FORESTS, RESERVOIRS, AND STREAM FLOW pyar 
theless be of the very highest value, if it could be made to hold, 
for it would give permanence to the banks, security to riparian prop- 
erty, and would largely prevent bar-building by training the river 
in a regular channel and relieving it of everything except its through 
load of sediment. 
The bank-caving problem of these valleys is unaffected in any 
appreciable degree by the influence of forests or cultivation on the 
water-sheds, and cannot be solved or materially assisted by any 
practicable changes in these conditions. The problem is strictly a 
local one, and the remedy must be a local one. Even if it were 
possible to bring the waters down from the uplands perfectly clear, 
it is not at all certain that the effect upon the bottom lands would not 
be injurious rather than beneficial; for then the caving soil, instead 
of being quickly deposited again, would in part be carried out to 
sea, and the bottom lands, unless protected, would be gradually eaten 
away. 
In addition to the four main propositions discussed above, a few 
subordinate features of the question will now be considered. 
A feature of the Forestry Service which is generally overlooked 
is the possible effect of culture upon the bed of humus so much relied 
upon in these discussions to prove the restraining action of forests 
upon run-off. Mr. Pinchot, in his statement to the Judiciary Com- 
mittee, said: 
“The effect of a forest on a steep slope is to cover that slope with 
leaves, rotten and half rotten sticks, and other mechanical obstructions 
which prevent the water from running below as rapidly as it would 
otherwise.” 
It is understood that the forest policy is to keep this litter cleared 
up as a measure of fire protection, and one frequently sees in articles 
on forestry, photographs of the typical forest culture in which the 
ground is thoroughly cleaned up. The result must be to diminish pro- 
portionately the retentive action of the forest bed and to increase its 
liability to erosion. In the light of the foregoing discussion, fire pro- 
tection is of much greater importance than the retentive effect of the 
forest bed on the run-off. The remarkable degree to which the forest 
bed will dry out in prolonged drought, making it one vast tinder-box, 
supports this conclusion, and is another proof of the extreme desic- 
eating effect of forest growth upon the soil. 
