752 ON THE ANCIENT ROUTE ALONG THE KONCHE-DARYA [Chap.XXI
the irrigation of their fields could most appropriately and conveniently proceed to pray for an adequate supply of water to fill their canals ; for the point where the Shindi river actually debouches from the foot-hills to make its way down to the Ying-p`an site across the thirsty gravel glacis is, as shown by Dr. Hedin's map, fully ten miles away, a distance too great both for cultivators and priests to attend upon shrines.6
From the top of the little plateau the remains of two more Stûpas were visible to the north-east. Proceeding there along the main flood-bed of the river, I found that about three-quarters of a mile above the ruins of Ying. i it divided into two shallow channels containing scrub and tamarisks. Between them ran a narrow tongue of raised ground, and on this we soon came upon distinct traces of an old canal. Its banks, worn down by the winds, were built of gravel and clay, just as Abdulmalik said he had seen them higher up near Bêjàn-tura where the river-bed debouches from the hills. Its bottom still showed up clearly as a band of hard clay, 4 to 5 feet wide, with boulders placed along it here and there to strengthen the canal banks. The direction of the canal was traceable also by the stumps of tamarisks which had once grown up along its sides. We followed it without difficulty for over half a mile and then crossed the eastern channel to its left bank.
Of the two small Stûpas which we reached at this point, the more southerly had decayed into a mere mound of brickwork, about 17 feet across at its debris-covered foot and about 13 feet high. But from about 3 feet upwards the circular shape of the superstructure could still be made out. The second Stûpa, standing about 40o yards farther north, was somewhat better preserved. It had a base i 5 feet square and close on 7 feet in height, surmounted by a dome rising 82 feet more to its broken top. The masonry in both Stûpas consisted of sun-dried bricks, very hard and containing but little straw, measuring about 15" X 12" X 3". In the northern one the lowest portion of the base showed two layers of stamped clay, about 6 inches thick, separated by a single brick course, just as in the base of the Stûpa Ying. i. The north-eastern foot of the base showed some undercutting by wind-erosion ; but the general level of the gravel-covered surface had not been lowered since the construction of these Stûpas, though exposed to the full force of the winds sweeping across the bare Sai. There was nothing now to suggest the particular reason why the Stûpas had been erected just at this spot. On ground so remarkably uniform as the Sai of this alluvial fan, the position of canal heads would necessarily be shifted with the lapse of centuries, and the abandoned canal we had previously traced on the opposite side of the flood channel did not appear to me of great antiquity.
That the site of Ying-p`an, in fact, had been reoccupied at a more recent period was a conclusion that had been already suggested by Dr. Hedin's notice of a Muhammadan burial-place to the north-east of the circumvallation to be described farther on. So I was not surprised when on our return towards the ruined shrine Ying. i we came, at a point about three-quarters of a mile above it, upon a number of graves, manifestly Muhammadan, occupying a small clay terrace below the edge of the Sai and on the right bank of the united flood-bed. There were about thirty-three of them, all marked by small oblong mounds of kisek or brick-like lumps of hard clay and orientated from north to south in orthodox Muslim fashion. Where some recent flood had cut into the edge of the small terrace two or three lay half open. In one the head of the body was seen clearly turned to the west, towards Mecca, and in none was the coarse cotton of the shrouds decayed, notwith-
6 For that actual debouchure, see Plate 19 in Hedin, Central Asia, Maps, vol. i. In the name Budschentubulak there marked we have another instance of the transformation which the name Bèjân-tura has undergone in careless Turfânlik pronunciation ; cf. above, ii. p. 720, note 3. Abdulmalik clearly indicated to me Bèjân-tura as the true name of the
place, also known as Aghiz-aghzi, ` mouth of the gorge'.
I may note here that the conjectural representation in Map No. 25. D. 2, of the Shindi defile below Inkur-otak, the lowest point reached by me, is an error of compilation which, I regret, escaped my attention when correcting the hill drawing of this sheet.