The Women of Teviec: Ancient DNA and the Last Hunter-Gatherers of Brittany

Hero image: AI generated illustration of two Late Mesolithic women of the Breton coast, based on the ornaments found at Téviec and on the pigmentation predicted from the Hoëdic and Téviec genomes. It is not a facial reconstruction of any individual.

In the summer of 1928, Marthe and Saint-Just Péquart, a couple of amateur archaeologists from Nancy, began digging into a mound of shells on Téviec, a rocky islet of a few hectares off the Quiberon peninsula in Morbihan. Under the debris of countless meals of limpets, mussels and oysters they found the dead. The first grave they opened held two women, lying together, adorned with shell necklaces, bracelets and anklets, under a structure of red deer antlers.

The grave was lifted in one block and sent to the Muséum de Toulouse, where it still lies today. The two women of Téviec became one of the most reproduced images of European prehistory: the tender double burial, the antler crown, the shells. They are also, as later study suggested, possibly two victims of violence.

In 2024, a team led by Luciana G. Simões and Mattias Jakobsson at Uppsala University, working with Grégor Marchand and Amélie Vialet in France, published complete genomes of ten of the last hunter-gatherers of Atlantic France. Many readers have asked us whether the two women of Téviec are among them. The short answer is no. The longer answer is more interesting, because the genomes that were sequenced include several women and girls from the same cemeteries, and they tell a story about how these island communities found partners, raised children and kept farmers at a distance.

Islands of the dead

Téviec and its neighbour Hoëdic, a small island further out to the southeast, are two of the very few Mesolithic cemeteries known on the Atlantic coast of Europe. The Péquarts excavated both in a few seasons. At Téviec, 10 graves held at least 23 people, adults and children together, packed into about 50 square metres among hearths and food waste. Some graves were opened again and again over generations, the earlier dead pushed aside for new ones.

The bodies were placed in a flexed position, often sprinkled with red ochre, with flint blades, bone points and boar tusks. Twice at Téviec, a frame of red deer antlers was built around the head of the dead. Small hearths, perhaps for funeral meals, were lit on stone slabs above some graves.

The ornaments are the most striking part of the record. Solange Rigaud, who restudied the shells, found that they were dominated by two kinds of small sea snail: periwinkles of the genus Littorina, tiny shells in yellow, orange, pink and red, and cowrie shaped Trivia, nicknamed "coffee beans". Each was carefully perforated. They were found around the skull, the neck, the wrists and the ankles, and some were probably sewn onto a net or a headdress rather than strung. With them came pendants of perforated red deer teeth and long bone pins, made from boar, dog or deer bone, some finely engraved. Éva David showed from wear marks that the pins had been used in life, probably to fasten clothing or hair. These were people who cared a great deal about how they looked.

The two women of grave A

The double burial in Toulouse holds two women aged roughly 25 to 35. When the skeletons were re-examined around the time of their restoration and new exhibition in 2010, anthropologists reported multiple impacts on the skulls. In the reading summarised by Grégor Marchand, both women appear to have been killed with blunt objects. One of them had several blows to the head, at least two of which could have been fatal. Not everyone accepts this: some specialists argue that part of the damage could come from the weight of the overlying midden pressing on fragile bones over seven millennia.

Violence is not an isolated case at Téviec. In grave K, a man the Péquarts called the "great hunter" had two flint arrowheads lodged in his vertebrae, most likely the cause of his death, and a broken jaw that had healed long before. Whoever these people were, the coast was not always peaceful.

Grave A is also the one that the public knows best, thanks to the antler crown and the photographs of 1928. It is easy to forget that it is only one grave out of ten, and that its occupants have never been sampled for DNA.

Do we have their DNA?

Not yet. The Téviec and Hoëdic collections were divided long ago. The two women of grave A went to Toulouse in 1937. Most of the other skeletons are kept at the Institut de paléontologie humaine in Paris, and that is where the Uppsala team took their samples. The ten genomes published in 2024 come from Téviec grave K, from four graves at Hoëdic, and from Champigny near Paris, an older inland site used as a comparison.

Among those ten people are five females: one woman from Téviec, three from Hoëdic and a girl of three to seven buried at Hoëdic. They are the closest we can get today to the women of grave A: buried in the same way, wearing the same kind of shells, on the same coast, within a few generations of them.

What the genomes say

All ten people belong to the population that geneticists call Western Hunter-Gatherers, the descendants of the groups that spread across western Europe at the end of the Ice Age. Every maternal line is a branch of haplogroup U5, the classic lineage of European foragers. Every man whose Y chromosome could be read carries I2a1. The pigmentation genes point to the combination already known from Cheddar Man and Loschbour: most of these people had dark skin and blue eyes. Two of them, hoe003 and the girl hoe005, were predicted to have had paler, intermediate skin.

The most surprising results concern families. These communities were small, yet the genomes show no sign of inbreeding. No pair of first degree relatives was found anywhere in the sample. People buried in the same grave were often unrelated or only distantly related. In Téviec grave K, the three individuals tested were second or third degree relatives, and the man at the bottom of the grave, most likely the "great hunter", was more closely linked to at least two of the people buried above him than they were to each other. At Hoëdic, the woman and the girl of grave J, laid to rest together, were not related at all.

As Amélie Vialet put it, there were "strong social bonds that had nothing to do with biological kinship". The authors interpret the pattern as the trace of a network of distinct social units that exchanged partners. Each group stayed small, but marriage partners came from outside, which kept relatedness low over the generations. Isotopes add a detail: the Téviec and Hoëdic people ate differently, with more seafood at Hoëdic, as if two neighbouring communities with their own habits lived a few hours apart by boat.

No farmers in the family

This is where the genomes overturned an older idea. In 2001, Rick Schulting and Michael Richards measured carbon and nitrogen isotopes in the Téviec and Hoëdic skeletons and noticed that some women had a more terrestrial diet than the men they were buried with. They proposed that these women had grown up elsewhere, possibly among the first farming communities arriving in western France, and had married into the coastal groups. It was an elegant hypothesis, and it shaped the debate for twenty years.

The genomes do not support it. Simões and colleagues found no genetic ancestry from Neolithic farmers in any of the analysed hunter-gatherers, including those whose dates overlap with the first farmers in neighbouring regions. In a commentary in the same issue of PNAS, Peter Rowley-Conwy concluded that the foragers of Hoëdic and Téviec "showed no signs of adopting agriculture or inter-marrying with farmers". Partners were exchanged, but only between hunter-gatherer groups.

We ran the same test with Global25 coordinates. Each genome was modelled as a mixture of a hunter-gatherer source (Loschbour, Luxembourg) and an early farmer source (Neolithic Anatolia). To know what "zero" looks like in this kind of model, we first ran it on twelve hunter-gatherer genomes from England, Belgium, Germany, Luxembourg and Doggerland that are older than any farming in western Europe and cannot contain farmer ancestry. They come out between 1.7 and 6.2%, with a median of 3.5%: that is the background noise of the method.

Every Téviec and Hoëdic genome falls at or below that background, from 0% for the man of Hoëdic grave D to 2.4% for the woman of grave J and the boy of grave C. When we added 5% of Anatolian farmer ancestry artificially to each genome, the model read it back as 4.8 to 7.3%, and a 10% addition as 9.8 to 12.2%. A woman born among farmers would carry about 50%, and her child about 25%. Our honest detection floor is about 5%. Below it, Global25 cannot separate a trace from noise, but nothing in the data suggests that a trace is there. This agrees with the published genome wide result.

Two outsiders in grave J?

If the women of Téviec and Hoëdic did not come from farming villages, where did they come from? The genomes cannot give a birthplace, but they can show whether some individuals are genetically closer to other hunter-gatherer groups.

To test this, we placed each genome on a single axis. At one end (value 0) is the average of the Mesolithic hunter-gatherers of Nouvelle-Aquitaine, on the Atlantic side of France. At the other end (value 1) is the man buried at Champigny, near Paris, around 6200 BC, representing the inland hunter-gatherers of northern France.

Six of the seven island genomes cluster at the Atlantic end, between 0.08 and 0.24, among them all the men and the low coverage women of Téviec K and Hoëdic C. Two individuals stand apart: the woman of Hoëdic grave J (hoe004, 0.80) and the girl buried with her (hoe005, 0.76). Both genomes have high coverage in the Global25 set (96% and 99%), and they are not related to each other, so this is not one family accident repeated twice. They sit with the inland hunter-gatherers of northern France and Belgium, in the same zone as Malonne in Belgium (0.93) and Ranchot in the Jura (1.01).

This fits a detail of the isotope data. The Uppsala team noted that the woman and the girl of grave J had a diet that differed from the Hoëdic average and looked more like that of Téviec, and suggested they may have come from elsewhere. Our axis adds a genetic hint in the same direction, with a caution: the two poles are only 0.040 apart, single genomes are noisy, and a spike in test shows that it takes roughly 75% of Champigny-like ancestry to move a Téviec genome to 0.80. We read this as a real difference in background, not as a precise percentage. These two burials may represent exactly the kind of partner exchange the genomes reveal, reaching beyond the islands towards the interior.

Not Iberians, despite the shells

The shell midden cemeteries of Brittany are often compared with those of the Muge valley in Portugal, where hunter-gatherers buried their dead in similar mounds in the same millennium. The resemblance is cultural. Genetically, the two worlds are far apart. The average of the eight Téviec and Hoëdic genomes is 0.021 from the hunter-gatherers of Nouvelle-Aquitaine and Cheddar Man, 0.024 from Loschbour and 0.025 from Mesolithic Ireland, but 0.090 from Iberian hunter-gatherers and 0.127 from the Geometric Mesolithic of Portugal. Iberian foragers carry a large share of older Magdalenian ancestry that the Breton islanders lack. Téviec and Hoëdic belong to the same northwestern population as Britain, Ireland and Luxembourg.

Is their ancestry still in Bretons?

A natural question for readers in Brittany: do Bretons today carry the blood of the women of Téviec? Let us look at it with the standard three source model of European ancestry: hunter-gatherers (here the Téviec and Hoëdic average), Anatolian farmers, and steppe herders of the Yamnaya culture.

Modern Bretons come out at about 13% hunter-gatherer ancestry, 41% Anatolian farmer and 46% steppe. That is exactly the level of the Normans, the Cornish, the Welsh and the Irish, and only slightly above Parisians (11%). There is no special Breton reservoir of Mesolithic ancestry. The French Basques, at 18%, carry more.

Two details matter here. First, most of this ancestry did not come down directly from the islanders. The Middle Neolithic farmers of northwestern France already carried 15 to 19% hunter-gatherer ancestry, absorbed in the centuries after farming arrived, so the forager share of modern Bretons largely came through farming populations. Second, and this is a limit of the method that we prefer to show rather than hide: Global25 cannot tell which hunter-gatherers contributed. When we replace the Téviec and Hoëdic average with Loschbour from Luxembourg, the results change by less than half a point and the fit is the same. When we put both sources in the same model, the condition number jumps from 2.4 to 42 and the model gives all the hunter-gatherer share to one or the other, depending on the population, a classic collinearity trap. The two sources are only 0.024 apart. Saying that Bretons descend "from Téviec" rather than from any other western forager group would be reading noise.

What they looked like

Popular images of the women of Téviec now circulate widely, many of them AI generated. Some are attractive, but several details do not match the evidence. The ornaments of Téviec were not large scallop shells: they were hundreds of tiny perforated sea snails, Littorina periwinkles in warm colours and Trivia cowries, worn as necklaces, bracelets and anklets, and probably sewn onto nets or caps. The dead also wore pendants of red deer teeth and engraved bone pins. There is no evidence for woven plant fibre dresses at Téviec; hide and fur clothing are a safer assumption for the Atlantic coast in the sixth millennium BC.

As for the faces, the closest guide is the genomes of their neighbours: dark skin and blue eyes in most individuals, with a paler, intermediate skin predicted for two of the Hoëdic people. Our own illustration at the top of this article follows those rules, and it remains an illustration. No facial reconstruction of the two women of grave A based on DNA exists, because their DNA has not been studied.

What is still missing

The obvious next step is to sample the two women of grave A in Toulouse. A genome from each of them would tell us whether they were related, sisters, mother and daughter, or strangers buried together like the woman and the girl of Hoëdic J. It would also tell us whether they belonged to the Atlantic group of the island men, or to the inland profile of grave J. The site itself is today private property and closed to researchers, which makes the collections in Toulouse and Paris all the more precious.

Until then, what the genomes do say is already remarkable. In the last centuries before farming swept over Brittany, the people of Téviec and Hoëdic lived in small, independent communities that ate differently, buried strangers together as kin, and found their partners in other hunter-gatherer groups, not in the farming villages that were appearing on the horizon. When those villages arrived, the islanders' descendants were absorbed into them. A little of their ancestry, roughly one eighth, still runs in the people of Atlantic Europe today, in Brittany as much as anywhere else.

Methods

We used Global25 scaled coordinates by Davidski and the Moriopoulos 2026 collection, merged with first occurrence priority, and the individual dated ancient samples file for the nine Late Mesolithic genomes available in Global25 (tev001, tev002, tev003, hoe001 to hoe005, stp001; hoe006 is not included). Coverage in the Global25 set ranges from 18% (hoe001) and 20% (tev002) to 99.9% (tev003); results for the two low coverage women are indicative only. Distances are Euclidean in the 25 dimensional space. Admixture models use non negative least squares with a sum to one constraint enforced by a heavily weighted extra row (weight 1000). The farmer test uses Luxembourg_Loschbour and Turkey_N as sources (separation 0.495, condition number 1.8). Its background was estimated on twelve pre-farming genomes from England, Belgium, Germany, Luxembourg and Doggerland with at least 40% coverage (two Doggerland samples flagged as contaminated were excluded), and its detection floor by spike in tests adding 2, 3, 5 and 10% of Turkey_N to each genome. The Atlantic to inland axis is the projection onto the vector from France_Mesolithic_Nouvelle-Aquitaine (n=3) to France_Mesolithic_Champigny (n=1); the poles are 0.040 apart. The three source model uses the Téviec and Hoëdic average (8 genomes), Turkey_N and Russia_Samara_EBA_Yamnaya (separations 0.49, 0.41 and 0.34; condition number 2.4; fit distances 0.027 to 0.050). With Loschbour as the hunter-gatherer source, all results change by less than 0.5 points. A four source model with both hunter-gatherer sources has a condition number of 42 and is not interpretable. Sex, haplogroups, kinship, dates, isotopes and phenotypes are taken from Simões et al. (2024) and are not recomputed here.

Global25 coordinates used in this article

The nine Late Mesolithic genomes, their average, the model sources and modern Bretons. They can be pasted directly into the Vahaduo tool.

tev001_Teviec_K_male,0.116100,0.113739,0.196103,0.191217,0.163107,0.057451,0.014101,0.043844,0.102466,0.023326,-0.020461,-0.016935,0.015907,-0.009221,0.063517,0.063643,0.003651,0.011022,-0.010684,0.059654,0.114049,0.009398,-0.053490,-0.187376,0.022154
tev003_Teviec_K_male,0.119514,0.108662,0.193086,0.192186,0.165262,0.054663,0.018331,0.032076,0.099808,0.015490,-0.018999,-0.013638,0.013677,-0.006881,0.062974,0.065499,0.001043,0.014062,-0.008045,0.056652,0.114548,0.010510,-0.054229,-0.187376,0.018920
tev002_Teviec_K_female,0.130897,0.104600,0.188183,0.201230,0.166185,0.064145,0.006815,0.027922,0.096331,0.020410,-0.019324,-0.019633,0.011150,-0.012661,0.062974,0.065897,0.003651,0.010642,-0.010056,0.061154,0.117792,0.007666,-0.055954,-0.176410,0.020956
hoe001_Hoedic_C_female,0.113823,0.105615,0.193086,0.196385,0.145258,0.062193,0.009400,0.035768,0.097967,0.023691,-0.015914,-0.006444,0.021556,-0.005918,0.068403,0.068151,0.008345,0.010388,-0.011313,0.054651,0.119165,0.017064,-0.058050,-0.182315,0.027183
hoe002_Hoedic_C_child_male,0.126344,0.109677,0.191954,0.194124,0.163415,0.052431,0.016686,0.037383,0.093263,0.012574,-0.016076,-0.012289,0.016055,-0.012524,0.060396,0.065897,0.004694,0.008995,-0.016718,0.057152,0.112801,0.017064,-0.053983,-0.173760,0.025626
hoe003_Hoedic_D_male,0.122929,0.106631,0.199874,0.196062,0.161876,0.054663,0.023266,0.038768,0.098990,0.011663,-0.017213,-0.016186,0.011744,0.000688,0.062160,0.065632,0.005867,0.005954,-0.010936,0.061780,0.116794,0.012242,-0.057680,-0.187737,0.025507
hoe004_Hoedic_J_female,0.128620,0.118817,0.192709,0.198000,0.162492,0.055778,0.022796,0.043614,0.095308,0.022597,-0.018512,-0.016935,0.018583,0.005367,0.065553,0.047202,-0.025946,0.014442,-0.007542,0.054901,0.106188,0.022876,-0.053859,-0.176049,0.022034
hoe005_Hoedic_J_child_female,0.120652,0.113739,0.196480,0.191540,0.167416,0.059404,0.016451,0.045921,0.097967,0.012757,-0.015265,-0.018434,0.018880,-0.000688,0.068674,0.045080,-0.025686,0.012542,-0.006536,0.053526,0.121660,0.014591,-0.054845,-0.182435,0.025267
stp001_Champigny_male,0.125205,0.108662,0.190069,0.198646,0.161569,0.066655,0.019506,0.031383,0.100012,0.005832,-0.018675,-0.014387,0.012487,-0.006881,0.066639,0.045346,-0.025946,0.011402,-0.007919,0.055402,0.108059,0.008285,-0.060145,-0.179182,0.023710
Teviec_Hoedic_avg_8ind,0.122360,0.110185,0.193934,0.195093,0.161876,0.057591,0.015981,0.038162,0.097763,0.017813,-0.017721,-0.015062,0.015944,-0.005230,0.064331,0.060875,-0.003048,0.011006,-0.010229,0.057434,0.115375,0.013926,-0.055261,-0.181682,0.023456
France_Mesolithic_Nouvelle-Aquitaine,0.120652,0.109000,0.189315,0.195416,0.164851,0.053082,0.019349,0.038306,0.097217,0.016037,-0.009743,-0.021831,0.012240,0.000275,0.057048,0.062847,-0.002173,0.007981,-0.005405,0.066073,0.119414,0.010716,-0.053654,-0.184685,0.026105
Luxembourg_Loschbour,0.130897,0.109677,0.203645,0.198000,0.162492,0.059125,0.015041,0.038075,0.100217,0.016219,-0.015427,-0.017235,0.019921,-0.001239,0.061346,0.070670,0.002608,0.007348,-0.008925,0.065406,0.117543,0.010387,-0.049422,-0.173639,0.019519
Turkey_N,0.117902,0.180087,0.003426,-0.101059,0.051240,-0.047969,-0.003799,-0.006846,0.036167,0.080678,0.008261,0.011309,-0.024164,0.000579,-0.042712,-0.010370,0.022556,0.001388,0.013649,-0.010448,-0.014261,0.005693,-0.004904,-0.003750,-0.004436
Russia_Samara_EBA_Yamnaya,0.125838,0.089254,0.042908,0.115456,-0.027868,0.044685,0.004491,-0.002949,-0.054858,-0.072996,0.001858,0.000350,-0.001652,-0.023610,0.037263,0.015734,0.000000,-0.001478,-0.001704,0.012506,-0.003120,0.001374,0.011229,0.018436,-0.004524
French_Brittany,0.131466,0.138340,0.057850,0.039285,0.039169,0.015813,0.003155,0.004211,0.008186,0.009490,-0.005131,0.005590,-0.014368,-0.013102,0.018343,0.004399,-0.007302,0.001280,-0.000462,0.001773,0.004467,0.001787,-0.001254,0.010547,0.000659

References

1. Ancient DNA Simões LG, Peyroteo-Stjerna R, Marchand G, Bernhardsson C, Vialet A, Chetty D, Alaçamli E, Edlund H, Bouquin D, Dina C, Garmond N, Günther T, Jakobsson M (2024). Genomic ancestry and social dynamics of the last hunter-gatherers of Atlantic France. PNAS 121: e2310545121. doi.org/10.1073/pnas.2310545121
2. Commentary Rowley-Conwy P (2024). Hunter-gatherers and earliest farmers in western Europe. PNAS 121: e2322683121. pmc.ncbi.nlm.nih.gov/articles/PMC10927556
3. Press Uppsala University (2024). Stone Age strategy for avoiding inbreeding. uu.se
4. Isotopes Schulting RJ, Richards MP (2001). Dating women and becoming farmers: new palaeodietary and AMS dating evidence from the Breton Mesolithic cemeteries of Téviec and Hoëdic. Journal of Anthropological Archaeology 20(3): 314 to 344. doi.org/10.1006/jaar.2000.0377
5. Excavation A Mesolithic site in Brittany (1937). Nature 140: 329 to 331, on the excavations of M. and S.-J. Péquart and the study of the skeletons by M. Boule and H. Vallois. nature.com/articles/140329a0
6. Archaeology Marchand G (2016). Nécropole mésolithique de Téviec. Bécédia, Bretagne Culture Diversité. bcd.bzh
7. Ornaments Mesolithic engraved bone pins: the art of fashion at Téviec (Morbihan, France). Meso 2015 conference proceedings, chapter 93. shs.hal.science/halshs-03216415
8. Ornaments Espace des sciences. Des bijoux à la mode de Téviec. Sciences Ouest 319. espace-sciences.org
9. Museum Téviec, overview of the site and the double burial kept at the Muséum de Toulouse. en.wikipedia.org/wiki/Téviec
10. Data Global25 PCA coordinates by Davidski, and the Moriopoulos 2026 collection of Global25 coordinates.