A variety of non-xanthine heterocycles were found to be antagonists of binding of [3H]phenylisopropyladenosine to rat brain A1-adenosine receptors and of activation of adenylate cyclase via interaction of N-ethylcarboxamidoadenosine with A2-adenosine receptors in human platelet and rat phenochromocytoma cell membranes. The pyrazolopyridines tracazolate, cartazolate and etazolate were several fold more potent than theophylline at both A1- and A2-adenosine receptors. The pyrazolopyridines, however, were still many fold less potent than 8-phenyltheophylline and other 8-phenyl-1,3-dialkylxanthines. A structurally related N6-substituted 9-methyladenine was also a potent adenosine antagonist with selectivity for A1 receptors. None of several aryl-substituted heterocycles, including a thiazolopyrimidine, imidazopyridines, benzimidazoles, a pyrazoloquinoline, a mesoionic xanthine analog and a triazolopyridazine exhibited the high potency typical of 8-phenyl-1,3-dialkylxanthines. A furyl-substituted triazoloquinazoline was very potent at both A1 and A2 receptors. A pteridin-2,4-dione, 1,3-dipropyllumazine, was somewhat less potent than theophylline at A1- and A2-adenosine receptors, whereas 1,3-dimethyllumazine was much less potent. A benzopteridin-2,4-dione, alloxazine, was somewhat more potent than theophylline. Other heterocycles with antagonist activity were the dibenzazepine carbamazepine and beta-carboline-3-ethyl carboxylate. The phenylimidazoline clonidine had no activity, whereas a related dihydroxyphenylimidazoline was a weak non-competitive adenosine antagonist.