On the circulations caused by the vibration of air in a tube

There seems a need for accurate measurements of the velocity of sound in gases, both at ordinary temperatures with chemically allied series of gases and vapours, and at high temperatures in connection with the quantum theory. The method which suggests itself at once is to utilise in some way the stationary vibrations of a gas enclosed in a tube. One procedure is to use dust to indicate the modes of vibration, in the familiar way first demonstrated by Kundt. Two other methods which do not require the use of dust are now being tested by the writer, and will be described in due course. The successful application of any tube method demands, however, a rather more precise knowledge than has hitherto been available of the behaviour of a real gas when set into vibration in a tube of finite cross section. The present paper deals with certain features of this behaviour which do not appear to have received attention before. These are a general circulation of the air from antinode to node along the walls, returning up the centre of the tube, and a vortex motion which takes place round any obstacle, provided that the amplitude of the vibration exceeds a certain limiting intensity. The effect of this vortex motion is that the forces between two bodies placed in the vibrating gas are fundamentally different from those worked out by Koenig on the tacit assumption of vortex-free motion. It was observation of the behaviour of dust particles that first led to the experiments that revealed the general circulation and the vortex motion round obstacles. It is simpler, however, to discuss the nature of these air movements first and to defer the discussion of their bearing upon the formation of the dust figures described by Kundt, which are of a more varied nature than is always realised, to a second paper. Another series of investigations is devoted to the explanation of the dust movements in terms of these new observations.