Unusual Avian Vocal Mechanism Facilitates Encoding of Body Size

In this work we study the sound production mechanism of the raspy sounding song of the white-tipped plantcutter (Phytotoma rutila), a species with a most unusual vocalization. The biomechanics involved in the production of this song, and scaling arguments, allowed us to predict the precise way in which body size is encoded in its vocalizations. We tested this prediction through acoustic analysis of recorded songs, computational modeling of its unusual vocal strategy, and inspection of museum specimens captured across southeastern and south-central South America.

Figure 1

The song and a synthesis by our dynamical model. In the first panel, we show the waveform of the song of Phytotoma rutila. In the second panel, we display a detail of two sound segments. A rapidly decaying fast oscillation is compatible with a pulse being filtered by a dissipative resonator. A pulse, as synthesized by our dynamical model, is shown in the third panel. Time traces representing pulses, filtered by a damped oscillator of the appropriate resonant frequency, are shown in the bottom panel. In this simulation, $({\gamma }_1,{\gamma }_2,r,{\omega }_r)=(9\times {10}^3,18.0,0.3,148.2)$. For the simulation displayed in the bottom panel, $\mu \sim 1.005$.

Figure 2

Quantification of the decay time and frequency of the filter. We display a segment of the sound pressure time trace, after computing its absolute value. The maxima of ten consecutive peaks were computed. We fitted the points with a decreasing exponential function. The temporal differences between consecutive peaks were used to estimate the resonant frequency of the filter. The decaying amplitude of the peaks is attributed to the dissipative nature of the filter.

Figure 3

Computed values of the decay times and filter frequencies for the songs in our study. Notice the remarkably low variation of the resonant frequency for each song. A point represents a song, with its color indicating the singing subject. Each song consists of a sequence of segments, and from each one a set of decaying amplitude peaks can be extracted. Point location indicates the mean decay coefficient and filter frequency for the whole song, and the sizes of the bars indicate their ($2\sigma$) dispersions. The inset shows the sonogram of a song. The black square indicates the selected song.

Figure 4

Testing the predictions of the model. Mean peak frequency as a function of the altitude (top), and tarsus length as a function of the altitude (bottom). The regressions were computed using python’s library “Statsmodels.” The $p$ values of the regressions are listed in the insets, and the 95% prediction intervals are indicated as the gray bands’ widths along the fitted linear functions. The databases are described in the text.