The rate of release of transmitter quanta, elicited by variable depolarization pulses applied to a nerve terminal by means of a macro-patch-clamp electrode, was measured in muscles of crayfish and frogs. The electrode was perfused with solutions containing different Ca concentrations, Cae. The bath was superfused separately, usually with solutions containing nominally no Cab and elevated Mgb. A fixed depolarization pulse followed the variable test pulse within 7-10 ms, and facilitation, Fc, of release after the fixed pulse was determined as a measure of Ca-inflow during the test pulse. As described before, Fc always showed a peak, Fc, at depolarization amplitudes of the test pulse below the saturation level of release. When Cae was changed, the depolarization levels generating Fc shifted in a negative direction if Cae was lowered, and in a positive direction if Cae was increased. These shifts agreed with the known dependence of the effective membrane potential (controlling e.g. Ca inward current) on Cae which is due to shielding of surface changes by Ca2+ (cf. Hille 1984). Changes of Cab, at constant Cae, did not affect the depolarization dependence of Fc. It is concluded that Ca inflow is not the only factor controlling quantal release, and that at least in depolarizations beyond those eliciting Fc another potential dependent factor increases release while Ca inflow presumably falls.