Out of Eden: The Peopling of the World

by Oppenheimer, Stephen

  Are the various geographical sources of the first Americans suggested by the anthropologists consistent with those indicated by the genetic story? The stories told by mtDNA and the Y chromosome should have more power to resolve such questions. We can now look at all the American founder lines in terms of their individual Asian genetic homelands discussed in Chapter 6.

  Two or three Asian sources?

  In the last chapter I presented a tale of the Great Freeze driving Asians east, south-east, and north-east out of the increasingly inhospitable Mammoth Steppe, the Central Asian upland region north of the Himalayas. I also suggested that there were at least three different sets of peoples in East Asia, with different genes and technologies. First, already resident on the eastern coastline of Asia, and in particular in Japan, were the descendants of the old beachcombers. Second, there would have been the Mongoloids: the Southern Mongoloids in Southeast Asia, and the Northern Mongoloids to the north of the Yangtzi and most likely also inland on the Tibetan-Qinghai Plateau. Finally there were likely to have been North Eurasians spreading over from the Russian Altai and sharing gene types with North and Eastern Europeans. In terms of geographical locations of potential sources for migrations into America, there were correspondingly at least three (see Figure 7.6). In Central Asia there were two potential source regions. They overlapped extensively before the ice age, but were roughly represented in the south by the Tibetan-Qinghai Plateau and in the north by the Russian Altai and the southern Siberian steppe. So, there were two genetically, culturally, and physically different groups: mammoth-hunting Upper Palaeolithic West Eurasians to the north-west, and microblade-making, probably incipient Mongoloids to the south. As I mentioned in Chapter 6, the former would most likely have made their way east down the Uda River to the Sea of Okhotsk, and then and down the Amur River to Sakhalin, to the north of Japan. The latter might have moved down the two great rivers of China, the Yangtzi and the Yellow River, towards the coast.

  These three Eurasian peoples, with their three sets of genetic markers, stone tools, and distinctive physiognomies, mingled or staked out their territory before the LGM in the relatively temperate Asian coastal region stretching from Japan and Korea north-east to Okhotsk. They were all candidates for the next step into Beringia, and thence to America, before the ice caps slammed the door shut. The question is whether the evidence of the genetic trail supports that sequence. We can take the story of the male line first because it is relatively clear cut.

  A north Eurasian source for American males?

  Is the picture of regionally diverse genetic origins supported by the Y chromosome story? The answer is no and yes. No, because as far as the Y chromosome is concerned, one line dominates all the Americas. And yes, because, when we look more closely, we shall find that line and its geographical origin injecting a strong element of Asian regional differences into the parallel mtDNA picture, thus supporting multiple sources (see Figure 7.6). Furthermore, there are several curious interlopers. The most striking aspect of the male line in the Americas is that over 90 per cent of all modern Native American Y chromosomes derive from one North Eurasian line, Polo (Figure 7.7), and that Polo’s branch, as a whole, also accounts simultaneously for 50 per cent of Europeans, has a scattered Siberian tribal representation, and is virtually absent from East Asia and elsewhere.64

  Not surprisingly, given the passage of time, most of the subtypes of the Polo line present in Europe are not the same as the American subtypes. This distinction is close to what is found with the mitochondrial X line, in that the split was a long time ago, maybe more than 25,000 years, and the two separated lines have each gone their way, generating more differences as time has gone on. The position is not quite the same as with X, however, since the Polo line is still common in North Asia, was also the main American founder, and is also found in Europe. These findings support a common origin for European and American members of this male line somewhere in North or Central Asia, and a northern route of entry into the Americas. Having looked at the tribal distribution in Siberia, Russian geneticist Tatiana Karafet and colleagues suggested that the Lake Baikal region is the homeland source of the American migration of this marker (although of course high frequency in itself does not identify a homeland). (Figure 7.6.) The European–American links do not stop there. The dominant European subgroup of the Polo line, Ruslan, accounting for 30 per cent of Europeans, is also found among Native Americans at a rate of 12 per cent, higher than could be explained by recent European admixture. North Asia also holds the Ruslan root type, and thus could have been the source of the American Ruslan founder.65

  Figure 7.6 Possible Asian sources of American founder lines. The ‘Venn’ approach rather than precise mapping is used to emphasise the ultimate genetic sources of American lines from three overlapping ancient ethnic groups, the original Pacific Rim beachcombers, the Mongoloids and the Upper Palaeolithic Eurasian hunter-gatherers.

  Figure 7.7 The American Y-chromosome tree. While the bulk of Native American Y chromosomes derive from Polo’s son Quetzalcoatl, scattered other North Eurasian Seth offspring are represented. In contrast to the mtDNA picture, the solitary Cain founder is the only male hint of any provenance from further south or east in Asia.

  Far commoner in America than the Ruslan line is a direct branch derivative of the Polo root, Q – which I shall call Quetzalcoatl after the mythical feathered serpent. Quetzalcoatl and his sons account for 64 per cent of American Y chromosomes and are almost unique to the Native Americans. The only exceptions to this last statement found so far are three Siberian Inuits; and one individual each from a neighbouring Siberian group, the Chukchis; the ‘Even’ tribe, an isolated group farther south on the Siberian Pacific coast; the Manchu (of Manchuria); and the Uzbeks from Central Asia. The exceptions all share the special American Y-chromosome type Quetzalcoatl. Since Quetzalcoatl is found among the Na-Dene and the Inuit-Aleut, it is most likely that this American Y marker arose very early on, perhaps even before the founding colonists entered North America before the last ice age. In other words, Quetzalcoatl is a true founder line and not a first-generation derivative. Such a view is supported by dating of the Quetzalcoatl marker in America, which comes out at 22,000 years ago.66

  Another North Eurasian Y line that is found very rarely in Native Americans is the TAT type. Although it could have been a post-Columbian European admixture, a North Asian source is still statistically more likely.67

  This overwhelming dominance of North Eurasian Y lines in America might seem to point to a single southern Siberian source of Native Americans, and knock multiple migrations out of play. But we should remember that the Y chromosome is much more prone to drift than is mtDNA. Another way of putting this is to say that in many traditional societies, a few of the men tend to father many of the children, whereas the women tend to have more equal numbers of children. This enhanced drift effect is most likely why American Y-line types seem so lopsided.

  There is, however, a remaining American Y founder who is more likely to have come from farther east than southern Siberia, and from a beachcomber rather than a Central Asian hunter source. This is the Cain beachcomber line, which spread round the Indo-Pacific coast soon after the movement out of Africa and is identified by an extra mutation, M217. While this Cain line is present throughout North and East Asia, his highest frequencies are found nearer the Pacific coast, in Manchuria and Okhotsk, around the lower Amur River. He is present at 10 per cent or more right down to Southeast Asia (see Figure 4.5). With his beachcombing history, this Cain could be the Y type that matches the B mtDNA marker in its more southerly Asian origin (see below).68

  Maternal lines from other parts of Asia

  The geographical origins of the maternal American founders are more diverse than the male lines, particularly with respect to the mtDNA B group, which is derived from Southeast Asia and is absent from the regions of North Asia most favoured for American Y sources. This spread of geographical sources of migrants supports the view that
there were several founding sources for the New World.

  In the last chapter I mapped the distribution in Asia of the five American maternal genetic founder lines. Group C is found north of the Himalayas in Asia, and A is far commoner in north-east Asia than further south, so we should look north of the Himalayas for source populations for the peopling of the Americas with these two lines. Group C is widespread in this region, extending from Tibet in the south nearly as far west as the Urals, where it has its highest frequency among the Sel’kups, hunters and fishers who live near the lower Yenisei River, within the Arctic Circle. On the Asian Pacific coast and farther south, C has a much lower frequency. As I explained in Chapter 6, the effects of the last ice age would have produced bottlenecks and re-expansions, so gene frequency on its own may not tell us the exact site of a homeland. However, given that the only parts of northern Central Asia we are certain remained occupied throughout the last ice age were Lake Baikal and the Upper Yenisei River areas of southern Siberia, this seems as likely a region as any for the homeland source of the American Group C (see Figure 7.6).69 Group A, on the other hand, coming ultimately from South China, perhaps up the Yangtzi to the Qinhai Plateau, is uncommon in Central and North Asia, with the exception of the extreme northeast of Siberia, where it is the main type. Since this part of Siberia was also part of Beringia, and closest to North America, and since it was A that re-expanded in Beringia after the LGM, north-east Siberia seems to be the most likely immediate Asian source of American Group A. As we saw in Chapter 5, the ultimate origin of C was probably India and Pakistan, via the western end of the Himalayas. Group C is a key candidate for the eastern maternal marker of the Upper Palaeolithic peoples who ranged from France to Manchuria before the last ice age. Those hunters may have contributed their skills to the pre-glacial ancestors of the big-game, Clovis-point hunting cultures of North America, which expanded after the ice age.

  Like C, the American X line is also likely to have a more northerly Eurasian origin (Chapter 5), but because of the large X-free gap between Europe and America, where she is absent, her actual source regions for the peopling of the Americas is a puzzle. The X line has recently been unambiguously identified in North Asia among Altaic peoples of southern Siberia, suggesting a possible homeland (see Figure 7.6). The 30,000-year-old link with Europe strongly suggests that X moved across the Asian steppe with Upper Palaeolithic hunters as a low-frequency shared Caucasoid component in the peopling of the Americas, and subsequently nearly became extinct in Asia. Group D is another of the five American maternal founder lines which is common in north-east Asia today. Group D’s ancestor, on the other hand, could have come into North Asia from Southeast Asia, south of the Himalayas. Its present distribution all along the East Asian Pacific coast, as far south as southern China, suggests a mainly coastal route north.70

  In contrast to these four maternal lines, which have a more northerly distribution today, the B maternal group clearly has her origin and place in the south. Group B dominates Indo-China, Southeast Asia, and the Pacific, and has a very different world distribution from the northern lines. Although B is found in Japan, China, and Mongolia, she is absent from the Subarctic regions both in Eurasia and North America. This makes a strong case for the source population of American Group B having been very different from C or X.

  There are further constraints on the origins of Group B. In Asia, B is represented by two main branches, B4 and B5, found side by side in most regions except the Americas and the South Pacific, where only B4 is found. B4 is the only branch that got to America, and she also dominates the South Pacific. The B4 branch in East Asia is so huge, varied, and widespread that it would be like looking for a needle in a haystack to search for an Asian match for the precise American founder type. But by coincidence, there is a near match from the recent results of two independent, complete mtDNA sequencing studies, one conducted by geneticists in Japan and the other by geneticists from Sweden and Germany. This laborious method, rather like a mini Human Genome Project, on just one stretch of DNA, goes for complete sequencing of the whole, circular mtDNA molecule. The result is the most precise and specific maternal genetic fingerprint possible. The match I found when comparing these two studies was between a Piman Indian Group B4 and a Japanese Group B4, thus pointing to the Asian Pacific Rim coast as the source for at least one American B type. The matches are strong, although as expected they are not like identical twins, but they do exclude the rest of the deep branches of Asian B4 and they do establish a link between Japan and America (see Figure 7.7).71

  Rainbow coalition

  Arguments based on finding individual near matches cannot accurately identify multiple different Asian homelands for the five American founder lines in Asia any more than the process of looking for a region, such as Mongolia, which harbours A, B, C, and D can identify a sole American homeland in Asia. They do, however, show us a more open-minded view of the possible events leading up to the peopling of the Americas and the potentially diverse sources of that colonization. We saw in Chapter 6 how the archaeology of the Asian Pacific coast bears signatures of new incoming Upper Palaeolithic technology from the Eurasian steppe in the west, and probably also microblade cultures from Tibet and the Qinghai Plateau, during the build-up to the last ice age. These may well have represented refugees from Caucasian populations to the west and from a Mongoloid source farther south. Arriving in north-east Asia, Korea, and Japan, they may have found themselves rubbing shoulders with the older coastal East and Southeast Asian peoples who were descended from the beachcombers (and represented today by the Ainu). So there were possibly as many as three different-looking groups that moved north-east along the coast, living off the rich marine life and game. They and their diverse cultures, genes, and physiognomies would all eventually have arrived in Beringia.

  Wandering separately

  What might Beringian communities have looked like on the eve of their entry into Alaska, 22,000–25,000 years ago? Were they like Star Trek’s Star-Base 9, polyglot, polymorphic, and polychromatic? In other words, were they completely mixed communities? I do not think so. Beringia was huge, with a variety of resources and locales to offer different cultures. It is more likely that, in spite of inevitable mixing, some of the cultural and genetic diversity was preserved in separate ethnic groupings.

  When those groups ventured into the Americas and fanned out and multiplied, they did not proceed like some consolidated, multiethnic United Nations team. The cultures which sprang up when the newcomers put down roots in the New World preserved separate and unique physical, genetic, and cultural elements from the various Asian homelands. Is this picture supported by the archaeological evidence? I think so. The search for pre-Clovis artefacts turned up implements beneath the Clovis layers and in the early non-Clovis sites of North and South America which were from completely different traditions and had different analogues on the Asian side. This is just what we would expect from several parallel colonizations by different cultures.

  We might expect, given the varied physical appearance of early American communities, to see clear genetic differences between them. Such differences are there in the marked tendency for tribal differentiation in mtDNA subtypes as noted by Torroni and Wallace (see above); yet one of the arguments for a single colonization of the Americas has always been the finding of A, B, C, and D major groups present in North, Middle, and South America. The Americas are huge, however, and there are marked differences in the relative frequencies of the five founders across different American peoples.

  Present-day ethnic groupings are distinguished by particular tribal-specific mtDNA types. With certain exceptions, there is little sharing of individual mtDNA types between ethnic groups. This has been interpreted to mean that the single founding wave quickly split up into groups that remained separate and drifted to different frequencies of founders. But the evidence equally fits the alternative picture of the first wave of colonization containing multiple discrete genetic strands which can be traced all
the way back, through Beringia, to different Asian sources. We have already seen the most extreme cases in the Subarctic region, where the Na-Dene speakers of Alaska and the north-west coast only have Group A2 American maternal types. As we have seen, the reason for the single American A2 line in the Na-Dene and Inuit-Aleut is thought to result from the near extinction of their ancestors causing drift down to one line.72

  This extreme predominance of Group A extends throughout northern America north of the 50th parallel, and includes not only the Na-Dene and Inuit-Aleut but also the Amerind, Algonquian-speaking northern Ojibwa of Canada and the Great Lakes. Their Group A is characterized by three unique types not found anywhere else. Not only that, but the Ojibwa have the highest rates of the rare X type at 25 per cent (see Figure 7.3), and have the unique honour of having only the dominant Y founder rather than its commoner derivatives. Although they speak an Amerind language, the Ojibwa share their Subarctic location and even some northern cultural features with the Na-Dene speakers. Their unique but relatively diverse genetic make-up contains a high rate of true founding lines. This suggests that the Ojibwa and related groups, rather than being a population which nearly became extinct during the ice age, as happened to the Na-Dene, may through their northern isolation have preserved a particular genetic identity deriving from their Beringian source, and ultimately from north-east Asia.73

  The two Native Americans who volunteered, along with others, to be sampled in the film (The Real Eve) associated with this book (see Plate 25) were from the Great Lakes region. By some extraordinary coincidence, each had one of these unique and uncommon local types. Our Ojibwa participant had the rare A1 type, while our Cree participant (Cree also belongs to the Algonquian language group) had an X type, which had previously been found only in Northern Ojibwa people.74