Morphological and Geographical Evidence are Misleading with Respect to the Phylogenetic Position and Origin of the Narrow Endemic Polyploid <I>Androsace cantabrica</I> (Primulaceae)

Abstract

Allopolyploidy is recognized as an important mode of plant speciation. The allopolyploid origin of a taxon is often inferred from morphological intermediacy or a mixture of morphological characters from the putative parental taxa, especially if accompanied by geographical intermediacy of the polyploid. Here we show that in the case of Androsace cantabrica (Primulaceae), a narrow endemic from the central Cordillera Cantabrica (Spain), both lines of evidence are misleading. Instead of being an allopolyploid derivative of A. halleri (western Cordillera Cantabrica, eastern Pyrenees, Massif Central, Vosges) and A. laggeri (central Pyrenees), A. cantabrica is most likely an autopolyploid closely related to A. adfinis s. 1., endemic to the southwestern European Alps. This counterintuitive biogeographic connection not only adds to the group of species showing a Cantabrian-Alpine disjunction, but, together with the unexpected link between A. chaixii, also from the southwestern Alps, and A. halleri in form of the northern Spanish local endemic A. rioxana, underlines the important role of large-scale range shifts in shaping the current distribution of European mountain taxa.