STUDIES FOR STUDENTS. 845 



In some regions, and over considerable areas, river erosion 

 has so far modified the topography developed at the time of 

 deposition of the drift as to reproduce a topography comparable 

 in kind to that which affected the subjacent rock before the drift 

 was deposited. In this case the characteristic features of drift 

 topography have been destroyed. River erosion topography 

 has been superimposed on the topography of the drift-covered 

 area. Ridges and hills fashioned by streams working upon the 

 drift, stand in that relationship to adjacent valleys which the 

 laws of stream erosion determine. 



Plane tracts may be brought into existence in very different 

 ways. They occur in driftless as well as in drift-covered areas, 

 Flatness, therefore, cannot be regarded as diagnostic, either of 

 drift or of driftless regions. 



It follows from the foregoing that the drift forces must have 

 been such as were able to develop plane surfaces at some points, 

 surfaces marked by more or less symmetrical drift hills which are 

 measurably independent of valleys at others, and short, choppy 

 hills, in still others. They must have been such as were able to 

 modify, to all degrees, the pre-existent rock topography. 



Relation of Drift to the Underlying Rock. 



Where fresh exposures show the contact between any consid- 

 erable bed of drift and the underlying rock, it may be seen that 

 the drift does not generally grade into the rock. Especially 

 where the underlying rock is hard, the plane of contact between 

 it and the drift is generally sharply marked. Where the under- 

 lying rock is not indurated, the plane of contact is sometimes less 

 well defined. Furthermore, where the underlying rock is hard, 

 its surface is commonly firm and fresh. Signs of weathering are 

 absent. Any weathered surface it may once have had before the 

 drift was deposited was completely removed during that process. 

 This relationship is shown in the accompanying diagram (Figure 

 3). It should not be inferred that the relationship between the 

 drift and the underlying rock expressed in the diagram is uni- 

 versal, even where the drift is thick. It frequently does not 



