Table 2. Results of the correlation analysis between morphological features (Table 1) and key behavioral traits. Significant positive correlations are indicated in green, negative in red, and the behavior(s) most strongly correlated with a morphological feature, is in bold. Results of PC1 follow the format: P-value (positive pairs: negative pairs). Results of PC2 follow the format: P-value (positive pairs: negative pairs: neutral pairs). Positive pairs represent phylogenetically independent pairs of taxa that contrast in both the morphological feature and the behavior in a pattern indicating paired loss or paired gain of this feature and behavior. Negative pairs show the opposite pattern, indicating that when one character is lost the other is gained, or vice-versa. In neutral pairs, the phylogenetically independent taxa vary in just one of the two characters (neutral pairs are not included in PC1). In the ML analysis, a delta-AICc of 0 indicates the best-fitting model for that hypothesis, and in alternate models, the larger the delta-AICc value, the worse that model performs relative to the best model. Greater delta-AICc values in the independent model indicate stronger cases of correlation (because this indicates the relatively poor fit of the uncorrelated/independent model).