860 
writings that are hard to place unequivocally in one of 
the three groups. These things understood; however, we 
will find it easy enough to survey the present state of 
tropical meteorology. It will consist of the conflicts of 
these schools as revealed in the literature. The scope 
and detail of the conflicts is likely to astonish the 
meteorologist working on high-latitude problems, ac- 
customed as he is to a large measure of agreement on 
the fundamental descriptions of temperate and high- 
latitude weather. In the tropics there is not even agree- 
ment on the common forms of tropical clouds or on 
the meteorological conditions accompanying precipita- 
tion. It has become the custom to generalize in haste 
and to reject inconvenient observations at leisure. Large 
areas of the tropical atmosphere have not been explored 
by observation and even the better-known areas have 
not yet yielded statistics of sufficient scope or reliability 
to justify the tropical parts of the ‘‘models” which are 
incorporated into the textbook descriptions of the gen- 
eral circulation of the atmosphere. In synoptic meteor- 
ology, conditions are little better than in climatology. 
The role of water substance in the genesis of tropical 
depressions is a matter of dispute, the existence or non- 
existence of fronts the subject of irreconcilable specu- 
lations. A dynamical meteorology of the tropics can 
hardly be said to exist. The assumptions of such work 
as has been done are usually overlooked, and the results 
applied in a manner quite beyond the expectation of 
the original authors. Worse than this, however, is the 
tendency to apply dynamic theories depending for their 
validity on assumptions that hold in high latitudes but 
not in low latitudes. For example, the extent to which 
reasoning resting implicitly on the geostrophic-wind 
equation has been applied in equatorial meteorology is 
disquieting. This state of affairs, however, should not 
give occasion for pessimism; it is a sign that tropical 
meteorology, the oldest branch of scientific weather 
study, is undergoing a new development and that that 
development is a vigorous one. Each school of thought 
has had its achievements, has clarified some previously 
obscure field of knowledge, and has had its defects 
subject to sharp criticism. It is our task now to estimate 
the extent of these achievements and to mitigate the 
evil of those defects. 
THE CLIMATOLOGICAL METHOD 
There is no doubt that, at least over the oceans, the 
observed values of the meteorological elements deviate 
from their monthly, seasonal, and annual means to a far 
lesser degree in the tropics than in higher latitudes. The 
mean values therefore can be said, in a special sense, to 
be more representative of the synoptic values. The 
steadiness of the wind, for example, has long been 
known, and the representativeness, in the above sense, 
of the mean trades and mean monsoons is perhaps the 
_most ancient fact of modern meteorology, antedating in 
its origin the very notion of a synoptic map. 
The word ‘‘trade”’ itself is a contraction of “to blow 
trade,” that is, to blow constantly in the same direction, 
along the same track. The empirical discovery implicit 
in the expression has, of course, been well confirmed 
TROPICAL METEOROLOGY 
by scientific investigation. Gallé,! for example, has given 
an analysis of the persistence of the wind in various 
parts of the Indian Ocean. The persistence may be 
defined as the ratio of the mean vectorial wind speed 
to the mean speed without regard to direction. If, over 
the period covered by the means, the wind has the 
same direction, the persistence would be 100 per cent; 
on the other hand, if the directions were distributed at 
random, the persistence would be zero. According to 
Gallé, the persistence of the trade in the square 
10°S—20°S, 80°E-90°E does not fall below 69 per cent 
and for most months lies between 80 per cent and 90 
per cent. This can be contrasted with the corresponding 
figures for the square 35°S-45°S, 70°E-80°H, where the 
persistence over any month does not rise above 57 per 
cent and, in the mean, amounts to 45 per cent for the 
year. But Gallé’s figures merely give quantitative ex- 
pression to the conclusion which anyone can draw from 
the surface-wind maps that are published from time to 
time by national meteorological organizations [69]. Scru- 
tiny of the wind data on these maps confirms not only 
the great persistence of the trades but also a similar 
large persistence, within the seasons of their greatest 
development, of the monsoons. To a lesser degree the 
maps also give one the impression of uniformity: over 
a very large proportion of the tropical oceans, the pre- 
dominant component of the wind is the east component 
and this, combined with the known persistence of the 
trades, leads even careful workers to fall into the habit 
of regarding all tropical west winds occurring outside 
the monsoon regions as anomalous. 
At first sight the representativeness of the means of 
meteorological elements other than the wind appears 
not to be so striking. Temperature and pressure, for 
example, show synoptic variations that may depart 
widely from the monthly, seasonal, or yearly means. 
However, the periodic nature of these variations is 
evident, the pressure showing semidiurnal oscillations 
whose amplitudes are largest at or near the equator, 
while the temperature has a clear diurnal periodicity 
whose amplitude becomes less the more closely the 
conditions of observation approach those typical of the 
open sea. When allowance is made for the periodicity 
and for calculable orographical effects by correction of 
the synoptic observations, the mean values for a month 
or a season turn out to be as highly representative of 
the corrected synoptic values for temperature and 
pressure as for wind. Further, in the case of tempera- 
ture, remarkable uniformity is disclosed. Over vast 
stretches of the western Pacific, forinstance, the tempera- 
ture of the surface air during the whole “wet season” 
varies little from 85F [61]. Even in monsoon regions, 
where we might expect to find large seasonal variations 
of temperature, remarkable instances of uniformity are 
known. 
At Colombo the mean difference [sic] in surface tempera- 
ture between the North-East and South-West monsoon 
seasons is less than 2°F. and just about as much rain falls ~ 
1. Quoted in [86, p. 45]. The original paper is not available 
to me. 
