What defines an adaptive radiation? Macroevolutionary diversification dynamics of an exceptionally species-rich continental lizard radiation

Abstract

Background: Adaptive radiation theory posits that ecological opportunity promotes rapid proliferation of phylogenetic and ecological diversity. Given that adaptive radiation proceeds via occupation of available niche space in newly accessed ecological zones, theory predicts that: (i) evolutionary diversification follows an 'early-burst' process, i.e., it accelerates early in the history of a clade (when available niche space facilitates speciation), and subsequently slows down as niche space becomes saturated by new species; and (ii) phylogenetic branching is accompanied by diversification of ecologically relevant phenotypic traits among newly evolving species. Here, we employ macroevolutionary phylogenetic model-selection analyses to address these two predictions about evolutionary diversification using one of the most exceptionally species-rich and ecologically diverse lineages of living vertebrates, the South American lizard genus Liolaemus.

Results: Our phylogenetic analyses lend support to a density-dependent lineage diversification model. However, the lineage through-time diversification curve does not provide strong support for an early burst. In contrast, the evolution of phenotypic (body size) relative disparity is high, significantly different from a Brownian model during approximately the last 5 million years of Liolaemus evolution. Model-fitting analyses also reject the 'early-burst' model of phenotypic evolution, and instead favour stabilizing selection (Ornstein-Uhlenbeck, with three peaks identified) as the best model for body size diversification. Finally, diversification rates tend to increase with smaller body size.

Conclusions: Liolaemus have diversified under a density-dependent process with slightly pronounced apparent episodic pulses of lineage accumulation, which are compatible with the expected episodic ecological opportunity created by gradual uplifts of the Andes over the last ~25My. We argue that ecological opportunity can be strong and a crucial driver of adaptive radiations in continents, but may emerge less frequently (compared to islands) when major events (e.g., climatic, geographic) significantly modify environments. In contrast, body size diversification conforms to an Ornstein-Uhlenbeck model with multiple trait optima. Despite this asymmetric diversification between both lineages and phenotype, links are expected to exist between the two processes, as shown by our trait-dependent analyses of diversification. We finally suggest that the definition of adaptive radiation should not be conditioned by the existence of early-bursts of diversification, and should instead be generalized to lineages in which species and ecological diversity have evolved from a single ancestor.

Fig. 1

Phylogenetic relationships within the Liolaemus radiation showing variation in body size (snout-vent length obtained by averaging male and female SVLs) across species (black bars, in mm). Clade colours indicate the eight main groups (or subgenera) within the genus

Fig. 2

Tempo and mode of macroevolutionary diversification in Liolaemus lizards. The bottom plot shows the lineage through time (LTT) curve of species accumulation over time (solid line) and the 95 % (yellow area) to 50 % (red area) confidence intervals (note the most recent pulse is borderline). The phylogenetic tree in the background shows a maximum-likelihood phylogenetic reconstruction of ancestral body sizes (ln-transformed) along the branches and nodes of the tree, and the interspecific range is shown in the coloured bar with the frequency distribution of SVL of the entire genus. The top plot shows mean subclade disparity through time (DTT) for body size (solid line), compared with the median subclade DTT (calculated based on 10,000 simulations) of phenotypic evolution on the genus phylogeny under a Brownian motion model (dashed line). The grey shaded area represents the 95 % confidence interval of DTT range based on simulations of body size disparity

Fig. 3

Projection of the Liolaemus phylogeny into a morphospace defined by body size (ln-transformed, on y) and time since the clade’s origin (on x, in My elapsed since the root). Ancestral body size states are estimated using likelihood. The degree of uncertainty is indicated by increasing transparency of the plotted blue lines around the point estimates with the entire range showing the 95 % confidence interval. Red arrows indicate the position of the three body size peaks (in mm) identified by the surface analysis (see text for details)