Elevations in systemic metabolism brought about by exercise will raise the venous partial pressure of CO₂ and decrease venous Po₂ and pH. Despite these changes in venous blood composition, steady state arterial blood gas tensions and pH are retained at normal, ‘resting’ levels by a proportional increase in alveolar ventilation, thus demonstrating the sensitivity and functional importance of the respiratory system to changes in blood chemistry. The means by which metabolic rate is ‘sensed’ to initiate the appropriate hyperpnoea is, however, not known with certainty and a number of mechanisms have been implicated. These include both neural and humoural mechanisms, implicate both feedforward and feedback mechanisms and even include the possibility that exercise hyperpnoea is a learnt phenomenon. Additionally, a degree of redundancy may occur in the system with each proposed mechanism apparently able to account for most, if not all, of the ventilatory responses observed. We have raised metabolic rate in anaesthetised rats by means of insulin infusion and have shown that such a manoeuvre can raise ventilation without alteration in arterial blood gas tensions via alteration in carotid body CO₂^1,2. In the absence of carotid bodies, elevations in metabolic rate led to excessively high falls in blood pH. Additional experiments we performed on in vitro carotid body preparations showed that this organ was not sensitive to a fall in glucose concentration and afferent chemodischarge was not elevated, but could even be decreased by severe hypoglycaemia. We subsequently suggested that circulating adrenaline might act to augment rat carotid body sensitivity and other studies have revealed that a number of other hormones may account for the response seen in humans³. Elevations in ventilation and CO₂ sensitivity may be observed even with the relatively small increase in post-prandial metabolism. However, in contrast, to our findings, there are other reports to show that the rat carotid body has an in vitro sensitivity to falls in glucose concentration and it has been suggested that it may act as a peripheral glucosensor⁴, although the mechanism for sensing glucose appears to involve carotid body type 1 cell membrane channel activation rather than the inactivation known to occur during hypoxia. The reasons for the discrepancies are not known, but it has been suggested that an interaction between Po₂ and glucose may be important⁵ or that the particular preparation used can influence the findings. Whilst evidence to date regarding some role for the carotid body in mediating exercise hyperpnoea appears strong, more studies are warranted to establish precisely how metabolism is sensed so precisely.