The effects of chronic alcohol consumption on regional brain glucose metabolism were examined in Balb/c mice using the [14C]2-deoxyglucose autoradiographic technique. Animals were given a solution of 12% v/v ethanol as their only source of fluid for either 6, 12 or 18 months and compared to control groups receiving either an isocaloric solution or saccharose or tap water. Alterations of cerebral brain glucose metabolism were assessed in mice who were returned to a non-alcoholic diet and allowed to freely explore a T-maze. The results showed that chronic ethanol consumption induced reductions of regional metabolic activity which were functions both of the duration of alcohol treatment and of the structure studied. Whereas a six month period of alcoholization did not induce any significant effects on metabolic activity, 12 months of treatment were necessary to induce the first observable and significant reductions in [14C]2-deoxyglucose labelling. These effects were mainly limited to diencephalic structures such as the lateral mammillary nuclei and the anterodorsal thalamic nuclei. The cerebellum was also affected but to a lesser degree. After 18 months of alcoholization, a generalized spread of the metabolic reduction to the entire mammillary complex (lateral, medial and posterior nuclei) and to the thalamic nuclei was observed. This same duration of treatment was necessary to induce the first detectable decrease of metabolic activity in the hippocampus. In agreement with data from human neuropathology, these findings confirm the particular vulnerability of diencephalic structures to ethanol and suggest that damage limited to diencephalic regions rather than to hippocampal or cortical areas could be primarily responsible for the memory disorders observed in Korsakoff's syndrome.