THE ETHER AND MATTER POINCAEE. 209 



is necessary only to a])ply tlie principle of relativity of Lorentz. 

 Accordingly, the various bundles of rays of the same true wave length 

 relative to the molecule and striking the molecule from various 

 du'ections would not have the same apparent wave length to the 

 observer who supposes the molecule at rest. The apparent wave 

 length is not altered by the reflection, but the true wave length is. 



We thus come to an interesting result: The reflected or dift"used 

 energy is not equal to the incident energy. What is unaltered is not 

 the energy, but the product of the energy by the wave length. At 

 fii-st I felt satisfied. It seemed from tliis that an incident quantum 

 would give a diffused quantum since a quantum varies inversely as 

 the wave length. Unfortunately this brings us nothing. 



I was led by this analysis to Rayleigh's law. I laiew that in advance , 

 but I hoped that in seeing in detail how I was brought to it that I 

 could tell what modifications I must make in the hypothesis in order 

 to get Planck's law. In that hope I was deceived. 



My first thought was to look for something resembling the theory 

 of quanta. It would indeed be surprismg that tvv'o such entirely 

 di&'erent explanations should both take account of the same deroga- 

 tion of the law of equipartition of energy. What is the effect of the 

 discontuiuous structure of energy? We might suppose that this 

 discontmuifcy relates to the luminous energy itself when it passes 

 through the free ether and so when it falls upon a molecule does so 

 in a discontmuous manner, in little separate battalions. It is easy 

 to see that that would not alter our results. 



Or, we might suppose that the discontinuity occurs at the very 

 moment of diffusion, tliat the diffusing molecule does not diffuse the 

 light in a continuous manner, but by successive quanta. This again 

 will not do because if the light wishing to be transformed has to wait, 

 so to speak, in. an antechamber for its omnibus to fill before starting 

 out, there would result a forcible retardation. Now, according to 

 the theory of Lord Rayleigh, the diffusion by the molecules in the 

 direction of the incident ray produces simply ordinary reflection. 

 That is to say, it interferes regularly with the incident light, and this 

 would not be possible were there a retardation of phase. 



If we try, impartially, to choose which of our premises we must 

 abandon, we are none the less embarrassed. We can not see how we 

 can deny the principle of relativity. Must we then modify our law 

 for the diffusion of light by molecules at rest ? That would be very 

 difficult. We surely could not stretch our imagination into believing 

 that the sky is not blue. 



I will stop in this 'embaiTassment and close with the following 

 reflections. As science progresses, it becomes more and more difficult 

 to fit in the new facts when they will not fit in spontaneously. The 

 older theories depend upon the coincidences of so many numerical 



