SECTION II.-THE VALLEY OF BESH-TOGHRAK
On the morning of March loth I was able to discharge the camels hired from Mirân and Discharge Charchan, which had helped us so well on our explorations in the Lou-lan region and along the of hired .
ancient Chinese route, and to allow their owners to return westwards to their homes. Along with them honest Tokhta Akhûn and Niâz, his young Loplik companion, now also took their leave, well pleased with the rewards that their useful services had earned them. Turdi, with my return mail and the antiques he was to carry to Khotan, was entrusted to their care as far as Charkhlik (Fig. 189).
The task immediately before us was to complete our examination of the line that the Lou-lan route had once followed beyond the terminal basin of the Su-to-ho, by a survey of the ground along the foot of the hill range on the northern side of the valley right up to its head at Bésh-toghrak. A special geographical interest attached to this ground in view of the relation, as I shall explain below, between the Bésh-toghrak valley and the Mesa-filled area adjoining it eastwards, which I believe to represent an earlier terminal basin of the Su-to-ho. It was on account of this interest that I had previously detached Surveyor Muhammad Yâqûb from Mirân to Kum-kuduk, with instructions to carry thence a line of exact levelling along the bottom of the valley to the nearest portion of that basin. In order to obtain full details about the topography of the northern side of the valley and also to ascertain exactly where the arm of the ancient sea that once filled its depression terminated to the eastward, I thought it advisable again to divide our party. I therefore let Lai Singh proceed on March loth direct to the north-east from Kum-kuduk, and with him I sent Afraz-gul, whom I could trust to keep a careful look-out for any antiquarian indications or physical features of interest.
I myself, with the heavy baggage, marched on the same day by the caravan track as far as the eastern extremity of the winding plateau at the foot of which lie the several wells known as Yantak-kuduk (Map No. 35. A. 4). In Serindia I have recorded the reasons which lead me to believe that the route along the southern side of the Bésh-toghrak valley, as marked by the present caravan track from Tun-huang to Mirân, was already in use in Han times.' It certainly was followed by Fa-hsien in A.D. 400, by Hsüan-tsang in A.D. 645, and more than six centuries later by Marco Polo.2 It is therefore of interest to note here that Abdurrahlm, before leaving Kumkuduk, handed me the well-preserved bronze arrow-head Kum. o1 (Pl. XXIII). It is different from the Chinese ammunition of Han times, but closely corresponds, in respect of its barbed narrow blades and the triangular depressions in the ferrule between them, to the type, probably indigenous, of the arrow-heads, Lai S. 015 and C. xcvi. o16 (Pl. XXIII), found near the Kuruk-daryâ, and also of one found at the Niya Site.3 He stated that he had found it on coarse sand when looking after
his camels at no great distance from our Kum-kuduk camp. The pottery fragment he had picked up near the same place, Kum. 02, affords no chronological indication.
Leaving the main camel train to follow the caravan track, I then struck off with light baggage to the north-north-east. After passing through fairly thick reed-beds for about two miles, we came upon ground covered with hard salt-impregnated clay lumps. As we crossed this shôr I could see that it extended to the west as a continuous, gradually widening belt, while eastward it came to an end within about a mile and a half or less, being completely edged in by reed-beds. Here then the eastern extremity of the arm of the sea-bed could be definitely determined. The belt of shôr
1 See Serindia, ii. p. 555. 3 See above, pp. 274, 279 ; Serindia, i. p. 250 ; iv. Pl.
2 Cf. ibid., ii. pp. 558 sqq. XXIX, N. xxv. oo8.