The initial velocity of adenine nucleotide exchange catalyzed by the reconstituted ADP-ATP carrier from beef heart mitochondria was measured under the influence of membrane potential and with different nucleotide distributions between the internal liposomal and the external buffer volume. Both Vmax and Km of adenine nucleotide uptake not only changed due to the applied potential but also depended on the respective nucleotide distribution. The rate equations for the ADP-ATP exchange under the various conditions were derived. These equations were simplified by assuming two alternative situations; either (a) af affinity type model, where the membrane potential influences only the affinity of the adenine nucleotide carrier toward ATP and ADP, or (b) a velocity type or distribution model, where the membrane potential modulates the rate constants of the ADP-ATP exchange. On the basis of several simplifications in the reconstituted system, the rate equations could be solved and the rate constants and dissociation constants of the exchange in the "energized" and in the "deenergized" state could be calculated. These values were used to derive prediction tables for normalized exchange rates under different nucleotide distributions, which were then compared with the experimental data. Only the exchange rates predicted by the velocity-type model agreed with the measured values. On the basis of this model a definite asymmetry caused by the membrane potential could be seen. Whereas this asymmetry is not very pronounced in the case of carrier-ADP complexes, about 40 times more ATP-loaded binding sites face the outside of the vesicles in the energized state.