Please repost - thanks!

There is a PhD version of this too - www.ualberta.ca/en/graduate-...

By extending the comparative model we can assay the evolutionary signal in evolvability itself and test a number of biologically interesting trait configurations (codivergence, contra-divergence, ceiling effects, phenotypic integrations and more!)

The models are notably more complex but fittable with modern Bayesian GLMM routines. Think of them as looking at mean and scale (usually variance) evolution along the tree, here in an example of two concurrent Brownian Motion dynamics driving diversification of both variance and means.