measure its general warmth, depends upon the movement of the Gulf
Stream from the tropics to high latitudes. If by any geographical
change, such as would occur if Central America were lowered so as to
make a free passage for its waters to the westward, the glaciers of
Greenland and of Scandinavia would disappear, and at the same time the
temperature of those would be greatly lowered. Thus the most evident
cause of glaciation must be sought in those alterations of the land
which affect the movement of the oceanic currents.

Applying this principle to the northern hemisphere, we can in a way
imagine a change which would probably bring about a return of such an
ice period as that from which the boreal realm is now escaping. Let us
suppose that the region of not very high land about Bering Strait
should sink down so as to afford the Kuro Siwo, or North Pacific
equivalent of our Gulf Stream, an opportunity to enter the Arctic Sea
with something like the freedom with which the North Atlantic current
is allowed to penetrate to high latitudes. It seems likely that this
Pacific current, which in volume and warmth is comparable to that of
the Atlantic, would so far elevate the temperature of the arctic
waters that their wide field would be the seat of a great evaporation.
Noting once again the fact that the Greenland glaciers, as well as
those of Norway, are supplied from seas warmed by the Gulf Stream, we
should expect the result of this change would be to develop similar
ice fields on all the lands near that ocean.

Applying the data gathered by Dr. Croll for the Gulf Stream, it seems
likely that the average annual temperature induced in the Arctic Sea
by the free entrance of the Japan current would be between 20° and 30°
Fahr. This would convert this wide realm of waters into a field of
great evaporation, vastly increasing the annual precipitation. It