52 THE ICE AGE IN NORTH AMERICA. 



the inlet which accounts for the more rapid recession of the 

 glacier-front there than at the middle point projecting into 

 the water south of the line joining the east and west corners. 

 It was also noticeable that the falls of ice were much more 

 frequent near these corners, and the main motion of the ice 

 as afterward measured was, not toward the middle point 

 projecting into the inlet, but toward these corners where the 

 subglacial streams emerged below the water. 



No small difficulty was encountered in securing direct 

 measurements of the motion ; and, as the results may be 

 questioned, I will give the data somewhat fully. As it 

 was impossible to cross the main current of the glacier, we 

 were compelled to take our measurement by triangulation. 

 But even then it seemed at first necessary to plant flags as 

 far out on the ice as it was safe to venture. This was done 

 on the second day of our stay, and a base-line was established 

 on the eastern shore, about a mile above the mouth, and the 

 necessary angles were taken. But, on returning to repeat the 

 observations three or four days afterward, it was found that 

 the ice was melting from the surface so fast that the stakes 

 had fallen, and there were no means at command to make 

 them secure. Besides, they were not far enough out to be 

 of much service. It appeared also that the base-line was 

 on a lateral moraine, which was, very likely, itself in motion. 

 But by this time it had become evident that the masses of 

 ice uniting to compose the main stream of motion retained 

 their features so perfectly from day to day that there was no 

 difficulty in recognizing many of them much farther out 

 than it was possible to plant stakes. Accordingly, another 

 base-line was established on the east side opposite the pro- 

 jecting angle of ice in the inlet. From this position eight 

 recognizable points in different portions of the ice-field were 

 triangulated — the angles being taken with a sextant. Some of 

 the points were triangulated on five different times, at inter- 

 vals from the 11th of August to the 2d of September. Others 

 were chosen later and triangulated a less number of times. 



The base-line finally chosen (marked B on Fig. 24) was at 



