Fig 1.
Oscillograms (whole sound and enlarged section), spectrogram and power spectrum (750 point FFT; Hanning window; 64 Hz filter bandwidth) for representative examples of; a) Gru1 sound, b) Gru2 sound.
Fig 2.
Oscillograms (whole sound and enlarged section), spectrogram and power spectrum (750 point FFT; Hanning window; 64 Hz filter bandwidth) for representative examples of; a) Short-growl sound and, b) Long-growl sound. Amplitude and power are shown as relative units.
Table 1.
Classification of sound types to four categories by linear discriminant analysis.
Table 2.
Linear discriminant functions for the validation of sound type classification.
SD = sound duration; PN = number of pulses; PD = pulse duration; PP = pulse period; PI = pulse interval; PRR = pulse repetition rate; Pf = peak frequency; Lf = lower frequency; Uf = upper frequency 3dBBW = 3 dB bandwidth; Q = damping coefficient.
Table 3.
Acoustic features measured in Gru1, Gru2, Short-growl and Long-growl sounds.
Values are means ± SEM and range is given in parentheses below. For Pf, Lf and Uf the median value is given in parentheses on the right. SD = sound duration; PN = number of pulses; PD = pulse duration; PP = pulse period; PI = pulse interval; PRR = pulse repetition rate; Pf = peak frequency; Lf = lower frequency; Uf = upper frequency 3dBBW = 3 dB bandwidth; Q = damping coefficient. * = significant difference between means within each of the two different types of grunts and growls using t-tests at (P < 0.05) with sequential Holm-Bonferroni correction for Type I error inflation.
Fig 3.
Diel variation in the proportion of grunt and growl type sounds.
Grey vertical bar are growls and black vertical bar are grunts. Bar represents mean values observed in the hourly interval over five non-consecutive days of sampling. Crossed bar at 0900–1000 hrs indicate no data due to activation of water flow.
Fig 4.
Daily pattern of vocal activity showing mean hourly vocalisation rate (mean ± SEM; n = 5).
Black bar represents grunt sounds and grey bar represents growl sounds. The single shaded vertical bar at 0900–1000 hrs indicates the activation of water flow.
Fig 5.
Comparison of vocalisation rate (mean±SEM; n = 5) during different periods of the day.
Dark square represent grunt sound and white square represent growl sounds. Different letters indicate statistically significant differences between individual means within either growls or grunts (Tukey test; p<0.05).
Fig 6.
Total vocalisation counted in consecutive 15 min periods from 2 h prior to feeding, during feeding and 2 h after feeding commenced.
5a) Grunts and 5b) Growls. Counts of vocalisations were conducted for 5 non-consecutive days. Solid line and symbol = midnight; dotted line and hollow symbol = noon feeding. Vertical arrow indicates time of feeding. Data represent the mean values ± SEM.