BRACKETT. — TEMPORAL CLAUSES IN HERODOTUS. 187 



Sturm's change of irpiv ye brj to irpiv av in G, 82, has much less to 

 recommend it. His note (op. cit, p. 298) shows that he is aware that the 

 reading irpiv av does not explain the origin of the ye at all. The conjec- 

 ture of Luth on Ant., 1, 29, by which Sturm tries to support his pro- 

 posal, is not approved by Blass, and not mentioned by Jernstedt. The 

 meaning of ye Sy here is not less appropriate than with the indicative in 



6, 79 : ovk wpoiv . . . o ti £Trpr)(T<Tov, it plv y e 8 rj civTiov tis dvaftcis eiri 8ev8pos 

 Karei8e to noievpevov. Cf. also ye 817 with the optative, 2, 1 1 : kov ye 8rj ev 

 T&3 npoavaidipuijxevw xpovco nporepov *j ipe yeveadai, ovk av xcocrfleiT] koXttos, k.t.X. 



The MS. reading in the passage under discussion should be retained. 



is o is used with the subjunctive in Herodotus nine times, — three 

 times of future action (2, 115; 4, 43; 8, 108), six times of generic 

 action (1, 93; 3, 31, 82; 4, 30, 19G; 6, 8G y). av is everywhere used 

 with the subjunctive except in 3, 31 : ol 8i fiao-iXfjioi 8tKao-Ta\ KtKptpevoi 



av8pes ylvovTiu JJepo-eav, is b dnoddvcocri, K.r.X. ; and 8, 108: [Evpij3ia8r]s 8e 

 ttjv evavrirjv yvaprjv iTi8eToj eariov Z>v eivai tyevyeiv es b e\6rj (fievyav is ttjv 

 e<i)VTov. 



Let us first consider the passage in 8, 108. The MSS. read here as 

 follows : e\0j}, PRs, and all editors except van Herwerden ; e\6oi, ABC, 

 and van Herwerden.* The omission of av with is is a different matter 

 from its omission with other clauses meaning " until," and for this reason : 

 outside of Herodotus, is o occurs as a temporal conjunction in classical 

 Greek only in Homer (50 times), Hesiod (1), and Horn. Hymns (1), 

 and in each of these cases it is accompanied by Ke'.f The use of is S in 

 Herodotus seems to be a literary convention. If now the usage whereby 

 <e is required has become so firmly settled in the time of the composition 

 of the Iliad and the Odyssey that it was never omitted, it seems strange, 

 to say the least, if Herodotus was able to omit it at will. It can of 

 course be said that av may have been omitted with is S on the analogy 

 of axpi ov (1, 117), pexpt (4, 119), and possihly vplv (see above); but 

 inasmuch as av is used with is o much oftener than it is omitted with the 

 conjunctions just mentioned, this does not seem likely. 



As regards the reading eXdoi, it gives a wholly appropriate meaning, 

 and is grammatically above reproach, representing is 6 av e"\8r) of the 

 oratio recta. To be sure, as we have seen above, $ Herodotus does not 

 often change the subjunctive to the optative in oratio obliqua, but he 

 does do this in 5, 63 (okws ehdoiev), and there is no cogent reason why 



* Stein in his school edition (5th ed., 1893) reads «A0oi. 



t Cf. Fuclis, op. cit., pp. 129-30. J See p. 175. 



