In the spontaneously breathing animal, respiratory responses to chemical stimuli are influenced by phasic proprioceptive inputs from the thorax. We have compared the effects of hypercapnia and hypoxia on the level and timing of phrenic nerve activity while these phasic afferent signals were absent. Progressive hyperoxic hypercapnia and isocapnic hypoxia were produced in anesthetized paralyzed dogs by allowing 3-5 min of apnea to follow mechanical ventilation with 100% O2 or 35% O2 in N2, respectively; during hypoxia, isocapnia was maintained by intravenous infusion of tris(hydroxymethyl)aminomethane buffer. The peak height (P) of nerve bursts, inspiratory time (TI), and expiratory time (TE) were measured from the phrenic neurogram. With the vagi intact or severed, hypoxia decreased TI, whereas hypercapnia did not; both stimuli decreased TE. At the same minute phrenic activity (P x frequency), P, TI, and TE were all less during hypoxia than during hypercapnia. The decreases in TI and TE with hypoxia were significantly less after carotid sinus denervation. The results indicate that the patterns of phrenic nerve activity in response to hypoxia and hypercapnia are different: hypoxia has a greater effect on respiratory timing, whereas hypercapnia has a greater effect on peak phrenic nerve activity. The effect of hypoxia on respiratory timing is largely mediated by the peripheral chemoreceptors.