Agmatine, an endogenous clonidine-displacing substance, has been shown to have an affinity for both alpha 2-adrenoceptors and imidazoline receptors (IR). In conscious rabbits, we have examined the cardiovascular effects of agmatine and its interaction with clonidine, a presumed agonist and 2-methoxyidazoxan, an antagonist at alpha 2-adrenoceptors. We have also examined the effect of agmatine on agents having high affinity for I1-imidazoline receptors namely moxonidine (agonist) and efaroxan (antagonist). Initial dose-response studies showed that agmatine administered in low doses (0.01-10 micrograms/kg) into the fourth ventricle did not change mean arterial pressure but did produce a dose-dependent bradycardia (maximum -16 +/- 3 beats/min). A higher dose of 100 micrograms/kg produced an adverse reaction in the conscious animals accompanied by a marked increase in mean arterial pressure and a reversal of the bradycardia. This is in contrast to the effects of fourth ventricular clonidine and moxonidine, which caused a dose-dependent fall in both mean arterial pressure and heart rate. Agmatine when administered at the highest well-tolerated dose of 10 micrograms/kg did not further alter the clonidine-induced hypotension but produced a greater bradycardia (-12 +/- 4 beats/min clonidine; -29 +/- 4 beats/min clonidine plus agmatine; p < 0.05). Similarly, the hypotension induced by moxonidine was not altered by agmatine but heart rate was reduced after the addition of agmatine (p < 0.01). Efaroxan and 2-methoxy-idazoxan, at doses which produced no effects when given alone, similarly reversed the fall in heart rate elicited by agmatine and caused a small but significant rise in mean arterial pressure. We have previously shown that the doses of these antagonists used in this study produce an equal reversal of the bradycardia induced by fourth ventricular alpha-methyldopa (alpha 2-adrenoceptor agonist) and clonidine and hence have similar alpha 2-adrenoceptor blocking effects. Our results show that agmatine produces bradycardia as does moxonidine and clonidine but does not mimic or block the hypotensive responses to these agents. These findings do not support the hypothesis that agmatine is an endogenous ligand for IR. However, the bradycardia induced by agmatine may be mediated via alpha 2-adrenoceptors since it was equally blocked by efaroxan and 2-methoxy-idazoxan. Thus while alpha 2-adrenoceptor actions of agmatine on heart rate are evident at relatively low doses, the reason for the lack of alpha 2-adrenoceptor mediated hypotension is not known.