Abstract
Aim A major biogeographical hypothesis, the ‘niche‐breadth’ hypothesis, explains species range sizes based on the extent of a species’ niche (e.g. diversity of habitats occupied; range of environmental conditions tolerated). An alternative hypothesis explains range sizes using metapopulation theory (e.g. colonization dynamics; dispersal ability). Both niche breadth and colonization ability may be related to plant species’ reproductive characteristics. We evaluate both hypotheses by examining the relationship of plant range size to mating system and genomic structure (ploidy status). Location Western North America. Methods Using a data set of 60 taxa in the genus Clarkia (Onagraceae), we use three analytical techniques to examine the effect of reproductive characteristics on range size. We conduct cross‐species analyses of present‐day taxa to examine both mating system and polyploidy in relation to range size. We also conduct a phylogenetically independent contrasts analysis (using caic software) on the relationship of mating system to range size in diploid species. Third, we compare closely related taxon pairs that differ in mating system as an alternative method to control for phylogeny. Results Polyploid species have significantly larger ranges compared with diploid species. When considering only diploid taxa, no significant relationship is found for mating system in both cross‐species and phylogenetically independent contrasts analyses. The diploid pairwise analysis, comparing only the range sizes of direct sister taxa with alternative mating systems, does show a relationship, with outcrossing species having larger ranges than self‐fertilizing species. Main conclusions We argue that genetic diversity, colonization ability, or a combination of both factors may influence plant species’ range sizes. The significant pairwise analysis suggests that both the independent contrast and the cross‐species analysis may be confounded by polytomies of species at the terminal nodes of the phylogeny, indicating the importance of comparing the results of multiple analytical techniques. We propose that the range sizes of self‐fertilizing species have a bimodal distribution, obscuring the effect of mating system on range size, and that a broader survey of plant taxa will resolve the two modes from that of the outcrossing species. Lastly, polyploid species appear to show significantly larger range sizes than diploid species, irrespective of mating system.