Absorption spectroscopy with nanosecond time resolution shows that myoglobin undergoes conformational relaxation on the same time scale as geminate rebinding of carbon monoxide. Ligand rebinding following photodissociation of the heme-CO complex was measured from the amplitude of the average difference spectrum, while conformational changes were measured from changes in the detailed shape of the Soret spectra of the deoxyhemes. Experiments in which the solvent viscosity was varied between 1 and 300 cP and the temperature between 268 and 308 K were analyzed by fitting the multiwavelength kinetic data with both empirical and molecular models. Novel numerical techniques were employed in fitting the data, including the use of singular value decomposition to remove the effects of temperature and solvent on the spectra and of a Monte Carlo method to overcome the multiple minimum problem in searching parameter space. The molecular model is the minimal model that incorporates all of the major features of myoglobin kinetics at ambient temperatures, including a fast and slow rebinding conformation and two geminate states for each conformation. The results of fitting the kinetic data with this model indicate that the geminate-rebinding rates for the two conformations differ by at least a factor of 100. The differences between the spectra of the two conformations generated from the fits are similar to the differences between those of the R and T conformations of hemoglobin. In modeling the data, the dependence of the rates on temperature and viscosity was parametrized using a modification of Kramers theory which includes the contributions of both protein and solvent to the friction. The rate of the transition from the fast to the slow rebinding conformation is found to be inversely proportional to the viscosity when the viscosity exceeds about 30 cP and nearly viscosity independent at low viscosity. The viscosity dependence at high viscosities suggests that the two conformations differ by the global displacement of protein atoms on the proximal side of the heme observed by X-ray crystallography. We suggest that the conformational change observed in our experiments corresponds to the final portion of the nonexponential conformational relaxation recently observed by Anfinrud and co-workers, which begins on a picosecond time scale. Furthermore, extrapolation of our data to temperatures near that of the solvent glass transition suggests that this conformational relaxation may very well be the one postulated by Frauenfelder and co-workers to explain the decrease in the rate of geminate rebinding with increasing temperature above 180 K.