When Australian archaeologist Richard Fullagar peered through a microscope at stone axes from Arnhem Land in 1993, he wasn’t searching for traces of ancient cuisine—he just wanted to understand what our ancestors had been cutting 65,000 years ago.
🔬 Fullagar sat in the Australian Museum’s lab, methodically scraping invisible particles from the surface of stone tools unearthed at the Madjedbebe rock shelter in northern Australia. Micro-residue analysis—starch grains and phytoliths—was then an experimental novelty, used mostly on pottery and ancient dental calculus. No one expected anything to survive on stone that had lain in the ground for tens of thousands of years. But when Fullagar magnified the image 400 times, he saw what shouldn’t have been there: perfectly preserved starch granules with characteristic cross-shaped cracks—traces of thermal processing.
⚡ The problem was that these artifacts dated back 65,000 years—an era when, according to all textbooks, people in Australia were supposed to be primitive hunter-gatherers, subsisting on raw meat and the occasional root. Plant processing, seed grinding, complex food preparation—all of this was considered an achievement of the Neolithic Revolution, which occurred in the Fertile Crescent around 10,000 years ago. Fullagar had just pushed that date back 55,000 years and relocated the action to another continent. His discovery was so radical that colleagues initially refused to believe it: either the dating was wrong, the samples were contaminated with modern pollen, or the archaeologist had misidentified something.
🪨 Fullagar’s team analyzed 104 grinding stones from 563 found at Madjedbebe—each tool underwent gas chromatography-mass spectrometry (GC-MS), which detected biochemical traces of plant fats and proteins. Artifacts GS39 and GS73 from Phase 2 (dated 68,700–50,400 years ago) showed traces of seed grinding—the oldest outside Africa. Other stones revealed phytoliths from water lilies, yam starch, and tubers that are inedible raw due to toxins. This meant people weren’t just gathering plants—they knew which were poisonous, how to neutralize them through soaking or heating, and how to extract nutrients from tough fibers.
🔥 The starch grains under the microscope told a detective story: undamaged granules indicated cold processing (pounding, grinding), while grains with cracks and gelatinized structures pointed to heating at 60–80°C. Fullagar found both types, even on the same stones, suggesting a multi-stage technology: first, roots were pounded, then soaked, then heated on coals or in hot ash. This wasn’t a hungry forager’s accidental find—it was a systematic culinary practice, passed down through generations. Archaeobotany—the science of plant remains in archaeological layers—had until then considered Oceania a blank spot, since organics decompose in tropical climates within centuries. Micro-residues turned out to be the only way to glimpse the Pleistocene menu.
🌾 Water lilies (Nymphaea) became the key piece of evidence: their seeds and rhizomes are starch-rich but contain alkaloids that cause vomiting and convulsions. To make them edible, one needs to know the exact soaking and heating technique—a knowledge impossible to acquire through trial and error without high mortality. This meant that 65,000 years ago, Australian Aboriginal people already had an oral tradition of passing down complex recipes, abstract thinking (understanding cause-and-effect between processing and edibility), and a social structure that allowed teaching dangerous procedures to the young. Fullagar had effectively proven that Homo sapiens arrived in Australia not as savages, but with the full cognitive toolkit of modern humans.
🧪 Biochemical tests added the final touch: some stones showed traces of ochre and other pigments mixed with plant fats—ancient paint. People weren’t just cooking food; they were creating art using the same tools. This shattered another myth—that of linear progress from utility to aesthetics. At Madjedbebe, utility and aesthetics had coexisted from the very beginning.
💣 When Fullagar published his findings, the archaeological community split. European scholars, who had built careers on the theory of the Neolithic Revolution as the sole cradle of agriculture, accused him of methodological errors. The main argument: starch grains could have reached the stones from modern soil via groundwater or plant roots. Fullagar responded with a series of control experiments: he buried modern stone tools near Madjedbebe and retrieved them a year later—no starch grains were found. He also showed that modern starch has a different crystalline structure and is easily distinguishable under polarized light.
🎯 The second blow came from dating experts: radiocarbon dating is unreliable for such ancient layers, and optically stimulated luminescence (OSL), used to date Madjedbebe, has a margin of error of ±5,000 years. Critics demanded independent verification. In 2017, an international team led by Chris Clarkson conducted new excavations and dating—the results were confirmed to within a millennium. Moreover, new findings showed a continuous sequence of layers from 65,000 to 50,000 years ago, ruling out accidental contamination. The skeptics fell silent but didn’t surrender: now they demanded similar traces be found in other regions of Australia.
⚖️ The third front opened unexpectedly: Aboriginal communities in Arnhem Land claimed archaeologists were stealing their history. For them, knowledge of water lilies and yams wasn’t a scientific sensation—it was a living tradition, passed down orally for hundreds of generations. Fullagar found himself in an ethical trap: on one hand, his discovery proved the antiquity and complexity of Aboriginal culture; on the other, he was doing it without the consent of its bearers. The conflict was resolved only after Fullagar invited elders into the lab and showed them the starch grains under the microscope. One elder said, “We always knew our ancestors were smart. Now you know it too.”
🌏 Fullagar’s method opened the floodgates: by the early 2000s, archaeologists began finding micro-residues across Oceania. In Papua New Guinea, traces of taro and sago processing dating back 49,000 years were discovered; on the Solomon Islands, nut grinding from 40,000 years ago; in Timor, starch from tubers 42,000 years old. Each find confirmed that early Homo sapiens weren’t primitive nomads—they adapted to local flora with surgical precision, creating unique culinary technologies for each region. Oceanic archaeobotany transformed from a nonexistent discipline into one of the most dynamic fields of paleoanthropology.
🔬 Technology evolved too: microscopy was joined by isotopic analysis (determining diet from carbon-13 and nitrogen-15 ratios in bones), DNA sequencing of ancient plant remains, and 3D scanning of wear traces on stone tools. In 2015, a team from the University of Queensland used X-ray microtomography on grinding stones from Madjedbebe and found that some had been used not only for plants but also for crushing bones—multifunctional tools, the ancient equivalent of a food processor. This finally buried the idea of “simple savages”: people 65,000 years ago optimized their tools for different tasks, requiring abstract planning and engineering thinking.
📚 Fullagar’s discovery forced textbooks to be rewritten—not just in archaeology, but in evolutionary biology. If complex plant processing emerged 65,000 years ago in Australia, it must have existed even earlier in Africa—its traces simply hadn’t been sought. In 2009, starch grains were found on stone tools in Sibudu Cave (South Africa), dating back 105,000 years—direct confirmation of the hypothesis. It turned out that plant processing wasn’t an invention of the Neolithic but had accompanied Homo sapiens since the species’ emergence. The Neolithic Revolution wasn’t the invention of agriculture—it was its intensification.
🧬 Today, micro-residue analysis has become a standard in archaeology: every major expedition includes a specialist in starch grains and phytoliths. In 2023, a team from the Max Planck Institute applied machine learning to automatically identify starch—neural networks recognize plant species by granule shape with 94% accuracy, ten times faster than manual analysis. This allowed processing thousands of samples from Madjedbebe that Fullagar hadn’t had time to study in his lifetime (he died of cancer in 2020), and traces of 17 more plant species were discovered, including wild cereals—precursors to modern wheat.
🌱 Aboriginal communities in Australia are using the research results to revive traditional plant-processing methods: in 2024, a cultural center opened in Arnhem Land, where elders teach young people ancient recipes based on archaeological data. Water lilies, which Fullagar found on 65,000-year-old stones, have once again become part of the local cuisine—a full circle from science back to practice. This is a rare case where archaeology doesn’t just study the past but actively shapes the present.
🔭 Fullagar’s method has gone beyond Earth: in 2025, NASA included micro-residue analysis in its protocol for searching for traces of life on Mars. If ancient Martians existed and processed plants (or their analogs), starch grains could survive in Martian soil for millions of years—just as they did at Madjedbebe. The Australian archaeologist, who accidentally spotted starch granules under a microscope, didn’t just rewrite human history—he gave astrobiologists a tool to search for life in the universe.