abbot: solar constant of radiation 97 



diminution of radiation attending the decline of the sun from 

 zenith to horizon, or the descent of the observer from a high alti- 

 tude to a lower one. Accordingly several empirical formulae of 

 more complexity were proposed, which owing to their more nu- 

 merous constants, could be made to fit the observed variation of 

 the total intensity of radiation under different conditions more 

 closely. By the aid of such empirical formulae higher values 

 of the solar constant have been obtained. Some of these in our 

 own time have gone as high as 4 calories. Radau however says 

 ''it is clear that the intensity of the solar radiation outside the 

 atmosphere cannot be certainly obtained from experiments which 

 have been made [prior to 1878] for the result depends essentially 

 on the manner of calculation." This conclusion is still applica- 

 ble to pyrheliometer measurements not supported by spectrum 

 observations. 



The tendency toward high values of the solar constant was 

 powerfully stimulated by the publication of the report of the 

 Mount Whitney expedition by Langley in 1884. As Forbes and 

 Radau had stated, so Langley emphasized and acted upon the 



B 



fact that the formula A = Aqp'*^^ applies only to a homo- 

 genous bundle of rays in a pure atmosphere; and the intensity 

 of solar radiation outside the atmosphere can be exactly deter- 

 mined only when the atmospheric transmission coefficients of the 

 rays of all wave lengths, which go to make up the complex beam 

 of the sun, are separately determined and allowed for. Langley 

 was the first to determine and apply atmospheric transmission 

 coefficients for numerous rays of different wave lengths in the 

 solar spectrum. For this purpose he invented the bolometer, a 

 delicate electrical thermometer, and observed with it the varia- 

 tion of the intensity of each ray of the spectrum from low sun 

 to high. He found it impracticable to determine the transmis- 

 sion coefficients in the water vapor bands of the infra-red, but 

 assuming that there were no water vapor bands in the solar 

 spectrum outside our atmosphere, he avoided this difficulty by 

 smoothing the spectrum energj^ curve, which he computed from 

 his bolometric observations to represent the distribution of solar 



