1. This study was designed to investigate the role of airway receptors in the responses of a range of inspiratory muscles to airway occlusion. The occlusion had a rapid onset (< 10 ms), lasted 250 ms and produced only a slight impediment to inspiration. 2. Based on analysis of single trials and averaged rectified electromyographic responses (EMG) in six subjects, there was a major inhibition (IR) with an onset at 34 +/- 2 ms and a trough at 65 +/- 2 ms, and an excitation (ER) with a peak at 105 +/- 2 ms. These two responses are reflex given that voluntary reaction times to a tap on the chest wall occurred at latencies longer than the peak of ER. 3. The responses to airway occlusion did not appear in limb muscles which contracted phasically with inspiration. 4. Anaesthesia of the surface receptors of the upper airway did not attenuate the responses to occlusion. Because this procedure does not eliminate the inputs from muscle and deep laryngeal pressure receptors, two subjects were tested when intubated with a cuffed endotracheal tube so that the occlusion was delivered only to structures below the level of the trachea. Responses to airway occlusion were preserved when all upper airway receptors were 'bypassed'. 5. Responses to airway occlusion also remained after prolonged inhalation of nebulized lidocaine (lignocaine) sufficient to block the cough reflex. 6. The receptors mediating the responses to airway occlusion are therefore likely to reside in inspiratory muscles acting on the chest wall. If so, the short-latency inhibition contrasts with the excitatory stretch reflex responses observed in limb muscles.