Abstract
As reported by experimental studies, changes in pollination regime are expected to drive plant mating system evolution. In natural populations facing pollinator decline, plant-mating system is thus susceptible to evolve. We analyzed the demographic consequences of such evolution and if such evolution can rescue populations. We developed a quantitative genetic model for evolution after a pollinator crash and we analyzed the demographic consequences over a few dozens of generations. The model considers two sources of stochasticity. Contrary to classical models, inbreeding depression is considered as a probabilistic event affecting differentially inbred and outbred individuals (demographic stochasticity). Pollination is also considered as a probabilistic event (environmental stochasticity). The model is derived under (1) infinite population size and (2) finite population size. The results highlight three generic evolutionary scenarios. The evolution of selfing after a pollinator crash can rescue populations but can sometimes lead to evolutionary suicide. While the genetic variance of mating system traits determines the pace of evolution, initial population sizes determine the countdown for evolution to rescue population making stochastic extinction likely in small populations. Our model shows that evolution may not save populations due to frequency-dependent selection acting on mating system. We propose an alternative interpretation for the higher extinction rate of selfing taxa and we discuss its implications for plant conservation.