This was near the base of the _seracs_, and the streams here were abundant. Comparing the shapes of some of them with that of the ice-bands lower down the glacier, a striking resemblance was observed. Fig. 60 is the plan of a deep-cut channel through which a stream flowed on the day to which I now refer. Fig. 61 is the plan of a seam of white ice sketched on the same day, low down upon the glacier. Instances of this kind might be multiplied; and the result, I think, renders it certain that the white ice-seams referred to are due to the filling up of the channels of glacier-streams by snow during winter, and the subsequent compression of the mass to ice during the descent of the glacier. I have found such seams at the bases of all cascades that I have visited; and in all cases they appear to be due to the same cause. The depth to which they penetrate the glacier must be profound, or the _ablation_ of the ice must be less than what is generally supposed; for the seams formed so high up on the Glacier du Geant may be traced low down upon the trunk-stream of the Mer de Glace.[A] [Sidenote: STREAMS AND SEAMS.] [Illustration: Fig. 60. Plan of a Stream on the Glacier du Geant.] [Illustration: Fig. 61. Plan of a Seam of White Ice on the Glacier du Geant.] [Sidenote: SCALING OFF BY PRESSURE.] These observations on the white ice-seams enable us to add an important supplement to what has been stated regarding the origin of the dirt-bands of the Mer de Glace; The protuberances at the base of the cascade are due not only to the toning down of the ridges produced by the transverse fracture of the glacier at the summit of the fall, but they undergo modifications by the pressure locally exerted at its base. The state of things represented in Fig. 57 is plainly due to the partial pushing of one crumple over that next in advance of it. There seems to be a differential motion of the parts of the glacier in the same longitudinal line; showing that upon the general motion of the glacier smaller local motions are superposed. The occurrence of the seams upon the faces of the slopes seems also to prove that the pressure is competent, in some cases, to cause the bases of the protuberances to swell, so that what was once the base of a crumple may subsequently form a portion of its slope. Another interesting fact is also observed where the pressure is violent: the crumples _scale off_, bows of ice being thus formed which usually span the crumples over their most violently compressed portions. I have found this scaling off at the bases of all the cascades which I have visited, and it is plainly due to the pressure exerted at such places upon the ice. FOOTNOTES: [A] The more permanent seams may possibly be due to the filling of the profound crevasses of the cascade. (33.) [Sidenote: COMPRESSION OF GLACIER DU GEANT.]