Chapter 4 Effect of stabilizing selection on trait distribution



Figure 4.1 Changes in a hypothetical bivariate trait distribution induced by stabilizing selection. All of the distributions are normal before and after selection; 95% confidence ellipses are shown before selection (blue) and after selection (black). Selection is imposed by a bivariate stabilizing Gaussian function (ISS); 95% confidence ellipse shown in orange. The principal axes (principal components or eigenvectors) of the three ellipses are also shown. The bivariate trait mean is centered about the optimum of the ISS, so selection changes the variances and covariance of the trait distribution, but not its bivariate mean. Omega is a matrix that describes the intensity of stabilizing selection. Omega is approximated by a gamma-matrix whose elements are the coefficients in a regression of relative fitness on trait values.