Melander's experimental results on the apparent production of salt nuclei by evaporation of salt solutions are examined both theoretically and experimentally. It is concluded theoretically that the rate of evaporative escape of molecular sodium chloride or other salts from sea water is too low by a factor of 1018 to account for the formation of the nuclei active in cloud condensation. It is shown experimentally that the rate of collection of salt particles of diameter in excess of 0.05 micron over evaporating salt solution is too low by a factor of at least 106 to account for the observed rate of atmospheric condensation. Therefore, there appears to be no basis for accepting Wright's suggestion that Melander's work provides an answer to Simpson's long-standing objection to the sea-salt hypothesis of condensation nuclei.