TROPICAL METEOROLOGY 
different theory for the origin of the waves he describes; 
so far no one has attempted to reconcile the two 
dynamic explanations. 
The confusion that results from attempts to explain 
the dynamics of the perturbations in the tropics is 
probably temporary; at all events it is negligible com- 
pared with the greater confusion of ideas concerning 
the general circulation in low latitudes. Here the work- 
ers have followed Rossby with implicit confidence, and 
it is in his writings that we find the most complete 
expression of their views. We shall therefore examine 
in some detail those parts of his papers that relate to 
the tropics. 
In 1941 Rossby published ‘The Scientific Basis of 
Modern Meteorology”’ [56], summarizing his views on 
the dynamics of the general circulation; presumably he 
held the same views in 1945, since the same paper was 
republished in the Handbook of Meteorology [57). In 
this paper Rossby gives a masterly exposition of the 
climatological explanation of the tropical circulation. 
It is in fact Hadley’s explanation in modern dress. 
First, we have the direct circulation cell, on a non- 
rotating earth. As the result of the surplus of coming 
over outgoing radiation in the lowest layers of the 
atmosphere near the equator, the temperature in very 
low latitudes tends to increase, with consequent expan- 
sion of equatorial air columns; the contrasting radiative 
regime in higher latitudes produces a tendency to con- 
traction of the vertical air columns. ‘“Thus,” says 
Rossby, “‘at a fixed level of, say, 5 km (3 miles) above 
sea level, a greater portion of the total atmospheric air 
column would be found overhead near the equator than 
near the poles.”’ A direct circulation converting poten- 
tial into kinetic energy would result; the air motion 
would have a component toward the equator in the 
lower layers and away from the equator aloft. On a 
rotating earth, zonal components would be introduced. 
The motion, the same on every longitude, would be 
subject to the conservation of angular momentum. 
“A ring of air extending around the earth at the equator 
at rest relative to the earth, spins around the polar 
axis with a speed equal to that of the earth itself at the 
equator. If somehow this ring is pushed northward over 
the surface of the earth, its radius is correspondingly 
reduced; and it follows from the principle set forth 
that the absolute speed of the ring from west to east 
increases.” Here we have the origin of the antitrades. 
The trades, in like manner, acquire their easterly com- 
ponent as a result of the conservation of angular mo- 
mentum. The final result is that though east winds 
prevail between 30°N and 30°S in the lower atmosphere, 
“Above 4 to 5 km (214 to 3 miles), westerly winds pre- 
vail in all latitudes.” Thus the observed features of the 
mean circulation in the tropics, after the abstraction of 
diurnal, seasonal, and orographical perturbations, is 
completely explained. 
In 1947 Rossby completely rejected this model of the 
tropical circulation, for the following reasons [58]: 
1. He quotes Vuorela [71], Kuhlbrodt [39], and the 
wartime data from the Marianas and Marshalls as 
showing that the easterlies near the equator extend 
farther upward and poleward than would be compatible 
871 
with the direct-cell theory. He says, ‘‘...it is fairly 
clear that the simple thermal circulation outlined by 
Hadley cannot be the sole, or even the dominating, 
feature in the mechanism of the trade winds.” 
2. He then goes on to point out the great uniformity 
of the meteorological elements, especially temperature, 
in the tropics, and the difficulty of reconciling these 
observations with the simple thermal explanation. 
3. He next mentions the “discovery”? of Fletcher 
[28] that in some longitudes there is a mean west, not 
east, wind near the equator, and this in oceanic regions. 
Defant also is mentioned in this connection. Further, 
the well-known zone of divergence in the equatorial 
central Pacific is cited in support of the thesis that the 
direct-cell theory is inadequate. 
4. “The remarkably small annual march of the dol- 
drum belt’’ seems to Rossby to contradict the direct- 
cell theory. 
Each of these points has long been known to tropical 
meteorologists. For example, both van Bemmelen [70] 
and Kuhlbrodt [39] have emphasized point (1) and it 
was thoroughly established by 1939 im the Pacific as a 
result of pilot-balloon observations; point (2) is as old 
as the earliest climatological maps (see also [61]); point 
(8) has been discussed by Deppermann [23]; and, 
further, the explanation of persistent west winds at the 
equator was first advanced by Piddington and Reid 
[52] one hundred years ago. The dry zone in the central 
Pacific has been discussed by Schott [62] and the New 
Zealand meteorologists, particularly Seelye. All these 
points, and others I have already mentioned, long ago 
led many tropical meteorologists to reject the extreme 
climatological view and some to embrace the frontal 
and air-mass theory as applied to the tropics. 
After summing up these objections, Rossby gener- 
alizes from them, implying that because the direct-cell 
theory fails to apply over certain longitudes it fails over 
all. But this is to commit the typical fallacy of the cli- 
matological school: what I may call fallacious gener- 
alization from a given longitude. It arises when we 
become acquainted with new observations restricted to 
a narrow longitudinal band. Thus, when the older 
climatologists first became acquainted with the very 
persistent trade winds typical of the eastern border of 
the subtropical anticyclones, they generalized to form 
a model in which, along every longitude, these winds 
and the associated antitrades prevailed. We have re- 
cently become more aware of the mean state at the 
western ends of the subtropical highs, where “double 
equatorial fronts” and west winds on the equator are 
common; there is therefore a tendency to generalize 
from these longitudes to the whole tropical belt. The 
same remarks apply to the temperature distribution: 
at the eastern ends of the anticyclones, horizontal 
temperature gradients may be found both on the syn- 
optic and on the mean maps, aloft and at the surface 
[18, 61]. The older climatologists could therefore justify 
their model of the trades by appealing to observations 
from these longitudes. However, if we concentrate our 
attention on the western parts of the tropical oceans, 
particularly of the Pacific, and generalize to all longi- 
