And Paris to Enone;
And Iason to Isiphile;
And eft Iason to Medea;
And Ercules to Dyanira;

For he lefte hir for Iöle,

That made him cacche his deeth, parde!
How fals eek was he, Theseus;

That, as the story telleth us,

How he betrayed Adriane;

The devel be his soules bane! 1

For had he laughed, had he loured,

He moste have be al devoured,

If Adriane ne had y-be!' &c. Id. 387.

Here we already have an outline of the Legend of Phyllis; a reference to Briseis, Jason and Hypsipyle and Medea, and Deianira; a complete sketch of the Legend of Ariadne; and another version of the Legend of Dido.

We trace a lingering influence upon Chaucer of the Roman de la Rose; see notes to ll. 125, 128, 171. Dante is both quoted and mentioned by name; ll. 357-360. Various other allusions are pointed out in the Notes.

In ll. 280, 281, 284, 305-308 of the A-text of the Prologue (pp. 26, 27), Chaucer refers us to several authors, but not necessarily in connection with the present work. Yet he actually makes use of Titus (i. e. Livy, l. 1683), and of the 'epistels of Ovyde.' He also takes occasion to refer to his own translation of the Roman de la Rose (B. ll. 329, 441, 470), and to his Troilus (II. 332, 441, 469); besides enumerating many of his poems (417428).

I. THE LEGEND OF CLEOPATRA. The source of this legend is by no means clear. As Bech points out, some expressions shew that one of the sources was the Epitome Rerum Romanarum of L. Annæus Florus, lib. iv. c. II; see note to ll. 655, 662, 679. No doubt Chaucer also consulted Boccaccio's De Claris Mulieribus, cap. 86, though he makes no special use of the account there given. The story is also in the history of Orosius, bk. vi. c. 19; see Sweet's edition of King Alfred's

1

Cf. L. G. W. 2177, 2227.

2 Cf. L. G. W. 1952-8.

Orosius, p. 247. Besides which, I think he may have had access to a Latin translation of Plutarch, or of excerpts from the same. See the notes.

It is worth while to note here that Gower (ed. Pauli, iii. 361) has the following lines :

'I sigh [saw] also the woful quene
Cleopatras, which in a cave
With serpents hath her-self begrave
Al quik, and so was she to-tore,
For sorwe of that she hadde lore
Antonie, which her love hath be.

And forth with her I sigh Thisbe'; &c.

It is clear that he here refers to Chaucer's Legend of Good Women, which, in fact, is certainly the model on which the Confessio Amantis was based. Compare L. G. W. 11. 695-697; and note that, both in Chaucer and Gower, the Legend of Thisbe follows that of Cleopatra; whilst the Legend of Philomela immediately follows that of Ariadne. This is more than mere coincidence. See Bech's essay; Anglia, v. 365.

II. THE LEGEND OF THISBE. This is from Ovid's Metamorphoses, iv. 55-166, and from no other source. Some of the lines are closely translated, but in other places the phraseology is entirely recast. The free manner in which Chaucer treats his original is worthy of study; see, as to this, the excellent criticism of Ten Brink, in his Geschichte der Englischen Litteratur, ii. 117. Most noteworthy of all is his suppression of the mythological element. The story gains in pathos in a high degree by the omission of the mulberry-tree, the colour of the fruit of which was changed from black to white by the blood of Pyramus; see note to 1. 851. This is the more remarkable, because it was just for the sake of this very metamorphosis that Ovid admitted the tale into his series. See also notes to 11. 745, 784, 797, 798, 814, 835, 869, &c.; and cf. Gower's Confessio Amantis, ed. Pauli, i. 324.

III. THE LEGEND OF DIDO. Chiefly from Vergil's Aeneid, books i-iv. (see note to 1. 928, and compare the notes throughout); but ll. 1355-1365 are from Ovid's Heroides, vii. 1–8.

quoted at length in the note to 1. 1355. And see, particularly, the House of Fame, ll. 140-382. Cf. Gower, C. A. ii. 4-61.

IV. THE LEGENDS OF HYPSIPYLE AND MEDEA. The sources mentioned by Morley are Ovid's Metamorphoses, bk. vii., and Heroides, epist. vi. ; to which we must add Heroides, epist. xii. But this omits a much more important source, to which Chaucer expressly refers. In 1. 1396, all previous editions have the following reading,-' In Tessalye, as Ovyde telleth us'; but four important MSS. read Guido for Ovyde, and they are quite right. The false reading Ovyde is the more remarkable, because all the MSS. have the reading Guido in l. 1464, where a change would have destroyed the rime. As a matter of fact, ll. 1396-1461 are from Guido de Colonna's Historia Troiana, book i. (see notes to ll. 1396, 1463); and ll. 1580–3, 1589–1655 are also from the same, book ii. (see notes to ll. 1580, 1590). Another source which Chaucer may have consulted, though he made but little use of it, was the first and second books of the Argonauticon of Valerius Flaccus, expressly mentioned in 1. 1457 (see notes to 11. 1457, 1469, 1479, 1509, 1558). The use made of Ovid, Met. vii., is extremely slight (see note to 1. 1661). As to Ovid, Her. vii., xii., see notes to ll. 1564, 1670. The net result is that Guido is a far more important source of this Legend than all the passages from Ovid put together. Chaucer also doubtless consulted the fifth book of the Thebaid of his favourite author Statius; see notes to ll. 1457, 1467. Compare also Boccaccio, De Claris Mulieribus, capp. 15, 16; and the same, De Genealogia Deorum, lib. xiii. c. 26. Observe also that Gower gives the story of Medea, and expressly states that the tale 'is in the boke of Troie write,' i. e. in Guido. See Pauli's edition, ii. 236.

V. THE LEGEND OF LUCRETIA. Chaucer refers to Livy's History (bk. i. capp. 57–59); and to Ovid (Fasti, ii. 721-852). With a few exceptions, the Legend follows the latter source. He also refers to St. Augustine; see note to 1. 16902. Cf.

1 Gower is amusing, when he turns Ovid's ' Ad uada Maandri' (Her. vii. 2) into a reference to 'King Menander'!

2 In fact, St. Augustine tells the whole story; De Ciuitate Dei, cap. xix. And it was copied from St. Augustine's version into the Gesta Romanorum, Tale 135.

Boccaccio, De Claris Mulieribus, cap. 46, who follows Livy. Several touches are Chaucer's own; see notes to ll. 1812, 1838, 1861, 1871, 1881.

Gower has the same story (iii. 261), and likewise follows Ovid and Livy (p. 263).

VI. THE LEGEND OF ARIADNE. From Ovid, Met. vii. 456-8, viii. 6-182; Her. Epist. x. (chiefly 1-74). But Chaucer consulted other sources also, probably a Latin translation of Plutarch's Life of Theseus; Boccaccio, De Genealogia Deorum, lib. xi. capp. 27, 29, 30; also Vergil, Aen. vi. 20-30. Cf. House of Fame, 405-426; and Gower, ii. 302.

VII. THE LEGEND OF PHILOMELA. Chiefly from Ovid, Met. vi. 424-605; and perhaps from no other source, though the use of the word radevore in 1. 2352 should be accounted for. Cf. Boccaccio, De Genealogia Deorum, lib. ix. c. 8; and Gower, Conf. Amantis, ii. 313, who refers us to Ovid.

VIII. THE LEGEND OF PHYLLIS. Chiefly from Ovid, Her. Epist. ii.; cf. Remedia Amoris, 591-608. But a comparison with the story as told by Gower (C. A. ii. 26) shews that both poets consulted some further source, which I cannot trace. The tale is told by Hyginus and Boccaccio in a few lines. Cf. House of Fame, 388-396. A few lines are from Vergil, Æn. i. 85-102, 142; iv. 373.

IX. THE LEGEND OF HYPERMNESTRA. Chiefly from Ovid, Her. Epist. xiv. But Ovid calls her husband Lynceus, whereas Chaucer calls him Lino. Again, Ovid does not give the name of Lynceus' father. Chaucer not only transposes the names of the two fathers1, but calls Ægyptus by the name of Egiste or Egistes. Hence we see that he also consulted Boccaccio, De Genealogia Deorum, lib. ii. c. 22, where we find the following account. 'Danaus Beli Prisci fuit filius, ut asserit Paulus 2, et

1 We must remember that, in olden times, writers often had to trust to their memory for details not always at hand. Hence such a mistake as this was easily made.

2 The reference seems to be to Paulus Orosius, Hist. i. II; but Belus is not there mentioned. Yet Hyginus (Fab. 168) has: 'Danaus Beli filius ex pluribus coniugibus quinquaginta filias habuit.'

illud idem affirmat Lactantius, qui etiam et ante Paulum Orosium, dicit Danaum Beli filium ex pluribus coniugibus .I. filias habuisse, quas cum Ægistus frater eius, cui totidem erant melioris sexus filii, postulasset in nurus, Danaus oraculi responso comperto se manibus generi moriturum, uolens euitare periculum, conscensis nauibus in Argos uenit .. Ægistus autem, quod spretus esset indignans, ut illum sequerentur filiis imperauit, lege data ut nunquam domum repeterent, ni prius Danaum occidissent. Qui cum apud Argos oppugnarent patruum, ab eo diffidente fraude capti sunt. Spopondit enim se illis iuxta Ægisti uotum filias daturum in coniuges, nec defuit promisso fides. Subornatae enim a patre uirorum intrauere thalamos singulis cultris clam armatae omnes, et cum uino laetitiaque calentes iuuenes facile in soporem iuissent, obedientes patri uirgines, captato tempore iugulauerunt uiros, unaquaeque suum, Hypermestra excepta, quae Lino seu Linceo uiro suo miserta pepercit.' We may note, by the way, that Chaucer's spelling Hypermistre is nearer to Boccaccio's Hypermestra than to the form in Ovid.

THE METRE OF THE LEGEND.

The most interesting point about this poem is that it is the first of the 'third period' of Chaucer's literary work. He here, for the first time, writes a series of tales, to which he prefixes a prologue; he adopts a new style, in which he seeks to delineate characters; and, at the same time, he introduces a new metre, previously unknown to English writers, but now famous as 'the heroic couplet.' In all these respects, the Legend is evidently the forerunner of the Canterbury Tales, and we see how he was gradually, yet unconsciously, preparing himself for that supreme work. In two notable respects, as Ten Brink remarks, the Legend is inferior to the Tales. The various legends composing it are merely grouped together, not joined by connecting links which afford an agreeable relief. And again, the Prologue to the Legend is mere allegory, whilst the famous Prologue to the Tales is full of real life and dramatic sketches of character.

C