Pre-print, Working paper: I comment on the claim by John Roemer that in games of pure coordination, Kantians drive Nashers to extinction". Using an explicit dynamic model of evolution, I notice that in these games, Kantian optimizers do not always drive selsh optimizers to extinction.