MONT BLANC AND THE MER DE GLACE. 43 



the lower portions, it assumes, after awhile, an inclina- 

 tion which gives it a dip up the valley. Its depth may 

 be ten or a hundred feet, and its width, which is a few 

 inches at first, may grow to fathoms.* The two walls 

 generally approach each other downward, and we may 

 sometimes safely descend to the bottom. The wall-ice is 

 absolutely immaculate, with a greenish-blue transpar- 

 ency. Down in the crevasse we hear the rills coursing 

 through the substance of the glacier, and sometimes the 

 central torrent rumbling along the bottom. The sur- 

 face of the glacier is white and granular from the action 

 of the sun. Pools of water rest here and there, pure, 

 cold and refreshing, and numerous rills flow over the 

 surface, discharging themselves through perforations in 

 the ice-mass, into some subglacial stream. 



All goes well. Now we reach the median moraine, 

 which, from Montanvert, we had mistaken for the oppo- 

 site side. This is a longitudinal ridge of icy fragments 

 and commingled boulders and sand. The remaining half 

 of the glacier is strewed with rocks and glacial debris. 

 Now the real difficulties begin. The crevasses grow into 

 immense yawning chasms, and the ice-masses between 

 them rise up like mere knife-edges, on which one must 

 balance himself in the transit. The mutual intersections 

 of these crevasses continually interrupt the pathway, 

 and we are compelled sometimes to descend by a series 

 of ten or twenty steps cut in the ice by the guides, and 

 then to ascend by as many more. We cannot disguise 

 the fact that the element of danger enters into the pas- 



*In 1824, Forbes measured a crevasse at the base of the Glacier du Ge'ant 

 -which had a breadth of not less than 1214 feet (370 metres). Payot, Guide 

 Jtinerah-f, p. 146, note. 



