250 Prof. J. H. Jeans on Temperature-Radiation 



factor e~ 2 P@, where ft is a positive quantity. If the origin 

 of heat-radiation is to be found in the collisions between 

 electrons and atoms of matter, then ft will be comparable 

 with the time of collision, say, 7 x 10 -14 sec. at ordinary 

 temperatures. For yellow light p = 3xl0 15 , so that 

 e -2p0 = g-42o a Thus, purely as a matter of calculation, and 

 apart from any special hypotheses or assumptions, we find 

 that the rate at which visible light would be emitted, as the 

 result of heat-radiation, from matter at ordinary temperatures,, 

 must necessarily be very slow. Thus, although the final law 

 of partition of radiant energy in a perfectly reflecting en- 

 closure containing some matter would be that of the "normal 

 state," given by formula (32), yet it would require centuries 

 to reach this final state, and the smallest departure from 

 perfection in the reflecting power of the walls would result 

 in this final state being impossible of attainment, even if 

 infinite time were available. 



So far from it being possible to assume infinite time or 

 perfect reflecting power under experimental conditions, we 

 find that we must assume exactly the reverse when light of 

 short wave-length is concerned. For light of sufficiently 

 small wave-length, the densest matter must be as transparent 

 as is the atmosphere for ordinary light. Even for light of 

 the wave-length of Rontgen rays, the walls of the experi- 

 mental enclosure must be regarded as practically transparent. 

 Thus energy of short wave-length disappears entirely from 

 the enclosure within, say, 10~ 8 seconds after its emission, 

 while the emissions of centuries would have to accumulate 

 before the " normal state " could be established. 



Conclusion. 



26. It now appears certain that the observed partition of 

 energy which Planck's formula (4) attempts to represent 

 cannot be that of the "normal state"" — at least for short 

 wave-lengths. 



We are at once confronted with the question : Why is it 

 that the partition of energy given by this formula is, to all 

 appearance at least, that of a final state which the system 

 tends to assume, independently of the state from which it 

 started ? For the existence of such a final state, different 

 from the "normal state," would at first sight seem contrary 

 to the theorem of § 10. 



In answering this question it is necessary to emphasize 

 the distinction between "free" and "forced" oscillations 

 The " normal state" is one in which the oscillations are free; 



