7G4 EXPLORATION IN CHINESE TITRKESTAN. 



days, hence they had never l)een able to attack the ruins more 

 deeply covered ])y the sand. Thus, when I commenced with m}' little 

 force of laborers the systematic excavation of structures half buried 

 by low dunes, most interesting archa3ological results soon began to 

 reward me. From the cellas of little Buddhist shrines there came to 

 light in large numbers stucco images and relievos, frescoes and painted 

 wooden tablets, all showing representations of saints and legends of 

 sacred Buddhist lore (pi. iv). In style and technical treatment they 

 exhibit a close resemblance to that period of ancient Indian art which is 

 best known to us from the latter Ajanta cave paintings. Wherever 

 protected by the dr}' desert sand, the colors have survived in remarkable 

 freshness. Here, then, were rising from their tomb long-lost relics 

 of that Indian art which had found a second home in Buddhist Central 

 Asia before spreading farther into the Far East. 



Great was my jo}- when, on excavating what must have been the 

 ground-floor room of a small monastic dwelling place, the men came 

 upon the first leaves of paper manuscripts. Carefully extracted with 

 my own hands and cleared, thev proved to contain portions of a liudd- 

 hist canonical text in Sanskrit. Judging from the palisographic char- 

 acter of the writing, these and su))sequent finds of fragmentary 

 tSanskrit manuscripts from Dandan-Uilik ruins may approximately be 

 assigned to the sixth or seventh century of our era. In addition to 

 such texts in the classical language of India, the literary discoveries 

 of this site include a considerable num}>er of manuscript folia and of 

 detached documents on paper, written in Indian Brahmi characters, 

 ])ut in a non-Indian language. Taking into account that the same 

 strange language appears in inscriptions athxed to some frescoes, it 

 seems probable that w^e have here records of the indigenous tongue 

 actually spoken by the Khotan people of that period. Only the close 

 study of all these documents — a task which may take years — is likely 

 to lead to a decipherment, and thus to a solution of this interesting 

 (juestion. 



In the meantime it is fortunate indeed that the discovery of Chinese 

 paper documents in other small monastic dwellings permits us to 

 determine with accuracy the period when the settlement represented 

 by the settlement of Dandan-Uilik was finally abandoned. Among 

 the neatly folded small paper rolls containing letters, records of loans, 

 petitions, and similar matter, there are three at least which alreadv, 

 on pieliminar}" examination at Kashgar, proved to be dated with pre- 

 cision, the Chinese years indicated corresponding to the years 778, 

 782, 787 of our era. There are good reasons for assuming that these 

 petty records do not precede l)y any great length of time the date 

 when the dwellings were abandoned. We thus obtained the end of 

 the eighth century as the approximate chronological limit for the 

 existence of Dandan-Uilik as an inhabited localit}^ This dating is 



