The functional significance of chemoreflexive hypoxic bradycardia was explored in Atlantic cod Gadus morhua L. (mean mass approximately 800 g, acclimated to a seawater temperature of 11 degrees C) by investigating responses to progressive hypoxia following section of the cardiac branches of cranial nerve X. Cardiac denervation had no effect on oxygen uptake rate (MO(2)), gill ventilation rate (f(G)) or opercular pressure amplitude (P(OP)) under normoxic conditions, but caused a significant increase in heart rate (f(H)), to 50+/-1 beats min(-1) by comparison to 40+/-2 beats min(-1) in sham-operated cod (mean+/-s.e.m., n=9). Sham-operated cod exhibited transient profound bradycardia following oxygen chemoreceptor stimulation by bolus injection of sodium cyanide into the buccal cavity (2 mg in 2 ml seawater), but this cardiac chemoreflex was abolished in denervated cod. Both groups, however, exhibited similar marked transient chemoreflexive hyperventilation following NaCN. When exposed from normoxia (PO(2) approximately 18 kPa) to progressive hypoxia at nominal water PO(2)'s of 8, 6, 5, 4 and 3 kPa, both groups exhibited the same pattern of homeostatic regulation of MO(2), with no significant difference in their mean critical PO(2) (P(crit)) values, which were 7.40+/-0.81 kPa and 8.73+/-0.71 kPa, respectively (n=9). Both groups exhibited significant bradycardia during progressive hypoxia, although denervated fish always had higher mean f(H). The incipient threshold for bradycardia coincided with P(crit) in sham-operated cod whereas, in denervates, the threshold was below their P(crit) and bradycardia presumably reflected direct effects of hypoxia on the myocardium. The sham-operated group displayed a significantly more pronounced ventilatory response than denervates in hypoxia, in particular for P(OP). In sham-operated cod, peak ventilatory responses occurred in deep hypoxia below P(crit) whereas, in denervates, more modest peak responses coincided with P(crit) and, in deep hypoxia, they exhibited a significant decline in f(G) below their normoxic rate. Only a minority of shams lost equilibrium in hypoxia whereas a majority of denervates did, some of which failed to recover. The results indicate that chemoreflexive bradycardia plays no role in the homeostatic regulation of oxygen uptake by cod in hypoxia, but does contribute to maintenance of overall functional integrity below P(crit).