Using stochastic dynamics to model multispecies communities

We discuss a model which describes the dynamics of ecological communities on an evolutionary timescale. The fundamental constituents of the model are a large set of characteristic features, each of which has a fixed score against each of the others. Species consist of smaller sets of features; the score of one species against another derives from the scores of their constituent features. At regular time intervals a speciation event occurs: one feature in one species is chosen at random and replaced by another feature. The resulting species is then introduced into the system, new scores calculated and resources distributed accordingly. The complexity of the resulting web of species interactions is a result of the adaptive nature of the model. The distribution of resources amongst species leads to many quantitative predictions about the structure of the resulting food web. These are in good agreement with data collected by ecologists studying community food webs.