TORNADOES AND RELATED PHENOMENA 
than an average tornado. Yet, in special cases, they have 
been known to change into tornadoes. 
Dust Devils. The dust devil [14, 25], ranging from 
less than 10 ft to more than 100 ft in diameter, is about 
Yoo to Yo as large as a tornado. It has no pendant 
cloud, but a whirling column of dust or sand, in which 
the size of the particles increases with the distance from 
the center. Dust devils occur most frequently on very 
hot days over dry terrain, and are caused by strong 
surface convection in a slightly irregular wind field. 
Since the direction of rotation is accidental, it may be 
clockwise as often as counterclockwise. The horizontal 
rotation and upflow within a dust devil usually exceed 
20 mph. The dust devil is reported to have an average 
height of 600 ft and to last approximately the same 
number of hours as it is thousands of feet high. If the 
prevailing wind is less than 3 mph, a dust devil tends 
to seek higher terrain rather than to travel with the 
wind. 
Waterspouts. Waterspouts [13] fall into two classes: 
tornadoes over water, and fair-weather waterspouts. 
Most waterspouts are tornadoes, with cyclonic rota- 
tion and the characteristic pendant from a cumulonim- 
bus cloud. The spout may appear more transparent 
through its center than through its edges, because of the 
presence of a small eye, as in a tornado. Sometimes a 
waterspout appears to be double, the pendant of con- 
densed water vapor being inside a spreading column, 
which appears to be a fountain of spray from the 
surface. Waterspouts frequently appear in groups, as 
many as thirty having been visible from a ship in one 
day. Owing to the small amount of friction over a water 
surface, the motion in a waterspout is more nearly 
tangential, with less inflow and upflow than in a tor- 
nado. The reduced friction and larger moisture supply 
over a water surface are not the chief factors deter- 
mining intensity, for, if they were, waterspouts would 
be stronger than tornadoes. 
The fair-weather waterspout, like the dust devil, is a 
low-level whirl, rotating in either sense. It is caused by 
irregular air flow and pronounced surface instability, 
more the result of high humidity than of high tem- 
perature. 
The water surface under a waterspout is either raised 
or lowered depending on whether it is affected more by 
the atmospheric pressure reduction [17] or by the wind 
force. The waterspout does not lift any significant 
amount of water from the surface, as shown by the 
precipitation of mostly fresh water on a ship passing 
through a waterspout on the ocean. It often dissipates 
on reaching a shore line because of the absence of suffi- 
cient inward acceleration to compensate for frictional 
losses. 
Concluding Outlook 
The subject of tornadoes is very old—the ancient 
savants Pliny the Elder, Seneca, and Lucretius wrote 
about them. Since then, progress in the study of tor- 
nadoes has been slow. Everdingen [5] noted that the 
literature on tornadoes in the United States in the last 
half century has, for the most part, been confined to 
679 
compilations of such statistics as the distribution and 
frequency of occurrence and description of resulting 
damage. What is lacking is the application of satis- 
factory hydrodynamical theories to the frictional vor- 
tices of tornadoes, dust devils, and waterspouts. For 
example, the horizontal velocity profile could be found 
by an approximate solution of the differential equation 
expressing the rate of change of angular momentum of 
a fluid particle in terms of the net torque from surface 
and internal friction, as has been done for gaseous spray 
nozzles. 
Suggested items for study are (1) the outflow aloft 
(method of removal and destination of removed air), 
(2) the formation by action of hail, and (3) the sense 
of rotation, as examples of explaining facts, checking 
proposed theories, and verifying accepted ideas, re- 
spectively. For this work, many more surface and upper- 
air observations are needed. 
Since the tornado is a local circulation, the meteor- 
ological data ought to be gathered over micronetworks 
which could detect the development of a small second- 
ary cyclone and would accurately locate squalls, thun- 
derstorms, and hail showers with respect to the tornado. 
After a tornado occurs, careful surveys should be 
made of the damage to determine winds and atmos- 
pheric pressure drops. A standardized questionnaire 
[11] should be used in personal interviews and should 
be published in local newspapers with a request for 
replies. Copies of local photographs should be obtained 
for analysis, the best photographs beg those which 
include the top of the cumulonimbus cloud formation 
above the pendant cloud rather than the pendant alone. 
A more complete knowledge of the small-scale cir- 
cumstances attending tornadoes is needed for the under- 
standing of the nature and causes of tornadoes. Such 
knowledge will put tornado forecasting and tracking on 
a firmer basis. 
REFERENCES 
A valuable bibliography, which was published after this 
list of references was prepared, is that by Miss H. P. Kramer, 
“Selective Annotated Bibliography on Tornadoes.”’ Meteor. 
Abstr. & Bibliogr. (publ. by Amer. Meteor. Soc.) 1:307-332 
(1950). 
1. Baupwin, J. L., ‘Preliminary Report on Tornadoes in the 
United States during 1943 and Totals and Averages, 
1916-42, by States.” Mon. Wea. Rev. Wash., 71:195-197 
(1943). 
2. Brooks, E. M., ‘‘The Tornado Cyclone.’ Weatherwise, 
2:32-33 (1949). 
3. Brun, D., Physical and Dynamical Meteorology, 2nd ed. 
Cambridge, University Press, 1939. (See pp. 303-306) 
4. Espy, J. P., The Philosophy of Storms. Boston, Little 
Brown, 1841. (See pp. 304-373) 
5. Everpinacen, E. van, “The Cyclone-Like Whirlwinds of 
August 10, 1925.” Verh. Akad. Wet., Amst., 28:871-889 
(1925). 
6. Ferrer, W., A Popular Treatise on the Winds, 2nd ed. 
New York, Wiley, 1893. (See pp. 347-449) 
7. Finuey, J. P., Tornadoes. New York, C. C. Hine, 1887. 
8. Frora, 8. D., ‘Tornadoes in Kansas.”’ Mon. Wea. Rev., 
Wash., 57:97-98 (1929). 
