The Xeromphalina campanella/kauffmanii complex: species delineation and biogeographical patterns of speciation

European, North American and northeastern Asian collections of Xeromphalina section Xeromphalina were studied by sequencing the nuc rDNA ITS1-5.8S-ITS2 and 28S 5' regions and partial RNA polymerase II second largest subunit gene (RPB2). Previously designated Xeromphalina campanella I is designated X. campanella s. str. and a neotype for this species from the topotype region is established. This species is shown to be a discrete, cold-tolerant organism that is distributed across North America and Eurasia and does not exhibit significant geographical partitioning. A second closely related phylogenetic species previously designated X. campanella II, proposed as X. enigmatica, cannot be distinguished from X campanella morphologically but is reproductively isolated and is sympatric with X campanella across much of Eurasia and North America. Unlike X. campanella it shows geographical partitioning and some of the geographical populations likely have become reproductively isolated. Phylogenetic and geographical evidence suggests that X. enigmatica may have given rise to the eastern North American endemic, Xeromphalina kauffmanii, which also is reproductively isolated and is characterized by a hardwood substrate and a difference in basidiospore shape. Two putatively interbreeding haplotypes are evident for both eastern North American X kauffmanii and eastern North American X enigmatica and might be contributions from different glacial refugia. Cryptic taxa related to X. enigmatica are identified but not named due to small sample sizes including Asian taxa 1-5 and an apparent endemic from Idaho and British Columbia. Several species-delineation procedures were attempted and compared with this complex molecular dataset. Rosenberg's P-AB statistic and P-ID (liberal) were the most liberal, assigning species status to haplotypes or interbreeding clades within species. P-ID (strict) and P-RD (randomly distinct) were more stringent. Ability to intercross was the most stringent criterion for species delineation and did not correlate well with P-AB, P-ID and P-RD delineations.