4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I48 



the solar spectrum, with a reading of the pyranometer on a limited 

 sky area surrounding the sun, a reading of the pyrheliometer on the 

 sun itself, and a theodolite measurement of the sun's zenith distance, 

 the observations are complete. 



We made three or sometimes five determinations of the solar con- 

 stant a day by the short method, and used their mean value for the 

 day's result. From 1924 to 1952 over 9,000 mean daily values of the 

 solar constant were obtained. Many of them were observed from 

 three independent stations on the same day. It is shown by table 1, 

 page 13 of Smithsonian Publication 4545,^ that by comparing 1992 

 pairs of solar constant observations of the same day, made individu- 

 ally at stations separated by thousands of miles, some in the Northern, 

 others in the Southern Hemisphere, and at all times of the year, the 

 probable error of a single day's observation of the solar constant at 

 one station is ^ of 1 percent. All of the approximately 9,000 daily 

 determinations, published in volumes 6 and 7 of Annals, were scru- 

 tinized at Washington, and unanimously approved for publication by 

 a committee comprising L. B. Aldrich, W. H. Hoover, and Mrs. 

 A. M. Bond. 



TOWER TELESCOPE 



Langley hoped that these 9,000 daily determinations would prove 

 that a strong correlation could be found between daily solar constant 

 values and simultaneous observations of the distribution of the 

 intensity of solar radiation along the east-west diameter of the sun's 

 disk. Some preliminary observations of that distribution were made 

 in Washington in 1908. Later a tower telescope to form an 8-inch 

 image of the sun was erected on our observatory on Mount Wilson, 

 about 1913. From 1913 to 1920, on all observing days, the telescope 

 also was adjusted between observations to form a solar image on the 

 slit of the spectrobolometer. The telescope clock was then stopped, 

 and the solar image was allowed to drift centrally over the slit. Thus 

 was recorded D -shaped distribution curves in five wavelengths on 

 each day that a solar constant value was obtained. See figure 52, 

 Pub. 4545.2 



Owing to inexactness of solar constant values due to the fluctuation 

 of atmospheric transmission during the several hours required for 

 solar constant measures by Langley's method (used exclusively on 



^ Solar variation and weather, by C. G. Abbot. Smithsonian Miscellaneous 

 Collections, vol. 146, No. 3. 1963. 



