Halfway human

SHE IS a woman of mystery-small, but unusually clever, perhaps the Dorothy Parker of her day. No one can say precisely when that was, but some time between 100,000 and 800,000 years ago this bulbous-browed young creature lived, or at least died, somewhere in Central Java. Today, all that remains of her is the top of her head, and it's not talking. However, a new study of her brain structure suggests she could have done exactly that when she was alive.

That conclusion alone would be provocative enough, considering the age of this Java Gal-who goes under the prosaic name of SM3. But the real bombshell may lie in a second study that places her about halfway between Homo erectus-our supposedly dim-witted ancient cousins-and modern Homo sapiens. Cries of "missing link" and "smoking gun" have already gone up from a group of evolutionary anthropologists known as "multiregionalists". Unlike most of their peers, they do not accept the "out-of-Africa" theory, which says that our species evolved in Africa and began migrating across Asia and Europe around 100,000 years ago, ousting all other archaic humans as it went. According to the multiregionalists, hominids that started leaving Africa as early as 2 million years ago were never replaced wholesale by more recent migrations. Instead, these far-flung groups of hominids gradually evolved into us. And halfway specimens from places such as Java are just the sort of proof they've been looking for.

"[SM3] shows the clearest evidence of evolving in the direction of Homo sapiens, and so it becomes a powerful refutation of the replacement theory," says leading multiregionalist Milford Wolpoff of the University of Michigan at Ann Arbor. The findings couldn't have come at a better time for Wolpoff and his minority cause, coming hard on the heels of several offensives the multiregionalists have launched since the beginning of this year. These include the DNA analysis of a 62,000-year-old Australian nicknamed Mungo Man, which indicated that this modern-looking individual had decidedly unmodern genes. And another study, by Wolpoff himself, compared archaic and modern skulls from Australia and from central Europe. This research uncovered a local family resemblance in both cases, which suggests regional continuity of descent.

Any finding that challenges the boundaries between H. erectus and H. sapiens is grist to the multiregionalists' mill. The SM3 results are particularly intriguing, though, because they come from avowed out-of-Africa supporters and non-partisans. But SM3 herself is no stranger to intrigue. In 1999, she made headlines when she turned up in a New York City antiquities shop. She had been smuggled out of Indonesia and had gone missing for 22 years before shop-owner Henry Galliano spotted her and carted her off to the nearby American Museum of Natural History.

Even Galliano recognised how peculiar SM3 looks. She has a conspicuous brow-ridge, a classic sign of H. erectus, but it is not continuous like others of that species. Above the ridge, her forehead is markedly more vertical than any known H. erectus, and wider than it should be. The back of her skull also echoes this broader, more rounded shape, creating an overall globular appearance more characteristic of H. sapiens than H. erectus. Faced with such an oddity, experts at the museum were keen to take a closer look before returning SM3 to Indonesian officials. So Ian Tattersall, curator of anthropology, put Sam M�rquez and Ken Mowbray onto the job.

Determined to be thorough, the two graduate students took a barrage of measurements and CT scans. But it was an interior cast of the skull that immediately piqued their interest. Such endocasts indicate the overall shape of an individual's brain as well as capturing impressions of its surface left in the bone. Once dismissed as so much phrenology, the science of interpreting endocasts has gained accuracy and legitimacy from advances in neuroscience. The SM3 endocast was exceptionally clear and complete, so M�rquez and Mowbray showed it to fellow graduate student Doug Broadfield, a specialist in neuroanatomy. He, in turn, took it to his instructor at Columbia University, Ralph Holloway.

Holloway has made every other existing endocast of Indonesian H. erectus specimens, as well as several of Africans and Neanderthals, so the others awaited his verdict with interest. Two things about the Java Gal struck him. First, there was a distinct asymmetry between the brain's two hemispheres. This pattern, common in many H. erectus skulls, is considered a sign of increasing specialisation for certain tasks, just as it is in modern humans. Then there was the exceptional development of a small bulging area of the frontal lobes, which in modern humans is specific to language. "I think the degree of asymmetry that I saw in Broca's Cap is as much as I've seen in a fossil hominid," says Holloway. Together, he concludes, these features indicate the woman's brain was equipped to put thoughts into words.

Not everyone was thrilled with this analysis. "Both Doug and I are sort of flabbergasted that we've had to take so much flak on this matter," says Holloway. "It's just an open speculation, and why anybody would want to disown us for that, I don't know." But the idea that ancient hominids may have been smarter than many people like to think clearly rankles. It blurs what was a crystal-clear distinction between "us" and "them". After all, stark differences in appearance and demonstrated abilities are what set H. sapiens apart.

But SM3 seems to straddle the divide, appearing even at a glance to combine features of both H. erectus and H. sapiens. It fell to Eric Delson, an anthropologist affiliated with the museum, to classify her. He decided to try a new technique that essentially creates a 3D computer representation of an object, which can then be analysed using statistical methods.

Halfway Human
First Delson recreated the shape of SM3's skull using coordinates from some 200 points on its surface. He did the same for 23 other specimens-11 H. erectus, 10 modern H. sapiens, and 2 archaic H. sapiens. With this information, his computer program arranged the specimens according to their similarity, placing SM3 just about squarely between the erectus and modern humans-while even the archaic H. sapiens fell in among the erectus specimens.

Next, Delson told the computer that there were categories called "erectus", "sapiens" and "archaic", and asked it to find features that would force known members of each group closer together. Again the program lumped archaic H. sapiens in with the erectus specimens. And when Delson asked the computer to categorise the unknown SM3 it still ended up floating between the two main groups, though a bit closer to the erectus cluster. Delson therefore feels comfortable designating SM3 as an erectus, albeit a strange one. He suggests three possible explanations for her odd looks. She could have been an aberrant individual. She could be a member of a hitherto unknown population of Indonesian H. erectus with bulging foreheads. Or, he acknowledges, she could represent "a population evolving in the direction of modern humans". The last explanation is music to the ears of multiregionalists. "That is very daring for him to say," says Wolpoff. "If you thought Homo sapiens began in Africa, as I know [Delson] does, then something in Indonesia has no business evolving in that direction."

But SM3 doesn't even have to be the long-awaited missing link to be troublesome. The finding that she fits so poorly into existing species categories seems to highlight an ongoing problem in paleoanthropology. The experts still argue over what constitutes H. erectus, an enormous group that has been called "the muddle in the middle" because it spans more than a million years and includes many specimens, from Indonesia, Asia and Africa, that don't necessarily resemble one another. Tattersall and other "splitters" are inclined to recognise four or five separate species within the umbrella category of H. erectus. Multiregionalists, by contrast, see this diversity within species as evidence that there really is just one species of Homo, going back at least 2.5 million years, with only relatively superficial variations in appearance over time.

No sex please
The difference in interpretation is crucial to debate between the multiregionalists and supporters of the out-of-Africa theory, because a species is defined as a group that is reproductively separate. By lumping all H. erectus and H. sapiens together, multiregionalists reinforce their claim that modern humans gradually emerged by a process of parallel evolution and interbreeding between populations of hominids spread throughout Africa, Asia and Europe. And by splitting Homo specimens into erectus and sapiens, out-of-Africa proponents slam the door on interbreeding.

Recent studies have only added fuel to the debate. For example, evidence from the three Neanderthals whose DNA has been analysed so far indicates that genetically there is more in common between us and them than there is between members of individual chimpanzee social groups. Few palaeoanthropologists now doubt that early modern humans were biologically capable of interbreeding with other hominids. But the other half of the species definition is whether two groups would in fact interbreed, or whether they would consider themselves too different from one another and remain aloof. "You can't tell that for fossils. You can't tell that in many cases for living animals unless you have an army of graduate students to go out and watch them," says Delson. "For example, chimps and gorillas live in the same forest and they don't interbreed, although, if you take a small female gorilla skull and a large chimp skull, they're very hard to tell apart."

But Wolpoff's recent study did look for proof of interbreeding in bones. He used a checklist of diagnostic features to rate a group of modern and ancient skulls from Europe and Australia as well as archaic H. sapiens from Africa and nearby. He then ranked them all in order of decreasing similarity. In an out-of-Africa scenario, the modern specimens should most closely resemble the archaic sapiens. Instead, Wolpoff found regional similarities-particularly between a 14,000-year-old Australian H. sapiens known as WLH50 and Indonesian "Ngandong" H. erectus specimens. That is as it should be, according to his regional continuity theory. The alternative, he says, means you have to believe "that people came from Africa, replaced the natives around the world, but then came to look just like the natives. And that, to me, is not a probable explanation."

Many researchers remain unconvinced that skull shapes can never resolve this debate. Todd Disotell, a palaeoanthropologist turned molecular evolutionist at New York University, points to the stout frames of cold-adapted Inuit and Andean natives as evidence of how malleable the human skeleton can be, over relatively short periods of time, in response to environmental pressures. "You can tell a lot from fossils about how a population behaved, what kinds of food they ate, and so on," he says, but regional similarities between fossils may simply come from similar adaptations to local conditions. Disotell's genetic analysis reveals that looks can be very deceptive. "I've found monkeys that looked nearly identical that were completely unrelated. Then I found other monkeys that looked completely different, who were the most closely related pairs," he says. "You cannot use these morphological cranial features to infer relationship. The only proof is in the genes."

Here, at least, the out-of-Africa camp feels it is on solid ground. Since the landmark Mitochondrial Eve study was published in 1987, most researchers have been convinced that our remarkably homogeneous genome indicates that ours is a relatively young species with its origins in a single place-Africa, some 200,000 years ago. What's more, recent analyses of Neanderthal DNA suggest that they were different enough from us to rule out the possibility of significant interbreeding between them and our direct ancestors. But multiregionalists choose to interpret these findings in a different way.

Earlier this year, for example, Wolpoff published a paper casting doubt on one of the Neanderthal specimens. The infant, discovered in Mesmaiskaya cave in south-western Russia, was dated at 29,000 years old, by which time modern humans were living in Europe. Wolpoff argued that because an infant does not yet display physical traits that would allow it to be reliably identified as one species or another, the baby could in fact be H. sapiens rather than Neanderthal. Tattersall uses similar logic to dismiss a purported Neanderthal-human hybrid discovered in Portugal as "just a chunky modern kid". And if the Mesmaiskaya cave baby is a modern human, Wolpoff argues, its ancient-looking mitochondrial DNA sequence could simply prove that we all once carried a similar sequence until natural selection favoured its replacement with a more advantageous version.

The misfit of Oz
Echoes of this idea are found in another study by Alan Thorne of the Australian National University in Canberra and his colleagues. They analysed mitochondrial DNA from a 62,000-year-old fossil found near Lake Mungo in New South Wales. Mungo Man looks like a modern human and is classified as H. sapiens, but his mtDNA differs substantially from a modern human's. Thorne concludes that this genetic material is a vestige of his earlier hominid ancestors-probably H. erectus from Indonesia-who contributed to the blood lines of modern Australian Aborigines. This particular genetic sequence has disappeared, perhaps, says Thorne, because Mungo Man's clan died out or selection favoured an alternative.

Since publication, Thorne and his colleagues have been widely criticised for failing to show that they'd protected their ancient genetic sample from contamination. Indeed, only one of the several possible evolutionary trees that can be derived from their data supports an ancient age for the sequence in question. But there are other tantalising studies that claim to have found old genes buried in our genetic closet, which have so far not been explained away. One example is the work of Rosalind Harding of Oxford's Institute of Molecular Medicine. She has analysed two variations of a beta-globin gene, one common in Asia, the other unique to the people of Papua New Guinea and Vanuatu, and Australian Aborigines. Harding calculated that both variants arose more than 200,000 years ago, long before H. sapiens migrated to the region from Africa.

At the very least, research like Harding's and the other recent contributions from the multiregionalist camp have shaken up an academic field that was in danger of becoming fossilised. The out-of-Africa theory is still undoubtedly dominant, but researchers are prepared to admit that this matter is far from settled. Even Rebecca Cann from the University of Hawaii, who headed the Mitochondrial Eve team, acknowledges that we need a lot more evidence. In a special issue of Science, published last month, she stressed the need for more fossil evidence and more information about population sizes and movements to improve the accuracy of genetic analysis. She also called for more cooperation between different disciplines to fit all these pieces of the puzzle together. Only then might one of the models triumph, or some more complex truth emerge about our most recent origins. If only Java Gal could talk now.

Christine Soares is a freelance writer living in New York

New Scientist issue: 14th April 2001

PLEASE MENTION NEW SCIENTIST AS THE SOURCE OF THIS STORY AND, IF PUBLISHING ONLINE, PLEASE CARRY A HYPERLINK TO: http://www.newscientist.com

Further reading: The SM3 findings are published in The Anatomical Record, vol 262:4, p 341 (2001) "Mitochondrial DNA sequences in ancient Australians" by Gregory Adcock and others, Proceedings of the National Academy of Sciences, vol 98, p 537 (2001)
"Modern Human Ancestry at the Peripheries" by Milford Wolpoff, Science, vol 291, p 293 (2001)