REPORT OF THE SECRETARY. 79 



on Mount Wilson, both in 1905 and 1906, so that there can be little question that 

 the large changes noted there are really solar changes and not of atmospheric 

 or accidental origin. 



The reality of the supposed solar origin of the changes of radiation observed 

 is attested by many other evidences stated in Chapter VI, Part I, and Chapter 

 III. Part III. 



The temperature of the earth is shown to be in good agreement with the as- 

 sumed value of the "solar constant," 2.1 calories. Indeed, it is shown that un- 

 less the albedo, or reflection, of the earth exceeds 37 per cent (a value here 

 determined for it and based on observations at Washington and Mount Wilson), 

 then the mean value of the solar constant can not exceed 2.33 calories, else the 

 earth must be a better radiator than the " absolutely black body " or perfect 

 radiator. 



It is shown that the surface of the earth can radiate only very slightly to 

 space, on account of the interference of clouds and water vapor to terrestrial 

 radiation ; and that the substance which maintains the earth at nearly con- 

 stant temperature, by emitting to space radiation equal to that received by the 

 sun, is principally the water vapor layer at 4,000 to 5,000 meters in elevation, 

 whose mean temperature is 10° or more below 0° C. 



There is introduced the conception of an " hypothetical earth," similar in di- 

 mensions and motions to the real earth, but hollow and like a soap bubble in 

 thickness of wall ; perfectly absorbing for solar radiation, and a perfect radiator 

 for long waves ; perfectly conducting for heat along parallels of latitude, but 

 perfectly non-conducting along meridians of longitude. The temperature of this 

 " hypothetical earth " is calculated for all times of the year, and for all lati- 

 tudes, by the aid of the known value of the " solar constant " and the laws of 

 radiation of perfect radiators. 



A comparison is made between the annual march of temperature of the 

 " hypothetical earth " and the observed annual march of temperature for 64 

 stations on the real earth. It is thereby shown that a given fractional change 

 of solar radiation running its cycle in a year produces one-fourth the given 

 fractional change in the absolute temperature of the " hypothetical earth," 

 one-fourteenth of the given fractional change in the temperature of most inland 

 stations, one twenty-fifth for coast stations ; and one-fiftieth for small islands 

 in great oceans. For a fluctuation of 5 per cent in solar radiation having a 

 period of about a year there would be produced a change of only about 1° C. 

 in the mean temperature of inland stations and only about 0.3° C. for island 

 stations. The effects of more rapid changes of solar radiation would be less 

 readily discernible in their effects on mean temperatures, but may nevertheless 

 be of meteorological importance as promoters of atmospheric circulation. 



From a comparison extending over thirty years of the temperatures of 47 

 well-distributed inland stations it appears probable that changes of solar 

 radiation do produce, not infrequently, well-marked and recognizable changes 

 of temperature over the continental areas of the world. Such changes of tem- 

 perature would be predictable if accurate measurements of the solar radiation 

 were systematically continued at a few favorable stations. 



Numerous measurements of the comparative brightness of the center and 

 edge of the solar disk indicate that the observed changes in solar radiation are 

 attended by a variation of the transparency of the solar envelope, and perhaps 

 are caused by it. 



Many results of observation not here enumerated, such as the mean trans- 

 parency of the upper and lower strata of air, the reflecting power of the clouds, 

 the probable temperature of the sun, and the quality of the radiation of sun- 

 spots, will be found set forth both in words and by charts; and also a full 



