SOLAR RADIATION ABBOT 181 



Since the formula holds strictly only for rays of homogeneous 

 Avave length, we require spectrum measurements. If from a series 

 of holographs of the solar spectrum, in which the partial transmissi- 

 bility of the optical instruments has been allowed for, we determine 

 atmospheric transmission coefficients for about 40 selected wave 

 lengths between 0.34 and 2.5 microns, we can compute the intensity 

 which each of these rays would represent if observed outside the 

 atmosphere. This determines the sun's spectrum energy curve as it 

 would be observed in free space. If we compute the area included 

 under the curve thus determined, and divide it by the area included 

 by the curve observed by the spectrobolometer, the quotient is the 

 factor by which we must multiply the total intensity of the solar 

 beam as measured by the pyrheliometer to give the intensity which 

 the pyrheliometer would have read if in free space. Correcting the 

 result to mean solar distance, we have the solar constant of radiation. 



TRANSMISSIBILITY OF OPTICAL APPARATUS 



As required for solar-constant determinations, we have made many 

 measurements of the relative transmissibility of our optical instru- 

 ments for the different spectral rays. Our procedure involves two 

 spectroscopes, of which the auxiliary one delivers its spectrum upon 

 the slit of the main one used for the holographic work. Under these 

 circumstances, we measure with the bolometer the intensity of many 

 spectral rays both before and after they traverse the main spec- 

 troscope. Their relative transmissibilitj^ appears in the ratios of 

 these measurements. 



WAVE-LENGTH DISTRIBUTION OF SOLAR RADIATION, AND THE SUN'S 

 EFFECTIVE TEMPERATURE 



Such determinations of transmission in the optical train, together 

 with the determinations of transmission in the atmosphere, enabled 

 us to represent and tabulate the distribution of energy in the solar 

 spectrum as it is outside our atmosphere. Our best results in this 

 line were published in 1923 in a paper entitled " The distribution of 

 energy in the spectra of the sun and stars." ^ They have been of use 

 to other investigators for various purposes. 



If we assume that the sun is approximately a perfect radiator, our 

 work on the solar constant yields three methods of estimating his 

 effective temperatures : First, from the spectral position of maximum 

 intensity. Second, from the general form of the curve of distribu- 

 tion of energy in the spectrum. Third, from the sun's distance and 

 diameter combined with the value of the solar constant of radiation. 



1 Smithsonian Misc. Coll., vol. 74, No. 7. 



