Ever since Kohlrausch (1922) measured dark-adaptation by plotting the logarithm of the threshold against the time in the dark, it has been clear that the recovery of visual sensitivity occurs in two stages. For, in general, the curve plotted exhibits two quasi-independent branches which meet at a pretty sharp kink. As Kohlrausch demonstrated, the early and faster process is associated with day vision: the later process with night vision. An enormous amount of subsequent work has substantiated this interpretation, but we are on much less secure grounds if we conclude (as is often done) that the first branch is concerned with cones only, the second with rods only and that there is no rod-cone interaction.The 'cone branch' is obtainable from the rod-free fovea; it has the spectral sensitivity of daylight vision, monochromatic test flashes are usually seen as coloured, acuity is good, and a large Stiles-Crawford effect is found. This is strong evidence that cones are involved, but weak evidence that rods are not involved. To be sure, they are not involved at the central fovea, but the cone branch there is not quite the same shape as that found at the parafovea, where rods are anatomically present (Hecht, Haig & Wald, 1935;Sloan, 1950).The 'rod branch' is obtainable from all parts of the retina where rods are present and is absent from the only places (fovea and blind spot) where they are not. This branch exhibits the spectral sensitivity of twilight vision which (after correction for absorption by the ocular media) corresponds very closely to the absorption spectrum of rhodopsin (Crescitelli & Dartnall, 1953), a photosensitive pigment which can be seen contained in the living rods (Boll, 1876). Test flashes appear colourless whatever their wave-length, acuity is poor and the Stiles-Crawford effect is small or absent. This is strong proof that rods are involved, but evidence is not so clear whether cones contribute or not. If rhodopsin is the only pigment which absorbs quanta at threshold throughout the second branch, then after any * Present address: Physiological Laboratory, University of Cambridge.