Optimal Seed Size under Mixed Mating Systems When Predispersal and Postdispersal Inbreeding Depression Is Decoupled

Inbreeding depression (ID) has different components, and how these affect selection on seed size is currently unknown. Using an optimality model, I find that when pollen limitation selects for mixed mating systems, increased predispersal ID (abortion of selfed ovules) increases the optimal sizes of both selfed and outcrossed seeds, whereas increased seed size-independent postdispersal ID (reduced survival of selfed seeds) only increases the size of outcrossed seeds. The effect of decreased efficiency of maternal investment in selfed seeds depends on details of the model assumptions, but in many cases it will increase resource allocation to both types of seeds. If seed size-independent postdispersal ID is less than 0.5, predispersal ID will select for selfed seeds receiving more maternal investment than outcrossed seeds. The same is true for decreased efficiency of maternal investment. If decreased efficiency of maternal investment leads to a reduction in selfed seed size, selfed seeds may end up being smaller than outcrossed seeds. The model highlights the complex roles ID plays in determining seed size and indicates that empirical measurement of postdispersal ID may underestimate its extent. Tests of the model may help us understand whether plants can adaptively allocate resources differentially between selfed and outcrossed seeds.