Previous work has shown that the fluorescent styryl dye FM1-43 stains nerve terminals in an activity-dependent fashion. This dye appears to label the membranes of recycled synaptic vesicles by being trapped during endocytosis. Stained terminals can subsequently be destained by repeating nerve stimulation in the absence of dye; the destaining evidently reflects escape of dye into the bathing medium from membranes of exocytosing synaptic vesicles. In the present study we tested two key aspects of this interpretation of FM1-43 behavior, namely: (i) that the dye is localized in synaptic vesicles, and (ii) that it is actually released into the bathing medium during destaining. To accomplish this, we first photolyzed the internalized dye in the presence of diaminobenzidine. This created an electron-dense reaction product that could be visualized in the electron microscope. Reaction product was confined to synaptic vesicles, as predicted. Second, using spectrofluorometry, we quantified the release of dye liberated into the medium from tubocurarine-treated nerve-muscle preparations. Nerve stimulation increased the amount of FM1-43 released, and we estimate that normally a stained synaptic vesicle contains a few hundred molecules of the dye. The key to the successful detection of released FM1-43 was to add the micelle-forming detergent 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), which increased FM1-43 quantum yield by more than two orders of magnitude.