426 W. M. DAVIS OBSERVATIONS IN SOUTH AFRICA 



by the master floods — a slow process truly, but not slower than the penul- 

 timate erosion of a land surface should be. It may be added that we saw 

 in the Vryheit district of Natal many gulches in hillside alluvium, which 

 were explained as storm-flood channels by Mr Anderson, who there accom- 

 panied us. I am tempted to think that the occasional branching channels 

 in the alluvium on the Veld are best explained in the same way. 



Some of the railway cuttings near Johannesburg exhibited a succession 

 of soil and rock fragments which seemed abnormal and which may come 

 to be interpreted as indicating a change of climate; this is merely put 

 on record here, but not discussed, as it was only seen from the passing 

 trains. 



UND RAINED HOLLOWS OR "PANS" 



% There are certain parts of the Veld in which shallow un drained de- 

 pressions or "pans" occur, often holding lakes or pools of variable area. 

 Not many of these were seen from our route of travel, and we were not 

 able to give special attention to the few which we passed. No satis- 

 factory explanation for the depressions has been offered. The underlying 

 formations are not known to contain soluble minerals whose removal 

 might cause a settling of the surface. The action of the wind may per- 

 haps be appealed to, for the scanty vegetation of the Veld permits the 

 wind to reach a large part of the surface of the ground, particularly in the 

 dry season, but I am not aware of any special features of the pans which 

 demonstrate their origin by wind action. The long continued removal of 

 a significant amount of fine soil during each wet season by animals has 

 been suggested, and when the large numbers of antelopes that formerly 

 roamed here is considered, this suggestion seems to be not without value. 

 But important as the pans are in certain districts, the action of surface 

 wash and sheetfioods seems to be dominant over the greater part of the 

 Veld, for the surface very generally exhibits that systematic arrangement 

 of drainage lines and slopes which long continued washing must produce 

 and which long continued wind action must destroy. Further reference 

 to this matter is made in the closing section of this article. 



RIDGES AND TABLES OF DOLERITE 



We may next consider the effects of variation of rock structure in caus- 

 ing variations of form in the Veld. In the ideally simple scheme of the 

 geographical cycle a land mass of uniform structure is usually postulated, 

 because in such a case the action of the erosive processes is not compli- 

 cated by the unlike resistance of strong and weak rocks. This simple 



