________________
No. 20.]
with a small percentage of copper, which had formed an efflorescence on the surface of the extremely brittle band, with the result that I could neither unroll it without breaking it, nor subject it to the usual chemical treatment. By the use of strong acid, however, applied with a zinc pencil, I was able to remove the copper efflorescence and expose, one by one, the punctured dots of the lettering on the back of the scroll, and then, having transcribed these with the aid of a mirror, to break off a section of the scroll and so continue the process of cleaning and transcription. In this way I succeeded in making a complete copy of the record from the back of the scroll, while the letters were yet intact. Afterwards I cleaned in like manner and copied the face of each of the broken sections, and was gratified to find that my second transcript was in accurate agreement with the first." Later on Sir John, by means of other methods, was able to clean the scroll more effectively, and succeeded in presenting a photographic reproduction of practically the whole inscription.
The characters are Kharoshthi of the early Kushana type. As in other Kharoshṭhi inscriptions, the length of vowels and diphthongs has not been marked. In the case of a, i, aud u it is usual to retain this peculiarity in transliterating, while e and o are often marked as long, if we have no special reasons for assuming that they are short. The only consistent transliteration is, however, not to use the sign of length in the case of e and o any more than in the case of a, i, and u. E and o are, of course, always long in Sanskrit; in the Prakrits, on the other hand, they are sometimes long and sometimes short, just like other vowels. I shall therefore write e and o throughout, though most of the o-s and all the e-s occurring in the inscription are probably long. With regard to individual letters we may note the sign for ña in -ñati-, 1. 4, and f in Imdafria, 1. 2. The voiced and voiceless dentals are carefully distinguished; compare pradisthavita, 11. 1 and 2; rajatirajasa, 1. 3, etc. The compound letter tva occurs in bodhisatva-, 1. 3. The same sign has hitherto been transliterated tm in atmano, 1. 5. The only reason for this latter writing is that the corresponding Sanskrit word atmanah contains a tma. If that had not been the case, nobody would have thought of reading otherwise than tva. Prakrit is not, however, Sanskrit, and I have no doubt that the correct reading is atvano. In the first place we know from the Ara inscription! that there was another compound tma, and it is not at all likely that there should be on the one side two different compounds for tma, and on the other two different meanings of one of them. Moreover the usual Prakrit forms atta and appa cannot well be derived from Sanskrit atman, but both seem to go back to an intermediate form atvan, just as sattva becomes satta as well as sappa. Another compound occurs in pradisthavita, 11. 1 and 2; vasthavena, 1. 2. It will be seen that it is used both instead of Sanskrit sta and of Sanskrit stha. As a rule, it can be stated that the compound is rarely used in Kharoshthi inscriptions except in foreign words and such as can be considered as Sanskrit loanwords. St as well as sth regularly became tth in the dialect. It is a well known feature of the Prakrits that an a causes aspiration of a neighbouring consonant, and the necessary link between st as well as sth and tth is sth. Similarly we find shth in the Iranian name Hashthuna in the Wardak vase inscription.3 It is therefore just as natural to transliterate the compound stha as sta; and, if we consider its shape, it is evident that the lower portion is identical with the letter tha and has nothing to do with ta. I therefore write stha and consider this as a semi-learned writing which does not represent the phonetical features of the dialect.
TAXILA INSCRIPTION OF THE YEAR 136.
285
With regard to orthography, we may note, in addition to the use of the compound stha, that ya ropresents a Sanskrit ja in puyae, 11. 4 and 5. The writing is not over careful. Thus we find a fore in putrana, 1. 2, prachaga-, 1. 4, and perhaps in ma, 1. 5. Syllables have bren left out in Tachhasie, for Tachhasilae, 1. 3; arahana, for arahantana, 1. 4; sarvasana, for sarvusatvania, 1. 4, and -saluhina, for salohidana, 1. 4.
1 Above F. 143.
Cf. Lüders, SBAW., 1914, p. 416. J. Seuart, Jo. Asiat, XI, iv, p. 574. Konow, 8.B.4, 1916, p. 808. Mr. Pargitor roads Kashfuna.
