Habitat-related differences in song structure and complexity in a songbird with a large repertoire

BMC Ecology

Table 2 General linear models assessing variation in song thrush song characteristics that differed significantly between studied habitats

Model	AIC_C	ΔAIC_C	w _i	ER
Whistle minimum frequency
HABITAT	740.66	0.00	0.39
HABITAT + MALES	742.05	1.40	0.19	2.01
HABITAT + HOUR	742.10	1.44	0.19	2.05
Whistle peak frequency
HABITAT	776.56	0.00	0.34
HABITAT + HOUR	777.53	0.97	0.21	1.62
HABITAT + DAY	777.79	1.23	0.18	1.85
HABITAT + DAY + HOUR	778.32	1.76	0.14	2.41
HABITAT + NOISE	778.40	1.84	0.13	2.51
Syllable repertoire
NOISE + DAY	684.54	0.00	0.32
NOISE	685.04	0.50	0.25	1.28
NOISE + DAY + HOUR	685.72	1.18	0.18	1.80
NOISE + DAY + MALES	686.24	1.70	0.14	2.34
NOISE + MALES	686.43	1.89	0.12	2.58
Whistle repertoire
NOISE	567.04	0.00	0.30
NOISE + MALES	567.17	0.13	0.29	1.07
NOISE + HOUR	568.48	1.43	0.15	2.05
NOISE + DAY + HOUR	568.74	1.69	0.13	2.33
NOISE + DAY	568.76	1.72	0.13	2.36
Twitter repertoire
NOISE + DAY	664.54	0.00	0.41
NOISE + DAY + HOUR	666.03	1.49	0.19	2.11
Twitter fraction
NOISE + DAY	736.76	0.00	0.27
NOISE	737.13	0.37	0.22	1.20
HABITAT + DAY	737.42	0.66	0.19	1.39
DAY	737.58	0.83	0.18	1.51
HABITAT	738.10	1.34	0.14	1.96
Linearity index
NOISE	− 85.43	0.00	0.36
NOISE + DAY	− 84.26	1.17	0.20	1.80
NOISE + HOUR	− 83.78	1.65	0.16	2.28
NOISE + MALES	− 83.73	1.70	0.15	2.34

Models with the highest probability (ΔAIC_C < 2) are shown. The Akaike weight (w_i) and evidence ratio (ER) were calculated on the basis of Akaike’s Information Criterion corrected for small sample size (see “Methods” for details). Predictor codes: DAY, day of season; HOUR, hour after sunrise; NOISE, background noise level; HABITAT, habitat type; MALES, other singing males in hearing range during recording

ISSN: 1472-6785