Sec. iv] REMAINS OF A RURAL SETTLEMENT AND CITY OF ETZINA ' 459
their first move from Dzungaria to the Etsin-gol, said to have taken place four centuries ago, ` their ancestors found the ruins just as they are now ', it appears very doubtful whether any historical value whatever can be attached to this legendary tradition. But there is one point in it that deserves to be noted, because it indicates a correct perception of what must have been a determinant factor, if not in the actual abandonment of the site, at any rate in preventing its reoccupation and the reclamation of the area once cultivated.
According to the legend ` the imperial army, being unable to take Khara-khoto by assault, determined to deprive the beleaguered city of water ' by diverting the river Etsin-gol, and effected this object by blocking the original channel with bags of sand. Of these bags remnants were said to have been found even in recent times.13 We did not succeed in clearly locating the head of the old channel when we searched for it on our return journey near Borgasu, where according to our Mongols it had been. But in the light of our observations along the Etsin-gol and of the corresponding experience gained at almost all deserted sites along the southern edge of the Taklamak5.n, there is much to support the belief that difficulties in connexion with water had played a great part in the final abandonment of the site, though not quite in the way related by the legend. No artificial diversion of the river-bed could have lowered the level of subsoil water at the town site so quickly and to so great an extent as to render its wells useless during a siege. But it is quite certain that the shifting of the river from the old bed passing Khara-khoto to the bed now followed by the Ikhe-gol would inevitably cut off irrigation from the once cultivated area, which lies on an average fully six miles to the east of Khara-khoto, and fourteen miles from the nearest point of the present river channel ; for it is clear that the canals upon which its cultivation depended must have taken off from the Etsin-gol branch, which is still clearly visible at Khara-khoto and which was traced by Lai Singh for more than five miles farther to the south-east.
Such changes of the river-bed are bound periodically to threaten all canals maintained in a deltaic area, and if for some reason the settlement at the time was unable to cope with a change serious enough to affect the volume of water received at the head of its canals, gradual abandonment of the previously cultivated lands would necessarily ensue.13a An attempt to determine now whether such a change had actually occurred in the case of the settlement east of Khara-khoto would have involved a very detailed survey of whatever traces of the old canals still survive. Neither the advanced season nor the time available would have permitted us to undertake this task. But even if it had been otherwise, I question whether the result would have yielded evidence sufficiently definite to exclude consideration of another possible reason for the abandonment of the area, a reason most potent in a climate so arid as that of the Etsin-gol delta : I mean ` desiccation ', whether general or local.
All my observations during our marches along the Etsin-gol and its delta combined to impress on me the probability that ` desiccation ', i. e. in this case a diminution of the water-supply brought down by the river, had played a very important Part in producing the conditions now displayed by the old settlement. It may not have been the sole or immediate cause of its abandonment. But it certainly is the reason why the site has not been reoccupied since, and why no reclamation of the once cultivated ground could, I believe, be successfully attempted at present. Even at the oasis of Mao-mei, more than 150 miles higher up the river, and with conditions of ground far more favourable for the maintenance of canals, serious difficulty had been experienced for some years past in securing a sufficient discharge in the canals early enough in the season, and much land that had once been cultivated appeared to have been abandoned recently.l3b
13 See Geogr. J., October, r9o9, p. 388. 13a Cf. above, p. 14o sq. ; Serindia, i. pp. 203 sqq.
131) See above, p. 409.
Alleged diversion of river.
Change of river affecting irrigation.
Diminution of water-supply in river.