Forty-three ASA Grade I patients scheduled for elective abdominal surgery received at random either 25% or 50% inspired oxygen for intra-operative mechanical ventilation lasting 4-6 h. Pulse oximetry was monitored continuously. Venous admixture was assessed from the PaO2/FIO2 ratio and was measured twice intraoperatively: at the time of incision, and during surgical wound closure. PaO2 was measured 1 h after extubation having breathed room air for 10 min, if tolerated. The patients in the two groups were similar in regard to general characteristics, and had similar operations. Patients given oxygen 50% had operations that lasted longer, which made the trial more sensitive. The inspired oxygen did not affect pulmonary gas exchange either within each group or between groups under the conditions of the study. In no patient did pulse oximetry record an oxygen saturation below 95% intra-operatively.