This study analysed the inhibition produced by the agonists moxonidine (imidazoline I(1) receptors>alpha(2)-adrenoceptors) and agmatine (endogenous ligand of imidazoline I(1)/I(2) receptors), using B-HT 933 (6-ethyl-5,6,7,8-tetrahydro-4H-oxazolo[4,5-d]azepin-2-amine dihydrochloride; alpha(2)-adrenoceptors) for comparison, on the rat cardioaccelerator sympathetic outflow. Male Wistar rats were pithed and prepared to stimulate the cardiac sympathetic outflow or to receive i.v. bolus of exogenous noradrenaline. Sympathetic stimulation or noradrenaline produced, respectively, frequency-dependent and dose-dependent tachycardic responses. I.v. continuous infusions of moxonidine (3 and 10 microg/kg min), agmatine (1000 and 3000 microg/kg min) and B-HT 933 (30 and 100 microg/kg min) inhibited the tachycardic responses to sympathetic stimulation, but not those to noradrenaline. The cardiac sympatho-inhibition by either moxonidine (3 microg/kg min) or B-HT 933 (30 microg/kg min) was not modified by i.v. injections of saline or the antagonists AGN192403 [(+/-)-2-endo-Amino-3-exo-isopropylbicyclo[2.2.1]heptane hydrochloride; 3000microg/kg; imidazoline I(1) receptors] or BU224 (2-(4,5-dihydroimidazol-2-yl)quinoline hydrochloride; 300 microg/kg; imidazoline I(2) receptors) and abolished by rauwolscine (300 microg/kg; alpha(2)-adrenoceptors). At the same doses of these compounds, the sympatho-inhibition to moxonidine (10 microg/kg min) and agmatine (1000 microg/kg min) was: (1) not modified by saline, AGN192403 or BU224; (2) partially blocked by rauwolscine or the combination of rauwolscine plus BU224; and (3) abolished by the combination of rauwolscine plus AGN192403. These results demonstrate that the cardiac sympatho-inhibition to: (1) 3 microg/kg min moxonidine or 30 microg/kg min B-HT 933 involves alpha(2)-adrenoceptors; and (2) 10 microg/kg min moxonidine or 1000 microg/kg min agmatine involves alpha(2)-adrenoceptors and imidazoline I(1) receptors.