1. The effect on ventilation of airway anaesthesia, produced by the inhalation of a 5% bupivacaine aerosol (aerodynamic mass median diameter = 4.77 μm), was studied in 12 normal subjects. 2. The dose and distribution of the aerosol were determined from lung scans after the addition to bupivacaine of 99m Tc. Bupivacaine labelled in this way was deposited primarily in the central airways. The effectiveness and duration of airway anaesthesia were assessed by the absence of the cough reflex to the inhalation of three breaths of a 5% citric acid aerosol. Airway anaesthesia always lasted more than 20 min. 3. Resting ventilation was measured, by respiratory inductance plethysmography, before and after inhalation of saline and bupivacaine aerosols. The ventilatory response to maximal incremental exercise and, separately, to CO 2 inhalation was studied after the inhalation of saline and bupivacaine aerosols. Breathlessness was quantified by using a visual analogue scale (VAS) during a study and by questioning on its completion. 4. At rest, airway anaesthesia had no effect on mean tidal volume ( V T ), inspiratory time ( T i ), expiratory time ( T e ) or end-tidal P co 2 , although the variability of tidal volume was increased. On exercise, slower deeper breathing was produced and breathlessness was reduced. The ventilatory response to CO 2 was increased. 5. The results suggest that stretch receptors in the airways modulate the pattern of breathing in normal man when ventilation is stimulated by exercise; their activation may also be involved in the genesis of the associated breathlessness. 6. A hypothesis in terms of a differential airway/alveolar receptor block, is proposed to explain the exaggerated ventilatory response to CO 2 .