Theory of Radiation. 263 



direction. But supposing now some parts of the enclosure to 

 be hotter than others, conduction of heat takes place ; and 

 although the thermometric temperature of the element we are 

 considering may be the same as before, yet the molecules will 

 move more quickly in one direction than in another, and the 

 temperature of the element, without regard to the temperature 

 of the surrounding elements, will not indicate the direction in 

 which the molecules move fastest. The temperature at a given 

 point is only an average indication, and only completely 

 defines the state of the gas if the temperature is uniform. We 

 should be wrong to assume that, because a certain distribution 

 of velocities holds when the temperature is uniform, it must 

 also hold when conduction of heat is allowed to take place. 

 What holds in the case of conduction of heat may also hold in 

 the case of radiation; and we cannot assume, without further 

 proof, that the radiation of a body is independent of the tem- 

 perature of the surrounding bodies. It admits, on the con- 

 trary, I believe, of proof that the radiation of a body into a 

 cooler medium, being a phenomenon of dissipation of energy, 

 involves altogether different considerations from those we are 

 accustomed to apply when the temperature does not change. 

 In order to fix the attention, imagine a Bunsen burner with a 

 soda bead in it. The spectroscope will show us the well- 

 known yellow sodium-lines ; but we know that the sodium- 

 molecule is also capable of sending out certain rays in the 

 green part of the spectrum*; and these rays are so weak in the 

 case under consideration, that they escape observation. Fol- 

 lowing backwards the rays which enter the spectroscope, we 

 finally reach the internal radiation of the Bunsen flame; and 

 we may therefore say that the internal radiation is strong for a 

 certain set of yellow vibrations, and weak for a certain set of 

 green radiations. About the vibrations of the molecule the 

 spectroscope does not directly reveal any thing; but there 

 must be some relation between them and the intensity of radia- 

 tions of the surrounding medium. If there is more internal 

 radiation for the yellow light, it can only be because the vibra- 

 tions of the molecules for the yellow light are stronger. Suppo- 

 sing, now, the flame to be suddenly placed in an enclosure of 

 constant temperature, which temperature shall be the same as 

 that of the flame itself. The internal radiation is constant 

 within that enclosure; or rather it only depends on the refrac- 

 tive index of the medium. 



For the yellow light we should find that the internal radia- 



* 1 do not wish to imply any definite opinion on the sodium-snectrum 

 and its changes with temperature, but only take a certain view ot it as an 

 illustration. 



