ACTION OF RITEES ON ROCKS. 93 



4. By entering the fissures of rocks and freezing, so as to separate them by 

 expansion. This is one of the most powerful modes in which the work of excava- 

 tions is carried forward. 



5. By ice floods. In these cases the stream becomes choked with ice, with only 

 water enough to make it plastic, and enable gravity to urge it forward. The 

 moving mass does, indeed, very strikingly resemble a glacier, and it moves forward 

 with a similar immense power; ploughing up the loose surface, tearing off the pro- 

 jecting rocks, and sometimes forming new channels for the river. 



6. Where there are cataracts in rivers, all these modes of erosion usually act 

 with a maximum intensity : and at this day probably the principal amount of 

 erosion by rivers takes place where there are cataracts. These cataracts are con- 

 stantly receding, although when measured by the life of man, the rate of retroces- 

 sion is scarcely perceptible ; but measured by geological periods, it becomes very 

 manifest, and we find evidence, that in this manner long and deep gorges have 

 been produced, and lofty barriers removed. The consequence of this latter process 

 is, that the river below and above the barrier, thus partially or wholly removed, 

 will excavate a deeper bed in the loose materials there accumulated. 



5. Without attempting to determine the precise amount of erosion by rivers, I 

 wish to state distinctly that I do not impute to this agency the whole, or even the 

 larger part, of the formation of the valleys through which rivers now run. Much 

 less do I maintain that present rivers have produced these valleys : for there is 

 proof in some cases, that other streams once flowed through valleys now occupied, 

 perhaps, by rivers totally unable to have eroded them. For the formation of most 

 of our present valleys we may assign the following agencies : — 



1. The original upheaval and dislocation of the strata. 



2. Long continued oceanic action. 



3. The drift agency. 



4. Rivers on former continents. 



5. Existing rivers. 



I impute to rivers only such a part in the work of erosion as can be proved by 

 an application of the preceding principles. 



Caution in the application of the preceding Rules. — 1. The older the rock through 

 which rivers have cut their way (caiteris paribus), the greater should we expect 

 the amount of erosion. 



2. But, secondly, the position of the strata, if the rock be stratified, and the 

 amount of water acting upon them, or the number and direction of the fissures, if 

 the rock be unstratified, will greatly modify the amount of erosion. If the strata 

 cross the stream and dip in the same direction as the slope of the river, the action 

 of the water will be much more powerful than if the dip is in the opposite direc- 

 tion. Or if the inclination of the strata corresponds with that of the stream, the 

 erosion will obviously be slower. Again, some unstratified rocks present but few 

 fissures, while others are full of them, and this fact will make a great difference in 

 the erosion. 



3. Bocks, essentially alike in chemical composition, may yet vary very much in 

 hardness, and in the ease with which they might be disintegrated. How great 



