Effects of TI-233 (4-isopropyl-1-[N2-(5,6-dimethyl-aminonaphthalene-1-sulphonyl)-L-arginyl ]- piperidine) on glutamate-induced responses and nerve-evoked synaptic responses were compared at the crayfish neuromuscular junction. Intracellularly recorded excitatory junctional potentials (e.j.ps) were markedly augmented by TI-233 when they were evoked at long intervals, whereas the unit size of extracellular e.j.ps was hardly affected by TI-233 and, at that stage, the glutamate-induced current was markedly reduced by TI-233. The decay rate of extracellular e.j.ps was slightly increased 3 min after the addition of TI-233 at concentrations higher than 0.05 mM. Repetitive stimulation of the excitatory axon at a high frequency caused a gradual decrease in the amplitudes of extracellular e.j.ps in the presence of TI-233. After prolonged application of TI-233 with repetitive nerve stimulation, the glutamate-induced response became significantly smaller than the control. TI-233 increased the input resistance of the crayfish muscle fibre and facilitated transmitter release at the excitatory neuromuscular junction. These two effects would entirely explain the augmentation of intracellular e.j.ps by TI-233. TI-233 (greater than 3 microM) reduced the amplitude of current responses to trains of glutamate pulses in a dose-dependent manner, but this reduction by TI-233 was time- and activity-dependent. The effect of TI-233 on glutamate-induced responses was voltage-dependent and hyperpolarization increased this effect. Pretreatment of the muscle fibre with concanavalin A did not affect the gradual decline, caused by TI-233, of the successive currents evoked by a train of glutamate pulses. The apparent differences between the glutamate-induced current and nerve-evoked synaptic response revealed by TI-233 can be explained by open-channel block of the glutamate-activated ion-channel, and do not confute the hypothesis that glutamate is the natural transmitter substance at this junction.