Table 1.
Multivariate PGLS regression results for the shape PCs on size and diet.
| Shape PC1 | R2 = 0.570 | ||
|---|---|---|---|
| b | r | P | |
| CL | 0.0000 | 0.0022 | 0.9886 |
| Diet PC1 | 0.0303 | 0.2242 | 0.1340 |
| Diet PC2 | -0.0903 | 0.5631 | < 0.0001 |
| Diet PC3 | 0.1162 | 0.6312 | < 0.0001 |
| Shape PC2 | R2 = 0.807 | ||
| b | r | P | |
| CL | -0.0010 | -0.2116 | 0.1581 |
| Diet PC1 | 0.0190 | 0.2998 | 0.0429 |
| Diet PC2 | -0.1047 | -0.8650 | < 0.0001 |
| Diet PC3 | -0.0754 | -0.7549 | < 0.0001 |
| Shape PC3 | R2 = 0.444 | ||
| b | r | P | |
| CL | -0.0008 | -0.1592 | 0.2907 |
| Diet PC1 | 0.0545 | 0.6420 | < 0.0001 |
| Diet PC2 | 0.0079 | -0.1201 | 0.4264 |
| Diet PC3 | -0.0226 | -0.3049 | 0.0394 |
| Shape PC4 | R2 = 0.440 | ||
| b | r | P | |
| CL | -0.0020 | -0.4482 | 0.0018 |
| Diet PC1 | -0.0233 | -0.4031 | 0.0054 |
| Diet PC2 | -0.0003 | -0.0056 | 0.9705 |
| Diet PC3 | -0.0016 | -0.0273 | 0.8573 |
| Shape PC5 | R2 = 0.159 | ||
| b | r | P | |
| CL | -0.0008 | -0.1974 | 0.1884 |
| Diet PC1 | 0.0151 | 0.2877 | 0.0525 |
| Diet PC2 | -0.0007 | -0.0142 | 0.9255 |
| Diet PC3 | 0.0140 | 0.2484 | 0.0960 |
The partial regression (b) and partial correlation (r) coefficients and their significances (P) are depicted for each independent variable: CL - Condylobasal Length, and diet principal components. Diet PC1 is a contrast between animalivores (positive) and frugivores (negative). Diet PC2 separates sanguivores (positive scores) from the rest. Diet PC3 separates nectarivores (positive) from the rest. The variance inflation factors were small, ranging from 1.249 (Diet PC1) to 1.007 (Diet PC2). The multivariate model was significant (Wilk's Λ = 0.0092, F = 15.23, df1 = 25, df2 = 150.1, P < 0:00001).