1. The mechanisms responsible for the depolarization of the hepatocytes secondary to anoxia have been studied in isolated perfused dog liver. It was attempted to elucidate the role of the inhibition of the sodium pump following exhaustion of the energy reserves and of the modifications of membrane permeability. Anoxia was compared to ouabain and to a reduction of the cellular ATP level. 2. Membrane potentials were measured with micro-electrodes. Potassium, sodium and chloride were determined in plasma samples and liver tissues. Extracellular space was measured with tritiated inulin or with an electrical impedance method. Adenine nucleotides were also measured in liver biopsies. 3. The fall in membrane potential produced by administration of ouabain (0-1 mM) is greater than the effect of the redistribution of sodium + potassium ions; this suggests that the sodium pump is functioning, at least partially, electrogenically. The administration of dinitrophenol (10 mM), which causes a 74% fall in the ATP level in 15 min, produces, as does ouabain, a depolarization which also corresponds to stopping an electrogenic pump. 4. A partial reduction in the level of ATP brought about by hypoxia, by an inhibitor of cellular respiration, antimycin (10 mM), or by fructose (20 mM) results in a hyperpolarization which may be attributed to an elevation of potassium permeability (PK) since it is concomitant to a loss of K from the liver. The change in membrane permeability could be related to a rise in the free calcium in the cells which has not been documented. Other possible hypothesis include a facilitated transport for potassium. 5. The administration of amobarbitone (10 mM) produces immediately a depolarization which is independent of the progressive reduction in the level of ATP. The depolarization has been attributed to a direct effect of amobarbitone on the membrane reducing the permeability for potassium ions. 6. The depolarization observed in ischaemic anoxia is greater than that produced by ouabain for the same variation in ions concentration. In addition to a likely inhibition of the electrogenic sodium pump, changes in membrane permeability inducing a rise in the PNa/PK ratio must also occur. 7. After ischaemic anoxia for 24 hr at 3 degrees C, the ratio of PNa/PK rises to 0-68 which indicates abolishment of the selective character of membrane permeability. The augmentation in cell volume produced by anoxia might result in an opening of membrane pores, which could entail the augmentation of sodium permeability; the latter would be responsible in part for the depolarization produced by anoxia. 8. According to the severity and length of oxygen deprivation an increase in PK, a cessation of the sodium pump activity and finally an increase in PNa will occur.