122 Professor Bi'iiLER on the Sanskrit Linguals. 



become stringent in the Prakrit dialects, an r r' immediate- 

 ly preceding an 'n' must be assimilated to the latter. In 

 the Prakrit dialects of the dramas an nn thus produced 

 remains unchanged, e. g. kanno ' the ear' for karnah. In Sans- 

 krit, however, as well as in some of the vernaculars, the 

 first n is dropped and a preceding short vowel may be length- 

 ened. Thus we find in the Rigveda dunaca ' imperishable' 

 for durnaca, and dunaca, ' difficult to obtain' for dur- 

 naca. In both these cases either form actually occurs. Ani 

 ' the pin of the axle of the wheel,' which I believe to be de- 

 rived from the gunated form of the root 'ri,' ' ar,' by the 

 affix 'ni,' as well as pani ' hand,' are other instances of this 

 change, common to the Vedic and classic Sanskrit.* The lat- 

 ter word ' p&ni' seems to stand for ' parni' and to be derived 

 from the root ' pri,' ' to transact business, act' just as kara 

 'hand' from kri 'to do'. More frequently however the first 

 n of the group nn deduced from 'in' is simply dropped, as 

 in the roots ghun, cun, cun, pun, pun for ghurn, curn and 

 purn. The u of the roots pum etc. is originally short and 

 is lengthened in consequence of a special rule in Sanskrit. 

 This accounts for the double forms c'un and c'un. The n of 

 ani, a rarer form for ani, of anu 'small,' pan 'to buy and sell' 

 seems to owe its origin to the same source. 



Anu, I think, is derived from an Indo-European root, 

 'ar 'to hurt,' 'to grind,' from which the Sanskrit irma, 

 arus ' wound' are derived and which in Greek is re- 

 presented by ' af in aXe'w to grind, aXtvpov ' flour.' The 

 etymological meaning of anu (arnu) would therefore be 

 ' diminished by grinding,' (compare the Greek repev from 

 T€tpw=Latin tero.) Regarding pan it seems to be derived 

 from an Indo-European root par, which is preserved in the 

 Sanskrit pri ' to transact business' and in the Greek 



* Many of f lic following conjectures have already been brought forward 

 by other philologists, especially bv Ben fey and Boethlin?k and Roth. 

 Bee Beirfey'aVollBt. Ski. gram. sb< 121, ff. B.&R's Skt. dictionary, 5. w. 



