of the MSS. The Arabic, as I learn from Professor Margoliouth, is literally 'and in so far as he does not introduce (or, there do not enter) into these compositions stories which resemble.' This version appears to deviate both from our text and from Dacier's conjecture. There is nothing here to correspond to συνήθεις of the MSS.; on the other hand, though συνθέσεις may in some form have appeared in the Greek original, it is not easy to reconstruct the text which the translation implies. Another conjecture, communicated privately to me by Mr. T. M'Vey, well deserves mention. It involves the simpler change of ὁμοίας το οἵας. The sense then is, 'and must not be like the ordinary histories' ; the demonstr. τοιούτους being sunk in οἵας, so that οἷαι ἱστορίαι αἱ συνήθεις becomes by attraction, οἵας ἱστορίας τὰς συνήθεις.
I subjoin a few other notes derived from correspondence with Professor Margoliouth:—
(a) Passages where the Arabic confirms the reading of the MSS. as against proposed emendation:—
iv. 14. 1449a 27, ἐκβαίνοντες τῆς λεκτικῆς ἁρμονίας: Arabic, 'when we depart from dialectic composition.' (The meaning, however, is obviousy misunderstood.)
vi. 18. 1450b 13, τῶν μὲν λόγων: Arabic, 'of the speech.' The μέν is not represented, but, owing to the Syriac form of that particle being identical with the Syriac for the preposition 'of,' it was
