FORESTS, RESERVOIRS, AND STREAM FLOW 297 
of the Ohio; the total cost is estimated at $125 000000; the income 
from resulting water-power at $20 per horse-power, and a certain 
computed lowering of flood heights on the Ohio and Mississippi Rivers, 
and a corresponding increase in low stages, are given. ‘The full 
details of the scheme are set forth in quite elaborate form. So far 
as the present criticism is concerned, the practicability of finding the 
necessary sites will be accepted, and only the estimate of costs and 
revenues, and the deductions as to benefits will be called into question. 
In their effect upon floods, admitting that all the reservoirs pro- 
posed can be built, the result must fall short of the claims put forth. 
If built at all, they must be built, as will be shown later, primarily 
for power development. Now it will never be possible, until Science can 
forecast the weather more perfectly than it is yet able to do, to reg- 
ulate reservoirs for the maximum benefit of both purposes. This 
consideration is sometimes made light of, but nevertheless it is one 
of real importance. For industrial purposes the reservoirs should be 
full before the rainy season ends; for flood protection they should 
be so far empty that they may be able to hold back any flood-produc- 
ing storm that is likely to come. While, doubtless, in a majority of 
years, a middle course could be pursued that would not involve much 
risk on the flood side of the question nor much loss on the power 
side, yet there would surely come exceptional seasons—the seasons 
of flood-producing rains or the seasons of great drought—when the 
reservoirs would be caught too full on the one hand or too empty on 
the other. Their full calculated capacity would not then be available 
for either purpose, and it is difficult to conclude that this would not 
happen frequently. In particular, if the reservoirs are really operated 
to prevent floods, it must often happen that dry weather will find 
them only partially filled, and that their full capacity will not be 
available either for power or navigation. This would not apply, 
of course, to a reservoir great enough to store all the run-off from 
its water-shed in the greatest known flood, unless considerable storage 
were left over from previous years—as is often done in the Upper 
Mississippi reservoirs. Mr. Leighton’s estimates are based upon the 
mean discharge of the streams, which is, of course, greatly exceeded, 
possibly doubled, in very wet years. In any case it would seem to 
be necessary to hold ample capacity in the reservoirs as late as the 
end of March each year to provide for possible emergencies; but if 
