though of these three particular octave-forms not one appears in the sonnets of 1807 (in which indeed there is but one irregular octave to be found, viz., ABBA: ABAB, in "I grieved for Buonaparte"), they all occur, along with many others, in the collective body of Wordsworth's sonnets-ABAB: BAAB four times, and the other two forms twice.' Indeed, of the twenty-four distinct combinations of rhyme-endings which (admitting an alternative rhyme cc in the second quatrain) are possible in the octave, Wordsworth employs from first to last no fewer than nineteen. The following table shows these possible combinations, the five marked * being those not found in Wordsworth:

1 References to the examples from Dante and Petrarch had better be given. ABAB: ABAB: Dante, Sonetti viii. ix. x. xvi. xviii. xix. xxxii. Petrarch (Vita), Sonetti xxxvi. li. xc. cxxxv. (Morte) xii. xiii. xxxix. xlii. xliii. 1. ABAB: BABA: Petrarch (Vita) Sonetto ccii. (Morte) Sonetto xi. ABAB: BAAB: Petrarch (Vita) Sonetto clvi. (Morte) Sonetto xxvii. It is noteworthy that, in the first and the last of these forms, what Archbishop Trench calls the strong outward framework of the sonnet (▲ ** A** A) is dislocated.

The figures subjoined to each form denote the number of times that particular form is used by the Poet out of a grand total of 523 sonnets. The figures in the table amount to 520. Add one Shakespearian sonnet and two with irregular octaves.

Wordsworth, it is to be observed, used what we may call the degenerate octave-forms very sparingly at first, but more and more freely as time advanced. It is convenient for the purpose of studying the forms to divide Wordsworth's sonnets

into four periods: (1) the sonnets written before and during 1807; (2) those written between 18081821; (3) those that fall within 1822-1840; and (4) those belonging to 1841-1846. In the first period, as we have seen, but one degenerate form occurs, while the normal two-rhymed octave is found thirtyseven times to twenty times of ABBA: ACCA (the normal three-rhymed type). In the second period (1808-1821) there are one hundred and fifty-eight two-rhymed, to one hundred and thirty-three threerhymed, normal types; while of degenerate forms, ABAB BAAB Occurs four times, ABAB: ABBA thrice, and ABBA: ABAB twice. In the third period (1822 -1840) the relative proportions of the two-rhymed and three-rhymed types are reversed—they are now forty-six to sixty-eight; while degenerate forms of octave to the number of forty-eight crowd in. Lastly, from 1840 to 1846, while the degenerate forms actually amount to more than half the total (thirty-three out of sixty-one), the two-rhymed type resumes its old place, being slightly in advance of the three-rhymed (fifteen to thirteen).

A word remains to be said of the sestet-forms.

Of these Wordsworth employs as many as twentytwo-some of them, of course, far more frequently than others. Unhappily he fell into two mistakes with regard to the sestet; into the one unwittingly, into the other with his eyes open. Failing to keep carefully in view the contrast and opposition between octave and sestet, he missed the allimportant rule that "the disposition of the rhymes in the tercets [sestet] must be such as not to reproduce the disposition of those in the quatrains [octave]" (Pattison), and consequently admitted freely such forms as abba ba, a babba, a b ba a b, in which the cadences of the octave are monotonously reiterated. To do this was, in effect, to neglect the melodic means at his command; and thus Wordsworth has, by an unlucky oversight, sadly marred what might else have been "the ravishing division" of many a fine sonnet.1 The other mistake is not easily to be pardoned. Wordsworth was well aware of the

1 At the close of the octave the sonnet-wave, cresting, falters for a moment, to fall and hurry inshore in the sestet (prono præceps agitur decursu). The inflow of the sestet is checked by the adoption of the inverted rhymesequence (rima chiusa) of the octave (a b ba).

objections to a rhyming-couplet at the close of the sonnet, and yet he frequently indulged in this provoking vice-indeed, from 1822 to 1840 a b bacc and ab a b c c were his favourite sestet-sequences; and, even in the sonnets of the first period, a b b a c c is found five times. Subjoined is a table showing the Poet's principal sestet-forms, with the number of times each is used out of a total of five hundred and twenty-three sonnets:

*The favourite sestet-forms of Dante and Petrarch.

§ Favourite forms of Michael Angelo.

¶ Occurs eight times in Dante's forty-three sonnets.

It may be well to add a table giving the sestets