1. We have compared the potency of the putative platelet-activating factor (Paf) receptor antagonists (WEB 2086, L-652,731 and BN 52021) against Paf-induced aggregation of rabbit and guinea-pig platelets, aggregation of rabbit polymorphonuclear leukocytes (PMNLs) and prostacyclin generation by guinea-pig resident peritoneal macrophages. 2. On rabbit washed platelets and PMNLs WEB 2086, L-652,731 and BN 52021 each antagonized competitively Paf-induced aggregation. The rank order of potency was WEB 2086 congruent to L-652,731 greater than BN 52021 and was the same for the two cell types. 3. The pA2 values for each of the three antagonists were similar on rabbit washed platelets and PMNLs. Moreover, the pA2 for WEB 2086 on rabbit platelets (7.58) did not differ significantly from that on guinea-pig platelets (7.69). 4. On guinea-pig resident peritoneal macrophages WEB 2086 was 10 fold less potent for receptors mediating increased generation of 6-oxo-prostaglandin F1 alpha (6-oxo-PGF1 alpha) than for those mediating platelet aggregation. 5. The potencies of L-652,731 and BN 52021 were also markedly less (2 log units) for the macrophage receptors than for platelet or PMNL receptors and BN 52021 was more potent than L-652,731 in the macrophages. 6. WEB 2086 and L-652,731 significantly reduced basal 6-oxo-PGF1 alpha produced by macrophages, but none of the antagonists affected 6-oxo-PGF1 alpha production during stimulation by A23187. 7. These data raise the possibility that there may be a Paf receptor-subtype mediating prostacyclin generation in macrophages that is different from that on the platelet and PMNL. Hence, the potency of Paf antagonists against platelet aggregation would not be a good predictor of antagonist potency in disorders involving macrophages.