I give above, in parallel columns, the scheme proposed, and those of Delius, Malone, Drake, and Chalmers. I quote these three latter from Allibone. Before remarking on the scheme, I desire to add that it is by no means final: I have myself several other tests in course of application, and I have no doubt other workers in the same field will find additional ones, now that the subject is ventilated. As a provisional scheme, however, 1 place some confidence in it, for this reason: that, although it is based solely on the metrical table, it in no instance contradicts any external evidence. It also distributes the work much more equally than the other schemes; never requiring more than two plays to be written in one year.

Remarks on the Position of certain Plays in this List :—

First Period.-These five plays are distinctly marked off as a separate class by the vast preponderance of rhyming lines. Love's Labour's Lost has more than 1000, Midsummer Night's Dream 850, Romeo and Juliet 650, Richard II. 530, and Comedy of Errors (though a very short play) 380, which is equivalent to 600 in a play of ordinary length. Now, no other of Shakespeare's plays reaches to the number of 200 rhyming lines; and as the battle between rhymed and unrhymed compositions was fierce at this time, I feel that there is no doubt that Shakespeare joined the advocates of rhyme at first, and gradually learned to feel the superi ›rity of blank verse; at any rate, the difference between these Five Plays of the first period, as to amount of rhyme, is too great, in my opinion, to admit any other play, however inferior, to be ranked with them. I know how strongly some think that the Two Gentlemen of Verona must have preceded Midsummer Night's Dream, because this latter is so beautiful a "work." I do not say a "play;" for I agree with N. Drake and others in the view that the Two Gentlemen is superior as an acting piece, however inferior as a poem. I must, for want of space, refer to Drake's Treatise for a full statement of his arguments. For myself, I find it impossible to believe that the Two Gentlemen was not written some two years before as the Merchant of Venice, which is so like it in metrical handling; and equally impossible to regard the Midsummer Night's Dream as a production of any but the earliest period, when fancy was strong, and the sense of the prose realities of life comparatively weak. Note also that the three

comedies in this first period all observe the unity of time, no action extending to the second day, and that they are all similar in their nature, turning on the solution, as it were, of an embroilment produced under circumstances barely or only hypothetically possible. The almost total absence of Alexandrines in Romeo and Juliet, and their absolute absence in the three comedies (the one instance in Love's Labour's Lost I think is corrupt), is another very striking difference from all the other plays. In Richard 11., however, these Alexandrines are admitted, and this is therefore the play in which this, the first sign of the Second Period, begins to show itself. In other respects this play is, to my thinking, far removed from John or Henry IV. It bears something of the same relation to Marlowe's Edward II. that the Two Gentlemen of Verona does to the Taming of a Shrew, or Richard III. to Henry VI. Shakespeare in it seems not to move with the same freedom that he does in his later plays, and the whole work has an artificial air. Another point that distinctly separates the earlier from the later historic plays is, the absence of prose: Richard II. and John have none, Richard III. only one bit, but that reads like, and I believe is, a portion of Peele's work; and of these earlier plays Richard III. is the only one that is absolutely devoid of Comedy. This also marks its position.

Second Period.-The positions I have given to the plays in this period so nearly coincide with those generally assigned (except as to the Two Gentlemen of Verona, which has already been noticed) that no special remarks seem needed. One general characteristic of the period is the diminution of the number of rhyming lines, which number about 100 to 200 in this period for a full play. Much Ado and Merry Wives have a smaller number, but are almost entirely in prose, being quite exceptional in this respect. Also the number of short lines is considerably increased, though not nearly so much as in the next period. Alexandrines are admitted from 5 to 20 in a play (but in Richard II. there are 33 ?); the number of feminine endings increases, but not in any regular progression, and doggerel lines, stanzas, sonnets, and alternate rhymes (which abound in the earliest plays) gradually die out: there are not many in any plays of this period.

Third Period.-In this a few words may be needed on the position I have assigned to Macbeth. I agree with Clark and Wright that this play has been much altered since its first composition, and has now many interpolations in it; in Act iv. Scene 3-the passage on the touching for the evil the marks of interpolation are palpable; 1. 159 follows metrically on l. 139,

--

"'Tis hard to reconcile-See who comes here"

making a perfect line: the Doctor (unknown elsewhere in the play) is dragged in for the express purpose of introducing the subject; and "Comes the king forth, I pray you," 1. 140, is inconsistent with "Come, go we to the king," 1. 236.

In this Third Period the metre is much freer; prose and verse are intermingled in the same scene; tri-syllabic feet abound; short lines are very abundant; double endings very greatly multiplied; Alexandrines not composed of two lines of six syllables are introduced; the Alexandrines with regular cæsura increase greatly; the number of rhyming lines gradually falls off. The plays are difficult to test, as to metre, in this period; partly from the similarity of style in the great plays, partly from the great variations in the Quarto and Folio texts. I am, however, applying further tests, which I hope will be decisive.

Fourth Period. In this the rhymes fall off rapidly, and in the Comedies actually disappear; the metre becomes more regular and less impassioned,' and the general impression left by these later works is, that they were produced at greater leisure, and more carefully polished. The dates given by metrical considerations agree too nearly with those assigned on external evidence to need com

ment.

And here I think I may fairly point out how singular the coincidence of the order here given is with that assigned by the best English critics on external evidence. This order was first made out from the rhyme-test only; and, except in the instances of plays which are not undoubtedly authentic or written at two different periods, I have not changed the relative position of one since I first

Professor Ingram has since shown that "weak endings" are the characteristic test of the Fourth Period.

sent this list to Messrs. Clark and Wright in 1870. At that time I had not read any treatise on the external evidences, and was not even aware that any attempt had been made to classify the plays into periods. I own this with some shame; but claini, at the same time, some additional confidence in the results of the rhythmical tests. It may seem to some ludicrous to speak even of the application of mathematics to such a subject; but it will be seen from the table that the plays assigned to the period ending in 1598 by the rhyme test, exactly agree with those in Meres's list (setting aside questions of genuineness). Now, the doctrine of chances gives us as the odds against these 10 plays being selected out of the 30 which are undoubtedly more or less Shakespeare's, more than 20 millions to one: in exact numbers, one chance only out of 20,030,010 would hit on this exact selection of plays. To a mind accustomed to the exact sciences, this fact alone is conclusive as to the immense value of the rhyme test.

I might go into detail concerning the reasons for the position of each particular play; but I think it better to consider all special matter separately. The table itself is subjoined. It is only necessary here to add a caution as to the amount of subjectivity to be expected in such a table as this; there must be some until the laws of metre are more definitely laid down than they are at present. (I.) As to the rhymes, it is sometimes doubtful if a rhyme is intentional or accidental. In such cases the rhyme is counted in this table. (II.) It is sometimes doubtful if one line of six feet, or two lines, one of five feet, and one of one, be intended. In the following table the line is, if possible, reckoned as divided. (III.) In some instances, lines of four feet in the Globe Edition can be avoided by re-arranging the lines without altering the reading; this has been sparingly done in Pericles, and in cases where the arrange. ment of the lines is made by modern editors without authority in the original texts. (IV.) All Alexandrines proper with cæsura at the end of the third foot are counted in the six-measure lines, and not as two lines of three measures, except where, as in The Two Gentlemen of Verona and Richard III., lines of six syllables are repeated many times together. In the larger tables from which this one is abridged, all peculiarities are noted for each scene; perhaps when our text is settled, it may be worth while to print such