4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 87 



of the sky about the sun, which we named the pyranometer. By its 

 aid we have devised a brief empirical method for estimating the at- 

 mospheric transparency in all wave lengths. We have also devised a 

 spectroscopic method for estimating the quantity of precipitable water 

 held in the form of vapor in the atmosphere. From our determinations 

 of atmospheric transparency we have checked exactly with other 

 methods on the determination of the number of molecules per unit 

 volume. 



As the temperature of the earth and the fundamental factors of 

 climate and weather depend on the intensity of solar radiation, we 

 have made earnest efforts over a long period of years to secure accurate 

 measurements of it. When we began this work in 1903, authorities 

 were in doubt over the entire range as between PouiUet's value of 1.76 

 calories, and Angstrom's value of 4-0 calories for the solar constant of 

 radiation. As a result of our work, carried on at all seasons, at stations 

 ranging from sea-level to 4,500 meters elevation, and checked by auto- 

 matic apparatus exposed from sounding balloons at 25,000 meters 

 elevation, there is now no doubt that the true value lies certainly within 

 one per cent of 1.94 calories per square centimeter per minute. 



We have discovered evidences of variability of the sun's emission. 

 Having devised a brief method of measuring the solar constant, we 

 have applied it several times a day on all favorable days over a long 

 term of years. We have occupied mountain stations in desert lands m 

 Arizona and southern California, in northern Chile, and in South West 

 Africa. Plate i, Figure i shows our station at Mount Montezuma in 

 northern Chile, 9,000 feet above sea-level. Plate i, Figure 2 shows a 

 closer view of the apparatus. The pyrheliometers and the pyranometer 

 are exposed outside, and the solar altitude is measured with a theodo- 

 lite. A beam of light is reflected into a cave observatory where the 

 spectrobolometric work is done. Figure i shows the close accord 

 attained in the monthly mean values of the solar constant at three 

 widely separated stations. It is clear that if the observations at the 

 earth's surface and the estimates of losses in the earth's atmosphere 

 were correctly made, then determinations of the solar constant (that 

 is, the intensity of solar radiation outside the atmosphere) ought to 

 agree exactly wherever made on the earth's surface. In fact we have 

 so far refined our determinations that our two best widely separated 

 observatories, Montezuma, Chile, and Table Mountain, Calif., do 

 agree in their monthly mean values over a period of five years within 

 an average difference of 0.08 per cent. The probable error of the 

 mean curve shown in Figure i is well below o.i per cent. 



