The acoustic role of the enlarged, bony, nasal cavities and rigid tracheal chambers in the horseshoe bat, Rhinolophus hildebrandti (Fig. 2) was investigated by determining the effect of their selective filling on the nasally emitted sonar pulse and on the sound traveling backwards down the trachea. Normal sonar signals of this bat contain a long constant frequency component with most energy in the second harmonic at about 48 kHz. The fundamental is typically suppressed 20 to 30 dB below the level of the second harmonic (Fig. 1). None of the experimental manipulations described affected the frequency of the sonar signal fundamental. Filling the dorsal and both lateral tracheal chambers had little effect on the emitted vocalization, but caused the level of the fundamental component in the trachea to increase 15 to 19 dB in most bats (Table 2). When only the dorsal chamber or only the two lateral chambers were filled, the effect was less striking and more variable (Tables 3 and 4), suggesting that the tracheal fundamental is normally suppressed by acoustic interaction between these three cavities. Filling the enlarged dorsal nasal cavities had no effect on the tracheal sound. The effect of this treatment on the nasally emitted sonar pulse was inconsistent. Sometimes the fundamental increased 10 to 12 dB, other times the intensity of all harmonics decreased; in still other cases the second, third or fourth harmonic increased, but the fundamental remained unchanged (Tables 5, 6, and 7). When bats were forced to vocalize through the mouth, by sealing the nostrils, there was a prominent increase in the level of the emitted fundamental (10 to 21 dB) and in the fourth harmonic (6 to 17 dB). In one instance there was also a significant increase in the level of the third harmonic (Tables 8 and 9). The supraglottal tract thus filters the fundamental from the nasally emitted sonar signal, although the role of the inflated nasal cavities in this process is unclear. We conclude that a high glottal impedance acoustically isolates the subglottal from the supraglottal vocal tract. The tracheal chambers do not affect the emitted sonar signal, but may attenuate the fundamental in the trachea and prevent it from being reflected from the lungs back towards the cochlea. It may be important to prevent the reflected fundamental from stimulating the cochlea, via tissue conduction, along multiple indirect pathways which would temporally smear cochlear stimulation.(ABSTRACT TRUNCATED AT 400 WORDS)