INTRODUCTION

In the time of my residence in Cambridge, I have seene Tragedyes, Comedyes, Historyes, Pastorals and Shewes, publickly acted, in which the Graduates of good place and reputation have bene specially parted: this is held necessary for the emboldening of their iunior schollers, to arme them with audacity against they come to bee imployed in any publicke exercise, as in the reading of the Dialecticke, Rhetoricke, Ethicke, Mathematick, the Physicke, or Metaphysicke Lectures. It teacheth audacity to the bashfull Grammarian, being newly admitted into the private Colledge, and after matriculated and entered as a member of the University, and makes him a bold Sophister.

Thomas Heywood,
Apology for Actors (1612)

I. THE TRILOGY

1. According to Francis Mere’s Palladis Tamia (1598) Shakespeare, Marlowe, Peele, Kyd, Jonson, and (together with a couple of other academic playwrights) “Doctor Leg of Cambridge” are identified as “our best for tragedy.” Curious about the inclusion of Legge along with the leading lights of the English stage, one turns to that great repository of Elizabethan biography, Fuller’s The Worthies of England, NOTE 1 in which are set forth the basic facts of Legge’s life, together with an account of his his literary career.
2. Born in Norwich in or about 1535, Legge came to Cambridge in 1552, where he entered Corpus Christi College. After transferring to Trinity College he took his Bachelor of Arts in 1556/7 and his Master of Arts in 1560. He was made Fellow of Jesus College in 1568 and elected Master of Gonville and Caius College in 1573 (he was personally selected for this post by his predecessor, Dr. John Caius). Legge received the degree of Doctor of Laws in 1575. Appointed University Commissary in 1579, he served as Vice-Chancellor of the University in 1578-9 and again for part of the academical year 1592- 3. Legge also served as a Mastery in Chancery in 1593, and as a Justice of the Peace in 1597. He died in 1607, and his tomb is to be seen in the chapel of his College. By all accounts he was a tolerant and easygoing Master, even though he was faced with the task of guiding his College through the severe difficulties and sectarian dissentions faced by Cambridge during his lifetime, and was a tutor with a charismatic personality, highly respected and liked by his colleagues and pupils. He was deeply devoted to the interests of his College, where his memory is still cherished.
3. Both for recreational purposes and because of a Humanistic belief that such activity had educational value, the production of Latin plays, both classical works by Plautus, Terence, and Seneca, and also original compositions, was a popular enterprise during the Tudor period. NOTE 2 The account-books of Trinity College show that Legge was involved in dramatic productions at that college during his entire period there. NOTE 3 As distinguished as his academic career was, he his chiefly remembered as the author of two dramatic works. The Destruction of Jerusalem is mentioned by Meres, and Fuller NOTE 4 gives an anecdote that “[Legge] having at last refined it to the purity of the Publique Standard, some Plageary filched it from him, just as it was about to be acted.” The work in question, entitled Solymitana Clades, is also available in The Philological Museum.
4. The other work, which made Legge famous in his own day and secured his lasting reputation, was the Richardus Tertius of 1579.NOTE 5 Towards the beginning of his extensive discussion of this work, G. B. Churchill makes some large claims:NOTE 6

The position of Richardus Tertius in the development of the drama in England is of the highest interest and importance…according to our present knowledge Richardus Tertius appears to have been the first history-play, or “Chronicle History” written England…To Legge, therefore, was due the turning of the drama in England in an entirely new direction. It was he who first perceived that English history as related by the chroniclers possessed as great a store of dramatic material as the classical saga, or Biblical story…It was he who while the national consciousness and pride was swelling under the great events of Elizabeth’s reign, first gave [playwrights] a new object and incitement in the history of England’s past, presented in living form upon the stage.

5. We may probably imagine that the lesson that the English Chronicle writers contained plenty of material apt for dramatization was not lost on certain Cambridge undergraduates of the time who may well have been the audience, possibly read the play, and certainly must have heard about it: Lodge and Green. Two other students who came up to Cambridge not long after 1579, Marlowe and Nashe, NOTE 7 would have had ample opportunity to hear about, and perhaps read, a work which had obviously made a deep impression, and if it indeed was replayed in the next decade — its possible performance history will be discussed below — may also have seen it acted.7
6. Churchill bases his claim for taking Richardus Tertius seriously on the grounds of historical importance, not intrinsic literary merit. But the work does not lack its modern enthusiasts. A. L. Rowse has discussed its value in more absolute terms:NOTE 8

Despite some obvious shortcomings, it is not difficult to share this favorable reaction. In some ways, one must admit, the writing of a Latin dramatic work on an English historical subject produces results that strike a modern reader as merely silly. Legge repeatedly indulges in wild anachronisms in attributing Roman customs, institutions, and beliefs to his fifteenth century characters. They have domestic lares et penates, the status of a great man is measured by the number of clientes flocking to his door,NOTE 9 Richard is frequently referred to as “the Claudian,” and so forth. In no other sphere does this anachronizing tendency produce as strange results as in that of religion: are his fifteenth century Englishmen Christians or do they believe in the machinery of the Roman pantheon, in one sense presided over by Jupiter Tonans and the Olympic gods, in another by Fortune and the Fates? The somewhat schizophrenic answer seems to be that they are simultaneously Christian and Pagan. Then too, some characters (notably Brackenbury and King Edward’s eldest daughter, both of whom we meet in Actio III) are imbued with Stoic philosophy of the Senecan, suicide-glamorizing variety: one does not have to do evil, or exist in evil circumstances, as long as one knows how to die.
7. But if one is willing to overlook such flaws, at a deeper level the experiment is solidly successful. Richard III, at least as portrayed by Thomas More and the chroniclers who depend on him, is the English archetype of the Renaissance tyrant, of Machiavellian Man in action.NOTE 10 It is for this reason, and not just because his downfall paved the way for the rise of the Tudor dynasty and so in some sense for that of the modern English nation, that his career exerted such fascination for the Elizabethans. As a writer of Renaissance Latin tragedy, Legge of course draws much inspiration from the only available Roman prototype, Seneca. The problem of tyranny looms large in Senecan tragedy, in plays such as Hercules Furens, Thyestes, the unfinished Phoenissae, and the pseudo-Senecan Octavia, and therefore Senecan tragedy provides an effective instrument for the dramatic examination of Richard and his career. And the characteristic Senecan atmosphere of oppressive menace and intense dread, the Senecan psychological analysis of the tyrant, and of those who must live, suffer and die under his rule, are altogether suitable for this subject. But Legge does not simply appropriate this Senecan tool for studying the dynamics of Richard’s tyranny. He introduces a major improvement. Everyone who has ever mentioned Richardus Tertius in print has described it as a play (and some have severely criticized it for its monstrous length and fifteen-Act structure). It is in fact a trilogy (each one identified as an Actio), and the result is interesting. A self-contained classical tragedy focuses tightly on a single critical moment, but a trilogy can tell a more complex story spanning a large amount of time. Seneca provides a portrait gallery of tyrants, but he can only represent tyranny in full bloom. By adopting the trilogy format, Legge can show the tyrant’s rise to power in the first two plays, and trace the trajectory of his downfall in the third.
8. There is another reason for paying attention to Richardus Tertius. Over the past couple of centuries it has been relatively easy to take a dismissive attitude towards Seneca-based Renaissance tragedy precisely because readers (including many professional Latinists) have been so willing to dismiss Seneca himself. During the past two or three decades a major revolution has taken place in Latin studies. Classicists have revised their opinion of Seneca sharply upward, in large part because they perceive him grappling with the problem of tyranny, which looms so large in our own times. Modern readers know full well that such monsters of evil as Lycus and Atreus do exist, that the kind of overwrought, neurotic, even hysterical psychology of Seneca’s characters is what one indeed does find in those condemned to live under a tyrant’s rule. The elements of Senecan tragedy that in some ages of the world might have been dismissed as products of morbid fantasy, or as examples of overreaching sensationalism, are now recognized for what they are — exercises in accurate, penetrating political and psychological analysis. Modern readers are also more prepared to acknowledge the highly dramatic quality of Seneca’s plays. Senecan tragedy is no longer regarded as intolerably static and rhetorical, and even the canard that his plays were not written for dramatic production (which was originated by literary critics in the early nineteenth century as little more than an expression of contemporary taste, and only later adopted as an article of faith by classicists) is increasingly being called into question. The resurgence of interest in Seneca creates conditions under which neo-Latin tragedy can better be studied and appreciated.
9. One of the chief objects of Churchill’s lengthy examination of Richardus Tertius was to reveal Senecan influence. But at the most fundamental level of all, that of dramatic conception, more influence is exerted by Aeschylus. NOTE 11 Legge superimposes the structure of an Aeschylean trilogy on his ostensibly Senecan models.NOTE 12 Each Actio (in an appended note to Solymitana Clades written in English the word “Action” is used to identify a single play in that trilogy, and surely the word is suggested by the Greek δρᾶμα) is a separate play divided into five Acts; many of these are subdivided into scenes, some quite short, with changes of location occurring between the latter: the Octavia Praetexta included in the Senecan corpus provides classical precedent for this technique. The formal title inscription preserved by some manuscripts shows that the Actiones were performed on successive evenings.
10. For all Legge’s dependence on Greek and Latin models, many distinctive features of classical tragedy are abandoned. Each Actio ends with a musical passage, sung and instrumentally accompanied. But the classical tragic chorus, perpetually onstage (or at least nearly so), periodically intervening to reflect on the dramatic action, is eliminated. Legge’s “choruses” are really just crowd scenes. Choral passages are not used for such traditional purposes as delineating the play’s structure, marking the passage of time, or covering the interval while actors change into new costumes for the next episode. In view of the frequently dubious relevance of Senecan choral passages to the dramatic action, ΝΟΤΕ 13 and the rum-te-tum inflexibility of their meters, so at odds with the fascinating metrical variety of their Greek counterparts, many readers will doubtless regard the elimination of the tragic chorus as merciful.
10. In short, as was only natural for a humanistically educated writer of serious Latin drama, Legge wrote in the vein of Senecan tragedy, sometimes appropriating tags or even entire lines from that poet and using typically Senecan vocabulary, imagery, and rhetorical devices. He also liberally borrowed from the Senecan corpus in such matters as framing individual scenes, character-depiction, and dramatic technique. But, although the author maintained a certain outward resemblance to classical tragedy, he was in fact going about the business of inventing a new form of drama outstripping, in terms of originality and inventiveness, any other University plays, or, for that matter, any Tudor play written before his time.
11. That Legge was consciously creating a new kind of dramaΝΟΤΕ 14 is abundantly shown by the fact that he cheerfully ignored many of the real or supposed conventions of classical tragedy. Besides abandoning the classical tragic chorus, he makes no pretense of adhering to such conventions as the Unities or the three-actor rule. And so Legge’s relation to classical drama is an interesting one: he felt entirely free to pick and choose which precedents to follow and which to abandon. He reasoned that he was not obliged to adhere to classical models precisely because he was working in a new genre unbound by the strictures of classical drama. On the other hand, this new dramatic form was to have certain standards of its own. Certainly, the most important of these is moral earnestness coupled with ethical instruction. Legge’s usual (albeit not invariable) close adherence to his historical source has been disparagingly dismissed as “slavish.”ΝΟΤΕ 15 And, surely, critics would give the same opinion of Legge’s relation to his source for Solymitana Clades, Flavius Josephus’ History of the Jewish Wars. But in both trilogies Legge gives ample evidence that he can write an independent scene or speech when he wants. More likely, he stuck as closely to his sources as he did out of an intellectual conviction that his new kind of historical drama ought to be faithful to the historical record. Hence, if the usual division of Elizabethan plays into the three categories of tragedies, comedies, and history plays is a valid one, Legge’s two trilogies must be collectively viewed as an important milestone in the development of that dramatic form. If we set aside excursions into historical fantasy of the Gorboduc type, Legge comes as close as anybody to being the father of English historical drama.
12. The use of stage directions is foreign to ancient tragedy. Whatever indications for staging an ancient playwright cared to make had to be incorporated into the words of the actual text. The stage directions found in the manuscripts of Richardus Tertius are, at least at some points, unusually copious and detailed. They are an important clue to Legge’s intentions: far from writing dryasdust closet drama, he was aiming at an impressive coup de theatre.16
13. With one exception, the manuscripts indicate that Richardus Tertius was first performed at St. John’s College in 1579 (the suggestion has been made that they mean 1580 New Style, but Nelson II.919 pointed out that college account books strongly support a dating to 1578/9). The manuscript preserved at Gonville and Caius College and identified as B here dates the work to 1573. Although it is perhaps not out of the question that it was written earlier but for some reason not performed until 1579, Churchill (p. 267) was probably right in dismissing this latter date as a copyist’s error.
14. The performance of Richardus Tertius must have been an event of some magnitude. Both the lists of dramatis personae found at the beginning of each of the three Actiones and the stage directions indicate the work’s spectacular nature. Obviously, one of Legge’s concerns was to involve a large number of people in the production: actors, mute parts, “extras” in crowd scenes and processionals, and instrumental musicians. On the basis of the actor lists preserved by the mss. (reproduced by Nelson II.944-6) one gathers that the production probably involved the participation of over a hundred men and boys serving as actors, mutes, extras, singers and musicians. The interest in presenting an impressive visual spectacle in the coronation scene at the end of Actio II and the battle scene at the end of Actio III (in which the stage directions even call for gunfire) is scarcely out of keeping with the spirit of Senecan tragedy, and Cambridge certainly had the resources to turn its academic ceremonial pomp to theatrical purposes. A dining hall at one of the larger colleges, with its stage-like raised platform at one end built to accommodate the high table, provides an excellent venue for dramatic production. The reason why the trilogy was produced at St. John’s rather than Legge’s own College is obvious: St. John’s was more notable for its dramatic activity and possessed the acting talent and technical expertise necessary to mount this very demanding work — all the identifiable performers were individuals affiliated with St. John’s — and had a larger dining hall to accommodate a larger audience.. NOTE 16
15. With its vivid characters, dramatic tension, powerful emotional scenes, and electrifying developments, Richardus Tertius was meant to be theatrical in the fullest sense of the word, and there is every reason for thinking that it scored a notable “hit.” The trilogy must have exerted an especial impact on its audience insofar as its basic premise is that events of national history are worthy of serious dramatization than incidents from the Bible, classical mythology, or Graeco-Roman history. In an age of rising patriotic sentiment, this was a powerful message. The Fuller biography speaks of it being received “with great applause.” When defending tragedy in An Apologie of Poetrie (1598), Sir John HaringtonΝΟΤΕ 17 writes “…and for Tragedies, to omit other famous Tragedies: That, that was playd at S. Johns in Cambridge, of Richard the 3. would move (I thinke) Phalaris the tyraunt, and terrifie all tyrannous minded men, from following their foolish ambitious humors.” Harington was a Cambridge undergraduate in 1579; evidently he had been in the audience, and Richardus made a deep impression.
16. Another measure of our work’s success is evidence, admittedly less than conclusive, suggesting that Richardus Tertius was performed in one or more revivals. The Bodleian MS. Tanner 306 (here identified as G) states that it was performed at St. John’s College on March 17, 1582, in the presence of the Earl of Essex. Although, as Lordi (p. vi) pointed out, Essex had received his M. A. in July 1581 (he belonged to Trinity College, and after taking his degree at the advanced age of fourteen retired to his country seat in Pembrokeshire), and although G in its dramatis personae lists the names of the 1579 performers, these considerations do not absolutely exclude the possibility of a 1582 revival. The date (March 17, 1582) looks interestingly circumstantial. A British Library manuscript of Richardus Tertius, identified as F here, was written by Henry LacyNOTE 18 of Trinity College in 1586. Although Churchill (p. 395 — by a typographical error he dates the manuscript to 1856) states that this manuscript was executed “for presentation at Trinity College,” there is no internal evidence tending to suggest that this copy was prepared in connection with a revival performance. Finally, we may dismiss the theory of F. G. FleayNOTE 19 that Richardus Tertius was revamped, by the addition of the Actio III Epilogue, for performance during a proposed visit by the Queen during the 1592/3 Christmas season, that failed to eventuate. The most that can be said about this conjecture is that if Elizabeth had come to Cambridge, Richardus Tertius would have been a judicious choice for her entertainment. Certainly the theory that the Epilogue was added for this purpose was added for this purpose is wrong: it stands in dated manuscripts (identified as C, E, F and H here) executed before 1592.
17. This evidence for one or more revival performances in the 1580’s, inconclusive as it may be, is of potential importance for a prospective editor of Richardus Tertius. For certain phenomena found in some manuscripts are possibly to be explained by invoking the theory that the trilogy was somewhat reworked for one or more actual or at least proposed revivals. Some manuscripts (those identified as B, G, I, and J) contain lines and passages differing from those of the other manuscripts in ways that cannot be attributed to the normal operations of copying corruption. In G and J two scenes of Actio I (I.v.iii, I.v.iv) have been doctored so as to give Lord Lovell a role in that Actio.NOTE 20 The copyist of E, who took the trouble to collect several documents pertinent to the work’s production, appends a plan for dividing Actio III into two Actiones, so that the entire work might be performed over four nights, and begins this document with an obviously defective transcript of a new scene to begin Actio III, an interview between Brackenbury and Richard’s agent John Green (the text of this is given in Appendix I). One is tempted to suspect that these radical revisions were made in connection with a revival performance, but the possibility of course exists that the revival in question may have been planned but failed to occur.
18. Legge’s primary source is Edward Hall’s The Union of the Noble and Illustre Famelies of Lancastre and York (probably first published in 1548), although Churchill (pp. 484 - 524) shows that individual details are taken from Thomas More’s History of Richard the Third (Lordi argues that he used the 1557 Rastell edition of the English version of this biography) and other sources. We have seen that Legge’s dependence on Hall has been called “slavish.” This assessment misses the essential point. Hall’s Chronicle is a highly exciting work, rich in colorful circumstance and incident. He uses the almost novelistic technique of putting frequent speeches in the mouths of his characters. The book is, in a word, dramatic. Legge’s essential stroke of inspiration was the realization that this material could easily be converted into actual drama. This was all the more true since the speeches in Hall are themselves elaborately rhetorical and bombastic and (like Thomas More’s biography, upon which it ultimately depends) not untouched by classical influence and perhaps by Seneca himself, and so are highly congenial to the spirit of Senecan drama. Because of its exciting and readable nature, Hall’s Chronicle was a highly popular item of Tudor literature. Therefore, rather than dismissing Legge’s work as a merely slavish imitation of Hall, it would be closer to the mark to regard it as a kind of forerunner of a modern dramatic or cinematic adaptation of a familiar work of literature: the audience derives pleasure from it precisely because it can see the literary prototype reflected in the adaptation. (As evidence of the fact that Legge could confidently capitalize on his audience’s knowledge of Hall, one can point to passages that allude to facts retailed by Hall in such an offhand way that they would be completely meaningless in the absence of such familiarity; this is true, for example, of the reference to Richard’s visit to Gloucester at 2750, and of his diplomatic overtures to Scotland at 3822ff.) In this special subspecies of drama, a playwright’s originality perhaps does not count for as much as it normally would.
19. Legge aimed at a dramatic production of great sweep, vastly larger and more complex in structure and narrative scope than a Senecan tragedy. This sweep most visibly involves time and space. It is also sociological: the characters represent nearly every level of English society, and at many points we are allowed to observe how the average Englishman reacts to developments of high politics. Of course, Legge could only achieve the sweep of a Chronicle play at the cost of ignoring the classical Unities. It might therefore be argued that he was obliged to forego any substantial dramatic unity or effective plot construction, at least as understood by the ancients, in order to reproduce the complex sprawl of Chronicle narrative. This observation, while not entirely wrong, requires modification in the light of the fact that Richard Tertius is a trilogy rather than a single play. By the spirit of classical standards, the first two Actiones indeed do display reasonable unity and tightness of focus. Actio I (which could be entitled The Fall of the Woodvilles) shows Richard creating the conditions under which he can seize power. by his machinations the Woodville-Grey combination is progressively undone; the two little boys are brought under his control; and the few others who could oppose his capture of the throne are neutralized. The Actio covers a reasonably short span of time, the crucial weeks following the death of Edward IV in April - June, 1483. Its action begins in an atmosphere of dread and tension, which steadily builds until a genuine climax is reached: the Council meeting where Richard triumphs over his opponents and reveals his true nature. By classical standards of plot construction and unity, Actio I is the most successful of the three Actiones that comprise Richardus Tertius. Actio II is shorter and less exciting. This is probably deliberate. In the Oresteia trilogy the Choephoroi is in some sense a similar interlude between the higher-pitched Agamemnon and Eumenides. This Actio is also focused on a short period, covering the deposition of Edward V and the coronation of Richard in May - June, 1483, and it too leads up to a single climax, the coronation ceremony.
20. So the observation that Richardus Tertius is lacking in unity because it diligently follows Hall’s Chronicle really boils down to a criticism of Actio III. And here criticism is more cogent. The Actio dramatizes a variety of events: the killing of the little princes; the machinations leading to Buckingham’s abortive rebellion and its collapse; Richard’s attempts to shore up his tottering reign, centering on the poisoning of his wife in order to clear the way for a proposed marriage to Edward IV’s eldest daughter, which does not eventuate; his diplomatic maneuvers designed to stymy Henry of Richmond in Brittany; the events leading up to Bosworth; the battle itself. Actiones I and II between them dramatize Hall’s relatively short chapter and cover events that occurred during late April to early July, 1483, so that in Actio III Legge is obliged to get through all the significant material contained in Hall’s chapter on Richard’s reign, spanning a period of slightly less than two years. Here, one confesses, he has not solved the problem of imposing any genuine dramatic unity on all this material. Possibly he would have done better to make the killing of the boys the climax of his second Actio. Such a redisposition of the narrative material would have been quite possible. Actio I is just about 1800 lines long, and Actio III contains nearly two thousand. By contrast, Actio II is less than a thousand lines long, and even if we acknowledge that Legge made it so short in order to accommodate the extensive dumb-show of the coronation, such brevity is remarkable. If he had made the killing of the princes the climax of Actio II, he would have achieved a better structural balance and created the possibility for greater dramatic unity in Actio III. He then could have found a single thread which could have linked the material covered in that Actio. As we know from Shakespeare, such a thread is available: the deteriorating personality of the wrongdoer as his crimes undermine him spiritually and psychologically, even before he receives his just deserts. While Legge does portray Richard as gnawed by guilt and anxiety, and wildly vacillating between hope and despair in reaction to developing events, he does not quite manage to develop this into a genuine unifying theme. Something of the feeling of classical tragedy could have been achieved by stressing the inevitability of his downfall: the active hand of divine retribution, or the machination of the Fates, could at every turn be shown to be working his ruin. Or again, Shakespeare introduced another unifying theme in the working-out of the curse imposed on Richard by Queen Margaret. But in Legge’s defense it must be said that achieving adequate dramatic unity in Chronicle plays was a problem faced by Elizabethan playwrights in general. It is unfair to come down very hard on Legge, the pioneer of this genre, for failing to solve its inherent problems, when some of the Chronicle plays written, wholly or at least in part, by Shakespeare himself are more or less susceptible to the same criticism. A fairer observation would be that Legge set himself an impossible task because his ideal of faithfulness to the historical record is essentially irreconcilable with the requirements of sound dramatic construction.
21. To the reader familiar with other Tudor dramatic portraits of Richard, in The True Tragedy of Richard the Third (printed 1594), shown by Churchill to be a source for Shakespeare (pp. 484 - 524 — there is no chronological problem here because there was often a considerable time-lag between a play’s writing and its printing), and Shakespeare’s 3 Henry VI and Richard III, his portrayal by Legge will be in some ways familiar. Hall and the other Chronicle writers take their cue from the History of king Richard III commonly attributed to Thomas More, both in that they closely follow his account down to the point where it breaks off, and because More manufactured the orthodox Tudor account of Richard and his misdeeds that they uncritically reproduce. This account was repeated and embellished by Polydore Vergil in Book XXV of his Anglica Historia (1534), who provided the primary source for the Chronicle writers for that part of Richard’s career not covered by the More biography. As is only to be expected, Legge presents yet another standard portrait of Richard as a thoroughly evil king. Absence of mention of some of his supposed crimes (the assassination of Henry VI, the killing of the captives after the battle of Tewkesbury), and only passing references to his alleged responsibility for the death of his brother the Duke of Clarence, are more than counterbalanced by a fairly direct accusation that he poisoned his wife Anne; indeed, Legge credits this crime to him far more explicitly than any other Tudor writer (some, such as Polydore Vergil, write of it as a rumor). Richard’s chief characteristics are, predictably, lust for power pitched to such a high degree that it is described in the language of disease and madness, murderousness, guile, and hypocrisy. But Legge’s Richard differs greatly from the Richard we know from Shakespeare and, for that matter, the Richard of More, the Chronicle writers, and The True Tragedy. He is a thoroughly wicked man, but he is neither a monster nor a genius of evil. When I say he is not a monster, I mean this quite literally. The single most startling thing about Legge’s portrait of Richard, in which he is unique among purveyors of the Tudor account, is an astonishing absence of reference to his supposed physical deformities; the only such mention is a mention of his withered arm at 1669. This failure to describe him as a deformed hunchback is not without deep significance. In the orthodox Tudor story his repulsive physical appearance is an outer manifestation of his inner spiritual condition; physically as well as morally he is set apart from mankind. since Hall gives the standard picturesque description of Richard’s deformities (pp. 342ff.),NOTE 21 the decision not to represent him as a physical monster was Legge’s own.
22. Nor is Richard portrayed as a genius of evil. Churchill (p. 399) pointed out that “The Richard of The True Tragedy is not only central but dominating, not merely attracts the chief interest but absorbs practically all of it. The play is not the chronicle-history of a reign, it is purely the history of a character.” This Richard is not merely consumed and corrupted by a lust for power. In him, ambition is so great and so single-minded that he becomes the essence of ambition, ambition personified. As such, he is not so much a person as an elemental force and so he, surrounded by merely human characters, is the play’s sole focus of interest, the single source of its dynamism. This is of course equally true of his characterization in Richard III. Then too, Shakespeare's Richard is distinctly reminiscent of Machiavelli as represented in the popular imagine: he is crafty and good at using words to manipulate those around him. Here, it must be emphasized that the decision to make Richard an essentially weak man is was’s own, that cannot be ascribed to his dependence on Hall. Richard, as presented by More and the Chronicle writers who essentially reproduce his portrait, is quite anticipatory of the familiar Richard of the later dramatists. As Churchill observes (p. 375), More’s Richard is “strong: strong in fertility of resource, strong in craft, strong in courage.” Nor (despite what Churchill writes on p. 377) can the decision to make him a weak man be explained by his dependence on Seneca, although at first sight such might seem to be the case. In some ways Legge models Richard on the tyrants in the Senecan corpus: in his murder of his two young nephews he resembles Atreus in Thyestes, in his brutal courtship of young Elizabeth he looks like Lycus courting Megara in Hercules Furens, in his calculated reliance on violence, terror and criminality, he bears a generic resemblance to Atreus and also to Eteocles of Phoenissae and Nero of Octavia. But each of these Senecan tyrants is single-minded, strong and decisive.NOTE 22 Richard’s spells of self-doubt and indecision tend to come out in dialogues with his counselors, which usually have little or no basis in Hall. Such scenes bear a superficial generic resemblance to some in the Senecan corpus in which a tyrant engages in dialogue with an underling or advisor: the conversation of Atreus and the Satelles in Thyestes, that of Iocasta and Eteocles in Phoenissae, and Nero and Seneca in Octavia. In Seneca, these scenes are a means of plumbing the depths of the tyrant’s villainy and delineating his psychology and policy of government. But the resemblance of Legge’s counselor scenes to these Senecan passages is more apparent than real, since they are used for a diametrically opposite purpose: they reveal Richard’s psychological weak spots rather than explore the sources of his strength.
23. Then too, the reader cannot help being disturbed by Richard’s meaching oration to his soldiers on the day of Bosworth. Legge, to be sure, reproduces his speech in Hall, who makes him say (p. 415) “And although in the adepcion and obteining of ye Garlande, I being seduced and provoked by sinister counsail and diabolical temptacion did commyt a facynerous and detestable acte, Yet I have with strayte penaunce and salte teyres (as I trust) expiated and clerely purged the same offence, which abhominable crime I require you of frendship as clerely to forget, as I dayly do remember to deplore and lament thesame…” This public confession of the murder of the princes, coupled with a self-dramatizing display of contrition, is not what the theatergoer wants. Richard ought to go to his death defiantly, Richard-like to the end. Shakespeare had the good sense to discard this Chronicle speech in favor of something more aggressive and dramatically suitable. Legge would have been well advised to do the same. His portrait of Richard is less psychologizing than those of later Elizabethan stage Richards. It is surprising that, although he writes monologues or extended asides to probe the thoughts of a number of his other characters, he waits until Actio III to give any such passage to Richard, and what is revealed is the man's inner doubts and fears. This presents a sharp contrast to later dramatic treatments of Richard, where this technique plays a major role: both The True Tragedy and Richard III commence with a soliloquy in which his essential ambition and malice are laid bare, and there is a similar soliloquy in 3 Henry VI III.ii. Rather remarkably, in I.v.i Catesby is given a short monologue that is, by the standards of these later plays, far more Richard-like than any speech allowed Richard himself. English academic drama was not without its brilliantly delineated characters (such as the title character in Edward Forsett’s 1581 Pedantius and Oeneus in William Gager’s 1582 Meleager). By making his Richard a coward who kills out of fear rather than a towering figure of evil, Legge may well have had a special purpose: to use Richard as an example of Plato's observations in Book IX of the Republic that the tyrant is the unhappiest man alive because he is constantly hagridden by fear. But, as morally instructive as this presentation may be, it still seems artistically miscalculated. One cannot help concluding, Legge missed the opportunity of adding another figure to the gallery of memorable characters in academic drama.

II. INFLUENCE

24. If Richardus Tertius is an interesting landmark in the development of Elizabethan historical drama, it also commends itself to our attention as a possible source, direct or indirect, for Richard III. Churchill (pp. 393 - 5), upon whom later discussions essentially depend,NOTE 23 adopted the only reasonable strategy for addressing this question. Legge and Shakespeare depended on the same Chronicle sources, insofar as Grafton and Holinshed repeat Hall’s account in slightly embellished form and everybody is ultimately dependent on More and Polydore Vergil. Therefore, all resemblances between Richardus Tertius and Richard III, that can be accounted for by reference to their Chronicle sources must be set aside , and discussion ought to be limited to any shared features that cannot be explained in this way.NOTE 24 The attention of Churchill and later writers who have participated in this discussion has traditionally focused on Richard’s wooing of the eldest daughter of Queen Elizabeth in the final scene of Actio III, Act iv, and its resemblance to Richard’s wooing of Anne in Act I of Richard III. These scenes of course involve different women, and do not have the same outcome (Legge’s Richard fails to gain his object, Shakespeare’s succeeds). But in both scenes Richard tries to gain a political object by forcing his attentions on an unwilling woman, using bullying tactics to obtain his goal. And in the course of both scenes, he draws a dagger and offers to let the woman kill him. Legge’s scene is based on nothing in the Chronicles and is, as far as we know, his own invention. In a general way it is modeled on Lycus’ courtship of Megara in Seneca’s Hercules Furens, but the detail of the proffered dagger is not Senecan. Churchill observes other resemblances between these two scenes: “Notably alike in both plays is the plea with which Richard answers the charge of murdering his nephews; in Legge, sic fatis placet, in Shakespeare, ‘All unavoided is the doom of destiny.’ Notable too is the frequent use of Senecan stichomythia.”
25. Resemblances between Richardus Tertius and Richard III are not limited to the wooing scenes. Actio III, Act iii consists of a monologue by Gloucester followed by the entry, in quick succession, of two messengers who arrive with the glad news that the Buckingham revolt has collapsed. When the manuscripts use new lists of speakers to indicate entrances and exits rather than scene-breaks in the modern sense (a subject discussed below), they serve to obscure the interesting resemblance of this scene to the final portion of Richard III IV.v, in which a series of messengers enter in rapid order giving their reports to Richard. This resemblance looks significant, since the Chronicle writers give no information about how he learned that the revolt was ended. Also, both Legge and Shakespeare introduce occasional scenes in which anonymous Citizens converse about the doings of Richard and the aristocracy. Shakespeare may have learned from Legge this convenient dramatic technique for reporting contemporary public opinion. Churchill observes further resemblances. In the Chronicles “the counsaillers of kyng Richard” advise him not to kill Lord Strange until the battle of Bosworth has been won. In Legge and Shakespeare, the Duke of Norfolk offers this advice, and “Norfolk’s ‘after the battle let George Stanley die’ sounds very like Norfolk’s post bella gnatus patris expiet scelus.” Furthermore:

The wretched, bloody, and usurping boar,
That spoil’d your summer fields and fruitful vines

for which The True Tragedy, where “the boar” is not mentioned, will not account. On this, however, one cannot build much.

26. In his discussion, Bullough offers the generalizing observation that women play a very significant role in both Richardus Tertius and Richard III, in striking contrast to The True Tragedy of Richard III, where (save in the subplot about Jane Shore) they are unimportant. To explain the resemblance of the wooing scene in Richardus Tertius to the wooing of Anne in Richard III, Churchill suggested that Shakespeare may have been familiar with one or more dramatic treatments of the subject besides The True Tragedy, now lost, written by someone who had read Legge. Thomas Heywood wrote a prologue and epilogue for a Richard III play: “A young witty Lad played the part of Richard the third: at the Red Bull: the Author because hee was interessed in the Play to incourage him, wrot him this Prologue and Epilogue” (Pleasant Dialogues and Dramas, 1637, fol. 247).NOTE 25 Also, the suggestion has been made that an old play entitled Buckingham, which was revived by Sussex’ Men at the Rose in December and January 1593/4, may have been a prototype for Richard III. NOTE 26 But there seems to be no grounds for excluding the possibility that Shakespeare had read Richardus Tertius. Ben Jonson’s famous canard that he would not have known enough Latin to be able to read this work has been scientifically discredited, NOTE 27 and Bullough, at any rate, was willing to concede that he had direct knowledge of it.
27. Shakespeare was not the only Elizabethan dramatist familiar with Richardus Tertius. In the works of Legge’s Oxford contemporary William Gager [1555 - 1622] there are several echoes of the trilogy. In the Meleager of 1582 (964) we find the phrase procerum cohors, commonly used in Richardus, and at 1090 Gager uses the phrase invidiae capax, a phrase not used by any classical Roman poet but found at R. T. 2311. A passage in the same play (341ff.), in which Meleager tries to persuade Atalanta that a girl’s virginity does not belong entirely to herself, is based on Catullus lxx.62 - 4, but seems also to borrow from R. T. 3232ff. In Gager’s 1592 Ulysses Redux, when Penelope’s suitors realize that Ulysses is going to shoot them all down, they collectively shout arma, arma, cives; arma quis miseris dabit? (1679). Surely this echoes R. T. 579 urbs urbs, cives, ad arma, ad arma! Equally striking is a scene in Ulysses Redux (1479ff.) in which the Suitors plot Telemachus’ death. Since this stichomythic debate deals with the killing of an heir who is an obstacle to a proposed seizure of power, it resembles R. T. 1285ff. more than any Senecan scene. Likewise, Richard Burton appears to have known Richardus. At Anatomy of Melancholy I.ii.ii.xi (p. I.280 of the Everyman edition) he writes “Ambition, a proud covetousness…one defines it as a pleasant poison.” Since no ancient author calls ambition a dulce venenum, this would appear to be a reference to R. T. 8, unless, of course, both writers are echoing a common source I have not identified. More significantly, in Burton’s comedy 1617 comedy Philosophaster (most recently edited by Connie McQuillen, Binghamton N. Y., 1993) III.vi, Antonius unsuccessfully woos Camaena. He produces a dagger and offers to let her stab him as an alternative to satisfying his desires then and there, but she rejects him with disdain. This scene looks much like a comic equivalent, if not actual parody, either of Richard’s wooing of Anne in Shakespeare or of his wooing of Edward’s elder daughter in Legge. The fact that Antonius’ suit is unsuccessful perhaps suggests that the inspiration for this scene came from Legge, not Shakespeare.

III. THE NEED FOR A CRITICAL EDITION

28. In the title of the Clare College manuscript (identified as C here) copied by Anthony Cade in 1583 (New Style), the words impressum U. C. are written and then crossed out. Lordi (pp. vii - x) argued that Richardus Tertius was scheduled for publication at Cambridge, but that the hostility of the London authorities to the operation of a printing press there quashed the project. In 1844 Barron FieldNOTE 28 printed a text that was frankly no more than a transcription of a single manuscript (owned by Emmanuel College and identified as D here). His transcript was reprinted by W. HazlittNOTE 29 later in the century. Of Hazlitt’s text Fleay NOTE 30 wrote that “it sadly wants editing,” and that it contains “nearly a thousand errors.” Obviously, therefore what was needed was a sound text based on a collation of all available text sources. In 1979 Robert J. Lordi rose to the challenge and published a new text designated by its title as a critical edition. In point of fact, this was no real attempt at producing such an edition: Lordi presented a collation of nine of the eleven known manuscripts for all of Actio I and for the first hundred lines each of Actio II, Act III, and Actio III, Act III, and tried to establish the relation of these manuscripts to each other. He argued the supreme importance of C, which, according to its colophon, was copied by Cade from the author’s own holograph. For most of the trilogy his working method, therefore, was only to consult the readings of the other manuscripts when he had cause to doubt C’s veracity. Since (as can be seen from his accompanying translation) his knowledge of Latin was slight, he entertained few such doubts, and so the text he printed is little more than a transcript of C, and is superior to Barron Field’s text precisely to whatever degree that C is a better manuscript than D. What is lacking is any real editorial intervention and, as will now be demonstrated, C has no preeminent authority as a witness to what Legge may have written.
29. On the face of things, the need for a critical edition is not self-evident. If we have at our disposal a manuscript copied directly from the author’s holograph, there should be no need for much editorial effort unless Cade was a very incompetent copyist indeed. An inspection of the facts suggests that such is not the case: there are very large number of instances where an editor’s intervention _— sometimes at the cost of going against the unanimous evidence of the manuscripts — is still required to restore the sense, syntax, or meter. The internal evidence of C and the other manuscripts disproves C’s claim to depend directly on Legge’s holograph. For although this manuscript pretends to be a direct copy of a holograph, it is no way visibly superior to the other manuscripts in the manner that one would expect. It often shares erroneous readings and other defects with some or even all of the other extant ones: there is a common substrate of error preserved by all of them. Such, for instance, is the case in Actio I, Act I at lines 74, 85, 87, 99, 124, and 132, and examples could be multiplied in like degree though each Act of all three Actiones. Additionally, there are very many points where one or a very few manuscripts contain a right reading, or in any event an acceptable one, whereas C sides with the large majority in presenting an obvious corruption. There are a number of other ways in which C is defective in precisely the same way as all the others, exhibiting defects that would scarcely be found in a manuscript uniquely representing the contents of the author’s holograph. This consideration serves to establish that all eleven extant manuscripts are descended from a common archetype wherein the defects they share were already present. That can scarcely have been Legge’s personal text. In view of the importance of this question, it is worth itemizing these features.
i. There are several points where one or more lines appear to be missing in the text in all mss., after lines 314, 523, 1403, 1771, 2721, and 2938. At other points (such as 1712) the textus receptus is garbled to the point of unintelligibility in C no less than in the others.
ii. In the list of speaking parts that stands at the beginning of each scene, or are used to mark entrances within scenes, names of speakers are occasionally omitted in the mss., for example Gloucester at II.iv.iii, Lord Stanley at I.vi.vi, and the Sergeant-at-Arms at I.v.vi. Likewise, a few names are missing from the list of dramatis personae prefacing each Actio: in the list for Actio I the Servant of Gloucester and the second citizen are not included.
iii. Per contra, on one occasion the name of a character is wrongly inserted in such a list of speaking parts, when that character does not appear in the ensuing scene (Lovell in I.iv.ii).
iv. On several occasions there are more or less strong grounds for thinking that scene-divisions are incorrectly indicated by the mss. Cf. the initial note on III.iv.iii, and the notes on 4086ff., 4163, and 4246.
v. The tenor of at least one speech suggests that the mss. may attribute it to the wrong speaker (see the note on 306). For another possibly misattributed speech see the note on 770.
vi. In the list of speakers for I.iii.ii all mss. but GJ have CHORUS PROCERUM TUMULTUANTIUM (although in the initial list of dramatis personae some mss. have CHORUS CIVIUM TUMULTUANTIUM, some have CHORUS PROCERUM TUMULTUANTIUM, and some have both), but the appeal to cives in line 579 makes it seem likelier that this is a mob of London citizens, as does Hasting’s address to the chara civium cohors beginning at 604. The mention of artificers in the stage direction in DH reflects a similar understanding. Confusion about the identify of these individuals originates in the fact that for reasons of dramatic economy Legge has combined two separate but closely juxtaposed incidents in Hall: 1.) “diverse lordes, knightes, and gentilmen” rush out in their armor in response to the Queen’s retreat into asylum (p. 350), and 2.) Richard exhibits arms stored in barrels to the common people of London, claiming that these armaments had been prepared for a Woodville coup (p. 351). The misidentification of this group as a crowd of nobles looks like the responsibility of the copyist of the archetype, either because he remembered Hall’s account or because he was under the misapprehension that the Archbishop’s speech immediately following is directed to this group. The late mss. GJ have the identification right, probably because somebody saw this error and corrected it.
vii. In I.iv.i Richard and the Council request a prelate to fetch the young Duke of York from asylum. Who is this prelate? In the lists of speaking parts prefacing I.iv.i and II.iv.ii the mss. identify him as the Archbishop of York. But there are several reasons for thinking that he is the Archbishop of Canterbury. 1.) In Legge’s source Hall, the prelate charged with this duty is Canterbury, so there is a priori reason for expecting him to be the prelate. 2.) As soon as one looks past the initial lists of speaking parts in these two scenes, and the words used to indicate changes of speakers in some (but not all) of the mss.,NOTE 31 he will see that in the words of the text this prelate is never identified as the Archbishop of York. 3.) This character is frequently addressed as a Cardinal. The Archbishop of Canterbury was a Cardinal, the Archbishop of York was not. These arguments, one must admit, are less than wholly conclusive. At 4051f. Legge seems to identify the Archbishop of York as a Cardinal (see the note ad loc.) and at several points (e. g., 1553, 1687) he wrongly calls the Bishop of Ely a Cardinal (Bishop John Morton was not promoted to that rank until 1492, by which time Henry VII had elevated him to the See of Canterbury and made him Chancellor of the realm). Then too, it might be thought that Legge introduced this change of characters himself because the prelate in question is wrongly identified as the Archbishop of York in the English version of the More biography (although the internal evidence of Richardus Tertius shows that he was familiar with that work in its Latin version).NOTE 32 But there is another argument for identifying the prelate as the Archbishop of Canterbury that, while circumstantial, is more compelling. This has to do with the dramatic situation. The Archbishop of York is presently in disgrace and has just been removed from the office of Chancellor. By first giving the Great Seal of England to Queen Elizabeth and then cravenly taking it back, he has revealed himself as weak, vacillating, and exceedingly lacking in judgment. Richard makes a motion in Council to entrust the delicate and vital mission of retrieving the young prince from sanctuary to the prelate in question because his fides makes him peculiarly suitable for this errand (694ff.), and the other members of the Council echo his sentiment. Fides, at the moment, is hardly the most salient characteristic of the Archbishop of York. In the same speech, furthermore, Richard urges (690ff.):

Ergo viri mittantur assensu sacro
Quorum dubia nunquam fides regi fuit,
Matri minus suspecta.

[“Therefore let men be sent by the sacred will of this Council, whose loyalty to the King is undoubted, nor suspect to the Queen.”]

These words describe the Archbishop of Canterbury. Although a relative of the House of York, Cardinal Bourchier had presided over the See of Canterbury during both Lancastrian and Yorkist reigns and had proven himself as essentially neutral and above the fray. Since, by reclaiming the Seal, York has proven himself as disloyal to the Queen as to the Council, he scarcely fits this specification. It is true that in the course of Richardus Tertius Legge commits several historical slips, as will be observed in appropriate commentary notes. But if he did substitute York for Canterbury in these scenes, this would not be another slip, it would be a serious historical and dramatic blunder. So the balance of probability supports the conclusion that this character is the Archbishop of Canterbury. If the manuscripts do misidentify him, there are two ways this could have occurred. In the archetype his name may have been omitted from the list of speakers at the beginning of these scenes — we have already seen that this occasionally happens — or else the copyist of the archetype happened to be familiar with the English version of the More biography, and so took it upon himself to “correct” Legge’s original.
viii. An important consequence follows from the discovery that this prelate is in all probability the Archbishop of Canterbury. In our mss. Actiones I and II are both preceded by a prose Argument (the presence of a Prologue eliminates the need for an Argument to Actio III). The Actio I Argument states that the prelate who accomplished this mission was the Archbishop of York. Therefore these Arguments must have been present in the archetype, and if the prelate in question is supposed to be Canterbury, then the proposition that Legge wrote them is thrown into doubt, unless the Actio I Argument had subsequently been doctored. (For another factual problem in the Actio I Argument, see the note ad loc.)
iv. By the same token, both this Argument and the list of dramatis personae for Actio I state that Edward V was fifteen at the time represented by our trilogy. In fact, at the time of his father’s death Edward was nearly thirteen and his brother was two years younger. Although Legge must have known the truth, as it is reported by Hall (p. 345), without stating any specific ages he manages to convey the impression that they are substantially younger, so as to increase the dramatic pathos and to make Richard appear an infanticide (cf., for example, 498). Since any exaggeration of Edward’s age is directly contrary to this program, here is another possible for reason for doubting that Legge wrote this Argument. (It is reasonable to suppose that the author of the Actio I Argument also wrote that for Actio II).
x. The stage direction at the end of Actio III found in some mss. indicates that, like its predecessors, it ended with a song. The text of the song or at least an indication of what was to be sung is not preserved in any extant ms. Likewise, it is probably also significant that C lacks the full set of stage-directions, most completely preserved in D (and also in E as supplemented by its appendix). Unless we are to think that Cade chose to omit many of these when making his copy, it is difficult to explain their absence on the theory that he was copying an author’s holograph. This latter theory becomes in fact untenable when it is appreciated that K, a NS. very closely related to C but not, as I shall argue, copied from it, lacks precisely the same directions.
30. These defects, together with the textual errors and omitted lines previously mentioned, strongly suggest that the common archetype of our mss. can scarcely have been Legge’s holograph. At best it can only have been an imperfect copy thereof, made by someone who in a number of ways failed to understand the author’s dramatic intentions, and who also omitted such material as the ending of Actio II and many stage-directions. Indeed, the number of textual errors as well as the more important mistakes just noted suggest that the copyist responsible for the archetype did not do a very good job. The premise that this archetype was a direct copy of Legge’s holograph is not susceptible to proof. None of our mss. can be less than two or even three removes from what Legge wrote, and perhaps the actual number is greater. It is worth bearing in mind that a trilogy with such a huge cast of characters and that enjoyed such popularity must have generated a very large number of performance texts, actors’ “sides,” as well as mss. meant for subsequent reading;NOTE 33 and if it is indeed true that the trilogy received one or more revival performances in the next decade, the number of mss. would have been further multiplied. (It should perhaps be pointed out that among English academic dramas Richardus Tertius is represented by an almost unprecedented number of surviving mss., matched only by those of George Ruggle’s enormously popular 1615 comedy Ignoramus. This means that the kind of difficulties that inevitably beset a manuscript text tradition must have proliferated beginning at a very early date.
31. The presence of the Cade colophon in C of course requires explanation. If we are to exclude the possibility of deliberate misrepresentation, there seems only one possibility, that Cade (a sizar at Gonville and Caius College) was mistaken when he thought that the MS. from which he made his copy was Legge’s own. As an undergraduate, he may well have been unfamiliar with the handwriting of the Master of his college. Possibly he asked Legge for a text to copy, was fobbed off with an inferior one, and wrongly assumed it was a holograph. In any event, the lesson to be learned is that C is not so authoritative that an editor is relieved of his normal critical responsibilities. he cannot indiscriminately cling to C through thick and thin. Quite to the contrary, he must operate just as if he were editing any other manuscript-descended text. At every point where the text looks corrupt, he must evaluate the evidence of all the mss. regarding syntax, meter, and sense. Whenever there are grounds for thinking the received text to have been disrupted in transmission, he must not hesitate to intervene and put matters right, even in the face of the unanimous evidence of the mss.

IV. THE MANUSCRIPTS

32. The first step towards a satisfactory edition must be a new collation of the sources. The eleven extant manuscripts of Richardus Tertius may now be listed (Lordi’s system of sigla is retained, save that mss. J and K were not included in his collation):NOTE 34

33. It is now necessary to ascertain the interrelationship of these sources. Since Richardus Tertius is a trilogy, mss. of the three Actiones may have circulated independently (as is suggested by G), and this in turn might mean that the interrelationship of our sources might not be the same for the three Actiones. In the event, this proves to be no mere theoretical possibility, since the actual state of affairs proves to differ somewhat in the three Actiones (the situation with the three plays of the Oresteia trilogy is similar). In the first instance, therefore the mss. interrelationship for Actio I will be described. The differences for Actiones II and III will then be explained.
34. ACK are very closely related. These three mss. share a number of idiosyncracies: they present the full title page, they attribute the speech beginning at 1747 to Gloucester, whereas in the other mss. the Citizen continues to speak, they omit the Show of the Procession and the song festum diem colamus at the end of Actio II, and, except towards the end of Actio III, they are devoid of stage directions. In addition, they present very many unique readings. Illustrative examples may be cited: 110 dies omitted, 301, admone iam clam, 402 solum, 475 caelum, 495 spiraris, 551 torquet, 662 levae, 739, quod, 745 perimi, etc. AK present many shared idiosyncracies not found in C. Therefore C cannot be copied from either of these mss. Readings unique to AK, or at least not shared with C, include 36 ademptum, 55 caput praestans, 111 cum omitted, 137, credit sinistris, 169 sum, 182 multum iure, 205 honos, 222 respiunt, 227 suaserit, 236 tum omitted, 240 dextra, 242 quos, 250 luent, 256 dextram, 266 mutare, 273 secreti, 293 lubem, 296 mutus, 336 nubunt, 350 recti, 400 Belleni, etc. The extreme similarity of these two mss. makes it intrinsically likely that one of them is a copy of the other. There are no instances in which K exhibits an error or defect not also found in A, but A exhibits defects not found in K (exceptions to this generalization are where K contains a corruption and A a different one derived from it, but at no point does A preserve a right reading, or one common to the other mss., were K has a corruption): 314 patrum omitted, 480 line omitted, 523 puscitur, 543 laudem forte (where it is particularly clear that the A copyist has been misled by the handwriting of K), NOTE 35 557 alti tonans, 564 cuiquam, 577 his, 596 claudunt, 601 dicet, 638 Argolicos, 645 sibi, 663 manet, 691 dubio, etc. It is therefore clear that A is copied directly from K: it is a codex descriptus having no independent evidentiary value save for a limited number of marginal corrections introduced by a second hand. K and C are obviously very similar. Lordi (p. xxvi), who did not collate K, argued that A is copied from C. It is important to dispel a similar possible idea about the relation of K to C, since this would imply that K, too, has no evidentiary value. If K were a copy of C, the situation would be just as that regarding A and C: it would contain errors and variants not shared with C, but not vice versa. But in fact there are points where C (in both its corrected and uncorrected state) contains variant readings not reproduced by K. Significant examples of such divergence include title page trivespera, Argumentum infami, 426 staret, 436 duri ne, 600 magnum corrected to quisquam, 759 proprium omitted, 1063 parvulo, 1065 suo, 1142 sanctum, 1170 dedi, 1253 fragile, 1396 quique, 1426 advotae, 1458 iussu meo, 1495 sedes, 1508 aperive, 1557 novit, and 1675 luent. These examples, while not copious, serve to cast doubt on the proposition that K is copied from C. It is much more likely that they are related collaterally: they are both copied from a lost common exemplar, and any peculiarities common to them both can be assumed to have characterized that ms.
35. A close relationship obtains between D and F. Very commonly, in regard to variant readings or other idiosyncrasies, they are joined by one or more other mss. But we may particularly observe the occasions on which DF present uniquely shared features: 17 tenera, 87 cingant filii, 97 infaelicis, 118 nudabat, 145 ista, 161f. lines omitted, 227 evaserat, 239 dirisque, 280 motum malum, 299 abnuat, 314 line omitted, 326 irrumpat, 392 Riverium patrum, 447 cura, 570 ego, 575 fidem, 605 est amplexus, 624 nostra, 625 funere, 627 suadere, 645 negant, 654 chare, etc. D is a later MS. than F, so F can scarcely be derived from D. But can D be a copy of F? Points where D contains errors not found in F, and vice versa, exclude this possibility. Therefore D and F must be related collaterally. The truth of this proposition can be demonstrated from Act i. Readings unique to D include 1 laetus, 8 gestiebam, 14 nepotum, 18 haec, 25 versit, 32 commitatuum, 41 incidet, 60 duram, etc. On the other hand, readings unique to F include 13 filii maior 21 impia regis, 25 verset, 31 hinc, 53 turbida, lugibria in the same line, 58 miseriat, 62 Galliam, etc. This is scarcely the state of affairs we should find if D were copied from F. The large number of agreements between DF supports the conclusion that these mss., despite their very different dates, are descended from a common ancestor. It will be noted that this ancestor bears no special relationship to the common ancestor of ACK, as DF lack the title page, contain most stage directions and the song at the end of Actio II, and since the variant readings that characterize DF are on the whole different from those that distinguish CK. CK and DF must therefore be regarded as subgroups belonging to two separate branches of the tradition.
36. At first sight it appears that GIJ represent yet a third branch of the tradition (Lordi, p. xviii, thought that this was true of GI, but he failed to collate J). For they form a group characterized by many substituted and interpolated lines or passages, many of which are shared by two or even all three. The relation of GJ is especially close: they contain variants and interpolations not found in I; the text has been modified so as to give Lovell a part in this Actio, and towards the end of the Actio a large number of lines, containing the entire speech of the Sergeant-at-Arms, have been deleted, either because of a decision to shorten the Actio, or more probably because both mss. are copied from a common exemplar, out of which a page was missing. Substituted and interpolated lines common to GIJ are found at 111, 114, 123, 213, 215, 233 (also present in B), 235, 237, 693, 860, 904, and 978. Further substitutions and interpolations are found in GJ but are absent from I: 110, 147, 272, 277, 482, 524, 580, 708ff., 821, 835ff., and 1533f. At only one point does I have a variant passage not shared by GJ (1792ff.), and at only one point does J have an interpolated line not found in G (1269a, which also appears in B). On the other hand, there are a few places where GI have variant lines not found in J: 13, 56, 300f., 308, and 324, which suggests that G stands in a somewhat closer relation to I than does J. It is not easy to determine the origin of these variants and interpolations. Clearly, they do not result from simple textual corruption, although in some cases they may represent attempts to correct corruptions in the standard text. Nor, usually, can they be regarded as improvements on the lines they replace (although one might feel that a few originals, such as 272, are superior). Since they almost uniformly say the same thing as does the standard text, but use different words (the variant at 2596 is the one substantial exception), it is difficult to believe that they are added for dramaturgic reasons. The fact that GJ have a somewhat different set than those found in I strongly suggests that they were not all manufactured at the same time, by the same person (for example, revisions introduced in connection with a revival performance). It seems likelier that they were created at different times for different reasons: to substitute for corrupted lines, to fill lacunae created by omitted ones, possibly as actors’ emendations, or even as Legge’s own revisions. This picture of a gradual invasion of the standard text receives support from the fact that a few of the variant lines of the GIJ group are also found in B, admittedly a MS. with a very eclectic background. B also contains an interpolation (1784a) not found in GIK or GJ, and in the margin of E is written a single line that substitutes for 787f.
37. These variations and interpolations are not the only features that distinguish the GIJ group. Some readings are unique to these three mss.: 104 sepulto…odio, 106 pacata, 130 tuos, 141 magna nunquam, 197 ut omitted, 211 placeret hoc regi nisi, 299 non, 232 cesserit, 285 quam dura semper fata res nostras premunt, 290 medicina, 675 fuit, 686 si fama, 730 nunc, 751 culpabitur, 792 aut omitted, 804 iusserit, 807 huic, 1542 plures, 1743 ac omitted, and machinatores in the same line. These shared features support the conclusion that GIJ are derived from a common ancestor. Several reasons have already been given for thinking that GJ enjoy an especially close relationship: besides those already mentioned, GJ share a number of unique readings not found in I: 110 bis caelo stellifero excidit sera dies, 174 quando, 534 omne istud, 551 metu, 552 coronent, 611 parata, 660 cardinali, 681 colludat, 703 protervia odium, 705 dei namque spirante spiritu, 727 muniet, 728 sic, etc. The similarities between GJ are sufficiently notable that it is probable that one of these MS. is copied from the other, or that they are copied from a common source. That the latter is the true state of affairs is shown by the fact that each of these mss. contains variants not found in the other. Although the first 96 lines of Act I are missing from J, plenty of illustrative examples can be found by examining the remainder of the Act. Variants found in G but not in J are 101 tremor, 112 at, 153 ventis; example of J variants not found in G are 98 falso, 107 monebit, 107 qui timet priores, 108 priora omitted, 109 quis, 109 detenet, 141 sperare nunquam magna virtus parum quisquis, and 150 excelso. Furthermore, each MS. accidentally omits lines that are present in the other: G omits 372ff., 581, 714f., 1235 and 1487, while J omits 217 and 1537. But, albeit GIJ constitute a separate group of mss. linked by many special features, it would be wrong to regard them as a distinct branch of the tradition. The GIJ group ought to be considered a special subdivision of the same branch represented by DF. For these two groups share many variant readings and other traits, such as more or less complete stage directions and the inclusion of the song at the end of Actio II. In addition, there is a very large number of times when variants are uniquely shared by one or more members of the DF group and one or members of the GIJ group. By my count, these may be enumerated as follows: DFGIJ 12, DFGI 12, DFG 1, JIG 3, DFI 22, DFJ 2, DGI 3, DGJ 1, FGI 1, FGJ 1, DG 3, DI 5, DJ 1, FG 1, FI 5 and FJ 9. The high frequency of such shared variants goes to show that DF and GIJ share a common ancestry. The relative statistics for the various possible permutations are instructive. The high number of unique DFI features suggests that I, characterized by fewer variants and interpolations than GJ, stands closer to the common ancestor of this branch than do GJ, and that the direction of change is towards these latter mss. It may also be noticed that D shares unique variants with GIJ eighteen times. The fact that no greater imbalance exists between these figures suggests that neither D nor F has drifted further away from their mutual ancestor than has the other.
38. B is also a member of this same branch of the tradition, occupying a place midway between the DF and GIJ groups. Again, combinatory statistics tell the story. B is in agreement with DFGIJ 15 times. Agreement with various mss. within this group are as follows: BDFGI 7, BDFGJ 2, BDFIJ 1, BDGIJ 1, BDGIJ 2, BDFG 2, BDFI 4, BDFJ 1, BDGI 3, BFGI 2, BFJI 1, BFJI 1, BGIJ 4, BDF 4, BDI 3, BDG 2, BDJ 2, BIJ 1, BD 1, BF 1, BG 6, and BI 11. These figures show that B has strong affinities with the GIJ group and most particularly with I. Another way of showing this special relationship is to count the number of agreements between B and each of these five mss.: D 49, F 44, G 54, I 63, J 27. Again, the total number of agreements between B and I is considerably higher than with any of the other four. We may best suppose that B is derived from a common ancestor of the GIJ group, and this theory accounts for its special affinity with I.
39. On 79 occasions E shares readings with one or mss. of the BDFGIJ family (on eight of these with all six) that are not found in the CK group, and on 27 occasions with one or more mss. of the CK family (on 20 with both). This strongly suggests that E belongs to the former family. Its frequent agreement with CK is to be explained by the supposition that it, or a MS. from which it has descended, has been collated against some MS. related to CK. E agrees with one or both members of the DF group sixteen times, and seventeen with one or more members of the BGIJ group. So it would appear that E represents a position in the history of the text where the two subgroups within this branch of the family had started to separate, but before variants and interpolations had begin to infiltrate the BGIJ branch.
40. The position of H in the tradition is ambiguous. On 51 occasions it contains CK readings (in 21 joined by E), and on 71 it has BDEFGIJ ones. The reader will be surprised to find E counted in both camps — the reason for this will be explained immediately below — and for the moment it suffices to note that there are 19 variants unique to EH. When these are subtracted, H shares 53 readings with one or more mss. of BDFGIJ. These facts are liable to one of two possible explanations. First, H could belong to the CK group but be contaminated with BDEFGIJ readings for jut the same reason that E presents a number of CK ones. When H shares variants with CK, it is joined by E on 21 occasions. According to this theory, CEHK readings and at least some EH readings can be understood by thinking that they represent lections that originated in the CHK branch of the tradition and were imported into E by the process of contamination described above. Or, per contra, one might suppose that H is essentially a member of the BDFGIJ group that has been contaminated by CK readings. Of the two possible explanations, the first appears preferable, for two reasons. It provides an efficient explanation of the frequency with which EH tend to side with one or mss. of the CK group. Then too, on only three occasions does H agree with all seven mss. of the BDEFGIJ group, whereas on 16 it agrees with CK. Readings unique to CH occur eleven times, and ones unique to CEH five times; there are no readings unique to HK or EHK. But it seems doubtful whether any weight ought to be placed on the absence of similar special links with K. C contains many unique variants not reproduced by H, so H cannot be derived directly from C. The relation of H to CK must therefore be collateral. By the same token, the fact that EH share a number of readings, as described above, scarcely warrants the conclusion that E or a direct ancestor of E was collated against H itself, since a number of CEK readings are not found in H; it is much more likely that the MS. that contributed these shared readings stood reasonably near to H in the CEHK side of the tradition. Another consideration tends to support the diagnosis that H belongs to the branch of the tradition represented by CK. This MS. lacks the formal title inscription otherwise characteristic of those mss., but it is a copy executed in 1583 and bound together with several other documents towards the end of the seventeenth century. It is quite possible that the title page was lost before or during the binding process. On the other hand, H concludes with an ending colophon that is virtually identical to that found in C — it is curious that both mss. have an ending colophon with a dating to 1583 written in Greek numerals — save that the name of the copyist is omitted. The probable explanation is that a similar colophon stood at the end of a MS. that was the mutual ancestor of CK and of H, but that it was omitted by the copyist of K.
41. By this point, the broad outline of the Actio I mss. tradition is visible. All of our extant mss. derive from a common archetype (identified as Ω here) which, as we have seen, already contained various kinds of error and so cannot have been Legge’s own holograph. Ω was copied twice by two others, here designated a and b. From a is descended the CHK branch, and b is the ancestor of BDEFGIJ. This latter branch split into subsections, DF, which bear the most resemblance to the mss. derived from a, and BEGHIJ, which began to branch off and became increasingly invaded by variants and interpolations. The history the text of Actio I can be represented diagrammatically by a stemma codicum, in which lost manuscripts that must be assumed are represented by lower-case letters. It should be understood that this stemma is conceptual and does not purport to be historically accurate: it postulates the minimum number of lost mss. from which the existing state of affairs can logically be derived. In all probability, this stemma fails to reveal the complexity of the prehistory standing behind our extant sources, but it serves to summarize the main trend of events and the interrelationship of our mss.

The MS. identified here as g is that which Anthony Cade thought was Legge’s holograph; the evidence fails to clarify wither it was in fact a itself (so that CK are better witness to a than is H, with the latter MS. being derived from a by a more indirect route), or whether CK, like H, are all descended from intermediate copies ultimately derived from a.
42. We may next consider the differences in the mss. tradition for Actio II. These are several. 1.) G is unavailable for Actiones II and III. 2.) E is no longer heavily contaminated by readings characteristic of the a branch of the tradition. Its derivation from b is now fully visible. 3.) B acquires an entirely different character. It now closely resembles I. Both mss. misidentify Dr. Shaa’s interlocutor in IV.i. BI contain a large number of shared variant and interpolated lines (as well as textual readings): 1924, 2002ff., 2026ff., 2124, 2204, 2214, 2229, 2304, 2316, 2359, 2417, 2570, and 2596. B contains at least a couple of significant such variants not found in I: 1925, 2031. 4.) But if B joins that section of the tradition characterized by variants and interpolations, J leaves it. For Actio II, J is wholly devoid of variants and interpolations, and shows no especial kinship with BI (save insofar as all three mss. belong to the b branch of the tradition). On the other hand, it shows strong affinities with DEFJ. Again, combinatory statistics reveal this: there are twelve readings unique to DEFJ, thirteen DFJ readings, eleven DH readings, three EJ readings, and thirteen FH readings. The mss. tradition for Actio II may also be represented diagrammatically:

t43. In Actio III J reverts to its original character and again displays a large number of variants and interpolations, and B remains in this special group. There are many such features, as well as special readings, held in common by BIJ. Again, if we are to be guided by the evidence of this kind of line, I stands considerably closer to J than it does to B insofar as IJ share a large number of them not found in B (2891, 2901. 2940f., 2946f., 2967, 3149, 3545, 3568, 3677, 37379, 3781, 4487a, 4490, and 4495). The number shared exclusively by BI is much less (3066ff., 3788), and the number shared exclusively by BJ is only slightly larger (2925a, 3542a, 3549, 3707, 3777f.). It therefore looks as if IJ are more closely related to each other than to B, and that of the two J is somewhat more closely related to B than is I. This too may be summarized by a diagram:

44. In the stemmata representing the MS. tradition for the three Actiones, lower case letters represent the minimum number of lost mss. that must be postulated in order to reconstruct the tradition. There is no reason to assume that the mss. represented by lower case numbers for each Actio are necessarily the same as those represented by lower case letters for the other two. In most instances, where the tradition remains stable throughout all three Actiones, doubtless they are. But in the case of that sector of the tradition characterized by variants and interpolations, this is very doubtfully so. The reader must be particularly warned away from making one possible assumption. G is available for Actio I only. The relation of GIJ in Actio I bears a striking resemblance to that of BIJ in Actio III, so that there is an initial temptation to think that G is a direct copy of an earlier MS. that became separated in such a way that G follows it for Actio I and B follows it for Actiones I and III. But a comparison of the stemmata for Actiones I and III quickly dispels any such assumption: if the theory were true, in Actio III B would stand in a closer relationship to J than to I. But we have seen that the reverse is true.
45. In examining the apparatus criticus the reader can constantly observe evidence that appears to contradict the conclusions made here: variants shared by mss. belonging to different families. Although clearly a member of the b branch, for example, in Actio I B contains characteristic CHK readings (sometimes shared with E) at 31, 91, 149, 222, 403, 421, 741, 770ff., 968, and 1639. These apparently contradictory facts are not disturbing. Such evidence is liable to two explanations, and so should not be used to cast doubt on the understanding outlined here. First, there is the mechanism of contamination: instances in which a MS. is collated and corrected against another, sometimes belonging to a different branch of the tradition. In the surviving mss., evidence of such contamination is visible. We have seen that, to different degrees, B, E, and H are contaminated because the copyist had more than one text in front of his eyes as he worked. Others have been more or less heavily corrected, presumably as the result of collation with other mss., by a second hand. The actual number of mss. standing behind our extant ones is doubtlessly greater than the minimum numbers postulated in the above stemmata codicum, and it is scarcely unlikely that similarly corrected and even more deeply contaminated ones stand in the background of the eleven we have, which serves to blur somewhat, but scarcely obscure, their positions in the tradition. It is also of course possible that different copyists may independently make the same error. This is particularly common regarding copyists’ misinterpretations of their exemplars, for in the handwritings of the mss. certain letters are constantly liable to confusion: f and long s, crotchet c and r, u and v, v and b. An undotted i creates confusion between such words as hunc and hinc, minus and nimis. A poorly formed a can look like a u, or vice versa, an e written too tightly can be taken for an i, a badly drawn o can look line an a. Ligatured ae and a are very frequently confused. The number of strokes used in combinations of i, m, and n can be difficult to interpret (crede experto). Letters can be obscured by neighboring upstrokes and downstrokes, and at word-ends flourishes are sometimes made with such verve that they can readily be mistaken for a vowel/tittle combination. Then too, problems are frequently created by misused or misinterpreted abbreviations. A barred p, for example, is supposed to be used for per, but some copyists also use it for prae or pro (so that, for example, perdo can become metamorphosed into prodo). A copyist can fail to see a tittle over a vowel indicating it should be followed by m or n, and he can just as easily imagine he sees one where none exists. Use of abbreviations to indicated inflectional endings can cause trouble, and abbreviated final que is often ignored. Certain words (e. g. ego and ergo, libens and lubens) are constantly confused. Letters corrected or improved by overwriting can be confused. Other copying mistakes (for instance, making a subjunctive out of an indicative or vice versa) are very common. Such garden-variety confusions account for a large number of variants, and are liable to occur independently. On the other hand, most if not all of these mss. were produced at the University of Cambridge, which was by definition a community of Latin scholars. At times, copyists spotted mistakes and their exemplars and either fixed them or at least introduced conjectural emendations; this too has the effect of blurring the affinities of the mss.
46. The prospective editor of Richardus Tertius must take a deep interest in the results of this codicological investigation. The most important lesson to be learned is that there is no reason in theory why any of our mss. (except the apograph A) cannot contain readings that stood in the archetype. Therefore an edition cannot be based on any single ms., or even on a few of them: it must be eclectic and consider the evidence of all ten independent witnesses. Nevertheless, when the editor is confronted by alternative readings of equal intrinsic value, he can select between them on grounds of probability. The stemma codicum given above clearly shows that as mss. become infiltrated by variant and interpolated lines, they progressively depart from the archetype. Those members of the b family that have not been affected by this process bear the closest resemblance to members of the a family, and therefore may be deemed to be superior witnesses to the archetype. In Actio I, E, B, I, and GJ progressively separate from DF, and this suggests a decreasing order or probability that an archetype reading will be uniquely retained. Similarly, in Actiones II and III those manuscripts characterized by variant and interpolated lines may be assumed to stand at the most distant remove from the archetype. Another important principle is that there is no reason for thinking that either a or b, insofar as they can be reconstructed from their descendants, was a better witness to the archetype than the other, and when some one or more a- descended mss. and some of the better b-descended ones share a reading, it is likely to have stood in the archetype. At many points, these considerations provide useful guidance for selection between alternative readings.

V. THE PRESENT EDITION

47. My object is to produce a sound and readable edition of the Latin text, which may then serve as the basis for an English translation and further study and enjoyment of the trilogy. The guiding assumption of the present edition is that Legge was, by the standards of his time, a competent Latinist. As will be explained in the next section, certain apparent solecisms are to be understood as reflecting the imperfect understanding of the rules of classical Latin syntax and versification characteristic of the author’s day, and I do not propose to do Legge the dubious favor of improving on what he wrote. Nevertheless, the mss. present a large number of instance where the received text is disrupted, sometimes to the point of incoherence, and where such disruption is clearly the result of the copying process (including a number of corruptions that must have been present in the archetype of the extant mss.). Such instances do require editorial emendation, and in this sense I do my work no differently than if I were editing a classical text. Nevertheless, in line with my second aim of producing a text that is both readable and searchable electronically, I shall impose modern punctuation, spell out abbreviations, introduce distinctions between i and j, u and v, eliminate the long s, separate ligatures, and remove diacritical markings. Since the mss. provide a wide range of orthographical variants, one cannot make any pretense to a scientific reconstruction of Legge’s own spelling of Latin words (the same is true of his punctuation), and shall be content to present a text using the orthography typical of his time and place. The mss. are no more uniform regarding the orthography of his stage directions, which he wrote in English, and when not otherwise stated I have reproduced that of D, merely because that manuscript preserves the fullest set.
48. Another form of editorial intervention is required. Most Acts are subdivided into two or more unnumbered scenes (except in G, but since this is a late manuscript and stands at the farthest remove from the archetype in the stemma codicum, there seems no reason to place great faith in G’s testimony). At the beginning of most onstage groupings that involve more than a single character, the mss. present a list of speaking parts (but not mutes and extras). It might therefore seem that such speaker-lists constitute a device for indicating new scenes, and indeed this was standard Renaissance practice:ΝΟΤΕ 36

Baldwin could have added that, just as a Renaissance “scene” did not necessarily involve any change of setting, so it did not necessarily involve any intervening passage of time.
49. Legge’s use of these speaker-lists is not entirely consistent. In many instances they identify the characters who appear in given “scene,” whether or not they are present at its inception. For example that for I.i includes the Messenger, even though he does not make his entry for more than a hundred lines. But on other occasions the list includes only those characters initially present, and a new list is subsequently inserted as a means of indicating the entry of a new character. In I.ii.ii, to cite one such instance, Gloucester’s Servant is not listed at the beginning of the scene, and a new list is inserted to indicate that when the nobles exit the King’s Servant remains behind and converses with Gloucester’s Servant, who enters at this point. The setting is the same and there is no break in dramatic time, so that it would be wrong to regard their dialogue as a separate scene in the modern sense of the word. Had Legge numbered his scenes, I would probably have felt obliged to follow his system, and this would have created a great deal of confusion for the reader since Richardus Tertius contains a great number of shifts of setting, some following upon each other very rapidly. Fortunately he did not, and this allows me to impose the modern concept of the scene without going against his explicit wishes. On the level of dramatic organization, this editorial policy, which is admittedly rather unhistorical, corresponds to the application of modern systematic punctuation to the words of the text. In both cases, the gain in intelligibility vastly outweighs the loss of scholarly accuracy created by the introduction of these anachronisms. In the particular case of this trilogy, this decision yields a handsome reward: it renders visible the resemblance of the way Richard learns of the defeat of the Buckingham revolt to the similar sequence of events in Richard III, discussed in an earlier context as important evidence that Shakespeare used Richardus Tertius as a source. So the policy of this edition is to keep the number of scenes to a minimum. They are allowed to flow on uninterrupted as characters enter and exit, so long as the setting does not shift and in the absence of any obvious passage of intervening time. Usually this rule can be applied without difficulty, and the occasional problems are acknowledged and discussed in individual commentary notes.

VI. LEGGE’S LATIN

50. Justus Lipsius’ appraisal of Legge’s abilities as a Latinist is worth quoting. In a 1585 letter to Legge,NOTE 37 he wrote In antiquitatis studio tam egregie versatus es, ut id te teipse potest quod de se Apollo Enni, a me omnes Cantabrigienses consilium expetunt in literis incerti, quos ego, meo ope, incertis certos, compotesque consilii dimitto [“Your classical learning is so outstanding that you can say the same as Ennius’ Apollo, that ‘all those at Cambridge doubtful about literary matters seek my advice; they are unsure when they come to me, but by my assistance they go away sure and certain of mind.’”] This assessment might appear to be in severe conflict with the received text of Richardus Tertius as represented by any single ms. and sometimes by all of them combined. But such discrepancies are liable to two types of explanation: errors introduced into the manuscript tradition as a result of the normal process of corruption, and ways in which the principles of writing Latin were, in Legge’s days, imperfect by classical standards. I shall briefly review these.

A. Imitation of Seneca

51. Throughout his discussion of Richardus Tertius, Churchill makes much of his borrowings from Seneca. Undeniably, Legge not only uses the characteristic vocabulary, rhetoric, and imagery of Seneca’s tragedies; he also borrows a number of phrases or even complete lines (sometimes with minor modifications) from these plays. But Churchill’s relentless pursuit of this theme leaves a very wrong impression that Legge has achieved little more than compiling a cento of Senecan borrowings. Then too, more careful analysis shows that he borrows phraseology from other standard classical poets. But in his study of secondary education in Elizabethan England T. W. BaldwinNOTE 38 has studied the way Latin verse composition was taught in the sixteenth century. The student was supposed to quarry phrases out of the texts he read (by such canonical authors read in schools as Lucretius, Vergil, Horace, Juvenal, Persius, Lucan, and above all those by Ovid), supplemented by ones garnered from phrase-books such as the Flores Poetarum, Plegromius’ Sylva Synonymorum, and Textor’s Epitheta, and liberally lard his own efforts with these. Analysis of virtually any Anglo-Latin poet or playwright of the Renaissance reveals a use of this method. Naturally, the author in question will tend to quarry phraseology from Latin poets who wrote in the same genre in which he is working, so that writers of comedy took from Plaits and Terence, writers of tragedy from Seneca, and so forth. And a good playwright such as Legge could put this standard method of composition to artistic purpose, working his literary borrowings into the fabric of his verse with clever originality (and at least one case, genuine brilliance — see the note on 1211), doubtless intended to be appreciated as such by the cognoscenti among his audience.

B. Syntax
i. Moods of verbs

52. The most striking syntactical habit found in Richardus Tertius is a predilection for using the imperfect subjunctive, for an obvious reason: this verbal form produces many word-shapes useful for writing iambic senarii. They are especially handy at the end of the line. In order to use them freely, Legge is willing to stretch the rules (to whatever extent he was aware that rules existed). Most notably, he repeatedly uses an imperfect subjective in final clauses after a primary verb. While unusual, this construction is allowable by the rules of good classical Latin (Allen and Greenough, New Latin Grammar §485.i); evidently the justifying idea is that of an action begun in the past and continuing in the present, where the purpose of the action is considered from the viewpoint of the time it started. In actuality, however, in some instances the application of this rationale is considerably more obvious than in others, so that it would be fairer to say that the imperfect subjunctive can be freely substituted for the present subjunctive in clauses where purpose is stated or implied. At other times, the normal rules for the selection of tenses of the subjunctive are ignored with no possible justification (again, most commonly by the use of irrational imperfect subjunctives). More accurately, Legge was unaware of these rules, which were not fully enunciated until the nineteenth century. He almost certainly learned his Latin out of the standard Tudor manual, William Lilly’s A Short Introduction of Grammar Generally to be Used,NOTE 39 which does not mention them. Nor does Priscian’s Institutiones Grammaticae, the grammar read in the Universities. For this reason, even when the meter admits conversion of an apparent solecism to a desirable tense, an editor should leave the received text undisturbed as long as the mss. evidence is unambiguous (but when the good mss. are divided, the grammatically correct form is selected).
53. As Allen and Greenough also explain (§575c), the use of indicative verbs in indirect questions is permissible in poetry. Legge freely avails himself of this liberty. Hence even when a metrically equivalent subjunctive is available, it seems best to leave such indicatives undisturbed (but when the mss. are divided between an indicative and a subjunctive, the subjunctive is selected — so too in other cases were irregularities are at stake).
54. The author also uses imperfect or even present subjunctives with dum in purely temporal clauses, or even mixes indicatives and subjunctives when such clauses are strung together (e. g. at 1896ff. and 2132ff.). In this context it might be worth noting that, while uncommon, this usage has classical precedent (Allen and Greenough §556a.). On the other hand, another syntactical peculiarity found in the mss. evidently has no precedent in good classical authors: temporal clauses consisting of postquam and the imperfect subjunctive (in his later Solymitana Clades Legge extends this usage to ubi). In each such case the meter admits alteration to a perfect or pluperfect indicative, but this usage is encountered in Latin of the later Roman period (Thesaurus Linguae Latinae article on postquam, col. 251) and is frequent in Medieval and Renaissance Latin, and so emendations are not made.
55. Legge had a penchant for using the conjunction licet (“although”) in concessive clauses. In the works of good authors of the classical period, some tense of the subjunctive is required and indicatives are not used in this construction (Thesaurus Linguae Latinae article on licet cols. 1365f.). The mss. present a number of instances where this rule is not followed (as well as plenty where it is). It would be an easy enough matter to rewrite all of these examples to the appropriate present or perfect subjunctive since in each of these cases the meter admits such emendation, but the substitution of some other tense or mood of the verb occurs often enough that one wonders whether Legge failed to observe the rule. In this case we may apply the same test as with postquam: are examples of concessive licet with indicative verbs also found in Solymitana Clades? The answer is yes, and some of those examples cannot be changed to subjunctives because of the meter. Hence all examples of concessive licet with the indicative should be left undisturbed.

ii. Other Syntactical Peculiarities

56. Legge uses infinitives with an imperative force, a device not found in classical Latin that is presumably borrowed from Greek. Occasionally we encounter unjustified use of reflexive pronouns, a familiar characteristic of postclassical Latin (Lilly’s Latin textbook failed to cover reflexive pronouns, so many Tudor Latinists reared on it misuse them).

iii. A General Remark

57. In view of what I have written about the imperfect knowledge of the rules of Latin syntax in the sixteenth century, the reader may be surprised that in commentary notes I furnish a number of syntactical explanations taken from a modern manual, Allen and Greenough’s New Latin Grammar. At the risk of seeming anachronistic (and absent any such thing as a comprehensive modern handbook of Renaissance Latin grammar), I supply these to assist the reader’s understanding of Legge’s sometimes difficult Latin. There is of course no implication intended that Legge would have been capable of articulating these rules and principles: some he may at best have sensed intuitively on the basis of his no doubt good knowledge of the classics, and others may have been mechanically copied from his readings in classical Latin.

C. Vocabulary

58. Legge uses vocabulary appropriate to Latin high poetry in general, and to Senecan tragedy in particular. There are only rare cases in which this generalization does not hold true. In his vocabulary iussus is a fourth-declension noun that can be used in any case, not just the ablative singular. At 1033 he uses the notoriously colloquial caballus for “horse.” At 1642 the mss. present the nonexistent word satellitium, but the defective meter of this line strongly indicates that the text is corrupt at this point. There is a special category of words where he could not be guided by classical precedent. He (like all Englishmen writing about their own nation in Latin) was perforce obliged to press a number of classical words into service with special meanings in order to render the names of various English social and political institutions, titles, places, etc. Examples are praetor Londonensis (the Lord Mayor of London), senatus (usually the Privy Council and never — despite Lordi’s persistent mistranslation — Parliament), arx (the Tower of London), heros (Lord). Phrases as well as words can be used in the same way: procerum cohors, for instance, designates the Peerage. There is no need to compile a mini-lexicon of such words and phrases, since this is effectively accomplished in the English translation.

D. Meter

59. With the exception of the songs at the ends of Actiones I and II, Richardus Tertius is written entirely in iambic senarii. At many points, the received texts (even as preserved in all ten sources) display such metrical problems as impermissible resolutions in even-numbered feet, cretic feet, and even worse offenses. When the mss. are divided, it is easy to prefer MS. readings that scan over ones that do not. When readings presented by all mss. fail to scan, we may assume that the archetype was liable to the same kinds of problem constantly encountered in the extant mss.: these include omission or addition of words or syllables, transpositions of words, or more thoroughgoing reshuffling of word-order within the line, and not infrequently the failure of a line to scan properly serves as a signal to the editor that some other problem (such as a syntactical one) requires his attention. But it is of course impossible to emend Legge’s lines to procure proper scansion without ascertaining how he understood the rules of iambic versification. Certainly the Elizabethans were well aware of the basic principles of Latin iambics, as can be seen from such metrical handbooks as George Buchanan’s De Prosodia Libellus (a reprint of the 1590 edition was published at Menston, England, in 1970). In the earlier sixteenth century, certainty about the use of anapests and spondees in even-numbered feet did not exist, but in his discussion of the subject,NOTE 40 John Hazel Smith argued that awareness of such rules grew as the newN Learning advanced, especially after Terentianus’ treatise on metrics became available at Cambridge in the 1530’s (as attested by John Cheke).
60. On the page of Anthony Cade’s MS. (C) that contains the first lines of Actio I, Act I, there is a more or less accurate table of the permissible resolutions in an iambic senarius and an improvement over Lilly (although Cade thought the sixth foot could contain a tribrachic resolution). In the line of his table devoted to spondees, these resolutions were included for the second and fourth feet, but then (correctly) crossed out. It is a sign of the kind of advance in comprehension of the rules described by John Hazel Smith that Cade, who had learned his Latin recently, knew they were impermissible. But Legge, who had received his education a generation earlier, did not. In all probability he thought that spondaic even-numbered feet were allowed because such is stated in section of William Lilly’s grammar manual devoted to prosody.
60. In the light of his use of spondaic second and fourth feet, the reader might wonder whether Legge also employed anapests in these positions. To be sure, anapestic even-numbered feet are found in earlier Tudor iambics (such as those of Watson’s Absalom of ca. 1540), but the greater part of Smith’s metrical discussion centers on anapests and demonstrates that well before Legge’s time they were acknowledged to be impermissible. In Richardus Tertius there are 24 such instances in which the received text, or at least that of the majority of mss., contains anapestic second or fourth feet (I mean in addition anapests employed in a special circumstance described below). If Legge had thought such anapests permissible, it is difficult to see why he would have employed them so sparingly, quite unlike the situation with second- and fourth-foot spondees. Some of these 24 instances occur in lines that look corrupt for other reasons, and in almost every case the anapest can easily be removed by such simple remedies as word-transposition.
61. Two special features of Legge’s metrical technique remain to be discussed. The first is that, as D. S. Raven put it,NOTE 41 “Final o is commonly shortened in many ‘silver age’ poets, e. g. Juvenal, except in dative and ablative cases.” Legge has observed this feature and not infrequently imitates it. The second feature is more surprising and will no doubt cause controversy. Raven (p. 26) describes another Latin metrical feature: “Literally the term brevis brevians implies a short syllable shortening a succeeding long, and the effect of the law is that a long syllable is shortened if it is both (a) directly preceded by a short syllable and (b) directly preceded and/or followed by a syllable with word-accent.” It is a remarkable fact that both Richardus Tertius and Solymitana Clades contain a number of lines which only scan properly if this rule of brevis brevians is invoked, and which otherwise do not appear to be corrupt. Take, for example, R. T. 498, irate pater, innocens quid admisit puer? Here brevis brevians converts an illicit second foot anapest into a satisfactory tribrach. Or again, to choose another example at random, at 1526, fraudes abominatur, atrox quassans caput, brevis brevians has the same effect on the fourth foot. The objection that brevis brevians is more characteristic of early Latin comic writers such as Plautus and Terence, but not of Seneca, is not disturbing. This wouild attribute to Legge an entirely anachronistic awareness of the historical development of Latin poetry. By the time he wrote Solymitana Clades he had developed a taste for Ennian alliteration, a stylistic device of the earliest phases of Latin poetry, and he cheerfully grafts these onto a fabric of generally Senecan poetics, creating an effect that strikes a modern, history-conscious reader as farouche. A second objection is considerably more serious: is it not highly implausible that anybody in the sixteenth century — particularly someone who was under the impression that spondaic even-numbered feet were permissible — would have grasped the principle of brevis brevians, far in advance of Bentley’s work on Terence in the eighteenth? I must admit I am incompetent to provide a satisfactory answer to this question. Either Legge read Plautus and Terence and noted empirically that seemingly false quantities were permissible in certain situations (if so, I have no idea how he would have rationalized his observation), or perhaps what I have written above about anapests is wrong. Legge, after all, was a considerably older man than other academic playwrights of the late Elizabethan period, and may be preserving the understanding of metrics prevalent when he was a schoolboy, so that his versification is “old fashioned” in comparison with that of his literary contemporaries. In any event, the undeniable fact is that both Legge’s trilogies contain a considerable number of lines which scan properly only if brevis brevians is applied. Even though I am unsure of how Legge justified them to himself, there is no visible reason for doubting that they are what he wrote, and so I have left them undisturbed.
62. Another kind of question may seem to be raised by some emendations which are based on the assumption that Legge’s understanding of the proper quantification of individual Latin words was correct. For example, the received text of 593 is ambitio thronum et poscit in praedam sibi. This can only stand if it is assumed that Legge wrongly scanned the first syllable of thronum long. But he uses this word often, otherwise always with a properly short syllable. The occasions are rare on which he demonstrably erred in his quantifications, using a wrong one twice or more, or scanning the same word differently in different lines. In this matter, therefore, “innocent until proven guilty” would appear to be the soundest editorial policy. Instances of erroneous quantification — most memorably a tendency to give mereo(r) a long first syllable, especially in the perfect tense — are indicated in notes.

VII. ACKNOWLEDGEMENTS

63. It is never possible to complete a large project without incurring large obligations, both to individuals and to institutions. First, the authorities of the various libraries in the United Kingdom and the United States that possess manuscripts of Richard Tertius not only answered requests for photographic copies, but also a number of them exerted themselves to furnish background information about the items in question. I hope that they feel repaid by learning a bit more about the manuscripts in their care. I am grateful to Peter Lang Verlag, the copyright holder of the print edition, for transferring publication rights to me, facilitating this present electronic edition. Professor C. N. L. Brooke, the modern historian of Gonville and Caius College, provided me with interesting information about Legge, and encouraged me to persevere in my work. Above all, I am grateful to my wife, Dr. Kathryn A. Sinkovich, both for discussing innumerable questions about Legge’s Latin with me and for tolerating me as I worked on this project.

POSTSCRIPT (NOVEMBER, 2022)

Looking back on my presentation of Richardus Tertius, I remain generally satisfied (with a little lingering unease about what I wrote about brevis brevians, but something had to be said on the subject and no better alternative has come to mind). But all along I have been conscious of one important weakness. With its trilogic form and ending on a strong upbeat note of a reconciliation of conflicting opposites, this work obviously takes its inspiration from Aeschylus's Oresteia. But how in the world could Legge have read that great work? It is true that Euripides was beginning to make an impression on early modern English dramatists, NOTE 42 but at least the dialogue portions of his plays make for fairly easy reading and can be comprehended after a few semesters of Greek, whereas Aeschylus in general and the Oresteia in particular make for much tougher going and only a remarkable undergraduate is capable of dealing with that trilogy. One would therefore naturally expect that the Attic tragedians would begin to attract the attention of English readers in the order of their difficulty. NOTE 43 I therefore wrote off claims that echoes of the Oresteia can be detected in Hamlet as merely silly. NOTE 44
I am therefore deeply grateful to Professor Tanya Pollard for kindly setting me strait. She taught me that it is most likely that Legge had somehow laid hands the a copy of the set of Latin translations by Johannes Sanrovius, Aeschyli Poetae Vetutissimi Tragoediae Sex, printed at Basel in 1555 (Ben Jonson owned a copy of his versions as reprinted in Pierre de la Rovere's Poetae Graeci Veteres, printed at Paris in 1614). If Shakespeare on any other writer of the period indeed had any familiarity with Aeschylus, in all probability this would have been thanks this the same source.

Notes

NOTE 1 Thomas Fuller, The Worthies of England (edited by P. Austin Nutall, London, 1840, repr. New York, 1965) II.491. The principal source for this biography is the obituary in William Moore’s continuation of the Gonville and Caius College Annals (written in 1655-6), entry for the year 1607, edited by John Venn (London, 1904) 213f., partially reedited by Alan H. Nelson, Records of Early English Drama: Cambridge (Toronto, 1989) I.282f. Definitely based on Fuller is the entry mentioning Legge in Anthony à Wood’s Athenae Fasti (1692) for the year 1586. Cf. also John Venn, Caius College (University of Cambridge College Histories, London, 1901) 76 - 89 and Biographical History of Gonville and Caius College 1349 - 1897 (Cambridge, 1897) III.64 - 69 and C. N. L. Brooke, A History of Gonville and Caius College (Woodville, Suffolk-Dover, New Hampshire, 1985), especially 84 - 88. Brooke, Plate 9c, provides a photograph of the effigy on Legge’s tomb in the College chapel. An interesting specialized study is P. G. Stein, “Thomas Legge: A Sixteenth Century English Civilian and his Books,” Satura Roberto Feensta (ed. J. A. Ankum et al., Fribourg, 1985) 545-56. Various anecdotes about Legge coping with the religious contentions afflicting Cambridge during his life are given in the course of H. C. Porter, Reformation and Reaction in Tudor Cambridge (Cambridge, 1958), cf. index s. v. I have not had the opportunity to read the unpublished monograph about Legge by Lord McNair in the Gonville-Caius College Library.

NOTE 2 Cf. G. B. Churchill and Wolfgang Keller, “Die lateinischen Universitäts-Dramen Englands in der Zeit der Königen Elizabeth,” Jarhrbuch der Deutschen Shakespeare-Gesellshaft 35 (1898) 231 - 323, Frederick C. Boas, University Drama in the Tudor Age (Oxford, 1914, repr. New York, 1966), and C. G. Moore Smith, College Plays Performed in the University of Cambridge (Cambridge, 1923); also Smith’s “The Academic Drama at Cambridge: Extracts from the College Records,” in Collections (The Malone Society, 1923) II.2.159 - 166.

NOTE 3 Cf. Robert J. Lordi, Thomas Legge’s Richardus Tertius: A Critical Edition with a Translation (New York-London, 1979) iif.

NOTE 4 Again, this is based on William Moore’s continuation of the college Annals.

5 The title page found in some mss. states that Richardus was performed over three nights of the comitiis bacchalaureorum anno Domini 1579. This is verified by archival evidence. The list of St. John’s College rentals expenses for October quarter of 1578/9 list a total expenditure of 5 s. for paper “to write out the bookes for the tragedy,” and the similar records for January quarter of the same year record an expenditure of 5 s. for a dinner for Dr. Legge and for various repairs to the College hall. Since these expenditures are immediately juxtaposed, it looks as if these preparations (new whitewashing, nets to darken the windows, etc.) may be preparations for the production of Richardus. Likewise, College records for July quarter of the same year record the expenditure of 35£ “layde out about the playes of diverse uses sicut patet per billam.” This is an extremely large sum, and may pertain to the costs both of Richardus and of the comedy Hymenaeus, which may have been produced at St. John’s in the same year (we have a list of the actors who appeared in that play, and many of the same individuals acted both in Richardus and Hymenaeus, although this list equally well supports the idea that Hymenaeus was produced the preceding year, as argued by Boas, p. 393). In any event, it is easy to imagine that Richardus required the lion’s share of the money, and this is further testimony to the enormous scale on which this stage spectacular was conceived and produced. Thus these records seem to provide circumstantial evidence that Richardus was produced in 1579. So they are interpreted by Lordi, vii. n. 11. (These records are reproduced by Nelson I.284f.)

NOTE 6 G. B. Churchill, Richard III before Shakespeare (Berlin, 1900, repr. Gloucestershire - Totowa N. J., 1976) remains the seminal discussion of this work; virtually all subsequent discussions, such as that of Boas 109 - 32, are derived from his treatment. The quote is from pp. 269 - 72.

NOTE 7 Nashe afterwards recalled how an actor fluffed a line in performing the play (cf. the note on 579ff.), which seems to suggest that he saw a performance, but this anecdote could be based on hearsay.

8 NOTE 8 Bosworth Field (New York, 1966) 266f.

9 NOTE 9 Both these phenomena are illustrated at 1381ff.

NOTE 10 His most recent biographer, Desmond Seward, Richard III: England’s Black Legend (New York, 1984) finds the historic Richard guilty as charged (my own view, for what it is worth, is that he was indeed guilty but acted out of a conviction that the Woodville-Neville combination was intent on his destruction and reckoned that he needed to get the knife in first). For Richard as the quintessential Renaissance tyrant cf. pp. 15 - 21.

NOTE 11 This is not the least interesting thing about Legge. In the course of this edition I note some ways in which Richardus Tertius reflects Aeschylus, and probably this is a subject that would repay further study (one could pursue, for example, the way in which Legge seems to use such leitmotifs as disease and madness in a way somewhat reminiscent of the Oresteia.) In the same way, in I.I.i of Solymitana Clades a single character remains onstage throughout and confronts a series of interlocutors, and this looks like a borrowing of the dramatic technique of the Prometheus Bound. See further the Postscript at the end of this Introduction).

NOTE 12 (This note has been withdrawn).

NOTE 13 A notorious problem in Senecan criticism: see W. Mark, Funktion und Form der Chorlieder in den Seneca-Tragödien (diss. Hamburg, 1932), especially pp. 6ff.

NOTE 14 If challenged, he could have claimed that he was following classical precedent insofar as there existed a Roman dramatic form called fabula togata that dramatized episodes of national history, but since the only extant specimen of that form is the pseudo-Senecan Octavia, he was not unduly bound to imitate such works in detail.

NOTE 15 By Churchill, p. 272, a verdict echoed by Lordi, p. x.

NOTE 16 Cf. David Greenwood, “The Staging of Neo-Latin Plays in Sixteenth Century England,” Research Opportunities in Renaissance Drama 12 (1969) 33 - 42.

NOTE 17 Thomas Heywood in his Apology for Actors (1612, sig. F4^v) borrowed this assessment from Harington.

NOTE 18 At least partially so: see the description of the manuscript below. The name is sometimes erroneously given as Lacey in the secondary literature.

NOTE 19 A Chronicle History of the London Stage, 1559 - 1642 (London, 1890) 79.

NOTE 20 Lovell’s name is wrongly added in the initial list of Actio I dramatis personae in all mss. that contain the list, and is also wrongly added in the list of speakers for Act IV, scene ii. It is not impossible that somebody wrote in a part for him merely on the basis of these notices (compare how the copyist of E started to attribute a speech to him at the end of I.iv.ii). The fact that DG supply the name of the actor playing Lovell, as well as that of the character in the Actio I dramatis personae, does not constitute evidence that Lovell’s inclusion in this Actio is right. Despite this initial list, after all, D does not include him as a character in the play itself. It is difficult to decide whether the omission of a long passage of I.v.vi in G and J, which has the effect of eliminating the part of the Sergeant-at-Arms, represents a deliberate “cut” in the text or merely results from the loss of a page from a common ancestor of these two closely related manuscripts.

NOTE 21 Throughout this edition Hall is cited from the 1809 edition (repr. New York, 1965).

NOTE 22 Churchill (p. 375) raises the interesting possibility that Senecan portraits of tyrants exerted influence on More’s characterization of Richard.

NOTE 23 Most notably Geoffrey Bullough, Narrative and Dramatic Sources of Shakespeare (London - New York, 1960) III.234 - 37. Cf. also Robert J. Lordi, “The Relationshiop of Richardus Tertius to the Main Richard III Plays,” Boston University Studies in English 5 (1961) 139 - 53.

NOTE 24 Churchill (pp. 494 - 79) found evidence that Legge was known and used as a source by the author of The True Tragedy. At pp. 474 - 524 he discusses Shakespeare’s use of The True Tragedy as a source. Therefore this anonymous play was an indirect route by which Legge assuredly influenced Shakespeare. In the present discussion, I focus on ways in which Legge may have influenced that cannot be explained by reference to The True Tragedy.

NOTE 25 For this play cf. also W. W. Greg, A Bibliography of the English Printed Drama to the Restoration (London, 1962) IV.1744.

NOTE 26 Cf. the anonymous editor of the Malone Society text of The True Tragedy, printed for the Society at Oxford, 1929 (under the general editorship of W. W. Greg), p. vi.

NOTE 27 By T. W. Baldwin, William Shakespere’s Small Latine & Lesse Greeke (Urbana, 1944).

NOTE 28 Appendix to his edition of The True Tragedy of Richard III (London, 1844).

NOTE 29 In his enlarged edition of Collier’s Shakespeare’s Library II.i (London, 1875).

NOTE 30 F. G. Fleay, A Biographical Chronicle of English Drama, 1559 - 1642 (London, 1891) II.36.

NOTE 31 In some mss. the designation of this individual in initial lists of speaking characters and the marginal identifications used to show changes of speakers merely identify him as “the Cardinal.” This, perhaps, suggests that the memory that this individual was in fact Canterbury was never quite expunged. In mss. I and J the Archbishop of York is identified as a Cardinal in each scene in which he appears, so these mss. have no evidentiary value at this point.

NOTE 32 Lordi pointed out (p. 376 n. 92) that II.iv.ii dramatizes an episode found only in the Latin version of the More biography.

NOTE 33 Nelson (II.887) differentiates formal copies, meant for binding, from fascicles “of an informal character, perhaps not intended for binding originally, and roughly contemporary with the play production…a fascicle seems more likely to represent the original state of the text, and may even have served as a performance text.” It is important to note that, according to this definition, all eleven of our manuscripts are formal copies. Obviously, the question of the relation of the extant mss. to the large number of performance texts and “sides” that surely needed to be prepared is entirely moot.

NOTE 34 For another description of the mss. see Nelson II.918f. On the flyleaf of C is a pencilled note in (I think) a nineteenth century hand stating that a ms. is owned by St. John’s College. I am assured by the Librarian of that College that no such copy exists.

NOTE 35 K was written in a very rapid, scrawling hand, and in comparing the microfilms one can often see how the copyist of A was similarly misled. A particularly clear example is found at 806, where K has the correct reading condidit furta templo. Condidit is not very legibly written, and so a second hand has rewritten the word in the margin to clarify the reading. A’s copyist duly writes condidit furta templo condidit. Again, at 1279ff. the copyist of K has not made a good job of aligning the names used to indicate changes of speakers, and so these are confused in A.

NOTE 36 T. W. Baldwin, On Act and Scene Divisions in the Shakspere First Folio (Carbondale, Ill., 1965) 15; the first two chapters of that study are relevant and instructive. See also Mark Damen's provocative article "French Scenes in Greek Tragedy: the Scenic Structure of Classical Drama," Theatre Journal 55:1 (2003) 113 - 134.

NOTE 37 Preserved in the Fuller biography and also in the Gonville and Caius College Annals (ed. John Venn, London, 1904) 214. I have published this letter at “Justus Lipsius to Thomas Legge, January 1, 1585,” Humanistica Lovaniensia 40, 275 - 81.

NOTE 38 Baldwin Chapter XLI (II.380 - 416).

NOTE 39 Reprinted with an Introduction by Vincent J. Flynn (New York, 1945).

NOTE 40 A Humanist’s “Trew Imitation”: Thomas Watson’s Absalom, A Critical Edition and Translation (Urbana, 1964) 70 - 80.

NOTE 41 D. S. Raven, Latin Metre: An Introduction (London, 1965) p. 23.

NOTE 42 See Tanya Pollard, “Greek Playbooks and Dramatic Forms in Early Modern England,” in Forms of Early Modern Writing (edd. Allison K Deutermann and András Kisẻry, Formal Matters: Reading the Materials of English Renaissance Literature (Manchester 2013) 99 - 123. See also Aaron J. Kachuchk, “Exit Pursued by Bears: Shakespeare and the Classical Tradition,” Classical Reception Journal 13 (2021) 86 - 106.

NOTE 43 Thomas Watson's 1581 Latin version of Sophocles' Antigone at best comes only partially into consideration. As remarked in the Introduction to my Philological Museum edition, it has been pointed out that this is in large part a doctored version of Thomas Naogeorgius' Latin translation, published at Basle in 1558.

NOTE 44 Louse Schleiner, “Latinized Greek Drama in Shakespeare's Writing of Hamlet,” Shakespeare Quarterly 41 (1990) 29 - 48.