A.--Arch. Selden B. 24; Bodleian Library (_First class_). Add.--Additional 9832; British Museum (_First class_). Additional 12524; British Museum (_First class_). B.--Bodley 638; Bodleian Library (_Second class_). C.--Cambridge Univ. Library, Gg. 4. 27 (_First class_). F.--Fairfax 16; Bodleian Library (_Second class_). P.--Pepys 2006; Magd. Coll., Cambridge (_First class_). T.--Trinity College, Cambridge, R. 3. 19 (_First class_). Th.--Thynne's edition, pr. in 1532 (_Second class?_). Tn.--Tanner 346; Bodleian Library (_Second class_). [alpha].--Additional 28617; British Museum (_First class_); but only a fragment, viz. ll. 513-610, 808-1105, 1306-1801, 1852-2110, 2125-2135, 2151-2723. [beta].--Cambridge Univ. Library, Ff. 1. 6 (Thisbe _only_). [gamma].--Rawlinson C. 86; Bodleian Library (Dido _only_). They may be thus described. C. (Camb. Univ. Lib. Gg. 4. 27) is the famous Cambridge MS., containing the Canterbury Tales, denoted by the symbol 'Cm.' in the footnotes to vol. iv (i.e. throughout the Canterbury Tales); also by the symbol 'Gg.' in vol. i., i.e. in the Minor Poems; see p. 49 of the Introduction to vol. i. It also contains some other pieces by Chaucer, viz. the A. B. C., Envoy to Scogan, Truth, Troilus, and the Parlement of Foules. It is of early date, and altogether the oldest, best, and most important of the existing copies of the Legend. I shall call all those that resemble it MSS. of the _first class_. Its great peculiarity is that it possesses the unique copy of the early draught of the Prologue; see p. xxi. Upon comparison of it with the Fairfax MS. (the best MS. of the _second class_), it is found to offer slight differences in many places throughout the various Legends, besides presenting large differences throughout the Prologue. The variations are frequently for the better, and it becomes clear that the first class of MSS. is of an older type. The second class is of a later type, and differs in two ways, in one way for the worse, and in another way for the better. In the former respect, it presents corrupted or inferior readings in several passages; whilst, on the other hand, it presents corrections that are real improvements, and may have been due to revision. No doubt there was once in existence a correct edition of the revised text, but no existing MS. represents it. We can, however, practically reconstruct it by a careful collation of MS. C. with MS. F.; and this I have attempted to do. Throughout the Prologue, I take MS. C. as the basis of the 'A-text,' correcting its eccentricities of spelling, but recording them in footnotes wherever the variation is at all important; such a variation as _hym_ for _him_, or _yt_ for _hit_, I regard as being of no value. At the same time, I take MS. F. as the basis of the B-text, and correct it, where necessary, by collation with the rest. Throughout the Legends themselves, I take MS. F. as the basis of the text, collating it with C.