172 Dr. Mitra— On JEkotlbhdrn. [July, 



There is no positive rule for the treatment of two or more simple 

 non-initial letters in the same word, but according to usage the two 

 letters are sometimes subjected to the two optional rules successively, 

 and sometimes to the same rule. Accordingly the Sanskrit word 

 TJpadesa, is in some places written as Ubadesa, at others Uadesa or 

 Uaesa. It is obvious that a rule acknowledged to be optional and go- 

 verned by so many conditions cannot be accepted as proof in any 

 particular case without an amount of information which it is hopeless 

 to expect in regard to so obscure a term as the hypothetical eka-uti. 

 There is nothing to show that Mr. Growse is in possession of such in- 

 formation. 



But to proceed a step further, and bring the rule with its governing 

 conditions to bear upon the term. Mr. Growse holds that ' the elision 

 of the syllable ka (he means the letter k) would no doubt be an ano- 

 maly,' but I see no reason why such should be the case. That k is 

 1 simple and non-initial,' and legitimately comes under the purview of 

 the rule, and Prakrit writers have always acted upon this belief. Va- 

 raruchi, in his grammar has cited earn as the equivalent for the Sanskrit 

 ekam, and in the ' Venisanhara ' I find eai for ekdki. In ' S'akuntala ' we 

 have eaini for ekdkini. In the ' Prakrita-Prakasa ' edraha stands for 

 ekddasa. In other works egdralia is also met with, shewing that the 

 second rule is that which is in operation. When neither of the optional 

 rules is accepted the eka remains unchanged ; and we find ekadara for 

 ekatara in the ' Malavika-agnimitra.' And with such evidence before 

 me I fail to perceive why it would be ' an anomaly ' to elide the k 

 of eka. 



Making the elision, the first form of the hypothetical word would be 

 eci-koti. Applying the first rule to the second k also we get ea-oti. If 

 the rule be limited to the second member of the compound, the re- 

 sult becomes eka-oti. The second rule would give us egakoti, ega-oti, 

 eka-goti or e'dgoti. There is no rule in any Prakrit or Sanskrit grammar 

 under which the k of koti can be elided and its associate vowel o carried 

 over to the syllable ka of eka and produce ekoti, sandhi being, as 

 already stated, absolutely prohibited. Were it otherwise, the union 

 of o with the a of eka would make ekauti and not ekoti. Thus under no 

 circumstance can the rule appealed to, produce ekoti. The rule in fact 

 is quite fatal to the validity of the conjecture set forth. 



In the above remarks I have confined myself to the effect of the 

 rule cited by Mr. Growse. I should add here that that gentleman has 

 quoted no rules to show how the cerebral t of koti is metamorphosed into 

 the dental letter. As far as I am aware of, there is no rule in any Sans- 

 krit or Prakrit grammar to provide for the change. Under these cir- 



