Out of Eden: The Peopling of the World

by Oppenheimer, Stephen

  I should clarify that just because 20–34 per cent of modern European mtDNA lines have been retained from before the LGM, that does not mean that the rest of the lines found today had to have entered Europe from outside after the LGM. No, they were mostly locals. Of all modern European lines, 55 per cent originated in the period just after the ice age (the Late Upper Palaeolithic) but these, like the V haplogroup, probably derived from pre-existing European lines and simply reflect the post-glacial re-expansions from the refuges – in other words, new shoots off an old stock. Real fresh immigration into Europe from the Near East during the Neolithic period (from 8,000 years ago onwards) perhaps accounts for only 15 per cent of modern lines.10

  An interesting recent discovery about the genetic make-up of the south-eastern European region has to do with Adam’s markers in Romania. The Carpathian Mountains were glaciated at the LGM and thus formed an effective, jagged barrier between south-eastern Europe and the regions bordering the Black Sea. This barrier splits Romania in two from top to bottom. Without their ice, the Carpathians hardly constitute an insurmountable physical barrier today, but they still mark a clear genetic boundary. This is revealed by Y-chromosome markers characteristic of north-eastern Europe and the Ukraine occurring at higher rates to the east of the Carpathians, while the markers more characteristic of Central Europe are found to the west. But this micro-regional boundary is very much obscured by the great dominance of the main Eastern European Y lineage M17, which may characterize the original preglacial Eastern Gravettian intrusions from the east. This lineage is found at very high rates throughout Eastern Europe – from Poland, through Slovakia and Hungary, to the Ukraine. M17 is still found at high frequencies among the Slavic peoples of the Balkans, which could support the existence of the Balkan ice-age refuge.11

  Northern Asia in the cold

  The Himalayas grew an ice cap, and deserts expanded in the band round the 40th parallel and to the south, in Central Asia, during the LGM. The population of southern Central Asia was severely reduced, although probably not completely wiped out – at least in Tibet if not elsewhere. This may have effectively separated South, East, and Southeast Asia from Central and North Eurasia, but it does not mean that life stopped in the centre and the north. During the LGM, the permafrost boundary extended down to about the 50th parallel. Permafrost did not prevent human occupation of the steppe tundra. There is even evidence of human activity north of the permafrost line at the LGM, particularly in the Russian plain and also further north-east in Talitskogo, Siberia. Hunter-gatherers of southern Siberia continuously occupied some sites, especially Afontova Gora on the Yenisei River, developing their own unique southern Siberian culture. Strips of steppe tundra and forest steppe, rich in large herbivores, still stretched eastward from the Russian plain through to north-eastern Siberia (Figure 6.3). In spite of the severe weather extremes, there were rich pickings for those hunters who could adapt. And they did, as is shown by evidence from scattered archaeological sites in those areas that date to this period.12

  Again, the genetic record tells us that these hardy hunters of the permafrost retained some of their ancient maternal lines from before the LGM, and that there were subsequent expansions of certain subgroups after the thaw. As I mentioned in Chapter 5, A, C, and Z are characteristic of North Asia, while D also extends farther south in China. Of these lines, A clearly survived from well before the LGM, as did various Manju and Nasreen lines. Then there was a gap of 10,000 years until about 17,000 years ago, before the re-expansion of lines such as variants of D, A, C, and finally Z.13

  Figure 6.3 East Asian refuges at the LGM. Most of Northeast Asia was uninhabitable save for narrow corridors of steppe tundra, but some hardy hunters stayed on there. In contrast, Japan, China and Sundaland increased their habitable areas. As the refugees left the northern steppe, intrusive Upper Palaeolithic technology appeared in East Asia at the peak of the LGM, particularly in Japan and Korea.

  The genetics of North and Central Asia has simply confirmed what we might expect from the archaeology: when it got cold, the hunters of the Mammoth Steppe shrank in range and hunkered down, and when it got warm again they bounced back and re-expanded. The more interesting questions are what they looked like, and who their closest local descendants are today. Why and when did Mongoloids become the dominant types throughout most of Asia and the Americas? Does the genetic story tell us anything that cannot be inferred from the archaeological record?

  Driven off the high cold steppe

  In the last chapter I started to scratch away at several of these questions, particularly the genetic origins of the Mongoloids. I showed how several pieces of evidence point to southern Siberia, the Upper Yenisei River, and the Lake Baikal region as possible sites for further physical specialization of the Northern Mongoloid peoples during the Palaeolithic. The earliest Eurasian skeletal evidence for the Mongoloid physical type was claimed for part of a skull found in Afontova Gora, in southern Siberia, dated to 21,000 years ago, just before the LGM (see Chapter 5). Even this evidence is weak though, and no other place in the Old World, including East Asia, shows such types from before the LGM or for a long time afterwards. Admittedly, human skeletal evidence of any description from the Far East before 11,000 years ago is scanty. This gap in the record leaves the impression – sometimes taken as dogma – that there were no Mongoloid types in East Asia until around 7,000–10,000 years ago, and none in Island Southeast Asia until well after that.14

  The big problem with the fossil gap is how to explain that the Mongoloid physical type accounts for 99 per cent of peoples living in these regions today. Not only that, but they are the most numerous type in the world. In fact, since most Native Americans are generally also regarded as Mongoloid and are, with a couple of exceptions, extreme Sinodonts, they were geographically the most widespread physical type before Columbus reached America and changed its history irrevocably.

  It takes no real stretch of the imagination to view the Northern Mongoloid peoples as evolving a specialized physique adapted to hunting in harsh, windy, and cold conditions in a geographically isolated region of southern Siberia. But hunting and gathering was what all humans did at that time. So how did these hunter-gatherers come to rapidly occupy and dominate more than half the world after the ice age? An extraordinary story waits to be unlocked, perhaps by our genes. Another problem we face, if we accept this late dispersal of Mongoloids, is posed by the date of colonization of America by Mongoloids. In the next chapter we shall look at both the archaeological and the genetic evidence for the date of their first entry, and shall find that the latter suggests that all the main founder genetic lines had entered America before the LGM.

  With the American pioneers in mind, I think the concept that Mongoloids dispersed to East and Southeast Asia only within the past few thousand years is also likely to be wrong, mainly because most of the gene lines that are characteristic of Southern Mongoloids appear to have their origins in that region. But there is no obvious reason why a Central Asian group of highly specialized hunters should disperse so rapidly to dominate East Asia at the beginning of the Neolithic less than 10,000 years ago. I think that the Mongoloid presence, expansions, and dispersals started much earlier and continued in pulses up until recently. But before going into detail, let me sketch in the next two paragraphs what I think happened in East Asia. In the absence of a strong archaeological record, much of my speculation is based on the genetic trail.

  Throughout the Upper Palaeolithic period, from 40,000 years ago up to the LGM, there was a huge partly wooded hunting ground between the Tibetan Plateau and southern Siberia. It was occupied by skilled hunters who derived their genetic heritage from widely separated parts of Asia (see Chapter 5). First, I think Dale Guthrie is partly right in that the Northern Mongoloids overdeveloped the Mongoloid physical type in adapting to the extreme cold and wind at high altitude in the southern and eastern part of this homeland. But farther to the west, at lower altitudes, the hunters may have looked more li
ke Europeans. As the glaciers began to take their grip, all the hunters on the high steppe were squeezed. Those in the south of the homeland in south-eastern Tibet were forced off the high, cold, and dry Tibetan Plateau and down the great Asian rivers to China and Indo-China. They escaped down the Brahmaputra, to Assam, down the Mekong to Indo-China, and east down the Yangtzi to southern China. Those on the Qinghai Plateau to the north-east of Tibet were squeezed down the Yellow River to northern China. Those who had hunted farther north at the 40th parallel in the uplands of Xinjiang and Mongolia were forced even farther north by the rapidly expanding Taklamakan and Gobi Deserts of southern Central Asia and into the lowland refuges of southern Siberia, such as Afontova Gora on the Yenisei River (Figure 6.4).

  The problem with the hunters going north to find warmer lowlands was, of course, that the nearer they got to the Arctic Circle, the more chance there was of it getting colder and windier. There was one escape from this northern cold trap – to the north-east.

  They could have moved along a narrow tongue of steppe tundra either side of the Lena River to Yakutsk, then east to the Sea of Okhotsk, where the coastal climate was more temperate. Low sea levels at the time exposed a wide lowland strip creeping out into the Sea of Okhotsk. Another north-easterly escape route from the Lake Baikal region to the Pacific coast lay a little farther south down the great Amur River. This temperate north-west Pacific coastline was the route to America. Such dispersals from Central Asia, to three points of the compass, would in several instances have taken Mongoloids back to East Asian coastal regions from which their ancestors had originally trekked inland 20,000 years or more before.

  South China may have had an expanding Mongoloid population as refugees left the high steppe, but judging from the genetic story, most of the mtDNA lines in Southeast Asia were local and had always been there. Groups such as B and F, which are regarded as typical of Mongoloid populations, have great local antiquity in the south.15 This time round, however, those coasts looked very different.

  More land on the coast

  The period around the LGM was not just a time of territorial contraction and humans seeking refuge. Paradoxically, more land was opened up for human colonization by the lowering of the sea level than was closed off by the freeze-up. Large parts of the Asian coastline from India to Siberia now extended hundreds of kilometres farther out towards the edge of the continental shelf. Siberia joined with America to create the new landmass of Beringia. The East China Sea and the Yellow Sea were drained. Japan became connected with the Asian mainland through the island of Sakhalin. These northern coastal regions were warmer than the Central Asian steppe. Australia and New Guinea joined to form the continent of Sahul. The Indian subcontinent acquired new land, and joined up with Sri Lanka. The region that gained most out of the LGM was Southeast Asia, where the South China Sea, the Gulf of Bangkok, and the Java Sea all dried up to link Indo-China, Malaysia, and the largest islands of Indonesia into one great patchwork of woods and savannah, a vast continental landmass called Sundaland, twice the size of India (Figure 6.3).16

  Figure 6.4 Centrifugal migrations from Central Asia at the LGM. The LGM forced Central Asians back down the same rivers they had come up 20–40,000 years before. This meant Siberians mainly going to Northeast Asia and America (Polo and Cain). People from Xinjiang and the Tibet-Qinghai Plateau went to China and Indo-China and the latter on into Southeast Asia (Cain and Ho). Only a few Cain and Ho genes would have crossed the Wallace Line to form colonies. The mtDNA picture is similar.21–24

  It was not just coastland reclaimed from the sea that beckoned humans to recolonize as the ice sheets grew. The joining of Siberia with Alaska gave humans their first chance to enter the great continental landmasses of the Americas without getting their feet wet. This they did during the last stages of the build-up to the LGM. That fits with the view that the movement across Beringia (now the Bering Strait) at such a cold time was in response to worsening conditions in Asia.

  The centrifugal dispersal of Mongoloid hunter-gatherers down the great Asian rivers to the long Pacific coastline (see Figure 6.4) did not necessarily result in the immediate replacement of the descendants of the beachcombers who were living there. It is more than likely that a patchwork of colonies was established which lasted for a long time, and that eventually the new genetic pool merged with the old. Northern China, Korea, Japan, and Sundaland had huge expanded coastlines consisting of wooded temperate and subtropical lowlands into which new refugee immigrants from the Central Asian steppe would most naturally have expanded. So, the initial settlement by the newcomers may have partially flowed over, or avoided, pre-existing relict peoples of the old beachcombing route.

  When the sea returned like a huge tide after the ice had melted, however, much of the archaeological evidence of such coastal occupation would have been washed away. We would expect to see such ice-age archaeological evidence only in regions with a steep coastline and continental shelf, such as Japan, and some of the Indonesian islands. Remnants of preglacial beachcombing populations would also be found only in regions that have remained isolated by geography or rising sea levels, such as Japan.

  I should stress that the Mongoloid dispersal view I am presenting here differs from orthodox archaeological reconstruction not just in dates and timescale but also in the movements of people. Some prehistorians of the Southeast Asian region, such as Australian Peter Bellwood, favour a recent one-off Mongoloid replacement of older Australo-Melanesian hunter-gatherers in Island Southeast Asia by Taiwanese from farther north along the Pacific Rim. In this view, the Australo-Melanesian hunter-gatherers would have corresponded most closely to the Semang ‘Negritos’ (see Chapter 5). Bellwood argues, moreover, that the expanding population of Mongoloids were Neolithic rice farmers from southern China who arrived by sea via Taiwan and the Philippines only 4,000 years ago. I have presented the evidence against this theory elsewhere.17

  What I am saying is first that Mongoloid expansions and re-expansions in East and Southeast Asia started earlier, around the LGM at least 18,000 years ago. They mainly used land routes, and have continued on and off ever since. In fact, possibly the largest population expansion in Island Southeast Asia before the present era may have occurred as recently as the arrival of the Metal Age in the islands, around the last few centuries BC, when rice agriculture greatly expanded.18

  The second point is that there were not just two groups, the Mongoloids migrating down from the north and the Semang-like hunter-gatherers, who were vying for the landmass of Southeast Asia. There was another indigenous expanding group, the Southern Mongoloids, who may have occupied the Southeast Asian continental shelf and much of the Pacific coastline from very ancient times, and may indeed have been ancestral to all Mongoloids themselves.

  This indigenous Southern Mongoloid group is identified best by its unique and associated southern genetic markers, which dominate the region.19 They probably looked rather like some minorities of south-eastern Indo-China today, and in particular like one of the groups of Malaysian aboriginals farther south in the Malay Peninsula, known as the Aboriginal Malays. The latter also resemble the majority Malay population in having some Mongoloid features. From this perspective, aspects of the so-called Southern Mongoloid appearance may have been around for a much longer time. As we shall see, their ancestors can be identified by more than just their genes and teeth.

  In common with the great majority of other Southeast Asians, the Aboriginal Malays have been identified as Sundadonts. The Aboriginal Malays are even closer dentally to the Pacific Rim populations, including the Polynesians, and on a dental basis could be regarded as ancestral to the Sinodonts further north (see Chapter 5 and Figures 5.3 and 5.4).20 In short, Southern Mongoloids could have been around in the South since before the LGM.

  Post-glacial Asian invasions of Southeast Asia and Oceania

  As I have said, my reconstruction is based more on the genetic record than on the scanty skeletal and archaeological record. In fact, my own in
terest in genetic prehistory was first awakened by medical genetics, while I was working in the south-west Pacific in the early 1980s. My findings suggested that early migrations of local genetic types out of Southeast Asia to New Guinea and Island Melanesia were more likely than huge late Mongoloid migrations from China replacing Australo-Melanesians throughout Australasia and the Pacific, as previously thought.21 So I shall start with genetic evidence for the dates of Mongoloid expansions into Southeast Asia and Oceania.

  In Chapter 3 we saw how the some of the best combinations of archaeological, physical, and genetic evidence for the early route out-of-Africa come from intermediate and end points of the trail in Malaysia, Australia, and New Guinea. I also described how the deep branching genetic lines for the Australia and New Guinea were different from each other, indicating the antiquity of their colonization. The unique genetic identity of such early arrivals allows us to measure and clearly identify all latecomers and to see where they came from. Equally useful is the ability to date and estimate the size of later migrations.

  By taking this approach, the English geneticist Martin Richards and I were able to suggest that while, in agreement with current views, the ancestors of the Polynesians – the great Pacific pioneers of 3,500 years ago – were mostly different from the oldest inhabitants of the New Guinea region, they did not come from Taiwan, farther north, as generally believed but from Southeast Asia. Furthermore, although the Polynesian lines were newcomers to the Pacific, their own forebears had probably already arrived in eastern Indonesia from Indo-China by 17,000 years ago.22 The lines carried by the Polynesians were branches of those found on the Southeast Asian mainland, and one ancient line in particular (mtDNA Haplogroup B4) was shared with the American dispersal via the northern Pacific Rim. Apart from genetically linking the Americans with the Southeast Asian dispersals, this also suggests a date near the LGM for the earliest migrations out of mainland Southeast Asia. Richards and I have continued to explore other markers of Southeast Asian migration, leading up to the work on the aboriginal peoples of the Malay Peninsula described in this book.