Using voltage-clamp techniques spontaneously occurring miniature end-plate currents (mepc) and nerve-evoked end-plate currents (epc) were recorded in frog glycerol-treated or cut muscle preparations. Epcs were induced by pairs of stimuli (the delay of the 2nd stimulus, delta t being 6 ms-30 s; one pair was delivered every 60-90 s). The decay time constant of the epc (tau epc) was longer, the larger its quantal content despite the presence of active acetylcholinesterase (AChE). After treatment with anticholinesterases (prostigmine or armin, an irreversible inhibitor) this increase in tau epc became more pronounced. When AChE was fully active the decay of the 1st epc (tau 1) was slightly faster than the decay of the 2nd epc (tau 2) only when the interstimulus interval was rather short (delta t less than 20 ms). Following treatment with anticholinesterases this difference between tau 2 and tau 1 could be determined even when delta t was as long as 30 s. In anticholinesterase-treated preparations delta tau was found to be inversely proportional to log delta t: a 50% increase in the decay time-constant of the 2nd epc occurred with delta t = 120 ms. During continuous stimulation (10 impulses/s) tau epc increased from the 1st to the 5-6th responses, but then decreased in parallel with the fall in the epc amplitude. The phenomenon of postsynaptic potentiation we observed could be readily abolished when quantal content was decreased by the presence of magnesium ions, but it was relatively unaffected when the receptor density was decreased by alpha-bungarotoxin (alpha-BuTX).(ABSTRACT TRUNCATED AT 250 WORDS)