国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

Sec. iiii   THE ANCIENT REMAINS OF NAN-HU   617

which were picked up by Chiang Ssti-yeh in the course of a careful search, mainly to the east of the ruined town, and which can be recognized, there are, as shown in Appendix B, only a single Sung coin with the nien-hao of A. D. 1038-4o, nine coins with the legend Kai yiiait belonging to the Tang period, and no less than eleven which certainly belong to pre-Tang issues. It is interesting to note that these last, besides three pieces of Wang Mang's issue of A. D. 14-19 and three Wu-chu coins, comprise a copper coin bearing the legend Pan-liang (' half an ounce ') which is of a type of the second century H.C. not otherwise represented in my collection.

Before I discuss the identification which Chinese learned tradition assumes for the site of the ` old town ' of Nan-hu, and which, as we shall see, receives much support from my archaeological observations and finds, it will be convenient to notice what other old remains I examined at and near the site. After moving on to the north-east for about three-quarters of a mile across the Ku-lung-ran' Tati ', a ruined mound is reached which obviously marks the position of an ancient watch-tower. It measures about twenty feet square at its base and, built with carefully stamped layers of clay about 2-2â inches thick, still rises to some twelve feet in height. It is known to the Nan-hu people by a distinctive name,2 and was said to have stood by the side of the old road to Tun-huang where it crossed the dry river-bed already mentioned towards the western end of the ancient embankment on the ' Sai '.2 That road was declared to have remained in regular use until the great flood of 1893 had, as stated above, here transformed the dry bed into a deep ' Yâr' and made its passage impossible for cart traffic. I found in fact the actual bed cut into the soil to a depth of about fifty feet and the very steep banks showing clearly its recent formation. I may note in passing that the well-marked stratification, observed in this cutting, of alternate layers of red alluvial clay and of sand or fine gravel gives plain evidence of a succession of wet and dry periods which must have affected the formation of this alluvial fan during geological times.'

Continuing to the north-east for another mile or so across a sandy area, where growth of tamarisks and reeds hid more ' Tati ' remains and the line followed by the old cart track was still traceable in places, I reached the margin of the riverine depression. In a conspicuous position above the edge of the bare gravel plateau rose the ruined watch-tower which I had already noticed on my first approach to Nan-hu. Manifestly old in its main structure, built with solid ftisé layers of three to four inches in thickness, it showed plentiful repairs of relatively modern look, executed in sun-dried bricks of small size. Its base measured thirty-six feet four inches square, and its height twenty-two feet. My ' guide ', an old village headman of Nan-hu, who in time grew somewhat less secretive than the rest, declared that the tower had until about seventy years before my visit been used for a post guarding the route. A small domed structure, badly decayed, which I found close to the north-east of the tower, together with a large heap of refuse, seemed to bear out this statement. The ancient embankment, which the road had followed from the great bend of the Tang Ho, was clearly seen to end at the tower, and this, in conjunction with what has been observed above as regards the track crossing the ' Tati ' towards the ruined town and used as the route to Tun-huang down to 1893, makes it appear practically certain that there existed a close connexion from early times between the embankment and the direction of the road from Tun-huang.

The point merits special consideration with regard to the question as to the origin and character of the embankment. According to the local belief, as communicated by my informant, it was intended

Remains at eastern edge of ' Tati'.

Ruined watch-tower at end of embankment.

Origin of ancient embankment.

The name sounded like Pan-chi-tun, but I regret that I did not obtain a record of it in Chinese characters. Hence its transcription is doubtful.

3 See above, pp. 6 r o sq.

' About half a mile further up I noted the following layers

1s„

in the banks of the river-bed, there from 12 to x5 feet deep : red clay at bottom, 2-3 feet thick ; coarse rubble, i foot ; fine gravel, 2 feet ; red clay, 6 inches ; coarse sand, 2 feet ; red clay on top, 4 feet.

4

K