1. To determine if negative upper airway pressure causes reflex pharyngeal dilator muscle activation, we used intra-oral bipolar surface electrodes to record genioglossus electromyogram (EMG) activity in response to 500 ms duration pressure stimuli of 0, -2.5, -5, -15, -25 and -35 cm H2O (0-90% rise time less than 30 ms) in ten normal, conscious, supine subjects. 2. With the subjects relaxed at end-expiration, stimuli were applied in each of three conditions: (i) glottis open (GO), (ii) glottis closed (GC) and (iii) controls with the mouth and nose closed. 3. Six rectified and integrated EMG responses were bin averaged for each pressure in each experimental condition. Response latency was defined as the time when the EMG activity significantly increased above pre-stimulus levels. Response magnitude was quantified as the ratio of the EMG activity for 80 ms post-stimulus to 80 ms prestimulus; data from after the subject's voluntary reaction time (for tongue protrusion) were not analysed. 4. Negative airway pressure activated the genioglossus. The median latency of activation (34 ms) was much faster than the time for voluntary activation (184 ms) indicating a reflex response. 5. Significant activation, compared to 0 cmH2O controls and controls with mouth and nose closed, occurred with pressures of at least -5 cm H2O (GC) and -15 cm H2O (GO). At -25 and -35 cm H2O, responses with GO were significantly greater than with GC. 6. The magnitude ('strength') of the responses differed between subjects; these differences were repeatable. 7. We conclude that negative airway pressure causes reflex pharyngeal dilator muscle activation in man. Responses with GC suggest that upper airway receptors can mediate the response but larger responses with GO indicate a contribution from subglottal receptors.