DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 419 
point farther along in July, and affording thereafter a subsidence at Mr. Leighton. 
a rate somewhat approximating the subsidence of the computed nat- 
ural-flow curve. It is apparent, therefore, that, although the controlled 
flow at Pokegama is in one point actually greater than the maximum 
natural flow would have been, there is not the slightest reason for 
believing that this controlled flow came anywhere near adding to the 
severity of the flood which would have occurred without reservoir 
control. 
The author has based his conclusions on the relative natural and 
controlled flows at Pokegama, whereas the actual test of the matter: lies 
in the record of the Aitkin gauge. His apparent tendency to regard 
the actual discharge from the reservoir as the important feature, rather 
than the actual effect on the river below, is further indicated by 
Fig. 3. 
Fig. 3 certainly shows, as the author contends, that the controlled 
flow at Pokegama was not as uniform as the natural flow would have 
been; but is this not the whole purpose of the reservoir system? In 
the practical operation of reservoirs one does not seek to secure uni- 
form discharge from those reservoirs, but rather to adapt the discharge 
so that, in connection with the uncontrolled discharge from lower 
tributaries, it may finally produce a condition of uniformity in the 
river many miles below, and not, indeed, in the river just below the 
reservoir outlet. It would have been eminently fair, in presenting 
Tig. 8, to have added thereto two more curves, one showing the dis- 
charge at Aitkin, Brainerd, Little Falls, or St. Cloud, as controlled 
by the upper reservoirs, and the other, a computation of what that 
flow would have been had the reservoirs not been installed. The 
reason reservoirs are installed may be, in some cases, that the flow at 
the point of installation is too uniform to combine satisfactorily with 
conditions in lower tributaries. Therefore, uniformity of reservoir- 
discharge flow is decidedly not the sole object. The fact that such 
discharge is not as uniform as the natural flow would have been is of 
no consequence in summing up the usefulness of the reservoir system. 
The difficulties that Colonel Chittenden expresses, with reference 
to the effective operation of a system of reservoirs like that proposed 
by the speaker on the Ohio basin, apparently arise from the fact that 
he has not considered the great capacity of these reservoirs. Their 
effective capacity, in every case, is computed upon a total conservation 
of the water on 100% of a certain area. This is discussed at length 
in the paper. Of course, exceptional seasons would come when the 
floods would fill these reservoirs higher than at other seasons; but, if 
we take the actual records of such exceptional seasons, we shall find 
it very rare indeed when the flood run-off in any season, however pro- 
longed, is as great as the total mean run-off of that area for an entire 
year. Therefore, before a reservoir, the capacity of which has been 
