1. The mechanism of rabbit muscle pyruvate kinase was investigated by measurements of fluxes, isotope trapping, steady-state velocity and binding of the substrates. All measurements were made at pH8.5 in Tris/HCl buffer and at 5mm-free Mg(2+). 2. Methods of preparing [(32)P]phosphoenolpyruvate from [(32)P]P(i) in high yield and determining [(32)P]-phosphoenolpyruvate and [8-(14)C]ADP are described. 3. The ratio Flux of ATP to ADP/Flux of ATP to phosphoenolpyruvate (measured at equilibrium) increased hyperbolically with ADP concentration from unity to about 2.1 at 2mm-ADP, but was unaffected by phosphoenolpyruvate concentration. Since the ratio is greater than unity, one pathway for the addition of substrates must involve phosphoenolpyruvate adding first to the enzyme in a rate-limiting step. However, the substrates must also add in the alternative order, because of the non-linear increase in the ratio with ADP concentration and because the rate of increase is very much less than that predicted from the steady-state velocity data for an ordered addition. The lack of influence of phosphoenolpyruvate on the ratio is consistent with the rapid addition of ADP in the alternative pathway. At low ADP concentrations the alternative pathway contributes less than 33% to the total reaction. 4. Isotope trapping was observed with [(32)P]phosphoenolpyruvate, confirming that when phosphoenolpyruvate adds first to the enzyme it is in a rate-limiting step. The release of phosphoenolpyruvate from the ternary complex must also be a slow step. Trapping was not observed with [8-(14)C]ADP, hence the addition of ADP to the free enzyme must be rapid unless its dissociation constant is very large (>20mm). 5. Binding studies showed that 4mol of [(32)P]phosphoenolpyruvate binds to 1mol of the enzyme, probably unligated to Mg(2+), with a dissociation constant appropriate to the mechanism indicated above. Binding of [8-(14)C]ADP could not be detected, and hence the binding of ADP occurs by a low-affinity step. The latter is also demanded by the steady-state velocity data. 6. The ratio Flux of phosphoenolpyruvate to ATP/Flux of phosphoenolpyruvate to pyruvate (determined from the incorporation of label into phosphoenolpyruvate from [3-(14)C]-pyruvate or [gamma-(32)P]ATP during the forward reaction) did not differ significantly from unity. Steady-state velocity data predicted grossly different flux ratios for ordered dissociations of the products, and the results indicate that the dissociation must be rapid and random. The data also exclude a Ping-Pong mechanism. 7. Permissible rate constants for the above mechanism are calculated. The results indicate a high degree of cooperativity in binding, whatever the order of addition of substrate.