BRACKETT. — TEMPORAL CLAUSES IN HERODOTUS. 185 



action, the pure optative is generally used, and that the use of nev here 

 may be accounted for by the change from the plural to the singular. Cf. 

 further Od., 21, 293, and Monro, Horn. Gram., 283, c. 



But in our passage there is no element of contingency whatever ; all 

 the maidens, whether handsome or homely, necessarily at some time 

 became yaaav <opaiai. Herodotus uses as with the optative in general 

 past conditions four times, and okcos 51 times, and nowhere (except in 

 1, 196) is av used with the optative ; and it cannot be shown that there 

 is any more reason for an av here than in any of the 54 other cases. 

 There is, therefore, every reason to believe that the text here is corrupt. 

 Stein's suggestion gives a very apt meaning, but does not account well 

 for the MS. reading, us av at. 



The expression oaos 8r] is used by Herodotus in the sense of "a certain 

 number or quantity," as in 1, 160: ol 8e MvTiXijvaloi . . . iie8i86vai tov 

 IlaKTvrjv Trapeo-K6vd£ovTo enl uiada> oaco 8i], "they were preparing to give 

 up Paktyas for a certain price.'' oaos 8rj means in Herodotus, as 

 Schweighauser says, " quantitatem aliquam per se definitam sed quam 

 is qui loquitur exaete definire non potest aut non vult ; " 6'croy 8fj is 

 used in this sense also in 1, 157; 2, 103; 3, 52, 159; 4. 151. It 

 seems quite possible that Herodotus wrote here : ws Caai 8fj napOevoi 

 yivo'iaro yaiiw wpalai, " whenever a certain number of girls became of 

 marriageable age," etc., which number he was doubtless unable to state 

 exactly. This meaning is exactly appropriate here. For oaos 8f) placed 

 before its noun, cf. 4, 151 : atria KaTaXtnovres oawv Si) /j.t]vuiv. The reading 

 of the MSS. may then have arisen thus: first 817 (AH) may have been 

 changed to av (AN) ; that this mistake is a common one in MSS. is well 

 known.* Then ws oaai av was changed to w? av oaai, and later Ss- was 

 dropped. 



I pass now to consider the use of av in clauses of subsequence. The 

 statistics of the use of av with irplv have been given by Sturm in Schanz's 

 Beitrage, Bd. 1, pp. 296 ff. The subjunctive with -rvpiv occurs in clauses 

 of generic action six times (1, 140, 197, 198; 3, 109; 4, 117, 196); in 

 five cases all MSS. give av ; and in 3, 109, where the MSS. differ, av is 

 doubtless to be read, npiv with the subjunctive referring to future time 

 is found eight times; in three cases all MSS. .give av (1, 32. 159; 

 5, 106) ; in three cases all MSS. omit av (4, 157; 6, 82 ; 7. 220). but 7 ; 



* Cf. Gildersleeve, Syntax, 450; Voemel, Dem. Contiones, p. 228, and crit. 

 notes on De Clierson., 43, and Phil., 4, 12; Porson, Tracts and Miscellaneous 

 Criticism, London, 1826, p. 182; and especially, H. Richards, Class. Rev., 6 (1892), 

 pp. 340-342 



