the adjacent country, and in a much greater degree by glacial action, as it is a well-established fact that the river had cut its way (except for a short distance below St. Anthony) prior to the glacial epoch. On account of this lateral furrowing the bluff walls present a broken and serrated appearance, but this, when rightly considered, does not in the least militate against the correctness of the cataract theory. The valley gorge, which is from two to six miles, is at present somewhat wider at the top where the cliffs appear, and where wind and frost have been free to act, than the water originally cut it, and the débris falling below has formed a talus which, increased and modified by glacial action, has to a considerable extent effaced the wall-like appearance which is such a marked feature of the comparatively freshly-cut canon of Niagara.
Another fact which has hitherto received no satisfactory explanation is the deep accumulation of sand in the valley-bottom. There are no data sufficient to determine the depth of this deposit, but as the great river nowhere flows upon a rock-bed, but everywhere, except in its extreme northern section, has a sand bottom; and, as cities and villages are built within the bluffs on the compact accumulations at the sides of the present channel, we conclude that it must have a depth of several hundred feet.
If the gorge had been chiseled out by a process of gradual wear (which would have been the case if the strata had been of uniform resisting power), then the river should flow upon a rock-bed, and not upon sand, for the latter would, as it now does, protect the underlying strata from all wear. Upon the cataract hypothesis this peculiar condition can be met with an easy and satisfactory explanation. The descent of the Mississippi is very gradual. Directly at the base of the falls, wherever they may have originated, the sand-rock would be cut down to a depth determined by the comparative hardness of the rocks and the volume of water. For a short distance below the descent the rock would be swept clean of sand and débris, except, perhaps, the large limestone chunks fallen from above, but as the stream came to flow more evenly below the falls, sand from the erosion above would be deposited at the bottom. Two other existing conditions would assist in the deposit: first, the very slight fall of the river; and, second, the detached masses of limestone broken from the crest of the falls would help to collect and retain the sediment, and thus the accumulation would begin. All these facts, taken in connection with the further fact that no bluffs appear above St. Anthony Falls, while they are continuous below, except where broken by the lateral erosion, have a tendency to establish the theory advanced. Just how far to the south the requisite conditions of stratification exist, we have not as yet ascertained, but they probably exist wherever the limestone-capped bluffs bound the river. If the limestone formation is wanting at any place, there rapids would have taken the place of falls,
