In recovery from exercise, phosphocreatine resynthesis results in the net generation of protons, while the net efflux of protons restores pH to resting values. Because proton efflux rate declines as pH increases, it appears to have an approximately linear pH-dependence. We set out to examine this in detail using recovery data from human calf muscle. Proton efflux rates were calculated from changes in pH and phosphocreatine concentration, measured by 31P magnetic resonance spectroscopy, after incremental dynamic exercise to exhaustion. Results were collected post hoc into five groups on the basis of end-exercise pH. Proton efflux rates declined approximately exponentially with time. These were rather similar in all groups, even when pH changes were small, so that the apparent rate constant (the ratio of efflux rate to pH change) varied widely. However, all groups showed a consistent pattern of decrease with time; the halftimes of both proton efflux rate and the apparent rate constant were longer at lower pH. At each time-point, proton efflux rates showed a significant pH-dependence [slope 17 (3) mmol x l(-1) x min(-1) x pH unit(-1) at the start of recovery, mean (SEM)], but also a significant intercept at resting pH [16 (3) mmol x l(-1) x min(-1) at the start of recovery]. The intercept and the slope both decreased with time, with halftimes of 0.37 (0.06) and 1.4 (0.4) min, respectively. We conclude that over a wide range of end-exercise pH, net proton efflux during recovery comprises pH-dependent and pH-independent components, both of which decline with time. Comparison with other data in the literature suggests that lactate/proton cotransport can be only a small component of this initial recovery proton efflux.