1. This study describes attempts to compare prostacyclin (IP-) receptors in human, pig, horse, rabbit and rat platelets and in circular muscle of human, rabbit and dog mesenteric and pig gastroepiploic arteries. Three stable prostacyclin analogues, iloprost, cicaprost and 6a-carba-prostacyclin (6a-carba-PGI2) and a prostaglandin endoperoxide analogue EP 157 (previously shown to mimic prostacyclin on human platelets) were used. 2. Our main conclusion is that prostacyclin receptors on human, pig and horse platelets are similar in nature, but distinct from those on rabbit and rat platelets. Functional studies (inhibition of aggregation) showed that iloprost and cicaprost always had similar potencies whereas 6a-carba PGI2 was much more potent than EP 157 on rabbit and rat platelets (300 and 1000 fold on a molar basis) compared with human, pig and horse platelets (2, 7 and 7 fold respectively). Measurement of initial rates of cyclic AMP production confirmed these orders of potency. 3. Although pig platelets were quite sensitive to inhibition by EP 157 (threshold = 10 nM in some experiments), maximal inhibition of aggregation was not always achieved (20 microM). EP 157 also produced only small elevations of cyclic AMP and inhibited rises in cyclic AMP induced by iloprost. It is possible that EP 157 has a lower efficacy than iloprost at the IP-receptor and on pig platelets it can sometimes act as a partial agonist. 4. Human, pig and horse platelet membranes bound [3H]-iloprost at 30 degrees C and this binding was inhibited by the four prostanoids. On human and pig membranes the order of potency was cicaprost = iloprost greater than 6a-carba PGI2 greater than EP 157. The order of potency may be similar on horse platelet membranes, but the analysis is complicated by the presence of a second component of [3H]-iloprost binding that is inhibited by iloprost and 6a-carba PGI2 but not by cicaprost. This binding may be due to the presence of an EP1-receptor, since iloprost and 6a-carba PGI2 but not cicaprost are known to have potent EP1-receptor agonist actions on smooth muscle preparations. IC50 values for cicaprost inhibition on human, pig and horse membranes were 110, 90 and 165 nM respectively. The need for IP-receptor radioligands of greater specificity is apparent from these studies. 5. Minimal binding of [3H]-iloprost to rabbit and rat platelet membranes was obtained at 30 degrees C. Lowering the incubation temperature to 4 degrees C and ensuring that the temperature did not rise during the filtration process increased binding and allowed inhibition curves to be obtained. The results suggest a lower binding affinity for [3H]-iloprost, associated with a higher dissociation rate for the radioligand-receptor complex. IC50 values for cicaprost were 900nm for rabbit and 640nm for rat platelets. In a similar manner to horse platelet membranes, the presence of a second binding site for [3H]-iloprost was detected on rabbit platelet membranes. 6. Sensitivity to the relaxant action of iloprost on the arterial smooth muscle preparations decreased in the order: human mesenteric, dog mesenteric, rabbit mesenteric, pig gastro-epiploic. Cicaprost was always slightly more potent than iloprost (1.2-2.8 fold). On the pig vessel preparation 6a-carba PGI2 did not produce complete relaxation. The possibility that this is due to an opposing contractile action mediated via EP1 or EP3 receptors is discussed. 7. EP 157 relaxed the human, pig and rabbit arterial preparations at concentrations 100-200 times those of iloprost. This correlates well with its IP-receptor agonist potency on human, pig and horse platelets. The results obtained with EP 157 further demonstrate the potential difficulties in separating platelet inhibitory and vasodilator properties of prostacyclin mimetics in man.