____________________________ _n_λ_{argon} = _c__{argon} = √γ(_p_/_d_)(1 + _at_)_{argon}, where _n_ is the number of vibrations per second, =λ= the wave-length of sound, and _a_ the coefficient of the expansion of a gas for a rise of 1° in temperature, _t_, viz. 0·00367. Now if the expression _p_ (1 + _at_) can be shown to be identical for argon and for air, the value of =γ= for argon can be calculated by the very simple proportion-- λ^{2}_d__{air} : λ^{2}_d__{argon} :: 1·408 : γ_{argon}. This involved a measurement of the rate of rise of pressure of argon, _p_, per degree of rise of temperature, _t_; or, in other words, the verification of Boyle’s and Gay-Lussac’s laws for argon; and this research was successfully carried out by Dr. Randall of the Johns Hopkins University of Baltimore, U.S.A., and Dr. Kuenen, of Leyden, working in Professor Ramsay’s laboratory.[29] They made use of a constant volume thermometer, and measured the rise of pressure corresponding to a definite rise of temperature, comparing the gases argon and helium in this respect with air. The values found between 0° and 100° for air, argon, and helium were-- One volume air, heated from 0° to 100°, raises pressure in the proportion of 1 to 1·3663 Argon 1·3668 Helium 1·3665 It may therefore be taken for certain that, within the limits of experimental error, the value of the expression _p_(1 + _at_) is identical for all three gases. We see, then, that for argon, as for mercury gas, the value of =γ=, the ratio between the specific heats at constant volume and at constant pressure, is 1 to 1·66, whereas for air, hydrogen, oxygen, nitrogen, carbon monoxide, and nitric oxide, it is 1 to 1·4. We have now to consider what conclusion can be drawn from this difference. On the usually accepted theory of the constitution of matter, it is held that atoms may be regarded as spheres, hard, elastic, smooth, and practically incompressible. True, we really know little or nothing regarding the properties of such particles, if particles there be; but in considering their behaviour it is necessary to make certain suppositions, and to see whether observed facts can be pictured to our minds in accordance with such postulates. If, from the known behaviour of large masses, conclusions can be drawn regarding small masses, and if these conclusions harmonise with what is found to be the behaviour of large numbers of small masses acting at once, the justice of the supposition is, although not proved, at least rendered defensible as one mode of regarding natural phenomena.