> 
232 Journal of the Asiatic Society of Bengal. [N.S., XVIII, 
had not yet reached its next definite stage of development. 
It is always a difficult task to determine the duration of one 
alphabetic epoch—if one might use the term—and it might be 
anything like a hundred to a hundred and fifty years, or even 
more. 
In the absence of other evidence, therefore, the Safichi in- 
scriptions might have been roughly assigned to the period 250 
to about 100 B.C., but a comparison with the inscription of 
Heliodorus seems to carry back this lower limit 
This inscription seems to usher in the next stage of 
development in the Brahmi alphabet. It shows distinctly the 
beginnings of the principal characteristics of the later period 
in the advanced forms of individual letters as well as in a 
marked tendency towards the angularisation of the letters and 
the equalisation of their verticals. Cf. e.g. v in déva (1. I), 
vadhamanasa (1. 7), and s in va(sudé)vasa (1. 1) Diyasa (1. 3), 
dasémna (1.7); angular h in maharajasa (1. 4), Héliodorena 
(1. 2); and the equalisation of the verticals of s in pul(r)asa 
(1. 6), and tratarasa (1. 6). 
The presence of so many decidedly later characteristics 
in a short record of seven lines marks it as belonging toa later 
epoch. It is true that the beginnings of the tendency towards 
angularisation and the equalisation of the verticals are just 
perceptible in a few inscriptions from the railings of Stipa I 
and II at Safichi and also possibly in those of Asoka. (Cf. 
e.g.hin No. 4, pl Ill; s in Nos. 6 and 8, pl.. IV; Bihler’s 
Table II, Col. V. 39, and Col. VI. 40.) But a comparison of 
these letters with those of Heliodorus’ inscription referred to 
above would at once reveal the fact that the forms in the 
latter are far more advanced. ‘This fact together with the 
other advanced forms such as the triangular v and angular s, 
which are practically absent in the former group, point to the 
Heliodorus’ inscription as marking a new stage in the develop- 
ment of the Brahmi alphabet.. If it is ultimately proved, 
when the impressions of the whole series of the Safichi inserip- 
tions are available for comparison, that all these characteristics 
are altogether absent from them, it would follow that the 
original construction of the railing! was separated by a pretty 

! The massive Stiipas like those at Safichi and Bharhut were con- 
structed, as the short donative inscriptions conclusively prove 
co-operation of a number of individuals. Now the different ind 
eee . 
over which their names are engraved or they could pay an amount of 
money sufficient to cover the expenses of building them. The : 
try of the different parts of the building is, however, incompatible with 
the first view, and makes it almost certain that although the donors refer 
to their gifts of the different parts of the building, they reaily paid their 
cost. This again, could have been done in two ways:—(i) by collecting 
subscriptions for the structures and engraving, as so hey W 
completed, the names of the subscribers over the different parts selected 



