Atmospheric sodium distributions derived according to the recent Liu and Reid model for the sodium photochemistry, which includes reactions leading to the formation of NaOH, are compared with lidar measurements, and good agreement is found. Attempts to reproduce the observed vertical distribution of atmospheric sodium by photochemical models are hampered by a lack of precise knowledge of the appropriate reaction rates. It is easy to show, however, that on the basis of the usual Chapman photochemistry only a most unreasonable choice of rate coefficients would lead to the observed vertical distribution, even when a source is included in the model. Models of this sort, apparently compatible with the observations, have invariably contained the implicit assumption of an unexplained sink, unreasonable distributions of the reacting species, or have not been in equilibrium. A number of these models also suffer from the defect of leading to values of the sodium airglow emission much smaller than that observed. It is concluded that, unlike previous models, the Liu and Reid model for the sodium photochemistry is capable of explaining both the observed vertical distribution of sodium, and the intensity of the sodium nightglow.