Three healthy young males were maintained for sessions of about 1 h in a state of mild asphyxia (Pa,O2 [arterial O2 tension] .apprx. 55, Pa,CO2 [arterial CO2 tension] .apprx. 45 torr), i.e., with moderately strong drives from both arterial and intracranial chemoreceptors. Tidal volume (VT), breath duration (VT) and duration of inspiration (TI) wre recorded, and ventilation (VE) and duration of expiration (TE) were derived breath by breath. The arterial chemoreceptor component of the drive was briefly and abruptly reduced, perhaps silenced, by 3 separate procedures: the inspiratory pathway was connected for 2 breaths to a 2nd gas supply line containing, B, hypoxia with PI,CO2 [inspiratory CO2 tension] zero (removal of hypercapnia with maintained hypoxia); C, pure O2 (removal of asphyxia); and D, O2 with 40 torr added PCO2 (removal of hypoxia with maintained hypercapnia). In controls, A, the 2nd inspiratory line contained the maintenance mixture so that the switch involved no change of inspiratory gas composition. Each type of test was repeated 24 times on each subject. Responses attributable to silencing of arterial chemoreceptors (i.e., with 1 1/2-3 breath latencies, about equal to the lung-to-ear circulation time) are reported elsewhere. Very small responses, occurring only half a respiratory cycle after 1st inhalation of the test mixture, were detected by pooling all responses of each kind from all subjects. When hypoxia was withdrawn, with (C) or without (D) simultaneous withdrawal of hypercapnia, VT and .ovrhdot.VE were reduced by 3 and 2%, respectively, probably because gas mixtures containing high O2 concentrations are appreciably more viscous than hypoxic mixtures and so require more effort to breathe in and out. When hypercapnia was withdrawn with (C) or without (B) simultaneous withdrawal of hypoxia, TE was significantly lengthened (mean, +65 .+-. 18 ms). The change of TE was discussed in relation to known effects of CO2 on airway receptors in the dog.