THE RATIO OF THE SPECIFIC HEATS OF GASES. 287 



to be in favour of the theorem as a question of abstract 

 dynamics. The difficulty lies in its application. 



The physical objections are serious, but perhaps not 

 insuperable. Let us consider the one most often quoted. 



Mercury has a very brilliant spectrum consisting of a 

 large number of lines. The vibrations of the ether that 

 give rise to this spectrum are presumably caused by the 

 motion of something or other within the molecule. 



Every separate motion in the molecule implies at least 

 one degree of freedom, so that the value of q for mercury 

 should be very large. Hence |3 should be large, and y 

 very little greater than unity, meaning that when heat is 

 communicated to mercury vapour nearly all the energy 

 should go into internal motions of the molecules, and very 

 little to increase the kinetic energy of translation of the 

 molecules, that is, the temperature. But we know from 

 Kundt and Warburg's experiments that j3 is zero for 

 mercury vapour, or none of the energy communicated to 

 the gas takes the form of internal vibrations. 



This seems conclusive against the theorem if the 

 degrees of freedom involved in the production of spectra 

 are really independent dynamical degrees of freedom at all. 

 Boltzmann, however, suggests {Nature, 28th Feb., 1895) 

 that the mercury molecule is only indirectly concerned in the 

 production of the spectrum ; that it is the ether which has 

 a structure, and the numerous degrees of freedom belong 

 to it, and not to the molecule. Schuster {Nature, 24th Jan., 

 1894) and Watson (Nature, 3rd Jan., 1895) have also 

 suggested possible ways oP reconciling the existence of a 

 complicated spectrum with observed values of y. 



It should be noticed that so far as experimental evidence 

 goes at present, a gas can probably not be made to give a 

 line spectrum by mere heating. Wherever we have such a 

 spectrum produced, there is some electrical or chemical 

 action going on, 1 so that in the absence of evidence to the 

 contrary, it may be inferred that the vibrations in the 

 molecule which cause the spectra are not heat effects — 



1 Much work has been done on the subject recently by Pringsheim, 

 Paschen, Smithells and others, but the question is still under discussion. 



