Untangling the influence of ecological and evolutionary factors on trait variation across hummingbird assemblages

Abstract

Abstract. Phylogenetic community ecology combines phylogenetic hypotheses with local species composition and functional-trait information to evaluate historical and contemporary mechanisms influencing local assemblage structure. Most studies assume that, if functional traits are conserved, then patterns of trait variation should match patterns of phylogenetic structure within local assemblages. Here we evaluated if we could predict trait structure by assuming that environmental filtering or biotic interactions work primarily on phylogenetically conserved functional traits. We investigated patterns of phylogenetic assemblage structure and functional-trait variation in bill length, wing length, and body mass in 236 hummingbird assemblages (126 species) across two major gradients in northern South America: elevation and precipitation. While mean trait values for assemblages vary predictably based on empirical knowledge of hummingbird biology, the distribution of trait values within assemblages do not correspond to those predicted based on phylogenetic signal and phylogenetic structure. Instead, we were able to identify instances where assemblages have high levels of morphological variation despite their close evolutionary relatedness and vice versa. Our results provide support for both filtering and biotic interactions across gradients, as has been documented in other studies.

Process-Based Species Pools Reveal the Hidden Signature of Biotic Interactions Amid the Influence of Temperature Filtering

“…In addition, we found that the number of assemblages exhibiting clustered body mass dispersion was approximately equal to the number of assemblages with even dispersion. Clustering of body mass tended to be more common at higher elevations (12,000 m), consistent with the findings of Graham et al (2012), whereas even dispersion was more common at lower elevations (!3,000 m). Hummingbirds tend to be large on averageand the range and variation in weights narrow-at high elevations.…”

“…Our study reveals a clear discrepancy between results from phylogenetic and trait-based analyses, a result that is increasingly observed in studies of community phylogenetics (Kraft et al 2008;Graham et al 2012). This lack of consistency might suggest that traits other than the two we measured here enable niche differentiation among cooccurring species of hummingbirds.…”

“…C, Constraints on the probability of resampling are imposed with varying strength. All georeferences of assemblages were checked using the elevation recorded and then confirmed using digital elevation data or gazeteers (for details, see Graham et al 2009Graham et al , 2012. We considered only mainland assemblages with more than three species and for which all members were represented in our phylogeny.…”

“…Specifically, since temperature explains much variation in hummingbird assemblage structure (Graham et al 2009, we predicted that a temperaturebased species pool would have the strongest influence on the perceived structure of hummingbird communities. Concordantly, we hypothesized that (2) patterns of beak length dispersion, which might reflect competition for food resources (Stiles 2008;Graham et al 2012;Maglianesi et al 2014), would be evenly dispersed and that this pattern would become stronger after accounting for temperature filtering. Finally, we hypothesized that (3) patterns of body mass dispersion, a trait that is under strong selection pressure in cold Andean environments (Brown and Bowers 1985;Stiles 2008;Graham et al 2012), would appear clustered when using a broadly defined species pool but random when using a species pool that accounts for temperature filtering.…”

What Is on the Horizon for Ecophylogenetics?

“…AC/DC (Revell et al 2008;Cooper et al 2010;O'Meara 2012) to describe trait variation can also provide relevant information. A straightforward example of this issue is demonstrated by the results of a recent study (Graham et al 2012) reporting that phylogenetic signal of species traits detected with Blomberg's K did not reflect community organization in relation to traits. Specifically, the authors showed that phylogenetically clustered communities of hummingbirds harbored an even set of traits, yet these traits showed strong phylogenetic signal (Graham et al 2012).…”