ESS germination strategies in randomly varying environments. I. Logistic- type models (Q1079531)

The paper studies the germination strategies for a finite number of different morphs of a parthenogenetic annual plant species growing in a temporally varying habitat. The point is that seeds in the ground may germinate or not; if they remain dormant, they forgo an opportunity to reproduce, but they are able to survive conditions fatal to germinated seeds. For a single-morph population, let X(t) denote the number of seeds in year t just before germination, and let G (0\(\leq G\leq 1)\) be the fraction of seeds which germinate. The per capita yield of germinated seeds, \(Y(X(t),\alpha_ t)\), depends on the density of germinated seeds, X(t), and on the state of the environment \(\alpha_ t\) (an i.i.d. random variable). Nongerminating seeds remain dormant until next year, with survivorship probability s \((0<s<1)\). Thus the population dynamics is governed by the stochastic difference equation \[ X(t+1)=X(t)[GY(X(t),\alpha_ t)+(1-G)s]. \] Under certain assumptions (mainly on the function Y) it is shown that X(t) converges to a stationary distribution, provided \[ E[\log (GY(0,\alpha_ t)+(1- G)s)]>0. \] (Otherwise, X(t)\(\to 0\) and the morph is not viable). Denote by \(\tilde X_ G(t)\) the limiting stationary process for germination rate G. \(G_ 1\) is an evolutionary stable strategy (ESS) for morph 1 if morph 1 is viable, and any invading morph 2 with \(G_ 2\neq G_ 1\) becomes extinct, provided the initial density of morph 2 is low enough, i.e. \[ X_ 2(t)\to 0\quad with\quad probability\quad one,\quad where\quad X_ 2(t+1)=X_ 2(t)[G_ 2Y(\tilde X_{G_ 1}(t),\alpha_ t)+(1-G_ 2)s]. \] The paper gives conditions under which \(G=1\) resp. \(0<G<1\) is an ESS \((G=0\) is not viable), and discusses their biological implications.