The behaviour of air saturated with aqueous vapour and allowed to expand suddenly, has been investigated by Coulier, Aitken, Kiessling, and R. v. Helmholtz. As is well known, if the moist air has been previously freed from “dust,” no condensation takes place except on the walls of the vessel, even if the expansion be sufficient to produce considerable supersaturation. For convenience, the term “dust” is here used to include all nuclei which can be removed either by filtering or by repeatedly forming a cloud by expansion and allowing it to settle. What is the limit, if such exists, to the degree of supersaturation which can be attained without condensation taking place throughout the moist air, is a question of considerable meteorological as well as purely physical interest. It was primarily with the object of finding an answer to this question that the experiments to be described were undertaken, such experimental evidence as already existed on the subject being of a very incomplete and contradictory character.