made violent changes; on the contrary, he seldom or never 'emended' by mere conjecture, and, though he marked many lines as spurious, he did not omit them. The greatest divergences which we find between Aristarchus and the vulgate are not so great as those between the quartos and the folios of Hamlet.
Yet we can see that he had before him a good many recensions which differed both from the vulgate and from one another. He mentions in especial three classes of such MSS.—those of individuals, showing the recension or notes of poets like Antimachus and Rhiânus, or of scholars like Zenodotus; those of cities, coming from Marseilles, Chios, Argos, Sinôpe, and in general from all places except Athens, the city of the vulgate; and, lastly, what he calls the 'vulgar' or 'popular' or 'more careless' texts, among which we may safely reckon 'that of the many verses' (ἡ πολύστιχος).
The quotations from Homer in pre-Alexandrian writers enable us to appreciate both the extent and the limits of this variation. They show us first that even in Athens the vulgate had not established itself firmly before the year 300 B.C. Æschines the orator, a man of much culture, not only asserts that the phrase φήμη δ'ἐς στρατὸν ἦλθε occurs 'several times in the Iliad,' whereas in our texts it does not occur at all; but quotes verbally passages from Θ and Ψ with whole lines quite different. And the third-century papyri bear the same testimony, notably the fragment of A in the Flinders-Petrie collection published in 1891 by Prof. Mahaffy, and the longer piece from the same book published by M. Nicole in the Revue de Philologie, 1894. The former of these, for instance, contains the beginnings or endings of thirty-eight lines of Λ between 502 and 537. It omits one of our lines; con-
