POMS Te fh Par ey ee ee, Ie RP TOMS ey Gee ty 7 me MARU ert RM 
96 SCIENCE. 
Now let us look for the explanation of these 
varied effects, and discover, if possible, the 
reason of the extremely local development of 
such intense motions. 
The explanation given for sand-whirls in the 
desert fails to provide for the excessive force 
of the tornado. A thin, warm surface-stratum 
of air would be prevented by friction with the 
ground from attaining any very excessive ve- 
locity ; and, moreover, it is often excessively 
hot without tornadoes following, and tornadoes 
often happen when the air is not perfectly still. 
Yet, as they occur most frequently on warm or 
hot afternoons, surface-warmth very probably 
re-enforces other causes up to the point of 
violent storm development. 
The existence of conflicting winds, as already 
noted, gives us more aid. So long as the cold 
wind passes under the warm, there will be 
no ereat disturbance, for the equilibrium will 
remain stable; but, if the warm wind advances 
under the cold, an unstable equilibrium may 
result. We have already seen that warm satu- 
rated air requires the smallest vertical difference 
of temperature to destroy its stability; and 
also that the saturated condition may often 
be met in the cloud-stratum, although absent 
below it. For these two reasons we may infer 
that a tendency to upset will be more frequently 
reached a few hundred or thousand feet above 
the earth than closer to the ground. Suppose 
that such a condition is reached when a mass 
of warm southerly wind has pushed itself 
below the colder north-westerly stratum: the 
surface-air will often rest quiet and become 
warm below such a meeting, for the same rea- 
son that calms occur along the equator at the 
meeting of the trades ; and a change must soon 
relieve this unnatural arrangement. The warm 
wind, feeling about for a point of escape through 
its cold cover, soon makes or finds a vent where 
it can drain away upwards ; and then the entire 
warm mass, even a mile or more in diameter, 
and often more than one thousand feet in 
thickness, begins the rotary motion already 
described in whirls and cyclones, rises at the 
centre, and passes away. Before describing 
the peculiar tornado features, let us contrast 
the storm as now developed with the two 
other kinds of storms already explained. The 
desert-whirl arises from a thin layer of hot 
dry air, warmed at the place where the whirl 
begins, ascending in a small column through a 
considerable thickness of colder air. Friction 
with the ground prevents the attainment of an 
excessive velocity ; and the ascending current 
ean lift only sand and light objects. As soon 
as the bottom-air is drained away, the whirl 
[Vou. ITLGe 0. 
stops. The cyclone is fairly compared, on 
account of its great horizontal extension, to a 
broad, relatively thin disk, with a horizontal — 
measure several hundred times greater than 
its thickness, having a spiral motion of much — 
rapidity, inward below and outward above, but. 
a central ascending component of its motion - 
so gentle that raindrops can ordinarily fall 
down through it. Its continuance depends 
largely on heat derived from vapor condensa- 
tion: it is therefore self-acting after it has 
once begun, and goes on drawing in new air 
long after the original supply is exhausted. — 
The tornado is like a cylinder, with a height 
equal to or greater than its diameter. Its 
warmth is chiefly imported to the point where 
its action begins, partly as sensible, partly as 
‘latent’ heat; but, unlike the cyclone, its ac- 
tion ceases as soon as the original mass of 
warm air escapes upward through its warm 
cover. On apprehending these peculiarities, 
we may better appreciate its farther develop- 
ment. 
The tornado has two rtione to be consid- 
ered, in addition to its general progression, — 
the spiral rotation, and the central updraught. 
The latter cannot, except under special condi- 
tions yet to be mentioned; become very rapid, 
for it depends primarily, simply on differences 
of temperature insufficient to produce very 
active motion; but the former attains a great 
velocity near the centre in virtue of the me- 
chanical principle already quoted, —the ‘ pres- 
ervation of areas.’ When a whirling body is. 
drawn toward the centre about which it swings, 
its velocity of rotation will increase as much as 
its radius of rotation decreases ; the centrifugal 
force will also increase, and with the square of 
the velocity, or inversely as the square of the 
radius. This law claims obedience from air, 
as well as from solid bodies: hence, if the air 
of a tornado mass have a gentle rotary veloci- 
ty of twenty or thirty feet a second at a thou- 
sand yards from the centre, this velocity will 
increase as the central air is drained away 
and the outer particles move inward; so that, 
when their radius is only one hundred yards, 
they will fly around at the rate of two or three 
hundred feet a second, or over one hundred — 
and fifty miles an hour. It must be under-— 
stood, however, that this requires that there 
should have been no loss of motion by friction, 
and hence can be true only for the air at a dis- — 
tance above the ground; and, further, that, in — 
spite of the great horizontal rotary motion, | 
there is still only a moderate vertical current. 
And consequently we have not yet arrived 
the cause of the violent central and upward 
“—— ' 2 
