1. To determine whether discomfort associated with breathing (dyspnoea) is related to the chemical drive to breath, three subjects were totally paralysed while fully conscious. Subjective responses to a rising CO2 stimulus were obtained during rebreathing, rebreathing with CO2 added, and breath holding. Dyspnoea was measured with a 10-point Borg scale. 2. Following nasotracheal intubation and ventilation (oxygen saturation, O2,Sat, 98-100% and end-tidal CO2, PET,CO2, 30-40 mmHg), total neuromuscular blockade was induced by a rapid injection of atracurium (> 2.5 mg kg-1) and complete paralysis was maintained with an infusion (5 mg (kg h)-1). Paralysis was confirmed by abolition of the compound muscle action potentials of both the diaphragm and abductor hallucis evoked by supramaximal electrical stimulation of the relevant nerves. Communication via finger movement was preserved for the first 20-30 min following paralysis by inflation of a sphygmomanometer cuff on one arm. 3. Before and during complete paralysis, dyspnoea increased progressively during hypercapnia produced by rebreathing (with or without CO2 added to the circuit at 250 ml min-1). The mean PET,CO2 eliciting 'severe' dyspnoea was 46 mmHg during rebreathing, 42 mmHg during 'breath holding', and 52 mmHg during rebreathing with added CO2. There were no significant differences between the values obtained during paralysis and in the control study immediately before paralysis. The duration of breath holding was not prolonged by paralysis and the PET,CO2 at the 'break point' was not altered by paralysis. 4. Thus, dyspnoea is preserved following total neuromuscular blockade. This suggests that chemoreceptor activity, via the central neuronal activity which it evokes, can lead to discomfort in the absence of any contraction of respiratory muscles. 5. During paralysis, attempted contraction of arm, leg and trunk muscles increased heart rate and blood pressure. For attempted handgrip contractions, the increases in heart rate (range, 7-15 beats min-1) and mean arterial pressure (range, 20-32 mmHg) were similar to those recorded with actual contractions in trials immediately before paralysis. In one subject, graded increases in heart rate and blood pressure occurred for attempted contractions of 45 s duration over a range of intensities (0-100% maximal effort). 6. During complete paralysis, transcranial electromagnetic stimulation of the motor cortex produced illusory twitch-like movements of the wrist and digits. This also occurred in separate studies during complete ischaemic paralysis and anaesthesia of the forearm and hand.(ABSTRACT TRUNCATED AT 400 WORDS)