204 PROCEEDINGS OF THE AMERICAN ACADEMY. 



We now pass to clauses of subsequence which express limit. In these 

 clauses the main verb is generally of extensive aspect. First, as to 

 clauses witb irpiv and trplu fj, followed by a finite verb. In all these sen- 

 tences the main verb is formally or virtually negative. This negative 

 introduces a disturbing factor in determining what tense stem shall be 

 used, and as a result no general law can be laid down, In the clauses 

 of this kind which refer to past time the imperfect is found four times 

 (1, 14; 6, 79, 211 ; 7, 239), and the aorist three times (6, 45 ; 7, 137 ; 

 8, 8). In the clauses of limit introduced by ea>s, p^xP 1 - p-'XP 1 °^ "XP 1 °^ 

 a verb of extensive action is used in the main clause in every case but 



one. 5, 28 : [rj Mi'X^tosj Karinrepde 8e toiitcov iir\ 8vo yeveas dvbpav voarjcraaa 

 es tu fxaXicrTci crracri, pexp 1 °^ P iv Ilapioi KarrjpTicrav. Here the temporal 

 clause depends upon and limits the participle poija-aaa ; but that the action 

 of voijaaa-a extended over a considerable time is shown by the eVi Svo 

 y eve lis dvBpwv. 



In clauses with es o where a choice between the imperfect and aorist 

 was possible the aorist is found in the main clause only eight times "}" (1, 

 158; 2, 143, 143, ; 4, 105, 1G0, 201 ; 6, 24, 83). In four of these 

 places, 6, 83 (e(rx ov . . . apxovres) ', 2, 143 (dnebfiKvvcrav . . . 8ie£lovres) ', 

 2. 1 4o 2 (dvTeyeverjkoyrjcrav ■ . . (pdpevoi) ; 1, 158 (eV^e . . . dotcecav K(n aTvta- 



ricav), though the verb is aorist, the notion of extensive action is conveyed 

 by the present participle. In 4, 160: '6s (3a<ri\ev<Tas npara rola-i eavroi 



d8e\(peoicri e'aracriacre, es o piv ovrot dnoTuTTOi'Tes o'i^ovto, k.t.X., WC should 



rather expect the imperfect, but the expression is condensed and almost 

 elliptical ; the meaning fully expressed would be, " he first raised a 

 quarrel with his own brothers, and this continued until" etc. Another 

 explanation would be that es is used here in a loose sense, meaning 

 little more than " and finally." Cf. above, p. 178. 4, 105: o'0uis ydp rr(pi 



7ToXXous pev T) x^P 1 ! dvtcpatve. 01 8e nXevves avu>6ev cr(pi en tu>v eprjp.Gov eire- 



irecrov, is oitie^opevoi o'lKrjaav pera BovSivav, k r.X. Here the temporal clause 

 refers both to dve(paive and to eTvenecrov, but it cannot be denied that the 

 temporal relation is loose and of secondary importance. The aorists in 

 the main clause, 4, 201 (vm)Kovaav) and 6, 24 (n^iVeTo) are correctly ex- 

 plained by Fuchs (op. cit., p. 70). 



The future tense stem is found in the main clause, 7, 158 : eor av 8m- 



* Hence correct the statement of Fuchs (Schanz's Beitrage, Heft 14, p. 78), 

 who says that in clauses with pexp 1 °" nl Herodotus the imperfect is always found 

 in the main clause. 



t Fuchs (op. cit., p. 70) seems to be in error in saying that the aorist occurs five 

 times in these circumstances. 



