Did Dinosaurs Live in Herds? Bone Beds, Trackways and the Social Lives of Giants

A mass death assemblage is a deposit in which many individuals of one species are buried together, and dinosaurs left many. The Centrosaurus beds of Alberta, the Maiasaura beds of Montana with an estimated 10,000 animals, the Edmontosaurus beds of the Dakotas and the Protoceratops and Psittacosaurus groups of Mongolia all contain one species in numbers, with juveniles and adults mixed, which is the signature of a herd caught by a single event, a flood, a drought, a volcanic ashfall, rather than of animals dying one by one over years. The pattern is commonest in the plant-eaters, the horned ceratopsians and the duck-billed hadrosaurs, and it appears more often in the Late Cretaceous than earlier, which may reflect the animals or the rocks.

Footprints record the living. Trackway sites on every continent show multiple dinosaurs of one kind moving in the same direction across the same surface, and some show more. At Davenport Ranch in Texas, 23 sauropod trackways run parallel, with the smaller animals' prints among the larger, which was read as young in the centre of the herd, though a re-examination found the pattern less tidy. In the Yanguoxia site in China and along the Purgatoire River in Colorado, sauropods walked in groups whose members turned together. Theropod tracks are more often solitary, and a set of parallel tracks of the large carnivore Acrocanthosaurus at Glen Rose in Texas, running beside sauropod tracks, has been read as a hunt and as a coincidence of animals using the same shore. The evidence in brief:

• •Single-species bone beds with mixed ages: strong evidence of herds in ceratopsians, hadrosaurs, Psittacosaurus and some sauropods
• •Parallel trackways with consistent spacing: groups moving together, mainly in sauropods and ornithopods
• •Nesting colonies: hadrosaurs and titanosaurs laying in the same ground in numbers, with nests spaced as in a bird colony
• •Juvenile-only groups: clusters of young Sinornithomimus and Psittacosaurus, suggesting that the young of some species stayed together apart from adults
• •Solitary finds: most large theropods and the armoured ankylosaurs are found alone, which may mean solitary lives or may mean preservation

The pack-hunting raptor of the films rests on one site: a Deinonychus bone bed in Montana where several of the sickle-clawed predators lie with a large plant-eater, Tenontosaurus, which was read in the 1960s as a pack that had brought down prey and died in the attempt. The reading has weakened. The isotopes in the teeth of Deinonychus of different ages show that the young ate different food from the adults, which in modern animals is a sign that the young were not fed by a pack; the crocodile and Komodo dragon model, in which several predators converge on a carcass and squabble over it, fits the bone bed as well. Tyrannosaur bone beds in Alberta and Utah contain several individuals of different ages together, which is suggestive, and some workers argue for family groups; the honest answer is that the evidence for coordinated hunting in any dinosaur is thin, and that gathering at a kill is not the same thing.

Birds are dinosaurs and crocodiles are their closest living cousins, and both bracket the possibilities. Crocodiles gather in numbers, guard nests and carry their young, but do not cooperate in any organised way; birds flock, colony-nest, form family groups and, in a few species, hunt together. A behaviour found in both, such as nest guarding, was probably present in the dinosaurs between them, and the brooding oviraptorosaurs and the crocodile-style buried nests of sauropods fit that. Beyond it, each species is its own case. The herbivores that lived in numbers were probably doing what herbivores do now, finding safety and finding each other; the predators were probably doing what large predators mostly do now, which is hunting alone or in loose company, and the raptors of the films remain a hypothesis with one bone bed behind it.

The fossil record shows where animals died, not how they lived, and a mass death tells of one bad day rather than of daily habit; animals that lived alone can die together at a shrinking waterhole. Trackways record a few minutes, and a group walking together may have been a herd or a queue at a river. The colours, calls, displays and courtship that the crests of the hadrosaurs and the frills of the ceratopsians imply, and that would have been the fabric of any social life, are inferred and not observed. The dinosaurs that lived in herds are those whose bones say so, the ones that lived in packs are those whose bones might, and the rest are silent.

Deciding whether a pile of bones records a herd that died at once is detective work, and the science behind it, called taphonomy, studies everything that happens between death and discovery. Bones left exposed on the ground crack and flake in the sun in recognisable stages, so a bed in which all the bones show the same slight weathering suggests the animals died and were buried within a short time of each other. A bed with fresh bones mixed with heavily weathered ones points instead to a place where carcasses collected over many years.

Other clues come from arrangement and damage. Long bones that all point the same way were lined up by flowing water, which tells researchers a flood moved the remains and may have carried them some distance. Tooth marks show that scavengers fed before burial. The mix of ages is read from the bones themselves: size alone can mislead, but growth rings in thin sections give ages in years, so a bed can be shown to hold yearlings, adolescents and adults together. Only when several such clues agree do palaeontologists feel confident calling it a herd.

Living in a crowd gives a plant-eater advantages that ecologists can measure in animals today. More eyes and ears mean a predator is spotted sooner, so each animal can spend more time feeding and less time looking up. A hunter that attacks a group usually takes one victim, so the odds for any single member fall as the group grows, an effect called dilution. The biologist W. D. Hamilton added a sharper idea in 1971, the selfish herd: each animal gains by pushing towards the middle and putting others between itself and danger, which can pull scattered animals into tight groups without any cooperation at all.

Groups also have costs, and those help explain which dinosaurs formed them. A herd strips the plants around it quickly and shares its diseases and parasites, so big groups suit animals that keep moving, and there is evidence that some dinosaurs did. Oxygen isotopes in the tooth enamel of the sauropod Camarasaurus, from the Morrison Formation of the western United States, record water drunk in the lowlands for part of the year and on higher ground for the rest, which suggests seasonal journeys of a few hundred kilometres. Among living grazers such as wildebeest and caribou, the longest journeys are made by the largest herds.

Some of the best evidence for how dinosaur bone beds formed comes from watching living herd animals die. Every year the great wildebeest migration in East Africa crosses the Mara River, and at difficult crossings large numbers of animals drown. One study estimated that thousands of wildebeest drown there in a typical year, leaving carcasses piled along the banks and caught on rocks.

Those carcasses rot, scatter and are buried in river sediment, the same sequence palaeontologists infer for many dinosaur bone beds packed with individuals of one species. Watching it happen helps researchers recognise the signs: bones sorted by size and shape by the current, a narrow range of ages, and damage from scavengers. It also supports the idea that some dinosaur herds crossed rivers together and paid the same price.

Most of the strongest evidence for herds comes from the Late Cretaceous, near the end of the dinosaurs' reign, which left open whether group living arose late. A site in southern Patagonia changed that. In 2021 a team led by the Argentine palaeontologist Diego Pol described a nesting ground of Mussaurus, an early long-necked plant-eater from the group that gave rise to the sauropods, containing more than a hundred eggs and the skeletons of around eighty individuals, from embryos to adults. Volcanic ash in the same rocks was dated to about 193 million years ago, early in the Jurassic.

The arrangement of the bones was as telling as their number. Eggs lay at several levels, suggesting the animals came back to the same ground over successive seasons, as colonial nesting birds do. The skeletons were also grouped by age: young animals of similar size were found together, while adults turned up alone or in pairs, a pattern called age segregation that is seen in living herding animals. One cluster of juveniles appears to have died together, perhaps in a drought. If the interpretation holds, organised group living existed some 40 million years earlier than the previous oldest evidence, close to the origin of the long-necked dinosaurs.

Many plant-eating dinosaurs lived in groups, as single-species bone beds of thousands of horned and duck-billed dinosaurs, parallel sauropod trackways and nesting colonies show, and some young stayed together apart from adults. Evidence that predators hunted in coordinated packs rests on one Deinonychus site and has weakened, with gathering at a carcass fitting the bones as well; birds and crocodiles bracket what is likely, and for most species the record of a single death or a few minutes' walking cannot say more.