Ring doves vocalize with their beaks and nostrils closed, exhaling into inflatable chambers in the head and neck region. The source sound produced at the syrinx contains a fundamental frequency with prominent second and third harmonic overtones, but these harmonics are filtered out of the emitted signal. We show by cineradiography that the upper esophagus, oral and nasal cavities collect the expired air during vocalization and that the inflated esophagus becomes part of the suprasyringeal vocal tract. The level of the second and third harmonics, relative to the fundamental frequency (f 0 ), is reduced in the esophagus and emitted vocalization compared with in the trachea, although these harmonics are still considerably higher in the esophagus than in the emitted signal. When the esophagus is prevented from fully inflating, there is a pronounced increase in the level of higher harmonics in the emitted vocalization. Our data suggest that the trachea and esophagus act in series as acoustically separate compartments attenuating harmonics by different mechanisms. We hypothesize that the trachea behaves as a tube closed at the syringeal end and with a variable, restricted opening at the glottal end that lowers the tracheal first resonance to match the f 0 of the coo. The inflated esophagus may function as a Helmholtz resonator in which the elastic walls form the vibrating mass. Such a resonator could support the f 0 over a range of inflated volumes.