﻿M. A. Heim on Glaciers, 505 



longitudinal section would have given the form shown in 

 fig. 14. 



The " sliding of the filaments towards the middle " is some- 

 what plausible if we consider that at the margins of the gla- 

 cier, where the motion is much slower, a heaping up of masses, 

 and in consequence a movement of them toward the middle 

 must take place. This, however, is the case only at the upper- 

 most part of the glacier, where the wide channel of the neve 

 opens into the narrower glacier-valley. There the chief mass 

 moves toward the middle, and at the margins only that is 

 driven after which has plenty of room. In the glacier itself, 

 as the undisturbed course of the central and lateral moraines 

 shows, it is otherwise. The glacier behaves as a number of half- 

 cylinders lying one within another, the inner sliding more quickly 

 down the valley than the outer ; there is no trace of heaping-up 

 at the margins. The only question is, are these cylinders infi- 

 nitely thin ? that is, does the velocity increase continuously to- 

 ward the middle ? or does each set of infinitely thin ones, which 

 do not admit of much displacement of one over another, form a 

 thicker cylinder ? that is, does the velocity increase per saltust 

 do displacements take place as in the gypsum stream (fig. 8)? If 

 I have correctly indicated the displacement-fissures in the Rhone 

 Glacier, they are, in this comparison, the divisions between two 

 thicker half-cylinders. 



When, analogous the valleys of the Rhone, the Trift, and 

 many other glaciers, a very steep place is made in the valley, 

 the gypsum stream breaks down over it exactly as does a glacier, 

 in the form of stairs. At the foot of the fall the terraces gra- 

 dually become small transversely running arcuate elevations, 

 and at length vanish altogether : in glaciers, as is well known, 

 they leave the dirt-bands behind. 



If small pieces of wood are placed upright in the gypsum 

 stream, they very soon incline toward the bottom of the valley 

 — a proof that the velocity diminishes downwards. 



That in a winding valley the maximum of motion approaches 

 the concave side, that the glacier scarcely reacts on small lateral 

 glens in its valley, that it presses aside or overpowers small 

 lateral affluents, &c, all these can be very beautifully demon- 

 strated in the stream of killed gypsum. Of course the forms of 

 the surface are never those of the glacier, because there is no 

 ablation, and at the progressing lower end of the gypsum stream 

 the motion is a rolling ; while it is not so in the glacier, on 

 account of the ablation, or this rolling only occurs, quite sub- 

 ordinately, in winter, if even then. 



If killed gypsum is caused to move slowly from a sh low 

 cavity into two valleys in opposite lirections, at the summit 



