Table 3.
Half‐life of trait evolution toward the optimum (in mya)
| Treatment | |||
|---|---|---|---|
| Trait | Frost | Mild | Heat |
| t ½ ± SD | t ½ ± SD | t ½ ± SD | |
| Seed size | 33.25 ± 28.42 | ||
| Time to germination | 11.85 ± 11.41 | ||
| Growth | |||
| Initial growth rate | 5.40 ± 2.77 * | 6.51 ± 2.34 * | 6.66 ± 2.06 * |
| Maximal growth rate | 0.75 ± 2.74 | 0.14 × 104 ± 9.46 × 104 | 25.75 ± 667.31 |
| (–)Time to fastest growth | 1.21 × 104 ± 70.84 × 104 | 0.55 × 104 ± 30.15 × 104 | 2.9 × 104 ± 142.2 × 104 |
| Asymptotic size | 10.90 ± 4.99* | 15.49 ± 6.88* | 14.31 ± 9.51 |
| Number of leaves S2 | 2.58 ± 2.58 | 3.53 ± 3.36 | 3.52 ± 2.87 |
| Leaf traits | |||
| Leaf area | 12.09 ± 12.09 | 14.23 ± 11.59 | 25.49 ± 11.98 * |
| Specific leaf area | 7.48 ± 2.33* | 5.27 ± 1.78* | 5.25 ± 1.94 * |
| Leaf dry matter content | 2.62 ± 3.19 | 2.86 ± 2.71 | 349.66 ± 2.51 × 104 |
| Leaf thickness S2 | 9.55 ± 2.56* | 7.33 ± 1.57* | 7.81 ± 1.29 * |
| Leaf dissection index | 15.07 ± 6.87* | 23.36 ± 16.50 | 11.44 ± 8.89 |
| Thermal tolerance s.l. | |||
| Frost resistance | |||
| …acclimated (1 h at –6°C) S1 | 0.75 ± 1.49 | ||
| …nonacclimated (1 h at –5°C) S2 | 6.24 ± 4.47 | ||
| …acclimated (1 h at –11°C) | 3.43 ± 2.70 | ||
| …nonacclimated (1 h at –10°C) S2 | 1.74 ± 2.38 | ||
| Heat resistance | |||
| …acclimated (1 h at +47°C) S1 | 2.14 ± 2.54 | ||
| …nonacclimated (1 h at +45°C) S2 | 4.15 ± 3.32 | ||
| …acclimated (1 h at +51°C) | 1.69 ± 1.77 | ||
| …nonacclimated (1 h at +50°C) S2 | 24.86 × 104 ± 475.96 × 104 | ||
| Tolerance IGR | 5.74 ± 2.47 * | 5.51 ± 1.75 * | |
| Tolerance MGR | 0.16 ± 0.26 | 0.88 ± 0.70 | |
| Tolerance (–)XMID | 0.16 ± 0.39 | 1.46 ± 0.81 * | |
| Tolerance ASYM | 0.09 ± 0.15 | 0.71 ± 0.49 | |
Values of phylogenetic half‐life (under OU1, or OUM when it was the best model or among the best models by |ΔAICc| ≤ 2 to other such models) are based on ARD models and 100 independent stochastic character maps (full details in Table S3). Values are means ± SD of phylogenetic half‐life in mya for traits within treatments, calculated based on bootstrap replicates (i.e., the random removal of a third of the species, with N = 10,000 simulations per trait within environment). Significance in half‐life (*) was calculated by (mean – 1.64SD) > 0. If not specified, a trait was assessed both in sowing round S1 and in S2.