Birds have peramorphic skulls, too: anatomical network analyses reveal oppositional heterochronies in avian skull evolution

Abstract

In contrast to the vast majority of reptiles, the skulls of adult crown birds are characterized by a high degree of integration due to bone fusion, e.g., an ontogenetic event generating a net reduction in the number of bones. To understand this process in an evolutionary context, we investigate postnatal ontogenetic changes in the skulls of crown bird and non-avian theropods using anatomical network analysis (AnNA). Due to the greater number of bones and bone contacts, early juvenile crown birds have less integrated skulls, resembling their non-avian theropod ancestors, including Archaeopteryx lithographica and Ichthyornis dispars. Phylogenetic comparisons indicate that skull bone fusion and the resulting modular integration represent a peramorphosis (developmental exaggeration of the ancestral adult trait) that evolved late during avialan evolution, at the origin of crown-birds. Succeeding the general paedomorphic shape trend, the occurrence of an additional peramorphosis reflects the mosaic complexity of the avian skull evolution.

Fig. 1. Hierarchical organization of anatomical networks in extinct and extant theropod dinosaurs.

a UPGMA cluster of anatomical network and skull modularity of an adult Tyrannosaurus rex (skull reconstruction modified from Carr) modified after Werneburg et al.. b UPGMA cluster of anatomical network and skull modularity of an adult Gallus gallus (skull reconstruction modified from Jollie) modified after Werneburg et al.. c Ward.2D cluster of anatomical network, anatomical network and skull modularity of a juvenile Gallus gallus (skull reconstruction modified from Jollie) from the recent study. Horizontal dashed lines mark the partition into Q-modules, while circles at nodes mark the statistical significance of S-modules (white, p-value < 0.05; grey, p-value < 0.01; black, p-value < 0.001). All silhouettes are from http://www.phylopic.org/.

Fig. 2. Results of the anatomical network analysis (AnNA), principal coordinate analysis (PCoA) and phylogenetic flexible discriminant analysis (pFDA).

a Range in the number of skull bones N in non-avian archosaurs (n = 19), juvenile and adult crown birds (both n = 41). For juvenile crown birds, the range of N is plotted against relative skull size (ratio of skull box volume). Results of ordinary least square regression analysis (OLS) describing the correlation between N and relative skull size of juvenile and adult crown birds are given in the box. b, c Same as a, showing the range of PCo1 and PCo2 in non-avian archosaurs and juvenile and adult crown birds. d PCoA morphospace and biplot based on network parameters showing the distribution of juvenile and adult extant archosaurs and non-avian dinosaurs (Group 1). e pFDA plot showing the separation between juvenile and adult extant archosaurs (Group 1) and the distribution of non-avian dinosaurs. f pFDA plot showing the separation between juvenile and adult crown birds (Group 2) and the distribution of non-avian dinosaurs and Alligator missippisensis.

Fig. 3. Distribution of skull modules in some non-avian theropods, and ontogenetic pairs of Gallus gallus and Struthio camelus from the left and right (mirrored) side.

The modularity of non-avian theropods is very similar between different species (including non-avian Avialae), but also throughout ontogeny (see ontogenetic pair of Tarbosaurus bataar). Juvenile crown birds resemble the ancestral condition in number and distribution of modules, while adult crown birds show a severe reduction of modules due to ontogenetic bone fusion. Colours highlight different skull modules within species, but do not necessarily imply homology between species. Original sources of modified skull reconstruction: Scipionyx samniticus, Allosaurus fragilis, juvenile Tarbosaurus bataar, adult Tarbosaurus bataar, Archaeopteryx lithographica, Ichthyornis dispar, Gallus gallus, Struthio camelus (see Supplementary Data 2 file).

Fig. 4. Ancestral state reconstruction of PCoA data for adult and juvenile crown birds.

a PCo1. b PCo2. Based on ancestral values, skull network evolution shows a severe shift from Avialae (4) to Ornithurae (5) and Aves (6) as indicated in the histogram. Substituting adult crown birds with juveniles (see circles; histogram: *residual values; **original values) results in a more parsimonious evolution (see standard deviation SD), where the change from Avialae (4) to Aves (6) is rather continuous (see Supplementary Data 2 file). (1) Theropoda; (2) Coelurosauria; (3) Eumaniraptora; (4) Avialae; (5) Ornithurae; (6) Aves; (7) Neognathae; (8) Palaeognathae.