In the great depression between the quartzite ranges, the moraine bends westward, showing that the ice advanced farther on the lowlands than on the ridges. As the moraine of this low area approaches the south range, it curves to the east. At the point southwest of Baraboo where the easterly curve begins to show itself, the moraine lies at the north base of the quartzite range; but as it is traced eastward, it is found to lie higher and higher on the slope of the range, until it reaches the crest nearly seven miles from the point where the eastward course was assumed. At this point it crosses the range, and, once across the crest, it turns promptly to the westward on the lower land to the south. Here the ice advanced up the valley between the East bluff (east of the lake) and the Devil's nose (Plate XXXVII), again illustrating the fact that lowlands favor ice advance. The valley between the Devil's nose and the East bluff is a narrow one, and the ice advanced through it nearly to the present site of the lake. Meanwhile the restraining influence of the "nose" was making itself felt, and the margin of the ice curved back from the bottom of the bluff near Kirkland, to the top of the bluff at the end of the nose. Here the edge of the ice crossed the point of the nose, and after rounding it, turned abruptly to the west. Thence its edge lay along the south slope of the ridge, descending from the crest of the ridge at the nose, to the base of the ridge two miles farther west. Here the ice reached its limit on the lowland, and its edge, as marked by the moraine, turned southward, reaching the Wisconsin river about a mile and a half above Prairie du Sac. The course of the terminal moraine across the ridges is such as the margin of the ice would normally have when it advanced into a region of great relief. The great loop in the moraine with its eastern extremity at k, Plate XXXVII, is explained by the presence of the quartzite ridge which retarded the advancing ice while it moved forward on either side. The minor loop around the Devil's nose is explained in the same way. Both the main loop, and the smaller one on the nose, illustrate the point made earlier in the text. The narrow and curious loop at m, is of a slightly different origin, though in principle the same. It is in the lee of a high point in the quartzite ridge. The ice surmounted this point, and descended its western slope; but the thickness of the ice passing over the summit was so slight that it advanced but a short distance down the slope before its force was exhausted, while the thicker ice on either side advanced farther before it was melted. _Glacial Deposits._