University of Virginia Library


Page 131

Shakespeare's Troilus and Cressida: The Relationship of Quarto and Folio

FOR MOST OF SHAKESPEARE'S PLAYS OF WHICH pre-folio quarto texts as well as the folio text itself exist, the relationship between Q and F has been determined.[1] About certain of the plays, however, some question remains. Troilus and Cressida is such a play. Two early printed editions of Troilus and Cressida exist: the quarto printed in 1609 by George Eld for the publishers Richard Bonion and Henry Walley, and the text in the first Folio published in 1623. Conclusive evidence as to the relationship of the Q and F texts has not been presented; it is the purpose of this article to offer new evidence that will establish, on a bibliographical basis, the textual relationship of the two extant early printed editions of the play.

Serious study of the relationship of the Q and F texts of Troilus and Cressida began with the work of the editors of The Cambridge Shakespeare in 1865. W. G. Clark and W. A. Wright expressed a theory about the relationship that was widely accepted for the next fifty years:

The two texts differ in many single words: sometimes the difference is clearly owing to a clerical or typographical error, but in other cases it appears to result from deliberate correction, first by the author himself, and secondly by some less skillful hand. . . . On the whole we are of the opinion that the Quarto was printed from a transcript of the author's original MS.; that his MS. was afterwards revised and slightly altered by the author himself, and that before the first Folio was printed from it, it had been tampered with by another hand.[2]


Page 132
Charles Knight, however, in his edition of the play in 1875, offered a somewhat different theory:
From whatever secondary source it [the quarto] proceeded, there can be no doubt that it was printed from a genuine copy of the great poet. The slight variations between the text of the quarto and of the folio . . . sufficiently show that the original was most accurately printed. The alterations of the folio are not corrections of errors in the original, but, for the most part, slight changes in expression. We have no doubt that each text was printed from a different but genuine copy.[3]

The only detailed study of the 1609 quarto in the nineteenth century was made by H. P. Stokes and appeared in his introduction of the Griggs facsimile of the quarto.[4] He pointed out that "The Folio is careful to give a separate line to the commencement of each speech; indeed this fondness for fresh lines is so great that if Q. by mistake has a new paragraph, F. is sure to 'say ditto'."[5] But he then adds: "This [the stage directions] seems to suggest that the 'Troilus and Cressida' as it appears in the 1st Folio, was printed from the Theater copy."[6]

Appleton Morgan, in his preface to the Bankside edition of the play, wrote: "The variants in the quarto and folio texts (so carefully listed by Mr. Stokes) seem to me all chargeable to typographical sources . . . the later printer might easily have been responsible for them all."[7] Nine years later, Sidney Lee stated that the editors of the folio "evinced distrust of the quarto by printing their text from a different copy."[8] And in 1909, A. W. Pollard concluded that the quarto had not been used in the printing of the folio.[9] Pollard's view, supported by J. Q. Adams,[10] seems to have prevailed for the next twenty years, for in the Yale


Page 133
edition, N. Burton Paradise states, "It is now believed that the Folio was printed from a manuscript belonging to the theater and the Quarto from a copy made for the private use of some friend of the actors."[11]

The first important work on Troilus and Cressida after Pollard was Peter Alexander's study of the quarto in 1928.[12] In this article, Alexander differs with Pollard and comes to the conclusion that the folio text was printed from a copy of the quarto that had been corrected from a manuscript in the possession of Hemming and Condell.[13] In the same year (1928), W. W. Greg had stated in a lecture before the British Academy that the folio text of Troilus and Cressida was printed "not from the previous quarto text, but from an independent manuscript representing substantially the same version."[14]

Two years after Alexander's article, E. K. Chambers wrote that he was "inclined to think that F was set up from a copy of Q, not so much because of a few misreadings and abnormal spellings [Chambers must here refer to Alexander's article] which they have in common, since these might be derived from a common original, as because of a traceable resemblance in typography and the like."[15]


Page 134

In an attempt to refute the views of Alexander and Chambers, Dr. S. A. Tannenbaum published an extended textual study of Troilus and Cressida in 1934.[16] Dr. Tannenbaum explained the similarities between Q and F (many of which he lists) by assuming that these were found in two manuscripts from which Q and F were independently printed, and, having listed certain corruptions in the F text, concludes:

Such absurdities serve to establish the facts that F was not printed from Q, was set up from a manuscript which was difficult to read (or that the copyist could not decipher his original), that the copies for Q [sic] were prepared by different scribes, and that the F text, though better than Q's, is not careful and was not authorized.[17]

Reversing the opinion expressed in his British Academy lecture, W. W. Greg in 1939 came to support Alexander's view: "F appears to have actually been set up from a corrected copy of Q. I think there can be no doubt of this. Besides common errors and unusual spellings there are several points where the arrangement of F can only be explained by peculiarities in Q that the latter is unlikely to have taken over from its copy."[18]

That the textual relationship of Q and F remains obscure is shown by the remarks of two recent editors of the play. G. L. Kittredge, in 1936, wrote: "The relation between the text of the Quarto and that of the Folio is not clear, but the differences are unimportant."[19] And G. B. Harrison, the most recent editor of the play, states: "The quarto issued in 1609 is fairly well printed, but differs in many small points from the text printed in the folio. Each version contains short passages omitted by the other. From certain similarities in the setting of the two texts, it seems either that the folio text was printed from a copy of the quarto carefully but not uniformly corrected from a playhouse copy, or that both texts derive from a common original."[20]


Page 135

Although it is evident that since Alexander's article in 1928, the trend has, in general, been away from the view of Clark, Wright, Pollard, and Adams, the foregoing survey shows that there is still uncertainty about the relationship of Q and F. Until this relationship is settled, study of the text of Troilus and Cressida must rest on doubtful grounds.


The problem of determining that one edition served as copy for a second edition has interested bibliographers for some time. R. B. McKerrow discusses the bibliographical evidence that can be used to determine the relationship of editions, his general conclusion being that the best proof of the relationship of two editions is the demonstration that the second edition reproduces abnormalities that had their origin in the shop where the first edition was printed:

. . . whenever he [the editor] comes across anything abnormal in the typographical arrangement of a text, it will generally pay him to consider whether this may be due to the blind following of an earlier edition.[21]
Although no single piece of evidence that I shall now present can, in itself, be taken as indisputable proof that F was printed from a copy of Q, the cumulative effect of this evidence cannot, I think, be questioned: for this evidence demonstrates that F reproduces peculiarities found in Q that had their origin in George Eld's printing shop. Previous investigators have noted certain of the similarities between Q and F. As these similarities apparently have not been sufficient to clinch the relationship of Q and F, I shall confine myself to kinds of evidence that have not, for the most part, been previously considered.[22] This evidence will be drawn from (1) the use of roman and italic type, (2) speech-heading forms, and (3) significant spellings.


Page 136

Italic and Roman Type:

The fact that F agrees with Q in the use of italic or roman type when its use is dictated by the typographical conventions followed in each cannot, of course, be considered as evidence of their relationship; but when aberrancies in Q are reproduced in F, this agreement becomes significant. It is the normal practice in both Q and F to use italic type for proper names, both in the text and in the stage-directions. Exceptions to this general practice occur, in Q, and it is significant that F then reproduces the roman type used in Q:[23] Stage-direction preceding II.iii.1:

  • Q: Enter Thersites solus.
  • F: Enter Thersites solus.
Stage-direction preceding III.iii.38:
  • Q: Achilles and Patro stand in their tent.
  • F: Enter Achilles and Patroclus in their tent.

Proper names from mythology are, like the names of characters, generally in italic type in both Q and F. The following significant exceptions occur, with F reproducing the roman font employed in Q:

I.iii.89  Q: Sol  F: Sol 
V.ii.174  Q: Neptunes  F: Neptunes 
Although 'Trojan' is never italicized in Q (and only three times in F), 'Myrmidon' is (with one exception) italicized in both, whereas 'Phrigian' is never italicized in either Q or F:              
I.iii.378  Q: Myrmidon  F: Myrmidon 
V.v.33  Q: Myrmidons   F: Myrmidons  Q: Myrmidons   F: Myrmidons  
V.viii.13  Q: Myrmidons   F: Myrmidons  
IV.v.186  Q: Phrigian  F: Phrygian 
IV.v.224  Q: Phrigian  F: Phrygian 
V.x.24  Q: Phrigian  F: Phrygian 


Page 137
'Troy,' a word that occurs fifty-two times in this play, is normally in roman type in both Q and F. The following exceptions occur, with F each time reproducing the aberrant italics of Q:      
III.i.149  Q: Troy   F: Troy  
III.ii.193  Q: Troy   F: Troy  
III.iii.141  Q: Troy   F: Troy  
The following agreement in the typographical treatment of geographic names is significant:        
I.i.103  Q: India   F: India  
I.ii.80  Q: India  F: India 
I.iii.328  Q: libia  F: Lybia 
V.iv.20  Q: Stix  F: Stix 
And finally, F's repetition of Q's aberrant use of italics for the word 'Autumne' at I.ii.138 is significant, especially as elsewhere in F this word is invariably in roman type.


In the pre-1640 dramatic manuscripts that have survived, there is little consistency in the treatment of speech-headings save that they were written in the left margin of the page. In the manuscript of John of Kent, speech-headings are invariably written out in full, and generally so in Believe as Ye List and Ironsides; in other plays they are more often abbreviated.[24] W. W. Greg, describing extant dramatic manuscripts, says: "Speakers' names are usually abbreviated, but the practice varies: Massinger tends to write them in full, while on some pages of Thomas of Woodstock they are reduced to an initial. A scribe would often write a page of text first and add the speakers' names later: this tended to produce bad alinement, but ambiguity was generally avoided by the practice of drawing short lines on the left separating the speeches."[25] In the three pages of the Sir Thomas More manuscript written by Hand D, the thirteen speech-headings for the character Lincoln appear as follows: Lincolne, Linco, Linc, Lin, Lin, Lin, Linc, Linc, Linc, Linc, Lincolne, Lincolne, and Lincoln.

I do not think the question has ever been specifically studied, but it is my impression that compositors attempted, consciously


Page 138
or unconsciously, to normalize speech-headings. If the speech-headings in his copy were abbreviated fairly uniformly, the compositor would presumably follow his copy. If, however, the speech-headings were written out in full, or if they varied radically, the compositor would probably establish a norm abbreviation of his own, at least for the most frequently appearing speech-headings.[26] An analysis of the speech-heading forms in Troilus and Cressida shows a tendency to normalize the abbreviations.

In speech-headings Troilus' name is generally abbreviated in Q to Troy., but there are exceptions. Between F1v and the bottom of F3, the form Troy. appears seventeen times, but on F3v the three speeches of Troilus are each prefixed with Tro. On H1 we find the normal abbreviation Troy., but on the following two pages (H1v and H2) the abbreviation Troyl. is used seven times. On H3 (Troilus does not have a speech on H2v) the shift is back to Troy. On K3v and K4v we find the normal abbreviation Troy. appearing seven times, but on K4, the form Troyl. appears four times. Similar evidence of shifts at the end of the type-page is supplied by the speech-headings for Pandarus. On A4v, for example, we find the form Pand. used nine times, and the catch-word is Pand.; but on B1 there is a shift to Pan., which appears thirteen times. Similarly, the speech-headings for Hector are abbreviated Hect. for his four speeches on L1v, but on L2 the shorter form Hec. is used four times. On G3v the speech-heading for Thersites appears as Thersi. (five times); but on the following page (G4) the form Thers. appears eleven times. Other evidence of a similar nature can be drawn from the speech-headings of other characters.

It will be noted that the shift in these speech-heading forms coincides with the end of the type-page. Now unless we are willing to believe that similar shifts occurred in the MS from which Q was set and that these shifts each time happened to coincide with the end of a type-page of Q, we must accept the fact that these shifts in speech-heading forms represent the tendency to normalize by the compositors of Q.[27] In other words, the speech-headings


Page 139
in Q cannot all represent manuscript readings; for if we suppose that the Q forms reproduce manuscript readings, we must assume against all reason that these shifts in the manuscript coincided with the arbitrary (37 lines to a page) page division of Q. Such a hypothesis, involving such a chain of coincidences, seems to me to be untenable.

Analysis of the speech-headings in F affords additional evidence that F was set from Q. We have seen that the variations in the forms of certain speech-headings in Q cannot represent readings found in the manuscript from which it was set. If, therefore, we find the compositors[28] of F reproducing these variations, we shall have proof that F was set from Q. Although the folio compositors normalized speech-headings somewhat more uniformly than did the quarto compositors, there remains sufficient evidence to establish the relationship of the two texts.

In the quarto, the speech-headings for Pandarus are normally abbreviated to either Pan. or Pand. Both of the folio compositors generally normalize these speech-headings to Pan. The following significant exceptions occur, with F reproducing the longer form found in Q:

I.ii.129  Q: Pand.   F: Pand.  
III.i.110  Q: Pand.   F: Pand.  
III.ii.204  Q: Pand.   F: Pand.  
IV.ii.23  Q: Pand.   F: Pand.  
V.iii.97  Q: Pand.   F: Pand.  
V.iii.99  Q: Pand.   F: Pand.  
V.iii.101  Q: Pand.   F: Pand.  

The speech-headings for Diomedes offer additional evidence that F was set from Q. The speech-heading appears fifty-six times in F, forty-six times as Dio., and ten times as Diom. Eight of the ten cases in which the longer form is used in F reproduce the form found in Q.

There is some confusion in the text of Q as to the spelling of


Page 140
Calchas: it appears as Calcas and Chalcas. The 'h' spelling is used in one Q speech-heading (V.ii.1), and F significantly reproduces the Q form Chal.

It is not the practice in either Q or F to give names in full in the speech-headings, even for the first appearance of a character or at the beginning of a new scene. In the following cases, when the name is given in full in Q it is likewise given in full in F:

I.iii.31  Q: Nestor F: Priam
II.ii.97  Q: Priam F: Priam
III.i.61  Q: Paris F: Paris
V.iii.62  Q: Priam F: Priam
V.iii.71  Q: Priam F: Priam
V.iii.94  Q: Priam F: Priam

In Q the Speech-heading for Ajax is sometines written out in full. In F it is usually abbreviated Aia. F uses the longer form ten times, in each case reproducing the full form found in Q.

And finally, the absence of a speech-heading in Q and F alike at II.iii.1 suggests that F was printed from Q. In Q, the first line on D4v is the centered stage-direction, Enter Thersites solus. His speech, indented but without speech-heading, begins on the following line. In F, the stage direction Enter Thersites solus is centered and the speech, indented but without speech-heading, begins on the following line.


The fact that F reproduces a given spelling in Q cannot be considered evidence of the dependence of F upon Q. Such a single spelling, particularly if it were unusual, might represent a manuscript spelling independently reproduced in the two texts.[29] But if F is independent of Q, we should not find F reproducing indiscriminate spelling variants found in Q; nor should we find F reproducing the characteristic spellings of the two compositors of Q. In two groups of words, F reproduces these indiscriminate variations; and in a third group, F reproduces the characteristic spelling of the two Q compositors.

The use of final ie for y in polysyllables is not a characteristic spelling of the compositors of either Q or F. Seventy-two such


Page 141
spellings, however, are found in Q; the compositors of F reproduce forty-six of these spellings. The explanation of these F spellings must be that the F compositors were influenced by their copy to depart from their normal spelling habits.[30]

In Q, the exclamation 'Oh' is spelled indiscriminately 'O' and 'Oh.' In F, it is, with twenty-one exceptions, spelled 'O.'[31] Only seven times when Q has 'Oh' does F not reproduce the form; and only once does F use 'Oh' when the corresponding Q spelling is 'O.'

Compositors A and B of the folio occasionally depart from their normal spelling habits. Compositor B, who normally spells do and go without a final e, five times spells these words with a final e. And Compositor A, who normally uses a final e, thirteen times spells these words without the characteristic final e. These exceptions to the characteristic spelling practices of the two F compositors furnish the most striking evidence of the relationship of Q and F. Seventeen of these eighteen exceptions reproduce spellings found in Q, and these Q spellings are characteristic spellings of the quarto compositors X and Y. In the following cases, we find the folio compositors reproducing significant Q spellings that are the peculiar spellings of compositors X and Y and cannot therefore all have existed in any manuscript.

I.i.119  Q: goe  F: goe  II.iii.169  Q: do  F: do 
I.iii.308  Q: goe  F: goe  III.iii.90  Q: do  F: do 
II.i.97  Q: goe  F: goe  IV.i.27  Q: doo[32]   F: doo 
II.ii.112  Q: goe  F: goe 
IV.ii.28  Q: do  F: do  III.ii.56  Q: go  F: go 
IV.ii.28  Q: do  F: do  III.ii.56  Q: go  F: go 
V.i.30  Q: do  F: do  III.ii.62  Q: go  F: go 
I.i.42  Q: go  F: go  III.ii.204  Q: go  F: go 
III.i.73  Q: go  F: go  IV.ii.25  Q: go  F: go 


Page 142


The evidence presented here, supported by the evidence that previous investigators have brought to light, demonstrates, I hope, that F was indeed set up from a copy of Q. The implications of this fact may now be briefly considered.

We can, first of all, define the authority of the two extant early printed editions of the play. W. W. Greg writes:

"In seeking to determine which is the most authoritative edition, an editor should distinguish between 'substantive' editions, namely those not derived as to essential character from any other extant edition, and 'derivative' editions, namely those derived, whether immediately or not and with or without minor intentional modification, from some other extant edition. It may be taken that the most authoritative edition will be a substantive one, but the distinction is in practice sometimes difficult to draw, so that this has less significance than at first appears."[33]

The 1609 quarto of Troilus and Cressida is a 'substantive' edition, and the 1623 folio text of Troilus and Cressida is a 'derivative' edition. But it is demonstrable that F is not wholly a derivative edition, for it has some thirty various lines that are not found in Q.[34] These lines are all clearly genuine and have been accepted as such by all editors. Since these lines were not in Q, from which F was set, they must have been supplied from some outside source, i.e. a manuscript.[35] This manuscript supplied not only the lines wanting in Q but also almost certainly some of the substantive variant readings found in F.[36] The folio text of Troilus and Cressida must therefore be defined as a derivative edition containing some substantive readings emanating from an independent source.


Page 143

But defining the authority of Q and F solves only the first of a series of problems confronting the editor of the play. He must next investigate the origin and nature of the manuscript behind the substantive readings in F, as well as the origin and nature of the manuscript from which Q was set; he should, however, make Q—the only substantive edition of the play—the copy-text for his edition. The use of F's variant readings, except to repair obvious corruption in Q, would be dependent on the editor's estimate of the nature and authority of the manuscript from which they come.[37]



A convenient summary is to be found in W. W. Greg, The Editorial Problem in Shakespeare (1942), pp. 188-89.


The Works of William Shakespeare, 2nd. ed. (1892), VI, ix-x. The preface of the first (1865) edition is reprinted in this, the second edition.


The Works of William Shakespeare (1875), II, 369.


Shakespeare Facsimile Quartos, No. 13 [1886], pp. iii-xii. Stokes, writing at a time when the techniques of modern critical and analytical bibliography were almost unknown, made an astute analysis of the data in his possession. He was the first to demonstrate that the title-page without the reference to performance at the Globe had been designed to replace the title-page having this reference. For convenience, line references in this article are to the Griggs facsimile, which follows the Globe edition.


Stokes, op. cit., p. vii, n. 1.


Ibid. I infer that Stokes meant a MS prompt book and not a copy of Q that had been used in the theater.


Troilus and Cressida, ed. by Appleton Morgan (1889), p. 23.


A Life of Shakespeare (1898), p. 369.


Shakespeare Folios and Quartos (1909), p. 58.


A Life of William Shakespeare (1923), p. 538.


The Tragedy of Troilus and Cressida, ed. by N. Burton Paradise (1927), p. 168.


"Troilus and Cressida, 1609," The Library, 4th. ser., IX (1928), 265-86.


Alexander cites as proof that the folio text was printed from a copy of the quarto ten "errors" found alike in both texts, and seven "readings, which, though not recognized as errors by Clark and Wright, are unlikely in independent texts" (pp. 269, 271). Mr. Alexander, as we shall see, is correct in believing that the folio text was printed from a copy of the quarto, but as E. K. Chambers implies (see below), he does not refute the possibility that these seventeen readings existed in the original manuscript or manuscripts from which the quarto and folio were independently printed.


"Principles of Emendation in Shakespeare," Proceedings of the British Academy, XIV (1928), 153. Alexander (op. cit., p. 269, n. 2) says, "Dr. Greg this year in his Shakespeare Lecture to the British Academy has anticipated the conclusion here come to about the relation of the Quarto to Folio," but I can find no evidence to support this statement. Greg later (see below) supports Alexander's view, but in the lecture before the British Academy he clearly stated that the folio was not printed from the quarto.


William Shakespeare (1930), I, 440. An unfortunately worded sentence on the following page (441) has led at least one investigator (Dr. Tannenbaum, see below) to the erroneous conclusion that Chambers contradicts himself, saying first that F was printed from Q and later that F was printed from the author's original manuscript. The sentence reads, "If I am right as to iv.5.96, the manuscript used for F was probably the author's original, and the variations between Q and F are intelligible on the assumption that this was so and that Q was printed from a transcript, perhaps for a private owner." By the phrase "the manuscript used for F" Chambers means, of course, the manuscript used for correcting a copy of Q from which F was then printed.


"A Critique of the Text of Troilus and Cressida," The Shakespeare Association Bulletin, IX (1934), 55-74, 125-44, 198-214.


Tannenbaum, op. cit., p. 207.


The Editorial Problem in Shakespeare, p. 111. In his A Bibliography of the English Printed Drama to the Restoration, also published in 1939, Greg states (p. 414): "The text printed in this [the folio] and subsequent editions contains some lines not in that previously published, and clearly made use of an independent source, but whether it was actually printed from manuscript or from a corrected copy of (a) [the quarto] is not certain."


The Complete Works of Shakespeare, ed. by G. L. Kittredge (1936), p. 879.


Shakespeare Major Plays and the Sonnets, ed. by G. B. Harrison (1948), p. 656.


An Introduction to Bibliography (1927), pp. 184-89.


In my unpublished doctoral dissertation The 1609 Quarto of Troilus and Cressida and Its Relation to the Folio Text of 1623 (The University of Virginia, 1949), the evidence of the relationship of Q and F is presented in somewhat greater detail, and certain kinds of evidence, notably lining practices, are included which cannot be presented here. In addition to its bearing on the present problem, I hope that the evidence here presented will suggest techniques that can successfully be applied to other problems of a similar nature.


An examination of the comedy section of the folio reveals only eight other stage-directions in which roman type is used for names of characters: The Tempest (A5, B2v), Two Gentlemen (C2, C2v), and The Merry Wives (D2, D4, D5). It seems probable that these exceptions (like the two from Troilus and Cressida noted below) can be explained only by assuming that the compositor followed his copy in spite of the typographical conventions of the folio.


Chambers, William Shakespeare, I, 113.


Greg, The Editorial Problem, p. 34.


This can be demonstrated in at least one other dramatic quarto that I have studied. In The Puritaine, printed by Eld in 1607, the name of the character George Pye-board is variously spelled in the text: Pie-board, Pye-boord, Pi-Board, and Pyboard. Two compositors, working in relay, set the text of The Puritaine: one normalized the speech-heading to Pi(y)e., whereas the other used the abbreviation Pi(y)b. The speech-headings in Sir John Oldcastle (1600), at present being investigated, seem to exhibit similar variations.


Spelling tests, supported by the evidence of measures and certain typographical peculiarities, indicate that two compositors were employed in setting the Q text of Troilus and Cressida. I designate them X and Y to distinguish them from the two folio compositors, A and B. See below. The compositors of Q and F, and their work, is to be the subject of another paper.


Like Q, F was set by two compositors, A and B. For their identification and an analysis of their work, see E. E. Willoughby, The Printing of the First Folio of Shakespeare (1932), pp. 57-58.


See Chambers, William Shakespeare, I, 440-41.


Certain of these ie spellings are particularly significant. At I.i.174, the word 'fiftie' is the last word of the line in Q. The spelling with the final ie was probably adopted as a means of justifying the line. In F the word is the first word of the line and does not therefore represent a spelling adopted to justify the line. At II.iii.113, Q's spelling 'courtesie' is reproduced by F, and at IV.v.82 and V.v.15, Q's variant spelling 'curtesie' is likewise reproduced by F.


The spelling 'O' seems to be the normal form used throughout F. In Julius Caesar, a play set by the same two folio compositors (A and B) who set Troilus and Cressida, the exclamation occurs ninety times, each time as 'O'.


This spelling is particularly significant, representing as it does a unique spelling in Q and F. The final o instead of the expected final e in Q may represent foul case.


The Editorial Problem, p. xiii. Greg's definition may be compared with that of the late R. B. McKerrow: "Consideration of the various early printed texts . . . will show us that in the case of any play there is at least one edition which cannot have been derived from any other edition now extant . . . and that others of these texts are derived, with or without intentional modification, from earlier extant editions. Let us call the texts of the first group 'substantive' texts; those of the second 'derived' texts. It is evident that 'the most authoritative text' of which we are in search must be a 'substantive' one." Prologomena for the Oxford Shakespeare (1939), p. 8.


These lines are listed by Stokes, op. cit., p. vii.


That Shakespeare, as an afterthought, added these lines to a copy of Q need not seriously be considered. Most, if not all, of them are to be explained as omissions due to the carelessness of the Q compositors.


Greg, op. cit., p. xiii, distinguishes between 'essential readings' and 'accidents of spelling and punctuation.' In "The Rationale of Copy-Text" in this present volume he defines these categories more precisely, using the terms 'accidentals' and 'substantive readings.'


The present writer is engaged in a study of the variant readings and hopes to publish in the future a systematic survey of the evidence.


Page 144