Adenosine produced a concentration-related enhancement of antigen-induced 5-hydroxytryptamine (5-HT) release from rat serosal mast cells. This potentiation was maximal following the simultaneous addition of adenosine with antigen. Enhancement of 5-HT release was accompanied by potentiation of the adenosine 3':5'-cyclic monophosphate (cyclic AMP) response to challenge. The cyclic AMP response, which was antagonized by 8-phenyltheophylline, was characterized as an A2-purinoceptor-mediated effect by the use of 5'-N-ethylcarboxamideadenosine (NECA) and L-N6-phenylisopropyladenosine (L-PIA). Enhancement of 5-HT release, conversely, was not blocked by 8-phenyltheophylline suggesting it to be mediated by a cyclic AMP-independent mechanism. The effect of adenosine on 5-HT release was not reduced by the inhibition of the facilitated uptake of adenosine with dipyridamole, hexobendine or p-nitrobenzylthioguanosine, therefore, suggesting it to be mediated by a cell surface receptor. The receptor mediating enhancement of 5-HT does not appear to belong to the P2-purinoceptor subtype as adenosine was more potent than both adenosine monophosphate (AMP) and adenosine diphosphate (ADP) and alpha, beta-methylene ATP was inactive. Furthermore, the effects of AMP were blocked by alpha, beta-methylene ADP, which inhibits the conversion of AMP to adenosine. Adenosine, NECA, L- and D-PIA were all of equal potency in enhancing 5-HT release. Inosine and 3-deazaadenosine were also active. The rank order of potency of these adenosine analogues is not consistent with an effect at A1- or A2-purinoceptors. There appear to be two adenosine receptors on rat mast cells, an A2-purinoceptor which stimulates adenylate cyclase and a separate purinoceptor, stimulation of which produces enhancement of mediator release by an unknown mechanism. The effects mediated by these receptors appear to be independent of each other.