EASTERN ESCARPMENT 435 



descending slope is drawn on true scale a very gentle warping without 

 faulting seems to satisfy all the requirements of the case. 



Still farther north, where our route led from Salisbury on the highland 

 to Beira on the coast, we saw at Umtali, on the border of Ehodesia, a late 

 mature topography apparently eroded below the level of the Salisbury 

 highland by rivers that have a comparatively short course to the sea. On 

 passing Umtali we descended rapidly along the valley of a stream which 

 flows almost directly to the coast and whose young headwaters appeared 

 to be actively encroaching upon the broad valley floor in which Umtali 

 lies. The features at the divide between the mature and the young val- 

 ley would have afforded an admirable subject for study on the ground 

 during our nine-hour stop there ; but, as no one seemed to have taken note 

 of them before, our time was spent chiefly in a walk to one of the neigh- 

 boring hills, where we had a general view of the surrounding country, 

 and the statement here made regarding the young headwaters is based 

 only on what was seen from our train running fast down grade in the late 

 afternoon. 



Important as may be these several examples of marginal dissection by 

 the rivers of Natal and the eastern slope generally, and impressive as 

 may be the example of central dissection by the Zambesi, it remains true 

 that the highland peneplains are still undissected over large areas, and 

 that their streams do not usually follow valleys in the ordinary sense of 

 that word, but flow through broad shallow depressions which represent the 

 form assumed by valleys in their extreme old age. 



ORIGIN OF THE VELD 



Peneplains are normally developed as lowlands with respect to the 

 general baselevel of the ocean surface. Such an origin may be plausibly 

 suggested for the peneplain of the Yeld. It then becomes necessary to 

 conceive of the whole region of the South African highland as having 

 stood, without disturbance, at a lower level than now through the geo- 

 logical ages required for its widespread erosion, and to imagine that it 

 was afterward broadly elevated to its present highland altitude. In con- 

 sequence of such elevation its rivers would proceed to intrench them- 

 selves beneath its highland surface; and inasmuch as their intrenchment 

 is not yet complete, it must under this explanation be supposed that the 

 broad uplift of the region took place at a geologically modern date. 



Another explanation has, however, been lately suggested by Passarge 

 for high-standing plains of erosion. He conceives that an extensive arid 

 region, from which rivers do not flow to the sea, may be slowly reduced 

 to a peneplain of small relief, or even to a plain, at an altitude entirely 



