1897.] 
241 
Central Asian Manuscripts. 
obtained from that ruined vihara, only the veriest scraps. Some of these 
fragmentary manuscripts, e.g,, the Sanskrit vocabulary in Part VI of 
the Weber MSS., are sufficiently important to make us wish to obtain 
the complement. It is possible that the missing portions of these manus¬ 
cripts may have suffered destruction in the course of the excavation 
of those two treasure seekers; a good deal undoubtedly must have 
been destroyed ; but it is also quite possible that some further portions 
are still held back by the finders, and may come to light hereafter as a 
result of suitable inducement. 
I now proceed to a detailed account of the several sets of the 
Macartney MSS. 
Set I. This set consists of two entirely different manuscripts, 
specimens of which are shown on Plates XV and XVI respectively. 
They are written in two different types of Brahmi, Set Io. being in the 
Central Asian, while Set I b is in the Northern Indian type. 
Set la consists of 35 leaves, two of which are shown on Plate XV. 
They are all broken off on one side. Their width is complete, 21- 
inches. The existing length is 5 inches, and about 2-| inches must be 
broken off ; the total length, therefore, would be 74 inches. In the 
missing part there must have been the string-hole. This calculation 
can be easily proved. Comparing the Macartney MSS. fragment with the 
Weber MSS. fragment No. VII, 80 and with the Petroffski MSS. fragment 
No. VIII, 81 it will be seen at once that all these three fragments 
absolutely agree in all points of shape, size, and type of letters. If we 
add to this that all three fragments treat of the story of Manibhadra, 
there cannot remain the slightest doubt but that they are portions 
of the same manuscript, one of which has gone to St. Petersburg, while 
the other two are in my hands. Now, by a careful comparison of the 
eight leaves in his possession, Dr. von Oldenburg has been able to 
practically restore the text on the obverse side of his leaf No. 3. The 
restored transcript of this page he has published, as well as its original. 8L 
It will be seen from the transcript that the average number of aksaras 
on a full line is 34. On the second line of the page the existing aksaras 
number 23, and the line itself measures nearly 5 inches. Accordingly 
the. missing 11 aksaras, together with a small margin, would require 
a space of 2\ inches. Hence the page, when complete, would have 
measured 74 inches. Purther, the missing aksaras on the second and 
fifth lines number 11 and 12 respectively, while on the third and fourth 
SO See Journal, As. Soc. Beng., Yol. LXII, Part I, p. 31, and Plate II, fig. 3. 
8L See Journal, Imp. Russian Archeeological Society, Vol. VIII, pp. 13, 17, and 
Plate II, fig. 8. 
