Functional and ecomorphological evolution of orbit shape in mesozoic archosaurs is driven by body size and diet

Abstract

The orbit is one of several skull openings in the archosauromorph skull. Intuitively, it could be assumed that orbit shape would closely approximate the shape and size of the eyeball resulting in a predominantly circular morphology. However, a quantification of orbit shape across Archosauromorpha using a geometric morphometric approach demonstrates a large morphological diversity despite the fact that the majority of species retained a circular orbit. This morphological diversity is nearly exclusively driven by large (skull length > 1000 mm) and carnivorous species in all studied archosauromorph groups, but particularly prominently in theropod dinosaurs. While circular orbit shapes are retained in most herbivores and smaller species, as well as in juveniles and early ontogenetic stages, large carnivores adopted elliptical and keyhole-shaped orbits. Biomechanical modelling using finite element analysis reveals that these morphologies are beneficial in mitigating and dissipating feeding-induced stresses without additional reinforcement of the bony structure of the skull.

Fig. 1. Orbit shape morphospace occupation of all archosauromorph taxa (n = 410) and in individual groups.

Silhouettes in the main PCA plot represent extreme orbit shapes.

Fig. 2. Patterns of morphospace occupation through the Mesozoic.

Orbital morphospace plots shown for individual time intervals (a–g) and selected taxa representing specific orbit shapes (skull images redrawn based on respective specimen references detailed in the supplementary data 1).

Fig. 3. Orbital shape changes through ontogeny.

a Juvenile and adult morphospace position and orbit shape for Tyrannosaurus rex and Tarbosaurus bataar. b Juvenile and adult morphospace position and orbit shape for Proterosuchus fergusi and Germanodactylus cristatus. Morphospace plots as in Fig. 1 (skull images redrawn based on respective specimen references detailed in the supplementary data 1).

Fig. 4. Composite phylogenetic tree of analysed species.

Euclidean distances representing different orbit shapes (circular = 0, compressed, keyhole-shape, etc.=0.5) mapped onto phylogeny highlighting occurrences of non-circular orbit morphologies.

Fig. 5. Influence of skull size and diet on orbital shape.

a Skull length heatmap superimposed on orbital shape morphospace (as in Fig. 1). b Dietary regimes heatmap superimposed on orbital shape morphospace (as in Fig. 1).

Fig. 6. Biomechanical performance spaces.

Stress concentration factors (ratio between peak and reference stresses) for tested mechanistic planar models visualised as heatmaps with orbital shape morphospace superimposed. a Dorsoventral compression, b anterior shear, c anteroposterior compression, d dorsal shear.

Fig. 7. Three-dimensional deformation space.

Position of hypothetical skull models visualised for PCs 1-3. Distance between undeformed and deformed models indicated by arrows and calculated Euclidean distances. Von Mises stress contour plots for each model show in undeformed condition.

Fig. 8. Comparison of actual and hypothetical Tyrannosaurus models.

a, c, e Original orbit shape, b, d, f circular orbit shape. a, b Osteological models with reconstructed eyeball fitted to the size of the orbit; c, d von Mises stress contour plots; e, f compressive and tensile stress contour plots.