Out of Eden: The Peopling of the World

by Oppenheimer, Stephen

  After humans first walked across into Asia 2 million years ago, the desert barriers separating Europe and the Middle East from Asia were closed. Crossing from Africa by the southern route, they could move on and into India only by making their way along the coast, while those taking the northern route could only go into the Levant, the Caucasus, and Europe.

  These complex corrals and barriers to migration into Asia set the pattern for the human colonization of the rest of the world. As each new version of the genus Homo arose in Africa, some tribes would take the northern route during a warm interglacial, while others would take the southern route during an ice age. The first humans to take the northern route during an interglacial, around 1.8 million years ago, reached Dmanisi in Georgia. To look at, they were perhaps closer to Homo habilis, the earliest and most primitive African human, than to their East Asian erectus contemporaries. Somewhere between early African erectus (Homo ergaster) and Homo habilis they have been assigned a new species name, Homo georgicus. This parallel colonization by two different early human species taking different routes supports the view that human exits from Africa before the closing of the southern route had more to do with opportunities of geography available to any African mammal than with some special behaviour of Homo erectus, as was previously thought.26

  Figure 1.6 (see previous pages) Single southern route out of Africa. The full beachcomber route from the Red Sea along the Indo-Pacific coast to Australia and beyond, including likely extensions to China, Japan and New Guinea. Vegetation and sea-level shown as at 65–85,000 years ago. Note how the extent of desert throughout this period prevented access to Northern Eurasia.

  Dmanisi humans seem, on the face of it, to have died out long ago in the Caucasus and the Levant, when it got colder. As we have just seen, the same tragedy struck modern humans some 2 million years later, after the Eemian interglacial. When the Levant dried out after a few thousand years, the Dmanisi humans were unable to escape north, south, or east, and so perished. It is perhaps a tribute to the next emigrants, Homo heidelbergensis, that their offspring were more successful in repeatedly colonizing Europe, possibly three times over the last 800,000 years. The last of these northern migrations may have taken place during an interglacial as recently as 250,000 years ago. By this time the African source Homo species, although still beetle-browed, had evolved larger brains and now made the more advanced Middle Palaeolithic tools. Foley and Lahr call the ancestor of that last African emigrant Homo helmei.27 They postulate that Homo helmei evolved in Europe into our first cousin, Homo neanderthalensis (see Plate 4 and Chapter 2).

  Over nearly the same period (160,000–800,000 years ago) there could have been a similar number of human migrations to the Far East by the southern route. These large-brained Asians were similar, though not identical to, their European cousins. In the same way that the Neanderthals acquired bigger brains, later eastern migrations out of Africa 200,000 years ago may also have given rise to larger-brained types such as the Maba and Dali skulls in China,28 and even possibly ‘Madeleine’ in Java (see Figure 0.2).

  We can even approximately date the last of these southern emigrations from Africa to a time before the appearance of modern humans. The valley of Narmada in central India was home to a large-brained archaic human whose skull has been only poorly dated but who certainly lived over 150,000 years ago. Although it has yielded few pre-modern human fossils, India is much richer in the stone artefacts that they made. The older, so-called Acheulian (Lower Palaeolithic) tools in India date roughly from 160,000 to 670,000 years ago, while the newer Middle Palaeolithic tools from Africa appear in India at the end of that period, around 160,000 years ago.29 This suggests that the last movement of pre-modern humans out of Africa could have taken place as recently as 160,000 years ago, during the dry glacial age that preceded the dawn of modern humans.

  The final change from Middle Palaeolithic to the Upper Palaeolithic tools is also seen in the Central Narmada and Middle Son Basins in northern India, associated with a volcanic ash layer from the great Toba explosion on the island of Sumatra around 74,000 years ago. This suggests a major culture change around that time.30 Upper Palaeolithic tools were characteristic of the first modern humans to arrive in the Levant and Europe, 25,000–30,000 years later.

  Cold feet?

  A southern route across the Red Sea which is most accessible when the Earth is glaciated goes some way to explaining the mounting evidence for our own defining exodus from Africa. I think it came not during the warm Eemian interglacial period 125,000 years ago, but 45,000 years later, at the threshold of a prolonged glaciation. As we now look at the evidence for this founding migration and even try to narrow the date down, there are three obvious, related questions: why we crossed the Red Sea, when we did, and why we did not cross before.

  The answer to the last of these questions – why we waited so long before taking such a short sea-hop – 11 km (7 miles) during glacial periods, narrower than the mouths of many rivers – is perhaps the most revealing. We humans may be wilful, aggressive, and impatient, but we are also resourceful and smart. We would not think of migrating en masse to another planet, or overseas to the next continent, if we were not happy with conditions as reported by scouts. We would also like to be convinced of the chances of surviving the round trip, should we be forced to return. Reconstruction of the stepwise prehistoric migrations by the Polynesians in the Pacific over the past few thousand years tells us that much.

  Judging by how rapidly modern humans subsequently spread around the coast of the Indian Ocean to Australia, the first modern Asians could easily have taken rafts or boats across the narrow mouth of the Red Sea many times during the first 70,000 years that their forebears lived in East Africa. So what was stopping them? The answer, almost certainly, was the presence of other people on the other side of the water – perhaps detected by scouting expeditions. Clearly those other people would not have been modern humans, but more likely the nearest thing: Asian descendants of African Homo helmei. The Indian evidence referred to above supports this view, as do the abundant Middle Palaeolithic tools found across the water in the Yemen and southern Arabia.

  A recent book on human origins31 has used the multi-species bar scene in the movie Star Wars to emphasize that, in the few hundred thousand years before modern humans arrived, a number of human species had succeeded in cohabiting on this planet for long periods without exterminating one another. This probably had more to do with the fact that each subspecies usually occupied separate continents and islands than with their being good neighbours.

  When several human species did try to share a continent, the result was usually a rapid reduction to one species. Fossil evidence in Africa shows that the birth of modern humans 130,000 or more years ago signalled the rapid end of the record for our parents Homo helmei there. The last of these were dated to around 130,000 years ago. Whether the near extinction of all African human species during the very dry period 140,000–190,000 years ago simply left modern humans as the only survivor, or whether they ‘cleansed’ Africa of Homo helmei, we can make a good guess. This guess is based on our recent track record of successful and attempted genocides in Tasmania, Germany, Rwanda, and the Balkans. The Tasmanian extinction in the nineteenth century was the only complete racial annihilation achieved in these deliberate attempts at destroying an entire people, and took a relatively short time, presumably because the victims were all confined to and could not escape from what became their island grave. So, according to modern historical records, we are ready, willing, and able to exterminate a neighbour of our own kind. Given a few thousand years there is every reason to suppose we could have achieved the same result on another species at continent level.32

  The first steps into a new continent are the most difficult. The extraordinarily successful prehistoric expansion from Island Southeast Asia in which the ancestors of the Polynesians occupied the islands of the Indo-Pacific region tells us something about moving into new territory. Although they had sup
reme sailing technology, envied even by Captain Cook, they always avoided settling islands that already had people living on them. Where the length of their eastern sailing trips left no choice but to make a ‘guest’ landfall, for instance on the large inhabited island of New Guinea, their colonies have remained literally ‘on the beach’ to this day, thousands of years later.33 For a small band of invaders to cross by sea and, with superior military technology, easily take over a populated continent has become possible only in the past 500 years.

  As survivors, modern humans may have had the continent of Africa to themselves from 130,000 years ago. If, as is likely, the earliest modern East Africans were aware of archaic beetle-browed neighbours with similar abilities across the water in the Yemen, they may well have decided that their dominance in Africa was good enough for them, and that discretion was the better part of valour for the time being. We will know better what was waiting on the other side of the Red Sea when the Palaeolithic of the Yemen is more thoroughly investigated and dated. The early excavations conducted by Cambridge archaeologist Gertrude Caton-Thompson have, however, clearly established that there was human occupation there during the Palaeolithic.34

  Time to move

  Much has been made recently of a ‘beachcombing’ lifestyle among the first modern Africans as a motive for spreading out of Africa and around the Indian Ocean (see Figure 1.6). For most of their first 2 million years on Earth, humans were roaming the savannah as hunter-gatherers. Like the Kalahari of southern Africa, they exploited the rich nutritional value of group-hunted game, supplemented by salads of roots, fruit, and leaves. As the major glaciation of 130,000–190,000 years ago began to reduce their savannah range, someone had the idea of foraging and eating shellfish and other marine produce from the beach. It is always possible that beachcombing started even earlier, but since the beaches are now under water, we will never know. Such a diet, being rich in protein and Omega-3 fatty acids, is nutritious, good for the brain, and easier to obtain than game. Beach tucker has the added advantage of remaining available when the savannah dries up during an ice age.

  Evidence for such beachcombing is unexpectedly easy to assess, since characteristic piles of split shells (shell middens) are left behind. There is, however, a problem in knowing just how long humans have been doing it. Shell middens are generally found just above the high-tide line, but for most of the past 200,000 years sea levels have been many metres below today’s beaches. This means we would expect to miss most ancient middens except those laid down during the high sea levels of an interglacial, such as the one 125,000 years ago.

  Neanderthals combed beaches in Spain and Italy 60,000 years ago, so it is possible that they brought the practice with them from Africa. Until recently, however, the earliest evidence for African beachcombing came from Klasies River mouth in South Africa, dated to between 100,000 and 115,000 years ago. In 2000, however, new evidence was discovered for early beachcombing at Abdur (see Plate 8) on the Eritrean west coast of the Red Sea, just to the north of the Gate of Grief. Dated to 125,000 years ago, at the peak of the Eemian interglacial, the same beach site yielded butchered remains of large mammals, indicating a mixed diet. The implements, which included blades made from obsidian, a volcanic glass, are most likely to have been made by modern humans.35

  The great interest in this site on the Red Sea is twofold: it provides us with the oldest evidence for beachcombing anywhere and it is very close to the southern route out of Africa. Both aspects feed into an attractive model, which may be called the ‘beach-buggy to Australia’. We get a compelling story of beachcombers multiplying until their patch of beach could support them no longer, then moving on to the next unexploited beach, and so on. By such rapid progression, once over the Red Sea the vanguard would just have followed the coast of the Indian Ocean, eating their way right down to Indonesia within 10,000 years. The low sea levels of the time would have allowed a dry walk from Aden to the tip of Java, and then easy island hops to Australia, where shell middens are found from the earliest traces of human habitation.

  I am pretty sure that this model of the early colonization of Australia is correct, but the dates have to fit, not only for the archaeological evidence but also for the molecular clock on the gene tree for all the other Eurasian dispersals. If, on top of this beach buggy model, we impose only a single out-of-Africa exodus to colonize both Australia and the rest of the world, we can start to make strong predictions for the order and dates of colonization en route of India, Southeast Asia, and the parallel movement to New Guinea. These predictions should be the test of the theory.

  Why and when did we leave Africa?

  One problem with the genetic molecular clock is that we have to allow rather generous margins for error. Thus, although we can be confident that the gene tree is telling us there was only one successful exodus, the approximate genetic dates say that this could have taken place any time between 65,000 and 95,000 years ago.36 Our ancestors could thus have been collecting shells for some 60,000 years on the western shore of the Red Sea – from at least 125,000 years ago – before deciding to move on.

  Before looking at the archaeological evidence for that move, it might help to establish whether there was a precipitating climatic event, as in so much of our prehistory. Could a transient increase in the severity of glaciation have brought the sea level at the Gate of Grief to a point where our ancestors could have walked across from reef to reef? The increasing aridity of the East African coast might have reduced available drinking water, thus making the rather wetter monsoon conditions on the southern Yemen coast more attractive to the beachcombers. A ford across to Arabia might have been just the inducement to overcome any remaining fears of the neighbours. The latter, in turn, perhaps lacking beachcombing skills, may have left southern Arabia altogether because of the drought.

  In spite of the attractive biblical allusion of this scenario, the problem with a dry exodus across the mouth of the Red Sea 60,000–90,000 years ago is that oceanographic evidence denies that the Gate of Grief was ever completely dry during our time on Earth. Such an event nearly happened three times during major glaciations of the past half million years: 440,000 years ago, 140,000 years ago, and at the last glacial maximum 18,000 years ago.37

  The strait was very much narrower during glacial periods, allowing easy island-hopping across the shallows and the reef islands of the Hanish al Kubra, at the northern end of the isthmus. As we have seen, this must have been where Asian Homo erectus crossed, so it should not have been much of a problem for our pioneers to complete the journey to Australia within a few thousand years. Although the Red Sea did not part, there was a major cold event around the time of our first exodus from Africa (Figure 1.7). Measurements made on the Greenland ice cap show that the second-coldest time of the last 100,000 years was between 60,000 and 80,000 years ago.38 At its coldest, 65,000 years ago, this glaciation took the world’s sea levels 104 metres (340 feet) below today’s levels. The way in which this sea lowstand (as it is called) was measured in the Red Sea holds an unexpected clue for what might have been the stimulus to move.

  Oceanographers have been able to measure interglacial sea-level highstands by looking at coral reefs. Sea-level lowstands which occurred during glacial maxima are more difficult to confirm. Eelco Rohling, an oceanographer at Southampton University, has found a way to use the Red Sea to overcome this problem. He measured prehistoric levels of Red Sea plankton. During a glacial maximum, the Gate of Grief approaches closure and the Red Sea becomes effectively isolated from exchange with the Indian Ocean. Evaporation causes the Red Sea’s salinity to increase so much that plankton, the base of the marine food chain, disappears. During an interglacial period, like now, the high sea level allows the Red Sea to flush out its salt, and sea life can start again. Sea-level lowstands like the one 65,000 years ago did not completely block the mouth of the Red Sea, so the plankton, although severely affected, did not completely disappear. The low level of plankton allowed Rohling and colleagues to i
mprove the previous estimates for the sea-level depression at this time.39

  Depression of plankton is worsened by low oxygen levels and high water temperatures in the shallows, as happens near beaches at Abdur in Eritrea and at the mouth of the Red Sea. Low levels of plankton in the Red Sea are, in turn, likely to have affected the success of the beachcombers. By contrast, the beaches of the Gulf of Aden just across the water in the Yemen were outside the Gate of Grief and had nutrient-rich, oxygenated upwelling water from the Indian Ocean. In other words, over on the southern Arabian coast the beachcombing conditions were probably excellent.40

  So perhaps dwindling food resources on the western shores of the Red Sea, attractive beaches on the Gulf of Aden, and cool wet Yemeni uplands for refuge were what spurred our ancestors to take their momentous step. The question still remains of when: did they wait until the sea levels were lowest and the beachcombing was at its worst, 65,000 years ago, or did they move earlier, when things first began to deteriorate? Again, the Red Sea plankton may give a clue. Plankton and sea levels did not decline evenly between 100,000 and 65,000 years ago. Instead there was a short, sharp depression of both at 85,000 years ago, with sea levels falling to 80 metres below present, followed by an equally dramatic and brief improvement 83,000 years ago.41 Maybe that 85,000 year old dip was the spur that set our ancestors on their beachcombing trail.

  Evidence of another great climatic catastrophe during that sea-level trough comes from cores drilled in the Arabian seabed.42 This was the volcanic eruption of Toba in Sumatra, 74,000 years ago. Known to be by far the biggest eruption of the last 2 million years, this mega-bang caused a prolonged nuclear winter and released ash in a huge plume that spread to the north-west and covered India, Pakistan, and the Gulf region in a blanket 1–3 metres (3–10 feet) deep.