of the waters in their channels of circulation through the
ocean, will lead us irresistibly to the conclusion that always, in
summer and winter, there must be, somewhere within the arctic circle, a
large body of open water.

There is an under-current setting from the Atlantic, northward through
Davis' Straits, into the Arctic Ocean, and a surface-current setting
out.

The fact is proved beyond a doubt by the observations of arctic
explorers, who have seen immense icebergs drifting rapidly northward
against a strong current. This apparent anomaly could only be accounted
for by the fact that a powerful undercurrent carried them northward; and
as at least seven times more of these bergs must have been under than
above water, we can easily understand how the under-current, acting on
the larger mass of each berg, had power to carry it against the
surface-current.

This under-current is _warm_, while the upper-current is _cold_. Now we
know that according to Nature's laws, heated water, like heated air,
rises to the surface, and cold water sinks to the bottom. How, then,
comes this warm current to be underneath the cold, as soundings have
proved it to be? It is owing to the fact that the under-current is much
salter, and therefore heavier (despite its warmth), than the
surface-current; which latter, being mingled with the drainage and
ice-masses of the arctic regions, is comparatively fresh, and therefore
light as well as cold.

The hot and salt waters of the tropics are carried north by the Gulf
Stream. There are here two counteracting agents at work. Heat inclines
the Gulf Stream to rise; saltness inclines it to sink. During the first
part of its journey, as we know, its great heat prevails over the other
influence, and it flows as a surface-current. But, at a certain point
in its northward route, it meets with the cold, brackish, ice-bearing
currents that flow out of the arctic basin. Having lost much of its
heat (though still possessing a great deal more than the arctic
currents), the saltness of the Gulf Stream prevails; it dips below the
polar waters, and thenceforth continues its course as an under-current,
salt, and comparatively warm.

To state the matter briefly: The hot water, which _ought_ to keep on the
surface because of its heat, is sunk by its superabundant salt; and the
cold water, which _ought_ to sink because of its cold, is buoyed on the
surface because of its want of salt.

Now arises the