Introduction
Seed plant populations may be linked through two key ecological processes: pollen and seed dispersal (Krauss et al. 2008, Ballesteros-Mejia et al. 2016, Gelmi-Candusso et al. 2017, Cortés et al. 2018). Pollen dispersal by animals in particular is considered as major driver of plant population structure, often more important than seed dispersal (Yu et al. 2010, Clavino-Cancela et al. 2012, Kartzinel et al. 2013, Skogen et al. 2019, Gamba & Muchhala 2020, Nazareno et al. 2021). Animal pollinators differ markedly in their mobility, flower-visitation behaviour and rates of pollen transfer (Bawa 1992, Breed et al. 2015, Krauss et al. 2017, Dellinger et al. 2021). These differences may result in distinct within-population mating patterns and among-population differentiation (Wessinger 2021). Less mobile pollinators (i.e. small insects, territorial hummingbirds) may generate localized mating patterns between neighbouring individuals, increased inbreeding, and high population differentiation (Schoen & Clegg 1984, Opedal et al. 2017, Schmidt-Lebuhn et al. 2019, Wessinger 2021). Highly mobile pollinators (i.e. flying vertebrates, large bees), in contrast, may promote outcrossing and gene flow over larger geographic distances (Hughes et al. 2007, Whelan et al. 2009, Ballesteros-Mejia et al. 2016, Gamba & Muchhala 2020). Understanding the extent to which different animal pollinators drive genetic differentiation, ultimately affecting the potential for adaptive evolution and speciation, is essential for better resolving macroevolutionary processes of angiosperm diversification, but it is also of major relevance for choosing appropriate management strategies in human-altered landscapes and under current climate change (Hadley et al. 2012, Toon et al. 2014, Castilla et al. 2017). Surprisingly, however, the impact of different pollination strategies on population genetic parameters across related plant species (accounting for shared macroevolutionary background) has rarely been assessed (Barbará et al. 2007, Kramer et al. 2011).
Besides (pollinator-mediated) gene flow, a species’ present-day population genetic structure may be influenced by climatic and demographic history, landscape features, and adaptation to distinct (and changing) abiotic environmental conditions (Helmstetter et al. 2020). Mountains in particular, with their highly heterogeneous, rugged terrain, create dispersal barriers and strong environmental gradients across small spatial and temporal scales, reinforcing population differentiation and, potentially, speciation (Kisel & Barraclough 2010, Surget-Groba & Kay 2013, Nevado et al. 2019). Accordingly, mountains across the world represent biodiversity hotspots with an exceptional number of recent (plant) radiations (Luebert & Weigend 2014, Rahbek et al. 2018, Rangel et al. 2018). Several of these radiations coincide with phases of mountain uplift or Pleistocene glacial cycles (Cortés et al. 2018). Particularly the latter, inducing repeated periods of population isolation, range shifts and habitat recolonization, have left imprints in the current population genetic structure (Ornelas et al. 2013, Ortego et al. 2014, Escobar et al. 2020).
The tropical Andes represent one of the planet’s biodiversity hotspots, encompassing lowland rainforests, montane cloud forests and Páramo grasslands (Gentry 1982). The effects of Pleistocene climatic fluctuations in these habitats are complex, with idiosyncratic responses documented among different groups of organisms and geographic regions (Ramírez-Barahona & Eguiarte 2013, Vasconcellos et al. 2019). Some studies suggest that (cloud) forests were fragmented into small refugia at mid elevations during dry glacial periods, leading to marked population structuring and a loss of genetic diversity (‘dry-refugia hypothesis’, Haffer 1969, Gutíerrez-Rodríguez et al. 2011). Other data indicate that Neotropical mountains remained moist, with a mere downward and consecutive upward migration of forest zones, resulting in diffuse population structuring and increased genetic diversity in continuously forested areas (e.g. Colombian Chocó; ‘moist-forest hypothesis’, Carnaval et al. 2009, Valencia et al. 2010, Ornelas et al. 2019, Hernández-Langford et al. 2020).
Clearly, with their vital role in mediating gene flow, pollinators differing in mobility have the capacity of exponentiating or offsetting isolating effects of geographic or climatic barriers. Importantly, pollinators are not independent of their abiotic environment either. The flower visitation activity of ectothermic insect pollinators in particular may be strongly reduced by adverse weather conditions prevalent in tropical mountains, while endothermic vertebrates are less affected by weather fluctuations (Cruden 1972, Dellinger et al. 2021). To date, however, we largely lack studies from related Neotropical plants differing in pollination strategy to evaluate the relative contributions of biotic interactions, the (current and historic) abiotic environment, and landscape features to structuring population genetic diversity. Connecting these factors, however, will ultimately improve our understanding of the microevolutionary dynamics underlying (Neo)tropical diversity.
Here, we explore the effects of multifarious factors (geographic distance, topography, habitat suitability, Pleistocene climatic instability, environment) on population genomic diversity and differentiation across six related Neotropical plant species (Merianieae, Melastomataceae) differing in pollination strategy (but not in seed dispersal, see Material and Methods). We aim to understand whether mating patterns are more localized (i.e. lower heterozygosity and nucleotide diversity, higher inbreeding, stronger within-population isolation-by-distance) in localities of bee-pollinated species (two species) than of vertebrate-pollinated species (four species), and whether flying vertebrate pollinators consistently promote higher levels of gene flow across larger (geographical, environmental) distances than bees. We further expect stronger effects of variation in climatic conditions in bee- than in vertebrate-pollinated species given the strong link between abiotic conditions and the activity of ectothermic insect pollinators. For each species, we contrast two adjacent (< 12 km) localities against three to four more distant localities and directly address differences in distribution ranges and ecosystems colonized through niche modelling approaches.