Megaevolutionary dynamics and the timing of evolutionary innovation in reptiles

Abstract

The origin of phenotypic diversity among higher clades is one of the most fundamental topics in evolutionary biology. However, due to methodological challenges, few studies have assessed rates of evolution and phenotypic disparity across broad scales of time to understand the evolutionary dynamics behind the origin and early evolution of new clades. Here, we provide a total-evidence dating approach to this problem in diapsid reptiles. We find major chronological gaps between periods of high evolutionary rates (phenotypic and molecular) and expansion in phenotypic disparity in reptile evolution. Importantly, many instances of accelerated phenotypic evolution are detected at the origin of major clades and body plans, but not concurrent with previously proposed periods of adaptive radiation. Furthermore, strongly heterogenic rates of evolution mark the acquisition of similarly adapted functional types, and the origin of snakes is marked by the highest rates of phenotypic evolution in diapsid history.

Fig. 1. Distribution of phenotypic and molecular relative evolutionary rates in different reptile groups.

Rates are relative to the mean posterior estimate of base of the clock rate value (= 0.00345 substitutions per character per myr). a distribution of phenotypic rates among early evolving diapsid lineages (median=1.9; mean=2.1). b distribution of phenotypic (median=0.51; mean=1.0) and molecular rates (median=0.54; mean=0.83) among lepidosaurs. c linear regression between phenotypic and molecular rates (log-transformed) among lepidosaurs (R-squared: 0.0001425; p-value: 0.8615; two-sided). n = 109 (a) and n = 211 (b, c) for molecular and phenotypic evolutionary rates pooled from all unique bipartitions from the posterior trees. Shaded grey area = 95% CI. Source data are provided as a Source Data file.

Fig. 2. Relative phenotypic evolutionary rates and disparity through time in early diapsid reptiles.

Branch widths proportional to posterior probabilities for clades stemming from the respective branches. a phenotypic rates among the major early evolving diapsid reptile lineages from the Early Permian to the Middle Triassic. Grey area represents 95% CI around LOESS regression line segmented by geological time bins. Carboniferous rates are not considered here due to low sample size and extremely broad confidence intervals. b phenotypic disparity in early diapsids through time. Box plots represent distribution of 100 bootstraps at each time bin. Lower bound = 1st quartile (Q1); upper bound = 3rd quartile (Q3); notching = median; diamond = mean; minima = Q1-1.5xIQR; maxima = Q3 + 1.5xIQR; outliers = dots. Grey area represents one SD around the means (connected by blue trendline). n = 109 (a, b) evolutionary rates pooled from all unique bipartitions from the posterior trees. c phenotypic rates of evolution in reptiles plotted on the time-calibrated maximum compatible tree from Mr. Bayes. For full tree see Supplementary Figs. 5, 10.

Fig. 3. Relative phenotypic and molecular evolutionary rates and disparity through time in lepidosaurs.

Branch widths proportional to posterior probabilities for clades stemming from the respective branches. a phenotypic (blue) and molecular (red) rates among the major lepidosaur lineages from the Jurassic to the present time. Grey area represents 95% CI around LOESS regression line segmented by geological time bins. Triassic rates are not considered here due to low sample size and extremely large confidence intervals. b phenotypic disparity in early diapsids through time. Box plots represent distribution of 100 bootstraps at each time bin. Lower bound = 1st quartile (Q1); upper bound = 3rd quartile (Q3); notching = median; diamond = mean; minima = Q1-1.5xIQR; maxima = Q3 + 1.5xIQR; outliers = dots. Grey area represents 95% CI around LOESS regression line segmented by geological time bins (phenotypic = dark blue and molecular = dark red). n = 211 (a, b) for molecular and phenotypic evolutionary rates pooled from all unique bipartitions from the posterior trees. c phenotypic rates of evolution in reptiles plotted on the time-calibrated maximum compatible tree from Mr. Bayes. For full tree see Supplementary Figs. 5, 10.

Fig. 4. Relative molecular rates of evolution in lepidosaurs.

Branch widths proportional to posterior probabilities for clades stemming from the respective branches. Molecular rates are plotted for all sampled extant taxa and internal branches until their most recent common ancestor. For full tree and rate values see Supplementary Figs. 5, 11.

Fig. 5. Phylomorphospace of early diapsid reptiles and lepidosaurs.

The first two axes of phenotypic variation among principal coordinates. Early diapsid reptiles occupy a distinct region of the morphospace from lepidosaurs, which in turn, have rhynchocephalians, non-serpentiform squamates, and serpentiform squamates (snakes and amphisbaenians) occupying different regions of the morphospace defined by PCo 1 and 2.