Bekk€r''s Digammated Text of Homer. 1 85 



counted right, and the unconformable cases only 10, or about 8 per 

 cent. The perfect of ce»'^(i»'w (root FaJ) occurs but twice (/, 173, a, 422). 

 Bekker both times writes naaiv iFadura. We hohl that he should have 

 written na<n rsFotddTa : the pi'esumption is that the digamma was reg- 

 ularly repeated in the reduplication, as Bekker gives it in Ft'Fotxa. 

 The perfect middle of edo) (root peA) occurs four times, twice after v 

 movable, and twice after a hiatus, which however is at the feminine 

 caesura of the third foot. Bekker everywhere writes eT^\ue6a, hTe)^fii- 

 vog : we hold, as before, that he should have w^ritten FfFt'A.Mci^ce, fcfcV^- 

 vog. The next case to be considered — that of Uno/jai (root Fsln) to 

 hope — is attended Avith more difficulty. The perfect £oA7ra and plu- 

 perfect lialnBiv occur twelve times in all : 3 times with hiatus, 4 times 

 with V movable before them : therefore 7 times where digamma is 

 admissible ; leaving 5 cases which resist it. This large proportion 

 of unconformable cases might make us doubt whether we ought to 

 recognize digamma at all in these forms. But the f of the root is 

 unquestionable, and gives a strong presumption for F in the redupli- 

 cation. And besides, the three cases of hiatus occur in a part of the 

 verse (at fem. caes. of 2d foot) where hiatus is inadmissible. We 

 hold therefore that there is sufficient evidence of Homeric Yt-eolna and 

 ve-pdiXneiv (or peFoATren'), and that these forms should have been given, 

 according to Bekker's principle, wherever the verse allows them. He 

 has in fact given them only in the 3 cases of hiatus, while in the 4 of 



V movable he retains that letter and writes svolna, iFaknstv ; thus con- 

 travening both his general method and his procedure in the parallel 

 case of socxa, iaxeiv. In sooya, itooyscv, ^\Q find very much the same 

 state of things — 12 passages in all, of which 5 resist digamma. The 



V movable, however, occurs here in only one case (I, 289) : 



TowxTJj,, o; Jij 710^0! xdx' OLvOqihTioiaiv ifhgyei. 

 Here, from the analogy of his procedure in reference to t(!>\neiv, we 

 may presume that Bekker would have written ^vOqibnoiaiv iFd^yeiv, if 

 he had not followed Voss in making a greater change, altering the 

 dative to an accusative in accordance Avith the usual construction of 

 the verb, making Cipdqimovg elpwoyeiv. It might be questioned, how- 

 ever, Avhether we ought not in this case to have i/Ff^^j^etj' for efsfoqysiv, 

 in the same manner as ^jptxro for sfstixto. In the perfect middle of 

 I'^yw (root Tsoy) to shut, we find a different state of things. Here we 

 have LQxajai, and b^xuto occurring 7 times. They are evidently forms 

 without reduplication, like oida I hioio (i. e. Totda, not repot Ja), sT/uai 

 am clothed (i. e. reoftat, not peFeor.Mwt), and in stems beginning with 

 other letters, de/uTat have received (for deds/ntai)^ araiycc I command 

 (for TJvwj'a), Hence, Avhen we find ie^yuitai and iiQ/axo occurring each 

 24 



