Local superfusion of limited dendritic areas with hypertonic or hyperkalemic solutions stimulates the release of quanta from a small population of synapses made on rodent hippocampal neurons maintained in primary culture, and each quantal event can be detected in the postsynaptic neuron. With maintained stimulation, the initial release rate is about 20 quanta per sec per synapse, and this rate declines exponentially to a final low level. These observations can be interpreted as depletion of available quanta and, with this interpretation, a bouton would contain one to two dozen quanta in its readily releasable pool. Tests with a second application of the solution that produces release reveal that the pool of readily releasable quanta is replenished with a time constant of about 10 sec (36°C). The pool of quanta defined in this way may correspond to the population of vesicles docked at the bouton's active zone.Because synapses are so densely packed in brain, the properties of individual, identified boutons are very difficult to investigate. Specifically, no physiological information is available on how many quanta are readily available for release by a single central bouton. Earlier investigators have approached the study of individual central synapses in two ways. First, a giant presynaptic terminal from goldfish bipolar cells has been investigated electrophysiologically with exocytosis and endocytosis detected by measuring changes in membrane capacitance (1, 2). Although this approach has many advantages, a disadvantage is that individual exocytotic events cannot be detected, and the giant terminal contains a large number of specialized release zones that are not typical of most other central synapses. Second, hippocampal synapses have been studied in primary cultures with confocal microscopic techniques that permit rates of exocytosis and endocytosis to be followed optically by the uptake and release of fluorescent dye (3), a method that has been used to address similar issues for the amphibian and mammalian neuromuscular junctions (4,5). This method makes individual boutons available for investigation but has limited temporal and spatial resolution so that single quantal releases cannot be detected.We have exploited the accessibility and low density of synapses made in primary cultures of rodent hippocampal neurons to estimate the number of readily releasable quanta per bouton and the time it takes this pool to be replenished once it is depleted. The method we have employed is to superfuse locally a small, identified population of synapses on a single neuron with a solution that produces exocytosis and to count every miniature excitatory postsynaptic current (mEPSC) that occurs. The high resolution of whole cell recording, together with the relatively low release rates that result when a sufficiently small population of synapses is active, permits us to detect virtually every transmitter quantum released. Two different superfusion solutions ("release solutions") have been employed to produce transmitte...