1911.] ABBOT— SOLAR CONSTANT OF RADL\TION. 241 



about three hours of observation, under a cloudless and uniformly 

 clear sky. and about three days of computing. 



We began to make solar-constant observations in Washington 

 at the Smithsonian Astrophysical Observatory, in October, 1902, 

 and continued them there whenever a favorable opportunity was 

 presented, until !May, 1907. In all this time we made only 44 

 tolerably satisfactory determinations at W'ashington, for cloudless 

 days were rare, and many days that promised fairly proved disap- 

 pointing, by reason of the appearance of smoke, haze or clouds. 

 Four important results came from the W^ashington observations. 

 First, no apparently good determinations yielded solar-constant 

 values above 2.38 of our then provisionally adopted scale of calories, 

 or 2.25 true calories. Second, the mean value in the true calories 

 from 44 determinations was 1.960. Third, the transmission of the 

 atmosphere was determined on many days, and for many wave- 

 lengths.^^ Fourth, a strong probability was raised by the results of 

 observations of 1903 that the sun is a variable star.^- This variation 

 seemed to reach 10 per cent, in its extreme range, but no tendency 

 towards a regular period was then found for it. A dependent 

 variation in terrestrial temperatures seemed indicated. 



Primarily in order to make spectro-bolometric determinations 

 of the solar constant, suitable to test the supposed variability of the 

 sun, an expedition under my charge went out to ]\Iount Wilson in 

 1905, by invitation of Director Hale of the ]Mount W^ilson Solar 

 Observatory. The site proved excellent for the purpose, on account 

 of its considerable altitude, cloudless sky and freedom from wind. 

 Much aid and comfort was furnished by Director Hale and his 

 staff. The expedition was repeated in 1906, 1908, 1909 and 19 10. 

 We now occupy a cement observing shelter and living quarters there, 

 on ground leased from the Solar Observatory. Our observations 

 have generally occupied the six months, May 15 to November 15, 

 and in the last years we have made practically daily determinations 

 of the solar constant of radiation during this interval. 



" Astronomers have not yet very generally availed themselves of the 

 accurate coefficients of atmospheric transmission obtained in our researches 

 for all parts of the spectrum, and from Washington, Mount Whitney and 

 Mount Wilson. 



'^ See S. P. Langley, Astrophysical Journal, Vol. 19, p. 305, 1904. 



