950 FORESTS, RESERVOIRS, AND STREAM FLOW 
for the control of floods in the River Rhone would have acted, if built. 
These reservoirs were to have open outlets, not capable of being 
closed, which were intended to restrain only a portion of the flow. A 
careful study of their operation in certain recorded floods showed that 
they would actually have produced combinations more dangerous than 
would have occurred without them. 
Consider now periods of extreme drought, and grant that, as a gen- 
eral rule, springs and little streams dry up more completely than when 
forests covered the country, although this difference is very greatly 
exaggerated in the popular mind.* At first thought one would con- 
clude that, since the springs and streams make up the rivers, these 
also ought now to show a smaller low-water flow than formerly. This, 
however, is not the case. The difference between the former low-water 
flow of a spring or rivulet and what it is now is relatively an insig- 
nificant quantity. Most of such water sources yield but a small frac- 
tion of a cubic foot per second. Whether these small quantities are 
a trifle more or less cuts very little figure in the aggregate; and so it 
counts but little in the flow of a great river whether some of its 
extreme sources lose a portion of a volume that is already inappre- 
ciable. When the summer showers come, however, there is a marked 
difference. At such times the forests not only hold the water back— 
they often swallow it completely. Small showers that make a per- 
ceptible run-off in the open are often practically all absorbed in the 
leaves of the trees. Heavier showers, that make freshets in the open, 
are largely absorbed in the leaves and forest bed and pass off in evap- 
oration; so that, contrary to the general view, the evaporation from 
the forest is greater at such times than in the open country, and the 
run-off from summer precipitation is less. A single shower may pro- 
duce a sufficiently greater run-off in a deforested area to more than 
offset the diminished low-water flow for several weeks.t Now on most 
of the smaller streams quantity of flow is a more important matter than 
*The term ‘‘as a general rule” is used, for it is by no means absolute. In 
particular the drainage of low swamp lands leads off into the streams, in ary 
weather, waters that formerly remained or passed off in evaporation, and in such 
cases even the low-water flow is greater than it used to be. In 1895 the writer 
saw an example of this on the Scioto River near the outlet of the great Scioto 
swamp which had recently been drained. A small mill was able to operate during 
the low-water season more regularly than formerly. Tile drainage, now so widely 
used, has the same tendency. 
tSo far as the writer is aware, Colonel T. P. Roberts, of Pittsburg, Pa., was 
the first to call attention to this characteristic of stream flow. , 
