Bekker's Digammated Text of Homer. 187 



would take the more primitive formation seen in Uoixa, with its inter- 

 change of tx, eiK, oix. For such reasons the rise of eoixa from a root 

 dik seems to me scarcely more than a possibility. It must be observed 

 too that unless the connection of ^oixa with a root dlk is rendered 

 probable, there is no more reason for writing it with y than with f. 

 And if 710 7)} ore reason, we may justly say that there is less reason for 

 y than for f. Because the very regularity with which this word in 

 the Homeric text gives evidence of a consonant initial, is a circum- 

 stance in favor of F. In this respect it ranks, as we have seen, among 

 the most regular words, the unconformable cases being only 10 in 

 126, or about 8 per cent. If we consider how much earlier and more 

 complete was the disappearance of y from the language when com- 

 pared with that of f, we shall be slow to believe that the former 

 should maintain itself in the Homeric verse with the same constancy 

 as the latter. 



Another case in which Curtius recognizes traces of an initial y in 

 Homer, is the deponent le/nai to be eager, to desire, to long. It occurs 

 in 61 instances, of which 22 by hiatus give evidence of a consonant 

 initial, and the unconformable cases are only 3, or about 5 per cent. 

 Bekker writes it Avith digamma. To this Curtius objects, asserting 

 that the verb t'/.w* is a reduplicated form of the root yd, which appears 

 in Sanskrit, and is itself an extended form of the root i, to go. Thus 

 lr},in z= yi-ya-mi, to cause to go, to send. In the middle this would 

 mean to send oneself, and hence to hasten, to pursue eagerly, to aim 

 at, to long /or. To this no objection can be made as regards the 

 meaning. But it is a remarkable circumstance that "';,«« to send shows 

 no traces in Homer of anything but a vowel initial. Of simple forms 

 in the present and imperfect — these I take for comparison because 

 Ufiat, to desire is confined to those tenses — I find in the sense of send- 

 ing 29 (all active except ^, 11, M, 274, x-, 304, which show the middle 

 or passive) ; and of this 29, not less than 24, or more than 80 per 

 cent, refuse to admit a consonant initial. Compound forms, such as 

 (kcplijfii, /xe6li]iii, etc., I have not taken into the account : I believe, how- 

 ever, that all of them w^hich are capable of furnishing any evidence 

 on the point, testify against a consonant initial for the simple verb. 

 It may be regarded as perfectly certain that ^/^t to setid was sounded 

 by Homer with an initial vowel, and "e/nai to desire Avitli an initial 

 consonant. We have here a distinction of the same kind as that 

 which we before proved to exist betAveen iQvw to draw and inio/uai, to 

 guard. Now it might be said by one who maintained the original 

 identity of i)?."* to send and Isnai, to desire, that the y, which once be- 

 longed to both alike, was retained, and that with uniformity, in the 



