420 



LAKE CHELAN 



in elevation. So it is with a river at the point of junction of 

 branches. The surface of the water must be practically at the 

 same level in all cases, but the bottoms of the channels differ 

 b}' the difference in depth of the streams at their point of junc- 

 tion. This fact affords us a measure of the minimum thickness 

 of the ice at any place. It cannot have been less than the ver- 

 tical distance between the bed of the main glacier and that of 

 the tributary, and, indeed, must in all cases have been greater, 

 owing to the thickness of the tributary. 



LAKE CHELAN, AT THE NARROWS 



To extend the comparison between a river and a glacier, it 

 may be added that the central portion of the glacier flows faster 

 than the bottom and sides, as they are retarded by friction, just 

 as in the case of a stream. This is demonstrated by the gradu- 

 ally increasing curvature of the lines crossing the glacier, such 

 as transverse lines of dirt or crevasses. In the upper portion of 

 the glacier these may be straight, or nearly so, but lower down 

 become more and more curved, with the convexity downward. 



A glacier is constantly receiving upon its surface rock, gravel, 



