1148 
of the thesis. An analysis by Brunt [14] of surface dust 
counts for one year taken over a distance interval of 
ten miles downwind from the center of Norwich, Hng- 
land, shows that the pollution is approximately inversely 
proportional to the distance from the center of the 
city and that it diminishes with increasing wind speed. 
Suggestions as to research on various aspects of the 
problem of pollution by cities have been made [24, pp. 
139-145]. Studies of the following are suggested: the 
three-dimensional distribution of pollution under vari- 
ous meteorological conditions; the distribution and char- 
acteristics of country pollution, including the heights 
reached by it; the processes of natural removal of pollu- 
tion; atmospheric turbulence in its association with 
pollution; pollution in city parks, where there is no 
emission of pollution, for the detailed information to be 
gained; and special aspects of the problem which could 
readily be investigated by local authorities without the 
services of a highly trained staff. 
The theoretical aspects of pollution by multiple 
sources have as yet received little attention. The prob- 
lem to be solved is essentially that of the pollution 
caused by a large number of point sources of a widely 
varying nature. A start has been made by Sherwood 
[75], who applied Sutton’s concepts to study theoreti- 
cally the screening produced by the smoke emitted 
from a number of equidistant continuous point sources 
arranged along a line lying across wind. The problem 
presented by an industrial city is of course extremely 
complex: the point sources are irregularly located over 
an area, at varying heights, are emitting a variety of 
contaminants at differing rates, which also vary with 
time, into an atmosphere whose diffusing properties 
vary with both periodic and aperiodic components. 
Such a statement may merely induce deep pessimism, 
but it also suggests that an attempt to handle the 
problem of pollution from such an array of poimt 
sources by statistical combinations of individual point 
sources may prove fruitful. It is obvious that a complete 
theoretical treatment cannot be expected in the near 
future, but a start could be made by singling out one of 
the numerous variables and determining, if possible, 
the patterns of pollution resulting from various statis- 
tical representations of the chosen variable. Some of 
the other variables might then be handled by other 
appropriate statistical simplifications, leading gradually 
to a theory of increasing scope and validity. The meas- 
urement of the distribution of tracer substances emitted 
according to plan from a number of suitably chosen 
sources would be invaluable for checking theoretical 
distributions. 
SECONDARY EFFECTS ON METEOROLOGICAL 
ELEMENTS 
The possibility that atmospheric pollution has an 
effect on some of the meteorological variables and on 
the microfeatures of weather must be considered. For 
example, the question of the influence of a pall of smoke 
over a city under stagnant atmospheric conditions on 
both solar and terrestrial radiation is an interesting one 
ATMOSPHERIC POLLUTION 
which merits attention. It is possible that under such 
conditions differential heating or cooling between the 
polluted city air and the cleaner air over surrounding 
country districts leads to local convergence or diver- 
gence which in itself may act so as to increase con- 
centrations over certain areas and at certain heights 
while causing decreased concentrations elsewhere. 
Rainfall. Climatological studies of rainfall at Roch- 
dale, England, indicate that both the average amount 
and average rate of rainfall are slightly smaller on 
Sunday than on other days of the week [2, 3]. Since the 
factories were closed on Sunday, it is inferred that 
condensation on the nuclei produced by industrial proc- 
esses leads to the increased rainfall on the other days 
of the week. It has been shown that the Ruhr region of 
Germany gets measurable rain or drizzle on twenty 
more days per annum than do nearby less industrialized 
areas [103]. 
Fog. Measurements of relative humidity in fog often 
give values substantially less than 100 per cent, partic- 
ularly in industrial areas. From this we may infer that 
atmospheric pollution increases the incidence and dura- 
tion of fogs. There is also direct evidence: a long series 
of observations in Prague, Czechoslovakia, show that 
there were, on the average, about eighty-two days with 
fog per year in the period 1800-1880, and that since 
that period the number has nearly doubled [41]. The 
observations were made continuously at the same point 
and by the same rules. The effect of increasing in- 
dustrialization and its attendant pollution is thus 
marked. It has been suggested [43] that sulfur dioxide 
may contribute directly to fog formation under certain 
conditions when photochemical oxidization by ultra- 
violet light is sufficient to produce sulfur trioxide, 
which is highly hygroscopic. On the other hand, there 
are indications [62] that such combustion nuclei alone 
cannot be responsible for the difference between relative 
humidities in fog in industrial and rural districts. The 
possibility that radiational cooling of contaminating 
particles, and thus of the ambient air, is a significant 
factor has been considered [49]. Other phases of the 
behavior of fog in a polluted atmosphere have been 
discussed [20]: in industrial cities it has been observed 
that a dirty fog at, say, 9 p.m. changes to a clean one 
by about 5 a.m. next morning, without a clearing of the 
fog, suggesting that the dirty particles fall out and are 
replaced by clean ones; fog droplets in town air dissolve 
the sulfur dioxide present, forming sulfurous acid which 
will oxidize to sulfuric acid which in turn acts to retard 
the clearing of such a fog. In a laboratory study of 
artificial fogs [65] it was found that, as pollution in- 
creases, a corresponding increase of fog density occurs 
only until the pollution attams a value characteristic 
of a small town. On the other hand, there is a con- 
tinuous increase in fog duration with increasing air 
pollution. Further laboratory investigations [64] have 
shown the relative effectiveness of various combustion 
products and of the quality of the combustion process 
in prolonging the fog and in increasing its density. 
Incomplete combustion tends to increase both fog den- 
