A forearm bone came out of the Aurora layer at Gran Dolina this summer, a fairly complete ulna from a collection made up mostly of fragments. It belonged to a member of Homo antecessor, the species that occupied this cave in the Sierra de Atapuerca in northern Spain about 850,000 years ago, and it was one of at least 24 human fossils recovered in a single excavation season.

IPHES-CERCA, the Catalan institute that runs the Gran Dolina excavation, announced on 24 July that the season’s haul raises the minimum number of individuals identified in level TD6.2 to between 13 and 15. What came out is mostly small and broken: a fragment of temporal bone, cervical and lumbar vertebrae, hand and foot phalanges, three teeth, the ulna, and pieces of tibia, femur and humerus. Many of those bones carry cut marks left when the bodies were processed, and at least six of the individuals were adults, according to the institute, an age group it describes as having been clearly underrepresented until now.

Why an age distribution became an argument

Gran Dolina has been the reference site for early human cannibalism since the mid-1990s, and the age of the dead there has been carrying the explanation. The human bones show the same treatment as the deer bones found alongside them: skinning, dismembering, defleshing, and breakage to reach the marrow.

The question was never whether it happened. It was why, and the answer the excavation team published in 2012 leaned heavily on who the victims were.

In that paper, in the Journal of Human Evolution, Palmira Saladé and colleagues reported a minimum of 11 individuals from level TD6. The age breakdown was lopsided: six children, four of them younger than five, three individuals between 10 and 15, and two young adults. Signs of human modification appeared on 67 Homo antecessor specimens, 41 percent of those recovered by then.

Set against that age profile, the team argued for intergroup predation, one group of hominins periodically hunting and eating members of another. The age skew was the load-bearing part. The paper drew an explicit analogy with chimpanzee border patrols, where attacking parties pick their moments so that the odds are uneven and the attackers go uninjured.

“The ages represented in the H. antecessor group at TD6 suggest that the individuals that were captured involved the least risk for the predators, which implies a low-cost territory control strategy,” the authors wrote. Change the age distribution and you change how much work that argument can do.

Two young adults were already in the count

The 2012 paper listed two young adults among its 11 individuals, though its own Table 4 sorts those 11 into eight infants, two sub-adults and one adult.

What that paper never did was attribute a single cut mark, percussion scar or peeling fracture to an adult specimen. Working bone by bone through the butchered material, it attached an age to material in exactly two places: an infant’s axis vertebra with three transverse incisions, and a pair of radii it identified as sub-adult. Its own summary was that the consumed individuals were mostly immature, not that they were all children.

What the new material adds is butchery marks on adult bone at the level of the individual specimen, where the 2012 paper could only infer adult consumption across the assemblage as a whole.

Saladé, who coordinates the TD6 work, framed it this way in the institute’s release: “Until now, we had mainly infant and juvenile remains, which conditioned our interpretation of the assemblage. The discovery of more adult individuals, some of whom also bear cut marks, demonstrates that cannibalism was not restricted to younger individuals.”

The 2012 authors also looked for ritual and did not find it. “In the TD6 record, no material evidence has been found that would allow us to infer any kind of ritual,” they wrote. Their conclusion was nutritional.

Six adults, and a count that has not settled

The 24 fossils, the range of 13 to 15 individuals and the six adults were announced at a press conference and in an institutional release. No peer-reviewed paper reports any of them. The taphonomic and paleoanthropological studies, in the institute’s own words, “are now underway”, and they are working on material from a season that reopened the Aurora layer and dug only part of its upper section. Until they are published the numbers are provisional, and they are visibly still moving.

Two days before the Gran Dolina release, the same institute’s campaign-wide announcement gave the individual count as between 12 and 15. At the closing press conference the team put the adults lower still, at between two and six, and stressed that the analyses were preliminary. The Gran Dolina release is then internally inconsistent on the same point: its summary line says up to six could be adults, while the body of the same text says the presence of at least six adults is confirmed. The Spanish and Catalan versions carry the identical discrepancy, so it is not a translation artefact. A 25th fossil turned up while the closing press conference was underway, which is why some accounts, including one by a member of the team, say 25 rather than 24.

The ulna is a smaller instance of the same looseness. IPHES calls it the first ulna ever recovered from the site; the 2012 paper describes an ulna shaft fragment cut by transverse incisions, plus a second, doubtfully attributed ulna fragment with scraping and percussion marks, so what is new is the completeness of the bone rather than its existence.

Ages given for the adults are shakier again. The ICAL wire, reporting the team’s preliminary analysis at the press conference, put them at roughly 27 to 30 years. Marina Lozano Ruiz, a dental anthropologist at Rovira i Virgili University and IPHES, wrote in her own account of the season that teeth recovered since the 2024 season indicate adults of 20 to 26. Neither figure appears in any institutional release, and the two do not agree.

There is also the previous year to account for. In July 2025 the institute announced a child’s cervical vertebra from the same level, bearing cut marks it read as deliberate decapitation. Twelve months on, that specimen has still not appeared in a peer-reviewed journal.

And a cut mark is evidence of cutting. Michael Pante, a paleoanthropologist at Colorado State University who was not involved in the discovery, made the point about the 2025 vertebra: “This decapitation doesn’t mean they consumed that individual.” The people at Atapuerca were “obviously cutting up a child for some reason, but there are a number of reasons they may have done that,” he said. National Geographic notes a funeral ritual as one possibility. Pante was unpersuaded by the hunting-rivals reading too: “There is not a lot of evidence of that.”

The narrow thing that is now true

Human bone lifted from level TD6.2 carries marks made by stone tools, and some of it is adult. The collection is now representative enough that the excavators say they can begin to approach the group’s demographic structure, rather than read it off a handful of children.

That is a narrower claim than the one circulating this week, and it is the one the evidence currently supports.