878 
twenty years ago, in spite of much work [44, 55, 66]. 
All forecasting methods so far suggested suffer from 
the defect that exceptions to the deduced rules occur 
too frequently in practice. Some methods suffer from 
the defect that they merely shift the forecasting prob- 
lem from one field to another and often more difficult 
field. Thus all methods based upon isotherm steering 
suffer from the fact that it is frequently more difficult 
to forecast the isothermal field at upper levels than to 
forecast the track of the storm itself. 
Above all we need an extensive investigation of the 
necessary and sufficient conditions for the formation of 
rain in the tropics. In spite of published accounts of the 
precipitation of cumulus clouds whose tops are well 
below the zero isotherm, and in spite of the use of this 
fact by all forecasters operating in oceanic tropical 
regions, one is constantly surprised by the scepticism 
concerning it among high-latitude meteorologists. The 
first worker to publish a complete atlas of precipitating 
tropical clouds will be performing a great service to 
meteorology, for though he may solve no problems he 
will at least draw the attention of the scientific world 
to the fact that the Bergeron-Findeisen theory of pre- 
cipitation gives a very incomplete account of the causes 
of colloidal instability in clouds. In that atlas, I hope, 
will appear good pictures of the complete precipitation 
of a cumulus, 7.e., precipitation not only from the bot- 
tom but also from the sides and overhanging top of the 
cumulus—the almost explosive upsetting of the col- 
loidal stability throughout the cloud mass. Deppermann 
has already published photographs of such clouds, 
which he calls “melting” cumulus—a very descriptive 
term [22]—but apparently he did not realize the great 
theoretical importance of the phenomenon he was de- 
picting. The problem of precipitation has its synoptic 
aspects also. The present observations, as transmitted 
in the international code, give us very little clue to the 
type of rain that is falling, whether it be from an upper 
sheet of altostratus, unconnected with cumulonimbus, 
from altostratus connected with very distant cumulo- 
nimbus pillars, from cumulonimbus, from cumulus of 
great vertical extent, or from comparatively shallow 
cumulus. The result often enough is that the forecaster 
cannot tell whether the rain is connected with an ex- 
tensive region of convergence in the lower atmosphere 
or whether it is to be attributed to the colloidal in- 
stability of cumulus clouds in a region which is in total 
divergent and subsiding. 
Connected with this problem is that of the lapse rates 
of temperature and humidity within tropical clouds. It 
has been claimed that, level for level, tropical clouds 
are colder than their environment. A recent paper by 
Barrett and Riehl [4] has, however, pointed out the 
possible origin of such claims and has rightly empha- 
sized the observational difficulties. More work, how- 
ever, needs to be done. On the theoretical side, the 
recent work on entrainment [68] of air in tropical clouds 
seems to be capable of leading to the solution of an 
old problem of tropical forecasters—the problem of 
what to do with the radiosonde observations. Many 
soundings in the tropics seem to be completely unin- 
TROPICAL METEOROLOGY 
formative. The lapse rate approaches very closely to 
the moist-adiabatic over a deep layer, and this can be 
clearly correlated with existing weather. However, with- 
out noticeable change m the weather one may next 
receive a sounding that differs from the first in the 
lapse rate both of temperature and of humidity, with- 
out apparent reason. It might be suggested that these 
vagaries are connected with the passage of the ap- 
paratus through different parts of cumulus and cumulo- 
nimbus clouds and that the soundings reflect the 
microstructure of the tropical air, not the broad-scale 
features we wish to study synoptically. Theoretical 
knowledge of the factors affecting the lapse rates in 
clouds derived from entrainment studies, together with 
information on the path of the sonde with respect to 
existing cloud masses, might do much to remove the 
perplexity that the present soundings sometimes cause. 
Conclusion. One would like to conclude as a result 
of this survey that tropical meteorology, the most 
ancient branch of scientific meteorology, has shown 
more progress in the last ten years than in the previous 
fifty. However that may be, it is certain that it now 
offers a rich field for research to the climatologist, syn- 
optician, and dynamic meteorologist alike. To the cli- 
matologist it offers the challenging task of collecting, 
reducing, and studying the unique wartime tropical 
data before they become lost forever; to the synoptician 
it offers the prospect of new empirical discoveries in a 
field from which the major errors have just been cleared 
away—discoveries, moreover, that cannot fail to react 
on the theory of fronts and air masses in high latitudes; 
to the dynamic meteorologists it presents a series of 
problems in perturbation theory, applicable to an at- 
mospheric region which, more closely than any other 
on this earth, resembles his ideal: the homogeneous, 
horizontally moving, frictionless, and incompressible 
fluid. 
REFERENCES 
1. Auprrt, L., “The Intertropical Convergence Zone of the 
Eastern Pacific Region.’’ Bull. Amer. meteor. Soc., 26: 
426-432 (1945). 
2. ARAKAWA, H., ‘“The Formation of Hurricanes in the South 
Pacific and the Outbreaks of Cold Air from the North 
Polar Regions.” J. meteor. Soc. Japan, Ser. 2, Vol. 18, 
No. 1, pp. 1-6 (1940). 
3. Barnert, M. A. F., The Cyclonic Storms in Northern New 
Zealand on the 2nd February and the 26th March, 1936. 
New Zealand Meteor. Off. Note No. 22, 34 pp., Welling- 
ton, 1936. 
4. Barret, HE. W., and Rieu, H., “Experimental Verifica- 
tion of Entrainment of Air into Cumulus.”’ J. Meteor., 
5: 304-307 (1948). 
5. Bererron, T., “‘Richtlinien einer dynamischen Klima- 
tologie.”’ Meteor. Z., 47: 246-262 (1930). 
6. Bserxnes, J., “La circulation atmosphérique dans les 
latitudes sous-tropicales.’’ Scientia, Sér. phys.-math., 
Paris (1935). 
7. —— “Theorie der aussertropischen Zyklonenbildung.” 
Meteor. Z., 54: 462-466 (1987). 
8. —— and Hormsog, J., ‘“‘On the Theory of Cyclones.” ./. 
Meteor., 1: 1-22 (1944). 
9. Bsrrxnes, V., and others, Dynamic Meteorology and Hy- 
