Diversification, biogeography, and classification of Amsinckiinae (Boraginaceae), with an emphasis on the popcornflowers (Plagiobothrys)

Amsinckiinae is a diverse and ecologically important subtribe of annual herbaceous or perennial suffrutescent taxa with centers of distribution in western North America and southern South America. Taxa in the subtribe occur in all major ecosystems in California and more broadly in western North America, from the deserts of Baja California in the south where Johnstonella and Pectocarya are common, north to the ephemeral wetland ecosystems of the California Floristic Province where a majority of Plagiobothrys sect. Allocarya taxa occur, and east to the Basin and Range Province of western North America, where Cryptantha sensu stricto (s.s.) and Oreocarya are well represented. The subtribe minimally includes 9 genera: Amsinckia, Cryptantha s.s., Eremocarya, Greeneocharis, Harpagonella, Johnstonella, Oreocarya, Pectocarya, and Plagiobothrys; overall minimum-rank taxonomic diversity in the subtribe is ca. 330-342 taxa, with ca. 245--257 taxa occurring in North America, 86 in South America, and 4 in Australia. Despite their prevalence on the landscape and a history of active botanical research for well over a century, considerable research needs remain in Amsinckiinae, especially in one of the two largest genera, Plagiobothrys. Taxonomic concepts in the genus have fluctuated through the 1900s and to date no critical reappraisal of the accepted taxonomy in Plagiobothrys has been performed using modern molecular phylogenetic methods; evolutionary patterns have been neglected as well. This dissertation focuses on evaluating patterns of diversification in the subtribe with an emphasis on Plagiobothrys. The objectives of the dissertation were three-fold: (1) to generate a hypothesis of evolutionary relationships among target Amsinckiinae and if necessary, revise classification in Plagiobothrys so that only monophyletic groups are recognized taxonomically; (2) to evaluate patterns and processes involved with the American Amphitropical Disjunction (AAD) biogeographic pattern in Amsinckiinae; and (3) to examine patterns of diversification within Plagiobothrys with respect to habitat affinity. Chapter 1 examines evolutionary relationships and classification in Amsinckiinae with a focus on Plagiobothrys. Molecular sequence data were used to estimate a hypothesis of evolutionary relationships. Plagiobothrys was found to be non-monophyletic, with some members more closely related to Amsinckia or Cryptantha s.l. than to Plagiobothrys in the strict sense. Two new genera were established and one older genus name was recommended for usage so that only monophyletic groups are recognized taxonomically. Morphology was examined in light of the new estimate of phylogenetic relationships. These analyses demonstrate that the trait traditionally used to delimit the large genus Cryptantha s.l., a grooved adaxial surface, is a shared ancestral feature and is in conflict with the evolutionary history inferred in this study. Chapter 2 focuses on AAD, an interesting but understudied New World biogeographic pattern wherein close relatives grow in temperate North America and temperate South America, but are lacking in the intervening New World tropics. Molecular sequence data and fossil and secondary calibrations were used to generate a time-calibrated phylogram on which biogeographic patterns could be assessed using maximum likelihood and Bayesian inference approaches. This study shows that sampled members of Amsinckiinae have moved between North America and South America over 10 times, and always from north to south. Timing of inferred dispersals ranges from Plio-Pleistocene to near to the present. Statistical models suggest that taxa that dispersed from North America to other continents had smaller fruits and more highly ornamented fruits, a pattern that supports birds as the probable dispersal vector. Chapter 3 evaluates diversification in Plagiobothrys with an emphasis on examining the role of habitat and ecological opportunity in promoting increased net diversification rates. Once again, molecular sequence data and fossils calibrations were used to generate a time-calibrated phylogeny of Plagiobothrys and the subtribe. Likelihood-based methods were used to reconstruct the ancestral habitat in the genus, in particular to assess the number of times that vernal pools and other ephemeral aquatic ecosystems were invaded. This analysis was unequivocal in estimating a single transition to ephemerally aquatic ecosystems. A character (habitat) independent diversification analysis identified a major net diversification rate shift following the invasion of vernal pools and other ephemeral aquatic ecosystems. A character dependent analysis of diversification with respect to habitat resulted in an estimate of diversification rates over three times greater in ephemeral aquatic lineages than in terrestrial lineages. These findings support the hypothesis that increased rates of diversification accompanied the invasion of vernal pool and ephemeral aquatic ecosystems in the Mediterranean-type climate regions of western North America and western South America. Diversification did not begin until these systems formed ca. 4-3 Ma, suggesting that the evolution of these ephemeral water bodies over millions of acres presented an ecological opportunity that was exploited by members of Plagiobothrys from the Plio-Pleistocene to the present day