Table 2. Model selection for survival (S), recapture (p) and emigration (psi) probabilities.

Model	K	AICc	ΔAICc	AICc Weights	Model Likelihood	Deviance
1 {S(state+sr+d)p(t+sex)psi(sex+sr+d)}	22	9704.53	0	0,199	1	9659.86
2 {S(state+sr)p(t+sex)psi(sex+sr+d)}	21	9705.04	0.512	0,154	0.774	9662.44
3 {S(state)p(t+sex)psi(sex+sr+d)}	20	9705.30	0.775	0,135	0.679	9664.75
4 {S(state+d+sr)p(t+sex)psi(sex*sr+d)}	23	9705.43	0.901	0,127	0.637	9658.70
5 {S(state+d)p(t+sex)psi(sex+sr+d)}	21	9706.08	1.550	0,092	0.461	9663.47
6 {S(state+d+sr)p(t+sex)psi(sex*d+sr)}	23	9706.59	2.061	0,071	0.357	9659.86
7 {S(state+sr+d)p(t+sex)psi(sex+d)}	21	9707.15	2.624	0,054	0.269	9664.55
8 {S(state+d+sr)p(t+sex)psi(sex*(sr+d))}	24	9707.49	2.964	0,045	0.227	9658.70
9 {S(state+sr)p(t+sex)psi(sex*(sr+d))}	23	9707.95	3.421	0,036	0.181	9661.23
10 {S(state*sr+d)p(t+sex)psi(sex*(sr+d))}	25	9708.90	4.374	0,022	0.112	9658.05
11 {S(state+d)p(t+sex)psi(sex*(sr+d))}	23	9709.04	4.508	0,021	0.105	9662.31
12 {S(state*d+sr)p(t+sex)psi(sex*(sr+d))}	25	9709.45	4.919	0,017	0.086	9658.59
13 {S(state*(sr+d))p(t+sex)psi(sex*(sr+d))}	26	9710.69	6.158	0,009	0.046	9657.76
14 {S(state)p(t+sex)psi(sex*d)}	20	9710.89	6.366	0,008	0.042	9670.34
15 {S(state+sr+d)p(t+sex)psi(sex+sr)}	21	9711.90	7.370	0,005	0.025	9669.29
16 {S(state+sr)p(t+sex)psi(sex+sr)}	20	9714.06	9.538	0,002	0.009	9673.52
17 {S(state)p(t+sex)psi(sex)}	18	9715.29	10.763	0,001	0.005	9678.84
18 {S(state)p(t+sex)psi(sex*sr)}	20	9716.33	11.802	0,001	0.003	9675.78
19 {S(state+sex)p(t+sex)psi(sex)}	19	9716.90	12.376	0	0.002	9678.41
20 {S(state)p(t+sex)psi(.)}	17	9718.48	13.949	0	0.001	9684.08

Factors: state (resident or disperser), t = time (day), and sex (male or female).

Covariates: sr = sex-ratio and d = density.

K = number of parameters.

Competing best models, having ΔAICc values lower than 2, are models 1 to 5 and the one including all significant effects is model 3 in bold (see Table 3 for LRT statistics associated). The 20 best models are shown.