STORAGE RESERVOIRS. 109 



to such a height that the discharge at the outlet is equal to the entire inflow. The same 

 is true of the restraining influence of forests. A sudden and heavy precipitation of 

 short duration, which might produce a severe freshet in a deforested region, would 

 probably experience considerable retardation, and even reduction, if it should fall upon 

 a forest-covered region ; but if the rains continue long enough to exhaust the retentive 

 capacity of the forest soil, to fill all the springs and replenish the ground storage, then 

 forests cease to be any protection whatever. In fact, the presence or absence of forests 

 in a vast watershed like that of the Mississippi River is without appreciable influence 

 upon the great floods. 



" In the case of floods, which are the results of combinations of discharges from 

 the various tributaries, reservoirs may actually operate to increase the combination. 

 Take for example the natural reservoirs at the sources of the Mississippi. While they 

 restrain the flood excess in that stream, they keep up a heavy flow for some time after 

 the flood has passed. If this larger flow happens to come in with a flood crest at the 

 junction of some tributary below, it will actually increase the combination over what 

 would have been the case without the reservoirs. In the French investigations, presently 

 to be described, the dams proposed for restraining floods were to have open sluiceways 

 without means of closing them. In the ordinary flow of the stream all the water could 

 pass through. But they were to be so proportioned that when the flow should pass a 

 certain point the surplus would be retained in the reservoir, the outflow being always 

 limited by the capacity of the open sluices. The arrangement was, therefore, precisely 

 like that of a natural lake without a dam across the outlet. The outflow could never 

 be entirely restrained, and it would increase in proportion to the height of water in the 

 reservoir. Now, in the case of a large stream like the Rhone, where flood combination 

 is the really dangerous thing, it was found that these reservoirs, had they actually been 

 constructed, would have increased certain floods. They would have maintained a 

 heavy retarded flow on some tributaries which in their natural condition would have 

 entirely run out before the arrival of floods from other tributaries. As it happened, 

 this retarded flow in the one case would have come upon a flood crest in the other, and 

 would actually have increased the natural combination. This, of course, could not be 

 true of reservoirs with closed sluices, unless, as above stated, the reservoirs were entirely 

 filled with the flood passing over them. 



"It is, therefore, clear that the efficiency of reservoirs in moderating great floods 

 would have to be a matter of judicious management in controlling combinations quite 

 as much as of actual capacity. 



"Another matter to be noted in this connection is that flood protection and indus- 

 trial use are not entirely compatible objects. To serve the former purpose alone the 

 reservoir should be kept empty until the flood arrives, so that its whole storage may 

 be available. But this might leave the reservoir only partly filled when its supply is 

 needed for other purposes. Generally, therefore, the whole capacity of reservoirs built 



