1922.] Dates of the Votive Inscriptions. 227 
to expect at least six distinct stages of alphabetical develop- 
ment in the series of inscriptions which begin in the third 
year of Kanishka and terminate in the year 99 of his era. 
For there cannot be any valid ground for the supposition that 
changes in the alphabet took place at much longer intervals 
immediately or shortly after the Safichi toranas were erected 
himself recognises the fact that the differences in the inserip- 
tions on the railings of Stipa I and Stipa II at Safichi are com- 
paratively very slight and that “ the latter may be somewhat 
later than the former.’ He even puts the two under one general 
chronological division. Yet he sets down the interval between 
the two as that of one generation (p. 4) which can by no 
means be considered to have been less than twenty-five years, 
According to his own views, therefore, the differences between 
the broad chronological divisions must be considerably more 
than twenty-five years each. But leaving aside the fourth 
group which, with its four sub-divisions, comprised about one 
hundred years on the above calculation, there remains a 
period ot only fifty years or a little over that, for accom- 
modating the six broad chronological divisions mentioned by 
him. It thus appears to be difficult to reconcile the different 
views put forward by the author and one can only hope that 
he will satisfactorily explain his position. In any case these 
considerations seem to suggest grave doubts about the 
correctness of the fundamental principles underlying Mr. 
Chanda’s work and the impression gains ground that there is 
something w te in his basic assumptions, an impression 
which I regret to say, is strengthened rather than weakened 
a detailed examination of the work to which we next proceed. 
There is a preliminary difficulty in any detailed examina- 
tion of Mr. Chanda’s theory, which I think must be stated at 
the outset. Mr. Chanda has not utilised the brilliant oppor- 
tunity he had, thanks to Sir John Marshall, of making a chart 
of all the different types of letters used in the Safichi Inserip- 
tions on the lines laid down by Bithler. In the absence of the 
facsimile of all the inscripticns, this alone would have enabled 
his readers to judge for themselves the whole ea als 
question in general, and the correctness of his own views 
particular. Instead of doing that Mr. Chanda has given us 
merely the facsimiles of a few typical examples. It is 
needless to point out that the typical examples chosen by an 
author labour under the same defects as the eye copy of an 
inscription ; for, in both, the author is prepossessed by his own 
