In all cases the numbers are too large, and this is a serious difficulty, because any tendency to rotate round the central line would cause the values to be less, not greater than 1·4. For triatomic molecules the calculated value of γ is 1⅓, but in actual fact the ratio in the case of triatomic molecules, such as H_{2}O, CO_{2}, N_{2}O, etc., is always less than 1⅓. These speculations stand on a basis very different from the first conception, namely, that all heat must be employed in communicating translational motion to molecules of mercury gas, argon, and helium, and it appears that the atoms of these three elements must necessarily be regarded as having the properties of smooth elastic spheres. The atoms and the molecules must in their several cases be identical. And, inasmuch as the chemical evidence regarding mercury leads to the same conclusion, it appears legitimate to infer that argon and helium must also be monatomic elements. [28] _Méchanique céleste_, vol. v. p. 123. [29] _Proc. Roy. Soc._ vol. lix. p. 63. CHAPTER VII THE POSITION OF ARGON AMONG THE ELEMENTS From what has been said in the preceding chapter there can be no doubt that the molecular weight of argon is 39·88. We have now to consider what this conclusion involves. Taken in conjunction with the fact that the ratio between its specific heat at constant volume and that at constant pressure is 1⅔, it follows that energy imparted to it is employed solely in communicating translational motion to its molecules. In the case of mercury gas such behaviour is taken as evidence that the conclusion following from the formulae of its compounds, from the density of its compounds in the gaseous state, and from its own vapour-density, as well as from its specific heat in the liquid state, namely, that its molecules are monatomic, is correct. Is it legitimate to conclude that because argon in the gaseous state has the same ratio of specific heats, therefore it also is a monatomic gas?