We have utilized immunofluorescence techniques to look for synaptic vesicle antigens on the plasma membrane of resting and active nerve terminals. Rabbit antiserum was raised against purified cholinergic synaptic vesicles from the electric organ of Narcine brasiliensis, a marine electric ray. Antibodies to synaptic vesicles were shown to bind selectively to nerve terminals in cryostat sections of frog nerve-muscle preparations. Binding was demonstrated indirectly by using fluorescein-labeled goat anti-rabbit antibodies. Structures in cross sections that bound antiserum were identified as nerve terminals because of their size, shape, and position and because they coincided with sites that bound rhodamine-conjugated alpha-bungarotoxin and had acetylcholine esterase activity. Presumably, sectioning gave antibodies access to binding sites within the nerve terminal. However, when antibodies to synaptic vesicles were added to the bathing medium of intact neuromuscular preparations prior to sectioning, antibody binding was marginal or undetectable, suggesting that few vesicle antigens were normally accessible on the outer surface of resting nerve terminals. When intact preparations were stimulated to release their vesicular acetylcholine by the addition of 1 mM LaCl3, antibody binding to the intact nerve terminals became striking. These findings suggest that the synaptic vesicle membrane and the synaptic terminal plasma membrane differ in composition. They also provide further support for the exocytotic hypothesis of neurotransmitter release, which predicts that vesicle markers should be exposed on the outside of nerve terminals when vesicles fuse with the plasma membrane during stimulation.