question the conclusion; but, as regards the _principle_ laid down, that
rotation can produce no motions when once the form of equilibrium is
attained, we must unequivocally dispute it. If our atmosphere were of
uniform density, the rotation of the earth would cause no current such
as we have described; with our atmosphere as it is, the result will be
different. The momenta of two portions of matter are the products of
their inertiae by their motions, and, in the present case, we must take
the inertiae of equal spaces. A cubic inch of air at the surface, and at
three miles above the surface, is as 2 to 1; but their centrifugal
velocity varies only as the radii of the respective spheres, or as 1320
to 1321. In the polar regions, therefore, the momentum of the surface
air preponderates, and, in this case, the _surface_ current is towards
the equator, and the upper current towards the poles. When, however, the
centrifugal velocity is considerably increased in a lower latitude, and
the curvature of the surface becomes more and more inclined to the
direction of that resolved part of the centrifugal force, which is
always _from_ the axis, the surface layers will evince a tendency to
leave the surface, and an intermingling will then take place in the
space between latitude 70d and 50d, or in latitude 60d. As this layer is
continually urged on in the same direction by the surface layer of
latitudes above 60d, the upper layer now becomes a current setting
_towards_ the equator, and, consequently, the back current occupies the
surface. Now, considering that the rarefying action of the sun is
elevating the air under the equator, there must necessarily be an upper
current from the equator to the poles; so that if we conceive the two
currents to meet about latitude 30d, there will be a second
intermingling, and the current from the poles will again occupy the
surface. Thus, we regard a part of the effect of the trades to the
rotation of the earth, which is the chief impelling power at the poles,
as the sun is at the equator; and the latitudes 60d and 30d will be
marked by some especial phenomena of temperature, and other
meteorological features which do actually obtain. These would be much
more marked if the irregular configuration of land and sea, the
existence of mountain chains, and the different heating power of
different latitudes, owing to the unequal distribution of the land, did
not interfere; and the currents of the air (disr