THE underlying assumption implicit in work concerned with the propagation of stress waves in liquids is that any pressure disturbance originating at some point in the liquid travels unchanged in form with a velocity c = √(dp/dρ), which is independent of the density. Such a wave is an ‘acoustic’ wave, and it is well known that this assumption is amply justified for the small variations of density and pressure usually developed by sound sources under water. But is it justified when the pressure at a point in a liquid rises suddenly by 200 atmospheres (say) due to the passage of a transient stress pulse ? The purpose of this communication is to consider this question on the basis of the type of conditions prevailing in previous experiments described by me and other workers1–3.