218 
[No. 4, 
R. Hoernle —Three further Collections of 
letter. The last step of the evolution is reached in No. XII, which 
shows all the characteristics of the Central Asian type of Brahmi, just 
like Part IX of the Weber MSS. ( ibidem , Plate III, tig. 3—5); but note 
especially the full biannular form of ya in ya XI16 s , also the angular 
form of dha in XII6 3 . 
I proceed to notice some details of this collection of fragments. 
No. I. This is written on palm-leaf, in a very neat, clear, and 
careful hand, so that it is a pity that not more has been preserved 
of the manuscript. The characters are of the Northern Indian Gupta 
class, and their type indicates a rather early Gupta period. The 
letter ya is used in its tridental form ; even the intermediate form is 
absent; see yd , yai in I, Id 3 , ye in If 3 . The superscribed conjunct r is 
formed within the line, see rda I/r 2 , rtta Ic l . A date before 350 A.D. 
suo-o-ests itself. There is nothing in the fragments to indicate the size 
of the leaves, or the extent and contents of the work. The fragment h , 
however, shows the number 2 on its margin, which would seem to 
indicate it as the remnant of the second leaf. 
No. II. These fragments are written on bircli-bark and might be of 
a work of the same age as the Bower MSS. From the style of the 
characters they might, indeed, be fragments of that work, though there 
is nothing in them to indicate the nature of the work to which they 
may have belonged. Fragment lie is written in a larger hand than the 
others, and probably belonged to a separate work. 
No. III. All these fragments are written on paper. The five 
pieces a, b , c, e, f are written in Northern Indian Gupta, while piece 
d is written in Central Asian. The latter, therefore, belonged to a 
work quite different from the others. But the handwriting in the 
pieces a and b is a little different from that in the pieces c , e, /; 
and these two sets, therefore, may have belonged to two different 
manuscripts, though their purport is the same: they treated of 
spells. Pieces a and b are still connected with the original thread ; 
and other pieces of thread which I received together with this 
collection of fragments are shown in the centre of the Plate. I 
would place the date of the manuscript to which pieces a and b be¬ 
longed early in the 5th century, contemporary with the Bower MS., 
on account of their showing the intermediate form of ya in yd III a. 8 
But it must be noted that the tridental form also occurs in yd III c. 3 
The superscribed conjunct r is formed within the line ; see rnnl III e l . 
No. IV. Written on thin paper, in bold and clear Central Asian 
of a very early type, as shown by the tridental form of ya, and the 
straight form of the medial u in asnlca , line 1. Both forms point to a 
date not later than (say) 450 A.D. The curious appendage to the foot 
