1. We set out to elucidate the pharmacological mechanisms by which alpha 2-adrenoceptor and 5-HT-receptor ligands affect the haemodynamic response to acute central hypovolaemia in conscious rabbits. 2. Acute central hypovolaemia was produced by inflating an inferior vena caval cuff so that cardiac output fell at a constant rate of approximately 8.5% of its baseline level per min. 3. Drugs were administered into the fourth cerebral ventricle in either 154 mM NaCl (saline) or 20% w/v 2-hydroxypropyl-beta-cyclodextrin (beta-CDX). After vehicle treatments, the haemodynamic response to acute central hypovolaemia had the usual two phases. During Phase I, systemic vascular conductance fell in proportion to cardiac output so that mean arterial pressure fell by only 8 mmHg. Phase II commenced when cardiac output had fallen to approximately 60% of its baseline level, when vascular conductance rose abruptly and arterial pressure fell to < or = 40 mmHg. The haemodynamic response was not dependent on the vehicle used (saline or beta-CDX). 4. Methysergide delayed the occurrence of Phase II in a dose-dependent manner, and prevented it at a dose of 30- 600 nmol (geometric mean = 186 nmol). The effects and potency of methysergide were not dependent on the vehicle used, indicating that beta-CDX can be used as a vehicle for fourth ventricular administration of lipophilic drugs to conscious rabbits. Clonidine (10 nmol) reversed the effects of a critical dose of methysergide. 5. Phase II was also prevented by 8-hydroxy-2-(di-n-propylamino)tetralin (5-HT1A-selective agonist, geometric mean critical dose (range) = 13.1 (10-30) nmol), sumatriptan (5-HT1D-selective agonist, 72.1 (10-300) nmol), mesulergine (5-HT2/1C-selective antagonist, 173 (30-1000) nmol), idazoxan (alpha 2-adrenoceptor-selective antagonist, 548 (100-3000) nmol), and mianserin (5-HT2/1C-selective antagonist, 548 (100-3000) nmol). It was not affected by MDL 72222 (5-HT3-selective antagonist, 300 nmol) or ketanserin (5-HT2/1C-selective antagonist, 3000 nmol). 6. To characterize the nature of alpha 2-adrenoceptors in rabbit brainstem, we examined the binding of [3H]-rauwolscine to membrane homogenates of whole brainstem. [3H]-rauwolscine bound to a population of sites with the characteristics of alpha 2A-adrenoceptors. 7. From these results we suggest that activation of 5-HT1A receptors in the brainstem can prevent Phase II of the response to acute central hypovolaemia in conscious rabbits. Our results do not support the notion of an endogenous 5-hydroxytryptaminergic mechanism mediating Phase II. The mechanism by which the alpha 2-adrenoceptor antagonists yohimbine and idazoxan prevent Phase II remains to be elucidated. However, their potency relative to other 5-HT-receptor ligands indicates that an agonist action at 5-HT1A-receptors is more likely than an antagonist action at alpha 2-adrenoceptors.