Châtelperronian bone fragments identified as Neandertal based on protein residues

Frido Welker and colleagues have applied a new method to identify tiny bone fragments as Neandertal remains, based on the protein residues that they contain.

The study uses a fairly simple concept—take a lot of unidentifiable bone fragments and test them with a protein mass spectrometry toolkit (ZooMS) to identify them to genus if possible. They were pretty successful. The real innovation is that they were able to identify a protein polymorphism for which 99% of living humans have a derived amino acid, and all known Neandertals have the ancestral amino acid. In the context of a Châtelperronian archaeological layer, finding the ancestral allele gives a strong indication (but not an infallible one) that a bone fragment probably has a Neandertal heritage.

I think the result is pretty cool because Welker and colleagues were able to identify a large number of bone fragments that probably belong to an infant skeleton, which they already knew was in the site based on a handful of identifiable remains.

Despite their spatial proximity within Layer X and squares C7 and C8, none of the newly described specimens could be fitted together to form larger fragments. The morphologically informative specimens, however, seem to represent small fragments of an immature cranium and an unfused vertebral hemiarch of neonatal age (SI Appendix, Fig. S10 and Table S13). This is supported by isotopic evidence suggesting that these fragments belonged to a nonweaned infant and by proteomic evidence in the form of proteins present in bone before bone remodeling has started. All the newly identified specimens were found in close spatial association with a previously described hominin temporal bone from square C7, Layer Xb, assigned to an infant around 1 y old, as well as 10 dental specimens from squares C7 and C8 (6, 7). These dental specimens overlap in developmental age, suggesting they represent one or possibly two individuals between 6 and 18 mo old (7). The 28 newly identified specimens, together with already described specimens, may therefore represent the skeletal remains of a single infant.

Châtelperronian makers

The reporting on this paper has emphasized the conclusion that the bones are Neandertal, which in addition to the protein evidence is also indicated by some mtDNA sequence fragments from some of the remains. To me this is basically a failure to reject a long-substantiated association: Neandertals made the Châtelperronian industry.

The Châtelperronian has been found across much of France and into the Pyrenées of Spain, and dates to around 40,000 years ago. It is termed a “transitional” industry because although it is based upon flaking strategies typical of earlier Mousterian assemblages, the Châtelperronian includes elements usually attributed to Upper Paleolithic toolkits including many bone artifacts and elongated flake tools that share many details with the blades of later Aurignacian assemblages.

The identity of the makers of the Châtelperronian has been a white whale for some archaeologists. The only diagnostic skeletal remains ever found with Châtelperronian tools have been Neandertals, notably the Saint-Césaire skeleton but also more fragmentary skeletal remains from Arcy-sur-Cure. Yet despite this association, some archaeologists have doubted that Neandertals made such apparently “sophisticated” tools. They have examined whether Châtelperronian assemblages are a palimpsest or mixture of elements from earlier Mousterian and later Upper Paleolithic contexts, whether Neandertal remains might have moved out of stratigraphic position, whether Châtelperronian layers may have slumped in a way not recognized in old excavations.

Today it would be truly surprising if an unambiguous modern human skeleton turned up in a Châtelperronian archaeological layer. If a specimen with modern human traits emerged, or a tooth that shared traits with modern humans, I would not be surprised, but that’s because I expect there was likely population mixture in Europe at the time of the Châtelperronian some 40,000 years ago.

Ten years ago, a debate erupted about the type locality of the Châtelperronian industry, Grotte des Fées du Châtelperron. I wrote about that episode here (Interstratified palimpsests), and it was never really satisfactorily resolved.

Why was it such a big deal? The last refuge of scoundrels has been the notion that Neandertals were incapable of making Châtelperronian artifacts on their own and needed the enculturating influence of modern humans to do so. This idea was popularized by Jared Diamond in his book, The Third Chimpanzee and has hung around within European archaeology for more than thirty years. The importance of possible interstratification of Châtelperronian and Aurignacian industries is a test that Neandertals who made Châtelperronian artifacts and modern humans who made Aurignacian artifacts might really have encountered each other on the same landscape. Without the possibility of such contact, it is hard to conceive of the idea that Neandertals were “acculturated” by their encounters with modern humans.

What is my opinion? I have to say, two or three thousand years is a lot of time, and it is hard for me to imagine that populations did not exchange a good amount of information across Europe during that time, whether they lived in overlapping areas or not. I doubt that the early Aurignacian people were very much different from late Neandertals in their cultural and technical abilities. Even if they were different on average for some cognitive abilities, which I can well imagine, I expect their abilities would have overlapped to a substantial degree.

The conclusion of the paper sums up my feelings fairly well.

Our biomolecular data provide evidence that hominins contemporaneous with the Châtelperronian layers have archaic nuclear and Neandertal mitochondrial ancestry, supporting previous morphological studies (6, 7). They are therefore among some of the latest Neandertals in western Eurasia, and possible candidates to be involved in gene flow from Neandertals into AMHs (or vice versa) (48). Future analysis of the nuclear genome of these or other Châtelperronian specimens might be able to provide further insights into the direction, extent, and age of gene flow between Late Pleistocene Western European Neandertals and “incoming” AMHs (49, 50).

More results coming

Doing more to uncover the hominin remains at a site is valuable even if the anatomical information to be gained is minimal, and even if genetic information cannot be reliably obtained from the fragments with today’s techniques.

Still, in the short term at many sites, what the method reveals about the fauna will be even more interesting. Identifying rare species in faunal assemblages from non-diagnostic fragments should greatly increase the representativeness of the archaeological record. As reported by Welker and colleagues in an earlier paper (2015), they can reliably identify bone fragments to genus in a large fraction of cases, even for bone fragments that have passed through a carnivore’s digestive tract.

This result is published this week just after a number of studies were presented in talks at an international conference on ancient biomolecules. That other work is reported by Ann Gibbons (“Oldest-ever proteins extracted from 3.8-million-year-old ostrich shells”), which includes a statement that protein has now been recovered from tooth enamel of extinct fauna from Dmanisi. Protein analysis of ancient fossils suddenly looks like a very big deal.


Brown, S., Higham, T., Slon, V., Pääbo, S., Meyer, M., Douka, K., ... & Derevianko, A. (2016). Identification of a new hominin bone from Denisova Cave, Siberia using collagen fingerprinting and mitochondrial DNA analysis. Scientific reports, 6. doi:10.1038/srep23559

Welker, F and others. 2016. Palaeoproteomic evidence identifies archaic hominins associated with the Châtelperronian at the Grotte du Renne. Proceedings of the National Academy of Sciences, U.S.A. (early edition). doi:10.1073/pnas.1605834113

Welker, F., Soressi, M., Rendu, W., Hublin, J. J., & Collins, M. (2015). Using ZooMS to identify fragmentary bone from the late Middle/Early Upper Palaeolithic sequence of Les Cottes, France. Journal of Archaeological Science, 54, 279-286. doi:10.1016/j.jas.2014.12.010