202 RUINED SITES NEAR DOMOKO [Chap. V
the main oasis of Domoko and investigating the peculiar irrigation conditions which have played an important part in its modern history. As the changes produced by them offer distinct interest for the archaeological student no less than for the geographer, the observations already set forth in my personal narrative may justly here also receive a record.'
When in March, 1901, I first touched the northern edge of the Domoko tract I was struck at once by the shifting which its irrigated area had undergone within living memory. In the zone
overrun by desert vegetation and light drift-sand, to the north-west of the outlying colony of Malak-
âlagan, I had then come across the abandoned fields and dwellings of an extensive village site known as ` Old Domoko '.2 From there, according to the villagers' uniform statements, the area of cultiva-
tion had been transferred to its present position, about nine miles farther south, near the Khotan-
Keriya high road, only some sixty years earlier. The gradually increasing difficulty of conducting the irrigation water sufficiently far was declared to have been the cause of the migration. Whatever
was the origin of that difficulty, it was clear that the migration had served to bring the cultivated area of Domoko nearer to the springs in which the water of the mountain streams about Nitra and Tört-Imam comes to the surface again at the foot of the great glacis of piedmont gravel southwards. Upon these the oasis entirely depends for irrigation before the summer floods bring down the ak-su (` white water ') of the melting glaciers and high snow beds.
It was thus tempting to connect this transference with that general process of desiccation or gradual drying up of the climate which, as is becoming clear to competent observers, has affected
the physical conditions of Central Asia very extensively during the historical period. No one has
done more to investigate this great physical change than Professor Ellsworth Huntington who has studied its problems over widely distant parts of Asia as well as in Southern Europe and on the
American continent. With this object in view also he devoted several weeks in the autumn of 1905 to the systematic examination of the physiography of the plain between the Khotan and Keriya Rivers, and did not fail to call special attention in his fascinating volume, The Pulse of Asia, to this well-authenticated change in the area of cultivation.3 Shrinkage in the amount of available water-supply was accepted by him as the obvious explanation.
Yet there were not wanting indications of special local conditions which might have influenced the change. I remembered what a tradition heard on the occasion of my first visit in 1901 asserted
about such transferences of the cultivated land, backwards and forwards, having occurred repeatedly
in the case of Domoko. The opening of the Malak-alagan colony, fully six miles to the north of the centre of the present main oasis and dating back only to about 189o, looked curiously like an
illustration of the reverse movement northward having already set in. When stopping there on my
way to Khâdalik, I learned that the number of settlers at Malak-âlagan had increased considerably since 1901, and that its irrigation was now being extended even towards the long-abandoned fields
of Old Domoko. But it was only when marching from Khâdalik south to Domoko, and to the alleged old site near Mazar-toghrak, that I obtained a clear idea of the peculiar conditions of water-supply to which this interesting colony owed its origin and Old Domoko its chance of resumed cultivation.
We first struck the Domoko-yar to the west of Khâdalik near the point where its stream is dammed up in order to feed the Malak-alagan Canal (Map No. 31). Then we followed the broad reed-covered depression upwards until the eastern edge of the present oasis was reached near the hamlet of Ak-kol. By making a short détour eastwards I was shown, at a distance of less than