame gravity.) For Clifford's entrance, wounded, at the start of 2.6, Cox and Rasmussen borrow O's colourful stage direction "with an arrow in his neck" that follows Hall, while Martin eschews it as "perhaps faintly ludicrous", not quite what Hall has ("striken in the throte"), and unsuited to the action of the scene. It is worth observing that Cox and Rasmussen's long expression of the uncertainty regarding the nature and origin of the printers' copy for O and F gives them greater freedom to emend and conflate than Martin's more definitive account of the textual situation permits him. Before Clifford's line "And whither fly the gnats but to the sun?" (2.6.9) Cox and Rasmussen insert from O the line "The common people swarm like summer flies" that F omits, presumably thinking that the sense requires it. Martin lets F stand, points out that "summer flies" and "gnats" do not seem to be the same thing to Shakespeare (the former conjure up heat, the latter light), and persuasively argues that because "summer flies" are mentioned by Clifford 8 lines later in F, conflating O and F (as Cox and Rasmussen do) produces "lame repetition" not present in either early printing. The moment when the Yorkists find dying Clifford is almost the same on both editions. Martin follows F exactly in having "Clifford groans / RICHARD Whose soul is that which takes her heavy leave? / A deadly groan, like life and death's departing. / See who it is. / EDWARD And now the battle's ended, / If friend or foe, let him be gently used. / RICHARD Revoke that doom of mercy, for 'tis Clifford" (2.6.41-45), while Cox and Rasmussen move Edward's speech prefix back three words to give him the command "See who it is". Both editors resist arguments by C. J. Sisson, amongst others, that O's distribution of these lines gives markedly superior staging and both think that O's "Clifford grones and then dies" would rob the scene of the horrible abuse of a dying man by the Yorkists. Martin observes that although there is some difficulty in Richard answering his own interrogatory command ("See who it is . . . 'tis Clifford"), his giving orders in the presence of his brother (F's version) hints at his future ambition.

    Just before he is captured by the gamekeepers, F has King Henry say "Let me embrace the sower Aduersaries" (TLN 1422, 3.1.24). Martin points out that the stress falling on 'ver' sounds wrong (it should fall on the first syllable) and although 'sweet adversity' is proverbial he follows Sisson (1956, 84) in rejecting as unlikely a misreading of "aduersitie" as "aduersaries" and instead favours Pope's emendation to "adversities". Once the plural "adversities" is accepted, there is no reason to suppose that F's "the" is a form of the personal pronoun 'thee' and so no need to put a comma before it. Cox and Rasmussen make no emendation to F and rightly point out that "polysyllabic words often vary in emphasis in Shakespeare"; moreover they have what they believe to be a Biblical analogue for the line as it stands (Matthew, 5.25 as discussed above). Another of Richard Proudfoot's happy suggestions appears in Cox and Rasmussen's alteration of F's "Whom thou obeyd'st thirtie and six yeeres" (TLN 1831, 3.3.96) to "six and thirty years" on the grounds that it scans properly and the underlying manuscript might well have had a numerical "36" that was badly expanded by the compositor. Martin follows F here, as he does for Rivers's question to Queen Elizabeth "Madam, what makes you in this sudden change" (4.4.1) for which Cox and Rasmussen follow Collier's alteration of "you in" to "in you". Yet again we see the New Bibliographical Oxford Shakespeare editor being reluctant to emend if there is any hope of making sense of the base text, while the Arden Shakespeare editors, who offered lengthy reasons to be editorially cautious (since we do not really know much about the origins of O and F) and were scathing of New Bibliography, are in practice more cavalier in their interventions. Both editions reluctantly let stand F's version of the proclamation at 4.769-75, which rather awkwardly uses a common soldier to do the public reading, rather than following O which has Montgomery do it, and both reject out of hand the complicated theory offered by the editors of the Oxford Complete Works in which the soldier was invented by a compositor trying to make sense of Hastings's phrase "fellow soldier". There remains the problem that the proclamation ends awkwardly ("Edward the Fourth, by the grace of God, King of England and France, and Lord of Ireland, etc"), which neither edition can explain and which both think might indicate that the actor playing the soldier could be relied upon to fill in the rest, whether from a property document or common knowledge.

    Neither edition starts a new scene with the Folio's "Exeunt" of Henry's supporters at 4.8.32, Martin adding "all but Henry" and then bringing on Exeter to talk to him, while Cox and Rasmussen add "all but King Henry and Exeter", the latter already on since the beginning of the scene because they followed Capell in replacing Somerset (who has nothing to do or say in this scene) with Exeter in the opening stage direction. Before the Yorkists burst in on Henry there is in F a "Shout within, A Lancaster, A Lancaster" (TLN 2653, 4.8.50) that both editions retain, explaining that perhaps it is a Yorkist plot to confuse the Lancastrian guards or else the cry of the guards when then realize what is afoot. For the scene in which George of Clarence switches allegiance back to his Yorkist family, both editions have him accompany his line "Father of Warwick, know you what this means?" (5.1.81) with some business with a red rose: showing it to Warwick in Martin's edition, taking it out of his hat and throwing it at Warwick in Cox and Rasmussen's edition which takes its stage directions from O, which has a quite different version on the scene. In O (as in Hall), a parley with his brother Richard of Gloucester's convinces Clarence to change sides, while in F Clarence enters having already made this decision. Martin makes a case for F's version being superior because it focusses on the Clarence/Warwick relationship without the distraction of 'whispering Vice' Richard, while Cox and Rasmussen think O's version "more dramatic" in its portrayal of "Richard's rhetoric affecting a reconciliation". Finally, F has Somerset speak of Montague's dying voice "Which founded like a Cannon in a Vault" (TLN 2846, 5.2.44) and both editions keep "cannon" in preference to O's "clamor", Cox and Rasmussen silently ignoring the common emendation while alone Martin goes to the trouble of refuting McKerrow's somewhat wild suggestion (actioned by the Oxford Complete Works) that the word should be the musical term "canon".

    Only one monograph of relevance appeared this year, David Scott Kastan's 2001. Kastan begins with a conviction that stage and page are incommensurable, that performance makes a new thing rather than enacting an existing one: "Hamlet is not a pre-existent entity that the text and performance each contain, but the name that each calls what it brings into being" (p. 9). Thus we should not always think in a stage-centred way, for the stage tends to dehistoricize, making him our Shakespeare, everybody's Shakespeare, while print conserves him. Because Shakespeare clearly intended performances and seems to have no concern for his books in print, a recurrent theme in Kastan's book is the ontological and epistemological status of extant early texts of Shakespeare, and although he never quite settles these matters Kastan is sure of the falsity of G. Thomas Tanselle's distinction that the 'work' is the set of unrealized intentions that the 'text' only approximate. Rather, it is the materialization that makes the 'work' possible in the first place (p. 4). Kastan sets out to chart the entire history of Shakespeare in print, beginning with the seldom-noted facts that Shakespeare was a best-selling published author in his own lifetime and that as early as 1638 a Folio of Shakespeare was represented in an oil painting (Van Dyck's painting of Sir John Suckling); already his prestige was a matter of print, not performance (pp. 10-1). Necessarily in a short book (136 pages), Kastan's narrative moves at a breathless pace and tends towards generalization, but for the sixteenth and seventeenth centuries in particular there is a wealth of information newly synthesized into a compelling argument. Citing the famous Q1 Hamlet line "To be, or not to be, I there's the point" (D4v), Kastan points out that this seems like textual corruption only if one is expecting to find "that is the question" and that "I there's the point" is perfectly good language, indeed it provides a "moment of unmistakably Shakespearean power along the tragic trajectory of the play" when it occurs in the Folio text of Othello (TLN 1855, 3.3.232) (pp. 26-7). From the evidence of early dramatic playtexts Kastan constructs a convincing case for thinking that the market for printed plays was the playgoing audience and that only gradually during the period did the name of Shakespeare as author come to be as (and eventually more) important than the name of the performing company. The defining event, of course, was the publication of the 1623 Folio to which Kastan devotes a central chapter displaying great breadth of historical knowledge and a virtuosity of compression. (Just one slip: Kastan gives the wrong date, March 1597 (p. 54), for Lord Hunsdon's promotion to Lord Chamberlain that changed the name of Shakespeare's company back to Chamberlain's men; it was in fact 17 April 1597, as Chiaki Hanabusa recently pointed out (Hanabusa 1999).)

    Repeating an argument he made in Shakespeare After Theory (1999), Kastan argues against one of the one founding principles of New Bibliography, A. W. Pollard's distinction of the "stolne and surreptitious copies" mentioned in the Folio preliminaries from the "all the rest" in order to form two categories of pre-Folio publication of Shakespeare: the 'bad' quartos and the good. In a footnote to page 73 Kastan assigns this distinction to Pollard's Shakespeare's Fight with the Pirates (1920) but in fact it appeared 11 years earlier in Pollard's Shakespeare Folios and Quartos (p. 4). Kastan thinks that, as actors used to textual instability, Heminges and Condell would not have made such as strong distinction between the existing quartos, but since Kastan earlier argues that good editions were produced to replace ones perceived to be 'bad' this appeal to casual theatrical sensibilities seems weak (p. 74). In an article to be reviewed next year (2002) Lukas Erne argues that Shakespeare's fellow actors actively supported the publication of his plays, in which case there is little reason to suppose with Kastan that the Folio preliminaries dismiss all previously published Shakespeare as "maimed and deformed". After a tour of Shakespeare publishing from the seventeenth to the nineteenth centuries (pp. 79-110), Kastan returns to matters philosophical in his final chapter on the impact of electronic text (pp. 111-36), which Kastan values most of all because it defamiliarizes the textual medium, the codex having become so familiar to us that we easily overlook its conventions. Although he thinks recovery of authorial intentions is laudable, Kastan aligns himself with Jerome McGann's view that texts do not exist independently of the media that carry them rather than Tanselle's platonic view that textualizations are imperfect representations of unembodiable work, and Kastan's position is implicitly nominalist: "Hamlet is perhaps best considered not something in itself at all but, rather, the name for what allows us comfortably to consider as some metaphysical unity the various instantiations of the play . . ." (p. 133). Kastan's excitement over the possibilities of electronic text is leavened with a caution about the shift of power that the Worldwide Web brings as readers become "dependent upon technologies . . . [over which they have] distressingly little control" (p. 130-1). They cannot take away our books, he observes, but they can shut down the websites we use; true, but they cannot take our CD-ROMs either. Kastan gets a little carried away on the euphoria of textual copiousness of polymorphously self-connected hypertext, and takes up George Landow's argument that it realizes the textual jouissance promised by post-structuralism (pp. 125-7), but in fact in the case of Shakespeare all you need is a collection of about 80 electronic texts to encompass the entire pre-Commonwealth cache of printings. Kastan imagines a huge hyperlinked archive including the early printings, theatre reviews, film versions, etcetera, but one might argue that we already have such a thing, it is called a library. Marvelling at the possibilities raised by such projects as Peter Donaldson's Shakespeare Electronic Archive at Massachussetts Institute of Technology, Kastan wonders if they take away the need to edit at all, for unedited early editions "are the most compelling witnesses to the complex conditions of their production" (p. 123). The economics of the print medium have denied readers cheap versions of the early printings in facsimile and Kastan, modestly omitting to mention that he is a general editor of the Arden Shakespeare, complains that today's editions of Shakespeare are too much alike and engender "wasteful duplication of scholarly energy" (p. 124). Kastan's optimism that electronic text might offer new potentialities unavailable in the print medium is properly guarded, and amidst ever-changing technology he ruefully asks how many of us cannot open electronic documents of our own that we made more than 10 years ago (p. 131). (To be fair, this is a matter of individual failing since the computer support departments of universities around the world have always provided the right advice about keeping one's personal archive in a machine-readable form; they report that academics in the humanities tend to ignore the advice while those in the sciences follow.) A tiny flaw that indicates the electronic origins of Kastan's book itself is the persistent use of the wrong kind of apostrophe (a right-facing instead of a left-facing one) at the start of words that begin with elision, as in "'em" (p. 85) and "'s" (meaning "us", p. 109); Microsoft Word bossily enforces this change to prevent 'error'. 

    Shakespeare Quarterly and Shakespeare Survey contained no articles of interest to this review. Review of English Studies contained three, and they will be taken in order of appearance. In a companion piece to a previous article (1999), MacDonald P. Jackson reaches the same conclusion by different means, namely that sonnets 104 to 126 are Jacobean while the rest date from the 1590s (2001c). Jackson's evidence for this is Shakespeare's use of particular words at particular times in his career, for example the almost total absence of "goodness" and "particular" from the early works can help date sonnets that contain these words. Eliot Slater showed that Shakespearian rare-words (ones used more than once but fewer than 11 times overall in the canon) cluster in time and that the sonnets are lexically linked to the play's Love's Labour's Lost, Romeo and Juliet, Richard 2, A Midsummer Night's Dream, Much Ado About Nothing, and Henry 5 (that is, to the period 1595-9), and Gregor Sarrazin's much earlier work found the sonnets to be linked to the period 1593-98 (p. 60). Slater and Sarrazin took the sonnets as a whole, and Jackson thinks it much better to follow A. Kent Hieatt, Charles W. Hieatt, and Anne Lake Prescott whose work on the occurrence of early (pre-1599) and late (post-1599) Shakespearian rare-words in the sonnets showed them that he was working on the sonnets after 1598 and so caused them to divide the sonnets into 4 zones: Zone 1 (sonnets 1-60), Zone 2 (sonnets 61-103), Zone 3 (sonnets 104-126), and Zone 4 (sonnets 127-154). Hieatt, Hieatt, and Prescott decided that zones 1, 2, 4 were written 1590-5, 3 was about 1600, and 1 was revised in the seventeenth century, but their analyses were not finely grained: rare-words were for them just 'early', 'late', or 'both periods', and they did not provide enough information about the distribution of different kinds of rare-words in their control texts for comparison with the distributions in the sonnets (p. 62). For Jackson, an important category of rare-words is 'middle'--say, from King John (1595-6) to Macbeth (1606) and although Hieatt, Hieatt, and Prescott do not use this category, Jackson manages to extract data about it from their tables. Jackson's Table One shows the occurrences of 'early', 'middle', and 'late' rare-words for the sonnets zones 1-4, although it has a line-wrapping problem that makes it unnecessarily hard to read and its own footnote ends mid-sentence: "'Early' means found in" (p. 63).

    This refinement broadly supports Hieatt, Hieatt, and Prescott's earlier conclusions about the zones, but Jackson observes that very rare words tend to cluster more than averagely rare words, so that Sarrazin's category of words that occur only twice or thrice in the canon ought to be a highly sensitive indicator of chronology. Indeed it is: if one divides the canon into four chronological periods, the 'Sarrazin links' mostly confirm that the plays in each group belong together, as Jackson showed in his book Studies in Attribution: Middleton and Shakespeare (1979). Jackson's groups are:

(1) Titus Andronicus, 1 Henry 6, The Comedy of Errors, 2 Henry 6, 3 Henry 6, The Taming of the Shrew, Richard 3, The Two Gentlemen of Verona, Love's Labour's Lost, A Midsummer Night's Dream

(2) Romeo and Juliet, Richard 2, King John, The Merchant of Venice, 1 Henry 42 Henry 4, The Merry Wives of Windsor, Much Ado About Nothing, Henry 5, Julius Caesar

(3) As You Like It, Twelfth Night, Hamlet, Troilus and Cressida, Measure for Measure, Othello, All's Well that Ends Well, Timon of Athens

(4) King Lear, Macbeth, Antony and Cleopatra, Pericles (acts 3-5 only), Coriolanus, Cymbeline, The Winter's Tale, The Tempest, Henry 8 (excluding the Fletcher parts)

Jackson combines categories (3) and (4) to get a general index of 'lateness': calculate how many 'Sarrazin links' a given play has with this 'late' category (as a percentage of how many links it has to all the categories) and one should get a simple indication of how 'late' it is. If one puts all the plays in order of this ratio they come out pretty much in the chronology we know from Karl Wentersdorf's work. Jackson's explanation of his interpretation of the plays' ordering is somewhat compressed (he calls it "reading their position on the vocabulary order as a position on Wentersdorf's chronological order"), and I presume his procedure is as follows: one notes that play A occupies position B on the vocabulary list (the plays in order of their 'lateness' index), one looks to Wentersdorf's list for the play occupying position B, which we may call play C, and look to the date, D, assigned to it by Wentersdorf. The question Jackson appears to ask is 'how close to D is the true date of play A?', and he reports that for 31 of the 37 plays the answer is not more than 3 years out and for half it is correct to within a year (p. 65). This provides a benchmark for an undated work, since one can calculate its 'lateness' index (from its 'Sarrazin links') and then read off the date from the known chronology of the plays. By this method Venus and Adonis comes out at 1592-3 and The Rape of Lucrece as 1593-4, as we would expect, and A Lover's Complaint comes out with such a high 'lateness' index that it has to be seventeenth-century. Thus Jackson's new method confirms results found by other methods, and should be reliable.

    What of the sonnets? Jackson shows the 'Sarrazin links' tests for the 4-zoned sonnets against the 4-zoned plays as his Table Two, and the most important thing is that Zone 3 shows lots of links with Groups 3 and 4, while for the other sonnets zones links with the first two play groups predominate. The pattern of 'Sarrazin links' for the 'Marriage' sonnets (numbers 1-17) is 3, 5, 2, 2, showing a bulge of links with the second group of plays and hence suggesting that these were written after rather than before 1595. Jackson does some other tentative reading of the detail, but has little confidence in it; the important thing is that Zone 3 sonnets are most likely Jacobean not Elizabethan, and the others probably were written within 1595-9 (p. 66). Having done this analysis for Sarrazin's twice-or-thrice-used words, Jackson does it again with the Hieatt-Hieatt-Prescott rare words, which are not nearly so rare and hence not such a good indicator of date. Jackson counts how many sonnets in each zone have their highest number of links with play groups 1, 2, 3, and 4, in other words for each sonnet he records which of the four play groups it has most links with, and then assigns the sonnet to that group. The results are as follows, with the four numbers for each zone showing how many sonnets in that zone are most strongly connected with play groups 1 (Titus Andronicus to A Midsummer Night's Dream), 2 (Romeo and Juliet to Julius Caesar), 3 (As You Like It to Timon of Athens, and 4 (King Lear to Henry 8) respectively:

Zone 1 is 9, 12, 10, 7
Zone 2 is 14, 9, 7, 1
Zone 3 is 4, 4, 4, 3
Zone 4 is 8, 4, 3, 3

This confirms that the sonnets in zones 1 and probably also zone 3 are later than those in zones 2 and 4, for the latter profiles are front-loaded with links to the play groups 1 and 2 (the early plays). Jackson repeats the analysis using the Hieatt-Hieatt-Prescott word links between the sonnets and the poems Venus and Adonis, The Rape of Lucrece, and A Lover's Complaint, and gets the following profiles for links to the poems in that order:

Zone 1 is 12, 33, 7
Zone 2 is 16, 28, 3
Zone 3 is 7, 17, 7
Zone 4 is 10, 12, 1

Of course, the three poems are different lengths whereas the plays are roughly-equally sized, and the specific ratios of length are 33½ : 56½ : 9.9 for Venus and Adonis, The Rape of Lucrece, and A Lover's Complaint respectively. (Something must be slightly adrift here, as the ratios sum to 99.9 instead of 100.) Thus 22.6% (7 out of 31) of the sonnets in Zone 3 have most links with A Lover's Complaint, a poem that occupies only 10% of the total size of the three poems taken together, which difference (22.6% being more than double 10%) is caused by the chronological factor that Jackson is attempting to isolate. Since we now think that A Lover's Complaint is seventeenth-century, this suggests that Zone 3 (sonnets 104-26) and perhaps also Zone 1 (sonnets 1-60) "are later, or contain more late writing" that the other sonnets (p. 68).

    In pursuit of a still more finely grained approach, Jackson counts the Hieatt-Hieatt-Prescott links between the sonnet zones and each individual play in the Shakespeare canon (rather than using four chronological groups as before), checking the observed frequency of the rare words against the background of each play's vocabulary, so that if a play contains 5% of the total number of different words in the canon, it should have 5% of the rare words found in a sonnet; any more than 5% suggests a chronological link. Taken together sonnet Zones 1 and 2 have unexpected links with A Midsummer Night's Dream, Henry 5, and King John, and although the first of these might be explained by the shared genre of love, the last two cannot. This suggests composition of the 'Young Man' sonnets (numbers 1 to 103) in 1596-9, and that these sonnets have fewer than expected links with the last 11 plays (All's Well that Ends Well to Henry 8) confirms this view. The 'Dark Lady' sonnets (numbers 127 to 152) have links with 2 Henry 6, The Comedy of Errors, and Richard 2, thus the Zone 4 sonnets are early. Jackson breaks the sonnets down into still smaller collections (p. 69), although he is cautious with his conclusions because cognizant of the problem that small datasets are subject to distortion by random fluctuation. The case of The Phoenix and Turtle Jackson offers as a warning: by the linguistic link analyses described above it would seem to have been written or revised after the few first years of the seventeenth century, but in fact we know it was printed in 1601. Thus we should not rely on the evidence here to say that Shakespeare definitely was involved with the sonnets up to (and including participation in) the publication of the 1609 quarto (p. 73). But if Jackson is right that the 'Marriage' sonnets (numbers 1-17) were written after 1595, they cannot have been written to encourage Henry Wriothesley to marry since from 1589 (his 16th birthday) to 1594 Wriothesley was being pressured by William Cecil, Lord Burghley, to marry Burghley's eldest granddaughter Elizabeth Vere, with the threat of a large fine when he was 21. Had Shakespeare written the 'Marriage' sonnets then, Wriothesley would have thought him Burghley's stooge. In 1594, when Wriothesley became 21, this fine was exacted, so addressing the 'Marriage' sonnets to Wriothesley then, when Wriothesley was impoverished and could not afford to marry, would be insulting. No, Jackson concludes, the 'Marriage' sonnets cannot have been addressed to Wriothesley at all; they were to Henry Herbert. Likewise, the 'Rival Poet' sonnets, if written 1596-1604, cannot be about Marlowe since he died in 1593 (p. 74). Conveniently, Jackson ends by restating his main conclusion: ". . . the majority, if not all, of the last twenty-odd of the sonnets to the Friend [numbers 104 to 26, in Zone 3] were written in the seventeenth century. A few other sonnets, in both the Friend and the Dark Lady series, may have been written equally late, but the bulk of them belong to the 1590s" (p. 75).

    Next from Review of English Studies is Charles Cathcart's attempt to demonstrate that Hamlet must have been written in 1599 because it is echoed in Marston's Antonio and Mellida and the anonymous Lust's Dominion, both written in the winter of 1599/1600 (2001). Antonio and Mellida features a portrait inscribed "Anno Domini 1599" and "Aetatis suae twenty-four", suggesting composition in 1599 since Marston's 24th birthday fell in early October 1599 (p. 342) A reference to "the new Poet Mellidus" in Jack Drum's Entertainment (late spring/early summer 1600) is the terminus ad quem of Antonio and Mellida and Lust's Dominion is "widely accepted" to be "the spaneshe mores tragedie" that Henslowe paid Dekker, Haughton, and Day for on 13 February 1600, and the view that Marston too wrote Lust Dominion, for which he was loaned £2 for on 28 September 1589, is strengthened in this article. Cathcart admits that the dating of Antonio and Mellida from the portrait is not entirely secure since it might be supposed to represent Marston's father and the date the year of his death, but Reavley Gair has argued that Marston's poor relationship with his father at the end makes this unlikely (p. 343). Michael Neill and MacDonald P. Jackson date Antonio and Mellida on a collocation of "morphews" and "Cousin german" (4.1.25-6) which also collocate in Philemon Holland's translation of Pliny, published 1601 but Cathcart points out that Jonson's Poetaster clearly satirizes both Antonio plays and was in performance by autumn 1601 (according to Tom Cain, earlier according to others), which does not leave enough time for the first Antonio play to be in performance by mid-1601, especially as the failure of Antonio's Revenge to fulfil the promise of Antonio and Mellida's induction has suggested to several commentators that the second part did not follow hard upon the first. Thus Cathcart rejects the "morphew . . . Cousin german" evidence as a piece of later revision not affecting the date of composition of Antonio and Mellida (p. 345). In Antonio and Mellida there is a ghost that Antonio can, and Mellida cannot, see, just like the closet scene in Hamlet except that in Antonio and Mellida it "has no significance beyond the merely local" and hence the influence runs from Shakespeare to Marston and not the other way (p. 346). This logic is faulty, since one might by the same thinking argue that Martin Amis's Time's Arrow must be the source for Kurt Vonnegut's Slaughterhouse Five because Vonnegut does not make much of 'time-in-reverse' idea, using it in just one paragraph. But in fact we know that Amis took this paragraph in Vonnegut's book and expanded it to make his novel. Having established the direction of influence, Cathcart sets out to marshal indisputable evidence of a Hamlet/Antonio and Mellida connection by numerous verbal parallels (pp. 346-8) such as "this distracted globe" (Hamlet 1.5.97) with "your distracted eyes" (Antonio and Mellida 2.1.267), "soul . . . hoops . . . steel" (Hamlet 1.3.63) with "hooped . . .  steel . . . soul " (Antonio and Mellida 5.2.210-12), "sliver" (Hamlet 4.7.145) with "sliftered" (Antonio and Mellida 1.1.219), "plausive" (Hamlet 1.4.29) with "applausive" (Antonio and Mellida 2.1.111), and "chop-fallen" (Hamlet 5.1.188) with "chap-fall'n" (Antonio and Mellida 4.2.1); this last example being a late 1590s coinage, so unlikely to be due to shared descent from the ur-Hamlet.

    The most thorough parallel is in Hamlet Senior's battlements speech to Hamlet (1.5) and Andrugio's speech to Antonio (4.2): "thy fathers spirit" (Hamlet 1.5.9) with "Thy fathers spirit" (Antonio and Mellida 4.2.21), "hold my heart" (Hamlet 1.5.93) with "my panting heart" (Antonio and Mellida 4.2.12), "freeze thy young blood" (Hamlet 1.5.16) with "heat thy blood; be not froze" (Antonio and Mellida 4.2.18). There are reports of drownings in both plays, with the sinking thing buoyed up and making noise (singing in Hamlet, sighing in Antonio and Mellida), and the prologue of Antonio and Mellida refers to pouring "pur'st elixed juice" of art into the audience's ears, just as distilled juice is poured in Hamlet Senior's ear. Indeed, both plays use imagery of damaged ears: "assail your ears" (Hamlet 1.1.29), "do mine ear that violence" (Hamlet 1.2.170), "the whole ear of Denmark . . . abused" (Hamlet 1.5.36-8), "cleave the general ear" (Hamlet 2.2.565), "daggers enter in mine ears" (Hamlet 3.4.85), "infect his ear" (Hamlet 4.5.88), "ravished the ear" (Antonio and Mellida 2.1.116), and "taint not they sweet ear" (Antonio and Mellida 2.1.193). Cathcart argues that Shakespeare's use is the more thoroughgoing--and there are previous examples in "Piercing the night's dull ear" (Henry 5 4.0.11) and "bite thee by the ear" (Romeo and Juliet 2.3.72)--and hence the earlier. Cathcart offers a number of verbal similarities in common words between Antonio and Mellida 5.2 and the ghost-on-battlements scene (1.5) in Hamlet as well as the thematic links of feverish sons deprived of their kingdoms, fathers in full armour exhorting sons to resist, and imagery of a poisonous snake. In Shakespeare the snake imagery follows from what precedes it (the story given out of Hamlet Senior's death by snakebite) whereas in Marston it does not fit well, which tells Cathcart that Marston is the borrower, likewise Balurdo and Ophelia sing different lines from the same song and the same line from another song, but whereas Ophelia's song has thematic links with the rest of the play, Balurdo's does not. There are verbal parallels between the description of the ship headed for England in Hamlet and Antonio and Mellida, but in the latter it is not integrated with the plot but rather "unanticipated and undeveloped" (p. 350). Cathcart uses the same logic for the many echoes of Hamlet that occur in Lust's Dominion (pp. 351-3) and observes that if the borrowings are accepted we would have to dislodge at least two other datings (Antonio and Mellida by internal evidence, Lust Dominion by external evidence) to avoid the conclusion that Hamlet was ready by the end of 1599. That particular year might seem already crowded with Shakespeare plays, but then the Chamberlain's men had the Globe to launch and as Leeds Barroll has shown Shakespeare wrote not at a steady rate but in bursts of high activity separated by lulls.

    If Hamlet is as early as Cathcart suspects, we might find other early echoes of it, and Cathcart offers "retrograde to our desire", with the word "retrograde" being mocked in Jonson's Cynthia's Revels which he thinks was performed in 1600, as the Jonson Folio claims. In fact, the Jonson Folio dates must be March-March not January-December since Volpone alludes to the sighting of a whale in the Thames that Stow's Annals dates to 19 January 1606 yet the Folio insists it was acted in 1605. Greg made an uncharacteristic slip in this regard (1925-6, 345), thinking the verbal parallel inconclusive because he overlooked a marginal note in Stow that used precisely the words ("as high as Woolwich") found in the play and presumably circulated in word-of-mouth transmission. Thus Hamlet could be opened in the first three months of 1601 and yet be echoed by Cynthia's Revels that by Jonson's March-March reckoning was performed in 1600, and Cathcart is wrong to claim that "A debt to Hamlet in Cynthia's Revels would exclude 1601 as a possible date for Hamlet . . ." (p. 355). Cathcart concludes with a consideration of the interpretative implications for a 1599 rather than 1600-1 date for Hamlet, such as bringing Hamlet's use of, and comments upon, satire closer to the Bishops' Ban of 1599, that "arbitrary and petty ruling" (p. 359) 

    Finally from Review of English Studies is Katherine Duncan-Jones's argument that the 1616 printing of The Rape of Lucrece was a clumsy attempt to make it look more accessible than it is (2001). Many readers have agreed with Gabriel Harvey that The Rape of Lucrece is like Hamlet in being too long and too difficult, and despite the word "rape" in the title, there is no titillation, no comedy, no farce, and none of Venus's sweaty physicality in "bloodless and bodiless" Lucrece; instead, Shakespeare takes us (and his original, predominantly male, readers) into her mind. There is no real narrative in The Rape of Lucrece (except the Argument's summary), all is introspection and dialogue. Venus and Adonis was printed at least 10 times by 1617, The Rape of Lucrece just 6 and the format of the 1616 edition suggests that it was thought to be in need of editorial intervention to increase its attractiveness to readers. The final line is often mispunctuated as "Tarquin's everlasting banishment" but Duncan-Jones points out that the Argument makes clear that the entire dynasty is meant (so it should be "Tarquins'") and thus it should have been a warning to the earl of Southampton about his desires causing revolution, but the poem fails to make the point clear. A mark of the poem's reception in its own time is that Thomas Heywood, Shakespeare's admirer, copied Venus and Adonis in his Oenone and Paris in 1594 but took at least a decade to copy The Rape of Lucrece in his The Rape of Lucrece: A Roman Tragedy, which did try to inject the necessary storyline and humour. In 1614 John Harrison sold his rights in The Rape of Lucrece to Roger Jackson, who presumably knew that the reprinting of Heywood's bawdy play on the subject (also in 1614) would probably boost interest (p. 519). (Actually, Duncan-Jones writes that it was Nicholas Okes, not John Harrison, but this is merely a slip, as is a slight mixing of January-January and March-March dating schemes.) Jackson was an old hand at reprinting others' works from the past, but in this case he waited a couple of years, perhaps until Shakespeare died so that his changes (advertised on the title-pages as "Newly revised") might pass as authorial, or perhaps this ruse simply occurred to him once Shakespeare had died. Either way, later seventeenth-century editors did act as though the changes were authorial, and respected them (p. 520). What are the changes? One was to change the title from Lucrece to The Rape of Lucrece, and another was to break the indigestible work into 12 number sections, each prefaced with a summary. But these are misleading, and the crucial moment when Lucrece is comforted by a painting about the sack of Troy and Hecuba's woes (a strong parallel with Hamlet) is not indicated. Overall, the section divisions look like an effort to make the work appear less hard-going to prospective buyers (p. 521). Duncan-Jones surveys the variants between Q6 (1616) and the earlier printings and concludes that they are mostly silly slips with no overall intention. The most obvious 'revision' is a extensive use of italics: it does not make the thing easier to read, but it might well make the casual peruser think that it does, that these are signposts in a renownedly difficult work. By then Shakespeare was already 'dead' in the modern sense regarding an author's loss of control over his writing, as well as in the standard sense (pp. 522-3).

    The journal Analytical and Enumerative Bibliography folded at the end of 2001 with a bumper double issue on "Shakespeare's Stationers". Paul Menzer makes a rather slight argument that Q1 Hamlet needs fewer properties than Q2 or F, so it might reflect a touring text (2001). Menzer's style is irritating in its jokiness and he is poorly served by the quality of typesetting, which renders parts of his own body text as inset quotations. In matters of substance that are slips and ambiguities, such as calling Q1 "a version of F1" (p. 31) which is, of course, impossible since Q1 was printed two decades earlier; he means that the underlying manuscripts might be so related and as ever Jowett's terminology (MSQ1, MSF) is what is needed. At the end of Q1 Horatio calls for a stage to be erected in the market place for him to tell the tragic story of what has happened, whereas in F he says the bodies should be set up on a stage and he will speak the events over them, which sounds to me much the same thing, but for Menzer Q1 is metatheatrical, reflecting its own "transient" conditions of performance. Menzer characterizes Q1 as "built for speed" (p. 32) without saying why he thinks touring performances were faster; if he means they were shorter he should at least address recent scholarship showing that shortened plays typically need more actors  because doubling opportunities are lost. Colloquialism gives the wrong impression when Menzer calls Philip Henslowe a "clotheshorse" (p. 33) as though he thinks that the theatrical impresario's costume purchases were all for himself, and Menzer's analysis of the properties called for in different versions of the play is contestable. In what way is the "Arbor" in which the victim lies in the Q1 dumbshow to The Mousetrap "less specific than F1's 'Banke of Flowers'" (p. 35)? I would have thought it equally specific and possibly calling for a more bulky property. Using the Records of Early English Drama (REED) volume for the town of Cambridge, Menzer attempts to work out the features of the stage at Queen's College Cambridge, one of the universities that Q1's title-page claims hosted a performance, and to relate these to the phrasing of stage directions in Q1. Unfortunately, Menzer does not cite REED page-numbers or even which volume he is using (there are two for Cambridge), and my check of the index entries failed to turn up any evidence of the 17 feet width of the stage claimed by Menzer. Menzer uses the occurrence of "gallery" in Jonson's Epicoene ("do you observe this gallery . . .with a study at each end?" 4.5) to argue that the two projecting tiring houses at Cambridge (known from REED) effectively made a 'gallery' between them and hence Q1's stage direction reference to a "gallery". But this overlooks the fact that Epicoene was written for the indoor Blackfriars playhouse where there was a flat frons scenae, not projecting booths, and in any case his 17-foot stage hardly has the room for what Menzer imagines to be staging. Assuming the two tiring houses were a minimum of 4 feet wide, there would be just 9 feet remaining for 'gallery', so Hamlet's characteristic claustrophobia is more than adequately justified if, as Corambis says, "The Princes walke is here in the galery" (p.36-8). Menzer's errors come thick and fast in the final pages, including his claim that in the play's opening moment, ". . . the wrong guard issue[s] the challenge" (p. 43)--a contemporary military manuals shows the pre-emptive challenge to be quite normal (Edelman 2002)--and Menzer thinks that Hamlet's "shall I couple hell?" puns on 'shall I have sex with hell' and he cites the OED entry for "couple" meaning 'come together sexually' without noticing that this entry is specifically intransitive (for reflexive) and Hamlet's usage is definitely transitive (p. 45). A choice ambiguity comes near the end: "As the first quarto is nearly 1,600 lines briefer than F1 Hamlet, there is scant material unique to that text" (p. 46); who can tell which text he means? In the citations of authority, Robert E. Burkhart's book loses part of its title ("designed for") and R. A. Foakes is wrongly credited with sole editorship of the standard edition of Henslowe's Diary (R. T. Rickert was of course he co-editor).

    Jean R. Brink adds to our knowledge of the stationers William Ponsonby and Robert Waldegrave (not actually "Shakespeare's Stationers", of course) and their rivalry in printing the works of Philip Sidney (2001). Terri Bourus argues that Q1 Hamlet is not a piracy because the men involved in printing it would not do that, and we can account for the printing of Q1 and Q2 with simple bibliographical knowledge of Nicholas Ling, Valentine Simmes, and James Roberts (2001). Bourus's knowledge of the printing industry is not always perfect: she claims that the Stationers' Company charter "confined printing, though not bookselling, to the City of London with the exception of the university cities of Oxford and Cambridge" (p. 206). In fact it was not the cities of Oxford and Cambridge that were allowed have printing presses, it was just the university presses that were allowed, and more importantly this was not in the company charter of 1557 but the Star Chamber act of 1586, which also tightened up licensing. Bourus has considerable biographical knowledge about particular stationers, but is not always able to marshal it into an argument and has not avoided some egregious errors, such as claiming that puritans did not go to the theatre (p. 210). Bourus thinks that plague closure of the theatres in 1603 probably hurt sales of Q1 Hamlet (p. 216) but one might just as easily argue that serious addicts of drama would buy plays to get their fix, and she speculates that during plague closure from March 1603 until April 1604 Shakespeare revised Hamlet into something that was unperformable (p. 217), as I suppose he might were he thinking of a readerly market. Shakespeare might have offered this long version to Roberts, who would again have approached Ling and the readerly potential of this version would have been apparent to them, but they would have worried about hurting sales of Q1 which still had not sold out (p. 218) At this point Bourus argues that the theatres being closed would have stimulated the demand for printed plays, but earlier she argued the opposite: "It seems reasonable to speculate that sales of this book [Q1 Hamlet] would have been below average, since Q1's release date coincided with the outbreak of the great plague of 1603" (p. 216). Bourus thinks that Roberts and Ling would have destroyed unsold Q1's (to help sales of Q2), which is why there is only one extant. Roberts gave Ling's compositor the new Shakespeare manuscript, plus a copy of Q1 to help where the manuscript was hard to read, which is why, as we know, Q2 follows Q1 in some accidentals in the first act. The compositorial errors in Q2 (by a man who did not make many errors in other work) probably show that the work was hurried, as we might suspect because with Q1 suppressed, Ling needed the money (p. 220). This is an inherently implausible conjecture: Ling suppresses his own text to increase the market for a new version, which is then mangled because he is in a hurry because he has lost the income from the first one. Bourus concludes that Ling and Roberts were respectable men with professional relations to the playing companies, and they would not have got involved in piracy, and Simmes, although he got in trouble, did high standard work so it is unlikely he would have done Q1 if it were mangled text; Q1 and Q2 are just different versions of Shakespeare's play. This is reasonable enough as a conclusion, but since the claim that 'bad' quartos are piracies is no longer commonly made, it is somewhat unnecessary.

    The final piece in Analytical and Enumerative Bibliography is a sequence of explanatory and emendatory notes on the Folio Henry 5 by Thomas L. Berger and George Walton Williams (2001). The prologue says "O pardon: since a crooked Figure may / Attest in little place a Million, / And let vs, Cyphers to this great Accompt, / On your imaginarie Forces worke" (TLN 16-19, 1.Pro.15-8). A "crooked Figure" is usually taken to mean zero (the cipher), but as Gary Taylor notes in his Oxford Shakespeare edition of the play, the word "crooked" nowhere else means a full circle. Some critics have argued that the "crooked Figure" is a one, which did have some finishing dashes that made it not a simple down stroke but a zig-zag. In The Winter's Tale Polixenes says "And therefore, like a cipher, / Yet standing in rich place, I multiply / With one 'We thank you' many thousands more" (1.2.6-8) and since a zero cannot go on the left end of a number, Shakespeare must think of the "rich place" as the right-side of a number and hence the "little place" of Henry 5 must be the left-side. In King Lear, the Fool's "Now thou art an 0 without a figure. I am better than thou art, now. I am a fool; thou art nothing" (1.4.174-6) distinguishes the "figure" from the zero, and Henry 5 does so too ("And let us" not "So let us") and there was indeed a "learned tradition", that zero is not a number and elsewhere in Shakespeare "figure" and "cipher" are contrasted. In all this, Berger and Williams are reading "O pardon" as an exclamation, not as a form of 'Pardon the O' as Humphrey Tonkin did (2000) and they not consider the force of Ernst Gombrich's suggestion that "wooden O" means 'wooden zero' (2000). In the prologue's "For 'tis your thoughts that now must deck our Kings, / Carry them here and there: Iumping o're Times" (TLN 29-30, 1.Pro.28-9) the audience might be said to carry its thoughts (imperative) or perhaps the thoughts are said to carry the kings (indicative). Berger and Williams decide that it is perhaps better for line 29 to be, like line 28, indicative and thus for both to contrast with lines 1-27 where the principal verbs are imperative. In "'Gainst him whose wrongs gives edge unto the Swords. / That makes such waste in briefe mortalitie" (TLN, 174-5, 1.2.27-8), it is the swords that make waste, not the wrongs. The problem is "whose wrongs gives" and the most likely explanation is that the repeated "-s" in "whose wrongs" made the compositor add one to "give" too, and perhaps also to "Sword". In "for God before. / Wee'le chide this Dolphin at his fathers doore" (TLN 458-9, 1.2.307-8), "God before" could be a prayer ('God going before') or an oath ('I swear before God'), and Berger and Williams advise that a prayer better suits the tone of the passage. The act 2 chorus ends with two rhyming couplets (TLN 501-4), violating the pattern of the other choruses, although the first chorus might have a rhyme on its penultimate pairing of "supply . . .history" and might be allowed to be anomalous. But the other three choruses end on a personal note with the chorus referring directly to the audience rather than the story. The penultimate couplet of act 2 chorus does this, and its final couplet does not, so Berger and Williams reckon this final couplet is a non-Shakespearian addition. For "No, to the spittle goe" (TLN 575, 2.1.71) the usual emendation to "spital", short for "hospital", makes it a place free of disease (as in Love's Labour's Lost 5.2.857), whereas the meaning each of the three times it is used in this spelling in Folio Henry 5 is the opposite, a place of disease, so Berger and Williams think we should retain F's "spittle" as Shakespeare seems to mean something different from hospital by it. It is odd that Berger and Williams should think of the hospital referred to in Love' Labour's Lost as a place free of disease, since it contains "the speechless sick" and "groaning wretches" (5.2.837-8), just as modern ones do.

    For "feare attends her not" (TLN 917, 2.4.29) Berger and Williams supply the somewhat obvious gloss that England is unfearful because she (in the form of her king) is giddy and vain. Emending "Dolph. For the Dolphin . . . what to him from England? / Exe. Scorne and defiance, sleight regard, contempt, / And any thing that may not mis-become / The mightie Sender, doth he prize you at" (TLN 1010-4, 2.4.117-9) Capell put a semi-colon after "defiance" since only up to there is Exeter answering the Dauphin's question; the rest is a sentence with "prize" its main verb. Taylor, on the other hand, put the semi-colon after "contempt" because "Scorne . . . contempt" is the answer. Berger and Williams distinguish the "externally directed" challenge of "Scorne and defiance" that cannot be the objects of "prize you at" from the "internally conceived" words that follow ("slight regard" and "contempt") that can be the objects of "prize you at". I would have thought that if you can prize someone at "slight regard" you can prize them at "scorn", and the waters are muddied here by a spurious negative in Berger and Williams's prose: "'he does not prize you at scorn and defiance' is not idiomatic". To Berger and Williams the obvious break is after "defiance" and Capell's semi-colon is right. Also, Exeter answers like this (a few syllables in direct response, then a longer comment) elsewhere in the play (TLN 970, 991). The end of the second act and the beginning of third go like this: "To answer matters of this consequence. Exeunt. / Actus Secundus. / Flourish. Enter Chorus". Here and at TLN 462 (wrongly given as TLN 1462 on page 268 where it appears in through-the-play sequence, as befits a glossarial notes), a flourish accompanies the chorus's entrance, but probably only because elsewhere in the manuscript a flourish seems to float between the end of one scene (a king's exit) and the beginning of the next (the chorus's entrance) and the compositor wrongly attached it to the latter. The flourish should, of course, always be moved to the king's exit. This third act chorus says "With silken Streamers, the young Phebus fayning" (TLN 1050, 3.Cho.6), which last word is commonly emended to "fanning", but "fayning" meaning feigning also appears in F and the sense is appropriate, the streamers producing a false illusion of the rising sun. There is also a parallel moment in 3 Henry 6: "See how the morning opes her golden gates / And takes her farewell of the glorious sun. / How well resembles it the prime of youth, / Trimmed like a younker prancing to his love!" (2.1.21-4). Fluellen's speech prefix at "Flu. Up to the breach, you Dogges; auaunt you Cullions / Pist. Be mercifull great Duke to men of Mould" (TLN 1137-9, 3.2.21-3) is suspicious; should it not in fact be one of the great dukes that enters and rouses them? Fluellen later describes Pistol as brave and furthermore this would be the first time we see Fluellen, and people usually are identified at their first appearance in a play. Fluellen becomes just "Welch" in his speech prefixes 50 lines later, suggesting that at this point in the play some revision has occurred; Fluellen was originally not individuated and then became so.

    There's a false exeunt in "Erping. The Lord in Heaven blesse thee, Noble / Harry. Exeunt. / King. God a mercy old Heart, thou speak'st cheare- / fully. Enter Pistoll" (TLN 1879-82, 4.1.33-34.1) since the king remains and speaks, and even if only Erpingham leaves there is still the problem of his being addressed after he has gone. Presumably, the exit for Erpingham was on the same line as Pistol's entrance, and the compositor, short of space and unused to such a collocation, moved Erpingham's exit up two lines. Taylor reassigns the speech beginning "Will. 'Tis certaine . . . " (TLN 2034 not "12034" as given here, 4.1.197) from Williams to Bates, but there is no need since it makes perfect sense as it is and leads smoothly enough to Williams's quarrel with Henry. At TLN 2157-63, 4.1.321-6 there are several short lines here that editors have tried to force into a metrical scheme, but Berger and Williams think it hopeless and that we should just leave them as unmetrical. The king says "My brother Gloucesters voyce? I: / I know thy errand" and T. W. Craik for Arden 3 deleted the first "I" on the grounds (explained in private correspondence to Berger and Williams) that it was Shakespeare's false start for "I know thy errand" and that the compositor took it as an interjection and added the colon himself. Berger and Williams prefer to see "I:" as an interjection and note that the question mark after "voyce" might indicate not interrogation but emphasis. The dialogue can stand as it is, no emendation necessary. Berger and Williams note that the preceding speech, an audible prayer by Henry, could become a silent prayer and hence the half-line "Imploring pardon". Like the end of Claudius's prayer, the half-line might indicate that the person praying continues to do so silently, and Berger and Williams think that rather than have Gloucester burst in and interrupt Henry's devotion, it is better if his call "My liege" is made off stage, answered by Henry speaking to himself ("My brother's voice? Ay. / Enter Gloucester. / [Rises] I know thy errand") or perhaps a fraction earlier so that Henry's "Ay" is addressed to him. There's a logical problem in "Dolph. Mount them, and make incision in their Hides. / That their hot blood may spin in English eyes" (TLN 2178-9, 4.2.9-10) since blood does not spin, but it does "spit" several times in Shakespeare, so that is the best emendation. Of course, the blood would have to spit itself for this emendation, and Craik (in private correspondence to Berger and Williams) points out that fire does that in "Spit, fire! Spout, rain!" (King Lear 3.2.14). Presumably what happened was compositorial error: the word "in" drove out the "-it" of "spit" to make "spin". Accent perhaps obscures meaning in "French. O perdonne moy. / Pist. Say'st thou me so? is that a Tonne of Moyes?" (TLN 2403-4, 4.4.22-3) and "Tonne" could be 'tun' (a container) or a 'ton' (unit of weight). In private correspondence Craik argued that since 'tun' was spelt 'tun' earlier, it is unlikely to be spelt 'tonne; here, but perhaps, say Berger and Williams, the previous line's "-donne" was supposed to be echoed in a disyllabic "Tonne". Also, the compositor might have changed the spelling to fill out his prose line and also to distinguish the English word from the French pronoun "ton" 5 lines earlier. It makes thematic sense of Henry to get a 'tun' from a Frenchman earlier and for swaggering Pistol to get a 'tun' from a Frenchman here too, so Berger and Williams favour reading 'tun'. There is unShakespearian repetition (of "throngs") in  "Orl. We are enow yet liuing in the Field, / To smother up the English in our throngs, / If any order might be thought vpon. / Bur. The diuell take Order now, Ile to the throng; / Let life be short, else shame will be too long" (TLN 2478-82, 4.5.20-4). The gap is a little to large for compositorial anticipation or recollection, and Berger and Williams think the second instance more right than the first since a throng generally has a purpose, so "our throngs" seems the problem. Since the nobles have left the field, there is no need for the first word to respect the ranks of those on the field, so the Berger and Williams suggest emending the first "throngs" to "swarms" and wonder if the "th-r" sound of "smother" is what caused 'swarms' to get turned into "throngs".

    There is an erroneous "Actus Quartus" at (TLN 2524, but something like "Alarums. Excursions" is needed to pass the time during which the English kill their prisoners and then the French attack the boys and the luggage. One cannot have Henry give the order to kill prisoners at TLN 2522 ("euery souldiour kill his Prisoners") and it be executed, together with the killing of the boys, by TLN 2526 when Fluellen reports it ("Kill the poyes"). The word "law" (TLN 2673, 4.7.150) is an Irish interjection, according to Berger and Williams, and is used several times by MacMorris, so its use here by Fluellen suggests that it was reassigned from MacMorris (given as "WacMorris" here) for casting reasons. For "Will. Sir, know you this Gloue? / Flu. Know the Gloue? I know the Gloue is a Gloue. / Will. I know this, and thus I challenge it. / Strikes him" (TLN 2720-3, 4.8.5-8) Berger and Williams sort out the exchanging of gloves and their passing between caps and hands. They reject Gurr's suggestion that Williams strikes the glove hanging from Fluellen's cap (a literal reading of "has struck the glove" 20 lines later) because it is unchivalric and does not mean the same as a "box o' the ear" as the blow is repeatedly called. Taylor's stage direction "plucking the glove from Fluellen's cap" has the advantage also of getting the gloves back into their rightful owners' hands. Just how much time is indicated by "S. Dauies day is past" (TLN 2899 not "12899" as given here, 5.1.2)? The battle of Agincourt was on 25 October 1415 and the next St David's day after that was 1 March 1416. Historically, 10 days after that St David's day (that is, on 11 March 1416) Henry's advance party of troops heading for Troyes arrived in France. In A Midsummer Night's Dream "St. Valentine is past" is said well after 14 February, so there is no problem with thinking that Shakespeare used "past" here to mean 10 or more days past and that the scene is set before Troyes where Henry's men are awaiting his arrival. In "The euen Meade . . . / Wanting the Sythe, withall vncorrected, ranke; / Conceiues by idlenesse, and nothing teemes, / But hatefull Docks, rough Thistles, Keksyes, Burres, / Loosing both beautie, and utilitie" (TLN 3035-40, 5.2.48-53), is "nothing" the subject of verb "teemes" or its object? Where 'teem' is used elsewhere in the canon it is transitive, so presumably it needs an object here and the only thing available is 'nothing'. The sense is thus: 'the mead conceives by idleness and teems [produces] nothing but'. At TLN 3041-4 (5.2.54-8) Henry lists the domestic matters that have been neglected because of the war ("for want of time") but Berger and Williams think that the fighting has not made less time available, so perhaps "time" is a misprint for 'care'. (Personally, I would have thought it acceptable to say that people preoccupied with one thing would have less time for others.) Finally, just before the epilogue there is a flourish at the king's exit that the compositor might have misunderstood as accompanying the chorus's entrance after it (TLN 3366-7 but wrongly given as 3266-7 here).

    Three relevant articles appeared in Papers of the Bibliographical Society of America this year. Eris Rasmussen argues from the evidence of a printer's device deteriorating that the undated Q4 Hamlet was printed 1619-23, probably within 1619-21, and was thus available to be the quarto used to help in the printing of the Folio text, as has been suspected from other evidence (2001). On 19 November 1607 Nicholas Ling transferred his right to print Hamlet to John Smethwick, who published Q3 in 1611 and then at some point Q4 with an ornament on the title-page (McKerrow's device number 376). Rasmussen sent out a team of researchers to find all the books Smethwick published 1599-1640 and (for books in Bodleian and Huntington) to get colour slides of each title-page on which this device appeared, which slides were electronically scanned at high resolution. The resulting images were taken on a laptop computer to the British Library to compare with its copies of Smethwick's book bearing the device. McKerrow reported no noticeable deterioration in the block in the books he saw it used in, but from their multiple copies of 28 extant books bearing it Rasmussen's team found progressive deterioration in two areas. One is the bit of drapery hanging from the dog's mouth in the upper-right-hand corner, which was attached to the scroll of the frame in every example up to and including STC 3670 (printed 1621) but is detached from STC 16672 (1623) and onwards. On the Bodleian, British Library, and Huntington Q4 Hamlets it is attached, so 1623 is the terminus ad quem. (In an article on deteriorating printers' device reviewed below, Lynette Hunter cautions against mistaking insufficient or excessive inking for a break or non-break in a device, but Rasmussen's use of so many printings of the device should guard against that.) Similarly, the central forelock of the cherub's hair (at the top of the device) was still attached to the brow in STC 7222 (1619) but had broken free, leaving an 'island' of hair fragment floating unaided, in STC 3670 (1621). This island is still visible in just one copy of STC 16672 (1623). Q4 Hamlet has the forelock detached from the brow so the terminus a quo is 1619, but the 'island' is not yet visible and there seems to be more of the forelock present than in STC 3670 (1621) so this is the probable terminus a quem (1623 is the certain terminus a quem, as already established).

    Rasmussen refers to the "elegant progression of damage" (p. 27) in his Illustration 6 but makes no mention of the obvious digital enhancement of the images from these books, for not only has he used enlargement (which is entirely reasonable) but clearly he has had his Adobe PhotoShop software (mentioned as doing the 1200 times enlargement, p. 24n4) perform line-edge detection which 'draws around' the inside and outside edge of a shape and then subtracts the shape, thus turning a filled-circle, say, into pair of unfilled concentric circles. According to how sensitive a setting one uses for this feature, it will find edges to artefacts that are merely 'noise' such as a dirt (not ink) on the paper, and this has clearly happened to some of the images in Illustration 6, most clearly the one of the Bodleian Q4 Hamlet copy in his Illustration 6 on page 28. Arguably, the 'island' of cherub's hair that Rasmussen refers to has been created by this electronic processing, for clearly in the case of the Bodleian copy of STC 16672 the software has found a whole archipelago of islands far off the coast of the cherub's head that are in fact merely dirty marks on the paper, not inkings. With no discussion of the use of the software's 'edge-detection' feature, and no assurance that the same 'sensitivity' and 'discrimination' settings were used for each image, the results of this analysis are entirely suspect. Particularly peculiar is the fact that the 'edge-detected' Huntington STC 16672 image is free from the falsely-identified islands yet its unprocessed source image suggests that this book is if anything more smudgy and dirty than the Bodleian STC 16672. With digital enhancement it is all too easy to make 'noisy' data seem more meaningful than it really is, and the variabilities of inking and of dirty paper are extremely loud 'noises' in this context. Rasmussen continues with the observation that some quarto has long been suspected of influencing the Folio, and the pattern of F/Q4 agreement against Q1-3 has made editors suspect it was Q4, if only we could be sure it was available. Now, claims Rasmussen, we do know that it was printed in time. This is not quite what Rasmussen has shown, however, for even if accepted in full his study indicates that Q4 Hamlet was printed before STC 16672 (Lodge's Euphues Golden Legacy) in 1623, for that is the certain terminus a quem Rasmussen announced on page 25; he attached an evasive "probably" to the terminus a quem of 1621 derived from the evidence of the cherub's forelock on page 27. Moreover, we do not know when in 1623 Lodge's Euphues Golden Legacy was printed, and we know that F Hamlet was set in type beginning in the late spring of 1623, so Q4 might still not have been printed by then. The slippage in Rasmussen's prose is palpable: Q4 Hamlet was printed "certainly before Euphues in 1623" (p. 27) transforms into "published before 1623" (p. 29). Those are not the same things at all.

    The most important and impressive of articles this year is Paul Werstine's demonstration that the alleged differences of habit between Folio compositors D and F do not stand up to an analysis of the influence of copy; when one takes Ralph Crane's spellings into account, it is impossible to tell them apart (2001). It has long been suspected that the first four Folio plays, The Tempest, The Two Gentlemen of Verona, The Merry Wives of Windsor, and Measure for Measure (occupying quires A-G) were printed from Crane transcripts, on the evidence of punctuation, spelling, elision and other features that match those in extant Crane dramatic manuscripts. Thus the compositors who set these plays cannot have been so fixed in their habits as to "escape the influence of manuscript copy". Hinman called them compositors A, B, C, and D, but Howard-Hill showed that this A could not be the same A that Hinman found working on The Winter's Tale and the histories and tragedies, so he called the four men B, C, D, and F. John O'Connor showed that there was not much evidence to distinguish this new man F from D (perhaps he really was  D?) so he tried to find some evidence from spelling. Howard-Hill discriminated D/F by the habit of indenting the second line of an overflowing line and thought the absence of this phenomenon from quires A-E shows D's absence, but Werstine counters that this proves nothing since there is no consistent idiosyncratic pattern, only collective ones: when D set indented flow-overs, so did others, and apparently when he refrained so did others. Moreover, D seems influenced by use of indented flow-overs in his copy, but again not consistently. This negative evidence is not convincing, but positive evidence is: the only use of flow-overs in quires A-G come in D's stints, so the single one on G5v is probably his (p. 318). O'Connor too rested his argument on weak negative evidence, the absence of other compositors' habits. Four comedies were set from printed copy, Much Ado About Nothing, Love' Labour's Lost, A Midsummer Night's Dream, and The Merchant of Venice, and from D's habits in his stints (since source could be compared with results) O'Connor worked out D's spelling habits, although he admitted that D followed copy spelling so often that it is hard to tell just what his preferences are. O'Connor nonetheless came up with some preferences (setting -ie endings where his copy has -y endings, and also preferring the unusual spellings sweete, meete, maide, eie, and praier) and because these are almost entirely absent from pages of quires A-G not set by B or C, O'Connor deduced that D worked on none of these quires except F. It is not hard to tell compositors' habits if they are setting from printed copy still extant, but where the copy is a lost manuscript it is usually impossible. Quires A-G were set from lost manuscript copy, but we have the advantage of knowing that his manuscript was in Crane's hand, and we know his spelling habits from other extant manuscripts (p. 319).

    In Ralph Crane and Some Shakespeare First Folio Comedies (1972) Howard-Hill summarized Crane's spelling habits, so we can reliably guess at the spelling in the copy for quires A-G and see how far compositor D was influenced by copy in his spelling of words that we know his personal preferences about from those F passages we know were set from extant printed copy. Werstine's Table One records what compositors B, C, D, and F did with 17 words--any, beutie, body, company, deny, happie, heavy, pittie/pitty, presently, try, very, euery, Lady, many, carry, marry, weary--that Connor thought compositor D liked to end in -ie, starting with the Crane spelling (the one presumed to have stood in the lost manuscript copy for quires A-G), and then showing how often this word was made to end -y and how often -ie by each of the four compositors in Crane-manuscript-copy plays (The Tempest, The Two Gentlemen of Verona, The Merry Wives of Windsor, and Measure for Measure), in the Q-copy plays (Much Ado About Nothing, Love's Labour's Lost, A Midsummer Night's Dream, and The Merchant of Venice), and in four other comedies (The Comedy of Errors, As You Like It, The Taming of the Shrew, and All's Well that Ends Well) that we know were set from lost manuscript copy. For each group Werstine's table shows how often each man set -y and -ie endings for the given word, taking care to separate out occasions when this was done in long lines, whose spelling might have been changed to achieve justification. (Werstine agrees with O'Connor that such cases tell us nothing about real spelling habits, for justification needs might well have taken precedence over personal preference.) Werstine summarizes from his own table: for 12 of the words (the first 11 plus the last one), compositor D showed a marked tendency to prefer -ie endings even where his printed copy had -y. For the same 12 words, where compositor F was working from Crane-manuscript he only once set -ie (for companie) and 11 times set -y. However, for some words (euerie/euery, Ladie/Lady, manie/many) compositor D was inconsistent in his habits when setting from Q-copy, sometimes changing the -ie ending to -y and sometimes vice versa, even in short lines where justification could not have been forcing his hand (p. 324). There are 2 occurrences of euerie in the Q copy, 1 of which compositor D changed to euery. Looking only at short F lines, there are 6 occurrences of euery in Q and of these compositor D left 5 alone and turned one into euerie. Again, looking only at short F lines, there are 14 occurrences of Lady in the Q copy, 5 of which compositor D changed into Ladie. There are 12 occurrences of Ladie in the Q copy, 9 of which D changed into Lady. (Actually, these changes are wrongly given in Werstine's prose, which states the Ladie-to-Lady figure as 14/5 and the Lady-to-Ladie as 12/9; in private communication Werstine has confirmed that the table is right and the prose is wrong.) Werstine goes on in his prose to describe in like manner what compositor D did with manie-to-many and many-to-manie but it is impossible to check because for the word 'many' in his Table One there is no row for compositor D at all. So, compositor D's preference for euery, Lady, and many seems stronger than the previous examples, but since D's preference is not clear we cannot make much use of this; indeed for the words 'carry' and 'marry' compositor D actually changed -ie endings in his Q copy to -y endings, just as compositor F does. For 12 words compositor D behaves in a way that distinguishes him from compositor F, but for 3 words he is inconsistent, and for a further 2 words he behaves like (and therefore could be) compositor F (p. 324).

    Of those crucial 12 words for which compositor F distinctively uses -y and compositor D distinctively uses -ie, 10 are spelt -y by Crane, so perhaps all compositor F is doing is following his scribal copy. There are some minor problems with Werstine's reporting of Howard-Hill's discovery of Crane preferences. Howard-Hill gives "heauy", not "heavy" as Werstine has it, which cannot be simple modernization since Werstine preserves "euery". Howard-Hill has not got "presently" (as Werstine has), and Howard-Hill gives both "very" and, "verie" whereas Werstine just has "very", although it is not clear why Howard-Hill does not report this as 'very/ie' as he does for other words that Crane is inconsistent about. Werstine provisionally concludes that the lack of compositor D's -ie spellings in quires A-G does not prove D did not set the parts of quires A-G usually attributed to F since he might simply have tolerated the -y endings (Crane's preference) in his manuscript copy and set them even though his own preference when setting from printed copy was to use -ie endings. Certainly, in the part of quire F (from Crane manuscript copy) that compositor D did set, he used 18 -y endings in short lines, to just one -ie ending. (This is not verifiable from Werstine's Table One because it does not break quires A-G down into compositor D and compositor F parts.) O'Connor's observation that from quire F the number of -ie endings rises--and hence compositor D started there--is invalid because compositors B and C also demonstrably increased their usage of -ie endings from quire F onwards, and looking at their spelling habits when setting from quarto copy, compositor B and C also seem to have been strongly influenced by Crane's preferences when setting from manuscript copy in his hand (p. 325). Thus O'Connor's spelling tests that contain compositor D's work to just quire F amongst quires A-G would also deny compositors B and C hands in quires A-G too. There were a further five spellings that O'Connor thought were distinctive of compositor D--eie, maide, meete, praier, sweete--but Werstine excludes praier and meete on the grounds that compositor D showed no preferences regarding these outside of long lines, and he tabulates the other three in his Table Two. O'Connor's reliance on Compositor D's alleged preference for eie is mistaken: he did change some eye spellings in his Q copy to eie spelling, but then not only did he set eye 27 times where his Q copy had that spelling, but 3 times he set eye even though his Q copy had eie, so if his Crane manuscript copy had a mix of eie and eye spellings, as seems likely, then compositor D might (like compositor F) have set only eye (p. 326). Regarding the maid/maide preference, we know that compositor B frequently used maide in setting from manuscript copy later in the Comedies, and that compositor C showed his "tolerance of maide spellings" when setting from Q copy (following his copy's maide spelling 6 times), so the fact that these two men (B+C) set only maid (never maide) in quires A-G probably means that they were following Crane's spelling in their copy, for he spells it maid, never maide, in a number of manuscripts that Werstine has checked electronic versions of (it is not in Howard-Hill's list of Crane spellings). Therefore, the two times that maide does turn up in the parts of A-G assigned to compositor F might instead be where compositor D is setting, because compositor D only ever set maide in his Q copy work, even where the Q copy had maid. Regarding the sweet/sweete preference, O'Connor observed that compositor F used only the short form sweet while compositor D used both sweet and sweete, but in fact looking at compositor D's work from Q copy we can see that he never changed sweet to sweete (as always, we are concerned only with short lines) and mostly what he did was follow his copy. So, since we know that Crane used the spelling sweet, the assignment of pages in quires A-G to compositor F on the grounds that they contain the spelling sweet is faulty: compositor D would have set sweet every time given that Crane copy, or in other words perhaps he only flitted between sweet and sweete when his copy had sweete (p. 328).

    Another type of available evidence is elision of future tenses of the kind he'll. O'Connor argued that compositor D characteristically used -le (so, heele), changing Q copy spellings of 'll to -le, while compositor F favoured 'll. Werstine tabulates the behaviour of all the compositors of the first 12 Folio plays (B, C, D, F). This table is badly presented: there are lines of data for someone whose identificatory letter is not given (there is just a blank cell) and some cells are blank while others are filled with a hyphen, and nowhere does Werstine explain the difference between these two. Also, confusingly Werstine refers to what happens in "the bottom half" of the table, but he means the rows of data for quires H1-Y1v and because the whole table has been wrapped to fit the journal's printed page, those rows are not simply "the bottom half" but rather occur twice in the table. Moreover, the labels seem to have gone wrong, for the first time the quires are called "Hl-Ylv" (both alphabetic 'l' not numeric '1') and the second time "Hl-Y1v" (one l, one 1). Nonetheless, correctly interpreted the table shows that from knowledge of Crane's habitual spellings we can say that compositor D never changed a quarto copy spelling of a future elision long line (p. 328). What of compositor D's practice in the manuscript copy plays? In those compositors B and C used -le against their usual practices ('l and 'll respectively), so perhaps they were just following copy, in which case compositor D's setting of -le in the same plays was just a matter of his following copy too. Elsewhere compositor D used -le and 'll in work for which we do not know what the copy was, but he might still have been simply following copy in each case, and indeed in this he would (if it were true) be like compositor F setting quires A-G, who used Crane's preference ('ll). Crucially, then, Howard-Hill's and O'Connor's work in distinguishing compositors D and F failed to take into account how copy might have influenced the spacing and spelling evidence they were using. What we need is positive evidence for F's habits not influenced by copy, but it is not available because apart from the evidence adduced by Werstine so far, the working habits of D and F appear to have been the same or were inconsistent (so not distinguishing ones) or the influence of copy obscures the evidence (p. 330).

    To attempt to distinguish compositor D and compositor F, Werstine repeated O'Connor's collation of D's setting from Q copy with those quartos, excluding changes due to eye-rhymes (those like quay/day that strike only the eye) and excluding long-lines. In order not to be fooled by ignorance of spellings drawn straight from copy, Werstine excluded words where we do not know Crane's spelling, and this left 60 words charted in the Appendix to the article. We should be able to see compositor F's distinctiveness in spellings he prefers in quires A-G that are not Crane's nor compositor D's, but of the 322 times compositor F set one of these 60 test-words in short lines, 317 times he chose a spelling that was either the same as Crane's or the same as compositor D's, or the same as both of them. 5 distinct spellings in 322 spellings is 1/64 (p. 331). For compositor B and compositor C's stints in quires A-G, the same 'distinctiveness' ratio, how many times in short lines they used a distinct (not Crane, not D) spelling out of how many times they set the word at all, is 1/14 for compositor B and 1/9 for compositor C. Of course, Werstine is not claiming that compositor D is compositor F, rather he has most convincingly shown that we cannot tell them apart by their spellings. On a range of words they do not distinguish themselves by their spelling, and on specific words they actually do exactly the same things: both were weak about been, deare, and sense and would sometimes follow copy and sometimes apply their own spellings. Yet both were strong (to the point of absolutism in some cases) about heart, diuell, grieue, answere, and indeede and overruled their copy on these spellings (p. 332). As Werstine observes, two men might share some preferences, but it is hardly likely that they "would maintain these preferences with equal strength" (by being weak on the same three words and strong on the same five words). But what if we are being fooled by dependence on the incomplete list of Crane spellings found by Howard-Hill and there are other distinguishing words we could turn to? To discount this possibility, and to see if there is another way to distinguish compositor D from compositor F, Werstine has looked in compositor F's pages in quires A-G for spellings that we know--from his overruling of Q copy in quires H-V--compositor D did not share, and he compared them with spellings that compositor D actually used in quire F, the result being Werstine's Table Four. The results are inconclusive, with compositor D never showing strong preferences that can be contrasted with compositor F's preferences; indeed we cannot be sure that we are rightly distinguishing by this means which pages of quires A-G compositor F set and which pages compositor D set, for they share spellings in those quires (adieu, already, lyon). Comparing compositor F's spellings in quires A-G with compositor D's spellings in quires H-V there is some variation, but it is no greater than one man might show in his habits; indeed, it is no greater than the variation compositor D certainly did show, given current attribution of pages, between his work on quire F and his work on quires H-V. In short, Werstine is able to announce that "Exhaustive analysis of Compositor D's habits thus produces almost no evidence to distinguish him from Compositor F, and much to associate him with Compositor F." (pp. 333-4). The main Folio compositors (A, B, C, E) are still in place, but the "peripheral" ones (definitely, since this article, F, and most probably H, I, and J too) are not solidly grounded in evidence. As Werstine rightly concludes, this casts yet more gloom on New Bibliography's dream of recovering what Shakespeare himself wrote.

    William Searle argues that, contrary to Taylor, the manuscript used to annotate a copy of Q Troilus and Cressida to make the Folio text was not necessarily scribal, for the phenomena displayed in the Folio could be created by compositorial error (specifically the misunderstanding of an annotator's instructions about placement and substitutions) and subsequent miscorrection of these errors (2001). Phillip Williams showed the F Troilus and Cressida was printed from an annotated copy of Q and Taylor attempted to show that the annotation was probably by reference to a scribal transcript of authorial papers. There are errors in F that Taylor claimed could only come from "clearly legible misreadings" in the manuscript used to annotate Q (hence it was not an authorial manuscript), and this Searle sets out to disprove: they are in fact much more likely to be errors by one the compositors of F, for they all fall in one man's stint. (p. 503). Alice Walker, who would not accept the idea of authorial revision, blamed compositor B for F Troilus and Cressida's deficiencies, but Werstine has shown that B was a careful man. Compositor H, hired at the last minute as Troilus and Cressida was finally squeezed into F, worked only on this play so we cannot test his care in other plays in F, as we can for B. Taylor listed a number of miscorrections in F--things that Q got right that F got wrong--and Searle reprints it. Taylor has us imagine the work of the person who annotated Q by reference to an unknown manuscript. If that manuscript was hard-to-read foul papers, would he really have preferred his own stab at what it said over what Q said? For that is the phenomenon we're dealing with, the preference for wrong words over perfectly good ones in Q. Surely, Taylor argued, the annotator would have preferred his (wrong) manuscript only if that were a clearly-written manuscript, the clarity of the writing giving it a spurious authority. Of course, this argument depends on Q being right in the first place (else F's reading is not a "miscorrection"), and F being definitely wrong, and at least one of Taylor's claimed miscorrections is not right in Q to begin with. Also, Taylor did not consider the possibility that the error in F came not from the annotator but from the compositor misreading the annotator's handwriting. Also, we should not count cases where the apparent F error appears in a section for which the annotator did nothing to Q, and Taylor found four sections where F was printed directly from (unannotated) Q, some of which might be because the manuscript used to annotate Q lacked some sheets (p. 505). Any reprint will necessarily introduce new errors, and Searle's argument is essentially that the list of F miscorrections used by Taylor is really a list of random errors introduced in a reprinting of Q. It is noticeable that 3 of Taylor's 28 examples were set by compositor B and 24 by compositor H (one Searle rejects as not being right in Q in the first place), and even allowing for H setting more than B did, that makes H's work much more frequently in error (p. 506). Worse for Taylor's theory, several of the Folio 'errors' on his list have recently been accepted into modernized editions of Troilus and Cressida, so they are not clearly errors in F at all; leaving them out changes the tally to just 1 'error' in compositor B's stint, and 21 in compositor H's, so it really looks like the phenomenon is compositorial, not manuscriptural.

    First Searle sets out to suggest that 3 of the items on Taylor's list are in fact "faulty substitutions" (p. 508), not misreadings, and to these Searle adds 4 more of his own not on Taylor's list and then proceeds through these 7 cases. The first, and one not discussed by Taylor, is "He sate our messengers and we lay by. / Our appertainings , visiting of him / Let him be told so, least perchance he thinke, / We dare not moue the question of our place" (Q E1r) versus " He sent our messengers and we lay by. / Our appertainments, visiting of him / Let him be told of so perchance he thinke, / We dare not moue the question of our place" (F ¶4r). Searle suggests that the annotation was supposed to make Q read "He sent our messengers off, so we lay by" and the "sent" made it, but the "off" did not, and the "off, so" of this correction got put in two lines further down, turning Q's "so, least" into F's "of so". The extra word "off" need not be extra-metrical, for "messengers" could be two, not three, syllables. Searle's second example, Taylor's number 8, is "To subtill, potent, tun'd to sharp in sweetnesse" (Q F1v) versus "To subtill, potent, and too sharp in sweetnesse" (F ¶5v). Searle suggests that the annotation was supposed to make Q read "Too subtle-potent, and tuned too sharp in sweetness" (that is, the intention was simply to get an "and" in before "tuned" but instead the "and" replaced the "tuned"). Searle admits that this involves an extra-metrical 'and', but the play has plenty of those. Searle's third example is Taylor's number 11, "And violenteth in a sence as strong / As that which causeth it" (Q H3r) versus " And no lesse in a sence as strong / As that which causeth it (F ¶¶2v). Searle suggests that the annotation was supposed to make Q read "And violenteth in a sense no less / Than", which is to say "no less" was supposed to replace "as strong" but instead was taken as a substitution for "violenteth", and "Than" was supposed to replace "As" in the next line but was omitted. Taylor's claim is that the scribe reading the authorial manuscript (to make the scribal transcript that was used to annotate Q to make the copy for F) read "no lesse" where the Q compositor (reading the same authorial manuscript) saw "violenteth", but Searle objects that these two things are not graphically close (p. 511). Why should the same word get set properly by the Q compositor (for "violenteth" is undoubtedly right) yet totally confuse the scribe, causing him to write nonsense? Searle invokes the same 'what's-likelier?' principle as Taylor (that if the annotator had Q and an illegible authorial manuscript why would he allow his uncertain reading of the latter to overrule the possibility of what was in Q?; no, he must rather have had a fair copy manuscript and trusted it) but asks this of the scribe making the transcript from foul papers to fair copy rather than the annotator using the fair copy to mark-up Q. After all, why would this scribe misread in the authorial foul papers a word that the printers of Q, using the same authorial foul papers, got right? Taylor sought a reason  for the seemingly illogical behaviour of the annotator (the overruling of Q) and found it in a fair copy manuscript being used to make the annotations, but we do not have to accept this illogical behaviour on the part of the annotator since the Folio compositor misreading the annotative markings on the Q used as F's copy is the likelier source of the error.

    Searle's fourth example, not discussed by Taylor, is "To shame the seale of my petition to thee:" (Q H4v) versus " To shame the seale of my petition towards" (F ¶¶3r). Searle suggests that the annotation was supposed to make Q read "To shame the zeal of my petition towards thee" (that is, to change "to" into "towards", which could be monosyllabic) but the Folio compositor might have thought it had to be disyllabic and hence cut "thee". Searle's fifth example, not discussed by Taylor, is " Astronomers foretell it, it is prodigious" (Q K2r) versus " Astronomers foretell it, that it is prodigious" (F ¶¶4v). Searle suggests that the annotation was supposed to make Q read "Astronomers foretell it, that is prodigious" (that is, to change the second "it" to "that", but instead the "that" was put between the two "it"s. Searle's sixth example is a combination of Taylor's 18th and 19th, "Vlis. Shee will sing any man at first sight. / Ther. And any man may sing her, if hee can take her Cliff, / She's noted" (Q K2r) versus "Vlis. She will sing any man at first sight. / Ther. And any man may finde her, if he can take her life: she's noted" (F ¶¶4v). Searle suggests that the annotation was supposed to make Q read "And any man may sing her if he can find her cliff: she's noted", that is to change "take" to "find", but instead the word "sing" was changed to "find". Taylor claimed that both variants (sing/finde and Cliff/life) are misreadings, but Searle points out that a man who could read "sing" in "Shee will sing" should be able to read it in the next line also, and if he could not he must have been a dunce not to read around for context to make sense of the tricky word. The King's men would not have used a dunce for a scribe. Of the second change ("Cliff" to "life") Searle does not give his view, but since "take her Cliff" does not make sense, presumably compositor H altered it to "take her life"; or perhaps the cliff/clef/cleft bawdy pun works as well with "take" as "find". Searle's seventh example, not discussed by Taylor, is " Bid me do any thing but that sweete Greeke" (Q K2v) versus " Bid me doe not any thing but that sweete Greeke" (F ¶¶4v). Searle suggests that the annotation was supposed to make Q read "Bid me do anything but not that, sweet Greek", that is the word "not" was supposed to go between "but" and "that" but instead it got placed between "do" and "any", making a double negative (so, positive) that Cressida does not mean. Searle concludes this section by observing that errors like these are what we would expect "a not-very-bright compositor" to do when confronted with "the annotated text of a difficult Shakespearean play", so the problem resides in the compositor's mistaken interpretations of the interlineations, deletions, and other marks on the annotated Q he was given, whereas Taylor thought that the manuscript used to annotate Q had the errors because the scribe of this manuscript misread the authorial papers (p. 512).

    The foregoing reduces Taylor's list of what he thought were graphic errors in the making of the fair copy later used to annotate Q to form copy for F, but under what circumstances would a compositor make mistakes that someone like Taylor could misunderstand as graphic errors? Searle answers: when his sheets are not getting proof-correction, or are miscorrected without reference to copy, and in these cases if the error generates a word--or what resembles a word and gets altered to one by press-correction-- we cannot see what went wrong and are likely permitting these interlopers in our editions (p. 513). Having revised downwards the number of graphic errors that happen to fall in compositor H's stints (by instead calling them compositorial error), we can recalculate the average number of times we should expect graphic error to hit compositor H's stints, and the answer is 1 line in 103 (down from 1 line in 82 calculated on page 507). But are they spread evenly? No, on quire ¶ he sets 3 putative graphic errors (just 1 graphic error for every 208 lines), but everywhere else he sets 1 for every 84 lines. Of those 3 on quire ¶, one of them (Taylor's number 7) is "greater hulkes" (Q) versus "greater bulkes draw deepe" (F) and so arguably not an error at all since a bigger ship is elsewhere in the play called a "bulk" (pp. 514-5). Likewise, of the other two putative graphic errors (in the transcript used to annotate Q) set by compositor H on quire ¶, Taylor's number 5 ("flexure" in Q to "flight" in F) is an example of a common word replacing an unusual one and hence "may readily enough have resulted from somebody's ignorant attempt at improvement". The last one of these 3 on quire ¶, "Fam'd" in Q to "Fame" in F, might well be the compositor's "insecurity--typical in the period--in the handling of terminal d" (p. 515), which is an explanation I do not understand. It is surely unlikely that compositor H read the quarto word "Fam'd", thought it must be wrong, and so changed it to "Fame". From Hinman's analysis we know that compositor B and compositor H worked together on quire ¶, the first for Folio Troilus and Cressida, and they continued working together until the middle sheet of the second quire, ¶¶3-4, where compositor A set nearly 200 lines. Then compositor B was called away and compositor H had to work alone, and thus he did his own proof-reading, which is why his error-rate dramatically rises. Looking at different states of the Folio separated by press-correction, we can see what errors presumably made by H were caught by B's proof-reading, including a couple that cannot be explained by misreading but only "the compositor's insecurity in the handling of dramatic materials". One is the failure to give to give a speech prefix for Pandarus's "In good troth it begins so" (TLN 1587) in the uncorrected state of F (¶5v) so that it looks like the continuation of Paris's preceding speech. Presumably the whole line ("Pan. In good troth it begins so"), absent in Q (F1r), was added as a marginal annotation to the copy of Q used to set F, and compositor H--not good with drama--missed out the speech prefix. Likewise in the uncorrected state of F (¶5v) there is an obviously wrong "Exeunt Pandarus" that in the corrected state becomes "Exit Pandarus", so presumably the annotation to Q, which lacks this direction entirely (F1v), was something like "ex. Pan." and compositor H did not know enough about drama to expand this correctly, but the proof-corrector did (p. 516). Likewise setting the last line of the play, compositor H gives Pandarus an "Exeunt" which was allowed stand because no-one checked it. Compositor H made 5 errors of simple repetition of words, all in quire ¶¶, against compositor B's total of one in all, and one of them seems to have got miscorrected: Q has "And like dew drop from the lions mane, / He shooke to ayre" G3r, and Folio has "And like a dew drop from the Lyons mane, / Be shooke to ayrie ayre" (¶¶1r), so presumably in F "ayre" got repeated by dittography and then the "ayre ayre" mistake was sophisticated into "ayrie ayre". There are two items remaining on Taylor's list that are similar compound errors: Q's "He that takes that doth take my heart withall" (K3r) that becomes in the Folio "He that takes that, rakes my heart withall" (¶¶5r), of which the take/rakes error might simply be a compositor misreading of his Q copy (p. 517).

    Taylor's error number 18 is this " With the rude breuity, and discharge of one," (Q H3v) versus " With the rude breuity and discharge of our"
(F ¶¶2v). Searle suspects that compositor H, reading his own proof, found he had a turned letter making "one" into "oue" and realizing that "oue" is nonsense changed it to "our". All the rest of the items in Taylor's list could be caused by misreading of printed copy, or are single-letter errors ('literals') that any compositor can make, or miscorrected literals. The problem has been that compositor H's incompetence was covered up by the good proof-reading of compositor B who checked his work, and in any case his incompetence only shows up when he has annotated Q for copy. When he is working from clean Q, compositor H he is mostly fine, although when has not got compositor B reading his proofs he again fails, setting "yong Diomed" for "yond Diomed" at TLN 2563. Three currently available editions of Troilus and Cressida (Taylor's, Foakes's, and Bevington's) accept Taylor's view that the manuscript used to annotate Q to make copy for F was a scribal manuscript, and hence although they prefer F for substantive variants (because it is a revised version), they prefer Q for indifferent ones on the grounds that its stemma is shorter and thus less likely to have "merely casual or accidental error". But Searle has shown by overwhelming evidence that there is no reason to suppose that the manuscript was scribal rather than authorial (p. 519).

    Studies in Philology contains two articles of interest. Kenji Go argues for the acceptance of a reading from the 1609 quarto of the sonnets (2001). The quarto version of sonnet 106 reads "So all their praises are but prophesies / Of this our time, all you prefiguring, / And for they look'd but with deuining eyes, / They had not still enough your worth to sing", and usually "still enough" is emended to "skill enough". Go defends "still" as an adverbial use (meaning 'nevertheless') that predates the OED's first example by a century. But "enough" of what? Go thinks the referent is either all the descriptions that the sonnet mentions, or "praises" earlier in the same sentence. Indeed, "praises" is the key idea of the sonnet, but Go admits that the two words are quite far apart. On the other hand, they are in the same quatrain and sentence, and the conceit is 'not praises enough to sing your worth', and that praise/worth link runs throughout the sonnets to the fair Youth, and Go traces it in seven other sonnets, The Two Gentlemen of Verona and the Shakespearian part of Edward 3 (pp. 118-21). Go explores why previous defenders of "still", George Wyndham and C. J. Sisson, failed: they did not spot "praises" as the implied complement of "enough" and they saw a sharp antithesis in the sonnet between the talents ("tongues" or skills) of the old poets and the talents of the modern ones. But for Go the real antithesis is between those being praised in the past and the impossibly perfect boy being praised in the present, and this also deals with the objection that the word "For" beginning the final couplet makes it a non sequitur (since it really needs "But"), since the final couplet repeats the excuse of lines 11-12 that the Youth is too perfect to be praised according to his worth (pp. 122-5). There remains the objection that "still" wasn't used to mean 'as yet' before 1632 or to mean 'nevertheless' or 'however' until 1722, but there was 'as formerly' which would roughly do here, although it would be more likely if the word order were "They still had not enough" rather than "They had not still enough" as Q has it (p. 127). Go, however,  has clinching evidence from Cymbeline: "[INNOGEN] The thanks I give / Is telling you that I am poor of thanks, / And scarce can spare them. CLOTEN Still I swear I love you. / INNOGEN If you but said so, 'twere as deep with me. / If you swear still, your recompense is still / That I regard it not. CLOTEN This is no answer" (2.3.86-91). This first of these uses of "still" is obviously in the sense of 'nevertheless', which is why editors since the eighteenth century have put a comma after it, and Innogen's answer is not very witty unless "still" carries the senses of 'nevertheless' and 'always' and she is quibbling (pp. 133-4). Go explores the less-than-compelling arguments against reading "skill" in sonnet 106 and decides that the best reason to keep "still" is the "inexpressibility topos" that runs through quite a few sonnets, contrasting the dumb I-poet of the sonnets with other loquacious poets, of which 106 is the last and culmination (pp. 137-40). Since the point of the sonnet is the lack of praise of the Youth, the ellipsis ("still enough" meaning 'still enough praises') is entirely appropriate as another example of the I-poet being dumb (p. 141). In an appendix Go considers that fact that two early manuscripts read "skill", but they are not of value, being from the 1630s or later and possessed of obvious errors.

    John Klause argues that Shakespeare's King John was influenced by a number of works by the Jesuit Robert Southwell. (2001). Klause finds verbal parallels between Louis the Dauphin's language about loving Lady Blanche because he is reflected in her eyes and Southwell's poem Saint Peters Complaint on Christ's eye, and perhaps the Bastard's mocking of it with images of hanging, drawing, and quartering (1.2.497-510) draws on Shakespeare's knowledge that Southwell himself was hanged, drawn, and quartered in 1595 (pp. 404-5).  Klause lists a collection of collocations linking Saint Peter's Complaint and King John (pp. 406-7) and argues that the latter also owes something to Southwell's Epistle of Comfort, since King John's use of a couple of biblical quotations (from Psalms and Galatians) is odd until we realize that Southwell too put them together. Likewise the language of the scene in front of the walls of Angiers follows Epistle of Comfort's description of the destruction of Jerusalem, and there are some looser connections too (pp. 408-17). Cardinal Pandulph's speech to the French king in 3.1 about which of several obligations in an oath must be kept comes from the Epistle of Comfort and Shakespeare's writing just after John's defiance of Pandulph (3.1) borrows a lot of words and phrases from Epistle of Comfort, none of which matches what is in Troublesome Reign, although for the actual defiance Troublesome Reign matches King John closely. Klause explores some phrases that Epistle of Comfort, Troublesome Reign, and King John have in common and observes that Epistle of Comfort "shares nothing of significance with TR except what King John has in common with both works", so there's no possibility of descent by Epistle of Comfort to Troublesome Reign to King John, but there might be linear descent by Epistle of Comfort to King John to Troublesome Reign or else Epistle of Comfort to King John and Troublesome Reign to King John (p. 417n21). Another Southwell work, An Humble Supplication, circulating in manuscript also "scatters its language throughout Shakespeare's play" and it was written in response to a government proclamation against Catholics of November 1591, and Klause lists the (rather weak) verbal parallels (pp. 419-22). Klause suspects that the putting out of Arthur's eyes (as a means to kill him) with hot irons came from Southwell too. In Troublesome Reign Pandulph says that whoever kills the king will be forgiven the sin, but that is doctrinally flawed from a Catholic point of view since forgiveness requires contrition of the sinner, religious authority is not enough, and it certainly cannot forgive before the fact; thus Shakespeare (presumably informed by Southwell) changed this so that killing the king is a virtue not a sin at all (p. 424). If these borrowings are accepted, the earliest date for King John is whenever An Humble Supplication was written, and since An Humble Supplication was a response to a proclamation of November 1591, King John cannot be earlier than, say, December 1591. Just possibly, King John was written early in 1592 (counting January-December) and was imitated in Troublesome Reign, which got into performance and print before 25 March 1592, in which case its title-page dating of 1591 is counting March-March (p. 425n34). For Klause, the most important conclusion of all this is that Shakespeare "welcomed a Jesuit into his mind" (p. 426).

    This reviewer's stint began with work done in 1999, and it has been until now impossible to obtain volumes 12 (1999) and 13 (2000) of the journal TEXT, so relevant essays in those volumes will be reviewed here. Volume 12 contained nothing of interest to this section, but Volume 13 had three articles. Volume 14 (2001) would normally be reviewed here, but because it was unavailable at the time this review was written it will be noticed next year. W. Speed Hill considers the impact Shakespeare has had upon editing (2000). If there were only the pre-1593 Shakespeare plays in existence, he presumably would not have become the national poet, and Hill thinks that the works of Spenser, Sidney, Jonson, Beaumont and Fletcher, and Middleton would not have generated the textual theory we have, either because until recently they were not thought to be important enough, or the early texts just are not difficult enough to edit. Indeed it is unlikely that Shakespeare' contemporaries would have been edited in the monumental editions we have of them, for this was in many cases preparatory work for doing Shakespeare himself. Other candidates for 'national poet' in Shakespeare's absence, such as Milton and Wordsworth, do not need much work to establish the text, rather the energy goes into the glossing and commentary. Donne left us mostly manuscripts (whereas Shakespeare left us primarily books), but the importance of books in Shakespeare studies made editors of Donne prefer his 1633 and 1635 printings even though they knew that the lost manuscripts used as copy for printing them were further down the stemma than some surviving manuscripts. This is an important effect the editing of Shakespeare has had on the editing of others (pp. 6-7). Looking to earlier writings, Tim William Machan argues that the Shakespeare editing tradition has imposed an inappropriate notion of authorship upon medieval literature (2000). For seventeenth- and eighteenth-century readers, Shakespeare was more easily made the father of English poetry than his medieval forebears since he was not tainted by coming before the Reformation and he also, conveniently, wears his learning more lightly. Shakespeare was valued for creating "cultural empathy", what Johnson called the "faithful mirrour of manners and life", while medieval literature apart from Chaucer was virtually apologized for (in phrases about rustic charm and simplicity) when published (pp. 12-4). For medieval literature the problem was straightforwardly and narrowly how to make the difficult language accessible without losing accuracy, whereas the editorial problem in Shakespeare was beginning to span many cultural domains, including conceptualizations of the author, reader, and critic (pp. 16-7). Because of the textual-critical tradition, Shakespeare comes down to us "in a resolutely unilinear fashion, monogenetically descending from either the first folio or one of the early quartos", which leads to an irony. Early editors of Shakespeare began by treating the Folios like classical manuscripts with independent authority (as Mowat also observes in a chapter reviewed below), and only later did editors fully realize the fact of monogenetic descent that made this invalid. Then, when later editors approached medieval works they wrong treated these as monogenetic when in fact they are more like classical works in their polygenesis (p. 19n23). Despite their polygenesis, medieval works got edited like Shakespeare works in the nineteenth century and multiple readings in manuscripts of equal authority have been reduced to single readings, and the singular author invented as the source of authenticity (p. 20). Without Shakespeare we would now understand medieval literature via multiple manuscript descendance and with the originating agency of not one but a collection of voices (p. 22). Rather than focussing on the print culture that produced Shakespeare, we would be thinking about manuscript culture and its relation to "late-medieval England's diglossia" of the simultaneously existing low-English-oral tradition and high-Latin-literary tradition. When we retrospectively apply our notion of authorial paternity to Chaucer and Langland, treat their agency as a given, we obscure the fact that Chaucer and Langland were insisting on the validity of their authorial consciousness precisely in opposition to a medieval culture that denied this to vernacular writers. Of course, authors do not alone create the idea of authorial agency (whole cultures do that), and the sense of relatively stable agency that we take back with us to Chaucer is "more the creation of a Shakespearian-focused textual criticism than a historical medieval reality" (p. 23). This anachronism is being addressed in medieval textual studies, and the conception of authorship used by editors is rightly being historicized; this reorganization is not mere relativism or decadent postmodernism, but good historicism (pp. 24-25).

    In a major work on the history of twentieth-century Shakespeare editing, Paul Werstine shows that New Bibliography was never a matter of consensus and that McKerrow in particular rejected Greg's theory of copy-text; thus the New Textualists (not Werstine's term) are squarely within rather than without New Bibliography's diverse field (2000).  In the matter of 'accidentals' (by which he meant spelling, punctuation, word-division, and so forth) Greg, in "The rationale of copy-text", advised ignoring your copy text and following whichever text you think best represents the author's habits, for there were no standard habits for the sixteenth and seventeenth centuries and Greg denied competence outside this period. Yet Fredson Bowers promoted Greg's 'Rationale' principles for editors of nineteenth-century works and thus a Shakespeare-centered way of thinking dominated other periods, even though Greg's ideas did not even get realized in a Shakespeare edition, the closest thing to a realization of them being Bowers's edition of Dekker. Rather, Shakespeare continued to be presented in modernized spelling and punctuation rather than the 'accidentals' of early printing as Greg advocated (p. 27-8). Greg's ideas generally, and in particular his confident distinction between printings made from foul papers and those made from promptbooks, continue to inform modern Shakespeare editions. Recently the editing of Shakespeare has been decentred, starting with the Oxford Complete Works which was "edited without reference to poststructuralism" and yet was a decentring project nonetheless (p. 29). (Actually, I would argue that Wells and Taylor got to poststructuralism by a relentless pursuit of the empirical, as Derrida's ideas would indeed predict.) Wells and Taylor largely followed Greg, but crucially they admitted all readings that might have got onto the stage, not merely the authorial ones; in this they were effectively following Jerome McGann's A Critique of Modern Textual Criticism (1983) in their conception of the socialized text. Since McGann was actually concerned with the Romantic poets in this book, we might say that Shakespeare was thus edited by reference to Byron (p. 30). Greg got his confident author-centered editorial principles from his belief that he could see authorial foul papers in Hand D of Sir Thomas More, and although he occasionally mentioned the possibility of scribal transcripts, he suppressed this possibility until he had formed his model of textual transmission based on the binarism of foul papers and promptbooks, the former too untidy to use directly but kept in the playhouse (and maybe annotated lightly) where they were transcribed to make the latter. Although in discursive writing Greg would admit that the binary was not terribly stable (since promptbooks might contain foul-paperish features), in practice, when determining underlying copy for particular printings, he discounted this possibility (pp. 31-2). Thus Greg would argue that the absence of evidence for promptbook copy means that the copy must have been foul papers, and since a company would guard its promptbook ("which may have been" the one containing the censor's licence), Greg assumed that what went to the printers must have been foul papers, which therefore take us close to the Shakespearian manuscripts. Bowers, on the other hand, argued that nothing in Henslowe's Diary showed a dramatist handing over his foul papers, rather he probably handed over a fair copy and kept the foul papers for himself. Bower suggested that there were more than two types of possible copy for printers, and Orgel has since argued that dramatic collaboration does not even locate the author at its centre, the text is just a "working model" of the play (p. 33).

    The expression 'foul papers' could mean specifically those that were incomplete, so Greg's sense of them as "a complete authorial manuscript" is quite wrong (p. 34). This is not quite fair, for there is only Knight's transcript of Fletcher's Bonduca standing as an example of 'foul paper' incompleteness. Greg's term 'promptbook' was anachronistic and not analogous to 'book' in the period (as he claimed), and indeed the entire binary classification Greg erected has been shown to be invalid, for manuscripts he called promptbooks have features that he claimed were unique to foul papers. Werstine thinks that there is "no consensus" now about whether Hand D of Sir Thomas More is Shakespeare and he cites Howard-Hill's tentative summary of the situation: "none of them [the contributors to this collection of essays] believes that the case for Shakespeare's presence in the More manuscript is less strong than that which could be made to deny it or to identify another playwright". This is not a fair quotation of Howard-Hill's view about the consensus of opinion, for he goes on to say that the hypothesis of Shakespeare's authorship and ownership of Hand D is supported by "separate but convergent lines of enquiry conducted by scholars of pre-eminent skills and authority" and "cannot be met by simple denial or doubts as to its adequacy", which is precisely what Werstine is doing (p. 34). Some modern editors have, then, abandoned Greg's narrative and his hope of determining underlying copy for printings, and of determining the relationship of that copy to Shakespeare's manuscripts. Werstine quotes Henry Woudhuysen's Arden 3 Love's Labour's Lost on the uncertainty about theatrical manuscripts generally and how, if the 'foul paper' features could get past the printers and into print (as New Bibliography insists) then equally they could be transmitted into scribal copies. This is true, but the argument is usually run in the opposite direction: scribal transcripts contain noticeable things that would not be in authorial or theatrical copy, such as act intervals and Latinate labels. Werstine lists some recent editions that show the influence of his type of uncertainty, which leads to giving the reader more than one early text (as with Jill Levenson's Oxford Shakespeare Romeo and Juliet reviewed here last year) or to conflation with clear markers to show provenance and to show editorial construction work, as with R. A. Foakes's Arden 3 King Lear and Mowat and Werstine's New Folger Shakespeare series. Thus, claims Werstine, editing Shakespeare has become usefully decentred from the man and recentred on the early printed texts themselves (p. 36). Is this lack of an editorial consensus a new or old condition? Tanselle says it is new but Werstine shows that it is old. Tanselle represents New Bibliography as an evolution: McKerrow was opposed to eclecticism but in his Prolegomena for the Oxford Shakespeare he showed that he would be willing to emend an early text with variants from a later one if the variants were accepted as a unit, and then Greg (in The Editorial Problem and later "Rationale") went the next step and argued for allowing editorial judgment to choose on each variant individually. But this is merely to repeat Greg's version of the story, and he understated the extent to which he and McKerrow disagreed and the divergences are all the more apparent if one brings in John Dover Wilson (p. 37). Although McKerrow often left implicit who he was disagreeing with, we can reconstruct the ongoing progression of the arguments.

    McKerrow in Prolegomena disagreed with Greg's notion of foul papers (holograph copy of a play in its final form) for he thought that if the players had it, they would write on it. McKerrow was not convinced Hand D was Shakespeare, so unlike Greg he did not think that he knew what a Shakespeare holograph would look like, and hence he did not think we could know Shakespeare's habits in spelling, capitalization, italicization and pointing (pp. 38-9). McKerrow also distanced himself from John Dover Wilson and Arthur Quiller-Couch's idea that many surviving printings of Shakespeare's plays incorporated parts of lost plays and from Wilson's view that for some plays one person copied the text and another the stage directions. Wilson developed his elaborate model of repeated interference in the text of Folio Hamlet via scribe P who made Shakespeare's manuscript (which underlies Q2) into a promptbook, and scribe C who subsequently revised this promptbook to make the text that underlies F1; Greg wholeheartedly embraced his narrative. Overall, in Prolegomena McKerrow put as much distance as he could between himself and Wilson and himself and Greg about the possibility of inferring copy from printed text: to McKerrow it was hopeless (pp. 40-2). McKerrow alluded to a careless playwright sending individual sheets to the theatre as he writes them, which must be Robert Daborne's letter to Henslowe ("J send you the foule sheet & ye fayr I was wrighting as your man can testify") that Werstine interprets as showing how unusual it is to be giving Henslowe a foul sheet and "how he feels obligated to provide better copy" (p. 43). I cannot see these things in the letter, only the sense that Daborne had to be pressed for it. Another cause of non-authorial (and hard to remove) errors and inconsistencies in a play imagined by McKerrow was incomplete revision, and Werstine says that Sir Thomas More and Sir John van Olden Barnavelt are the most famous examples of "incompletely revised manuscripts" (p. 44). It would be fairer to call these 'allegedly' incomplete, since Werstine himself has long insisted on the impossibility of our knowing just what it took for a manuscript to be considered ready for the theatre. McKerrow, having no confidence in our ability to determine the underlying manuscripts of the early printings, felt that copy text should be chosen on the basis of the overt qualities of the extant printings themselves, specifically "carefulness and freedom from errors" rather than covert features of them informing a doubtful theory about their provenance. Thus, ". . . McKerrow's understanding of authority is documentary rather than metaphysical" (p. 46), which is a odd opposition; has "metaphysical" become again an abusive epithet to throw at people that you think are being unreasonable idealistic? Unlike Greg, McKerrow thought it impossible to tell if revision apparent in a later reprinting is authorial or not, and overall his advice to editors was surprisingly close to modern New Textualism: one should determine "the most authoritative text" that we have, and "reprint this as exactly as possible save for manifest and indubitable errors" (Prolegomena p. 7). After of year of claiming that New Bibliography is self-deluding, Werstine now appears to think that McKerrow was a New Textualist, but only by splitting McKerrow from Greg and Dover Wilson; I suspect if he looked closely at the latter pair he would find something of the New Textualist acceptance of indeterminacy in their work. To be fair, Werstine explicitly denies that he thinks McKerrow as a proto-New Textualist and he now turns McKerrow's inconsistencies.

    Greg and Bowers pointed out that McKerrow's claim that we cannot really know what the copy text of a printing was is in contradiction to his "Suggestion" (1935), alluded to in Prolegomena where he repeats its principle. They were right, but Greg was inconsistent too: in The Editorial Problem in Shakespeare he claimed that for Hamlet we can know that Q2 was printed from foul papers and F from promptbook and that this can be known for other works too, yet elsewhere in the same book he says that we do not know this for any other play. In the case of Hamlet Greg dithered between accepting Wilson's certainty (Q2 is from foul papers, F is from promptbook) and remarking on the problem that Q2 has signs of the prompter: "Drum, trumpets, and shot. Flourish, a piece goes off" in Q2 looks like theatrical annotation creating repetition, while in F there are stage directions not easily derived from Q2 (which they should be for Wilson's theory) and moreover the stage directions are indeterminate in a way a promptbook's should not be (pp. 48-50). Greg's analysis showed that Wilson's categorization of Q2 Hamlet being printed from foul papers and F Hamlet being printed from promptbook was deeply problematic, yet this was the only example Greg could offer of our being able to tell underlying copy for two substantive versions of a Shakespeare play. Bowers strongly attacked Wilson's methods and also attacked McKerrow for overreacting to Wilson and being too conservative, but Bowers too seems to flit between demolishing Wilson's edifice about Hamlet and adopting it. Twentieth-century editors are really back where McKerrow was in having to reject Greg's theory of the copy-text and falling back on the pragmatism that McKerrow shows in his Prolegomena (pp. 51-2). Thus, we should reject Tanselle's narrative of early-twentieth New Bibliography being a consensus that spread beyond Shakespeare editing and see it as a conflictual field from the start. Otherwise, those who reject Greg's theory (New Textualists) seem cut off from the New Bibliography tradition, whereas in fact they are within its diverse field (p. 53)

    Werstine's narrative of New Bibliography's capacity to incorporate the New Textualism is markedly at odds with that told in Barbara Mowat's survey of editorial thinking from the sixteenth century to the present day (2001). Mowat begins by insisting on the consensus Werstine denies: "For much of the twentieth century . . . editors and textual critics accepted and depended upon a single larger story . . ." (p. 13) and she turns to Thomas Kuhn's model of 'paradigm shifts' to distinguish the pre-twentieth century pessimism regarding our ability to determine the underlying copy for printed texts from the New Bibliographical confidence about this (p. 18). Yet the New Bibliographical paradigm "maintains its hold on the reproduction of Shakespeare's text" (p. 24)  which can only be explained, Mowat claims, by the absence of a new paradigm to take its place. Where Werstine hopes to show that New Bibliography can accommodate the disagreements that he and others have with Greg's binary thinking--which is admittedly a surprising position for Werstine to adopt after years of apparently self-imposed exile from the tradition--Mowat looks to the New Textualists for "a future in which a new paradigm may be established" (p. 26). If there has to be accommodation, it is New Bibliography that must give ground and "find a way to explain and absorb the factual and theoretical challenges to its hegemony", or else "editing may flourish in the absence of any accepted paradigm" (p. 26), which last comment rather suggests that she does not fully accept the implications of what Kuhn meant by a 'paradigm', which was something we cannot do without.

    Jeffrey Masten argues that the binarism of author's hand and alien hand is deconstructed in Sir Thomas More and was not at all stable in Shakespeare's time (2001). Like Werstine, Masten notes that the rising importance attached to the manuscript of Sir Thomas More was closely related to New Bibliography's category distinctions of good and bad quartos and foul-papers versus promptbooks, and he thinks that we seek in this manuscript an authorial integrity witnessed in a hand that in fact the document denies in its dispersal of authority (p. 112). The problem is our failure to think up new ways to deal with editing collaborative works (as most plays of the period are) and we continue to edit the person (Shakespeare) not the work, and we think of that person as singular (p. 113). Attribution studies, which takes identity as a fact, does not take seriously enough the ways in which habits are emulated, adopted, adapted, and thrown off; 'hands' in writing should not be understood as synecdoche for persons but as metonymy for writing (the process, not the person) (p. 115). We are still treating collaboration in an Old Historicist way: if only to keep the unity of the persons, we carve up the play, we disintegrate it to differentiate them. Rather, Masten exhorts, we should be New Historicist in historicizing not only the text but also "our models of agency, individuality, style, corporate effort, contention, influence, and so forth" to put those "within the realm of the discursively social", and we might have to invent new editorial apparatuses and practices to do this. This would be to get back some of the sense of "individual" that Raymond Williams drew our attention to in Keywords, that of indivisibility from the group, and would concentrate on the social whirl of interpenetrating texts and practices (as the New Historicists have insisted on) that not only made texts but also made text-makers (p. 116). These are undoubtedly laudable aims, but while Masten gives a brilliant reading of the play Sir Thomas More (which will be only briefly noted here) he signally fails to invent new editorial apparatuses and practices that might achieved the desired ends.

    Sir Thomas More  begins with a scene about distinguishing English property (women, food) and foreigners' property and in the quelling scene More does a "radical denaturalization" that "places its hearers (onstage and off) in a position of cross-identification that resonates throughout the play" (p. 117). As Masten admits, this might in fact seem "a simple and transparently conservative move", and I would agree; conservatives know the power of the injunction to 'do unto others'. But the particular reversal of places entails a breaking of sumptuary laws ("you in ruff of your opinions clothed") since commoners were not supposed to wear ruffs (p. 118). Masten is wrong: law limited the size of ruffs, not who could wear them. The "cross-class-dressed 'ruff' comes back to redress itself" when More argues that other "ruffians" will "shark on" the rebels if the topsy-turvydom of rebellion is allowed succeed. Of course, More himself does cross-dressing, shape-shifting (in the play-within-the-play) and finally becomes a 'stranger' on the scaffold. Stripped of his titles, he becomes "only More", a name that "signifies a cross-identification between lack and excess" and all the more so pronounced, as it was then, "one-ly More" (p. 119). The point of this reading is that traditional scholarship, which divides the text into 'hands', "has stopped us from reading its continuities and theirs" (for the themes run throughout the play), but I would counter that Masten undoubtedly did his sparkling reading on the basis of existing editions--which served him well--and indeed that he could not assert that the continuities were continuities had not previous editors divided up the hands in the first place. In this period before coherent national identity, the various terms for otherness (alien, denizen, foreigner) were vague and they overlapped, and foreigner could mean just 'not from this parish', or 'recently arrived in town from the countryside', as indeed Shakespeare was (pp. 120-1). A particularly fraught notion was the 'denizen', which meant a 'native' and also someone who lived 'within' (dans) and had certain limited rights to work; geographically as well as in the play, the boundaries of strangeness are not clear and Tilney's efforts in rewriting the play are at least partly to make the play specific (he alters "stranger" to "Lombard") and yet also general ("English" to "man"). Masten ends with a series of summarizing bullet points (pp. 122-3): there are native Secretary hands and foreign Italian hands in the manuscript and indeed literate men learned to write either and might mix them up (as Greg complains Hand S does), so in form too it encodes the deconstructive stranger/native theme. (Masten does not call it deconstructive, but that is the essence of what he argues.) The word 'stranger' was itself an only-partly assimilated foreigner at this time, having come from France in the late-fourteenth century, and Hand D's spelling of it as "strainger" was a strange Scottish spelling used by James 1, the foreigner monarch possibly on the English throne by the time of the play's revision. Such an interpenetration of the play and its linguistic and political contexts is, for Masten, what we should be concentrating on instead of parceling up work into originating author(s) and "alien hands" as though these were stable terms.

    N. W. Bawcutt argues against the wrong-headed and anachronistic sense of the 'socialized' text posited by Stephen Orgel, David Scott Kastan, W. Speed Hill, and Paul Werstine (he might have include Jeffrey Masten too), pointing out that the modern idea that theatre men had no literary ambitions is contrary to the evidence (2001). In fact, dramatists did want their plays printed and without the actors' cuts, as shown by the title-page of Every Man out of His Humour (1600) which claims that the contents show the play "As It Was First Composed by the Author B. I. Containing more than hath been Publickly Spoken or Acted", the title-page of The Duchess of Malfi (1623) which claims that it contains "diuerse things Printed, that the length of the Play would not beare in the Presentment", and the title-page of Barnes's Devils Charter (1607) which claims that the contents were "more exactly reuewed, corrected, and augmented since [performance] by the Author, for the more pleasure and profit of the Reader". Jonson, Webster, and Barnes might be dismissed as bookish and literary, but Brome cannot and his The Antipodes (1640) ends with a note to the reader saying that the printing included all the things left out of performance, "inserted according to the allowed Original". To be fair, all these examples are discussed in David Scott Kastan's book reviewed above, and he holds his views in spite of them. Orgel claims that "If the play is a book, it's not a play" but that is not so, and indeed in The White Devil (1612) Webster praised the performers while condemning the ignorant audience and in The Devil's Law Case (1623) he asserts the mutual dependency of writer and performers (p. 5). Likewise Marston in his preface to The Fawn (1606), insisted also that while comedy did not read well, his tragedy would withstand "the most curious perusall". This interest in printing is common in the period, Fletcher and Shakespeare being the exceptions to the rule: Brome, Heywood, and Shirley wrote about printing as a natural succession to performance, the classical drama was known only because it could be read, and a number of dramatists referred to their plays as poems, a genre normally published (pp. 6-7). Commendatory poems by fellow playwrights preceding the 1623 text call it "his Dutchesse of Malfy", Webster's not the company's, and treat the printing of it as an appropriate monument to secure Webster's posterity (p. 11).

    So much for ownership, what of fidelity? Jasper Heywood's preface to his translation of Seneca's Thyestes (1560) complains that an early work of his was "corrupted" in the printing house, so accuracy is not a recent concern being foisted onto the period; it was already there, and Bawcutt cites complaints by Jonson, Chapman, and Heywood against bad printing (pp. 12-3). True, printers mixed corrected and uncorrected sheets, but this was due to thrift and inefficiency, not a poststructuralist concern for fluidity, and the address of 'The Printers to the Reader' of Thomas Urquhart's Epigrams Divine and Moral (1641) shows that, contrary to Orgel, the printers did idealize the final, perfected text even if they could not, for reasons of economy, make one. The printers explain that they include a full list of errata even though (because of the press correction) not every copy will have all the errors, for they were "willing rather to insert the totall, where the parts are wanting in their distinguish't places, then by omitting any thing of the due count, to let an errour slip uncorrected". Equally aimed at perfection were the requests in many books that the reader go through and make the necessary corrections in pen (pp. 14-15). There was a tradition of authors going through and making corrections in copies of their books they wanted to present to someone, as Massinger did for The Duke of Milan given to his patron Sir Francis Foljambe. We can summarize that plays were considered by their writers as their own intellectual property, they were not always happy with what the players did to them, they frequently arranged to publish them, and they would be amazed at our modern veneration of printing errors (pp. 16-7). The new notions of the socialized text overthrow "two centuries of patient and disinterested efforts to purify and clarify texts" that the Renaissance authors would have thanked us for (p. 20).

    John Jowett argues that the discussion about foreign war news in Measure for Measure 1.2 would have been highly topical in 1621, which is when Middleton added it to Shakespeare's old play for revival (2001). By 1621 the play's disguised-ruler topicality--arising from the accession of the unknown quantity James 1 in 1603--would have been decidedly old-fashioned. Middleton's addition of the material at the beginning of 1.2 about the King of Hungary's peace alludes to Protestant resistance to Counter-Reformation Catholicism in Europe, the Thirty Years' War and the Palatinate wars, and Middleton did this sort of thing to other works at the time. (p. 230). Vienna was ruled by the Catholic Habsburg emperor Ferdinand 2 but in 1621 it was under attack from the Protestant King of Hungary (Bethlen Gabor) who wanted to free it from the Catholic Habsburg empire; Middleton changed the location from Ferrara to Vienna to take advantage of these events. James 1's son-in-law Frederick, the Elector Palatine, was proclaimed King of Bohemia in 1620 and formed a league with Bethlen, King of Hungary (p. 231). So, the politics of this Austrian war were topical in London, and what seems to have been of special interest was detail of the accords between princes, hence Lucio's reference to the dukes coming to "composition with the King of Hungary" (p. 232). But the alliance of Bethlen Gabor with Frederick the Elector Palatine was not viewed with complete equanimity in England, because Bethlen was Turkish. James 1 asked Frederick to break off the alliance, and his hostility to Bethlen lies behind "the tension surrounding Lucio's allusion to a truce between Vienna and the King of Hungary" (p. 233). Negotiations to end the conflict were reported as in-process in a news-sheet of 6 October 1621, and presumably it was just after this that Middleton wrote his lines about "If the Duke with the other dukes comes not to composition", since if the audience knew the outcome of the real negotiations the allusion would "fall flat". Ordinarily the delays of scripting, licensing, and rehearsal, make such topicality hard to achieve in a play, but in the case of a revival of Measure for Measure these delays did not apply (p. 234). The location of the stage is both London and Vienna, for "Heaven grant us its peace, but not the King of Hungary's" can be read from the Protestant English point of view (an ally of England making an unwanted unilateral peace with the enemy) and from the Catholic Austrian point of view (Bethlen's offer of peace is not to be trusted). Thus the stage is metaphoric, representing what is happening, and metonymic, partaking in the events depicted (pp. 236-7). The freedom to debate the matter of possible war with Spain was a prerogative that parliament was insisting upon and James was resisting as an encroachment on his power to make foreign policy, and indeed James tried to suppress corantos. Middleton's adaptation of Measure for Measure defies his efforts to control public discussion of the matters, and the old play presumably appealed because it relates sexual freedom with freedom of speech (p. 238-9). In an appendix (pp. 240), Jowett conveniently summarizes his earlier-published work on Middleton's hand in Measure for Measure which showed that the unShakespearian oaths and the act divisions must post-date original composition (being later than 1606 and 1609 respectively), that the song "Take oh take those lips away" originated in Rollo, Duke of Normandy written in 1617-20 and was apparently imported to the play with localized revisions to "lock it into the dramatic action", that the "news" passage in 1.2 has Middletonian preferences (has instead of hath, whilst instead of while, ay instead of yes, between instead of betwixt), and that the idea of razing "Thou shalt not steal" from the Ten Commandments occurs nowhere else in Shakespeare or any other dramatist, but crops up thrice in two other Middleton plays. The Middleton parallels come from works written before and after Measure for Measure, so it cannot be that Measure for Measure influenced Middleton to write like this.

    Arthur Ing Freeman and Janet Ing Freeman set out to dispel the myth that James Orchard stole and mutilated a Q1 Hamlet belonging to his father-in-law (2001). The first Q1 was found in 1823, lacking its final leaf, and was reprinted in 1825 in "quasi-facsimile", and the second (presumably recognized for what it was because of the facsimile) turned up in 1856, with its final leaf but not the title-page. Halliwell bought this complete copy at an inflated price through an intermediary having originally turned down the seller in person (pp. 350-1). Because a Catalogue of Printed Books at Middle Hall by Sir Thomas Phillipps (Halliwell's future father-in-law) names a 1603 Hamlet quarto and yet none was known, William Alexander Jackson supposed that this was the one Halliwell later owned, that he stole it from Phillipps, and that he cut out its title-page to conceal the Middle Hall stamp on it. Since Halliwell was barred from his father-in-law's house after he married Henrietta Phillipps, he would have to have pocketed the Q1 Hamlet 14 years (1842-56) before he first showed it to anyone. Moreover, Halliwell's copy had a most distinguishing feature, the interleaving of blank pages on which someone had written some notes from Theobald's edition; if he removed the title-page to disguise his theft, Halliwell would surely have removed these too. Also, since Phillipps was always accusing Halliwell of dishonourable deeds, he would hardly have put up with the loss of his Q1 Hamlet without blaming his son-in-law, especially once that son-in-law publicly announced he had one (pp. 354-7). In fact, Phillipps did not have a Q1 Hamlet even if Halliwell wanted to steal it: the catalogue entry is almost certainly for the quasi-facsimile reprint (pp. 358-9). Whoever interleaved the pages in the first complete Q1 Hamlet used an existing printed edition of the 'good' text and indicated that he sometimes preferred a Q1 reading over others (p. 360). He seems to have worked out, without a title-pages to guide him, that what he had preceded the 1605 and 1611 editions--presumably because their title-page claims to be enlarged texts--so he must have been quite knowledgeable about Shakespeare texts; the annotating is from the period 1726-33 (p. 362).

    In the first of two essays reviewed here, Lynette Hunter argues that the date of Q4 Romeo and Juliet cannot be precisely determined from deterioration of its tailpiece device (as George Walton Williams claimed) and it could lie anywhere within 1616-28, although probably 1618-28 (2001a). In 1965 George Walton Williams suggested that the anonymous and undated Q4a Romeo and Juliet was printed in 1622 by William Stansby (for John Smethwick, as the title-page says), but since Smethwick was involved in the Folio, Hunter thinks this would be odd. After all, the Pavier quartos were stopped to remove competition for the Folio, so why would Smethwick be doing something to hurt his own bigger project? (The rather obvious answer is that Smethwick was invited to join the Folio project late in its development, once it was clear that he already had the rights to a number of the plays.) The dating of Q4 to 1622 was done by looking at the degradation of the tailpiece device on L4v and comparing it to other Stansby books from 1615 to 1623, but Hunter has had several Q4s looked at and "the endpieces are not uniform in appearance in respect of these breaks" and nor are they in other books that use the device. As such, we can only say from the tailpiece that Q4 is not earlier than 1616 (when one certain break does first appear in other books using the device) and not later than 1628 when Stansby stopped using the device, something we know because in 1629 he used another one of his favourite devices in the reprinting of a work that had previously used this one (p. 282). Hunter and her Arden 3 co-editor on the play, Peter Lichtenfels, and their team have found in Q4s only the consistent presence of Williams's break number 1 and the inconsistent presence of the other breaks. Although the evidence is not as clear as Williams first claimed, a cluster of his breaks (numbers 1, 2, and 4) does seem to emerge in 1618 and is present in a few Q4 copies, so that is the earliest likely date for Q4 (p. 283). Watermark evidence is inconclusive, and now that the date of Q4 is up for grabs (anywhere within 1618-1628) the relationship to the Folio is anybody's guess; where we previously thought it was printed around the same time as F, Q4 Romeo and Juliet might in fact have been 4 years earlier or 5 years later and had nothing to do with it (p. 284). Now that we do not know whether Q4 was printed before or after F, the possibilities about how their underlying copies are related, and what this tells us about printing house practices, are all up in the air. There are no parts of Q4 that seem dependent on F as their copy, but there are parts of F that might be influenced by Q4, but if so we have to explain the absence from F of corrections that Q4 makes to Q3 (p. 285). At this point Hunter misuses the expression 'begs the question' to mean 'invites the question' and the reader is pointed to her second paper, about Q4 Romeo and Juliet itself.

    Having claimed that we cannot know what was the copy for Q4 Romeo and Juliet, Hunter contradicts this assertion in a book chapter regarding the variants between Q1, Q2, Q3, and Q4 Romeo and Juliet that is by turns vague, ambiguous, and apparently ignorant of how variation and corruption can occur (as by compositorial slip), and puts everything down to the putative 'intelligence' of a supposed editor of Q4 (2001b). Editors of the play, Hunter notes, usually adopt Q2 but then go on to edit it in ways that make it like Q4 or Q3, so why not use Q3 as copytext for modern editions? Perhaps, she wonders, because it is merely a reprint of Q2 (indeed); nonetheless Q4 has been intelligently edited and should at least be collated in modern editions (p. 9). The signs of intelligent editing are, as we might expect, that it makes the changes modern editors make. There are points where Q4 agrees with Q2 against Q1/Q3, but not enough to suppose that the editor of Q4 had a copy of Q2 in front of him: Q4 is essentially an intelligent reprint of Q3, and it almost always follows Q3 where Q3 has already changed something from Q2. That is to say, where there is a problem that a smart person might try to fix, Q4 uses Q3's fix. Perhaps that is because the theatre people told the editor to trust Q3, or else because the editor of Q4 was the editor of Q3, Hunter speculates. Q3 was printed from Q2, but with a number of changes that suggest that an 'editor' marked up the copy of Q2 first, and this man was not from the playhouse (since speech prefixes and stage directions retain their errors) nor a compositor (since layout and catchword errors get through). As well as these errors there are intelligent changes, some showing access to Q1. At this point Hunter's articles becomes hard to read because she uses line-number references adjectivally, producing such gibberish as "For example, at 5.2 Q2 mistakenly prints 5.2.23 (Exit. when Friar John leaves. Q3 slavishly reproduces this clear error, even adding a bracket to Friar Lawrence's 5.2.30 Exit. (2) (Exit. (3)" (p. 12). I cannot tell which words in those sentences are quotations nor which edition the line-numbers refer to; possibly these sentences have been mangled in their printing. In the next paragraph quotations from quartos are represented by a tiny change in the size of typeface and no quotation marks at all, so it is virtually impossible to see where the quotations of clauses start and finish. There is also fatal vagueness, as in "Approximately 75 per cent of Q1 covers the same ground as Q2, and over half the lines in that 75 per cent are the same if not very similar (Irace)"; what does "covers the same ground mean" if, as seems implied, the lines are not "very similar"? Perhaps this is meant to imply a paraphrasal relationship, or merely the conveying of the same events of the plot. The Q2/Q3 variants are categorized by Hunter under three heads: (A) those occurring in the 25% of Q2 which "has no counterpart" in Q1, (B) those which follow Q1, and (C) those for which Q1 is different again from Q2 and Q3. This last category Hunter confusingly describes as "those which occur where Q1 makes a change" but here the tense and agency are wrong: Q1, being the earliest text, did not change or disagree with anything that we know of.

    Of the Q3/Q2 variants for which we have Q1 text also, nearly half are at places where Q1 has something different again, which statistically is an unlikely coincidence unless the person making Q3 had access to Q1; in other words were the locations of Q3/Q2 variants picked at random, they would not half the time turn out to be places where there are also Q2/Q1 variants. (Actually, one cannot be sure this is significant until one has determined the influence of Q1 on Q2--something Hunter has not done--because if Q1 heavily influenced Q2 then Q1/Q2/Q3 variants might merely be examples of repeated attempts to correct error in a single line of linear descent.) More significantly, of the 100 or so differences between Q3 and Q2, nearly half occur in the quarter of the Q2 for which there is no corresponding section in Q1, so ". . . Q3 will change Q2 twice as many times where Q1 is not there to corroborate as when it is". This does seem significant evidence for consultation of Q1 during the making of Q3, since were Q1 not used we expect only a quarter of the changes to fall in places for which it lacks lines, and the chance of half the changes randomly falling in these places is extremely small. Yet, despite respecting Q1 (in the sense of emending more ambitiously when it is not there to contradict him) the editor of Q3 is not afraid to overrule it even where Q1-Q2 agree (p. 14). From this, however, Hunter leaps to the assertion that Q4 seems to "recognize the authority of the skill involved in editing Q3", following it rather than Q2 for nearly all Q3/Q2 variants. This is a mistake since Hunter has not shown that "skill", only confidence in and respect for Q1, which is not the same thing. One would have to argue that a number of changes are intelligent improvements over previous printings to establish skill, and Hunter has not engaged in any serious debate of the value of particular Q3 variants. Hunter categorizes the 'decisions' of Q4 with respect to variants and agreements between the three preceding quartos and constructs a shaping intelligence at work. For example, she thinks that where Q3 shows a change from Q2, Q4 respects this change and follows it, and only very seldom does it overrule a Q3/Q1 agreement. Looking at occasions when Q4 agrees with Q1 against Q3/Q2, Hunter thinks the policy is one of "allowing Q1 to advise but not dictate", which is pretty woolly, but worse is to come: there are occasions when Q1 differs from Q3 and although Q4 does not abandon Q3 to follow Q1, it "moves in the direction of the spirit of Q1" (p. 15). In all this there is much counting of variation but little effort to account for it other than by a shaping intelligence, yet without detailed consideration of the particular changes the reader has no reason to suppose that the changes are improvements rather than merely errors in transmission.

    Hunter attempts to show that the 'editor' of Q4 might have had some theatrical experience by the variant at 2.1.10: "dove" (Q1), "day" (Q2 and Q3), "die" (Q4). Hunter discusses the Great Vowel Shift, but without using the International Phonetic Alphabet so it is hard to know just what sounds she is trying to indicate, and she argues that Q4's not following Q3 and Q2's "day" but instead putting "die" would not have happened "if the editor had not 'heard' something different" (p. 16). Of course, this could easily have happened by any one of the many ways that error can creep into a text, including compositorial slip, something that Hunter does not mention as a source of variants. The "most radical changes in Q4" (by which she means the most radical of Q4 departures from its copy Q3) occur in the quarter of the text for which there is no corresponding part in Q1 (just as was the pattern with Q3) but some of the changes "make only a little difference to significance, for example 5.3.8 something (3/4) some thing (4)". I am not sure that that makes any difference to the meaning in early modern English. Hunter lists changes (such as "murd'red (2/3) murdered (4)") that one might expect a compositor to make in justifying a line, but without letting the reader know whether the line in question is full, and she sees "an exceptionally attentive mind" making "subtle adjustments to punctuation that radically affect or effect significance". This insistence on a shaping mind reaches absurd heights with the attribution to "the Q4 editor" of the mistake of turning at Q2/3's correct "mouth of outrage" (5.3.215) into Q4's "moneth of outrage" and then claiming that it might be a variant spelling of "moans" because "mones" appears in Q4 (p. 17). Hunter overlooks the possibility of imperfect press-correction: perhaps during proofing it was spotted that "month" had been set, and since (as OED month1 confirms) the spelling "moneth" was almost universal in this period, it was wrongly changed to that. Hunter sets out to sketch the biography of her 'intelligent' editor of Q4, who "seems to have had knowledge of the play in performance" because speeches are correctly reassigned against the advice of earlier printings (p. 17). There is an obvious objection to this line of reasoning: Hunter's own knowledge that Q4 is right is not based on seeing the original performances, so if she can work it out from the surviving printings then her imagined Q4 editor could have. Hunter thinks that Q4's making clear that Balthazar returns with Romeo from Mantua is also a sign of the theatre, but again one could 'fix' that from reading the play. Where Q3 names "Will Kemp", Q4 has the character name "Peter" and Hunter thinks that "Q4 firmly deletes the actor's name presumably because it is no longer a selling point, Kemp must have faded from people's minds". Would a name buried within a book ever have been a selling point in the first place? A final error: Hunter notes that F lacks oaths that would violate the 1606 Act to Restrain Abuses of Players whereas Q1, Q2, Q3, and Q4 retain them and wonders if that is because "they are less formal texts that [sic] F?" It is well known that the act did not cover printing at all and the Folio has other plays containing oaths that could not be said on stage after 1606. Hunter begins to recognize the image "of an editor working just like a modern editor on the text itself, with the addition of a theatrical understanding" and concludes that the "good practice" of her kind of theatre-centred editing "goes back a long way" (p. 20).

    Carl James Grindley provides a facsimile and transcript of an under-examined version of the King Lear story (2001). It appears in John Hardyng's Chronicle, a work written between 1450 and 1470 and surviving in 11 complete manuscripts and many  fragments, and two printings in 1543 (not based on one of the extant manuscripts), and a bad edition of 1812. Because of the variation between manuscripts, there is no establishable stable text. Tapan Kumar Mukherjee aims to clear up a calendrical obscurity when friends mock Benedick by seeming to complete an aural 'letter' of his with "CLAUDIO To the tuition of God, from my house if I had it - / DON PEDRO The sixth of July, / Your loving friend, / Benedick" (Much Ado About Nothing 1.1.265-7) (2001). Phillip Clayton-Gore argued that this referred to the quarter-days--25 March (Lady Day), 24 June (Midsummer day), 29 September (Michaelmas), and 25 December (Christmas day)--or rather the variant used by the Crown Estate Commission in which Midsummer Quarter-day falls on 5 July. This would make Don Pedro's "sixth of July" the start of a new quarter and hence an occasion of merriment with no rent to worry about for three months. Mukherjee reckons we need the Julian system of counting days before the Kalends (the 1st day of the month), the Nones (7th day in March, May, July, and October, the 5th in other months), and the Ides (the 15th day in March, May, July, and October, the 13th in other months). By this system, he claims, 24 June (Midsummer day) is "the sixth day of the Kalends of July", and hence the reference is to Midsummer. Unfortunately he must have miscounted, for having already insisted that the days are counted inclusively, 24 June must be the 8th of the Kalends of July (24th, 25th, 26th, 27th, 28th, 29th, 30th, 1st).

    This year's Notes and Queries is much better printed that the last and it contains the typical crop of about two dozen pieces on matters textual. Geoff Wilkes considers the problem in Macbeth of "As whence the Sunne 'gins his reflection, / Shipwracking Stormes, and direfull Thunders: / So from that Spring, whence comfort seem'd to come, / Discomfort swells (TLN 44-7, 1.2.25-8) (2001). The sense has to be that the same place that gives hope brings discomfort, but how does the sun do reflexing? Perhaps by reaching the equinox or solstice and turning back again, but that is strained and the OED does not support the use of 'reflect' (turn back) until 1662. Wilkes thinks that the moon is where the sun "'gins his reflection", thereby giving comfort but also causing shipwrecking storms. David Lake and Brian Vickers decide that the 1622 quarto of Othello was set from a transcript made by two scribes (2001). E. A. J. Honigmann argued that Q was probably printed from a scribal transcript on the basis of some hypothesized contractions that caused misreadings ("ha" for "have", and "tho" for "though), and they are typical late Jacobean contractions, so non-authorial. Had Honigmann counted the contractions he would have found that Q has many more occurrences of "ha' and "'em" than we would expect from Shakespeare's habit around 1602 (the date accepted by Lake and Vickers), and none of his usual uses of "'a" meaning 'he'. So, Q is almost certainly set from scribal copy, not authorial papers and the likely explanation is that the scribe, working around 1621, consciously or unconsciously made the play "sound more contemporary, more Jacobean" (p. 286). In The Stability of Shakespeare's Texts (1965) Honigmann showed evidence that two scribes did the work that made the copy of Q-1622, on the evidence of spelling preferences that seem to follow a preference for either marginal or centered entrance stage directions, although Honigmann makes little of this in his more recent book Texts of 'Othello' (1996). Reviewing the latter for Shakespeare Studies, MacDonald P. Jackson pointed out that the spelling switches that Honigmann noted in 1965 (though/tho, bin/beene, and use of "ha" and "em") are so clustered that statistically they are unlikely to be produced by randomness. So, the two-scribe hypothesis for the copy underlying the 1622 quarto should again be taken seriously.

    Continuing his excavation of Nashe in Shakespeare, J. J. M. Tobin argues that there are borrowings from Christ's Tears Over Jerusalem in Measure for Measure (2001). There Shakespeare would have found "the hoode makes not the Moncke" that Lucio says in Latin at 5.1. 260, although Tobin does not observe that Shakespeare had already used cucullus non facit monachum in Twelfth Night two years earlier. Also, Angelo's "The tempter or the tempted, who sins most, ha?" (2.2.169) comes from Nashe's "both the person of the tempted and the tempter" and so does the play's collocation of "preserved" and "temptation" (2.2.157, 164). Tobin has even more tenuous Measure for Measure links to Christ's Tears Over Jerusalem via syntactical structure ("One thing . . . another thing") and more distant collocations. Escalus's "Some run from brakes of ice" (Riverside text) Tobin explains as "an abbreviated reference" to a usurer's trick that Nashe describes, in which having "broke the Ise" of borrowing once, the young gentleman-victim, will repeatedly borrow on increasingly unfavourable terms, which links with the Duke's reference to Nature's good usury at the beginning of the play ("nature never lends . . . thanks and use" 1.1.36-40). G. Blakemore Evans notices that in the sources of Hamlet (Saxo Grammaticus, Belleforest, and The Hystorie of Hamblet) the letter that condemns Hamlet is cut on a wooden board, but in Robert Parry's 1595 chivalric romance Moderatus there is a scene of a letter being opened and then resealed using a copy of the signet that originally sealed it (2001). Maybe Shakespeare knew Moderatus, but Blakemore Evans thinks it more likely that Parry ("who had visited London several times before 1595") recalled something he had seen in the Ur-Hamlet. Eric Sams finds the anonymous plays The Chronicle History of King Leir and Edmund Ironside so full of verbal correspondences that they must be by the same person (2001). Sams shows that there are over 140 verbal parallels between then, but many are too common to be significant--"be advised" might be shared by any two works, and so might "Enter . . . disguised" and "fountain[s] . . . spring"--and it is hard to know why he goes further and claims they were written "at about the same time" (p. 266). Sams claims that he omitted some additional examples because "they resist tabulation", such as the archaisms 'quondam' and 'whilom' (why those, especially?), and some links he does list are just silly, such as "'my gracious lord' used when addressing a king. Likewise Sams finds a parallel in the plays having "antithesis . . . references to the law . . . proverbs, puns and word-play . . . and usages antedating OED citations", which surely must be true of much literature of the period. There are technical problems with the note too, for Sams uses an 'author-date' style of citation but nowhere are the full bibliographical details provided and the Oxford Shakespeare's Textual Companion is misdated to 1988. Sams ends with a complaint that all his evidence has never "convinced anyone".

    Paul Vincent thinks that 2 Henry 6 was written by Shakespeare and person(s) unknown, since it is not all of a piece in spelling or in classical knowledge (2001). Gary Taylor showed that 1 Henry 6 probably was written by 4 or more hands, and this note tries to show that likewise 2 Henry 6 is probably the work of at least two dramatists. Where a play is set from "directly from authorial papers" (decided on the basis of variability in speech prefixes and imprecision or faulty stage directions) as Folio 2 Henry 6 seems to be, the spelling choice O/Oh can tell us a lot. (No mention here of the suspicion recently cast on the New Bibliographical idea that variability and imperfection in speech prefixes and stage directions indicate authorial copy, nor does Vincent mention the role of Q3, a Q1 reprint, in the printing of Folio 2 Henry 6.) Vincent tabulates O/Oh preference in Folio 2 Henry 6 and it is pretty clear: apart from 3.2 and 3.3, "O" predominates (26 O's to 5 Oh's). In 3.2 and 3.3 it is the other way round: 12 Oh's to 1 O's. 3.1 has no O's or Oh's, which is odd because it is tense stuff plot-wise, but it does have 7 Ah's, something Shakespeare rarely used. (Vincent does not mention it, but Shakespeare's preference was O, of course.) The abnormally high occurrence of "ye" (something Shakespeare avoided) over "you" in the play points away from it being all by Shakespeare and if we look at one type of compound adjective--the highly original and Shakespearian "conjunction of a noun, adjective, or preposition with a present participle"--then we find them clustering in 3.1 and 3.2, suggesting these were written by someone who did not write the rest of the play. (Placed together with the previous evidence this is problematic: there are unShakespearian Ah's in 3.1 and unShakespearian Oh's in 3.2, yet these are the places Vincent finds Shakespearian compound adjectives; Vincent does not juxtapose the evidence in a way that would make this problem readily apparent.) The level of learning that we can infer from classical allusions also varies noticeably. In 3.2 there are two errors of classical mythology known to educated men from the Aeneid: Aeolus keeping his winds in "brazen caves" (3.2.89), a faulty adjective, and Ascanius "witch[ing]" Dido and telling her about the fall of Troy (3.2.116-8) when it should be Cupid-as-Ascanius doing the witching and Aeneas, Ascanius's father, relating the fall of Troy. These could not have been made by Green, Nashe, Peele, or Marlowe, and again, act 3 marks itself out as different. (But since this act contains the unShakespearian preference for "Oh", it is not his either; we are running out of candidates.)

    Being led off to execution, Suffolk comments on some famous deaths: "A Roman sworder and banditto slave / Murdered sweet Tully; Brutus' bastard hand / Stabbed Julius Caesar; savage islanders / Pompey the Great; and Suffolk dies by pirates" (4.1.137-40) which is wrong on all counts, and the idea of Brutus being a bastard son of Caesar is especially interesting as Shakespeare makes nothing of it in Julius Caesar, although he presumably knew of it because it is in Plutarch and was notorious. The implication--and here Vincent is following and quotes J. A. K. Thomson (misspelled Thompson, p. 273n22)--is that Shakespeare knew the claim, but chose not to use it in Julius Caesar and therefore probably did not use it in 2 Henry either, so this part is someone else's writing. There are plenty of correctly-made classical allusions in 2 Henry 6, so presumably the faulty ones are by someone who did not write the good ones. An example of a good one is a correct distinction of the famous Ajax from the lesser Locrian Ajax (son of Telamon, called Telamonius Ajax): "like Ajax Telamonius, / On sheep or oxen could I spend my fury" (2 Henry 6 5.1.26-7). Vincent thinks that Shakespeare elsewhere distorts the story Telamonius Ajax in "he's more mad / Than Telamon for his shield" (Antony and Cleopatra 4.14.1-2) since it was Telamonius Ajax (son of Telamon) and not Telamon himself that ran mad for not winning the shield of dead Achilles. (Actually, the latest Oxford Shakespeare and Arden Shakespeare editors of Antony and Cleopatra agree to let "Telamon" mean "Telamonius Ajax" and do not treat this as an error. Thus Vincent concludes that the classical allusions in 2 Henry 6 do not all come from the same hand.

    Anthony Mortimer wonders about the following lines in Venus and Adonis: "'Long may they kiss each other, for this cure! / O, never let their crimson liveries wear, / And as they last, their verdure still endure / To drive infection from the dangerous year" (505-8) (2001). If the lips are "crimson", how do they also have "verdure", greenness? Answer: 'crimson' (or at least the related French word cramoisi) like 'scarlet', might mean bright colours other than red (although red was the dominant meaning), so these marginal senses come into play in Shakespeare's shift from red lips to green herbs used to ward off the plague. John M. Rollett has more evidence to add to his theory that parts of the 1609 Sonnets quarto was set by someone reading the manuscript to the compositor (2001). Two years ago Rollett argued that sonnet 146 shows what can happen by aural transmission (1999) and now he cites Hardy M. Cook's work on the sonnets for Ian Lancashire's Renaissance English Texts (RET) project. (Actually, Rollett does not mention Lancashire or RET, and he wrongly gives the Internet URL as "http://www.library.utoronto.ca/utel/ret/shakespeare/1609pre.html" when it should be "http://www.library.utoronto.ca/utel/ret/shakespeare/1609int3.html"; the only way to get from the first, which has no links, to the second is by browsing the directory tree, not something for the uninitiated.) Cook thinks that at times only one of the two compositors was setting type and Rollett thinks the other might have been free to do the copy reading. Rollett claims that "Cook states that the only other serious error in Q occurs in line 6" of sonnet 144", whereas in fact Cook wrote that "the only emendation of a substantive universally followed by modern editors is Malone's "{{s}i}de" for "{{s}i}ght" in line six of Sonnet 144", which is not the same thing. (Cook's curly braces represent the long 's' and ligatures, but most readers will not in any case find the reference since Rollett gives it as "Hardy M. Cook, ibid., 1, 1" whereas in fact one has to load a different file into one's web-browser (it is at http://www.library.utoronto.ca/utel/ret/shakespeare/1609int1.html) and look for its first paragraph.) Rollett points out that "side" to "sight" is more likely to be a hearing error than a reading one and that the errors in sonnets 144 and 146 occur on signature I3r and were done by compositor Eld A, who together with the person reading (perhaps Eld B) just had a bad day.

    A. D. Nuttall observes that there is an old chestnut of inverse nomenclature in lucus a non lucendo (it is called a wood, lucus, because there is no light, lux, there) written about by many ancient grammarians (2001). This is what Bottom's "It shall be called 'Bottom's Dream', because it hath no bottom" (A Midsummer Night's Dream 4.1.212-3) alludes to, appropriately because the play's wood is dark. Deanne Williams thinks that from John Lydgate's The Fall of Princes, where Herod's cities of similar names are mentioned, Shakespeare got the names Sebastian and Cesario in Twelfth Night, and also got the play's relationship between gloom and celebration (2001). The twin cities of Sebaste and Cesarea were built by Herod whose Massacre of the Innocents is commemorated 8 days before Twelfth Night, the celebration of the arrival of the Magi. Orsino's court is like Herod's (as represented in "medieval Christmas plays", none of which Williams cites) in its "exotic-erotic luxuria" and Malvolio's outburst is like Herod's rage, as is failure. The city of Cesarea was situated most unfavourably for ships, hence the shipwreck. Shakespeare probably got the names from John Lydgate's Fall of Princes (1431-8), where "Cesaria" and "Sebasten" are described as "tweyne" (which Williams thinks "hints at their status as twins" although it really just means two) and are named there alongside "Antipadra", which Williams thinks gave Shakespeare the idea of 'after' (anti) 'Father' (padra), their father is dead. From The Fall of Princes, which is about Petrarch cheering up gloomy and lethargic Bochas, Shakespeare got Viola's cheering up melancholic Illyria, manifested in Orsino's moping, Olivia's mourning, and Feste's "the rain it raineth every day". The dichotomy of the Feast of the Holy Innocents (a celebration despite the slaughter) was about getting on with things despite the disaster, and this is the dichotomy of Twelfth Night: Viola shows how to carry on despite losing a brother and a father. The Feast of the Holy Innocents was a Christian version of Roman Saturnalia, with inversions like that of the play, which of course ends with a gloomy song (the point is to accept life and death). In the first of two pieces, Thomas Merriam argues that all of Sir Thomas More is by Shakespeare and not Antony Munday, who used feminine endings and deviant lines much less often (2001b). Contrary to Philip W. Timberlake's assertion that Munday's use of feminine endings rose during the 1590s, it fell. Timberlake found that Munday's average use of feminine endings was in 13.7% of his lines (the minimum in some scenes being 7.8 and the maximum 24), and Sir Thomas More comes out at 20.7%. There is not agreement how to define feminine endings, and Merriam chooses "the notion of deviant lines which accommodates 8, 9, 11, 12, and 13 syllables per line, irrespective of final stress", which is much more broad than the usual identification by an additional unnecessary and unaccented syllable. Even with this wider net Merriam cannot get Munday's percentage of 'deviant lines' up to Sir Thomas More's level. Here the article becomes a compressed and incomprehensible argument, not least because Merriam omits "Koenig"'s first name and draws upon data in Halliday's Shakespeare Companion and Chambers's WS: Study of Facts and Problems without explaining how those data were compiled or for what purpose. The data should really be the texts we now have, in named editions, something Merriam omits even for Sir Thomas More where choice of edition makes a huge difference. Merriam names the percentages that other people have given for feminine endings in Shakespeare plays, and then reports that he did some "extrapolating from the figures given by Chambers" without say what he means by extrapolating or how it was done, other than a formula with two detailed constants that he does not explain the derivation of. I presume he compared all the Timberlake figures with all the Chambers figures, and derived a formula for converting one into the other, and then fed Chambers's count for Merry Wives of Windsor into it to get a figure that Timberlake would have got had he analyzed that play.

    Merriam offers a table comparing the proportions of feminine endings in The Two Gentlemen of Verona, Munday's John a Kent and John a Cumber, Sir Thomas More, The Merry Wives of Windsor, and an average for three Munday texts: Downfall of Robert Earl of Huntingdon, Death of Robert Earl of Huntingdon, and the Munday parts of 1 Sir John Oldcastle. Merriam observes from comparing John a Kent (early 1590s) at 13.8% with the composite Munday figure for the late 1590s, 10.8%, that his proportion of feminine endings/deviant lines did not rise in the 1590s, but we should note that the latter is the average of three widely different tallies of 8.1%, 10.7%, and 13.2%, and that incidentally these average to 10.7% not 10.8% as given in Merriam's table. Merriam also notices that Sir Thomas More and The Merry Wives of Windsor are alike in their high proportion of "feminine endings" (20.7% and 21.8%). Merriam plots how Shakespeare's plays increasingly used feminine endings through his career, using Chambers's data adjusted to make it nearer to what Timberlake would have counted had he been looking at them, compared to other plays looked at by Timberlake (who stopped at 1595), and concludes that Sir Thomas More has far too much metrical deviancy (20.7% of lines) to be by anyone but Shakespeare. However, Merriam's data excludes everyone after 1595 except Shakespeare and Munday, and even then Munday is not terribly far off at 13.8% in John a Kent. Perhaps everyone started doing more deviancy towards the end of the 1590s, when Timberlake's data stops. All this is unpersuasive since Merriam says nothing about the revisions in Sir Thomas More (which could well date from the 1600s) nor about which edition he used, which matters because different editors incorporate the revisions in different ways.

    In his second article, Merriam counts uses of the word 'hath' to distinguish Shakespeare's work from Fletcher's (2001a). Merriam begins by quoting Stanley Wells on the problem that authorship attribution studies use tricky mathematics, which Merriam thinks untrue. I disagree: Merriam's own explanation of "principal component analysis" of data about "logometric habits" last year was a model of obscurity in advanced mathematics, as when he wrote that ". . . the eigenvalues and eigenvectors of the characteristic equation of the correlation matrix are derived by an algorithm"  (2000, 439). To show how easy it all is, Merriam here patiently explains that "relative word frequency" means the number of times a word occurs divided by the total number of words in the text, as though that were the sort of thing that stumps the non-specialist. The substance of this piece is that the frequency of 'hath' in the Shakespeare Folio is stable across the plays, and likewise in 13 of 14 Fletcher plays, and these frequencies are different so it is a good discriminator. Merriam gives a table of 'hath'-counts for Shakespeare's "First Folio plays" and many Fletcher ones but does not indicate the precise provenance of the electronic texts other than "the Oxford Text Archive and/or Professor Ward Elliott" and the Chadwyck-Healey English Verse Drama database. The obvious questions that Merriam leaves unanswered are whether they are in original spelling and what has been done to represent the long 's' and ligatures, these things being widely known only for the Chadwyck-Healey texts, which have original spelling, do not attempt to represent ligatures, and modernize the long 's'. Most surprisingly, for the plays from Chadwyck-Healey's database "no total word count was established" so a single figure of 22264 was used, it being the average word-count for the Fletcher plays taken from the Oxford Text Archive. It is a trivial matter to do total word counts from the Chadwyck-Healey texts, so I have made my own. I agree with Merriam's counts for occurrences of 'hath', but we can discard his assumption of a single figure for total word counts for the plays and supply the actual figures in each case. Here are the counts for each play, followed by the relative frequency of 'hath' in round brackets, correcting Merriam's where necessary: A Wife for a Month 23,335 (0), Wild Goose Chase 23,150 (0); Women Pleased 22,005 (0.000136 not 0.000135); Wit Without Money 23,204 (0.000215 not 0.000225); Love's Pilgrimage 24,858 (0.000443 not 0.00049); Elder Brother 22521 (0.000533 not 0.00043); Beggar's Bush 20,586 (0.000729 not 0.00067); Fair Maid of the Inn 22,580 (0.000797 not 0.00081); Queen of Corinth 21,825 (0.001054 not 0.00103); Noble Gentleman 20,355 (0.001916 not 0.00175); Faithful Shepherdess 21,616 (0.002174 not 0.002111). The only difference these revised figures make is to push Noble Gentleman down Merriam's list into a slot in between Shakespeare's Julius Caesar and Coriolanus; Merriam would doubtless explain this as being due to its being a collaboration. But Merriam says nothing about the textual provenance, which matters because if all his group A plays were printed (or indeed transcribed) by one person and all of group B by another, it would be entirely possible that the differences between them on a matter like 'hath' reflect the printers' or scribes' habits, not the authors'. Merriam turns to the outlying anomaly, Faithful Shepherdess (a Fletcher-only play that appears right in the middle of the Shakespeare's in his table), and offers another table that merely shows it outlyingness in a different way; it is much more like Shakespeare than Fletcher according to his test. Having taxed the intelligent reader's patience with a description of how to calculate relative frequency, which is trivial, Merriam suddenly assumes the background knowledge needed to make sense of "the middle two quartiles or central half of each distribution of relative frequency values" and indeed this second table is bewildering for a number of reasons. Not least of these is its vertical scale being labelled -0.0005 (at the origin) and then 0.0004, 0.0013, 0.0022, 0.0031, 0.0040, or intervals of 0.0009. Starting with a negative origin point is absurd since there can never be a negative number of haths in a play (nought is the minimum) and the true origin is the bottom of the Fletcher group since two of his plays never used the word. Merriam ends with a dig at "Traditionalist literary scholars" who posit a wide gulf between themselves and the number-crunchers; on this evidence they are quite right.

    Roger Stritmatter argues that Henry's 5 speech about the king not being responsible for the souls of his men comes from the Geneva Bible Ezekiel 18:20 (2001). Stritmatter begins by declaring his hand with an announcement that the notes he has published in Notes and Queries over the last 5 years about Shakespeare's knowledge of the Bible came from his PhD thesis on a 1568-70 Geneva bible owned by Edward de Vere, 17 Earl of Oxford, in which these passages are underlined. Ezekiel 18:20 is the hitherto unknown Biblical source for Henry 5's theological lecture at 4.1.146-184, which Stritmatter gives as 4.1.130-305, much too long a stretch and he does not state which edition he is using. Stritmatter claims that because Ezekiel 18 is about "the heritability of sin", its lines "the same soule that sinneth, shal dye: the sonne shal not beare the iniquity of the father" are alluded to in Macduff's "Sinful Macduff, / They were all struck for thee. Naught that I am, / Not for their own demerits but for mine / Fell slaughter on their souls" (Macbeth 4.3.226-9) and again in "Let sin alone committed light alone / Upon his head that hath transgressèd so; / Let guiltless souls be freed from guilty woe. / For one's offence why should so many fall, / To plague a private sin in general?" (The Rape of Lucrece 1480-4). The latter is slightly more plausible than the former, being at least about punishing only the sinner. There are two further references to parental; sin alighting on the child: "LAUNCELOT Yes, truly; for look you, the sins of the father / are to be laid upon the children" and "JESSICA So the sins of my mother should be visited upon me" (The Merchant of Venice 3.5.1-2, 11-2). Stritmatter says that these express "the opposite moral, that the sins of the parents should be visited on the children" (his emphasis), but surely Jessica means 'should not'. The two-way detachment of "the sonne shal not beare the iniquity of the father, nether shal the father beare the iniquitie of the sonne" (Ezekiel 18:20) is used by Henry to absolve himself for the state of the souls of his men that die in battle, and he uses the word "iniquities" too. Stritmatter neglects to mention that, if the parallel is right, Henry's argument is not merely derived from scripture but is a perversion of it. The initial question is whether doing the sinful bidding of a king ("if the cause be not good", 4.1.133) attracts eternal damnation, whereas Ezekiel is about whether being merely related to a sinner (although free from sin oneself) is enough. Henry elides this distinction and brings in the obfuscation that many soldiers have serious previous sins to their names. I would say that the difference between Henry and Ezekiel is so great that it is more likely to be theological commonplace than allusion.

    Howard Jacobson observes that "[CASSIUS] The fault, dear Brutus, is not in our stars, / But in ourselves, that we are underlings" (Julius Caesar 1.2.141-2) is close to Seneca's non locorum vitium esse quo laboramus sed nostrum (de Tranquillitate 2.15), and even if Shakespeare did not get it from Seneca directly, the latter appeared in a collection of Senecan sententiae that came out in 1597 (2001). MacDonald P. Jackson finds a way that Julius Caesar can be emended to take account of Jonson's mockery about Caesar wronging with "just cause" without having Caesar make the absurdly megalomaniacal claim that his 'wrong' is 'right' (2001a). The Folio reads "[CAESAR] Thy Brother by decree is banished: / If thou doest bend, and pray, and fawne for him, / I spurne thee like a Curre out of my way: / Know, Caesar doth not wrong, nor without cause, / Will not be satisfied" (TLN 1251-5) and in the posthumously-published Discoveries Jonson mocked Shakespeare's "Caesar did never wrong but with just cause" as a response to "Caesar, thou dost me wrong"; the latter line is not in F. John Dover Wilson argued for emending F to "Caesar did never wrong but with just cause" as typically dictatorial and (on the evidence of Jonson's comment) as probably what Shakespeare wrote but was changed by the actors after Jonson's mockery. But this still leaves the problem that Caesar's comment is not prompted by anything and does not really mean anything: what is the "cause", what might make him "satisfied" and of what? Jackson suggests that Metellus Cimber throws himself at Caesar's feet and gets kicked, and complains "Caesar, thou dost me wrong", so that the exchange goes like this: "[CAESAR] Thy brother by decree is banished. / If thou dost bend, and pray, and fawn for him, / I spurn thee like a Cur out of my way. / METELLUS CIMBER Caesar, thou dost me wrong. / CAESAR Know, Caesar doth not wrong, / But with just cause will he be satisfied". In this suggestion "But with" means 'Only with', as it is used in Hamlet ("But with the whiff and wind of his fell sword / Th' unnerved father falls" 2.2.476-7). This emendation has the advantage of making Caesar speak sense and not megalomaniacally (as "Caesar did never wrong but with just cause" would) since nothing else in the play makes him megalomaniacal. An actor speaking Jackson's version and wrongly pausing after "just cause" would make the gaffe that Jonson seized upon, and in altering the script to deflect Jonson's mockery they changed "But with just" into "Nor without" and accidentally omitted Metellus's interjection. (Actually, they must have also changed "will he be satisfied" into "will not be satisfied", so turning "But with just" into "Nor without" is not "the sole change", as Jackson claims.)

    Steven Doloff points out that Elizabethans would have a sense of the Italian, especially Venetian, life as corruptly bestial from (amongst others) Roger Ascham's The Scholemaster (1570), which refers to "Circe's Court" (2001b). There is a distinct Circean motif (from men to beasts) in Othello, with its endless animal imagery, so there is Italian beastly otherness as well as Moorish racial otherness in the play, and indeed the play makes the former more dangerous (via Iago) than the latter. In a second note, Doloff argues that Lear's "Howl, howl, howl, howl!" while carrying dead Cordelia (5.3.232) is not only (as W. R. Elton has it) a rejection of Stoic philosophy but also an allusion to Diogenes the dog-like Cynic (2001a). Edgar-as-Tom calls himself a "dog in madness " (3.4.87-8), the usual Cynic-Dog association, and Lear calls him a "philosopher" (3.4.144) and identifies with him to the extent of wanting to disrobe too ("unbutton here" 3.4.103), but only at the end of the play, stripped of everything (not just clothes) does Lear truly become like Diogenes and howl like a dog. Roger Prior thinks that for Othello and Love's Labour's Lost Shakespeare borrowed from Ariosto's Italian poem Orlando Furioso and not Sir John Harington's English translation of it (2001). We can surmise that Shakespeare read Italian, for he seems to have read Othello's source, Giraldi Cinthio's Hecatommithi, in the original, and he also seems to have read in Italian Ariosto's Orlando Furioso, since Cassandra's "furor profetico" is echoed in Othello's "A sibyl . . . In her prophetic fury sewed the work" (3.4.70-2), which was already known but can now be confirmed by further borrowings from Ariosto that Priors describes. The handkerchief in Cinthio does not have the "supernatural qualities" that Shakespeare wanted, so he instead took these from the tent that in Ariosto Cassandra made for Hector (a soldier) and which passed to Menelaus (a soldier-cuckold) and which much later was owned by another sorceress, Melissa, and given away by her as a wedding present. Likewise the handkerchief in Othello is made by a "sybil" and later is given by an "Egyptian . . . charmer" (3.4.56-7) to Othello's mother and given by Othello as a wedding present. Shakespeare could have got these narrative details from Sir John Harrington's 1591 English translation of Orlando, but there are many verbal parallels that suggest that he used the Italian text. For example, "fece un bel don di quello" (she made a fine gift of it, canto 46 stanza 80) became "made a gift of it" (3.4.61) and in stanza 82 there is "poi che . . . ebbe la morte" (on Hector's death) the cloth passed "in sorte" (by fate) to Menelaus who had an unfaithful wife Helen ("la moglie"), which becomes Othello's mother passing on the handkerchief on her deathbed (3.4.63) with the instruction that "when my fate would have me wived, / To give it her" (3.4.64-5), which he did and later thought the wife, Desdemona, unfaithful. Prior gives some lesser examples of Shakespeare's verbal borrowing from Ariosto, and admits that they might just have come from Harington instead. Shakespeare had already used Ariosto in Armado's speech as Hector in Love's Labour's Lost ("Mars . . . gift . . . of Ilion . . . pavilion . . . flower" 5.2.644-8), which comes from canto 46 stanza 80 of Orlando Furioso: "padiglion" (pavilion), "un bel don" (a splendid gift), "d'Ilia" (of Ilion), and stanza 85: "Marte . . . fiori" (Mars . . . flower). Ariosto wrote his poem for Hippolytus of Este who claimed descent from Hector, so Ariosto has his Cassandra prophesy the appearance of his descendant and work his image into the cloth of the tent that she makes, so "Hippolytus, Hector's heir, thus becomes part of the gift" and likewise Armado's line means the same: "Gave Hector a gift, the heir of Ilion". Again, Prior shows some lesser links that might have come via Harington, but also some verse-structure parallels (Armado's 11 or 12 syllable lines, like Ariosto, their similar trochaic endings, and their rhyming abab like the beginning of one of Ariosto stanzas) which cannot have. Coincidentally, Robert Tofte who we know saw Love's Labour's Lost was a translator of Ariosto.

    Catherine Loomis finds a contemporary historical analogue for the transmission of news in Macbeth (2001). Sir Robert Carey carried the news of Elizabeth 1's death to King James, and a brief account of this journey was published in 1603, his full memoirs being published in 1759. The news of Elizabeth's death, like that of Lady Macbeth, is conveyed first by crying women. Carey got a bloodied head from a kick from his horse on the long journey to Scotland, and the detail appeared in the 1603 pamphlet. Thus an audience hearing Duncan's opening line "What bloody man is that?" (1.2.1) would think of Carey's story of a king of Scotland receiving news of a death from a bloodied man, and Carey's story also had belligerent porters and a knocking at a gate (James's) presaging death. (George Steevens's names is misspelled "Steven's" here, and not in a possessive context, p. 297n4). MacDonald P. Jackson dates All's Well that Ends Well to before mid-1606 by Parolles's use of the name "Spurio" (2001b). Only two other plays use this name, one is Middleton's The Revenger's Tragedy and the other is Thomas Nabbes's The Unfortunate Mother. Nabbes is much too late, but Middleton's play was printed 1607-8 and all its names are appropriate to their characters; Middleton's Spurio is important, while in All's Well that Ends Well it is just a casual reference. Jackson thinks that for this reason "Shakespeare must surely be the debtor". (I do not think this follows at all; Middleton could have heard the name as used in All's Well that Ends Well and added it to his list of suitable names to be used in the composition of The Revenger's Tragedy.) Around this time Middleton and Shakespeare worked together on Timon of Athens. Jackson is editing The Revenger's Tragedy for the Oxford Complete Middleton and thinks the date of composition is early spring 1606 since it is indebted to Volpone (which was written in the first few months of 1606) and King Lear (not completed before the autumn of 1605). (Again, these give termini a quo but not termini ad quem; The Revenger's Tragedy could have been written any time after these and before its Stationers' Register entry on 7 October 1607.) So, if (Jackson is careful to place this 'if') The Revenger's Tragedy was first performed in early 1606, All's Well that Ends Well cannot have been composed before mid-1606. David George thinks that Caxton's The Recuyell of the Historyes of Troye (reissued in a new version in 1607) is the source of Volumnia's image of Hector's bleeding forehead in Coriolanus (2001). Volumnia pictures her son's bloody brow like "Hector's forehead when it spit forth blood / At Grecian sword, contemning" (1.3.44-5), which is not from Homer's Iliad since there Hector's fight with Grecian Achilles is not with swords when he contemns him, and his face does not bleed. In any case, there was no complete English Iliad for Shakespeare to read until 1611 and although George Chapman's partial version was out the only part that seems even close to Volumnia's allusion is in book 12 that Chapman did not publish until probably 1609, which is probably too late to influence Coriolanus. The likeliest source is Caxton's The Recuyell of the Historyes of Troye that Shakespeare used for Troilus and Cressida and which was published in a new version amended by William Fiston in 1607; this has a fight between Achilles and Hector (who uses a sword), with Hector contemning and with his facing bleeding. On the other hand, "contemning" might not be the correct emendation of the Folio's "Contenning", and we can wonder what is doing the contemning and at what is it directed; F2's "contending" is also possible. George discusses the relative merits of these and does not recommend one over the other. David Roberts notes that in King Leir Ragan reads a letter and "bytes her lip, / And stamps" which is like how Wolsey receives bad news in a letter in Henry 8 ("bites his lip, and starts" 3.2.114), although he admits that this is a pretty tenuous link and it is hardly unusual to bit one's lip (2001)

    Rodney Stenning Edgecombe notes that Macbeth has a difficult image in "pity, like a naked new-born babe, / Striding the blast" (1.7.21-2), which is virtually oxymoronic is its soft limbs striding (2001b). There is the same paradoxical mixture of weakness and strength as the baby Jesus has in Robert Southwell's Epistle of Comfort, so that is a source of the image. In As You Like It 1.2 Le Beau picks out Celia by saying "the taller is his daughter" (TLN 440) but later it is clear that she is shorter than Rosalind; in preference to such emendations as "shorter" and "lower" (which are not graphically close), Edgecombe proposes "tawnier" since later Celia-as-Aliena is called "browner" (2000). (The obvious objection, and the reason no-one has suggested it before, is the unlikelihood of "wni" being mistaken for "ll"; in combination "w" and "ni" might pass for each other, but all three letters lack the top-loop that characterized "l" in Secretary and Italic handwriting.) Later Celia says that Orlando has "bought a paire of cast lips of Diana" (TLN 1725, 3.4.14) and the problem is "cast" (2001a). Edgecombe thinks that it is a spelling of "cased", that the lips are enclosed like a relic; the context supports this being a reference to Catholic relic worship. (He might, but does not, point out that the spelling "cast" for "cased" occurs in the first quarto of Pericles (sig. H4r).

Bawcutt, N. W. 2001. "Renaissance Dramatists and the Texts of Their Plays." Research Opportunities in Renaissance Drama. 40. 1-24.

Berger, Thomas L. and George Walton Williams. 2001. "Notes on Shakespeare's Henry V." Analytical and Enumerative Bibliography. 12. 264-87.

Bourus, Terri. 2001. "Shakespeare and the London Publishing Environment: The Publisher and Printers of Q1 and Q2 Hamlet." Analytical and Enumerative Bibliography. 12. 206-28.

Bradley, David. 1992. From Text to Performance in the Elizabethan Theatre. Cambridge. Cambridge University Press.

Brink, Jean R. 2001. "William Ponsonby's Rival Publisher." Analytical and Enumerative Bibliography. 12. 185-205.

Cathcart, Charles. 2001. "Hamlet: Date and Early Afterlife." Review of English Studies. 52. 341-59.

Doloff, Steven. 2001a. "Lear's Howl and 'Diogenes the Doggue'." Notes and Queries. 246. 292-93.

Doloff, Steven. 2001b. "Shakespeare's Othello and Circe's Italian Court in Ascham's The Scholemaster (1570)." Notes and Queries. 246. 287-89.

Duncan-Jones, Katherine. 2001. "Ravished and Revised: The 1616 Lucrece." Review of English Studies. 52. 516-23.

Eccles, Christine. 1990. The Rose Theatre. London. Nick Hern.

Edelman, Charles. 2002. "Hamlet, Soldier Manque." Around the Globe. 21. 44-45.

Edgecombe, Rodney Stenning. 2000. "'The Taller is His Daughter' in As You Like It." The Upstart Crow. 20. 153.

Edgecombe, Rodney Stenning. 2001a. "'Cast Lips of Diana in As You Like It III.iv." The Shakespeare Newsletter. 51. 63.

Edgecombe, Rodney Stenning. 2001b. "Southwell's 'Burning Babe' and the 'Naked New-born Base' in Macbeth." Notes and Queries. 246. 295-96.

Erne, Lukas. 2002. "Shakespeare and the Publication of His Plays." Shakespeare Quarterly. 53. 1-20.

Evans, G. Blakemore. 2001. "An Echo of the Ur-Hamlet?" Notes and Queries. 246. 266.

George, David. 2001. "Hector's Bleeding Forehead: Coriolanus, I.iii.34-9." Notes and Queries. 246. 299-302.

Go, Kenji. 2001. "Unemending the Emendation of 'Still' in Shakespeare's Sonnet 106." Studies in Philology. 98. 114-42.

Gombrich, Ernst. 2000. "'Wooden O': Letter to the Editor." Times Literary Supplement. Issue 5058 (10 March). n. pag..

Greg, W. W. 1925-6. "The Riddle of Jonson's Chronology." The Library (=Transactions of the Bibliographical Society). 6. 340-47.

Greg, W. W. 1955. The Shakespeare First Folio: Its Bibliographical and Textual History. Oxford. Clarendon Press.

Grindley, Carl James. 2001. "The Story of King Lear in John Hardyng's Chronicle." Cahiers Élisabéthains. 59. 77-80.

Hanabusa, Chiaki. 1999. "A Neglected Misdate and Romeo and Juliet Q1 (1597)." Notes and Queries. 244. 229-30.

Hill, W. Speed. 2000. "Where Would Anglo-American Textual Criticism be If Shakespeare Had Died of the Plague in 1593?" TEXT: An Interdisciplinary Annual of Textual Studies. 13. 1-7.

Hunter, Lynette. 2001a. "The Dating of Q4 Romeo and Juliet Revisited." The Library. 2. 281-85.

Hunter, Lynette. 2001b. "Why Has Q4 Romeo and Juliet Such an Intelligent Editor?" Re-constructing the Book: Literary Texts in Transmission. Edited by Maureen Bell, Shirley Chew, Simon Eliot, Lynette Hunter and James L. W. West. Aldershot. Ashgate. 9-21.

Ing Freeman, Arthur and Janet Ing Freeman. 2001. "Did Halliwell Steal and Mutilate the First Quarto of Hamlet?" The Library. 2. 349-63.

Jackson, MacDonald P. 1999. "Rhymes and Shakespeare's Sonnets: Evidence of Date of Composition." Notes and Queries. 244. 213-19.

Jackson, MacDonald P. 2001a. "'But with Just Cause': Julius Caesar, III.i.47." Notes and Queries. 246. 282-84.

Jackson, MacDonald P. 2001b. "Spurio and the Date of All's Well That Ends Well." Notes and Queries. 246. 298-99.

Jackson, MacDonald P. 2001c. "Vocabulary and Chronology: The Case of Shakespeare's Sonnets." Review of English Studies. 52. 59-75.

Jacobson, Howard. 2001. "Julius Caesar, I.ii.39-40." Notes and Queries. 246. 282.

Jowett, John. 2001. "The Audacity of Measure for Measure in 1621." Ben Jonson Journal. 8. 229-47.

Kastan, David Scott. 2001. Shakespeare and the Book. Cambridge. Cambridge University Press.

Kennedy, Richard F. 1998. "Speech Prefixes in Some Shakespearean Quartos." Papers of the Bibliographical Society of America. 92. 177-209.

King, T. J. 1992. Casting Shakespeare's Plays: London Actors and Their Roles, 1590-1642. Cambridge. Cambridge University Press.

Klause, John. 2001. "New Sources for Shakespeare's King John: The Writings of Robert Southwell." Studies in Philology. 98. 401-27.

Lake, David and Brian Vickers. 2001. "Scribal Copy for Q1 of Othello: A Reconsideration." Notes and Queries. 246. 284-87.

Long, William B. 1985. "'A Bed / for Woodstock': A Warning for the Unwary." Medieval and Renaissance Drama in England. 2. 91-118.

Loomis, Catherine. 2001. "'What Bloody Man is That':? Sir Robert Carey and Shakespeare's Bloody Sergeant." Notes and Queries. 246. 296-98.

Machan, Tim William. 2000. "'I Endowed Thy Purposes': Shakespeare, Editing, and Middle English Literature." TEXT: An Interdisciplinary Annual of Textual Studies. 13. 9-25.

Martin, Randall. 2000. "Rehabilitating John Somerville in 3 Henry VI." Shakespeare Quarterly. 51. 332-40.

Masten, Jeffrey. 2001. "More or Less: Editing the Collaborative." Shakespeare Studies. 29. 109-31.

McKerrow, R. B. 1931. "The Elizabethan Printer and Dramatic Manuscripts." The Library. 12. 253-73.

McKerrow, R. B. 1935. "A Suggestion Regarding Shakespeare's Manuscripts." Review of English Studies. 11. 459-65.

McMillin, Scott. 1988. "Building Stories: Greg, Fleay, and the Plot of 2 Seven Deadly Sins." Medieval and Renaissance Drama in England. 3. 53-62.

Menzer, Paul. 2001. "'Tis Heere, Tis Here, Tis Gone': Q1 Hamlet and Degenerative Texts." Analytical and Enumerative Bibliography. 12. 30-49.

Merriam, Thomas. 2000. "An Unwarranted Assumption." Notes and Queries. 245. 438-41.

Merriam, Thomas. 2001a. "A Simple Discriminator of Shakespeare and Fletcher." Notes and Queries. 246. 306-09.

Merriam, Thomas. 2001b. "Feminine Endings and More." Notes and Queries. 246. 278-80.

Mortimer, Anthony. 2001. "'Crimson Liveries' and 'Their Verdour': Venus and Adonis, 505-8." Notes and Queries. 246. 274-75.

Mowat, Barbara A. 2001. "The Reproduction of Shakespeare's Texts." Shakespeare. Edited by Margreta De Grazia and Stanley Wells. Cambridge Companions to Literature. Cambridge. Cambridge University Press. 13-29.

Mukherjee, Tapan Kumar. 2001. "The Sixth of July: Much Ado About Nothing: I.i.274." English Language Notes. 38.4. 16-18.

Nuttall, A. D. 2001. "Bottom's Dream." Notes and Queries. 246. 276.

Prior, Roger. 2001. "Shakespeare's Debt to Ariosto." Notes and Queries. 246. 289-92.

Rasmussen, Eric. 2001. "The Date of Q4 Hamlet." Papers of the Bibliographical Society of America. 95. 21-29.

Roberts, David. 2001. "Henry VIII and The True Chronicle History of King Leir." Notes and Queries. 246. 302-03.

Rollett, John M. 1999. "'Repel These Rebel Powers': Shakespeare's Sonnet 146 Emended." Notes and Queries. 244. 228.

Rollett, John M. 2001. "The Compositor's Reader: Shakespeare's Sonnet 146 Revisited." Notes and Queries. 246. 275-76.

Sams, Eric. 2001. "King Leir and Edmond Ironside." Notes and Queries. 246. 266-70.

Searle, William. 2001. "'By Foule Authority': Miscorrection in the Folio Text of Shakespeare's Troilus and Cressida." Papers of the Bibliographical Society of America. 95. 503-19.

Shakespeare, William. 2001a. Henry VI, Part Three. Ed. Randall Martin. The Oxford Shakespeare. Oxford. Oxford University Press.

Shakespeare, William. 2001b. King Henry VI Part Three. Ed. John D. Cox and Eric Rasmussen. The Arden Shakespeare. London. Thomson Learning.

Sisson, C. J. 1956. New Readings in Shakespeare. Vol. 2: The Histories, The Tragedies. Cambridge. Cambridge University Press.

Stritmatter, Roger. 2001. "The Biblical Source of Harry of Cornwall's Theological Doctrine." Notes and Queries. 246. 280-82.

Tobin, J. J. M. 2001. "Nashe and a Crux in Measure for Measure." Notes and Queries. 246. 262-64.

Tonkin, Humphrey. 2000. "'Wooden O': Letter to the Editor." Times Literary Supplement. Number 5063 (14 April). 21.

Vincent, Paul. 2001. "Unsolved Mysteries in Henry the Sixth, Part Two." Notes and Queries. 246. 270-74.

Warren, Roger. 2000. "The Quarto and Folio Texts of 2 Henry VI: A Reconsideration." Review of English Studies. 51. 193-207.

Wells, Stanley, Gary Taylor, John Jowett and William Montgomery. 1987. William Shakespeare: A Textual Companion. Oxford. Oxford University Press.

Werstine, Paul. 1990. "Narratives About Printed Shakespeare Texts: 'Foul Papers' and 'Bad' Quartos." Shakespeare Quarterly. 41. 65-86.

Werstine, Paul. 2000. "Editing Shakespeare and Editing Without Shakespeare: Wilson, McKerrow, Greg, Bowers, Tanselle, and Copy-text Editing." TEXT: An Interdisciplinary Annual of Textual Studies. 13. 27-53.

Werstine, Paul. 2001. "Scribe or Compositor: Ralph Crane, Compositors D and F, and the First Four Plays in the Shakespeare First Folio." Papers of the Bibliographical Society of America. 95. 315-39.

Wilkes, Geoff. 2001. "A Textual Problem in Macbeth, I.ii." Notes and Queries. 246. 293-95.

Williams, Deanne. 2001. "Herod's Cities: Cesarea and Sebaste in Twelfth Night." Notes and Queries. 246. 276-78.