uption of Sumbawa, in 1815, ashes were carried to the islands of
Amboyna and Banda, which last is about 800 miles east from the site of
the volcano. Yet the southeast monsoon was then at its height.[188] This
apparent transposition of matter against the wind, confirmed the opinion
of the existence of a counter-current in the higher regions, which had
previously rested on theoretical conclusions only.

That a corresponding interchange takes place in the seas, is
demonstrated, according to Humboldt, by the cold which is found to exist
at great depths within the tropics; and, among other proofs, may be
mentioned the mass of warmer water which the Gulf stream is constantly
bearing northwards, while a cooler current flows _from_ the north along
the coast of Greenland and Labrador, and helps to restore the
equilibrium.[189]

Currents of colder and therefore specifically heavier water pass from
the poles towards the equator, which cool the inferior parts of the
ocean; so that the heat of the torrid zone and the cold of the polar
circle balance each other. The refrigeration, therefore, of the polar
regions, resulting from the supposed alteration in the distribution of
land and sea, would be immediately communicated to the tropics, and from
them its influence would extend to the antarctic circle, where the
atmosphere and the ocean would be cooled, so that ice and snow would
augment. Although the mean temperature of higher latitudes in the
southern hemisphere is, as before stated, for the most part, lower than
that of the same parallels in the northern, yet, for a considerable
space on each side of the line, the mean annual heat of the waters is
found to be the same in corresponding parallels. If, therefore, by the
new position of the land, the formation of icebergs had become of common
occurrence in the northern temperate zone, and if these were frequently
drifted as far as the equator, the same degree of cold which they
generated would immediately be communicated as far as the tropic of
Capricorn, and from thence to the lands or ocean to the south.

The freedom, then, of the circulation of heat and cold from pole to pole
being duly considered, it will be evident that the mean temperature
which may prevail at the same point at two distinct periods, may differ
far more widely than that of any two points in the same parallels of
latitude, at one and the same period. For the range of temperature, or
in other words, the curvature