150 The American Geologist. September, i^« 
glacier ;i large area is fed from a relatively small center. It 
is true that this area] expansion is the result in part of a dim- 
inution in thickness; but this in itself would be hardly suffi- 
cient to account for the broadening of the ice-sheet away 
from the center. 
From these centers of dispersion the material is dragged 
and carried out toward the margin of the ice, where that part 
of it remains which does not come within the grasp of the 
streams. Jt accumulates here in greater abundance than else- 
where for the reason that the source of supply is the entire 
area from center to margin, and because the ice can carry it 
no Farther. Leaving out of consideration the extreme terminal 
deposits. this drift can be carried no farther than the peripheral 
parts of the ice-sheet because the glacial motion is too slow. 
and the thickness of the ice too slight, but chiefly because time 
enough is not allowed for its slow removal. Like the flood 
plain of a river, drift is here laid aside temporarily, to be re- 
moved when a smaller burden of work is given the ice to do. 
This vast accumulation of detritus about the margin of our 
drift-covered area earty led to the belief that the glacier had 
cut deeply into the region over which it had passed; but that 
this is not so is well proved b} T the drainage system and the 
perfection of development of preglacial topography. Upon 
the basis of the above argument it seems to signify, rather, a 
brief occupancy of the land by the ice-sheet. 
How long, in years, this may have been, seems hardly 
capable of approximate estimation in the present state of our 
knowledge of the rate of glacial erosion. There are too many 
uncertain factors. The ice is retarded by friction on the bot- 
tom and its motion here must be slower than above. Then, 
too, the ice is less resistant than the rock over which it passes, 
and the passage of an ice-sheet over a given distance would 
not uecessarily mean that an included rock fragment moved 
at the same rate, for in all probability it lagged behind, scor- 
ing not only the bed rock but also the ice. The time of 
passage of material from center to periphery may therefore 
l»e long and the work of erosion slow. Moreover, if there is 
considerable material being thus dragged along beneath the 
ice. much of the energy of erosion is exerted in grinding this 
material, particle against particle, along the numerous slipping 
