Historical bottlenecks decrease genetic diversity in natural populations of Dryopteris cristata

The reconstruction of recent historical population sizes allowed us to investigate the influence of random evolutionary processes on present-day genetic diversity in populations of Dryopteris cristata . This long-lived, allotetraploid fern is rare and endangered in the study area at the southwestern border of its European distribution. Random amplified polymorphic DNA (RAPD) diversity of 280 individuals from 14 populations of D. cristata was extraordinarily low, suggesting an ancient bottleneck in the species’ history. Analysis of molecular variance ( AMOVA ) of 25 different RAPD multiband phenotypes revealed significant genetic variation among three geographical regions (15%) and among populations within regions (34%); 51% of total variance was attributed to variation within populations. High population differentiation indicated limited gene flow among populations, and genetic divergence was not correlated with geographical distance. There was no relationship between genetic variation within population, estimated as molecular variance, and present-day population size. Populations with recent historical bottlenecks of fewer than 25 individuals showed a substantial and significant reduction in genetic variation, compared with populations without bottlenecks. Comparatively high levels of genetic variation were still maintained in small remnants (60–110 individuals) of formerly large populations. Average deviations of frequencies of widespread polymorphic markers within populations from their frequencies in the whole dataset were significantly higher in small or recently bottlenecked populations than in constantly large populations, thus providing evidence for random sampling effects during genetic bottlenecks and drift in small populations. The present investigation demonstrates the importance of population history for understanding present-day genetic diversity within natural populations, as well as for conservation biology.