It was thus placed beyond doubt that the point of maximum motion of the Mer de Glace, at the place referred to, is not the centre of the glacier. But, to make assurance doubly sure, I examined the comparative motion along three other lines, and found in all the same undeviating result. This result is not only unexpected, but is quite at variance with the opinions hitherto held regarding the motion of the Mer de Glace. The reader knows that the trunk-stream is composed of three great tributaries from the Geant, the Lechaud, and the Talefre. The Glacier du Geant fills more than half of the trunk-valley, and the junction between it and its neighbours is plainly marked by the dirt upon the surface of the latter. In fact four medial moraines are crowded together on the eastern side of the glacier, and before reaching the Montanvert they have strewn their debris quite over the adjacent ice. A distinct limit is thus formed between the clean Glacier du Geant and the other dirty tributaries of the trunk-stream. Now the eastern side of the Mer de Glace is observed on the whole to be much more fiercely torn than the western side, and this excessive crevassing has been referred to _the swifter motion of the Glacier du Geant_. It has been thought that, like a powerful river, this glacier drags its more sluggish neighbours after it, and thus tears them in the manner observed. But the measurement of the foregoing three lines shows that this cannot be the true cause of the crevassing. In each case the stakes which moved quickest _lay upon the dirty portion of the trunk-stream_, far to the east of the line of junction of the Glacier du Geant, which in fact moved slowest of all. [Sidenote: LAW OF MOTION SOUGHT.] The general view of the glacier, and of the shape of the valley which it filled, suggested to me that the analogy with a river might perhaps make itself good beyond the limits hitherto contemplated. The valley was not straight, but sinuous. At the Montanvert the convex side of the glacier was turned eastward; at some distance higher up, near the passages called _Les Ponts_, it was turned westward; and higher up again it was turned once more, for a long stretch, eastward. Thus between Trelaporte and the Ponts we had what is called a point of contrary flexure, and between the Ponts and the Montanvert a second point of the same kind. [Sidenote: CONJECTURE REGARDING CHANGE OF FLEXURE.]