STEPPES, DESERTS, AND ALKALI LANDS. 605 



of Egypt been ascribed to the mud carried down by the flood 

 waters of the Blue Nile from the uplands of Abyssinia during 

 the season when torrential rains prevail there. 



Without denying a certain efficiency of this cause, a closer ex- 

 amination easily proves it to be inadequate to account for the 

 millennially undiminished fertility of the Nile Delta. The average 

 annual mud deposit of the Nile floods amounts to less than the 

 thickness of a common pasteboard.* Were this the best of stable 

 manure well worked in, it could not produce the effect claimed. 

 But examination proves it to be simply a rich soil, such as thou- 

 sands of farmers could haul and spread upon their lands if it could 

 produce the effect ascribed to this Nile sediment. Besides, per- 

 petual fertility belongs equally to the lands of the neighboring 

 Fay 00m, which, being irrigated only with the clear water of 

 Lake Mceris, do not receive the benefit of any sediment. But the 

 analysis of that water, or of the clear Nile water itself, does not 

 show it to contain any unusual amount of fertilizing matter in 

 solution. ' 



There are other examples, too, of lands perpetually productive 

 for thousands of years without fertilization. One is the " regur " 

 lands region of the Deccan, forming part of the plateau of south 

 central India ; another is the " loess " region of China, drained by 

 the head waters of the Yellow River, and for ages the granary of 

 that empire. In both these cases no alluvial fertilizing deposits 

 come into play, and there is little or no irrigation. But in both 

 cases we have a semiarid climate, the rainfall being close to the 

 limit of actual deficiency ; in the case of Egypt the deficiency is 

 extreme during nine months out of the twelve. 



What, then, is the effect of a deficient rainfall upon the nature 

 of soils formed under its influence ? And can these effects serve 

 to explain, in any measurable degree, the choice of the ancient 

 civilizations ? 



In the course of his investigations of the soils of the United 

 States, the writer has had occasion to make extensive compari- 

 sons of the soils of the Atlantic humid region with those of the 

 arid and semiarid West. A summary of the results of these com- 

 parisons was given in Bulletin No. 3 of the United States Weather 

 Bureau in 1892. They led to important conclusions of a general 

 nature, some of which could have been readily foreseen on gen- 

 eral principles. Continued investigations made since have given 

 additional confirmation, and have developed new facts having 

 important bearings upon the possible utilization and productive 



* Taking this at one twenty-fifth of an inch, it would amount to about five tons per acre, 

 or about two good two-horse loads. Three times that amount of stable manure is about the 

 usual dressing for an acre. 



