1837.] to the No. 1 Inscriptions of the Ldts. 469 



of Upper India ; and the recent accession of the Girnar inscription of 

 Gujerat, transmitted by Mr. Wathkn, and of the Aswastuma inscrip- 

 tion of Cuttack executed with such fidelity by Lieutenant Kittob, has 

 proved that it belongs equally to them, although in other respects both 

 these texts differ from those already known to us. Thus from the 

 very numerous examples of this passage, we have an opportunity of 

 observing all the variations it undergoes either from carelessness of 

 the sculptor, from grammatical license, or from mistakes of the 

 copyist. The most usual reading of the text, and the equivalent 

 according to my alphabet, are as follows : 



Devdnamapiya piyadasi laja hevam aha. 



Here we perceive at once that the language is the same as was 

 observed on the Bhilsa fragments, — not Sanskrit, but the vernacular 

 modification of it, which has been so fortunately preserved for us in 

 the Pali scriptures of Ceylon and Ava. Devdnam piya (often er piye) piya- 

 dasi laja, is precisely the Sanskrit, ^T«rf fsrsf fasj^f*! KT*T[, ' the lovely 

 raja Devdndmpriya ;' or, with equal propriety, ' the beloved of the 

 gods king Piyadassi ;' for either or both, may be the prince's name. 

 Hevam aha, (or rather evam aha for the h belongs to the word laja,) 

 I recognized at once as an old friend in the Pdli version of the Bud- 

 dhist couplet ye dharmma, &c. so thoroughly investigated in the 

 Journal for March, 1835 : evam dha, 'thus spake.' 



Many of the repetitions of this initial sentence abound in trifling 

 errors, especially in the vowel marks, and in the letters of nearly 

 similar form, as p and h. These it is not worth while to notice, 

 except as a caution against too implicitly following the text in other 

 places, where such slight alterations will restore intelligibility. But 

 Ratna Paula the Pali scholar, whom I immediately invited to assist 

 me in reading the inscription, could critically take objections to other 

 inaccuracies which were repeated in every instance of the pillar text. 

 Thus the double s was wanting in dasi ; the nominative laja should be 

 written rdjd ; hevam, evam ; and aha, dha. Satisfied that these were 

 but the licenses of a loose vernacular orthography, as particularly 

 evinced by the interchange of the liquids / and r, I was little abashed 

 in finding the same errors on the Bakra and Betiah lats, and even 

 on the Cuttack cave inscription : — and it was with a degree of surprise 

 and joy proportionate to the absence of expectation, that on looking 

 over the Girnar version, I found all three of the grammatical errors 

 removed ! The Girnar text is thus conceived : 

 3 p 



