EROSION B Y UNCONCENTRA TED WASH 751 



the hilltops. To avoid this improbable supposition it might be 

 assumed that the gravels were laid down on a plain whose elevation 

 is represented by that of the present exceptional hills and that a post- 

 Lafayette peneplain was developed 50 or more feet lower. 



Hypothesis of rill-wash applied to the St. Louis region. — As a 

 modiiication of, or substitute for this last hypothesis, the following is 

 suggested: After the deposition of the Lafayette gravels on a nearly 

 flat surface, uplift followed and the area was maturely dissected by a 

 drainage system which was to some extent ready made, having held 

 over from former conditions and which, therefore, to a certain degree, 

 began its work simultaneously on the entire area. The gravel was in 

 the main removed, except from a few patches between the head 

 waters of streams flowing northwest to the Missouri and others flow- 

 ing southeast to the Mississippi. It remained in these places partly 

 because they were flatter and therefore less subject to erosion, and 

 partly because, lying between headwaters, they were the last to be 

 reached by erosion. With the exception of these spots the topography 

 of the area was then one of comparatively even-topped ridges and 

 valleys, but without fiat uplands. Both while this dissection was in 

 progress and subsequently, unconcentrated wash (as described above) 

 lowered these ridges fifty or more feet. This was not effective on the 

 gravel-covered hills, partly because of their relative flatness, but 

 largely because percolation obviated wash. 



The difficulty in such a conception lies in the uniformity of the 

 down-cutting of all the ridges and the consequent preservation of 

 the typical form of a dissected peneplain. This arises from our habit 

 of thinking of rills as small rivers, each incising its own little valley 

 and absorbing its neighbors as soon as a shght advantage has been 

 gained. If the above theoretical reasoning pertaining to rill-wash be 

 correct, this difficulty disappears. The struggle for existence (so 

 characteristic of gullies) disappears from the community of rills as 

 soon as each is seen to be overloaded. Stability then takes the place 

 of instability and there is no longer any difficulty in maintaining an 

 undissected slope while degradation proceeds. Under these conditions 

 a large number of subequal ridges constituting a dissected plain will 

 be degraded at a subequal rate. 



It should be made clear that the correctness of the theoretical 



