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. ij   REMAINS BETWEEN TUN-HUANG AND NAN-HU   609

completely dry. Parallel to it, on the east, there extended a network of wind-eroded clay terraces, marking what in ancient times may have been fertile land. To the west of it the cultivated area extends even now much further south, receiving irrigation from a main canal which takes off from the Tang Ho some thirteen miles above Tun-huang-hsien. A modern watch-tower by the roadside, together with its adjoining quarters half in ruin (Fig. 167), which was reached after some six miles from Chên-fan-hsien, was of interest to me as an illustration of what the ancient watch-stations along the Limes might have looked like at one time.

To the west of it, across the bare gravel ` Sai ', my attention was attracted by remnants of walls

rising here and there above the plain. Riding across the canal towards these walls, I soon noticed that they invariably flanked what seemed to be gateways to large rectangular enclosures marked by low mounds of gravel. It was impossible to believe that these low mounds represented completely decayed walls when I saw that the walls on either side of the gateways still rose quite solid to a considerable height. In the two enclosures that I was able to examine more closely, these flanking walls were from five to seven yards long and rose to a height of 18-2o feet, with a thickness of eight feet. The bricks in their well-set masonry were sun-dried but massive, and of distinctly ancient appearance, measuring 17 by 10 inches, with a thickness of 10 inches. Beyond this the front of the enclosures, usually facing south, as well as the other sides, were merely low ridges of gravel. The relief they presented under the slanting rays of the setting sun made them just perceptible.

Along the line of these ` walls ' I could nowhere find any traces of brickwork or even of the

familiar fascines. One of the quadrangles measured about 75 by 70 yards. The enclosing ridges were invariably orientated, though but roughly. Within the enclosed areas there were always to be found several low tumuli, the largest usually facing the entrance from the north, and the rest scattered in small groups roughly aligned. In one quadrangle the tumuli were from about 21 to 45 feet in diameter, with a height of from five to six feet. In another some of these roughly circular mounds rose to a height of about eight feet.

The local Chinese with us, of course, professed complete ignorance as to the character of the

remains. But the idea that these were ancient places of burial soon occurred to me and to Chiang Ssû-yeh as well. Neither of us. knew at the time of a closely corresponding practice, old or modern, which would support this surmise. So it remained for my explorations of 1915 at ancient cemeteries near Kara-khoja and other sites of the Turfân district to furnish me with definite evidence of its correctness.' There I found an extensive series of Chinese tombs belonging to Tang times that had been cut into the hard clay of alluvial fans presenting the same surface appearance as this ` Sai'. Their position was similarly marked by low circular mounds within rectangular enclosing ridges of gravel ; only the walls flanking the entrance were absent. From what I subsequently observed at a much smaller cemetery near Ying-p`an, at the foot of the western Kuruk-tâgh, I am inclined to infer that the custom of arranging burial-places in this fashion, with groups reserved perhaps for particular families, etc., may date back to an earlier period.

On this question, as on other antiquarian points connected with my observations in Kan-su,

competent Sinologues could probably throw light from Chinese literary sources. What mattered for me at the time of surveying these remains near Tun-huang, and also soon after ne tr Nan-hu, was the practical certainty that, if the mounds proved to mark burial-places, I should not be able to get Tun-huang people, particularly orthodox in their superstitious awe of graves, to help in the systematic opening, or even to tolerate it on the part of foreign ' barbarians'. Any attempt of this kind was bound to produce local ill feeling against us, if not worse, and this was likely to interfere seriously with more attractive and fruitful archaeological operations of mine in the desert and elsewhere. So

' See Third Journey of Exploration, Geogr. Journal, xhviii. p. 204.

Gravel mounds with walled gateways.

Tumuli within enclosures.

Ancient places of burial.

Difficulties about exploring graves.

1314   41