428 THE ENDERE RUINS [Chap. XII
tained by the Tangs 13, these finds of Chinese documents would suffice to show that the Endere temple and the structures surrounding it must have been abandoned not later than the end of the eighth century.
Fortunately a dated Chinese record inscribed on the walls of the shrine itself supplements this terminus ad quern by more precise chronological information. I mean the Chinese sgraffito in three rows of characters which the clearing of the west wall of the cella brought to light close to the left of the north-west corner image, as seen in the photograph reproduced in Plate XI. Owing to the decay of the upper portion of the wall the top of the second and third rows is lost, but luckily the shorter first row on the right, containing the commencement of the record,
has escaped mutilation. As shown by M. Chavannes' translation in Part iii. of Appendix A, it states the date when the record was inscribed as the seventh year of the K`ai-yiian period,
or 719 A. D. Owing to the original surface of the rough and friable wall plaster, into which the characters had been scratched with some blunt-pointed instrument, having peeled off in parts the first character Ug Kai has become indistinct. It was, therefore, at first thought .
possible that it might read fi (Clzêng yüan), which would have made the year intended correspond to 791 A.D. But M. Chavannes' reading of it has been confirmed by three Chinese
literati as well as by Dr. Bushell, who all, after careful examination of my photographs, arrived independently at the same reading, and it is supported by internal evidence in the text of the sgraffito itself.
The latter, as far as it can be interpreted in its incomplete state, distinctly mentions the ` Four Garrisons' and the ` Great Fan ', i, e. the Tibetans, as well as the latter's officers. It
further records that ' the high dignitary of the tai-clz`ang (court of sacrificial worship), Ch°in
Chia-hsing, returned to the district placed under his orders '. Whether it was this personage or some other Chinese dignitary who, in a preceding passage, is said to have ' heard that his
commissioner of troops and of the cavalry was dead ' must remain uncertain. Nor has
M. Chavannes been able to discover Chin Chia-hsing's name in Chinese historical records. We have seen above that in 790 A. D., or at the latest in 791 A. D., Chinese authority in
Eastern Turkestdn, after having from 'about 766 A. D. maintained itself only with difficulty and
in complete isolation from the rest of the empire, finally succumbed to Tibetan invasion 14. It seems difficult to believe that in the very year when the effacement of Chinese dominion had
been completed and the ' Four Garrisons ' had ceased to exist in name as well as in reality,
a Chinese dignitary should have recorded on the walls of a Buddhist shrine on the confines of Khotan his return ' to the district placed under his orders '. On the other hand, we know
that from 714 A. D. onwards the Tibetans had been annually harassing the Chinese borders,
and that in alliance with the Arabs and the rebellious Western Turks they had invaded the Four Garrisons ' about 717 A. D.15 From 719 A. D. onwards, however, a series of records in the Tang Annals attest the gradual reassertion and extension of Chinese power in the Târim
Basin and the neighbouring regions 16. These successes of the Emperor Hsüan-tsung's policy seem to have been attained rather by diplomacy than by armed power. Possibly we have some trace of this in the respectful reference our sgraffito makes to the ` Great Fan ', the very title by which the Tibetans designate themselves in the Lhasa inscription of 822 A. D. recording a treaty with China.