FORESTS, RESERVOIRS, AND STREAM FLOW 295 
of the practical difficulties that stand in the way of any comprehensive 
reservoir scheme for controlling floods. 
In 1903 the great flood of the Kaw River brought up the reservoir 
question again. Ex-Senator Burton, of Kansas, advocated the plan 
very urgently, stating in a speech at Kansas City that he “would have 
tens of thousands of reservoirs beginning at the head-waters of the 
stream and coming right down.” <A Board of Engineer Officers was 
appointed to investigate the practicability of providing against future 
disasters such as this flood had caused. The reservoir idea had made 
so deep an impression upon the public mind that a specific considera- 
tion of that feature of the problem was requested. In its report* the 
Board found adversely to the scheme on the ground that its great 
cost, conservatively estimated at $11 000000, and the annual loss from 
the withdrawal of the necessary lands from occupancy, conservatively 
estimated at nearly $600000, would not be justified on the ground 
of flood protection alone. Owing to the character of the country, 
this last consideration was particularly strong. The only real justifi- 
cation of so extensive a system in a country so largely devoted to 
agriculture would be its use in irrigation and power, and, if it became 
necessary for these purposes, doubtless a portion of it would be built. 
The most elaborate study of this subject ever undertaken until 
very recently was made by the French Government, to determine 
whether reservoirs could be utilized to prevent the recurrence of such 
great disasters as the floods of 1856 in the valleys of the Rhone and 
other streams. A full résumé of these studies is given in the writer’s 
report already referred to on Reservoir Sites in the Arid Regions. 
The conclusion was the same that has been reached in every similar 
investigation. An interesting feature of the system then considered 
was that the reservoirs were to have sluices permanently open so 
that it would not be possible to close them entirely. They would 
operate, it was expected, to hold back a definite percentage of flood 
discharge—enough to keep the floods below the dams within safe 
limits. They would thus act automatically, just as forests are sup- 
posed to do. This was all right so far as the individual tributaries 
were concerned, but it was found, when the possible effect upon 
tributary combination in the main stream was considered, that, by 
holding back earlier portions of freshets and prolonging their run- 
*Senate Doc. 160, F8th Cong., 2d Session, pp. 14-17. i 
