TROPICAL METEOROLOGY 
now fashionable to suppose actually occurs in the atmos- 
phere. In “‘The Movement of Air Across the Equator” 
[83], he enunciates this principle and applies it, under 
very restrictive special assumptions as to the nature of 
the motion. In brief, the motion must be horizontal, 
nonviscous, and autobarotropic, with a constant meridi- 
onal component. Although as far as I know no investiga- 
tion has been undertaken to find out whether vertical 
motion, frictional forces, solenoids, and variations in the 
meridional component of the wind can be neglected in 
the tropics, the results of this paper are frequently 
quoted without mention of the assumptions and applied 
as if they were perfectly general. Moreover, the efficacy 
of the deduction from the principle of the conservation 
of vorticity, as assigning a cause for cyclone formation 
in this context, depends upon the validity of the frontal 
theory in the tropics. If that theory is found to be in- 
applicable, the principle of the conservation of vorticity 
would have to be invoked in some entirely different 
context, with results that cannot at present be antici- 
pated. 
Although there are still a large number of meteor- 
ologists who adhere to the principles outlined above, 
with greater or lesser modification, the synoptic experi- 
ences of World War II, especially in the Pacific and the 
Caribbean, gave a death blow to the air-mass theory in 
its extreme form. This theory failed to pass the test to 
which all theoretical work in meteorology must finally 
come (though the evil day may be put off for almost a 
decade) ; it was found almost useless as a guide to short- 
period forecasting in low latitudes. The writer can 
vouch for this through personal experience, through 
conversation with many tropical meteorologists in the 
Pacific, the Caribbean, and the United States (where 
people with experience im all tropical areas are now to 
be found), and through the answers to the questionnaire 
issued by the Institute of Geophysics. But the reader 
need not depend on these sources of information; the 
recent literature reflects the disillusionment of a whole 
generation of tropical meteorologists. This is shown in 
attempts to doctor and patch the frontal theory in such 
a way as to take into account the many anomalous 
movements and the sudden appearances and disap- 
pearances of the fronts, especially of the equatorial 
front, or in frank abandonment of the air-mass theory 
altogether. The following modifications of the frontal 
theory, as previously outlined, have been proposed: 
1. The polar front enters the tropics, not as a typical 
surface front but as a front aloft, in the westerlies. It 
persists in that situation but may come down to the 
surface in lower tropical latitudes and at a later time.* 
Forsdyke (29, p. 82] says, 
Waves in the cireumpolar westerlies occur at the surface 
on the poleward sides of the subtropical high-pressure belts, 
and at higher levels they extend to low latitudes where the 
westerlies overlie the trades. It is tempting to identify these 
with the upper parts of cold fronts associated with the de- 
pressions of middle latitudes. At Mauritius they are frequently 
3. See Rep. No. 600-50, Publications of the Weather Divi- 
sion, USAAF, p. 15. 
867 
associated with nimbostratus cloud and rain which spread 
from the southwest above the lower current from the east- 
south-east. They should be indicated on the charts as upper 
fronts. It is possible, therefore, to have in the same part of 
the chart a lower front moving from an easterly, and an upper 
front moving from a westerly point, and crossed lower and 
upper fronts may occur. 
2. On entering the tropics, the polar front separates 
into two portions: (a) the density discontinuity or shear 
line, and (b) the polar trough (see [16, 53, 65]). 
3. The polar front is destroyed soon after entering 
the tropics. The very rational Memo 131/44 of the 
British Naval Meteorological Branch [46, p. 28], says, 
So far we have considered only discontinuities which orig- 
inate in the tropics, notably the I.T.F. What happens to 
fronts which move into the tropics from higher latitudes? 
Generally speaking, they lose their well-defined slope and 
become diffuse: their movement becomes slow and indefinite 
and by the time they get near the equator, if they get that 
far, they bear little resemblance to their text-book proto- 
types. Most of them ultimately lose their identity in the 
equatorial low pressure trough which is a great leveller of 
air mass disparities. A good many of them get no further 
than the subsidence zone of the sub-tropical highs where 
they “dry-out.” 
4. There is little or no temperature contrast at the 
equatorial front which is, indeed, simply a more or less 
narrow zone at which the trades of the two hemispheres 
converge and rise. This represents a return to the origi- 
nal position of Brooks and Braby, and the climatological 
school. Brooks and Braby were careful to point out that 
no temperature contrast could be found across the 
equatorial front in the regions where, judged by the 
_ mean winds and precipitation, it was strongest. Numer- 
ous quotations are available on this point [15, Chap. 
10; 46, 65]. Garbell [32, p. 70] says, 
The discontinuity between the two tropical air masses 
coming from the two hemispheres usually consists not so 
much in a difference in surface temperature as in a difference 
mainly in the vertical distributions of temperature, moisture, 
stability and wind velocity between the two easterly air 
streams. 
5. The equatorial front is single and continuous [14]. 
6. The equatorial front is double [28, 72]. 
7. The equatorial front is part single, part double [32]. 
8. The equatorial front is discontmuous. There are 
many equatorial fronts [54]. 
9. The equatorial front may slope toward the summer 
hemisphere [13]. 
10. The equatorial front may slope one way in the 
lower atmosphere and the opposite way in the upper 
atmosphere [34]. 
11. The equatorial front does not slope at all [46]. 
12. The equatorial front moves discontinuously. It 
may jump from one position to another without passing 
through intermediate points [29]. 
13. The equatorial front may move by disappearing 
in one position while another forms in a different 
position [17]. 
The list of suggested modifications to the frontal 
