been seen forcing their way southward at considerable speed through
ordinary surface ice, which was either at rest or moving in the
opposite direction. The train of these bergs, which moves upward from
the south polar continent, west of Patagonia, indicates also in a very
emphatic way the existence of a very strong northward-setting current
in that part of the ocean.

* * * * *

We have now to consider the causes which could bring about such great
extensions of the ice sheet as occurred in the last Glacial period.
Here again we are upon the confines of geological knowledge, and in a
field where there are no well-cleared ways for the understanding. In
facing this problem, we should first note that those who are of the
opinion that a Glacial period means a very cold climate in the regions
where the ice attained its extension are probably in error. Natural as
it may seem to look for exceeding cold as the cause of glaciation, the
facts show us that we can not hold this view. In Siberia and in the
parts of North America bordering on the Arctic Sea the average cold is
so intense that the ground is permanently frozen--as it is, for
instance, in the Klondike district--to the depth of hundreds of feet,
only the surface thawing out during the warm summers. All this region
is cold enough for glaciers, but there is not sufficient snowfall to
maintain them. On the other hand, in Greenland, and in a less though
conspicuous degree in Scandinavia, where the waters of the North
Atlantic somewhat diminish the rigour of the cold, and at the same
time bring about a more abundant snowfall, the two actions being
intimately related, we have very extensive glaciers. Such facts, which
could be very much extended, make it clear that the climate of glacial
periods must have been characterized by a great snowfall, and not by