592 THE TUN-HUANG OASIS AND ITS NORTHERN LIMES [Chap. XV
The wall which the tower was intended to guard passed to the north of it with a bastion-like projection, keeping at a distance of nineteen feet or so from the north foot of the base. Erosion by wind and driving sand, the force of which we felt only too well amidst the bitterly cold blasts of these days, chiefly coming from the east and north-east, had long ago carried off all but the lowest layer of fascines in the wall (Fig. 161). Here they were made up entirely of tamarisk branches, a clear proof that the character of the vegetation on the adjoining ground towards the Su-lo Ho had undergone no great change since the time when the wall was constructed. But the ends of these lowest fascines cropped out so clearly from the overlying stratum of pebble-filled clay and gravel on the level flat that the line of wall, thus marked as a low but distinct swelling, could be followed by the eye with ease and to a considerable distance. The next tower to the east, T. xxxi, towards which the line of this a;ger ran quite straight, proved to be only one and a quarter miles away. Beyond it three more towers, T. xxxii—xxxiv, were within sight ; but I had to leave the examination of them till later.
Returning to tower T. xxvi, I had a close search made of the ground immediately adjoining it. Unpromising enough it looked, as the gravel surface was perfectly bare and level except for some clay débris fallen from the tower close to its east foot. But a kind chance provided encouragement at the outset. At a spot about twelve feet from the south-east corner of the base, careful examination of the surface showed slight refuse cropping out among the pebbles. After the ground had been scraped here (see Fig. 150), it proved to be the last remnant of the miscellaneous rubbish that once filled a small apartment about eight feet square. Of its walls, built with clay and faced with reeds and plaster, only traces survived. But even this shelter, scanty as it was, had sufficed to preserve relics of interest and obvious antiquity. The first to turn up, and almost on the surface, was a wooden tablet, T. xxvi. 1, over ten inches long and close on one inch in its actual width, bearing Chinese characters neatly inscribed in five small columns and a larger single line below them. The document, which Chiang Ssit-yeh at once recognized as part of an account, will be found deciphered and reproduced in M. Chavannes' Documents chinois, No. 702, Plate XX, and has proved to contain part of a multiplication table. No evidence of date such as I was eagerly looking out for was to be found either in this tablet or in two other fragmentary Chinese records on wood. One, T. xxvi. 3 (Documents, No. 703, Plate )(IX), with very clear writing, was part of a broken label which, as M. Chavannes' decipherment has shown, refers to a cross-bow and arrows of a certain military detachment ; the other was a piece from a record of the slip' type, so familiar to me among the Chinese documents of the Niya and Lou-lan Sites, showing merely two characters. A fourth record, also a ` slip', was less broken, but its numerous characters were rendered illegible by salt that had permeated the wood.
Chiang Ssit-yeh declared that the writing bore a strangely ancient look, and scanty as were these records, the mere fact of their material being wood, and their discovery at a spot of so little apparent promise, were enough to justify further hopes. The ground near the tower was scraped eagerly clown to the natural soil by the labourers, whom a prompt reward in silver had now roused from their torpor. But in addition to numerous pieces of broken pottery, all of black, well-burnt clay, marked on the outer surface with narrow parallel ridges due to the matting in which the ware was moulded, there turned up only a small wooden knob, painted black, T. xxvi. oo1, of uncertain use, and a much-worn shoe of woven hemp string, T. xxvi. 002, of which details will be found in the Descriptive List below. The pottery and the shoe have since proved to belong to types which I can now safely associate with Han times. But at the time I gave a more grateful welcome to the chronological evidence supplied by two Han copper coins of the Wu-chu type, much clipped and corroded, which were discovered adhering to each other at five yards' distance to the west of the
IIIF Color HighRes
