Aim: The bipolar disjunction, a biogeographical pattern defined by taxa with a distribution at very high latitudes in both hemispheres (> 55° N; > 52° S), is only known to occur in about 30 vascular plant species. Our aim was to use the bipolar species Carex arctogena to test the four classic hypotheses proposed to explain this exceptional disjunction: convergent evolution, vicariance, mountain‐hopping and direct long‐distance dispersal.Location: Arctic/boreal and temperate latitudes of both hemispheres.Methods: A combination of molecular and bioclimatic data was used to test phylogeographical hypotheses in C. arctogena. Three chloroplast markers (atpF–atpH, matK and rps16) and the nuclear ITS region were sequenced for all species in Carex sections Capituligerae and Longespicatae; Carex rupestris, C. obtusata and Uncinia triquetra were used as outrgroups. Phylogenetic relationships, divergence‐time estimates and biogeographical patterns were inferred using maximum likelihood, statistical parsimony and Bayesian inference.Results: Carex sections Capituligerae and Longespicatae formed a monophyletic group that diverged during the late Miocene. Two main lineages of C. arctogena were inferred. Southern Hemisphere populations of C. arctogena shared the same haplotype as a widespread circumboreal lineage. Bioclimatic data show that Southern and Northern Hemisphere populations currently differ in their ecological regimes.Main conclusions: Two of the four hypotheses accounting for bipolar disjunctions may be rejected. Our results suggest that direct long‐distance dispersal, probably southwards and mediated by birds, best explains the bipolar distribution of C. arctogena.
Polyploidy among plants is most frequent in the arctic, where glaciation cycles put selective pressures on populations by repeated fragmentation and fluctuation in climate. Polyploids should have been more fit in the novel habitats created as glaciers receded because of increased genetic material and novel gene products, which results in phenotypic plasticity and rapid adaptation. Higher ploidy is then expected to confer a broader tolerance of environmental conditions. Eutrema edwardsii R. Br. (Brassicaceae) is an arctic-alpine mustard with a near circumpolar distribution that occurs as a tetraploid, hexaploid, and octaploid. We used flow cytometry to document the distribution of polyploid cytotypes using herbarium tissue, and modeled the niche of each cytotype to test for niche differentiation. Flow cytometry revealed four cytotypes among 85 individuals. Notably, 60% of the herbarium tissue assays were successful using tissue up to 50 years old. Principle components analysis was performed on 20 climatic variables, of which, the first four axes were used as environmental variables for niche modeling. Niche models were created for tetraploid and hexaploid populations and used to calculate niche overlap (Shoener’s D). Overlap between tetraploid and hexaploid models (D = 0.534) is lower than the null distribution (D = 0.681–0.944) supporting the hypothesis of niche divergence.