Mr. Swain. 
370 DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 
layer of sand on an inclined surface practically impervious to water. 
The ground under the forest bed is not impervious to water, but is in 
fact kept by that bed in a highly absorbent condition. 
The fact must be emphasized that those who believe in the bene- 
ficial effect of forests upon flow do not urge the preservation of forests 
on lands needed for agriculture and topographically suited for it. The 
beneficial effects of forests on flat lands, in modifying the violence of 
freshets and increasing the low-water flow, is much less clear than in 
the case of forests in steep mountain regions. It is the preservation of 
these last—forests upon steep slopes not suitable for agriculture, where 
the erosion on cleared land is very great—that is believed to be espe- 
cially important for the streams. 
Fig. 6. 
In fact, this part of the discussion seems rather academic. Probably 
few if any engineers will deny that lakes, even if unregulated, are or 
may be of great benefit in regulating the discharge of streams which 
flow out of them, and yet even here we may conceive of a case where the 
reverse would be true. Suppose, for instance, that the two streams, A and 
B, rise in the same mountain region, a, and unite at d, Fig.6. Ifa heavy 
rainfall oceurs at the head-waters, the part which flows via A will 
reach d much sooner than that which flows via B. Now, if a reservoir 
is placed on A, it is conceivable that it may check the flow on that 
stream sufficiently to make the rise of the water at d somewhat greater 
than before. But of course the flood has been reduced all the way 
from ato d. This simply illustrates the fact that if artificial reservoirs 
are to be located, their location should be carefully studied. 
When it is claimed, as the writer does, that the preservation of the 
mountain forests will be of great benefit in regulating the flow, this does 
not mean that if a region is deforested the maximum flood will be 
greater than ever before, for the conditions are, as has been abundantly 
shown, extremely variable. On the Connecticut and Merrimac Rivers, 
for instance, the highest water ever known occurred before the present 
rapid cutting of timber in the upper parts of the basins. Nevertheless, 
it may be equally true that even those freshets would have been still 
worse if the forests had not been there; and it is quite certain that, if 
the height of extreme floods has not been increased by deforestation, 
the frequency and duration of great floods have. The statement* that 
*See page 262. 
