Aim We evaluated the hypothesis that, given niche conservatism, relatedness of co-occurring hummingbird species of a given clade will increase at greater distances from the elevation where it originated. We also used prior knowledge of flight biomechanics and feeding specialization of hummingbird species (family Trochilidae) to evaluate which environmental variables were important predictors of changes in phylogenetic structure for each hummingbird clade.
Methods We compiled species lists for 189 local hummingbird assemblages across major environmental gradients in Ecuador from a variety of published and non-published sources. For the entire family and each of the major hummingbird clades (hermits, emeralds, mangoes, coquettes and brilliants) we quantified the phylogenetic structure of each assemblage using the net relatedness index (NRI). This index calculates the standardized mean of all possible pairwise phylogenetic distances between co-occurring species. We related NRI for each clade to elevation, precipitation and vegetation-related variables using generalized additive models.
Results Our findings support the prediction of an increase in the co-occurrence of close relatives away from the elevation of origin at the family level and for assemblages of mangoes and brilliants. The opposite pattern was found for assemblages of coquettes and emeralds. For the hermits, variation in phylogenetic structure was not explained by elevation. Clades with high levels of feeding specialization (hermits and brilliants) always included a vegetation-related variable as an important predictor of change in phylogenetic structure.
Main conclusions We found no overall support for the conservatism and zone of origin hypotheses. Knowledge of each clade’s natural history proved useful for predicting which environmental variables correlated with phylogenetic structure of local assemblages. Clades with the same elevation of origin appear to have radiated along the elevational gradient in association with different environmental factors.