1838.] Revision of the Bactrian Alphabet. . 641 



the Chaldaic ; I find that in every case this letter may be best represented 

 by the Sanskrit 5rf j, and indeed in the early coins of Apollodotus, 

 &c. its form 3 seems to be copied from the ancient Sanskrit P, reversed 

 in conformity with the direction of the writing. The only inflection I 

 have met with of this letter is ^ ju. 



I can make no discrimination between cerebrals and dentals ; because 

 the Greek names translated have of course no such distinctions, but 

 from the variety of symbols to which the force of d and t must be 

 ascribed, I incline to think the alphabet is provided with a full comple- 

 ment, though it is in the first place indeed almost a matter of option 

 which letter to call d, t, r, or n, they are all so much alike — thus for 

 t we have 1, *1, ""l, and *i, and with the vowel i, \ % f . 



As the equivalent of d again we have the same % *l, % and also 

 i, £, f> : and for dhi \, and ^, the former evidently *1 with 1 subjoined ; 

 the latter quasi tti or ddi ; sometimes it is nearer S> ri. 



I do not attribute this ambiguity to the letters themselves so much as 

 to the carelessness and ignorance of the writers, who might pronounce 

 the foreign name Apollodotus, indifferently Apaldtada, Apaladata, 

 and even Apalanata, Being obliged to make a choice, I assume as in 

 my former paper ; — 



H, ^\, for ta, whence *1 td, *\ ti, j or 7 te, and 7\ tra f 



J, tta, tha, ^ thi, J, or ^, the, but in fact these forms are as com- 

 monly used for dh, and its inflections. 



t 5 *1, i, for da, nda; \^, di ; *]■ de, 2 du ; *?, dh, \ dhi. 



"L, i, na. I 4o not perceive any indications of the other nasals, and 

 indeed they seem to be omitted when joined to another consonant : but 

 I find some thing corresponding to the anuswara attached below the 

 vowel a, and before consonants it seems represented by m, as t mcha f 

 H, mri, %> mba ? 



J 1 pa. The first of the labials is one of the best established letters. It 

 has been discovered also inflected as h pi, P pe ; J 1 pu ; and united 

 with either h or s in Oi pha or spa; also with li in *b,pli, and in other 

 combinations which will be noticed as they are brought forward. I 

 suspect further that in vh, -J 1 , we have pa, and in t;, pra : but the data 

 are uncertain. 



¥ , If, pha or fa ? I have no stronger reasons than before for continuing 

 this value to If : — it seems in some few cases to usurp the place of v ; 

 it is inflected also, as ty fe, f fu, ^tfra. 



9. or ex, ba ? is still undetermined ; in the doubtful name above 

 quoted ABArA20T> it seems to be replaced by *1 or % — the aspirate is 

 also unknown. 

 4 L 2 



