thought of the other plays of Shakespeare that I would class together as twin productions of his genius, I hold that nothing can be more certain than that the two plays of Friendship which contain the stories of faithful Antonio and faithless Proteus were meant as pendants to each other; and that the two plays of youthful pission, with the stories of "True and Faithful Juliet" and "False and Faithless Cressid,” were also meant as pendants. Try the experiment; prepare your ear by reading straight off a couple of Acts of Romeo and Juliet, and then read this Troylus story without a word of the other parts of the play, as far as Act iv. Sc. 4, and I confidently await the verdict.

But we must pass on to the second story, that of Hector. This is contained in the following parts :

Act. Scene. Line.

I. I, 108-119.

I. 3, 213-309.

II. 2, all.

III. 1, 161-172.
IV. 4, 142-150.
IV. 5, 1.11.

IV. 5, 54-276.

*V. I, all.

V. 3, I-97.

(Æneas and Troylus.)
(Æneas and Agamemnon.)

(Priam, Hector, Troylus, Paris, Helenus,
Cassandra.)

(Paris, Helen.)

(Paris, Æneas.)

(Grecian Generals.)

(Hector, Æneas, Greeks, &c.)

(except Thersites and Patroclus part.)

(Hector, Andromache, Cassandra, Priam, Troylus.)

V. 5, 1—end of play—(Troylus's last speech, &c.)

(But Sc. 7, 8, 9, and perhaps Epilogue, probably spurious.)1

It

The spurious part of the last Act is probably débris from Dekker and Chettle's Troylus and Cressida, written in 1592, and reproduced in a revised form as Agamemnon in 1599. If any one doubts that such an amalgamation of plays by different authors could take place let him refer to The Tragedy of Cæsar and Pompey, by Chapman, Act ii. Scene 1, which is clearly not a piece of the play. but the remains of an old play of the same title acted in 1594 at the Rose. alludes to the Knack to Know a Knave, published in the same year, 1594, and acted two years earlier. There are other instances of this: I hope the French scene in Henry V. between Alice and Katharine is one of them. At any rate, no play should be edited without careful consideration of all evidence obtainable as to other plays on the same subject. The play entered in the Stationers' books in 1602 for publication was probably Shakespeare's, not including the Achilles' story; it could not have been Dekker and Chettle's, as they did not write for the Chamberlain's company. It may have been, however, partly made up from their play in the catastrophe, as the Quarto Hamlet was from the early Hamlet of 1588-9. Editors have been misled in this matter by not noticing that the 1602 entry was of a play belonging to the Chamberlain's men.

This part of the play, which contains everything connected with the war, with Hector's challenge to Ajax, with his combat with him, with his final encounter with Achilles, and his death, was, in my opinion, added to the early sketch of Troylus' loves (which was not enough to make a five-act play), not long after the writing of the first part.

;

The style of this second part is more advanced than the first but not so much so as many of the Second Period plays. It reminds us most of The Merchant of Venice or John. It also in parts has an echo of Marlowe, just as we might expect, if I am right as to its date of composition. See my edition of Henry VI.1

For instance :

"Is she worth keeping? why she is a pearl
Whose price hath launch'd above a thousand ships
And turn'd crown'd kings to merchants.

If you'll avouch 'twas wisdom Paris went,
As you must needs, for you all cried 'Go, go:'
If you'il confess he brought some noble prize,
As you must needs, for you all clapt your hands,
And cried 'Inestimable,' why do you now

The issue of your proper wisdoms rate?" &c.—Act ii. Sc. 2. Compare with the first three of these lines Marlowe, Faust— "Was this the face that launch'd a thousand ships

And burnt the topless towers of Ilium?"

I must also notice that, according to Theobald, the Margarelon of Act v. Sc. 5 and Act v. Sc. 7, with the Sagittary of Act v. Sc. 5, are derived from Caxton's Three Destructions of Troy: according to Malone, the Knights, Act iv. Sc. 5, 1. 158, are from the same source. All these references are from the Hector story: which confirms my opinion that that is an integral and separable part of the play. I have not seen the above-named story-book, nor Lydgate's poem on this subject: Lad I had an opportunity of doing so, I should probably have had something to add to what I have here stated as it is, I must be content at present to give the evidence derivable from the materials at my command. I cannot

:

Unfortunately still in MS. (Sept. 1875).

66

omit, however, one little confirmation of my theory that lies on the surface. In Act i. Sc. 2, Hector goes to the field and figh、s. In Act i. Sc. 3, after this, we find him grown rusty in the longcontinued truce." Surely these passages were not written at the same period.

Of course this Hector part will not read as complete in itself as the Troylus story does, inasmuch as it had to be fitted on to it: it is, however, wonderfully near completeness, taking all circumstances into account.

In this part, and this only, we have the style of Shakespeare's third-fourth manner in metre; in word-coining; in metaphor, in development of character. I need not dwell on this, it is the extreme palpability of this fact that has caused all this play to be usually assigned to the date of 1608, or thereabouts: what has not been seen is, that these characters do not run through the whole play, but only this Achilles and Thersites part. I must, however, say a few words on the alterations Shakespeare must have made, if he wrote this play in the way I say he did.

It will have been noticed that I asterised some parts of the Troylus story. This is the reason. These parts are in their present shape evidently remodelled in the last revision. Ulysses's speeches are clearly in the latest manner. The scene between Diomed and Cressida, however, if Troylus, Ulysses, and Thersites are cut out, falls into regular metre rather more than the scene as it stands does: and is in the earliest style. I think it is part of the first Troylus sketch: I am sure that Cressida's rhymed soliloquy is. Readers of

Chaucer will remember, that Troylus in his version discovers Cressida's faithlessness by finding a brooch in a cloak he wins from Diomed in battle. I believe that Shakespeare followed Chaucer, as his only authority, in his first sketch, and so did not take Troylus to the Greek tents at all: this scene being given between Diomed and Cressida only to show that Troylus's suspicion from the brooch was a true one. But finding afterwards how easily he could make him see instead of suspect by sending him with Hector to the Greek tents, he cut out the fighting scene and the brooch, and put in the additions to this scene. So we explain all the difficulty under this head. The other asterised bits are all of the Third Period, put in to match the new version of this scene. There are other little links too minute to note here, which I should point out in editing the play.

But there is one point noticed by the Cambridge Editors that so strongly confirms my theory, that I must give it in full. It will be seen from the tables given above that the Troylus story ends at Act v. Sc. 3, the Hector story at the end of the present play: while the final additions as to Ajax, Ulysses, &c., are all inserted in the previously existing parts, and do not reach to the end; either as we have it now, or as it existed in either of the two earlier stages. Now Shakespeare would not in all probability write even so incomplete a sketch as the Troylus story without contriving an end for it and writing this end. This is the practice of all great writers, as far as we can trace their manner of work: and we find it exemplified in Twelfth Night, the only other play of Shakespeare composed in the same way as this one, at two distinct periods: the end there is clearly of the early work. We ought, therefore, to find some trace of the first ending of the Troylus story, if anywhere, at the close of Act v. Sc. 3. Of course Shakespeare may have obliterated it, but if he has not, it can be only looked for where the love story is closed. Now exactly at that point we read in the Folio three lines,

"Pan. Why but hear you?

Troy. Hence, brother lackey, ignomy and shame,
Pursue thy life and live aye with thy name;"

which three lines are evidently meant for the original end of the play, as they occur again just before Pandarus's final epilogue. The

occurrence of these lines in both these places cannot be explained by supposing a second author for the last scenes: for Act v. Sc. 5, and Act v. Sc. 10, which occur after the first insertions of them in Act v. Sc. 3, are undoubtedly Shakespeare's, although the piece from the entrance of "one in sumptuous armour" (Act v. Sc. 6), to the end of Act v. Sc. 9, is of dubious authenticity, and perhaps the Pandar epilogue. I do not, however, discuss this question here. It is of more importance to our present subject to see if the metrical tests will bear out our previous conclusions. And before giving statistics, I must observe that the use of these tests seems to be misunderstood even by those who have used them as supporting their views; or are using them to obtain conclusions on disputed points. I lay down, therefore, some canons of method relating to them.

I. No conclusion can be drawn from an insufficient number of instances. This number varies with the test. A dozen instances of weak-ending in a page of ordinary 8vo. would stamp a play at once as Massinger's; but to any conclusion drawn from less than 1,000 lines as to number of rhymes or double-endings I should attach very little value.

II. Tables of ratios must not be used without considering the positive amounts of the numbers from which the Latios are calculated: thus, in comparing The Tempest with Winter's Tale, the ratios of rhymes to blank verse lines come out as I: 729 and 1: infinity respectively. This looks like an enormous difference, but it means only that there is one rhyming couplet in The Tempest and none in Winter's Tale. No conclusion could be based on such a ground. Again, in The Merry Wives of Windsor the addition of one rhyme would alter the proportion from I: 22 to I: 20, so that if anyone unacquainted with Shakespeare's metre were to count

"Fear you not that! Go, get us properties

And tricking for our fairies,"

as a rhyme, he would displace the position of the play considerably in the table. It is clear that, in plays chiefly prose, conclusions cannot be drawn from these tests in cases where the numbers are close together.