High selfing and high inbreeding depression in peripheral populations of Juncus atratus

The mating system of a plant is the prime determinant of its population genetic structure. However, mating system effects may be modified by postzygotic mechanisms like inbreeding depression. Furthermore, historical as well as contemporary ecological factors and population characteristics, like the location within the species range can contribute to genetic variability. Using microsatellite markers we assessed the population genetic structure of the wind-pollinated Juncus atratus in 16 populations from peripheral and nearly central areas of the distribution range and studied the mating system of the species. In three peripheral populations, outcrossing rates at seeds stage were low (mean t(m) = 5.6%), suggesting a highly autogamous mating system. Despite this fact, on adult stage both individual heterozygosity (mean H-O = 0.48) and gene diversity (mean H-E = 0.58) were high even in small populations. Inbreeding coefficients were consistently low among all populations (mean F-IS = 0.15). Within the three peripheral populations indirect estimates of lifetime inbreeding depression were surprisingly high (delta(eq) = 0.96) and inbreeding depression could be shown to act mostly on early seedling establishment. Similar conditions of autogamy combined with high inbreeding depression are typical for plants with a large lifetime genomic mutation rate that cannot avoid selfing by geitonogamy. However, the results presented here are unexpected for small-statured, herbaceous plants. Substantial genetic differentiation among all populations was found (mean F-ST = 0.24). An isolation-by-distance pattern was apparent on large scale but not on local scale suggesting that the overall pattern was largely influenced by historical factors, e. g. colonization, whereas locally genetic drift was of greater importance than gene flow. Peripheral populations exhibited lower genetic diversity and higher inbreeding coefficients when compared with subcentral populations.