Consumers limit the abundance and dynamics of a perennial shrub with a seed bank.

For nearly 30 years, ecologists have argued that predators of seeds and seedlings seldom have population-level effects on plants with persistent seed banks and density-dependent seedling survival. We parameterized stage-based population models that incorporated density dependence and seed dormancy with data from a 5.5-year experiment that quantified how granivorous mice and herbivorous voles influence bush lupine (Lupinus arboreus) demography. We asked how seed dormancy and density-dependent seedling survival mediate the impacts of these consumers in dune and grassland habitats. In dune habitat, mice reduced analytical lambda (the intrinsic rate of population growth) by 39%, the equilibrium number of aboveground plants by 90%, and the seed bank by 98%; voles had minimal effects. In adjacent grasslands, mice had minimal effects, but seedling herbivory by voles reduced analytical lambda by 15% and reduced both the equilibrium number of aboveground plants and dormant seeds by 63%. A bootstrap analysis demonstrated that these consumer effects were robust to parameter uncertainty. Our results demonstrate that the quantitative strengths of seed dormancy and density-dependent seedling survival--not their mere existence--critically mediate consumer effects. This study suggests that plant population dynamics and distribution may be more strongly influenced by consumers of seeds and seedlings than is currently recognized.