Many experimenters who have used cold phosphorus for the absorption of oxygen from air must have noticed the appearance of flickering clouds of luminosity when the action is nearly complete. These occur in the gas space and obviously represent a delayed action between the slight oxygen residue and the vapour of phosphorus. This action has been discussed by Joubert, who recognised it as due to the propagation of combustion in an explosive mixture. He obtained it in a more striking form by the slow leakage of air into an exhausted vessel containing phosphorus: his conclusion was that below a certain pressure, too small to measure, phosphorus vapour would not combine with oxygen. As the mixture of phosphorus vapour and oxygen became richer in oxygen by the inflow of the latter, the point was reached when combustion became possible and an explosion was propagated. From the point of view of kinetic theory it seemed very strange that phosphorus vapour should behave thus. Joubert’s view of the facts would seem to imply that the reaction of a molecule of phosphorus with oxygen was not dependant solely on the character and energy of molecular collisions, but also on the absolute value of the interval of time between them. These theoretical difficulties, and also the fascination of the experiments themselves, led me to attempt a further investigation.