Population Replacement Evidence

What a population replacement looks like, and what we actually see in Chalcolithic–Early Bronze Age Britain

Updated and corrected version

With thanks to Tom Booth for drawing attention to the Orkney evidence and the modelling of patrilineal systems.

By around 2000 BC, more than ninety per cent of the ancestry of people in Britain derived from continental sources that had arrived only a few centuries earlier. The question is what that number actually measures, and what mechanisms are consistent with the evidence.

(A note on names. “Beaker people” assumes a pot equals a population. In Iberia, Beaker-associated individuals derive most of their ancestry from local Neolithic farmers; in central Europe they carry substantial steppe ancestry. Same material culture, two population histories. I use Beaker Complex for the material culture. For the genetics of the British transition I refer to Steppe-related or, more precisely, Lower Rhine–Meuse-derived ancestry. For the British period under discussion I use Chalcolithic–Early Bronze Age, or C–EBA.)

We have several well-documented episodes of settler colonisation — the Americas, Australia, Tasmania and New Zealand — where written records tell us roughly what happened. We also have useful prehistoric and early-historic parallels, notably the Dorset–Thule transition in the Arctic and the fifteenth-century conquest of the Canary Islands. Strip the documents away (or, in the prehistoric cases, recognise that none ever existed). Leave only bones, artefacts, settlement patterns, radiocarbon dates and ancient DNA. What survives? And how does that signature compare with third-millennium Britain?

What a population replacement looks like when we can check the documents

The Americas

The scale of post-1492 depopulation is contested but enormous; declines of the order of ninety per cent over a century and a half are routinely cited for many regions. Genetically this registers as a transient contraction of roughly half in effective population size — female effective size in the mitochondrial reconstruction of O’Fallon and Fehren-Schmitz (2011), autosomal effective size in the exome study of Lindo et al. (2016), who modelled a 57 per cent reduction together with sharp shifts in selection on immune genes.

One of the largest demographic catastrophes in recorded history therefore appears in the genome as something in the region of a halving of effective population size — and even that took two decades of work to see. It does not register as a ninety-per-cent ancestry replacement in the descendant indigenous population, because that is a different quantity. Native American ancestry persisted. The lineages that survived founded the modern populations. Effective population size is not a headcount; bottlenecks recover; the genome records the shape of the surviving lineage rather than the number of the dead. The genetic signal of mass mortality is weak, lagging and easy to miss.

Australia

The Colonial Frontier Massacres project has documented more than 400 frontier massacres between 1788 and 1930, with an estimated death toll above 10,000 Aboriginal and Torres Strait Islander people. The working definition is the deliberate killing of six or more undefended people in one operation. The record was assembled from settler diaries, newspapers, court records, parliamentary papers and survivor testimony. Massacre is characteristically planned, covert, and designed not to be discovered; most occurred on private land or at waterholes.

Almost none of this is archaeologically recoverable. Bodies were burned, dispersed or left unburied; sites are unmarked. Take away the archive and Australia’s frontier violence disappears almost completely. A future archaeologist would find an abrupt material-culture replacement, introduced fauna and flora, a new settlement system, and no massacre horizon whatsoever. Absence of a massacre horizon is therefore close to worthless as evidence for a peaceful transition. It is what we would expect to find either way.

Tasmania

The Aboriginal population at British settlement in 1803 is estimated by Ryan at around 7,000. By the mid-1830s around 200 survivors had been removed to Flinders Island. Between those two points lie disease, dispossession, the abduction of women and children, and the Black War. The entire episode occupies roughly thirty years — a single generation. No radiocarbon programme could resolve that as anything other than a horizon.

The claim that Tasmanian Aboriginal people were rendered extinct is a myth; descendant communities exist and are substantial, tracing largely through Aboriginal women in the Bass Strait sealing settlements. Genetically the shape is near-total replacement of the resident ancestry profile, with a minority indigenous residue transmitted disproportionately through women and indigenous male lineages largely gone.

New Zealand

Māori population estimates run from roughly 100,000 in 1769 to a low of about 42,000 in 1896, after which recovery occurred. Mortality was overwhelmingly from introduced disease rather than warfare. A colonisation in which disease dominated produced a far weaker genetic signal than C–EBA Britain shows: a bottleneck and substantial admixture, not replacement. Dispersed, low-density settlement limited epidemic spread.

Dorset–Thule

The closest prehistoric parallel is the Dorset–Thule transition in the Eastern Arctic. The Dorset (Paleo-Eskimo) population had occupied the Canadian Arctic and Greenland for roughly four thousand years, maintaining a genetically continuous and largely isolated lineage after an earlier migration from Siberia. Around the 11th–13th centuries AD, Thule groups expanded eastward from Alaska. Within a century or two the Dorset disappear from the archaeological record.

Ancient DNA shows the two populations were genetically distinct: there is essentially no Dorset ancestry in Thule or modern Inuit samples (though limited Palaeo-Eskimo ancestry has been detected in some other American Arctic and Na-Dene populations). Material culture was fully replaced. Direct evidence of face-to-face contact or violence is minimal to absent. Without contemporary written accounts, a future prehistorian would recover abrupt material-culture replacement, near-total genetic turnover, and no clear conflict horizon — a cleaner replacement signal than most documented colonial episodes.

Rapa Nui and the Canary Islands

Fifteen ancient genomes from Rapa Nui (Easter Island) went looking for the famous seventeenth-century “ecocide” collapse and found no bottleneck at all. Population appears to have grown steadily until European contact; the real demographic crash came in the 1860s with Peruvian slave raiding and introduced disease.

A documented fifteenth-century Spanish conquest of the Canary Islands involving warfare, enslavement and disease left modern Canary Islanders with approximately 18 % indigenous (Guanche) autosomal ancestry (Serrano et al. 2023), surviving more strongly on the maternal side than the paternal. A known military conquest therefore retained more indigenous ancestry than C–EBA Britain did.

Later British calibrations

The later history of Britain offers further calibration. The Roman occupation left almost no detectable population-level genetic signature. The Anglo-Saxon migrations of the fifth and sixth centuries contributed a substantial continental northern European component — often 50–75 % or more in early medieval eastern and southern England — that remains visible, though diluted, in the modern English gene pool. The Norman Conquest of 1066 produced a clear political and elite transformation but essentially no genome-wide discontinuity in the common population; it was an elite replacement rather than a demographic one. These three cases show that political takeover, cultural change and ancestry replacement are not the same thing.

What we actually see in Chalcolithic–Early Bronze Age Britain

The British record has three quantitative pillars.

First, the transition is multi-generational. Booth et al. (2021), re-examining the Olalde et al. 2018 data, showed that the shift from individuals with little or no, or substantial (20–40 %), Neolithic-related ancestry to the later homogenised population runs across 311–472 years (ten to sixteen generations) with up to 145 years of chronological overlap. This is fatal to every framing that requires an event. In the best-sampled region, 12 of 21 (57 %) sampled Wiltshire C–EBA burials were close genetic relatives; at Amesbury Down the figure was 8 of 11 (72 %).

Second, the sample is heavily biased toward families practising a visible inhumation rite. Cremation — the commonest recognised Late Neolithic rite — yields no workable genome-wide data. The ninety-per-cent figure is therefore the ancestry composition of the families we can sequence, not a simple population average.

Third — and this is the point that has not yet been fully absorbed — the residual local British Neolithic ancestry is smaller than the 2018/2021 figures suggested. Olalde et al. 2026 supply the proper source population: the Lower Rhine–Meuse Bell Beaker group. With that proxy the main British Beaker cluster (England_BB) is genetically cladal with the Lower Rhine–Meuse Bell Beaker group (single-source P = 0.61); no additional British Neolithic ancestry is required. For the later Chalcolithic–Early Bronze Age group the models assign 7.3–7.9 % Middle/Late Neolithic ancestry. Because it is impossible to determine whether that residual is British or continental, the bound is explicit: a maximum of eight per cent from local Neolithic populations of England, and a minimum of zero.

The sequence is now clearer. In the main Beaker-period population, local British Neolithic ancestry falls to a level indistinguishable from zero (compatible with 0–~3.5 %). A 7–9 % Neolithic-related component then reappears, or becomes detectable, in the following centuries. Whether that later component is residual British ancestry or further western continental input (or both) cannot yet be resolved. The quantity is well measured; its geographic source is not.

A small number of outliers with lower steppe ancestry, including the Amesbury Archer, provide a poorer fit to the main Lower Rhine–Meuse model. Olalde et al. 2026 note that their additional Neolithic-related ancestry may derive from local British Neolithic populations or from separate migratory streams. They remain real individuals of interest, but they do not constitute a population-level rebound.

Demographic proxies (cereal radiocarbon dates, settlement scarcity, woodland regeneration) indicate that the Late Neolithic landscape was already thinner and more pastoral than the Early Neolithic. Organised groups capable of raising Silbury and completing the sarsen phase of Stonehenge were still present; the landscape was not empty. But it was not densely occupied in the earlier fashion.

Britain had already undergone one near-total ancestry replacement. The arrival of Neolithic farmers around 4000 BC largely replaced the preceding Mesolithic hunter-gatherer population; that earlier transition is routinely discussed in demographic terms without the language of catastrophe. The C–EBA episode is the second such event.

Why the three usual explanations fail

Plague. The three British Yersinia pestis detections published by Swali et al. (2023) date to ~4000 cal BP — four to five centuries after Steppe-related ancestry appears in southern Britain. Those strains lack key virulence factors for efficient flea-borne transmission. An earlier presence is now documented: Sikora et al. (2025) report positive identifications of Y. pestis in two individuals from the Banks chambered tomb on South Ronaldsay, Orkney (NEO627 and NEO630, both newly reported), directly dated to 5291–4973 and 4961–4833 cal BP (roughly 3350–2880 BC). These are low-coverage detections (81 and 115 reads) rather than reconstructed genomes, and neither met the depth threshold for phylogenetic placement, so lineage assignment remains open. They nevertheless constitute the earliest evidence of plague in Britain, predating the southern samples by around nine centuries. Across the wider Sikora dataset, hits before 2500 BP are characterised by absence of the 19 kb region on pMT1 containing ymt. Detection is not evidence of a population-wide epidemic capable of driving the ancestry shift. Even where disease dominated in the comparative cases (New Zealand), the genetic outcome is far weaker than the one observed in southern Britain.

Violence. No mass graves, no skirmish sites, and no rise in traumatic injury mark the transition itself. Charterhouse Warren lies three centuries later. There is as yet no genetic evidence that it involved communities of differing ancestry, though the aDNA work is not yet published; the victims were isotopically local. Absence of skeletal violence proves nothing by itself — Australia shows that sustained frontier killing can leave precisely this signature. Yet the genetic evidence from Orkney makes a simple violent colonisation less tenable as the default explanation for Britain as a whole.

On the evidence of a single Westray cemetery (Links of Noltland), Bronze Age genomes from Orkney show a substantial autosomal shift toward Steppe-related ancestry, comparable in scale to the mainland. The paternal lineages, however, tell a different story: Neolithic male lineages, particularly I2a1b-M423, persisted for roughly a thousand years after the genome-wide change (Dulias et al. 2022). The pattern fits predominantly female-mediated gene flow into a local male population that remained in place. A rapid, male-dominated military takeover is difficult to reconcile with such continuity. Late Neolithic Orkney already displays strong indications of patrilineal and patrilocal organisation; in that setting, incoming ancestry appears to have been absorbed largely through women marrying in. The authors themselves caution that the result is a snapshot from one remote part of the archipelago.

Scandinavia provides the nearest large comparative dataset and does not show the same pattern. There the Corded Ware / Battle Axe horizon brought substantial Steppe-related ancestry accompanied by strong paternal turnover (early dominance of R1a, later rise of other lineages including I1). Neolithic farmer Y-lineages largely disappear. Orkney’s long persistence of local male lineages after a major autosomal shift is described by Dulias et al. as completely absent elsewhere in Copper Age / Bronze Age Europe. It stands out against both southern Britain and Scandinavia.

Modelling of segmentary patrilineal systems shows that variance in reproductive success between groups, combined with lineal fission, can produce a severe reduction in male effective population size without violence (Guyon et al. 2024). This removes the automatic inference from a Y-chromosome bottleneck to conflict. It does not, however, supply a non-violent mechanism for the specific pattern seen in southern Britain — the near-total replacement of one set of Y-lineages (Neolithic I2a) by another (R1b) that were previously absent. Social organisation can thin diversity within a lineage pool; full lineage replacement still requires a process that systematically favoured the incoming paternal lines.

Peaceful coexistence between equals. An ancestry replacement of this scale is still not what mutual accommodation between demographic equals looks like. Something systematically favoured one set of lineages over many generations. The comfortable reading continues to require special pleading.

Does Britain look like any of the comparative cases?

Set against the calibrations, the British evidence shows a stronger ancestry-replacement signal than the Americas, New Zealand or most of Australia; a residual local contribution at most comparable to (and probably lower than) the Canarian case; the same absence of a massacre horizon that documented Australian frontier violence produces; and chronological resolution too coarse to detect thirty-year regional collapses. On the limited evidence of one Westray cemetery, Orkney demonstrates that the same broad demographic horizon could be negotiated differently within Britain itself. Social structure appears to have shaped the outcome.

It also shows one feature that none of the colonial cases display in the same way. Beaker-period activity is closely associated with the Stonehenge landscape. Silbury Hill was completed inside the early horizon of the transition. The Amesbury Archer, an isotopic first-generation incomer of probable Alpine or central European origin, was buried a few kilometres from the monument with exceptional provision. Olalde et al. 2026 themselves note the continued building and use of Late Neolithic monuments and read this as cultural continuity, with substantial cultural change only in the twenty-third century BC. An alternative reading is selective association with an already powerful sacred geography rather than simple continuation of the Neolithic building tradition or deliberate ceremonial erasure. The two interpretations are not mutually exclusive, but the genetic discontinuity remains real.

What the evidence now allows us to say

The ninety per cent figure is not a mortality estimate. It is a statement about the volume of arrivals and their long-term reproductive success, drawn from a biased sample of families who buried their dead in a way that leaves a body we can sequence. The transition lasted ten to sixteen generations. The residual local British Neolithic contribution to the main later population is at most eight per cent and possibly zero. Disease is documented from ~4000 cal BP in southern Britain (Swali et al. 2023) and earlier still in Orkney (Sikora et al. 2025), but remains unevidenced as the primary driver of the ancestry shift. Absence of massacre evidence is exactly what sustained frontier violence can leave behind — yet the Orkney pattern and the modelling of patrilineal systems show that strong male-lineage turnover does not automatically require it.

The most coherent reading is therefore this. Incomers arrived in numbers into a landscape already thinned for reasons unconnected with them. They held demographic and subsistence advantages. Over four centuries they absorbed a small resident population through arrangements that produced large-scale reproductive asymmetry. Those arrangements were shaped by social organisation — patrilineal and patrilocal structures that, as the limited Orkney evidence shows, could produce different genetic outcomes in different regions. Scandinavia followed the more common northern European pattern of strong paternal turnover; the one well-sampled Orkney cemetery did not. Violence remains possible in particular places or phases. It is no longer the explanation demanded by the genetic numbers.

We cannot say the process was peaceful everywhere. We can say that a simple violent colonisation is less likely than the genetic data once seemed to suggest, and that social structure provides a sufficient mechanism for much of the patterning we actually observe.

See also https://www.sarsen.org/2026/08/how-much-of-neolithic-britain-survived.html for a statistical discussion.

References

Allentoft, M.E., Sikora, M. et al. 2024. Population genomics of post-glacial western Eurasia. Nature 625: 301–311.

Booth, T.J., Brück, J., Brace, S. & Barnes, I. 2021. Tales from the supplementary information: ancestry change in Chalcolithic–Early Bronze Age Britain was gradual with varied kinship organization. Cambridge Archaeological Journal 31(3): 379–400.

Dulias, K. et al. 2022. Ancient DNA at the edge of the world: Continental immigration and the persistence of Neolithic male lineages in Bronze Age Orkney. Proceedings of the National Academy of Sciences 119: e2108001119.

Gretzinger, J. et al. 2022. The Anglo-Saxon migration and the formation of the early English gene pool. Nature 610: 112–119.

Guyon, L., Guez, J., Toupance, B., Heyer, E. & Chaix, R. 2024. Patrilineal segmentary systems provide a peaceful explanation for the post-Neolithic Y-chromosome bottleneck. Nature Communications 15: 3243.

Lindo, J. et al. 2016. A time transect of exomes from a Native American population before and after European contact. Nature Communications 7: 13175.

Moreno-Mayar, J.V. et al. 2024. Ancient Rapanui genomes reveal resilience and pre-European contact with the Americas. Nature 633: 389–397.

O’Fallon, B.D. & Fehren-Schmitz, L. 2011. Native Americans experienced a strong population bottleneck coincident with European contact. PNAS 108: 20444–20448.

Olalde, I. et al. 2018. The Beaker phenomenon and the genomic transformation of northwest Europe. Nature 555: 190–196.

Olalde, I., Altena, E., Bourgeois, Q. et al. 2026. Lasting Lower Rhine–Meuse forager ancestry shaped Bell Beaker expansion. Nature 652: 938–946.

Pool, I. 1991. Te Iwi Maori: A New Zealand Population Past, Present and Projected. Auckland: Auckland University Press.

Raghavan, M. et al. 2014. The genetic prehistory of the New World Arctic. Science 345: 1255832.

Ryan, L. 2012. Tasmanian Aborigines: A History Since 1803. Sydney: Allen & Unwin.

Schulting, R.J. et al. 2025. ‘The darker angels of our nature’: Early Bronze Age butchered human remains from Charterhouse Warren, Somerset, UK. Antiquity 99: 101–117.

Serrano, J.G. et al. 2023. The genomic history of the indigenous people of the Canary Islands. Nature Communications 14: 4641.

Stevens, C.J. & Fuller, D.Q. 2012. Did Neolithic farming fail? The case for a Bronze Age agricultural revolution in the British Isles. Antiquity 86: 707–722.

Sikora, M., Canteri, E., Fernandez-Guerra, A. et al. 2025. The spatiotemporal distribution of human pathogens in ancient Eurasia. Nature 643: 1011–1019.

Swali, P. et al. 2023. Yersinia pestis genomes reveal plague in Britain 4000 years ago. Nature Communications 14: 2930.