674 
single tornado, repeated reports of two cumulonimbus 
clouds or thunderstorms coming together just before 
the tornado appears indicate that the junctions are a 
result rather than a cause of the tornado circulation [11]. 
In the absence of much light coming through the thick 
clouds from above, reflections from the ground pre- 
sumably cause the green- and yellow-tinted clouds near 
tornadoes. The cumulonimbus clouds are often very 
dark because of their great vertical extent. 
Often a tornado is preceded by a violent thunder- 
storm (northeast of the tornado) [11]. More often a 
tornado is followed by a thunderstorm (southwest of 
the tornado), often a severe one. Or a thunderstorm 
center may follow a path parallel to the path of the 
tornado (usually northwest of the tornado). Green- 
colored lightning and even ball lightning have been 
reported. On the other hand, sometimes a tornado 
occurs even though no thunderstorm is reported. 
Rain almost always accompanies a tornado [17], but 
heavy rain is not likely within the tornado itself. 
Heavy showers are common, especially following a 
tornado, and frequently yield the greatest precipitation 
just northwest of the tornado path. Hail is also common 
[17], sometimes occurring as much as two hours before 
and/or after a tornado, often northwest and sometimes 
southeast of the tornado track. Its duration is shorter 
than that of rain, but it may cause more damage than a 
minor tornado. Very large hailstones [6], ranging up to 
discs 7 in. in diameter (weighing 3 lb), and hail accu- 
mulating to a depth of several inches have been reported 
near or even along tornado paths. 
The air preceding a tornado is usually warm and 
humid, whereas the air following it is usually cool and 
dry. Research at Parks College of Aeronautical Tech- 
nology showed that 111 out of 142 tornadoes occurred 
on the line of highest surface dew point. The drop in 
temperature with the arrival of the first thundersquall 
usually exceeds the difference between the temperatures 
of the preceding and following air masses. 
Within a few miles outside the tornado path it is 
not unusual for pressure drops of over 3 mb to occur, 
and in extreme cases the change amounts to about 
10 mb. This indicates that the tornado is surrounded 
by a low-pressure area which, with its attendant winds, 
has a radius of about 5-10 mi [2], which is intermediate 
in size between the parent low and the tornado. In 
this respect, the tornado is similar to an ordinary thun- 
derstorm, which has an inflow current coming from as 
far as about 5 mi away from the edge of the active 
cloud [22]. 
A tornado is usually preceded by a general southerly 
wind and followed by a general westerly wind. Research 
at Parks College showed that 92 out of 132 tornadoes 
occurred on the line of highest surface wind speed. 
Winds in the vicinity of the tornado are likely to be 
strong and variable, sometimes reaching full hurricane 
force outside the tornado proper. The reasons for these 
strong winds are that (1) the parent cyclone causing 
the prevailing winds is more intense than is usual for a 
cyclone, (2) the secondary low surrounding the tornado 
has steep pressure gradients, (8) there may be heavy 
LOCAL CIRCULATIONS 
thunderstorm squalls, and (4) small local whirlwinds 
are often present. Any one of these effects alone or in 
combination with the others may account for a zone a 
few miles wide with scattered damage to buildings and 
trees. 
Properties of Tornadoes 
Central Pressure and Wind Speed. The pressures and 
winds within a tornado must be determined largely by 
indirect methods since there are very few instrumental 
measurements. These methods include a detailed analy- 
sis of the tornado damage and a theoretical study of the 
relationship between wind and atmospheric pressures. 
Two fundamental causes of damage are (1) the terrific 
winds, exerting pressures and suctions, and carrying 
missiles, and (2) the sudden changes of the prevailing 
atmospheric pressure. Since both are related to pres- 
sure, the combined pressure effects observed on struc- 
tures cannot be attributed either to atmospheric 
pressures or to wind pressures alone without the intro- 
duction of errors. Cracks in the ground [12] and com- 
pletely smashed objects cannot be used in estimating 
wind pressures when the true causes of the damage 
are unknown. 
Although Hazen [11] thought that tornadoes were 
local high-pressure areas, barographs in tornadoes show 
pressure drops usually up to 25 mb. In one case an 
aneroid barometer fell 200 mb (Minnesota, Aug. 20, 
1904). The actual atmospheric pressure drops are prob- 
ably greater because the pressure changes inside build- 
ings are sluggish and because the buildings may not be 
in the exact center of the tornado path. Espy’s experi- 
ments [4] showed that low pressure was not the cause 
of the forced removal of feathers from chickens. Neither 
this denudation nor the breaking and peeling of bark 
from tree trunks can be used as an argument for such 
pressure. Explosions of windows, doors, walls, and roofs 
not under the direct influence of tornadic winds (when 
the vortex is just off the earth’s surface) are clear indi- 
cations of the sudden arrival of a very low atmospheric 
pressure and a very large horizontal pressure gradient. 
For a frictionless vortex with inflow, the gradient would 
be inversely proportional to the cube of the distance 
from the center (except very near the center). 
Determinations of maximum wind speeds in torna- 
does are rough approximations. Calculations of wind 
based on structural failures and impacts of flying ob- 
jects yield values ranging from less than 100 mph to 
one case of more than 300 mph, enough to demolish the 
strongest buildings. 
Ferrel [6] gives the following theoretical relationships 
between wind and atmospheric pressure in a frictionless 
vortex surrounded by a windless environment: The total 
kinetic energy of the wind is equal to the work that 
would be done by the pressure gradient in bringing the 
air into the vortex from the environment. The wind 
speed is the same as the free-fall velocity of an object 
dropped in a vacuum from the height having the 
same pressure in the environment as that at the ground 
in the vortex. For example, a pressure of 900 mb, or a 
decrease of 100 mb, would give a wind of about 300 
