The hepatic removal of the glutathione conjugate of bromosulfophthalein (BSPGSH) was studied in the single-pass perfused rat liver with the multiple indicator dilution (MID) technique against various background concentrations of BSPGSH (20 to 214 μmol/L) over which nonlinear binding to both plasma (albumin) and tissue proteins with two classes of binding sites was found. A bolus containing 51Cr-labeled red blood cell (a vascular reference), [125I]albumin and [14C]sucrose (large and small molecular weight interstitial references, respectively), D2O (a cellular space reference), and [3H]BSPGSH was injected into the portal vein during steady-state. The eliminated fraction of dose, obtained by subtracting the survival fraction of [3H]BSPGSH in plasma from one, corresponded to the steady state extraction ratio (E) with bulk data, which declined from 0.74 ± 0.04 to 0.27 ± 0.01 with concentration. The major portion of the tracer outflow profile was a throughput component, which is the proportion of tracer that did not enter liver cells during its transit through the liver. The influx, efflux, and sequestration coefficients, evaluated with previously developed barrier-limited models, provided the corresponding influx (k1), efflux (k-1) and excretion (kseq) rate constants. Concentration-dependent influx (Vmax = 83 nmol min−1 g−1 and Km = 3.7 μmol/L), efflux (Vmax = 15 nmol min−1 g−1 and Km = 1.8 μmol/L), and excretion (Vmax = 94 nmol min−1 g−1 and Km = 1.8 μmol/L) were obtained for BSPGSH, when Km values are expressed in terms of the unbound concentrations. In these calculations, the observed unbound tissue concentration was not used for estimation of the Vmax and Km for efflux and excretion because of overestimation, because of the presence of highly concentrated BSPGSH in ductular elements present in liver homogenates; rather, the unbound tissue concentration was calculated from the influx, efflux, and removal rate coefficients. Because of carrier-mediated entry, the unbound tissue concentration does not equal the unbound plasma concentration, and kinetic parameters for BSPGSH excretion could be alternately estimated when the rate of excretion or net rate of loss of BSPGSH from plasma was regressed against the estimated tissue unbound concentration. This yielded a Vmax of 97 nmol min−1 g−1 and a Km of 3.6 μmol/L, values similar to those obtained from MID. Regression of rates of biliary excretion or net plasma disappearance against the logarithmic average unbound concentration, however, yielded a Vmax of 50 nmol min−1 g−1 and a Km of 3.3 to 3.6 μmol/L, values which were, as expected, different from those obtained with MID for either influx, efflux, or removal. It was found that intracellular BSPGSH levels had not increased sufficiently enough to saturate efflux and excretion, and the limiting feature of the transfer process at highbulk concentration of BSPGSH appeared to be the influx process. (HEPATOLOGY 1995; 22:1188–1207.).