Peripheral 5-HT3 receptor mechanisms are involved in activation of gastrointestinal (GI) mucosal vagal afferent fibres. 5-HT3 receptor mechanisms in the central nervous system (CNS) may be involved in behavioural and reflex motility responses. This study investigates the processing of different sensory inputs in the CNS and the involvement of 5-HT3 receptors at these different levels. In Urethane (1.5 g/kg, i.p.) anaesthetized, splanchnectomized ferrets, the jugular vein was cannulated for intravenous (i.v.) drug injection, and the coeliac axis for intraarterial (i.a.) injection close to the upper GI tract. The carotid artery was intubated with a T-cannula for CNS-directed intracarotid (i.c.) injections. An intragastric cannula was used for fluid distension (40-50 ml), and an oesophageal catheter for balloon distension (2 ml). Efferent fibres were dissected from the right cervical vagus for single-unit recording. Nineteen single vagal efferent fibres were selected, with low frequency resting discharge (2.5 +/- 0.3 impulses/s), but no respiratory or cardiovascular phasic input. All responded rapidly (< 2.5 s) to gastric distension (532 +/- 230% change in firing rate) and oesophageal distension (300 +/- 170%). Gastric distension caused excitation in 14 fibres, inhibition in 4 fibres, and a biphasic response in 1. Oesophageal distension excited 16 and inhibited 3. Discharge was also influenced by i.a. injection of 5-HT or the 5-HT3 receptor agonist 2-methyl 5-HT (10-100 micrograms) in all fibres tested. These responses consisted of rapid (< 2.5 s) and powerful changes in firing rate, with excitation, inhibition or biphasic responses. 65% of responses to i.c. or i.v. injection were opposite in direction to those after close i.a. injection, indicating the activation of a different population of receptors. No differences were seen between effects of i.c. and i.v. injections. The 5-HT3 receptor antagonist granisetron (100 micrograms/kg, i.v.) blocked or reduced efferent responses to 5-HT receptor agonists, whereas responses to gastric and oesophageal distension were unchanged. Thus there is extensive convergence of inputs from gastric and oesophageal mechanoreceptors onto vagal motorneurones. These central effects of mechanical stimuli do not involve 5-HT3 receptor mechanisms. Other 5-HT3 receptor inputs are evident, probably peripherally from GI mucosal afferent fibres and from within the CNS.