The kinetics of carbon monoxide (CO) recombination in H93G myoglobin has been studied as a function of temperature in aqueous buffer solution and in 75% glycerol/buffer solutions. H93G adducts with imidazole (1m), 4-methyl imidazole (4-Me Im), 1-methyl imidazole (1-Me Im), and 4-bromoimidazole (4-Br Im) were analyzed by means of three-state and four-state sequential models. Rate constants for the inner process of CO forming a bond with iron and for CO ligand escape into the solvent were extracted from the analysis. At 3 10 K, where the largest differences are observed in buffer solution data, the rate constant for CO recombination to the heme iron differs by a factor of approximate to15 for the various H93G proximal adducts, while the rate constant for CO escape differs by a factor of only approximate to1.4. Thermodynamic analysis based on microscopic reversibility shows that the effect of proximal adducts on equilibrium constant for the intrinsic CO binding to heme iron, K-CO, is 15 times larger than the effect on the equilibrium constant for ligand escape, K-escape. Arrhenius plots of the CO recombination rate constant yielded an average activation enthalpy of approximate to17 kJ/mol for the rebinding rate constant for the CO recombination process in three of the adducts: Im, I-Me Im, and 4-Br Im. A Landau-Zener model of the rate constant for the CO recombination process is introduced to account for differences observed in the prefactor of the rate constants. Combination of this analysis with potential energy surfaces calculated using the BLYP functional in a density functional theory approach with a large basis set indicates that the barrier to rebinding should be the same for the adducts studied in the absence of interaction with the protein. The origin of differences in activation enthalpy or entropy for substituted imidazole adducts of H93G therefore should arise from steric interactions of the proximal ligand with the protein.