Steady-state concentration gradients across cell membranes have often been attributed to the associated leakage of solute down its electrochemical potential gradient, and active transport at an equal rate in the opposite direction. Several workers have evaluated the minimal energetic requirements of such a "pump-leak" model for sodium in muscle tissue, presuming that influx occurs only via the leak pathway and to no extent by way of the active transport pathway. The high energy requirements so predicted have led to the suggestions that either (a) sodium is not actively transported, being at equilibrium distribution across the cell surface, or (b) substantial sodium movement must be by means of exchange diffusion. The present treatment, based on the consideration that the active transport mechanism is bidirectional, demonstrates that the rates of influx and efflux associated with a given rate of active transport are explicit functions of two parameters: (1) the ratio of the exchange resistance of the active pathway to that of the leak pathway, and (2) the electrochemical potential difference across the cell surface. Lacking precise values for these parameters, the demonstration of a high rate of isotope flux is not compelling evidence either against active transport or for a discrete exchange diffusion mechanism. Various concepts and criteria of exchange diffusion are discussed.