Skip to main content

Wherefore Art Thou, Neanderthal?

Just a decade has transformed the debate over whether Neanderthals were practising what we have no better word for, than art. 

In 2007 a review paper by Marie Soressi and Francesco d'Errico critically discussed much of the until-then disparate finds, and showed that in fact there was a substantial body of 'legit' evidence. Since then, researchers have upped their game through not only developing more sophisticated analytical approaches and methods, but also just keeping their minds open to the possibility. Intentionally looking out for those tiny traces which have just managed to hang on over tens of millennia can make the difference between finding them or not. 

Two new papers are being released today, and even with the different world we now find ourselves in compared to ten years ago, aspects of them still represent genuinely jaw-dropping finds. [I was asked for comment by a media outlet which is why I'm able to write all this in advance]. 

So, what's the deal, and can we trust these results?

First, let's talk about Cueva de los Aviones, Murcia, Spain, where rather than new finds we have an effective re-dating of a site which was already presented as containing evidence of symbolic behavious in 2010.
In the earlier research, a quite impressive array of activity seemed to be going on: red, yellow and orange pigments were being mixed up sometimes with pyrite, a shiny rock, and were found in the same levels as perforated marine shells. Some of the shells were collected alive, and have a mix of human-made holes, and natural perforations selected by size, interpreted as personal ornaments. 

The Cueva de los Aviones shells were dated using radiocarbon to between 45-50 Ka, but because of the fact that chronologies for incoming Homo sapiens in Europe have been pushed back in recent years (including incredible art from German sites dated to before 40 Ka), dates that are around 45 Ka are not longer viewed as 'watertight' proof of Neanderthal activity. 

The new research claims to have blown that uncertainty away, at least at this site, through dating to c. 115 Ka of a flowstone deposit (speleothem) apparently capping the archaeological deposits. This would certainly be a date that is way older than any non-Neanderthal species we know about yet in Europe. But is it really dating the archaeology?

Cueva de los Aviones is an interesting site, with a very complex formation history that unfortunately has been mostly lost, which in archaeology is never a good thing. In the supporting online material (not the main paper), the authors give more details about the context of the archaeology and the flowstone. 

The base layer at the site is a marine-deposited cobble bed that's cemented into a conglomerate, and it's assumed to represent the most recent high sea level era– the Eemian– when temperatures were warmer than today. This took place around 122 Ka, and typically deposits like this are found about 7m above present sea-level, although here due to local subsidence they're right at the current shore. 

The archaeological layer is right on top of this marine deposit, and has only been preserved at one edge of the cave mouth because it too is cemented. The rest of the cave's filling has disappeared, presumably because of massive erosion during the current post-glacial high sea-level.

Excavations in 1985 left a section, which is what the current project was able to excavate to produce the shells, and now to date. If the flowstone formed on top of the archaeological layers, then they must be more ancient than its oldest dated samples, and therefore pre-115 Ka. 

But... because there's only a tiny amount of sediment left, some questions about how the site formed are vital to answer. In the online supporting material, there is a photograph of the layers beneath the flowstone, which shows a dramatic dip downwards from the cave towards the sea. There must have been a significant slope of archaeological material that extended way back and up to the higher part of the cave, in the past. In fact we know the cave itself was previously larger, because flowstone doesn't form outside like this.

The potential problem, which the paper does not give any data to disprove, is that we know flowstone can form on top of deposits which being softer are later eroded out, leaving it hanging; and afterwards other archaeological layers can then fill that void. So it's theoretically possible that the radiocarbon dates for the shells are correct; the archaeology is about 45-50 Ka, but the flowstone records an earlier phase of sedimentation, and the pre-115 Ka deposits it formed on top of are vanished. 
For this to happen however, there would need to be an erosional agent in the cave after after 87 Ka, which is the youngest part of the flowstone (and well after the Eemian high-sea levels about 120 Ka)

There is actually some data that suggests post-Eemian sea levels did fluctuate before dropping massively during the glacial advance after 75 Ka. While lower than today's, some kinds of data show a small peak after 87 Ka. The below crop comes from a paper by Sidall et al. 2007, and shows two small peaks: around 100 Ka, and another about 80 Ka.

Sidall et al. 2007, reproduced for discussion

But let's say no marine erosion was possible. You still see records of massive erosion and hanging flowstone remnants in many inland caves, including where Eemian age deposits have been removed and the cave refilled by much later archaeological layers. Creswell Crags in the UK includes sites where this seems to have happened.

I have no problem per se with the use of pigment or perforated shells going back to 115 Ka for Neanderthals. We know they were collecting raptor talons at Krapina which is also about the same date, and at the much earlier site of Maastricht-Belvédère a liquid haematite mix was identified in 2012, the pigment for which must have come from at least 40 km away. In fact at Cueva los Aviones, if the deposits formed during an era where sea-level was not dramatically lower than today, then the sea would have been much closer and the transport distance is much more feasible than the c. 7 km if the shells were really aged 45-50 Ka. But there needs to be more work done to demonstrate that the archaeological layers are in fact those the flowstone originally formed on top of; perhaps micromophological analysis could demonstrate interdigitation of the flowstone and underlying layer. Plus, extra dating methods for the artefact-bearing layers themselves is also a priority, to prove the radiocarbon results really are minimums.

Let's move onto the next paper, which is in my opinion much more "wow". It reports dates up to 65 Ka from three Spanish caves – La Pasiega, Maltravieso and Ardales– for red pigments used in painting. These sites are highly significant for three reasons:

  • The oldest dates, although not as old as Cueva de los Aviones, are well past the fuzzy 45-50 Ka zone, before which we (currently) believe only Neanderthals were present in Europe.
  • The fact that pigment is being used on actual cave walls and flowstone formations is entirely new- until now, the only direct altering of the surfaces of a site by Neanderthals was known from the Gibraltar engraving, which is older than 40 Ka (and covered by Neanderthal-associated archaeology).
  • Not only is there broad application of pigment across surfaces (at Ardales), but we also see for the first time paint being used to make images. At La Pasiega there is a straight line (part of a more complex panel, but we'll come to that), while Maltravieso is truly a spine-tingler: a negative hand impression, made by placing the hand against the wall and spraying or daubing paint around it, to leave an empty shape. 

For all three sites, its the flowstone covering the paintings that was dated, via the uranium-series method (same as at Cueva de los Aviones). In the case of Ardales where pigment was spread across stalagmite and flowstone 'curtain' formations, it was possible in some cases to bracket-date the painting: by taking samples from beneath and on top, you get a maximum and minimum age. But for the other sites they only sampled above the art, taking several in sequence where possible to get a micro-stratigraphy, which in nearly all cases gave a correct succession of increasingly old dates closer to the paintings.

 The headline dates are:

  • ARD-13 from Curtain 8 at Ardales, where the outer sample had an age of  >46 Ka, but the inner sample is >65.5 Ka. Other dates from different areas have minimums that go down to 39-36 Ka.
  • PAS 34 from the vertical line at La Pasiega had two upper samples that gave a minimum age of 52 Ka, and a lower date closer to the painting which was less pure, and with a high degree of uncertainty, which gave 79.7 +14.9 ka or minimum age of 64.8 Ka.  
  • Minimum age range form five samples near the thumb of the Maltravieso hand of 66.7 Ka – 55.2  

Until now, the oldest dated paintings in caves anywhere were a handprint from all the way in Borneo dated to 39.9 Ka (plus a probably slightly younger animal), and a red disc with an almost identical date from the site of El Castillo, literally round the hill from La Pasiega. These new dates therefore push the earliest yet known painting cave art back by at least twelve thousand years.

While the line at La Pasiega is striking, let's take a moment to indulge the incredibleness of the Maltravieso hand: we've got a few blurry footprints from European caves that are probably Neanderthal, plus a very cool thumb print from birch pitch at a site in Germany. But this is the real, ghostly impression of a whole living person's hand. 
It's impossible not to try and look at the shape for differences- are the fingers thicker (their finger-tip bones were more robust) or shorter? It's hard to say; it's quite obviously a hand, held sideways, fingers not especially spread, and the hand from Borneo certainly looks much more gracile in comparison. But the pigment is very faded (the researchers had to adjust the lighting in the image to bring out the details), and is nearly covered by flowstone.


Once again most of the information for the painted sites is stuffed into the supporting online materials, and there's a lot of data to pick through in there. Is any of it problematic? I'm not an expert in this or any dating method, but didn't understand why at Maltravieso they sampled cave dirt to provide correction values, but not at La Pasiega, where they used a generic value. This may not be significant, but it seemed strange not to use the same approach at all sites, especially when there were issues of detrital thorium there too. Beyond questioning the u-series method itself however, there aren't obvious contextual issues like there may be at Cueva de los Aviones. But I still have various thoughts and questions.   

First, the dating: at all three sites, there was rather a large range of ages from flowstone samples that were spatially separated across only a very small area, suggesting that the flowstone formation is quite uneven, and the place you choose to sample can really affect your results. This does however only imply that the possible 'real' ages might in fact be older. 

Second, the authors note that lines and handprints are incredibly common features in European caves, but when they are found as part of panels made at several points through time, they appear to have been done first. This raises the possibility that there's a hidden layer of Neanderthal art in many more sites, underneath Upper Palaeolithic additions. And that is an intriguing notion: the first Homo sapiens entering Europe did not find ‘empty’ caves, but walked into the dark to discover red markings and even handprints of others before them. And were then moved to add to this record themselves. 

All three sites would seem provide an astonishingly long chronology of human engagement with the caves; just taking La Pasiega, there are dates elsewhere in the cave for stylistically Upper Palaeolithic art ranging from around 22 Ka (the height of the last glacial maximum) through to late prehistoric; some panels are obviously multi-phased.
And the panel containing the line is a case in point. Described by early researchers in 1913 as “The Trap”, the whole thing includes not only a large grid-like design which the line in question is part of, but also partial animals in the middle, a fat inverted triangle below, many dots above, and undefined features to the right. But the u-series dates are incredibly disparate. The paper also reports for the right-hand vertical line, which appears to be no more vibrant in hue than the claimed Neanderthal line, a vastly younger minimum Bronze Age date (c. 3 Ka). 

Part of the 'Trap' panel at La Pasiega. The very anciently dated vertical line is the one the deer's head is facing. The much mrore recent date is for the other vertical line. The dots to the upper left are missing, as is an unclear image to the right. Source:

The dots too– which do seem to overlie the grid– have given drastically varying minimum ages too, from about 12 Ka to only 2 Ka; they do look less dark in colour than the lines; and the animal figures are also more finely drawn and consequently fainter, but remain undated. However to the left of the grid, there is a large deer-like outline; the flowstone here is only about a thousand years old. Is this a case where a Neanderthal-drawn line (and possibly other undated parts of the that panel) were embellished many tens of millennia later? Could the matching vertical line that visually looks so similar be almost 50,000 years younger? There is evidence that multi-phase painting might have been happening much earlier too at Ardales, with date ranges that could point to different phases of activity on different curtains (c. 45 Ka and much older c. 65 Ka).

It is possible that the flowstone on the "trap" panel simply grew at massively different rates in different parts of the painting. Yet if that’s true, why isn’t the oldest line more thickly covered than the other parts of the paintings? A
better understanding of the growth and chronologies of flowstone in caves like this would be helpful.

What’s clearly needed aside from much more detailed dating of the panel – in my opinion the destructive aspect is justified when there is one massively older measurement, and its so potentially important– together with technological analysis of the different parts of the images: are the pigments different recipes, or maybe applied differently? Does the dating match any apparent superpositions of the paint? And the same kind of analysis should be done for Ardales.

 Some other questions occur too, first about the archaeological context: 

·       Maltravieso does have Neanderthal-associated archaeology but it seems to be rather older than the dated hand image; Ardales does too and some of it is in a trench only a few metres from one of the painted areas. The pigment is described as 40-140 cm above the present-day cave floor level, so it would be interesting to know what level the Neanderthals were living at- how accessible were the painted areas ?

·       La Pasiega doesn’t have any clearly identified Neanderthal archaeological layers as far as I can see, although there is something from deep in the deposits that’s previously been called “possibly Mousterian”. But we do know from the Bruniquel discovery last year, where Neanderthals about 174 Ka were spending hours breaking up and constructing circles from stalagmites, that they sometimes did unusual things deep in sites that they didn’t seem to live in.

What does this all mean for Neanderthal archaeology ?

The Krapina raptor talons: Image: OA article- Radovčić D, Sršen AO, Radovčić J, Frayer DW (2015): "Evidence for Neandertal Jewelry: Modified White-Tailed Eagle Claws at Krapina." PLoS ONE 10(3): e0119802

If the dates are accurate, there are a number of implications.

  • Cueva de los Aviones, as the authors state, would now pre-date any evidence from early Homo sapiens contexts in the Near East and Africa for either processing of pigment recipes or the collection and piercing of shells in association with pigment use.  This matters for the ongoing dates about whether Neanderthals were on an independent trajectory to symbolically complex behaviours, and it joins the growing body of evidence from Europe that this was the case ranging from pigment on shell at other sites, stalagmite constructions at Bruniquel, France, collection of raptor feathers and talons in many regions.

  • Ardales seems to fit into what we see at Bruniquel- Neanderthals being drawn to interact in deep caves with the flowstone formations there. The use of pigment, even if it’s non-graphic, is a new development and increases the diversity of these activities from fragmenting, building and engraving to painting. And it also connects to a 2002 article reporting pigment inside natural stalagmite 'cups' from a Middle Palaeolithic context in Romania.

  •  The fact that at Ardales the pigment was sometimes only visible as a line within broken pieces of flowstone means that red-painted surfaces became hidden by later deposits. It must have required a painstaking survey of the site to find this, and the idea of more ‘entombed’ Neanderthal paintings existing out there is sort of spooky.

  • Both the hand impression and the line are potentially hugely important in showing complicated uses of pigments on cave walls, which is a first for Neanderthals. The line would fit with the linear markings we’ve seen engraved on some bones, and also the Gibraltar engraving; in that sense it also recalls the earliest markings we see in the African contexts.

  • But beyond that the Maltravieso hand is something even more astonishing. All of the rich symbolic meanings ascribed to later handprints and negatives impressions in Upper Palaeolithic caves must be considered for Neanderthals: sense of self, marking of places and more.
    I can also see it as potentially giving us a bridging concept to what would be the single remaining lacuna in the Neanderthals' symbolic repertoire: representational images. They might not have been drawing images of animals (that we know of!), but they were using paint to create a representation of the body, whatever the meaning was. 

The Bruniquel stalagmite constructions. Image: By Lhfage (Own work) CC BY-SA 4.0, via Wikimedia Commons

A last point is that the conclusions of both papers tack on the claim that their findings lay to rest the question of who made the so-called transitional cultures: the Chatelperronian and others found in Europe c. 45-40 Ka, just before the earliest accepted Upper Palaeolithic. This addendum is unneccessary and also rather misses the point. Those who will accept any form of Neanderthal symbolism have already been able to point for a while to similarities between some of the objects in the transitional industries and things Neanderthals were doing much earlier, such as making bone tools, stringing eagle talons together, or engraving various substances. There is also no known cave painting in the transitional industries. More fundamentally, they also don’t help us with the detailed debates over the lithic technology which still remain  for the 'transitional' industries, or the problematic complex taphonomy for key Chatelperronian sites with Neanderthal skeletal material which– despite recent re-dating– are still difficult to be completely certain don't represent mixed assemblages due to earlier research histories.

What these new finds do is potentially extend our knowledge of what Neanderthals were up to well before– as far as we know– H. sapiens arrived, and once again force us to consider them on their own terms.


D. L. Hoffmann, D. E. Angelucci, V. Villaverde, J. Zapata, J. Zilhão,. 2018. Symbolic use of
marine shells and mineral pigments by Iberian Neandertals 115,000 years ago. Sci. Adv. 4, 5255

D. L. Hoffmann,1* C. D. Standish,2* M. García-Diez,3 P. B. Pettitt,4 J. A. Milton,5
J. Zilhão,6,7,8 J. J. Alcolea-González,9 P. Cantalejo-Duarte,10 H. Collado,11 R. de Balbín,9
M. Lorblanchet,12 J. Ramos-Muñoz,13 G.-Ch. Weniger,14,15 A. W. G. Pike. 2018. U-Th dating of carbonate crusts reveals Neandertal origin of Iberian cave art. Science, 359, 6378

M. Siddall, J. Chappell and E.-K. Potter. 2007. Eustatic Sea Level During Past Interglacials. In The Climate of Past Interglacials.  Edited by F. Sirocko, M. Claussen, M.F. Sanchez Goni, T. Litt. Developments in Quaternary Science, 7: 75- 92


Popular posts from this blog

Adventures in Silcrete: "It's flint Jim, but not as you know it!"

Something that everyone who works in the archaeology of deep prehistory has to get to grips with is the technology of stone tools, or lithics. This includes thinking about the ways in which people made their tools, which techniques they chose to use, etc. It also means that Palaeolithic archaeologists, alongside needing to know stuff about climatology, palaeontology, and ecology, need to delve into the science of geology. People in prehistory might not have understood the origins of different kinds of rocks, but they certainly appreciated the diversity in stone qualities, not only between very different rock types but also within geological/mineral categories.

These two Neandertal tools that I studied for my PhD, called handaxes, are both very finely worked, but made from completely different rocks. The one on the left (Castle Lane, Bournemouth) is made from Cretaceous flint found in the south and east of Britain, and the one on the right (Coygan Cave) from rhyolite, a volcanic stone…

Geological Road Trip: Volcanic landscapes of the Massif Central

Geology and geography are fundamental to archaeologists in understanding the landscape contexts that people of the past lived within. While climate and environments have drastically altered over the time span of the Palaeolithic, the topography often, on a broad scale, remains relatively similar. Erosion can be extensive, river systems can change course (the Thames used to flow much further north than it now does for example), and the great depth of sediment accumulation in some areas changed local situations. But the big stuff made of rock like plateaux, mountains and watersheds have remained relatively static over the time hominins have been around. There are exceptions to this however, primarily in the form of volcanism and tectonic action, and the region I'm working in is a textbook example. Here in the Massif Central, there is a long history of volcanic action of many types, the most recent of which occurred less than 5000 years ago- well within the history of human settleme…