What Did Dinosaurs Eat? Teeth, Gut Stones and Fossil Dinners

The single most informative fossil is a tooth, because shape follows function closely. Blade-like teeth with serrations on both edges, the pattern in tyrannosaurs and allosaurs, are knives for slicing flesh, and the serrations are the same solution a steak knife uses. Leaf-shaped teeth with coarse cusps, as in the early plant eaters, strip and crush vegetation. The broad batteries of the duck-billed dinosaurs, in which hundreds of small teeth lock together into a grinding surface that renews itself continuously as the top wears away, are a mill for tough plants. Conical unserrated teeth, as in spinosaurs, grip slippery prey and belong to a fish eater. Microscopic scratches and pits on the worn surfaces, examined the way they are in living mammals, separate grazers from browsers, and in some hadrosaurs the wear indicates a diet including gritty material near the ground.

The long-necked sauropods present the hardest question, since an animal of thirty tonnes needs an enormous quantity of low-quality food and had teeth that could not chew. Their solution was to skip chewing. The peg-like or spoon-shaped teeth stripped leaves from branches and the material went down whole to ferment in an immense gut, where micro-organisms broke down the cellulose over days, which is what a cow does with a smaller stomach and more chewing. That arrangement makes the long neck useful in two ways, sweeping a large feeding area without moving the body and reaching into trees, and it explains why the teeth of some species were replaced every month or so. The plants available were not grass, which had barely appeared, but conifers, ferns, horsetails, cycads and, later, flowering plants, and experiments feeding modern equivalents to fermentation vessels have found horsetails and araucaria far more nutritious than their reputation suggested.

Occasionally the food itself survives, and these specimens are worth more than any amount of inference from teeth:

• •A small Sinosauropteryx from China preserved with a mammal jaw in its abdomen, and a Microraptor with a bird and, in another specimen, a lizard swallowed whole and head first
• •A Baryonyx from Surrey with fish scales etched by stomach acid in its ribcage, along with bones of a young Iguanodon
• •A mummified hadrosaur whose gut contents included conifer needles, twigs and, unexpectedly, fungi and rotted wood, suggesting a less choosy diet than assumed
• •Coprolites, fossil dung, including a large specimen from Saskatchewan packed with shattered bone, which indicates a tyrannosaur that crushed bone rather than stripping meat cleanly
• •Gastroliths, smoothed stones found within the ribcages of some plant eaters and birds, swallowed to grind food in a muscular gizzard, although many stones once called gastroliths have turned out to be ordinary pebbles
• •Tooth marks matched to a specific animal, most famously a Triceratops pelvis bearing scores of tyrannosaur bites, some of which healed, showing the victim survived

A long argument ran about whether the largest predator of the late Cretaceous hunted or scavenged, and the answer that has settled is that the question was badly posed. Every large carnivore alive takes carrion when it can, lions included, and none survives on it alone, because carcasses are too unpredictable. The evidence for active hunting is direct: healed bite marks on the bones of plant eaters mean the animal bitten walked away, which requires a living attacker. The evidence for scavenging is equally good, and the animal had the best sense of smell of any known dinosaur, judged by the size of the olfactory bulbs in its braincase, along with a bite force able to shatter bone and get at marrow. It did both, as everything of its size does.

Several groups broke the simple split. The therizinosaurs were theropods, from the lineage that produced the meat eaters, with enormous claws and a plant eater's gut and teeth, an anatomy that took years to interpret. The oviraptorosaurs, named unfairly after a specimen found on a nest that was assumed to be robbing it and was in fact its own, had beaks and no teeth, and probably ate eggs, shellfish, plants or all three. Some small theropods have the gut stones and reduced teeth of plant eaters. Insect eaters existed, with the stubby single-clawed arms of Mononykus interpreted as a tool for opening termite nests, and the earliest known mammals ate dinosaur young, since one specimen from China was found with the bones of a baby Psittacosaurus in its stomach, and another was fossilised in the act of attacking one.

Diet is reconstructed from tooth shape and microscopic wear, jaw mechanics, gut contents, fossil dung, gut stones and bite marks on bone. The giant sauropods did not chew but fermented conifers, ferns and horsetails in a huge gut, the duck-bills ground tough plants in continuously renewed dental batteries, and the great predators both hunted and scavenged, as the healed bites on their victims and the bone-filled dung both show. Several groups broke the pattern, including a theropod lineage that turned to plants and mammals that ate dinosaur young.