﻿Titer mod ynamical Theory of Radiation. 59 



([mass]-!- [length] 3 ) to the aether, it remains impossible to 

 construct a out of aether-constants. 



The evaluation of g can, however, proceed by taking 

 advantage of the circumstance that there are physical con- 

 stants associated with matter, which are common to all kinds 

 of matter — in particular, the charge of the electron, e. If 

 we make a depend on the charge of the electron, the 

 evaluation of a may be borrowed from the second theory. 

 We then find that a is proportional to e~ 6 , and the numerical 

 value of e deduced from the observed value of a supplies 

 strong confirmation that this is the right way of evaluating <r. 



6. Thus if we agree to extend the first (thermodynamical) 

 theory in the way here suggested, we may say that both 

 theories lead to the formula o-T 4 , that both lead to the same 

 numerical evaluation of cr, and that this evaluation is in 

 agreement with experiment. 



7. Next let us imagine an enclosui e with perfectly reflecting 

 walls, containing matter at temperature T. Let us refer to 

 this piece of matter as R, and let us suppose that the matter 

 is real matter in which each electron has the customary 

 charge e. Let us suppose that we have also a mass of 

 "ideal" matter I, in which each electron has a charge ^ e only. 



If we put R inside the enclosure we obtain, according to 

 the thermodynamical argument, a state of equilibrium in 

 which the aether has a volume density of energy o-T 4 . 



Remembering that the constant a has been found to be 

 proportional to e~ 6 , it appears that if we take R out of the 

 enclosure and insert I, we obtain equilibrium with a volume 

 density of energy 64o\T\ 



Now let us put R and I in the enclosure together. The 

 thermodvnamical argument might, I think, very fairlv be 

 used at this juncture to establish the equation 1 = 64. With- 

 out pressing it thus far, however, we may say that it shows 

 that the ideal matter 1 will give out more energy than it 

 receives, while the opposite will be true of the real matter R. 

 Thus the two bodies, originally at equal temperatures, will 

 tend to assume different temperatures. This might or might 

 not happen in nature if we had id- al matter to experiment 

 with. What is of more importance is that the thermody- 

 namical argument has led to a result which is opposed to the 

 fundamental principles of thermodynamics. Or, to state the 

 situation in another way, the thermodynamical argument first 

 supposes the second law to be of such wide applicability that 

 it may be applied to the aether, and then proves that it may 

 not legitimately be applied even to different kinds of matter. 



8. The difficulty may be illustrated by a second imaginary 

 experiment, suggested to the writer in a conversation with 



