TORNADOES AND RELATED PHENOMENA 
cause of it. Jones [15] finds that lightning in tornadic 
thunderstorms gives more intense radio static (sferics) 
than lightning in ordinary thunderstorms. 
Espy [4] theorized that heat and latent heat of 
vaporization supplied the energy for the strong con- 
vection in a tornado. To supply thermal energy fast 
enough, a violent updraft of warm moist air would be 
required. Since the process cannot get started by itself, 
the heat of vaporization is a maintaining factor rather 
than the initial cause, which must be an external source 
of energy. In other words, the pendant cloud formation 
aids the tornado, but does not cause it. A California oil 
fire (1923) heated the surface air sufficiently to generate 
hundreds of whirlwinds, one of which was violent 
enough to kill two people and wreck a house. 
Hail has been regarded by some meteorologists as 
being the result of vortex motion [17], and a tornado 
as being an unusually severe hailstorm cloud. On the 
other hand, Showalter [19] recently suggested that hail 
might be the initial mechanism that causes a tornado. 
Hail falling from the overhanging top of a cumulo- 
nimbus cloud would cool the layer of dry air below it 
by conduction and evaporation until the original dry 
inversion (or stable zone) would reach unstable equi- 
librium. As this instability would not extend above the 
cooled layer, the available convective energy would be 
concentrated in the lower atmosphere. 
Horizontal Motion as a Source of Energy. Showalter 
[19] pomts out the necessity of lifting or horizontal 
convergence for the establishment of free convection. 
In either case, horizontal inflow is required. Concerning 
frontal lifting, Humphreys [13] states that vortex mo- 
tion cannot usually be generated within the warm air 
mass being lifted, regardless of how much windshift 
there is at the front. For vortex motion to develop in 
the warm air it is necessary for the warm air itself to 
have angular momentum about the center of convec- 
tion. This requirement may be fulfilled in the trough of 
a squall line. Horizontal convergence at a front is 
effective in producing vortex motion if both air masses 
rise, for in this case the frontal windshift supplies the 
necessary angular momentum. Humphreys [13] says 
this condition is best fulfilled at an upper cold front, 
which Willett [24] qualifies as having a small horizontal 
temperature contrast. 
Taylor [20] regards a tornado as the extreme develop- 
ment of a small intense secondary low formed on a 
front. There is much evidence in favor of the existence 
of a secondary low (‘tornado cyclone”) around a tor- 
nado [2]. Before the tornado forms, there is ample 
angular momentum about the center of the tornado 
cyclone. To generate and maintain a tornado, a certain 
minimum rate of inflow is required to create an acceler- 
ation of the wind sufficient to outweigh by far the 
frictional losses. 
Garbell [9] mentioned that a tornado may follow a 
combination of (1) strong vertical motion indicated by a 
rapidly growing cumulonimbus cloud, and (2) cyclonic 
circulation. Incipient vortices are present in most thun- 
derstorms; it is only when the gyratory motion becomes 
677 
very intense that they reach from the cloud base to the 
ground. 
Wegener [23] introduced the theory that a tornado is 
the same as the rotating thunderstorm squall cloud, one 
end of which reaches down to the earth. Aside from 
insufficient observations to substantiate this theory, 
one objection is that the numbers of tornadoes rotating 
clockwise and counterclockwise are far from being equal, 
as would be required if there were no preference as to 
which end of the squall cloud would dip to the earth. 
Cause of the Low Pressure in a Tornado. The biggest 
difficulty in explaining the maintenance of a tornado 
is to account for the failure of the low pressure at its 
center to disappear in the face of strong surface inflow. 
A common explanation for the low pressure is that 
the pressure reduction is caused by the outward cen- 
trifugal force of the whirl. If the low pressure is caused 
by the whirl, the latter must exist first, probably 
developing between two strong countercurrents. For 
the low to develop there must be a slight outflow 
which, by the approximate conservation of angular 
momentum, would lead to a reduction in wind speeds 
to values less than those of the original countercurrents. 
The fatal objection to this explanation is that wind 
speeds in a tornado far exceed the speeds of the winds 
prevailing on either side of it. 
If inflow rather than a slight outflow is assumed, the 
centrifugal force of the inflowing air is greatly increased 
(approximately inversely as the cube of the radial 
distance). Some have argued that such a terrific in- 
crease in centripetal acceleration requires a great 
strengthening of the pressure gradient and, of course, 
a very low pressure at the center. Actually the mass 
inflow would have exactly the opposite effect on the 
central pressure, namely, the filling of the low. After 
a shght inflow, the increase in winds would make the 
centrifugal force larger than the pressure gradient force, 
making the inflow zero or even reversing it. The fallacy 
is in the original assumption of unlimited inflow. 
The foregoing discussion shows that a tornado cannot 
be generated by either inflow or outflow at the surface. 
Hare [10] proposed that the volume of the uprushing 
air exceeds that of the inflowing air, thereby leaving 
relatively low pressure. Redfield disagreed, saying there 
is no such thing as “suction.”’ A surface low-pressure 
area can be produced by excessive removal of air up- 
ward only if the pressure gradient force acting upward 
exceeds gravity before as well as during the tornado. 
This requires the formation of a low aloft, the generation 
of which still needs to be explained. 
Divergence aloft is necessary [3] not only to generate 
and maintain the low aloft, but also to prevent inflow 
from filling the low at the surface. Since it is known 
that there is inflow at the surface and vertical ascent in 
the core, continuity demands outflow aloft. Such out- 
flow may have been witnessed once before a tornado 
formed [17]. Espy thought there was enough eviction 
of air aloft to build up a ring of higher pressure in the 
surroundings. For deepening the surface low, the out- 
flow aloft must exceed the inflow at the surface. This 
is possible due to the dragging away (entrainment) of 
