Dinosaurs

What Is a Dinosaur?

A dinosaur is not simply any large, extinct reptile. To a palaeontologist the word names a single family tree: the last common ancestor of Triceratops and the house sparrow, plus all of its descendants. That tree took root on the Triassic plains of Pangaea around 233 million years ago, and one branch of it is still singing in your garden.

Dinosaurs belong to the archosaurs, the "ruling reptiles", which split around 250 mya into two lines. One led to crocodiles and their many extinct cousins; the other led to pterosaurs and dinosaurs. What sets dinosaurs apart lies mostly in the hips and legs. Their thigh bone slotted into a hip socket with an open hole through the middle, and their legs were held straight beneath the body like pillars rather than sprawled out to the side. That upright stance freed the chest for breathing while running, an engineering advantage that may help explain their long success.

In 1887 the British palaeontologist Harry Govier Seeley divided dinosaurs into two great groups by the shape of the pelvis. The Saurischia, or lizard-hipped, include the long-necked sauropodomorphs such as Diplodocus and the meat-eating theropods such as Allosaurus. The Ornithischia, or bird-hipped, include Stegosaurus, Triceratops and the duck-bills. The irony is that birds descend from the lizard-hipped theropods, not the bird-hipped line. In 2017 Matthew Baron, David Norman and Paul Barrett proposed regrouping theropods with ornithischians, a suggestion that palaeontologists still debate.

The definition also settles some famous cases of mistaken identity. Pterosaurs were close cousins that flew, but they were not dinosaurs. Plesiosaurs, ichthyosaurs and mosasaurs were marine reptiles from quite separate branches. Dimetrodon, the sail-backed star of many toy boxes, is more closely related to mammals than to any dinosaur.

Did you know? Dimetrodon died out roughly 40 million years before the first dinosaurs appeared, so it never met one.

The oldest secure dinosaurs come from rocks around 233 to 231 million years old in Argentina's Ischigualasto Formation and Brazil's Santa Maria Formation. From those dog-sized beginnings came animals more than 30 m long and others small enough to perch on a twig. Today the bee hummingbird of Cuba, around 5 to 6 cm long, is the smallest dinosaur alive, and the world holds roughly 11,000 species of birds.

Thinking of dinosaurs as a lineage rather than a monster style changed everything. It explains why fossils keep turning up with feathers and wishbones, and it means the story did not end 66 million years ago. Dinosaurs did not vanish; most of them did, and the survivors learned to fly.

How a Fossil Forms

Almost every animal that has ever lived has left no trace at all. Scavengers tear the body apart, bacteria and weather finish the job, and within a few years even big bones crumble to dust. A fossil is the rare exception, and the science of how a living thing becomes one, taphonomy, was named by the Soviet palaeontologist Ivan Efremov in 1940.

The first requirement is speed. A dinosaur that died in or beside a river, lake or shallow sea stood the best chance, because flowing water could bury it under sand and mud before scavengers scattered it. That is why so many dinosaur skeletons come from ancient river systems such as the Late Jurassic Morrison Formation of the American West. At Dinosaur National Monument in Utah, the Carnegie Quarry that Earl Douglass began digging in 1909 preserves a jumble of sauropod, stegosaur and theropod bones dumped by an ancient river on a sandbar; around 1,500 of them remain on show in the rock wall today.

Once buried, the bone is sealed off from oxygen and slowly sinks as more sediment piles on top. Over thousands to millions of years the sand and mud harden into sandstone and mudstone. Groundwater seeps through, carrying dissolved minerals such as silica, calcite and iron compounds. In permineralisation these minerals crystallise in the tiny pores and canals of the bone, turning it heavy and stony. Much of the original bone mineral often survives, which is why scientists can still slice fossils and study their growth rings. Sometimes the bone dissolves entirely, leaving a hollow mould that later fills to make a natural cast.

Did you know? At Lightning Ridge in New South Wales, groundwater rich in silica turned dinosaur bones into precious opal, including the small plant-eater Weewarrasaurus, named in 2018.

Exceptional sites, called Lagerstätten, go further. Fine volcanic ash or still, oxygen-poor water can preserve skin, feathers and stomach contents, as in the Jehol beds of China and the Solnhofen limestone of Bavaria. These rare windows are where feathered dinosaurs first came to light.

Finally the rock must return to the surface. Tectonic forces lift ancient basins into hills, and rain, frost and rivers wear the covering layers away. For a short window, perhaps a few years, a bone lies exposed before erosion destroys it too. It must be found in that window by a farmer, a quarry worker, a child on a beach or a trained field crew walking badlands with eyes on the ground.

Taphonomy matters because it shapes the whole picture. Animals living near water, with big, sturdy bones, are overrepresented; small forest dwellers and upland species are nearly invisible. Every museum hall is a sample biased by death, mud and luck, and good palaeontologists read the gaps as carefully as the bones.

Deep Time and the Geological Clock

Dinosaurs lived across a span of time so long that ordinary words fail. The first dinosaurs and the last Tyrannosaurus are separated by around 165 million years. Stegosaurus had been extinct for roughly 80 million years before T. rex evolved, which means T. rex lived closer in time to us than to Stegosaurus. To make sense of such numbers, geologists built a calendar out of rock.

The idea of deep time owes much to the Scottish farmer and naturalist James Hutton. At Siccar Point in Berwickshire in 1788 he showed friends tilted grey rocks capped by flat red sandstone, evidence of cycles of deposition, uplift and erosion that must have taken immense ages. He concluded that the Earth showed "no vestige of a beginning, no prospect of an end". In 1815 the surveyor William Smith published the first geological map of a whole country, tracing English strata by the fossils they held, and in 1841 John Phillips named the Mesozoic, the "middle life" era of the dinosaurs.

The Mesozoic has three periods. The Triassic, 252 to 201 mya, began after the worst mass extinction in Earth's history and saw the first dinosaurs, around 233 mya. The Jurassic, 201 to 145 mya, was the age of giant sauropods and the first birds. The Cretaceous, 145 to 66 mya, brought flowering plants, Triceratops and T. rex, and ended at the K–Pg boundary, a thin clay layer rich in iridium found from Italy to Montana.

Those dates are not guesses. In 1913 Arthur Holmes published The Age of the Earth, pioneering the use of radioactive decay to measure the age of rocks. Today the workhorse is uranium–lead dating of tiny zircon crystals in volcanic ash. Uranium decays to lead at a known, steady rate, so the ratio of the two acts as a stopwatch that started when the crystal formed. Fossils lying between two dated ash beds are bracketed in time, often to within a few hundred thousand years. Argon dating of ash in Montana places the end-Cretaceous boundary at about 66.0 mya.

Did you know? If Earth's 4.54-billion-year history were squeezed into one day, dinosaurs would appear at about 22:46 and the big ones would vanish at around 23:39, while modern humans arrive in the last six seconds.

Deep time is also a lesson in humility. The Mesozoic was not a single world but a long sequence of them, with drifting continents, shifting climates and casts of animals that never met. Reading it correctly means asking not just "what lived?" but "when, and alongside whom?" The rock column turns a jumble of monsters into history.

Megalosaurus: The First Named Dinosaur

On 20 February 1824 the Reverend William Buckland, Oxford's first reader in geology, stood before the Geological Society of London and described the bones of a "great fossil lizard". He called it Megalosaurus. It was the first dinosaur to be scientifically named, eighteen years before the word "dinosaur" existed.

The bones came from the Stonesfield Slate, thin limestone beds quarried near Woodstock in Oxfordshire for roofing tiles. They are Middle Jurassic, around 166 million years old. Buckland's material included vertebrae, part of a pelvis, limb bones and, most impressively, a piece of the right lower jaw now kept at the Oxford University Museum of Natural History. The great French anatomist Georges Cuvier had examined the specimens when he visited Oxford in 1818 and agreed they belonged to a giant reptile.

The jaw told Buckland a great deal. Its teeth were curved, flattened blades with fine serrations along their edges, like steak knives. Inside the bone, new crowns can be seen growing beneath the old ones, showing that the animal shed and replaced its teeth throughout life as crocodiles do. Clearly this was a predator, and a large one. Buckland estimated a length of about 12 m, scaling up from lizards. Modern studies, based on more complete relatives, suggest Megalosaurus bucklandii (the species name came in 1827) was a theropod around 6 m long and perhaps 0.7 tonnes, walking on two legs, not four.

Did you know? A fossil bone end from Oxfordshire, probably from a Megalosaurus thigh, was illustrated by Robert Plot in 1677 and later labelled Scrotum humanum because of its shape.

Megalosaurus became a celebrity. Charles Dickens opened Bleak House in 1852 with London's mud so thick that "it would not be wonderful to meet a Megalosaurus, forty feet long or so, waddling like an elephantine lizard up Holborn Hill." For decades afterwards, almost any large meat-eating dinosaur tooth found anywhere in the world was dumped into Megalosaurus, until the name became what palaeontologists call a wastebasket taxon. Only in recent years has the genus been trimmed back to the Oxfordshire material and its close kin.

The Stonesfield bones also showed how the science should work. Buckland reasoned from fragments to a whole animal by comparing them with living reptiles, made a bold estimate, and later finds corrected it. The method, not the first guess, is the legacy. A single jaw, read carefully, put a giant predator in the English countryside and opened the door to a lost world.

Mary Anning and the Jurassic Coast

The most important fossil hunter of the early nineteenth century never found a dinosaur. Mary Anning, born in Lyme Regis, Dorset, in 1799, spent her life searching the crumbling cliffs between Lyme and Charmouth for the remains of Jurassic sea reptiles, selling them to make a living. Her finds forced the scientific world to accept that whole kinds of animals had once existed and then vanished.

The cliffs are the Blue Lias, alternating limestone and shale laid down in a warm sea around 200 to 190 million years ago, early in the Jurassic. Winter storms and landslips at places such as Black Ven expose new fossils each year, and also destroy them, so collectors must work fast and in dangerous conditions. Mary's father Richard, a cabinetmaker, sold "curiosities" to tourists; after his death in 1810 the family depended on what Mary and her brother Joseph could find.

In 1811 Joseph dug out a long-snouted skull, and around a year later Mary, aged about 12, uncovered much of the skeleton behind it. It was an ichthyosaur, a dolphin-shaped marine reptile. In 1823 she found the first nearly complete skeleton of a plesiosaur, its neck so absurdly long that Georges Cuvier at first suspected a fake until he saw the evidence. In 1828 came a flying reptile, later named Dimorphodon, the first pterosaur found outside Germany. She also noticed that some stony "bezoar stones" contained fish scales and bones, insight that helped William Buckland identify them as fossil dung, which he named coprolites in 1829.

Did you know? As a baby in 1800 Mary survived a lightning strike that killed three women standing beside her under a tree.

None of these animals were dinosaurs, a common confusion. But they mattered to the dinosaur story because they showed extinction was real and that the ancient world held reptiles beyond imagination. Henry De la Beche's 1830 watercolour Duria Antiquior, the first scene of prehistoric life, was based largely on her finds.

Anning was barred, as a woman, from the Geological Society of London, and many of her fossils entered museums under the names of the gentlemen who bought them. She died of breast cancer in 1847, and the society later marked her death in its proceedings. The popular claim that she inspired the tongue-twister "She sells seashells" is unproven.

Since 2001 her coast has been the Jurassic Coast World Heritage Site, 155 km of cliffs from Exmouth to Studland recording around 185 million years of Earth history. A bronze statue of Mary, with her hammer and dog Tray, was unveiled in Lyme Regis in 2022. The cliffs are still giving up fossils; her reputation, at last, has caught up with them.

Gideon Mantell and Iguanodon

Gideon Mantell was a busy country doctor in Lewes, Sussex, who spent his nights on fossils. Around 1822 he acquired several large, worn teeth from a quarry in Tilgate Forest near Cuckfield. Family tradition says his wife, Mary Ann, spotted the first of them while he visited a patient; historians debate the story, but the teeth themselves changed palaeontology.

They were strange: ridged, chisel-like grinders worn flat by chewing plants, unlike anything in the meat-eating Megalosaurus. Experts were unimpressed. Georges Cuvier at first suggested a rhinoceros, and others thought fish. Then, at the Hunterian Museum in London, the curator's assistant Samuel Stutchbury showed Mantell the skull of a modern iguana. Its little teeth matched his in shape, only many times smaller. In 1825 Mantell published his giant herbivore as Iguanodon, "iguana tooth", the second dinosaur ever named and the first known plant-eater.

Mantell scaled up from the iguana and imagined a colossal reptile, at one point estimating it at 12 to 18 m long. In 1834 a partial skeleton was blasted out of a quarry at Maidstone in Kent. Friends bought it for him, and the "Mantell-piece" became his key specimen. It included a conical spike, which Mantell placed on the animal's nose like a rhinoceros horn. His sketch of a horned, sprawling giant on all fours became the model for the life-size statue later built at Crystal Palace.

Did you know? Maidstone's coat of arms has included an Iguanodon since 1949, in honour of the 1834 find.

The real shape of Iguanodon came from Belgium. In 1878 coal miners at Bernissart, digging around 322 m below ground, broke into a clay-filled fissure packed with skeletons: at least 38 individuals, many complete. Under the direction of Louis Dollo at the Royal Belgian Institute of Natural Sciences in Brussels, the first was mounted in 1882. The skeletons showed the "horn" was a thumb spike, a conical claw on each hand, and that the animal could rear up on its hind legs. Today Iguanodon bernissartensis is reckoned around 10 m long and 3 to 4 tonnes, a beaked plant-eater that probably walked on all fours at times but could move on two. Dollo's upright, tail-dragging kangaroo pose has since been corrected too, with the back held level.

Mantell's life ended in pain and some bitterness. A carriage accident in 1841 injured his spine, and he felt overshadowed by the powerful Richard Owen. Yet his reasoning from a handful of teeth was sound: a big reptile that chewed plants had once roamed southern England. Better skeletons fixed the details, as they still do.

Richard Owen and the Crystal Palace Monsters

By 1840 Britain had three large fossil reptiles that fitted no existing group: Megalosaurus, Iguanodon and the armoured Hylaeosaurus, described by Gideon Mantell in 1833. The anatomist Richard Owen, then at the Hunterian Museum of the Royal College of Surgeons, studied their bones and noticed they shared features, above all a sacrum of five fused vertebrae anchoring the hips and legs held beneath the body like a mammal's. In a report published in 1842, following a lecture at Plymouth in 1841, he united them as Dinosauria, from the Greek for "fearfully great lizards".

Owen had motives beyond anatomy. He imagined dinosaurs as advanced, almost mammal-like reptiles, upright and powerful, and he used them to argue against the evolutionary ideas then circulating, since the reptiles that followed looked to him less sophisticated. The fossils would later serve the opposite argument rather well.

The public met Owen's dinosaurs in 1854. After the Great Exhibition of 1851, the Crystal Palace was moved from Hyde Park to Sydenham in south London, and its new park needed attractions. The sculptor Benjamin Waterhouse Hawkins, advised by Owen, built around 30 life-size models of extinct animals on islands in the lake, arranged roughly in geological order. The dinosaurs were stout, scaly quadrupeds: Megalosaurus on four legs, Iguanodon with a horn on its nose, Hylaeosaurus bristling with spines. They were made with brick and iron frames under a skin of cement.

On 31 December 1853 Hawkins held a famous dinner for scientists and backers inside the mould of one of the Iguanodon models, with Owen at the head. When the park opened on 10 June 1854, crowds flocked to see the monsters, and cheap printed guides and souvenir models spread "dinosaur mania" for the first time.

Did you know? Hawkins later began a similar set of models for New York's Central Park, but in 1871 workmen linked to the Tammany Hall political machine smashed them and buried the pieces.

The statues are wrong in many details, as later fossils showed: no horn on the nose, and Megalosaurus walked on two legs. But they are exactly right as a record of what the best science thought in 1854. The Crystal Palace itself burned down in 1936; the sculptures survived, were listed Grade I in 2007 and have been restored. Owen's other monument is the Natural History Museum in South Kensington, which he campaigned for and which opened in 1881. Between a word and a lake island, he gave the dinosaurs both their name and their public.

The Bone Wars

In the 1870s the American West became the scene of the most bitter feud in the history of science. Edward Drinker Cope of Philadelphia and Othniel Charles Marsh of Yale began as friends, meeting in Berlin in 1864 and even naming species after each other. Within a few years they were enemies, and their rivalry, later called the Bone Wars, filled museums with dinosaurs while tarnishing both men's reputations.

One early spark was Elasmosaurus, a plesiosaur Cope described in 1868 with its head placed on the end of the tail. The error was pointed out publicly, and Marsh delighted in repeating it. Disputes over collecting sites in Kansas and Wyoming followed, and in 1877 the war became a race.

That year three major discoveries were made in rocks of the Late Jurassic Morrison Formation. The schoolteacher and artist Arthur Lakes found huge bones near Morrison, Colorado, and wrote to Marsh. At Garden Park near Cañon City, O. W. Lucas sent finds to Cope. And two Union Pacific railway workers, William Reed and William Carlin, wrote to Marsh about bones weathering out at Como Bluff, Wyoming, which became one of the richest dinosaur sites ever found. Marsh hired them, and his field collectors included the young Samuel Williston. Cope's collectors included Charles H. Sternberg, who worked with him in Montana in 1876.

The tactics grew ugly. Each side poached workers, spied on the other's quarries and bribed railway agents, and both camps were accused of smashing bones to keep them from rivals. The papers were rushed out, sometimes describing an animal from scraps and a telegram. In 1890 the feud spilled into the New York Herald, with each man accusing the other of incompetence and plagiarism.

Did you know? Marsh named Apatosaurus in 1877 and Brontosaurus in 1879, and for over a century the second was treated as the same animal as the first.

The haul was astonishing. Marsh named around 80 new dinosaur species and Cope around 56. Marsh's list includes Stegosaurus, Allosaurus, Diplodocus and Triceratops; Cope's includes Camarasaurus, Coelophysis and Monoclonius. Many were valid and are household names. Others were duplicates or fragments, and palaeontologists have spent more than a century untangling the synonyms both men created in their haste.

Cope died in 1897, short of money after bad investments; Marsh died in 1899, his fortune largely spent. Their war turned the American West into the centre of dinosaur science and gave the world its first truly complete skeletons of giants. It also stands as a warning: rivalry can drive discovery, but haste and spite leave a mess that others must clear up.

Trace Fossils: Footprints, Eggs and Dung

Bones tell us what an animal was. Trace fossils tell us what it did: where it walked, how fast it moved, where it laid its eggs and what it had for dinner. Footprints, nests, burrows, tooth marks and droppings are the frozen moments of dinosaur lives, and they are often far more common than skeletons.

Footprints form when an animal crosses soft mud or sand that then dries and is buried. They were being studied long before anyone understood them. In 1836 Edward Hitchcock described three-toed tracks from the Connecticut Valley as the prints of giant ancient birds, which was wrong at the time and turns out to be oddly prophetic. Tracks are named as their own genera, such as Eubrontes and Grallator, because it is rarely possible to say exactly which species made them. In 1976 the zoologist R. McNeill Alexander showed that stride length and hip height give a walking or running speed, so a trackway can work as a fossil speedometer. Most dinosaur tracks record animals walking at an unhurried 3 to 8 km per hour. Sites such as Cal Orck'o in Bolivia, where a quarry wall holds thousands of prints, show herds, solitary hunters and changes of pace across a single afternoon.

At Lark Quarry in western Queensland, rocks around 95 million years old preserve more than 3,300 small footprints. In the 1980s they were read as a stampede of small dinosaurs fleeing a large predator. A 2013 study argued that many of the prints were made by animals swimming or wading in shallow water instead, and the debate continues.

Eggs were first recognised as dinosaur eggs on a large scale in 1923, when Roy Chapman Andrews's American Museum of Natural History expedition reached the Flaming Cliffs of Mongolia's Gobi Desert. Nests since found on every continent except Antarctica show clutches arranged in rings or pairs, and some still hold embryos.

Did you know? In 1998 Karen Chin described a T. rex coprolite from Saskatchewan 44 cm long, packed with crushed bone from a young plant-eater.

Coprolites, fossil dung, reveal diets directly. William Buckland named them in 1829 after Mary Anning noticed bones inside them. Other traces include gastroliths, polished stones swallowed to help grind food in the gut, and tooth marks scored into bones, which show who bit whom.

Trace fossils have reshaped the picture of dinosaur behaviour more than any other evidence. They show social animals travelling together, careful parents tending nests, and predators that hunted and ate meat bone and all. They also keep palaeontologists honest, because a footprint cannot be moved, rearranged or mounted in the wrong pose. It is a direct record of one moment in a living animal's day.

New Eyes on Old Bones

For most of its history palaeontology meant hammers, brushes and a good eye for anatomy. Those still matter, but since the 1990s a new tool kit has transformed what a fossil can reveal. The same bone that once gave a name and a length can now yield growth rates, brain shapes, colours, body temperatures and a place on a family tree tested by computer.

CT scanning, borrowed from hospitals, lets researchers look inside fossils without breaking them. Lawrence Witmer's team at Ohio University has scanned dozens of dinosaur skulls to reconstruct their brains, sinuses and inner ears. In 2004 a CT study of the London Archaeopteryx showed its brain and inner ear were already bird-like. More powerful synchrotron X-rays, from facilities such as the European Synchrotron in Grenoble, can image embryos still curled inside unhatched eggs.

Bone histology takes the opposite approach: a thin slice of bone, ground until light passes through it, viewed under a microscope. Like tree rings, lines of arrested growth mark each year. In 2004 Gregory Erickson and colleagues used them to show that Tyrannosaurus grew explosively in its teens, putting on up to around 2 kg a day, and reached adult size at about 20. The same technique detects medullary bone, the calcium store that egg-laying birds build, which identified some specimens as females.

Melanosomes are microscopic pigment bodies. In 2010 a team led by Fucheng Zhang and Mike Benton used a scanning electron microscope to find them in the feathers of Sinosauropteryx, revealing a ginger, banded tail. Isotope chemistry reads diets and climates from teeth and bone. Clumped-isotope studies published in 2011 suggested sauropods ran at around 36 to 38 °C, and oxygen isotopes in Camarasaurus teeth point to seasonal journeys to higher ground.

Did you know? In 2005 Mary Schweitzer reported flexible, vessel-like structures inside a T. rex thigh bone from Montana, a claim that still divides palaeontologists.

Finally, phylogenetics turns anatomy into data. Researchers score hundreds of features across many species and let software find the most likely family tree. That is how the theropod origin of birds became overwhelmingly supported, and how the 2017 proposal to regroup the main dinosaur branches was put to the test.

These methods do not replace fieldwork; they multiply its value. A specimen collected in 1900 can answer questions nobody then thought to ask, which is one reason museums keep even the dullest-looking fragments. Old bones keep giving new answers as the tools improve.

The Great Dying

About 251.9 million years ago, at the boundary between the Permian and Triassic periods, life came closer to being wiped out than at any time before or since. The end-Permian extinction, nicknamed the Great Dying, removed something like four in five marine species, perhaps more, and a large share of life on land. Reefs vanished. The last trilobites, a group that had lasted some 270 million years, disappeared for good.

The clearest record lies in a quarry at Meishan in Zhejiang Province, China. There the pale Changhsing Limestone, crowded with corals, brachiopods and single-celled forams, ends abruptly beneath a thin clay known as Bed 25. That clay is altered volcanic ash, and its zircon crystals can be dated by uranium–lead methods. In 2014 Seth Burgess, Samuel Bowring and Shen Shuzhong used them to show that the main killing took no more than about 60,000 years, a geological instant. In 2001 the official boundary marker, the "golden spike", was fixed just above, at the base of Bed 27c, where the tiny conodont Hindeodus parvus first appears.

The prime suspect is the Siberian Traps, a vast pile of flood basalts in what is now Russia. Magma forced its way through coal seams and salt beds, cooking them and releasing carbon dioxide, methane and sulphur gases in pulses. The fingerprints are everywhere: a sharp plunge in the proportion of carbon-13 in rocks worldwide, oxygen isotopes in conodont teeth suggesting tropical seas heated to around 40 °C, and chemical signs that oceans lost their oxygen and turned more acidic. Zircon dates from Siberia match those from China.

On land, the survivors were an odd assortment. In South Africa's Karoo Basin, the rocks just above the crisis are dominated by Lystrosaurus, a stocky, beaked plant-eater about a metre long. It was a synapsid, a distant cousin of mammals' ancestors, not a dinosaur, and it spread across much of Pangaea as a classic "disaster taxon" thriving in an emptied world. Recovery was slow: complex reefs and forests took millions of years to return.

Did you know? For millions of years after the Great Dying no coal formed anywhere on Earth, a hole in the record that geologists call the "coal gap".

The disaster reset the board. Into the empty niches of the Early Triassic came new kinds of reptile, among them the archosaurs, whose descendants would include crocodiles, pterosaurs, dinosaurs and birds. The Great Dying is also the closest ancient parallel to a rapid release of carbon into the atmosphere, which is why climate scientists study Meishan as closely as palaeontologists do. Every dinosaur story begins in that thin grey clay.

Pangaea

Throughout the Triassic, almost all of Earth's land was gathered into a single supercontinent, Pangaea, "all land". It had come together through continental collisions in the late Palaeozoic, and by the time the first dinosaurs appeared it stretched from pole to pole in a giant C shape. Its northern half, Laurasia, held the cores of North America, Europe and Asia; its southern half, Gondwana, held South America, Africa, India, Antarctica and Australia. Around it lay Panthalassa, a single world ocean, and into its eastern side cut the warm Tethys Sea.

The idea came from Alfred Wegener, a German meteorologist, whose book The Origin of Continents and Oceans appeared in 1915. Wegener pointed to coastlines that fit like jigsaw pieces, matching rock belts on either side of the Atlantic and identical fossils on continents now oceans apart: the small Permian reptile Mesosaurus in Brazil and South Africa, and the Glossopteris seed-fern flora across the southern lands. Most geologists rejected him because he had no convincing engine. That came in the 1960s, when magnetic stripes on the sea floor and fossil magnetism locked into rocks showed that plates really do move, carrying continents with them.

The dinosaur era supplied a clincher. In 1969 an expedition that included the American palaeontologist Edwin Colbert found bones of Lystrosaurus at Coalsack Bluff in Antarctica's Transantarctic Mountains. The same beaked plant-eater was already known from South Africa and India. A land animal about a metre long could not have swum between them; the continents had been joined.

One landmass meant a strange climate. Much of Pangaea's interior lay thousands of kilometres from the sea, and the contrast between hot land and cooler ocean drove a megamonsoon: drenching seasonal rains on the coasts and searing aridity inland. The rocks show it in red beds, salt deposits and fossil dunes, from the Keuper of Germany to the Chinle Formation of the American south-west. A 2015 study of the Chinle at Ghost Ranch argued that wild swings in drought and wildfire kept large plant-eating dinosaurs out of the tropics for tens of millions of years.

Did you know? Wegener never saw his idea accepted: he died on the Greenland ice sheet in 1930, three decades before plate tectonics showed he had been broadly right.

Pangaea explains why Triassic and Early Jurassic dinosaur faunas look so similar on every continent: with no ocean in the way, animals could walk from Argentina to Germany. The break-up began around 200 million years ago, when rifting between North America and Africa started to open the Atlantic. As the pieces drifted apart over the next hundred million years, their dinosaurs went their separate ways. The map of the Mesozoic is the key to its family tree.

The Archosaurs

The archosaurs, or "ruling reptiles", are one of the great success stories of the vertebrate world. Edward Drinker Cope coined the name in 1869. Today the group has just two surviving branches, crocodylians and birds, but in the Triassic it held a far wider array of shapes and lifestyles. Archosaurs share a set of telltale features: teeth set in sockets, an extra opening in the skull in front of each eye called the antorbital fenestra, and another in the lower jaw.

Their story starts in the wreckage of the Great Dying. Early relatives such as Proterosuchus, a sprawling predator from South Africa's Karoo Basin, appear in the very first Triassic rocks, and by about 245 million years ago the big-headed Erythrosuchus was among the largest hunters on land. Soon afterwards the true archosaurs split into two lines, a division first worked out from the structure of the ankle. The crocodile line (Pseudosuchia) kept a rotating, "crocodile-normal" joint; the bird line (Avemetatarsalia) developed a simple hinge, with the legs held straight beneath the body. Modern family trees rest on computer analyses of hundreds of skeletal features, notably Sterling Nesbitt's 2011 survey of archosaur relationships.

Some of the most surprising evidence comes from Tanzania's Manda Beds. Bones collected there in 1933 by the British palaeontologist Rex Parrington were joined by new finds in 2015, and in 2017 Nesbitt and colleagues named Teleocrater rhadinus. Around 245 million years old, it is the oldest known member of the bird line, yet it was a long-necked animal some 2 to 3 m long that walked on four legs and had a crocodile-like ankle. The ancestors of dinosaurs, it turns out, did not start out as nimble bipeds.

The bird line also produced the pterosaurs, the first vertebrates to fly under their own power. They are close cousins of dinosaurs but not dinosaurs, a distinction popular culture still blurs. In 2020 Martín Ezcurra and colleagues used CT scans of skulls and inner ears to show that the pterosaurs' nearest relatives were the lagerpetids, small, quick-footed animals known mainly from the Americas. On the other side of that branch sat the silesaurids, such as Silesaurus from Krasiejów in Poland, and the dinosaurs themselves.

Did you know? A crocodile is more closely related to a sparrow than it is to any lizard or snake.

The living members do a quiet but vital job. Because crocodylians and birds bracket the dinosaurs on the family tree, anything they share was probably present in their common ancestor and its descendants. Both have four-chambered hearts, build nests and guard their young, and alligators turn out to have the same one-way airflow through their lungs that birds do. This "phylogenetic bracket" is how palaeontologists infer soft anatomy and behaviour that never fossilises. The crocodile basking on a riverbank is a family portrait of the Triassic.

The Crocodile-Line Rulers

For most of the Late Triassic the big animals on land were not dinosaurs. They were their cousins on the crocodile line, the pseudosuchians. In the Chinle Formation of Arizona and New Mexico, as in Argentina's Ischigualasto and Brazil's Santa Maria rocks, their bones far outnumber those of early dinosaurs, and they filled almost every role a large land animal could play.

At the top were the rauisuchians, deep-skulled predators with serrated, blade-like teeth. Postosuchus kirkpatricki, named by Sankar Chatterjee in 1985 from the Post Quarry in Texas, reached about 4 to 5 m; Saurosuchus from Ischigualasto was larger still, at 6 m or more, and Prestosuchus stalked southern Brazil. Many of them held their legs beneath the body in a pillar-like stance, evolved quite independently of the dinosaurs' upright gait.

Plant-eating was the business of the aetosaurs, armoured in rows of bony plates called osteoderms and equipped with shovel-like snouts. Stagonolepis comes from the sandstones around Elgin in Scotland; Desmatosuchus of Texas and Arizona carried long spikes over its shoulders, and Typothorax was broad and flat as a coffee table. In the rivers lurked phytosaurs, long-snouted ambushers such as Smilosuchus, around 7 m long, which looked like crocodiles but had their nostrils perched near the eyes rather than at the tip of the snout. Whether they belong inside the crocodile line or just outside it is still debated.

The strangest of all were the shuvosaurids. In the early 2000s Sterling Nesbitt, then a student, opened a plaster jacket collected at Ghost Ranch, New Mexico, in 1947 and found a toothless, beaked animal that ran on two legs. Named Effigia okeeffeae by Nesbitt and Mark Norell in 2006, it looked astonishingly like the ostrich-mimic dinosaurs that evolved tens of millions of years later. Its Texan relative Shuvosaurus had first been mistaken for a dinosaur. Measured by the range of body shapes, a 2008 study led by Stephen Brusatte found the crocodile line more varied than the dinosaurs throughout the Late Triassic.

Did you know? Effigia okeeffeae is named after the painter Georgia O'Keeffe, who lived at Ghost Ranch and painted the very cliffs its fossil came from.

Almost all of this diversity vanished at the end of the Triassic. Only the small, slender crocodylomorphs, such as Hesperosuchus from the Chinle and long-legged Terrestrisuchus from Wales, made it through, and from them come today's crocodiles, alligators and gharials. Their story has rewritten the old picture of dinosaurs as natural winners that outcompeted everything around them. For thirty million years the crocodile line ruled, and the dinosaurs' chance came only when that rule was broken.

Eoraptor and Herrerasaurus

In the dry north-west of Argentina, in San Juan Province, lies Ischigualasto, the "Valley of the Moon": a badland of grey, green and red mudstones laid down by rivers and floodplains in the Late Triassic. A volcanic ash near the base of the Ischigualasto Formation has been dated by argon methods to about 231.4 million years ago, and the park, together with neighbouring Talampaya, is a UNESCO World Heritage Site. Its rocks hold some of the oldest well-preserved dinosaurs on Earth.

The first to be recognised was Herrerasaurus ischigualastensis. Its bones were noticed in 1959 by Victorino Herrera, a local farmer, and the animal was named in his honour by Osvaldo Reig in 1963. For decades it was known only from partial skeletons, until an expedition in 1988 led by Paul Sereno with Argentine colleagues found a nearly complete skeleton with its skull. It was a fast, two-legged predator up to about 6 m long, perhaps 350 kg, with serrated, backward-curving teeth, grasping hands with three long, clawed fingers and a sliding joint within the lower jaw that helped it hold struggling prey.

Eoraptor lunensis, the "dawn plunderer", was found in 1991 by Ricardo Martínez on a Sereno-led team and named in 1993. It was tiny by comparison: about 1 m long and 10 kg, light enough to carry under one arm. Its teeth are a mix, hooked at the front of the jaw and leaf-shaped further back, suggesting a generalist that ate small animals and plants alike. Hailed at first as an early theropod, it was moved in 2011, when Martínez and colleagues described the related Eodromaeus, to the base of the long-necked sauropodomorph line.

These two were not alone. The same rocks have yielded Eodromaeus, an early theropod, the sauropodomorph Panphagia, and Sanjuansaurus, a relative of Herrerasaurus. Even this early, then, the dinosaurs had split into recognisable branches. Exactly where Herrerasaurus fits remains contested: many analyses make it an early theropod, others place it just outside the main saurischian groups.

Did you know? In Ischigualasto's lower beds the commonest large animal is not a dinosaur but Hyperodapedon, a beaked, pig-sized plant-eater called a rhynchosaur.

What Ischigualasto changed is our sense of proportion. The earliest dinosaurs were neither giants nor rulers. They were rare, mostly small members of a crowded community dominated by rhynchosaurs, mammal-like cynodonts and the crocodile line. Yet the basic design was already in place: upright legs, a light, agile build and a range of diets that would later be pushed to extremes. The dawn of the dinosaurs was quiet, and it happened in Argentina.

Hunting for the First Dinosaur

Nobody has found the first dinosaur, and nobody expects to recognise it if they do. Dinosaurs are defined by a package of features, among them an open hip socket, a long crest on the upper arm bone for powerful muscles and legs held straight beneath the body. Their closest relatives, the dinosauromorphs, acquired those features piece by piece, so the boundary between "nearly a dinosaur" and "dinosaur" is blurred. The hunt is really a search for the moment the package came together, and for where.

Some of the earliest clues are footprints. In 2011 Stephen Brusatte, Grzegorz Niedźwiedzki and Richard Butler described tiny tracks called Prorotodactylus from the Holy Cross Mountains of Poland, about 250 million years old. The narrow trackways and close-set toes show a small dinosaur relative walking on upright legs within a few million years of the Great Dying.

Bones push the story forward in southern Pangaea. Tanzania's Manda Beds have yielded Asilisaurus, named in 2010, a silesaurid close to the dinosaur line whose existence implies that dinosaurs must already have branched off. From the same rocks comes Nyasasaurus parringtoni, around 243 million years old: an upper arm bone and a few vertebrae collected by Rex Parrington in the 1930s, studied by Alan Charig in the 1950s and formally named in 2013 by Sterling Nesbitt and colleagues. Its bone tissue shows fast, dinosaur-like growth. It is either the oldest dinosaur known or its closest cousin.

In Argentina, the Chañares Formation of La Rioja produced small, long-legged relatives such as Lagerpeton and Marasuchus during Alfred Romer's Harvard expeditions in the 1960s. Zircon dating published in 2016 made these rocks about 236 million years old, younger than once thought. The oldest unquestioned dinosaurs follow soon after: from Brazil's Santa Maria Formation, dated to about 233 million years ago, come Staurikosaurus, named by Edwin Colbert in 1970, Saturnalia and Buriolestes; from Ischigualasto, a little younger, come Eoraptor and Herrerasaurus.

Did you know? Nyasasaurus waited more than 75 years for a formal name: its bones were collected in the 1930s but not described until 2013.

Africa entered the record in 2022, when Christopher Griffin and colleagues described Mbiresaurus raathi from Zimbabwe, a sauropodomorph about 230 million years old. Its age and setting support the idea that early dinosaurs spread along a humid southern belt of Pangaea before reaching the drier tropics. Meanwhile the family tree itself is in question: in 2017 Matthew Baron, David Norman and Paul Barrett proposed regrouping theropods with ornithischians as the Ornithoscelida, a challenge to the 130-year-old division that is still being tested. The first dinosaur probably lies in unexplored Middle Triassic rocks of Gondwana, waiting.

Coelophysis at Ghost Ranch

Coelophysis bauri was a slender, lightly built predator about 3 m long, much of it neck and tail, weighing perhaps 20 kg. Its name means "hollow form", after its thin-walled bones. It lived in the Late Triassic, around 210 million years ago, in the river country of what is now New Mexico, then part of tropical Pangaea. It is one of the earliest well-known theropods, the group that would later include Tyrannosaurus and the birds.

The name is older than the famous fossils. Fragments collected by David Baldwin in 1881 went to Edward Drinker Cope, who named the animal in 1889. The breakthrough came in June 1947, when a team from the American Museum of Natural History led by Edwin Colbert visited Ghost Ranch in northern New Mexico. Field assistant George Whitaker found bones weathering out of the Chinle Formation, and the quarry turned out to hold hundreds of skeletons packed into a single layer, from juveniles to full-grown adults, many of them complete and articulated. It remains one of the richest dinosaur bonebeds known.

So many skeletons could not be dug out one by one. The bonebed was cut into large blocks, wrapped in plaster jackets and shipped to museums including the Carnegie Museum in Pittsburgh, the New Mexico Museum of Natural History in Albuquerque and the Ruth Hall Museum of Paleontology at Ghost Ranch itself, where visitors can still watch preparators at work. How the animals died is debated. Most researchers think drought drew them to shrinking pools and floods then swept the carcasses together into a stream channel.

The quarry also produced a famous myth. Two skeletons seemed to contain the bones of baby Coelophysis in their bellies, and for decades the dinosaur was cited as a cannibal. In 2006 Sterling Nesbitt and colleagues re-examined them: one "meal" lay beneath the skeleton rather than inside it, and the other belonged to a small crocodile-line reptile, probably related to Hesperosuchus. Studies of the many individuals have since shown that animals of the same size could be at very different stages of growth.

Did you know? In 1998 a Coelophysis skull from the Carnegie Museum flew aboard the Space Shuttle Endeavour to the Mir space station, making it one of the first dinosaurs in space.

Ghost Ranch matters because it offers something palaeontology rarely gets: a population rather than a single specimen. Hundreds of individuals let researchers measure growth, variation between sexes or morphs, and the range of body sizes in a single species. New Mexico made Coelophysis its state fossil in 1981. A single bad season, 210 million years ago, became one of the best windows on any dinosaur.

Plateosaurus and the Early Long-Necks

Plateosaurus engelhardti is the best-known early long-necked dinosaur, a member of the sauropodomorphs, the lineage that would later produce the giant sauropods. It was named in 1837 by the German palaeontologist Hermann von Meyer from bones found near Nuremberg. It lived around 214 to 204 million years ago, in the Norian age of the Late Triassic, across what is now Germany, Switzerland and France, and adults reached between about 5 and 10 m in length.

What makes it famous is quantity. At Trossingen in south-west Germany, excavations between 1911 and 1932, led first by Eberhard Fraas and later by Friedrich von Huene, uncovered dozens of skeletons. Halberstadt in Germany and Frick in Switzerland have added many more. Early researchers imagined herds dying on desert treks, but in the 1990s Martin Sander argued that heavy animals were bogged down in soft mud and died where they sank.

The anatomy is a sketch of what was to come. A small skull carried leaf-shaped, serrated teeth suited to shredding plants, perhaps with the occasional small animal. A long neck of ten vertebrae let it browse high without moving its body. The hands bore a large thumb claw. For a century artists drew Plateosaurus on all fours, but in 2010 Heinrich Mallison built a digital model of the skeleton and showed that its forearm could not rotate to turn the palm down. It walked on two legs, with its hands free for grasping and the palms facing inward.

Bone histology added a twist. In 2005 Martin Sander and Nicole Klein cut thin sections of Plateosaurus bones and counted their growth lines. Individuals of the same age differed greatly in size, and fully grown adults ranged from about 4.8 to 10 m. Their growth seems to have depended on food and conditions, a flexibility called developmental plasticity that is more typical of modern reptiles than of birds or mammals.

Did you know? Plateosaurus was named five years before Richard Owen coined the word "dinosaur" in 1842.

Relatives turn up all over Pangaea: Thecodontosaurus from Bristol, named in 1836; Massospondylus from South Africa, whose eggs with tiny embryos were described in 2005; Lufengosaurus from Yunnan, named by Yang Zhongjian in 1941; and Ingentia prima from Argentina, described in 2018, which already reached around 10 m in the Late Triassic. Together they show that the long-necked body plan, and the drive to great size, began earlier than once thought. In the small head and long neck of Plateosaurus lies the blueprint for the largest animals ever to walk the Earth.

The Carnian Pluvial Episode

Around 234 million years ago, in the Carnian age of the Late Triassic, the climate of Pangaea flipped. For a million years or more, pulses of heavy rainfall swept a supercontinent better known for red deserts. Geologists call it the Carnian Pluvial Episode, "pluvial" meaning rainy, and over the past few decades it has grown from an oddity in British rocks into one of the turning points of the Triassic.

It was first described in 1989 by two British geologists, Michael Simms and Alastair Ruffell. In the Mercia Mudstone of southern England, layer upon layer of dry red sediment is interrupted by grey and green clays and sandstones laid down by water. Simms and Ruffell matched this to similar changes in the Alps, such as the Lunz sandstones of Austria and the dark Reingraben shales, and argued that the whole world had turned wet. Many colleagues were sceptical. Since then the signal has been found from the Dolomites of Italy to China, and a 2020 review led by Jacopo Dal Corso pulled the evidence together.

The likely trigger was Wrangellia, a large igneous province of flood basalts erupted on an oceanic plateau whose remains now lie in Alaska and British Columbia. Sharp drops in the share of carbon-13 in Carnian rocks record repeated injections of light carbon into the atmosphere, matching the eruptions. The world warmed by several degrees, the water cycle intensified, and the oceans suffered. Reef-building slowed, and ammonoids, crinoids and other marine groups lost many species.

On land, the effects show up in plants and resin. Conifer forests spread, and trees produced resin in unusual quantities: the Dolomites hold some of the oldest abundant amber in the world, about 230 million years old. In 2012 Alexander Schmidt and colleagues found two tiny mites and a fly inside droplets of it, among the oldest animals ever found in amber. Some researchers link the episode to the decline of rhynchosaurs and big dicynodonts, though how much of that the rains caused is still debated.

Did you know? The two mites found in the Dolomites amber were gall mites, a group that still feeds on plants today, though these ones lived on conifers.

The dinosaurs did well out of it. In 2018 Massimo Bernardi and colleagues showed that in the Dolomites dinosaur footprints went from rare to common across the episode, and dinosaurs became far more abundant and widespread in its aftermath. The rains also coincide with the appearance or spread of modern conifer families and of other groups that would shape the Mesozoic world. The Carnian Pluvial Episode is now seen as one of the moments when the modern world was set in motion, and when the dinosaurs got their first break.

The End-Triassic Extinction

About 201.5 million years ago the Triassic ended in one of the five great mass extinctions of the past half-billion years. Roughly three in four species disappeared. In the sea, the conodonts, tiny eel-like animals whose teeth date rocks around the world, vanished for good, and reefs and ammonoids were hit hard. On land, almost the whole crocodile-line cast that had ruled the Late Triassic, the aetosaurs, phytosaurs and rauisuchians, was wiped out, along with most of the big amphibians. The dinosaurs came through.

The cause lies in the birth of the Atlantic. As Pangaea began to rift, enormous volumes of magma erupted as the Central Atlantic Magmatic Province, or CAMP, whose remains now cover parts of four continents: the cliffs of the Palisades on New Jersey's Hudson shore, the North Mountain basalt of Nova Scotia, the High Atlas of Morocco and large areas of Brazil. In 2013 Terrence Blackburn, Paul Olsen and colleagues dated zircons from these rocks and showed that CAMP erupted in pulses over about 600,000 years, with the first matching the extinction horizon. Rhythmic lake sediments in the Newark Basin of the eastern United States, tuned to cycles in Earth's orbit, provided the timescale.

The killing worked through the atmosphere. Fossil leaves from East Greenland studied by Jennifer McElwain and colleagues in 1999 have fewer breathing pores across the boundary, a sign that carbon dioxide levels rose steeply. Warming and more acidic seas struck reef builders and shelled animals, while sulphur gases probably brought short, sharp volcanic winters. On land, pollen of conifers drops away and fern spores surge, the "fern spike" recorded from the Newark Basin to St Audrie's Bay in Somerset. The base of the Jurassic itself was formally fixed in 2010 at Kuhjoch in Austria, at the first appearance of the ammonite Psiloceras spelae.

Why dinosaurs survived is not settled. One idea, put forward by Olsen and colleagues in 2022, comes from the Junggar Basin of north-west China, where lake deposits with ice-rafted pebbles lie close to dinosaur footprints. At high latitudes, they argue, dinosaurs had already adapted to freezing winters, perhaps with insulating fuzz, and so were ready for volcanic cold snaps that killed their rivals. Others stress their fast growth and efficient, bird-like breathing.

Did you know? The Palisades cliffs that face Manhattan across the Hudson River are made of magma from the same eruptions that probably triggered the end-Triassic extinction.

The aftermath was dramatic. In the Newark Basin footprint record, larger three-toed Eubrontes tracks appear just after the boundary, as Olsen and colleagues reported in 2002, a sign that bigger theropods quickly spread into the emptied landscape. Within a few million years the Early Jurassic held crested hunters, the first armoured dinosaurs and ever-larger long-necks. The crocodile line had stumbled. The dinosaurs walked through the door it left open.

Sauropods: Life at Full Stretch

Sauropods were the long-necked, long-tailed plant-eaters that became the largest animals ever to walk on land. Their body plan is one of the most successful designs in the history of life: a small head, a neck that in some species reached 15 m, a barrel of a gut slung between four pillar-like legs, and a long tail to balance it all. The group appeared in the Late Triassic, around 210 mya, and lasted until the asteroid struck 66 mya, a run of some 140 million years that left fossils on every continent.

The early chapters come from Argentina and southern Africa. Ingentia prima, described by Cecilia Apaldetti and colleagues in 2018 from the Quebrada del Barro Formation of San Juan Province, shows that giant size had already begun before the Jurassic. By the Early Jurassic, Vulcanodon from Zimbabwe, named by Michael Raath in 1972, already stood on straight, column-like limbs, the stance every later sauropod kept.

For much of the 20th century sauropods were pictured wallowing in swamps, supposedly too heavy to stand on dry land. The evidence buried that idea. In 1938 Roland T. Bird of the American Museum of Natural History followed sauropod trackways along the Paluxy River near Glen Rose, Texas, showing animals walking across firm mud around 113 mya. Limb bones built like columns, compact feet and deep, narrow ribcages are those of land walkers; in the early 1970s Robert Bakker compared them with elephants, not hippos. The hands are telling too: the palm bones stood upright in a horseshoe, which is why sauropod hand prints are crescent-shaped.

Sauropods did not chew. Peg- or spoon-shaped teeth stripped and cropped foliage, and digestion happened in a vast gut where plant matter probably fermented for days. The stomach stones once thought to grind it are rare, and in 2007 Oliver Wings showed they were too few to work like a bird's gizzard. Growth was fast. Thin sections of limb bone, studied by Martin Sander and colleagues in Bonn, show fibrolamellar bone laid down at rates like those of large mammals, so a hatchling small enough to hold in two hands could reach adult size in perhaps two or three decades.

Did you know? Some late titanosaurs had lost the finger bones of their hands entirely, walking on a horseshoe of bare palm bones with no fingers at all.

Sander's team argued that sauropod size came from a cascade of features working together. Not chewing let the head stay small; a small head allowed a long neck; a bird-like breathing system lightened the skeleton; and laying many small eggs let populations recover quickly. Each feature on its own is unremarkable. Together they made sauropods the benchmark every claim about giant land animals still has to meet.

Diplodocus

Diplodocus is the classic sauropod: about 25 m of neck, body and whip-like tail, yet weighing perhaps 12 to 15 tonnes, light for its length. It lived on the floodplains of western North America around 154 to 152 mya, in the rocks now called the Morrison Formation, alongside Allosaurus, Stegosaurus and several other giant plant-eaters.

The first bones were dug up in 1877 near Cañon City, Colorado, by Benjamin Mudge and Samuel Williston, collecting for Othniel Charles Marsh of Yale. Marsh named Diplodocus longus in 1878. The name means "double beam", after the twin-runnered chevron bones under the tail, which may have protected its blood vessels when the tail pressed on the ground. The best skeleton came later. In July 1899 a crew funded by the steel magnate Andrew Carnegie found a partial skeleton at Sheep Creek, Wyoming, and in 1901 John Bell Hatcher named it Diplodocus carnegii.

The skeleton is a lesson in engineering. Around 15 neck vertebrae and some 80 tail vertebrae balanced over the hips like the two arms of a suspension bridge, the tail tapering to a slender rod of bones. In 1997 Nathan Myhrvold and Philip Currie used computer models to suggest the tail tip could crack like a bullwhip, faster than sound. A 2022 study by Simone Conti and colleagues modelled the tail again and found the tip would have moved at around 30 m per second, far too slow for a sonic boom.

The long, low skull carried pencil-like teeth only at the front of the jaws, ideal for stripping leaves from branches. Wear marks suggest it raked vegetation rather than biting it. How high it held its head is still argued over. In 1999 Kent Stevens and Michael Parrish built digital models that kept the neck close to horizontal, sweeping low plants like a vacuum hose; in 2009 Mike Taylor, Mathew Wedel and Darren Naish pointed out that living animals hold their necks raised, so sauropods probably did too. The family was not small: Diplodocus hallorum from New Mexico, first named Seismosaurus by David Gillette in 1991, may have passed 30 m.

Did you know? By counting daily growth lines in the teeth, Michael D'Emic and colleagues found in 2013 that each Diplodocus tooth was replaced roughly every 35 days.

Thanks to Carnegie's plaster casts, sent to museums around the world, Diplodocus became the first dinosaur many people ever saw. For scientists it remains the reference animal for the whole diplodocid family, the skeleton against which new finds are measured and the test case for every model of how a 25 m animal could stand, feed and move. A century and a half after Cañon City, it still has arguments left in it.

Brachiosaurus and Giraffatitan

The brachiosaurids were built like giraffes: front legs longer than the back, shoulders high, neck carried upward. Brachiosaurus altithorax, the "arm lizard with a deep chest", was named in 1903 by Elmer Riggs of the Field Columbian Museum in Chicago, from a partial skeleton his team dug up in 1900 in the Grand River Valley near Fruita, western Colorado. It came from the Morrison Formation, around 154 to 150 mya.

The best material came from Africa. Between 1909 and 1913 a German expedition led by Werner Janensch and Edwin Hennig excavated Tendaguru Hill in what was then German East Africa, now Tanzania. Hundreds of local workers, with foremen such as Boheti bin Amrani, dug, plastered and carried the bones on foot to the port of Lindi, several days' walk away. More than 200 tonnes of fossils reached Berlin. In 1914 Janensch named the biggest sauropod Brachiosaurus brancai.

Its skeleton, mounted in Berlin's Museum für Naturkunde in 1937 and remounted in 2007, became the tallest mounted dinosaur in the world. Weight estimates show how much method matters: in 1962 Edwin Colbert put it at about 78 tonnes, while in 2009 Mike Taylor, using a careful volumetric model, arrived at about 23 tonnes. The skull told a story too. Its high bony arch once suggested nostrils on top of the head, used as a snorkel, but in 1951 Kenneth Kermack showed that water pressure at depth would have made breathing impossible, and in 2001 Lawrence Witmer showed the fleshy nostrils sat far forward, near the tip of the snout.

Did you know? The Berlin skeleton stands 13.27 m tall, recognised by Guinness World Records as the tallest mounted dinosaur skeleton anywhere.

Were the American and African animals really one genus? Gregory Paul thought not, placing the Tanzanian species in a subgenus he called Giraffatitan in 1988. In 2009 Taylor compared the bones one by one and found many differences: Brachiosaurus had a longer trunk and differently shaped back vertebrae and hip bones. Most palaeontologists now accept Giraffatitan brancai as a genus of its own, though a few still lump the two.

The giraffe shape mattered in the Morrison and Tendaguru forests. While diplodocids cropped vegetation at low and middle heights, brachiosaurids browsed treetops perhaps 9 m or more up, a division of the food supply that helped many giants live side by side. Brachiosaurus also had a second career: in 1993 it was the first dinosaur seen in full in Jurassic Park, rearing for the leaves. Its story is now a reminder that a famous name can hide two animals, and that it took a bone-by-bone comparison to tell them apart.

Brontosaurus Returns

For most of the 20th century the best-known dinosaur name in the world was officially wrong. In 1877 Othniel Charles Marsh named Apatosaurus ajax from a partial skeleton found in the Morrison Formation of Colorado. Two years later, in 1879, he named Brontosaurus excelsus, the "thunder lizard", from a bigger skeleton collected at Como Bluff, Wyoming, by his collector William Reed. These were the years of the Bone Wars with Edward Drinker Cope, and Marsh named new genera at speed.

In 1903 Elmer Riggs of the Field Columbian Museum looked again and concluded that the two were too similar to separate. Under the rules of zoological naming the older name wins, so Brontosaurus became Apatosaurus. The public never quite accepted it. Yale's Peabody Museum mounted its skeleton in 1931 under the name Brontosaurus, a sauropod became an oil company logo, and when the United States Postal Service issued a Brontosaurus stamp in 1989, palaeontologists complained about the "invalid" name.

The head was wrong too. Marsh's 1883 reconstruction used a short, boxy skull like that of Camarasaurus. At the Carnegie Museum in Pittsburgh, William Holland suspected that a slender, Diplodocus-like skull found near the skeleton of Apatosaurus louisae at what is now Dinosaur National Monument was the real one, but he never published the case, and after his death a Camarasaurus-style head was placed on the mount. In 1975 John McIntosh and David Berman showed Holland had been right, and in 1979 the Carnegie mount finally got the correct head.

Did you know? The 1989 United States stamp set was criticised twice over: for using the name Brontosaurus and for including Pteranodon, which is not a dinosaur.

Then, in 2015, the thunder lizard came back. Emanuel Tschopp, Octávio Mateus and Roger Benson published a specimen-by-specimen analysis of the diplodocid family, scoring 81 specimens for 477 anatomical features. Their tree split the apatosaurines into two groups as different from each other as other accepted genera are. Apatosaurus, with an extraordinarily broad, massive neck, stood on one branch; Brontosaurus, with three species, B. excelsus, B. parvus and B. yahnahpin, stood on the other. These were animals of around 20 m and 15 tonnes or more, living around 155 to 152 mya.

Not everyone agrees. Where one genus ends and another begins is partly a matter of judgement, and some specialists still prefer to keep everything in Apatosaurus. But the method is the real story: the decision now rests on measurable features and a published tree that anyone can test, rather than on one expert's eye. A name sunk in 1903 was raised 112 years later by data, and the thunder lizard is back in the textbooks.

Argentinosaurus and the Titanosaurs

In 1987 the rancher Guillermo Heredia found what he took to be a petrified log on his land near Plaza Huincul, in Argentina's Neuquén Province. It was a sauropod leg bone. Excavations followed, and in 1993 José Bonaparte and Rodolfo Coria named Argentinosaurus huinculensis. It lived around 95 mya, in the rocks of the Huincul Formation, and it is among the strongest candidates for the heaviest land animal ever.

The trouble is how little survives: a handful of back vertebrae, part of the hip region, ribs and leg bones. One vertebra stands about 1.6 m tall. From these fragments, and comparisons with better-known relatives, most estimates put Argentinosaurus at around 30 to 35 m long and roughly 65 to 80 tonnes. The uncertainty is honest: with so few bones, every figure leans on assumptions about the missing ones.

Argentinosaurus belongs to the titanosaurs, the last great radiation of sauropods. They spread through the Cretaceous, especially across the southern continents, and many walked with their limbs set wide apart, leaving "wide-gauge" trackways. In 1980 José Bonaparte and Jaime Powell described Saltasaurus loricatus from Salta Province, which proved that some sauropods carried bony armour plates in their skin. In 2017 José Luis Carballido, Diego Pol and colleagues named Patagotitan mayorum from at least six individuals found at La Flecha ranch near Trelew, Chubut, around 101 mya; a revised estimate puts it near 57 tonnes. In 2014 Kenneth Lacovara's team described Dreadnoughtus from Santa Cruz Province, one of the most complete giant titanosaurs known.

They were not only South American. Rapetosaurus, described by Kristina Curry Rogers and Catherine Forster in 2001, comes from Madagascar, and Alamosaurus roamed New Mexico and Texas right up to the end of the Cretaceous. In India, fossil dung from the Lameta Formation, attributed to titanosaurs such as Isisaurus, turned out to contain microscopic silica bodies from grasses.

Did you know? Those Indian dung fossils, published by Vandana Prasad and colleagues in 2005, showed that grasses existed in the age of dinosaurs, millions of years earlier than many botanists had assumed.

Titanosaurs overturned two old assumptions. The southern continents, once seen as a fossil backwater, turned out to hold the biggest dinosaurs of all, which is why Argentina is now a world centre of sauropod research. And sauropods did not dwindle away after the Jurassic: titanosaurs were thriving, diverse and enormous when the asteroid hit 66 mya. The largest animals ever to walk the Earth were still walking at the very end.

How Big Could They Get?

Weighing an animal that died 95 million years ago is harder than it sounds. Palaeontologists have bones, often only a few, and two main ways to turn them into tonnes. The first is scaling: in 2012 Nicolás Campione and David Evans showed that the combined circumference of the upper arm and thigh bones predicts body mass well in living four-legged animals, so the same equation can be applied to dinosaurs. The second is volumetric: build a three-dimensional model of the body, measure its volume and multiply by an assumed density, lowered for sauropods by their air-filled bones. In 2012 Bill Sellers and colleagues used laser scans of mounted skeletons and wrapped them in the tightest possible "convex hull", then added a correction learned from living animals.

The results can differ by a factor of three for the same animal. Giraffatitan was put at about 78 tonnes by Edwin Colbert in 1962 and about 23 tonnes by Mike Taylor in 2009. For the true heavyweights, Argentinosaurus and Patagotitan, careful estimates mostly fall between about 55 and 80 tonnes. For length, the long, slender Supersaurus from Dry Mesa Quarry in Colorado, named by Jim Jensen in 1985, may have reached 33 to 35 m.

Beyond these lie the ghosts. In 1878 Edward Drinker Cope described Amphicoelias fragillimus from a single, enormous vertebra found at Garden Park, Colorado. In 2006 Kenneth Carpenter estimated the animal at 58 m and more than 120 tonnes. But the bone was lost, and in 2018 Carpenter himself reassessed Cope's measurements, renamed it Maraapunisaurus and placed it among the rebbachisaurids, at perhaps 30 to 32 m. India's Bruhathkayosaurus faces similar doubts: its bones were reported in 1989 and have since been lost. The rule is simple: the more extreme the claim, the less bone there tends to be.

Did you know? The only record of Cope's giant vertebra is his published description and drawing; the bone itself vanished, perhaps crumbling to pieces in storage.

Why not bigger still? Weight grows with the cube of length, but bone strength only with the square of its cross-section, so legs must thicken faster than the body grows. A huge animal makes heat faster than it can lose it through its skin, and a raised head 10 m above the heart demands high blood pressure, a problem physiologists still argue over. Sauropods eased these limits with air sacs, which lightened the body and shed heat, and with small eggs laid in large numbers, so populations could survive while each giant took decades to grow. The blue whale, at around 150 tonnes or more, is heavier still, but it is held up by water.

On land, the evidence runs out somewhere around 70 to 80 tonnes. That ceiling may reflect real limits or simply the rarity of giant skeletons. Either way, the question of size has turned palaeontology into a science of method, where the honest answer always comes with a range.

Air Sacs and Hollow Bones

Birds breathe in a way quite unlike mammals. Their lungs are small and rigid, and air flows through them in one direction, pumped by a set of thin-walled air sacs in front of and behind them. The sacs also send branching tubes into the skeleton, hollowing out vertebrae and other bones. This skeletal pneumaticity leaves telltale holes, called pneumatic foramina, that can be seen in fossils. Sauropods have them everywhere.

The first to notice was the Cambridge palaeontologist Harry Seeley, who in 1870 described sauropod vertebrae from the Isle of Wight and remarked on their bird-like cavities. Othniel Marsh noticed the same hollow, lightly built vertebrae in his Morrison sauropods, and in 1947 Werner Janensch argued that the cavities in the Tendaguru giants had been filled with air. For decades, though, the idea stayed a footnote.

Medical scanners changed that. From the late 1990s Mathew Wedel and colleagues put sauropod vertebrae through CT machines and found the insides divided into a honeycomb of air cells, separated by bony plates sometimes only millimetres thick. The neck vertebrae of Sauroposeidon, a brachiosaurid from Oklahoma described by Wedel, Richard Cifelli and Kent Sanders in 2000, are over 1.2 m long and built more like a bird's bones than a lizard's. On the outside, thin bony ridges called laminae braced each vertebra like the struts of a girder. Wedel estimated that air spaces may have lowered a sauropod's overall density by around a tenth.

The system was not unique to sauropods. In 2005 Patrick O'Connor and Leon Claessens showed that the predator Majungasaurus from Madagascar had a bird-like air-sac system too, and similar holes appear in Tyrannosaurus and its relatives. Air sacs seem to have evolved early in the dinosaur family tree, long before the first bird.

Did you know? A predatory dinosaur from Argentina was named Aerosteon, "air bone", in 2008 because its skeleton, including its wishbone and hips, was so riddled with air spaces.

For sauropods the benefits were enormous. Pneumatic bones made a 10 m neck far lighter to hold out. One-way flow solved the problem of "dead space": a windpipe several metres long holds a lot of stale air, which a mammal would have to breathe in and out again, but a flow-through lung keeps fresh air moving across the gas-exchange surfaces. And a body full of air sacs could dump heat, vital in an animal that might weigh 50 tonnes. Air sacs are now seen as one of the keys to sauropod gigantism, and the clearest sign that the long-necked giants breathed more like a sparrow than like an elephant.

Mamenchisaurus and the Longest Necks

In 1952 workers building a highway in Sichuan, China, uncovered a sauropod skeleton. Two years later the pioneering palaeontologist Yang Zhongjian, known in the West as C. C. Young, named it Mamenchisaurus constructus, after the Mamingxi ferry crossing near the site. Young's strong accent is said to have turned "Mamingxi" into "Mamenxi", and the slip stuck.

The genus became famous with a second species. Mamenchisaurus hochuanensis, named by Young and Zhao Xijin in 1972 from a near-complete skeleton found at Hechuan in Sichuan, had 19 neck vertebrae, a neck making up around half of its 22 m length. Most mammals, from mice to giraffes, are locked at seven neck vertebrae. Sauropods escaped that limit by adding extra vertebrae and stretching each one, and the mamenchisaurids of Late Jurassic China took both tricks further than almost anyone.

The record holder is Mamenchisaurus sinocanadorum. Its bones were found in 1987 by the Sino-Canadian Dinosaur Project, led by Dong Zhiming and Philip Currie, in the Shishugou Formation of the Junggar Basin, Xinjiang, around 162 mya. Dale Russell and Zheng Zhong named it in 1993. In 2023 Andrew Moore and colleagues used CT scans of its vertebrae and comparisons with better-known relatives to estimate the neck at about 15.1 m, the longest known for any animal and more than seven times a giraffe's, rivalled perhaps only by Supersaurus.

How did such a neck work? The vertebrae were extremely light, their insides full of air cells from the bird-like air-sac system. Beneath them ran long, thin cervical ribs that overlapped in bundles several vertebrae long, probably acting like tensioned rods that stiffened the neck and helped it carry its own weight. The scans suggest the bones were largely air, with walls only a few millimetres thick.

Did you know? The name Mamenchisaurus honours a ferry crossing that does not share its spelling: it comes from a mispronunciation of Mamingxi.

Why grow such a neck at all? In 2013 Mike Taylor and Mathew Wedel argued that a long neck let a sauropod stand still and sweep a vast feeding area, saving the energy of moving a body of many tonnes. Phil Senter suggested in 2007 that necks may also have been shaped by competition for mates, an idea most researchers find less convincing. Other long-necked giants, such as Sauroposeidon from Oklahoma, with a neck of perhaps 11.5 m, show the same trend evolved more than once.

The mamenchisaurids show what happens when evolution finds a body plan with room to stretch. With no need to chew, air in their bones and extra vertebrae to add, sauropods built necks that no mammal could ever match. A neck of 15 m remains the longest that any animal is known to have carried.

Auca Mahuevo: A Sauropod Nesting Ground

In November 1997 an expedition led by Luis Chiappe, Lowell Dingus and Rodolfo Coria was prospecting the badlands of Neuquén Province in Patagonia when they walked onto a hillside littered with fragments of fossil eggshell. There were thousands of eggs, spread over several square kilometres. They named the site Auca Mahuevo, a play on the nearby volcano Auca Mahuida and the Spanish más huevos, "more eggs". The rocks belong to the Anacleto Formation, laid down by rivers around 80 mya.

The eggs are nearly spherical, about 13 to 15 cm across, roughly the size of a grapefruit, and they lie in clutches of a few dozen. Inside some of them the team found something never seen before: unquestionable sauropod embryos, reported in Nature in 1998. The tiny skulls carried teeth, and patches of fossilised skin showed rows of small, bead-like scales. The anatomy identified the parents as titanosaurs, the group that included giants such as Argentinosaurus.

How we know the nesting story comes from the rock itself. In 2004 Chiappe and colleagues described nests as shallow, rimmed hollows about a metre or so across, apparently scraped out with the hind feet before the eggs were laid. Rivers that overflowed their banks then buried whole clutches in mud, sealing the embryos inside. Most telling of all, the eggs occur at several distinct levels stacked one above another. Generation after generation of titanosaurs came back to the same floodplain to lay, much as sea turtles and many colonial birds return to traditional nesting beaches today.

Did you know? In 2020 synchrotron X-ray scans of a titanosaur embryo from Patagonia, led by Martin Kundrát, revealed a small horn-like prong on the tip of its snout, probably used to break out of the egg.

The site also shows what titanosaur parents probably did not do. There is no sign of adults guarding nests or feeding young, and an animal of many tonnes could hardly sit on a clutch. More likely the eggs were left to the warmth of the sun and soil, and tiny hatchlings fended for themselves from the first day. Other Argentine nesting grounds, such as Sanagasta in La Rioja, where eggs were laid in ground warmed by hot springs, support the same picture.

That strategy helps explain sauropod size. A giant that lays many small eggs every season can afford huge losses among its young, while a mammal of similar size, carrying one calf for a year or more, cannot. Auca Mahuevo turned that idea from a theory into a hillside of evidence. Every giant on the plains began as an egg that would fit in your hand.

Dippy and the Museum Giants

In 1898 the steel magnate Andrew Carnegie read a newspaper story about a gigantic dinosaur found in the American West and decided that his new museum in Pittsburgh must have one. He paid for expeditions, and in July 1899 a crew led by Jacob Wortman, with the veteran collector William Reed, found a partial skeleton at Sheep Creek, Wyoming. In 1901 John Bell Hatcher named it Diplodocus carnegii. Under the museum director William Holland, the preparator Arthur Coggeshall and his team assembled the skeleton, adding bones from other specimens to fill the gaps.

The story then crossed the Atlantic. King Edward VII, seeing a sketch of the skeleton at Carnegie's Skibo Castle in Scotland, is said to have asked for one. Carnegie had the bones moulded and a complete plaster cast made. On 12 May 1905 it was unveiled at the Natural History Museum in London: 292 bones, about 21.3 m long, the first sauropod skeleton most Britons had ever seen. Londoners soon named it Dippy. The original bones went on display in Pittsburgh in 1907.

Carnegie, who hoped science could help keep the peace, then gave casts to heads of state across Europe and beyond. Holland and Coggeshall travelled to install them: Berlin and Paris in 1908, Vienna and Bologna in 1909, St Petersburg in 1910, La Plata in Argentina in 1912, Madrid in 1913 and Mexico City in 1930. The historian Ilja Nieuwland has traced how the casts became gifts of state. For many people across two continents, Diplodocus became simply what a dinosaur looked like.

Did you know? Dippy is not a fossil at all. Every one of its 292 bones is a plaster cast, which is why identical twins of it stand in museums from Madrid to Mexico City.

Dippy has changed shape with the science. Early mounts let the tail drag on the ground; in 1993 London's cast had its tail raised and curved to match trackways that show no tail-drag marks. In 1979 it moved into the museum's great central hall, now Hintze Hall, where it stood until 2017, when it was replaced by the skeleton of a blue whale named Hope. Dippy then toured the United Kingdom from 2018 to 2020, from Dorset to Glasgow, drawing large crowds at every stop.

New giants have taken up the role. In 2016 the American Museum of Natural History in New York installed a 37 m cast of Patagotitan, so long its head pokes out into the corridor, and Chicago's Field Museum unveiled its own Patagotitan, Máximo, in 2018. In Berlin, Giraffatitan still towers over its hall. These casts and skeletons are where most people meet the sauropods. Carnegie's plaster made the science of dinosaurs public, and it turned one Wyoming skeleton into the most widely seen dinosaur on Earth.

A Greenhouse Planet Splits Apart

The Jurassic Period began 201 million years ago with a supercontinent cracking at the seams and ended 145 million years ago with a world of islands, seaways and young oceans. Through all of it the planet ran hot. There were no lasting polar ice caps, carbon dioxide levels were several times today's, and forests of conifers, ginkgoes, cycads and ferns grew far towards both poles. This was the greenhouse in which dinosaurs became the dominant large animals on land.

The break-up of Pangaea had started with the volcanic outpourings that closed the Triassic. Rift valleys opened between North America and north-west Africa, filled with lakes, then with seawater, and by about 180 mya the Central Atlantic was a true ocean floored with new basalt. Further south, around 183 mya, the Karoo–Ferrar eruptions spread lava across what are now southern Africa and Antarctica, and the first cracks began to divide Gondwana into east and west. To the north, the Tethys Ocean pushed westwards and the Gulf of Mexico began to open.

How do we know? The oldest crust on the Atlantic floor, dated by its magnetic stripes and by drilling, is Jurassic, so the ocean cannot be older. Coastlines and rock belts still match across the water, and palaeomagnetism frozen into lavas records how far each continent sat from the pole. Climate leaves chemical traces too. The ratio of oxygen isotopes in the bullet-shaped shells of belemnites records sea temperature, and the black, oxygen-starved shales of Whitby in Yorkshire and Holzmaden in Germany mark the Toarcian Oceanic Anoxic Event of about 183 mya, when warming triggered by the Karoo–Ferrar eruptions starved the seas of oxygen.

Rising sea levels mattered as much as drifting plates. Shallow seas flooded the edges of continents, turning Europe into an archipelago of low islands and pushing the Sundance Sea deep into western North America. Those warm, quiet waters laid down the limestones and mudstones that now yield so many Jurassic fossils, from the ichthyosaurs of Holzmaden to the lagoons of Bavaria and Skye.

Yet the continents were not cut off at once. In the Late Jurassic, Tanzania's Tendaguru beds held long-necked Giraffatitan and plated Kentrosaurus, close cousins of animals in the American West, while Portugal's Lourinhã cliffs have yielded Allosaurus, Torvosaurus and Stegosaurus almost identical to their American relatives. Land bridges and short sea crossings still let herds and hunters move between regions.

Did you know? The Atlantic is still widening today, by a few centimetres a year, roughly the pace at which a fingernail grows.

The slow fragmentation set the stage for the rest of dinosaur history. Separated continents grew their own faunas, which is why the Cretaceous would bring tyrannosaurs to the north and titanosaurs and abelisaurs to the south. The Jurassic map is where that divergence began, one rift valley at a time.

The Morrison Formation

Across the American West, from New Mexico to Montana, runs a band of striped mudstone and sandstone in soft greens, greys and purples. This is the Morrison Formation, named after the town of Morrison, Colorado, where the geologist and teacher Arthur Lakes found giant bones in 1877. It covers roughly 1.5 million square kilometres and records about ten million years of the Late Jurassic, from around 157 to 147 mya. No other rock unit has given the world so many famous dinosaurs.

The Morrison was laid down on a broad, seasonal plain east of a chain of volcanoes. Rivers wandered across it, lakes swelled and shrank, and a long dry season alternated with rains. Conifers and tree ferns lined the channels; ferns and horsetails carpeted the open ground. Volcanic ash drifted in and settled as clay beds, and the zircon crystals in that ash now give precise uranium–lead dates. Geologists divide the formation into members, of which the colourful Brushy Basin Member at the top holds many of the richest quarries.

The fossil sites read like a roll call of palaeontology. Como Bluff in Wyoming and Garden Park in Colorado were battlegrounds of the Bone Wars. At what is now Dinosaur National Monument in Utah, Earl Douglass found the quarry wall in 1909 that still displays around 1,500 bones in place. The Cleveland-Lloyd Quarry, also in Utah, has produced over 12,000 bones, most of them Allosaurus. Barnum Brown's Howe Quarry in Wyoming yielded a tangle of sauropod skeletons with skin impressions in 1934.

What makes the Morrison remarkable is its cast. The sauropods alone include Camarasaurus, the most common, along with Diplodocus, Apatosaurus, Brontosaurus, Barosaurus and Brachiosaurus. Hunting them were Allosaurus, horn-nosed Ceratosaurus and the heavy Torvosaurus, with small Ornitholestes chasing lizards and mammals. Stegosaurus, Camptosaurus, Dryosaurus and the early armoured Gargoyleosaurus browsed lower down. Around their feet lived crocodile relatives such as Goniopholis, turtles, lungfish, frogs, pterosaurs and shrew-sized mammals like Docodon.

Did you know? So many giant plant-eaters lived together that scientists still debate how the Morrison fed them all, and the favourite answer is that each sauropod browsed at a different height and on different plants.

Isotopes in teeth and the wear on them support that idea of niche partitioning. Diplodocus, with its peg-like teeth and low-slung neck, probably raked ferns and low growth, while Camarasaurus and Brachiosaurus, with broader, spoon-shaped teeth, cropped tougher leaves higher up. Sauropod tooth enamel even hints that some herds migrated seasonally to the uplands in search of water.

The Morrison gave museums their first mounted giants and gave science its reference ecosystem for the Jurassic. Whenever a new site turns up in Portugal, Tanzania or China, palaeontologists compare it with this one. It is the yardstick against which the Late Jurassic world is measured.

Allosaurus

For about ten million years of the Late Jurassic, the commonest big predator in western North America was Allosaurus. Othniel Charles Marsh named it in 1877 from a few bones collected at Garden Park, Colorado, calling it the "different lizard" because its vertebrae were lighter than those of other known reptiles. A typical adult was around 8.5 m long and weighed about 2 tonnes, with a few individuals approaching 10 m. It walked on two strong legs, balanced by a long tail, and carried three-fingered arms ending in large, curved claws.

Its skull was long and narrow, with a pair of short bony crests in front of the eyes and blade-like, serrated teeth. In 2001 Emily Rayfield and colleagues used finite-element analysis, an engineering technique that loads a digital model with virtual forces, and found that the skull was strong but the bite surprisingly modest. Their interpretation: Allosaurus opened its jaws wide and struck downwards like a hatchet, slashing flesh rather than crushing bone in the manner of T. rex.

Few dinosaurs are known from so many specimens. The Cleveland-Lloyd Quarry in Utah holds the remains of at least 46 individuals, from juveniles to adults, packed among a much smaller number of plant-eaters. Whether a drought, a predator trap or repeated floods gathered them there is still argued. Bone histology from these and other specimens shows that Allosaurus grew fast as a juvenile, reaching adult size in its mid-teens and rarely living beyond about 30 years.

The most famous individual is Big Al, found in Wyoming in 1991 by a Swiss team and excavated by the Museum of the Rockies. The skeleton is unusually complete, and in 2002 Rebecca Hanna counted 19 bones with injuries or infections: broken ribs, a damaged shoulder blade and a badly infected right foot that must have made every step painful. Life as a big predator was clearly rough. In 2020 Daniel Chure and Mark Loewen made Big Al part of a new species, Allosaurus jimmadseni, the earlier of the two Morrison allosaurs.

Did you know? An Allosaurus tail vertebra from Utah has a hole that matches the size and shape of a Stegosaurus tail spike, so at least one fight between them is written in bone.

Allosaurus was not confined to America. Portugal's Lourinhã Formation has yielded Allosaurus europaeus, more evidence that Iberia and North America were still connected. Studies of how its neck and skull worked, of its tooth marks on sauropod bones and of its injuries suggest a hunter that tackled young sauropods and other mid-sized prey, and perhaps scavenged giants.

Because it is so well sampled, Allosaurus has become a laboratory animal for palaeontology, a large theropod in which growth, variation and injury can be studied across dozens of individuals rather than guessed from one. Long before the tyrannosaurs, it showed what a Jurassic top predator looked like up close.

Stegosaurus

With a double row of tall plates along its back and four long spikes on its tail, Stegosaurus is one of the most recognisable animals ever to have lived. It browsed the Morrison plains of western North America around 155 to 150 mya. Large adults reached around 7 to 9 m in length and several tonnes, yet the head was small and narrow, held barely a metre off the ground, with a toothless beak at the front and small leaf-shaped cheek teeth behind.

Othniel Charles Marsh named it in 1877, meaning "roof lizard", because he first imagined the plates lying flat over the back like tiles. Better skeletons soon showed them standing upright in two staggered rows, usually 17 in all, the largest around 60 cm tall. The plates were not solid shields but thin, bony structures grown within the skin and grooved by blood vessels. Studies of their internal structure under the microscope suggest they were mainly for display and recognition, perhaps flushing with colour, while the old idea that they shed heat remains possible but secondary.

The tail was the real weapon. Its four spikes, each up to about 90 cm long, are nicknamed the thagomizer, a word that began as a joke in a 1982 Far Side cartoon by Gary Larson and was adopted by palaeontologists. In 2005 Kenneth Carpenter and colleagues described an Allosaurus tail vertebra pierced by a wound that matches a Stegosaurus spike, and several stegosaur spikes show broken and healed tips, evidence that they were used in earnest.

The best skeleton of all is Sophie, a young adult found in 2003 at the Red Canyon Ranch in Wyoming and acquired by the Natural History Museum in London, where it went on display in 2014. About 5.6 m long and estimated by Susannah Maidment and colleagues at around 1.6 tonnes, Sophie is the most complete Stegosaurus ever found, complete enough to be laser-scanned and modelled to work out how the animal moved and fed. A few Stegosaurus bones from Portugal show that the genus also roamed Late Jurassic Europe.

Did you know? The story that Stegosaurus had a "second brain" in its hips comes from a large space in the spinal canal, which most scientists now think held a glycogen body like those found in birds today.

The brain inside the skull was genuinely small for an animal of its size, a fact that made Stegosaurus a byword for dimness in the early twentieth century. But brain size alone says little about behaviour, and a small-brained plant-eater with a lethal tail and good defences lasted for millions of years.

Stegosaurus remains the defining member of the stegosaurs, a group that began in the Middle Jurassic of China and spread across the globe. Sophie's bones have turned a cartoon icon into an animal whose posture, weight and fighting style can be tested. That is the quiet revolution of modern palaeontology.

Scelidosaurus and the First Armour

Long before the tail club of Ankylosaurus or the plates of Stegosaurus, an animal about 4 m long was plodding across the coastal lowlands of Early Jurassic Europe with rows of bony studs set into its skin. This is Scelidosaurus, and it is both the earliest well-known armoured dinosaur and one of the first dinosaurs ever found almost complete.

It comes from the crumbling cliffs around Charmouth in Dorset, part of the same Jurassic Coast where Mary Anning had worked a generation earlier. The rocks are marine mudstones about 193 to 191 million years old, so the carcass must have been washed out to sea before sinking. In the 1850s the local quarry owner and surgeon James Harrison collected bones from the foreshore and sent them to Richard Owen, who named Scelidosaurus harrisonii in 1859 and described an almost complete skeleton in the early 1860s. It was the first nearly complete dinosaur skeleton known to science, a decade before most of the American finds.

Its body armour is the reason it matters. Rows of keeled bony plates called osteoderms ran along the neck, back, flanks and tail, and a distinctive three-pointed scute sat behind each side of the head. Skin impressions preserved alongside the bones show smaller scales between the larger studs. Scelidosaurus was a low, heavy plant-eater that walked on all fours, with a small head and leaf-shaped teeth suited to cropping ferns and cycads.

For more than 150 years Owen's skeleton was poorly studied, embedded in hard limestone nodules. Between 2020 and 2021 David Norman of Cambridge published a detailed re-description, bone by bone, of that specimen and others found at Charmouth since. Its place in the family tree is still argued: most analyses put it near the base of the Thyreophora, the "shield-bearers", close to the split that led to stegosaurs on one side and ankylosaurs on the other, though some place it on the ankylosaur line itself.

Did you know? The bones of the first Scelidosaurus were so tightly bound in limestone that some were only freed in the twentieth century by dissolving the rock away in weak acid.

Its relatives show how armour evolved. Scutellosaurus from the Kayenta Formation of Arizona, a little older, was only about 1.2 m long, lightly built and partly bipedal, with hundreds of small scutes. Emausaurus from northern Germany sits nearby on the tree. By the Middle Jurassic, Huayangosaurus in China was already a true stegosaur with plates and spikes.

Scelidosaurus shows that heavy armour arrived early in dinosaur history, within a few million years of the end-Triassic extinction, and that the later tanks and plated giants grew from a modest, studded ancestor. A beachcombing surgeon in Dorset found the root of a whole branch of the dinosaur family tree.

Cryolophosaurus: Antarctica's Crested Hunter

In the summer field season of 1990–91, the geologist William Hammer of Augustana College, Illinois, was working high on Mount Kirkpatrick in the Transantarctic Mountains, close to the Beardmore Glacier. At around 4,000 m above sea level, in thin air and bitter cold, his team found the bones of a large meat-eating dinosaur in a band of Early Jurassic rock. Quarrying them out took jackhammers and rock saws. In 1994 Hammer and William Hickerson named the animal Cryolophosaurus ellioti, the "frozen-crested lizard".

The crest is the giveaway. Instead of running along the length of the skull, as in Dilophosaurus, it rises above the eyes and runs across the head from side to side, curved and furrowed like a comb. The resemblance to a 1950s quiff earned it the nickname "Elvisaurus". The crest was thin and fragile, useless as a weapon, so it was almost certainly for display to rivals and mates. The animal itself was around 6 to 7 m long, a lightly built two-legged hunter and the largest known predator of its time and place.

The bones lie in the Hanson Formation, sandstones and mudstones mixed with volcanic ash. Dating of volcanic material in and around it puts these beds at roughly 194 to 188 mya, and the formation sits just below the dark lavas of the Kirkpatrick Basalt, part of the Karoo–Ferrar outpourings. Hammer returned in 2003 and recovered more of the skeleton, and in 2007 Nathan Smith and colleagues published a full description. Its exact branch on the theropod tree is debated: some studies link it to Dilophosaurus, others place it closer to the line leading to later big theropods.

Did you know? Cryolophosaurus was the first dinosaur from Antarctica to receive a formal scientific name, although an armoured dinosaur had been found on James Ross Island in 1986.

What did its world look like? Antarctica then lay at high southern latitudes, still joined to Africa, India and Australia in Gondwana, but there was no ice cap. Plant fossils from the same region show forests of conifers and ferns that grew through months of summer daylight and survived the dark of winter. Alongside Cryolophosaurus, the Hanson rocks have produced the long-necked sauropodomorph Glacialisaurus hammeri, named in 2007, a pterosaur bone and the remains of tritylodonts, plant-eating relatives of mammals.

Working there is still hard. The field season is a few weeks long, every tool arrives by aircraft, and specimens must be cut out in blocks and flown thousands of kilometres before they can be cleaned. That is why so little of Antarctica's dinosaur record is known, and why each find carries weight.

Cryolophosaurus proved that dinosaurs, including large predators, lived near the poles and thrived in a world of long darkness and mild temperatures. It turned the coldest continent from a blank on the dinosaur map into a place with its own Jurassic ecosystem, and it hinted at how much still lies under the ice.

Guanlong and the Early Tyrannosaurs

The family that produced Tyrannosaurus rex did not begin with giants. Its early members were small, long-armed hunters that lived in the shadow of bigger predators for some 80 million years. The best-known of them is Guanlong wucaii, the "crowned dragon of the five colours", from the Late Jurassic of north-western China.

Guanlong comes from the Shishugou Formation of the Junggar Basin in Xinjiang, around 160 mya. In 2006 a team led by Xu Xing of Beijing's Institute of Vertebrate Paleontology, working with James Clark of George Washington University, described two skeletons: an adult of about 3 m and a younger animal, one lying on top of the other. Growth rings in their bones suggest ages of around 12 and 6 years. The adult's head carried a tall, thin crest running from the snout to the back of the skull, hollow and full of air spaces, far too delicate to fight with. It was a signal, like a cockerel's comb.

How the two came to be stacked was explained in 2010 by David Eberth, Xu and Clark. The pits that held them, roughly 1 to 2 m deep, seem to have been footprints of giant sauropods punched into soft, ash-rich mud. Small animals wandered in, became stuck and were buried on top of one another. The same "death pits" trapped Limusaurus and other small creatures.

Why call Guanlong a tyrannosaur at all? It had long arms and three fingers, not the two tiny ones of T. rex. But its skull shares tell-tale features with its famous relatives: fused nasal bones along the snout and front teeth with a D-shaped cross-section. Phylogenetic analysis, which compares hundreds of such features across many species, places it firmly at the base of the Tyrannosauroidea. Even older is Proceratosaurus from Gloucestershire, England, about 167 mya, another small crested hunter first described in 1910 and only recognised as a tyrannosauroid a century later.

Did you know? The same Shishugou death pits held Limusaurus, a small theropod whose young had teeth but whose adults were completely toothless and beaked.

Later relatives from China's Yixian Formation, around 125 mya, revealed something else. Dilong paradoxus, named in 2004 and under 2 m long, preserved simple filament-like feathers. Yutyrannus huali, named in 2012, was about 9 m long and weighed perhaps 1.4 tonnes, yet it too was covered in filaments up to around 20 cm long. Feathers ran in the family long before the giants arrived.

Only in the last 20 million years of the Cretaceous, after other big predators had faded in Asia and North America, did tyrannosaurs grow into 12 m, multi-tonne bone-crushers. Guanlong is the reminder that the most famous dinosaur family spent most of its history as bit-part players, and that evolution often builds giants from small beginnings.

The Solnhofen Lagoon

Around 150 million years ago, much of southern Germany lay under a warm, shallow sea dotted with low islands. Behind barriers of sponge and coral reefs, a string of quiet lagoons formed near today's towns of Solnhofen, Eichstätt and Langenaltheim in Bavaria. Their floors gathered layer upon layer of fine lime mud, which hardened into the pale, flat-splitting Solnhofen limestone. It is one of the most celebrated fossil deposits on Earth.

The secret of its preservation lies in the water. Cut off from the open sea, the lagoon bottoms became salty and starved of oxygen, toxic to almost anything that sank into them. No worms churned the mud and few scavengers survived to pick carcasses apart. Animals that drowned, or were swept in by storms from the islands and the open sea, settled undisturbed and were sealed by the next film of lime. The result is a deposit, known to geologists as a Plattenkalk, in which jellyfish, dragonfly wings and the outlines of feathers survive.

The stone itself made history. Its grain is so fine and even that in 1796 Alois Senefelder used it to invent lithography, printing from flat stone. Quarrymen have split the slabs by hand for centuries, sorting hard beds, called Flinz, from the soft layers in between, called Fäule. Most of the great fossils were found by these workers, and the quarries still supply museums such as the Jura-Museum in Eichstätt.

The fauna is astonishingly varied. Pterodactylus, described by Cosimo Collini in 1784 and named by Georges Cuvier, was the first pterosaur known to science; long-tailed Rhamphorhynchus is common. Among dinosaurs, the chicken-sized Compsognathus, named in 1859, has the bones of a small lizard preserved in its gut, and Juravenator (2006) and Sciurumimus (2012) preserve traces of skin and filaments. Then, of course, there is Archaeopteryx, the feathered dinosaur that made Solnhofen famous. Fish, ammonites, shrimps and the floating crinoid Saccocoma are found by the thousand.

Did you know? One Solnhofen slab shows a horseshoe crab's trackway almost 10 m long ending at the crab itself, a "death march" across the poisoned lagoon floor described by Dean Lomax and Christopher Racay in 2012.

Such finds work like a series of photographs. Each slab records a single moment, and careful study of how bodies lie, which parts are missing and which way they point tells researchers about currents, salinity and how long carcasses floated before sinking. The lagoon is a lesson in taphonomy, the science of how living things become fossils.

Solnhofen opened a window onto the soft tissues and small animals that ordinary rocks destroy. It gave the world its first pterosaur and its first feathered dinosaur, and it remains the model for every exceptional fossil site found since, from China's Liaoning to Brazil's Crato. A poisoned lagoon became one of the best-lit rooms in the history of life.

Archaeopteryx

In 1861, just two years after Charles Darwin published On the Origin of Species, quarry workers near Langenaltheim in Bavaria split open a slab of Solnhofen limestone and found a skeleton with feathered wings and a long bony tail. Named Archaeopteryx, "ancient wing", it was the transitional fossil that Darwin's critics said did not exist, and it has stayed at the centre of the story of birds ever since.

Archaeopteryx was about the size of a raven, roughly 50 cm from snout to tail tip. It lived around 150 mya on the islands of the Solnhofen archipelago. That first skeleton was bought by the British Museum in 1862 and described by Richard Owen, and it is now in London's Natural History Museum. The finest specimen, found in the 1870s and now in Berlin, preserves the head, spread wings and a fan of tail feathers. Roughly a dozen skeletons are now known, including the Thermopolis specimen in Wyoming, described in 2005 by Gerald Mayr and colleagues.

What makes it extraordinary is the mix. Like other theropod dinosaurs, it had teeth in its jaws, three clawed fingers on each hand and a long tail of more than 20 vertebrae. Like birds, it had broad feathered wings with asymmetric flight feathers, the vanes narrower on the leading edge, an arrangement that works in airflow. Some features it shared with both: a wishbone and feathers are now known in many dinosaurs that never flew. In the 1860s Thomas Henry Huxley argued that birds were descended from dinosaurs. Feathered fossils from China since 1996 have confirmed it beyond reasonable doubt.

New tools keep extracting more. In 2010 X-rays from a synchrotron picked up chemical traces of the original feathers. A 2012 study by Ryan Carney and colleagues found pigment-bearing melanosomes in an isolated Solnhofen feather suggesting it was matte black, though whether that feather belongs to Archaeopteryx is disputed. In 2018 Dennis Voeten and colleagues scanned its wing bones and found walls built like those of birds that fly in short bursts, such as pheasants, suggesting it could flap, not only glide.

Did you know? One specimen sat in a Dutch museum for over a century labelled as a pterosaur, until John Ostrom recognised feather impressions on it in 1970; in 2017 it was moved to a genus of its own, Ostromia.

Debate continues over exactly where it sits. Older feathered dinosaurs such as Anchiornis, from China around 160 mya, blur the line, and a few analyses have placed Archaeopteryx among the raptor-like dinosaurs rather than on the bird branch. Most still treat it as an early bird, one of many experiments in feathered flight.

Archaeopteryx turned evolution from argument into evidence. A single slab of limestone showed a reptile and a bird in one animal, and later finds proved that the line between them was never sharp. Every sparrow on a garden feeder is its living relative.

The Trackways of Skye

At low tide on the north-east coast of the Isle of Skye, flat ledges of rock emerge from the sea, and on them are footprints. Some are three-toed and sharp, left by meat-eating theropods. Others are broad, rounded hollows big enough to sit in, left by sauropods. They were pressed into the mud of coastal lagoons around 170 to 165 million years ago, in the Middle Jurassic, a slice of time from which dinosaur fossils of any kind are rare worldwide.

The rocks belong to the Great Estuarine Group, a pile of sandstones, shales and limestones laid down in a shallow basin of lagoons, deltas and brackish bays in a sinking rift basin that lay between what is now Scotland's mainland and the Outer Hebrides. Fossil shells, ostracods and algal mats in the same beds show water that swung between fresh and salty, and mudcracks show surfaces that dried out between floods.

Several sites have reshaped the picture. In 2015 Steve Brusatte of the University of Edinburgh and colleagues described sauropod trackways at Duntulm, in rock laid down in a lagoon about 170 mya. The prints, up to around 70 cm across, show long-necked giants wading through shallow water, overturning an old idea that sauropods were confined to dry land. In 2020 a team led by Paige dePolo described dozens of footprints at Brothers' Point, Rubha nam Brathairean, and at another site nearby, including theropod prints that point to a greater variety of dinosaurs on Skye than its scarce bones suggest. At An Corran beach near Staffin, three-toed prints appear in a bed revealed only at low tide.

The methods are part of the story. The ledges are exposed for a few hours a day and pounded by waves, so teams work fast. They photograph each surface from many angles, often by drone, then use photogrammetry to build 3D models accurate to millimetres. From those models they measure print length, stride and the angle of each step, which give estimates of hip height, body size and walking speed.

Did you know? The same Skye rocks have yielded some of the most complete Middle Jurassic mammal skeletons anywhere, including the shrew-sized Borealestes, described in 2021.

Dinosaur bones from the same rocks are scarce: a few sauropod and theropod teeth and limb bones, alongside small reptiles, salamanders and early mammals from the Kilmaluag Formation. The trackways therefore carry much of the evidence. They show who was present, how they moved and which habitats they used, at a moment when the major dinosaur groups were diversifying.

Skye's trackways have turned a remote island into one of the key windows on the Middle Jurassic, and the local Staffin Museum displays casts for visitors. Every winter storm strips away more rock and uncovers new prints, while erasing older ones. The record is renewed and destroyed at the same time, which is why someone always needs to be out on the ledges at low tide.

What Makes a Theropod?

In 1881 Othniel Charles Marsh gave the meat-eating dinosaurs he was digging out of the American West a name: Theropoda, or "beast feet". The group first appears around 231 mya in the Ischigualasto Formation of north-western Argentina, with small, quick hunters such as Eodromaeus, and it has never died out. Every bird alive today, from the 2 g bee hummingbird of Cuba to the ostrich, is a theropod. So were Compsognathus, barely 1 m long, and Spinosaurus, at around 14 m the longest predator ever to walk on land.

Theropods walked on two legs, and their feet carry the family signature. Three long toes faced forward and took the weight, while the first toe was a small dewclaw held off the ground. That is why a three-toed trackway, such as the thousands of prints called Grallator in the Connecticut Valley, is usually read as a theropod's walk. The hands tell a story of loss: early forms kept four or five fingers, most later ones grasped with three, and tyrannosaurs managed with two.

Inside, the skeleton was built light. Many vertebrae and limb bones were hollow and thin-walled, invaded by air sacs from the lungs through small openings called pneumatic foramina. CT scans of skulls and backbones map these chambers in detail, and they match the flow-through breathing system of living birds. Theropods also had a furcula, or wishbone, formed from fused collarbones; it has been found in Coelophysis, Allosaurus and Tyrannosaurus. Most had blade-like, backward-curving teeth with serrated edges, much like steak knives, and many had a flexible joint halfway along the lower jaw.

How do palaeontologists decide who belongs? In 1986 Jacques Gauthier published a cladistic analysis that sorted animals by shared, inherited features rather than overall resemblance, and it placed birds firmly inside the theropods. Today's studies score hundreds of such characters across hundreds of species in computer programs, and the result has held up as new fossils pour in from China, Mongolia and South America.

Did you know? A 198-million-year-old trace at St George, Utah, shows where a theropod crouched in the mud with its palms facing inwards, as a bird folds its wings, not downwards like a lizard's.

One popular idea needs correcting: "theropod" does not mean "carnivore". Ornithomimids such as Gallimimus had toothless beaks and probably ate plants as well as small prey, oviraptorosaurs cracked seeds and shells, and therizinosaurs turned almost wholly to leaves. The theropod body plan was a toolkit, and evolution used it for hunting, browsing and, in the end, flying.

Seeing birds as theropods changed the questions we ask. Breathing, brooding, growth rates and even behaviour can now be tested against living relatives that share the same skeleton. Look closely at a pigeon's foot and you are looking at the family badge.

Tyrannosaurus rex

For the last two million years of the Cretaceous, the largest predator in western North America was Tyrannosaurus rex. It lived from about 68 to 66 mya on Laramidia, the island continent west of a great inland sea, and its bones turn up in the Hell Creek Formation of Montana and the Dakotas, the Lance Formation of Wyoming and the Scollard and Frenchman formations of Alberta and Saskatchewan. Barnum Brown found the first good skeleton at Hell Creek in 1902, and Henry Fairfield Osborn named it in 1905: "tyrant lizard king".

The largest adults reached about 12 m long, with hips around 3.7 m off the ground, and most mass estimates land between 6 and 9 tonnes. The skull was up to 1.5 m long and unusually broad at the back, so both eyes could look forward: Kent Stevens's 2006 study put its overlapping field of view at about 55 degrees, wider than a hawk's. The idea that it could see you only if you moved is pure film. Casts of its braincase show enormous olfactory bulbs, so smell was keen too. Its arms, about 1 m long with two fingers, look comic but carried heavy muscles.

Growth rings in bone make T. rex one of the best-understood dinosaurs as an animal. In 2004 Gregory Erickson and colleagues sectioned ribs and other bones from several individuals and found that a young tyrannosaur grew slowly, then sprinted through a teenage growth spurt, adding around 2 kg a day before reaching full size near 20. Few lived past 30. Medullary bone, the calcium store laid down by egg-laying birds, has been reported in one specimen and used to identify it as female.

Did you know? Skin impressions published in 2017 from the neck, hips and tail of large tyrannosaurs show small, pebbly scales, so a grown T. rex was probably mostly scaly even though some of its smaller relatives were feathered.

That last point is where science and popular images part company. Feathered tyrannosauroids such as Yutyrannus show the family had feathers, yet adult T. rex skin looks scaly; a light covering on the back or in chicks remains possible. Other arguments are live too. In 2022 Gregory Paul and colleagues proposed splitting the species into three; a reply led by Thomas Carr found the variation too small to support it, and most researchers agree. In 2024 skull bones from New Mexico was named Tyrannosaurus mcraeensis, a possible older relative, though its age and status are still debated.

With more than 30 partial skeletons known, T. rex has become the standard against which other predators are measured. It is a reminder that a famous animal can still be studied afresh: growth, senses and skin were all rewritten in the last 25 years.

Sue, Stan and the Famous Rexes

Most dinosaurs are known from one skeleton or a handful of bones. Tyrannosaurus rex is different: more than 30 partial skeletons have been found, and the best of them have names, catalogue numbers and life stories of their own. Together they let palaeontologists compare individuals the way a doctor compares patients.

The story begins with Barnum Brown of the American Museum of Natural History, who found the original specimen in Montana's Hell Creek badlands in 1902. That type specimen, now CM 9380, was sold to the Carnegie Museum in Pittsburgh in 1941 to keep it safe during the Second World War. Brown's second find, AMNH 5027 from Big Dry Creek in 1908, gave New York the skull and the upright, tail-dragging mount that defined the dinosaur for most of the 20th century; it was remounted in the 1990s in today's horizontal pose.

The most famous of all is Sue, FMNH PR 2081. Sue Hendrickson spotted bones in a cliff near Faith, South Dakota, on 12 August 1990 while working with the Black Hills Institute. A long legal fight over who owned the fossil ended with the land's owner, Maurice Williams, and in 2000 Sue went on show at the Field Museum in Chicago. About 90 per cent complete by volume and some 12.3 m long, the skeleton is from an animal that died at about 28 and carries healed broken ribs, an injured shoulder and holes in the jaw that may record an infection. Nobody knows whether Sue was male or female.

Stan, BHI 3033, was found by amateur collector Stan Sacrison in South Dakota in 1987 and dug out by the Black Hills Institute in 1992. Casts of its superb skull stand in museums worldwide, which made Stan one of the most studied tyrannosaurs even before the original was bought in 2020 for the Natural History Museum Abu Dhabi. In Saskatchewan, Scotty (RSM P2523.8), found in 1991, was estimated in 2019 at about 8.8 tonnes, the heaviest T. rex known.

Did you know? The Wankel rex, found by Kathy Wankel near Montana's Fort Peck Reservoir in 1988, provided the first complete T. rex arm, and is now on a 50-year loan to the Smithsonian as the Nation's T. rex.

Other individuals matter for what was inside them. In 2005 Mary Schweitzer reported flexible, vessel-like structures in the thigh bone of MOR 1125, nicknamed B-rex, and identified medullary bone, a calcium store for eggshell, that marked it as female. Juveniles such as Jane at the Burpee Museum in Illinois fuel a long debate over whether small skulls named Nanotyrannus are young T. rex or a separate animal; a 2025 study of a juvenile preserved alongside a Triceratops argued for a separate species, and the argument continues.

Named skeletons turned a species into a population, with growth, injury and variation that a single bone could never show. Each new rex adds another patient to the file.

Spinosaurus: The River Monster

Few dinosaurs have been rebuilt as often as Spinosaurus aegyptiacus. It lived about 100 to 94 mya along the coasts and river systems of North Africa, a world of mangroves, sawfish and coelacanths, and at around 14 m it was probably the longest meat-eating dinosaur known. Yet for most of the 20th century almost nobody had seen a bone of it.

The first skeleton was dug up in 1912 at Bahariya Oasis in Egypt by Richard Markgraf, a collector working for the German palaeontologist Ernst Stromer. Stromer named it in 1915 and described a crocodile-like lower jaw and backbones with spines up to 1.65 m tall, which must have carried a sail along the back. The bones went to the Bavarian State Collection in Munich, and on the night of 24 April 1944 an Allied bombing raid destroyed them. Only Stromer's drawings and photographs survived.

The modern picture comes from the Kem Kem beds of south-eastern Morocco. In 2014 Nizar Ibrahim, Paul Sereno and colleagues described a partial skeleton bought from a local fossil hunter and traced back to its site, with short hind legs, dense limb bones and long nostrils set far back on the snout. In 2020 Ibrahim's team returned and excavated the tail, whose tall, thin spines formed a fin-like paddle. They proposed an animal that swam after prey, the first truly aquatic dinosaur.

Did you know? Spinosaurid teeth are so common in the Kem Kem beds that local miners recover thousands of them, while bones of the animal itself remain rare.

Here the scientists genuinely disagree. In 2022 Matteo Fabbri and colleagues compared bone density across hundreds of living and extinct animals and argued that the thick, dense bones of Spinosaurus point to diving for food. Later that year Sereno, with Nathan Myhrvold and others, modelled the body in water and concluded it would have been buoyant, slow and unstable, more like a giant heron wading in shallows than an otter. David Hone and Thomas Holtz also argued for a shoreline wader. The sail itself may have been mostly for display, since a tall, thin fin is a poor engine.

What everyone accepts is that Spinosaurus fed heavily on fish. Its conical, unserrated teeth suited gripping slippery prey, and chemical studies of spinosaurid teeth by Romain Amiot and colleagues in 2010 found oxygen isotope ratios close to those of crocodiles and turtles, implying a semi-aquatic life. The Kem Kem rivers held fish several metres long, including the sawfish Onchopristis and the coelacanth Mawsonia.

Spinosaurus shows how a dinosaur can be rebuilt from scratch when bones are lost and new ones found, and how much of a reconstruction is argument. Whether it swam or waded, it stretched the idea of what a predatory dinosaur could be.

Giganotosaurus and Carcharodontosaurus

Long before T. rex, the southern continents had giant predators of their own. The carcharodontosaurids, or "shark-toothed lizards", were cousins of Allosaurus that grew as long as any tyrannosaur, and two of them, one from Argentina and one from North Africa, became famous in the 1990s as rivals for the title of largest meat-eater.

Giganotosaurus carolinii was found in 1993 by Rubén Carolini, a car mechanic and amateur fossil hunter, in the badlands near Villa El Chocón in Neuquén Province, Argentina. Rodolfo Coria and Leonardo Salgado named it in Nature in 1995. The skeleton, about 70 per cent complete, comes from the Candeleros Formation, around 99 to 97 mya. Estimates put it at 12 to 13 m long and roughly 6 to 8 tonnes, and a second, larger lower jaw suggests some individuals were bigger still. Its skull was long and narrow, about 1.6 m, with thin blade-like teeth.

Carcharodontosaurus saharicus has a more tangled history. Teeth from Algeria were described by Charles Depéret and Justin Savornin in 1925. Ernst Stromer named the genus in 1931 from a partial skeleton found in Egypt, and like his Spinosaurus it was destroyed in Munich in 1944. In 1995 Paul Sereno's team found a skull about 1.6 m long in the Kem Kem beds of Morocco, giving the animal a face again; a second species, C. iguidensis, was named from Niger in 2007 by Stephen Brusatte and Sereno.

The two hunted differently from T. rex. Their teeth were flat and finely serrated, built to slice flesh rather than crush bone, and their skulls were lighter and narrower. Their prey included sauropods: Giganotosaurus shared its world with the titanosaur Andesaurus, and its close relative Mapusaurus, from a bonebed at Cañadón del Gato that held at least seven individuals of different ages, lived alongside the colossal Argentinosaurus. Whether those animals hunted in groups or simply died together is still debated.

Did you know? Meraxes, a carcharodontosaurid described from Argentina in 2022, had tiny arms like a tyrannosaur's, so short arms evolved more than once in giant predators.

A popular myth needs correcting. Giganotosaurus and T. rex never met: they lived some 30 million years and a continent apart. And "biggest" is a slippery prize. Size estimates rest on a few incomplete skeletons, and Giganotosaurus, Carcharodontosaurus, T. rex and Spinosaurus overlap within the margins of error.

The carcharodontosaurids showed that gigantism in predators evolved several times, in different families and on different continents, wherever there were giant plant-eaters to eat. Big prey made big hunters.

Baryonyx: The Fish-Eater from Surrey

In January 1983 William Walker, a plumber and amateur fossil hunter, was searching the Smokejacks clay pit near Ockley in Surrey when he split open a nodule and found a huge, curved claw. It measured about 31 cm along its outer curve. Walker took it to the Natural History Museum in London, and within months a museum team was back at the brickworks, digging out the bones of a meat-eating dinosaur from the grey Weald Clay, laid down in river floodplains around 125 mya.

The excavation, led by Alan Charig and Angela Milner, recovered much of a skeleton from hard clay and ironstone, encased in plaster jackets. Freeing it took years of careful preparation. In 1986 Charig and Milner named it Baryonyx walkeri, "heavy claw", in honour of its finder. The specimen, NHMUK PV R 9951, belonged to an animal perhaps 7.5 m long and around 1.5 tonnes, and some unfused bones suggest it was not yet fully grown.

It was not built like other big predators. The snout was long, low and narrow, ending in a spoon-shaped rosette of teeth, much like a crocodile's. The lower jaw held 32 teeth on each side, about twice the usual number for a theropod, and they were conical with only very fine serrations, better for gripping than slicing. The forelimbs were powerful, and the great claw probably sat on the thumb, suited to hooking prey from water.

Did you know? Baryonyx was the first dinosaur shown by direct evidence, acid-etched fish scales in its stomach, to have eaten fish.

That evidence came from the body cavity: scales and teeth of Lepidotes, a large bony fish, partly dissolved by stomach acid, together with bones of a young iguanodontian. So Baryonyx fished, but it also took or scavenged dinosaurs. Later chemical studies of spinosaurid teeth agreed that the group spent much of its time around water.

Baryonyx also solved an old puzzle. Its skull matched the fragmentary jaws Ernst Stromer had described for Spinosaurus in Egypt, and it was soon placed in the same family, the spinosaurids. Relatives followed: Suchomimus from Niger in 1998, and in 2021 two new spinosaurids from the Isle of Wight, Ceratosuchops and Riparovenator, the second named in honour of Angela Milner. Isolated Baryonyx-like teeth and bones have also been reported from Spain and Portugal, showing the family was spread across Early Cretaceous Europe.

A single morning's find by an amateur gave Britain one of its most important dinosaurs and gave the world the key to a whole family of fish-eating predators. Many great discoveries still begin with a curious person and a hammer.

Carnotaurus and the Abelisaurs

While tyrannosaurs ruled the north at the end of the Cretaceous, the top predators of South America, Africa, Madagascar and India belonged to a different family: the abelisaurids. Its most striking member is Carnotaurus sastrei, the "meat-eating bull", with two thick horns above its eyes, a short, deep skull and arms so small they make T. rex look well equipped.

Carnotaurus was found in 1984 by an expedition led by the Argentine palaeontologist José Bonaparte, in the La Colonia Formation of Chubut Province, Patagonia, on a farm belonging to the Sastre family. Bonaparte named it in 1985. The single skeleton, MACN-CH 894, is from rocks about 72 to 69 million years old and belonged to an animal around 7.5 to 8 m long, weighing roughly 1.5 to 2 tonnes. Its forearm bones were only about a quarter the length of the upper arm, and the fingers were stubby nubs.

It is also one of the very few big theropods with extensive skin. Impressions from the side of the body show a mosaic of small scales with rows of larger, cone-shaped scales, and no sign of feathers. The horns may have served in display or head-to-head shoving; nobody knows for certain. In 2011 Scott Persons and Philip Currie reconstructed its huge tail muscles and argued that Carnotaurus was among the fastest large theropods in a straight line, though probably poor at turning.

Did you know? Carnotaurus is known from just one skeleton, yet it preserves more skin impressions than almost any other large meat-eating dinosaur.

Its relatives fill the southern continents. Abelisaurus, which gave the family its name in 1985, came from Río Negro Province. Aucasaurus was found in 1999 at Auca Mahuevo, the Patagonian sauropod nesting ground. Rajasaurus, described in 2003 by Jeffrey Wilson, Paul Sereno and Indian colleagues, comes from the Narmada valley, and Majungasaurus from Madagascar. In 2003 Raymond Rogers and colleagues showed that Majungasaurus bones carry tooth marks made by Majungasaurus, the first strong evidence of cannibalism in a dinosaur. All share short, rough-textured skulls, often with bumps or horns, and reduced arms.

The wider group, the Ceratosauria, includes the Jurassic Ceratosaurus of the Morrison Formation and the odd little Limusaurus from China, whose young had teeth and whose adults lost them for a beak. This spread fits the break-up of Gondwana: abelisaurids diversified as the southern continents drifted apart, while tyrannosaurs held Asia and North America.

The abelisaurs show that there was no single "age of tyrannosaurs". Different continents ran different experiments in being a large predator, and the southern answer was short-faced, horned and almost armless.

Therizinosaurus: The Scythe Lizard

In 1948 a Soviet–Mongolian expedition working in the Nemegt Formation of the Gobi Desert collected some enormous, flattened claws and fragments of ribs. They were around 70 million years old and unlike anything known. In 1954 the Soviet palaeontologist Evgeny Maleev named their owner Therizinosaurus cheloniformis, the "turtle-shaped scythe lizard", because he thought it was a giant turtle-like reptile.

The claws are still astonishing. The bony cores of the largest are more than 50 cm long along the curve, and in life a horny sheath would have made them longer still, the longest claws of any known animal. More of the skeleton turned up slowly: forelimb bones in the 1970s showed arms about 2.5 m long, and hind-limb bones described in 1982 revealed broad feet with four weight-bearing toes. Even so, the animal remained a mystery wrapped around three claws.

The puzzle grew when related animals appeared. In 1979 the Mongolian palaeontologist Altangerel Perle named Segnosaurus, with a toothless beak, leaf-shaped cheek teeth and a hip whose pubic bone pointed backwards, like those of bird-hipped dinosaurs. For years these "segnosaurs" were argued to be late-surviving prosauropods, ornithischian cousins or something in between. Then in 1993 Dale Russell and Dong Zhiming described the more complete Alxasaurus from Inner Mongolia, which showed beyond reasonable doubt that segnosaurs and Therizinosaurus were all theropods: plant-eating relatives of the dinosaurs that led to birds.

Did you know? Beipiaosaurus, a therizinosaur from China's Yixian Formation named by Xu Xing and colleagues in 1999, was covered in simple, filament-like feathers, so the giant Gobi scythe lizard probably had a fuzzy coat too.

Put the clues together and Therizinosaurus becomes a pot-bellied browser, perhaps 9 to 10 m long and around 5 tonnes, with a small head, long neck and a wide gut for fermenting leaves. In 2014 Stephan Lautenschlager tested therizinosaur claws with computer models and found they worked best for hooking and pulling branches, not digging or fighting, though a claw that size would also have discouraged a predator. Nesting sites from the Javkhlant Formation of Mongolia, published in 2019 by Kohei Tanaka and colleagues, show related therizinosaurs laying more than a dozen clutches together in colonies.

Therizinosaurus is a lesson in how easily a single part can mislead. A claw looked like a weapon, a weapon looked like a predator and a predator looked like a turtle, until new skeletons showed a feathered theropod that ate plants. It proved that the meat-eaters' family tree held vegetarians too.

How Fast Could T. rex Run?

In the film Jurassic Park, a Tyrannosaurus keeps pace with a speeding jeep. It is one of cinema's great chases and, according to nearly every study since, physically impossible. Working out how fast an extinct 8-tonne animal moved is a problem in engineering, and palaeontologists have attacked it from several directions.

The first serious test came in 2002, when John Hutchinson and Mariano Garcia published a study in Nature asking how much leg muscle a T. rex would need to run fast. Using the physics of living runners such as chickens and humans, they found that a sprint at 72 km/h would require leg muscles making up an impossibly large share of its body. Later work by Hutchinson and colleagues put plausible top speeds somewhere between about 18 and 40 km/h, and even the upper end was uncertain.

In 2017 William Sellers and colleagues built a computer model of the skeleton and muscles and used optimisation software to find its fastest gait. When they added the stresses on the bones, the answer narrowed: above about 20 km/h, the leg bones would have broken. A grown T. rex could walk briskly, and perhaps manage a fast stride, but it probably never had both feet off the ground at once.

Did you know? An African elephant, the closest living animal to an adult T. rex in weight, reaches about 25 km/h but never gets all four feet off the ground at the same time.

Walking is easier to pin down. In 2021 Pasha van Bijlert and colleagues modelled the tail as a hanging, swinging beam and calculated that the natural rhythm of a T. rex stride gave a comfortable walking pace of about 4.6 km/h, close to a human's. Real footprints agree: in 2014 Richard McCrea and colleagues described three tyrannosaurid trackways side by side near Tumbler Ridge in British Columbia, made by animals walking at roughly 6 to 8 km/h. In 2016 Scott Persons and Philip Currie found that tyrannosaurs' long legs were better at saving energy over long distances than at sprinting.

Size changes everything. A juvenile T. rex was slimmer, with relatively longer shins, and was probably much quicker than its parents, so young and old may have hunted different prey in different ways. Scientists still disagree on the exact numbers for adults, since each model rests on estimates of muscle mass and posture, but the range has shrunk to "slow jog at best". Even so, only a sprinter in the class of Usain Bolt, peaking at about 44.7 km/h, could have been sure of outrunning an adult over a short dash.

The speed question changed how dinosaurs are studied, bringing robotics, biomechanics and computer simulation into palaeontology. The film's jeep would have escaped easily; its passengers might not have.

Bite Force and the Hunt

Tyrannosaurus rex had the most powerful bite of any land animal known. Measuring it is impossible directly, so palaeontologists reconstruct it from the skull, the muscles and the damage those jaws left behind. The evidence also settles a long-running question: did T. rex hunt, or did it only scavenge?

The first estimate came from bitten bone. In 1996 Gregory Erickson and Kenneth Olson described a Triceratops pelvis from Montana scored with nearly 60 bite marks, some punctures deep enough to show a tooth had been driven in and then dragged. Pressing a cast T. rex tooth into cow bone in the laboratory, they estimated more than 6,000 N for a single tooth. Later teams rebuilt the jaw muscles in computer models. In 2012 Karl Bates and Peter Falkingham calculated a bite of between about 35,000 and 57,000 N at the back teeth, and in 2017 Paul Gignac and Erickson estimated about 35,000 N, with pressures at the tooth tips high enough to shatter bone. For comparison, Erickson measured a saltwater crocodile, the hardest-biting living animal, at about 16,400 N.

The skull was built for this. It was broad and reinforced at the back, where the jaw muscles attached, and the teeth were thick, rounded blades up to about 30 cm long including the root, closer to railway spikes than steak knives. Erickson and Olson called the feeding style "puncture and pull": bite deep, then wrench back, cracking bone to reach the marrow inside.

Did you know? A T. rex coprolite, or fossil dropping, found in Saskatchewan and described by Karen Chin in 1998 is 44 cm long and packed with crushed bone from a young plant-eating dinosaur.

Did it hunt living animals? For years Jack Horner argued that T. rex was mainly a scavenger, pointing to its small arms and large smell centres. The best answer lies in healed wounds. In 2013 Robert DePalma and colleagues described Edmontosaurus tail vertebrae from the Hell Creek Formation of South Dakota with a T. rex tooth crown embedded in them, surrounded by new bone. The duck-billed dinosaur had been bitten and escaped, and it lived long enough to heal. A Triceratops skull with a healed bite on its horn and face tells the same story.

The consensus is that T. rex behaved like a hyena or lion: it hunted when it could and scavenged when it found a carcass. Adults were too slow to chase down prey over distance, so ambush and short rushes were likely, perhaps with juveniles doing more of the running.

The bite studies show how far evidence can reach in the absence of a living animal: tooth marks, healed scars and dung became measurements of behaviour. They proved the tyrant king was a hunter, and that it ate the bones too.

Ankylosaurus and the Tail Club

Ankylosaurus magniventris lived at the very end of the Cretaceous, between about 68 and 66 mya, on the same floodplains of western North America as Tyrannosaurus and Triceratops. It was a low, broad plant-eater some 6 to 8 m long and probably 5 to 8 tonnes, built like an armoured coffee table. Its back was paved with bony plates called osteoderms, its skull was capped with thick, fused armour, and its tail ended in a club of solid bone.

The animal is known from surprisingly little. Barnum Brown of the American Museum of Natural History named it in 1908 from a skull roof and bits of skeleton found in the Hell Creek Formation of Montana. In 1910 Brown collected a second specimen, with the tail club, along the Red Deer River in Alberta, and the largest skull, found by Charles M. Sternberg in 1947, is about 74 cm wide. No complete skeleton exists, so much of its shape is reconstructed from better-known Asian and Canadian cousins such as Pinacosaurus, Euoplocephalus and Zuul.

The club itself is a masterpiece of engineering. Two large osteoderms wrap the last tail vertebrae to form the knob, while the vertebrae in front interlock in long, overlapping spurs and are bound by ossified tendons, turning the rear of the tail into a stiff handle. The flexible base of the tail did the swinging. In 2009 Victoria Arbour modelled clubs of different sizes and concluded that the biggest could deliver blows strong enough to break bone, though the smallest were probably too light to do serious harm.

Who was it hitting? The obvious answer is a predator's legs, and an ankylosaurid swinging at shin height would have been a nasty prospect for a tyrannosaur. But the best evidence points at its own kind. Zuul crurivastator, a 76-million-year-old ankylosaurid from the Judith River Formation of Montana, preserves its skin and armour in place, and a 2022 study found broken and healed spikes clustered along its flanks, exactly where a rival's club would land. Clubs, like antlers, may have mattered most in contests over mates and territory.

Did you know? Zuul crurivastator is named after the demon dog in the 1984 film Ghostbusters; its species name means "destroyer of shins".

Armour in ankylosaurs went further than most visitors realise. Even the eyelids carried a small bony plate, and CT scans of ankylosaurid skulls reveal looping nasal passages that may have helped warm the air, save water or amplify calls. The heavy armour also left them heavy to move, and their wide bodies and hoof-like toes suggest a slow, steady pace.

Ankylosaurus shows that the last dinosaurs were not passive victims waiting for a tyrannosaur. Plant-eaters evolved weapons as sophisticated as any predator's teeth, and the tail club remains the most specialised bludgeon ever grown by a land animal.

Borealopelta: The Dinosaur Mummy

On 21 March 2011, a heavy-equipment operator named Shawn Funk was digging through the oil sands of Suncor's Millennium Mine near Fort McMurray, Alberta, when his bucket struck something harder than the surrounding rock. It was the front half of an armoured dinosaur, preserved in three dimensions with its skin and armour still in place. Described in 2017 by Caleb Brown and colleagues at the Royal Tyrrell Museum, it was named Borealopelta markmitchelli: the "northern shield", honouring the technician who spent more than five years and around 7,000 hours freeing it from the rock.

Borealopelta was a nodosaur, a club-less cousin of Ankylosaurus, about 5.5 m long and roughly 1.3 tonnes. It lived around 112 to 110 mya, in the Early Cretaceous, and yet it was found in marine rock of the Clearwater Formation, laid down in a sea that then covered much of Alberta. The likeliest story is "bloat and float": the animal died on land or in a river, swelled with gas, drifted out to sea, and finally sank. It landed on its back, pressing into soft seafloor mud, and minerals quickly sealed it in a concretion before scavengers or decay could take it apart.

The result is the best-preserved armoured dinosaur ever found. The osteoderms still wear their keratin sheaths, the horny covers that made the spines longer and sharper than the bone alone, and the skin between them is preserved as a thin dark film. The longest spines jut sideways from the shoulders like a pair of bull's horns.

Chemistry did the rest. Mass spectrometry found traces of benzothiazole, a breakdown product of the reddish pigment pheomelanin, and the pigment was concentrated on the back and sparse on the belly. That pattern, called countershading, is how many modern animals hide from predators, so even a tonne-and-a-half tank was apparently worth camouflaging. Tyrannosaur-sized hunters such as Acrocanthosaurus roamed North America at the time.

Did you know? Borealopelta's last meal survives in its gut: mostly fern leaves, with about 6 per cent charcoal, suggesting it browsed regrowth after a forest fire in late spring or summer.

That stomach mass, studied in 2020, is the most detailed dinosaur meal yet described. It shows a picky eater that chose soft fern fronds over the abundant cycads and conifers, and growth rings in tiny twigs date its death to the growing season.

Borealopelta changed what "fossil" can mean. Instead of a skeleton to be clothed by an artist, it is an animal with its own surface, colour and dinner, and it shows that even heavy armour did not make a dinosaur safe.

Triceratops

Triceratops is the three-horned face of the end of the Cretaceous. It lived between about 68 and 66 mya on the coastal plains of Laramidia, the western landmass of North America, and it is the most common large dinosaur in the Hell Creek Formation of Montana and the Dakotas and the Lance Formation of Wyoming. Adults reached 8 to 9 m long and perhaps 6 to 12 tonnes, and the largest skulls approach 2.5 m, among the biggest heads of any land animal.

The first find was misread. In 1887 a pair of horn cores turned up near Denver, and Othniel Charles Marsh named them Bison alticornis, a giant Ice Age bison. Within two years John Bell Hatcher, collecting for Marsh in Wyoming, sent back complete skulls, and in 1889 Marsh named Triceratops horridus, the "rough three-horned face". Hatcher went on to collect around thirty skulls, and the specimens still crowd the museums of America.

That abundance has made Triceratops a case study in how dinosaurs grew. In juveniles the brow horns are short and curl backwards, the frill is edged with sharp triangular bones and the snout is stubby. As the animal matured the horns straightened and swung forwards and the frill edge flattened. Bone histology, the study of growth lines in thin slices of bone, matches the changes to age. In 2010 John Scannella and Jack Horner argued that the even larger, holed frill of Torosaurus was simply the oldest stage of Triceratops; other studies have found adult Torosaurus specimens that do not fit, and the question remains unsettled. Scannella also showed that T. horridus gives way to T. prorsus higher in the rock, so the genus was evolving in place.

The horns were used, though not only for show. Skulls studied by Andrew Farke and colleagues in 2009 carry healed wounds on the cheek and frill exactly where locked horns would gouge, a sign of fighting between rivals. Predators left their own record: a Triceratops pelvis from Montana bears dozens of Tyrannosaurus bite marks, and puncture marks on frills show that the neck shield was pulled at during feeding.

Did you know? A Triceratops jaw held dental batteries of hundreds of stacked teeth, with up to about 800 in the whole mouth, constantly replaced as they wore down.

Those batteries sliced tough plants like scissors, and a few specimens preserve skin with large, rounded scales. Triceratops survived to within the last few hundred thousand years of the Cretaceous; its bones lie close beneath the iridium-rich layer left by the Chicxulub impact. It was not a lone gladiator but the cattle of its world, grazing, growing, fighting and being eaten right up to the end.

Protoceratops and the Griffin Legend

Protoceratops andrewsi was a sheep-sized horned dinosaur without the horns. About 2 m long and probably 60 to 100 kg, it had a parrot-like beak, powerful jaws and a bony frill, and it lived between about 75 and 71 mya in the dune fields of what is now the Gobi Desert of Mongolia. It belongs to the same family tree as Triceratops, but it is smaller and older, and it walked on all fours across a dry land of sand seas and seasonal streams.

It was discovered in 1922 at the Flaming Cliffs, or Bayn Dzak, by the photographer J. B. Shackelford of the American Museum's Central Asiatic Expeditions, led by Roy Chapman Andrews. Walter Granger and William King Gregory named it in 1923, and the cliffs proved so rich that Protoceratops is now known from hundreds of skeletons, from hatchlings to adults. That makes it one of the best-sampled dinosaurs in the world, ideal for studying growth and variation.

The expedition also found eggs, which were promptly credited to Protoceratops. Seventy years later an embryo inside a matching egg turned out to be an oviraptorid, and the "egg thief" Oviraptor found lying on a nest was cleared of its crime. The real eggs of Protoceratops were only identified in 2020, when Mark Norell and colleagues showed that a clutch from Ukhaa Tolgod had soft, leathery shells, more like a turtle's than a bird's, which explains why so few were ever found.

Then there is the griffin. In 2000 the folklorist Adrienne Mayor proposed that Scythian gold prospectors crossing the Gobi and Altai saw beaked, four-legged skeletons weathering out of the rock, and that their stories reached Greek writers as the griffin, a beaked, four-legged beast that guarded gold. The idea is elegant and widely repeated. In 2016 palaeontologist Mark Witton and Richard Hing challenged it, pointing out that griffin imagery appears in Egypt and the Near East well before Greek contact with Scythians, that the main gold fields lie hundreds of kilometres from the Protoceratops beds, and that griffins have wings and no frill. Historians and palaeontologists still debate the link.

Did you know? A 2011 study described a nest of 15 young Protoceratops buried together in sand, perhaps siblings still sheltering near where they hatched.

Whatever its role in myth, Protoceratops changed how we picture early horned dinosaurs. Large samples showed that frills varied hugely between individuals as they grew, its soft eggs reshaped the story of dinosaur nesting, and its encounter with a famous predator is told in its own chapter. A small animal from a remote desert ended up rewriting both science and, perhaps, legend.

Styracosaurus and the Frill Puzzle

Styracosaurus albertensis, the "spiked lizard", had one of the most extravagant heads of any animal. A single nose horn up to about 60 cm long rose from its snout, and the back of its frill sprouted six long spines, the biggest around half a metre. The body behind was that of a typical horned dinosaur: about 5.5 m long and perhaps 3 tonnes. It lived around 75 mya in what is now Dinosaur Provincial Park, Alberta, in the swampy coastal lowlands recorded by the Dinosaur Park Formation.

Charles M. Sternberg found the first skull in 1913, and Lawrence Lambe of the Geological Survey of Canada named it that year. In 1915 Barnum Brown's team collected a nearly complete skeleton, and a further Alberta skull was named S. parksi before being recognised as the same species. Bonebeds in Alberta have since produced individuals of different ages, which turned out to be the key to its strangest feature.

What were the frills for? The first guess was defence: a bony shield to protect the neck from tyrannosaurs. But Styracosaurus's frill, like that of most centrosaurines, was pierced by two large openings, the parietal fenestrae, and in life was covered only with skin. It would have made a poor shield. A second idea was that it anchored huge jaw muscles; muscle scars show the jaw muscles stopped well short of the frill's edge. A third was temperature control, but frill shapes vary wildly between close relatives living in the same climate, which a radiator would not need.

That leaves signalling. Growth series show that the spines and hooks of centrosaurine frills appear late, as animals approached adulthood, the same timing as antlers in deer. In 2012 David Hone, Darren Naish and Innes Cuthill argued that such ornaments were shaped by mutual sexual selection, with both sexes advertising quality. A 2018 study by Andrew Knapp and colleagues tested the rival idea that frills helped species tell each other apart and found little support: ornaments did not diverge more where related species lived side by side. Most researchers now see the frill as a billboard.

Did you know? No two Styracosaurus frills are quite alike: spine length, curvature and number of small extra points vary from skull to skull, perhaps as distinctive as faces.

That conclusion matters beyond one animal. Horned dinosaurs evolved new frill shapes faster than almost any other part of their skeletons, and in Alberta and Montana species followed one another every few hundred thousand years, each with its own headgear. If the ornaments were social signals, then dinosaur evolution was driven not only by predators and plants but by the oldest pressure of all: being noticed.

The Fighting Dinosaurs

On 3 August 1971, a Polish–Mongolian expedition working at Tugrikin Shiree in the Gobi Desert found two skeletons that seemed to have stopped mid-fight. A Velociraptor mongoliensis lies on its side, its left hand clutching the head of a Protoceratops andrewsi and the great sickle claw of its right foot driven into the soft throat region below the plant-eater's frill. The Protoceratops, crouched over it, has its beak clamped on the predator's right forearm, which appears broken. Both died in that position, about 75 to 71 mya, in the dune fields of the Djadochta Formation.

The expeditions, run jointly by the Polish Academy of Sciences and the Mongolian Academy of Sciences, were led by Zofia Kielan-Jaworowska, one of the great field palaeontologists of the century. The block they lifted is now a national treasure of Mongolia, kept in Ulaanbaatar, and it remains one of the very few fossils to show two dinosaurs interacting at the moment of death.

What stopped the fight? Both animals were buried quickly and completely, in sand, while still locked together. The favoured explanation is that a rain-soaked dune collapsed on them, or that a violent sandstorm buried them alive; sedimentologists have found evidence of both kinds of event in the Djadochta rocks, and the exact mechanism is still debated. Some researchers note that parts of the Protoceratops's limbs are missing and suggest scavengers pulled them away before the sand closed in completely.

The pair also corrects a popular picture. The Velociraptor here is about 2 m long, much of that tail, and perhaps 15 to 20 kg: roughly the size of a large turkey, not the man-sized hunter of Jurassic Park, whose film monsters were modelled on the bigger American Deinonychus. In 2007 Alan Turner, Peter Makovicky and Mark Norell described quill knobs, the bumps where wing feathers anchor to bone, on the forearm of another Velociraptor, so the animal in this fight was almost certainly feathered.

Did you know? Another Velociraptor from the Gobi has a pterosaur bone in its gut, showing it would also eat, or scavenge, flying reptiles.

The Protoceratops was no easy target either. Of similar or greater weight than its attacker, with a sharp beak and powerful jaw muscles, it was capable of inflicting real damage, and in this case it seems to have done so. The fight reads less like a hunter overpowering a victim and more like a close, desperate struggle between near-equals.

Fossils usually record death, rarely the moments before it. The Fighting Dinosaurs show predation as it actually happened, in a specific place and posture, and they gave scientists a direct test of how a small feathered hunter used its claws. It is behaviour turned to stone.

Pachycephalosaurus

Pachycephalosaurus wyomingensis, the "thick-headed lizard", was a two-legged plant-eater of the very end of the Cretaceous, living around 70 to 66 mya in the Hell Creek and Lance formations of Montana, the Dakotas and Wyoming. It was about 4.5 m long and perhaps 400 kg, and its skull carried a smooth, rounded dome of bone up to 25 cm thick, ringed by knobs and short spikes at the back and on the snout. It shared its world with Triceratops and Tyrannosaurus.

Like many dome-heads, it is known chiefly from skulls. The thick domes survive rolling in rivers far better than the rest of the skeleton, so isolated domes are common while bodies are rare. Charles Gilmore first described material in 1931, and Barnum Brown and Erich Schlaikjer named the genus in 1943 from a much more complete skull found in Montana. The body is reconstructed largely from smaller relatives such as Stegoceras and Homalocephale, with stout hips, stiff tail and short arms.

The obvious reading of the dome is a battering ram, and for decades Pachycephalosaurus was drawn charging like a bighorn sheep. Then the doubts came. In 2004 Mark Goodwin and Jack Horner sliced domes and found spongy, fast-growing bone with radiating structure, and argued it was too fragile to butt and better for display. Others suggested flank-butting instead, with the dome swung into a rival's side. In 2011 Eric Snively and Jessica Theodor used CT scans and computer models to compare dome-heads with sheep and duikers and found the domes could absorb head-on impact well.

The clinching evidence came from injuries. In 2013 Joseph Peterson, Collin Dischler and Nicholas Longrich examined more than 100 domes and found healed lesions, the marks of bone infection after trauma, in about 22 per cent of them, concentrated on the top of the dome where blows would land. Flat-headed specimens showed none. Most researchers now accept that pachycephalosaurs did fight with their heads, though whether head-to-head or head-to-flank is still argued.

Did you know? Dracorex hogwartsia, a spiky, flat-headed relative named in 2006, honours Hogwarts School in the Harry Potter books; it may be a juvenile Pachycephalosaurus.

That suggestion, made by Horner and Goodwin in 2009, links Dracorex, the spiky Stygimoloch and Pachycephalosaurus as growth stages of one animal, with spikes shrinking as the dome swelled. Not everyone agrees, but it shows how growth can masquerade as diversity. The small ridged teeth suggest a diet of leaves, seeds and fruit, perhaps with insects.

Pachycephalosaurus turned a strange skull into a test of method. It took sectioned bone, scanners and a census of scars to show what one structure was for, and the answer was the simplest: they really did use their heads.

Psittacosaurus and the Colour of Camouflage

Psittacosaurus, the "parrot lizard", was one of the earliest and most successful horned dinosaurs, though it had no horns and no frill to speak of. About 1.5 to 2 m long and perhaps 20 kg, it had a deep, beaked skull with bony cheek flares, and it lived across Mongolia and China between about 125 and 100 mya. More species have been named in Psittacosaurus than in any other dinosaur genus, and hundreds of specimens are known, from hatchlings to adults.

Henry Fairfield Osborn named it in 1923 from finds by the Central Asiatic Expeditions in Mongolia. The modern revolution came from the Yixian Formation of Liaoning, north-east China, where lakes and volcanic ash preserved animals in astonishing detail. One specimen, now in the Senckenberg Museum in Frankfurt and described in 2002 by Gerald Mayr and colleagues, has a row of long, quill-like bristles standing up from its tail, up to about 16 cm long. A plant-eater far from the bird line, it still had hollow filaments in its skin.

The same specimen preserves its skin and pigment as dark films. In 2016 Jakob Vinther and colleagues mapped the distribution of melanin across the whole body. The back was darker and the belly lighter, with dark patches on the shoulders and face. This is countershading, and its exact gradient carries information. The team built a full-size model, painted it to match the fossil and photographed it under different kinds of light. The shading worked best under diffuse light, as found beneath a forest canopy, suggesting Psittacosaurus lived among trees rather than on open plains.

Other fossils fill out its life. Thin sections of limb bones, studied in 2013, show that hatchlings walked on all fours and switched to two legs as they grew. Groups of juveniles have been found buried together in volcanic debris, suggesting youngsters stayed in creches, and a 2021 study described the only well-preserved dinosaur cloaca, in the same Senckenberg animal.

Did you know? A badger-sized Chinese mammal, Repenomamus, was found with the bones of a juvenile Psittacosaurus in its stomach, proof that some Cretaceous mammals ate dinosaurs.

Psittacosaurus shows how far the study of colour has come. Pigment patterns, once a matter of artistic guesswork, can now be measured, mapped and tested against light and habitat. A dinosaur's camouflage can tell us where it lived, what hunted it and how it saw its world, all from a few dark stains on a slab.

Showing Off: Crests, Frills and Display

Walk through a dinosaur hall and you meet a parade of strange headgear: tubes, fans, horns, helmets, spines. Many of these structures were too weak to fight with and too costly to grow for nothing. Increasingly, palaeontologists read them as display structures, signals to rivals and mates, like a peacock's tail or a stag's antlers. Display is hard to prove in extinct animals, but the evidence has piled up from growth, sound, colour and comparison.

Growth is the first clue. In 1975 Peter Dodson showed that several "species" of crested duck-billed dinosaurs from Alberta were really one species at different ages, because the crests of Corythosaurus and Lambeosaurus appeared and enlarged only as animals neared adulthood. Ornaments that switch on at maturity are a classic sign of sexual signalling. The same timing turns up in the frills and horns of ceratopsians and the domes of pachycephalosaurs.

Sound is the second. The backswept crest of Parasaurolophus, up to about a metre long, holds hollow nasal passages that loop up to its tip and back. In the 1990s Robert Sullivan, Thomas Williamson and Carl Diegert used CT scans of a New Mexico skull to model the airways on computer and produced a low, resonant call, like a foghorn. The crest was a visual signal and a sound box at once.

Colour is the third. Since 2010, preserved melanosomes, the microscopic pigment packets in fossil feathers, have revealed real dinosaur colours. Anchiornis, from the Jurassic of China, had black-and-white spangled wings and a rufous crown; Sinosauropteryx had a ginger-banded tail and a dark mask; Microraptor was glossy black with a sheen like a starling's. Patterns like these were made to be seen. Fan-shaped tail feathers on oviraptorosaurs such as Caudipteryx suggest courtship dances that their bird descendants still perform.

Did you know? The plates of Stegosaurus were once thought to be radiators, but thin sections studied in 2005 suggest their blood vessels mainly fed growing bone, leaving display as their likeliest role.

There is honest disagreement. Sexes are rarely distinguishable in dinosaur skeletons, so some researchers prefer the idea of mutual sexual selection, proposed by David Hone, Darren Naish and Innes Cuthill in 2012, in which both sexes display. Others argue that species recognition, defence or temperature control contributed in particular cases. What has changed is the default: an odd structure is no longer assumed to be a weapon.

The Mesozoic, it turns out, was a noisy, colourful, competitive place. Dinosaurs strutted, called and flashed their plumage, and the birds singing and displaying outside your window are carrying on a tradition more than 150 million years old.

The Centrosaurus Bonebeds

Centrosaurus apertus was a mid-sized horned dinosaur about 5.5 m long and weighing perhaps 2.3 tonnes, with a single long nose horn and two curling hooks drooping forwards from the top of its frill. Lawrence Lambe named it in 1904. It lived around 76 to 75 mya on the humid coastal plain of the Dinosaur Park Formation in southern Alberta, where rivers wound towards the shallow Western Interior Seaway. What makes it remarkable is not any one skeleton but the sheer number of them, lying together in bonebeds.

The best-studied is Bonebed 43 in Dinosaur Provincial Park, a UNESCO World Heritage Site, excavated over many seasons by the Royal Tyrrell Museum under Philip Currie and David Eberth. Its mudstone layer holds thousands of disarticulated bones from dozens of Centrosaurus of all ages, mixed together, with almost nothing from any other plant-eater. Michael Ryan and colleagues analysed it in 2001: the bones were scattered and slightly sorted by flowing water, some were trampled, and shed teeth of tyrannosaurs such as Gorgosaurus and Daspletosaurus lie among them, left by scavengers feeding on the carcasses.

Bonebed 43 is not alone. Near the hamlet of Hilda, east of the park, Eberth and colleagues in 2010 mapped a mega-bonebed, a cluster of bonebeds at the same level spread over more than 2 square kilometres, all dominated by Centrosaurus. The total may represent hundreds and possibly thousands of animals. The rocks show no river channel big enough to explain such a spread. Instead, the team argued that a coastal flood, perhaps driven by a tropical storm, drowned a vast herd on the low plain, and that later floods reworked the carcasses into the layers we see.

How do we know it was one herd? Single-species bonebeds with juveniles, subadults and adults side by side are hard to explain except by animals living and dying together. Similar beds of Pachyrhinosaurus at Pipestone Creek in Alberta and of Einiosaurus in Montana's Two Medicine Formation suggest that many horned dinosaurs gathered in large groups, at least seasonally, perhaps migrating like modern wildebeest.

Did you know? A malformed Centrosaurus leg bone from Dinosaur Provincial Park was diagnosed in 2020 as osteosarcoma, an aggressive bone cancer still found in people today.

Bonebeds turn palaeontology from biography into demography. One skeleton tells us what an animal looked like; a thousand tell us how a population grew, how often it was injured or ill, and how it died. The Centrosaurus herds show that horned dinosaurs lived in crowds, travelled together and, sometimes, perished together on a single bad day.

Iguanodon and the Thumb Spike

Iguanodon bernissartensis was one of the commonest large plant-eaters of Early Cretaceous Europe, browsing the river plains of what is now Belgium, England, Germany and Spain around 126 to 122 mya. It reached about 10 m in length and perhaps 3 to 5 tonnes. Its most famous feature is also its strangest: on each hand, where we have a thumb, it carried a stout, conical thumb spike of bone.

The spike has had a long and confused career. In the 1830s Gideon Mantell, working from scattered bones, sketched it on the animal's nose like a rhinoceros horn, and it stayed there through the life-size models at Crystal Palace in 1854. The truth emerged underground. In 1878 coal miners at the Sainte-Barbe pit in Bernissart, Belgium, broke into a clay-filled fissure at a depth of 322 m and found complete skeletons, eventually at least 38 of them. Excavated block by block and taken to Brussels, they were studied for decades by Louis Dollo, who saw the spikes lying beside the hands, two per animal, exactly where thumbs belong.

The rest of the hand is just as remarkable. The three middle fingers end in blunt, hoof-like bones and are bound together, so they could bear weight on the ground. The little finger, by contrast, was long, flexible and could curl across the palm to hook branches and gather food. One limb thus served as foot, fork and dagger. What the spike was for is still argued over: stabbing a predator at close range is the classic answer, but breaking open seeds and fruit, or fighting rivals, have their supporters.

Did you know? The Bernissart bones are laced with pyrite, which swells and crumbles in damp air, and the Brussels museum has been fighting this "pyrite disease" for more than a century.

Posture changed as much as the hand. Dollo mounted his skeletons upright like giant kangaroos, tails trailing on the ground, and that image dominated museum halls for a hundred years. In the 1980s David Norman of Cambridge re-examined the Bernissart material and showed that the backbone was held horizontal, stiffened by a lattice of bony tendons, and that the heavy, hoofed hands were built for walking. Adults probably spent most of their time on all fours, rising onto their hind legs to run or reach.

The Bernissart skeletons still stand in a glass hall at the Royal Belgian Institute of Natural Sciences, nine upright and others lying as they were found. Together they turned a pile of mystery teeth into the first dinosaur known from complete bodies, and taught palaeontologists a lasting lesson: a single bone, placed with confidence, can be placed entirely wrong.

The Duck-Billed Dinosaurs

The hadrosaurs, or duck-billed dinosaurs, were the great herd animals of the Late Cretaceous. From roughly 86 to 66 mya they spread across North America, Asia, Europe and South America, crossed the sea to Africa and reached Antarctica. In many fossil beds of western North America they are the commonest large dinosaurs of all. They were ornithopods, plant-eating cousins of Iguanodon, with a broad toothless beak at the front of the jaws and a formidable grinding battery of teeth behind it.

Their story begins in New Jersey. In 1858 Joseph Leidy described Hadrosaurus foulkii, a partial skeleton dug from a marl pit at Haddonfield. It was enough to show that a dinosaur's hind legs were much longer than its arms, and in 1868 Benjamin Waterhouse Hawkins mounted it in Philadelphia, the first dinosaur skeleton ever assembled for display. Since then hundreds of skeletons, eggs, nests, trackways and skin impressions have made hadrosaurs among the best-understood dinosaurs of all.

The family splits into two great branches. The saurolophines had flat heads or solid crests: Edmontosaurus, at up to 12 m one of the last dinosaurs standing in North America; Maiasaura, the nesting mother of Montana; Brachylophosaurus, whose specimen "Leonardo", found in 2000, preserved much of its skin; and Shantungosaurus of China, around 15 m long and perhaps 13 tonnes or more, the largest ornithopod known. The lambeosaurines carried hollow crests built from their nasal bones, such as Parasaurolophus, Corythosaurus, Olorotitan from the Russian Far East and Tsintaosaurus from China. The crests differ so much between species, and grow so late in life, that they were almost certainly for display and sound.

Did you know? In 2020 a jawbone from a Moroccan phosphate mine was named Ajnabia odysseus: a hadrosaur that had somehow crossed open sea from Europe to reach Africa.

Older books show hadrosaurs wading in lakes, hiding from predators in the water, their "duck bills" dabbling for soft weeds. That picture rested partly on mitten-like skin around the hands of mummified specimens, once read as webbing. It has gone. Their limbs, hoof-like feet, trackways and the conifer twigs and wood found in their fossil dung show they were land animals that browsed and grazed low vegetation, and their bills were capped with a sharp keratin beak for cropping it.

Hadrosaurs matter because there are so many of them. Bonebeds holding thousands of individuals, growth series from embryo to adult and fossils from the Arctic to Patagonia let palaeontologists study dinosaur populations rather than single heroes. Much of what we know about how dinosaurs grew, nested and lived together comes from the humble duck-bill.

Parasaurolophus and the Trumpet Crest

Parasaurolophus was a lambeosaurine hadrosaur of western North America, living around 76 to 73 mya on the warm coastal plains of what are now Alberta, Utah and New Mexico. It was about 9.5 m long and weighed perhaps 2.5 to 3 tonnes, a big but otherwise ordinary duck-bill except for one thing: a curved, hollow crest sweeping back from the top of its head, in adults roughly a metre long.

William Parks of the University of Toronto named the first species, P. walkeri, in 1922 from a fine skeleton found along the Red Deer River in Alberta. Two more species followed from the south: P. tubicen, described by Carl Wiman in 1931 from New Mexico, and the short-crested P. cyrtocristatus, named by John Ostrom in 1961. What makes the crest special is not the outside but the inside. It is built from the nasal and premaxillary bones, and air breathed in through the nostrils ran up the crest through paired tubes, looped back down at the far end and returned to the throat.

Early ideas about the crest were imaginative. It was proposed as a snorkel, though there is no opening at the tip; as an air tank for diving; as a weapon; and as a prop for a tough flap of skin. In 1981 David Weishampel took a different approach and treated the crest as a musical instrument, calculating that its long tubes would resonate at low frequencies like a crumhorn or bassoon. In the late 1990s scientists at Sandia National Laboratories CT-scanned a P. tubicen skull from New Mexico and modelled the sound it could produce: a deep, rumbling, mournful tone.

Did you know? A baby Parasaurolophus found in Utah, nicknamed "Joe" and described in 2013, was less than a year old and had only a low bump where its crest would grow.

That growth pattern is a clue. Juveniles lacked long crests, and the crest's shape differs from species to species, which points to signalling: recognising your own kind, finding a mate, perhaps announcing age and status. CT studies of lambeosaurine braincases suggest their inner ears were tuned to low-frequency sound, which fits animals that called to one another across a forest. The crest may also have been brightly coloured, but no pigment evidence survives, and whether adults of both sexes had the same crest is still uncertain.

Parasaurolophus shows how far modern tools can stretch a fossil. Bone can tell us about breathing, hearing and even voice, provided we ask the right questions. Seventy-five million years on, we can make a fair guess at what a dinosaur sounded like, and it was not a roar.

Maiasaura: The Good Mother Lizard

Maiasaura peeblesorum, the "good mother lizard", was a saurolophine hadrosaur about 9 m long that lived around 76.7 mya on the uplands of what is now Montana, in the rocks of the Two Medicine Formation. It is the dinosaur that made parental care a serious scientific question, and the first dinosaur genus named with a feminine ending.

The discovery began in a rock shop. In 1978 Marion Brandvold, who ran a shop in Bynum, Montana, showed Jack Horner and his friend Bob Makela some tiny bones she had collected near Choteau. They were baby hadrosaur bones. Digging on the nearby Peebles ranch, Horner and Makela found a bowl-shaped nest, about 2 m across, holding the remains of fifteen youngsters roughly a metre long. They named the animal in 1979, honouring the landowners in its species name.

More nests followed on a low rise the team called Egg Mountain and at other levels nearby. The mound nests were scooped from mud, probably lined with rotting vegetation that warmed the eggs, and spaced around 7 m apart, about the length of an adult. That spacing suggests a colony, like a modern seabird rookery, where neighbours nested just out of pecking range. Clutches seem to have held several dozen eggs. The nestlings had worn teeth but leg bones whose ends were not yet fully formed, which Horner read as babies too weak to walk far, fed in the nest by their parents.

Did you know? In 1985 a piece of Maiasaura bone and some eggshell flew aboard the Space Shuttle Challenger on the Spacelab 2 mission, making it the first dinosaur in space.

That interpretation has been debated ever since, because worn teeth and soft joints can arise in other ways, and the details of Maiasaura family life cannot be watched directly. Later work added firmer evidence. A bonebed nearby, caught in volcanic ash, contains the remains of a vast herd, often estimated at around 10,000 individuals. In 2015 Holly Woodward and colleagues sliced the shin bones of some fifty individuals and read their growth rings: babies grew explosively, most died in their first year, and survivors reached adult size at around eight years old. Whether or not the parents carried food, the young clearly lived for weeks in the nests and the species lived in large groups.

Maiasaura changed the picture of dinosaurs as cold, solitary reptiles that laid eggs and left. It put families, colonies and fast-growing young into textbooks, and it is now Montana's state fossil. Not bad for a handful of bones in a shop window.

Edmontosaurus and the Dinosaur Mummies

Edmontosaurus was one of the last and largest duck-billed dinosaurs, a flat-headed saurolophine that grazed the coastal plains of western North America between about 73 and 66 mya. E. annectens, from the Hell Creek and Lance formations, reached around 12 m and several tonnes, and shared its world with Tyrannosaurus and Triceratops. Lawrence Lambe named the genus in 1917 after Edmonton, Alberta. What makes it famous, though, is not its bones but its skin.

In 1908 the fossil hunter Charles H. Sternberg and his sons were working the Lance Formation near Lusk, Wyoming. While their father was away fetching supplies, George and Levi uncovered an Edmontosaurus lying on its back, arms flung out, with the skin pressed against its ribs and limbs. Henry Fairfield Osborn bought it for the American Museum of Natural History, where it is still known as AMNH 5060, and in 1912 described it as a dinosaur mummy. A second Sternberg mummy went to the Senckenberg Museum in Frankfurt in 1910.

These are not mummies in the Egyptian sense. No soft tissue survives; what remains is skin texture and outline, preserved as impressions in the surrounding sediment. They show a pebbly mosaic of small, non-overlapping scales with patches of larger ones, and a fleshy covering on the hands that once suggested webbed paddles. A later specimen, "Dakota", found in North Dakota by Tyler Lyson in 1999, preserved skin wrapped around the bones of a forearm and tail.

Did you know? Bite marks on "Dakota" suggest that crocodiles and other scavengers opened the carcass, letting gases and fluids escape so the skin dried and settled onto the bones.

How such delicate surfaces survive has been argued for a century. A 2022 study of "Dakota" by Stephanie Drumheller and colleagues proposed a route through deflation: punctured by scavengers, the body dried out instead of rotting. In 2025 Paul Sereno and colleagues described new Edmontosaurus mummies from the same Wyoming area, arguing that a film of clay formed around the drying body and set into a thin mask only millimetres thick. Their specimens showed a fleshy crest running down the neck and back and hooves on the hind toes, details no skeleton could have revealed.

Edmontosaurus also carries evidence of danger. A specimen in Denver has a healed bite on its tail, apparently from a Tyrannosaurus, showing that it survived an attack. The mummies matter because they turn a skeleton into an animal with a surface, a silhouette and a body plan we can test. A century after the Sternbergs, the skin is still teaching us how to read the flesh.

Dental Batteries: The Best Chewers Ever

Most reptiles bite and swallow. Hadrosaurs chewed, and they did it with one of the most sophisticated tooth arrangements any animal has evolved. A dental battery is a block of teeth packed side by side and stacked in vertical columns, so that as the working tooth wears down, the one beneath it rises to take its place. Duck-billed dinosaurs had batteries in both upper and lower jaws, and the horned ceratopsians evolved their own version for slicing.

The numbers are startling. An adult Edmontosaurus could have around 60 tooth columns in each side of each jaw, with three or more teeth stacked in every column, giving many hundreds of teeth in the mouth at once. Only the top few were in use, worn into a single continuous grinding surface like a rasp. New teeth formed constantly below, so the battery never wore out, and fossil collectors find shed hadrosaur teeth by the bucketful in Late Cretaceous rocks across North America and Asia.

The real secret lies inside each tooth. In 2012 Gregory Erickson and colleagues cut thin sections of Edmontosaurus teeth and found six distinct tissues, including enamel, several kinds of dentine and cementum. Horses and bison, modern champions of grinding, have four. Because the tissues wear at different rates, the chewing surface stayed rough, with tiny ridges and valleys that sharpened themselves as the animal fed. It is a design no living reptile possesses.

Did you know? A hadrosaur tooth contained more kinds of tissue than the teeth of a horse or a bison, the mammals usually held up as the finest grinders alive today.

How the jaws moved is a long-running puzzle, because hadrosaurs could not grind side to side as cows do. In the 1980s David Weishampel and David Norman proposed pleurokinesis: as the lower jaw closed, the upper jaws hinged slightly outward, sliding the two batteries across each other. Critics thought the skull too rigid for it. In 2009 Vince Williams, Paul Barrett and Mark Purnell studied the microscopic scratches on Edmontosaurus teeth and found the patterns fitted a complex, multi-directional chewing stroke, broadly consistent with the idea, though details are still argued.

What went through those jaws is known too. Fossil dung from the Two Medicine Formation of Montana, attributed to hadrosaurs, is packed with fragments of rotting conifer wood, and some contains crustacean shell, perhaps eaten with the logs it lived in. Hadrosaurs could tackle tough plant food that other herbivores left alone, and that may help explain why they became the most numerous large herbivores of the late Mesozoic. Sixty-six million years before the first horse, dinosaurs had already perfected the chew.

Polar Dinosaurs

It sounds like a contradiction, but dinosaurs lived at the poles. The Cretaceous world was much warmer than today, with no permanent ice caps, yet high latitudes still had months of winter darkness, cold snaps and probably snow. Fossils from northern Alaska, south-eastern Australia and Antarctica show that dinosaurs not only visited these places but nested, hatched and grew up there.

The northern evidence comes from the Prince Creek Formation on Alaska's North Slope, laid down around 70 to 69 mya at a palaeolatitude of roughly 80° N, farther north than almost anywhere people live today. In 1961 the Shell geologist Robert Liscomb collected bones along the Colville River, but they sat in a store room, labelled as mammal, until the 1980s. The Liscomb Bonebed has since yielded thousands of bones of a duck-bill close to or within Edmontosaurus, along with the horned Pachyrhinosaurus perotorum and a small tyrannosaur, Nanuqsaurus. In 2021 Patrick Druckenmiller and colleagues reported tiny teeth and bones of hatchlings and embryos from at least seven dinosaur groups, which means these animals bred in the Arctic rather than visiting it.

In the south, Australia was joined to Antarctica and sat well inside the polar circle. At Dinosaur Cove in Victoria, Tom Rich and Patricia Vickers-Rich spent years in the 1980s and 1990s tunnelling into sea cliffs with volunteers and mining equipment to extract bones around 106 mya old. Their best-known find, the small ornithopod Leaellynasaura amicagraphica, named in 1989 after their daughter, had large eye sockets and an extraordinarily long tail, about three times the length of its body.

Did you know? Dinosaur Cove fossils were so hard to reach that the team blasted and drilled tunnels into the coastal cliffs, working with miners' equipment to follow the bone layer.

Antarctica itself has produced dinosaurs from Jurassic mountains to Cretaceous coasts, including ornithopods such as Trinisaura from James Ross Island, named in 2013, and the armoured Antarctopelta. How these animals survived the dark months is debated. Some may have migrated, though trackways and nestlings suggest many stayed put. Bone histology gives mixed answers: some polar species show growth rings marking hard winters, while early work on Leaellynasaura found steadier growth. Insulating feathers or fuzz, eating stored-up energy and slowing down are all plausible, and none is yet proven.

Polar dinosaurs knock down the old idea that dinosaurs were sun-loving reptiles that needed tropical heat. They point instead to animals with high metabolisms, able to keep warm and active through a polar night. The dinosaurs did not just tolerate the cold; they raised their families in it.

Herds on the Move

Did dinosaurs live in herds? For a long time the question could only be answered with a shrug and a painting. Today there are four kinds of hard evidence, and together they show that many plant-eaters, and some meat-eaters, lived in groups for at least part of their lives.

The first is the bonebed: a single layer packed with many individuals of one species, killed together by a flood, drought or ash fall. In Montana's Two Medicine Formation a bonebed of Maiasaura stretches for kilometres and is often estimated to hold around 10,000 animals. In Inner Mongolia, a group of more than a dozen young Sinornithomimus, ostrich-like dinosaurs, died together in a drying lake. Bonebeds of horned dinosaurs in Alberta tell the same story. One caution: rivers can pile up carcasses that never lived together, so each bed has to be read carefully.

The second is trackways. In 1940 Roland T. Bird mapped the Davenport Ranch site in Texas, where some two dozen sauropods walked the same way at the same time, about 113 mya. At Peace River Canyon in British Columbia, parallel trails of large ornithopods show animals moving side by side. Queensland's Lark Quarry, with over 3,000 small footprints, was famous as a "stampede" fleeing a big predator; later studies argued some tracks were made by swimming animals, and the site is still debated.

Did you know? The oldest good evidence of dinosaur herds comes from Mussaurus in Patagonia, about 193 mya, where juveniles and adults were found in separate age groups around a nesting ground.

The third is nesting grounds. In 2021 Diego Pol and colleagues described a site in southern Argentina with more than 100 eggs and the skeletons of around 80 Mussaurus, an early long-necked dinosaur. Youngsters of similar age were buried together, apart from adults, suggesting a herd organised by age, millions of years before the giant sauropods.

The fourth is chemistry. In 2011 Henry Fricke and colleagues measured oxygen isotopes in the tooth enamel of Camarasaurus from the Morrison Formation and found the animals had drunk water from distant uplands for part of the year, implying seasonal journeys of up to about 300 km. Whether Arctic hadrosaurs migrated thousands of kilometres to escape winter is disputed, because hatchling bones show many stayed in the north.

Herding changes how we picture dinosaur behaviour, ecology and even intelligence. Living in groups means sharing vigilance, competing for mates and managing young, all things we see in birds and mammals. The pattern now looks clear: for many dinosaurs, life was a crowd.

Growing Up Dinosaur

Every dinosaur, however huge, began in an egg no bigger than a football and usually much smaller. A sauropod hatchling weighing a few kilograms could become an adult of 50 tonnes or more, a ten-thousand-fold increase. How dinosaurs managed it, and how fast, is written in their bones and teeth.

The key tool is bone histology. Cut a thin slice through a leg bone and you often see lines of arrested growth, or LAGs: rings laid down each year when growth paused, much like tree rings. Count them and you can estimate age; measure the spacing and you can plot a growth curve. In 2004 Gregory Erickson and colleagues used this method on tyrannosaurs and found a teenage growth spurt, and the same S-shaped curve turns up across dinosaurs: slow at first, explosive in adolescence, levelling off at adulthood.

Growth starts before hatching. Embryonic teeth also lay down daily growth lines, and in 2017 Erickson's team counted them in embryos of Protoceratops and the duck-bill Hypacrosaurus. Incubation took about three months for the first and nearly six for the second, far longer than for birds of similar egg size and closer to reptiles. Long incubation would have left eggs exposed to predators and floods for a whole season.

Did you know? Young Maiasaura grew so fast that most of their adult size was reached in a few years, yet Holly Woodward's growth study found that around nine in ten died in their first year.

Adolescence brought change in shape as well as size. Juvenile Parasaurolophus had only a bump where the crest would grow, young horned dinosaurs had backward-curving horns, and baby sauropods had short necks and big heads. Some juveniles may have lived quite differently from adults, in their own groups and eating different food, which helped many species avoid competing with their own offspring.

Dinosaurs also started breeding before they finished growing. Female birds lay down a temporary calcium store called medullary bone inside their leg bones when producing eggshell. In 2005 Mary Schweitzer and colleagues identified it in a Tyrannosaurus femur, and in 2008 Andrew Lee and Sarah Werning found it in Tenontosaurus and Allosaurus that were still growing. Like crocodiles and unlike most modern birds, dinosaurs reached sexual maturity before reaching full size.

This life history, with fast growth, early breeding and high juvenile mortality, may be one reason dinosaurs could become so large and so successful. It also changes how fossils are named: more than one supposed new species has turned out to be a youngster of a known one. Read carefully, a single slice of bone is a dinosaur's whole biography.