The argument over when our lineage split from chimps is about to be settled, with colossal consequences for prehistory
LINE them up in your head. Generation after generation of your ancestors, reaching back in time through civilisations, ice ages, an epic migration out of Africa, to the very origin of our species. And on the other side, take a chimp and line up its ancestors. How far back do you have to go, how many generations have to pass, before the two lines meet?
This is one of the biggest and hardest questions in human evolution. We know that at some point we shared a common ancestor with chimps, but exactly when - and what that ancestor was like - have been maddeningly hard to pin down. Palaeontologists have searched for fossil remains, and geneticists have rummaged through the historical documents that are human and chimp DNA. Both made discoveries, but they did not see eye to eye.
No more. New estimates for when our lineage and chimps went their separate ways suggest that some of our established ideas are staggeringly wrong. If correct, they demand a rewrite of human prehistory, starting from the very beginning.
When was that beginning? The obvious first place to look for answers is in the fossil record. But fossil humans - or more strictly hominins, the group that includes us and all our extinct relatives from after the split - are notoriously thin on the ground and difficult to interpret.
Geneticists have more to work with. DNA contains telltale traces of events in a species’ past, including information about common ancestry and speciation. In theory, calculating the timing of a speciation event should be straightforward. As two species diverge from a common ancestor their DNA becomes increasingly different, largely due to the accumulation of random mutations. The amount of genetic difference between two related species is therefore proportional to the length of time since they diverged. To estimate when the human-chimp split occurred, geneticists can simply count the differences in matching stretches of chimp and human DNA and divide it by the rate at which mutations accumulate. This is known as the molecular clock method.
But there’s a catch. To arrive at the answer you have to know how fast the mutations arise. And that leads you back to square one: you first need to know how long ago we split from chimpanzees.
To get around this catch-22, geneticists turned to orang-utans. Fossils suggest that they split from our lineage between 10 and 20 million years ago. Using this fudge, geneticists arrived at a mutation rate of about 75 mutations per genome per generation. In other words, offspring of humans and chimps each have 75 new mutations that they did not inherit from their parents.
Fossils or DNA
This number rests on several big assumptions, not least that the orang-utan fossil record is a reliable witness - which most agree it is not. Even so, it led to a guess that human ancestors split from chimpanzees between 4 and 6 million years ago.
When fossil-hunters hear this number, they cry foul. The lower end of the estimate is particularly hard to swallow. Australopithecus afarensis - an early hominin from east Africa - already has distinctly human characteristics yet dates back at least 3.85 million years. Its canines were small, for instance. And it walked upright.
Both of these traits are considered hominin, meaning they evolved in our lineage after the split and did not appear on the chimp side. And yet it is hard to see how they could have evolved so quickly, in perhaps as little as 150,000 years after the split.
"Geneticists ignored the palaeontologists completely," says Owen Lovejoy of Kent State University in Ohio. "We would get estimates around 4 million years, and yet there are unmistakable and highly evolved hominins that go back almost 4 million years. To claim a 4 million year divergence date is just silly."
Even a 5 to 6-million-year split was met with scepticism. That’s largely because of three recently discovered fossils from Africa dating from around the same period. All three predate Australopithecus, but still bear unmistakable marks of humanity. Though the interpretation of the remains is controversial, many regard them as being post-split.
Simply put, the palaeontologists were sure there was little chance that the DNA results were accurate. Humanity, they affirmed, had to be older than the geneticists claimed.
History looks set to prove them right. In the past three years, researchers studying human populations have for the first time been able to observe mutations almost as they happen. And that makes all the difference. Instead of relying on an estimate based on rare fossils, we can now watch the molecular clock ticking in real time. “Until we were able to compare genomes of children with their parents, we could not estimate the mutation rate in humans,” says Aylwyn Scally of the Wellcome Trust Sanger Institute in Cambridge, UK.
In September, Augustine Kong of Decode Genetics in Reykjavik, Iceland, and colleagues published one such ground-breaking study. After scanning the genomes of 78 children and their parents to count the number of new mutations in each child’s genome, they found that every child carries an average of 36 new mutations (Nature, vol 488, p 471). Crucially, that is half what was previously assumed, meaning the molecular clock ticks more slowly than we thought - pushing the human-chimp split further back in time (see diagram).
How far back exactly? Earlier this year, Kevin Langergraber at Boston University and his colleagues solved another piece of the puzzle. Mutation rates in studies like Kong’s are measured per generation. To convert this into an estimate of when our ancestors split from chimps, you need to know how long a generation is - in other words, the average age at reproduction. We have a good handle on this for humans, but not in other primates. For chimps, estimates ranged from 15 to 25 years.
Using data from 226 offspring born in eight wild chimp populations, Langergraber found that, on average, chimps reproduce when they are 24-and-a-half (PNAS, vol 109, p 15716). Based on the new numbers, his team estimated the human lineage went its separate way at least 7 million years ago, and possibly as far back as 13 million years ago.
"It’s clear that if this is right, most textbooks dealing with the history of our species will have to be rewritten," says Klaus Zuberbühler at the University of St Andrews, UK, who helped collate data for the study. "The significance can hardly be overestimated."
John Hawks of the University of Wisconsin-Madison agrees. “I think that this will affect pretty much every event in human evolution, from the initial divergence of our lineage to the dispersal out of Africa.”
Perhaps the most significant implication is in the search for the earliest members of the human tribe. For now Australopithecus is the oldest accepted hominin, but an earlier split brings other species into the frame.
The late 1990s and early 2000s was a golden age of discovery for palaeoanthropologists. In the space of a decade, the remains of three potential new hominins were discovered in the deserts of east and central Africa. The most complete was Ardipithecus ramidus, a 4.4-million-year-old skeleton from Afar, Ethiopia, nicknamed Ardi. This was later joined by Sahelanthropus tchadensis, 6 to 7 million years old, and Orrorin tugenensis, about 6 million years old.
Ardipithecus is by far the best known of the three. Roughly the size of a chimp, the skeleton includes human-like teeth, a small skull and the lower limbs of an animal that could walk upright (though it also had an opposable big toe for clasping branches). A possible relative - Ardipithecus kadabba - has also been identified from teeth and a few bone fragments, pushing back the origin of the genus to around 5.8 million years ago.
Sahelanthropus is known from a single skull from Chad, nicknamed Toumai (seephoto). Like Ardipithecus, its teeth are small and human-like, and the middle of its face is short - another human trait. The shape of the hole where the spine would have inserted at the base of the skull hints that it could walk on two legs, although this is hotly debated.
Orrorin, meanwhile, is known only from a handful of teeth plus some leg and finger bones, which suggest it also walked upright but still climbed trees.
All together the bones would barely fill two shoeboxes, but they made a big noise. It was generally thought that when we finally managed to dig up the earliest hominins, we would find something that looked like a chimp. And yet Ardi, Toumai and Orrorin had distinctly human characteristics. “They upset the received wisdom,” says Tim White of the University of California, Berkeley, who led the Ardi discovery.
Some were quick to claim them as human ancestors. But the old molecular clock said otherwise: they were too early. And so they were dismissed as side-branches on the family tree, dead-end experiments in evolution with little or no relevance to the main event.
Now, with the new molecular clock estimates, they are being welcomed back into the fold. “The argument that they are too early has evaporated,” says White, who thinks all three are members of the same genus.
The timing certainly looks right. “If you look at the consensus of recent mutation-rate measurements, Sahelanthropus is just about on the boundary,” says Scally, who recently published a review of the revisions and their consequences for evolution (Nature Review Genetics, vol 13, p 745). “Whether it’s a human, a proto-human, or in a period when humans and chimpanzees are gradually separating, I don’t think anyone can say. But from a genetic perspective, I certainly don’t think you can rule it out, which people used to do.”
The anatomy makes sense too, says White. “It seems to those of us who study these fossils that the way you get from the last common ancestor toAustralopithecus is via something like Ardi. It had already evolved in the direction of Australopithecus. In other words, it’s post-split.”
"Does Ardi represent a species that is on the direct line?" he continues. "We don’t know because we don’t have enough fossils from other places yet. But we can’t rule it out."
Another possibility that cannot be ruled out is that the split is even further back in time. The slow accumulation of mutations means that new estimates of the mutation rate still have big margins for error. In general, geneticists and palaeoanthropologists seem comfortable with a revised figure of 7 to 8 million years. Some, however, go further.
"For me, a 13-million-year-old split could be right on the button," says Lovejoy. "If you go back 10 to 15 million years, the planet was covered in apes, many beginning to show the kinds of anatomical adaptations that you see in modern humans."
Lovejoy is out on his own, though. A week after Kong and colleagues published their new estimate, another team - including many of the same researchers - published another. They analysed DNA from more than 85,000 Icelanders, focusing on short stretches of DNA called microsatellites. According to co-author David Reich of Harvard University these are a more reliable record of mutations.
The rate they found was not quite as slow as Kong’s. As a result, their estimate of the timing of the split is a more constrained 7.5 million years (Nature Genetics, vol 44, p 1161).
There are a few other loose ends to tidy up. Another problem with Kong’s estimate, says Reich, is that if you use it to date the split between orang-utans and African apes - humans, chimps and gorillas - you get something in the range of 30 million years, wildly inconsistent with the maximum 20 million years suggested by the fossil record.
In an attempt to reconcile the two, Scally has proposed that as our ancestors evolved from small primates into large apes, the number of mutations they accumulated with each passing generation decreased. This is in keeping with what is seen in other mammals. “It is observed quite widely, including in primates, that species with larger body size tend to have longer generation times,” says Scally. Longer generations mean slower mutation rates.
This would be plausible, says Reich, if it weren’t that for it to be right, mutation rates in our ancestors and in orang-utans would have had to have dropped at exactly the same time. “I find such an extreme event hard to believe,” he says. Despite that, Reich says, “Scally’s hypothesis is probably the best one out there.”
Quibbles aside, it now seems certain that our lineage is considerably older than we once thought. And that has consequences for the rest of human prehistory. The molecular clock has been used to date a number of key events, not least when our ancestors left Africa. That has been estimated by looking at genetic differences between the Yoruba people of Nigeria and Europeans and Asians.
Early genetic estimates suggested this happened 50,000 years ago. So when fossil remains in Israel and archaeological sites in India were found to be around 100,000 years old, there was some explaining to do. The Israeli bones were dismissed as the remains of an early, dead-end excursion, and the Indian sites as an error, pure and simple. The new molecular clock resolves the discrepancy, pushing the departure back to between 90,000 and 130,000 years ago.
It does something similar for the split between us and Neanderthals. Bones found in a cave in Atapuerca, Spain, and attributed to the probable ancestor of Neanderthals, Homo heidelbergensis, date to between 400,000 and 600,000 years ago. But this created a problem as the molecular clock suggested H. heidelbergensis appeared after that. But the new estimates mean it is in fact around 500,000 years old.
Other key events await revision. But the main finding is clear. The human lineage is significantly older, and our closest living relatives more distant, than we once thought. We are used to thinking of ourselves as separate and distinct from the rest of the animal kingdom. We just got a bit more separate, and a bit more distinct.
Catherine Brahic is New Scientist's environment and life sciences news editor