Impact of the terrestrial-aquatic transition on disparity and rates of evolution in the carnivoran skull

Abstract

Background: Which factors influence the distribution patterns of morphological diversity among clades? The adaptive radiation model predicts that a clade entering new ecological niche will experience high rates of evolution early in its history, followed by a gradual slowing. Here we measure disparity and rates of evolution in Carnivora, specifically focusing on the terrestrial-aquatic transition in Pinnipedia. We analyze fissiped (mostly terrestrial, arboreal, and semi-arboreal, but also including the semi-aquatic otter) and pinniped (secondarily aquatic) carnivorans as a case study of an extreme ecological transition. We used 3D geometric morphometrics to quantify cranial shape in 151 carnivoran specimens (64 fissiped, 87 pinniped) and five exceptionally-preserved fossil pinnipeds, including the stem-pinniped Enaliarctos emlongi. Range-based and variance-based disparity measures were compared between pinnipeds and fissipeds. To distinguish between evolutionary modes, a Brownian motion model was compared to selective regime shifts associated with the terrestrial-aquatic transition and at the base of Pinnipedia. Further, evolutionary patterns were estimated on individual branches using both Ornstein-Uhlenbeck and Independent Evolution models, to examine the origin of pinniped diversity.

Results: Pinnipeds exhibit greater cranial disparity than fissipeds, even though they are less taxonomically diverse and, as a clade nested within fissipeds, phylogenetically younger. Despite this, there is no increase in the rate of morphological evolution at the base of Pinnipedia, as would be predicted by an adaptive radiation model, and a Brownian motion model of evolution is supported. Instead basal pinnipeds populated new areas of morphospace via low to moderate rates of evolution in new directions, followed by later bursts within the crown-group, potentially associated with ecological diversification within the marine realm.

Conclusion: The transition to an aquatic habitat in carnivorans resulted in a shift in cranial morphology without an increase in rate in the stem lineage, contra to the adaptive radiation model. Instead these data suggest a release from evolutionary constraint model, followed by aquatic diversifications within crown families.

Figure 1

Composite phylogeny used in this study. Extant relationships and branch lengths from [40], placement of fossils according to [41]. This shows the Otarioidea topology, with Odobenidae as sister taxon to the Otariidae. Analyses were also run on the same tree but with a Phocidae-Odobenidae sister grouping, following the Phocomorpha hypothesis. Branch colors: Feliformia, orange; non-pinniped Caniformia, red; stem pinnipeds and allodesmines, dark blue; Phocidae, mid-blue; Odobenidae, teal; Otariidae, pale blue.

Figure 2

11 landmarks used in data analysis, shown on a skull of Arctocephalus gazella . Landmarks 4–7 were taken bilaterally. Landmark descriptions can be found in Table 2. Wireframe used to present shape variation from PCA shown in red.

Figure 3

Scatterplots showing variation on PC1 - PC4. These axes represent 39.76%, 20.36%, 14.92% and 6.14% of variance. Based on species means. Fossil pinnipeds are as follows: En, Enaliarctos emlongi; Al, Allodesmus sp.; Po, Pontolis magnus; Pi, Piscophoca pacifica; Ac, Acrophoca longirostris. Polygons connect fissipeds (red) and extant pinnipeds (blue) and reflect groupings used in the disparity analyses. Extremal shapes are shown in Figure 4.

Figure 4

Wireframes showing shape variation on PC1-PC4 in lateral and dorsal views. Anterior is to the right of the image. Landmarks the wireframe was based on are shown in Figure 2.

Figure 5

Morphological variation in carnivoran skulls. A-C: representatives of the three families of pinnipeds. D-F: Examples of shape variation within fissipeds. Note the enlarged nasal opening typical of positive PC1 scores, found in phocids, odobenids, and the fissiped otter. Dog and cat represent dolichocephalic and brachiocephalic extremes respectively, reflected by PC2 score.

Figure 6

Evolutionary morphospace showing the reconstructed evolution of Carnivora on PC1 and PC3 through time. Based on IE analysis. Phylogeny and colors as shown in Figure 1. Points at the zero time point represent both nodes and tip values.

Figure 7

Multivariate IE analysis results showing shape change on each branch. Multivariate evolutionary distances are calculated from node estimates based on IE analysis. Thicker branches represent greater morphological change on that branch.