224 E. C. ANDKEWS. 



Two cases may here be considered : — 



(i) That in which the channel constriction is sym- 

 metrically disposed with respect to the main 

 chanel axis. 



(ii) That in which the constriction is not symmetrically 

 arranged with respect to the main channel axis. 



(i) Channel sides symmetrically arranged with respect to 

 the main axis. 



The composition of the stream need not concern us here. 

 So long as gravitative stress urges it toward the base level 

 so long will it corrade the channel structures with which 

 it comes in contact. In the ideal case we have chosen to 

 consider the stream action is the same on each side of the 

 main valley axis. 



In Fig. 3 (a) let P P' represent a longitudinal section of 

 a stream flowing through a channel constriction where an 

 increase of stream volume has been just set up, and where 

 the stream is dragging earth material over the live rock of 

 its channel base and sides. We will suppose the case of a 

 channel bottom of almost negligible slope. Now from 

 gravitative considerations the cross-section of the stream 

 when traversing the constriction is less than the cross- 

 section of the broader part of the channel farther upstream, 

 therefore the stream has its velocity increased while pass- 

 ing along the channel " narrow." But for a similar reason 

 the narrower portions of this general channel constriction 

 will cause the stream velocity to be relatively quickened 

 in turn at such spots, and rock material will here be dragged 

 more quickly over the channel base than at other points in 

 the constriction and the stream will cut more deeply at 

 these points. Increase again of the corrasive power of 

 streams is not related in a simple ratio to increase of 

 velocity, and therefore a slight increase of stream velocity 

 may be associated with a pronounced increase of corrasive 



