Table 2.
Results from the univariate random regression animal model analyses of relative red area from the onset of senescence (9th week onward) in the control and warm winter treatment groups. V I is the between‐individual variance in relative red area and V A is the additive genetic variance. PE × T and G × T denote the permanent environment and additive genetic variances in the rate of senescence. The significance of each variance component was tested by comparing between different hierarchical models (a) based on a likelihood ratio test. For example, V I was tested based on the comparison between model 1 and model 2. The REML estimated variances and covariances between elevation and slope of the best‐fit models are given with their SEs in brackets
| Model selection | ||||||||
|---|---|---|---|---|---|---|---|---|
| Modela | Tested component | d.f. | Control | Warm winter | ||||
| LogL | χ2 | P | LogL | χ2 | P | |||
| 1 | – | −1181·88 | −1196·11 | |||||
| 2 | V I (ind 0) | 1 | −1077·55 | 208·66 | <0·001 | −1055·04 | 282·14 | <0·001 |
| 3 | V A (a 0) | 1 | −1074·27 | 6·56 | 0·010 | −1052·47 | 5·14 | 0·023 |
| 4 | PE × T (pe 1) | 2 | −1024·69 | 99·16 | <0·001 | −985·22 | 134·50 | <0·001 |
| 5 | G × T (a 1) | 2 | −1019·43 | 10·52 | 0·005 | −984·16 | 2·128 | 0·345 |
| RRAM (co)variances of the best‐fit models | ||||
|---|---|---|---|---|
| Variance–covariance | Control (model 5) | Warm winter (model 4) | ||
| Parameter estimate | P | Parameter estimate | P | |
| Vɛ−1 | 9·266 (1·714) | 5·068 (1·068) | ||
| Vɛ−0·667 | 4·408 (0·907) | 4·370 (0·803) | ||
| V ɛ−0·333 | 4·995 (0·853) | 3·223 (0·561) | ||
| V ɛ0 | 6·892 (1·063) | 3·544 (0·562) | ||
| V ɛ0·333 | 2·987 (0·495) | 2·405 (0·404) | ||
| V ɛ0·667 | 1·029 (0·341) | 1·220 (0·351) | ||
| V ɛ1 | 3·423 (0·787) | 3·210 (0·745) | ||
| V pe 0 | 1·924 (1·293) | 3·339 (1·173) | ||
| Cov pe 0, pe 1 | 0·651 (1·030) | 0·527 | −0·863 (0·597) | 0·183 |
| V pe 1 | 1·648 (1·373) | 4·941 (0·862) | ||
| V a 0 | 3·818 (1·891) | 2·337 (1·422) | ||
| Cov a 0, a 1 | −2·685 (1·532) | 0·026 | – | |
| V a 1 | 3·656 (1·921) | – | ||
Model 1: T i,t = μ + timeF + compF + ɛi,t.
Model 2: T i,t = μ + timeF + compF + f(ind 0i ,t) + ɛi,t.
Model 3: T i,t = μ + timeF + compF + f(pe 0i ,t) + f(a 0i ,t) + ɛi,t.
Model 4: T i,t = μ + timeF + compF + f(pe 1i ,t) + f(a 0i ,t) + ɛi,t
Model 5: T i,t = μ + timeF + compF + f(pe 1i ,t) + f(a 1i ,t) + ɛi,t