We studied the effect of an experimentally augmented breath on the duration of expiration (TE) to test the hypothesis that within-breath reductions in alveolar partial pressure of CO2 (PACO2) delay the onset of the subsequent inspiration through a carotid chemoreceptor mechanism. Studies were performed in five carotid body intact (CBI) and three carotid body denervated (CBD) dogs during quiet wakefulness and nonrapid-eye-movement sleep. The vagus nerves were blocked in order to abolish vagal mechanisms capable of modulating TE. In CBI dogs, mechanically augmented breaths [mean tidal volume (VT), 149% of control] induced a mean decrease in PACO2 of 3.3 Torr and prolonged TE by 2.21 s to 127% of control (P less than 0.001). In contrast, in CBD dogs augmented breaths of comparable magnitude failed to change TE. Spontaneous sighs in fully intact dogs (mean VT, 369% of control) produced marked prolongation of TE (to 429% of control). In CBI dogs, there was a direct relationship between VT and the resulting TE across all experimental conditions (control, vagal blockade, augmented breaths, spontaneous sighs), whereas in CBD dogs, changes in VT failed to alter TE. The results indicate that the carotid bodies are able to induce within-breath alterations in TE and that such a mechanism can account for the prolongation of TE by vagal blockade and by spontaneous sighs.