but included it in a more complete form than our existing Greek MSS. of Barlaam and Josaphat. The sections of the Armenian directed against the worship of water and of mere men also strike me as better put and arranged than the corresponding sections of the Greek text, though it must be owned that the latter, so far as it here varies from the Armenian, is in general endorsed by the Syriac form of the Apology, which is a document quite independent of the Book of Barlaam and Josaphat.
Now no single clue of this kind would be enough to establish the independence of the Armenian, and its priority to Boissonade's Greek text, in the eyes of anyone who considers how closely allied in general are these two forms of text, if contrasted with the non-Christian Arabic, and even with the Georgian. But an entire series of such clues occurring together forbids us to regard the Armenian as an abridgement of our existing Greek text. Yet abridgement of some text it certainly is, and of one very closely similar to Boissonade's Greek. This text we have seen was a Syriac one; and in this Syriac text, now lost, we must perforce recognise an earlier stage of the text than Boissonade's Greek supplies. The Syriac was probably much shorter and less padded out with disquisitions on all points of Christian faith and morals. But in it the apologues had already assumed the order in which they follow in the Armenian and Greek; and it must have contained everything which the Greek and Armenian have in common. This Syriac text was, I am inclined to think, itself a translation of a Greek text, now lost, which was subsequently expanded into Boissonade's text. Boissonade's text was elaborated out of this earlier and shorter Greek form, by some one who freely expanded it and substituted everywhere for its simpler phrases the rhetorical flourishes of a later age. If this be a correct view, our existing Greek text of the Apology of Aristides must be regarded with some suspicion. Such an apology had already of course a place in
