We examined the effects of progressive hypercapnia and hypoxia on the efferent neural activity in a whole abdominal expiratory nerve (medial branch of the cranial iliohypogastric nerve (L1) in anesthetized, paralyzed dogs. To eliminate effects of phasic lung and chest-wall movements on expiratory activity, studies were performed in the absence of breathing movements. Progressive hyperoxic hypercapnia and isocapnic hypoxia were produced in the paralyzed animals 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 at a predetermined rate. To quantify abdominal expiratory activity, mean abdominal nerve activity in a nerve burst was computed by integrating the abdominal neurogram and dividing by the duration of the nerve burst. Hypercapnia and hypoxia both increased mean abdominal nerve activity and decreased expiratory duration. In contrast to the ramplike phrenic neurogram, the abdominal neurogram consisted of three phases: an initial rising phase, a plateau phase in which abdominal nerve activity was approximately constant, and a terminal declining phase in which the activity returned to the base-line level. The height of this plateau phase and the rates of rise and decline of abdominal nerve activity all increased with increasing hypercapnia and hypoxia. We conclude that, with proprioceptive inputs constant, both hypercapnia and hypoxia are excitatory to abdominal expiratory neural activity.