560 Ancient Indian Numerals. [No. 7. 



ta inscription at Sanchi.* The 2 and 3, in this form likewise appear 

 among the Nasic legends. It will be seen that the 2 and 3 are 

 essentially the same figures that we have in use at the present day, 

 modified merely by the cursive form given to them by the connect- 

 ing semicircles, by which the necessity of removing the pen from 

 the paper while expressing each separate stroke was avoided, and 

 these in themselves furnish us with a singular illustration of what 

 progressive modifications of Indian alphabets owe to the mechanism 

 of penmanship.f 



Dr. Stevenson contributes an independent form for the one, 

 which he likens to the Bactrian ^T but which under either of its 

 modifications, is simply and solely the compound Tf pt of the alpha- 

 bet, in the body of which it finds a place. He also has a new cypher 

 for 3 in his inscription No. 5, but I am not satisfied either with the 

 accuracy of the outline, or the low value assigned to the figure em- 

 ployed ; as, even accepting the symbol to stand for 3, the sum total 

 to be consistent with the associate system of notation, ought to be 

 read as three hundreds and not 300,000 as given in the text. 



The four is proved to us by the copper plate inscriptions noticed 

 above. The form of this cipher varies considerably in its different 

 examples, ranging from the simple v to the most common type JJ 

 and extending even to the possible ^f or ^? of Dr. Burns No. 1, 

 copper-plate, $ this last reading if admissible, might be thought to 

 ally it with the initial of *J3T, an age, one of the symbolical words for 

 four. Certain of the Sah coins give the outline of the symbol as a 

 ij with two backward horizontal lines issuing, parallel to each other, 

 from the prolonged limb of the second down-stroke of the letter, 

 and again, singular to say, at times the subjoined letter is formed 

 by one continuous stroke, carried round after the manner of the 

 Gupta «r, under which aspect the figure might easily have grown 

 into the a of modern days. 



Dr. Stevenson gives us four imperfect varieties in the outline of 

 this number commencing with the ^ above adverted to, passing on 



* J. A. S. B. VI. 455. Bhilsa Topes, No. 197. 



f Prinsep has already traced the gradual development of many of these alpha- 

 betical changes. J. A. S. B. VI. 1044. 



J J. A. S. B. VII. PI. XX. J. R. A. S. XII. 32. 



