438 



-NEW YORK STATE MUSEUM 



the Seine, the Rhone, the Loire, the Garonne and other import- 

 ant rivers resulted in a decision not to carry out the numerous 

 reservoirs which had been proposed', owing to the uncertainty 

 and doubtful efficacy of their action in floods. 



Further investigations were made after the inundations of 

 1875. These latter observations show that in the case of the 

 Garonne, a reservoir capacity of about 20,000,000,000 cubic feet 

 would be required to protect Toulouse, and one of 50,000,000,000 

 to 60,000,000,000 cubic feet to protect Agen and the rest of the 

 basin. Unfortunately, the capacity of the reservoirs which could 

 be constructed in the upper valley of the Garonne would amount 

 to only about one-sixth of that required for protecting Toulouse. 

 The investigations led to similar conclusions in the case of the 

 lower Garonne, and the other principal river basins in France. 1 



In one particular the French engineers seem to have been far 

 from right in their investigation of the utility of storage reser- 

 voirs. Thus the statement is made that such reservoirs, to be 

 useful against floods, must be kept empty throughout the whole 

 season when floods may occur. This, perhaps, may be true in 

 France, but it is not true, under the different conditions of rain- 

 fall, in a number of cases in the State of New York. Possibly, 

 the French engineers overlooked the moderating effect of a reser- 

 voir, with large water surface, upon a flood even when the 

 reservoir is filled to the flow-line. The writer has discussed this 

 question extensively in his several reports to the State Engineer 

 and Surveyor, and also in his report to the Board of Engineers 

 on Deep Waterways, and a computation is given in connection 

 with floods on Genesee river which shows that even under the 

 adverse condition of water at the crest, the temporary storage 

 on the water surface is in most cases great enough to practically 

 double the time of a flood and hence to greatly decrease its de- 

 structive effect. The conclusion on this head, therefore, is that 

 an extreme flood, which would not be effectually mitigated, even 

 1 hough a reservoir were full, would occur not oftener than once 

 in a century. 



Flood warnings. Since the flood wave in a river is progres- 

 sive, some idea can be formed in advance as to the stages of 



lAnnnlcs des Ponts et Chaussees, sixth sorios. Vol. IT, 1881. p. 5, 

 lation in Proo. Inst. C. E. 



