6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 94 



quired. These are made at our stations several times on each observ- 

 ing day by means of the spectrobolometer. This instrument, shown 

 diagrammatically in figure 3, is explained in the Annals. Plate 3 

 shows a group of successive solar spectrobolometric observations made 

 at Montezuma, Chile, July 7, 1924. The relative losses of radiation 

 suffered at different wave lengths in transmission through the spec- 

 trobolometer are measured and allowed for as described in the Annals, 

 volume 2, pages 50-52, and volume 3, pages 27-29. 



Knowing the sun's altitude, and thereby the length of path of the 

 sun rays in the atmosphere compared to the length of a vertical path 

 therein, taken as unity, these several curves may fix the atmospheric 

 transmission coefficients at all wave lengths. Thereby the spectral 

 energy curves can be reduced in form and height to what they would 

 have been if observed outside the atmosphere. This reduction is ex- 

 plained in the Annals, volume 2, page 56, and volume 3, page 28. The 

 total area included under such a spectral energy curve is proportional 

 to the total energy of the solar beam as it would be observed with the 

 pyrheliometer. Hence, the ratio of areas included under two spectral 

 energy curves, one computed as of outside the atmosphere, and the 

 other observed as at the earth's surface, is the factor by which the 

 pyrheliometer measurement is to be multiplied to yield the intensity 

 of the sun's radiant energy outside the atmosphere. Including also, 

 as a factor, the square of the ratio of the earth's actual solar dis- 

 tance to its mean value, we arrive at the " solar constant of radiation." 



In the year 1919 it was discovered that a mere measurement of 

 the brightness of the sky surrounding the sun could be made to yield 

 closely enough the coefficients of atmospheric transmission at all wave 

 lengths. This measurement is made with the instrument called the 

 pyranometer. It thus becomes possible to make five solar-constant 

 determinations in one morning and reduce them within the time 

 formerly occupied with one determination. The method as now de- 

 veloped is explained in the Annals, volume 5, pages 1 10-120. 



3. The Variation of the Sun's Radiation 



Figure 4 shows superposed in the form of lO-day means the 

 solar-constant results obtained at Montezuma, Table Mountain, and 

 Mount Brukkaros from 1925 to 1930. The order of excellence of 

 the stations is the order just given. This is indeed plain from the 

 relative smoothness of the three curves of figure 4. But though diflfer- 

 ing in details, the three stations agree in showing in common certain 

 principal trends, and thereby indicate a real variation of the sun. 



II 



