Ventilatory response to graded external dead space (0.5, 1.0, 2.0, and 2.5 liters) with hyperoxia and CO2 steady-state inhalation (3, 5, 7, and 8% CO2 in O2) was studied before and after 4% lidocaine aerosol inhalation in nine healthy males. The mean ventilatory response (delta VE/delta PETCO2, where VE is minute ventilation and PETCO2 is end-tidal PCO2) to graded dead space before airway anesthesia was 10.2 +/- 4.6 (SD) l.min-1.Torr-1, which was significantly greater than the steady-state CO2 response (1.4 +/- 0.6 l.min-1.Torr-1, P less than 0.001). Dead-space loading produced greater oscillation in airway PCO2 than did CO2 gas loading. After airway anesthesia, ventilatory response to graded dead space decreased significantly, to 2.1 +/- 0.6 l.min-1.Torr-1 (P less than 0.01) but was still greater than that to CO2. The response to CO2 did not significantly differ (1.3 +/- 0.5 l.min-1.Torr-1). Tidal volume, mean inspiratory flow, respiratory frequency, inspiratory time, and expiratory time during dead-space breathing were also depressed after airway anesthesia, particularly during large dead-space loading. On the other hand, during CO2 inhalation, these respiratory variables did not significantly differ before and after airway anesthesia. These results suggest that in conscious humans vagal airway receptors play a role in the ventilatory response to graded dead space and control of the breathing pattern during dead-space loading by detecting the oscillation in airway PCO2. These receptors do not appear to contribute to the ventilatory response to inhaled CO2.