analyzing gaseous substrates. A compartmental model was found to describe the wash-in data adequately. This type of modeling assumes that the vascular and cellular compartments in the liver are well stirred and is a much simpler interpretation of liver physiology than the distributed and dispersion models that have been used to analyze multiple indicator-dilution experiments. Carbon monoxide disposition in the liver has not previously been reported; therefore we wished to explore the modeling of wash-in experiments by using a substrate whose disposition has been elucidated in the liver by means of other techniques. In this study, wash-in experiments were used to examine the disposition of D-glucose in the perfused rat liver. Several types of physiological models were used to analyze the data, and results were compared with those reported for other methods. The disposition of substrates in the liver has been studied extensively with the multiple indicator-dilution technique. This involves the injection of a bolus of indicators into the inflow of an organ and an analysis of outflow time-activity curves. The physiological modeling of this technique is well described and include the distributed models developed by Goresky and co-workers [1-3], the dispersion model [4, 5], and the Crone-Renkin early extraction model [6, 7]. By contrast, wash-in experiments have been studied infrequently, and this technique is less well characterized. Recently, we reported the use of wash-in experiments to determine the behavior of carbon monoxide in the liver [8], hind limb [9], and placenta [10]. Wash-in experiments were chosen to simplify the technical difficulties inherent in collecting and Abstract: Wash-in experiments, which may be useful for the study of the disposition of sub-strates in the liver, have not been well described. To investigate physiological models for wash-in curves, we performed experiments on the per-fused livers of male Wistar rats anesthetized with pentobarbital. Test perfusate contained 14 C-sucrose as the extracellular marker and 3 H-glucose. Liver perfusions were performed with background glucose concentrations of 5.7, 10.7, 51.2, or 108.5 mM. Outflow time-activity curves were analyzed with the use of four models. The V max and K m for the influx of glucose were 1.1 ± 0.03 µmol/s/g and 41 ± 3 mM with the Crone-Renkin early extraction model; 1.4 ± 0.04 µmol/s/g and 36 ± 3 mM with dispersion model analysis; 1.8 ± 0.1 µmol/s/g and 25 ± 4 mM with the Goresky distributed model to fit differentiated wash-in curves; and 2.7 ± 0.6 µmol/s/g and 28 ± 21 mM with compartmental analysis. There was reasonable agreement between the four models, and they yielded results similar to those reported for glucose uptake in other preparations. [The Japanese Journal of Physiology 54: 421-429, 2004]