\ 



476 Journal of the Asiatic Society of Bengal. [July, 1907 



witli respect to the actual date of the mairtiscripts themselves, 

 for, it must be borne in mind, copyists have always' had a 

 tendency to adapt notations to the systems in vogue in their own 

 times and to correct figures or alter them in accordance mth some 

 particular convention. Numerous examples of such changes 

 ^jottld be quoted, but the most striking one for my pui'pose is that 

 made by Dr. Fleet in the Corpus Inscriptionum Indicarum, volume 

 iii., page 73. He there quptes as the earliest epigraphical instance 

 in India of the use of ' numerical words' the record of " Saka 

 Samvat 867 (a.d. 945-46) for the accession of the Eastern 

 Chalukya king Amma II (Ind. Antiq. vol. vii., p. 16) in which 

 the date is expressed by the (eight) demi-gods called Vasu, the 

 (six) flavours, and thie (seven) mountains." lii tKe original 

 inscription the date is given by the words ^^ yiri-rasa-vasa, etc.^^^ 

 ^Nrhich is correctly translated' (by Fleet himself. Ind. Ahtiq.viir, 

 p. 16) by the words "The (seven) 'mountains, the (six) flavours 

 aiid the (eight) kinds of demi-gods called Vasu"'(i',e. the Saka 

 yeat* 867). Kote that Dr. Fleet, in the first quotation given above, 

 changed the order of imting the date'.' Instead of writing 

 "mountains, flavours, vasus " he wrote " vasus, flavours, 

 mountains" just as we might write 867 for 768 by mistake. 

 But with Dr. Fleet it is not a niistake, that is, it is not an 

 accidental mistake, for he gives other examples with just the 

 same inversion. What he actually did was to alter the Indian 

 plan of writing the smaller elements first to our convention of 

 writing the larger elements first, evidently attaching no im- 

 portance to this order of writing. As a matter of fact, this point 

 is of the utmost , importance in any investigation regarding the 

 origin of our notation and forms a valuable link in my present 

 Argument, indeed, forming the starting point of my essay. 

 Also first-hand evidence of every kind that is available has been 

 sbught-^evidence that cannot have been corrupted in transit. 

 It hals also been necessary to investigate certain other evidence, 

 not because of its real value but because of the importance that has 

 ^en attached to it by other investigators. 



h 





11. 



± '' 



1 



~^ Sanskrit and kindred scrints are now written from 



right 



3r*e was, indeed, m ancient times, an 

 btached the name Karoshthi, ii use ii 

 characters of which were written 



&stead of from 



iraneous Indian script ; but this affects the question only remotely- 



Sanskrit and kindred scrints are 



and have 'been, for centuries, written from 

 Arabic f amilv of 



^ pts are written i 



be naturalto expect number words ai 

 the mode or direction of the writii _ 

 appear strange to see numerical symbols wri 



It would 



conjunction 



If the development of the- 



