428 ~W. M. DAVIS OBSERVATIONS IN SOUTH AFRICA 



dolerite-capped mesas, and from this it must be understood that the talus 

 is not a thick accumulation, like the heavy fans that accumulate at the 

 base of oversteepened cliffs in glaciated districts; it is only a sheet of 

 waste whose fragments travel slowly down the slope, while the slope 

 slowly retreats before the slow expansion of the plain at its foot. The 

 talus slopes are seldom ravined, and from this it may be inferred that the 

 talus, like the plain, is acted upon rather uniformly over large areas in- 

 stead of on lines of concentrated water flow. We saw indications at 

 various points that the contrast of conditions between ridge and plain 

 acted as a control in the distribution of vegetation, concerning which the 

 botanists of the party could probably give details. It would appear also 

 that the stony slopes of the ridges and mesas might often serve to harbor 

 various forms of animal life which could find no satisfactory refuge on 

 the featureless plains. But it sometimes happens that a broad sheet of 

 dolerite lies at a level only slightly above that of the surrounding plain, 

 and in such case the flat surface of the sheet is covered with boulders of 

 decomposition to such an extent as to exceed in this respect the talus 

 slopes of the ridges and mesas. 



RELATION OF RIVERS AND RIDGES 



The relations of the drainage system to the dike ridges present some 

 interesting features. The plains may be conceived as more or less com- 

 pletely divided into irregular compartments, large and small, by the 

 ridges. Man) of the compartments are. drained by streams that escape 

 where the ridges are wanting; the plains of such imperfectly separated 

 compartments form a continuous surface of uninterrupted grade; but it 

 not infrequently happens that the stream by which a compartment of the 

 plain is drained enters the compartment or escapes from it through a 

 notch in one of the inclosing ridges. Then the plains up and down stream 

 from the notch are not at precisely accordant levels ; for the stream in the 

 notch is usually of a steeper slope, either on bed-rock or on large boulders, 

 than it is on the open plains. The plain upstream from a notch is of 

 course graded with respect to the notch; the plain below the notch is 

 graded with reference to some other local baselevel farther downstream. 

 This relation has probably obtained for ages past, the graded plains hav- 

 ing been worn down as fast as their local baselevels were lowered. Here 

 as elsewhere the highly developed organization of the drainage system is 

 exhibited. There is, for example, no excessive accumulation of alluvium 

 on a plain upstream from the notches which are so characteristic of the 

 region; hence there does not seem to have been either an undue deepen- 



