We have measured the time course of release of 42K and 86Rb from an occluded state of the Na,K-pump using a rapid filtration apparatus. We have found that at 20 degrees C and in the presence of ATP, 42K is released with a rate constant of approximately 45 s-1 and 86Rb with a rate constant of approximately 20 s-1; both ATP and ADP are effective at a low affinity site (Kd approximately 0.3 and 1 mM, respectively) with the rate of deocclusion being only half as great in ADP as in ATP. Mg2+ stimulates 2-fold at low concentrations probably by forming MgATP, and free Mg2+ is strongly inhibitory at high concentrations (Kd approximately 10 mM). Mg2+ also decreases the affinity for ATP, and the data are consistent with mixed type inhibition; from the analysis the dissociation constant is approximately 1 mM for the inhibitory Mg2+ and the Rb+-occluded form without ATP. The rate of 42K or 86Rb release increases monotonically with pH while ATPase activity decreases above pH 8, so that deocclusion is not rate-limiting in the overall cycle at high pH. This is reflected by a convergence of the rate of Na,K-ATPase and Na,Rb-ATPase activities at high pH and by a decrease in the observed steady-state level of the occluded 86Rb intermediate at high pH. K+, Rb+, Na+, and Cs+, but not Li+, increase the rate of 42K and 86Rb release at constant ionic strength, presumably at sites other than the transport sites. The spontaneous rate of deocclusion is only approximately 0.1 s-1 at low ionic strength in the absence of nucleotides, and it is increased markedly by all cations tested except Li+. Overall the data are consistent with deocclusion as a rate-limiting step in the Na,K-pump cycle.