National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

Sec. ii]   IN SEARCH OF THE `OLD WALL' NORTHWARD   589

in the towers guarding the old wall west of the oasis. The tower, measuring about twenty feet square at its base and rising to over eighteen feet, formed a remarkably compact mass, probably on account of the cementing effect of the salts which had been absorbed in the material. Its solidity, as well as its age, could be gauged from the way in which wind erosion had carried off the natural clay beneath the corners without any injury to the overhanging masonry.

Of the hoped-for line of wall there was, however, no trace to be seen here. Nor did I notice any old remains as we pushed on to the north-north-east, where I wished in the first place to locate the course of the Su-lo Ho. The ground we crossed for another four miles was covered with rich scrub and tamarisks, and two rough enclosures, made of very hard lumps of salty clay and evidently intended as sheep pens, showed that it had been used at one time for grazing. From the top of a low and narrow clay ridge then encountered I first sighted on the north a wide marshy expanse, indicating approach to the river. In the midst of it a succession of clay terraces was ranged in rows, all striking east to west. It was a very instructive sight, as it recalled at once the eroded clay Mesas I had passed in such numbers within the dry terminal basin of the Su-lo Ho, north of the present one, and again in the vicinity of the Khara-nôr.6

It soon furnished also the manifest explanation of their origin. A mile or so further on we came to the first freshwater lagoon, and then had to ford a succession of shallow watercourses, all flowing westwards and fed from the river. It was easy to realize here the cause which had determined the bearing of the rows of clay terraces. It was clearly the action of flood-water which, working on the bottom deposits of an earlier and wider alluvial bed, had carved out ridges parallel to its own line of drainage from east to west. Subsequent erosion by winds blowing mainly from the north-east had cut up these ridges into rows of isolated terraces, and this scouring still continued on the bare clay surface raised above the flood level. Here I had a striking illustration of that very process of erosion, by the combined forces of running water and wind, to which I had been led to ascribe conjecturally the formation of those strange Mesa ` witnesses ' previously found after leaving Besh-toghrak in basins now wholly or partially dry. We shall see what careful use had been made by those who constructed the Limes line of the excellent positions which these high clay terraces furnished for watch-stations. Hence this quasi-geological explanation of their origin may well find a record here.

For two and a half miles beyond the first clay ridge I succeeded in pushing on northward across belts of boggy ground and a network of shallow flood channels running between the chains o f clay terraces. After I had crossed with difficulty a channel about twenty yards wide and 4-5 feet deep, in which the water flowed briskly, the main course of the Su-lo Ho came in. sight at last, marked by a wide sheet of ice. But the ground had long before this proved quite impracticable for laden camels, and the intention of sending the Surveyor across the Su-lo Ho, in order to have the whole of its terminal course mapped from the right bank, had to be abandoned.? Camp was pitched by the side of the first lagoon reached, and next morning I retraced our route to the southernmost clay ridge, and thence started eastwards in search of the continuation of the Limes line that I was eager to locate. The ruined remains which Captain Roborovsky's map marked by the side of the route to Hâmi, and which, I conjectured, might possibly have some connexion with it, still lay a considerable distance away to the east and could not be expected to guide us. But within two miles or so in that direction rose a conspicuous tower, already sighted on the previous day's march, and to that I now led my party.

The reed-covered steppe which had to be crossed to it showed traces of fields and irrigation cuts

8 See above, pp. 552, 575 sq.   my instructions by R. B. Lai Singh, but in the reverse

7 This task was in 1914 successfully performed under   direction.

Approach Su-lo Ho.

Origin of clay

terraces.

Stopped by Su-lo Ho inundation.