JANUARY 25, 1884. ] 
winds that distinguish the tornado from other 
storms, but this cause is close at hand. 
Admit for a moment that there is no friction 
between the air and the ground. We should 
then have a tall vertical cylinder of air, spin- 
ning around near the centre at a terrific speed, 
at the base as well as aloft, and consequently 
developing a great centrifugal force. As a re- 
sult, the density of the central core of air must 
be greatly diminished. Most of the central air 
must be drawn out by friction into the whirl- 
ing cylinder, and prevented from returning by 
the centrifugal force. The core will be left 
with a feeling of emptiness, like an imperfect 
vacuum. If there were any air near by not 
controlled by the centrifugal force, it would 
rush violently into the central core to fill it 
again. Now consider the effect of friction with 
the ground. The lowermost air is prevented 
from attaining the great rotary velocity of the 
upper parts, and consequently is much less 
under the control of the centrifugal force, 
which is measured by the square of the ve- 
locity. The surface-air-is therefore just what 
is wanted to fill the incipient vacuum: so it 
rushes into the core and up through it with a 
velocity comparable to that of the whirling it- 
self; and this inward-rushing air is the destruc- 
tive surface-blast of the tornado. 
_ This explanation, first proposed by Mr. Fer- 
rel a few years ago, is most ingenious and 
satisfactory. Moreover, he has followed its 
several parts by close mathematical analysis, 
and shown that the moderate antecedent con- 
ditions are amply sufficient to account for all 
the violence of the observed results. 
_ There are still several points to be considered. 
The whirling motion has’ been described as 
corresponding in nearly all cases with that of 
northern cyclones; and yet it cannot be sup- 
posed that the indraught winds of a tornado are 
drawn from sufficient distances to show the ef- 
fect of the earth’s deflective force: it is more 
probable that the tornado is to be regarded as 
a small whirl within a larger one, for the warm 
and cold winds are probably part of a large 
cyclonic system in which differential and rotary 
motions are established ; and, when such winds 
form a small local whirl of their own, it will 
rotate in the same direction as they do, from 
right to left. Fora like reason the planets ro- 
tate on their axes in the direction in which they 
revolve around the sun. The constant direc- 
tion of rotation in tornadoes may therefore, by 
itself, be taken as evidence that their cause is 
not in a stagnant atmosphere, like that of the 
desert-whirls, but is connected with the con- 
flicting currents of a large, gentle cyclone. 
SCIENCE. 
97 
The progressive motion of the tornado- 
centre is so constant in its direction to the 
north-east or east, that it cannot depend on local 
conditions within itself, but must rather result 
from its bodily transportation by the prevail- 
ing winds, with which the tornado-tracks agree 
very well in direction and rate. It will last 
till the lower warm air, which constituted the 
original unstable mass, is exhausted. This 
generally happens in about an hour, when 
it has traversed a distance of nearly thirty 
miles. 
The tornado thus constituted may be likened 
to a. very active air-pump, carried along a few 
hundred feet above the ground, sucking up the 
air over which it passes. It is for this reason 
that the surface-winds are so nearly radial. 
For this reason an enclosed mass of air, as in 
a house, suddenly explodes as the vacuum is 
formed over it; and as the air rushes to the 
centre, and there expands and cools, its vapor 
becomes visible in the great funnel, or spout, 
pendent from the clouds above. No rain can 
fall at the centre. Bodies much heavier than 
rain are lifted there, instead of dropped: so 
the rain must rise through the central core, 
and fall to one side of the storm, or before or 
behind it. If the expansion be very great, 
and the altitude reached by the drops rather 
excessive, then they will be frozen to hail- 
stones before falling. Hail-storms and torna- 
does commonly go together: they mutually 
explain each other. Electricity has no impor- 
tant part to play in the disturbance. 
It was stated under cyclones that their cen- 
tral barometric depression had two causes, — 
the overflow caused by the central warmth, 
and the dishing-out of the air by centrifugal 
force. The first of these is ordinarily regarded 
as the effective cause of the wind’s inward 
blowing. It has already been pointed out that 
the second and greater part of the depression 
is also effective in drawing in the winds when 
friction decreases their rotary velocity. We 
may now call attention to a third cause of cen- 
tripetal motion in the cyclone already alluded 
to, in which it is like the tornado. The upper 
winds move with great rapidity, and cause a 
strong barometric depression at the centre of 
their whirling; but at the base of the storm, 
where friction with the sea, or still more with 
the land, reduces the lower wind’s motion, and 
so diminishes their centrifugal force, we may 
have an indraught of the tornado style, in 
which the centrifugal diminution of central 
pressure in the upper winds is an effective 
cause of centripetal motion in the lower winds. 
While this is not the principal cause of surface- 
