It was investigated whether the movement of chloride ions through the osmotic barrier of Ehrlich cells is passive. i.e. driven by a membrane potential, which again is controlled by an active cation pump, or is active. Therefore, the distribution and the unidirectional fluxes of Cl−, along with the distribution of K+, Na+ and water between cells and medium, were measured while the cation pump was active (37°, Krebs‐Henseleit buffer, pH 7.4, 95% O2+ 5% CO2) and, for comparison, while the cation pump was inhibited by gassing with N2 instead of O2 or by adding DNP to the medium. In the presence of 1% bovine serum albumin, DNP altered the electrolyte levels inside the cell only at relatively high concentrations, 0.5–1 mM, then having the same effects as has anoxia. In either case, cells shrank by about 10%, attaining a new steady state which lasted for at least 20 min. K+ and Na+ gradients were than reduced to 2/3 and 1/2 of their normal values respectively. With cells thus shrunken and with control cells, exchange of cell chloride for labelled chloride in the medium reached an equilibrium after about 20 min. In the controls, the exchangeable fraction of total cell chloride, calculated by logarithmic extrapolation, amounted to 60–70% only (Hempling and Kromphardt [1]); the same fraction exchanged with nitrate in the medium. In cells treated with DNP or N2, the exchangeable fraction (for 36Cl) was still smaller: 45–55%. It seemed to be proportional to the water content of the cell. On the other hand, the non exchangeable chloride did not depend on cell water content nor did it change on DNP treatment or anoxia. It stayed fairly constant at 100 μequiv per g dry weight. Total cell chloride, however, increased again in direct proportion to the cell water (which varied spontaneously in different batches at the same medium osmolarity). It is concluded that the concentration of exchangeable chloride with respect to its corresponding cell compartment equals that of total chloride with respect to total cell water. On the basis of this assumption, the distribution ratio (cell: medium) of free exchangeable chloride was calculated to be 0.65 in control cells and 0.57 in cells with inhibited energy supply. Both unidirectional Cl− fluxes decreased in the last case to 30–50% of control; the 42K influx, however, measured simultaneously, was depressed more strongly than the 36Cl influx. This suggests that the inhibition of the cation pump is primary and that of chloride fluxes is secondary, i.e. that the chloride fluxes are passive and that chloride distributes according to the membrane potential. A measured potential, as reported by Hempling [2], agrees with this view. But the distribution of chloride and that of the sum of [Na+] + [K+] do not follow Donnan's law, suggesting that either the passive permeabilities or the pumped fluxes of the two cations must be different. Therefore the cation pump, be it electroneutral or electrogenic, must determine the potential in either case. Furthermore, since reduced energy supply diminishes Cl− influx and efflux while leaving the influx of the small neutral thiourea molecule unchanged, the cation pump must influence the passive permeability for ions specifically.