Studies on genetic diversity help us to unveil the evolutionary processes of species and populations and can explain several traits of diploid-polyploid complexes such as their distributions, their breeding systems, and the origin of polyploids. We examined the allozyme variation of Dryopteris aemula and D. oreades, diploid ferns with highly fragmented habitats, and the allotetraploid D. corleyi to (1) analyze the putative relationship between both diploids and the tetraploid, (2) compare the levels of genetic variation among species and determine their causes, and (3) assess the breeding system of these taxa. The allozymic pattern of D. corleyi confirms that it derived from D. aemula and D. oreades. The lack of genetic diversity in D. aemula, a species of lowland habitats, may be due to genetic drift associated with the contraction of populations in the last glaciation. By contrast, the alpine D. oreades had moderate intrapopulation genetic variation, which may derive from the expansion of populations during the last glaciation. In the latter species, low interpopulational variation suggested effective gene flow (spore exchange), and genotype frequencies in Hardy-Weinberg equilibrium indicated cross-fertilization of gametophytes. Evolutionary history appears to be an essential element in the interpretation of genetic variation of highly fragmented populations.