Did Farmers Replace Hunter-Gatherers In Europe?

Did Farmers Replace Hunter-Gatherers In Europe? A month ago, I would have said, mostly not. A new study released this month suggests that the answer may be, mostly they did.

Prologue

About 1.9 million years ago, Homo Erectus left Africa and spread across Eurasia. About 200,000 years ago, a new species, called the Neanderthals, left Africa and also spread across Eurasia. About 100,000 years ago, modern humans left Africa, but the path to Europe was impeded by the ice age until something like 45,000 years ago when modern human hunter-gatherers started to fill the ecological space left open by the extinction first of Homo Erectus, and then Neanderthals. The Neanderthals may have been pushed over the edge to extinction by modern humans either through competition or war or both, as they co-existed in Europe until about 28,000 years ago.

Neanderthals replaced Homo Erectus (they also overlapped, although the Homo Erectus were gone almost everywhere by the time that modern humans left Africa), and modern humans replaced the Neanderthals.

Replacement Or Cultural Change?

This brings us to the next big question regarding ancestry of modern Europeans (a similar issue is present in the Indian subcontinent and is being examined with many of the same tools).

After the domestication of animals and crops in the Near East some 11,000 years ago, farming had reached much of central Europe by 7500 years before the present. The extent to which these early European farmers were immigrants or descendants of resident hunter-gatherers who had adopted farming has been widely debated.

The Early DNA evidence

One of the main tools that has been used to explore the question is genetic evidence. Brian Sykes in his book "The Seven Daughters of Eve" (2001) popularized the early findings made regarding question using mitochondrial DNA (which is simpler than nuclear DNA and inherited matrilineally), and in a brief mention, parallel confirming early findings made using an analysis of Y-chromosome genetics (which is simpler than other chromosomes and is inherited patrilineally).

Those findings suggested that somewhere between one in five (Y-chromosome data), and one in six (mtDNA) of modern Europeans were descended from Neolithic farmers whom immigrated to Europe, while rest of the Europeans were descended from earlier hunter-gatherer populations who took up farming by learning it from the Neolithic farmer immigrants from the Near East.

Sykes had good reason for his conclusion. He grouped mitocondrial DNA data into clusters, and seven of the clusters were found with some frequency in Europe, each assigned a letter. Six were widely distributed across Europe and were distinguished by mutations that probably happened before the Neolithic revolution. The oldest of the groups, U, was dated to roughly the time that hunter-gatherers entered Europe. A seventh group, J is the youngest and probably arose around the time of the Neolithic revolution. Group J is not found in Basque Country (in Northern Spain) which is home to the most distinct European language and has long been hypothesized to include some of the ancestrally oldest Europeans. But, group J was found to be much more common in the places where farming first appeared in Europe. Broken up into one subgroup along the European coast, and another subgroup up through the river valleys where farming first appeared in Continental Europe. Farmers, Sykes supposed, would be less mobile than hunter-gatherers. Group J also looked likely to have arisen in the Near East based upon where mtDNA diversity within the group is greatest (a large share of Berber nomads have mtDNA from Group J).

Skyes' conclusion was also backed by a sample from a single hunter-gatherer era skeleton found in a British cave twenty miles South of the town of Bath, dated to about 9,000 years ago (prior to British farming) called Cheddar Man, which belonged to mtDNA group H which makes up about 47% of modern Europeans.

A confirmation of the general proportions he found with Y chromosome data seemed to seal the deal.

The New DNA Evidence

A study released in the October 2, 2009 issue of Science, "Genetic Discontinuity Between Local Hunter-Gatherers and Central Europe’s First Farmers," by B. Bramanti, et al, casts doubt on this conclusion. It looked at the mitocondrial DNA of about four dozen skeletons, about half from late European hunter-gatherer skeletons and about half from the earliest Neolithic era farmers in Central Europe, which it compared to about four dozen modern European mitocondrial DNA samples and prior studies. It clustered the mtDNA sampled into essentially the same groups used by Sykes and the researches whose work he drew upon.

Rather than having a good mix of the six mtDNA groups that Sykes had believed had intermixed to make up the pre-Neolithic gene pool in Europe, 82% of the hunter-gatherer skeletons had mtDNA from group U (14 from subgroup U5, 2 from subgroup U4 and 2 from an unspecified U type with data not clear enough to distinguish a subgroup), the oldest of the groups, which is now rare in most of Europe. The U5 subgroup includes 1-5% of the modern European population on the Mediterranean coastline, 5-7% of the European population in most core European areas, 10-20% in northeastern European Uralic speakers, and over 40% in the Scandinavian Saami people who are now found in parts of northern Finland, Norway, Sweden and extreme northwestern Russia. The letter U for this mtDNA group probably comes from the of Uralic language group that these people speak, although their ancestral origins were probably further South (the far north was covered by an ice sheet when the mutations that define group U are believed to have arisen). Modern Europeans belong to subgroup U4 at frequencies of 1-5% in most of Europe, "with Western Europe at the lower end of this range and northeastern Europe and central Asia showing percentages in excess of 7%."

The four exceptions in the 22 hunter-gatherer skeleton sample all came from a site in Ostorf, Germany, near Germany's North coast on the Baltic Sea. One was from mtDNA group K, two were from mtDNA subgroup T2, and one was from mtDNA group J. At the eleven other hunter-gatherer sites, all more inland, every single skeleton contained mtDNA from group U. Ostorf, Germany's samples of hunter-gatherers were relatively young (about 4950 to 5200 years ago), and was in an "enclave surrounded by Neolithic funnel-beaker farmers." The other three skeletons from Ostorf were part of mtDNA subgroup U5. So, this may be an exception that proves the theory behind the rule.

Group T, believed to have originated around 17,000 years ago, is the mtDNA group to which about 9% of modern Europeans along the Mediterranean and Western edge of Europe belong, and is particularly common in the west of Britain and in Ireland, suggesting that some of the founding population of Group T was made up of coastal seafarers. It is not surprising then that half of the exceptions to the U group rule were T group members found at a coastal hunter-gather site.

The early Neolithic farmer skeletons in Central Europe, meanwhile, showed multiple mtDNA groups, "mainly N1a, but also H, HV, J, K, T, V and U3 types." The early central European farmer skeletons had no U5 or U4 subgroup members. There were no N1a or H types in the early hunter-gatherer sample.

The N1a subgroup did not become prominently known until a 2005 paper by Haak who is on of the co-authors of the current paper. "N1a became particularly prominent when Wolfgang Haak et al. studies on 7500 year old skeletons in Central Europe revealed that 25% of the Neolithic European population might have belonged to the N1a haplogroup. The skeletons were found to be members of the Linear Pottery Culture which is credited with being the first farming culture in Central Europe. . . . It is a rare haplogroup as it currently appears in only .18%-.2% of regional populations. It is widely distributed throughout Eurasia and Northern Africa and is divided into the European, Central Asian, and African/South Asian branches based on specific genetic markers. Exact origins and migration patterns of this haplogroup are still unknown and a subject of some debate." Group N is believed to be the genetic predecessor of mtDNA groups including groups U, K), HV, H, V, J and T. It is believed to have originated in either the Near East or East Africa, which the evidence for the Near East seeming somewhat stronger as it is not found in Africa outside East Africa despite being younger than subgroups of group N found elsewhere.

The recent study has its own issues. The sample size of skeletons is not huge. It doesn't have hunter-gatherer or Neolithic samples that are from outside Central Europe.

The hunter-gatherer samples in particular are from quite far North, well removed from the areas near the northern Mediterranean coast where mtDNA groups H (ca. 30,000 years ago), V (ca. 17,000 years ago), T (ca. 17,000 years ago) and K (ca. 15,000 years ago), which are common in modern Europe seem likely to have arisen (J probably arose in the Near East about 10,000 years ago), U probably arose in Greece about 45,000 years ago, and X probably arose in the Caucuses about 25,000 years ago).

There could, for example, have been several distinct hunter-gatherer populations in European in the early Neolithic, only one of which (with members overwhelming from the mtDNA subgroups U4 and U5) was found in North Central Europe at the dawn on the Neolithic era, who began to intermingle to a much greater degree around the time that Neolithic farmers immigrated, perhaps first uniting with each other against the invaders, and then adopting the invaders customs (in the paradigm of the Greeks who united to some extent against Roman invaders, with the Romans then adopting much of their culture).

But, this study also offers persuasive evidence that Sykes didn't have available to him. Sykes inferred ancient population patterns entirely from modern European populations and a single British hunter-gatherer skeleton. Bramanti, in contrast, used the direct evidence of mtDNA from actual members of these populations.

Bramanti's small sample size is compensated for by the degree to which his results are emphatically clear that the two populations were separate. In terms of mtDNA genetics, the hunter-gatherers and early farmers in Europe were as distinct as just about any modern non-African populations are today. The odds that this was due to any meaningful degree by to mere random sampling error is extremely low.

There are also a wealth of historical precedents for incoming farming populations almost wiping out previously uncontacted hunter-gatherer populations. It happened in North and South America, and continues today when isolated groups in South America make contact with outsiders. It happened in Australia (whose pre-European population was pre-Neolithic), and to a lesser extent in New Zealand (the Polynesians of New Zealand and the rest of Oceania were post-Neolithic in origins). It has happened in Papua New Guinea and the Andaman Islands. Indeed, small pox, which was an important cause in the great decline of hunter-gatherer populations in these examples would be a plausible culprit in the prehistoric, early Neolithic case in Europe. Small pox was a post-domestication of animals cross over from domesticated animals to which hunter-gatherer populations would not have been exposed. The degree to which modern Europeans show mtDNA links to the hunter-gatherer skeletons in the new study is in line with the experience of historical first contact episodes.

Bramanti doesn't pose an easy resolution to the question of how to resolve the seeming contradictions in different lines of reasoning from mtDNA evidence (although the study does confirm the previous conclusion that even the oldest modern human skeletons do not show genetic links to the Neanderthals). Indeed, the study makes clear that the modern European sample does not have a straight forward link to the early European hunter-gather population or the early European Neolithic farmer populations by descent or population mixing alone.

Comparing and Contrasting the Theories

The fact that even the extremely distinct hunter-gatherer population in Bramanti's study had a more than 50% overlap with the early farmer population in one location, makes clear that these populations were not separate species of modern humans. Neanderthal mtDNA is much more genetically distinct than any modern humans are from each other and there are no cases of Neanderthals with modern human mtDNA, and there are no cases of modern humans with Neanderthal mtDNA out of thousands of samples, carefully chosen to ferret out anomolies in genetically unusual populations, analyzed.

Bramanti's study is making distinctions between two waves of immigrants to Europe, ultimately from the Near East, separated by no more than 25,000-35,000 years.

The analysis that Sykes relies on also concludes that there were at least two wave of immigrants to Europe, ultimately from the Near East, who were separated in time by many thousands of years.

The difference is that Bramanti's analysis suggests that the first wave accounts came overwhelmingly from only one part of the type U mtDNA group, that their matrilineal descendants make up only a small minority (probably less than one in seven) modern Europeans, while Sykes analysis suggests that hunter-gatherers at the time of the Neolithic revolution were much more genetically diverse and that their matrilineal descendants account for 80%-84% of the modern European population.

Neither researcher proposes a complete replacement theory, but Bramanti suggests that pre-Neolithic hunter-gatherers of Europe were overwhelmingly replaced by at least one (and possibly more) waves of Neolithic farmer immigration, while Sykes analysis suggests that the Neolithic farmer immigrants added only a new minority layer to the existing mix, reather than predominantly replacing the existing inhabitants.

There is basically consensus involving mtDNA groups J and U, that together make up perhaps 25-30% of Europe's population. The former was clearly part of a wave of Neolithic farmer immigration to Europe. The latter was clearly part of the earliest hunter-gatherer populations in Europe.

The big issue apparent discrepency concerns what kind of people the matrilineal descendants of the other 70-75% percent of Europe's population in the other mtDNA groups were like and when they arrived. Sykes theory argues that they were hunter-gatherer populations who arose earlier and intermixed with each others. Bramanti's data suggest that they were part of the early Neolithic farmer populations the immigrated to Central Europe later, and it isn't clear at what point they intermixed. Neither approach entirely rules out multiple waves of immigration within hunter-gatherers, or by farmers, although Bramanti's data disfavors multiple waves of hunter-gatherer immigration to Europe.

No one thinks that either set of data is wrong. The question is how to interpret that data and what additional data might clarify the story.

A Possible Story

One possibility is that the Near Eastern people who took up farming in the Neolithic era were genetically diverse. Mutations in the mtDNA groups other than J that arose at various points among coast Mediterranean hunter-gatherers may have been exchanged by early coastal hunter-gatherer populations for many thousands of years and back migrated to the hunter-gatherer populations of the Near East and Asia Minor where farming would later arise in the Neolithic era. Most of the Near Eastern farmer immigrants to Europe in the early Neolithic would have been genetically indistinguishable from the Mediterranean coastal hunter-gatherers with whom they had formed a single population for thousands of years. But, one in four or five of the fairly small founder group of farmer immigrants to Europe would have been part of the very new mtDNA group J that marked this wave of immigration and arose in the Middle East. This scenario would make genetic patterns in current European populations all but useless in determining whether there was a replacement or cultural assimilation by a homogeneous caste of mostly J mtDNA type farmers, with regard to existing coastal hunter-gatherer populations in coastal areas. We would have to turn to other lines of evidence to resolve that issue.

The Neolithic colonists probably consisted of two main groups headed in different directions out of Anatolia and the Near East. One J subgroup may have been particularly strongly represented in the founder group that went out along the coasts of the Mediterranean and Europe. The other J subgroup and the N1a subgroup (an older group that had experienced little mtDNA mutation for a long time period in the Near East), may have been strongly represented in the founder group the moved up through Central Europe's river valleys.

This still doesn't explain why the N1a subgroup all but disappeared. Were early Neolithic farmers of the linear pottery culture wiped out or genetically overwhelmed by a later wave of farmers? Did this mtDNA group coincide with some metabolic frailty that was of no consequence in the Near East, but was a serious survival disadvantage in Central Europe (perhaps burning energy too quickly and not leaving fat stores to survive lean winters)? Did the mutation clock tick independently in most of the ancient N1a mtDNA genomes and produce new mtDNA types?

Perhaps the N1a subgroup was common because of a very small pre-Indo-European Neolithic farmer colonists founder group from the Near East, that stayed small because their farms were fairly marginal in success, and they were overwhelmed in numbers of later Indo-European farmer colonists from the Near East where the Central European branch of the J subgroup made up a larger share of the founder group, in parallel to the two waves of farmer colonization believed to have taken place in India, first a pre-Indo-European (probably animist rather than polytheistic) Dravidian group, and then an Indo-European Aryan group that brought the Hindu religion with them. This group might have been more a group of colonist than a warrior-priest caste that invaded India.

In contrast, the pre-Neolithic hunter-gatherers of inland Europe and central Asia, probably a more sparsely populated group, may have been largely isolated from their coastal ancestors after breaking off from them sometime between 45,000 and 30,000 years ago as the ice pack receded opening up Europe for the first time to modern humans. They would have shared ecological space with Neanderthals in this time period (the most recent Neaderthal sites are dated to about 28,000 years ago), and would have experienced a later ice age that may have forced them to retreat to the South. The last glacial maximum was 18,000 years ago and many of the locations where skeletons of hunter-gatherers were found were exposed by a receding ice pack between 18,000 and 9,000 years ago. Long periods of isolation and harsh conditions caused by extended ice packs may have made this group an unattractive one for coasta hunter-gatherers to interact with, as they might have had different customs and languages that further disouraged interaction.

When Neolithic farmers arrived in Central Europe, they may have partially seen their numbers decline and partially retreated further North towards colder climes where their life style continued to be the most adaptive.

The Big Picture

Less technically, Bramanti's study is consistent with (although it does not compel) the hypothesis that in much of the world, a very large percentage of all of the non-African descendants of modern human hunter-gatherers were wiped out. Their population were overwhelmed in most local gene pools (including Europe's) by a small number of distinct populations of Neolithic farmers.

This is also supported by evidence that a very large proportion of languages spoken today show common origins in the Neolithic era time frame, rather than the plausible far earlier time frame of the departure of modern humans from Africa. The implication is that the vast majority of the languages used by pre-Neolithic hunter-gatherers in Europe, Asia and Africa have gone extinct, something that the languages of people who are overwhelmingly outnumbered in their own lands by foreigners and suffer gross population falloffs tend to do.

Footnote On India

Lingistic, archeological and historical evidence point to multiple waves of immigration to the Indian subcontinent; genetic evidence of that history is in the process of being collected and has puzzles, just as the European evidence does.

The first wave of modern human immigration to India is sometimes described as a wave of hunter-gatherers, linguistically linked to the people of Southeast Asia and the related languages spoken in a few pockets of Northeast Central India, who probably came to India by a coastal route when sea levels were low and ice caps were large. There is evidence that these people have genetic links to the oldest modern inhabitants of the Andaman Islands.

The second wave of modern human immigration to India is associated with the Dravidian language family and the earliest Indus River Valley civilization.

The third wave of modern human immigration to India is described as Aryan (a term that has fallen into disfavor after appropriated by the Nazis but still used widely in English language discussions about ancient India) who are associated with the branch of the Indo-European language family exemplified by Sanskrit and Hindi, and with the Hindu religion.

Some mtDNA evidence suggests that India's genetic makeup is very homogeneous across caste, linguistic and geographic lines, something that suggests that the Aryan wave of immigration may have involved a fairly small number of poeple with a large cultural, but not a large genetic impact. This wouldn't be inconsistent with the examples of hunter-gather populations being devistated by first contact with peoples who have historically farmed. India's mtDNA homogenity, could come either from the Dravidian wave of people culturally transforming earlier hunter-gatherers, or from the Dravidian wave of people replacing earlier hunter-gatherers and being homogeneous themselves, either due to founder effects or due to blending of subpopulations over thousands of years. The Dravidian populations themselves might not have been as devistated by an influx of Aryan immigrants to India because as farmers themselves, they might have had greater immunity to the devistating diseases that come with agriculture from domesticated animals.

But, the Y-chromosome evidence seems to argue for a much greater Aryan influence, with as much as half of the male population in Hindi speaking areas showing Aryan influences, while a much smaller percentage of the population in areas that speak the Dravidian languages showing this link.

This too isn't necessarily inconsistent. We know that the armies of Ghengis Khan had an immense genetic impact on the Y-chromosomes found in Asia. It wouldn't be much of a stretch (and fits the historical and culture myth) to imagine the incoming Aryans as an army made up disproportionately of men who had disproportionate reproductive success with predominantly local women, sometimes on a polygamous basis. As rulers, their cultural impact would also be disproportional.

CORRECTED, EXTENED AND REVISED on October 22, 2009.