Outer Glacial Drift. — Upham. 159 
Only in norlluvcsUTn Washington, on the basin of Puget 
sound. occui)yinj;- a width of 75 to 150 miles between the 
Cascade range on the east and the Olympic mountains and 
the ranges of \^ancouver Island on the west, was a projection 
from the continental icefields of British Columbia able to 
flow into the United States, while of course the. far greater 
part of the ice even there was fomied by snowfall on the ad- 
joining mountains. 
So marked dependence of the Cordillcran ice accumulatior» 
upon the grand topographic features is a complete proof, as 
I think, that the icefields of the Puget Sound basin and of the 
Canadian province on the north were produced chiefly by local 
snowfall, not by invasion or inflow from the central part of 
the Cordilleran glaciated area. Similarly, as I also think, 
the southern marginal parts of the Keewatin and Laurentide 
icefields grew in depth and areal extent chiefly by snowfall on 
those outer tracts, within the first 20 or 50 or 100 miles back 
from the extreme ice boundary, in nearly as large degree as 
the central and thicker parts of the ice-sheet. Yet some gla- 
cial outflow took place from the central areas throughout the 
maximum stage of glaciation ; and there was much peripheral 
outflow during all the stages of growth, culmination, and de- 
cline. With the slow ice motion, the very long duration of 
the Glacial period can be well appreciated when we bear in 
mind the glacial transportation of boulders a thousand miles 
from the east side of James bay to southern Minnesota, and 
the equally long distance of the outermost glacial drift in 
southern Russia from the Scandinavian mountains that sup- 
plied some of its boulders. 
Why the ice-sheets shrank earlier from the flat country of 
the .Mississippi and Alissouri basins, and of Russia, than from 
the more hilly or mountainous regions in both North America 
and Europe, and why morainic drift hills and ridges were 
formed on each continent principally during the late and clos- 
ing Wisconsin stages o^ the Glacial period, are questions that 
may be probably answered, for the first, by the flat contour and 
far inland situation of these countries first permanently un- 
covered from the ice-fields ; and, for the second, by the long 
continuance of glaciation. whereby the ice became more charg- 
ed with ^Irift in its lower part, and by the more vigorous glacial 
