Print PageBromus tectorum (cheatgrass or downy brome) is probably the most successful plant invader in the modern history of North America. This inbreeding annual grass now dominates an estimated 40 million hectares in the interior West. Historically, its core area of dominance has been sagebrush steppe, but it has continued to expand its range into marginal habitats, including more xeric warm and salt deserts as well as more mesic foothill and montane environments. To address the question of the origin of populations in newly invaded environments, we used three molecular marker systems (allozymes, microsatellites or SSRs, and single nucleotide polymorphisms or SNPs) to compare B. tectorum population genetic structure in core versus marginal habitats. We examined three hypotheses: 1) Phenotypically plastic 'general purpose' genotypes dominate populations across all habitats, 2) Populations in marginal habitats are the product of natural selection on the outcrossed progeny of core genotypes (i.e., in situ evolution), and 3) Populations in marginal habitats represent novel introductions that are genetically distinct from core genotypes.
Results/Conclusions
Molecular marker genotypes that dominated populations in warm and salt desert environments were not closely related to genotypes in core environments, but instead represented unique and strongly divergent lineages or clades within the species. Populations in mesic foothill and montane environments were made up of a combination of core genotypes and related but genetically distinct genotypes that were largely restricted to montane environments. We encountered very low levels of heterozygosity and obtained no evidence in support of the hypothesis that marginal populations evolved directly from core populations through in situ evolution. We conclude that the most likely origin of populations in marginal environments, particularly xeric environments, is through the introduction of preadapted novel genotypes from the Old World range. Allozyme marker data were generally congruent with data from the other two marker systems but lacked sufficient resolution to detect major clade divisions. Four-locus SSR genotypes from marginal habitats were generally shown to be monophyletic across populations as determined from phylogeographic analysis of data from 100 SNP markers, whereas core SSR genotypes were more often polyphyletic. Quantitative phenotypic data demonstrated that novel lineages from marginal habitats possess traits, such as early flowering and salt tolerance, that preadapt them to these habitats.
