Among-individual variation in pollen limitation and inbreeding depression in a mixed-mating shrub.

BACKGROUND AND AIMS

Variation in inbreeding depression (δ) among individual plants is considered to play a central role in mating system evolution and population genetics. Moreover, such variation could be linked to individual susceptibility to pollen limitation (PL) because those individuals strongly affected by δ for seed production will require more outcross pollen for setting a given number of fruits or seeds. However, no study has tested explicitly for associations between PL and δ at the individual plant level. This study assesses the extent of among-individual variation in PL and δ, the consistency of δ across life stages, and the relationships between individual PL and δ in the mixed-mating shrub Myrtus communis.

METHODS

Controlled hand-pollinations were performed in a natural M. communis population. Marked flowers were monitored until fruit production and a greenhouse experiment was conducted with the seeds produced.

KEY RESULTS

Compared with selfing, outcross-pollination enhanced seed number per fruit, germination rate and seedling growth, but did not enhance fruit-set. Only seed number per fruit was pollen limited and, thus, cumulative pollen limitation depended more on pollen quality (outcross pollen) than on quantity. The effects of δ varied considerably across life stages and individual plants. Cumulative δ was high across individuals (mean δ = 0·65), although there were no positive correlations between δ values at different life stages. Interestingly, maternal plants showing stronger δ for seed production were more pollen limited, but they were also less affected by δ for seedling growth because of a seed size/number trade-off.

CONCLUSIONS

Results show a general inconsistency in δ across life stages and individuals, suggesting that different deleterious loci are acting at different stages. The association between δ and PL at the individual level corroborates the idea that pollen limitation may be 'genotype-dependent' regardless of other factors.