352 Examination of the Inscriptions [ApRit, 



and the visarga is further explained hy a passage in the Tantrd-bhid- 

 hdna, a more modern work still, dated in 1406 Saka. 



The name of visarga is ' two ths,' ' Swdhd t f analapriya,— became the visarga 

 has the form of the letter 1h(0). 



This merely alludes to the modern form of the 4, which exactly 

 resembles the Bengali visarga. 



The oldest allusion he could furnish, was the following on the form 

 of the 6 from Pingala's Prakrit Grammar. 



"The yuru mark* is like the figure 6, crooked, and of two strokes : it is called 

 also lahu (laghu), it is also denoted by one stroke or one minute/* 



This passage evidently alludes to a form of 6 more resembling the 

 Bengali than the present Nagari type. 



Another channel through which I was in hopes of tracing the ancient 

 ciphers was the numerical system of those Indian alphabets which bear 

 most resemblance to the forms of the earlier centuries, such as those of 

 Cashmir, &c. In the specimens of these, which I have introduced into 

 the plate for the purpose of comparison, it will be seen that the three has 

 certainly considerable affinity to our ^T ; while the one, and five ap- 

 proach nearly to our ^ and \o , There is a faint resemblance, in 



others of the group ; but some again are totally changed. 



The Tibetan numerals (of the seventh century) do not yield much 

 more insight into the matter. They are, we may say, one remove back- 

 wards from the Bengali numbers — the 1, 2, 3 and 5, only agreeing 

 better with the Nagari forms. The 1, however, agrees exactly with 

 one of the ancient figures on the coins, and this has been my inducement 

 to consider the latter as 1. 



Upon regarding attentively the forms of many of the numerals, one 

 cannot but be led to suppose that the initial letters of the written names 

 were, many of them, adopted as their numerical symbols. Thus in the 

 Tibetan 5 ^> we see the ^ or p of the same alphabet, the initial of 

 pancha. The same may be said of the Cashmirian, and the modern 



Hindi form u and indeed in some measure of the ancient forms D and ^ , 

 Again the Tibetan 6 ^>, resembles the ch ^ of that alphabet : the 

 Ceylonese form is exactly the ch of its' alphabet ; and there is an equally 

 marked connection between the Nagari ^ and the ^ chha, which is the 

 common name of this numeral. 



* i. e. The mark used to denote a short quantity in prosody and in music, which 

 is formed ^»» 



