Neolithic farmers in Greece faced no clear winter cold snap, limpet shells suggest

Oxygen isotopes in 121 limpet shells from Franchthi, a coastal site in southern Greece, show Early Neolithic winters were not clearly colder than later ones, though one site cannot speak for the whole Mediterranean.

Peer reviewed · Single-site archaeological and geochemical study · Main text not accessible to SciQuest

Did a cold snap about 8,200 years ago stall the first farmers’ spread across the Mediterranean? At one Greek site, the answer looks like no. Niklas Hausmann and Danai Theodoraki of the Leibniz-Zentrum für Archäologie (LEIZA) in Mainz report in PNAS that shells collected at Franchthi record no winter cooling of the size proposed for 8,600 to 8,000 years ago, and no strong link between climate and what people ate. The catch: one site cannot settle a question about a whole sea.

Evidence at a glance

Is this a big deal?

A solid test of one idea at one place. It weakens the case that harsh winters held back farmers at Franchthi. The Mediterranean-wide claim needs records from more sites.

What scientists found

Bernhard Weninger and colleagues argued in 2014 that farming’s arrival in the Aegean, around 6600 BC, and its move onward, around 6050 BC, lined up with the start and end of a period of rapid climate change. That window includes the 8.2 ka event, a sharp cold spell set off by meltwater pouring into the North Atlantic. As the new paper’s abstract puts the hypothesis, extreme cold would have cut harvests and kept new settlements from spreading.

The Franchthi shells do not show that. Winter oxygen isotope values in the Early Neolithic had a median of 1.45 per mil, against 1.36 in the Middle Neolithic. Higher values mean colder water, but the gap is small: about 0.4 °C by the temperature equation in the authors’ code, or about 1 °C if you compare averages instead. A one-sided test gave p = 0.09, and a Bayes factor of 1.95 counts as weak evidence either way. Early Neolithic winters were also less variable, not more (standard deviation 0.29 versus 0.50 per mil).

How they found it

Franchthi sits on Kilada Bay in the Argolid, and people there collected limpets from the late Palaeolithic into the Neolithic. A limpet adds shell in bands, a bit like tree rings, and the oxygen in each band shifts with the water’s temperature as it grew. In earlier work at Franchthi, the same team used a laser method that reads magnesium in the bands to tell winter growth from summer growth.

The authors’ data hold 205 season-labelled oxygen measurements from 121 shells, from the Final Palaeolithic to the Final Neolithic: 51 shells from the Early Neolithic and 19 from the Middle Neolithic. The abstract calls it the largest dataset yet of high-resolution temperature archives in molluscs.

The analogy has a limit. A tree ring is dated to its year; a shell is dated only by the layer it was dropped in, and SciQuest could not read how the paper dates each layer.

Then came the test that matters for farmers. For 17 Neolithic excavation units, the authors compared summer and winter temperatures with counts of animal remains, from sheep, goats, cattle and pigs to deer, tuna and molluscs. Climate explained about 12% of the variation, and a permutation test gave p = 0.42, so no reliable link.

Why it matters to you

This won’t change anything in your daily life. It speaks to where we came from: whether early farmers were pushed around by the weather or coped with it. The same question runs through our story on how the Kikai-Akahoya eruption hit Jomon foragers unevenly.

What the evidence does not show

It does not show that the 8.2 ka event skipped Greece, and it does not test other sites. A spell lasting a century or two can hide inside a mixed archaeological layer. Shells record sea temperature in the bay, not frost on a field inland.

The Middle Neolithic sample is small, so a modest difference could be missed. Winter spread also differed between the periods (Brown–Forsythe test p = 0.01), and SciQuest could not read how the authors interpret that.

How this fits what we already knew

Earlier on-site evidence points both ways. At Çatalhöyük in Turkey, Mélanie Roffet-Salque and colleagues found the 8.2 ka event recorded in animal fats from cooking pots, alongside changes in the site’s animal bones (PNAS, 2018). Pascal Flohr and colleagues reviewed early farming sites in Southwest Asia and titled their paper “Evidence of resilience” to the 9.2 and 8.2 ka events (Quaternary Science Reviews, 2016). Sarah Parker and Sandy Harrison found the 8.2 ka event in more than 70% of global cave records (Scientific Reports, 2022), so the event itself is not in doubt.

Independent perspective

SciQuest found no comment from a researcher outside the author team as of 3 November 2026, 08:46 UTC, and has not interviewed one. The context above comes from earlier peer-reviewed work by other groups.

What happens next

The SEAFRONT project behind the study is studying shells from Grotta dell’Uzzo and Grotta d’Oriente in Sicily, Cyclops Cave in Greece and Haua Fteah in Libya. Matching records from those sites would show whether Franchthi is typical. Farming communities later reached central Europe, as the 6,500-year-old Chleby enclosure shows.

In one sentence

Franchthi’s limpets show its early farmers faced no clear winter cold snap and no climate-driven change in diet, a challenge to, not a refutation of, the idea that cold halted farming’s spread.

Read next

• Kikai-Akahoya eruption and Jomon foragers: disaster versus prehistoric people.
• An 11,000-year-old human-on-leopard sculpture at Karahantepe: the Neolithic in Anatolia.
• What tools were made in the Stone Age? The toolkit behind early farming.

Sources and disclosure

1. Hausmann N, Theodoraki D. “On-site, subannual climate records decouple Neolithic expansion from climate instability.” PNAS 123(45), published online 2 November 2026. doi:10.1073/pnas.2607859123. Abstract read via Crossref; funding: Deutsche Forschungsgemeinschaft award 439799406.
2. Hausmann N, Theodoraki D. Data and code for the paper, OSF project dgjrq (statistical output, supplementary δ¹⁸O table, Isotopes.xls).
3. Theodoraki D, Hausmann N. “Seasonality at prehistoric Franchthi: the use of large sample sets.” Archaeological and Anthropological Sciences (2026). doi:10.1007/s12520-026-02509-2.
4. Weninger B et al. “Neolithisation of the Aegean and Southeast Europe during the 6600–6000 calBC period of Rapid Climate Change.” Documenta Praehistorica 41 (2014). doi:10.4312/dp.41.1.
5. Roffet-Salque M et al. “Evidence for the impact of the 8.2-kyBP climate event on Near Eastern early farmers.” PNAS (2018). doi:10.1073/pnas.1803607115.
6. Flohr P et al. “Evidence of resilience to past climate change in Southwest Asia: Early farming communities and the 9.2 and 8.2 ka events.” Quaternary Science Reviews (2016). doi:10.1016/j.quascirev.2015.06.022.
7. Parker SE, Harrison SP. “The timing, duration and magnitude of the 8.2 ka event in global speleothem records.” Scientific Reports 12 (2022). doi:10.1038/s41598-022-14684-y.
8. LEIZA, Emmy Noether Group SEAFRONT project page.

Disclosure: This article was drafted with AI assistance and checked against the sources above. The PNAS main text was not accessible to SciQuest, so details come from the abstract and the authors’ public data, code and supplementary table; the shell type (limpets), the Kilada Bay setting and the laser season method come from the team’s companion paper (source 3). Medians, standard deviations and the 0.4 °C gap are SciQuest calculations from the authors’ Isotopes.xls file and the temperature equation in their code (T = 16.9 − 4.38δ + 0.1δ²); treat the conversion as approximate. The graphic is a SciQuest original built from those values. Spotted an error? See our corrections policy.