A menu needs a place and a date
“Ancient humans” sounds like a group you could invite to dinner. It covers different species living in different environments over an enormous span of time. Asking what they ate without specifying a place and period is rather like asking what everybody in the world had for lunch.
Specific finds let us narrow the question. Cut marks can establish butchery, starch on a tool can reveal plant processing, and particles trapped in a tooth can record contact with particular foods. These clues answer different questions. Combining them can reconstruct aspects of a diet, but none supplies a complete shopping list or a universal ancestral ratio.
Teeth preserve several different clues
A 2023 review of dietary methods explains why researchers combine evidence. Tooth shape suggests what an animal could process. Microscopic wear records contact with food and abrasives, often over a much shorter interval. Calculus—the hardened material formed from dental plaque—can trap plant particles, although not everything entering a mouth was food.
Chemical evidence answers another question. Research on Turkana Basin fossils used carbon in tooth enamel to distinguish resources originating in plants with different photosynthetic pathways, called C3 and C4. Different hominin groups showed different patterns.
Those labels are not shorthand for “vegetarian” and “carnivore.” A plant’s carbon signature can reach a consumer through an animal that ate the plant. The measurements help trace food resources through an ecosystem; they cannot simply be converted into percentages of meat and vegetables.
Cut bones make meat visible
The Smithsonian’s account of tools and food describes evidence of animal butchery by at least 2.6 million years ago. Researchers compare ancient cut marks with marks produced in experiments using stone tools. The comparison connects a trace on a bone to a plausible action.
Butchery does not automatically establish how an animal died. A carcass could have been hunted or scavenged, and distinguishing those routes needs further evidence. Nor does one butchered animal establish that meat supplied most of a group’s meals.
An excavation photograph full of bones can therefore be misleading as a menu illustration. It shows the remains recovered at that site. Turning those remains into a calorie chart requires assumptions about the foods and activities that left no comparable trace.
Plants could leave evidence on the tools
Roots and seeds need not survive as recognizable food for their processing to be detected. Anna Revedin and colleagues examined grinding tools from Bilancino II in Italy, Kostenki 16 in Russia, and Pavlov VI in the Czech Republic. They reported starch grains from wild plants on tool surfaces at sites dating to roughly 30,000 years ago.
The findings suggest plant processing, including possible flour production, across these different settings. They do not turn the people using the tools into farmers. Collecting wild plants and grinding them is different from growing domesticated crops.
That distinction changes the familiar meat-versus-plants argument. A plant food can require deliberate preparation without agriculture. The important question is what happened between collecting the ingredient and eating it. We do not need to invent a recognizable loaf of bread to appreciate what these tools reveal.
Neanderthal plaque preserved plants too
Dental calculus from Neanderthals at Shanidar in Iraq and Spy in Belgium contained plant microremains. Amanda Henry and colleagues identified starches and phytoliths, microscopic structures produced by plants, including evidence associated with grass seeds, legumes, and date palms across the samples. Some grass-seed starch showed damage characteristic of cooking.
This supports plant consumption and preparation in the sampled individuals. It does not mean every Neanderthal ate all those foods, or that a few trapped particles measure the importance of plants relative to meat. The illustration shows the principle; it is not a picture of the specimens.
Processing changed what a meal required
Ingredients alone miss part of the story. A piece of meat and the same meat sliced into smaller pieces are different chewing tasks. So are an intact underground plant organ and one that has been pounded.
Experiments by Katherine Zink and Daniel Lieberman tested meat and underground storage organs in human participants. Slicing meat and pounding plant foods improved aspects of chewing performance, supporting a role for mechanical processing in early human diets.
This is experimental evidence about a mechanism, rather than a recovered prehistoric meal. It helps explain why a stone tool could matter beyond obtaining food: preparation changes the work needed to eat it. It does not establish that an ancient population followed the particular dietary proportions used in the researchers’ calculations.
Fish teeth offer a cooking example
At Gesher Benot Ya’aqov in Israel, researchers studied freshwater fish remains from archaeological horizons about 780,000 years old. Fish teeth were associated with clusters of burnt flint, and their mineral structure indicated exposure to relatively low heat. Together with the context and the unusual pattern of preserved fish remains, the authors interpreted this as evidence of fish being cooked and eaten.
The argument depends on several clues agreeing. A heated object alone would not tell the whole story, and the paper explicitly acknowledges that the first beginning of intentional cooking remains unknown. This site supplies an early example, not a date on which every human population acquired a kitchen.
One ancestral ratio would erase the differences
Work in the Malawi Rift illustrates why location matters. Isotope evidence for early Homo and Paranthropus in wooded settings showed substantial C3-derived resources, differing from patterns in other African environments. The authors interpreted the wider evidence as dietary flexibility across habitats.
These populations cannot be collapsed into one ancestral food chart. Nor can a much later European grinding stone stand in for all earlier African meals. Dates, environments, species, and methods are part of the answer, not inconvenient details to remove.
Our article on prehistoric hunting and gathering roles addresses another version of the same mistake: turning varied subsistence work into a universal rule. The useful habit is to ask whose food, where, when, and demonstrated by what evidence. Archaeology can answer a surprising amount once the question becomes specific.
