804 
>rR. LOUIS VESSOT KING ON THE SCATTERING AND 
as an nukiioAvn and variable quantity. Diagram 1. shows that for Mount Wilson and for 
Mount Whitney the effect of “ dust ” is small, i.e., we may take N' = 0 for x greater 
than 1780 metres. In this case (77) reduces t<j (70) and the attenuation may be 
taken to be due almost entirely to the effect of air-molecules, wOiile the existence 
of a small term, y, indicates that even in the comparatively dust-free air above Mount 
Whitney there is a small amount of attenuation by absorption, i.e., a direct conversion 
of solar radiation int(.) thermal agitation of atmospheric molecules. 
We notice from Table II. that for the comparatively dust-free air above the 
level of Mount W^ilson the value of y under standard conditions of pressure is 
y = a„ H = '032, or, since H = 7'988 x 10® cm. at 0° C., the value of is 
ao = 4'0 X 10 cm. 
(78) 
On referring to (8) we notice that a,, is tlie fraction of radiant energy converted per 
centimetre of path into thermal molecular agitation. This fraction is greatly increased 
by the presence of small solid particles such as “dust,” &c. From (9) we can estimate 
the rate of increase of temperature in a gas under atmospheric pressure due to solar 
radiation passing through it. 
a„E 
W 
, do 
- e liave —r- 
dt 
—from equation (9). 
Taking the value of a,, for dust-free air from (78) and writing E = 1'92 calories per 
minute, /7y = '001293, s = '237, we find 
= 2'5 X 10 ^ degrees C. 
dt ^ 
ner minute 
(79) 
In the case of ordinary air at sea-level, the Wasliington ol)servations from Table II. 
show that the rate of increase of temperature calculated in (79) must ])e increased to 
about six times this value. 
With regard to sea-level stations. Diagram I. seems *to indicate, both for the 
Washington and Potsdam observations, a marked change in the nature of the 
absorption due to “dust” in the neighbourhood of'010//. We therefore discuss 
separately the case of long-wmve radiation (\>'010 /x) and short-wave radiation 
(X<'010//). 
(1) Long-wave lladiation (X>'610//). 
For long waves the straight lines of Diagram I. show that the term (x) of (77) 
must be of tlie form 
(X) = /3"X-Hy", 
where fi" and y" are constants for the range of wave-lengths greater than 'Oltfo. This 
