The cosmopolitan reticulate complex centering on Dryopteris carthusiana has consistently fascinated and perplexed pteridologists. In particular, the origins of the allotetraploids D. carthusiana and D. cristata have been of great interest. In the mid 1950s, Walker proposed that these hybrids contained a shared genome ("B") based on his cytogenetic studies, but it was not until the late 1960s that Herb and Florence Wagner along with Dale Hagenah named the species contributing the "B" genome D. semicristata. These authors proposed that the diploid contributing this genome was extinct, lingering only as a component genome of allopolyploids. Meanwhile, Widen and Britton suggested that Walkerís "B" genome could have been contributed by the Asian diploid D. tokyoensis based on evidence from phloroglucinol analyses and Hickok and Klekowski developed elegant theoretical genetic arguments to demonstrate that D. ludoviciana could have contributed the"B" genome. We have analyzed isozyme patterns and restriction enzyme digests of chloroplast DNA to address this question. Our combined data representing both nuclear and organellar DNA support neither D. tokyoensis nor D. ludoviciana as progenitor diploids supplying the "B" genome. Instead, our results are consistent with the genome having come from an unknown diploid which is most likely extinct, and we support the recognition of this genome as an extinct species named D. semicristata. We suspect that other reticulate complexes may contain now extinct genomes in their extant alloploid representatives.

Key words: Dryopteris semicristata, genome