On the basis of data obtained in rabbits, the imidazoline receptor ligand rilmenidine has been suggested to decrease blood pressure in humans by activating central alpha(2A)-adrenoceptors. A prerequisite for this hypothesis was the unproved assumption that rabbit and human alpha(2A)-adrenoceptors are equally activated by rilmenidine. Because alpha(2A)-adrenoceptors in the brain and on cardiovascular sympathetic nerve terminals are identical, the latter were used as a model for the former to confirm or disprove this assumption. Human atrial appendages and rabbit pulmonary arteries were used to determine the potencies of alpha(2)-adrenoceptor agonists in inhibiting the electrically (2 Hz) evoked [(3)H]norepinephrine release and of antagonists in counteracting the alpha(2)-adrenoceptor-mediated inhibition induced by moxonidine. In the rabbit pulmonary artery, rilmenidine and oxymetazoline are potent full agonists, whereas in the human atrial appendages they are antagonists at the alpha(2)-autoreceptors, sharing this property with rauwolscine, phentolamine, and idazoxan. In contrast, prazosin is ineffective. In addition, a partial nucleotide and amino acid sequence of the rabbit alpha(2A)-adrenoceptor (a region known to substantially influence the pharmacological characteristics of the alpha(2)-adrenoceptor) revealed marked differences between the rabbit and the human alpha(2A)-adrenoceptor. The sympathetic nerves of both the human atrial appendages and rabbit pulmonary artery are endowed with alpha(2A)-autoreceptors, at which, however, both rilmenidine and oxymetazoline exhibit different properties (antagonism and agonism, respectively). The antagonistic property of rilmenidine at human alpha(2A)-adrenoceptors indicates that in contrast to the suggestion based on rabbit data, the hypotensive property of the drug in humans is not due to activation of alpha(2A)-adrenoceptors but other, presumably I(1)-imidazoline receptors, are probably involved.