It seems now experimentally certain that the absorption of sound waves in water, benzene, carbon tetrachloride and methyl alcohol is strictly proportional to the square of the frequency up to 50 megacycles, but considerably larger than can be explained by viscosity and heat conduction. It is shown here that an explanation of this increased absorption by slowness of energy exchange between internal and external degrees of freedom is not in disagreement with the facts for the three last named liquids, while the absorption in water must have a different origin.