DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 351 
derived if the system could be successfully and practically applied, but Mr. Harts. 
were based on its inherent defects, which were considered so serious 
as entirely to overwhelm its advantages. 
In their publications, the advocates of the reservoir system lay 
much stress on the increased depths that are practicable by releasing 
certain quantities of water at low river stages. No mention, however, 
is made of the the fact—and sight of it seems completely lost—that 
depths due to additional flow are invariably less the farther down- 
stream the observations go, and that the theoretical discharge neces- 
sary for navigable depths is not the same at all points along the same 
stream. The Ohio and Tennessee are not canals having uniform 
widths, depths, and slopes. 
Navigation depends, not on maximum depths, but on minimum 
depths; so that, when use by boats is referred to, it is of no value to 
say that we may have 5 ft. of water in our river, when at shoal places 
there may not be more than 2 or 3 ft. These minimum depths determine 
the limit of use to be made of the stream. Ordinarily, the depths 
of channels vary with natural conditions. Where rivers widen, the 
depths grow relatively less. To say that we may have a 5-ft. depth in 
the Tennessee River at Knoxville, Chattanooga, or Riverton, by no 
means proves that at Colbert Shoals, Muscle Shoals, and various other 
places, the depths would be as great. As a- matter of fact, the depths 
on the gauges at these towns bear only a remote relationship to the 
navigable depths on the shoals. As the discharge of the Tennessee 
increases, the slope at Colbert Shoals decreases; while at the “Suck” 
the slope increases, owing to the gorging effect of the banks. Local 
conditions thus affect the flow to an important degree. 
Then, too, it is believed that the reservoirs would soon fill with 
-silt. All reservoirs must necessarily be located near the head-waters 
of the streams, since nowhere else will land be cheap enough or other- 
wise available for this purpose. In such localities the erosion in the 
water-shed is at a maximum, the slopes of all streams being steeper 
in their upper than in their lower portions. In the case of the 
Tennessee River it is found that the area claimed to be available for 
furnishing water for the reservoirs specified covers a water-shed of 
19 520 sq. miles. The scour from this area is enormous, and is now 
carried down mainly at high water. It cannot be claimed by the most 
enthusiastic supporters of this theory that these reservoirs would not 
rapidly fill with silt due to this erosion. It is folly to claim that the 
hillsides draining into the reservoirs can ever be adequately pro- 
tected against this scour by forests. As mentioned above, the pro- 
duction of silt in river valleys is due far more to the cultivation of 
fields and farms than to the cutting away of the large trees of the 
forests, as these are promptly followed by a second growth almost as 
useful as the first for protective purposes. The cleared farms are 
