37o TO TUN-HUANG AND AN-HSI [Chap. X
the cart-track which our camels had followed to the south-east. After about four miles' march it brought us across shör-covered jungle ground to our camp, which we found pitched near the western edge of the cultivation of the hamlet of Êrh-kung.
Next morning we set out again northward in order to trace, if possible, the continuation of the Limes line beyond the ground where we had lost it. After having passed for about two miles through abandoned fields and along outlying patches of cultivation we came amidst luxuriant scrub upon a low mound which seemed to run in the direction of the last-mentioned tower. But the stretch along which it remained traceable in that direction was short, and I therefore decided to turn east towards An-hsi.
Here we moved across a scrub-covered waste which divides the scattered village lands of Ssti-kung and San-kung. R. Ram Singh had already passed through it in June, 1907, on his survey reconnaissance to the outlying western parts of the An-hsi oasis. After proceeding about four miles we reached the line of ruined towers, T. xxxv111. a—c, which, as my examination in 1907 • had shown, marks the line of the Limes where it runs south of the present An-hsi. A full description of this short but interesting section of the Limes, embodying also the supplementary observations I was able to make on my second visit, has already been recorded in Serindia.17 It will therefore be sufficient here to refer to Fig. 215, which shows the tower T. xxxviii. a of the Limes line preserved at a point, strangely enough, so close to a much-frequented highway.
With this renewed visit I completed the task which had brought me from Tun-huang, and filled up the last gap in our survey of the Han Limes so far as it lay to the south of the Su-to-ho. At An-hsi town I reached ground with which two prolonged stays on my former journey had already made me familiar. Once again this poor wind-swept place at one of the great crossways of innermost Asia was to serve me as a temporary base. I had therefore reason to feel gratified by the cordial welcome with which I was greeted on arrival both at my old temple quarters and at the modest local Ya-mên.
The same day R. B. Lal Singh safely rejoined me from the mountains. He had pushed his way up from Chien-fo-tung over the barren stony slopes of the westernmost Nan-shan to the foot of the high range that forms the divide between the trough of the Su-lo-ho and the elevated plateaus containing the head-waters of the Tun-huang river. But the snow steadily increased in depth as he approached the pass that might have given him access to the defile above the outfall of the latter, and stopped farther progress with camels. He reached, however, a height sufficient to enable him, as Map No. 39. B—D. I shows, to complete satisfactorily the survey of the great mountain barrier which forms the southern ruin of the lower Su-lo-ho basin. He also cleared up the orography of those outlying hill chains which we had crossed in 1907 to the south of An-hsi. He then made his way down past the little oasis of Tung-pa-t`ou (Map No. 38. D. 4) to T`a-shih, and finally surveyed the previously unknown ground over which the drainage from this big valley descends towards An-hsi. Two days later our concentration was completed by the arrival of Surveyor Muhammad Yâqûb, who had descended the Tun-huang river to its junction with the Su-lo-ho and then crossing the latter to its right bank had carried his survey up to An-hsi.