The quantitative autoradiographic 2-[14C]deoxyglucose method was used to measure the effects of a continuous infusion of the respiratory stimulants, caffeine or doxapram, 18 mg/kg per h, on local cerebral glucose utilization in the adult male rat. Local cerebral glucose utilization was measured in 54 cerebral structures from different systems. Caffeine induced widespread increases in energy metabolism, resulting in a significant increase in glucose utilization in 25 structures out of the 54 studied. These increases were distributed within all systems studied, sensory, extrapyramidal motor, limbic and hypothalamic systems. in addition, caffeine induced a non-significant, 10-15%, increase in local cerebral glucose utilization in central respiratory areas. Doxapram infusion did not change the rates of glucose utilization in any of the structures. The rates of local cerebral glucose utilization were significantly lower after doxapram than after caffeine exposure in five cerebral areas, among which were three central respiratory areas. The results confirm the absence of side-effects of doxapram as compared to caffeine when used as respiratory stimulant, especially in neonates. These results also favor a preferentially central action of caffeine on respiratory areas and a more peripheral action of doxapram on chemoreceptors, at least at therapeutic levels.