1897.] 
Central Asian Manuscripts. 
247 
passim. The only exception is in the case of the ligature rya , when 
r is formed on the line ; see fl. 22a 3 . All this, however, is a practice 
which goes as far back: as the first century A.D. . 
I may also note, that as a rule no marks of interpunctuation or 
division are used. Exceptionally, however, a circular mark occurs, to 
mark the end of a chapter ( adhydya:), as on fl. 21a 5 , 21b 3 , and an oblong 
mark to indicate the end of a paragraph as on fl. 23 6 4 > 6 . 
It may also be worth noticing that the leaves of this work are 
also numbered on the obverse pages. This a practice ou which I have 
already remarked on page 227. 
The language of the manuscript is Sanskrit, but of the well-known 
ungrammatical or mixed type which was peculiar to the earlier Buddhist 
writers. Examples of this are the prakriticisms aiko (for aikah) in fl. 
22a 4 , tasma (for tasmdt') fl. 22a 4 , bhave (for bhaved ) fl. 22a 3 > 6 , mantrena 
(for mantrena) fl. 22 b l , 23?A 8 . 
The work is written partly in verse (yloJca) and partly in prose. 
The ^lokas, however, are frequently, very irregularly formed, the padas 
being sometimes too short, sometimes too long by one syllable; some¬ 
times two padas are run into one uninterrupted half-verse; see fl. 226 6 , 
23fr 9 . In my transcript, below, I have indicated any ^loka that could 
be recognized by the insertion, within angular brackets, of the usual 
single and double lines of division. A clear prose passage can be 
distinguished in fl. 236 6> 7 . 
The work appears to have been divided into adhydyas or chapters. 
On fl. 22a 5 we have the end of the eighth chapter, and on fl. 225 9 the 
mutilated ending of the ninth chapter. The tenth chapter which 
follows seems to have been called gandharva-karma or ‘ business of 
Gandharvas.’ 
A point worth noting is the frequent occurrence of unusual or 
unknown words. I have noticed the following instances : kejala fl. 22a 4 
(for kinjala ?), yanti ‘ingredient’ fl. 226 4 , kanavira fl. 23a 6 (for kanayira 
or karavira ?) ; spandana and rdsabha fl. 23a 5 as names of two medicinal 
plants. This adds to the general archaic look of the work. On fl. 23a 2 
there occurs the word rajamatra or ‘ a person of princely position ’; 
it occurs together with the word raja. According to the St. Petersburg 
dictionary, the word rajamatra is extremely rare; it seems to occur 
but once, in Caraka, part I, chapter 15. The context in Caraka is 
different; but the coincidence is surprising. Could our manuscript have 
anything to do with the original Caraka, that is, the work of Agnive^a ? 
For the nature of the work in our manuscript is undoubtedly medical 
or semi-medical. 
