THE OZARKIAN AND ITS SIGNIFICANCE 533 
Now there is nothing in the way of supposing — but on the con- 
trary much reason to believe—that the period of great elevation 
was one of comparative dryness and that the climate became 
moister in the latter part when the elevation was not so great. 
This would make the ice accumulation lag behind the elevation 
and the cold. 
2. But again: We must, of course, make a wide distinction 
between the annual snowfall and the rate of accumulation ; for 
this latter is the result only of the annual excess of snowfall over 
waste by melting and evaporation, and this excess may be to any 
degree small. 
2 AGeitas, the wvceness atid therefore the wezgkt ot the 
snow that we are here concerned with, it must be remembered 
again that there is still another and much more important source 
of waste antagonizing local accumulation and therefore increase 
of thickness, viz., the run-off—not the run-off as water but as 
ice by glacial motion. After a certain thickness is attained this 
run-off completely balances the annual excess over waste by 
evaporation and melting, and prevents, farther increase of thick- 
ness. This is the case now in all glacial regions. In the Alps, 
the Himalayas, etc., there is no indefinite increase of thickness 
because the run-off by glacial motion completely balances the 
annual excess. Similarly in Greenland, and the Antarctic con- 
tinent although there is great excess of snowfall over waste 
by evaporation and melting, yet there is no increase in the 
thickness of the ice sheet because the excess is balanced by the 
run-off of the ice into the sea and the formation there of ice- 
bergs which are carried away by oceanic currents. 
So also in Glacial times, after a certain thickness of snow had 
accumulated on any given area— say the Canadian highlands — 
the farther increase would be prevented, because balanced by 
the ice-sheet motion in all directions. Any increase of thick- 
ness would require not only excess but zucrease of excess and 
would be extremely slow because always kept in check by the 
increased run-off. 
To sum up: Suppose, then, the highlands about Hudson 
