Out of Eden: The Peopling of the World

by Oppenheimer, Stephen

  Some are still convinced that Australian aboriginals represent an earlier migration out of Africa than that which gave rise to Europeans, Asians, and native Americans. Yet again our genetic trail tells us otherwise. Several studies of Australian maternal clans have shown that they all belong to our two unique non-African super-clans, Manju and Nasreen, and large studies of Y chromosomes show that male Australian lines all belong to the same Out-of-Africa Adam clan as other non-Africans, and belong to the Seth and Cain lines. The same pattern is seen with genetic markers not exclusively transmitted through one parent. In other words, the combined genetic evidence strongly suggests Australians are also descendants of that same single out-of-Africa migration. The logic of this approach, combined with the archaeological dates discussed above, places the modern human arrival in the Malay Peninsula before 74,000 years ago and Australia around 70,000 years ago. It is also consistent with the date of exit from Africa predicted on beach-combing grounds. As we saw in Chapter 1, the motivation for leaving the African continent may have been the failure of the beach harvest on the shores of the Red Sea caused by sharply rising salinity 85,000 years ago. My date estimates for the trek around the Indian Ocean en route from Africa suggest that the beachcombers could have taken as little as 10–6,000 years to eat their way down the coastline to Perak and another 10–4,000 years to reach Australia. Such a time requirement is fulfilled by the difference between leaving Africa around 85,000 years ago and arriving in Australia 65,000 years ago. The former date is consistent with a date of 83,000 years estimated for the African L3 cluster expansion using the molecular clock (see also Chapter 1).26

  To gaze upon the face of the mother who launched a billion families A question which has obsessed journalists and film-makers ever since Newsweek’s 1988 Mitochondrial Eve story is what she and her companions looked like. It is likely that they all looked very different from one another, more varied than any random selection of people from any non-African country might look today. They were, after all, the stock from which all non-Africans developed, and all the evidence suggests that the variation within individual groups decreases the farther away from Africa one goes, even today. Yet, it is also the case that differences between one group and another, between, say, Chinese, Native Americans, Polynesians, Melanesians, and Australians, increase the farther we move from Africa; so the total diversity between the so-called races might seem to increase with distance from Africa. Let me put this another way. My late and very dear mother-in-law, who was Chinese, once told me – and she wasn’t joking – that Europeans all looked the same to her. Europeans have said the same about the Chinese. There is a paradox here, which we need to resolve before we can continue to chart the genetic diaspora out of Africa.

  It is all a question of how the diversity is split up. What appears to be a paradox is actually the result of multiple unequal group splits followed by genetic drift in action. It can be explained thus: as the explorers split into more and more branches, the original genetic variation was parcelled into different groups which today live scattered around Eurasia and America. Over the past 50,000 years or so, members of these individual splinter groups in the ex-African diaspora have drifted to become more like one another than like members of other groups. The splinter groups have diverged from one another. While I shall argue later that many of these changes were true evolutionary adaptations to new environments, some of them were just down to the random and unequal way the cake was cut.

  How is this? The answer lies in the same processes that operated when that first small group left Africa. Imagine – as we did in an earlier chapter – our stock of human diversity as a bag of marbles. Let us say that in Africa 75,000 years ago there were 10,000 marbles in the bag. When the exodus crossed the Red Sea, that diversity was reduced to 250 marbles in the founding colony. That reduction is known as a ‘founder effect’. When that first colony had been living as an isolated colony for several thousand years somewhere in south Arabia, a number of marbles had failed to reproduce themselves and others became the dominant types. The bag of genetic variation became even smaller. As discussed in Chapter 1, that is genetic drift.

  As the colony moved on there were unequal splits. When the group size became too large for the beach to sustain it, smaller groups, ‘pioneers’, broke off and ventured inland, travelling up rivers. Those smaller groups carried even fewer of the original genetic marbles that left Africa. They were undergoing further founder effects and drift.

  Now we have to ask which was the tree and which was the stem. Were the beachcombers always in the majority, and the river explorers the minority, or did the main trunk veer inland at some stage? Which group kept most of the marbles? Such a question might seem academic, but if our journalists want to gaze upon the faces of a modern group which is as close as possible to the original exodus, this is the question we must answer: which modern non-African population has retained most varieties of the original ancestral African marbles?

  Nuclear genetic markers: a bigger bag of marbles to look for older markers

  To look at this question of diversity further, we cannot rely on our mitochondrial and Y-chromosome genetic markers. As far as they are concerned, only one original African marble from each single parent line ultimately survived to mother and father the rest of the world. As I have said, that is how we can argue there was only one exodus (see Chapter 1). There is no diversity in one line. No, if we want to compare ancestral African genetic diversity with the rest of the world, we need genetic marker types that survived the founder effect and genetic drift of the exodus in greater variety and larger numbers than one line each. So we have to look at the rest of our genetic make-up that resides in the nucleus of our cells and is not solely linked to one sex or the other.

  For clarity’s sake, my marbles analogy was grossly oversimplified. When we look at the surprisingly small proportion of our DNA that actually does anything useful, we find that for each gene there may be from one to a dozen slightly different versions. These versions can usually be linked together in a family gene tree. The base of each gene tree and its lower branches are usually predominantly African, whereas later branches and occasionally some of the earlier ones are found in different frequencies outside Africa. In addition, we can look at the non-functioning parts of our DNA that also progressively acquire extra packets of meaningless information which can, nonetheless, be utilized by geneticists as tracer junk mail. These extra packets may accumulate to the point where they constitute as much as 10 per cent of our entire DNA. The absence of specific extra packets in an individual generally means that they have retained the ancestral or African type.27

  When we look at which populations have preserved most of our original African genetic heritage, we come back to Australia and New Guinea. These two landmasses together make up the huge, remote, and partially submerged Sahul Continental shelf and have retained more original African gene types than any other part of the non-African world.28 Today, however, New Guineans and Australians do not look alike (see Plate 17), and superficially the only population in the region who look at all like modern Africans are Melanesians, with their very dark skin and tightly frizzy hair. Even then, New Guinean highlanders, who are regarded as the ‘least diluted’ Melanesians from the original Pleistocene populations, are much more robust-featured (heavy-boned with strong supra-orbital ridges – see Chapter 5) than present-day Africans. But this begs the question as to what Africans looked like at the time of the modern exodus around 80,000 years ago. Almost certainly they did not look exactly the same as the gracile (fine-boned) Bantus, Nilotics, Pygmies, or !Kung look today. The fossil Qafzeh and Skhul skulls, dated to between 90,000 and 120,000 years ago, are much more robust than those of modern African populations, suggesting that Africans may have changed more than New Guinean highlanders in this respect. Having lived for a number of years both in New Guinea and Africa, it is my guess that the reconstruction of one of the Qafzeh skulls commissioned for our documentary would today pass unnoticed i
n New Guinea, but might be harder to place in Africa (see Plate 18).

  Genetic traces of the exodus among relict populations nearer to Africa

  What of other peoples who may be genetically closer Africans as a result of their position on the old beachcombing trail – what can comparing the frequencies of retention of ancestral gene types in African and non-African peoples tell us? Two regions stand out. The closest (along with Australians and New Guineans) are those in Pakistan and the southern Arabian peninsula, followed by Europeans and Indians. Trailing a long way behind, with the least retention of African diversity and the most drift, are the peoples of East Asia and the Americas.29

  The root position of Pakistanis and southern Arabians in retaining ancient African genetic diversity is certainly what we would expect from our proposed southern route out of Africa, and there are other pieces of evidence to support this. Along the south coast of Arabia are the isolated Hadramaut peoples, described by some as Australoid. Their maternal genetic make-up includes 40 per cent of African genetic lines; but although some of these markers could be related to the founding Out-of-Africa Eve, the majority of such lines have arrived from Africa more recently. Farther along the Indian Ocean coast the peninsular Indian populations also group genetically closer to the African root than do more easterly Asian peoples. Indian ethnic groups, both caste and tribal, were included in a large study of nuclear autosomal (non-sex-linked) markers. They were found to retain a higher rate of the African ancestral types than do Europeans and other Asian groups.30

  There are other signs that the ancient African genetic diversity has been preserved in Pakistan. While the population of Pakistan in general shares some ancient mtDNA links with India, Europe and the Middle East, they also possess unique markers that are found nowhere else outside Africa. There are indeed populations that hark back to that ancient connection. One aboriginal so-called Negrito group, the Makrani, is found at the mouth of the Indus and along the Baluchistan sea coast of Pakistan. It is speculated that they took the coastal route out of Africa. They have an African Y-chromosome marker previously only found in Africa which is characteristic of sub-Saharan Africa. The same marker is found at slightly lower frequencies throughout other populations of southern Pakistan, Saudi Arabia, and the United Arab Emirates, and at higher rates in Iran. Another unique Y-chromosome marker appears outside Africa only in this region (see Note 31). One other ancient Y-chromosome marker points specifically to Pakistan as an early source and parting of the ways. This is an early branch off the Out-of-Africa Adam that is present at high frequency in Pakistan and at lower frequencies only in India (especially in tribal groups) and further north in the Middle East, Kashmir, Central Asia, and Siberia. The fact that this marker is not found farther east in Asia suggests that the only way it could have arrived in Central Asia was by a direct early northern spread up the Indus to Kashmir and farther north.31

  The origins of Europeans: where was Nasreen born?

  The prime position of South Asia in retaining African genetic diversity brings us to discuss the South Asian site of the parting of the ways. And we can now focus on the power of our maternal and paternal gene trees to leave trails of their ancient movements. Although the finding of roots and early branches of the Nasreen and Manju lines along the South Arabian and Baluchistan coastlines is strong evidence for the idea that a single southern out-of-Africa exodus continued first to South Asia, it does not necessarily identify the exact site of origin of Nasreen or her six Western daughters, including Rohani, or of Rohani’s own West Eurasian daughters (U, HV, and JT) (Figure 4.2). Although Rohani clearly derives from somewhere in South Asia (see below), the birthplace of her West Eurasian daughters is less clear but of great interest. However, when we look at the mitochondrial genetic make-up of the whole of the Near East (including the Levant, Anatolia, Armenia, Azerbaijan, northern Kurdistan, as well as more southerly regions such as the Yemen, Saudi Arabia, Iraq and Iran) as revealed by the European founder analysis of Martin Richards and colleagues we find the greatest genetic diversity of Rohani’s western daughters in the Gulf state of Iraq (see Chapter 3). Notable in Iraq are the high rates of unclassified root genetic types and the absence of other non-Rohani western daughter groups of Nasreen, such as W, I, or X. So it may be that Rohani’s western daughters were born farther south, in Mesopotamia or near the Gulf, either before or during the first northward migration up the Fertile Crescent.32

  Such questions can be sorted out only by a formal founder analysis, such as that used by Martin Richards and colleagues to trace the European founders, but this time comparing South Asian regions with the Levant. My own view is that for Rohani’s Indian granddaughter U2i to be of a similar age, around 50,000 years, to her sister U5 in Europe, it is simplest for both Rohani and her daughters (Europa, HV, and JT) to have been born down south in the Gulf region, halfway between India and Europe. Certainly, as we saw in Chapter 3, a strong argument can be made for several early non-African Y-chromosome groups having their roots in southern Asia rather than up in the Levant.33

  Figure 4.2 The beachcomber mtDNA tree. Shows mtDNA branches found along the Indo-Pacific coast, Oceania and the Antipodes. In each, genetic continuity of spread can be seen from South Asia to Japan and to the south-west Pacific. Horizontal lines are for regional representation, vertical arrows for specific local representation; names with a star (‘*’) indicate putative relict populations from the trek.

  South Asia: fount of all Asian lines?

  While little has been published on genetic markers for the southern Arabian peninsula, there is a magnificent body of information on the Indian subcontinent. For mitochondrial DNA we have an Estonian genetics team, with Asian collaborators, to thank for much of this, and several European and American groups are now collaboratively working on the Y chromosome. While this may bias our attention to India and Pakistan, the focus is deserved. Pakistan is the source of the Indus, the main route for direct access from South Asia to Central Asia west of the Himalayas. India, Bangladesh, and Burma also harbour the great southern Asian rivers, the Ganges, Brahmaputra, and Salween, draining the south and east of the Himalayas.

  With the diversity of its inhabitants’ physical appearance and their cultures, peninsular India is a rich transitional ethnic and geographical zone between West and East Eurasia. To the north and west we see only a very gradual change in people’s appearance, including pigmentation (see Chapter 5), as we move from Pakistan, through Afghanistan and Iran, and west into Europe. The similarities between many South Asians and Europeans are striking.

  This gradual east–west transition across northern India and Pakistan is paralleled in the relative frequencies of the two genetic daughter super-clans, Nasreen and Manju of the Out-of-Africa Eve founder line (L3). (Figure 4.3.) When we look at the Hadramaut in the Yemen, these frequencies are in the ratio 5:1 Nasreen to Manju, consistent with the view that Nasreen originated farther west than Manju, in the Gulf region. Between the Red Sea and the mouth of the Indus in Pakistan, West Eurasian genetic descendants of Nasreen continue to outnumber those of our Asian super-clan Manju. As we move east across the mouth of the Indus in Baluchistan and on towards India, the picture begins to change. The ratio of Nasreen to Manju lines decreases to 1:1 in the far-western states of Rajasthan and Gujerat. By the time we get to Bengal and Bangladesh, the ratio has reversed and Manju dominates. We thus find that the dominant clan in India is Manju, with an increasing frequency as we move from west to east.34

  To the north and east of India, however, the changes are even more abrupt. In Nepal, Burma, and eastern India we come across the first Mongoloid East Asian faces. These populations generally speak East Asian languages, contrasting strongly with their neighbours who mostly speak Indo-Aryan or Dravidian languages. By the time we get to the east of Burma and to Tibet on the northern side of the Himalayas, the transition to East Asian appearance and ethnolinguistic traditions is complete, as is the rapid and complete change of the mitochondrial sub-clans of Manju and
Nasreen. In Tibet, for instance, the ratio of Manju to Nasreen clans has evened back to 3:1, and there is no convincing overlap of their sub-clans with India (see Figure 4.2). Instead, Tibet shows 70 per cent of typical East and Southeast Asian Manju and Nasreen sub-clans, with the remainder consisting of as-yet unclassified Manju types of local origin. The north-eastern part of the Indian subcontinent therefore shows the clearest and deepest east–west boundary.35 This boundary possibly reflects the deep genetic furrow scored through India by the ash-cloud of the Toba volcano 74,000 years ago.

  Figure 4.3 Distribution of Nasreen and Manju in South Eurasia. In West Eurasia there is only Nasreen; in most of East Eurasia there are even mixtures of Nasreen and Manju, but on the east coast of India there is nearly all Manju. The latter is consistent with near local extinction following the Toba explosion with recovery only of Manju on the east coast.

  To the south of the Indian peninsula, the main physical type generally changes towards darker-skinned, curly haired, round-eyed so-called Dravidian peoples (see Chapter 5). Comparisons of skull shape link the large Tamil population of South India with the Senoi, a Malay Peninsula aboriginal group intermediate between the Semang and Aboriginal Malays (see above).36

  Manju born in India, Nasreen possibly a little farther west in the Gulf

  Manju, who is nearly completely absent from West Eurasia, gives us many reasons to suspect that her birthplace is in India. Manju achieves her greatest diversity and antiquity in India. Nowhere else does she show such variety and such a high proportion of root and unique primary branch types. The eldest of her many daughters in India, M2, even dates to 73,000 years ago. Although the date for the M2 expansion is not precise, it might reflect a local recovery of the population after the extinction that followed the eruption of Toba 74,000 years ago. M2 is strongly represented in the Chenchu hunter-gatherer Australoid tribal populations of Andhra Pradesh, who have their own unique local M2 variants as well as having common ancestors with M2 types found in the rest of India. Overall, these are strong reasons for placing Manju’s birth in India rather than further west or even in Africa.37