Laboratory measurements of sound velocity in sea water samples from the Caribbean and Middle Atlantic have been made by means of a three‐megacycle ultrasonic interferometer. The results at various salinities and temperatures show that the velocity values in current use (Kuwahara's tables) are too low by approximately three meters/second.In addition, variations in the dissolved air content were found to have negligible effect on sound velocity and absorption. The sulfate/chloride ratio was determined as quite constant in a given region of the sea. Finally, it was established that the velocity and adiabatic compressibility observed for sea water are explainable on the basis of a linear summation of the effects of the individual dissolved salts.