The potential direct and modulating effects of acute hypoxia on catecholamine secretion in rainbow trout (Oncorhynchus mykiss) were assessed in situ, using a perfused cardinal vein preparation, and in vivo, using chronically cannulated fish. Acute (10 min) perfusion with hypoxic (P(O2)<10 mmHg) saline or homologous hypoxic blood did not have a statistically significant effect on basal (non-stimulated) catecholamine secretion. A field stimulation technique was used to excite the sympathetic nerves innervating the chromaffin cells electrically in situ under conditions of high-P(O2) (saline P(O2)=152 mmHg; 1 mmHg=0.133 kPa) or low-P(O2) (saline P(O2)<10 mmHg) perfusion at constant P(CO2) (2.3 mmHg). The results demonstrated that neuronally evoked catecholamine secretion was significantly lowered by 50 % during perfusion with hypoxic saline. To assess whether the inhibitory effect of hypoxia during neuronal stimulation in situ resulted from modulation of nicotinic and/or muscarinic receptor-linked pathways, perfused posterior cardinal vein preparations were injected with selective nicotinic (10(-)(7) or 10(-)(6 )mol kg(-)(1) nicotine) or muscarinic (10(-)(3 )mol kg(-)(1) methacholine) receptor agonists. For both doses of nicotine, catecholamine secretion was significantly lowered during hypoxia by 55 %. During muscarinic receptor stimulation, perfusion with hypoxic saline caused a 42 % reduction in the rate of catecholamine secretion. In contrast, catecholamine secretion elicited by depolarising levels of KCl (60 mmol l(-)(1)) was unaffected by the oxygen status of the perfusate. In vivo, intra-arterial injections of nicotine (300-600 nmol kg(-)(1)) into normoxic (water P(O2)=155 mmHg) or moderately hypoxic fish (water P(O2)=80 mmHg) caused a dose-dependent elevation of circulating catecholamine levels. However, despite the inhibitory influence of localised hypoxia on chromaffin cell responsiveness previously demonstrated in situ, the increase in plasma catecholamine levels after intra-arterial injection of nicotine was significantly enhanced in the hypoxic fish. The differences between the results from the in vivo and in situ experiments may reflect the contribution of higher control centres and modulating factors in vivo that are absent in situ.