The chemical composition and the function of intramembranous particles (IMP) of erythrocyte membranes were examined. Firstly, the sulfhydryl (SH) groups detected by the Fast Blue BBN method on ultrathin sections of rabbit erythrocyte ghosts were found to show a correlative distribution with IMP. Secondly, after treating human erythrocyte ghosts with mercuric chloride, IMP appeared to have become larger than the controls. The mean diameter of IMP was approximately 165 A on the P face and 160 A on the E face for the ghosts treated with mercuric chloride, whereas it was 94 A on the P face and 100 A on the E face for fresh ghosts.This result indicates that protein molecules which have SH groups are included in IMP of erythrocyte membrane. Thirdly, acetylcholinesterase (AchEase) activity was detected cytochemically in human erythrocyte ghosts from which IMPfree vesicles were blebbing.AchEase activity was found to be localized only on the external surface of IMP-aggregated ghosts and not on the IMP-free vesicles, suggesting a close relationship of AchEase and IMP. A cytochemical method for detecting AchEase activity on freeze-etched replicas was also described.The present results indicate that IMP in erythrocyte membranes are proteinaceous in nature and that some of them are enzyme molecules.The intramembranous particles (IMP) observed on freeze-cleaved plasma membranes are generally believed to represent protein molecules (19,40). One of the most conclusive results that support this hypothesis is given in the experiments with reconstituted membranes, that is, IMP appear on artificial lipid membranes only after definite proteins are introduced, such as rhodopsin (13), cytochrome oxidase (20) and so on. The results obtained hitherto, including the reconstitution experiments, have shown that protein molecules could contribute to IMP. But no answer has been given to the following questions: 1) Do all IMP consist of protein molecules? Isn't there any heterogeneity among IMP? 2) What kind of roles are those IMP playing physiologically? Concerning the first problem, some investigators claim that IMP which have a complementary depression on the opposite fracture face are made up of lipid and/or lipopolysaccharide molecules (37). More over, they have shown particles in liposomal membranes which are devoid of protein (36). The composition of IMP may be different among membranes of various species. But it is very difficult to examine whether heterogeneous IMP could exist