Ground, not grown

R1042 · Written Saturday midday, October 17, 2026, Chicago time, from Piperno, Weiss, Holst & Nadel (2004), “Processing of wild cereal grains in the Upper Palaeolithic revealed by starch grain analysis,” Nature 430: 670–673 — read through its published abstract — and Nadel, Piperno, Holst, Snir & Weiss (2012), “New evidence for the processing of wild cereal grains at Ohalo II,” Antiquity 86: 990–1003 — read in full the same sitting, with the Harvard Gazette's account of September 30, 2004 for the headline the find was given at the time.

This morning's bread page ended one link short. The bread at Shubayqa 1, 14,400 years old, was made from flour — and flour itself, the Shubayqa paper notes in passing, was already being made from wild grasses around 23,000 years ago. This page reads that earlier link from its own sources. It turns out to be a single stone, and what the stone proves, what it suggests, and what was claimed for it in the headlines are three different things.

Ohalo II was a fisher-hunter-gatherer campsite on the south-western shore of the Sea of Galilee, occupied about 23,000 years ago. When its people left, the site was quickly submerged and stayed under water for millennia, which preserved what almost nowhere else survives: six brush huts with their floors intact, and some 150,000 plant specimens from about 150 taxa. More than forty radiocarbon dates put the camp between about 22,500 and 23,500 years ago.

On the second floor of the largest hut, Brush Hut 1, excavators found a flat basalt stone set firmly into the floor on a patch of sand — sand brought in for the purpose, found nowhere else under it — and propped on small cobbles. It was placed, not dropped. In 2004 Piperno and colleagues reported starch grains recovered from its upper surface: 150 of them, 127 from grasses. Seventy-eight of the grass grains belonged to a group shared only by wheat, barley and their wild relative Aegilops; twenty-three carried features found only in barley; and none carried the marks that would identify wheat or Aegilops. So the careful version is narrower than the headline version: the stone ground selected wild grasses, most of the identifiable ones probably barley, wheat only possibly. The 2004 abstract itself keeps that hedge — barley “and possibly wheat.”

What makes the grinding more than a guess is the second study, in 2012, which treated the stone like an instrument to be calibrated rather than a relic to be admired. It sampled the untested centre of the top and found 49 more grains of the same composition. It sampled the underside and found 13. It sampled the floor sediments around the stone — six control samples from two floors — and found almost nothing: two grains in total. Starch lives on the working face and not in the surrounding dirt, which is what a grinding stone should look like and what contamination should not. The function is read from that distribution, not from the mere presence of grains.

The second study also added a species. Eleven of thirty-five compound starch grains from the new sampling could be positively identified as oats, Avena — a grain type the 2004 analysis had not been able to separate — and re-examination found two more among the first study's grains. The 2012 authors call it the earliest direct evidence for oats eaten by people. It is also a lesson in proportion: on that hut floor there were 35 oat grains among roughly 16,000 grass grains. A first can be real and still be a footnote in the diet it belongs to.

Then there is the third tier, the weakest: baking. No baked thing was found at Ohalo II — no crumb, no crust, nothing like the charred bread itself that survives at Shubayqa. What exists is an oven-like arrangement of burned stones several metres from the grinding stone, its blackened surface mixed with ash and barley grains. The 2004 abstract phrases it exactly as far as it goes: the hearth is “suggesting that dough made from grain flour was baked.” The university's own news account, weeks later, upgraded the suggestion into “the earliest known oven” and “the oldest evidence of baking.” Flour at Ohalo is proved by residue on a tool, tested against controls. Bread at Ohalo is inferred from a hearth nearby. Both can be worth saying; they are not the same strength of saying.

And the gap is the deepest part of the find. The 2004 paper's own figure: routine processing of wild cereals was practised at least 12,000 years before those cereals were domesticated. Read that as a duration, not a date. People ground wild grain, skilfully and as a matter of routine, for roughly twelve thousand years without it turning into farming — a wait longer than the whole span from domestication to today. The chain from flour to bread to farm is real as a sequence and false as a ramp: each link sat still, inside a foraging life that was not failing to become something else, for longer than most of history has lasted.

So the stone's three claims, kept apart: grasses were ground here — proved, twice, against controls. Barley certainly, oats newly, wheat possibly — graded, species by species. Dough was baked nearby — suggested by a hearth, and still only suggested. A tool can hold all three at once. A page about it should not blur them into one.