The glucuronidation of morphine was investigated in 10 premature neonates (postnatal age < 24 h at initiation of treatment) following 24 h of therapy (2 h loading infusion, followed by a constant rate infusion). Morphine, morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G) were measured in plasma obtained at 24 h in all babies. Plasma concentrations of M3G and M6G correlated significantly with morphine concentration (P < 0.01 in both cases), and with each other (P < 0.001), suggesting that the capacity for morphine glucuronidation in premature neonates is not saturated at the infusion rates used in this study. M3G/morphine and M6G/morphine plasma concentration ratios were independent of morphine infusion rates (P > 0.05) and morphine plasma concentrations (P > 0.05), providing further evidence of linear kinetics. However, M3G/morphine and M6G/morphine plasma concentration ratios increased significantly with increasing birth weight (P < 0.05 in both cases). This probably reflects increase in liver weight with increasing birth weight. Although morphine glucuronidation is deficient in premature neonates, significant concentrations of the respiratory stimulant M3G are achieved rapidly (20% of morphine plasma concentrations at 2 h). At this time, the respiratory depressant M6G could not be detected.