A radioisotope method has been developed for measuring the stimulation-evoked release of acetylcholine without the use of cholinesterase inhibitors from the mouse hemidiaphragm preparation which had been loaded with 3H-choline. Evidence has been obtained that 3H-choline was taken up by and released from both innervated and non-innervated mouse hemidiaphragm preparations. However, it was released in the form of 3H-acetylcholine in response to electrical field stimulation only from the innervated preparations. Long lasting (51 min) S1 stimulation of the preparations exhausted the radioactive acetylcholine stores to the extent that S2 did not evoke any release of 3H. These data suggest that when the labelled acetylcholine stores become exhausted, the labelled choline, still present in the tissue, cannot be released by electrical stimulation. Tetrodotoxin (1 mumol/l) administration and Ca withdrawal inhibited, 20-100 mumol/l 4-aminopyridine enhanced the release of 3H-acetylcholine in response to electrical stimulation. Activation of the presynaptic muscarinic receptors by the agonist oxotremorine (50 mumol/l) decreased the liberation of 3H-acetylcholine. The muscarinic antagonist atropine (1 mumol/l) abolished the inhibitory effect of oxotremorine and by itself increased the evoked release of the newly formed 3H-acetylcholine. Adenosine (50 mumol/l) reduced the evoked release of radioactivity. Theophylline (30 mumol/l) prevented the inhibitory effect of adenosine and itself enhanced the release. Xylazine (1 mumol/l), an alpha 2-adrenoceptor agonist did not affect the release. It is concluded that the stimulation-evoked release of 3H-acetylcholine from the mouse phrenic nerve hemidiaphragm preparation preloaded with 3H-choline is derived from the motor nerves. The release of acetylcholine is modulated by activation of presynaptic muscarinic and adenosine receptors.