The effects of cholesterol, 4,4dimethylcho-lesterol, and lanosterol (4,4',14a-trimethyl-A8'4-cholestadiene-3,-ol) on some properties of lecithin vesicles have been compared. Unlike cholesterol, lanosterol retards the exit of trapped glucose from phospholipid vesicles only slightly. The 13C nuclear magnetic resonance spectrum of cholesterol/lecithin vesicles shows no resonances attributable to the sterol. By contrast, several resonances attributable to quaternary carbon atoms or methyl groups are seen in the 13C nuclear magnetic resonance spectrum of lanosterol/lecithin vesicles, indicating that lanosterol is much-less immobilized than cholesterol. Because the membrane behavior of 4,4-dimethylcholesterol is closely similar to that of cholesterol, it is concluded that the axial 14-a-methyl group is responsible for the lessened membrane immobilization of lanosterol. The results emphasize the importance of a planar sterol a-face for interaction with phospholipid acyl chains. A hypothesis on the evolution of sterols has recently been presented (1). Because the prebiotic atmosphere is assumed to have been essentially anaerobic and because oxygen is an obligatory electron acceptor in the contemporary biosynthesis of sterols, it was suggested that any chemical evolution of the sterol pathway, if it did indeed occur, must have stopped at the stage of squalene. Only when the terrestrial atmosphere turned aerobic and after the arrival of aerobic cells could squalene be oxidized to squalene oxide and then cyclized to lanosterol. However, intracellular lanosterol is rapidly metabolized to cholesterol, and cells that terminate sterol biosynthesis at the lanosterol stage are not known. We suggest that the apparently compulsory metabolic conversion of lanosterol to cholesterol can be explained on the basis of current models (1-6) for cholesterol-containing membranes. According to these models the planar a-face of cholesterol and related cholestane derivatives interacts hydrophobically with the phospholipid fatty acyl chains in the membrane bilayer. As space-filling models show, the axial methyl group at C14 of lanosterol obstructs these interactions (1). On the other hand, according to such models, the two methyl substituents at C4 of lanosterol do not interfere sterically with sterol-fatty acyl chain contacts. We provide here some evidence supporting the above hypothesis. Two types of experiments were performed. The effect of sterols on artificial membranes was examined by determining the exit of trapped glucose and independently by 13C nuclear magnetic resonance (NMR) spectroscopy. EXPERIMENTAL Phosphatidylcholine (PtdCho) was purified from egg yolk by a slight modification of the procedure of Singleton et al. (7). Cholesterol (Sigma) was dried at 560 at 0.5 mm Hg before use.[1-14C]Glucose (specific activity, 4.86 mCi/mmol) was from New England Nuclear. Lanosterol was purified according to the procedure of Bloch and Urech (8). 4,4-Dimethylcholesterol was prepared as described by Woodward et al. (9). All sterols were ...