The postsynaptic membrane of the neuromuscular synapse treated with antiacetylcholinesterase is depolarized due to nonquantal release of acetylcholine (ACh) from the motor nerve ending. This can be demonstrated by the hyperpolarization produced by the application of curare (H-effect). ATP (1 × 10−5 m) decreased the magnitude of the H-effect from 5 to 1.5 mV. The membrane input resistance and the ACh sensitivity were unchanged, and so changes in these cannot explain the ATP effect. Adenosine alone was without effect on the nonquantal release. On the other hand, both ATP and adenosine depressed the frequency of spontaneous miniature endplate potentials, to 56% and 43% respectively. The protein kinase A inhibitor Rp-cAMP or the guanylyl cyclase inhibitor 1H-[1,2,4]oxidiazolo[4,3-a]quinoxalin-1-one did not affect the inhibitory influence of ATP on the H-effect, whereas staurosporine, an inhibitor of protein kinase C, completely abolished the action of ATP. Suramin, an ATP antagonist, enhanced the H-effect to 8.6 mV and, like staurosporine, prevented the inhibitory effect of ATP. ATP thus suppresses the nonquantal release via a direct action on presynaptic metabotropic P2 receptors coupled to protein kinase C, whilst adenosine exerts its action mainly by affecting the mechanisms underlying quantal release. These data, together with earlier evidence, show that nonquantal release of ACh can be modulated by several distinct regulatory pathways, in particular by endogenous substances which may or may not be present in the synaptic cleft at rest or during activity.