IF YOU had a choice, what sort of people would you like to think your ancestors were? Ambitious, creative movers and shakers who sowed the seeds of Western civilisation? Or uncultured, unsophisticated people who thought no further than where their next meal was coming from? If you come from European stock, then the notion that you are descended from a band of adventurous farmers who swept through Europe some 10 000 years ago, dramatically transforming human prehistory has a certain attraction.
鈥淭his is certainly the way most people like to view things,鈥� says Marek Zvelebil, an archaeologist at the University of Sheffield. The alternative-that your forebears were simple types who foraged for food in the woods-doesn鈥檛 do quite as much for the ego, suggests Zvelebil. 鈥淲ho, after all, would wish to regard hunter-gatherers, whose main aim in life was to have ecological relationships with hazelnuts, as their ancestors?鈥�
Well, if new genetic research proves to be correct, prepare to get in touch with the hazelnut lover in yourself. Because, according to Bryan Sykes, Martin Richards and their colleagues at the University of Oxford鈥檚 Institute of Molecular Medicine, most Europeans are descended from the first modern humans who moved into Europe from the Middle East at the beginning of the Upper Palaeolithic period, as much as 50 000 years ago. These people made a living by hunting animals and gathering roots, fruits and other nutritious plants. Later, they took up farming, a skill they learnt mostly from their neighbours, rather than from domineering newcomers. The Oxford team鈥檚 genetic data and their conclusions are stirring up lively debate among archaeologists.
Advertisement
The spread of agriculture through Europe has been a contentious issue for many years, and it is tied up with a second, equally contentious issue: the spread of Indo-European languages. Inevitably, the new genetic data from Oxford has become ensnared in this debate, too. Indo-European is the linguistic superfamily that includes all but a handful of languages spoken in modern Europe. Until recently, most archaeologists favoured the notion that these languages followed in the wake of warriors on horseback who swept out of northern Europe with territorial conquest on their minds.
Following the plough
This picture changed a decade ago when Cambridge archaeologist Colin Renfrew published his book Archaeology and Language. In it, Renfrew made a strong case that if the carriers of the new languages were riding anything, it was ploughs (metaphorically speaking), not horses; that they came from the Middle East and in peace, not from northern Europe and belligerently; and that they arrived 10 000 years ago, at the beginning of the Neolithic period, rather than 5000 years ago, as the previous theory held. In other words, the spread of farming and the spread of Indo-European languages went hand in hand, carried by the same migrating populations that swamped the resident populations of hunter-gatherers.
The Oxford team鈥檚 findings challenge Renfrew鈥檚 theory. They suggest that very few farmers migrated from the Middle East into Europe 10 000 years ago, making their contribution to the gene pool of modern Europeans minor, not overwhelming, as the currently accepted theory suggests. Sykes says his view of prehistory 鈥渨ould give the genetic pattern as we see it in modern Europeans鈥�. Yet this view is still compatible with Renfrew鈥檚 notion that Indo-European came from the Middle East, because, as Sykes points out, 鈥渓anguage migration doesn鈥檛 require the migration of people鈥�.
This new view is based on an analysis of how variants of mitochondrial DNA are distributed among modern populations in Europe and the Middle East. Mitochondrial DNA, the genetic material packaged inside the energy-metabolising organelles inside cells, has characteristics that make it a useful tool in tracing the history of ancient peoples. Renfrew welcomes this line of analysis as having 鈥渢he potential for informing us about population history鈥�, but he is worried that, for technical reasons, the Oxford data might be misinforming us instead.
Plain wrong
He is not the most severe critic. Luca Cavalli-Sforza, the geneticist at Stanford University in California who is most closely identified with the notion of population migration, thinks the Oxford team鈥檚 work is just plain wrong. To Cavalli-Sforza, there are so many technical problems with using mitochondrial DNA to infer population history that it is dangerously unreliable. He prefers to rely on information from genes in the nucleus, which he and other geneticists have collected in great quantities over the past four decades.
The invention and spread of agriculture is central to the history of humanity. For more than 100 000 years, humans subsisted in widely varied environments by foraging for food. Then, in a brief moment of prehistory starting about 10 000 years ago, people began to domesticate animals and plants in half a dozen 鈥渃entres of origin鈥� in the Old and New Worlds.
Initially, farming was more akin to horticulture than to agriculture in both scope and scale, but within a few millennia it was the principal means of subsistence throughout much of the world. The stage was set for momentous changes in social complexity: first came villages, then city states, and eventually nation states.
The march of agriculture from the Middle East throughout Europe is reflected in the archaeological record. Pottery and other artefacts associated with farming life have been recovered alongside the remains of domesticated animals such as sheep and goats, and crop plants such as wheat and barley. Radiocarbon dating of such relics shows that farming spread from Anatolia, now Turkey, to northern Europe in less than 2000 years, at a rate of about 1 kilometre a year. Archaeologists identify two routes of migration, based on characteristic forms of pottery: one through central to northern Europe, and the other around the shores of the western Mediterranean and up the Atlantic coast to France and Britain.
But the archaeological evidence cannot tell us what was migrating along these routes-people or ideas. Either way, the pattern of relics would be deceptively similar. Several factors have influenced the interpretation of this evidence, some of which has more to do with the human psyche than with facts, suggests Zvelebil.
For instance, based on early colonial contact with surviving hunter-gatherer societies, anthropologists initially believed that such people were technologically and culturally impoverished. Farmers, by contrast, were regarded as technologically and socially complex and sophisticated. So it was quite natural to believe that expanding farming populations would move into the territories of hunter-gatherers and that any foragers who stayed put would quickly take up farming, says Zvelebil.
This perspective was cemented in the phrase 鈥淣eolithic Revolution鈥�, coined by the Australian archaeologist Gordon Childe in the 1920s. To Childe, and to the generations of archaeologists who followed, pre-Neolithic hunters and gatherers were mere spectators to the dramatic change that swept the continent, and contributed little, if anything, to it. So there was a strong inclination to believe that farming was spread by migrating farmers, but without any convincing evidence.
If relics of the past cannot provide an unequivocal answer, then perhaps clues were to be had in the present, in the genes of modern Europeans. This was the notion that Cavalli-Sforza dreamt up almost half a century ago. He reasoned that it should be possible to unravel the history of modern populations from patterns of genetic relatedness among them. Four decades later, Cavalli-Sforza reached his goal, published in 1994 in a monumental work, The History and Geography of Human Genes.
Cavalli-Sforza鈥檚 approach is to look at how often a particular form of a gene turns up in a population-the gene frequency. He reasons that population movement would be reflected in a smooth gradient of frequencies. For instance, the rhesus factor (a blood protein) comes in two forms, positive and negative. In Western Europe, the proportion of individuals who carry the rhesus negative factor is relatively high, and becomes steadily lower as you move eastward. Such gradients develop when populations expand beyond their native territories into new regions. When migrants lose contact with the people from their original homeland, genetic differences begin to accumulate.
Initially, Cavalli-Sforza analysed blood proteins as a proxy for genes. More recently, he has looked directly at gene sequences. By the late 1970s, Cavalli-Sforza had enough data to examine the population history of Europe, based on 39 genes. The genetic map showed a gradient of gene frequencies from the Middle East spreading northwest through Europe, a pattern he believed reflected the Neolithic migration of farmers. This became known as the 鈥渨ave of advance鈥� model of the spread of agriculture. A population genetics model implies that the rate of movement would be close to 1 kilometre a year, in startlingly good agreement with the archaeological evidence. Today, Cavalli-Sforza and his colleagues have amassed information on the frequency of about a hundred genes in populations worldwide-and the same pattern still holds.
Advancing front
Perhaps beguiled by the power of the phrase 鈥渨ave of advance鈥�, many thought of the migration as 鈥渁 uniform front of colonists sweeping across Europe in the manner of German Panzer divisions鈥�, as Zvelebil describes it. In fact, such a migration-if indeed there was one-would be much more complex, influenced by the availability of suitable land for farming and the social and economic development of the established populations of hunter-gatherers-and how easy they were to oust or subjugate.
Cavalli-Sforza agrees that ideas as well as people would have spread through Europe. The key issue, then, is the scale of the migration: was it overwhelming or minor? Even though Cavalli-Sforza has never put a figure on the relative contributions of Neolithic and pre-Neolithic people to modern Europeans, the notion that the new arrivals overwhelmed the more ancient inhabitants became the prevailing lore among archaeologists.
The credibility of the wave of advance model-or some version of it-was further strengthened when Renfrew proposed his link between the spread of farming and that of Indo-European languages. Because Indo-European languages are virtually ubiquitous in Europe, Renfrew鈥檚 model implies that farmers migrated into every part of the continent, with the exception of the Basque region in the west and Hungary, Finland and Estonia in central and northern Europe, where the people speak non-Indo-European languages. Like the wave of advance model, Renfrew鈥檚 idea implies that incoming Neolithic farmers overwhelmed the pre-Neolithic people, except in a few odd corners.
When Sykes and Richards started extracting mitochondrial DNA from Saxon skeletons six years ago, they had no intention of becoming embroiled in the debate over the Neolithic transition in Europe. Their aim was simply to explore some aspects of Saxon migrations. 鈥淎s part of the project, we began collecting modern samples from people in Britain and Germany, to give a context for the DNA markers we were going to look at,鈥� recalls Richards. 鈥淲e began to see interesting things, things we didn鈥檛 expect.鈥�
One of these surprises was that key aspects of the mitochondrial DNA of Basque people look very much like that of other Europeans. 鈥淚f the rest of Europe is supposed to be descended from Neolithic immigrants, while Basque people are supposed to be a remnant of a pre-Neolithic population, then they should look different,鈥� says Sykes. 鈥淭hey don鈥檛, and this made us begin to wonder whether what we had been brought up on-the wave of advance model-should be questioned.鈥� So the Oxford team extended its sampling throughout Europe and the Middle East, eventually testing 821 individuals from more than 15 geographical locations.
Handy tool
Mitochondrial DNA is a potentially powerful tool for reconstructing population histories, for several reasons. For instance, it accumulates mutations much faster than DNA in the nucleus, which means that even in a relatively short period-in this case, a few thousand years-genetic differences will appear between different subpopulations. These variations tell us something about the history of the groups.
Secondly, mitochondrial DNA is inherited only through the mother, so there is no scrambling of maternal and paternal genes through recombination, making reconstruction of population history straightforward-at least in principle.
Imagine that the members of the first immigrant group all had mitochondrial DNA with the same nucleotide sequence, or haplotype. As time passes, and as the population grows, mutations begin to accumulate in the DNA, producing slightly different haplotypes within the population. This collection of haplotypes is said to constitute a lineage group, because they all come from the same ancestral molecule of mitochondrial DNA. As more time passes, more mutations accumulate in the population, which gives a way of looking at population history.
If mutations accumulate at a roughly constant rate through time, as many researchers believe, it is possible to calculate when the immigrant population entered Europe. If one mutation accumulates in a certain stretch of DNA every 10 000 years, for example, and if the number of mutations measured in the DNA of modern people is five, then the founding population must have become established 50 000 years ago. This argument collapses, of course, if mutations don鈥檛 accumulate at a constant rate, or if the rate used in the calculation is wrong.
When Sykes and his colleagues compared the DNA sequences from their 821 individuals, they found five lineage groups, not one. This simply means that today鈥檚 European population has descended from a set of ancestors who were fairly genetically diverse. The next step was to look at the diversity of haplotypes within the lineage groups, and calculate the age of the groups. If modern Europeans are descendants of immigrant farmers, then most of the groups will date back no more than 10 000 years. They would also be very similar to lineage groups in the Middle East today.
Tangled roots
The Oxford group鈥檚 analysis produced a spread of dates for the lineage groups, some of which were young (6000 and 12 500 years), while others were older (varying between 23 000 and 50 500 years). The two young lineages in the European sample resemble lineages in the Middle Eastern sample, and their geographical distribution closely matches the two routes of Neolithic immigration inferred from archaeological evidence. These lineages, then, are indicators of the movement of Neolithic people into Europe.
However, these young lineages represent only a small proportion of the total lineages in Europe-5 per cent in some regions, 15 per cent in others-which means that the number of farmers who entered the continent in Neolithic times was relatively small. Eighty-five per cent of mitochondrial lineages in Europe are much older than the Neolithic Revolution, and link us back to the people who arrived in Europe as long as 50 000 years ago.
Most anthropologists believe that modern humans first moved into Europe about 50 000 years ago, replacing Neanderthals, who had lived on the continent for some 250 000 years. 鈥淥ur results are pretty clear,鈥� says Sykes. 鈥淢ost mitochondrial DNA lineages in modern Europeans are derived from Upper Palaeolithic people who settled Europe a long time ago. The contribution of genes by Neolithic people was minor, implying that there were relatively few immigrants.鈥�
The pattern of the Neolithic transition backed by the Oxford team鈥檚 results-with small numbers of immigrant farmers who spread their ideas widely-has been termed the pioneer colonisation model, and is supported by a number of archaeologists, including Zvelebil. But if this picture is correct, how are Cavalli-Sforza鈥檚 data from nuclear genes to be explained? It is possible, suggests Sykes, that they are related to an earlier population movement. For instance, the data may contain an echo of the original occupation of Europe by modern humans 50 000 years ago.
Numbers game
Cavalli-Sforza rejects this suggestion. Simulations show that genetic gradients that are established through population migration can be maintained for 6000 years, but probably not 50 000 years. 鈥淚t seems to me unlikely,鈥� he says.
In any case, Sykes thinks he sees a way to reconcile the two positions-through Cavalli-Sforza鈥檚 own data. The gene frequency data are scrutinised by what is known as principal components analysis, a statistical trick for teasing out trends within a large body of data. The trend that Cavalli-Sforza takes to indicate the Neolithic immigration, known as the first principal component, accounts for about 27 per cent of the variation in the data. 鈥淚f this means that the immigrant farmers represented some 27 per cent of the population from which we all descended, then that鈥檚 not so very different from what we are saying,鈥� comments Sykes. 鈥淏ut so far Cavalli-Sforza hasn鈥檛 been quantitative about it in terms of people.鈥�
With his colleague Eric Minch, Cavalli-Sforza has now prepared a response to Sykes and Richards. In a paper to be published later this year in the American Journal of Human Genetics, he deals with the question of how big a contribution immigrant farmers made to the European gene pool. Although he argues that the percentage variation associated with the first principal component is not necessarily the same as the farmers鈥� genetic contribution, he suggests that 鈥渋t is probably not very far from it鈥�. A figure of 27 per cent as the contribution of immigrant farmers to the modern gene pool is still higher than the 15 per cent that Sykes infers from his mitochondrial DNA data, but it means that the two may not be so very far apart.
At least on the face of it. Cavalli-Sforza鈥檚 paper goes on to argue that the mitochondrial DNA sequence data used by the Oxford group are 鈥渕isleading鈥� and of little help in solving this particular problem. There is a growing, but unresolved, unease about the rate of accumulation of mutations in mitochondrial DNA. A faster rate of mutation, which some observers believe might be possible, would make Sykes鈥檚 lineage groups much younger, perhaps compatible with a Neolithic migration.
Cavalli-Sforza also points out that in hunter-gatherer and farming societies, women often leave home at maturity and settle in distant communities. This degree of migration would blur the genetic picture. For this and other reasons, says Cavalli-Sforza, the Oxford team鈥檚 conclusions 鈥渁re not warranted by their data鈥�.
Not surprisingly, Sykes disagrees. 鈥淲e鈥檝e thought about the issues he raises, of course,鈥� he says. 鈥淎nd we certainly wouldn鈥檛 have put all this effort into the work if we thought there was a serious problem.鈥� And the Oxford team has found a way of perhaps testing its conclusions, by resorting to ancient European skeletons again, this time ones that predate the Neolithic Revolution. If the team鈥檚 conclusions are wrong, and Neolithic immigration was on a massive scale, then mitochondrial DNA extracted from those pre-Neolithic skeletons should look very different from that seen today. But if they are right, then it will be very similar to that seen today, because we are largely descended from these people.
The extraction and detailed analysis of DNA from old bones is a long, difficult process, and only a couple of pre-Neolithic skeletons from Britain have been tested so far. Sykes won鈥檛 be specific about the results just yet, but he does say 鈥渋t鈥檚 looking promising鈥�.
