Traits and fitness
Settings
R code
The calls that repeat this analysis in R with the settings above.
Selection differentials and gradients
S total selection; β directional; γ quadratic curvature (negative is consistent with stabilising selection, positive with disruptive); γij correlational. * p < 0.05, ** < 0.01, *** < 0.001.
Summary
Gradient plot
Assumption checks
Normality of the traits (Mardia, Shapiro-Wilk), the largest VIF, individuals per quadratic term, and the residuals or dispersion of the gradient model. For reporting only. With the performance package installed its tests are added.
Fitness function
Cubic spline (Schluter 1988) with bootstrap 95% band.
Adaptive landscape for this trait
Mean fitness of a population against its mean for this trait, with the fitness function dashed.
Fitness surface
A group's peak is the highest point of the surface within that group's own range: filled when it is a peak of the surface, open when the surface keeps rising past the group's range or the edge of the data.
The fitted fitness where there are data. Drag to rotate. Diamonds: interior peaks of the surface; with groups shown, circles mark the group means and squares their highest points, filled when that point is a peak of the surface and open when it is not.
Adaptive landscape
Mean fitness for each possible population mean. Diamond: optimum; circle: current mean.
Mean fitness for each possible population mean. Drag to rotate. Diamond: optimum; circle: current mean.