The wide-spread belief that the electric potential difference (DeltaPhi_m) across membranes with electrogenic proton transport varies nearly linearly with the pH difference (Delta pH) upon build-up of the proton-motive force is critically examined. First, we analyse experimental literature data concerning rat liver and yeast mitochondria, and E. coli. We then present a model describing the interrelation between the Delta pH and DeltaPhi_m as the activities of the proton pumps or H+-ATPase or the influence of the proton leak vary. It is based on the quasi-electroneutrality condition, the dissociation equilibrium of impermeant weak acids, a simple description of the cation-proton antiporters and cation leak, and the Nernst equation applied to all those ions subsisting in equilibrium. The model yields a nonlinear equation giving DeltaPhi_m as a function of Delta pH. In various situations this function is quasi-linear in physiologically relevant ranges of Delta pH. Thus, the linearity hypothesis can be substantiated theoretically, but is not necessarily justified under all circumstances. It is shown that the slope of the DeltaPhi_m vs. Delta pH curve is, in the quasi-linear regions, about -2.303 RT/F (thus having the same value, but the opposite sign as in the Nernst equation) when the cation-proton antiporters are absent or completely inhibited, and can be much higher in absolute value when these antiporters are operative.