^{page 75 note 1} The only cases I know which point to a variant in the archetype are 2, 63, 1 (see p. 78); 1, 24, 3 δύναμις CG—δύναμις πλις EM—πλις ABF; and 2, 21, 3 ŵργητο CEG and ὣρμητο ABFM, both recognized by POxy 853. But though the archetype at 1, 24, 3 might have had δύναμις πбλις or πбλις δύναμις it might equally well have had δύναμις πбλις; and as for 2, 21, 3, óργ· and δρμ- are everlastingly permuted, cf. 1, 140 1 (recc.); 3, 82 I (recc.); etc.

^{page 75 note 2} Only at 8, 44, 2 -ν Ανδου BCG)(-νΔινλου AEFM do the variants appear due to uncial; if so, they must have been already in the archetype. Ζενξλας)(Ζενξιλας 55, 24, 1 is not necessarily uncial.

^{page 75 note 3} Three alleged agreements in error between the fifth-century Fayum parchment and CG are produced by Pasquali (op. inf. cit. 326) and Bartoletti 55. Two concern matters where scribes continually vacillate: βουλεύετο)(βούλετο and ταρους) (τρους, and the third rests upon misrepresentation: it is claimed that Fay. omits το 8, 92, 2 wrongly with CG, whereas the actual position is: νομα το κελεύσαντος ABEFM and, from calculation of space, perhaps Fay.; νομα κελεύσαντος G; kappa;ελεύσαντος ỏνματος C.

^{page 76 note 1} 3, 63, 2 ύπρχον γε M; (ύπρχον τε cett. is irrelevant, as confusion of γε and τε occurs at all periods.

^{page 76 note 2} A list of them will be found in the conspectus sigiorum, and their readings in the apparatus, of my forthcoming second edition of the Oxford text.

^{page 76 note 3} Ignoring, with Bartoletti, those erroneously adduced by Pasquali or uncertainly inferred from considerations of space.

^{page 77 note 1} Some editors have argued the falsity of αι σπονδαι from its being unknown to ‘the scholiast’. But the ‘scholium’ in question (αι σπονδαι δηγοντι) is an ‘Aldine scholium’ (see C.Q. 1936, 146 ff.). That is to say, it is not a scholium at all, but a variant in Parisinus svppl. gr. 256, inserted by the second hand, which was collating with a B-text.

^{page 78 note 1} I ignore 2, 35, 2 αύτύν)(αύτύν; for the of Dionysius vary and MS evidence on such a point is worthless. Nor has the omission of αύτῷ after τῷ at 7, 71, 4 by Dion, and AEFM any significance because of the easy homoeoteleuton.

^{page 79 note 1} Carat Thueydidtae, Ind. Schol. Gött., 1885, 6f.; Berliner Sitz. Ber., 1915 608, n. 3; and frequently elsewhere.