Exposure of sensitised intestine to specific allergen is known to produce appreciable reduction in water and electrolyte absorption. The mediators participating in this process have not been fully characterised. The effects of the 5-hydroxytryptamine2 (5-HT2) and 5-HT3 receptor antagonists, ketanserin and granisetron, respectively, on water movement during intestinal anaphylaxis were studied. Hooded Lister rats (120-150 g) were sensitised to ovalbumen and 14 days later, intestinal water and electrolyte movement was assessed at 10 minute intervals by in situ jejunal perfusion with a plasma electrolyte solution (PES) or PES containing 20 mg/l ovalbumen. Within 20 minutes of exposure to PES+ovalbumen, net water secretion that could be completely prevented by the mast cell stabilising agent doxantrazole occurred compared with absorption with PES alone (median -20 microliters/min/g (interquartile range -43 to -5), n = 11), v (107 (86 to 113), n = 10; p < 0.01). Pre-treatment with subcutaneous ketanserin 200 micrograms/kg (n = 7) or granisetron 300 micrograms/kg (n = 8) partially inhibited the secretory response to PES+ovalbumen (18 (11 to 48) and 13 (6 to 32) respectively; both p < 0.01 compared with PES+ovalbumen control). After 40 minutes perfusion with PES+ovalbumen, the changes in water movement were less pronounced 24 (-3 to 43) and neither ketanserin or granisetron had any effect (ketanserin: 48 (28 to 87), granisetron: 41 (32 to 83); NS). In all experiments, sodium and chloride movement paralleled that of water. Thus, the profound water secretion that occurs in the early stages of intestinal anaphylaxis is partly 5-HT dependent because it can be reversed by 5-HT2 and 5-HT3 receptor antagonists. Other mediators must also be involved, especially in the late phase of anaphylaxis.