384 F. W. Very — Sky Radiation and Isothermal Layer. 



absorption occurs (an amount which might be inferred from 

 the lowest estimates of the absorption bands in the solar spec- 

 trum), 19'5/0'07= 279 minutes =4'65 hours will be required 

 for a rise of temperature of 1 ° C. in the entire mass, and the 

 average diurnal fluctuation of temperature should not exceed 

 3° C, subject to what variation may arise from unequal distri- 

 bution of absorbent substances. The large amount of heat 

 required to maintain this great mass of air at a temperature of 

 over 200° Abs. C, or to change its temperature by even a few 

 degrees, must tend to keep the thermal condition constant, but 

 the absence of a seasonal variation can only mean that the air 

 is still impervious to the greater part of the radiation from a 

 body of air at this temperature. The freer atmospheric radia- 

 tion, of which there is also evidence upon entering this region, 

 probably appertains to the radiant bands of shorter wave- 

 length emitted by the aqueous vapor and other atmospheric 

 constituents. These bands of shorter wave-length (for example, 

 those constituting great H) can not be very strongly emitted at 

 this low temperature, but they serve to limit the rise of temper- 

 ature attainable by absorption of solar radiation to a very 

 narrow range, for if the temperature were to increase to only 

 a small extent, the radiant power of aqueous vapor in these 

 bands would grow at a much more rapid rate, because the 

 temperature in question is nearly the critical one of greatest 

 sensitiveness and most rapid percentage increase for these 

 particular bands. 



Each emission band of a discontinuous emission spectrum 

 has some low temperature at which it begins to appear, suc- 

 ceeded after a certain amount of development by a critical 

 period of most rapid change as the temperature continues to 

 rise, and finally reaching a steady stage of maximum develop- 

 ment, when there is little further change of relative intensity, 

 but thereafter, each component homogeneous ray in narrowly 

 limited bundles within the band increases according to some 

 power of the absolute temperature not far from the fourth or 

 fifth through a considerable range, and matches the correspond- 

 ing radiation of a full radiator within these narrow limits. 



Note. — When developed to their greatest extent, the maxima of emission 

 in line or band spectra are found to agree with the radiation for corre- 

 sponding wave-lengths in the spectrum of a black body. The statement in 

 the text is only an approximation. The variation of homogeneous radia- 

 tions can not be expressed as simply as this except close to the spectral 

 maximum. 



Approximate rates of increase of homogeneous radiations with increasing- 

 temperature were first obtained throughout the infra-red spectrum at Alle- 

 gheny Observatory in 1885 by means of the spectro-bolometer, and the fact 

 that the maximum ordinate of a spectral energy-curve moves towards 

 the shorter wave-lengths with increasing temperature was discovered at 

 the same time. The experiments were described by Langley in Annales de 



