The partial molar volumes of 1-heptanol and 1-octanol in red cell ghosts, in egg phosphatidylcholine bilayers, and in water and phosphate buffer have been measured to a precision of better than 4% by using a density meter. In every case, the partial molar volume was independent of concentration in the range studied. In both membranes, the partial molar volume of each alcohol was close to its molar volume whereas in aqueous solution it was considerably less. Comparison of the two membranes suggests that the major contribution to the partial molar volume arises from alcohol-lipid interactions of each case. Further comparison with partial molar volumes in bulk solvents suggests that on average the alcohols retain a hydrogen bond in the lipid bilayer. The magnitude of the volume change in ghosts is some 5 times less than the corresponding area changes previously reported by others [Roth, S. H., & Seeman, P. (1972) Biochim. Biophys. Acta 255, 190-198]. These two observations can only be resolved by assuming either that the bilayer expands anisotropically, experiencing a decrease in thickness with increasing volume, or that conformational changes in membrane-associated proteins can occur at constant volume to increase membrane area. Finally, these data are used to test the critical volume hypothesis of general anesthetic action. A volume change of 0.15% in red cell ghost membranes is found to be associated with anesthesia, which compares with a value of 0.2% predicted previously from pressure reversal of anesthesia studies. In egg phosphatidylcholine bilayers, a volume change of 0.36% is associated with anesthesia. The larger change in the lipid bilayer compared to the biomembrane originates solely in their different membrane/buffer partition coefficients.