450 JMKOAINIS Sake SVLII NA ICA VOLO 
the oscillations. There is nothing inthe phenomena of the drift 
that requires it, nor has anything been discovered in the behavior 
of the Greenland ice-sheet that suggests it. The summers were 
seasons of melting during the phase of advance as well as during 
that of retreat, and it may be doubted whether the most skillful 
observer could have detected any difference in the summers of 
the two phases unless he had made the most refined gauge 
measurements on the volume of the water discharged. Whether 
the ice-front would advance a little or recede a little or remain 
stationary during a long period of years was a matter of the 
utmost delicacy of adjustment. I doubt whether the average 
annual temperature for a period of years need differ more than 
two or three degrees to determine whether the ice-front shall 
stand at Fort Wayne or at Defiance. 
Again, there seems every reason to suppose that the general 
advance of the ice-sheet was after the same manner as the retreat, 
only that the oscillations were reversed, and that it required 
the same duration of time. From the halt at Defiance the ice 
would retreat to Toledo, and then readvance to Fort Wayne, and 
so on. If this is a true assumption, then it would require us to 
suppose that the ice-sheet must be able to advance from Toledo 
to Fort Wayne in the same time, and hence at the same rate, as 
it is supposed to retreat over the same interval during the gen- 
eral recession. But it is manifestly impossible to suppose that 
the ice advanced at a rate of anything like 1000 or even 100 feet 
aday. The great wide lobe that advanced up the Maumee valley 
isnot to be compared with the Muir, or the Karajak glacier, but 
rather with some part of the Greenland ice-cap that ends on 
land. It was not breaking off and floating away as bergs. It 
seems certain that its motion must have been very slow, probably 
not over two to five feet a day, even at the maximum. 
But if it seems necessary to put the period of oscillation at 
3000 years or more, it becomes a matter of comparatively small 
importance whether it bea little more or less. We have no means 
of knowing just what it was, but if 3000 years seems to be an 
extreme minimum and 6000 years seems better, there is no pos- 
