1. The 16-phenoxy prostaglandin E analogue sulprostone consistently potentiates primary aggregation waves induced by adenosine 5'-diphosphate (ADP), PAF and 11,9-epoxymethano PGH2 (U-46619) in platelet-rich plasma from human donors. The effect is not blocked by the TP-receptor antagonists, EP 092 and GR 32191. The high potency of sulprostone (threshold concentration = 4-10 nM) and the weak block of sulprostone potentiation by the EP1-receptor antagonist, AH 6809 (pA2 = 4.3) suggest the involvement of EP3-receptors as opposed to EP1- or EP2-subtypes. 2. Eight prostaglandin E (PGE) analogues were compared against sulprostone for their effects on PAF-induced aggregation in human platelet-rich plasma (PRP) in the presence of GR 32191 and the DP-receptor antagonist, BW A868C. PGE2 and 11-deoxy PGE2-1-alcohol showed evidence of both potentiating and inhibitory actions and butaprost showed only inhibitory activity at high concentrations. The remaining analogues always elicited potentiation, with the following potency ranking: sulprostone = 16,16-dimethyl PGE2 > MB 28767 > misoprostol > GR 63779X = 17-phenyl-omega-trinor PGE2. The results again indicate that EP3- rather than EP1- or EP2-receptors are involved. However, relative potentiating potency could be affected by differences in plasma protein binding and the very high sensitivity of the human platelet to prostacyclin (IP)-receptor-mediated inhibition (IC50 for the specific IP-receptor agonist cicaprost = 0.8 nM). 3. On human washed platelet suspensions the PGE analogues, with the exception of butaprost,inhibited the rise in adenosine 3':5'-cyclic monophosphate (cyclic AMP) induced by cicaprost (8 nM).PGE2 produced a monophasic inhibition curve (IC50 = 5.4 nM, 92% inhibition at 600 nM). The potency ranking was 16,16-dimethyl PGE2> sulprostone>MB 28767 = PGE2> misoprostol> GR 63778X>17-phenyl-w-trinor PGE2> 1 1-deoxy PGE2-1-alcohol. AH 6809 inhibited the effect of sulprostone and 17-phenyl-c-trinor PGE2 with pA2 values of 5.75 and 5.32 respectively; these values are at least one log unit lower than those found for EP1-receptor block in smooth muscle.4. There is a statistically significant correlation between IC50 values for the PGE analogues on the human platelet cyclic AMP assay and the guinea-pig vas deferens (standard EP3 preparation): slope =1.00, r = 0.80, P <0.05. However the correlation is far from ideal and GR 63779X in particular has a lower potency in the cyclic AMP assay. At this time we suggest that it is prudent to describe the human platelet receptor as 'EP3-like'.5. We believe that our results provide further evidence for linking PGE-induced potentiation of aggregation to inhibition of adenylate cyclase. Sulprostone is a suitable agonist for further study of this system and in particular the nature of the G-protein linkage(s) involved. In addition the necessity to consider potentiation of platelet aggregation in -relation to the clinical use of PGE analogues in man is emphasised.