1. Data on the presence of S-100 protein in synaptic endings are revised, and evidence is given in favor of its localization inside mouse brain cortex synaptosomes and on the surface of their external membrane. 2. For identification of the S-100-specific polypeptide, proteins of external synaptosomal membranes were iodinated with lactoperoxidase fixed on cyanogen bromide (CNBr)-Sepharose, and after synaptosome lysis S-100-positive material was isolated by means of affinity chromatography antibodies to S-100 protein (a-S-100)-Sepharose. The molecular weight of the polypeptide obtained corresponded to that of S-100 subunits (10 kD), and iodine incorporation pointed to its localization on the surface of synaptosomal membranes. 3. With the help of antibodies labeled with horseradish peroxidase (a-S-100-HP) or 125I (a-S-100-125I), which do not penetrate into noninjured synaptosomes, the amount of S-100 protein on synaptosomal membranes was found to be 18.5 ng/mg total protein (as assayed with a-S-100-HP) or 95.33 ng/mg (as assayed with a-S-100-125I). 4. At the same time, the total S-100 protein content in synaptosomes measured by means of radioimmune analysis after their complete lysis turned out to be 284 +/- 0.84 ng/mg, i.e., a part of S-100 seemed to be inside synaptosomes. 5. Cosedimentation of water-soluble S-100 protein with the synaptosomal fraction during isolation was insignificant. Prefixation with glutaraldehyde or paraformaldehyde decreased the amount of material reacting with antibodies, possibly due to steric effects or denaturation of active centers. This could have influenced the earlier attempts to detect S-100 protein in synapses. Treatment of nonfixed synaptosomes with a conjugate of a-S-100 with colloidal gold made it possible to detect S-100-positive material on pre- and postsynaptic membranes, which confirms the biochemical data.