Niche evolution in plant polyploids remains controversial and evidence for alternative patterns has been reported. Using the autopolyploid Dianthus broteri complex (2×, 4×, 6× and 12×) as a model, we aimed to integrate three scenarios - competitive exclusion, recurrent origins of cytotypes and niche filling - into a single framework of polyploid niche evolution. We hypothesized that high polyploids would tend to evolve towards extreme niches when low ploidy cytotypes have nearly filled the niche space. We used several ecoinformatics and phylogenetic comparative analyses to quantify differences in the ecological niche of each cytotype and to evaluate alternative models of niche evolution. Each cytotype in this complex occupied a distinct ecological niche. The distributions were mainly constrained by soil characteristics, temperature and drought stress imposed by the Mediterranean climate. Tetraploids had the highest niche breadth and overlap due to their multiple origins, whereas the higher ploidy cytotypes were found in different, restricted, nonoverlapping niches. Niche evolution analyses suggested a scenario with one niche optimum for each ploidy, including the two independent tetraploid lineages. Our results suggest that the fate of nascent polyploids could not be predicted without accounting for phylogenetic relatedness, recurrent origins or the niche occupied by ancestors.
