As fast as the ice assumed this liquid or softened state, it would be squeezed out toward the region where, because of the thinning of the glacier, it would enter a field where pressure melting did not occur. It would then resume the solid state, and thence journey to the margin of the ice in the ordinary manner. We thus can imagine how such a glacier as occupied the northern part of this continent could have moved from the central parts toward its periphery, as we can not do if we assume that the glacier everywhere lay upon the bed rock. There is no slope from Lake Erie to the Ohio River at Cincinnati. Knowing that the ice moved down this line, there are but two methods of accounting for its motion: either the slope of the upper surface to the northward was so steep that the mass would have been thus urged down, the upper parts dragging the bottom along with them, or the ice sheet for the greater part of its extent rested upon pressure-molten water, or sludge ice, which was easily squeezed out toward the front. The first supposition appears inadmissible, for the reason that the ice would have to be many miles deep at Hudson Bay in order that its upper surface should have slope enough to overcome the rigidity of the material and bring about the movement. We know that any such depth is not supposable. The recent studies in Greenland supply us with strong corroborative evidence for the support of the view which is here urged. The wide central field of that area, where the ice has an exceeding slight declivity, and is unruptured by crevices, can not be explained except on the supposition that it rests on pressure-molten water. The thinner section next the shore, where the glacier is broken up by those irregular movements which its wrestle with the bottom inevitably induces, shows that there it is in contact with the bed rock, for it behaves exactly as do the valley glaciers of like thickness.