The crests of these hills reach an elevation of between 1,000 and 1,100 feet. They appear to correspond with the level of the first peneplain recognized in the Devil's lake region. It was in the second cycle of erosion, when their surroundings were brought down to the new base-level, that these hills were left. West of Camp Douglas, there are still higher elevations, which seem to match Gibraltar rock. The Friendship "mounds" north of Kilbourn City, the castellated hills a few miles northwest of the same place, and Petenwell peak on the banks of the Wisconsin (Plate XXXII), are further examples of the same class of hills. All are of Potsdam sandstone. In addition to the "castle" rocks and base-level plain about Camp Douglas, other features should be mentioned. No other portion of the area touched upon in this report affords such fine examples of the different types of erosion topography. In the base-level plain are found "old-age" valleys, broad and shallow, with the stream meandering in a wide flood-plain. Traveling up such a valley, the topography becomes younger and younger, and the various stages mentioned earlier in the text, and suggested in Plate XIX, Figs. 1 and 2, and Plate XX, Fig. 1, are here illustrated. [Illustration: WISCONSIN GEOL. AND NAT. HIST. SURVEY. BULLETIN NO. V., PL. XXXI. Stand Rock. Upper end of the Upper Dalles.] [Illustration: WISCONSIN GEOL. AND NAT. HIST. SURVEY. BULLETIN NO. V., PL. XXXII. Petenwell Peak.] CHAPTER V. THE GLACIAL PERIOD. The eastern part of the area with which this report deals, is covered with a mantle of drift which, as already pointed out, has greatly modified the details of its topography. To the consideration of the drift and its history attention is now turned. _The drift._--The drift consists of a body of clay, sand, gravel and bowlders, spread out as a cover of unequal thickness over the rock formations beneath. These various classes of material may be confusedly commingled, or they may be more or less distinctly separated from one another. When commingled, all may be in approximately equal proportions, or any one may predominate over any or all the others to any extent. It was long since recognized that the materials of the drift did not originate where they now lie, and that, in consequence, they sustain no genetic relationship to the strata on which they rest. Long before the drift received any special attention from geologists, it was well known that it had been transported from some other locality to that where it now occurs. The early conception was that it had been drifted into its present position from some outside source by water. It was this conception of its origin which gave it the name of _drift_. It is now known that the drift was deposited by glacier ice and the waters which arose from its melting, but the old name is still retained.