COERASION OF THE EIVEE BED. 
243 
and accumulates while floods are subsiding, but it is lifted and carried for- 
ward again when the waters increase. At low water many sand bars are 
formed in the still reaches, but they are changed and modified by every 
flood, and many of them are wholly swept away and reconstructed with 
each oscillation. The scouring action of the sand is highly efficient. It is 
seen on all the rock surfaces laid bare by the subsidence of the waters, and 
in all the large fragments rolled in from the side gorges. The rocks are 
full of pot-holes and covered often with bosses highly polished and ground 
by the sand-blast. The large bowlders which fall from the cliffs above, or 
which are brought in through the side gorges, are rapidly dissipated. The 
largest ones are deposited at the mouth of the gorge where the rapid is 
formed, and where they are exposed to the fullest energy of the current. 
The largest masses, if fully exposed to this action, can scarcely survive 
more than three or four years. Although the dissipation of the fragments in 
the Colorado has never, so far as I am aware, been watched and timed, we 
are not without the means of confidently inferring a very short duration of 
that process. In the hydraulic mines of California the water discharged 
against the gravel banks has been known to cut in a single year chasms 
from 12 to 20 feet deep in the hard basaltic pavements over which it flows. 
The actual time of operation is equivalent to not more than 100 days of unin- 
terrupted activity. No doubt the instance here cited is an extreme case of 
rapid corrasion, and is not strictly parallel to that of the river. But a large 
river always charged with sand, and careering swiftly over a dam of rocky 
fragments, is a very strong case, if not an extreme one, and we may per- 
haps be led to regard four years as a long life for a rock-mass, six feet in 
diameter, exposed to the full and uninterrupted action of the stream. 
Not all of the large fragments, however, are exposed to this continuous 
action, but many of them are out of water during the greater part of the 
year. At the opening of a side gorge the river is usually constricted, and 
in many cases the width is not more than one-third of the mean width. 
The coarse debris forms a deposit somewhat analogous to an alluvial cone. 
Undoubtedly the corrasion is most efficient during the stages of high 
water. In many and perhaps in most livers this is not true, or at least is 
not necessarily true, for the sediment brought down by floods in other 
