Marek Kohn

Savannahstan: Beyond Africa and Asia?

This is the 'author's edit' - the submitted version - of an article which appeared in New Scientist as 'Made in Savannahstan', 1 July 2006.

The archaeological excavations at Dmanisi, in the Republic of Georgia, are a glorious exception to the rule that if you are in a hole, you should stop digging. They began as an exploration of a medieval town and turned into a pivotal site for our understanding of human evolution. Five skulls and other remains have been unearthed so far: the individuals they represent are now the central characters in a story whose plot may be about to undergo a major twist.

The familiar story is known as Out of Africa. It tells us that Africa was the region where hominin novelties appeared, and that some of these novelties then dispersed to other continents. There are two main parts to the story. In the first, small upright apes, known as australopithecines, evolve into large and recognisably human creatures that have the wits and build to strike out into Asia. In the second, modern human beings arise in Africa, a million years or so later, and take over the world. The arguments that raged for years over the latter part have somewhat abated; but now the first part of the story is being challenged. A range of questions were marshalled into a provocative critique by Robin Dennell, of the University of Sheffield, and Wil Roebroeks, of Leiden University, that appeared in Nature last December. “Most probably,” it concluded, “we are on the threshold of a profound transformation of our understanding of early hominin evolution …”
At the heart of the earlier story, ‘Out of Africa 1’, lay the assumption that what made us human also propelled us out across the rest of the planet. This idea has a powerful romantic appeal, suggesting that exploration and settlement are primordial and defining human instincts. Homo erectus, the type identified as the first emigrant from Africa, had “a typically insatiable human Wanderlust’’, according to the palaeoanthropologist Ian Tattersall. What enabled it to satisfy the urge to boldly go was the package of characteristically human traits that distinguished it from the australopithecines: longer limbs, increased body and brain size, an omnivorous diet, and the use of stone tools.
Until the past ten years or so, ‘Out of Africa 1’ had timing on its side. The earliest remains of these larger hominins in Africa are about 1.8 million years old. Around 1.5 million years ago (Mya), simple flaked stone tools are joined in the archaeological record by teardrop-shaped handaxes, suggesting a new level of hominin sophistication. Hominin fossils in east Asia were thought to be no older than a million years or so. A loose story came together, in which a type designated either Homo erectus or ergaster evolved in Africa, emerged with the assistance of handaxes or changes in behaviour associated with them, and gradually moved eastwards across Asia.

This tale has now been unravelled, and the handaxes shown to be red herrings, by startlingly early dates from both ends of Asia. Fossils found at the Indonesian site of Sangiran are now believed to be over 1.5 and possibly as much as 1.8 million years old., The Dmanisi finds are dated at 1.7 million years or more. Homo erectus sensu lato (‘in a broad sense’) seems to materialize almost simultaneously in Africa, east Asia and a point in between. Its stone tools were no advance on the first ones to appear in the archaeological record, half a million years previously. And although its brain size, at an adult average of about 700cm3 , was an advance on the australopithecines’ brains, which were under half a litre, it was at the bottom of the H. erectus range.
A still more radical challenge to the supposed role of cognitive abilities in the dispersal of hominins came from a mid-1990s fossil discovery that Robin Dennell considers one of the most important of the past twenty years. Australopithecine fossils had hitherto been found in the Great Rift Valley of eastern Africa and in the south of the continent. Then one turned up in Chad, in the middle of the continent and 2,500 km away from the Rift Valley.

If australopithecines had colonized that region 3 to 3.5 million years ago, argues Dennell, there is no reason why their horizons should have stopped at the Red Sea. This was an era of grasslands, stretching from northern Africa to eastern Asia. “If you were a herbivore that took grass seriously,” Dennell remarks, “you could munch your way all across south-west Asia to northern China.” Such herbivores would later have suited Homo, with its increased reliance on meat; australopithecines also could in theory have spread throughout the great savannah belt. Researchers’ ideas about the different continents’ roles in human evolution have been shaped in the past by their loyalties and their opinions of the regions’ living inhabitants; but ancient hominins would not have distinguished between Africa and Asia. Neither should we, Dennell and Roebroeks suggest. Instead, we could usefully re-imagine this vast transcontinental band of grass as a zone throughout which fundamental hominin adaptations may have taken place. Dennell has dubbed it “Savannahstan”.

The savannahs were the product of climate change, in which global cooling dried out moist woodlands and shifted the balance to grass. At times the global climate altered quite abruptly; these shifts, as the South African palaeontologist Elisabeth Vrba has described, rearranged the fauna. Species vanished; new species emerged. One of these climatic pulses occurred around 2.5 Mya. In the Arctic ice-sheets spread; in eastern Africa forest-adapted antelopes were replaced by ones suited to savannah. New, robust australopithecines appeared, as did somewhat larger-brained hominins that have been designated the first members of the Homo genus, and the earliest known stone tools.

No such climatic disruption coincided with the earliest known examples of Homo erectus sensu lato, dated around 1.8 Mya. Nor does the type have any clearly identifiable immediate ancestors. “Not for nothing has it been described as a hominin ‘without an ancestor, without a clear past’,” observe Dennell and Roebroeks, quoting from Alan Walker and Pat Shipman’s book The Wisdom of Bones. The origin of this form, decisively human in its adaptations, is the problem at the heart of Savannahstan.

Although the Dennell and Roebroeks paper is a challenge to the familiar account rather than an alternative to it, Dennell does have the makings of a story set in Savannahstan. He believes that hominins probably entered Asia long before the dates of the oldest known hominin fossils. These first migrants were either australopithecines or early Homo; he regards these as much the same kind of creatures, noting that Bernard Wood and Mark Collard have suggested that the earliest Homo form, H. habilis, might be better classified as an australopithecine.

Dennell finds strong support in the Dmanisi hominins for his belief that size was no obstacle to dispersal. Their brain sizes fit within the habilis range; evidence from a femur and a tibia, as yet unpublished, indicates that one of them may have weighed about 54kg and been 1.4m in height. Hominins did not need long legs to stride out of Africa.

He does, however, see in them the effects of climatic pressures that are thought to underlie the shift from australopithecine to human adaptations. Australopithecines were adapted to open spaces in woodlands, ranging around relatively small areas, living off plants, seeds, small mammals and perhaps carcasses. As the open spaces expanded into grasslands, some hominins responded by increasing the distances over which they ranged, and turning more to animals as a source of food.

Conventional thinking assumes the site of these changes to have been African, but Dennell argues that Asia may be a more likely location. “Maybe the Rift Valley was a cul-de-sac,” he suggests. Tongue in cheek it may be, but there’s a hint behind the remark of a powerful sense that Asian perspectives have been unfairly neglected. In the 1980s Dennell found stone tools in Pakistan that were dated to 1.9 Mya. From the dominant perspective, which had yet to be confronted by a series of claimed Asian hominin dates nearer to two million years than one, they were in the wrong place at the wrong time. Now they are clearly part of a sparse but definite pattern.

The traces of the global cooling pulse that reordered the African antelopes have also been detected in the soils of China’s Great Loess Plateau. Beneath the silty loess are layers of red clay, which appear to have been blown there by westerly winds. The particles of loess decrease in size from north to south, indicating that they were deposited by northerly winds, the heavier particles falling to the ground first. It appears that the winds changed when the climate cooled, becoming monsoons. The years became polarized into seasons, with increasingly arid summers, causing the grasslands to expand. Asia was the core of this process and Africa was peripheral, argues Dennell, so therefore the selective pressures would have been greater upon hominins in Asia. Horses dispersed from Asia to Africa during this period: maybe hominins did too. African erectus might have Asian roots.
In this perspective the Dmanisi hominins may represent an intermediate form, responding to climatic pressures but still retaining much in common with habilis. They might even be ancestors of the later erectus whose most celebrated representative is the tall, long-limbed ‘Turkana Boy’, with body proportions that epitomise adaptation to a hot dry climate.

Since the Dennell and Roebroeks review was written, American and Georgian researchers studying the Dmanisi finds have published a paper, in the Journal of Human Evolution, that points in similar directions. Suggesting the finds be classed as Homo erectus georgicus, Philip Rightmire and his colleagues conclude that Dmanisi may be “close to the stem from which H. erectus evolved”. They acknowledge the possibility that the Dmanisi population’s ancestors were early Homo emigrants from Africa, and that the dates do not rule out the possibility that erectus evolved in Asia.

Rightmire, of Binghamton University in the state of New York, observes that the scenario in which habilis evolves towards erectus in east Africa has been called into question by the discovery there of fossils that muddle the distinction between the two forms. In the Caucasus, the Dmanisi skulls look like habilis-sized erectus. “For me, the evidence from Dmanisi is critical,” he says. “It seems to me that such a population could well be ancestral to H. erectus (ergaster) in Africa and also to H. erectus in the Far East.” But he anticipates that rewriting the origin and dispersal of Homo erectus will be a slow process. “We're not likely to see a major breakthrough immediately.”

Another recent paper that broadly chimed with Dennell and Roebroeks’s arguments was published in the American Journal of Physical Anthropology  by Alan Templeton, of Washington University, St Louis. By comparing clusters of DNA that vary between individuals and tend to be inherited together, geneticists can identify how long ago particular mutations arose, and use these to map relationships within or between species. Until the past few years, they had to rely on DNA from mitochondria or sex chromosomes, but it is now becoming possible to increase the resolution of such maps by using data from other parts of the genome. Comparing 25 DNA regions in the human genome, Templeton found evidence for an expansion out of Africa around 1.9 Mya, and that gene flow between African and Eurasian populations – in both directions – was established by 1.5 Mya. “The hypotheses they present are testable with molecular genetic data,” he observes, “so I think that the prospects for testing some of their alternatives to ‘Out of Africa 1’ will be excellent in the near future.” Only four years ago, when he first conducted an analysis of this kind, there were insufficient data for him to detect any expansion out of Africa between one and three million years ago. After fifteen further studies with sufficient geographical range and sample size had appeared, he repeated his analysis – and he’ll redo it again once he has another fifteen studies upon which to work. “Given the rate at which the genetic data sets are increasing, I anticipate greater and greater statistical resolution of these older events in human evolutionary history. I would strongly argue that genetics will play an increasing and important role in testing their ideas in conjunction with new fossil and archaeological discoveries.”

For Robin Dennell, however, the objects in the ground are what matters. He and Roebroeks argue that absence of evidence is not necessarily evidence of absence. But they need to find Asian evidence to balance the generous legacies of the Rift Valley, gently letting erosion expose its well-preserved contents, and the southern African caves. Dennell mentions a few countries in he would like to search, such as Saudi Arabia, Iran, Afghanistan – a list, in fact, of places not to visit these days unless absolutely necessary. The Pakistani site where he found the ancient tools is also off limits because of the political turbulence that has spread across the region. He will have to wait for a change in the climate.



The shifting spotlight

In The Descent of Man (1871) Charles Darwin observed that as the gorilla and chimpanzee “are now man's nearest allies, it is somewhat more probable that our early progenitors lived on the African continent than elsewhere”. But during his lifetime Europe was where the hominin discoveries were made, and the evidence to confirm his suggestion did not begin to accumulate until half a century later.

The first major hominin discovery outside Europe was made in Java, in 1891, by Eugène Dubois, who later named it Pithecanthropus erectus. Similar fossils were found during the 1920s and 1930s, both in Java and at a site near Peking (now Beijing). The latter, designated Sinanthropus pekinensis, were lost during the turmoil of war in 1941. In 1950, the evolutionary biologist Ernst Mayr proposed that these Asian specimens should be classified as a single species, Homo erectus.

Similar specimens were not found in Africa until the 1960s, after some forty years of australopithecine and early or uncertain Homo discoveries. The first finds were made by the Leakey family’s project at Olduvai Gorge, in Tanzania; the most spectacular, of the nearly-complete ‘Turkana Boy’ skeleton, was made by the Leakey group at Lake Turkana in Kenya in 1984. By this stage Africa’s founding role in human evolution was indisputable. Notwithstanding disagreement about whether the African fossils should be designated erectus or ergaster, it looked as though all hominin life was there.

In the 1990s, however, the possibility of a significant role for Asia was revived by the acceptance that H. erectus had appeared there much earlier than previously thought; and by the Dmanisi excavations, which yielded their first skulls in 1999.