Hyperoxia and hyperbaric hyperoxia increased the rate of cerebral hydrogen peroxide (H2O2) production in unanesthetized rats in vivo, as measured by the H2O2-mediated inactivation of endogenous catalase activity following injection of 3-amino-1,2,4-triazole. Brain catalase activity in rats breathing air (0.2 ATA O2) decreased to 75, 61, and 40% of controls due to endogenous H2O2 production at 30, 60, and 120 min, respectively, after intraperitoneal injection of 3-amino-1,2,4-triazole. The rate of catalase inactivation increased linearly in rats exposed to 0.6 ATA O2 (3 ATA air), 1.0 ATA O2 (normobaric 100% O2) and 3.0 ATA O2 (3 ATA 100% O2) compared with 0.2 ATA O2 (room air). Catalase inactivation was prevented by pretreatment of rats with ethanol (4 g/kg), a competitive substrate for the reactive catalase-H2O2 intermediate, compound I. This confirmed that catalase inactivation by 3-amino-1,2,4-triazole was due to formation of the catalase-H2O2 intermediate, compound I. The linear rate of catalase inactivation allows estimates of the average steady-state H2O2 concentration within brain peroxisomes to be calculated from the formula: [H2O2] = 6.6 pM + 5.6 ATA-1 X pM X [O2], where [O2] is the concentration of oxygen in ATA that the rats are breathing. Thus the H2O2 concentration in brains of rats exposed to room air is calculated to be about 7.7 pM, rises 60% when O2 tension is increased to 100% O2, and increases 300% at 3 ATA 100% O2, where symptoms of central nervous system toxicity first become apparent. These studies support the concept that H2O2 is an important mediator of O2-induced injury to the central nervous system.