BeJcJcer''s Digammated Text of Homer. 183 



igiofiai to guard nearly or quite half of them resist it. This is a very- 

 great difference, and cannot possibly be imputed to accident. I hold, 

 therefore, that Ahrens is fully justified in separating the two words 

 as he has done in his Grammar of 1852 : and I regard as highly prob- 

 able his conjecture that iovo^at began originally with a (compare the 

 Greek d conjunctive, whicli was originally sa), and that it is con- 

 nected with Latin servo. The primitive sense may have been that of 

 xoatching which we see in the compound observo, and from which we 

 readily derive the ideas of guarding and jyreserving. But whether it 

 once began with (x or not, Ave must in any case disapprove the pro- 

 cedure of Bekker in writing it, wherever he can, with digamma. 

 There ai"e in fact only 4 places out of more than 40 which give any 

 sign of an initial consonant. Two of these are in the 23d book of 

 the Odyssey (82, 229), which has in it much that is peculiar, while 

 the others are in a line that occurs twice (», 194, x, 444) : 



avTOv TiuQ vjjt ie fxivBiv xal vr^a sqvadav. 

 This shows another metrical irregularity, the short * of vi^t being used 

 for a long syllable, A2>parently it is only a variation of the perfectly 

 regular verse : 



airov ttuq v^saav {.livevv xal j/^ag tgvadai (|, 260, ?, 429). 

 the two plurals being changed to singulars, with little regard to met- 

 rical exactness. 



In speaking of iqiofiai, I have touched incidentally upon the ques- 

 tion whether digamma should be prefixed to the augmented forms of 

 digammate verbs. Wherever the augment makes a syllable by itself, 

 Bekker, no doubt w^ith correctness, writes the digamma after it : thus 

 Isuyri teas broken from -B&ywui, e-p&hj iras jyressed from -pei^o), sFsmov J 

 said, tense-stem Few from Fs-FSTt. But when the augment coalesces 

 with the root in the same syllable, he i:)laces digamma at the begin- 

 ning, and of course before the augment: thus he writes tbTSov I saw, 

 originally eviSov, Frivauae was i'lding, originally EFavaaae, F^vdavs was 

 pleasing, originally sFavSave. Now the temporal augment of y\vuaaB 

 must either have come from a stem which had ali'eady lost the digam- 

 ma ; or it must have arisen from a stem with digamma by dropping 

 that consonant between e and a with contraction of these vowels. 

 In either case the augmented form should be without digamma, which 

 could only appear by what must be regarded as an improbable trans- 

 position: BFavaaae, FBttvaaae, F^vaaoB ; bfiSop, fsCSov, FSidov. It is at 

 any rate a transposition which we should not accept without clear 

 indications in the Homeric verse. I must own that I have not looked 

 up the evidence myself on this point. But a writer in Jahn's Jahr- 



