could be added, and the only possible source of increased volume is by giving the Great Lakes an artificial outlet by way of the Illinois River. A certain amount of diversion or rearrangement of tributaries would be possible, though not very practicable, as in getting rid of the load of sediment from the Missouri, or in turning the Tennessee through the Big Hatchie River to reduce the Ohio floods. But in all of these radical schemes the possible benefits to be derived are far outweighed by the inevitable difficulties and disadvantages. It is unquestioned that the tributary system must remain as it is now.
The construction of artificial outlet channels to take off the excess volumes which produce floods has been suggested many times, and it seems likely that if constructed in sufficient numbers they would prove effective. There would, however, always be great difficulty in keeping the outlets sufficiently free from sand so that their usefulness should remain unimpaired. A second and more serious objection to the outlet scheme is that in the lower valley, where the flood control is most difficult and the flood damage most wide-spread, the outlets would have to be provided by turning the water over the low-lying bottom lands. Outlet reservoirs could not be maintained because of the sand and mud with which they would be speedily filled. Under such conditions, therefore, the outlet plan clearly defeats one of the chief objects of flood control—the protection of rich plantations covering thousands of square miles on the river bottoms.
The solution by building a series of reservoirs in the head-waters 'of the chief tributaries appears to be the cheapest and most certain remedy for all these difficulties. By the construction of reservoirs the excess of water which produces flood stages could be impounded and held up with these important results: excessive and destructive high-water stages could not occur, while, on the other hand, by regulating the discharge from the reservoirs, a more even flow of water could be maintained at all times, eliminating to a large degree the losses from diminished water supply, reduced power and fouling of streams incident to the low stages of late summer and early autumn. As soon as the irresistible rush of flood waters is stopped the sapping and caving of banks will be reduced to a minimum, with the efficiency of revetments increased many fold; finally, cutting down the flood volumes means a great diminution of the amount of sediment carried, and a marked alleviation of the sand-bar evil. The reservoirs would, moreover, eliminate floods from the whole system, not merely from the lower course. The prevention of the annual flood damage in the Ohio would in itself be worth the entire cost of the reservoirs, yet until the work of control is carried to the headwaters no relief can be secured for that populous valley.
The solution by head-water reservoirs, of all proposed plans, has probably provoked the most discussion—on the one side, those who
