Fig. 4.

The simple apparatus with which such determinations are made is due to the physicist Kundt. It consists of a glass tube, through one end of which a glass rod passes, so that half the rod is enclosed in the tube, while the other half projects outside it. In the experiments on argon, the rod was sealed into the tube; in other cases, it is better to attach it with indiarubber, or to cause the rod to pass through a cork. The open end of the tube is connected with a supply of the gas, so that, after the tube has been pumped empty of air, the gas, in a pure and dry condition, can be admitted. Some light powder (and for this purpose lycopodium dust—the dried spores of a species of fungus—is best) is placed in the tube, and distributed uniformly throughout it, so that when the latter is in a horizontal position, a streak of the powder lies along it from end to end. The portion of rod outside the tube is rubbed with a rag wetted with alcohol, when it emits a shrill tone or squeak, due to longitudinal vibrations; the pitch of the tone depends, naturally, on the length of the rod, a long rod giving a deeper tone than a short one. The vibrations of the rod set the gas in the tube in motion, and the sound-waves are conveyed from end to end of the tube through the gas. As the tube is closed at the end through which the gas was admitted, these waves echo back through it; and a great deal of care must be taken to make the echo strengthen the waves, so that the compressions produced by the back waves are coincident in position with the compressions produced by the forward waves travelling onwards from the rod. The gas, could we see it, would represent portions compressed and portions rarefied at regular intervals along the tube. Where the gas is compressed, it gathers the lycopodium dust together in small heaps, the position of each heap signifying a node of compression. Hence, comparing the distances between the nodes of compression for any gas and for air, we find the relative wave-lengths of sound in the two gases; and, as the velocity of sound in air has been accurately measured, we thus determine the velocity of sound-waves in the gas under experiment.

Such experiments were made by Kundt and by his co-worker Warburg on mercury gas, and they found that in this case the value of γ was 1·67; that is, in the equation

γ = c2d/p

the value 1·67 had to be ascribed to γ, in order to render it equal to the product of the square of the velocity into the density, divided by the pressure.

Similar experiments with argon led to the same result as Kundt and Warburg found for mercury gas; but the calculation becomes more simple if it is allowable to take for granted that the elasticity, or alteration of pressure produced by unit alteration of volume, is identical in the case of argon and air. The full equations are—

and