Dinosaurs Rediscovered

by Michael J Benton

  The Late Triassic rift valleys of eastern North America formed lakes, and great thicknesses of lake sediments were laid down, often containing fish, insect, and plant fossils and, every now and then, a trackway produced by a dinosaur as it hurried across the wet lakeside sediments.

  Asia, Europe, and North America were in the northern hemisphere in the Late Triassic, but all somewhat south of where they are now. London and New York were at Mediterranean–Caribbean latitudes, and so they were considerably warmer than they are today. The absence of ice caps also ensured there were no cold winters. South America, Africa, India, Antarctica, and Australia were all joined together in the southern hemisphere, and these formed continuous land across the equator to the northern continents. Dinosaurs could march over thousands of kilometres from South Africa to Arizona, or from Canada to north Africa. There were regional faunas in some places, defined probably by mountain ranges and climate belts, but mostly the plants and animals on land were much more capable of spreading worldwide in the Triassic than today.

  These conditions continued into the Jurassic. Even though the north Atlantic had begun to unzip, animals could walk across from Africa to South America, and there were still some connections from North America through Greenland to Europe. This was true right through to the Late Jurassic, 150 million years ago, when dinosaurs such as the huge sauropod Brachiosaurus is known from Tanzania in east Africa and Wyoming in the central United States. The huge predatory dinosaur Allosaurus is also known from Wyoming, and possibly from Tanzania, but also from Portugal. There was a wide ocean between the northern and southern continents along the line of the equator, but the traffic of dinosaurs seems to have been funnelled north and south through a little strip of land that linked Morocco in the south and Spain in the north.

  During the Cretaceous, the continents continued rotating and moving. The south Atlantic opened, and the traffic of land life between South America and Africa was eventually cut off. The southern continents separated, with Africa moving north and keeping some contacts with Europe, but the southern tip of South America linking to Antarctica to the east, and then Australia. In the Late Cretaceous, India broke free and began its long trek north to eventually dock with the rest of the Asian continent about 50 million years ago. India continues to drive north, forcing the Himalayas ever higher as it does so.

  In the Late Cretaceous, not only were the continents moving closer to their present positions, but also sea levels rose hugely, by some 100 metres (nearly 330 feet) or more, as a result of enhanced mid-ocean mantle activity and uplift. This sea level rise flooded coastlines round all the continents, and split Africa and North America with mid-continental seaways. This meant that dinosaurs in the Late Cretaceous had their opportunities for migration massively curtailed. For example, the famous Tyrannosaurus rex of this time is found only in North America, not elsewhere in the world, unlike many of its predecessors. For the first time, east coast dinosaurs could not even cross the North American continent to meet their cousins on the west coast.

  Changing climates and changing worlds. A few years ago, dinosaur palaeontologists thought that all the main outlines of the origin of the dinosaurs had been resolved. Then everything changed. New fossils pushed dinosaur origins back by 15 million years into the Early Triassic. Perhaps, if you had been transported back in a time machine, you would have barely noticed the first dinosaurs. Among all the abundant, hefty, and noisy rhynchosaurs, synapsids, and crurotarsans snuffling around, the few, small, bipedal dinosaurs nipping in and out of the undergrowth would have seemed like a sideshow.

  Their explosion onto the scene 15 million years later, after the devastation wrought by the Carnian Pluvial Episode, was dramatic. Dinosaurs replaced rhynchosaurs and others overnight, geologically speaking. Understanding how all this happened has involved some remarkable new fossil finds, but also extraordinary advances in our understanding of rock dating and our ability to reconstruct ancient climates and ancient worlds, and to use modern computational methods to crunch the data and test the models of large-scale evolution.

  This is a chapter that will definitely need rewriting in ten years’ time. I predict someone somewhere will find some of these oldest dinosaurs, which are somewhat elusive right now. New studies in the field will pin down the nature of the Carnian Pluvial Episode better, and new analyses will help us properly understand the big evolutionary and ecological changes that were tearing the Earth and life apart through the Triassic.

  Chapter 2

  Making the Tree

  Classifications cause controversy. Through my entire research career, palaeontologists have squabbled strenuously over the classification of their organisms of choice, whether it be dinosaurs, trilobites, or fossil plants. These fights might seem inconsequential, but we are considering the fundamentals of how to document the wonders of biodiversity, and we are also addressing origins.

  Documenting biodiversity and origins is big science now – indeed, it forms part of the modern techniques termed, rather forbiddingly, phylogenomics and bioinformatics. Phylogenomics is the new discipline of establishing evolutionary trees from molecular data. Bioinformatics is the field of managing large data sets in the life sciences and number-crunching those data to produce information on the genetic basis of disease, adaptations, and cell function, and has applications fundamental to medicine and agriculture. Practitioners of these methods block their university’s supercomputers for weeks while they run billions of repeat calculations to get their answers. The American National Science Foundation has invested millions of dollars in their ‘Tree of Life’ initiative – a programme to fund consortia of scientists to produce complete evolutionary trees of particular groups of plants and animals, such as all 11,000 species of birds or all 300,000 species of flowering plants.

  We need accurate inventories of species to plan for practical conservation measures. Understanding which of the many species of mosquito passes on malaria parasites was critical in seeking cures. Biomedical scientists study evolutionary trees of fast-evolving viruses such as AIDS and influenza. In the case of viruses, evolutionary trees span months or years, whereas the evolutionary trees constructed by dinosaur palaeobiologists span millions of years. Without evolutionary trees, sometimes called phylogenies, we cannot explore big patterns in evolution. So while in some ways the classification of life might seem trivial or merely an arcane branch of librarianship, in others it is crucial.

  In this chapter, we will follow the quest I shared with a number of other contemporaries to crack the family tree of dinosaurs. The story began in 1984, at a conference in Tübingen, southern Germany, where four of us turned up a little nervously with our independent first efforts. We were all then in temporary employment, having finished our doctoral theses a year or two before, and eking out livings variously as research fellows in the United States (Jacques Gauthier and Paul Sereno) and the United Kingdom (Dave Norman and myself). Were we on the right track or not? We had each tackled some aspect of dinosaur phylogeny and all agreed, independently, that dinosaurs had all evolved from a single ancestor, and between us we had a solid picture of the exact pattern of relationships of the main dinosaur groups. This was a first, but would these iconoclastic ideas be accepted?

  Since 1984, we have continued our quest for more and more complete understanding of the dinosaur family tree. In 2002 and 2008, teams in my laboratory produced the first supertrees of dinosaurs – each the result of hugely laborious number-crunching – purporting to show the complete set of relationships among hundreds of species. And, unexpectedly, the whole thing blew up again in 2017 when a radical new proposal tore apart the consensus on dinosaur relationships. This is a story of discovery of new fossil specimens, innovative new ideas about methods, the harnessing of improving computer power, and the continuing fascination of the dinosaurian tree of life. It’s not over yet.

  Back in 1984, the risk of rejection for our first tentative dinosaur family trees was heightened by the fact that we had all applied a re
volutionary new set of techniques called cladistics. We coded our data on primitive punch cards and sent them off to our university’s mainframe computer services, and then waited a few days for the results to come in. This was all massively controversial in the 1980s, and we have since lived through decades of refinement of methods. We need to look back to the beginning of the so-called ‘cladistic revolution’, and consider whether the risks we took paid off. First, though – why was it all so controversial?

  What was the cladistic revolution?

  The cladistic revolution was about methods. When I was trained in classification, we used textbooks from the 1960s by luminaries such as Ernst Mayr and G. G. Simpson. They agreed that the best approach to classifying species, whether living or extinct, was to apply a great deal of experience. As Mayr recounted, it took him decades as a researcher on bird biodiversity to learn that a character such as feather colour would not help much in determining the deep relationships of birds, but that more fundamental characters, such as the shape of the bill or some particular muscles of the wing, were much more useful. As Simpson said, ‘classification of species is more art than science’.

  While Mayr and Simpson were writing, the cladistic revolution had already begun, but neither they nor anybody else noticed. In the year 1950, Willi Hennig, a rather dour professor of entomology in Berlin, published a book he had been writing in the 1940s while a prisoner of war. He did not use the term ‘cladistics’; that was introduced about 1960 by other evolutionary biologists, from the Greek word klados, which means a branch, referring to the ‘branches’ of an evolutionary tree. Hennig, in fact, called his new discipline ‘phylogenetic systematics’, meaning that he wanted to explain that the process of reconstructing the tree of life should be more science than art – he wanted biologists and palaeontologists to delve into their data, and discriminate the true worth of the characters or traits that they wished to use for the fundamentals of classification.

  Hennig’s book was in German, but very few biologists, whether they could read German or not, paid much attention. It was only when it was translated into English in 1966 that his message was received. At first, the evangelists for the new method, researchers at the American Museum of Natural History in New York and Natural History Museum in London, were intensely excited and wrote eloquently on the topic. They also applied the new methods to their studies of mutual interest, namely the evolution of fishes. However, others struggled, because Hennig had a very dry prose style, and he invented a lot of new terminology, often compound words, for his new ideas; readers had difficulties with the text both in German and in English. Nonetheless, Ernst Mayr and G. G. Simpson both waded in with highly critical attacks on the new cladistics, and Mayr invented the term ‘cladist’, which he and others used to designate (and denigrate) the adherents of this new creed.

  The museum evangelists in England and the United States pushed the ideas and sought to explain them through publications and conferences. By 1983, when I was writing my doctoral thesis, it was far from clear that Hennig’s cladistics would prevail. Most biologists and palaeontologists were indifferent or hostile to the new idea. I remember attending a meeting of the Willi Hennig Society in London in 1984, just before the Tübingen meeting, where people were shouting at each other, and one speaker was twirling the microphone on its cord and threatening the chairperson, who was trying to shut him up. On other occasions, tempers became so frayed that public apologies had to be demanded and delivered.

  Why so much heat and so little light? Willi Hennig’s insight was rather straightforward: that we need a testable method for the construction of phylogenetic trees, and this should be based around phylogenetically informative characters. Palaeontologists were to stop searching for ancestors, because you can never test a hypothesis of ancestry, but instead to seek out sister groups: that is, nearest relatives.

  As we saw in Chapter 1, the Silesauridae family is sister to the Dinosauria. This is a big claim: that these two groups are closest relatives and that they share a close common ancestor. In a modern representation, we can show the relationships of the dinosaurs to silesaurids, and all their other close relatives among the archosaurs, as an explicit tree, or cladogram (see overleaf). Groups fit within groups, and each group has a single ancestor and is characterized by one or more phylogenetically informative characters. These unique groups with single points of ancestry are called clades: hence the terms cladistics, cladogram, and cladist. In the case of the hypothesis that silesaurids and dinosaurs are closest relatives, the evidence is that they both share six or seven unique anatomical characters not seen in any other animals, such as the proportions of their hip bones, an opening between the ischium and pubis in the hip area, modifications to the femur and tibia in the hindlimb, and a rising process on the front of the astragalus bone in the ankle. We don’t consider general characters such as their slender limbs, their pointed teeth, or other features that are seen widely among early reptiles – these are like Mayr’s feather colours, not informative in constructing the cladogram.

  The search for the phylogenetically informative character is a tough one, but it provides a focus for testing. Anyone who wishes to argue that Silesauridae is not the sister group of Dinosauria must present an alternative hypothesis, in the form of a different cladogram that is supported by better evidence comprising alternative anatomical characters. In general, the more informative anatomical characters that can be mustered, the more likely the hypothesis is to be correct.

  This is what many of the critics did not like. They were being forced to go further than they were comfortable in doing. It was no longer good enough to hide behind the fig leaf of a dashed line and a scattering of question marks. The dinosaur family trees in my university textbooks showed all the relationships of the main dinosaur groups confidently in the Jurassic and Cretaceous, but the lines then tailed off into a cloud of uncertainty as we got to their origins. How far back should you go?

  Cladogram showing the evolution of archosaurs, with key phylogenetic characters indicated. The very close relationship between Dinosauria and Silesauridae is clear.

  Discovery of the clade Dinosauria

  Understanding about dinosaur classification had gone back and forth over the years. Richard Owen first named the group Dinosauria in 1842, and he included the theropod Megalosaurus, the sauropodomorph Cetiosaurus, and the ornithischian Iguanodon – one of each of the fundamental subgroups. Then in 1887, Harry Seeley, a professor at the University of Cambridge, dropped a bombshell. He had studied all the dinosaurs known to Owen, plus many more that had been found since 1842, and had decided they did not form a natural group. Rather, he claimed they should be split into two groups, the Saurischia, for the theropods and sauropodomorphs, and the Ornithischia. He noted that the saurischians all shared the so-called ‘reptile hip’ and the ornithischians all shared what he called the ‘bird hip’.

  Seeley’s insights were partly right, partly wrong, but it was the wrong aspects that dominated the subject for nearly 100 years. The reviews and textbooks showed dinosaurs as having arisen from two, three, or even more distinct ancestors. This showed a lack of clarity of thinking. Of the two styles of dinosaurian hip, only that of ornithischians was unique to the group. The saurischian hip arrangement was seen in crocodiles, lizards, and, perhaps confusingly, also in birds. In cladistic terms, it’s impossible to demonstrate that the saurischian hip arrangement was not simply acquired from their ancestors without modification. By the way, the so-called ‘bird hip’ of Seeley was a misnomer – although a unique character complex of Ornithischia, it was independent from the hips of birds, which in fact evolved from theropod dinosaurs, as we shall see later.

  Comparing the pelvic arrangement of saurischian dinosaurs (with ‘reptile hip’) and ornithischian dinosaurs (with ‘bird hip’).

  The confusion was partly resolved in 1974, when the American palaeontologist Bob Bakker and the English palaeontologist Peter Galton argued for the validity of Owen’s Dinosauria. In
their opening words:

  Traditionally dinosaurs are classified as two or three separate, independent groups of reptiles in the Subclass Archosauria. But evidence from bone histology, locomotor dynamics, and predator/prey ratios strongly suggests that dinosaurs were endotherms [= warm-blooded animals] with high aerobic exercise metabolism, physiologically much more like birds and cursorial mammals than any living reptiles.

  However, their conclusion was expressed in terms of shared aspects of biology and ecology, and that was not enough to convince the doubters. After all, the shark and the dolphin share many features of their swimming mode and feeding style, but that doesn’t change the fact that one is a fish and the other a mammal.

  This was the background all four of us had before the 1984 Tübingen meeting: eighty years of rejection of the reality of dinosaurs as a natural group by all the experts, and a cheeky proposal against that view by Bakker and Galton. We were convinced Bakker and Galton were right, but what was needed was a properly worked-out cladistic hypothesis, with each branch in the tree supported by one or more bullet-proof anatomical characters. In my paper, I identified fourteen characters unique to Dinosauria, including a series of features of the hindlimb, such as the inturned head of the femur, the muscle-bearing processes of the femur, the roller-like astragalus in the ankle with a rising process in front of the tibia, the much reduced calcaneum in the ankle, the bunched toe bones, and posture up on tippy-toes. These all relate to the fact that dinosaurs had perfected a fully upright stance at the time of their origin. The fact that saurischians and ornithischians share all these detailed anatomical features was convincing evidence, we thought back in 1984, that all dinosaurs formed a single natural group, Dinosauria, with a single ancestor.