56 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



radiation goes out in all directions, the influence of a single cloud 

 will be nrore continuous than is the case for the solar radiation. As 

 soon as the cloud comes over the horizon it will begin to affect the 

 radiation to the sky, its influence growing as it approaches the zenith. 

 This will be rendered clearer, and details will be afforded, by the 

 observations on the radiation to different parts of the sky, given in 

 a later chapter. 



It is evident that, when the sky is cloudy, we can distinguish be- 

 tween three radiation sources f of the atmospheric radiation : First, 

 the radiation from the parts of the atmosphere below the clouds ; 

 secondly, the part of the radiation from the clouds themselves, which 

 is able to pass through the inferior layer, and, in the third place, the 

 radiation from the layers above the clouds, of which probably, for 

 an entirely overcast sky, only a very small fraction is able to penetrate 

 the cloud-sheet and the lower atmosphere. 



Some measurements were taken in the case of an entirely overcast 

 sky. Figure 7 shows two curves drawn from observations at Clare- 

 mont. In the beginning the sky was perfectly clear, at the end it was 

 entirely covered by a low, dense cloud-sheet : cumulus or strato- 

 cumulus. 



In general the following classification seems to be supported by the 

 observations : 



Average radiation 



Clear sky 0.14-0.20 



Sky entirely overcast by : 



Cirrus, cirrostratus and stratus 0.08-0.16 



Alto-cumulus and alto-stratus 0.04-0.08 



Cumulus and strato-cumulus 0.01-0.04 



Especially in the northern winter climate, the sky is very often over- 

 cast by more or less dense sheets of stratus clouds. They are very 

 often not dense enough to prevent the brighter stars being very easily 

 seen through them, and especially in the night it is therefore often 

 difficult to tell whether the sky is perfectly clear or not. Dr. Kennard 

 proposed to me that one should use the visibility of the stars (1st, 2d, 

 3d, and 4th magnitude, etc.) to define the sky, when it seemed to be 

 overcast or very hazy. This may be of advantage, especially when 

 observations are taken in the winter time or extended to hazy condi- 

 tions. 



