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

Sec. iv]   FIRST REMAINS OF THE OLD CHINESE LIMES   567

importance, as indicating the line of ` the great route, vainly sought after till now, which, under the Han dynasty, ran to China through Bactria, Pamir, Eastern Turkestan, the Desert of Gobi, and Kan Suh'.3 But his passing notice could not help me, in the absence of any map or sketch of route, to locate the remains beforehand.4 Fortunately I had been able to inquire about them, before leaving Abdal, from Mullah, the true pioneer of the reopened route, and the information recorded from the mouth of my observant old guide gave me hope that I might come across the first ` P`ao-t`ais ' on this march to the spring of Toghrak-bulak.

This hope proved well founded. The first tower-like mound (T. n in Plate 33) lay too far off to the north to be noticed until we had passed by it. But in the second mound, T. i, approached after another couple of miles, I could easily recognize an unmistakable and relatively well-preserved watch-tower. About its antiquity I felt no doubt when I found the familiar layers of tamarisk branches inserted at regular intervals between the courses of stamped clay.

The tower rose in an easily defended position, flanked by small, steeply eroded Nullahs, on the very edge of the gravel plateau where it fell off with precipitous banks to a winding depression. This was fully a mile wide here and contained what was an unmistakable old river-bed, running to the north-west in the direction of the Wadi through which the route had emerged from the dry terminal basin. An extensive reed-covered belt stretching along the bottom and a string of salt-encrusted pools, then dry but still holding moisture beneath their surface, showed that water at times must make its way into this old channel. Its bottom lay fully eighty feet or so below the level of the gravel ` Sai '. Close to the west side of the tower were the foundations of a small and badly-decayed structure which, as I thought, might have been the watchmen's quarters. Small fragments of iron, apparently from implements, and of carved wood, together with a piece of some stout woollen fabric, were found on the slope below and confirmed this conjecture. This first relic of human activity in the desert was enough to raise cheering thoughts of fresh archaeological work before me. But for closer search there was no time then, and it was in darkness that I hastened after my caravan along the track, now luckily well marked in the gravel.

The camping ground of Toghrak-bulak was reached about three miles further on. It proved to be situated in a narrow, sharply scarped valley where reeds and scrub grew plentifully by the side of a hard-frozen marshy streamlet. This seemed to be spring-fed, as the name Toghrak-bulak, used by our Loplik guide, suggested, and further evidence was given by the presence of Toghraks, some dead, some further north still alive. But there was nothing to make me realize at the time, even in the light of the morning, that this was the bed of a live river which within less than six weeks would become almost impassable.L

In my eagerness to get at more ruins, I secured an early start on the morning of March 8. Proceeding by the caravan track eastwards across an absolutely barren plateau of gravel, we had covered about three miles when, on a low ridge to the south-east and apparently not far off, I noticed rising what looked like a small ruin. It was the one now marked T. in in Map No. 74. B. 3 and Plate 33. So, sending the caravan ahead to the springs where our next camp was to be, I approached it with Chiang Ssû-yeh and a few of the men provided with Ketmans. The distance

First ruined watchtower.

Old river-
bed below
tower T: r.

Su-lo Ho bed at Toghrak- bulak.

Tower T. III sighted.

' Cf. Bonin, loc. czl., as quoted by M. Cordier in Yule, Marco Polo, i. p. 203.

' Nor are any indications of these ruins to be found in the maps showing the surveys of Captain Roborovsky's expedition.

6 As evidence of the very deceptive nature of this ground it may be mentioned that the map attached to Captain Roborovsky's Report shows at Toghrak-bulak a small stream

losing itself apparently in a depression on the north. Further to the east, detached marshes are shown by the map in places. But of the existence of the true terminal bed of the Su-lo Ho, which flows out of the Khara-nôr and, connecting those marshes, continues its course for about seventy miles further west than previously assumed, no indication could be found in that expedition's surveys.