Matthew G. Baron

A new hypothesis of dinosaur relationships and early dinosaur evolution

For 130 years, dinosaurs have been divided into two distinct clades—Ornithischia and Saurischia. Here we present a hypothesis for the phylogenetic relationships of the major dinosaurian groups that challenges the current consensus concerning early dinosaur evolution and highlights problematic aspects of current cladistic definitions. Our study has found a sister-group relationship between Ornithischia and Theropoda (united in the new clade Ornithoscelida), with Sauropodomorpha and Herrerasauridae (as the redefined Saurischia) forming its monophyletic outgroup. This new tree topology requires redefinition and rediagnosis of Dinosauria and the subsidiary dinosaurian clades. In addition, it forces re-evaluations of early dinosaur cladogenesis and character evolution, suggests that hypercarnivory was acquired independently in herrerasaurids and theropods, and offers an explanation for many of the anatomical features previously regarded as notable convergences between theropods and early ornithischians.

Baron et al . reply

For over a century, the standard classification scheme has split dinosaurs into two fundamental groups1: ‘lizard-hipped’ saurischians (including meat-eating theropods and long-necked sauropodomorphs) and ‘bird-hipped’ ornithischians (including a variety of herbivorous species)2–4. In a recent paper, Baron et al.5 challenged this paradigm with a new phylogenetic analysis that places theropods and ornithischians together in a group called Ornithoscelida, to the exclusion of sauropodomorphs, and used their phylogeny to argue that dinosaurs may have originated in northern Pangaea, not in the southern part of the supercontinent, as has more commonly been considered6,7. Here we evaluate and reanalyse the morphological dataset underpinning the proposal by Baron et al.5 and provide quantitative biogeographic analyses, which challenge the key results of their study by recovering a classical monophyletic Saurischia and a Gondwanan origin for dinosaurs. This shows that the Ornithoscelida hypothesis is not the final word, and that there is still great uncertainty around the basic structure of the dinosaur family tree. There is a Reply to this Comment by Baron, M. G. et al. Nature 551, http://doi.org/10.1038/nature24012 (2017). The size and scope of the Baron et al.5 dataset (457 anatomical features scored for 74 early dinosaurs and close relatives) are important advances on previous studies of early dinosaur phylogeny8–10. It combines previously published and new morphological characters, setting a standard for the field, which we applaud. With that said, however, the results of the new study5 differ so radically from all previous cladistic analyses, and decades of pre-cladistic research, that they deserve close scrutiny. Our main concern is that the authors were able to personally study fewer than half of the taxa in their analysis; the others were scored mostly based on published literature, which is problematic, because many characters relate to fine anatomical details, requiring first-hand study to be reliably documented. The taxon sample of Baron et al.5 is larger than any previous analysis and this represents one of the strongest aspects of the study. However, the lack of some important taxa (for example, the early thyreophoran Scutellosaurus, the possible theropod Daemonosaurus, the newly described Ixalerpeton and Buriolestes, and a broader sample of averostran theropods) may have a substantial effect on character optimizations near the base of the dinosaur tree, and thus on the interrelationships of early dinosaurs. Our international consortium of early dinosaur evolution specialists has come together to critically assess the Baron et al.5 dataset. We have personally studied nearly all included taxa, and some of us were the original authors of most of the characters incorporated in the dataset. Our aim was straightforward: check the scorings for each taxon in the analysis, rescoring them if necessary based on first-hand observations and adding a small number of taxa (see Supplementary Information). We did not add or rewrite characters, as this would go beyond our intention to provide a quality control check on taxon scoring. Although we note that character definition and delimitation are critical, these would be better addressed in a longer, more detailed study. Our rescored dataset produced a strict consensus tree (Fig. 1), showing the traditional arrangement of the three major dinosaur groups: sauropodomorphs and theropods united as Saurischia, with Ornithischia on a separate branch. This tree is less resolved than the one described in Baron et al.5, and the same basic arrangement is found when we analyse only those taxa included in the original study. Relationships are, however, not particularly well supported: it would take two (full dataset) or three (original taxon sample) additional steps to enforce an ornithischian–theropod clade as reported by Baron et al.5 and Templeton tests show no significant differences between the two hypotheses (see Supplementary Information). Character scoring changes explain our different results. They also alter the optimisation of the 21 putative ornithoscelidan synapomorphies proposed by Baron et al.5 (see Supplementary Information), revealing that many have a complex distribution among early dinosaurs. Some are not only present in ornithoscelidans, but can also be found more broadly among early dinosaurs, including herrerasaurids and sauropodomorphs. Others are absent in many early diverging ornithoscelidans and probably evolved independently in later ornithischians and theropods. Several of the characters used by Baron et al.5 have uninformative distributions, are poorly defined and/or completely or partially duplicate one another (see Supplementary Information). This may have resulted from a largely uncritical assembling of characters from previous analyses with different aims, without integrating or modifying their descriptions and states. Baron et al.5 also argued, without a quantitative analysis, that the placement of Saltopus and Northern Hemisphere silesaurids as close outgroups to Dinosauria suggests a northern Pangaean origin for the clade. Instead, our numerical estimate of ancestral states (using three different evolutionary models and with an over 90% log-likelihood in two of them) and two biogeographic analytical tools (statistical dispersal–vicariance analysis and implementation of the dispersal– extinction–cladogenesis model) predict that dinosaurs originated in southern Pangaea for all three possible solutions provided by Baron et al.5 regarding the positions of Saltopus, Silesauridae and Dinosauria (see Supplementary Information). Our reanalysis highlights three central issues: (1) There is currently great uncertainty about early dinosaur relationships and the basic structure of the dinosaur family tree. We did not recover the Ornithoscelida of Baron et al.5, but the more traditional saurischian–ornithischian dichotomy that we did recover is weakly supported. It seems that the flood of new discoveries over the past decades9–13 has revealed unexpected complexity. Homoplasy was rampant in early dinosaurian evolution, and the earliest members of the major subgroups were very similar in body size and morphology, which makes unravelling their relationships remarkably difficult. (2) Dataset construction is key. Morphological phylogenetic analyses hinge on taxon and character sampling and scoring, so careful consideration of the primary homologies, and careful construction and coding of characters are very important. Our critical revision of the scorings of Baron et al.5 indicates that the original version of that dataset is not reliable for testing the phylogenetic relationships of early dinosaurs. (3) It is important to use appropriate computational analytical tools before making macro-evolutionary claims. Such methods can provide a range of results, depending on models of evolution and tree reconstruction, and allow hypotheses to be explicitly tested against one another. In conclusion, the data we present here lead us to be sceptical of the new phylogeny proposed by Baron et al.5 We are excited about the Ornithoscelida hypothesis, which will certainly reinvigorate the study of dinosaur origins. However, we do not currently find strong evidence to discard the traditional Ornithischia–Saurischia division, and we must also entertain a third possibility that was articulated in the 1980s14,15, but rarely discussed since: that sauropodomorphs and ornithischians may form their own herbivorous group, separate from the ancestrally meat-eating theropods. Suboptimal trees show

A dinosaur missing-link? Chilesaurus and the early evolution of ornithischian dinosaurs

The enigmatic dinosaur taxon Chilesaurus diegosuarezi was originally described as a tetanuran theropod, but this species possesses a highly unusual combination of features that could provide evidence of alternative phylogenetic positions within the clade. In order to test the relationships of Chilesaurus, we added it to a new dataset of early dinosaurs and other dinosauromorphs. Our analyses recover Chilesaurus in a novel position, as the earliest diverging member of Ornithischia, rather than a tetanuran theropod. The basal position of Chilesaurus within the clade and its suite of anatomical characters suggest that it might represent a ‘transitional’ taxon, bridging the morphological gap between Theropoda and Ornithischia, thereby offering potential insights into the earliest stages of ornithischian evolution, which were previously obscure. For example, our results suggest that pubic retroversion occurred prior to some of the craniodental and postcranial modifications that previously diagnosed the clade (e.g. the presence of a predentary bone and ossified tendons).

Multiple optimality criteria support ornithoscelida

A recent study of early dinosaur evolution using equal-weights parsimony recovered a scheme of dinosaur interrelationships and classification that differed from historical consensus in a single, but significant, respect; Ornithischia and Saurischia were not recovered as monophyletic sister-taxa, but rather Ornithischia and Theropoda formed a novel clade named Ornithoscelida. However, these analyses only used maximum parsimony, and numerous recent simulation studies have questioned the accuracy of parsimony under equal weights. Here, we provide additional support for this alternative hypothesis using Bayesian implementation of the Mkv model, as well as through number of additional parsimony analyses, including implied weighting. Using Bayesian inference and implied weighting, we recover the same fundamental topology for Dinosauria as the original study, with a monophyletic Ornithoscelida, demonstrating that the main suite of methods used in morphological phylogenetics recover this novel hypothesis. This result was further scrutinized through the systematic exclusion of different character sets. Novel characters from the original study (those not taken or adapted from previous phylogenetic studies) were found to be more important for resolving the relationships within Dinosauromorpha than the relationships within Dinosauria. Reanalysis of a modified version of the character matrix that supports the Ornithischia–Saurischia dichotomy under maximum parsimony also supports this hypothesis under implied weighting, but not under the Mkv model, with both Theropoda and Sauropodomorpha becoming paraphyletic with respect to Ornithischia.

How has our knowledge of dinosaur diversity through geologic time changed through research history

Assessments of dinosaur macroevolution at any given time can be biased by the historical publication record. Recent studies have analysed patterns in dinosaur diversity that are based on secular variations in the numbers of published taxa. Many of these have employed a range of approaches that account for changes in the shape of the taxonomic abundance curve, which are largely dependent on databases compiled from the primary published literature. However, how these ‘corrected’ diversity patterns are influenced by the history of publication remains largely unknown. Here, we investigate the influence of publication history between 1991 and 2015 on our understanding of dinosaur evolution using raw diversity estimates and shareholder quorum subsampling for the three major subgroups: Ornithischia, Sauropodomorpha, and Theropoda. We find that, while sampling generally improves through time, there remain periods and regions in dinosaur evolutionary history where diversity estimates are highly volatile (e.g. the latest Jurassic of Europe, the mid-Cretaceous of North America, and the Late Cretaceous of South America). Our results show that historical changes in database compilation can often substantially influence our interpretations of dinosaur diversity. ‘Global’ estimates of diversity based on the fossil record are often also based on incomplete, and distinct regional signals, each subject to their own sampling history. Changes in the record of taxon abundance distribution, either through discovery of new taxa or addition of existing taxa to improve sampling evenness, are important in improving the reliability of our interpretations of dinosaur diversity. Furthermore, the number of occurrences and newly identified dinosaurs is still rapidly increasing through time, suggesting that it is entirely possible for much of what we know about dinosaurs at the present to change within the next 20 years.

Correction to ‘Multiple optimality criteria support Ornithoscelida’

[This corrects the article DOI: 10.1098/rsos.170833.].

Support for the placement of Chilesaurus within Ornithischia: a reply to Müller et al.

In our recent study [[1][1]], we recovered the Late Jurassic dinosaur Chilesaurus diegosuarezi within the clade Ornithischia. This result contrasted with those presented in the original study on Chilesaurus by Novas et al. [[2][2]], in which this enigmatic taxon was recovered as a tetanuran theropod

An investigation of the genus Mesacanthus (Chordata: Acanthodii) from the Orcadian Basin and Midland Valley areas of Northern and Central Scotland using traditional morphometrics

Mesacanthus is a common and speciose genus of acanthodian fish from Lower Old Red Sandstone and Middle Old Red Sandstone assemblages (representing the Lower Devonian and Middle Devonian respectively) and is well represented in many palaeoichthyology collections in the UK. Based upon descriptions given during the 19th century, specimens of the genus Mesacanthus from the Orcadian Basin and Midland Valley areas of Northern and Central Scotland have historically been referred to a number of different species; of these, the most frequently discussed in the literature are M. mitchelli, M. peachi and M. pusillus. In order to test the validity of these three species, traditional morphometric analyses were carried out on over 100 specimens of Mesacanthus, from both the Lower Devonian and the Middle Devonian, that cover the full range of known localities for these taxa in Northern and Central Scotland. Based upon morphological and morphometric comparisons, this investigation has found that at least two species of Mesacanthus are valid (M. mitchelli and M. pusillus) as specimens from the Lower Devonian and Middle Devonian have been shown to differ significantly in a number of important ways. However, no evidence has been found for the validity of the second and distinct Middle Devonian species, M. peachi.