Dinosaurs Rediscovered
Page 8
Dinosaur trees and the evolution revolution
The bombshell was a paper written by Matt Baron and Dave Norman at the University of Cambridge and Paul Barrett at the Natural History Museum in London that really set the theropod among the pigeons. In this paper, the authors claimed that palaeontologists had been wrong about the deep classification of dinosaurs, and presented a radical new dinosaur tree. A few months later, a flurry of papers re-ran tests of their new hypothesis, some agreeing with it wholeheartedly, and others (including one I co-authored) casting some doubts.
The Baron paper attracted huge attention. It was the subject of a front-cover painting in Nature, and was reported worldwide, with headlines such as ‘After 130 years, the dinosaur family tree gets dramatically redrawn’ in the Atlantic and ‘Shaking up the Dinosaur Family Tree’ in The New York Times. The Guardian reported, in classic Guardian style, that this was ‘A discussion, not a war: two opposing experts talk dinosaur family trees’. The subject rumbles on, without resolution yet.
The standard view, since 1984 at least, was that Dinosauria consists of two subgroups, Saurischia, comprising theropods plus sauropodomorphs, and Ornithischia. In their new paper, Baron and colleagues argued that the three main dinosaur groups were arranged differently, with Theropoda flipping to pair with Ornithischia, rather than Sauropoda. The new clade of Theropoda + Ornithischia was called Ornithoscelida, and the Sauropoda were left outside. In our riposte, led by Max Langer from the University of São Paulo, we checked through the huge data matrix assembled by Baron and colleagues and picked holes here and there, and then re-ran the analysis, and recovered the traditional arrangement of Ornithischia and Saurischia – but only just.
A word about how cladistics is done. In studies such as these, the input of effort is not trivial. Matt Baron included seventy-four species in his analysis – not every dinosaur ever named, of course, but more than enough to provide broad coverage of all clades, as well as some silesaurids and other relatives. Each of these seventy-four species was coded for 457 anatomical characters, meaning that Matt (or someone else) had visited numerous museums around the world, pulled out the drawers, and checked each character, usually for a yes/no answer – character present or absent. These are usually coded as ‘1’ for present and ‘0’ for absent. Therefore, Matt Baron ended up with a huge data sheet with 74 rows and 457 columns, for a total of 33,818 cells that had all been checked. In our revision, we ploughed through some of these, focusing on specimens we could access easily, and so were able to correct some of the coding.
What is the point of these huge data compilations? They are, in fact, the basis for the calculation of the best-fitting tree. There are many ways to number-crunch through big data sets like this, seeking the tree, or more likely set of trees, that most efficiently, or most probably, accounts for most of the information. With so much data, it is not likely that a single phylogenetic tree will emerge, nor that it will fit the data perfectly. Much more commonly, there may be 100 or more equally most likely trees and these have to be summarized.
The three possible arrangements of major dinosaurian clades.
So, in the end, the to-and-fro of papers in 2017, and later in 2018, commenting on the radical new Baron proposal, did not lead to a clear answer. What was agreed was that more work should be done – close scrutiny is needed of all those 457 anatomical characters to make sure they are all independent and phylogenetically informative, and then teams of experts need to go back to the museum drawers and check over the fossils in detail.
Even then, there may not be a clear result. This might sound shocking, or an indictment of the cladistic method. It is certainly no criticism of cladistics, though, because the alternative to cladistics (or an equivalent statistical approach) is…nothing. We’d be back to assertion and guesswork – ‘I think the ankle tells us the story’, ‘No, I think the skull characters are more important’.
What we may be seeing here is a star phylogeny, as it is sometimes called, a series of branching points in the phylogenetic tree that perhaps happened very fast, or for which we lack critical fossils. A star phylogeny is an explosion of diversity, evidence of fast evolution of a new clade, and in some cases, perhaps including this one, there was no time for any unique anatomical characters to arise; or they might have been overwritten by later evolution. Perhaps it will be forever difficult to identify the key features that demonstrate once and for all whether the nearest relatives of theropods are sauropodomorphs or ornithischians.
Evolutionary trees, or phylogenies, are the key to understanding evolution. The details of how trees are constructed may seem arcane, and indeed in the past fifty years mathematicians and computer scientists have contributed massively to improvements in the function and speed of the methods. However, the transition from evolutionary trees drawn by hand, and really little more than informed guesswork, to computer-generated cladistic trees has been one of the most profound examples of how dinosaurology has shifted from speculation to science.
The reader may find the to-and-fro of the debates and trees, whether cladistic trees or supertrees, heavy going, but the consequences are far-reaching. These trees are the essential underpinning of how we describe dinosaur evolution through the Triassic, Jurassic, and Cretaceous, and – importantly – how we make calculations of relative rates of change. Identifying the fact that dinosaurs did most of their evolving in the first half of their time on Earth, and then slowed down, is profound. It may be wrong, of course, but the counter-argument can only be made by identifying faults in the original analysis and providing a better hypothesis using better methods and better data.
It’s extraordinary, after our initial efforts back in 1984, and the long decades of reworking and stabilization of the dinosaur tree, that it has all been blown apart again. It will certainly take the combined efforts of numerous experts to check through the data and explore possible solutions between the Saurischia and Ornithoscelida models of fundamental dinosaur relationships.
1This first-found dinosaur bone was later named Scrotum humanum, the first formal Latin name ever given to a dinosaur, by Richard Brookes, in 1763. Sadly, this was to become a nomen oblitum (‘forgotten name’), and the dinosaur was later given the monicker Megalosaurus bucklandii in 1824. But for the fact that the name Scrotum humanum was not widely used, and so became forgotten, the Megalosauridae – the family of dinosaurs containing Megalosaurus bucklandii – might instead have been named the Scrotidae.
2We have made the tree available online, and you can see it in all its glory at http://zoom.it/JJLR – this website allows you to move the tree around, and zoom in to see details of each species.
Chapter 3
Digging Up Dinosaurs
I got excited about dinosaurs when most kids do, about the age of seven or eight. This enthusiasm has never gone away, and it is strongest when I climb into a bouncing four-wheel-drive, in some hot and exotic corner of the world, and we head out into the field. The thrill of planning, reading around, spreading out the maps, and deciding where to go, cannot be beaten. It is a privilege, too, not only to work with professional colleagues in so many countries and continents, but also to live among the local people, and to be there not as a tourist, but as a person on a mission. Most exciting of all is the thrill of wondering what you might find.
Fieldwork is a standard part of any university degree in geology or biology, and I had spent plenty of time squelching around Scotland, following loping professors, and looking at obscure bits of grey rock buried beneath damp fronds. Visiting Elgin, 105 kilometres (65 miles) northwest of Aberdeen, where I grew up, was more promising. Here, the rocks at least were yellow, almost honey-coloured in the watery sun. More importantly, they had produced skeletons of ancient reptiles, and in the coastal sandstone quarry at Clashach, you could still see some footprints on the rocks. There, reptiles large and small had tramped up the lee slopes of the ancient sand dunes, presumably looking for water and plant food, and their prints had survived for over 250
million years, as fresh in every detail as the day on which they had been made. But northeast Scotland was not Mongolia, or Australia, or Canada.
Footprints of an early reptile from Clashach in northeast Scotland.
My chance came when I was an undergraduate at Aberdeen University. I rather cheekily attended the conference of the Society of Vertebrate Palaeontology and Comparative Anatomy in 1976, held that year at University College, London. I was a mere undergraduate, but thought, why not? The professors attending the conference were kind, and did speak to the few gawky students such as myself who were there. During one of the tea breaks, I buttonholed a quiet American professor. He was J. Alan Holman – the ‘J.’ was just an initial – and, quite amazingly, he invited me to go into the field with him from his base at Michigan State University. This was my first trip abroad, at the age of twenty-one, and I spent July to September of the summer of 1977 in Michigan and Nebraska. Holman was then the leading expert on fossil snakes and lizards of North America, and he did a two-month field season each year working through fossil beds in the Valentine Formation. He employed me as his field assistant, and even paid me to dig tonnes of sediment and dump it into great sieves constructed in wooden crates, which we agitated in the river. This washed away the mud and left rocks, twigs, and fossils behind. We boxed up the concentrate and took it back east for sorting and classifying. The humid heat of Nebraska was a shock to a pallid Scot, and the minuscule fossils weren’t quite dinosaurs, but this was living.
After returning to the United Kingdom, I wrote to Phil Currie, then a young researcher at the University of Montreal, who had just got his first job at the University of Alberta in Edmonton. He was the dinosaur man, and is now arguably the greatest living dinosaur expert in North America, or at least one of the top two or three. Currie responded by similarly offering me a job, paid, as his field assistant for the summer of 1978, and we lived for two months in the remote desert-like parts of southern Alberta around Drumheller. This was in Dinosaur Provincial Park, which had been established in 1955, but before the Royal Tyrrell Museum of Palaeontology had been established (it opened in 1985). I have since worked in the field in Germany, Romania, Russia, Tunisia, and China, but the principles of finding and digging up dinosaurs are the same everywhere.
How do palaeontologists find dinosaurs?
The key to finding dinosaurs is to choose the right kind of rocks – they must be the right age and it helps if dinosaurs have been found there before. Dinosaur Provincial Park in Alberta was a good choice, as many skeletons had been excavated there over the previous century. Once you are in the right kind of territory, the secret is good prospecting.
We drove the 280 kilometres (174 miles) from Edmonton to Drumheller in our field truck, a white pick-up with room for three in the front and a flat bed at the back to carry a few tonnes of bones. In the cavalcade was also a sleeping trailer, with beds for six, and a basic kitchen where one of the staff prepared exceptionally salty food and soups. When we complained, he told us we needed salt to replace the electrolytes we were losing in the heat of the sun; you don’t argue with the cook.
The first task I learned was prospecting, walking up and down the coulées. These are deep ravines that have been washed into the landscape by the occasional heavy rains to which this part of Alberta is subject. They cut down through soil and sandstones. The rocks belong to the Dinosaur Park Formation – what else could it be called? Rock formations are units of sedimentary rock (usually) that have a definite bottom and top, stratigraphically speaking, and can be mapped.
The Dinosaur Park Formation is a unit about 70 metres (230 feet) thick comprising green-grey sandstones and mudstones deposited in the latest Cretaceous, some 75 million years ago, in terrestrial environments. The sediments have yielded leaves and trunks of trees, river-dwelling molluscs and fishes, as well as dinosaurs of course – some forty species of them, including the horn-faced ceratopsians Chasmosaurus, Centrosaurus (see overleaf), and Styracosaurus, the duck-billed hadrosaurs Gryposaurus, Lambeosaurus, and Parasaurolophus, the ankylosaur Euoplocephalus (see overleaf) with its tail club, the small, fast-moving predators Ornithomimus (see p. 90) and Dromaeosaurus, and the huge Gorgosaurus, 9 metres (30 feet) long, a close relative of T. rex.
The secret about prospecting for dinosaurs in the badlands is to look for scraps of bone, and follow them back upstream. The coulées had been washed out by erosion, and this happens repeatedly, so any trail of bone fragments in the bottom of a stream can be traced back up the branching streamlets to their source. Then, the job of the trip leader is to decide whether the prospect is worth excavating. Do we have a whole skeleton or just a fragment? You might have spotted just the final scraps of a skeleton, and nothing much would be left behind, or it could be the tip of the tail or a toe bone, and the rest is just waiting for you there, pristine, in the rock – unseen for 75 million years.
Genus:
Centrosaurus
Species:
apertus
Named by:
Lawrence Lambe, 1904
Age:
Late Cretaceous, 77–75 million years ago
Fossil location:
Canada
Classification:
Dinosauria: Ornithischia: Ceratopsia: Ceratopsidae
Length:
6 m (20 ft)
Weight:
2.5 tonnes (4,420 lbs)
Little-known fact:
One location yielding huge numbers of Centrosaurus, near Hilda, Alberta, is probably the richest dinosaur bone bed in the world.
Genus:
Euoplocephalus
Species:
tutus
Named by:
Lawrence Lambe, 1902 (species), 1910 (genus)
Age:
Late Cretaceous, 77–67 million years ago
Fossil location:
United States, Canada
Classification:
Dinosauria: Ornithischia: Thyreophora: Ankylosauridae
Length:
5.5 m (18 ft)
Weight:
2.3 tonnes (5,071 lbs)
Little-known fact:
This dinosaur was so heavily armoured that it even had bony eyelids to protect its eyes.
Genus:
Ornithomimus
Species:
velox
Named by:
Othniel Marsh, 1890
Age:
Late Cretaceous, 75–70 million years ago
Fossil location:
United States, Canada
Classification:
Dinosauria: Saurischia: Theropoda: Ornithomimidae
Length:
3.8 m (12½ ft)
Weight:
170 kg (370 lbs)
Little-known fact:
Ornithomimids did not have teeth so, although they are theropods, they may have had a mixed diet of small animals and plants.
When the first dinosaur bones were found in the American West about 1860, the excavators were not trained scientists. In fact, they were navvies, driving the railroads through the open plains and mountains, and paid by the distance they could advance in a week. They were skilled at shifting rock fast. So, a whack with a great sledgehammer, or leverage by a long-handled spade, and out the bones would pop. They were thrown on flat-bed waggons and hauled by horse to the nearest railhead and sent east to the museums in New Haven, Philadelphia, and New York. Now, these no-nonsense methods would be frowned upon.
Several hours later, dripping with sweat in the roasting temperatures, we had all identified likely prospects. Phil Currie came round to inspect. One I had identified was chosen for our first excavation. The projecting bones showed it was probably a hadrosaur, one of the duck-billed plant-eaters that were hugely common in the Late Cretaceous, but worth extracting as the skeleton seemed complete, and suitable for exhibition.
The bones were lined up along a rather steep slope, so the first job was to build a bench in the rock, by levering and hacking the rock out from above t
he layer that contained the skeleton. You use any means at hand – we even had a huge and uncontrollable pneumatic drill operated by its own engine. It took a week to smash down the overburden to create a bench above the skeleton that would allow us to work with finer tools. We used hammer and chisel and power drills to remove the fine sandstone from above the bones. As we reached the fossils, we had to slow down and be careful, but it’s hard to avoid any slip-ups, and on occasion the chisel gouged a chunk out of the bone – argh!
Digging with power tools to remove rock from above the dinosaur specimen.
How do we record the excavation?
The first priority at excavation is to clear the whole site, so that the skeleton is laid out for view. Once the overburden is removed sufficiently, the site can be properly assessed. We could see the backbone of the dinosaur laid out, including the tail, limbs, and ribs. The skull wasn’t there, however, and the neck was heading right into the cliff. So, we had to cover the site with tarpaulins, and push the cliff back further to extend the bench. For every foot of bench we cleared, we had to remove another yard height of cliff, the slope was so steep. Eventually, we had cleared the site back far enough (or at least we had compromised between shifting another 20 tonnes of overburden to retrieve one more bone and the risk of missing anything else buried under the cliff). At the end of a long day of digging in 30-degree temperatures, we welcomed the chance to rush down and jump into the Sandy River and soak in the cool waters.