Life history evolution of seed-bank annuals in response to seed predation

Summary We present a model of life history evolution for seed-bank annuals in temporally varying environments in which both the seed bank and the distribution of fecundity across year types evolve in response to seed predation. The fecundity distribution refers to the expected reproductive success of germinating seeds across a range of different year types. We assume that it is a function of traits pertaining to growth and survival under different environmental circumstances. Such traits are assumed to result in a trade-off between reproduction in favourable and less favourable years. The model is used to explore how seed predation selects for changes in the seed bank and fecundity distribution and how changes in each of these further select for changes in the other. The direction of selection is contingent upon: whether or not a seed bank exists; whether predation has a greater effect on fresh or buried seed; whether the predation rate differs in different year types, and if so, if it is positively or negatively density-dependent; whether or not predation rate is sensitive to individual variation in seed yield, and if so, whether and how such dependency varies in different kinds of year. Under a variety of predation regimes, seed predators select for a temporal clumping of reproduction; i.e. a specialisation on a favourable subset of year types. This effect usually requires negatively density-dependent seed predation of the sort created by predator satiation. In fact, the classic scenario favouring masting in perennials creates the strongest such effect in our model. Yet unlike the masting of perennial plants, this effect is favoured in a seed-bank annual. It can even occur in a strict annual without a seed bank, and it can occur in a seed-bank annual even if seed predation is density-independent.