﻿Influence of Density on the Spectra of Incandescent Gases. 191 



the formation of a spectrum of the light emitted from the gas 

 layer considered, the values X and X { correspond to two immedi- 

 ately adjacent parts of the spectrum ; we shall have 



> V _i-(i-A x r v 



i K, m = K — K ^- ■ J\ • 



It needs not particular mention that in the concrete case, in 

 which the breadth of the spectrum-parts compared may be not 

 infinitely small, the values of X and X t are to be taken as mean 

 values for the adjacent narrow bright bands of the spectrum. 

 For the ratio of brightness of these adjacent parts we obtain ac- 

 cordingly, putting ■— = J A and -± =J K :— 



&* _[i-a-A A nj; 



E A/Wi [l-(l-Aj w ]J Al 



(2) 



3. The magnitudes J A and J Al are two values of the ratio of 

 the emissive to the absorptive power for two very slightly differ- 

 ent values of the wave-length at the same temperature. But this 

 magnitude, multiplied by a constant, is nothing else but the 

 function of the wave-length and temperature denoted by J in 

 KirchhofFs above-mentioned treatise. 



This function has the following remarkable properties : — ' 



(1) J is independent of the peculiar properties of the body 

 (cf. Kirchhoff, /. c. p. 392). 



(2) At constant temperature, therefore, in one and the same 

 spectrum, J varies continuously with the wave-length till the 

 value of the latter is reached at which J vanishes {ibid. p. 393). 



(3) It may be pronounced in the highest degree probable that 

 the function J, at constant temperature, presents no conspicuous 

 maxima or minima with the change of wave-length {ibid. p. 393). 



From the last two properties of the function J, it follows that 

 " when, in the spectrum of an incandescent body, strongly pro- 

 nounced maxima or minima are seen, its absorptive power, con- 

 sidered as a function of the wave-length of the incident rays, 

 must have conspicuous maxima or minima with those values of 

 the wave-length," 



By the help of this principle, Kirchhoff infers theoretically 



the possibility of reversing the discontinuous spectra of flames ; 



and since this inference is confirmed by observation, conversely 



an empiric verification of the two properties mentioned of the 



function J may herein be seen. 



E F 



Hence, in the present case, the two values -— - and. ~, for the 



A A Aa 



