Tim White is standing with a group of restless men atop a ridge in the Afar desert of Ethiopia. A few of them are pacing back and forth, straining to see if they can spot fragments of beige bone in the reddish-brown rubble below, as eager to start their search as children at an Easter egg hunt. At the bottom of the hill is a 25-foot-long cairn of black rocks erected in the style of an Afar grave, so large it looks like a monument to a fallen hero. And in a way it is. White and his colleagues assembled it to mark the place where they first found traces, in 1994, of “Ardi,” a female who lived 4.4 million years ago. Her skeleton has been described as one of the most important discoveries of the past century, and she is changing basic ideas about how our earliest ancestors looked and moved.

More than 14 years later, White, a wiry 59-year-old paleoanthropologist from the University of California at Berkeley, is here again, on an annual pilgrimage to see if seasonal rains have exposed any new bits of Ardi’s bones or teeth. He often fires up the fossil hunters who work with him by chanting, “Hominid, hominid, hominid! Go! Go! Go!” But he can’t let them go yet. Only a week earlier, an Alisera tribesman had threatened to kill White and two of his Ethiopian colleagues if they returned to these fossil beds near the remote village of Aramis, home of a clan of Alisera nomads. The threat is probably just a bluff, but White doesn’t mess with the Alisera, who are renowned for being territorial and settling disputes with AK-47s. As a precaution, the scientists travel with six Afar regional police officers armed with their own AK-47s.

Arranging this meeting with tribal leaders to negotiate access to the fossil beds has already cost the researchers two precious days out of their five-week field season. “The best- laid plans change every day,” says White, who has also had to deal with poisonous snakes, scorpions, malarial mosquitoes, lions, hyenas, flash floods, dust tornadoes, warring tribesmen and contaminated food and water. “Nothing in the field comes easy.”

As we wait for the Alisera to arrive, White explains that the team returns to this hostile spot year after year because it’s the only place in the world to yield fossils that span such a long stretch of human evolution, some six million years. In addition to Ardi, a possible direct ancestor, it is possible here to find hominid fossils from as recently as 160,000 years ago—an early Homo sapiens like us—all the way back to Ardipithecus kadabba, one of the earliest known hominids, who lived almost six million years ago. At last count, the Middle Awash project, which takes its name from this patch of the Afar desert and includes 70 scientists from 18 nations, has found 300 specimens from seven different hominid species that lived here one after the other.

Ardi, short for Ardipithecus ramidus, is now the region’s best-known fossil, having made news worldwide this past fall when White and others published a series of papers detailing her skeleton and ancient environment. She is not the oldest member of the extended human family, but she is by far the most complete of the early hominids; most of her skull and teeth as well as extremely rare bones of her pelvis, hands, arms, legs and feet have so far been found.

With sunlight beginning to bleach out the gray-and-beige terrain, we see a cloud of dust on the horizon. Soon two new Toyota Land Cruisers pull up on the promontory, and a half-dozen Alisera men jump out wearing Kufi caps and cotton sarongs, a few cinched up with belts that also hold long, curved daggers. Most of these clan “elders” appear to be younger than 40—few Alisera men seem to survive to old age.

After customary greetings and handshaking, White gets down on his hands and knees with a few fossil hunters to show the tribesmen how the researchers crawl on the ground, shoulder to shoulder, to look for fossils. With Ethiopian paleoanthropologist and project co-leader Berhane Asfaw translating to Amharic and another person translating from Amharic to Afariña, White explains that these stones and bones reveal the ancient history of humankind. The Alisera smile wanly, apparently amused that anyone would want to grovel on the ground for a living. They grant permission to search for fossils—for now. But they add one caveat. Someday, they say, the researchers must teach them how to get history from the ground.

The quest for fossils of human ancestors began in earnest after Charles Darwin proposed in 1871, in his book The Descent of Man and Selection in Relation to Sex, that humans probably arose in Africa. He didn’t base his claim on hard evidence; the only hominid fossils then known were Neanderthals, who had lived in Europe less than 100,000 years ago. Darwin suggested that our “early progenitors” lived on the African continent because its tropical climate was hospitable to apes, and because anatomical studies of modern primates had convinced him that humans were more “allied” with African apes (chimpanzees and gorillas) than Asian apes (orangutans and gibbons). Others disagreed, arguing that Asian apes were closer to modern humans.

As it happened, the first truly ancient remains of a hominid—a fossilized skullcap and teeth more than half a million years old—were found in Asia, on the island of Java, in 1891. “Java man,” as the creature was called, was later classified as a member of Homo erectus, a species that arose 1.8 million years ago and may have been one of our direct ancestors.

So began a century of discovery notable for spectacular finds, in which the timeline of human prehistory began to take shape and the debate continued over whether Asia or Africa was the human birthplace.

In 1924, the Australian anatomist Raymond Dart, looking through a crate of fossils from a limestone quarry in South Africa, discovered a small skull. The first early hominid from Africa, the Taung child, as it was known, was a juvenile member of Australopithecus africanus, a species that lived one million to two million years ago, though at the time skeptical scientists said the chimpanzee-size braincase was too small for a hominid.

In 1959, archaeologist Louis Leakey and his wife Mary, working in Olduvai Gorge in Tanzania, discovered a bit of hominid jawbone that would later become known as Paranthropus boisei. The 1.75-million-year-old fossil was the first of many hominids the Leakeys, their son Richard and their associates would find in East Africa, strengthening the case that hominids indeed originated in Africa. Their work inspired American and European researchers to sweep through the Great Rift Valley, a geologic fault that runs through Kenya, Tanzania and Ethiopia and exposes rock layers that are millions of years old.

In 1974, paleoanthropologists Donald Johanson and Tom Gray, digging in Hadar, Ethiopia, found the partial skeleton of the earliest known hominid at the time—a female they called Lucy, after the Beatles’ song “Lucy in the Sky with Diamonds,” which was playing in camp as they celebrated. At 3.2 million years old, Lucy was remarkably primitive, with a brain and body about the size of a chimpanzee’s. But her ankle, knee and pelvis showed that she walked upright like us.

This meant Lucy was a hominid—only humans and our close relatives in the human family habitually walk upright on the ground. A member of the species Australopithecus afarensis, which lived from 3.9 million to 2.9 million years ago, Lucy helped answer some key questions. She confirmed that upright walking evolved long before hominids began using stone tools—about 2.6 million years ago—and before their brains began to expand dramatically. But her upright posture and gait raised new questions. How long had it taken to evolve the anatomy to balance on two feet? What prompted some ancient ape to stand up and begin walking down the path toward humanness? And what kind of ape was it?

Lucy, of course, couldn’t answer those questions. But what came before her? For 20 years after her discovery, it was as if the earliest chapter of the human story were missing.

One of the first teams to search for lucy’s ancestor was the Middle Awash project, which formed in 1981 when White and Asfaw joined Berkeley archaeologist J. Desmond Clark to search for fossils and stone tools in Ethiopia. They got off to a promising start—finding 3.9-million-year-old fragments of a skull and a slightly younger thighbone—but they were unable to return to the Middle Awash until 1990, because Ethiopian officials imposed a moratorium on searching for fossils while they rewrote their antiquities laws. Finally, in 1992, White’s graduate student, Gen Suwa, saw a glint in the desert near Aramis. It was the root of a tooth, a molar, and its size and shape indicated that it belonged to a hominid. Suwa and other members of the Middle Awash project soon collected other fossils, including a child’s lower jaw with a milk molar still attached. State-of-the-art dating methods indicated that they were 4.4 million years old.

The team proposed in the journal Nature in 1994 that the fossils—now known as Ardipithecus ramidus—represented the “long-sought potential root species for the Hominidae,” meaning that the fossils belonged to a new species of hominid that could have given rise to all later hominids. The idea that it was a member of the human family was based primarily on its teeth—in particular, the absence of large, dagger-like canines sharpened by the lower teeth. Living and extinct apes have such teeth, while hominids don’t. But the gold standard for being a hominid was upright walking. So was A. ramidus really a hominid or an extinct ape?

White joked at the time that he would be delighted with more fossils—in particular, a skull and thighbone. It was as if he had placed an order. Within two months, another graduate student of White’s, Ethiopian paleoanthropologist Yohannes Haile-Selassie, spotted two pieces of a bone from the palm of a hand—their first sign of Ardi. The team members eventually found 125 pieces of Ardi’s skeleton. She had been a muscular female who stood almost four feet tall but could have weighed as much as 110 pounds, with a body and brain roughly the same size as a chimpanzee’s. As they got a good look at Ardi’s body plan, they soon realized that they were looking at an entirely new type of hominid.

It was the find of a lifetime. But they were daunted by Ardi’s condition. Her bones were so brittle that they crumbled when touched. White called them “road kill.”

The researchers spent three field seasons digging out entire blocks of sedimentary rock surrounding the fossils, encasing the blocks in plaster and driving them to the National Museum of Ethiopia in Addis Ababa. In the museum lab, White painstakingly injected glue from syringes into each fragment and then used dental tools and brushes, often under a microscope, to remove the silty clay from the glue-hardened fossils. Meanwhile, Suwa, today a paleoanthropologist at the University of Tokyo, analyzed key fossils with modified CT scanners to see what was inside them and used computer imaging to digitally restore the crushed skull. Finally, he and anatomist C. Owen Lovejoy worked from the fossils and the computer images to make physical models of the skull and pelvis.

It’s a measure of the particularity, complexity and thoroughness of the researchers’ efforts to understand Ardi in depth that they took 15 years to publish their detailed findings, which appeared this past October in a series of 11 papers in the journal Science. In short, they wrote that Ardi and fossils from 35 other members of her species, all found in the Middle Awash, represented a new type of early hominid that wasn’t much like a chimpanzee, gorilla or a human. “We have seen the ancestor and it’s not a chimpanzee,” says White.

This came as a surprise to researchers who had proposed that the earliest hominids would look and act a lot like chimpanzees. They are our closest living relatives, sharing 96 percent of our DNA, and they are capable of tool use and complex social behavior. But Ardi’s discoverers proposed that chimpanzees have changed so dramatically as they have evolved over the past six million years or so, that today’s chimpanzees make poor models for the last common ancestor we shared.

In his lab at Kent State University, Lovejoy recently demonstrated why Ardi is so unusual. He gently lined up four bones from Ardi’s hand on his lab bench, and he showed how they fit together in a way that allowed Ardi’s hand to bend far backward at the wrist. By comparison, a chimpanzee’s wrist is stiff, which allows the animal to put its weight on its knuckles as it moves on the ground—knuckle walking. “If you wanted to evolve Ardi’s hand, you couldn’t do it from this,” he said, waving a set of bones from a chimpanzee hand in the air. If Lovejoy is right, this means Ardi—and our upright-walking ancestors—never went through a knuckle-walking stage after they came down from the trees to live on the ground, as some experts have long believed.

As evidence that Ardi walked upright on the ground, Lovejoy pointed to a cast of her upper pelvic blades, which are shorter and broader than an ape’s. They would have let her balance on one leg at a time while walking upright. “This is a monstrous change—this thing has been a biped for a very long time,” Lovejoy said.

But Ardi didn’t walk like us or, for that matter, like Lucy either. Ardi’s lower pelvis, like a chimpanzee’s, had powerful hip and thigh muscles that would have made it difficult to run as fast or as far as modern humans can without injuring her hamstrings. And she had an opposable big toe, so her foot was able to grasp branches, suggesting she still spent a lot of time in the trees—to escape predators, pick fruit or even sleep, presumably in nests made of branches and leaves. This unexpected combination of traits was a “shocker,” says Lovejoy.

He and his colleagues have proposed that Ardi represents an early stage of human evolution when an ancient ape body plan was being remodeled to live in two worlds—in the trees and on the ground, where hominids increasingly foraged for plants, eggs and small critters.

The Ardi research also challenged the long-held views that hominids evolved in a grassy savanna, says Middle Awash project geologist Giday WoldeGabriel of Los Alamos National Laboratory. The Ardi researchers’ thorough canvassing—“You crawl on your hands and knees, collecting every piece of bone, every piece of wood, every seed, every snail, every scrap,” White says—indicates that Ardi lived in woodland with a closed canopy, so little light reached grass and plants on the forest floor. Analyzing thousands of specimens of fossilized plants and animals, as well as hundreds of samples of chemicals in sediments and tooth enamel, the researchers found evidence of such forest species as hackberry, fig and palm trees in her environment. Ardi lived alongside monkeys, kudu antelopes and peafowl—animals that prefer woodlands, not open grasslands.

Ardi is also providing insights into ancient hominid behavior. Moving from the trees to the ground meant that hominids became easier prey. Those that were better at cooperating could live in larger social groups and were less likely to become a big cat’s next meal. At the same time, A. ramidus males were not much larger than females and they had evolved small, unsharpened canine teeth. That’s similar to modern humans, who are largely cooperative, and in contrast to modern chimpanzees, whose males use their size to dominate females and brandish their dagger-like canines to intimidate other males.

As hominids began increasingly to work together, Lovejoy says, they also adopted other previously unseen behaviors—to regularly carry food in their hands, which allowed them to provision mates or their young more effectively. This behavior, in turn, may have allowed males to form tighter bonds with female mates and to invest in the upbringing of their offspring in a way not seen in African apes. All this reinforced the shift to life on the ground, upright walking and social cooperation, says Lovejoy.

Not everyone is convinced that Ardi walked upright, in part because the critical evidence comes from her pelvis, which was crushed. While most researchers agree that she is a hominid, based on features in her teeth and skull, they say she could be a type of hominid that was a distant cousin of our direct ancestor—a newfound offshoot on the human family tree. “I think it’s solid” that Ardi is a hominid, if you define hominids by their skull and teeth, says Rick Potts, a paleoanthropologist at the Smithsonian’s National Museum of Natural History. But, like many others who have not seen the fossils, he has yet to be convinced that the crushed but reconstructed pelvis proves upright walking, which could mean that Ardi might have been an extinct ape that was “experimenting” with some degree of upright walking. “The period between four million to seven million years is when we know the least,” says Potts. “Understanding what is a great ape and what is a hominid is tough.”

As researchers sort out where Ardi sits in the human family tree, they agree that she is advancing fundamental questions about human evolution: How can we identify the earliest members of the human family? How do we recognize the first stages of upright walking? What did our common ancestor with chimpanzees look like? “We didn’t have much at all before,” says Bill Kimbel, an Arizona State University paleoanthropologist. “Ardipithecus gives us a prism to look through to test alternatives.”

After Ardi’s discovery, researchers naturally began to wonder what came before her. They didn’t have long to wait.

Starting in 1997, Haile-Selassie, now at the Cleveland Museum of Natural History, found fossils between 5.2 million and 5.8 million years old in the Middle Awash. A toe bone suggested its owner had walked upright. The bones looked so much like a primitive version of A. ramidus he proposed these fossils belonged to her direct ancestor—a new species he eventually named Ardipithecus kadabba.

In 2000, Martin Pickford of the College of France and Brigitte Senut of the National Museum of Natural History in Paris announced their team had found an even older hominid—13 fossils representing a species that lived six million years ago in the Tugen Hills of Kenya. Two of the fossils were thighbones, including one that provided the oldest direct evidence of upright walking in a hominid. They named this creature Orrorin tugenensis, drawing on a Tugen legend of the “original man” who settled the Tugen Hills. Informally, in honor of its year of discovery, they called it Millennium man.

Hot on the heels of that discovery came the most surprising one of all—a skull from Chad, about 1,500 miles west of the Great Rift Valley of eastern Africa where many of the most ancient hominids have been found. A Chadian student named Ahounta Djimdoumalbaye picked up a ball of rock on the floor of the Djurab Desert, where windstorms blow sand dunes like waves on a sea and expose fossils buried for millions of years. When Djimdoumalbaye turned over the stone, he stared into the vacant eye sockets of an ape-like face—the skull of a primate that lived six million to seven million years ago on the shores of an ancient lake. It had traits that suggested it was a hominid—a small lower face and canines and a skull that seemed to sit atop its spine, as in upright walkers. Paleontologist Michel Brunet, then of the University of Poitiers in France, introduced it as the oldest known hominid, Sahelanthropus tchadensis. (Its nickname is Toumaï, which means “hope of life” in the Goran language.) But proving that a skull walked upright is difficult, and questions linger about whether Sahelanthropus is a bona fide hominid or not.

Taken together, fossils discovered over the past 15 years have provided snapshots of several different creatures that were alive in Africa at the critical time when the earliest members of the human family were emerging. When these snapshots are added to the human family album, they double the time researchers can see back into our past—from Lucy at 3.2 million years to Toumaï at almost 7 million years.

One of the most sought-after fossils of that distant era was Lucy’s direct ancestor. In 1994, 20 years after Lucy’s skeleton was discovered, a team in Kenya led by Meave Leakey (the wife of Richard Leakey) found teeth and parts of a jaw as well as two pieces of shinbone that showed the creature walked upright. The fossils, named Australopithecus anamensis, were 4.1 million years old.

“This has been a fascinating 40 years to be in paleoanthropology,” says Johanson, “one of the great times to be in this field.” But, he adds, “there’s still enormous confusion” about the murky time before 4 million years ago.

One thing that is clear is that these early fossils belong in a class by themselves. These species did not look or act like other known apes or like Lucy and other members of Australopithecus. They were large-bodied ground dwellers that stood up and walked on two legs. But if you watched them move, you would not mistake them for Lucy’s species. They clung to life in the trees, but were poised to venture into more open country. In many ways, these early species resemble one another more than any fossils ever found before, as if there was a new developmental or evolutionary stage that our ancestors passed through before the transition was complete from ape to hominid. Indeed, when the skulls of Toumaï and Ardi are compared, the resemblance is “striking,” says paleoanthropologist Christoph Zollikofer of the University of Zurich in Switzerland. The fossils are too far apart in time to be members of the same species, but their skulls are more like each other than they are like Lucy’s species, perhaps signaling similar adaptations in diet or reproductive and social behavior.

The only way to find out how all these species are related to one another and to us is to find more bones. In particular, researchers need to find more overlapping parts of very early fossils so they can be compared directly—such as an upper end of a thighbone for both Ardi and Toumaï to compare with the upper thighbone of O. tugenensis.

At Aramis, as soon as the clan leaders gave the Middle Awash team their blessing, White began dispatching team members like an air traffic controller, directing them to fan out over the slope near Ardi’s grave. The sun was high in the sky, though, making it hard to distinguish beige bone among the bleached out sediments. This time, the team found no new hominid fossils.

But one morning later that week, the team members drove up a dry riverbed to a site on the western margin of the Middle Awash. Only a few moments after hiking into the fossil beds, a Turkish postdoctoral researcher, Cesur Pehlevan, planted a yellow flag among the cobbles of the remote gully. “Tim!” he shouted. “Hominid?” White walked over and silently examined the molar, turning it over in his hand. White has the ability to look at a tooth or bone fragment and recognize almost immediately whether it belongs to a hominid. After a moment, he pronounced his verdict: “very good, Cesur. It’s virtually unworn.” The molar belonged to a young adult A. kadabba, the species whose fossils began to be found here in 1997. Now the researchers had one more piece to help fill in the portrait of this 5.8-million-year-old species.

“There’s your discovery moment,” said White. He reflected on the fossils they’ve bagged in this remote desert. “This year, we’ve got A. kadabba, A. anamensis, A. garhi, H. erectus, H. sapiens.” That’s five different kinds of hominids, most of which were unknown when White first started searching for fossils here in 1981. “The Middle Awash is a unique area,” he said. “It is the only place on the planet Earth where you can look at the full scope of human evolution.”

Ann Gibbons is a correspondent for Science and the author of The First Human: The Race to Discover Our Earliest Ancestors.