398 W. M. DAVIS OBSERVATIONS IN SOUTH AFRICA 



cause the abandonment of the previously established graded condition 

 and the development of a new grade. It is true that the variation of 

 river slope thus determined may be very small, but it is also true that a 

 small variation of slope will cause a significant change in the way of 

 aggrading or of degrading in the upper course of a river several score or 

 several hundred of miles in length. It is, moreover, generally believed 

 that climatic changes of importance have taken place in the Pleistocene 

 time ; hence the effects of such changes should be looked for in the valleys 

 of graded rivers, just as they are looked for where glaciers have been 

 formed and where lakes have expanded and contracted. While these 

 principles are familiarly accepted, their application to the problem of 

 terracing is less general than it should be. One of the most important 

 examples of such application is that offered in W. D. Johnson's admirable 

 paper on the High plains of Colorado and Kansas, in which variation of 

 climate rather than variation of continental attitude is appealed to as the 

 effective cause of the aggradation of the plains by the rivers flowing 

 from the mountains and for the later dissection of the aggraded plains 

 by the same rivers. Another example of the same kind is to be found in 

 the first volume of Pumpelly's report on his Carnegie expedition to 

 Turkestan, in which Ellworth Huntington describes certain terraces asso- 

 ciated with glacial moraines in the valleys of the Tian Shan mountains 

 * of central Asia. I can not attempt to determine how far these authors 

 are right in thus appealing to climatic variations in explanation of the 

 phenomena that they describe; but it is safe to say that, until variation 

 of climate and variation of continental attitude are both duly considered 

 and some means of distinguishing between their effects is discovered, it 

 will be premature to assume that either cause is fully responsible for the 

 observed effects. 



The Buff els Eiver terrace is about 100 miles from the mouth of its 

 river system. When the terrace was formed Buffels river was probably 

 well graded all the way to its mouth, a tolerably smooth and gently 

 sloping channel being then established even in the mountain gorges that 

 it traversed. At present, however, the river is not very smoothly graded 

 in certain parts of the Leeuw Kloof notch, and hence is probably even less 

 smoothly graded in the harder rocks of the notch in the Klein Zwartberg. 

 When an even grade is fully established later in the present cycle, the 

 valley about Laingsburg will have been significantly deepened and greatly 

 widened. The total depth of erosion in the terrace or the former valley 

 floor to the equally well developed future valley floor, great as it may 

 seem when only the Laingsburg district is considered, would be brought 

 about by a change of river slope amounting to only 10 or 15 minutes of 



