(ii) Factor analysis of behavioral traits
We applied factor analysis to reduce the number of behavioral traits recorded and identify factors that may represent correlated behavioral traits (Budaev 2010). We used principal axis factoring and a criterion of minimum eigenvalue > 1 to extract the factors (Fabrigar et al. 1999; Budaev 2010). Factor scores were based on correlations after Varimax (= orthogonal) rotation of the factors. This method allows extraction of orthogonal factors, and more stable and replicable results in case of small sample size (Budaev 2010). Behavioral traits were considered to contribute significantly to the factors when factor loadings were > 0.80 (Table 2), which is above the recommended threshold of 0.5 or 0.7 for behavioral analyses (Budaev et al. 2010).