124 Notes on Sanskrit Inscriptions from llathurd. [No. 2, 



Some of the inscriptions, as already stated, are dated, and the 

 figures of these dates are by far the most interesting, and at the 

 same time the most puzzling elements in their composition. Ge- 

 neral Cunningham, some time ago, commented on them at great 

 length in this Journal,* but without coming to a satisfactory con- 

 clusion. Nor can I congratulate myself upon having raised the ques- 

 tion much above the region of mere conjecture, though the conclu- 

 sions I have come to, appear to be much more probable and consistent. 

 After the decypherment of the dated inscriptions of Nasik by the 

 learned Dr. Bhau Daji,f the values of most of the figures must now be 

 accepted as settled ; but they cannot be read in the ordinary decimal 

 style, without producing very doubtful results, I propose, therefore, 

 to read them from the right in arithmetical series as numerical 

 notations without reference to their local values. This may, at 

 first sight, appear objectionable in a writing which proceeds from 

 left to right, but seeing that the Arabs and the Persians read their 

 figures, borrowed from the Hindus, from left to right, though their 

 writing proceeds from an opposite direction, it may be presumed 

 that the ancient Buddhists, who evidently took their figures from 

 the Aryan type, did not alter the original style of the figures 

 and wrote them from right to left. Hence it is that even in modern 

 chronograms, a rule is observed which says " figures, proceed 

 to the left." ^f^^T ^TWlf?f i Eaghunandana, the author of the 

 28 Tattvas, in his treatise on astrology, Jyotis Tattva, three 

 hundred years ago, quoted a s'loka to the effect that " in writing 

 many figures of one denomination the progress should be to the left." 

 ^5jT<ffaRWfi[TsraiT ^T*mT 3lf?n ; and to this day all chronograms 

 in Sanskrit are read in that way. Brown, in his Essay on Sanskrit 

 Prosody, notices the practice, though he does not quote any autho- 

 rity. Following this rule, the four figures of No. 1, (plate IV.) 

 may be read as 40 -|- 10 -f- 5 ~\- 4 = 59. Eeading from left to right 

 the result would be 4 -J- 5 -J- 10 -J- 40; which would be absurd as 

 progressing from small to large figures. If the third and the fourth 

 letters be taken for 9 and 6, and the whole be read decimally accord- 

 ing to their relative position, the date would be 4596, which would 



* Ante Vol, XXXI, p. 426. 



f Journal, Bombay Branch Royal Asiatic Society, Vol. VIII. p. 228. 



