When at page 113 of volume i. the editor introduces the words ‘certayne cause’ to complete the sense, he is careful to call attention, in a foot-note, to the fact that these words are not in the original, but taken from the ‘second edition,’ by which I presume he means that of 1498. But when he subsequently supplies seventeen pages which were also not in his original, he gives no hint of the fact; and his reticence was so successful that for fifty years the interpolations passed as genuine among learned critics, who quoted from them passages wholly spurious as Caxton’s genuine text. It was only in 1867 that, in collating Earl Spencer’s copy with the edition of Southey, I discovered that these passages—to which my attention was directed by Messrs. Longman’s note above mentioned—did not correspond with Caxton’s text, as represented by Whittaker’s restorations: and on afterwards collating them with the Osterley text itself I found the like result. It remained to trace them to their real sources. This was not so easy as might be supposed, for though it was evident that Mr. Upcott must have had recourse to one or other of the existing editions, the interpolated passages in fact agree exactly with none of them. But a careful collation of the last four chapters of the book (which include more than half the interpolations, and may be taken as a fair specimen of the whole) with the old texts, leaves no doubt that, with the exception of the first thirty-six lines of chapter x, they were taken, like the two words mentioned above, from the first edition of Wynkyn de Worde, but with the spelling occasionally altered, and here and there a small word put in, left out, or changed. These alterations throw an ingenious disguise over the whole; but if we penetrate through this we find that in these four chapters there are only thirteen words differing from those in Wynkyn de Worde’s first edition, and these unimportant; while in his second edition, and in those of Copland and East, the variations from Mr. Upcott’s text of the same chapters are respectively fifty-seven, fifty-six, and fifty in number, and many of them important in kind: and if we go to the edition of 1634 we find the differences still greater, except as to those thirty-six lines, which are supplied from this edition, as they were wanting in the other copy. But the colophon, or concluding paragraph of the book, Mr. Upcott could not take from any of the editions which followed that of Caxton; for though Wynkyn de Worde might, and in fact did, supply at least one or two of the first words, the latter part of his colophon relates to his own edition, and departs widely from that of Caxton, while those in the later editions are still more unlike; and yet Mr. Upcott’s colophon is a tolerable, though not an exact, representation of that of Caxton. But his other materials can be ascertained beyond a doubt. They are, the colophon as given by Ames, and repeated by Dibdin in a modernised and