Eland collagen found in 60,000-year-old ochre, but its purpose is unclear

Collagen matching the common eland turned up in 2 of 5 archived samples from roughly 60,000-year-old ochre-rich layers at Klipdrift Shelter, South Africa, but which tissue it came from, and why, is unknown.

Peer reviewed · Laboratory analysis of archived samples · One site, two positive samples

Ancient protein from the common eland, a large antelope, turned up in red ochre-rich sediment laid down about 60,000 years ago on South Africa’s southern Cape coast. First author Zahra Haghighi of the University of Bergen and colleagues at Kyushu University and the University of Copenhagen report it in Proceedings of the Royal Society B. It is rare direct evidence of what early Homo sapiens may have combined with ochre. The catch: the protein cannot say which part of the eland it came from, or why it was there.

Evidence at a glance

Is this a big deal?

A solid methods step with a modest archaeological payoff. Recovering species-level protein from 60,000-year-old sediment is hard, and the eland match is well supported. What people were doing with that eland material is still a hypothesis.

What scientists found

Ochre is iron-rich earth, ground to a red powder. Its use in southern Africa’s Middle Stone Age was widespread and habitual. The red survives; whatever it was mixed with usually does not.

At Klipdrift Shelter, about 150 km east of Cape Town, the Howiesons Poort layers held a continuous ochre-rich surface in layer PBE. From photographs and site notes, the authors estimate it was possibly more than 1 m long and 40 cm wide, and 5 to 7 mm thick. It was dug out in 2011 to 2013 without its extent being recorded.

Collagen turned up in two samples, from layers PBD and PBE: 69 peptides covering 16.5% of the collagen chain called COL1A2. Every variant read matched eland.

How they found it

The samples were never collected with protein work in mind; they had sat in an archive in Cape Town since excavation. Each subsample weighed 30 to 150 mg, about the mass of one to several grains of rice.

Sediment grips proteins tightly, so the team used two steps. A new acid-then-base extraction on two samples showed the protein type: mammal type I collagen. Then all five samples were treated with strong acid and read by mass spectrometry, matching against a small collagen-only database.

Think of the collagen chain as a long sentence that every antelope writes slightly differently. At three positions, numbered 693, 720 and 878, the Klipdrift sequence used spellings that, among the reference species, only eland uses. Two of those were each covered six times by overlapping fragments and turned up in both layers. The analogy has a limit: it can only pick from species in the reference set. Giant eland has no collagen sequence on file; the authors rule it out because it lives in central Africa today.

Two checks support the result. Near-complete deamidation, a chemical change that builds up over time, fits ancient protein, though heating and acid can cause some. And under microscopes, only 4 of 15 samples held bone-like grains, one or two each, none from PBE.

Why it matters to you

This won’t change anything in your daily life. It speaks to where we came from: what early modern humans did with their materials 60,000 years ago. The paper notes the eland is deeply revered in San mythology, and asks whether that significance has deep roots. For the wider toolkit of the period, see what tools were made in the Stone Age.

What the evidence does not show

It does not show a hide, a paint or a ritual. The authors set out two scenarios, a mixture or a decayed ochre-treated hide, and allow that other practices may have formed the feature.

Two positive samples from one site is a small base. Three of the four samples with bone-like grains came from PBD, so bone is harder to rule out there than in PBE. People also skinned animals and extracted marrow at Klipdrift, and eland is already known from the site’s animal bones. On SciQuest’s reading, eland material in the deposit is not by itself proof of a deliberate mixture. The authors are direct about the symbolic reading: the single example “is not sufficient to support the ritual significance of eland in the MSA.”

How this fits what we already knew

At Blombos Cave, two abalone shells about 100,000 years old held an ochre-rich mix, with burnt and crushed bone nearby. At Sibudu, milk proteins were found in an ochre mixture on a stone flake about 49,000 years old. Klipdrift extends protein detection from artefacts to loose sediment. That matters in Africa, where the oldest DNA recovered is only about 18,000 years old. Compare the Neanderthal wooden tools preserved as moulds in Spain.

Independent perspective

SciQuest found no comment from an archaeologist outside the author team as of 22 October 2026, 15:12 UTC, and has not yet interviewed one.

How other headlines framed it

• Kyushu University’s release, reposted by Popular Archaeology, called it an “invisible eland” and gave both scenarios, though it misspells the site as “Kilpdrift”.
• The Debrief led with Stone Age technology and the case for proteins outlasting DNA, and noted that the use is not established.
• Anthropology.net gave the closest reading of the paper, including the two-of-five count and the authors’ caution on ritual.

The paper does not use the “invisible” framing, so SciQuest leaves it out.

What happens next

The authors call for repeating the work at other Middle Stone Age sites, and for reference data that can tell skin from bone. They also want future digs to record sample positions and micro-layering, so protein results can be tied to how a layer formed.

In one sentence

Protein in 60,000-year-old South African ochre points clearly to eland, but whether it came from a mixture or a hide, and what it meant to the people there, is still open.

Read next

• What tools were made in the Stone Age? The technology around the ochre.
• Neanderthal wooden spears survive as moulds in Spain: another way perishable materials leave traces.
• An 11,000-year-old human-on-leopard sculpture: animals in later symbolic art.

Sources and disclosure

1. Haghighi Z, Tsutaya T, Velliky EC, Armitage S, Miller C, Mackie M, d’Errico F, Troche G, Olsen J, van Niekerk K, Henshilwood C, Cappellini E. “Palaeoproteomic investigation of archaeological sediments associated with ochre use in the Middle Stone Age of southern Africa.” Proceedings of the Royal Society B 293(2081): 20253300, published 21 October 2026. doi:10.1098/rspb.2025.3300 (full text read via Europe PMC).
2. Haghighi Z et al. Mass spectrometry dataset. PRIDE PXD069507.
3. Kyushu University news release, “Invisible Eland Identified in 60,000-Year-Old Ochre Sediment”, as reposted by Mirage News, 21 October 2026.
4. Coverage read for framing: The Debrief, Anthropology.net, Popular Archaeology.

Disclosure: This article was drafted with AI assistance and checked against the sources above, including the full open-access paper. The authors declare no competing interests; funding sources are listed in the paper. Calculations by SciQuest: the paper’s 171 sequenced residues at 16.5% coverage implies a COL1A2 chain of about 1,040 residues, which places the marker positions in the graphic. The rice comparison assumes about 20 to 30 mg per grain, so 30 to 150 mg is roughly one to seven grains. The graphic is a SciQuest original built from the paper’s reported values. Spotted an error? See our corrections policy.