In humans, attenuating carotid chemoreceptor activity by hyperoxia does not alter arterial PCO2 (PaCO2) during submaximal exercise, yet a transient hypercapnia occurs in carotid chemoreceptor-resected (CBR) asthmatic subjects during submaximal exercise. We hypothesized that this difference was due to asthma and not CBR causing the abnormal response. Accordingly, we determined the temporal pattern of PaCO2 during mild and moderate exercise in chemoreceptor-intact asthmatic (n = 10) and nonasthmatic subjects (n = 10). We also hypothesized that hyperoxia alters PaCO2 during exercise if exercise already has disrupted PaCO2 homeostasis. Accordingly, we studied, during exercise, asthmatic subjects while hyperoxic; nonasthmatic subjects during loaded breathing of room air, which increased PaCO2; and nonasthmatic subjects during loaded breathing while hyperoxic. While breathing room air, neither asthmatic nor nonasthmatic subjects maintained arterial isocapnia during exercise. An increase in PaCO2 between rest and exercise and between mild exercise and 1st min of moderate exercise was greater in asthmatic than in nonasthmatic subjects (P < 0.05). In six asthmatic subjects that were hypercapnic breathing room air during exercise, hypercapnia was accentuated by hyperoxia. The ventilatory load in nonasthmatic subjects resulted in a work load-dependent hypercapnia (P < 0.01) accentuated (P < 0.01) by hyperoxia. We conclude that normally in humans the carotid chemoreceptors contribute minimally to the hyperpnea of submaximal exercise. However, when PaCO2 is increased from resting values during exercise, then the chemoreceptors serve to augment ventilation and thereby minimize the hypercapnia.