A brief review of the condensation theory of Becker & Doring (1935) and Frenkel (1946) is given. It is shown that the theory is subject to the limitation that it assumes that small molecular complexes behave like liquid droplets. This is not the case for complexes consisting of a few molecules only, so that the rate of condensation expected from the theory may be altered by an appreciable amount. The Becker & Doring expression for critical supersaturation is compared with the cloud-chamber measurements of Volmer & Flood (1934), taking into account the rapid heating of the gas after the fast expansion. It is found to be in approximate agreement for water and some of the alcohols, but methyl alcohol, benzene and carbon tetrachloride show large discrepancies. More measurements of critical supersaturations are needed.