Appendix A. Results of multiple regression models.
TABLE A1. Results of multiple regression models testing the effects of the host tree Cecropia schreberiana gender, basal area, height, and interaction factors on Phoradendron hexastichum mistletoe infection (Ordinal Logistic Regression), P. hexastichum loads per tree (Standard Least Square), total bird visits to uninfected Cecropia trees (Standard Least Square), and visits by Spindalis and Nesospingus birds to uninfected Cecropia trees (Standard Least Square).
Variable |
Probability of infection (logistic) |
Load of P. hexastichum (SLS) |
Total visitation (12 bird spp.) (SLS) |
Visitation by Spindalis and Nesospingus (SLS) |
Estimate (F ratio) |
Estimate (F ratio) |
|||
Cecropia gender |
- 0.940 (7.73**) |
0.282 (12.82***) |
0.023 (6.90*) |
0.294 (10.05***) |
Basal area |
0.0 00 (0.35) |
0.000 (0.91) |
0.041 (19.9***) |
- |
Height |
-0.360 (8.51**) |
0.093 (16.83***) |
-0.032 (0.99) |
- |
Gender × Height |
0.017 (0.03) |
- |
0.073 (7.43*) |
- |
Gender × Basal area |
-0.001(3.51^{m}) |
0.002 (4.00*) |
- |
- |
Basal Area × Height |
-0.000 (2.30) |
0.000 (2.82) |
- |
- |
Pair (bird visitation columns only) |
- |
- |
||
Fruit crop (bird visitation columns only) |
- |
- |
||
Model r^{2} |
0.27 |
0.36 |
0.60 |
0.26 |
Note: All least square models were selected using a backward stepwise procedure (all variables and interactions inputted) with probability to remove a variable set to 0.1 (stepped out variables are denoted by a dash). See Figs. 1, 2, and 4 in the main article for plots of relationships.
m = P ≤ 0.06; * = P ≤ 0.05; ** = P ≤ 0.01; *** = P ≤ 0.00.