Page:Whymper - Scrambles amongst the Alps.djvu/392

336 under the influence of cold, rends away grain by grain, until at length, as in the accompanying diagram at  and, little ravines are formed upon each side of the quartz-vein.

If, on the other hand, the eroded rocks continue to experience the grinding of glacier, nothing of this kind results. The tendency of the quartz to protrude is incessantly checked, because, at the slightest suspicion of protrusion, it is attacked by the ice with increased power. If by any chance it becomes elevated above the surrounding rock, it bears off the weight of the ice from the surrounding rock, and this condition of affairs continues until both quartz and gneiss are brought to the same level.

There is little difference of opinion about these matters. It is perfectly well known that projections in the bed of a glacier are attacked by the ice, and that depressions escape abrasion through the protection afforded by the eminences. Hence it is that ultimately all angles and almost all curves are obliterated from the surfaces of rocks upon which glaciers work. Hence it is that in a district which has been severely eroded by glacier we find the rocks more flat—that is, less convex—than in one which has been less eroded.

It is evident, then, that glacier does not and cannot dig away into soft places occupying limited areas. This is not a matter of opinion, but a certainty; and it seems to me to be entirely