DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 421 
be at Cincinnati and other points along the river. Had he done 50, mr. Leighton. 
then, indeed, would it have been necessary to account for all the 
phenomena of flood transmission. Fortunately, observations at lower 
points rendered this unnecessary. The regimen has been observed; 
what happened at Wheeling, Cincinnati, ete., is known and also what 
contributed to the conditions there. Now, what is the effect of the 
reservoir system as proposed? It is not that of prolongation of flow; 
no attempt is made to flatten the wave by allowing the same amount 
of water to discharge in a longer space of time. The element of time 
remains unaffected, save that which results from change of head. 
The real effect is to cut out from the drainage area a certain portion 
which otherwise would contribute to the flood. A flood wave at Pitts- 
burg would contain equivalently less water; and, knowing what actually 
happened there without flood storage, one can compute approximately 
what would take place as a result of the reduction. This reduction 
would not be the equivalent of the area cut off; it would be greater, 
because this area is steeper than the average for the whole, and, there- 
fore, its flood wave is the more quickly transmitted to the lower valley, 
as demonstrated by Colonel Chittenden. The special point that he 
makes, by the way, is not involved in the discussion, for no one is con- 
tending that upland streams are subject to over-bank floods. In 
establishing storage reservoirs, efforts are made to protect the low- 
lands, not the uplands, and it is self-evident that they are most in 
need of protection from these upland streams, which are of so large 
a unit discharge capacity and so swift in flow that they do not inundate 
their own lands, but quickly deliver their floods into the lower rivers 
of relatively small capacity. 
Proceeding now with the matter of flood propagation, let it be 
assumed that the problem has these elements in it. Colonel Chittenden 
starts with a flood at Pittsburg increasing from 100000 to 200000 
cu. ft. per sec. It is asserted that this would not produce a similar 
increase at Cincinnati. Very well, but is this not in our favor, and 
are we not on the safe side, as far as flood damage is concerned, if 
we assume that the increase will be as great in one case as in the 
other? The rule should work both ways. If an increase of 50% in 
the discharge at Pittsburg does not produce a similar increase at 
Cincinnati, should not a decrease of 50% at the one place effect a 
greater decrease at the other? ; 
Carry the matter up past Pittsburg, even to the reservoirs. If, by 
manipulating the gates, 50% of the flood is conserved, a reduction 
equivalent to that assured by the author would take place when the 
crest had reached Pittsburg, and, in proceeding down the Ohio Valley, 
it will be reduced still further by the time it reaches Cincinnati. Why, 
then, would not the reservoirs have 125% of the predicted results, 
instead of 75%, as Colonel Chittenden suggests? 
