hon; or by a gradual pouring in
at the upper surface.

From a slight motion at the commencement, affecting only that portion of
the fluid adjoining either of those places of diminution or repletion,
gradually all the water becomes influenced and acquires more or less
rapid movement. But suppose a long reservoir or canal of fluid which has
two such points of exhaustion or two of such repletion (as imagined
above), and that one of either is near each end of the vessel. If each
aperture be opened at the same moment, equal effects will be caused in
each half of the fluid towards either end of the vessel, but in the
middle there must be a neutral point at which the water falls, yet has
no horizontal motion. The converse takes place in raising the level. And
in the case of fluid drawn off or diminished in weight at one end while
increased by repletion at the other, the _whole_ body of water will move
similarly to that in the former vessel, but unequally. Hence it is
evident, that before horizontal motion occurs, an augmentation or a
diminution of pressure must take place somewhere more or less remote;
and so it is with the lighter fluid atmosphere,--which has centres,
lines, or areas of depression towards which currents flow.

Such considerations show in some degree why the barometric changes
usually precede, but sometimes only accompany, changes of weather: and,
though very rarely, occur without any sensible alteration in the wind
current of the atmosphere. An observer may be near a central point
towards which the surrounding fluid tends,--or from which it diverges.
He may be at the very farthest limit of the portion of fluid that is so
influenced. He may be at an intermediate point--or he may be between
bodies of atmosphere tending towards opposite directions.

It has been said, that "a whirlwind which sets an extended portion of
the atmosphere into a state of rapid revolution diminishes the pressure
of the atmosphere over that portion of the earth's surface, and most of
all at the centre of the whirl. The depth of the compressing column of
air will, at the centre, be least, and its weight will be diminished in
proportion to the violence of the wind." Yet this has been controverted
with respect to the _general_ effect of air in horizontal motion, and
the depth of the column in question.

Certainly there are two kinds of whirlwinds--one caused by rarefaction,
tending to lighten vertical pressure under the vortex, though n