562 Ancient Indian Numerals. [No. 7, 



figure is well represented by the modern Bengali ^ d. The original 

 character assimilates somewhat with the exceptional form of ^ of 

 the Gupta Alphabet of Allahabad, but does not by any means ap- 

 proach the formation of that letter in the Sah inscriptions, nor does 

 it bear any likeness to the cramped ^ on the reverse of the coins 

 of the Sah Dynasty. 



A well-fashioned ancient w is the symbol next in order of succes- 

 sion : were there any more distinct justification for attempting to 

 explain these letter-ciphers by the symbolical terms of the Sanscrit 

 texts, a somewhat forced construction might be put upon this figure 

 as ^^f a^ka nine ; but if outline similitudes have, as it would seem, 

 a preferential claim, this figure may well demand to be con- 

 sidered as the early model of the c — 8 which retains so much 

 of its identity to this day amid the numerals of the medern Deva- 

 nagari. Dr. Stevenson has a cipher for this number — classed under 

 the Satrap heading — which he likens to a Bactrian T. However, I 

 cannot but observe that Lt. Brett's original copy of the text of the 

 inscription scarcely represents the figure in such close identity as 

 the outline entered in the Table of Numerals compiled from his fac- 

 similes would make it. 



Under the catalogue for the variants of eight, perhaps, should be 

 reckoned the letter, not altogether unlike an old ?tor? dh,* but 

 the determination of this point is dependent upon the correct ascer- 

 tainment of the true form of nine, which is but indistinctly express- 

 ed in No. I. Nasik inscriptions, being fashioned, in the one instance, 

 after the manner of an imperfect ^, and in the second, or supple- 

 mentary portion of the same inscription, as ^ of the same character, 

 the former of which is not far removed from the normal form of the 

 Sawrastran v t. The Sah Coin No. 31, PL II. J. K. A. S. XII. 

 seems to furnish us with the correct model of the nine, but I hesi- 

 tate in conclusively classing all these "Z t, "g" dh, ^ d and fs[ di — as 

 the proved exponents of this number, only because I observe, that 

 Dr. Stevenson, in his table, gives the form in an inversed direction, 

 something after the model of a cursive Bactrian S. This outline 

 certainly does not agree with that given in Lt. Brett's accompany- 

 ing detail inscription, but the author may possibly possess other 

 * J. R. A. S. XII. PL II. fig 31. J. A. S. B. VII. p. 354, fig. 5. 



