Out of Eden: The Peopling of the World

by Oppenheimer, Stephen

  Figure 1.3 The multiple-exodus hypotheses. Most recent out-of-Africa syntheses argue for at least one northern exit to Europe and further to the rest of Asia within the past 50,000 years, while acknowledging the possibility of an earlier southern route to Australia (Stringer et al. (1996, 2000),14 Lahr and Foley (1998),15 Underhill et al. (2001)15).

  So, the insistence by many experts that Europeans came out via North Africa can be seen to depend on various rationalizations. These include North African refuges and either multiple out-of-Africa migrations or an early onward migration from the Levant towards the Far East. There are problems with all these rationalizations in their attempt to preserve the hallowed northern route for Europeans. Taking the most straightforward first, Jonathan Kingdon has suggested that the world was peopled by a single early northern out-of-Africa movement during the Eemian interglacial 120,000 years ago.16 Since many of the desert corridors in Africa and West Asia were green at that time, the would-be migrants to Australia could have walked briskly eastward, straight from the Levant to India. They could then have taken a rest in green parts of South Asia before proceeding to Southeast Asia, arriving by 90,000 years ago, and then colonizing Australia. (By ‘South Asia’, I mean those countries between Aden and Bangladesh that have coasts facing the Indian Ocean. This includes the countries of Yemen, Oman, Pakistan, India, Sri Lanka, and Bangladesh, and the southernmost parts of those that border the Arabian Gulf: Saudi Arabia, Iraq, Beirut, the United Arab Emirates, and Iran.)

  As evidence of an ancient skilled human presence east of the Levant, Jonathan Kingdon points to abundant Middle Palaeolithic stone tools found in India starting from 163,000 years ago. The problem with this view is that there is no skeletal evidence for modern human beings anywhere outside Africa of that antiquity. Kingdon acknowledges that these tools could easily have been made by late pre-modern or archaic humans (the Mapas, as he calls them) who were in East Asia at that very time.

  Clearly, to get to Australia, Australians must have gone through Asia, but this kind of logic is still no proof that Anatomically Modern Humans migrated across Asia before 90,000 years ago, let alone 120,000–163,000 years ago.

  The eastern barriers

  There is another problem with Kingdon’s 90,000–120,000-year bracket for the colonization of Southeast Asia. If his early migration into Southeast Asia had left the Levant any later than 115,000 years ago, they would most likely have perished in the attempt. Analysis of human and other mammalian movements from Africa to Asia over the past 4 million years suggests that, except during an interglacial, there were a variety of obstacles to movement between the Levant and the rest of Asia. When the world was not basking in the moist warmth of an interglacial, there were major mountain and desert barriers preventing travel north, east, or south from the Levant. To the north and east was the great Zagros–Taurus mountain chain, which combines with the Syrian and Arabian Deserts to separate the Levant from Eastern Europe in the north and the Indian subcontinent in the south.17 Under normal glacial conditions this was an impassable mountainous desert. There was no easy way round to the north because of the Caspian Sea and the Caucasus.

  As in Marco Polo’s time, the alternative overland route from the Eastern Mediterranean to Southeast Asia was to get to the Indian Ocean as soon as possible and then follow the coast round. To the south and east of the Levant, however, are the Syrian and Arabian Deserts, and the only option was thus to follow the Tigris Valley round from Turkey and down the western border of the Zagros range to the Arabian Gulf (see Figure 1.8). But this route, through the Fertile Crescent, was also extreme desert outside interglacial periods and was therefore closed.

  The practical impossibility of modern humans getting from the Levant or Egypt to Southeast Asia 55,000–90,000 years ago means that a northern exodus from Africa during that period could have given rise only to Europeans and Levantines, not to Southeast Asians or Australians. Now, Europe and the Levant were not colonized until around 45,000–50,000 years ago, and Australia, on the other side of the world, was colonized well before then. This in turn means that, in order to preserve the northern route for Europeans, Chris Stringer, Bob Foley, and Marta Lahr have all had to postulate separate southern routes for Australians and even for Asians. This tangle is resolved by the genetic story.

  What the genes say about the northern route

  All such speculations have previously rested on an archaeological record consisting of a few inaccurately dated bones separated by huge time gaps. From the turn of the century, published work on the European genetic trail by scientists such as Martin Richards, Vincent Macaulay, and Hans-Jurgen Bandelt has changed all that and allowed us to examine the first leap out of Africa with a much clearer focus on the timing and location.18 This work does two things. First, it confirms that the Levantine expedition over 100,000 years ago perished without issue, so that the first doomed exodus of modern humans, like the Neanderthals who in turn died out 60,000 years later, left no identifiable genetic trace in the Levant. Second, although sub-Saharan Africans may have much more recently left their genetic mark surviving in one-eighth of modern Berber societies, it reveals no genetic evidence that either Europeans or Levantines came directly from North Africa.

  How do I know this? The construction of a precise genetic tree using mitochondrial DNA makes it possible to do more than just identify our common ancestors. Figure 1.4 shows our mitochondrial tree, starting at the base as many clans in Africa, then sending a single branch (L3) out into South Asia (India). L3 became our Asian Out-of-Africa Eve and soon branched many times to populate Arabia and India, then Europe and East Asia. We can date the branches in this tree and, by looking at their geographical distribution, show when and where the founders arose for a particular prehistoric migration. On a grander scale, this method is how the out-of-Africa hypothesis was proved.

  Using this approach, we can see that Europe’s oldest branch line, designated U (50,000 years old and shown in Figure 1.4), which I have called Europa after Zeus’ lover of the same name, arose from somewhere near India out of the R branch, which I have called Rohani after its Indian location. R, in turn, arose out of the N branch, which I call Nasreen, and which arose from the Out-of-Africa Eve twig, L3. If Europeans were derived from North African aboriginal groups such as the Berbers 45,000 years ago, we would expect to find the oldest North African genetic lines deriving directly from the origin or base of the L3 branch.19

  Figure 1.4 A single expansion from Africa. Simplified network of mtDNA groups showing how only one of 13 African clans gives rise to all non-Africans. N (Nasreen) group is expanded to emphasize West Eurasian ramifications, while M (Manju) has no representation there and heavy presence in India. (For detail and dating of tree branches see Appendix 1.)

  Instead, when we look at North Africans, in particular the Berbers, who are thought to be aboriginal to that region, we find the very opposite. North African lines are either more recent immigrants or very far removed from the root of L3. First, there is no evidence in North Africa for the earliest out-of-Africa branch nodes, Nasreen and Rohani. Those are found instead in South Asia – as shown in the figure. In fact, we find that North Africa has been populated mainly by recent southward migrations of typical European and Levantine genetic lines. The oldest indigenous North African mtDNA line, sometimes referred to as the Berber motif, is dated to have arrived from the Levant around 30,000 years ago, and is a solitary derived sub-branch (U6 – see the figure) of the West Eurasian Europa clan (West Eurasia being Europe and the Middle East). U6 shows every sign of having come from the Levant rather than the other way round. About one-eighth of maternal gene lines in North Africa come from more recent migrations from sub-Saharan Africa, and over half are recent movements south from Europe. Finally, the other daughter clan of Out-of-Africa Eve, the Asian M supergroup which is found in India, is completely absent from Europe, the Levant, and North Africa. This makes it extremely unlikely that North Africa could have been the source of Asians, as put forwar
d in Jonathan Kingdon’s scheme.20

  All this adds up to a view of both Europe and North Africa as recipients of ancient migrations from farther east. In other words, there is no evidence of a first northern exodus from sub-Saharan Africa into North Africa – but rather the opposite. Just where in the East these ‘Nasreen’, ‘Rohani’, and ‘Europa’ branches came from, we shall see shortly.

  Just one exodus

  The strongest evidence against the north as the primary route for Europeans, or for any other modern humans for that matter, comes again from the structure of the maternal genetic tree for the whole world. As can be seen in Figure 1.4, only one small twig (Out-of-Africa Eve) of one branch, out of the dozen major African maternal clans available, survived after leaving the continent to colonize the rest of the world. From this small group evolved all modern human populations outside Africa. Clearly, if there was only one exodus, they could have taken only one of the two available exit routes from Africa. I cannot overemphasize the importance of the simple and singular fact that only one African line accounts for all non-Africans.

  In any exodus from Africa there would have been a mix of different potential founder genetic ‘Eve’ lines. The same applies for any random group of people. Yet only one of these genetic lines survived. How this came about holds the proof for how many sorties from Africa were successful. Let us say there were fifteen genetically different types or lines of mtDNA that left Africa in a particular group (it could be more or less, as in Figure 1.5). This is a realistic number; even today there are fifteen surviving African maternal lines older than 80,000 years (see Appendix 1). These could be viewed as fifteen different kinds of marbles in a bag. From those fifteen lines only one mitochondrial line would, over many generations, become the Out-of-Africa Eve line or the common ancestral ‘mother line’ of the rest of the world. In other words, there would be only one kind of marble left. This random selection and extinction process is called genetic drift because the original mix of lines has ‘drifted’ towards one genetic type.

  The mechanism behind drift is simple. From time to time, some mothers’ lines will die out because they have no daughters surviving to reproduce. These gene line extinctions are shown in Figure 1.5. In a small isolated population, this will eventually leave a single surviving ancestral line. Drift has strong effects in small groups. A common modern example of drift, seen through the male side, is that of small isolated Alpine or Welsh villages ending up after generations with just one family surname – Schmidt or Evans, perhaps – on all the shopfronts.

  Figure 1.5 Genetic drift. Diagram to show how, over the generations in a small isolated population of constant size (e.g. 100–400), there is a tendency for natural variation in reproductive success to reduce diversity – so from many varieties of mtDNA to a few or just one line (highlighted in black). Tracing back from the 16th generation here shows they all have the same ancestral mother.

  If we now take two identical African exodus groups and isolate them in different regions (two bags of marbles in our analogy), each group will drift down to one gene line. Because drift is due to random extinction, the two groups will not usually both drift to the same line (the chances of this happening are 1 in 15 if there are equal numbers of each type to start with). If we then take two different groups coming out of different parts of Africa (the northern route and the southern route – see below) at different times, tens of thousands of years apart, they would not be at all similar and would represent a completely different selection of African lines. Again, the two groups would be even less likely to drift down to the same line.

  It does not require much statistical knowledge to see that the odds against two such independent small migrant groups moving out at different times and in different directions from Africa, and each then randomly drifting to exactly the same maternal gene line, are vanishingly small. It would be the same as ending up with exactly the same kind of marble in each bag. Had there been two migrations, there would inevitably be at least two Eves for non-Africans. In other words, the genes confirm to us that if only one mitochondrial gene line survived outside Africa, there was only one group that made it out of Africa to populate the rest of the world. That same small group eventually gave rise to Australians, Chinese, Europeans, Indians, and Polynesians. And their genes came out of Africa at one point in time.

  Clearly, with this unitary argument, if the northern route could not give rise to Australians, then logically nor could it give rise to anyone else outside Africa. But before rejecting the northern route and exploring the feasibility of the single southern route, we should pinch ourselves and ask whether the rest of the genetic evidence really is consistent with this single exodus view. The short answer is yes, when we look at genetic markers other than mitochondrial DNA. If we look at the Y chromosome, we find that all non-Africans belong to only one African male line, designated M168. Using markers passed down through both parents, several studies have again shown evidence for only one expansion out of Africa.21

  Another issue that is very relevant here is the prediction of dates. When we look at the founding African genetic line, Out-of-Africa Eve, which gave rise to non-Africans, we find an age of around 83,000 years22 (see Figure 0.3). This does not fit with the archaeological and genetic evidence of a much later colonization of the Levant and Europe from only 50,000 years ago.

  In contrast to the very derived nature of genetic lines in North Africa, there is another African region that holds an extraordinary diversity of genetic lines, close and parallel to the very roots of the out-of-Africa branch. This is Ethiopia, with its green hills standing out among the surrounding desert and dry savannahs of the Horn of Africa. But Ethiopia is to the south-east of the Sahara, literally at the gates of the southern route. Geneticists are now in a race to study Ethiopians, some of whom may be descended from the source population of that single exodus. This leads us now to look at the southern route.

  The southern route

  To fully understand the alternation of corridors that potential emigrants from Africa had to contend with, we have to go back to the first human to leave Africa. Millions of years ago, when the super-continent Pangaea was in the process of breaking up, Africa shivered and cracked her flank. This huge fault in the Earth’s crust connects to the rift valley system of East Africa. Its most visible effect is a deep cleft between Africa and Arabia – the Red Sea. At the northern end of the Red Sea now lies Suez, and at the southern end is an isthmus, 25 km (15 miles) wide and 137 metres (450 feet) deep, known as the Gate of Grief (Bab al Mandab) from its numerous reefs (see Plate 7). Two million years ago, Africa was still joined to Arabia and the strait was dry. A wide range of Eurasian and African mammals were able to cross between Arabia and Africa at both the northern and southern ends of the Red Sea. At that time, however, Africa was already moving away from Arabia at a rate of 15 mm per year, gradually opening the isthmus and eventually closing the southern gate out of Africa.

  Recent evidence shows that one of the last mammals to walk out of Africa, before the Gate of Grief finally flooded and closed, was our second cousin Homo erectus (see Plate 2), carrying with them a few basic pebble tools. At this stage, since both the northern and southern gates were open, our cousin spread rapidly eastward through India to East Asia and also up through the Levant, reaching Dmanisi in the Caucasus by 1.8 million years ago.23

  The flooding of the Gate of Grief was not the only barrier raised to mammalian movement into Asia. Two more hurdles had been erected. The first of these, as referred to above, was a long range of mountains, the Zagros–Taurus chain. These had been pushed up in the previous few million years at the same time as the Himalayas. Stretching from Iran to Turkey, this rocky barrier has effectively prevented access from the Levant to India, Pakistan, and the Far East ever since humans first moved out of Africa. The second land barrier separated the northern and southern gates from each other (Figure 1.6) and is known today as the Arabian Desert, or by such grim terms as ‘the empty quarter’. One of
the driest deserts in the world, it increased in size to occupy the whole Arabian peninsula as our present ice epoch, the Pleistocene, began to bite 2 million years ago.

  Ice ages mean cooler weather and less evaporation from the sea; this, in turn, meant less rain in the desert belt. In a really severe ice age, like the one we have just had 18,000 years ago, huge volumes of water are locked up in ice sheets over a kilometre thick. During a glacial maximum the sea level falls sufficiently for the normal water exchange between the Indian Ocean and the Red Sea to almost stop.24 There have been two of these events in the past 200,000 years, during our own time on Earth, when the Red Sea effectively became an evaporating salt lake. Most of the plankton died off. Although the Red Sea was more or less sterile, there was still a very narrow water channel of a few kilometres at the mouth, broken up by reefs and islands.

  Like the greening of the Sahara, such extreme glaciations have been rare and rather short-lived. When they do happen, roughly every 100,000 years over the past 2 million, humans from Africa could easily make the southern crossing across the mouth of the Red Sea, with the aid perhaps of primitive rafts, island-hopping where necessary. (The raft option was of course not available to other big mammals.) Even when accessible like this, the main part of the Arabian Peninsula is not a very attractive prospect for emigrants, always being bone dry in the depths of an ice age. There are, however, green refuges in the Yemeni highlands above Aden. The south Arabian coast also benefits from the monsoons thus allowing the beachcombing trail towards the Gulf. Humans may have crossed out of Africa by this southern route at least three times in that past 2 million years.25