Sec. iv] AFRAZ-GUL'S SUPPLEMENTARY SURVEYS 743
drainage bed in which they are situated continues for a considerable distance to the north-west, and may be assumed to be connected with the large salt-encrusted depression which the routes from Turfân to Singer cross at Arpishme-bulak (Map No. 29. B. 1).
Scrub was met more frequently also on the next march. Leading SE. over stretches of gently rising gravel Sai it brought the travellers to a wide salt-encrusted bed fringed with tamarisk-cones, and beyond it to the point known as Bakri-changche. The spring marking it appeared to have dried up years before. Proceeding south from this point on February 14th the main range of the Kuruk-tâgh trending from the direction of Singer la was traversed on an almost imperceptible watershed. Beyond it the route led down a wide drainage bed, lined with scrub, which forms the head of the one passing Altmish-bulak. At the salt spring known as Kuruk-toghrak-bulak from a dead wild poplar close by, a night's halt was made, reeds and other scrub affording fair grazing for the camels. Thence following the bed down to where it passes in a narrow defile through the outermost hill range overlooking Altmish-bulak, this little oasis of desert vegetation was reached by February 15th.
It is clear that the route described above from the surveyor's account, though the most direct between Turfân and ` ancient Lou-lan ', can never during historical times have claimed importance as a line of regular communication between the two territories. The routes leading through Singer must always have been preferable owing to easier access to drinkable water and to grazing.
At Altmish-bulak the camels were left behind under Hassan Akhun's care for a much-needed rest. Afraz-gul with the two remaining men set out on the morning of February i6th for the first task indicated in his instructions, the examination of the remains to the NNE. of the ancient castrum L.E. which on the previous year's visit had been left unexplored.2 Four days' food and ice rations, besides plane-table, Ketmans, and other indispensable outfit, were carried by the three on their shoulders. The route taken was the same by which on February 25-26, 1914, we had gained this vicinity.3 Early on the morning of February 17th the little party arrived at the Mesa bearing the small burial-ground, L.Q., noticed by Afrdz-gul on his first reconnaissance (Map No. 32. A. 3).
The Mesa, c. 45 ft. high and 30o yards long at its foot, showed a surface of salt-encrusted clay. Its top bore a number of graves marked by closely set pieces of wood after the fashion previously noted at the graves of L. F.4 In the majority of cases the bodies and the coffins containing them were found badly decayed. But the finds made in the few better-preserved graves, together with the character of the remains surviving in the rest, made it quite certain that the methods of burial were identical with those observed at the indigenous burial-ground of L.F. In grave ii the body was found badly decayed ; but at its foot an interesting object was recovered in the shape of the carved wooden figure of a female, L.Q. ii. 01 (Pl. XV). It closely resembles the images of similar archaic type in wood and stone, L.T. 01 and L.S. 6. o1 (Pl. XXVI), recovered from graves on the Kuruk-dary1.5 In grave iii the rough coffin was found covered with narrow wooden boards and above them with sheepskins. The body was wrapped in a thick woollen shroud ; the head was that of an old man, with red moustache and without beard. The specimen L.Q. iii. o1, taken from the portion of the shroud covering the head, shows a small bunch containing broken twigs as recovered also in L.F. 1.03 and L.S. 3. 01.6 Other objects, corresponding to finds from L.F., are the woven grass basket L.Q. iii. 02 ; the felt head-dress o3 ; the wooden pins 04-9 (Pl. XXIV).
The graves traced, among which several were of small children, extended over a distance of about 4o yards. Some 20 yards to the NE. of this area there were found on the surface the fragment
la See above, ii. p. 722. s Cf. above, ii. pp. 735, 737.
2 See above, i. p. 267. 6 Regarding these twigs of the Epkedra plant, cf. above,
3 Cf. above, i. pp. 283 sq. 4 See above, i. pp. 264 sq. i. p. 265, note ioa; ii. p. 736, and Add. and Corr.