We determined the effects of denervating the hilar branches (HND) of the vagus nerves on breathing and arterial PCO2 (PaCO2) in awake ponies during eupnea and when inspired PCO2 (PICO2) was increased to 14, 28, and 42 Torr. In five carotid chemoreceptor-intact ponies, breathing frequency (f) was less, whereas tidal volume (VT), inspiratory time (TI), and ratio of TI to total cycle time (TT) were greater 2-4 wk after HND than before HND. HND per se did not significantly affect PaCO2 at any level of PICO2, and the minute ventilation (VE)-PaCO2 response curve was not significantly altered by HND. Finally, the attenuation of a thermal tachypnea by elevated PICO2 was not altered by HND. Accordingly, in carotid chemoreceptor-intact ponies, the only HND effect on breathing was the change in pattern classically observed with attenuated lung volume feedback. There was no evidence suggestive of a PCO2-H+ sensory mechanism influencing VE, f, VT, or PaCO2. In ponies that had the carotid chemoreceptors denervated (CBD) 3 yr earlier, HND also decreased f, increased VT, TI, and TT, but did not alter the slope of the VE-PaCO2 response curve. However, at all levels of elevated PICO2, the arterial hypercapnia that had persistently been attenuated, since CBD was restored to normal by HND. The data suggest that during CO2 inhalation in CBD ponies a hilar-innervated mechanism influences PaCO2 by reducing physiological dead space to increase alveolar ventilation.