Models with different random structures compared with a GLS Model.

The response variable used was number of fledglings. Data set indicates whether the data refer to the breeding success of pairs, females or males. Random term indicates the random structure depicting the model (all models had the same fixed term structure). AIC is the Akaike Information Criterion and AICc is the same corrected for small sample sizes. n stands for the number of data included and K for the number of parameters of the model. ΔAICc is the increase in AICc units respective to the best model, and wi the Akaike weight, or the weight of evidence in favour of that model being the best model in the set. The best model in every set is marked in bold.