Community assembly, from a trait-based perspective, is a direct analog to the dynamics of asexual genotypes in population genetics, and suggests that community assembly is simply the consequence of natural selection on fitness-related traits plus demographic drift. This is the mechanistic basis of my maxent model of community assembly. To explore this link I present a series of simulations based on modified Lotka-Volterra dynamics and apply the output to my maxent model. I explore the model behavior when: (1) assembly is completely neutral? (2) when traits affect probabilities of survival and reproduction? (3) when important traits are missing? (4) when species pools are of different sizes?
Results/Conclusions When traits do not affect probabilities of survival and reproduction (neutral dynamics) the maxent model has no predictive ability. Predictive ability increases when natural selection dominates demographic drift. Ignoring traits under strong selection severely decreases predictive ability but ignoring traits under weak selection does not. Increasing pool sizes decrease predictive ability.
