A minute collation of all Erasmus’ editions is a desideratum we may one day come to see supplied. The present writer hopes soon to publish a full comparison of his first and second editions with the Complutensian text(167), as also with that of Stephen 1550, of Beza 1565, and of Elzevir 1624. All who have followed Mill over any portion of the vast field he endeavoured to occupy, will feel certain that his statements respecting their divergences are much below the truth: such as they are, we repeat them for want of more accurate information. He estimates that Erasmus’ second edition contains 330 changes from the first for the better, seventy for the worse (N. T., Proleg. § 1134); that the third differs from the second in 118 places (_ibid._ § 1138)(168); the fourth from the third in 106 or 113 places, ninety being those from the Apocalypse just spoken of (_ibid._ § 1141)(169). The fifth he alleges to differ from the fourth only four times, so far as he noticed (_ibid._ § 1150): but we meet with as many variations in St. James’ Epistle alone(170). 3. In 1518 appeared the Graeca Biblia at Venice, from the celebrated press of Aldus: the work professes to be grounded on a collation of many most ancient copies(171). However true this must be with regard to the Old Testament, which was now published in Greek for the first time, Aldus follows the first edition of Erasmus so closely in the New as to reproduce his very errors of the press (Mill, N. T., Proleg. § 1122), even those which Oecolampadius had corrected in the list of errata; though Aldus is stated to differ from Erasmus in about 200 places, for the better or worse(172). If this edition was really revised by means of manuscripts (Cod. 131) rather than by mere conjecture, we know not what they were, or how far intelligently employed. Another edition out of the many which now began to swarm, wherein the testimony of manuscripts is believed to have been followed, is that of Simon Colinaeus, Paris, 1534, in which the text is an eclectic mixture of the Complutensian and Erasmian(173). Mill states (Proleg. § 1144) that in about 150 places Colinaeus deserts them both, and that his variations are usually supported by the evidence of known codices (Evan. 119, 120 at Paris, and Steph. ια᾽, i.e. Act. 8, Paul. 10, have been suggested), though a few still remain which may perhaps be deemed conjectural. Wetstein (N. T., Proleg. vol. i. p. 142) thinks that for Bogard’s Paris edition of 1543 with various readings Evan. 120 or Steph. ιδ᾽ might have been used, but his own references hardly favour that notion.