I HAVE found1 that surface wave velocities can be conveniently and accurately derived by measuring a critical angle at which, when an ultrasonic beam is reflected from a solid surface in a water bath, there is a violent change in the waveform of the reflected wave train. This occurs when the surface wave velocity and the wave velocity in the water are related by way of Snell's law for an angle of refraction of 90°. Although the indication is accurate and has proved valuable for elasticity measurement, the reason for this behaviour has not been entirely clear.