Klamath Mountains, California. — Hershcy. 155 
be continually grinding on the rock. The river bed is encum- 
bered with huge granite boulders, some weighing hundreds 
of tons, and they are wedged into the narrow bottom of the 
gorge. Between and under the boulders are gravel and sand 
that protect the bed-rock from erosion. The wearing out of 
the boulders is an extremely slow process and their movement 
down stream is practically imperceptible. The river is in- 
capable of moving them by impact, and progression is only 
effected by undermining and occasional rolling over. This oc- 
curs for an individual boulder at very rare intervals. I have 
watched the canon for several years and have noticed no 
change in the groups of boulders ; old settlers have noticed 
very little change in half a century. The placer min- 
ers have been accustomed to prop dams against groups 
of boulders and some of these dams have stood for 
many years. Portions of the river bed that were mined 
in 1852 are still identified by their original peculiarities. 
It is a mistake to lightly assert that because this is a mountain 
region and Salmon river falls 350 feet in a mile, erosion is 
here rapid. It is not. At the upper falls, about two miles far- 
ther up the river, nearly as large a stream has been dashing 
down over hornblende schist, descending probably 300 feet 
in an eighth of a mile, ever since the Salmon river glacier 
in its later stages beheaded the old Cofifee creek', and vet it has 
accomplished an excavation that is verv' much less than that 
of the gorge under discussion. It is indisputable that the 
latter is much older than post-glacial time. Incorporating 
the factor of a larger flow of water during the later stages of 
the glaciation than at the present, wie are still constrained to 
admit that the beginning of the gorge cutting was long anter- 
ior to the close of glaciation. To use the figures 2 or 3 in con- 
nection with this matter is too conservative. 
This gorge is apparently large enough to have had its in- 
ception as far back as the Illinoian (Red Bluff?) epoch. But 
that is too radical a conclusion to entertain without question 
at this early stage of the study and instead I will tentatively 
connect its beginning with the lowan epoch. At the same time 
I will not be positivd that some factor which has entirely es- 
caped my notice may not cause a final reference of the incep- 
tion of the gorge cutting to the Early Wisconsin stage. 
