last Book of Troilus, ll. 358-385. We may also observe the following lines in Troilus, from Book I, 517-8:—

These lines, jestingly applied to Troilus by Pandarus, are in the House of Fame, 639, 640, applied by Chaucer to himself:—

Again, the House of Fame preceded the Legend of Good Women, because he here complains of the hardship of his official duties (652-660); whereas, in the Prologue to the Legend, he rejoices at obtaining some release from them. We may also note the quotation from Boethius (note to l. 972). As Boethius and Troilus seem to have been written together, somewhere about 1380, and took up a considerable time, and the apparent date of the Legend is 1385, the probable date of the House of Fame is about 1383 or 1384. Ten Brink further remarks that the references to Jupiter suggest to the reader that the 10th of December was a Thursday (see note to 111). This would give 1383 for beginning the poem; and perhaps no fitter date than the end of 1383 and the spring of 1384 can be found.

§ 6. Metre. Many of Chaucer's metres were introduced by him from the French; but the four-accent metre, with rime as here employed, was commonly known before Chaucer's time. It was used by Robert of Brunne in 1303, in the Cursor Mundi, and in Havelok. It is, however, of French origin, and occurs in the very lengthy poem of Le Roman de la Rose. Chaucer only employed it thrice: (1) in translating the Roman de la Rose; (2) in the Book of the Duchesse; and (3) in the present poem.

For normal lines, with masculine rimes, see 7, 8, 13, 14, 29, 33, &c. For normal lines, with feminine rimes, see 1, 2, 9, 15, 18, &c. Elision is common, as of e in turne (1), in somme (6), in Devyne (14); &c. Sometimes there is a middle pause, where a final syllable need not always be elided. Thus we may read:—

Two short syllables, rapidly pronounced, may take the place of one:—

The first foot frequently consists of a single syllable; see 26, 35, 40, 44; so also in l. 3, where, in modern English, we should prefer Unto.

The final e, followed by a consonant, is usually sounded, and has its usual grammatical values. Thus we have think-e, infin. (15); bot-e, old accus. of a fem. sb. (32); swich-e, plural (35); oft-e, adverbial (35); soft-e, with essential final e (A.S. sōfte); find-e, pres. pl. indic. (43); com-e, gerund (45): gret-e, pl. (53); mak-e, infin. (56); rod-e, dat. form used as a new nom., of which there are many examples in Chaucer (57); blind-e, def. adj. (138). The endings -ed, -en, -es, usually form a distinct syllable; so also -eth, which, however, occasionally becomes 'th; cf. comth (71). A few common words, written with final e, are monosyllabic; as thise (these); also shulde (should), and the like, occasionally. Remember that the old accent is frequently different from the modern; as in orácles, mirácles (11, 12): distaúnc-e (18), aventúres, figúres (47, 48): povért (88): málicióus (93): &c. The endings -i-al, -i-oun, -i-ous, usually form two distinct syllables.

For further remarks on Metre and Grammar, see vol. v.

§ 7. Imitations. The chief imitations of the House of Fame are The Temple of Glas, by Lydgate^[6]; The Palice of Honour, by Gawain Douglas; The Garland of Laurell, by John Skelton; and The Temple of Fame, by Pope. Pope's poem should not be compared with Chaucer's; it is very different in character, and is best appreciated by forgetting its origin.

§ 8. Authorities. The authorities for the text are few and poor; hence it is hardly possible to produce a thoroughly satisfactory text. There are three MSS. of the fifteenth century, viz. F. (Fairfax MS. 16, in the Bodleian Library); B. (MS. Bodley, 638, in the same); P. (MS. Pepys 2006, in Magdalene College, Cambridge). The last of these is imperfect, ending at l. 1843. There are two early printed editions of some value, viz. Cx. (Caxton's edition, undated); and Th. (Thynne's edition, 1532). None of the later editions are of much value, except the critical edition by Hans Willert (Berlin, 1883). Of these, F. and B., which are much alike, form a first group; P. and Cx. form a second group; whilst Th. partly agrees with Cx., and partly with F. The text is chiefly from F., with collations of the other sources, as given in the footnotes, which record only the more important variations.

§ 9. Some emendations. In constructing the text, a good deal of emendation has been necessary; and I have adopted many hints from Willert's edition above mentioned; though perhaps I may be allowed to add that, in many cases, I had arrived at the same emendations independently, especially where they were obvious. Among the emendations in spelling, I may particularise misdemen (92), where all the authorities have mysdeme or misdeme; Dispyt, in place of Dispyte (96); barfoot, for barefoot or barefote (98); proces (as in P.) for processe, as in the rest (251); delyt, profyt, for delyte, profyte (309, 310); sleighte for sleight (462); brighte^[7], sighte, for bright, sight (503, 504); wighte, highte, for wight, hight (739, 740); fyn, Delphyn (as in Cx.), for fyne, Delphyne (1005, 1006); magyk, syk, for magyke, syke (1269, 1270); losenges, for losynges (1317), and frenges (as in F.) for frynges, as in the rest (1318); dispyt for dispite (1716); laughe for laugh (Cx. lawhe, 1809); delyt for delyte (P. delit, 1831); thengyn (as in Th.) for thengyne (1934); othere for other (2151, footnote). These are only a few of the instances where nearly all the authorities are at fault.

The above instances merely relate to questions of spelling. Still more serious are the defects in the MSS. and printed texts as regards the sense; but all instances of emendation are duly specified in the footnotes, and are frequently further discussed in the Notes at the end. Thus, in l. 329, it is necessary to supply I. In 370, allas should be Eneas. In 513, Willert rightly puts selly, i.e. wonderful, for sely, blessed. In 557, the metre is easily restored, by reading so