Reproductive allocation and the fitness consequences of selfing in two sympatric species of Epilobium (Onagraceae) with contrasting mating sytems

Selfing has evolved repeatedly in outcrossing taxa, and theory predicts that an increase in the level of self-fertilization should occur in concert with changes in reproductive allocation and the magnitude of inbreeding depression. Here we characterize the mating system of two sympatric congeners, Epilobium ciliatum and E. angustifolium, and compare the taxa for 1) reproductive allocation patterns and 2) the fitness consequences of self-fertilization. For E. ciliatum, autogamy rates were high, pollinator visitation was low, and electrophoresis revealed no genetic variation at 11 putative isozyme loci. For E. angustifolium, autogamy rates were low, pollinator visitation was relatively high, and electrophoresis generated an outcrossing rate estimate of t = 0.64 (SE = 0.08). The pollen/ovule ratio was ten times higher for E. angustifolium than for E. ciliatum, due to a decline in pollen production in the selfing species. The proportion of total flower biomass allocated to female function was significantly greater in E. ciliatum, while that allocated to male function and attractive structures was greater in E. angustifolium. We quantified the fitness consequences of selfing at three life stages: seed number, percent germination, and mature biomass. Relative performance (RP) measures indicated less inbreeding depression for E. ciliatum than for E. angustifolium at all stages; differences in RP between the species were significant for seed number and cumulative total, but not for germination or biomass. RP was correlated among life history stages for only one comparison, suggesting that the genetic basis of inbreeding depression differs among life history stages. Variation among maternal parents for RP was significant at almost all stages in both species, with the exception of seed number in E. ciliatum. The striking variation among maternal parents in E. angustifolium, ranging from strong inbreeding depression to strong outbreeding depression, may reflect both variation in the history of inbreeding and the long-distance migration of individuals from different populations.