Further Exploration of Centre Valley 



account the moulins would travel downstream with the 

 ice, so that the stream would become elongated, and 

 this doubtless happens to some slight extent. 



No substantial change of locality occurs, however, as 

 the original position of the moulin is determined by a 

 tendency of the ice to fracture or crevasse in a definite 

 position determined by a sudden change of slope in the 

 rocky floor below. Thus, before the moulin has been 

 carried far a new opening is produced in the old position, 

 and this, of course, swallows the waters of the stream, 

 leaving the lower part of the channel dry. Below our 

 particular moulin was to be seen a string of hollows in 

 the surface of the ice connected by the deserted and 

 partially obliterated channel of the stream, and there 

 could be no doubt that these represented successive 

 positions of the water-sink. On following this line of 

 deserted moulins it was found that they became less and 

 less distinct as they were traced farther and farther down 

 the stream, until eventually they were obliterated by the 

 surface melting, the lower parts of the shafts having 

 been long ago filled up by the flow of the ice. 



On climbing still farther up the glacier we came 

 upon another region of broken ice, but here, instead of 

 the pointed seracs of the great ice-fall, we found great 

 flat-topped masses separated by deep and altogether 

 impassable crevasses, and beyond these, as we after- 

 wards saw from the side of a neighbouring mountain, 

 stretched a great snowfield, rising at a low angle, and 

 from the surface of which there rose on the distant 

 horizon several snowclad peaks. This appeared to be 

 a great central snow-field filling this elevated basin 

 amongst the mountains and supplying not only the 

 s 273 



