The other view to which I have alluded takes little or no account of river-erosion, but maintains that the floods of the Ice Age were sufficiently great to reach the highest levels at which river-gravels and loams occur. It is likely enough that, under present conditions, we can form but a very inadequate idea of the vast bodies of freshwater which formerly swept down our valleys, but we may be pardoned if we express our inability to conceive of our European rivers flowing with a breadth of many miles, and a depth of two or three hundred feet. A few years before his death, Mr. Darwin made a suggestion to me, which I think gives us the true solution of the problem. He thought that during an Ice Age great beds of frozen snow might have accumulated over the low-grounds outside of the glaciated areas (in the manner I have already described), and that many valleys might have been filled to a considerable depth during a large part of the year with blown snow, afterwards congealed. In autumn, when the running water failed, the lines of drainage might in many cases be more or less choked, and it would be a mere chance whether the drainage, together with gravel, sand, and mud, would follow precisely the same lines during the next summer. Such action being repeated year after year, it might well happen that many river-valleys might become largely filled with rudely alternating layers of frozen snow and fluviatile detritus. And if this were so, the flooded rivers in summer would be enabled to overflow much wider and more elevated tracts than they could otherwise have reached. As the climate became less excessive, we can conceive of the frozen snows gradually melting, and of river-detritus being deposited at lower and lower levels in the valleys. The probability of such frozen masses having choked up valleys and impeded the drainage during the Ice Age is not a mere plausible conjecture. In the far north of Alaska--in a region which was certainly not overflowed by the North American ice-cap--extensive sheets of ice occur, more or less deeply buried under thick soil. Nor can there be much doubt that these ice-masses date back to the Glacial period itself, seeing that in the soils which overlie them we meet with remains of the mammoth and other contemporaneous mammalian forms. Here, then, we have direct proof of the fact of frozen snow and ice having accumulated in the hollows of the land outside of the glaciated areas.[N] [N] I have given Mr. Darwin's views, and discussed the origin of the Pleistocene fluvio-glacial deposits at some length in _Prehistoric Europe_, chaps, viii. and ix. To this work I refer for detailed geological evidence in support of the view advocated above.