A sheep preparation was used to examine the regional formation and extraction of morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G), relative to the regional extraction of morphine, at four morphine dose rates. On separate occasions, four ewes received a 15-min loading infusion of morphine sulfate, followed by a constant infusion at 2.5, 5, 10, or 20 mg/hr for an additional 5.75 hr. During the Sth to 6th hr of infusion, blood samples were collected simultaneously from the aorta, pulmonary artery, hepatic vein, hepatic portal and renal veins, posterior vena cava, and coronary and sagittal sinuses. Urine was collected for 48 hr. Morphine, M3G, and M6G in plasma and urine were determined simultaneously by HPLC. The blood/plasma concentration ratio (lambda) for morphine, M3G, and M6G was determined in spiked ''blank'' blood. Steady-state plasma concentrations were achieved during the sampling period, and dose-normalized. concentrations were independent of the infusion rate. There was significant (p < 0.05) extraction (mean +/- SD) of morphine by the liver (0.676 +/- 0.014) and kidney(0.602 +/- 0.039), not extraction of M3G (0.106 +/- 0.046) and M6G (0.104 +/- 0.030) by the kidney, and net formation of M3G (-0.057 +/- 0.017) by the gut. The mean lambda for morphine, M3G, and M6G was 1.25 +/- 0.17, 0.80 +/- 0.03, and 0.82 +/-0.09, respectively. The mean total body clearance of morphine with respect to blood was 1.58 +/- 0.27 liters/min. Mean (+/-SD) percentage urinary recoveries as morphine, M3G, and M6G were 14.7 +/- 8.5, 75.4 +/- 11.1, and 0.49 +/- 0.39, respectively. Data indicate that the liver and kidney are the major organs for the elimination of morphine and that the kidney is the chief site for the elimination of the glucuronide metabolites. Furthermore, the lack of mass balance between the sum of morphine, M3G, and M6G in blood entering and leaving the liver, together with the net formation of M3G by the gut, suggests enterohepatic cycling of morphine.