66o MARIUS R. CAMPBELL 



would probably not take place, for the stream A B would prob- 

 ably corrade its channel about as rapidly as the land rose, and 

 so would prevent the lower stream from affecting its capture. 



If, however, the arrangement shown in Fig. lo prevails at 

 the close of a period of quiescence in which the surface is worn 

 down close to baselevel, the streams would be in a position to 



Fig. 10. 



take advantage of any opportunity to extend their drainage 

 basins. If at such a time a tilt occurred in the direction indi- 

 cated by the arrow the point a on the stream A C would remain 

 at about the same elevation as A. The stream A B would be 

 accelerated, but active corrasion would, for a long time, be mainly 

 limited to its lower course. The branch a b would be greatly 

 accelerated and might be able to cut headwards to the point c 

 and capture c B before A B cuts back to the point c. Whether 

 this diversion would be accomplished or not, depends upon the 

 rapidity of the uplift, the volume of water in A B, and the char- 

 acter and attitude of the rocks between the two streams. 



Thus the migration of divides on a tilted surface tends to 

 produce a system of drainage, the small branches of which are 

 perpendicular to the axial line, and flow in the direction of the 

 dip of the surface ; and also a series of branches of the second 

 order, flowing parallel with the axis, and in general coinciding 

 with the downward pitch of the same. When the axis of uplift 

 coincides with a divide already established, it is obvious that 

 the divide will be preserved, but a rearrangement will take place 

 in the head branches of the contending streams. The increased 

 gradient will induce the small branches to extend their upper 

 courses directly toward the axis, and the larger streams will 



