216 R. Hoernle —Three further Collections of [No* 4>, 
An equally good test-letter is the aksara ma. The Central Asian 
form of it is made in two distinct divisions : an open square (like the 
ordinary Brahmi ipa) above and a horizontal line below. On the other 
band the Nortli-Indian (Gupta) form consists only of an open square, 
the left side of which is serpentine. The two forms can be distinctly 
seen and compared in IYa* {mi Central Asian) and VII 3 (ma Northern 
Indian). 5 It will be noticed here that the Central Asian form origi¬ 
nated by prolonging the dent of the left perpendicular line of the 
North-Indian form so far as to touch its right perpendicular line. 
Another minor test is the general slant of the writing in Central 
Asian, contrasted with the upright writing of the Northern Indian ; this, 
however, makes itself not so apparent in single letters or words as in 
a whole page, where the difference of the two types of Brahmi characters 
forces itself at once on one’s attention. There are other minor points 
of difference between the two types : thus the medial long a , made by a 
long liorn-like projection or curve as in ha (Villa 1 ), ta (IXa 4 , IXc 4 ), 
sva (1 Va*) y ya (III d\ XI 6 s ). 
A further good test is the form of ya, which in the Northern 
Indian is distinctly tridental, whence it passes, through an intermediate, 
into the modern square form. In the Central Asian, on the other 
hand, the old tridental form of ya gradually passes into a bi-annular 
form. It is particularly to be noted, as a land-mark for chronological 
purposes, that the Northern Indian intermediate form only existed for a 
comparatively short time. It first appears in engraved documents about 
370 A.D., and disappears again about 540 A.D. 6 It lasted in round 
numbers for (say) 200 years, and w T as only used in conjunction with the 
superscribed vowels e, ai, o, au. It was clearly an attempt at producing 
a more convenient cursive form. It consisted in the closure of the left 
side of the trident, producing an irregular circlet. By the gradual 
broadening of this circlet, and the concurrent atrophy of the right side 
of the trident, the modern form of ya was produced. The latter is 
practically dominant in Northern India from (say) 600 A.D. It is 
curious to observe that the subscribed conjunct ya passed through a 
very similar course of evolution, though several centuries earlier than 
the non-conjunct ya. There the process occurred in the 1st and 2nd 
centuries A.D., the period being also about 200 years, and there was 
also the same intermediate form of ya. An instance of the latter is 
given by Prof. Buliler, from the 1st century A.D., in Plate III (Column 
5 Raised numbers indicate lines. Thus IYa2 means the second line on fragment 
a, belonging to No. 4 on Plate YII. 
6 See detailed proof in my paper on the date of the Bower MS. in Journal, 
As. Soc. Beng., Yol. LX, pp. 83, ff. 
