DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 363 
stored water downward, as well as outward, into the subjacent soil Mr. Smith. 
and rock as well as into the surface rivulets and streams. Similar 
diversion of a part of the precipitation into the ground-water supply 
takes place in cleared land, but not, it is believed, to the same degree, 
except where, under exceptional circumstances, the soil and subsoil is 
as open as it is under the forest mulch. The ability of the forest bed 
to delay run-off is admitted by Colonel Chittenden, who considers it 
likely to aggravate floods, in the logic of which deduction, however, the 
writer cannot follow him—at least this aggravation of flood conditions 
certainly cannot be regarded as an “unquestionable tendency of forests.” 
Rather this retarding action of the forest mulch will tend to increase 
the proportion of the rainfall contributed to the underground reservoir, 
thus reducing the amount immediately converted into run-off. De- 
creasing the rate of surface flow increases the time and opportunity 
for absorption by the pervious soil and rock beneath the forest mulch. 
This absorptive action must to a large extent make complete saturation 
of the forest bed of rare occurrence, and the only general areas where 
“the percentage of retention in the forest bed is 0” would be the tree- 
covered swamps, swamps ordinarily indicating the area where the 
ground-water table reaches the surface. This increased contribution 
to the ground-water reservoir, from which the progress of water to 
its point of issue, where it joins the surface drainage, is indirect and 
slow, cannot fail to counteract in some degree, sometimes in a very 
marked degree, the fluctuations in that part of stream flow originating 
in direct run-off. The persistence of springs during the period of 
minimum run-off is a matter of common observation, and is explained 
by the long and tortuous character of the underground route fol- 
lowed by the water, as contrasted with the surface route of direct 
run-off, 
The question may be raised whether this function of the forests 
in increasing the contribution to the underground flow is quantitatively 
important. The drying up of springs, with the extensive cutting off 
of the forests, is a well-known and widespread phenomenon, and one 
that bears directly on this point, though Colonel Chittenden doubtless 
places slight value on its importance. The most definite data that 
have come under the writer’s observation, as to the rate at which 
pervious rock will absorb surface water, indicate that in a short 
distance the amount so absorbed from a small stream is sufficient to 
be readily detected and even measured, being several second-feet in the 
case under observation.* Moreover, the conditions in a stream bed are 
not altogether favorable for such absorption—by reason of the silting 
up of the pores of the rock—probably less favorable than under the 
forest mulch. Indeed it appears reasonable that as a general rule the 
* U.S. Geological Survey, Water-Supply Paper No. 55, p. 56. 
