Crustacean motor axons provide a model in which activity-dependent changes in synaptic physiology and synaptic structure can be concurrently observed in single identifiable neurons. In response to a train of stimulation, crustacean neuromuscular junctions undergo pronounced facilitation of transmitter release. The effects of maintained high-frequency stimulation may persist for at least several hours ("long-term facilitation"). Electrophysiological studies suggest that the number of "active" synapses contributing transmitter quanta at low frequencies of stimulation increases during and after a train of high-frequency stimulation. However, at different terminal recording sites the effect of stimulation varies, and it was observed that not all released quanta produce a voltage change in the postsynaptic muscle cell. Electron microscopic examinations of serial sections from nerve terminals subjected to stimulation were made to determine whether changes in synaptic structure could be correlated with activity-induced long-lasting enhancement of transmission. A procedure was introduced for marking a recorded terminal with fluorescent polystyrene microspheres, which are visible in electron micrographs of the recording site. Crustacean nerve terminals possess a large number of discrete synapses, a small fraction of which have multiple presynaptic "active zones" (dense bodies with clustered synaptic vesicles, thought to represent sites of evoked transmitter release). In terminals previously stimulated, the proportion of synapses with multiple "active zones" is greater than in control unstimulated terminals. Total synaptic vesicle counts and readily releasable vesicles at synapses are not significantly different in previously stimulated terminals and controls. In terminals fixed during stimulation a few synapses show evidence of division in "active zones," and synaptic vesicle counts are lower than in controls.(ABSTRACT TRUNCATED AT 250 WORDS)