no. 4 



SOLAR RADIATION ABBOT, FOWLE, AND ALDRICH 



7 



the product of fourth power of wave length by logarithm of trans- 

 mission coefficient should be constant. As shown by one of us, 1 the 

 coefficients of atmospheric transmission obtained on Mt. Wilson 

 depend slightly on the total atmospheric humidity included between 

 Mt. Wilson and the sun. The transmission coefficients may be 

 reduced to dry air conditions by applying a very small correction to 

 them. These corrected coefficients, a , are found to be in close 

 harmony with Lord Rayleigh's theory, as is shown by the following 

 table. The observed values of a are means for September 20 and 

 September 21, 1914: 



Wave lengthAin^ 



Observed trans- 

 mission a 



Corrected trans- 

 mission a 



A 4 log a 



0.3504 

 .610 



.632 

 —30.0 



0.3709 



.671 



.686 

 —3i- 1 



0-3974 



•744 



• 752 

 —30.9 



0.4307 



-31.8 



0.47530.53480.5742 



.851 



.863 

 —32-7 



. 892 . 893 



—38.2 



•905 

 -46.7 



0.68580.7644 



.95o! -969 



.959 -979 

 -40.31— 31-4 



The deviation from a constant ratio in the yellow and red spectrum 

 is doubtless due to the very large number of atmospheric absorption 

 lines in this part of the spectrum. 



By the aid of Lord Rayleigh's theory of the scattering of light, 

 Mr. Fowle has determined from the Mt. Wilson experiments the 

 number of molecules per cubic centimeter of dry air at standard 

 temperature and pressure. He finds the value ( 2.70 ± 0.02) x io 19 , 

 while Millikan obtained, by wholly dissimilar methods, (2.705 ± 

 0.005) x io 19 . 



In the course of our experiments at Mt. Wilson, we found the 

 solar radiation outside the atmosphere variable in short irregular 

 periods of from five to ten days, and to have a variable range of from 

 2 to 10 per cent. That this variability is really solar was confirmed 

 by independent simultaneous observing at Bassour in Algeria and 

 still more recently by as yet unpublished experiments on the distribu- 

 tion of brightness over the sun's disk. This latter method is quite 

 independent of atmospheric disturbances. It seems to us that if our 

 solar constant results were erroneous to the extent that the solar 

 constant is really 3.5 calories instead of 1.93, as some of our critics 

 would persuade us, the probability of finding these real solar varia- 

 tions of from 2 to 10 per cent by simultaneous observing at stations 

 separated by one-third of the circumference of the earth would be 

 very small. We should suppose that if there are atmospheric con- 



1 F. E. Fowle, Astrophysical Journal, 38, 392, 1913 ; 40, 435, 1914. 



