A rock hill was sometimes left without covering of drift after having been severely worn by the ice. Such a hill is known as a _roche moutonnée_. An example of this type of hill occurs three miles north of east of Baraboo at the point marked z on Plate XXXVII. This hill, composed of quartzite, is less symmetrical than those shown in Figs. 28 and 29. Its whole surface, not its stoss side only, has been smoothed and polished by the ice. This hill is the most accessible, the most easily designated, and, on the whole, the best example of a _roche moutonnée_ in the region, though many other hills show something of the same form. [Illustration: Fig. 29.--The same hill after it has been eroded by the ice. A the stoss side. B the lee side.] It was not the hills alone which the moving ice affected. Where it encountered valleys in its course they likewise suffered modification. Where the course of a valley was parallel to the direction of the ice movement, the ice moved through it. The depth of moving ice is one of the determinants of its velocity, and because of the greater depth of ice in valleys, its motion here was more rapid than on the uplands above, and its abrading action more powerful. Under these conditions the valleys were deepened and widened. Where the courses of the valleys were transverse to the direction of ice movement, the case was different. The ice was too viscous to span the valleys, and therefore filled them. In this case it is evident that the greater depth of the ice in the valley will not accelerate its motion, since the ice in the valley-trough and that above it are in a measure opposed. If left to itself, the ice in the valley would tend to flow in the direction of the axis of the valley. But in the case under consideration, the ice which lies above the valley depression is in motion at right angles to the axis of the valley. Under these circumstances three cases might arise: (1) If the movement of the ice sheet over the valley were able to push the valley ice up the farther slope, and out on the opposite highland, this work would retard the movement of the upper ice, since the resistance to movement would be great. In this case, the thickness of the ice is not directly and simply a determinant of its velocity. Under these conditions the bottom of the valley would not suffer great erosion, since ice did not move along it; but that slope of the valley against which the ice movement was projected would suffer great wear (Fig. 30). The valley would therefore be widened, and the slope suffering greatest wear would be reduced to a lower angle. Shallow valleys, and those possessing gentle slopes, favor this phase of ice movement and valley wear. [Illustration: Fig. 30.--Diagram showing effect on valley of ice moving transversely across it.]