1. Application of an amino-residue-modifying reagent, 2,4,6-trinitrobenzene-1-sulphonic acid (TNBS), to the frog neuromuscular junction in high-magnesium Ringer solution rapidly increased both the amplitude of nerve-evoked end-plate potentials (EPPs) and the frequency of miniature end-plate potentials (MEPPs). These showed a similar initial time course and reached a maximum 3-7 min and about 10 min, respectively, after the start of application of 2 mM-TNBS. Then, the EPP amplitude decreased, while the MEPP frequency maintained its plateau value. The increase in transmitter release and the decrease in EPP amplitude by TNBS may have been due to different modes of action. 2. The distribution of MEPP amplitude was unchanged by TNBS treatment. 3. The carbachol-induced postsynaptic potential and the extracellularly recorded presynaptic action current were not affected by TNBS treatment for up to 30 min, indicating that the change in EPP amplitude produced by TNBS was not due to either a postsynaptic effect or a change in action potential at the presynaptic terminal. 4. The frequency of MEPPs was increased by TNBS application even when Ca2+ was omitted from the external Ringer solution or when a specific calcium channel blocker, synthetic omega-conotoxin, was added. This indicates that Ca2+ inflow to the nerve terminal is not necessary for TNBS action. 5. When a calcium chelator, BAPTA, was loaded into the presynaptic nerve terminal, the facilitation of EPPs by trains of nerve stimuli was scarcely observed. This suggested that the cytosolic free Ca2+ in the presynaptic terminal was buffered by BAPTA. Under this condition, the amplitudes of EPPs were increased by TNBS application to the same extent as in the control without BAPTA, but were accompanied by little facilitation. The MEPP frequency was also increased by TNBS to the same extent as in the control. These results suggest strongly that augmentation of transmitter release by TNBS was not due to an increase in cytosolic Ca2+ concentration. 6. These observations suggest that TNBS might react with specific protein(s) on the outer surface of the presynaptic membrane and accelerate the exocytosis of synaptic vesicles.