[(14)C]Mannitol was administered i.v. as a bolus injection to five postcholecystectomy patients with indwelling T-tubes and re-established enterohepatic circulations to evaluate the biliary clearance of [(14)C]mannitol as a means of estimating canalicular bile flow in man. [(14)C]Mannitol appeared in collections of bile 9-22.5 min after intravenous injection, rose to a peak, and thereafter paralleled the plasma [(14)C]mannitol disappearance curve. Bile-plasma [(14)C]mannitol ratios and [(14)C]mannitol clearances were determined during control and choleretic periods after correction of the bile [(14)C]mannitol points for the transit time of a given sample. After i.v. injection of sodium dehydrocholate in five studies, bile flow and mannitol clearance increased proportionately. However, when ductular secretion was stimulated with an i.v. bolus of secretin in three other studies, [(14)C]mannitol clearance remained essentially unchanged, indicating that [(14)C]mannitol entered bile at the level of the hepatocyte and could be utilized as a marker of canalicular flow in man. During control studies, when bile drained spontaneously from biliary fistulae in fasting patients, bileplasma [(14)C]mannitol ratios averaged 0.62+/-0.18 and canalicular flow, as estimated by [(14)C]mannitol clearance. (0.27+/-0.16 ml/min) accounted for 44-95% of total bile production (0.43+/-0.12 ml/min). When the rate of bile flow was plotted as a function of bile salt excretion after correction for the effects of biliary dead space, linear regression analysis revealed that approximately 7 mul of bile were secreted with each mumol of bile salt. Estimates of bile salt-independent canalicular flow accounted for at least one-third of the estimated 24-h bile production (604 ml) in these patients, indicating that this fraction of canalicular flow is a significant source of bile secretion in man.