G2

77/A' ALPINE PEG IONS.

Thus streams are formed, which rush down the slope of theglaciers, working out a channel for themselves, till at last theyare engulfed in a crevasse. From the depths of this, the plash offalling water comes up with a pleasant and occasionally almostmusical note. At times, however, in the morning, when thesestreams burst the bonds in which the night’s frost has imprison-ed them, they are apt to startle unaccustomed travellers. Iwas walking one morning over a glacier in Dauphine, when 1heard a sharp crack close at hand; this was followed by ariving bursting sound, as though the glacier itself were goingto open under my feet: then came a louder crash, and then arush of falling water, which explained the mystery. Thechannel of one of these streams had been frozen up during thenight in several places, and the noise was caused by theaccumulating waters forcing their way along their wontedcourse.

As we walk up the glacier we shall probably get a sight ofone or two moulins. These are deep well-like .cavities intowhich a stream plunges with a loud roar. They are formed inthe following way: if, owing to some inequality in the bed of aglacier, a small surface crack is opened, intercepting one of thechannels of drainage, the stream of course plunges into it, andin falling scoops away the surrounding ice. The sides of thecrack, as the glacier advances, will be pressed together again,and united by regelation; and the same motion forwards willtend to neutralise the effect of this water saw, and retain themoulin in nearly the same position on the glacier. One of thebest-proofs which I have ever seen of the correctness of thisexplanation was on the Roseg glacier near Pontresina . Wewere walking along one of the surface streams, and came toa place where its course was intersected by a long narrowcrevasse, which extended for a considerable distance across theglacier, and engulfed all the water. At this point the chasmhad been widened by the friction of the falling stream, so that