Modelling abiotic and biotic drivers of a plant radiation

Mountains worldwide are among the most species rich habitats on our planet. The rugged terrain of mountains, featuring a diversity of different habitat types across small distances, is believed to have driven the rapid evolution of these species. To date, most studies have investigated the impact of climatic conditions or single traits such as the growth form of a plant on its evolutionary and ecological history in mountains. We know little, however, about how interactions with other organisms such as pollinators have affected the diversification of these species. This is surprising because the change in climatic conditions from lowland habitats to mountains does not only impact plant species, but also their pollinators. In tropical mountains such as the Andes, for example, bees are common pollinators in lowlands, but are replaced by vertebrates such as hummingbirds at high elevations. In my project, I will investigate the relative importance of abiotic (e.g. climatic) conditions and biotic (e.g. pollination) processes on the evolution of the plant group Merianieae. The group is most diverse in the tropical Andes and most species have originated within the past five million years. Most species are pollinated by big carpenter bees and bumblebees through a specialized pollination strategy called buzz-pollination. In buzz-pollination, bees apply vibrations produced by their wing muscles to flowers to extract pollen as a reward. In Merianieae, buzz-pollinated species dominate in lowland tropical rainforests, but do also occur in montane cloud forests (ca. 2400 m to 3300 m). In these cloud forests, bees are generally scarce, however, given the often rainy, windy and cold weather conditions. It is thus not surprising that many of the montane species have adapted to pollination by vertebrates such as birds, bats and rodents, which are less impeded by the in hostile climate. We do not know, however, to what extent these pollinator shifts were a crucial precondition for Merianieae to colonize tropical mountains. Alternatively, adaptations to montane climatic conditions may have come first, and only after Merianieae had colonized the uplifting Andes less than five million years ago, these pollinator shifts may have happened. Using a series of different modelling approaches, I aim at disentangling the relative contribution of historic processes such as Andean uplift, climatic conditions, plant traits and pollinators on the rapid evolution of Merianieae. I will combine this modelling work (which happens in the computer) with empirical fieldwork in tropical rainforests in Colombia and Ecuador. Amongst other things, I will employ artificial vibration experiments mimicking buzzing bees to better understand how some bee- pollinated species can persist in mountains. I have assembled a diverse international team for this project and will spend one year at the University of Colorado Boulder (USA) to deepen my expertise in macroevolutionary analyses.

Interactions between animals and plants are ubiquitous and have shaped the evolution of both groups of organisms. Flower-pollinator interactions play a particularly important role in this context, as pollinators directly influence the reproductive success of plants. Accordingly, many plants show adaptations in their floral traits to different pollinator groups: for instance, bird-pollinated flowers are often red, produce large amounts of nectar, and have narrow, deep floral tubes, while bee-pollinated flowers are typically blue or yellow, produce only small amounts of nectar, and have broader or open shapes. While our understanding of floral adaptations to different pollinator groups is relatively well developed, we still know little about why plants switch pollinators, how such pollinator shifts occur, and what impact they have on diversification (i.e., the formation of new species). In this research project, we focused specifically on the historical drivers of pollinator shifts and their consequences, applying various modeling methods and phylogenetic analyses of different plant groups. We were able to show that shifts between bee and vertebrate pollinators do not occur randomly but are closely linked to abiotic climatic factors. In the tropical regions of South and Central America, hummingbird-pollinated plants occur significantly more often in cool mountain regions and cloud forests, whereas bee-pollinated species are disproportionately found in lowland rainforests. In temperate North and South America, these patterns become less distinct-here, hummingbird-pollinated species are often found at lower elevations, but in moister habitats. We also demonstrated that pollinator shifts themselves do not affect the diversification rate of a tropical plant group. Rather, the colonization of the rising Andes around 10-15 million years ago was the key factor behind a rapid increase in species numbers-a trend that continues to this day, with high diversification rates still observed in the Andean cloud forests. For the tropical plant family Melastomataceae, we were further able to show that mountain colonization generally preceded pollinator shifts, and therefore likely triggered them. Physiological differences between insects and vertebrates (many bee species are significantly limited in their flower visitation by the often cool and moist conditions of mountainous environments, while vertebrates are able to visit flowers even in cooler weather) are the most likely explanation for the observed association between pollinator shifts and climate. Additionally, both macroecological and theoretical studies showed that certain floral traits, such as flower color, are influenced not only by pollinator-driven selection but also by climatic conditions. Specifically, pink and red flowers occur more frequently in cooler regions, while yellow and purple flowers are more common in dry, sunny areas. Furthermore, we demonstrated through a series of studies that bee-pollinated Melastomataceae exhibit a surprising diversity of flower forms-an unexpected finding, considering that all these species are visited by the same pollinator group.