1296 FROM MAZAR-TAGH TO MARAL-BASHI [Chap. XXXII
SECTION II.—THROUGH AK-SU AND UCH-TURFAN
The eight marches by which I covered the 15o odd miles from Mazar-tagh to. the Tarim,
or Yarkand River, offered no opportunities for direct antiquarian observations. But as I passed down the steadily widening course of the Khotan-darya, between the numerous branching beds which the river from below Korla-ayaki (Map No. 25. c. 4) has formed at different times and in turn deserted as they silted up, I had before my eyes the best possible illustration of what the ancient deltas of the Keriya River and the Kuruk-darya in the Lop Desert must have looked like before
they finally dried up.1
Route along On April 28 we crossed the Tarim within a mile or so below the junction of the rivers of Yar-
Ak-ste R. kand and Ak-su. The former was almost dry at that season, whereas the latter filled a bed fully 300 yards wide and carried a considerable volume of water. The large size of the Ak-su River is explained by the great extent and relative nearness of the high snow-bed portions of the Tien-shan main range which it drains. Its headwaters stretch for a length of over four degrees of longitude from the western slopes of the great peak of Khan-tengri to the Terek Pass north of Kâshgar. But with this abundance of water available for irrigation there contrasted in a very striking fashion the scanty and careless cultivation which is carried on in the narrow village belt along the river's left bank. I had ample opportunities to notice this on the three long marches which brought us to the ` New Town' of Ak-su, and the recollections still fresh of the thriving' lands of Khotan necessarily deepened the impression.
Dolanpopu- That this undeveloped condition of what might become a large and flourishing tract could not
Ak-suat be due to an inadequate water-supply was here clear. In the end I was led to connect it with
a marked difference in the ethnic character of the population. This consists in the riverine parts of the Ak-su district to a very large extent of settlers of genuine Turk extraction, known as ` Dolans'. In speech, racial type, and original habits of life they appear to be.closely allied to the Kirghiz who occupy the grazing-grounds in the adjacent parts of the Tien-shan and are to be found also as cultivators in the valley of the Tushkan-darya above Uch-Turfan. That the Dolans who form the bulk of the population along the Yarkand River from above Maral-bashi to Ak-su are different in stock from the inhabitants of the oases to the south, east, and west is well known, and it is also
certain that. their conversion from semi-nomadic ways to settled agricultural life is of relatively recent date.2 The wave of migration which brought them from across the true Turk territories north of the Tien-shan into the Tarim Basin is not likely to have been an old one. Yet, as we shall presently see, the geographical factors which facilitated the Dolan immigration may help also to explain certain historical observations about Ak-su.
' Among various instructive features it will suffice to mention one. On nearing the Tarim there was striking confirmation of what I had repeated occasion to note before about the evidence afforded by lines of dead trees, or kölek, as to the direction which ancient beds, no longer otherwise traceable, must have followed. As we were striking across from the actual bed of the Khotan-darya below Zil (Map No. 24. c. 4) to the north-west towards the Tarim, I found an old bed of the former known as Ghaz-kum, though dry for many years past and further down completely choked by big dunes, still lined by living Toghraks growing on what were once its banks.
After passing about ix miles from Camp 376 all
trace of the old river-bed was lost. Yet, in the broad belt of bare drift-sand we had to cross further on, the lines of dead trees emerging between the dunes still kept the same direction from south-south-east to north-north-west as observed along the Ghaz-kum, until we had come within a few hundred yards of the belt of jungle lining the Yarkand River branch known as Körtiklik-akin. There the trees were all living and ranged in rows invariably rdnning from west to east and thus parallel to the river, the water of which accounted for their growth. The change of bearing was as sharp and sudden as if the alignment had been due to the hand of man.
2 Cf. Forsyth Mission Report, pp. 54 sq.