ENERGY, ESPECIALLY IN RADIATION 95 



3. Jeans has recently discussed Larmor's view, and arrives 

 at the conclusion (Phil. Mag., Dec. 1910) that it is neither 

 possible to avoid finiteness of the element of energy nor 

 ultimate discontinuity of ether structure in relation to radia- 

 tion if Planck's law corresponds to the true final condition of 

 equilibrium. That is to say, radiation can only be regarded as 

 capable of existing in the ether in amounts which are multiples 

 of a finite unit. 



While Jeans' own view ( 1) must be recognised as indicat- 

 ing a possible solution of the fundamental difficulty regarding 

 the partition of energy, it is not possible, because of our 

 ignorance of the intrinsic nature of matter, of ether, and of the 

 connection between these, to be quite certain that Larmor's 

 view, or even Planck's, is inadmissible. It is not inconceivable 

 that the nature of these entities may impose identity between 

 the distribution which obtains in the steady state under 

 experimental conditions, and that which would obtain in the 

 final state of a strictly conservative system. I venture there- 

 fore to indicate the following mode of considering the problem. 

 It leads to an expression which differs slightly in form from 

 that of Planck, but which can practically be identified with it 

 throughout the range of observed wave-lengths, and which 

 with it reduces to Rayleigh's form when the wave-length is of 

 suitable magnitude. 



4. Interchange of energy amongst freedoms of the same 

 type constitutes ordinary transmission of energy of the 

 type involved ; interchange of energy amongst freedoms of 

 distinct types constitutes that transmission of energy which is 

 ordinarily called transformation. When different subsystems, 

 in the equilibrium condition, are freely open to interchanges 

 of energy, a universal generalised temperature or potential, 

 possessing a definite statistical value throughout the total 

 system when that system possesses a definite total amount of 

 energy, must exist. 



Let there be altogether v subsystems, let JVj . . . N, be 



