220 Facsimiles of Ancient Inscriptions. [March, 



is one step further removed, but it springs directly from the Hala 

 Canara, and retains many of the Andhra letters still unchanged, par- 

 ticularly the dk and th. In the accompanying plate (xii.) we have 

 thought it worth while to exhibit these resemblances, and point out 

 the peculiarities noted, that no means may be neglected of facilitating 

 the examination of other inscriptions that may link on naturally at 

 either end of this fragment of the chain of our Indian palaeography. 



After having made the transcript according to the assumed value 

 of each letter, it was revised and corrected in all doubtful points by 

 reading it over with Ma'dhora'y* pandit, the aged librarian of the 

 Sanskrit college, who, from having been with Colonel Mackenzie, i» 

 better versed in the varieties of the Nagari alphabets than any pandit 

 in Calcutta. Where the context did not make sense, the letters were 

 carefully analyzed and all possible variations of each letter suggested, 

 until the true or most probable reading was apprehended. Although 

 some few doubtful passages remained, and many orthographical errors 

 were detected, the context was sufficiently intelligible, and satis- 

 factory. In some few instances (as in lines 6, 8, and 17) the distin- 

 guishing stroke or dot of the letter n has been omitted either by the 

 sculptor or by the transcriber. The omission can be supplied without 

 hesitation, as no other letter occurs at all similar in form. The cross 

 of the k in lines 7 and 8 is also wanting. 



For the translation we are indebted to the Rev. Mr. Yates, whose 

 critical knowledge of the Sanskrit enables him to give it the correct 

 grammatical construction which might evade an oral interpreter de- 

 pending upon a vernacular explanation by the pundits. 



Transcript of the Amardvati Inscription. 



* It was Ma'dhora'y who aided Captain Troyer in tbe Allahabad inscrip* 

 tion, J. A. S. vol. ii. 



