200 PROCEEDINGS OF THE AMERICAN ACADEMY. 



pies stated above (p. 173). Two passages only demand notice. 1, 164 : 



ev id wo "Apnayos dno rov rei'^f oy dnrjyayf rfjv (TTpaTifjV, ol ^to/ccuee? iv tovtw 



. . . fnXeov eVt Xiov. The use of eV w with the aorist is strange, and yet 

 no editor seems to have commented upon it. The phrase iv £, used in 

 a temporal clause, everywhere else in Herodotus means " while," and we 

 should naturally expect the meaning here to be, " While Harpagus was 

 leadinc away his army," etc. But it is impossible in the light of Herodo- 

 tus's usage elsewhere to believe that the action of the imperfect en\eov is 

 inserted in that of the aorist dnrjyaye. Nothing like it is found elsewhere 

 in Herodotus. The explanation is this. By a confusion of ideas — 

 another example of Herodotus's " genial Jaissez aller " ■ — iv <a was used 

 with a verb with which properly a conjunction meaning "after" should 

 have been used, iv a> stands here therefore in a clause which is strictly 

 one of prior action. Cf. Time, 7, 29, 4 : t6 yap yevos to roe QpaKwv . . .eV 

 w av 6apcrr)<TT), (poiviKwraTov eariv. The only other explanation that can be 

 given is that the true reading here is dniiye, but this is not probable. 



In Herodotus, 3, 150: iv oo-co yap o re pdyos rjpx e KaL ol eVra inavio-rrivav, 

 iv TOVTa iravTi to> xpoixa nal rfj Tapani] ey rfjv TroXinpKirjv TraptcrKevddaTO, It IS 



clear that we have an example of coincident extension. As to napeo-Kev- 

 ddaro, MSS. and editors read thus : Trapeo-Kevddaro, CP, Kr., Ab., vH., Kal. ; 

 7r;ip«TK(:vd{ovTo, ABR, St., H., TrapeaKevdcravTOi d z. The pluperfect is here 

 impossible. If we keep the pluperfect, the meaning must be that during 

 all the time indicated by the temporal clause the preparations were already 

 completed. But we know from the context that during the rule of the 

 magus and the insurrection of the seven the prepai'ations were going on. 

 Stein and Holder are undoubtedly right in adopting the imperfect. Bahr's 

 citations (5, 34, and 7, 218) have no bearing on this question, except as 

 to the possibility of the form Tvapto-KevdbaTo. 



Sometimes in clauses which formally are temporal clauses of contem- 

 poraneity the temporal relation is loose and inexact, and cannot with 

 accuracy be called either that of coincidence or that of insertion. Such 

 a case is found in 6, 10G : Tore Se nepcjidfls imo tS>v <jTpaTr]ya>v 6 <J?ei8t7T7ri'fi»7? 

 ovtos, ore nip ol (<pi] Kal rov Yldva (pavijvai, k.t.X. It is obvious that the tem- 

 poral clause defines the time of Tr(p<p8fis ; but it is not less obvious that 

 the action neither of e'qbr] nor (pavijvai is either coincident with that of the 

 main verb or inserted in it; neither is the temporal clause one of ante- 

 cedence. The fact is that by stating in the temporal clause a fact pre- 

 sumably well known to his hearers, the author thus identities the occasion 

 upon which the main action took place, ore here is equivalent to "on the 

 decision when." A similar usage is found with 6Ve in 6, 65, and 2, 156. 



