This paper describes the pharmacology of the novel 5-hydroxytryptamine3 (5-HT3) receptor antagonist GR38032F. 2. On the isolated vagus nerve and superior cervical ganglion of the rat, R,S-GR38032F behaved as a reversible competitive antagonist of 5-HT-induced depolarization with pKB values of 8.61 +/- 0.08 (n = 19) and 8.13 +/- 0.07 (n = 16), respectively. The resolved R- and S-isomers of GR38032F were approximately equipotent as 5-HT antagonists on the rat vagus nerve: the pKB values were 8.95 +/- 0.05 (n = 16) and 8.63 +/- 0.08 (n = 17), respectively. R,S-GR38032F was also an effective antagonist of 5-HT on the rabbit isolated vagus nerve: in this case the pKB value was 9.40 +/- 0.14 (n = 4). 3. On the rabbit isolated heart, low concentrations of R,S-GR38032F (3 X 10(-11)-1 X 10(-9) M) antagonized the positive chronotropic effect of 5-HT and 2-methyl-5-hydroxytryptamine (2-methyl-5-HT). However, the effects of the compound did not appear consistent with simple reversible competition. 4. On the longitudinal smooth muscle of the guinea-pig ileum, R,S-GR38032F caused concentration-dependent parallel rightward displacement of the 2-methyl-5-HT concentration-contraction response curve; in contrast, a portion of the response to 5-HT appeared resistant to R,S-GR38032F. pKB values estimated from the effects of the compound against 2-methyl-5-HT or the inhibitable portion of the response to 5-HT were 7.31 +/- 0.06 (n = 8) and 7.33 +/- 0.13 (n = 8), respectively. Against 2-methyl-5-HT, R-GR38032F seemed more potent (pKB 7.20 +/- 0.10; n = 6) than S-GR38032F (pKB 6.30 +/- 0.05; n = 6). 5. R,S-GR38032F is highly selective for 5-HT3 receptors, and at concentrations of 3 X 10(-6)-3 X 10(-5) M, had negligible agonist or antagonist activity on other 5-HT or non-5-HT receptor-containing tissues on which it was tested. 6. The potency and duration of action of R,S-GR38032F in blocking 5-HT3 receptors in vivo were assessed by measuring its ability to antagonize the bradycardic response to 5-HT or 2-methyl-5-HT administered intravenously (i.v.) to anaesthetized animals. For i.v. administration to the rat, the ED50 for R,S-GR38032F against 2-methyl-5-HT (100pgkg-1) was 0.4 (95% confidence limits 0.18- 0.87) ygkg-1 (n = 10); the corresponding value for oral administration to this species was 7.0 (3.0- 22.0)pgkg-' (n = 8-10 per dose level). R,S-GR38032F was similarly effective in the anaesthetized cat. 7. The present results are discussed with reference to the postulated existence of subtypes of the 5-HT3 receptor.