244 ABBOT— SOLAR CONSTANT OF RADIATION. [April 21. 



ception. We find it to be the rule, hence we must exclude accidental 

 errors as the main source of the apparent variability of the sun. 

 As for the other explanation suggested, we find no material differ- 

 ence in the result derived for the solar constant on a good day 

 whether we observe at sea-level, at one mile, or at nearly three 

 miles elevation, though the pyrheliometer readings on the ground 

 dififer by 25 per cent, between Washington and ]\Iount Whitney. 

 Hence we ma}- reasonably conclude that we do, in fact, correctly 

 estimate the loss which occurs in the atmosphere. The fluctuation 

 in the solar-constant results therefore seems to indicate either a true 

 variability of the sun, or else the interposition of meteoric dust, or 

 other cosmic hindrance to the passage of radiation from the sun to 

 the earth. These fluctuations, while not of regular periodicity, 

 generally run their courses within five or ten days.^*' 



It is now proposed to test this conclusion by conducting solar-con- 

 stant measurements simultaneously at Mount Wilson and in southern 

 Mexico. If the results of a long series of daily observations at 

 these remote stations should agree, it would seem quite unlikely 

 that any apparently simultaneous fluctuations of the solar constant 

 of radiation could be attributed to terrestrial influences. 



SUMM.'XRY. 



Special apparatus, including the silver-disk secondary pyrhelio- 

 meter, the absolute water-flow p}rheliometer and the recording 

 spectro-bolometer, has been employed by the writer and his col- 

 leagues at Washington and Mount Wilson and Mount Whitney, to 

 determine the mean value of the solar constant of radiation, and 

 its possible fluctuations. 



The observations, exceeding 400 in number, have been made in 

 all the years since 1902 to 1910. but most plentifully and accurately 

 in 1908, 1909 and 19 10. The mean value of the intensity of solar 

 radiation outside the atmosphere, at mean solar distance, is found 

 to be 1.922 (i5°C.) calories per square centimeter per minute, but 

 might prove i or 2 per cent, higher in years of less sun-spot activity. 

 The solar-constant values do not appear to depend on the altitude of 



"See Abbot and Fowle, Astropliysical Journal, April, 191 1. 



