For those only reading the RSS: Magdalenian mtDNA H in Cantabria confirmed by RFLP testing

Just a short heads up mostly for RSS-only readers to an important update/edit of one of today's entries:

Ancient mitochondrial DNA from the Basque Country and Cantabria

Now renamed as: Ancient mitochondrial DNA from the Basque Country and Cantabria: unmistakable mtDNA H in Magdalenian Cantabria

... on Hervella et al., 2012 (oa).

The key element of novelty is that the authors themselves have confirmed to me by email that they did test for RFLP markers, specifically Alu I in the case of mtDNA H. 

Therefore, against all rumors and confusion, we finally have unmistakable confirmation of the presence of mtDNA H (and H6) in Magdalenian Europe, specifically in Cantabria. It is not any Neolithic arrival to Europe and the studies (not peer-reviewed) of Chandler 2005 and Kéfif 2005, both reporting high mtDNA H in pre-Neolithic Portugal and Morocco can now be seen with a bit more credibility, with less systematic doubts.

For more details read the edited section in the relevant entry.  

Thanks again to the authors for the quick clarification.

Ancient mitochondrial DNA from the Basque Country and Cantabria: unmistakable mtDNA H in Magdalenian Cantabria

This seems a very important paper because, if everything is correct, it is the first peer-reviewed publication to establish conclusively that mitochondrial haplogroup H, the most common matrilineage of Europe today, existed in Paleolithic populations in Europe, specifically in Cantabria, Spain.

There have been previous reports of such important presence of mtDNA H in the area, notably Chandler 2005 (on Epipaleolithic and Neolithic Portuguese) and Kéfi 2005 (on Upper Paleolithic Moroccans). However both studies were informal for some reason, what has been used by some to grasp for straws in their claim of this lineage having arrived to Europe (and North Africa) only after Neolithic and oversimplify the Paleolithic genetic pool of Europe to almost only variants of haplogroup U (never mind that U is such a large, ancient and widespread haplogroup that saying "U" is almost like saying nothing).

Montserrat Hervella et al., Ancient DNA from Hunter-Gatherer and Farmer Groups from Northern Spain Supports a Random Dispersion Model for the Neolithic Expansion into Europe. PLoS ONE 2012. Open access.

Edited section: the authors did test for RFLPs, the haplogroup assignation is as good as it can be.

After the first read to the paper my impression was that they had only tested the HVS (both subsectors I and II), leaving some, even if smaller, doubt about the adscription of the Cantabrian Magdalenian lineages to mtDNA H.

But, luckily for truth, Jean Lohizun (see comments) found that the authors, who are partly the same team as Izagirre & De la Rúa 1999 and are using the same methodology up to a point, have actually tested not for coding region SNPs but for a standard list of RFLP markers. However the mention is so cryptic that it is most difficult to find: 

In order to classify the mitochondrial variability of the individuals analyzed in this study, we proceeded to amplify 11 markers, which are required for defining the 10 Caucasian haplogroups described [63]. The protocol and primers are described in [17], [20], [42]. The digestion patterns were verified using a fragment Bioanalyzer (Agilent Technologies).

This seemed to imply that when the authors say "H" they mean -7025a (aka -Alu I) but this marker as such are listed nowhere that we can see easily.

So I squeezed my mind until I concluded that Jean must be correct but a very slight doubt remained, so I emailed the authors, who quickly responded confirming Jean's keen reading.


Excerpts and translations:


Concepción De la Rúa (corresponding author) said:

Para conocer con más seguridad el haplogrupo, se hicieron además  RFPLs, que es una tecnica que permite evaluar la existencia de mutaciones puntuales en la region codificante del mtDNA.

Translation: In order to know with greater certainty the haplogroup, RFLPs were performed, which are a technique that allows for evaluation of punctual mutations in the coding region of mtDNA.

She added in relation to a previous study of her authorship:

Ambos analisis se realizaron en el paper recien publicado, pero no en el paper de Izagirre & de la Rúa de 1999, en donde solo se hicieron RFPLs.

Translation: Both analysis [RFLPs and HVS] were made in the recently published paper but not in that of Izagirre & de la Rúa from 1999, where only RFLPs were tested for.

Montserrat Hervella (lead author) said:

También, hemos realizado PCR-RFLPs con el fin de determinar los SNPs de la región codificante y así reconfirmar el haplogrupo al que pertenecen cada uno de los haplotipos que hemos obtenido mediante secuenciación. Los enzimas que hemos utilizado aparecen en varias citas bibliográficas en el articulo (17, 20, 42, 63).

La cuestión que nos plantea en torno a la determinación del haplogrupo H, la respuesta es que si hemos usado la enzima Alu I para determinar el SNP en la posición  7025 de la region codificante del ADNmt.

Translation: We have also performed PCR-RLFPs with the purpose of determining the SNPs from the coding region and therefore confirm the haplogroup to which each of the haplotypes obtained by sequenciation belong. The enzimes we used appear in several bibliographic citations in the article (17, 20, 42, 63).

About the question that you ask on the matter of the determination of haplogroup H, the answer is that we have indeed used the enzyme Alu I to determine the SNP in position 7025 of the coding region of mtDNA.

This should dispel any doubt: there was mtDNA H (and H6) in Magdalenian Cantabria.

(End of the edited section)

In any case these are the results as reported by the researchers (and synthesized by me in a single image):

Notice that the dots are not really precise in their geographic locator role (I may try creating a better map later).

Instead of using numerals I repeated the name of the reported haplogroup with the intention of giving a more visually intuitive impression (without going through the process of creating charts). 

The results for Paternabidea (tested for coding region markers) were discussed in November. They basically seem to establish that modern Basques are almost identical to Neolithic Basques.

But what about Paleolithic ones? The results of this one study rather suggest that a little bit of both: on the side of continuity, the same U5 sequence is reported for Erraila (Magdalenian) and Marizulo (Neolithic) but, on the side of change, all Paleolithic Basque sequences (n=2) are U5 lineages, while U5 today is a much less important matrilineage. 

It may well be argued that this result is a fluke, a coincidence, but it is consistent with what we see in Epipaleolithic Central and North Europe (mostly U5 with some U4, Bramanti 2009), suggesting that, while the model of recolonization of much of Europe from the Franco-Cantabrian region may well be correct, the story does not end there at all.

On the other side we see that the dominant lineage today among Basques and Western Europeans in general, haplogroup H, was already present in Magdalenian times in Cantabria, what is consistent with it being found in Epipaleolithic Portugal and Oranian Morocco and really puts to rest the hypothesis that promoted that it had arrived from West Asia with Neolithic colonists (something argued as a matter of fact but without any clear evidence). 

Food for thought (a hypothesis)

As of late, although I have yet to formalize it somehow, I've been thinking that a serious possibility is that mtDNA might have spread partly with Dolmenic Megalithism. While there is some apparent Neolithic expansion of H in the Basque area (apparent in this paper and resulting in an almost modern genetic pool), in other parts of Europe this is less obvious, with H showing up but not reaching modern levels just with Neolithic. In fact the loss of other Neolithic lineages like N1a strongly suggests that the populations of Central Europe were largely replaced after the Early Neolithic. Where from? Again from Southwest Europe, I suspect but not in the context that was once believed of Magdalenian expansion, but maybe in the context of Megalithic expansion instead, with origin not in the Franco Cantabrian region but in Portugal. 

This is just a draft hypothesis that I have mentioned before only in private discussions or at best in the comments section somewhere, and certainly it would need more research. My main argument is that before these results for Cantabria, the only pre-Neolithic location where the data clearly suggested very high levels of mtDNA H (near 75%) was Portugal (Chandler 2005) and Portugal played a major role in Neolithic and Chalcolithic Western Europe. As I said above, ancient Portuguese apparently developed the Dolmenic Megalithic pehnomenon (culture, religion...), later they developed some of the earliest Western European civilizations, since the Third millennium BCE, specially Zambujal (also at Wikipedia), which were central elements of this long-lasting Megalithic culture and later of the Bell Beaker phenomenon as well.

And we need very high levels of mtDNA H in a colonizer population to changed the genetic landscape from c. 20% H into c. 45% H, we need something like a 70-80% H ancestral population unless replacement was total, what I think unlikely. 

What happened to that overwhelmingly H population of Portugal (assuming that the hypothesis is correct)? They were probably colonized in due time, possibly in the Bronze Age (the mysterious archaeological "horizons" that replace urban life in much of Southern Portugal in that period with their strange crab-shaped elite tombs, vaguely resembling Mycenaean circular walled ones)  and/or in the period of Celtic invasions from inland Iberia later in the Iron Age (Hallstat periphery).

It's just a tentative hypothesis so far but one that I believe I must state for your consideration.

Air view of some the acropolis of Zambujal (the roofed area is a modern farm that sits atop)

Swedish Neolithic autosomal comparisons

In brief: it suggests Iberian-like and Russian-like origins for two different Neolithic populations of Scandinavia.

Pontus Skoglund et al., Origins and Genetic Legacy of Neolithic Farmers and Hunter-Gatherers in Europe. Science 2012. Pay per view.

Abstract

The farming way of life originated in the Near East some 11,000 years ago and had reached most of the European continent 5000 years later. However, the impact of the agricultural revolution on demography and patterns of genomic variation in Europe remains unknown. We obtained 249 million base pairs of genomic DNA from ~5000-year-old remains of three hunter-gatherers and one farmer excavated in Scandinavia and find that the farmer is genetically most similar to extant southern Europeans, contrasting sharply to the hunter-gatherers, whose distinct genetic signature is most similar to that of extant northern Europeans. Our results suggest that migration from southern Europe catalyzed the spread of agriculture and that admixture in the wake of this expansion eventually shaped the genomic landscape of modern-day Europe. 

I don't have access to the paper as such but the substance seems to be, as so often, in the supplementary materials, which are freely available (direct PDF link). 

The study analyzes the autosomal genetic signature of four prehistoric Swedes from the Neolithic period. Three are from Götland and belong to the regressive (neo-hunter, forest Neolithic) Neolithic culture of Pitted Ware, which has its roots probably in the Neolithic of Eastern Europe (Dniepr-Don culture). While they have been used as proxy for Paleolithic peoples, the comparison is not really valid: they were probably recent immigrants from Eastern Europe (or admixed with the Paleolithic locals maybe) with a Neolithic background although, because of the peculiarities of the Baltic frontier, they were not much into farming (it seems), although they did keep pigs. 

At the right, meet what is left from one of them, known as Ajvide 70. We can see some of the characteristic burial traits common in Pitted Ware culture but also in their apparent ancestors from Dniepr-Don culture (farmers): burial in extended position with ochre (visible at the legs in this case). 

The fourth one is from South-Western Sweden, near Göteborg and is consistently described as "farmer". This individual belonged to the Megalithic cultural phenomenon (or more properly Dolmenic Megalithism, as there are some megalithic structures that may not be related or only tangentially so) that began in SW Iberia c. 5000 BCE and which is characterized by collective (clannic?) burial in dolmens, where the older burials were pulled aside to make room for the new ones. They also buried their people in extended position although the use of ochre is less common. 

Dolmenic Megalithism extended since the fourth millennium (at least a thousand years after it was conceived in SW Iberia) through much of Atlantic Europe, being these settlements of Southern Sweden in their further North-Eastern reach (together with Denmark and much of Low Germany). Later, after the loss of those areas to the Indoeuropeans of Kurgan cultural background (Corded Ware) the phenomenon would also expand into the West and even Central Mediterranean (North Africa, Italy)... but that's another story.

Both burial traditions stand in contrast with those of the Neolithic cultural arrivals from the Balcans (and possibly earlier from West Asia), the Balcano-Danubian tradition and the Cardium Pottery one of most European Mediterranean areas, which practiced burial in flexed or fetal position. 

So let me emphasize my point: both groups studied here are Neolithic but both belong to cultures which are not exactly what we could describe as "core Neolithic" but rather cultures that appear to have at least partial roots in Paleolithic Europe: one from the East and the other from the Southwest and later all the West. 

Interestingly the autosomal genetics of these two groups are consistent with what I just outlined: the Pitted Ware people resemble Eastern Europeans, while the Megalithic farmer clusters best with Iberians and other SW Europeans.

Visually (all from the supplemental materials):

Cluster analysis (K=4):

click to enlarge

We can see that the dominant component of the Megalithic farmer is orange, which is dominant in Sardinia but almost lacks the yellow component dominant in West Asians (and to lesser extent North Caucasus and Italy). The not compared Iberians could be a good match as we can see below.

Two of the three Pitted Ware individuals cluster best with Russians, while the third does instead with some Finnish maybe. 

The main difference with modern populations in general is that the yellow (West Asian) component is residual, exactly as happens among modern Basques, Sardinians and Finns, which may have been rather isolated all this time. 

Several Principal Component analyses:

Judge yourselves because these comparisons are never 100% conclusive, but my impressions is that the Megalithic farmer (red) clusters mostly with Iberians, except in the last graph where it does with the British-CEU sample.

The Pitted Ware samples (cyan or sky blue) are less precise (they are three after all) but where Eastern European samples other than Finnish are available for comparison, they tend to cluster with them (or between them and Sweden and Finland). In the last panel, where Swedes have been oversampled and non-Finnic Eastern Europeans removed, Ire8 and Ajv52 cluster withing modern Northern Swedes, while Ajv70 is anomalous. 

So I do not think that the conclusions publicited in the press release (which are probably also those of the paper) can be taken at face value but rather with a good pinch of salt: the factual data seems to suggest that the genetics of these prehistoric people are quite related to that of modern Europeans from Scandinavia, Eastern and Western Europe. The comparison cannot be expected to produce identical individuals because, no matter how hard we might try or how much we may owe them in a sense, we are not our ancestors.

Ancient DNA from Santimamiñe (Biscay): not yet published but on the way

Researchers from the University of the Basque Country (UPV-EHU) have already sequenced but not yet published, it seems, mitochondrial DNA from the Paleolithic inhabitants of Santimamiñe, a cave with a lengthy sequence from the Gravettian (and earlier Mousterian and Chatelperronian but these are supposed to be Neanderthals) to the Iron Age. 

The cave is located not far from Gernika, in the municipality of Kortezubi and hosts some rock art of bisons, horses and bears, all located in a small hidden corner that I had the luck to visit as a kid (now it's not accessible to the general public for preservation reasons).

At the moment of writing this, I am not very sure to which period the human remains belong. Publications usually emphasize the techno-cultural and artistic aspects and unless a skull would have been found, what is not the case, there is little emphasis on human remains. However nowadays, with the advent of genetics, they get a new significance as, at least potentially, links to modern and other prehistoric people can be established or discarded, producing potentially valuable information for the understanding of Prehistory. 

On May 6th the research team is heading to the area to collect samples from maternally-rooted locals, in order to compare with the already decoded but not yet published ancient DNA sequence. 

I can't way to read the results, you can imagine. But even more because one of my ancestral lineages, the patrilineal one, is from a nearby village: Ibarrangelua.

I just hope that they have done the testing properly and not just sequenced the control region (HVS), which is often more confusing than revealing. 

Source[es]: Deia (found via Pileta).

Update (Jun 8): published in French but it is from the Neolithic, discussed here.

Key remains of American Prehistory stolen

The Tulum man in situ before the robbery

The skeleton known as Tulum man, found in a cenote (flooded underground chasm) in Quintana Roo, Mexico, and which was not yet retrieved from site by archaeologists, has gone missing. 

The Tulum man was dated to at least 10,000 years ago and therefore was one of the key pieces in the reconstruction of the Prehistory of America.

The managers of the research project, dependent of the National Institute of Anthropology and History (INAH) have denounced having been repeatedly the victim of thieves. That in spite of having kept a policy of strict discretion about the exact location of the findings. 

Requests of help have been distributed among divers because such a heist would have needed the participation of several of them.

Source: Milenio[es].

A qualified opinion on the Eastern Gravettian

The always interesting Aggersbach's Paleolithic Blog offers today an entry pondering the Eastern Gravettian and the so-called Willendorf-Kostenki complex. You probably want to read it in full but in any case I appreciate a pondered opinion on these matters, which may help us to understand better the overall process of the European Upper Paleolithic. 

Stone saws from Kostenki

My pick is:

In my view, the extremely rich archaeological record of the east European plain clearly supports the two-stage concept of an eastern Gravettian with occasional leaf point production, followed by a Gravettian with (Micro)-Gravettes, backed microliths and shouldered points (Mitoc Malu Galben in the Pruth valley; Molodova 5, layers VIII and VII; Molodova 1, lower layer; Korman’ 4, layers VII and VI; Voronovitsa 1, upper layer; and Babin I in the Dnester river basin, Khotylevo 2 in the Desna river basin).

More at Aggsbach's Paleolithic Blog.

Also mentioned, as the best approximation to a non-existing online paper in English about the Willendorf-Kostenki complex is an article at another fabulous and veteran site: Don's Maps, from which I'm borrowing some illustrations to complement this entry.

Reconstructed Kostenki tent/home, one of those not built with mammoth bones

Cis-regulatory variability has almost no geographic structure

That is what a new paper has found:

Barbara E. Stranger et al., Patterns of Cis Regulatory Variation in Diverse Human Populations. PLoS Genetics 2012. Open access.

Cis-regulatory elements are key pieces of the overall codification of the genome. The main finding of this paper is that they show almost no geographic structure, that virtually all variation in these key elements is inter-individual and not among populations.

From Supplemental Figure 1 - PC analysis

As you can see in the PCA plot, the inter-population variation is most subtle, almost impossible to discern with so much overlap. The lack of structure repeats for the other components.

Laos inhabited by H. sapiens since maybe 56,000 years ago?

Just a quick note because I really could not find too much.

I just read at a very succint note at Science Daily that:

Researchers have discovered the oldest known human remains in Southeast Asia, a partial human skull dating to at least 40,000 years ago. Excavations at Tam Pa Ling cave in northern Laos produced a dozen pieces from a Stone Age person’s skull, including a skullcap and a lower jaw, anthropologist Laura Shackelford of the University of Illinois at Urbana-Champaign reported April 14. Small front teeth, a rounded brain case and other traits identify the reassembled fossil as a modern Homo sapiens, Shackelford said. The find supports proposals that at least some human migrations out of Africa around 100,000 years ago followed a southern route that led to Southeast Asia.

Nothing more of relevance. I could find a paper (ppv) by Dr. Shackleford on Laos but it discusses another cave and much more recent dates (c. 15 Ka ago).

On the other hand I found a paper on another Laotian cave by different authors which mentions undefined human presence c.56 Ka ago:

Valéry Zeitoun et al, Multi-millennial occupation in northwestern Laos: Preliminary results of excavations at the Ngeubhinh Mouxeu rock-shelter. Comptes Rendus Palevol 2011. Pay per view.

Abstract

With over half a century of political instability, resulting from armed conflicts, decolonisation and the Cold War, archaeological investigations in Laos have been rare, leaving little more than a blank page in the chapter of Southeast Asia's prehistory. Recent research has shown that Laos holds a rich prehistoric heritage. In conjunction with the research initiated by J. White who conducted the first professional archaeological survey of northern Laos since decades, we have extended the investigations to the Luang Namtha province. This work allowed us to gather important data about Hoabinhian stone tool assemblages and former cultures. In particular, the archaeological remains and dating from the Ngeubhinh Mouxeu rock-shelter indicate that this mountainous region of Laos has been inhabited over a long period of time that possibly spans as far back as 56,000 ± 3000 BP.

While, by the moment, I ignore many details, I believe it is very interesting to mention as the early colonization of SE Asia by our species is still not well understood (and yet may be critical to understand all, or at the very least much, of Eurasian, plus aboriginal Australasian and American origins).

With due caution, I think we can infer from these and other less direct (Indian and Arabian archaeology) or more controversial materials (Liujiang skull, Luzon foot bone, speculations about very early colonization of Australia) that the colonization of SE Asia by our kind can be traced to at least that date of 56 Ka ago, probably even earlier.


Update (Aug 21): see this newer entry.

Epipaleolithic paddles found in Danish waters

Paddles are only part of the hoard of objects made of organic materials rescued from what once was an Epipaleolithic settlement of the Ertebølle culture (c. 5400-3900 BCE): undamaged antler axes (right), wooden knife handles and the skull of a dog had also been preserved underwater at that location, with low oxygen levels, just off the coast.

A quite spectacularly preserved paddle:

Video of cleaning a piece of wood at the Moesgård Museum (no sound):

Source: Science Nordic.


Update (Nov 2012):

A new article also at Science Nordic discusses the paddles and how they have marks (three bars at the side) that are similar in concept but different in the detail of the execution. The shape of the paddles is also different.

My intuition wanted to see in this the execution by people connected by culture and maybe even direct family relationship but separated in time. Although not too much, maybe just one century... Just speculating a bit - other interpretations are perfectly possible, of course.

Claim that Japanese are 60-72% Neolithic

Jomon clay head

An open access letter claims that modern Japanese are 2/3 of Neolithic ancestry (except Ryukyuans, who'd be 2/3 Paleolithic instead).

Yungang He et al., Paleolithic Contingent in Modern Japanese: Estimation and Inference using Genome-wide Data. Nature 2012. Open access. 

The explanation is however not really clear for me and, looking at their own data, I can't really accept such conclusions easily:

Only the yellow component (at K=4), almost totally absent in the Ryukyuans and dominant among North Chinese and Koreans, the likely parent populations of the Yayoi farmers, can be considered to inform the input of such immigrants to Japan. The exact apportions are not detailed anywhere in the letter but it seems to be c. 35% among Koreans (KR-KR) and Shanghai Chinese (CN-SH) and slightly above 20% among North Han (NHan, CHB). 

By comparison mainland Japanese (Japanese, JPT, JP-ML) show c. 10% in most cases. IF the parent proto-Yayoi population would be Koreans, then Japanese would have less than 1/3 Yayoi blood, while if the proto-Yayoi is equated to Northern Han instead, then the result would be at most 50%. 

In the case of Ryukyans, the Yayoi input would be negligible, almost zero. They'd be almost 100% Jomon, assuming this concept applies to the Ryukyu islands at all.

In truth I do not know what to think of this article other than it seems confusing and inconsistent with its own data.

Mitochondrial DNA revision and the Reconstructed Sapiens Reference Sequence (RSRS)

The authors and maintainers of PhyloTree have produced a new build (no. 14) with a long list of changes. However the most important part is probably in an associated paper:

Doron M. Behar et al., A “Copernican” Reassessment of the Human Mitochondrial DNA Tree from its Root. AJHG 2012. Freely accessible when writing this. 

The Copernican adjective is because they hope to replace the use of the rCRS (revised Cambridge Reference Sequence, with the highly derived haplogroup H2a2a1, and arbitrary yet convenient pick back in the day) with the Reconstructed Sapiens Reference Sequence (RSRS) which is also the haplotype of the matrilineal most recent common ancestor of extant Humankind, alias 'Mitochondrial Eve'.

The RSRS is also quite directly comparable to the RNRS (the 'Neanderthal Eve', see Fig. S1), being separated by 122 mutations of the coding region and a few more of the control region (HVR). However while between the 'Neanderthal Eve' and actual Neanderthal sequences there are some 17 coding region mutations (14-21), in our species the distance to the most recent common ancestor is longer, 40-50 coding region mutations (most commonly, from memory) and even more when we count the HVR, as illustrated in figure S2:

Fig. S2 - Distances in Substitution Counts from the RSRS to Extant Haplotypes

This image is particularly interesting because it illustrates what the authors describe as indications for violation of the molecular clock: differences in the in the length of the various branches, which vary wildly from circa 40 to above 70, a difference of almost 2:1 between the extremes.

Oddly enough, in spite of M being closer to the root than N and having on average more mutations to present day sequences, the authors manage to somehow assign a younger age to this haplogroup than to its sister N. But in any case their estimates must be wrong overall because the age attributed to L3'4'6 (71 Ka BP), L3'4 (64 Ka) and L3 (67 Ka) are more recent than the known archaeological evidence for  the migration out-of-Africa which, at the latest, must have happened c. 90-80 Ka ago. The L3 node must be older than these dates therefore.

The authors also propose some nomenclature terminology, notably the use of superscript n for the nodal haplotypes (as opposed to just unclassified sequences, for which we will still use asterisk). That way we can more easily discern a nodal or root H haplotype (Hⁿ) from a mere random unclassified H haplotype (H*). Other proposals for the nomenclature may be more debatable and in some cases they manage to violate them themselves in the supplemental materials.

Whatever my criticisms (everything is debatable and I love a good discussion), I must say that the PhyloTree team deserves my utmost appreciation and respect: before them understanding the human mtDNA landscape was a total mess, now it is possible even for amateurs with a keen interest like myself. Thanks a lot.

See also: Mitochondrial DNA and molecular clock.

Acheulean use of fire confirmed

That is what Francesco Berna and colleagues argue in a new paper:

F. Berna et al., Microstratigraphic evidence of in situ fire in the Acheulean strata of Wonderwerk Cave, Northern Cape province, South Africa. PNAS 2012. Open access.

The findings are dated to at least one million years ago and correspond to the Acheulean period, characterized by bifacial axes and generally attributed to Homo erectus or H. ergaster.

Partly charred bone from Wonderwerk

As the authors explain, previous claims of fire use in this period in East Africa,  the Levant or China have been plagued with controversy. In most cases the controversy stems from the relatively low resolution of the research and the possibility of the fires being natural occurrences. In the case of Zhoukoudian, near Beijing, what was once claimed as product of fire ended being sedimentary deposits. 

This research actually continues the findings of Peter B. Beaumont last year, who claimed (JSTOR, pay-per-view) that the South African cave hosted evidence of fire dated to c. 1.7 Ma.

The evidence included charred vegetal remains and animal bones, these last not completely calcinated, meaning that combustion was under 700 °C. Also remains of ironstone brought intently to the cave and worked in situ (main material of artifacts) show signs of manipulation under heat above 500 °C.

Apparently the fires were produced almost exclusively with light materials such as dry grasses and leaves. While it is not impossible that they might have used wood and then the evidence got lost, the data from bones indicates combustion under 700 °C, what is compatible with fire produced with lighter materials exclusively.

See also: Boston University's site, BU Today.