276 
Proceedings of Royal Society of Edinburgh. [sess. 
seemed quite sufficient to practically represent the facts, a 0 denot- 
ing the sun’s energy when the sky is clear, n the amount of cloud 
on the scale 0 to 10. Theoretically, we should perhaps expect an 
exponential function of the form 
- fin 
a n = a 0 e , 
and it is not at all impossible that this formula may correctly repre- 
sent the facts considering the conditions of heat. But, assuming 
the amount of water in the soil to he in some way proportional to 
the cloudiness, we have certainly to deal with a change in the 
specific heat of the earth’s crust, by which obviously the logarith- 
mic curve, applied to temperatures , must be bent more and more 
towards a straight line. 
It was found convenient to derive the value of the solar constant 
for a mean state of cloudiness equal to 5‘0, and the mean latitude 
of the places considered 50° north. So, if we call a m the solar con- 
stant under these conditions, a n the numerical coefficient derived 
from the daily curve of a certain place 
a n — a n cos cf) cos 8 , 
we have the equation 
* sec <£ 1-5/? 
a m -a n sec8—^ 
n being the mean state of cloudiness at the time considered. 
Now, from the above quoted work by H. Wild were taken 
20 different stations in various latitudes, from which were formed 
three normal places with the respective latitudes 59° *4, 51° *3, and 
42° -9. For every month from April to September a m was derived 
from that part of the hourly curve between sunrise and sunset, the 
result being the following 
If we combine the three places in every month separately, we 
obtain : 
April. May. June. July. Aug. September. 
a m = 6°-23 6°*23 6°-27 6°*31 6’43 6° -23 Centigrade. 
