An analysis is made of errors which occur due to diffusion in measurements of the drift velocity of electrons in gases using the methods of Bradbury and Nielsen, Pack and Phelps, and Hornbeck. This analysis includes an account of the influences of back diffusion and the effect of the boundary condition imposed by the receiving electrode. It is shown that, in the case of the method of Bradbury and Nielsen, effects due to diffusion can lead to spuriously high measurements of the drift velocity intro ducing a relative error of the order of 3/h(W/D) (h being the distance between the shutters and WID the ratio of the drift speed to the diffusion coefficient).