1922. ] Dates of the Votive Inscriptions. 231 
Safichi group, would, when applied to ka lead to a diametrically 
opposite conclusion. 
As regards the fourth test letter ha Mr. Chanda himself 
admits that the Safichi type of ha occurs pretty frequently 
in the Asokan records but that it is found almost exclusively 
in some of the pillar edicts. ‘ But,” remarks Mr. Chanda, “ if 
ha is derived from the Aramaic He turned upside down and 
from right to left, the first type of ha should be considered as 
more archaic. In the Nagarjuni hill cave inscriptions of Dasa- 
_Tatha all the has are of this archaic type. So the total ab- 
sence of this type of ha in the older votive inscriptions of 
Safichi, as in other decidedly post-Mauryan inscriptions is 
not without chronological significance”. Mr. Chanda has also 
advanced similar arguments in order to prove the antiquity of 
the first type of a. But speculations on this line are of no use, 
as the derivation of the Brahmi alphabet from the Phoenician, 
far less the derivation of the individual Brahmi character from 
far from being conclusively proved as yet. Quite recently the 
theory has been vigorously challenged by Prof. D. R. Bhandar- 
kar and he has produced facts and figures which go a great 
way towards demolishing it. 
ere is, however, a strange inconsistency in the above 
argument. If all the has in the inscription of Dasaratha are of 
archaic type, it must, on the principle adopted by Mr. Chanda, 
be looked upon as earlier in date than even the Asoka inscrip- 
tions wherein the has of the later type occur as frequently as 
the earlier ! 
The above discussions will make it quite clear that a, bha, 
ra, and ha are not really test letters, or, in other words, the 
forms of these letters found in Safichi inscriptions do not lead 
to any definite inference regarding their age. As Mr. Chanda 
has based the chronological divisions of the first four groups 
of inscriptions solely on the basis of these tests this part of his 
theory requires, in my opinion, further evidence before it can 
be accepted even as a working hypothesis 
It is indeed difficult to form any opinion on the age of the 
letters of the Safichi inscriptions till the estampage of the 
whole series is available for comparison. But if an opinion 
can be hazarded on the accessible materials, there does not 
seem to be any valid reason for rejecting the views of Cun- 
ningham and Bihbler that the inscriptions on the railing belong 
approximately to the age of Asoka; for so far as I can see, the 
facsimiles published by Mr. Chanda do not contain any sure 
trace of a distincily developed stage of alphabet. 
me be not misunderstood do not mean to suggest 
that the Saiichi inscriptions were incised during the reign of 
Asoka or even within ten or fifteen years after its close, but 
that they were engraved at a time when the Brahmi alphabet 
