Breeding systems and population structure in Limnanthes.

The breeding systems of seven Limanthes (Limanthaceae) populations, including one "inbreeding" and three "outbreeding" taxa, were quantified using a multilocus outcrossing rate estimator (tm) and autofertility estimates. Along with the assays of heterozygosity levels, these data were used to separate components of "effective" outcrossing in terms of Wright's equilibrium inbreeding coefficient (Fe) and adult (FA) and zygotic (FZ) fixation indices. The patchy distribution of alleles as a potential source of "substructure inbreeding" was tested from the allelic frequencies mapped along a linear transect. Evidence for consanguineous matings in restricted neighborhoods and for selection at two different life cycle stages, and the efficiency of the protandrous breeding system were noted and discussed. Multilocus estimates of outcrossing are useful for their greater precision and unbiased nature while single locus estimates can help in detecting the effects of selection and population substructure. The data generally support the "heterozygosity paradox" noted by Brown (1979) but further suggest that the paradox may often result from a lack of precision of outcrossing estimates and from overlooking the stages of the life cycle being sampled.