On the EigJith Verb-Class in Sanskrit. 5 



separated from it (cf. apna- : dpnas ' possession,' drdvina : 

 dravinas 'chattel'). As for I'nd 'debt,' its derivation is quite 

 uncertain, but if we must resort to ar ox arn, nothing better 

 can be suggested than the participial form r-nd ' hurt, bur- 

 dened, in-debted, (n.) debt.' Finally, not one of the kindred 

 words in the sister-tongues — and the number of such words 

 is quite considerable — shows any trace of a root-nasal, unless, 

 indeed, as suggested by some, it be opvu/jn, whose v, however, 

 is much more likely to belong to the suffix -vu (cf. wp-op-ov, 

 etc.). 



As regards gharii, which is said to mean 'shine,' no verb- 

 form that could be referred to such a root has been met with 

 in the extant literature. It might then be left out of consid- 

 eration here, 'were it not that a couple of nominal forms, 

 crJirnd 'heat' and s^h'rjii 'heat,' seem referable to the root 

 gJiani. As the native root-lists give also the root gJiar {ghr) 

 ' shine,' as belonging to the i-z/-class, it is evident, however, 

 that this root is the only acceptable form, and that gJiarn 

 sustains to ghar precisely the same relation as arii to ar 

 {gJir-nd, ghr-ni being quite regular formations). This sup- 

 position is decidedly favored by kindred words in related 

 tongues (cf. Zend gar-evia, Gr. Oep-o^at, Oep-ixi), dep-o^, etc. ; 

 \j>X. for-inns, etc. ; Goth, var-ni-jain ; SI'an. gr-e-ti, etc.). 



With reference to tani, said to mean 'graze,' it is, like 

 ghani, entirely unauthenticated ; and as there seems to be no 

 support for the acceptation of any similar root in any 

 other Indo-Germanic tongue, it may well be considered 

 as wholly fictitious, and invented to furnish an etymology 

 for the noun irna {trnd) 'grass.' This noun may possibly 

 be a participial form (parallel with trnd) of tar in the sense 

 of 'broken through,' — viz. the soil (cf. tinia-padi name of a 

 plant). 



We come next to the root ksin 'destroy,' which has no 

 more right to appear in the root-list than c?/-//, above. Even 

 here we find a shorter and well-authenticated root ksi 'de- 

 stroy ' inflected according to the j-//-class {ksj-no-nii, etc.). No 

 certain or authenticated example of a nasal is found either 



21 



