282 SCIENCE PROGRESS. 



gases of the same atomicity, may there not be another class 

 of diatomic gases possible with |3 zero, or very small ? 



It has already been mentioned that, throughout the 

 table, gases with the same atomicity have in most cases the 

 same y, but there is one class of gases that uniformly 

 prove exceptions. These are the gases with more than 

 one halogen atom in the molecule. Their proving an 

 exception cannot be due to the halogens having intrinsically 

 a higher heat capacity than other atoms, for in several 

 cases, such as H 2 , C 2 H 6 and C 3 H 8 , one H can be replaced 

 by CI without change in y, and the natural conclusion 

 to draw from this is that the CI has itself the same heat 

 capacity as the H that it replaces. We seem, in fact, to 

 have in the case of gases something similar to Dulong and 

 Petit's law of equality of atomic heats for elementary solids, 

 but the divergences cannot be explained in the way in which 

 the anomalous values for carbon, etc., were explained, for 

 chlorine, for instance, was shown by Strecker to have a 

 value of y that does not vary with the temperature. 



Since then the number and nature of the atoms in the 

 molecule do not fix the value of y, we must probably look 

 to their relations to each other or the configuration as an 

 important element, and from the many cases where atoms 

 of different kinds are interchangeable without altering y, it 

 may very well be that the distribution of energy depends 

 on nothing but the configuration. Of course the term 

 configuration must be taken to include the size as well as 

 shape, otherwise we should have no explanation of the 

 difference between H CI and Cl 2 , for in a diatomic molecule 

 the atoms must simply lie side by side, and no variation in 

 shape is possible. On this hypothesis the atoms in the 

 chlorine molecule would be farther apart than the atoms in 

 most diatomic molecules. Also methyl chloride would have to 

 be taken as having a different configuration from methane. 

 Chemical facts tell us that we can oet but one CH, CI 

 whichever H in CH 4 we replace by CI, but this does not 

 preclude the possibility of the chlorine distorting the mole- 

 cule to a shape that is the same in every case. 



We might go a step farther still, and ask whether the 



