The kinetic behaviour of [3H]DOPA in living rat brain was investigated by compartmental modelling of measured activities from combined metabolite pools in a time-series (180 min) of static autoradiograms from right cerebral hemispheres. Two models of [3H]DOPA uptake and metabolism that incorporated the removal of the decarboxylation product, [3H]dopamine, from brain were significantly more accurate than a model in which [3H]dopamine accumulated irreversibly in situ. Present estimates of [3H]DOPA kinetic constants were compared to previously published results based on the analysis of measured activities from individual metabolite pools separated by chromatographic fractionation of [3H]DOPA metabolites in the left cerebral hemispheres of the same rats. Autoradiographic estimates of DOPA decarboxylase activity with respect to [3H]DOPA in brain (k3DOPA) were under-estimated several-fold relative to chromatographic estimates; this discrepancy is explained by post-mortem enzyme activity and omission of biological compartments from the models. However, autoradiographic estimates of the unidirectional blood-brain clearance of [3H]DOPA (K1DOPA) and monoamine oxidase activity with respect to [3H]dopamine in brain (k7DA') agreed with chromatographic estimates. This concordance represents the first empirical validation of compartmental modelling of autoradiographic data as a method for quantitatively investigating the kinetic behaviour of radiolabelled L-DOPA in living mammalian brain.