~G. K. Gilbert—The Colorado Plateau Province. 99 
In the cafions corrasion is favored by a quantity of water which 
belongs to the mountain sources of the streams and not to the 
plateaus which they divide. It is favored by a great declivity 
of bed, for which it is indebted to the magnitude and recenc 
of the uplift. It is favored by a moderate supply of debris, 
always sufficient for the work of excavation, but not so great 
as to consume the entire energy of the current. 
The contrast between the degradation of the upland and the 
cutting of the water ways is strongest where the rocks are best 
fitted to resist disintegration. The rivers sink their channels 
into the land in a harmonious and interdependent system, and 
cannot excavate soft beds more deeply than hard. But the 
only downward limit to the degradation of the tables is the 
level of the draining river system ; and the varying retardation 
which it suffers from the resistance of different rocks, is ex- 
pressed in the varying height of the cafion walls. 
_A second problem which has arisen in the study of the ero. 
sion of the Plateaus may be calle 
The Problem of Waterfalls. 
_ Where rivers descend a slope that is terraced by the alterna- 
tion of hard and soft strata, they are apt to leap from the edges 
of the hard beds in waterfalls. But the Colorado, notwith- 
standing the structure of its bed presents the most favorable 
conditions, makes no leap. At the head of Marble Cafion, for 
stance, the river crosses a great bed of limestone, lying nearly 
level and underlaid by a great bed of friable sandstone. The 
tmestone resists all erosive agents as strongly as does the N lag- 
ara limestone, and the sandstone yields to them as easily as 
does the Niagara shale. But, instead of plunging from one to 
the other in a great cataract, the Colorado cuts the two with 
nearly equal grade of channel. Its average descent in the hard 
rock is ten feet to the mile, and in the soft, less than five feet. 
_Itis evident that for the production of waterfalls some con- 
dition is involved beside that of the constitution of the rock- 
system which the stream traverses,—some condition that per- 
tains to the constitution of the stream itself. Such a condition 
18 to be found in the relation of corrasion to transportation. 
Let us suppose that a stream, endowed with a constant sup- 
ply of water, is at some point — supplied with as 
8reat a load as it is capable of carrying. For so great a dis- 
tance as its velocity remains the same, it will neither corrade 
hor deposit, but will leave the grade of its bed unchanged. But 
if in its rogress it reaches a place where a less declivity of bed 
Ge a diminished velocity, its capacity for transportation will 
me less than the load, and part of the load will be depos- 
