866 
though the explanation of their relations to one another 
and to the field of motion would be different. 
I have labored the distinction between the empirical 
and the theoretical front because it was undoubtedly 
not made by the early workers of the frontal school in 
tropical meteorology. They succeeded in establishing 
the following facts by observation and synoptic 
analysis: 
1. In all tropical latitudes, organized lines of cumulo- 
nimbus cloud, closely resembling the cold-frontal cloud 
systems of high latitudes, are encountered by airmen 
and are observed to pass over fixed observing stations. 
2. In the great majority of cases, these lines of cloud 
are accompanied by zones of rapid wind change. The 
total shift is not as pronounced as in high latitudes and 
is generally more gradual. The shear at the shift is often 
cyclonic. These facts have been emphasized by Depper- 
mann [20], by Harmantas [35], by Frolow [31], and by 
Alpert [1]. From personal experience I can assure the 
reader that they are facts. 
3. Differences of temperature can sometimes be found 
between the air on either side of the cloud line. This is 
reported by Sawyer [60] and by Solot [67]. However, in 
certain oceanic regions, though there is often an abrupt 
drop in temperature at the surface with the passage of 
the cloud line, it may recover its former value when the 
weather clears. 
4. There may be very marked differences in specific 
humidity across the cloud lme—not, it is true, at the 
surface, but m the middle atmosphere. 
5. The lines are frequently but not always associated 
with a low-pressure center or trough at the surface, but 
owing to the weak pressure gradients and the great 
amplitude of the semidiurnal pressure oscillation, it is 
difficult to correlate any sharp change in the barogram 
with the passage of the cloud line, unless a violent thun- 
derstorm occurs at the station. 
6. One can sometimes find these cloud and wind- 
shift lines day after day in the region of lighter winds 
around and outside typhoons, hurricanes, and tropical 
cyclones [21]. Moreover, the storms sometimes appear 
to originate in the regions where such lines have previ- 
ously been discovered. 
To most tropical meteorologists, after 1933, these 
discoveries were sufficient to identify the lines as fronts, 
and since they made no distinction between the empir- 
ical and the theoretical front, they adopted the entire 
high-latitude apparatus of frontal and air-mass analysis 
and forecasting. The way had been prepared, as we 
have seen, by Brooks and Braby, by Bergeron and, most 
authoritatively, by the authors of Physikalische Hydro- 
dynamik [10]. The frontal wave theory of storm forma- 
tion was adopted without question; indeed it seemed 
to be confirmed by (5) and (6) above. A few meteor- 
ologists were concerned by the frequent lack of 
temperature differences across the fronts and were care- 
ful to distinguish “tropical fronts” where the differences 
were very small (in many cases they were, in fact, ab- 
sent) from extratropical fronts where there was usually 
no difficulty in detecting the difference, at least by 
radiosonde. Deppermann in particular was always care- 
TROPICAL METEOROLOGY 
ful to make the distinction. Having decided that fronts 
and air masses existed in the tropics, the frontal analysts 
of the decade 1933-1943 were concerned to describe and 
classify the major discontinuities of the tropics. First, 
and most important, came the equatorial or intertrop- 
ical front, whose synoptic existence was never doubted 
during this period. But there were other major systems, 
particularly in the western Pacific. Thus Deppermann 
distinguished a ‘‘tropical’’ front between the winter 
monsoon of eastern Asia and the trade of the North 
Pacific [20]. Other “major fronts” were described in 
different parts of the tropics [17] but these are now of 
only historical interest. In addition, much attention was 
given to the passage of polar fronts from the belt of the 
westerlies on into tropical latitudes. 
Some extraordinarily detailed frontal anatyses of hur- 
ricanes and tropical storms were produced during the 
thirties. In the South Pacific the full occlusion process 
leading to the development of tropical cyclones was 
illustrated by Kidson and Holmboe [88]. Similar analy- 
ses were published for storms in the Caribbean [87, 
64] and in the Bay of Bengal [59]. An elaborate triple- 
point theory due originally to Scherhag was applied to 
North Pacific storms [20]. If one were to judge by the lit- 
erature up to 1940, one could not doubt that the appli- 
cation of the frontal and air-mass theory to the tropics 
was completely justified. The chief problem seemed to 
be, not to account for the origin and development of trop- 
ical storms, but to explam why many more than were 
actually observed did not develop. The tentative theory 
advanced in Physikalische Hydrodynamik, accounting 
for the maintenance of the equatorial front in terms of 
temperature differences between the trades of the winter 
and the summer hemispheres, was extended to account 
for this also. The equatorial front, on this view, tended 
to become a sharp discontinuity, ‘“‘to become active” in 
the words of the air-mass school, after the air in a great, 
polar outbreak in the winter hemisphere reached the 
equatorial regions. This would communicate an impulse 
of an unspecified kind to the equatorial front which 
would thereupon become unstable and form a tropical 
cyclone. As early as 1935 De Monts had published a 
paper with the significant title “Role catalyseur de 
Vair polaire dans la génése d’un cyclone tropical” [19] 
describing this process in the South Indian Ocean. 
Japanese meteorologists seemed particularly attached 
to this explanation [2]. Later Grimes [33] endeavored 
to specify the nature of the impulse that would be given 
to the equatorial front by the arrival of winter-hemi- 
sphere air after a polar outbreak. He considered that a 
“cyclone only forms when there is a surge across the 
equator of air which in the hemisphere of origin had 
sufficient anticyclonic vorticity, which is conserved as 
cyclonic vorticity in the other hemisphere.” 
Since shearing motion at the frontal surface would be 
the only destabilizing agent leading to waves of increas- 
ing amplitude, Grimes was evidently looking for some 
process which would bring about a sudden increase in 
cyclonic shear at the supposed frontal surface. The 
process was made to depend upon the conservation of 
the vertical component of absolute vorticity, which it is 
