660 R. PUMPELLY EVOLUTION OF OASES AND CIVILIZATIONS 



probably of northern Persia as well, we see correspondence with the con- 

 ditions that permitted the existence of the herds of ruminants and horses 

 that in Pleistocene time ranged from Mongolia to southeastern Europe; 

 and that these animals existed in a wild state at Anau at the time when 

 the north kurgan was settled is proved by Doctor Duerst's study of the 

 bones collected during our excavations. 



Introduction op Irrigation 



If we look at the present climatic conditions ruling throughout 

 Turkestan, we find that irrigation is now almost everywhere essential. 

 The only exceptions are the high valleys and the piedmont borders of 

 the more lofty ranges. 



At a few points, as near Samarkand, grain is planted on the mere 

 chance of there coming once in two or three years enough rain to mature 

 a scanty crop; but along the piedmont plain of the Kopet range there 

 is no local precipitation after March. 



The arid extreme of the climatic fluctuations, which coincided with 

 the disappearance of the different cultures of the kurgans and seem to 

 have caused these interruptions, were very probably less dry than is now 

 the case, but they were doubtless sufficiently severe to render the pre- 

 viously practiced system of agriculture useless for the maintenance of 

 population and of domestic animals. 



Not until the introduction of the artificial distribution of water was 

 it possible thenceforth to maintain a continuity of civilized life. 



The introduction of irrigation reversed the order of the delta-building 

 processes. By bringing all the water under control through the season 

 in which it carries sediments and distributing it evenly over the delta, 

 the aggrading shoreline was kept back at the apex, instead of receding 

 toward the desert, and the delta was continually built up over its whole 

 extent. That this has been the case ever since irrigation began is shown 

 by the fact that since the first layers of irrigation sediments were depos- 

 ited over the old channel shown in shaft B, there has been no recurrence 

 of dissection. 



Had irrigation not come to the rescue, the aggrading shoreline would 

 have receded desertward and the prolonging channels would have car- 

 ried the sediments onward to form barren takyrs, or mud-flats, on the 

 dune-covered plain. 



Since the greater part of the fine sediments brought from the moun- 

 tains is now retained on the delta, the rate of growth of the irrigation 



