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

568   THROUGH THE LOP DESERT TO TUN-HUANG   [Chap. XIV

proved greater than estimated ; for what had seemed a flat level expanse of ` Sai' was found to be broken by a broad sandy depression containing luxuriant tamarisk scrub and plenty of dead Toghraks. Wheel-marks, the first seen since leaving Khotan, showed that Chinese from the Tunhuang oasis were apparently coming so far in search of timber. At last, ascending a steep scarp of about a hundred feet to the edge of the gravel plateau that skirted the depression along the south, I found myself at the ruin (Fig. 149).

It proved to be a solid square tower (see plan in Plate 36) with the faces receding towards the top, which still rose to a height of about 20 feet. At the base it measured 16-17 feet square. The solid masonry consisted of carefully set courses of hard sun-dried bricks, measuring, as at the first noticed tower, T. 1,14 by 7 inches, with a thickness of 5 inches. At regular intervals, each comprising three courses, thin layers of reeds were inserted between the masonry to give increased cohesion. The ruined watch-tower, for only as such could it have been intended, had suffered little damage in its brickwork, and even that was confined mainly to its northern face. On the south, the side least exposed to the winds which, as experience showed us, sweep this bleak expanse for the greater part of the year, the plaster covering of the brickwork still survived in places. There, too, were found some scanty traces of a broken wall adjoining. Apart from these, there was no indication of any structural remains except the solid tower itself, nor were any brought to light by a subsequent clearing of the ground, which spread its uniformly flat gravel surface all round, and here and elsewhere showed practically no trace of wind erosion.

But as I was carefully examining the neighbouring ground, my attention was soon attracted by a straight line of bundles of reeds, with their ends cropping out from under a slight swelling of the gravel soil, about twenty yards north of the tower and close to the edge of the plateau. I traced this line eastward without much difficulty to the top of a small knoll near by. Arrived there, I could see the line stretching away perfectly straight towards another tower (T. v11) visible some three miles to the east, and assuming the form of an unmistakable wall where it descended rather steeply on the gravel slope to a lower terrace of the plateau and then to the depression already mentioned. It was clearly part of that early ` Chinese wall ' for which M. Bonin's observation further east had made me look out, and a little digging on the low knoll, shown by Fig. 149 in the foreground, soon proved that I actually stood on remains of it. By clearing away the gravel and fine drift-sand which had accumulated below on the side sloping towards the depression northward, there was revealed a regular wall or agg-er constructed in a fashion which at first sight seemed strange indeed, but with which I was soon to become very familiar.

Here the wall had been built with alternate layers of bundles or fascines of reeds and of clay mixed with gravel. The former layers were 2 to 4 inches, and the latter 6 to 7 inches, in thickness. The bundles of reeds were placed horizontally and always at right angles to the direction of the wall. Their length was quite uniform and close upon seven feet. Their original thickness was here difficult to determine, as the accumulated weight of the successive layers had greatly compressed, and in fact almost solidified, the fascines. Yet each individual reed still retained full flexibility and, apart from the salt permeating the fibres, showed no obvious mark of great age. On the outside there was a revetment consisting wholly of fascines of reeds, fixed in the direction of the wall, and hence at right angles to the packed bundles within it. These fascines, which from their position could be detached and examined more easily, were found to show the same uniform length of seven feet, with an original thickness up to about eight inches. They were bound with bark twists at intervals of about six inches, and appeared to have been ` anchored ' by means of ropes of tamarisk twigs passing right through the thickness of the wall.

The preservation of these facing, or revetting, fascines was obviously due to the protection

Con- struction of watch-tower T. ill.

Discovery of line of wall.

Reed fascines used in agger.