388 W. M. DAVIS 



decreases with the decrease of relief and of slope, the activity of the 

 winds is hardly affected as maturity advances. The winds do not 

 depend on the gradient of the land surface for their gravitative accel- 

 eration; they may blow violently and work efficiently on a level sur- 

 face. Whirlwinds are, indeed, most active on true plains. It may 

 be that smooth plains are never swept by winds so violent as the 

 blasts which attack highlands and mountains; but it is probable that 

 the effective action of the winds is greater on a generally plain surface 

 than on one of strong relief, where the salient ridges and peaks con- 

 sist largely of firm rock, and where the loose waste is sheltered in 

 re-entrant valleys. Moreover, it is in very great part on the plains 

 that the winds of ordinary strength drift the sand about, and from the 

 plains that whirlwinds and dust-storms raise the finest waste high 

 enough for exportation. It may therefore be concluded that the 

 work of the winds is but little, if any, impaired by the general decrease 

 of relief that characterizes advancing maturity; and hence that their 

 relative importance increases. Moreover, the scanty rainfall of an 

 arid region will be decreased as its initial highlands, which originally 

 acted as rain-pro vokers, are worn down ; hence, as the relief weakens, 

 the winds will more and more gain the upper hand in the work of 

 transportation. It is conceivable that the rate of exportation of sand 

 and dust by the winds in maturity and all the later stages of an arid 

 cycle is more rapid than the removal of fine soil, partly or largely in 

 solution, from a plant-covered peneplain in the later stages of a 

 normal cycle; thus the slower work of the earlier stages of an arid 

 cycle may be partly made good by the relatively more active work in 

 the later stages. 



As the processes thus far described continue through geological 

 periods, the initial relief will be extinguished even under the slow 

 processes of desert erosion, and there will appear instead large, rock- 

 floored plains sloping toward large waste-floored plains; the plains 

 will be interrupted only where parts of the initial highlands and 

 masses of unusually resistant rocks here and there survive as isolated 

 residual mountains. At the same time, deposits of loess may be 

 expected to accumulate in increasing thickness on the neighboring 

 less arid regions. The altitude at which the desert plain will stand 

 is evidently independent of the general baselevel — or sea-level — and 



