1855.] Ancient Indian Numerals. 557 



Rev. J. Stevenson.* Among these records are to be found no less 

 than twenty-eight figures or combinations of figures, usually append- 

 ed to the written exposition of the given value defined at length in 

 the body of the text ;t the lower numbers are sufficiently simple and 



144 B. C. the Indians understood the art of the definition of sums by the sequent 

 arrangement of the ten units, each of which acquired value from its place in the 

 general total. His efforts seem to have been confined to the ascertainment of the 

 limited functions of each figure or the derivation of its normal type. 



Hence in noticing Dr. Stevenson's very valuable contributions to our knowledge 

 of the subject he remarks [note p, 704 J. A. S. B. 1854] Dr. Stevenson, in Bom- 

 bay Journal, vol. V. p. 38, found "a striking resemblance between the character 

 denoting a thousand and the Bactrian S reversed," " and after an examination of 

 the rest he" " thought it exceedingly probable that they were all derived from that 

 source." 



Major Cunningham thereupon proceeds to congratulate himself on the result, 

 that " our independent deductions are the more satisfactory as they were obtained 

 from different sources." 



It is certainly singular, that while acknowledging the correctness of Dr. Steven- 

 son's attributions and even pressing it into support of his own argument, the 

 writer did not perceive that the very admission of the conclusiveness of the one 

 determination necessarily compromised the other, the simple concession that a 

 separate and independent character of the Bactrian alphabet was borrowed, as 

 carrying that value to express the sum of one thousand, was directly opposed to 

 the notion that the unit literal cypher of that system of writing were adequate to 

 acquire value from relative position, as his published inscriptions purported to 

 prove ; if the science of numeration had advanced so far as to determine that the 

 act of locating the unit 1 in the fourth place of a line of figures sufficed to repre- 

 sent one- thousand, what need was there of complicating the operation by the use 

 of a special and separate cypher to define the requisite amount ? It is possible 

 that Major Cunningham understood that the adoption of Arian letters into the 

 Gujrat alphabets, as exponents of given numbers, took place prior to the elabor- 

 ation of the system of which he illustrates, but from the^tenor of his observations 

 it would certainly appear that he had lost sight of the difficulty above suggested. 



* J. A. S. B. 1854, Note p. 407. Journ. Bombay branch R. A. S. July, 1853, 

 p. 35. 



f I could have desired that the facsimiles of these inscriptions should have been 

 more calculated to command our faith in their exact rendering of the originals, 

 but I observe that Dr. Stevenson himself does not place any very great reliance 

 upon the transcripts, as he remarks, " I trust also to be able to compare all the 

 published copies of the facsimiles with the inscriptions themselves, which, in 

 respect to those at Nasik, I have been unable as yet to do, so as at least to get as 



