We will carry out a quantitative genetic research on the evolution of anti-predator traits in damselfly larvae. We will reconstruct genetic variance-covariance matrices and compare these among species pairs of damselflies that did or did not experience a habitat shift. (cfr. habitats with fish or dragonfly larvae as top predators). We will thus test the hypothesis of constancy of G-matrices, and explore the extent to which habitat shifts are made possible by patterns of evolutionary potential.