Table 3. Generalized linear mixed models relating size of prey consumed by black-tailed godwits with prey density, culmen length, sex and date.

Scrobicularia plana
Model selection	df	AICc	Test	L. ratio	p-value
1 Prey density+Culmen*Sex+Date	8	2542.2
2 Prey density+Culmen+Sex+Date	7	2542.1	2 vs 1	1.031	0.310
3 Prey density+Sex+Date	6	2541.1	3 vs 2	0.0003	0.986
4 Sex+Date	5	2540.2	4 vs 3	0.029	0.865
5 Date	4	2540.8	5 vs 4	1.663	0.197
6 Null model	3	2557.8	6 vs 5	18.018	0.000

The table presents comparisons of increasingly simpler nested models, using likelihood-ratio tests (L.ratio) and coefficients of best significant models, fitted by Restricted Maximum Likelihood. Bird identity was treated as a random factor and “*” stands for interaction between variables. The best model is presented in italics and the estimation of the coefficients (estimate) is given with standard errors (SE). AIC values corrected for finite sample sizes (AICc) are presented for comparative purposes only (differences in AICc values lower than 2 between candidate models were not valorised; [59]). t-tests and the respective p-values are used to test the significance of each tern of the final model.