1. The pharmacokinetics of morphine were studied in 26 newborn premature neonates (26-38 weeks gestational age) who were given a loading dose of 50 micrograms kg-1 of diamorphine followed by an intravenous infusion of 15 micrograms kg-1 h-1 of diamorphine. Plasma concentrations of morphine were measured during the infusion at steady-state and for 24 h after the cessation of the diamorphine infusion. 2. The mean steady-state plasma morphine concentration (+/- s.d.) for a diamorphine infusion rate of 15 micrograms kg-1 h-1 was 62.5 +/- 22.8 ng ml-1. 3. Morphine clearance was 3.6 +/- 0.9 ml min-1 kg-1, the elimination half-life was 8.9 +/- 3.3 h and the volume of distribution was 2.7 +/- 1.01 kg-1. 4. Morphine elimination kinetics were described by a mono-exponential function. 5. There was a direct relationship between the gestational age of the patients and the clearance (r2 = 0.31, P = 0.003) and half-life (r2 = 0.35, P = 0.01) of morphine, but no relationship was found between gestational age and volume of distribution. 6. The results suggest that the currently used dosing regimen of diamorphine achieves a safe and effective morphine concentration in the premature newborn but that the loading dose could be modified to achieve a more rapid onset of analgesia.