The retardation of the ice is most evident near the sides; in most cases, the ice for a considerable distance right and left of the central line moves with a sensibly uniform velocity; there is no dragging of the particles asunder by a difference of motion, and, consequently, a compact centre is perfectly compatible with fissured sides. Nothing is more common than to see a glacier with its sides deeply cut, and its central portions compact; this, indeed, is always the case where the glacier moves down a bed of uniform inclination. But supposing that the bed is not uniform--that the valley through which the glacier moves changes its inclination abruptly, so as to compel the ice to pass over a brow; the glacier is then circumstanced like a stick which we try to break by holding its two ends and pressing it against the knee. The brow, where the bed changes its inclination, represents the knee in the case of the stick, while the weight of the glacier itself is the force that tends to break it. It breaks; and fissures are formed across the glacier, which are hence called _transverse crevasses_. [Sidenote: GRINDELWALD GLACIER.] No glacier with which I am acquainted illustrates the mechanical laws just developed more clearly and fully than the Lower glacier of Grindelwald. Proceeding along the ordinary track beside the glacier, at about an hour's distance from the village the traveller reaches a point whence a view of the glacier is obtained from the heights above it. The marginal fissures are very cleanly cut, and point nearly in the direction already indicated; the glacier also changes its inclination several times along the distance within the observer's view. On crossing each brow the glacier is broken across, and a series of transverse crevasses is formed, which follow each other down the slope. At the bottom of the slope tension gives place to pressure, the walls of the crevasses are squeezed together, and the chasms closed up. They remain closed along the comparatively level space which stretches between the base of one slope and the brow of the next; but here the glacier is again transversely broken, and continues so until the base of the second slope is reached, where longitudinal pressure instead of longitudinal strain begins to act, and the fissures are closed as before. In Fig. 27A I have given a sketchy section of a portion of the glacier, illustrating the formation of the crevasses at the top of a slope, and their subsequent obliteration at its base. [Sidenote: COMPRESSION AND TENSION.] [Illustration: Fig. 27A, B. Section and Plan of a portion of the Lower Grindelwald Glacier.]