XC THUCYDIDES jectural restoration of the missing portions. It should be observed however that the inscription appears to have been vi^ritten o-Totxq'^ov, i. e. in equal lines, so that each letter fills up the same space ; hence it can in some places be restored with tolerable certainty. The six variations between the existing part of the inscription and the text of Thucydides are as follows. In p. 346, 1. 16 (Bekker) the inscription inserts Trpos aXXrj- Aous after 'HAeiot. In 11. 25, 27, the inscription has [Majvra/eas KoX 'HA[etovs] and Mavrn/eas /cat for Kai Mairtveas, showing that the names of these two contracting powers were trans- posed; and similarly in p. 346, 1. 36, and 347, 1. i, t^v 'Apyetiiiv r] for rj ttjv 'Apycttov. In p. 347, 1. 20, it has -VT] rfj cr-, showing that some other word than rrjv rjyeixoviav followed fXiTaTrefxij/afxevr] [rfj arpaTLo. ^ycrOM yjye/xovevovaa, K.), and in L 21 -IS Tais [TToAecri], showing that some other word than S6$rj (d;raa-ats, K.) preceded rats. These differences are very slight. On the other hand, it may be remarked that they occur in a fragment which amounts only to about a twelfth of the whole treaty. Assuming for the moment the correctness of Kirchhoff's conjectures, we may compare the text of Thucydides with the whole treaty as restored by him. Of thirty-one varia- tions (see Classen, Intr. to Thuc. Bk. viii. p. xxv ff.) thirteen are merely orthographical {idv for i]v, or ^aAaTxav for OaXaa- aav). In three cases the order of the names 'Eleans, Mantineans, Argives,' is different In four cases the in- scription inserts or repeats, probably for the sake of clear- ness, words which do not occur in our text. On the other hand, our text inserts rais TroAecrtj/ after SoKy in p. 347, 1. 7. In eight cases the variations occur in places where the restoration has little or no ground on which to rest. If we set these aside, the variations reduce themselves to two more or less probable conjectures, eVl rrjv yrjv for h ttjv yrjv (p. 346, 1. 36), and wv dpxova-L for wv av apxioa-i. (p. 347, 1. 9). '■ It varies in different parts of our present text.
