1917] on Recent Developments of Molecular Physics 



81 



A numljer of instances, such as that of the atomic heats of metals 

 ah'eady referred to, j^rove that the theorem is true m nature, 

 at least within certain limits. But the limits are easily discovered, 

 and it is readily seen that the theorem is not of universal applica- 

 bility. For example, in the instance just taken, the atom of silver 

 has four times as many electrons in its structure as the atom of 

 alumininm, so that its internal structure has four times as many 

 degrees of freedom. AVhy, then, does it not get four times as much 

 energy ? This and similar cases of failure of the theorem formed 

 the cloud over the dynamical theory of heat. 



A good deal of information can be obtained by investigating in 

 what ways the energy is distributed in the cases in which the theorem 



Silver 



Copper 



T =- 100 200 



Aluminium 



Fig. 1. 



of equipartition is found to fail. In the case of sohd elements such 

 as silver and aluminium, it is found that all the heat-energy resides 

 in the motions of the atoms as a whole ; the internal motions of the 

 electrons get none at all. The same is true in a monatomic gas, 

 such as helium or mercnry vapour, as is shown by the fact that the 

 ratio of specific heats is 1§. In a diatomic gas at ordinary tempera- 

 tures, three-fifths of the total energy resides in the motion of 

 translation of the molecules, while the whole of the remaining two- 

 fifths resides either in the motion of rotation of the molecules or in 

 YoL. XXII. (Xo. Ill) G 



