Another Aurignacian Neandertal, or just dinner?

I said I was going to do my best to scoop the press this week. How about this piece of undernews: at one of the few early Aurignacian sites to preserve skeletal remains, Les Rois, France, one of the Aurignacian-associated mandibles looks like it may have been a Neandertal.

Before I tell the whole story, let me telegraph the bottom line: Do I think this specimen was really an Aurignacian Neandertal?

My opinion has always been that Europeans in the time span from 40,000 to 25,000 radiocarbon years presented a varying mixture of “Neandertal” and “modern” morphological features. From that standpoint, it is not surprising to find a mandible that has the combination of features reported here. In this case, the most significant mandible (which is really quite a small fragment) shows one very interesting characteristic: a perikymata count and packing pattern similar to Neandertals and different from other Upper Paleolithic European teeth. But as I’ll point out below, living humans are variable in their enamel formation in ways that reduce the significance of the differences between Neandertals and later Europeans.

But the story is significant – not only do these remains extend the biological variability of known Aurignacian-associated people to include Neandertal-like developmental patterns, but also they help to inform us about the potential of cultural associations at other sites, including Vindija.

The morphology

Let’s consider what the authors wrote about the specimens. Here’s most of the abstract of the paper, in Journal of Anthropological Sciences, by Fernando Ramirez Rozzi and colleagues:

Here we reassess the taxonomic attribution of the human remains, their cultural affiliation, and provide five new radiocarbon dates for the site. Patterns of tooth growth along with the morphological and morphometric analysis of the human remains indicate that a juvenile mandible showing cutmarks presents some Neandertal features, whereas another mandible is attributed to Anatomically Modern Humans. Reappraisal of the archaeological sequence demonstrates that human remains derive from two layers dated to 2830 kyr BP attributed to the Aurignacian, the only cultural tradition detected at the site. Three possible explanations may account for this unexpected evidence. The first one is that the Aurignacian was exclusively produced by AMH and that the child mandible from unit A2 represents evidence for consumption or, more likely, symbolic use of a Neandertal child by Aurignacian AMH. The second possible explanation is that Aurignacian technologies were produced at Les Rois by human groups bearing both AMH and Neandertal features. Human remains from Les Rois would be in this case the first evidence of a biological contact between the two human groups. The third possibility is that all human remains from Les Rois represent an AMH population with conserved plesiomorphic characters suggesting a larger variation in modern humans from the Upper Palaeolithic (Ramirez Rozzi et al. 2009:153).

So what is this “child mandible from unit A2”? Here’s a picture showing pretty much every view:

Les Rois mandible B

Figure 3B from Ramirez Rozzi et al. 2009, picturing the Les Rois mandible B

As you can see from the picture, the mandible is far from complete. It has its adult premolars it lacks any posterior teeth and the base of the mandibular corpus. If I were looking for a diagnosis, I would not necessarily expect to find one. In that respect, the mandible is similar to the Kent’s Cavern maxilla. As in that case, for Les Rois B I don’t think you can do much to substantiate either Neandertal or non-Neandertal affinity based on external morphology alone. The absolute dimensions of the teeth overlap with both Neandertals and modern humans, as do the root dimensions (as determined by scans).

The text also includes this:

The change in orientation of the mandibular surface at the canine level evokes a flat or slightly arched anterior mandibular surface, characteristic of Neandertals (Schwartz & Tattersall, 2000) (Ramirez Rozzi et al. 2009:161).

This is correct but not strongly probative; the morphology is hard to judge and overlaps between these groups of humans.

Another mandible from the side, mandible A, the authors diagnose as a modern human with no specific Neandertal-like characteristics. Confusingly, mandible A comes from unit B, which overlies the unit A2 where the mandible B was found. And the site has a number of isolated teeth from both these units, some of which figure into the story.

The radiocarbon dates for the units A1 through B are clustered in a range around 30,000 radiocarbon years. That makes them far from the earliest Aurignacian, and they postdate substantially any Neandertal remains in France – really, they overlap only with the latest Mousterian sites in southern Spain.


So given the scant morphological evidence, why does the paper conclude so strongly in favor of some Neandertal affinity for the specimen?

The answer has to do with enamel formation. The authors examined the perikymata counts and packing patterns on the Les Rois teeth – an observation that was simply unavailable to earlier scientists who examined and reported on the remains. Whereas the eyeball-level morphological features of the specimens are relatively undiagnostic, the perikymata patterns appear to be more interesting. The slightly later specimens including mandible A, in unit B of the site, all look like other Upper Paleolithic non-Neandertal specimens. But the teeth in mandible B have enamel development profiles like Neandertals and unlike “anatomically modern” specimens from the Upper Paleolithic of Europe.

Mandible B is not alone in having a Neandertal-like developmental profile. From the same unit of the site, A2, there is one other canine tooth and three incisors, representing at least two individuals, all of which also have low perikymata counts. The paper represents these teeth as falling within the Neandertal distribution and outside the range represented by modern humans.

Unlike mandible B, the teeth present in mandible A all have high perikymata counts, there are no nonmetric characters present that would suggest Neanderthal affinity. If you found this at a much later site, you wouldn’t notice anything unusual about it. Does it matter that the “modern” looking specimens are the later ones, and the “Neandertal” looking specimens are earlier? Not clear – there are really too few remains to make this into a significant story, particularly in the context where the two units do not differ significantly in radiocarbon ages.

Should we believe that the dental remains from unit A2 are Neandertal? The dental development information is directly relevant to the variability of early Upper Paleolithic Europeans – the Les Rois specimens here extend that variability into significant overlap with Neandertal dental development schedules. I think that’s quite important – there’s no clean break denoting the demise of the Neandertals. That observation reflects other early Upper Paleolithic European samples, many of which also present Neandertal-like morphological characters.

But it’s unclear to what extent enamel formation profiles, reflected by perikymata counts, accurately inform about phylogeny. Modern humans are quite variable in these perikymata counts and packing patterns. When it comes to total counts of perikymata, Neandertals cannot be distinguished from the variability among recent human populations (Guatelli-Steinberg et al. 2005; 2007). In this study, molars are not an issue, because they are not preserved for the relevant teeth. Perikymata packing patterns do separate known Neandertal specimens from samples of recent humans (Guatelli-Steinberg et al. 2007), and in that respect the Les Rois A2 teeth are similar to Neandertals.

How important are these observations of dental development? That’s a broader question than I am prepared to answer here, except to note my earlier posts on dental development in Neandertals (“Neandertal teeth: the other shoe”, “How modern is ‘modern tooth development’?”). I can also point to a current review of the issue by Debbie Guatelli-Steinberg (2009), which introduces the recent literature. The short answer is that nobody really knows.


The mandible B is also cutmarked:

Cutmarks on mandible B consist of three parallel striations located on the lingual aspect, below the right lateral canine and P3 (Fig. 5 and Fig. 4 Suppl. Mat.). Two of them bear diagnostic features of flint cutting-edge generated marks in form of v-shaped cross sections, barbs and, in one case, a typical splitting (Fisher 1995) (Ramirez Rozzi et al. 2009:170).

The authors point out that many of the faunal remains are also cutmarked, including mandibles apparently smashed open. I suppose this may be construed as evidence for cannibalism – at the extreme, that the fearsome modern humans were hunting down the last Neandertals. And there’s no particular reason to think that this isn’t cannibalism at Les Rois, but given the scarcity of the sample, it’s not nearly so strong as the evidence at some other sites.

The authors suggest that this may fit in with a pattern evident at other Upper Paleolithic sites, in which human remains were deliberately altered or processed for symbolic purposes. There is a perforated human tooth at the site, evidently created for use as a pendant. Some kind of mortuary practice is probably just as consistent with the scanty information we have as cannibalism.


Regardless of whether the Les Rois hominids are Early Aurignacian or somewhat later in date, they appear to represent a population that includes substantial variability not present in later Europeans, but overlapping with earlier Neandertals. That observation of variability is consistent with a mixture of populations, possibly representing a declining fraction of Neandertal-derived genes over time.

So I would guess that Les Rois represents part of a larger range of variation. Further, we should keep in mind that the morphology of Neandertals in late Mousterian or Châtelperronian contexts also has variability that overlaps with contemporary human populations elsewhere.

On that topic, the Les Rois dental remains should make us return to the other Aurignacian-associated Neandertals: the Vindija G1 Neandertals. These remains are also fragmentary, but much more substantial and numerous than those from Les Rois. The anatomy of the specimens is covered in the review paper by Karavanic and Smith (1998), which itself reiterates earlier observations made by Fred Smith, Milford Wolpoff and others. This is a younger assemblage than the G3 layer where the Vindija genetic samples were taken, and represents the final Neandertals at the site.

The dating of the G1 layer has fluctuated back and forth, as I discussed in 2006. The most recent date puts the G1 Vi 207 and Vi 208 specimens at approximately 33,000 radiocarbon years – possibly overlapping in date with Les Rois specimens, but probably a couple of thousand years older.

In addition to the physical remains, Vindija G1 is notable for the presence of several bone artifacts, including a split-base bone point, type artifact of the Aurignacian. The layer also includes some leaf-shaped bifacial points. There has been considerable controversy about whether these Aurignacian-like elements of the assemblage are actually associated with the Neandertal remains, or whether some mixing of the layers occurred due to cryoturbation (freezing-induced sediment disturbance). The strongest argument in favor of disturbance is that individuals with Neandertal-like morphologies have never before been clearly associated with Aurignacian tools. It seems to me that the Les Rois remains pretty well demolish that argument.

Vindija deserves a longer discussion than this short note. Karavanic and Smith (1998) argued that the G1 assemblage should not be called “Aurignacian” but that the Upper Paleolithic elements of it be recognized as novel parts of a regional cultural tradition with roots in the local Mousterian. That would accord with other “transitional” technocomplexes like the Châtelperronian and Bohunician, which appear to combine new Upper Paleolithic tool forms with lithic procurement and processing strategies common in earlier Mousterian assemblages.

In that light, the concept of “Early Aurignacian” deserves close examination: it seems that almost everywhere in Europe we look, we find evidence of conceptual mixture. That’s certainly true of the biological remains, and when we consider that these sites and assemblages cover thousands of years of time – extending up to hundreds of human generations – it seems hard to believe that we can’t make some more sense out of them.

Anyway, this post has gone on long enough. I have a lot more notes as background, and I’ll see if I can’t shape them up for posting over the next couple of weeks.


Guatelli-Steinberg D, Reid DJ, Bishop TA, Larsen CS. 2005. Anterior tooth growth periods in Neandertals were comparable to those of modern humans. Proc Nat Acad Sci USA 102:14197-14202. doi:10.1073/pnas.0503108102

Guatelli-Steinberg D. 2009. Recent studies of dental development in Neandertals: Implications for Neandertal life histories. Evol Anthropol 18:9-20. doi:10.1002/evan.20190

Ramirez Rozzi FV, d'Errico F, Vanhaeren M, Grootes PM, Kerautret B, Dujardin V. 2009. Cutmarked human remains bearing Neandertal features and modern human remains associated with the Aurignacian at Les Rois. J Anthropol Sci 87:153-185.