During oxidation of a pure metal the transfer of atoms from metal to oxide lattice may involve the generation of a counterflow of vacancies into the substrate. Such a process, known as vacancy injection, has received widespread attention in recent years. In this paper, it is argued that much of the evidence quoted in support of injection is capable of alternative explanation. In particular, it is suggested that the growth of intergranular voids in metal substrates during oxidation is not due to vacancy injection from the interface but is a consequence of the induced tensile stress causing void growth by the Hull-Rimmer mechanism. It is also proposed that changes in the rate of vacancy loop growth in electron microscope foils upon changing the annealing atmosphere is due to alterations in the level of induced stress. If vacancies are not injected then they must be destroyed at the interface and this may seem difficult to reconcile with the fact that the interfaces of inert, non growing, oxides are not normally good sinks for metal vacancies. It appears that in order for vacancies to be created or destroyed at a two phase interface, atomic movements in each phase is required; it may be for this reason that the diffusional properties of an oxide-metal interface changes so radically when the oxidation process itself has ceased to occur.