ABBOT, FOWLE, ALDRICH : VARIATION OF THE SUN 313 



line if the sun is variable, but will fall uniformily distributed about 

 one point on that line if the sun's radiation is constant. There 

 is no difficulty in deciding that the line and not some single point 

 of the line best represents the results here given. 



The variation of the sun shown between the extreme obser- 

 vations amounts to 11 per cent, and many observations unite in 



<n 



u 



5' 

 o 



_i 



<, 

 o 



/i 



D 

 _l 



z 



o 



z 



D 

 O 



z 



6AS50UR VALUES CALORIES 



showing a variation of 7 per cent. The average deviation of the 

 separate determinations at Bassour from those of the same days 

 at Mount Wilson is 1.6 per cent. 



Hence the average deviation of a single day of solar-constant 



(-1 /^ \ 



--=-=) 1.1 per cent, and the 



probable error of a single solar-constant measurement at one 



