1. An investigation has been made into the 5-hydroxytryptamine (5-HT) receptor mediating relaxation of rat oesophagus in preparations precontracted with carbachol. 2. In tissues treated with pargyline (100 microM) and in the presence of corticosterone (30 microM) and cocaine (30 microM) the potency of 5-HT and 5-methoxytyramine (5-MeOT) was not changed but the maximum response to these agonists was reduced. Thus there was no evidence of metabolism and/or uptake through an amine depleting mechanism. 3. The relaxant concentration-effect curves to 5-HT were shifted to the left in a concentration-related manner by isobutylmethylxanthine (1 and 10 microM), suggesting the involvement of adenosine 3':5'-cyclic monophosphate in these responses. 4. 5-HT produced concentration-related relaxations of rat oesophagus with an EC50 value of 0.24 microM. Several indole agonists were tested and the following rank order of potency of key agonists obtained: 5-HT greater than alpha-methyl-5-hydroxytryptamine = 5-carboxamidotryptamine (5-CT) greater than 5-MeOT. In contrast, 2-methyl-5-hydroxytryptamine, sumatriptan and 8-hydroxy-2-(di-n-propylamino) tetralin were weak or inactive. 5. The substituted benzamides, metoclopramide, cisapride, renzapride and R,S-zacopride acted as partial agonists, producing 60-70% of the 5-HT maximum. 6. The relaxation responses to 5-HT were neither inhibited by antagonists selective for 5-HT1 or 5-HT2 receptors nor by the 5-HT3 receptor antagonists, ondansetron, granisetron or MDL 72222. 7. The relaxation responses induced by 5-HT, 5-CT, 5-MeOT and renzapride were selectively inhibited by high concentrations of ICS 205-930 with pKB values of approximately 6. 8. The 5-HT receptor mediating relaxation in rat oesophagus cannot be designated 5-HT1, 5-HT2 or 5-HT3 under the current 5-HT classification, but the observed effects are consistent with stimulation of the putative 5-HT4 receptor.