Should plants in resource-poor environments invest more in antiherbivore defence?

The 'resource availability hypothesis' proposes that low resource availability favours plants with inherently slow growth rates, which in turn favours large investments in antiherbivore defence. Here we present a simple model of evolution of plant defence allocation in a system incorporating plant-resource dynamics, and show that taking into account the feedback of plants on the amount of a limiting resource leads to very different predictions than those from the classical resource availability hypothesis. The latter is bused on the assumption that fitness is determined bu the potential growth rate, hence, implicitly, on the assumption of unlimited growth in a transient environment where plants do not deplete soil resources. In an environment where plants have accumulated enough biomass to control resource concentration, fitness is determined by the ability to deplete limiting resources if the environment is homogeneous or by other traits correlated with biomass or productivity if the environment is spatially structured. In this case, resource supply and maximum growth rate may either increase, decrease, or in most cases not affect at all, optimal defence investment, Much more detailed data about what determines fitness in 'climax' environments mould he necessary to make even qualitative predictions about the direction of evolution of antiherbivore defence investment as a function of resource availability and maximum growth rate.