Adaptive radiation and the evolution of nectarivory in a large songbird clade

Abstract

The accumulation of exceptional ecological diversity within a lineage is a key feature of adaptive radiation resulting from diversification associated with the subdivision of previously underutilized resources. The invasion of unoccupied niche space is predicted to be a key determinant of adaptive diversification, and this process may be particularly important if the diversity of competing lineages within the area, in which the radiation unfolds, is already high. Here, we test whether the evolution of nectarivory resulted in significantly higher rates of morphological evolution, more extensive morphological disparity, and a heightened build‐up of sympatric species diversity in a large adaptive radiation of passerine birds (the honeyeaters, about 190 species) that have diversified extensively throughout continental and insular settings. We find that a large increase in rates of body size evolution and general expansion in morphological space followed an ancestral shift to nectarivory, enabling the build‐up of large numbers of co‐occurring species that vary greatly in size, compared to related and co‐distributed nonnectarivorous clades. These results strongly support the idea that evolutionary shifts into novel areas of niche space play a key role in promoting adaptive radiation in the presence of likely competing lineages.

The association between morphological and ecological characters across a global passerine radiation

“…Given the strong positive correlation with the original morphological measurements (Table 1), we consider PC1 to be a good proxy for overall body size. In birds and throughout the animal kingdom more generally, body size is a core axis of niche differentiation, due to its robust association with prey size, the size of other consumed resources and/or resource monopolization (Lack, 1947;Hutchinson, 1959;Marki et al, 2019;Richman & Price, 1992;Wilson, 1975). Size (PC1) alone is a relatively weak predictor of the membership of the ecological character states ( Figure 6), with the incorporation of shape dimensions necessary to improve classification accuracy when attempting to predict ecological differences among species (Figures 5,6).…”

“…Assuming ecological selection is a significant influence upon morphological evolution, statistical differences in the distribution of trait values should be expected among sets of species that differ in their ecological strategies (e.g. Marki, Kennedy, Cooney, Rahbek, & Fjeldså, 2019;White, 2016). However, even if ecological selection pressures influence morphological evolution, statistical models with low predictive ability may result (e.g.…”

Peripheral eco‐morphology predicts restricted lineage diversification and endemism among corvoid passerine birds

“…Divergent selection pressures that are the outcome of interspecific competition are a likely driver of these trends (Darwin, 1876; Parsons, 2005). Given that Australasia is consistently recovered by biogeographical analyses to represent the ancestral area of the Corvides, and oscine passerines more broadly (Boles, 1995; Jønsson et al., 2011; Kennedy et al., 2017; Moyle et al., 2016; Oliveros et al., 2019), many closely related passerine lineages are likely to have been co‐distributed with one another throughout their evolutionary histories, making competitive interactions highly probable (Marki et al., 2019). Interactions with competitively superior lineages are theorized to lead to the progressive evolution of ecological specialization and increasingly restricted geographical distributions (Pepke et al., 2019; Ricklefs & Bermingham, 1999, 2002).…”

Evolution of the locomotor skeleton in Anolis lizards reflects the interplay between ecological opportunity and phylogenetic inertia

“…While this does not retain original trait variances, and absolute rates therefore are not meaningful 40 , it allows a comparison of relative rates and rate shifts between the three focal groups (see refs. 24,52,53 for similar approaches). We evaluated the overall levels of relative evolutionary rates as well as the number of rate shifts.…”

“…As input data, we used the principal components (PCs) rather than raw data since the orthogonality between PC axes fulfils the assumption of the model that traits are independent (see refs. 24,52,53 for similar approaches). PCs were derived from the full dataset, the full dataset without centroid size, or subsets (girdle or limb-length data), and we kept the number of PCs that were deemed explanatory by the 'broken stick' method as implemented in the R package vegan (version 2.5-6).…”