What did herbivorous dinosaurs eat?

Herbivorous dinosaurs did not live on modern grass. For much of the Triassic and Jurassic, the landscapes they inhabited were dominated by ferns, horsetails, cycads, ginkgoes and conifers. Flowering plants diversified later and became widespread in the Cretaceous, but they did not instantly replace older plant groups. What an animal could eat depended on its anatomy, age, habitat and the plants available in its own time.

Direct traces of a meal are rare. Most diets are reconstructed by comparing tooth shape, jaw movement, wear and the plants known from a fossil locality. The broader account of how palaeontologists infer dinosaur diets explains why each clue has limits.

A changing menu through the Mesozoic

Ferns, horsetails, cycads and conifers formed important parts of many early Mesozoic plant communities. Ginkgoes were also present. In the Cretaceous, flowering plants became increasingly diverse and widespread, adding new food sources without making every ecosystem or dinosaur diet the same. Fossils of leaves and wood help reconstruct the local vegetation, but their presence near a dinosaur does not prove that the animal ate each plant.

One unusually informative find is the preserved gut content of a juvenile Diamantinasaurus from Australia. It included remains attributed to conifers, seed ferns and flowering plants. That direct sample suggests a broad diet for this individual. It cannot establish what every sauropod ate, or whether the same animal chose the same foods throughout the year.

Different teeth, different processing strategies

Hadrosaurs and ceratopsians had rows of tightly packed teeth that formed dental batteries. As upper and lower teeth met, their surfaces wore into working edges that could crop or grind tough plant material. Multiple tissues and ongoing tooth replacement helped maintain those surfaces. The anatomy reveals a durable way to process food, but it does not by itself identify an exact list of plants.

Sauropods used other strategies. In many forms the teeth mainly gathered or stripped vegetation, while much of the processing happened after swallowing. Tooth replacement rates differed among sampled genera: estimates from studied Diplodocus teeth are about 35 days, compared with about 62 days in Camarasaurus. Those figures describe particular specimens and should not be treated as a single rate for all long-necked dinosaurs.

Microscopic scratches and other wear patterns can help estimate how teeth contacted food. They may record hard or abrasive items, but grit can also contribute to the marks. Researchers combine these observations with tooth anatomy, jaw mechanics and plant remains. A single polished tooth rarely settles a diet by itself.

Gastroliths and the limits of the stomach-mill idea

Stones called gastroliths occur with some dinosaur remains. Their presence alone does not show that every sauropod used stones to grind food. To support a gastric mill interpretation, researchers examine the number, size, polish and geological association of the stones. Some collections of stones associated with sauropods have too little combined mass to work like the powerful grinding systems of many birds.

Stones may have been swallowed incidentally or served another function, but each proposed role needs evidence from the particular specimen. Their absence also tells little: stones can move after burial, and stomach contents rarely fossilise in the first place.

Foraging height and seasonal changes

A long neck could extend a sauropod's reach, but it does not prove that the animal spent most of its time feeding in the tallest tree crowns. Neck flexibility, posture, body size and the plants available all affected the range an individual could reach. Different herbivores might have used different heights or parts of a plant community, yet fossil evidence rarely preserves a complete picture of that partitioning.

Food supply also changed among habitats and seasons. River margins, uplands and wetter ground could support different plant communities. Seasonal shifts are plausible, but fossils seldom preserve a calendar of meals. Precise claims about a species' favourite plant or seasonal menu usually go beyond what the available remains can show.

What can be said with confidence?

Plant fossils establish which vegetation grew in a region. Teeth and jaw mechanics show how a dinosaur could crop or process food. Wear records recent contact, and rare gut contents preserve actual meals. Coprolites can retain digested plant tissues or phytoliths, although their maker may be uncertain. Each line answers a different question.

The safest picture is diverse: herbivorous dinosaurs used several ways to gather and process plant foods, and their diets changed across time and place. A fossil meal can make that picture more concrete, but one specimen cannot stand in for every member of its group.