368 Scientific Proceedings, Loyal Dublin Society. 
It is very instructive to glance over the determinations of y in 
the Table on p. 364. Take, for example, the diatomic gases. Six 
of them are transparent, and the other four are coloured. The 
transparent gases furnish values for y, which all lie in the neigh- 
bourhood of 1:4, which is the value which would correspond to a 
molecule having five degrees of freedom if the conditions of 
Boltzmann’s theorem were completely fulfilled. In transparent 
gases they are probably not much interfered with, since in them 
the electrons are associated with Bd, events, and it is only after 
“many woillions of encounters that the ether can, through them, 
sensibly effect the A and Ba events with which the theorem is 
concerned. On the other hand, in the coloured gases one or more 
electrons (in these special gases probably something like six or 
eight electrons, judging from their spectra) are associated with Bo, 
events, and largely affect the events with which the theorem is 
concerned. They, accordingly, cause y to be different from what 
it would be if the dynamical interactions stood alone, which, 
judging from the transparent gases, we may conclude would be 
1-4, furnished by 5 degrees of freedom. The Bd motions in ques- 
tion would probably add something like three times 6 or 8 more 
degrees of freedom if the molecule stood alone, but by reason of 
the linkage only some fraction of this addition has to be made. 
It appears from the Table that in chlorine, which is the least 
deeply-coloured gas, that is, the gas in which the linkage is most 
effective, the addition to be made is of 1:2 degrees of freedom ; 
while in bromine and iodine an addition of 1-9 degrees of freedom 
has to be made, indicating that the linkage is more lax in these 
gases. 
Since in so many bodies the electrons seem to be associated 
with events which are very much isolated from those that are 
chiefly affected by the encounters, the modification of the dyna- 
mical condition which is introduced through them may be 
regarded as a perturbation. It would perhaps not be impracti- 
cable to discover in what way perturbating forces, not obeying the 
conditions of the Boltzmann-Maxwell theorem, can influence the 
results of that theorem. This would be of much value; and, if it 
can be made out, will, perhaps, explain why, in some transparent 
diatomic gases, the value of y is above 1:4, while it is less than 
that value in others. The reason probably is that in hydrogen, 
