What the Inner Ear Remembers

AFor most of the twentieth century the movements of prehistoric peoples were reconstructed from pots, and the reconstruction was contested every time a new style appeared. Since 2010 a second archive has opened. The petrous bone, the dense wedge of the inner ear, holds up to a hundred times more endogenous DNA than a long bone from the same skeleton, and its discovery turned a trickle of ancient genomes into a flood: fewer than twenty whole genomes had been published by 2011, against more than twelve thousand by 2024. The genomes appear to answer in a single measurement the question that pots could only circle, namely whether a new manner of burial marks new people or merely new fashions. Yet a genome is a fact about one person, and an archaeologist wants a fact about a population. Whether the two can be joined without a bridge of assumptions has divided the field since the first steppe results appeared, and the argument has grown sharper as the samples have grown larger.

BMarit Ellingsen, a population geneticist at the Bergen Centre for Human Population History, holds that the signal is too large to be explained away. Her group assembled 640 genomes from 47 cemeteries in northern and central Europe dated between 3200 and 2200 BC. Before 2900 BC, ancestry traceable to the eastern steppe accounted for four per cent of the pooled genome; within the following three centuries it accounted for sixty-eight. Ten generations, on her arithmetic, is too short an interval for drift or for intermarriage at the margins to produce a figure of that size. The pattern is also lopsided by sex: steppe ancestry makes up seventy-eight per cent of the Y-chromosome lineages in her post-2900 sample but only thirty-one per cent of the mitochondrial ones, an asymmetry she reads as evidence that the incomers were overwhelmingly male and that local women were absorbed rather than displaced.

CThe reading yields a prediction and Ellingsen has tested it. If migrating men rather than migrating households drove the turnover, the asymmetry ought to be sharpest where the newcomers had most to gain and faintest where they had least. Across her sample the pattern was confined to the cereal-farming lowlands; at upland sites above six hundred metres the two figures converged, at forty-four per cent on the Y chromosome against thirty-nine on the mitochondrial side, a gap her own statistics call insignificant. She therefore describes the migration as a process whose shape changed with the landscape it entered rather than as a wave of uniform height. She adds a limit of her own. Where the dead were cremated, nothing survives that can be sequenced at all, and cremation was the dominant rite across perhaps a fifth of the regions on her map; for those regions, she accepts, no genetic statement of any kind can be made.

DAnselm Reiter, who studies settlement archaeology at the Danube Institute in Regensburg, answers that every one of these genomes was taken from a grave, and that graves are not a sample of the living. His 2021 survey compared burial counts with house counts at thirty-one sites and concluded that formal burial was extended to roughly one person in six. If the rite is that selective, a change in who is admitted to the cemetery will look exactly like a change in who inhabits the valley, and the genome cannot tell the two apart. Reiter concedes two points. His house-count models carry margins of error he calls uncomfortable, since the occupancy of a dwelling must be estimated rather than counted. And the sex-linked asymmetry, he agrees, resists explanation by burial custom alone: no known rite admits incoming men and local women to the same ground while excluding everybody else.

EThe dispute left the seminar room some time ago. Ancestry results have been quoted in support of claims about who belongs where, and since 2021 several journals have demanded documented community consultation before a bone may be drilled, because sampling a petrous bone destroys it. Ellingsen notes, awkwardly for her own case, that the cemeteries yielding the cleanest sequences are the monumental ones built for the few, so the evidence that most favours her reading is the evidence least likely to describe a whole community. Reiter answers that a single appetite created both difficulties at once, since a discipline that rewards spectacular results will also drill the most spectacular graves first. Both regret one habit above all: that the field still reports individuals sequenced rather than communities represented, a figure that flatters everybody who cites it.

FA settlement of sorts is emerging in which the two accounts are assigned to different questions rather than ranked against one another. For the biography of an individual, and for the web of kinship inside a single cemetery, ancient DNA is unmatched and its bias is beside the point, since two brothers buried side by side are not made less related by the neighbours who were left out. For any statement about population size or about the proportion of a community replaced, the genomes must be calibrated against settlement evidence before they can be used at all, and across a fifth of regions they cannot be calibrated. The unresolved question is whether DNA preservation itself varies with the chemistry of the soil, which would make the bias geographical as well as social; no survey yet covers enough contrasting ground to settle it. What both researchers reject is the assumption behind the earliest surveys, that the individuals easiest to sequence are the ones most worth counting.