enough to accomplish his task well, he was too good a scholar not to know that he had done it very ill. The text is broken up into paragraphs, the numbers of the modern chapters and verses being placed in the margin; the peculiar divisions of the Codex Vaticanus (_see_ p. 56) sometimes omitted, sometimes tampered with. The Greek type employed is not an imitation of the uncials in the manuscript (of which circumstance we do not complain), but has modern stops, breathings, accents, ι _subscript_, &c., as if the venerable document were written yesterday. As regards the orthography it is partially, and only partially, modernized; clauses or whole passages omitted in the manuscript are supplied from other sources, although the fact is duly notified(146); sometimes the readings of the first hand are put in the margin, while those of the second stand in the text, sometimes the contrary: in a word, the plan of the work exhibits all the faults such a performance well can have. Nor is the execution at all less objectionable. Although the five volumes were ten years in printing (1828-38), Mai devoted to their superintendence only his scanty spare hours, and even then worked so carelessly that after cancelling a hundred pages for their incurable want of exactness, he was reduced to the shift of making _manual_ corrections with moveable types, and projected huge tables of errata, which Vercellone has in some measure tried to supply. When once it is stated that the type was set up from the common Elzevir or from some other printed Greek Testament, the readings of the Codex itself being inserted as corrections, and the whole revised by means of an assistant who read the proof-sheets to the Cardinal while he inspected the manuscript; no one will look for accuracy from a method which could not possibly lead to it. Accordingly, when Mai’s text came to be compared with the collations of Bartolocci, of Mico, of Rulotta, and of Birch, or with the scattered readings which had been extracted by others, it was soon discovered that while this edition added very considerably to our knowledge of the Codex Vaticanus, and often enabled us to form a decision on its readings when the others were at variance; it was in its turn convicted by them of so many errors, oversights, and inconsistencies, that its single evidence could never be used with confidence, especially when it agreed with the commonly received Greek text. Immediately after the appearance of Mai’s expensive quartos, an octavo reprint of the New Testament was struck off at Leipsic for certain London booksellers, which proved but a hasty, slovenly, unscholarlike performance, and was put aside in 1859 by a cheap Roman edition in octavo, prepared, as was the quarto, by Mai, prefaced by another graceful and sensible epistle of Vercellone(147). This last edition was undertaken by the Cardinal, after sad experience had taught him the defects of his larger work, and he took