The respiratory responses to intravenous morphine sulphate (0.12 mg/kg), morphine-6-glucuronide (M6G: 0.03 mg/kg) and placebo were assessed in 6 healthy volunteers, using a single blind randomised crossover design. Five of these subjects underwent an additional study of M6G at 0.06 mg/kg. Respiratory rate, minute volume and end-tidal CO2 were continuously measured using a low resistance non-rebreathing circuit, a mass spectrometer and a dry gas meter. The ventilatory responses to CO2 exposures (5.5% for 4 min) were assessed 40 and 20 min before, and 20, 40 and 80 min after drug administration. Following placebo and M6G (at both doses) no change in end-tidal CO2 occurred whilst the subjects were breathing air, whereas following morphine a significant rise was seen (P less than 0.05). Morphine reduced the ventilatory response to 5.5% CO2 at all times tested (P less than 0.05) and M6G (at both doses) reduced the response to CO2 at 20 and 40 min after administration, but to a lesser degree than did morphine (P less than 0.05).