________________
NOVEMBER, 1893.] ASOKA'S SAHASRAM, RUPNATH AND BAIRAT EDICTS.
305
27. Etiya must not be changed to étáya, as M. Senart proposes; it is the dative of the feminine stem étí, which appears in élissá, étisset, and so forth. The use of the feminine for the masculine is common enough in these inscriptions; compare e. 9. above 1. 2, imeyu kaliya.
28. B proves most distinctly that atá not and is the reading. The form alu for la occurs also in the Kalsi Rock-Ed. XIII. 2, 6, atésu, and is protected by numerous analogies like magala for marigala, leiti for learnti, and so forth.
29. The vowel of this word is not distinguishable. It probably was palare, and may be a mistake for pakuwd, as M. Senart thinks, or equivalent to pakure, " manner" (of actiny).
30. Read vadhisati.
31. The 14-stroke of the last syllable of paratisu is very short, but unmistaknble, especially in B. The correction pavatész, which M. Senart proposcs, seems to me unnecessary, as in l'ali i frequently appears for Sanskrit e.
32. Hadha is either a mistake or a ricarious form for kidla. The words 1&lopéta-rálatu are as plain as possible on the new materials, and B shews that the rock has not been worn away. On the supposition that ráluta stands for pálata, i. e. paratru, with the in lali not unusnai softening of the pa, the clanse may be translated : " This matter has been incised by my order in the far distance in the districts) and here in Vagacha)." The last words remind one of the phrase in Rock Ediet V, hide whilésu chú nagalésu, (K. 1. 16). With this interpretation the sense is unobjectionable, but it may be nrged that the parallelism of the next phrase and tlie corresponding passage of the Sahasram cdict make it probable, that there should be a future puerticiple passive instead of the past participle passive. If that seems indispensable, it will suffice to insert one single syllable and to write lékhápetara-válata. Lekhipetava, i. c. lékhápetavca is as good as lékhápetuviya. M. Senart's extensive changes seem to me neither necessary nor even advisable as they destroy the sense of the passage.
33. In B the dental tha and the final i of athi are perfectly recognisable. In A these signs look exactly like those of the old facsimile. Cha has been inserted as a correction. Silathublé is not very plain on the old facsimile, but unmistakable both in A and B. The change of a to u has been caused by the influence of the labial; compare E. Müller, Simplified Pali Grammar, P. 6.
34. Read lékhápētavaya. With the termination raya for tiya compare such words as paddlayé, duparipadaye, and so forth. The final ta stands for li, i.e. iti. It is, however, not absolutely necessary to correct ta to li, as M. Senart does. For, the Maharashtri ia, which appears for iti in the beginning of a verse or of a sentence, points to the former existence of a vicarious form itu, which might be shortened to ta.
35. The vowel of the penultimate syllable, which is much injured, is doubtful; that of the antipenultimate [ is clearly á, not i, as M. Senart's transcript makes it.
36. Vynthena, not Tyathena, is the reading; but the u-strcke is very short, and the semicircular stroke of the ya very thick.
37. The final i of vasdni is at least probable, and it is certain that there is no Anusvara after the na. Possibly yan hakan to be read.
38. The Anusvåra of sasinghe is not certain. Both the impressions have clearly upayáté not payite.
39. The second syllable of amisd stands above the line. The following syllable may have been non, but the stone is just here very ropgh. Dôvéhi, now known as the reading of the Mysore versions, is tolerably distinct with the exception of the last consonant, which is rather faint.