e tropical districts. Their effect is,
however, probably small as compared with that massive drift which we
have now to note.

The tropical waters when they attain high latitudes are constantly
cooled, and are overlaid by the warmer contributions of that tide, and
are thus brought lower and lower in the sea. When they start downward
they have, as observations show, a temperature not much above the
freezing point of salt water. They do not congeal for the reason that
the salt of the ocean lowers the point at which the water solidifies
to near 28 deg. Fahr. The effect of this action is gradually to press down
the surface cold water until it attains the very bottom in all the
circumpolar regions. At the same time this descending water drifts
along the bottom of the ocean troughs toward the equatorial realm. As
this cold water is heavier than that which is of higher temperature
and nearer the surface, it has no tendency to rise. Being below the
disturbing influences of any current save its own, it does not tend,
except in a very small measure, to mingle with the warmer overlying
fluid. The result is that it continues its journey until it may come
within the tropics without having gained a temperature of more than
35 deg. Fahr., the increase in heat being due in small measure to that
which it receives from the earth's interior and that which it acquires
from the overlying warmer water. Attaining the region of the tropical
current, this drift water from the poles gradually rises, to take the
place of that which goes poleward, becomes warm, and again starts on
its surface journey toward the arctic and antarctic regions.

Nothing is known as to the rate of this bottom drift from the polar
districts toward the equator, but, from some computation which he has
made, the writer is of the opinion that several centuries is doubtless
required for the journey from the Arctic Circle to the tropics. The
speed of the movement probably varies; it may at times require some
thousand years for its accomplishment. The effect of the bottom drift
is to withdraw from seas in high latitudes the very cold water which
there forms, and to convey it beneath the seas of middle latitudes to
a realm where it is well placed for the reheating process. If all the
cold water of circumpolar regions had to journey over the surface to
the equator, the perturbing effect of its flow on the climates of
various lands would be far greater than it is at present. Whe