|
Trait
|
Year
|
β
|
γ
|
|---|
|
Bud set
|
2009
|
0.255***
|
− 0.043***
|
|
2010
|
0.386***
|
−0.027
|
|
Canopy duration
|
2009
|
0.121***
|
−0.032
|
|
2010
|
0.240***
|
−0.042
|
|
Growth period
|
2009
|
0.187***
|
−0.036*
|
|
2010
|
0.347***
|
−0.102***
|
|
Leaf drop
|
2009
|
0.232***
|
−0.023
|
|
2010
|
0.342***
|
0.201***
|
|
Post-bud set period
|
2009
|
−0.162***
|
−0.047**
|
|
2010
|
−0.316***
|
−0.108***
|
|
Height: diameter
|
2009
|
0.082***
|
−0.060**
|
|
2010
|
0.008***
|
−0.073*
|
|
Active growth rate
|
2009
|
0.293***
|
−0.032**
|
|
2010
|
0.399***
|
−0.023
|
|
Amax
|
2009
|
−0.077***
|
−0.024
|
|
2010
|
0.002
|
−0.092
|
|
Amax/mass
|
2009
|
−0.053**
|
−0.042*
|
|
2010
|
−0.054
|
−0.038
|
|
C: N
|
2009
|
− 0.025
|
0.027
|
|
2010
|
0.045
|
−0.065
|
|
∆leaf
|
2009
|
−0.054**
|
−0.050
|
|
2010
|
−0.110***
|
−0.037
|
|
δ15N
|
2009
|
−0.002
|
−0.019
|
|
2010
|
−0.028
|
−0.038
|
|
gs
|
2009
|
−0.093***
|
−0.035
|
|
2010
|
−0.089**
|
−0.090*
|
|
LMAsummer
|
2009
|
−0.037*
|
−0.018
|
|
2010
|
0.027
|
−0.051
|
|
Narea
|
2009
|
0.006
|
0.010
|
|
2010
|
0.007
|
−0.046
|
|
Nmass
|
2009
|
0.031
|
0.010
|
|
2010
|
−0.059
|
−0.062
|
|
NUE
|
2009
|
−0.082***
|
−0.006
|
|
2010
|
−0.011
|
−0.047
|
|
WUE
|
2009
|
0.029*
|
−0.007
|
|
2010
|
0.097***
|
−0.093**
|
- Significance: * P < 0.05, ** P < 0.01, *** P < 0.001
- The signs and magnitudes indicate the direction and strength of the linear (β) or non-linear (γ) selection on each trait in either year. Significant F-tests indicates nonzero selection differentials (Additional file 1: Table S4) on a trait in a given year. The selection differential describes both direct and indirect selection on each trait. It equals to the regression coefficient of relative fitness onto standardized trait values after controlling the effect of any unmeasured traits on fitness by including a random intercept term, ‘Genetics’ (i.e., average Euclidean genetic distance using genetic marker data). For quadratic selection, a negative, significant value of γ indicates stabilizing selection, while a positive value is evidence for disruptive selection (tests detailed in Additional file 1: Table S5). Both selection analyses were visualized in Additional file 1: Figures S9 and S10
