What Did Dinosaurs Eat? Herbivores, Carnivores, and the Evidence

Figuring out what a living animal eats is easy — watch it. Figuring out what an extinct animal ate from fragments of bone and rock takes chemistry, biomechanics, and a methodical approach to fossilized excrement. Paleontologists have developed at least seven independent lines of evidence for reconstructing dinosaur diets, and when those lines converge on the same answer, the conclusion is robust. The picture they reveal is a rich, ecologically layered world where giant herbivores reshaped entire landscapes and carnivores occupied finely divided predatory niches — and where the boundary between plant-eater and meat-eater was sometimes blurry.

The Seven Lines of Evidence Paleontologists Use

Reconstructing diet from fossils is never a single-source conclusion. Teeth tell one part of the story, chemistry tells another, and fossilized gut contents tell a third. Each line of evidence has its own blind spots, which is why multiple independent methods converging on the same answer carries far more weight than any one alone.

Tooth morphology — the shape, size, and arrangement of teeth — is the most immediately readable signal. Blade-like, serrated teeth with recurved tips (called ziphodont teeth) indicate a predator that slices through flesh. Broad, flat, grinding teeth indicate a plant processor. Peg-like, cylindrical teeth work for nipping vegetation. Conical, slightly recurved teeth suit fish-catching. A paleontologist can often infer broad dietary category from tooth shape alone, the same way a wildlife biologist can infer a mammal's diet from its dentition.

Tooth wear patterns go further. Scanning electron microscope (SEM) analysis of microscopic scratches and pits on enamel surfaces reveals what the teeth ground against. Tough, gritty plants (like ferns growing close to the ground) leave different scratch patterns than soft leaves. Bone-crunching leaves distinctive crushing pits. SEM of dinosaur tooth enamel has been used since the 1980s to distinguish browsers from grazers and to identify bone-processing in theropods.

Gastroliths — smooth, polished stones found inside or immediately adjacent to ribcages — were swallowed deliberately by some dinosaurs to aid digestion, much as modern birds and crocodilians swallow grit. They are associated almost exclusively with herbivores and indicate a stomach that needed mechanical grinding to break down tough plant material.

Coprolites (fossilized dung) preserve actual dietary content — bone fragments, plant material, seeds, pollen, scale fragments from fish — in direct, unambiguous form. A coprolite full of crushed bone tells you its producer was a bone-eater. A coprolite containing grass seeds dates the evolution of grass-eating. Coprolites are rarer than bones but have provided some of the most precise dietary information available.

Gut content fossils are exceptionally rare but extraordinarily valuable. When the fossilization process captures material inside the body cavity — actual last-meal contents preserved in place — there is no interpretation required. Several dinosaur specimens have preserved gut contents, ranging from plant material in hadrosaurs to fish remains in Microraptor.

Bite marks on bones identify both predator behavior and which animals consumed what. A bone carrying bite marks with a particular cross-section size and spacing can be matched to the jaw anatomy of contemporary species. Bone surface texture under the marks reveals whether the animal was dead before it was bitten (scavenging) or show evidence of perimortem trauma. Force calculations from bite-mark depth and morphology allow mass estimates of bite force.

Isotope chemistry uses the fact that different foods carry different ratios of stable isotopes — particularly carbon-13/carbon-12 ratios (which distinguish C3 plants from C4 grasses) and nitrogen-15/nitrogen-14 ratios (which increase predictably as you move up trophic levels, from plant to herbivore to carnivore). Measuring isotope ratios in fossilized enamel or bone collagen (where it survives) can confirm broad dietary category and, in some cases, distinguish food sources.

Source: Fiorillo, A.R., Palaios (1991) — Early application of SEM microwear analysis to dinosaur tooth surfaces, establishing dietary inference protocols for fossil reptiles.

Source: Chin, K. & Gill, B.D., Nature (1996) — First confirmed description of a large theropod coprolite containing bone fragments, providing direct evidence of bone-processing behavior in large carnivorous dinosaurs.

Herbivores: The Majority and Their Strategies

Roughly two-thirds of named dinosaur species were herbivores — a proportion that mirrors most modern terrestrial ecosystems, where plant biomass vastly exceeds animal biomass and more species fill herbivore niches than predator niches. But "herbivore" is not a single strategy. Dinosaur plant-eaters evolved radically different approaches to the same nutritional problem: how to extract energy from cellulose-rich plant material that resists digestion.

Sauropods — the long-necked giants like Brachiosaurus, Diplodocus, and Patagotitan — were bulk feeders. Their small, peg-like or spatula-shaped teeth were not designed for grinding; they stripped and cropped vegetation, swallowing it whole. The actual mechanical breakdown happened in a muscular gizzard, aided by gastroliths. Gastroliths have been found inside the body cavities of Seismosaurus (Diplodocus hallorum) and other sauropods, confirming gizzard-based digestion. A sauropod's digestive strategy resembles a large modern bird more than a mammalian herbivore like a cow.

Because sauropods could not chew, they compensated with throughput. Estimates for the daily food intake of large sauropods (based on metabolic scaling from their estimated body mass) suggest consumption on the order of hundreds of kilograms of vegetation per day. At that intake rate, even low-quality plant material yields adequate energy — but requires an enormous fermentation chamber to extract it. Sauropod body cavities were large enough to house very long digestive tracts, and microbial fermentation in the gut would have done much of the chemical work.

Ornithopods — including Iguanodon, hadrosaurs (duck-billed dinosaurs), and their relatives — evolved one of the most sophisticated dental systems in vertebrate history: the dental battery. In hadrosaurs like Edmontosaurus, each "tooth position" in the jaw contained a vertical column of replacement teeth stacked above the functional one, so as teeth wore down they were continuously replaced from below. A single hadrosaur jaw could contain hundreds of individual teeth packed into a continuous grinding surface. SEM analysis of hadrosaur tooth wear reveals complex, multi-directional scratches consistent with processing tough, fibrous plant material — probably conifer needles, twigs, and coarse vegetation near the end of the Cretaceous.

Ceratopsians like Triceratops had tall, blade-like teeth arranged in shearing batteries — not grinding surfaces but slicing surfaces. This indicates a diet of tougher, woody material that needed to be cut through rather than pulped. The powerful jaw musculature suggested by ceratopsian skull anatomy, combined with the shearing teeth, points toward feeding on palm fronds, cycads, or similar tough vegetation.

Ankylosaurs had relatively simple, leaf-shaped teeth and small heads, suggesting selective feeding on low-growing, soft vegetation rather than bulk processing of coarse material. Their ground-hugging posture (their heads were typically less than 1 m off the ground) constrains where they fed.

Stegosaurs had surprisingly small, weak teeth for their body size — the beak-and-tooth combination seems adapted for cropping soft vegetation at low levels. There is good evidence from isotope analysis that stegosaurs fed on different vegetation than the large sauropods they coexisted with, occupying a distinct dietary niche in the same ecosystem.

Source: Whitlock, J.A., Journal of Vertebrate Paleontology (2011) — Functional analysis of diplodocid sauropod teeth, establishing that peg-like teeth were adapted for stripping rather than grinding, requiring gastrolithic or fermentative digestion.

Source: Erickson, G.M. et al., Science (2012) — Biomechanical analysis of hadrosaur dental battery function, demonstrating that the complex tooth arrangement allowed more sophisticated processing of vegetation than any other known reptile.

The Gastrolith Evidence: Swallowing Rocks on Purpose

The discovery of smooth, polished stones inside or near dinosaur ribcages has fascinated paleontologists since the 19th century. A gastrolith (from Greek: gastro = stomach, lithos = stone) is a stone swallowed by an animal and retained in a muscular stomach (gizzard) to grind food mechanically. Modern birds — the living descendants of theropod dinosaurs — swallow grit and small stones routinely. Crocodilians do the same.

Seismosaurus (Diplodocus hallorum), excavated in New Mexico in the 1980s, produced over 240 smooth, polished stones from inside its ribcage — the best direct evidence of gastrolithic digestion in a sauropod. The stones were physically distinctive from the surrounding rock: smoother, shinier, and made of a harder mineral than the sandstone matrix, indicating they had been transported to the site inside the animal rather than accumulated naturally.

The key diagnostic feature is polish. Stones tumbled in a muscular gizzard against each other and against food material develop a characteristic surface texture — a uniform, high-gloss polish — distinct from the rounded but dull finish of water-tumbled river stones. Gastroliths also tend to cluster in a location within the body cavity that corresponds to where a muscular gizzard would sit.

Not all smooth stones near dinosaur skeletons are gastroliths, and some early claims have been revised. The criterion of polish is critical — a smooth but unpolished stone could have been swallowed as ballast (as modern crocodilians sometimes do) or accumulated by geological processes. The confirmed gastrolith record is strongest for sauropods and some ornithopods.

Source: Gillette, D.D., New Mexico Museum of Natural History Science Bulletin (1994) — Detailed description of the Seismosaurus specimen and its gastroliths, establishing diagnostic criteria for identifying true gastroliths versus accidentally associated stones.

Coprolites: What Dinosaurs Left Behind

A coprolite is, bluntly, fossil excrement. And because excrement contains the undigested remnants of what an animal ate, coprolites provide dietary evidence that no other fossil type can match. The challenge is attribution: a coprolite rarely comes with a name tag identifying its producer. Paleontologists infer the producer from size, shape, preservation context, and the fauna known to have inhabited the same formation.

The most famous dinosaur coprolite — in terms of scientific impact — was described by Karen Chin and colleagues in a 1998 Nature paper. Found in Saskatchewan's Frenchman Formation (Late Cretaceous, approximately 68–66 Ma), the coprolite was enormous: approximately 44 cm long and 13 cm wide, making it the largest known coprolite from that formation. Its size, combined with the fact that Tyrannosaurus rex was the only theropod in the Frenchman Formation large enough to have produced it, made attribution to T. rex highly probable — though not provable with certainty.

Inside the coprolite: crushed bone fragments comprising approximately 30–50% of the sample volume. The bone was from a large herbivore (likely a hadrosaur or ceratopsian), and the crushing pattern indicated that the producer's jaw had been powerful enough to crack bone completely, extracting marrow. This confirmed what bite-force estimates already suggested: T. rex was a bone-processor, not merely a flesh-stripper. Its gut could handle bone as a food source.

Herbivore coprolites provide equally precise information. A titanosaur coprolite from the Late Cretaceous of India, described in 2005, contained grass phytoliths (microscopic silica structures found in grass cells) — the first direct evidence that at least some dinosaurs ate grass. This was scientifically significant because grasses were long thought to have diversified primarily after the end-Cretaceous extinction. The coprolite evidence pushed the presence of grass-eating in dinosaur ecosystems back into the Cretaceous.

Source: Chin, K. et al., Nature (1998) — Description and analysis of the large T. rex coprolite from Saskatchewan, providing direct evidence of bone-crushing behavior.

Source: Prasad, V. et al., Science (2005) — Titanosaur coprolite from India containing grass phytoliths, demonstrating grass consumption by dinosaurs in the Late Cretaceous and revising the dietary timeline.

Carnivores: Specialists in a Layered Ecosystem

Carnivorous dinosaurs were not a monolithic group. They ranged from insectivores the size of a crow to bone-crushing megapredators massing over 8 tonnes, and the ecological niches they occupied were as finely divided as those in any modern predator guild.

Large theropods — Tyrannosaurus, Allosaurus, Carcharodontosaurus, Giganotosaurus — were apex predators targeting large herbivores. Their dentition is blade-like, serrated, and often recurved: the serrations (denticles) are microscopic ridges running along the front and back edges of each tooth, functioning like a steak knife to slice through muscle and connective tissue. Bite marks on hadrosaur and ceratopsian bones from Cretaceous formations in North America have been matched to T. rex dentition with precise cross-sectional measurements.

Medium theropods — dromaeosaurids (the "raptors"), troodontids, and their relatives — occupied a different niche: fast, maneuverable pursuit predators of medium prey. Velociraptor (actual size: roughly turkey-sized, not the human-height movie version) preyed on animals like Protoceratops. The famous "fighting dinosaurs" specimen from Mongolia, preserving Velociraptor and Protoceratops locked in combat at the moment of death, is the most direct predation evidence in the dinosaur fossil record.

Spinosaurids specialized in fish and aquatic prey, as discussed in Spinosaurus: The Aquatic Dinosaur Bigger Than T. Rex. Their conical, slightly recurved teeth and isotope profiles distinguish them from purely terrestrial carnivores.

Small theropods competed for small prey: lizards, mammals, insects, eggs. Compsognathus — about the size of a turkey — was found with the skeleton of a small lizard inside its ribcage, providing direct gut-content evidence for small-prey predation.

Carnivore Group	Size Range	Primary Prey	Key Evidence
Tyrannosaurids	8–14 tonnes	Large hadrosaurs, ceratopsians	Bite marks, coprolites with bone
Allosaurs	2–6 tonnes	Sauropods, ornithopods	Bite marks on sauropod bone
Spinosaurids	3–20+ tonnes	Fish, aquatic prey	Tooth isotopes, fish bones in gut
Dromaeosaurids	10–200 kg	Medium vertebrates, eggs	Fighting fossils, gut contents
Compsognathids	1–15 kg	Lizards, insects, small mammals	Preserved gut contents

Source: Erickson, G.M. & Olson, K.H., Journal of Vertebrate Paleontology (1996) — Bite mark analysis on hadrosaur bone matched to T. rex dentition, with force calculations indicating bone-crushing capability.

Source: Norell, M.A. et al., American Museum Novitates (1994) — Description and interpretation of the "fighting dinosaurs" specimen from Mongolia preserving Velociraptor mongoliensis and Protoceratops andrewsi.

Were Any Dinosaurs True Omnivores?

The clean herbivore/carnivore division was not absolute. Several dinosaur groups show anatomical evidence of mixed diets, and a small number of taxa are classified as probable omnivores based on convergent lines of evidence.

Oviraptorosaurs present a puzzling case. Their beaked, toothless (in most species) skulls suggest neither a strict meat diet nor a standard plant diet. Gut content fossils from Sinocalliopteryx and isotope data from oviraptorosaur specimens have indicated fish, eggs, and plant material in the diet of different taxa within the group. Some oviraptorosaurs appear to have been opportunistic omnivores — eating whatever was available seasonally, much as modern corvids do.

Ornithomimosaurs — the "ostrich dinosaurs" like Gallimimus and Ornithomimus — had toothless beaks, long necks, and a body plan almost identical to modern ostriches. Ostriches are omnivores, eating plants, insects, and small vertebrates. Isotope analysis of ornithomimosaur specimens from Canada suggests a primarily plant-based diet in adults but a more omnivorous diet in juveniles — a pattern seen in some modern bird species that shift dietary strategy as they grow.

Heterodontosaurids are named for their heterodont teeth (teeth of different shapes in the same jaw) — a feature diagnostic of omnivory in mammals. They had chisel-like front teeth for cropping, large tusk-like caniniforms, and molariform cheek teeth for grinding. This dental combination is consistent with a mixed diet of plants, insects, and possibly small vertebrates.

Source: Zanno, L.E. & Makovicky, P.J., Proceedings of the Royal Society B (2011) — Systematic analysis of dietary indicators across herbivorous and omnivorous theropod dinosaurs, identifying oviraptorosaurs and ornithomimosaurs as likely omnivores.

Plant Evolution and the Dinosaur Feedback Loop

Dinosaurs did not exist in a static world of plants — they actively co-evolved with their food sources over 165 million years, and the relationship ran in both directions. Dinosaur feeding behavior was a powerful selective force on plant evolution, and plant diversity in turn shaped dinosaur diversity.

During the Triassic and early Jurassic (approximately 230–175 Ma), the dominant vegetation was conifers, cycads, ferns, and horsetails. The first sauropods evolved into a world where these low-diversity, tough-leaved plants dominated. Sauropod dental and digestive anatomy — strip-and-swallow with gizzard processing — is well-adapted to this type of food.

The mid-Jurassic and Cretaceous saw the rise of angiosperms (flowering plants), which diversified explosively from roughly 125 Ma onward. Flowering plants invested heavily in high-nutrient seeds and fruits to attract animal dispersers — and some paleontologists have argued that the diversification of ornithopods and ceratopsians in the Cretaceous tracks the diversification of angiosperms, with these dinosaurs evolving ever more sophisticated dental machinery to process the new, tougher angiosperm vegetation.

The grass-coprolite evidence mentioned earlier suggests that grass — today one of the most nutritionally productive plant families on Earth — was present and being eaten by dinosaurs in the latest Cretaceous, at least 66 Ma. Grass would go on to reshape the world's ecosystems after the extinction, but it had begun its ecological trajectory while dinosaurs still walked the Earth.

Source: Barrett, P.M. & Willis, K.J., Proceedings of the Royal Society B (2001) — Analysis of dinosaur-plant co-evolutionary dynamics, examining whether angiosperm rise correlates with ornithopod diversification.

Source: Butler, R.J. et al., Proceedings of the Royal Society B (2009) — Macroevolutionary study of herbivorous dinosaur diversity through time, tracking ecological diversification against plant record.

What People Get Wrong About Dinosaur Diets

The most pervasive error is treating all large theropods as equivalent meat-eaters. T. rex and Spinosaurus were both large theropods, but their diets were as different as a wolf's and a heron's. Tooth shape, body posture, habitat, and isotope chemistry all reveal finely divided dietary specializations within the "carnivore" category.

The second common mistake is assuming that large size always means large prey. Some very large theropods — including several ornithomimosaurs and therizinosaurids — had reduced or absent teeth and appear to have been predominantly herbivorous despite their substantial body size. Therizinosaurus had enormous hand claws (over 90 cm long) used not for killing prey but likely for pulling down tree branches to feed on leaves.

A third error concerns T. rex specifically: the debate over whether it was a pure predator or a habitual scavenger. The honest answer, supported by bite marks on bones that show healing (meaning the prey survived the attack and lived afterward), confirmed predation injuries on living animals, and the biomechanics of its forward-facing binocular vision, is that T. rex was primarily an active predator that also scavenged opportunistically — the same mixed strategy used by modern lions. For more on T. rex biology, see Tyrannosaurus Rex: The Science Behind the Legend.

For context on the ecological scale of these interactions across the full span of the Mesozoic, see The Three Periods of Dinosaurs.

The Takeaway

Dinosaur diets are not reconstructed from imagination — they are inferred from seven independent lines of physical and chemical evidence, each cross-checked against the others. The picture that emerges is ecologically sophisticated: a world where large sauropods strip-fed and fermented their food in bulk, hadrosaurs ground tough vegetation with dental batteries that rival anything in mammal evolution, apex theropods bone-crushed their prey for marrow, and fish-specialized spinosaurids occupied the rivers alongside giant sawfish. The boundary between herbivore and carnivore was genuine but not absolute. And the plants themselves evolved under 165 million years of dinosaur pressure — a feedback loop that shaped the terrestrial biosphere we inherited.

Frequently Asked Questions

Q: How do paleontologists know what a dinosaur ate?
A: Seven main lines of evidence: tooth morphology and wear patterns (SEM microwear), gastroliths (polished gizzard stones), coprolites (fossil dung with undigested food remains), preserved gut contents in exceptional fossils, bite marks on bone, carbon and nitrogen isotope ratios in enamel and bone collagen, and comparison with the anatomy of living relatives. No single line is decisive; convergence across multiple lines produces reliable conclusions.

Q: What percentage of dinosaurs were herbivores?
A: Estimates based on the known fossil record suggest approximately two-thirds of named dinosaur species were herbivores, consistent with the ratio seen in modern terrestrial ecosystems where plant biomass far exceeds animal biomass. However, herbivores are also overrepresented in the fossil record because large-bodied, bone-heavy animals fossilize more readily than small carnivores.

Q: Did T. rex eat plants?
A: No. T. rex had blade-like, serrated teeth up to 30 cm long — unambiguously adapted for slicing flesh. Its coprolites contain crushed bone. Its isotope ratios are consistent with high trophic-level carnivory. There is no credible evidence for plant consumption. The debate is not whether T. rex was a carnivore but whether it was primarily a predator or a scavenger (the evidence strongly favors active predation with opportunistic scavenging).

Q: What were gastroliths and which dinosaurs had them?
A: Gastroliths are stones swallowed deliberately and retained in a muscular gizzard to grind food mechanically. They are identified by their high-gloss polish (from being tumbled against other stones and food) and their clustering in the ribcage area. Confirmed gastroliths are most common in sauropods (including Seismosaurus/Diplodocus hallorum) and some ornithopods. The practice mirrors modern birds, which are direct descendants of theropod dinosaurs and routinely swallow grit.

Q: When did grasses evolve, and did dinosaurs eat them?
A: Grasses (Poaceae) evolved by at least 66–70 Ma in the latest Cretaceous, earlier than previously thought. A 2005 study of titanosaur coprolites from India found grass phytoliths directly inside the dung, proving that some large sauropods ate grass. Grasses were not the dominant vegetation of the Mesozoic — they became ecologically dominant only after the end-Cretaceous extinction — but they were present and consumed in the final phase of the dinosaur era.

Q: Were any dinosaurs omnivores?
A: Yes, though confirmed omnivory is difficult to prove from fossils alone. Oviraptorosaurs show gut contents and isotope profiles consistent with mixed diets of plants, eggs, and animal matter. Ornithomimosaurs appear to have shifted from omnivory as juveniles to herbivory as adults, based on isotope data. Heterodontosaurids had differentiated teeth (including canine-like structures) consistent with mixed diets. True omnivory — consuming both plant and animal matter regularly — was probably more common than the fossil record currently reflects.

Source: Therrien, F. & Henderson, D.M., Journal of Vertebrate Paleontology (2007) — Skull biomechanics analysis confirming T. rex as an active predator with bone-crushing capability, addressing the predator-vs-scavenger debate.

Source: Chin, K. et al., Nature (1998) — Saskatchewan T. rex coprolite with bone fragments, confirming bone-processing diet.

Source: Prasad, V. et al., Science (2005) — Grass phytoliths in titanosaur coprolite, revising Cretaceous plant ecology.

---

Read next:

• How Dinosaurs Laid Eggs and Nested
• Spinosaurus: The Aquatic Dinosaur Bigger Than T. Rex
• The Biggest Dinosaurs Ever Explained
• Tyrannosaurus Rex: The Science Behind the Legend
• The Three Periods of Dinosaurs