National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

Sec. i]   THE RUINED SITE T. )(Iv   687

at vi on the south-west slope, where a cutting was found to have been made into the hard clay, as at ii, evidently to accommodate some quarters.

A particularly curious discovery was made at v, a spot about 4o feet to the south-east of the well and on the highest portion of the hillock. But for the topographical facts explaining the peculiar character of this site, it might have greatly puzzled me at the time. There, under less than a foot of gravel and earth, we found a layer of bundles of reeds about Io feet square which evidently once had served as the foundation for the floor of a small structure. That this was of later date than the rest of the remains at T. xiv was indicated by the fact subsequently disclosed that the layer of reeds covered an ancient refuse-heap containing fragments of Han records on wood. When the layer was being cleared, there came to light ten fragments of paper leaves, evidently belonging to some Chinese Buddhist texts, T. xiv. v. a, b (Doc., Nos. 710-19, Plate xxI). The paper distinctly resembled that of my manuscript finds of 1900 at Dandan-oilik, and the writing as well as the text that of the Chinese Prajna-paramita translation found at Khadalik 4 So, taking into account this evidence, I felt justified in concluding that these were relics from some modest shrine which had existed at this otherwise long-abandoned site during Tang times. I could recognize other relics of it in a number of fragments evidently from miniature banners, T. xiv. v. 003, 001 I. a-c (Plate CXVII), made up of fine silk fabrics, including damasks, with triangular tops and wooden stiffeners, such as I had found at the temple of Endere excavated on my first journey.°

This conclusion was strikingly confirmed when continued clearing next morning disclosed a small wooden bowl, T. xiv. v. ooi (Plate LII), embedded below the reed flooring and containing, besides two brazen hairpins, 005-006, eighty Chinese copper coins. With two exceptions, which were Wu-chu issues apparently of the first-second century A. D., all bore the legend Kai ynan, which we know to have been introduced in the first reign of the Tang dynasty (A. D. 618-27) and continued through a great portion of the Tang period.° These Tang coins showed scarcely any wear resulting from circulation.6a So the attribution of the fragments of the Buddhist text to the early part of the Tang period became highly probable. It was then that the route to Lop had last been an important line of communication, before the Tibetan invasion of the Kan-su marches, after the middle of the eighth century A. D., closed it to direct intercourse between China and the Tarim Basin.

What archaeological evidence had thus led me to conclude on the spot has since been fully borne out by M. Chavannes' examination of the fragments of Buddhist manuscript found at the little shrine. By a painstaking scrutiny and reconstitution of the texts found on the torn pieces of paper he has established the fact that the fragment, Doc. No. 710 (Plate XXI), contains the beginning of a Chinese translation of the Prajna-paramita Sutra, which is known to have been made by Hsüan-tsang between A. D. 645-64.' This constitutes, as M. Chavannes has duly pointed out, a terminus a quo for the manuscript remains of T. xiv. v, and proves that they certainly are later than A. D. 650 and probably belong to the eighth century A. D. Nine other fragments, Doc. Nos. 71I-19, belong to a Sutra text translated about A. D. 400 by the Indian Buddhist scholar Kumarajiva. Another larger piece, Doc.

Remains of Tang shrine.

Paper MS. of Buddhist text.

Discovery of Tang coins.

Buddhist MS. fragments of Tang period.

• See above, p. 164.

• Cf. Ancient Kholan, i. pp. 429 sq.; ii. PI. LXXVII, E. i. oz6, 017.

6 It is of interest to note that the only two other coins found at T. xiv, which came from the refuse-heaps of the hillock, are of a Wu-chu type attributed to the first century A. D. ; see App. B, xv.

I may here in passing call attention to the curiously small number of coins found along the whole of the Limes west of

Tun-huang; apart from the deposit at T. xiv. v, they are eight in all ; see App. B, xv. The difference from the great number of coins picked up at the Lou-lan Site is striking. I am inclined to explain it by the fact that the surface conditions of the ground, far less affected by wind-erosion, do not make it so easy to find the coins which were lost by ancient wayfarers, etc.

" See the specimen, T. xiv. v. ooio, PI. CXL.

7 Cf. Chavannes, Documents chinois, p. 152.