Repeated evolution of vertebrate pollination syndromes in a recently diverged Andean plant clade

Abstract

Although specialized interactions, including those involving plants and their pollinators, are often invoked to explain high species diversity, they are rarely explored at macroevolutionary scales. We investigate the dynamic evolution of hummingbird and bat pollination syndromes in the centropogonid clade (Lobelioideae: Campanulaceae), an Andean-centered group of ∼550 angiosperm species. We demonstrate that flowers hypothesized to be adapted to different pollinators based on flower color fall into distinct regions of morphospace, and this is validated by morphology of species with known pollinators. This supports the existence of pollination syndromes in the centropogonids, an idea corroborated by ecological studies. We further demonstrate that hummingbird pollination is ancestral, and that bat pollination has evolved ∼13 times independently, with ∼11 reversals. This convergence is associated with correlated evolution of floral traits within selective regimes corresponding to pollination syndrome. Collectively, our results suggest that floral morphological diversity is extremely labile, likely resulting from selection imposed by pollinators. Finally, even though this clade's rapid diversification is partially attributed to their association with vertebrate pollinators, we detect no difference in diversification rates between hummingbird- and bat-pollinated lineages. Our study demonstrates the utility of pollination syndromes as a proxy for ecological relationships in macroevolutionary studies of certain species-rich clades.

Adaptation to hummingbird pollination is associated with reduced diversification inPenstemon

“…; Lagomarsino et al. ; Serrano‐Serrano et al. ), where hummingbird pollination has sometimes been associated with increased diversification rates.…”

“…One exception is the genus Castilleja, where a hummingbird syndrome lineage has diversified into a species-rich clade (Tank and Olmstead 2008). By contrast, reduced diversification rates associated with hummingbird pollination have not been observed in neotropical groups (e.g., Givnish et al 2014;Lagomarsino et al 2017;Serrano-Serrano et al 2017), where hummingbird pollination has sometimes been associated with increased diversification rates.…”

Beyond buzz‐pollination – departures from an adaptive plateau lead to new pollination syndromes

“…Armbruster, 1988;Ne'eman et al, 2010;Ashworth et al, 2015;Fenster et al, 2015). Although a large body of literature reports pollination syndromes for certain plant lineages (L azaro et al, 2008;Armbruster et al, 2011;Lagomarsino et al, 2017), and a recent quantitative evaluation of the concept found strong support even across angiosperms (Rosas-Guerrero et al, 2014), other studies have raised concerns about the utility of the concept (e.g. Waser et al, 1996;Kingston & McQuillan, 2000;Ollerton et al, 2009).…”

Greater pollination generalization is not associated with reduced constraints on corolla shape in Antillean plants

“…; Lagomarsino et al. ). Unlike many investigations performed at the population level, such approaches aim at characterizing constraints on morphological variation over macroevolutionary scales and, as such, should be informative to understand the forces that have been determinant in modeling the morphology of large groups of species.…”

“…Specifically, we test if increased generalization in pollination strategies is associated with relaxed selection constraints, or greater diversification (disparity), for corolla shape in species of the subtribe Gesneriinae of the Gesneriaceae family in the Caribbean islands. The recent development of powerful phylogenetic comparative methods allows the estimation of historic selective constraints on large groups of species (e.g., Hansen and Martins 1996;Beaulieu et al 2012;Butler and King 2004) and thus testing specific hypotheses regarding the role of pollinators on floral trait evolution (Smith et al 2008;Gómez et al 2015;Lagomarsino et al 2017). Unlike many investigations performed at the population level, such approaches aim at characterizing constraints on morphological variation over macroevolutionary scales and, as such, should be informative to understand the forces that have been determinant in modeling the morphology of large groups of species.…”