Appendix B (Table B1). Results of general linear models using Type I sums of squares, testing the effects of functional group richness, species richness, and abundance on the pollination success of C. arabica and C. canephora.
Model and predictor variables |
df |
SS |
F |
P |
Coffea arabica |
||||
1. Functional richness-pollination |
||||
Pollinator functional richness |
1 |
2390.30 |
21.8 |
0.0002 |
2. Richness-pollination |
||||
Pollinator richness |
1 |
2377.10 |
21.7 |
0.0003 |
3. Abundance-pollination |
||||
Pollinator abundance |
1 |
1638.25 |
14.9 |
0.0012 |
Pollinator richness |
1 |
1023.94 |
9.4 |
0.0071 |
4. Abundance-pollination |
||||
Pollinator abundance |
1 |
1638.25 |
14.9 |
0.0012 |
Pollinator functional richness |
1 |
1137.31 |
10.4 |
0.0050 |
Coffea canephora |
||||
1. Functional richness-pollination |
||||
Pollinator functional richness |
1 |
894.60 |
6.1 |
0.0381 |
2. Richness-pollination |
||||
Pollinator richness |
1 |
840.34 |
5.8 |
0.0429 |
Notes: only significant variables are shown. We changed the order of entered variables as follows: Model 1: functional richness (entered fist), species richness (entered second), and abundance (entered third, followed by the interaction terms among the three variables); Model 2: species richness (entered first), functional richness (entered second), and abundance (entered third, followed by interaction terms); Model 3 (only significant for C. arabica): abundance (entered first), species richness (entered second), and functional richness (entered third, followed by interaction terms); Model 4 (only significant for C. arabica): abundance (entered first), functional richness (entered second), and species richness (entered third, followed by interaction terms). Results of the overall models are given in Table 1 of the main text.