Drainage Systems of New Mexico. — Tarr. 267 
and it is at present engaged in cutting through this barrier. 
From the structural features of the outcrop I infer that to the 
west about a mile the Rio Grande would not have encountered 
this rock, at least not for several hundred feet. Other similiar 
accidents are to be found at various places in the course of the 
river. I believe that a careful study of the river west of Santa 
Fe will show that it has been turned many miles to the east of 
its early course by a great thickness of lava ; but I was unable 
to spend the time necessary for the verification of this supposi- 
tion. 
It is evident that the development of the caiion of the Rio 
Grande was rapid, first because it is only slightly effected by 
subaerial denudation ; and, secondly from the evidences fur- 
nished by the side streams. The greater part of the work in 
the formation of the Rio Grande caiion was the removal of the 
very uppermost layers which were hard basalt. After this was 
cut through the remaining work was, with some local exception, 
the removal of the soft deposits which could be cut away as 
fast as the stream could dispose of the load. Hence the first 
hundred feet of cutting was slowly done and the next nine 
hundred feet of rock was quickly removed. While the river 
was busy cutting through the lava the side streams had very 
little opportunity to cut channels since they were retarded 
in their work of erosion by the main stream to which they 
were tributary. When once the Rio Grande succeeded in erod- 
ing through the lava crust and began to eat rapidly down 
through the sands, the tributaries began to cut into the lava, 
first near their mouths then progressively up stream farther and 
farther. Erosion at the mouths of these tributaries very nearly 
kept pace with the deepening of the Rio Grande itself; but 
from this point upstream for a few miles the slope is very rapid 
until the lava-capping is reached where the stream is laboring 
hard to get through the rock into the soft sands below and gain 
upon the stream near the mouth. 
This condition is excellently shown on the Rio Grande at 
the mouth of Taos creek. For a few miles from its mouth the 
creek is deep and canon-like and at its mouth is as deep as the 
Rio Grande itself, whi(;h, at this point, is at least a thousand 
feet below the plateau surface. The canon grows progressive- 
ly shallower upstream, and a little more than five miles from 
its mouth the stream is on the lava ; and one or two miles up- 
