Table 3.
Evolutionary transition model selection in a likelihood framework for stochastic character mapping (SCM) using the corrected Akaike information criterion (AICc).
| Character | Model | lnL | AICc | dAICc | wAICc |
|---|---|---|---|---|---|
| Habit | ARD | −3.5849 | 11.169853 | 1.61901 | 0.18203 |
| SYM | −3.7754 | 9.550848 | 0 | 0.40898 | |
| ER | −3.7754 | 9.550848 | 0 | 0.40898 | |
| Leaf arrangement | ARD | −9.1369 | 30.27383 | 6.54656 | 0.02992 |
| SYM | −10.341 | 26.6812 | 2.95393 | 0.18033 | |
| ER | −10.864 | 23.72727 | 0 | 0.78976 | |
| Calyx teeth base | ARD | −2.7632 | 9.526412 | 1.13493 | 0.22087 |
| SYM | −3.1957 | 8.391486 | 0 | 0.38957 | |
| ER | −3.1957 | 8.391486 | 0 | 0.38957 | |
| Inflorescence | ARD | −2.585 | 9.16992 | 1.08775 | 0.22495 |
| SYM | −3.0411 | 8.082168 | 0 | 0.38752 | |
| ER | −3.0411 | 8.082168 | 0 | 0.38752 | |
| Anther texture | ARD | −3.9023 | 11.804559 | 1.8105 | 0.16821 |
| SYM | −3.997 | 9.994056 | 0 | 0.4159 | |
| ER | −3.997 | 9.994056 | 0 | 0.4159 | |
| Capsule shape | ARD | −2.5665 | 9.133063 | 1.19576 | 0.21568 |
| SYM | −2.9687 | 7.937306 | 0 | 0.39216 | |
| ER | −2.9687 | 7.937306 | 0 | 0.39216 | |
| Capsule texture | ARD | −6.9027 | 25.80548 | 8.84777 | 0.01005 |
| SYM | −7.1922 | 20.38433 | 3.42662 | 0.1512 | |
| ER | −7.4789 | 16.95771 | 0 | 0.83875 | |
| Capsule wing | ARD | −2.5665 | 9.133063 | 1.19576 | 0.21568 |
| SYM | −2.9687 | 7.937306 | 0 | 0.39216 | |
| ER | −2.9687 | 7.937306 | 0 | 0.39216 | |
| Capsule dehiscence | ARD | −2.5665 | 9.133063 | 1.19576 | 0.21568 |
| SYM | −2.9687 | 7.937306 | 0 | 0.39216 | |
| ER | −2.9687 | 7.937306 | 0 | 0.39216 | |
| Seed wing | ARD | −8.0014 | 28.00270 | 8.84776 | 0.01005 |
| SYM | −8.2908 | 22.58155 | 3.42661 | 0.1512 | |
| ER | −8.5775 | 19.15494 | 0 | 0.83875 |
lnL, log-likelihood of the model; AICc, corrected Akaike information criterion; dAICc, the difference between the AICc of the model and the best model; wAICc, AICc weight of the model. The dAICci, AICci – min AICc follows the following criteria: dAICci ≤ 1, substantial support for a model with AICci; 4 ≤ dAICci ≤ 9, low support for a model with AICci; dAICci > 10, no support for a model with AICci. The AICc weights indicate the probability that the model is the best among the whole set of candidate models (Burnham and Anderson, 2002; Sullivan and Joyce, 2005). ER, equal rates model; SYM, symmetrical model; and ARD, all-rates-different model. Bold font in the Model columns indicates the selected model by the likelihood ratio test.