As has been said, the first announcement of the Cincinnati ice dam was thought to give a natural and sufficient explanation for certain high-level gravel terraces occurring in the upper Ohio Valley. Subsequent investigations have brought to light other considerations which must more or less modify the first conclusions. It still remains true, however, that the ice dam accounts most naturally for many of the slack-water deposits which occur in the valley of the upper Ohio and its tributaries, while there are many areas which are yet but inadequately explored, but which promise important light upon the problem when the facts are all obtained. At the same time it appears that some of the terraces in the Alleghany and Monongahela Rivers are slightly higher than the obstruction at Cincinnati, compelling the advocates of the ice-dam theory to suppose some very probable changes of level since the deposition of the terraces which were at first supposed by Prof. Lesley to be so completely explained by it.
But more important is the bearing of recent discoveries upon the extent to which glacial gravels accumulated in the gorge of the upper Ohio and Alleghany Rivers, as shown in the section in the lower right-hand corner of Map. I. All along the Alleghany and Ohio Rivers there are remnants of gravel accumulations, from fifty to sixty feet deep, resting upon rock shelves about three hundred feet above the present rock bottom of the Ohio. There is now little reason to doubt that during the Glacial period the floods of water which poured into the Alleghany and the Ohio from all their northern tributaries brought along silt, gravel, and bowlders enough to fill up this rocky gorge with great rapidity, down as far probably as Wheeling. As the Alleghany River received glacial floods and glacial débris in great quantities, while the Monongahela did not receive any, it will be seen that the Monongahela must have been dammed by both the silt and the water which came down the Alleghany.
Instances in which the water of a tributary is dammed by that of the main stream will occur to any one upon a little reflection. Whenever one large tributary perceptibly rises, it raises the water level of the main stream as well above as below the junction, while a large rise in the main stream may temporarily reverse the current in a tributary. The Columbia River, for example, in Oregon, is subject to very extensive floods at seasons of the year when the Willamette is comparatively low. At such times a current sets up stream past the city of Portland. I remember, also, hearing, when a boy, the story of a June freshet on the Poultney River, in Vermont, caused by a succession of thundershowers about its head waters. The rise in the lower part of the stream amounted to thirty or forty feet. The thing which fixed itself most deeply in my mind was that a milldam upon Hub-
