Post-tetanic potentiation (PTP) was elicited at the frog sartorius and cutaneous pectoris neuromuscular junctions. A 30-sec, 30-Hz tetanus produced a 2- to 3-fold post-tetanic increase in endplate potential (EPP). In surface-recorded responses this PTP decayed in a double exponential way with time constants of 12.7 sec +/- 2.4 (SEM) and 146.8 sec +/- 36.6. In acute experiments 0.2 to 0.8 mM phenytoin (5,5-diphenylhydantoin, DPH) dramatically and reversibly reduced the early component. The late component was also reduced, although to a lesser extent and often not reversibly. DPH reduced PTP even when there was no failure of the EPP during the tetanus. Thus, the DPH effect did not require a complete block of the presynaptic action potential. At longer exposures and higher DPH concentrations EPP failures did develop, and this was associated with a more profound suppression of PTP. PTP was also elicited in tetrodotoxin (TTX)-containing solutions using electronic stimulation of nerve terminals to elicit transmitter release. This PTP had a much shorter duration (about 30 sec) than that seen in normal Ringer's solution and was followed by depression of EPP amplitudes. Thus, sodium entry into nerve terminals enables a mechanism which greatly prolongs PTP. DPH had no effect on PTP in TTX. These results, together with others in the literature, suggest that the reduction of PTP by DPH involves a graded reduction of sodium influx into nerve terminals during high rates of axon stimulation. The development of all-or-none failures of the presynaptic action potential results in even greater suppression of PTP.