208 ANNUAL EEPOET SMITHSONIAN INSTITUTION, 1912. 



piston the final distribution must be that corresponding to maximum 

 entropy, which is the same as that of black radiation. Such reasoning 

 has no value. 



The distribution of the radiation has a tendency to approach that 

 of black radiation. It can then no longer change, for wc can not 

 pass heat from a cold to a hot body; that is to say, without the 

 expenditure of external work. But here external work is supplied 

 through the strokes. of the piston, which appears as an augmentation 

 of the luminous energy in the cavity of the pump. The work is 

 changed into heat. 



The same difficulty would no longer exist if the bodies in movement 

 on which the light suffered reflection were infmitely small and infi- 

 nitely numerous ; then* kinetic energy would not correspond to mechan- 

 ical work but to heat. We could not then compensate the diminu- 

 tion of entropy which corresponds to the change in the distribution 

 of the energy by a transfonnation of work into heat, and then we 

 would be right in concluding that, if the initial distribution were 

 black that distribution would remain indefinitely. 



Let us now imagme an inclosure with fixed and perfectly reflecting 

 walls; let us inclose not only luminous energy, but also a gas. The 

 molecules of the gas will act as moving mirrors. If the distribution 

 of the energy among the wave lengths is that of the black radiation 

 corresponding to the temperature of the gas, then that distribution 

 should be stable; that is: First, whatever action the light has upon 

 the molecules should not alter the temperature of the gas; second, 

 whatever action the molecules have upon the light should not alter 

 its distribution of energy. 



Einstein examined this action of the light upon the molecules. 

 The latter suffer something which resembles the pressure of radiation. 

 Einstein, hov/ever, does not state the matter so simply. He com- 

 pares the molecules to very small mobile resonatore possessing at the 

 same time not only the kinetic energy of translation, but also the energy 

 of electric oscillations. The result at any rate would have come out 

 the same — he reached Rayleigh's law. 



Personally, I would have done the reverse and studied the action 

 of the molecules upon the light. The molecules are too small to 

 produce regular reflection. They could produce only a diffusion of it. 

 It is this diffusion, when we neglect the molecular movements with 

 which we are acquainted both in theory and experiment. It indeed 

 produces the blue of the sky. 



This diffusion does not alter tlie wave length, but is much greater the 

 smaller the wave length. 



We must pass now from the case when the molecules are at rest 

 to when their motion must be accounted for, taldng into account 

 their agitation which produces then' temperature. That is easy; it 



