FORESTS, RESERVOIRS, AND STREAM FLOW 249 
corresponding influence upon the run-off. There is another important 
condition not exemplified in the illustration, and that is that the forest 
areas are scattered everywhere, the ground has an infinite variety 
of slope, the showers never fall uniformly over an entire water-shed, 
and the final result in the total run-off is the summation of: thousands 
of tributary results. 
It is true, therefore, as popularly understood, that, in periods of 
ordinary rainfall with sufficient intervals for the forest bed to dry 
out somewhat, forests do exert a regulative effect upon run-off. They 
modify freshets and torrents, and prolong the run-off after storms 
have passed, and thus realize in greater or less perfection the com- 
monly accepted theory. 
This result utterly fails, however, in those periods of long-con- 
tinued, widespread and heavy precipitation, which alone cause great 
floods in the large rivers. At such times the forest bed becomes com- 
pletely saturated, its storage capacity exhausted, and it has no more 
power to restrain floods than the open country itself. Moreover, the 
fact that the forest bed has retained a portion of earlier rainfall and 
is yielding it up later to the streams, produces a condition that may be 
worse than it would be in a country cleared of forests. Really great 
floods in large rivers are always, as is well known, the result of com- 
binations from the various tributaries. It is when the floods from these 
tributaries arrive simultaneously at a common point that calamitous 
results follow. Any cause which facilitates such combinations is, 
therefore, a source of danger. Now, unquestionably, in a heavily 
wooded water-shed, forests do have a tendency in this direction. When 
a period of heavy storms occurs, spreading over a great area, continu- 
ally increasing in intensity, the forests, by retaining some portign of 
the earlier showers and paying them out afterward, do produce a gen- 
eral high condition of the river which may greatly aggravate a sudden 
flood arising later from some portion of the water-shed. That the for- 
est does promote tributary combinations, there would seem to be no 
question, and that it may therefore aggravate flood conditions neces- 
sarily follows. It is not contended that this increase is ever very great, 
but it is contended that forests never diminish great floods and that 
they probably do increase them somewhat. The forests are virtually 
automatic reservoirs, not subject to intelligent control, and act just 
as the system of reservoirs once proposed by the French Government 
