Sec. in EXCAVATION OF THE ENDERE TEMPLE 429
However this may be, the prolonged presence of Tibetans at the site is attested not merely by the votive offerings of Tibetan MSS. already described, but also by a series of Tibetan sgraffiti found scratched by various hands, often very cursive, into the north and south walls of the cella. Those on the latter side have been reproduced in Plate XII, while all of them, as far as they are decipherable from the photographs and eye-copies taken by me, will be found translated and learnedly commented upon by the Rev. Mr. Francke in Part iv. of Appendix B. They appear to be votive, recording offerings made to certain divinities, and indicating the benefits piously hoped for in return. A journey ' to the land on the other side (Tibet ?)', on which the donors of a wool-ox' expect to meet with `wealth, food, and grass', is indicated as the special occasion of the offering recorded in the sgraffito towards the southwest corner (to the right in Plate XII). These Tibetan scrawlings looked on the whole less well-preserved than the Chinese sgraffito, and there was certainly nothing about them to indicate their being later in date. On the contrary, the difference in state of preservation between the Chinese sgraffito on the west wall and an almost completely effaced and illegible Tibetan scrawl on the same wall suggested the reverse conclusion.
No direct chronological clue is furnished by the Tibetan sgraffiti ; but it can be considered quite certain that the date when they and the Chinese inscription were scratched into the walls could not have preceded by many years the deposition of the various manuscripts and the subsequent abandonment of the shrine. The rough and friable plaster of these cella walls was not a material that could, when exposed, remain intact for a long period without repair, and with its renewal all these casual scratchings as well as the Chinese record would, no doubt, have vanished. Taking into account the plainly attested date of the latter, this consideration fixes approximately the middle of the eighth century as the latest possible time for the abandonment of the shrine, and implicitly also for the production of the manuscripts found in it. This conclusion fully accords with the fact that among the eight Chinese copper coins found either within the Endere fort or in its close vicinity there is not a single specimen of the Tang issues so common at Dandân-U iliq 17. It also accounts for the distinctly older appearance which the sculptural and pictorial remains of Endere seem to bear as compared with those of the Dandân-Uiliq shrines.
The possibility of approximately exact dating invests with special interest a series of votive offerings of a humbler kind, which I found scattered in front of the various image bases, and which still remain to be mentioned. The rectangular piece of paper (E. i. 19.a, see Plate LXXIX) found in the sand close to the north-west facet of the central base shows a rapid but spirited sketch in colours of a Bactrian camel suckling her calf. Good observation is displayed in the drawing. On the hindquarter appears a brand. Could it be that of the owner, who, perhaps, deposited the sketch as an ex-voto to recover a lost animal ? Two smaller pieces of paper (E. i. 25. a, b, see Plate LXXIX), found rolled up together at the base of the north-west corner image, are mere fragments of larger sketches, one of them interesting on account of the Chinese look of the head represented.
But far more numerous were the small strips of textiles, sometimes shaped into little pennons, or else mere shreds, evidently torn from garments, which were found mixed up with the MS. pieces in front of the image bases. They comprised many fabrics, from elaborately-woven silk brocades to simple but strong cotton stuffs resembling the modern ' Khâm', and already referred to in Hsüan-tsang's account. Sometimes pieces of the same material turned
17 Among these coins from Endere, most of them fragmentary, three are wu-chu pieces, while the rest bear no legend ; see App. D.
Tibetan sgraffiti on cella walls.
Date of abandonment of temple.
Coloured sketches on paper.
Strips of textiles deposited as ex-votos.