Erectus Rising

Oh no. Not this. The hominids are acting up again.

By James Shreeve|Thursday, September 01, 1994
Just when it seemed that the recent monumental fuss over the origins of modern human beings was beginning to quiet down, an ancient ancestor is once more running wild. Trampling on theories. Appearing in odd places, way ahead of schedule. Demanding new explanations. And shamelessly flaunting its contempt for conventional wisdom in the public press.

The uppity ancestor this time is Homo erectus--alias Java man, alias Peking man, alias a mouthful of formal names known only to the paleontological cognoscenti. Whatever you call it, erectus has traditionally been a quiet, average sort of hominid: low of brow, thick of bone, endowed with a brain larger than that of previous hominids but smaller than those that followed, a face less apelike and projecting than that of its ancestors but decidedly more simian than its descendants'. In most scenarios of human evolution, erectus's role was essentially to mark time--a million and a half years of it--between its obscure, presumed origins in East Africa just under 2 million years ago and its much more recent evolution into something deserving the name sapiens.

Erectus accomplished only two noteworthy deeds during its long tenure on Earth. First, some 1.5 million years ago, it developed what is known as the Acheulean stone tool culture, a technology exemplified by large, carefully crafted tear-shaped hand axes that were much more advanced than the bashed rocks that had passed for tools in the hands of earlier hominids. Then, half a million years later, and aided by those Acheulean tools, the species carved its way out of Africa and established a human presence in other parts of the Old World. But most of the time, Homo erectus merely existed, banging out the same stone tools millennium after millennium, over a time span that one archeologist has called "a period of unimaginable monotony."

Or so read the old script. These days, erectus has begun to ad- lib a more vigorous, controversial identity for itself. Research within the past year has revealed that rather than being 1 million years old, several erectus fossils from Southeast Asia are in fact almost 2 million years old. That is as old as the oldest African members of the species, and it would mean that erectus emerged from its home continent much earlier than has been thought--in fact, almost immediately after it first appeared. There's also a jawbone, found in 1991 near the Georgian city of Tbilisi, that resembles erectus fossils from Africa and may be as old as 1.8 million years, though that age is still in doubt. These new dates--and the debates they've engendered--have shaken Homo erectus out of its interpretive stupor, bringing into sharp relief just how little agreement there is on the rise and demise of the last human species on Earth, save one.

"Everything now is in flux," says Carl Swisher of the Berkeley Geochronology Center, one of the prime movers behind the redating of erectus outside Africa. "It's all a mess."

The focal point for the flux is the locale where the species was first found: Java. The rich but frustration-soaked history of paleoanthropology on that tropical island began just over 100 years ago, when a young Dutch anatomy professor named Eugène Dubois conceived the idée fixe that the "missing link" between ape and man was to be found in the jungled remoteness of the Dutch East Indies. Dubois had never left Holland, much less traveled to the Dutch East Indies, and his pick for the spot on Earth where humankind first arose owed as much to a large part of the Indonesian archipelago's being a Dutch colony as it did to any scientific evidence. He nevertheless found his missing link--the top of an oddly thick skull with massive browridges--in 1891 on the banks of the Solo River, near a community called Trinil in central Java. About a year later a thighbone that Dubois thought might belong to the same individual was found nearby; it looked so much like a modern human thighbone that Dubois assumed this ancient primate had walked upright. He christened the creature Pithecanthropus erectus--"erect ape-man"--and returned home in triumph.

Finding the fossil proved to be the easy part. Though Dubois won popular acclaim, neither he nor his "Java man" received the full approbation of the anatomists of the day, who considered his ape-man either merely an ape or merely a man. In an apparent pique, Dubois cloistered away the fossils for a quarter-century, refusing others the chance to view his prized possessions. Later, other similarly primitive human remains began to turn up in China and East Africa. All shared a collection of anatomical traits, including a long, low braincase with prominent browridges and a flattened forehead; a sharp angle to the back of the skull when viewed in profile; and a deep, robustly built jaw showing no hint of a chin. Though initially given separate regional names, the fossils were eventually lumped together into one far-flung taxon, a creature not quite like us but human enough to be welcomed into our genus: Homo erectus.

Over the decades the most generous source of new erectus fossils has been the sites on or near the Solo River in Java. The harvest continues: two more skulls, including one of the most complete erectus skulls yet known, were found at a famous fossil site called Sangiran just in the past year. Though the Javan yield of ancient humans has been rich, something has always been missing--the crucial element of time. Unless the age of a fossil can be determined, it hangs in limbo, its importance and place in the larger scheme of human evolution forever undercut with doubt. Until researchers can devise better methods for dating bone directly--right now there are no techniques that can reliably date fossilized, calcified bone more than 50,000 years old--a specimen's age has to be inferred from the geology that surrounds it. Unfortunately, most of the discoveries made on the densely populated and cultivated island of Java have been made not by trained excavators but by sharp-eyed local farmers who spot the bones as they wash out with the annual rains and later sell them. As a result, the original location of many a prized specimen, and thus all hopes of knowing its age, are a matter of memory and word of mouth.

Despite the problems, scientists continue to try to pin down dates for Java's fossils. Most have come up with an upper limit of around 1 million years. Along with the dates for the Peking man skulls found in China and the Acheulean tools from Europe, the Javan evidence has come to be seen as confirmation that erectus first left Africa at about that time.

There are those, however, who have wondered about these dates for quite some time. Chief among them is Garniss Curtis, the founder of the Berkeley Geochronology Center. In 1971 Curtis, who was then at the University of California at Berkeley, attempted to determine the age of a child's skull from a site called Mojokerto, in eastern Java, by using the potassium-argon method to date volcanic minerals in the sediments from which the skull was purportedly removed. Potassium-argon dating had been in use since the 1950s, and Curtis had been enormously successful with it in dating ancient African hominids--including Louis Leakey's famous hominid finds at Olduvai Gorge in Tanzania. The method takes advantage of the fact that a radioactive isotope of potassium found in volcanic ash slowly and predictably decays over time into argon gas, which becomes trapped in the crystalline structure of the mineral. The amount of argon contained in a given sample, measured against the amount of the potassium isotope, serves as a kind of clock that tells how much time has passed since a volcano exploded and its ash fell to earth and buried the bone in question.

Applying the technique to the volcanic pumice associated with the skull from Mojokerto, Curtis got an extraordinary age of 1.9 million years. The wildly anomalous date was all too easy to dismiss, however. Unlike the ash deposits of East Africa, the volcanic pumices in Java are poor in potassium. Also, not unexpectedly, a heavy veil of uncertainty obscured the collector's memories of precisely where he had found the fossil some 35 years earlier. Besides, most paleontologists were by this time firmly wedded to the idea that Africa was the only human-inhabited part of the world until 1 million years ago. Curtis's date was thus deemed wrong for the most stubbornly cherished of reasons: because it couldn't possibly be right.

In 1992 Curtis--under the auspices of the Institute for Human Origins in Berkeley--returned to Java with his colleague Carl Swisher. This time he was backed up by far more sensitive equipment and a powerful refinement in the dating technique. In conventional potassium-argon dating, several grams' worth of volcanic crystals gleaned from a site are needed to run a single experiment. While the bulk of these crystals are probably from the eruption that covered the fossil, there's always the possibility that other materials, from volcanoes millions of years older, have gotten mixed in and will thus make the fossil appear to be much older than it actually is. The potassium-argon method also requires that the researcher divide the sample of crystals in two. One half is dissolved in acid and passed through a flame; the wavelengths of light emitted tell how much potassium is in the sample. The other half is used to measure the amount of argon gas that's released when the crystals are heated. This two-step process further increases the chance of error, simply by giving the experiment twice as much opportunity to go wrong.

The refined technique, called argon-argon dating, neatly sidesteps most of these difficulties. The volcanic crystals are first placed in a reactor and bombarded with neutrons; when one of these neutrons penetrates the potassium nucleus, it displaces a proton, converting the potassium into an isotope of argon that doesn't occur in nature. Then the artificially created argon and the naturally occurring argon are measured in a single experiment. Because the equipment used to measure the isotopes can look for both types of argon at the same time, there's no need to divide the sample, and so the argon-argon method can produce clear results from tiny amounts of material.

In some cases--when the volcanic material is fairly rich in potassium--all the atoms of argon from a single volcanic crystal can be quick-released by the heat from a laser beam and then counted. By doing a number of such single-crystal experiments, the researchers can easily pick out and discard any data from older, contaminant crystals. But even when the researchers are forced to sample more than one potassium-poor crystal to get any reading at all--as was the case at Mojokerto--the argon-argon method can still produce a highly reliable age. In this case, the researchers carefully heat a few crystals at a time to higher and higher temperatures, using a precisely controlled laser. If all the crystals in a sample are the same age, then the amount of argon released at each temperature will be the same. But if contaminants are mixed in, or if severe weathering has altered the crystal's chemical composition, the argon measurements will be erratic, and the researchers will know to throw out the results.

Curtis and Swisher knew that in the argon-argon step-heating method they had the technical means to date the potassium-poor deposits at Mojokerto accurately. But they had no way to prove that those deposits were the ones in which the skull had been buried: all they had was the word of the local man who had found it. Then, during a visit to the museum in the regional capital, where the fossil was being housed, Swisher noticed something odd. The hardened sediments that filled the inside of the fossil's braincase looked black. But back at the site, the deposits of volcanic pumice that had supposedly sheltered the infant's skull were whitish in color. How could a skull come to be filled with black sediments if it had been buried in white ones? Was it possible that the site and the skull had nothing to do with each other after all? Swisher suspected something was wrong. He borrowed a penknife, picked up the precious skull, and nicked off a bit of the matrix inside.

"I almost got kicked out of the country at that point," he says. "These fossils in Java are like the crown jewels."

Luckily, his impulsiveness paid off. The knife's nick revealed white pumice under a thin skin of dark pigment: years earlier, someone had apparently painted the surface of the hardened sediments black. Since there were no other deposits within miles of the purported site that contained a white pumice visually or chemically resembling the matrix in the skull, its tie to the site was suddenly much stronger. Curtis and Swisher returned to Berkeley with pumice from that site and within a few weeks proclaimed the fossil to be 1.8 million years old, give or take some 40,000 years. At the same time, the geochronologists ran tests on pumice from the lower part of the Sangiran area, where erectus facial and cranial bone fragments had been found. The tests yielded an age of around 1.6 million years. Both numbers obviously shatter the 1-million-year barrier for erectus outside Africa, and they are a stunning vindication of Curtis's work at Mojokerto 20 years ago. "That was very rewarding," he says, "after having been told what a fool I was by my colleagues."

While no one takes Curtis or Swisher for a fool now, some of their colleagues won't be fully convinced by the new dates until the matrix inside the Mojokerto skull itself can be tested. Even then, the possibility will remain that the skull may have drifted down over the years into deposits containing older volcanic crystals that have nothing to do with its original burial site, or that it was carried by a river to another, older site. But Swisher contends that the chance of such an occurrence is remote: it would have to have happened at both Mojokerto and Sangiran for the fossils' ages to be refuted. "I feel really good about the dates," he says. "But it has taken me a while to understand their implications."

The implications that can be spun out from the Javan dates depend on how one chooses to interpret the body of fossil evidence commonly embraced under the name Homo erectus. The earliest African fossils traditionally attributed to erectus are two nearly complete skulls from the site of Koobi Fora in Kenya, dated between 1.8 and 1.7 million years old. In the conventional view, these early specimens evolved from a more primitive, smaller-brained ancestor called Homo habilis, well represented by bones from Koobi Fora, Olduvai Gorge, and sites in South Africa.

If this conventional view is correct, then the new dates mean that erectus must have migrated out of Africa very soon after it evolved, quickly reaching deep into the farthest corner of Southeast Asia. This is certainly possible: at the time, Indonesia was connected to Asia by lower sea levels--thus providing an overland route from Africa--and Java is just 10,000 to 15,000 miles from Kenya, depending on the route. Even if erectus traveled just one mile a year, it would still take no more than 15,000 years to reach Java--a negligible amount of evolutionary time.

If erectus did indeed reach Asia almost a million years earlier than thought, then other, more controversial theories become much more plausible. Although many anthropologists believe that the African and Asian erectus fossils all represent a single species, other investigators have recently argued that the two groups are too different to be so casually lumped together. According to paleoanthropologist Ian Tattersall of the American Museum of Natural History in New York, the African skulls traditionally assigned to erectus often lack many of the specialized traits that were originally used to define the species in Asia, including the long, low cranial structure, thick skull bones, and robustly built faces. In his view, the African group deserves to be placed in a separate species, which he calls Homo ergaster.

Most anthropologists believe that the only way to distinguish between species in the fossil record is to look at the similarities and differences between bones; the age of the fossil should not play a part. But age is often hard to ignore, and Tattersall believes that the new evidence for what he sees as two distinct populations living at the same time in widely separate parts of the Old World is highly suggestive. "The new dates help confirm that these were indeed two different species," he says. "In my view, erectus is a separate variant that evolved only in Asia."

Other investigators still contend that the differences between the African and Asian forms of erectus are too minimal to merit placing them in separate species. But if Tattersall is right, his theory raises the question of who the original emigrant out of Africa really was. Homo ergaster may have been the one to make the trek, evolving into erectus once it was established in Asia. Or perhaps a population of some even more primitive, as-yet-unidentified common ancestor ventured forth, giving rise to erectus in Asia while a sister population evolved into ergaster on the home continent.

Furthermore, no matter who left Africa first, there's the question of what precipitated the migration, a question made even more confounding by the new dates. The old explanation, that the primal human expansion across the hem of the Old World was triggered by the sophisticated Acheulean tools, is no longer tenable with these dates, simply because the tools had not yet been invented when the earliest populations would have moved out. In hindsight, that notion seems a bit shopworn anyway. Acheulean tools first appear in Africa around 1.5 million years ago, and soon after at a site in the nearby Middle East. But they've never been found in the Far East, in spite of the abundant fossil evidence for Homo erectus in the region.

Until now, that absence has best been explained by the "bamboo line." According to paleoanthropologist Geoffrey Pope of William Paterson College in New Jersey, erectus populations venturing from Africa into the Far East found the land rich in bamboo, a raw material more easily worked into cutting and butchering tools than recalcitrant stone. Sensibly, they abandoned their less efficient stone industry for one based on the pliable plant, which leaves no trace of itself in the archeological record. This is still a viable theory, but the new dates from Java add an even simpler dimension to it: there are no Acheulean tools in the Far East because the first wave of erectus to leave Africa didn't have any to bring with them.

So what did fuel the quick-step migration out of Africa? Some researchers say the crucial development was not cultural but physical. Earlier hominids like Homo habilis were small-bodied creatures with more apelike limb proportions, notes paleoanthropologist Bernard Wood of the University of Liverpool, while African erectus was built along more modern lines. Tall, relatively slender, with long legs better able to range over distance and a body better able to dissipate heat, the species was endowed with the physiology needed to free it from the tropical shaded woodlands of Africa that sheltered earlier hominids. In fact, the larger-bodied erectus would have required a bigger feeding range to sustain itself, so it makes perfect sense that the expansion out of Africa should begin soon after the species appeared. "Until now, one was always having to account for what kept erectus in Africa so long after it evolved," says Wood. "So rather than raising a problem, in some ways the new dates in Java solve one."

Of course, if those dates are right, the accepted time frame for human evolution outside the home continent is nearly doubled, and that has implications for the ongoing debate over the origins of modern human beings. There are two opposing theories. The "out of Africa" hypothesis says that Homo sapiens evolved from erectus in Africa, and then--sometime in the last 100,000 years--spread out and replaced the more archaic residents of Eurasia. The "multiregional continuity" hypothesis says that modern humans evolved from erectus stock in various parts of the Old World, more or less simultaneously and independently. According to this scenario, living peoples outside Africa should look for their most recent ancestors not in African fossils but in the anatomy of ancient fossils within their own region of origin.

As it happens, the multiregionalists have long claimed that the best evidence for their theory lies in Australia, which is generally thought to have become inhabited around 50,000 years ago, by humans crossing over from Indonesia. There are certain facial and cranial characteristics in modern Australian aborigines, the multiregionalists say, that can be traced all the way back to the earliest specimens of erectus at Sangiran--characteristics that differ from and precede those of any more recent, Homo sapiens arrival from Africa. But if the new Javan dates are right, then these unique characteristics, and thus the aborigines' Asian erectus ancestors, must have been evolving separately from the rest of humankind for almost 2 million years. Many anthropologists, already skeptical of the multiregionalists' potential 1-million-year-long isolation for Asian erectus, find a 2-million-year-long isolation exceedingly difficult to swallow. "Can anyone seriously propose that the lineage of Australian aborigines could go back that far?" wonders paleoanthropologist Chris Stringer of the Natural History Museum in London, a leading advocate of the out-of-Africa theory.

The multiregionalists counter that they've never argued for complete isolation--that there's always been some flow of genes between populations, enough interbreeding to ensure that clearly beneficial sapiens characteristics would quickly be conferred on peoples throughout the Old World. "Just as genes flow now from Johannesburg to Beijing and from Melbourne to Paris, they have been flowing that way ever since humanity evolved," says Alan Thorne of the Australian National University in Canberra, an outspoken multiregionalist.

Stanford archeologist Richard Klein, another out-of-Africa supporter, believes the evidence actually does point to just such a long, deep isolation of Asian populations from African ones. The fossil record, he says, shows that while archaic forms of Homo sapiens were developing in Africa, erectus was remaining much the same in Asia. In fact, if some erectus fossils from a site called Ngandong in Java turn out to be as young as 100,000 years, as some researchers believe, then erectus was still alive on Java at the same time that fully modern human beings were living in Africa and the Middle East. Even more important, Klein says, is the cultural evidence. That Acheulean tools never reached East Asia, even after their invention in Africa, could mean the inventors never reached East Asia either. "You could argue that the new dates show that until very recently there was a long biological and cultural division between Asia on one hand, and Africa and Europe on the other," says Klein. In other words, there must have been two separate lineages of erectus, and since there aren't two separate lineages of modern humans, one of those must have gone extinct: presumably the Asian lineage, hastened into oblivion by the arrival of the more culturally adept, tool-laden Homo sapiens.

Naturally this argument is anathema to the multiregionalists. But this tenacious debate is unlikely to be resolved without basketfuls of new fossils, new ways of interpreting old ones--and new dates. In Berkeley, Curtis and Swisher are already busy applying the argon-argon method to the Ngandong fossils, which could represent some of the last surviving Homo erectus populations on Earth. They also hope to work their radiometric magic on a key erectus skull from Olduvai Gorge. In the meantime, at least one thing has become clear: Homo erectus, for so long the humdrum hominid, is just as fascinating, contentious, and elusive a character as any other in the human evolutionary story.
Comment on this article