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. ii]   IN SEARCH OF THE ' OLD WALL' NORTHWARD   591

first given expression, and of a belief which a series of antiquarian observations had steadily helped to foster in me, though in a somewhat different form. But there was no evidence yet as to the correctness of the dating which the Tungan's statement implied, nor even was it possible to make sure whether it was derived merely from a shrewd guess or from some genuine tradition. He did not claim himself to have followed this ` old Han road ' much further west. All he would assert was that an elder brother of his, since trading at Hsi-ning-fu, had taken a strange, i. e. non-Chinese, kuan, or official, along it soon after the old route from Tun-huang to Lop-nor was reopened. This detail makes it appear to me very probable that the foreign traveller meant was the ill-fated M. Martin, who, as related above, was the first European in modern times to make his way through the desert from Tun-huang to Abdal and Khotan, but did not live to record his story.8

Far away to the north-north-east and beyond the river there was visible a large group of ruined buildings, of which our Tungan spoke as heathen temples. Owing to the flooded state of the river, they were not accessible from our camp. Later information leads: me to believe that these ruins were of recent origin, belonging to an abandoned roadside station on the route to Hami. My endeavours to retain the hardy Tungan as a guide in these regions proved of no avail. As related elsewhere, he soon left us with a promise to return in the morning, which he did not keep, and was then searched for in vain. Thus there vanished from my horizon for good the only man who could or would tell me of the ' old wall '.

Ruined station on Mimi road.

SECTION III.—FIRST DISCOVERY OF DATED HAN RECORDS AT T. xxvn

On the morning of March 27 I set out with my Indian assistants and half a dozen Chinese labourers to the south-east. There I hoped to strike approximately the middle of the line of towers I had sighted. From the low-lying, scrub-covered plain across which we had to move they were invisible. But another conspicuous clay terrace helped to guide. On reaching it after some two and a half miles, we found on it a small troglodyte dwelling, occupied by an old Chinese woodcutter.' He professed, as was to be expected, total ignorance of old towers and everything else. Beyond this, a belt of exceptionally thick scrub and low tamarisk-cones, among which a small channel coming from the Su-lo Ho was steadily spreading inundation, impeded both view and progress. At last we emerged at the foot of a gently sloping, gravel-covered ` Sai ' with much dead wood on the ground and a few stunted Toghraks still alive. Then, in front of me, I saw rising the truncated cone of an old watch-tower, in shape and construction just like those previously examined in the desert westwards. As I galloped towards it, my eye soon caught the line of a low mound, with the familiar fascines exposed on the eroded surface, stretching away across the bare gravel to the nearest tower on the east, and continuing also with a divergent angle to the south-west. Then I felt quite assured that I was back again on my ` old wall ' or chic c/t'ianç, as our Chinese soon learned to call it.

The watch-tower, 'I'. xxvi (Fig. 150), was strongly built of regular courses of hard clay, each from 32 to 4 inches in thickness, between which were embedded the characteristic thin layers of tamarisk brushwood. The four faces of the solid square structure slanted slightly inwards. In order to give additional cohesion, the stamped clay had been reinforced by the insertion of vertical posts, which were probably joined up within by other timber, and of which one was conspicuously exposed on the top (see Fig. I so). At its base the tower measured approximately twenty feet square, and its height, in spite of the broken top, still rose to about twenty-five feet.

$ Cf. above, p. 565.   ' This point is marked in Map No. 8i. A. 3 by our subsequent Camp 166.

Limes wall
discovered
at T. xxvi.

Watch- tower T. milt.