6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 10/ 



gen for air, and adjusting the hydrogen pressure for largest deflection 

 at a time of single swing of 12 seconds; 2, by excluding much more 

 of the light from the reading-beam lamp ; 3, by having means to set 

 the focus of the image-forming lens from a distance, vi^ithout approach- 

 ing personally near the radiometer (we found indications of a slight 

 temperature effect from this cause in 1947). We intend also to choose 

 other wave lengths, so as to cover the spectrum of blue and white 

 stars in the ultraviolet, and of yellow and red stars in the infrared ; 

 4, by taking twice as many readings at each wave length; 5, by ob- 

 serving the same star on several more nights. 



We now give as a sample the readings on the star Arcturus at 

 wave length 4710 A. The numbers express the positions on the scale 

 in whole turns and fractions thereof, and the differences caused by 

 shifting the spectrum from one vane of the radiometer to the other. 



Table 2. — Sample readings at wave length 4/10 A. on Arcturus 



Such mean observations were first multiplied by correcting factors 

 for prismatic dispersion, so as to reduce them to a uniform scale of 

 wave lengths. It was then needful to correct the results for atmos- 

 pheric transmission, to reduce them to outside the atmosphere. For 

 this purpose all the solar-constant observations of the month of Sep- 

 tember made at Mount Wilson in the years 1915, 1916, 1917 were 

 considered. From Annals, Astrophysical Observatory, volume 4, table 

 37, the mean values of atmospheric transmission for September in 

 these three years were computed as follows: 



Table 3. — Mean atmospheric transmission 



Wave length 0.35 .40 .45 .50 .60 .70 .80 i.oo 1.20 1.60 



Transmission 587 .7x8 .797 .847 -890 .935 -956 -970 -975 -980 



These values were plotted, and values were interpolated at the wave 

 lengths observed in the stellar spectra. Air masses were computed 

 corresponding to the median time of observation of each wave length 



