The most physiologically important sensors for systemic glucoregulation are located in extracranial sites. Recent evidence suggests that the carotid body may be one such site. We assessed rat carotid body afferent neural output in response to lowered glucose, indirectly by measurement of ventilation, and directly by recording single or few-fibre chemoafferent discharge, in vitro. Insulin (0.4 U kg-1 min-1)-induced hypoglycaemia (blood glucose reduced by ca 50% to 3.4 ± 0.1 mmol l-1) significantly increased spontaneous ventilation (̇VE) in sham-operated animals but not in bilateral carotid sinus nerve sectioned (CSNX) animals. In both groups, metabolic rate (measured as ̇VO2 was almost doubled during hypoglycaemia. The ventilatory equivalent (̇ VE/̇ VO2) was unchanged in the sham group leading to a maintained control level of Pa,CO2, but ̇ VE/̇ VO2 was significantly reduced in the CSNX group, giving rise to an elevation of 6.0 ± 1.3 mmHg in Pa,CO2. When pulmonary ventilation in sham animals was controlled and maintained, phrenk neural activity increased during hypoglycaemia and was associated with a significant increase in Pa,CO2 of 5.1 ± 0.5 mmHg. Baseline chemoreceptor discharge frequency, recorded in vitro, was not affected, and did not increase when the superfusate [glucose] was lowered from 10 mm to 2 mm by substitution with sucrose: 0.40 ± 0.20 Hz to 0.27 ± 0.15 Hz, respectively (P > 0.20). We suggest therefore that any potential role of the carotid bodies in glucose homeostasis in vivo is mediated through its transduction of some other metabolically derived blood-borne factor rather than glucose per se and that this may also provide the link between exercise, metabolic rate and ventilation.