Experiments were designed to evaluate the effect of cyclic AMP on the electrically-induced release of noradrenaline from vascular sympathetic nerve terminals. The possible implication of the inhibition of adenylate cyclase in the negative feed-back control by prejunctional alpha 2-adrenoceptors of neurotransmitter release was also investigated. Rat isolated tail arteries were preincubated with [3H]-noradrenaline; the preparations were subsequently perfused/superfused with [3H )-noradrenaline-free medium and their perivascular nerves were field stimulated with 24 pulses at 0.4 Hz (0.3 ms, 200 mA). 2 compounds known to enhance the intracellular concentration of cyclic AMP, namely the membrane permeant analogue 8-Br-cAMP (10-300 mumol/l) and forskolin (0.3-10 mumol/l), an activator of adenylate cyclase, concentration-dependently enhanced the stimulation-evoked tritium overflow. The 1,9-dideoxy derivative of forskolin, which does not stimulate adenylate cyclase, was ineffective. Exposure to the cyclic AMP phosphodiesterase inhibitor rolipram 30 mumol/l produced a moderate increase (about 20%) in tritium overflow. However, in the presence of rolipram the facilitatory effect of forskolin was significantly more pronounced than in its absence. Whereas 8-Br-cAMP produced a slight concentration-dependent enhancement of the stimulation-induced vasoconstriction, forskolin and rolipram depressed it. The alpha 2-adrenoceptor agonist B-HT 933 (3-30 mumol/l) concentration-dependently inhibited the tritium overflow. The effect of B-HT 933 30 mumol/l was slightly, but significantly reduced in the presence of 8-Br-cAMP 100 and 300 mumol/l, but was not changed in the presence of forskolin 3 mumol/l. The facilitatory effect of rauwolscine 1 mumol/l was enhanced in the presence of 8-Br-cAMP 100 mumol/l. During perfusion with 8-Br-cAMP 100 mumol/l, the current strength and frequency were decreased to 150 mA and 0.2 Hz, respectively in order to obtain similar amounts of tritium overflow to those observed in the absence of the cyclic AMP analogue with the initial stimulation parameters. Under these conditions, the inhibition of the overflow by B-HT 933 30 mumol/l and the facilitation by the alpha 2-adrenoceptor antagonist rauwolscine 1 mumol/l were unaltered as compared to controls under initial stimulation conditions. It is concluded that, in the rat tail artery, the terminals of perivascular sympathetic nerves are endowed with an adenylate cyclase system. Cyclic AMP is able to modulate noradrenaline release, but does not appear to play a role in the initiation of the release process itself. In addition, the results do not support the hypothesis that prejunctional alpha 2-adrenoceptors depress noradrenaline release through the inhibition of adenylate cyclase.