more reached the low grounds at the base of the Alps, but did not flow so far as their predecessors of the preceding or second glacial epoch. The snow line of this third glacial epoch stood at an average level of about 4,400 feet below the present.
Each glacial epoch was necessarily marked by profound glacial erosion, and the consequent formation of massive sheets of ground moraine in the lower reaches of the great valleys, and of huge terminal moraines at or opposite their mouths. Enormous quantities of shingle and gravel were at the same time swept outward by the rivers escaping from the ice—each series of terminal moraines being thus closely associated with its separate and distinct set of fluviatile deposits. No difficulty is found in separating those successive accumulations of gravel. They form terraces lying one within the other at three successive levels. The highest rises upon an average 250 to 300 feet above the present rivers; the surface of the middle terrace is about 100 feet below the surface of the highest, and about the same distance above the level of the lowest terrace. Each terrace rests upon solid rock, and it is obvious, therefore, that the several epochs of gravel accumulation have been separated by epochs of active river erosion. This remarkable valley-within-valley formation is clearly the result of climatic changes. The highest terrace indicates the action of flooded rivers escaping from the glaciers of the first glacial epoch. These glaciers then disappeared or shrank into comparative insignificance, and an interglacial epoch of active valley erosion succeeded—the rivers cutting their way down for a hundred feet or more into the solid rocks. Next came the second glacial epoch, and the lowered valley bottom was again deeply covered with gravel. The glaciers of this stage then in their turn retired, and a second interglacial epoch supervened, when the rivers as before deepened their channels, working down through the older gravels and excavating the underlying rocks. Thereafter the third glacial epoch ensued, and a new series of gravels was deposited at a lower level than the preceding accumulation. Lastly, this third glacial epoch passed away and the rivers again trenched the fluvio-glacial gravels, the upper surface of which is now much above the reach of the greatest floods.
What relation, then, does the löss bear to the glacial and interglacial accumulations of the alpine lands? Fortunately to this question a definite reply can be given. It is dovetailed with the glacial deposits in such a manner as to show that its formation has taken place at successive epochs. Thus it occurs occupying an interglacial position between the accumulations of the first and second, and between those of the second and third glacial epochs. When we pass down the valley of the Rhine a similar succession is encountered. In the wide plain lying between the Vosges and the Black Forest, löss is met with on the same geological horizons, overlying the gravel terraces of the first and the second glacial epochs. Not only so, but even the youngest or lowest
