SOLAR OBSERVATORY. 153 



mean value of about 2.03 calories. The accuracy of the work depends upon 

 the uniformity of the transparency of the sky for the hours when the sun's 

 altitude is rapidly changing, and all the internal evidence of the observations 

 indicates that on most days the conditions above Mount Wilson were excel- 

 lent in this respect. It is believed that the separate determinations of the 

 "solar constant" are usually accurate, relatively to one another, within i per 

 cent; so that the considerable range of values from 1.93 to 2.14 calories 

 indicates a real variability of the solar radiation outside the earth's atmos- 

 phere. This conclusion has been tested in every way yet thought of, and 

 seems unshaken. 



The variations do not depend on changes of the transparency of the earth's 

 atmosphere or of the amount of its humidity, for there is no connection, 

 either direct or inverse, which holds consistently as between variations of 

 the "solar constant" values and variations of the transparency or humidity 

 of the atmosphere, as observed on different days. 



There is a real change of solar radiation at the limit of the atmosphere 

 which depends on the change of distance between the earth and the sun. 

 This change, which amounted to between 3 and 4 per cent during the 

 interval covered by the measurements, was accurately shown by the results, 

 so that confidence in the reality of the other changes observed is increased 

 by this circumstance. Measurements were made in Washington nearlv simul- 

 taneously with some of those on Mount Wilson. The Washington values of 

 the "solar constant" of radiation are evidently less trustworthy than the 

 Mount Wilson values, but differ from them by only about 3 per cent on the 

 average. Accordingly, great confidence is felt in the accuracy of the method 

 of determination, which leads to substantially identical results when applied 

 at two stations separated by 3,000 miles in longitude and i mile in elevation. 



The transmission of the atmosphere for monochromatic rays of 44 differ- 

 ent wave-lengths was determined on Mount Wilson on 130 dififerent days. 

 Comparing average values with average values for Washington, the average 

 transmission of the mile of air nearest sea-level has been obtained. It proves 

 that about as much light is scattered from the direct beam in the mile of air 

 nearest sea-level as in the many miles of atmosphere above Mount Wilson. 



The exact form of the solar-spectrum energy curve outside the atmos- 

 phere has been delineated, and also the forms resulting at the surface of the 

 earth for dififerent altitudes of the sun. 



In 1906 a third small observing shelter, surmounted by a tower 50 feet 

 in height, was erected on a point overlooking deep canyons on three sides, 

 and suitable apparatus was installed there to measure, at given altitudes and 

 azimuths, the reflecting power of the level ocean of cloud which rose some- 

 times to within a hundred feet of the base of the tower. From these meas- 

 urements it has been computed that the earth, if it were wholly clouded over, 

 would reflect to space 65 per cent of the sun's radiation. 



Measurements were made from this tower of the intensity of radiation 

 scattered from the sky at dififerent altitudes and azimuths, and of the com- 

 parative quality of sunlicht and skylight. 



With the data secured at Mount Wilson and at Washington a new deter- 

 mination of the albedo of the earth has been made, and it is found that 

 37 per cent of the solar radiation is reflected to space. 



The expeditions to Mount Wilson have been jn personal charge of Mr. 

 C. G. Abbot, and Mr. L. R. Ingersoll, of the University of Wisconsin, assisted 

 in the work for 3 months of each season. 



