¹ E. Moore, Textual Criticism of the D.C. (Cambridge, 1889), pp. 686–687.

² See L. G. Blanc, Gramm. d. ital. Spr. (Halle, 1844), p. 781; C. F. Schneider, Ueber d. Reim in D's. D. C. (Bonn, 1869), p. 2.

³ There is a curious parallel in Chaucer; when he uses a dubious rime in his ababbcc stanza in the Man of Law's Tale he tries to put the words as far apart as he can, in the second and fifth lines. On what follows see figures in N. Zingarelli, Vita di Dante (Milan, 1914), p. 114.

⁴ How deliberately the size of the several members of the Commedia was planned is shown at the end of the Purgatorio,—

Ma perchè piene son tutte le carte

Ordite a questa Cantica seconda, …

⁵ Romanic Review, xxiii, 38–40, with the suggestion that the grim lines Inf., xx, 1–3, are a spurious addition. To him and Dr. Rudolph Altrocchi I am grateful for suggestions and cautions.—I add that nowhere could a whole canto have been inserted without disturbing the numerical symmetry of the poem.

⁶ Moore, Textual Crit., pp. xviii, 686–687.

⁷ Moore, pp. 706–711, esp. 709.

⁸ Among many critics see Schneider, op. cit., p. 4.

⁹ On the Provençal, Gröber's Grundriss d. rom. Ph., ii, ii, and Grundr. d. prov. Lit.; F. Witthoeft, Sirventes Joglaresc, in Ausg. u. Abh., 88 (Marburg, 1891); K. Bartsch in Jahrbuch f. rom. u. engl. Philol., i, esp. 178–182. The sapphic serventese appears in French as late as about 1400; see Collection des Chron. Nat. Franç., xxiii, 323–405, in Buchon's Froissart.

¹⁰ E. Monaci, Crestomazia Ital. dei Primi Sec. (Città di Castello, 1889),i, 295–296 (thirteenth century); another sapphic serventese in Monaci, pp. 406–411 (one of the oldest—a narrative), and one in Scelta di Curiosità, xi (Bologna, 1865), pp. 11–24. See N. Zingarelli, Vita … di Dante (Milan, 1931), pp. 772–773, who points out that terza rima in essence is formed on the principle of the sestet, or the canzone.

¹¹ Here and later I use E. Moore's Opere di Dante, re-edited by Toynbee (Oxford, 1924); two such anonymous sestine in this form are on pp. 182c and 182d, and see Blanc, op. cit., pp. 765–766.

¹² The first suggestion of this origin seems to have been made by H. Schuchardt, Ritornell u. Terzine (Halle, 1874), pp. 122–133, also 3, 6, 17, and apparently by V. Imbriani. See also M. Barbi in Studi … dedicali a Pio Rajna (Milan, 1911), p. 97. The subject seems confused.

¹³ Schuchardt, p. 122; Barbi, loc. cit., pp. 97–98.

¹⁴ Schuchardt (pp. 133–34) gives one fourteenth century case of six ritornelli strung together, without interlinking; later cases pp. 17, 134, and in O. Marcoaldi's collection of nineteenth-century Umbrian folk-poetry (p. 60), etc.

¹⁵ The sonnets in these two groups are those numbered, in the Moore-Toynbee Opere, 54, 26*, 29, 30, 33*, 36, 41, 47*, 49, 51*, 51, V.N. 12 and 13; 43, 46, 44, 49*, and V.N. 1. Poems of over fourteen lines called sonnets are disregarded (but show how he experimented with other forms than the terzina), and also dubious and spurious ones, in the Moore-Toynbee edition; but of these quite a number are of the first sort above.—A minor American poet has lately realized the kinship between the two forms of verse, and published a volume Terza-Rima Sonnets, each having four terzine with a final couplet resuming one of the earlier rimes.

¹⁶ Altital. Lesebuch (Halle, 1886), pp. 88–92, 101; Crestomazia, i, 193–310. I disregard a half-dozen or more sonnets in which all the six rimes are the same word. One or two sonnets in the two books may be identical. Among the early Sicilian poets about a third of the sonnets have this scheme, almost all the rest abcabc (Langley in PMLA, xxviii, 518, 520).

¹⁷ The same scheme exists in the first six lines of ottava rima, used in popular poetry, some believe (Volpi, Il Trecento, p. 188), earlier than Boccaccio, the first literary poet to use it.—I do not know, nor seemingly do others, the origin, history, and exact significance of the phrases terza, ottava rima, instead of rima di tre, di otto, which seem more idiomatic for the present sense. But the “third rhyme,” la terza rima, is exactly what and only what distinguishes this verse from the common early form of sestet. For want of evidence I do not pursue this subject.

¹⁸ Melodic and mnemonic repetitions of rimes and other words at ends and beginnings of adjacent stanzas are far from being confined to Romance poetry, and are a marked feature of certain English poems, narrative and especially lyric, and especially in the fourteenth century in northerly England. Such are the Song of the Husbandmen in T. Wright, Polit. Songs, Camd. Soc. (1839), pp. 149–152, Pearl, Awntyrs of Arthure, a legend of the B.V.M. in C. Horstmann, Altengl. Leg., N.F. (Heilbronn, 1881), pp. 499–502, several lyrics by Lawrence Minot. On these and others see J. E. Wells, Manual of the Writings in Mid. Engl. (New Haven, 1916), index. No doubt the repetitions are for melodic, mnemonic, and sometimes structural purposes; whether ever to prevent losses is hard to say.—This is by no means the only case of reappearance in the far north of very special usages of the far south. I would specially mention the use by Burns in To a Louse, To a Mouse, To a Mountain Daisy, and many others, of a very peculiar and quaint stanza not uncommon in medieval Provençal; see C. A. F. Mahn, Werke d. Troubadours (Berlin, 1846), i, 3, etc., etc., and C. Appel, Prov. Chrest. (Leipzig, 1920), p. 80. Some of these reappearances of forms may be coincidences, but not all; the explanation I do not know.—Adjacent final-initial repetitions occur in Welsh poems, of date unknown to me, but some professedly early; see Iolo MSS (Llandovery, repr. Liverpool, 1888), pp. 240–242, 246–247, 250–251, 313, 645, 650, 693. Rimes linking stanzas occur in middle Breton, as Loth shows (Métr. Galloise, ii, ii, 188–189, in d'Arbois de Jubainville, Cours de Litt. Celt., xi); and the other repetitions as early as ninth- or tenth-century Irish (D. Hyde, Lit. Hist. of Irel., pp. 413–414), in a poem ascribed to Angus the Culdee.—The alphabetical poems, stanzas opening with successive letters, not uncommon in medieval Latin, French, and English, may be mnemonic as well as tours de force.—Of course none of this is much like Dante's scheme.

¹⁹ Inf., xx, 19, xxii, 118; Purg., ix, 70, xxxiii, 136–41; Par., xxii, 106.

²⁰ M. P., xviii, 625–626; PMLA, l, 105, and passim.

²¹ Epist. x, 10; De Vulg. Eloq., ii, 4; commedia used of Dante's poem in Inf., xvi, 128, xxi, 2, and in Epist. x, 6, 13.

²² On the history of the Divina, see Rajna, Bull. della Soc. Dant. Ital., n.s., xxii, 107–115, 255–258; also iii, 9–10.

²³ xxiii, 62, xxv, 1 ff. But in the sublime heavens commedia would be out of place. There is significance in his ways of speaking of the poem in the several parts of it.

²⁴ M. P., xviii, 631–632. In De Vulg. Eloq., ii, 4 he defends the use of poetae for some vernacular writers—but in contrast with magni poetae.

²⁵ Inf., i, 87, xvi, 129; Purg., ix, 70–71; etc.

²⁶ Novelle 114 and 115, mentioned to me by Dr. J. B. Fletcher, and quoted in various books on the Dante-tradition, such as G. Papini's (Lanciano, 1910), pp. 25–27.

²⁷ Moore, Text. Crit., pp. xviii f.