THE classical conception of the mechanism of the propagation of sound in gases presents two alternative expressions for the velocity, according as one assume with Newton that the elastic processes are isothermal, or with Laplace that they are adithatic It is well known that experiments show the fatter to be appropriate down to the lowest frequencies at which measurements have been made. The ratio of specific heats at constant pressure and constant volume respectively is characteristic of the molecules of a gas, and for this reason was much sought by Kundt and later experimenters on the velocity of sound in gases contained in tubes. Since the advent of convenient sources of high-frequency sound waves (supersonics) in the past twenty years, results have come to light which show that the ratio of specific heats may not be constant at all frequencies of vibration. Thus, Pierce1 in 1925 found that the velocity of sound in carbon dioxide rose at a frequency in the neighbourhood of 100 kc./sec. by about 5 per cent. The effect has since been confirmed for other-mostly triatomic-gases, little change being noted in diatomic gases at ordinary temperatures up to the upper limit of frequency (about 5 megacycles per second) as yet attained.