1. The effects of physostigmine and neostigmine were compared at frog neuromuscular junctions using intracellular microelectrodes. 2. Both drugs increased the amplitudes of miniature endplate potentials (MEPP) in a dose-dependent manner because of their ability to competitively inhibit cholinesterase activity. 3. In addition, physostigmine (1.0-20.0 mumol/l) decreased by approximately 20% the quantal content (determined by two methods), indicating a decrease in the amount of transmitter released upon nerve stimulation, whereas neostigmine did not alter quantal content. 4. Because the endplate potential (EPP) amplitude is a composite of pre- and post-junctional effects, the dose-response curve for neostigmine on EPP amplitude approximately paralleled that obtained on MEPP amplitude, whereas the effect of physostigmine on EPP amplitude was depressed because of its action to decrease transmitter release. 5. Upon washout of neostigmine, the increase in EPP amplitude reversed more rapidly than effects on MEPP amplitude, a difference dependent on the magnitude of quantal content and explained by an action on the postjunctional membrane. Similar results were obtained with physostigmine. 6. Neither drug significantly affected MEPP frequency although small decreases were sometimes observed. 7. These results suggest that at the frog neuromuscular junction the effects of neostigmine are explained primarily by its inhibition of cholinesterase, whereas physostigmine are explained primarily by its inhibition of cholinesterase, whereas physostigmine has an additional action on nerve terminals of decreasing nerve-stimulated release of acetylcholine.