On July 9, 2011, Sonia Harmand took a left turn instead of a right among the dry riverbeds that substitute for roads on the western shore of Kenya’s Lake Turkana, and promptly got lost.
“I accept responsibility for that,” says Jason Lewis, a paleoanthropologist at Rutgers University, who—with the help of a geological map and GPS—was supposedly playing the role of navigator. He and Harmand, an archaeologist at Stony Brook University, were spearheading a team hunting for traces of human ancestral behavior in sediments millions of years old.
The misdirection led to an enormous payoff: a discovery that pushes the date of the earliest known stone tools back by some 700,000 years, according to a report published Wednesday in the journal Nature. (Go on a virtual tour of the Turkana site and fossil collection.)
The serendipity began when Harmand steered their 1989 SUV straight into a dead-end riverbed. Rather than simply turn around, the team hopped out to survey the area—a shadeless, Badlands moonscape where the wind felt like a blast from a blow dryer. After about an hour, Harmand’s walkie-talkie crackled. Sammy Lokorodi, a sharp-eyed native of the Turkana area, had found some curious rocks on the next hill over.
Perched on the otherwise fine-grained surface, the large stone hunks bore fracture marks that to a trained archaeologist’s eye appeared to be caused by purposeful manipulation, rather than natural forces. While most of the fractured rocks were on the surface, a trio of partially buried ones suggested that the surface finds were eroding out of the sediments—and that more artifacts might be hidden just below.
At the time, the oldest known stone tools, from a site called Gona in Ethiopia, were dated to 2.6 million years ago. Those artifacts belong to a technology known as the Oldowan, named after Olduvai Gorge in Tanzania, where Louis Leakey first found such tools in the 1930s. But Harmand and Lewis, among others, thought the Oldowan technology was too carefully knapped to represent our ancestors’ first crack at tool-making. Some form of older, even more primitive technology must be out there, waiting to be found.
Given that some deposits in the area of that dead-end Turkana river bed were known to be at least 2.7 million years old, there was at least a chance that the wait was over.
It would take more than a year to pin down the age of the sediments bearing the tools. In the meantime, with funding from National Geographic the team returned for a second field season at the site, called Lomekwi 3. They collected nearly 150 artifacts and three dozen animal fossils. But a key discovery occurred not in the field, but in the National Museums in Nairobi, where Harmand discovered that at least one of the edges of a surface stone fit neatly into one of the buried artifacts, like two jigsaw puzzle pieces clicking into place. This confirmed a connection between the finds on the surface and the sediments.
By the end of 2013, the dating of those deposits was finally complete. The sediments, and thus the tools themselves, were 3.3 million years old—almost three-quarters of a million years older than the Oldowan tools from Gona. Because the Lomekwi 3 stones are so crudely fashioned, they could represent the critical transition in our early ancestry between tool use—banging on things with rocks, as chimpanzees do with nuts—and tool making: the deliberate reshaping of a stone to create a sharper edge.
“To me it makes a hell of a lot of sense that this is what predates the Oldowan,” says archaeologist Curtis Marean of Arizona State University. “It looks different from a lot of other technologies that people are used to looking for, which is why I think people didn’t find it. They had the wrong search image in their minds.”
For a long time, tool-making was thought to originate with the bigger-brained genus Homo, specifically with a species called Homo habilis, or “Handy Man,” so named by Leakey because bones of the species were found near Oldowan tools. But the discovery at Lomekwi 3 makes it nearly impossible to implicate a known Homo species in the striking of the first stone flake, because the oldest known Homo fossil is just 2.8 million years old.
"It was textbook knowledge, almost, that Homo habilis was the first handyman," says Zeray Alemseged, a paleoanthropologist at the California Academy of Sciences. "Now, clearly, we had more handy men and handy women out there on the landscape, prior to 3 million years ago."
Just who made the tools remains an open question. One possibility is Australopithecus afarensis—the species best represented by Lucy—which inhabited East Africa at the right time. Particularly important to that argument is an A. afarensis specimen found in Dikika, Ethiopia, just 200 yards from where Alemseged’s team later found animal bones dated to 3.4 million years ago they believe bear cut marks made by stone tools.
Another, even more intriguing candidate for the earliest tool maker is a controversial species called Kenyanthropus platyops. (Not all scientists accept that it represents a separate genus from australopithecines like A. afarensis.) The original specimen of K. platyops, a skull, was found less than a mile from Lomekwi 3. Other fragments attributed to the species have been found even closer to the site.
Neither of these hominins are in the genus Homo. But it wouldn’t be all that surprising for them to have been crafting tools.
“I think [australopithecines] would have had the cognitive capabilities to do it,” Nick Toth, of the Stone Age Institute, previously told National Geographic, “and even though their hands were probably not as dexterous, they probably would have had no problem flaking stone.”
Not all scientists are convinced, however, that the Lomekwi 3 artifacts are as old as Harmand and her colleagues suggest. Tim White, a paleoanthropologist at UC Berkeley who played a key role in the original analysis of A. afarensis, questions whether they considered carefully enough the complex geology of the region when they dated the sediments. White contends the age of the tools on the surface is particularly suspect. “They could have been made at any time, even a few years ago by a passing nomad,” he says.
University of Utah geologist Frank Brown, an authority on the geology of the Turkana Basin, is not similarly troubled.
“I wouldn’t worry about the age of the beds,” Brown says. Some of the big rocks were still tucked into their original sediments, Brown says. "And if they’re artifacts, then yes, they have 3.3 million-year-old tools.”
Other scientists, however, are puzzled by the sheer size of the Lomekwi 3 tools. Some weigh as much as 30 pounds—more than an order of magnitude bigger than a typical Oldowan artifact. While Oldowan tools appear to have been used to cut meat from bones, it’s unlikely the Lomekwi 3 tools would have served the same purpose, says David Braun, an archaeologist at George Washington University. Especially if they’re being wielded by creatures whose hands aren’t that different from ours.
“What the heck something with our manual anatomy is doing with something this enormous is really difficult to get a grip on,” Braun says. “That, I think, is still a mystery.”
Follow Nadia Drake on Twitter.