Convergent Evolution and Species Boundaries in Lomatium Species (Apiaceae)

Speciation is a complex and ongoing process caused by a variety of forces; this has in part led to a proliferation of species concepts ranging from arbitrary similarity-based concepts such as the morphological species concept to ancestry-based concepts that incorporate the complex nature of gene inheritance. Ancestry-based concepts rely on homologous characters, either molecular or morphological, that share a common origin in order to draw inferences about evolutionary histories and help understand species boundaries. Convergence in morphology and paralogous loci in molecular datasets can cause confusion in understanding if characters are functionally homologous. Other challenges in inferring species boundaries and phylogenies include hybridization, incomplete lineage sorting, and species with a high degree of phenotypic plasticity. Species boundaries and phylogenies complicated by the previously mentioned factors can be understood more readily by incorporating molecular data and associated phylogenetic analysis. The Perennial Endemic North American clade of Apiaceae (PENA) is one of the largest and least understood plant radiations in Western North America. Lomatium is the largest genera in PENA and as traditionally defined is paraphyletic with many cases of homoplasy and morphological convergence. Species boundaries, evolutionary relationships, and field identification are notoriously difficult in this genus. This research proposes to use molecular, morphological, and ecological data sources to investigate species boundaries and evolutionary relationships in two groups within PENA: three alpine endemics found within the Wallowa and Elkhorn Mountains of eastern Oregon (Lomatium greenmanii, Lomatium erythrocarpum, and Lomatium oreganum) and the Lomatium packardiae/anomalum subcomplex of the larger Lomatium triternatum complex found throughout the inland Northwest, intermountain West, and Northern Rocky Mountains. Convergent evolution is a common phenomenon among alpine plants due to the extreme selective pressures of the environment and the limited number of successful plant body designs that can withstand this environment. In two isolated mountain ranges in eastern Oregon, three morphological similar alpine endemic Lomatium exist; these similarities in geographic distribution, habitat preference, and morphology can be ascribed to either shared ancestry or convergent evolution. Previous studies based on limited morphological and molecular data suggested that shared ancestry was the cause of similarities in at least two of these lineages (L. greenmanii and L. oreganum). In the present study, newly sequenced samples of L. oreganum L. greenmanii and L. erythrocarpum are included in a phylogeny based on Sanger-sequence data (five plastid markers and two nuclear ribosomal) of the greater PENA clade including about ~200 samples spread across ~70 taxa. Each alpine Lomatium occupies a different phylogenetic position indicating that convergence and not shared ancestry is the major force influencing the evolution of similar phenotypes in this group. Increased conservation efforts are urged for these rare taxa as this research indicates that the phylogenetic diversity of the Wallowa and Elkhorn ranges is higher than originally hypothesized. Inferring species boundaries is a complex effort complicated by many factors including theoretical issues dealing with the nature of the species unit itself and biological realities which confound field or morphological-based identification. Next-generation sequencing has resulted in a 10 to 100-fold increase in the amount of molecular data available to researchers. The combination of next-generation sequencing and the recent development of complex statistical software that can model the speciation process while incorporating coalescent theory and accounting for incomplete lineage sorting allow researchers to investigate species boundaries and evolutionary relationships accurately in previously recalcitrant taxa. Morphological, ecological, and geographic differences in the L. anomalum/packardiae subcomplex do not always agree with monophyletic groups in Sanger-sequence based studies. Hypotheses for the recalcitrant nature of this subcomplex include that the complex is one phenotypically plastic taxa or incomplete lineage sorting is causing the incongruences between morphological, ecological, and molecular data. Fifty-four low copy nuclear introns and whole plastome data sets generated by target-enriched next-generation sequencing were used to generate three types of phylogenies: a concatenated approach for the intron data (MrBayes), a coalescent-based analysis of the intron data (STACEY; ASTRAL-III), and a Bayesian analysis of the whole plastome data (MrBayes). Further, morphological, ecological, environmental, and geographic data were investigated to uncover any unique characteristics which correspond with monophyletic groups. The STACEY coalescent-based phylogeny was the only analysis to uncover monophyletic groups (seven fine scale clades included in three coarser clades) that agree with non-molecular external data including geography and some previously recognized taxonomy. Lomatium packardiae was resolved as monophyletic and was the only clade distinct in a PCA of reproductive characters with shorter umbel rays and fruits than other members of the subcomplex. Incipient speciation with a high degree of incomplete lineage sorting and some phenotypic plasticity appears to be the most likely explanation for the incongruences observed in earlier Sanger-based studies. Currently, geography is the most reliable indicator of phylogenetic placement in the L. packardiae/anomalum subcomplex