Behker'^s Bigammated Text of Homer. 181 



doubt and for difference of opinion. A striking instance of this kind 

 is found in tlie word lAw^^)?'^//, wliicli occurs 8 times, with tlie con- 

 nected Ehx'fXK and ilihQiu, eacli of which occurs once (the last in A^ 4). 

 In 3 of these 10 passages (P, 667, ^, 4, £■, 684) tliere is a hiatus before 

 the \vord, but it is at the feminine caesura of the tliird foot, where hi- 

 atus has scarcely any Aveight in proving an initial consonant. In 5 

 passages (£■, 488, -P, 151, y, 271, e, 473, o», 292), a v movable precedes, 

 which neither hinders \is from assuming digamma, nor furnishes any 

 proof of its existence. The remaining two passages give evidence of 

 an initial vowel. In one, naxqixloio d' thiga (2", 93), the e of ^^ is elided 

 before sKw^ : in the other, jn^ ncag /joi ihiQ (j/, 208), the diphthong of (Uot. 

 is made short. In the fii'st case, Bekker gets rid of the difficulty by 

 reading naiooxlov ds Fihoga : in the second, he yields to the difficulty, 

 and writes i'Aow without digamma. We cannot here lay much stress 

 on the proportion (20 per cent) of unconformable passages, the whole 

 number being so small. But as digamma is not required in any single 

 passage of the 10, and is excluded by 2, it is certainly hazardous to 

 assume it without other proof of its existence. Such proof one might 

 perhaps find in its derivation. It is natural to take it from the root 

 which appears in the second aorist of algio)^ infin. Uer*', indie. eUov^ 

 where the augment affords evidence of an original consonant initial. 

 If sllor is for eFelov, IXelv for Fe^etv, we might connect them with Latin 

 vello, to pluck. But the 2d aor. of algiw is never written Avith digam- 

 ma by Bekker, and it is quite clear that it w^as not digammated in the 

 Homeric language. It furnishes therefore a very feeble presumption 

 for the digamma of the siibstantives ; and we cannot but conclude 

 that it would have been the safer and wiser course to leave the sub- 

 stantives also without digamma. 



Another word in which we must question the propriety of the di- 

 gamma that Bekker gives it, is the deponent verb igvoijai to watch, 

 guard, jyreserve. This verb in many of its forms is apparently identi- 

 cal with egvo) to draio ; and it has been assumed almost universally 

 that they come from the same root. Buttmann in his Lexilogus ar- 

 gues the question at length, maintaining their essential identity. 

 Apart from the indications of a digamma, there are other reasons for 

 separating the two verbs. Thus, as to form, ig<m to draw shows e* 

 only where it would arise from augment or reduplication; while hgio- 

 fuat to guard shows slqiaaovxav in the future, elovaaaadai, in the aorist 

 infinitive, and other like forms. Again, igvouav to guard is sometimes 

 inflected according to the ,Mt-form, as in sqvto, sXqwxo, etc., which is 

 never the case with igi^a to draw. And yet again, with igvoftai to 

 guard, there is a verb ^iofxat, with initial g, which has the same mean- 



