1. To determine the afferent pathways mediating pharyngeal dilator muscle activation in response to negative airway pressure in man, we recorded genioglossus electromyogram (EMG) activity (via intra-oral bipolar surface electrodes) in response to 500 ms duration pressure stimuli of -15 and -25 cm H2O in normal, conscious, supine subjects relaxed at end-expiration; responses were compared before and after upper airway anaesthesia. 2. Six rectified and integrated EMG responses were bin averaged for pressure stimuli applied with the glottis open (GO) and closed (GC) and to the outside of the face only (controls). Response magnitude was quantified as the ratio of the EMG activity for an 80 ms post-stimulus period (before the subject's reaction time for tongue protrusion) to an 80 ms pre-stimulus period. 3. In eight subjects, upper airway anaesthesia reduced the EMG responses with GC to a level indistinguishable from controls. After anaesthesia, responses with GO remained higher than those with GC. 4. With GC, the mean EMG responses decreased by 43% after selective anaesthesia of the nasal mucosa (trigeminal nerves) in two subjects, 32% after selective anaesthesia of the laryngeal mucosa (superior laryngeal nerves) in six subjects and by 21% after selective anaesthesia of the oropharyngeal mucosa (glossopharyngeal and lingual nerves) in four subjects. 5. We conclude that upper airway afferents mediate pharyngeal dilator muscle activation in response to negative pressure with GC and that subglottal receptors caused the increased activation with GO. With GC, the trigeminal and superior laryngeal nerves mediate an important component of the responses with the glossopharyngeal nerves playing a less important role.