1. The glutamatergic synapses formed between the unbranched giant reticulospinal axons onto spinal neurons in lamprey offer a central vertebrate synapse in which the presynaptic element can be impaled with one or several microelectrodes, which may be used for recording as well as microinjection of different substances. To provide a basis for the use of this synapse in studies of release mechanisms, we have examined the use- dependent modulation of the synaptic response under conditions of conventional cell body stimulation, and during direct stimulation of the presynaptic axon. 2. To examine the stability of the mixed electrotonic and chemical reticulospinal excitatory postsynaptic potential (EPSP) over time, action potentials were evoked at a rate of 1 Hz for 800-1000 trials. In three out of seven synapses the chemical component remained at a similar amplitude, while in four cases a progressive decrease (up to 35%) occurred. The electrotonic component remained at a similar amplitude in all cases. 3. During paired pulse stimulation of the reticulospinal cell body (pulse interval 65 ms) the chemical EPSP, component showed a net facilitation in all cases tested [from 0.64 ± 0.35 to 0.89 ± 0.48 (SD) mV, n = 13], while the peak amplitude of the electrotonic component was unchanged (1.37 ± 0.68 and 1.36 ± 0.66 mV, respectively). Recording of the axonal action potential during paired pulse stimulation showed that the width of the first and second action potential did not differ [ 1/2 width (2.48 ± 0.39 ms and 2.48 ± 0.42 ms, respectively; n = 8)]. 4. The degree of facilitation varied markedly between different synapses, ranging from an increase of a few percent to a two-fold increase (24 ± 16% mean change of total EPSP amplitude, corresponding to 44 ± 26% mean change of chemical EPSP amplitude). This type of variability was also observed in synapses made from the same unbranched reticulospinal axon onto different postsynaptic cells. 5. When paired pulse stimulation was applied to the reticulospinal axon in the very vicinity of the synaptic area (0.1-1 mm) a net depression of the chemical component occurred in 11 out of 19 cases, and in the remaining cases the level of net facilitation was lower as compared with cell body stimulation (range between +17 and -23% change of total EPSP amplitude; mean -5%; n = 19). 6. To test if the change of the EPSP plasticity during local stimulation correlated with an increased transmitter release, two microelectrodes were placed in the same reticulospinal axon at different distances from the synaptic area. The total EPSP amplitude was consistently larger, if the action potential was triggered with a depolarizing current pulse (1-2 ms) applied close to the synapse (0.1- 1 mm), than at a longer distance. During local stimulation the EPSP amplitude increased with the magnitude of the depolarizing stimulus pulse. If the axon was impaled closer than 0.1 mm from the synaptic area, however, the chemical EPSP became depressed, but a reversal occurred after the electrode had been removed. 7. The increase of the total EPSP amplitude during local stimulation was accompanied by an increase of the peak amplitude of the initial electrotonic component, indicating that the amplitude of the presynaptic action potential was increased. Recording of the axonal action potential confirmed that both its amplitude and duration increased when it was triggered with depolarizing current pulses applied from an adjacent microelectrode. 8. To examine the anatomic correlate of a given reticulospinal axon-motoneuron EPSP with an initial electrotonic component of 1.0 mV amplitude and a chemical component of 0.25 mV amplitude (distant stimulation), both elements were labeled intracellularly and a complete reconstruction of the synaptic connections was performed. Four axon- motoneuron contacts, located on distal dendrites at 310-340 μm distance from the motoneuron soma were observed, which altogether contained seven active zones, but only two gap junctions. Analysis of serial ultrathin sections showed that the different active zones were separated from each other by glial processes. 9. The present results show that the use-dependent modulation of unitary glutamatergic EPSPs can vary between individual synapses along an unbranched axon, in which no activity-dependent changes in the shape or propagation of the presynaptic action potential can be detected. If the stimulation electrode is placed within a distance of 1 mm from the synaptic area, however, the shape of the presynaptic spike, as well as the amplitude and use- dependent modulation of the EPSP, can be markedly altered. These data should be taken into account when large vertebrate axons are employed in studies of the mechanisms underlying synaptic glutamate release in the CNS.