________________
No. 28.)
SO-CALLED TAKHT-I-BAHI INSCRIPTION OF THE YEAR 103.
263
Persian, i.e., the result was probably & voiced sound. The Kharoshthi siga itself seems to be derived from v and not from p, by adding an upward stroke to the right. A similar stroke is found in the Kharðshthi documents discovered in Chinese Turkistan, where we sometimes find the letter which is usually transliterated at provided with a perpendicular, rising from the right end of the cross-bar. Of. plate XCII in Sir Aurel Stein's Ancient Khotan, where Messrs. Boyer, Rapeon and Senart read the akshara as sth in sthishyadi, 1. 9. I am, therefore, inclined to think that vh, i.e., an aspirated voiced spirant, was the sound meant, and it would probably be more correct to transliterate vh. I do not think, however, that we can be certain about the pronunciation, and, for practical reasons, I prefer to write f.
I may further draw attention to the compound letters tsa in sarbatfarao, 1, 2, and rjh in erjhana, 1. 5.
With regard to orthography and phonology we may note the change of intervooalic j to y in maharayasa, 1. 1 ; puyao, 11. 5 and 6; the softening of intervocalic t to d in madr, 1. 5: pidu, 1. 6; the change of v to b after an anusara in sa nbatéarae, 1, 2; of ts to tf in the same word; of shy to fin isa, 1, 2, and of śr to sh in shadhadana, 1. 4. Such changes are, as is well known, of frequent occurrence in Indian Kharðshthi inscriptions.
In spite of all the care that has been bestowed on the reading and interpretation of the epigraph, several points are still far from being finally settled. Dr. Thomas has, therefore, rendered a real service to scholars in preparing the new facsimile plate published together with this article.
The first line can be read with perfect certainty and has never presented any serious difficulty. Dowson read the second akshara of the king's name m instead of du, but & comparison with du in madu 1. 5, pidu, 1. 6, conclusively shows that the letter is du. The third akshara has usually been rendered as pha, but I have already given my reasons for transliterating fa. We must accordingly read: maharayasa Gudufarasa vasha 20-4-1-1, (during the reign of the Mahārāja Gudufara, in the 26 year. Vasha is also used in the Kaldarra inscription of the year 113 and the Skārah Dheri epigraph of the year 399, while the Machai record of the year 81 has vashē, which is probably intended in the other dates 28 well. There is not, however, any trace of an e-mātra. No certain inference can be drawn from the use of varsha instead of samvatsara irf these instances; the word may have been chosen in our record in order to distinguish between the two dates in 11. 1 and 2, or in order to indicate that the year used in the first one began with the rains.
Cunningham and Dowson identified Gadufars with the king Gondophares of Christian tradition, and this identification has been generally accepted and may be considered as certain. The date of 1. 1 has, further, always been interpreted to mean that the inscription belongs to the 26th year of the reign of King Gudufara. Epigraphiste will, however, agree with me that we are oply informed that it was issued during Gadufara's reign, while the year can just as well be referred to some era which may have been introduced by some of Gudafara's predecessors. If We compare the dating of Brahmi inscriptions of the Kushåna period, we might be inclined to think that the latter was the one. Of. o.g. No. 149a of Professor Lüders' List of Brahmi Inscriptions : maharajasya rajdtirdfanya dēva putrasya shahēr-Vudvishkasya rajya-samvatsari 24 and other similar records where an is used instead of rajya-samvatsaro. Nobody would here think of the regnal year of the king mentioned in the inscriptions, but ur hesitatingly refer the date to the Kanishka era. I shall state below why I think the same to be the case ip our epigraph.
Then follows, in l. 2 and the beginning of 1, 3, a new date.
