An elephant humerus that has been knapped into a tool. (Credit: CSIC)
In a nutshell
- Archaeologists discovered 27 bone tools dating back 1.5 million years at Olduvai Gorge in Tanzania, pushing back the timeline of systematic bone tool production by over a million years.
- Early humans preferentially selected elephant bones for toolmaking despite their scarcity, demonstrating sophisticated knowledge of material properties and advanced planning skills.
- The transfer of stone tool knapping techniques to bone materials reveals cognitive abilities in early human ancestors that were previously thought to emerge much later in human evolution.
LONDON — The past is often written in stone, but in this case, it is etched into bone. A team of researchers has uncovered evidence that early humans in Tanzania were crafting tools from animal bones 1.5 million years ago, overturning the long-standing assumption that bone technology emerged only much later in human history.
The findings, published in Nature, show that our ancestors during the transition between the Oldowan and early Acheulean periods were much more technologically capable than we ever imagined.
Rewriting the History of Ancient Craftsmen
For decades, the conventional wisdom held that while early humans began crafting stone tools well before the emergence of the genus Homo, complex bone technology was the exclusive domain of much later hominins. The earliest systematic production of bone tools was previously thought to be limited to European Acheulean sites from merely 400,000-250,000 years ago.
The site in Tanzania’s T69 Complex yielded 27 bone tools all dated to 1.5 million years ago and found in a single layer, proving they weren’t random but part of a consistent manufacturing tradition. These ancient toolmakers were creating objects that supposedly wouldn’t exist for another million years.
“By producing technologically and morphologically standardized bone tools, early Acheulean toolmakers unraveled technological repertoires that were previously thought to have appeared routinely more than 1 million years later,” the researchers explain in their paper. This suggests that the innovation and cognitive abilities of early humans have been significantly underestimated.
The ancient craftspeople weren’t just grabbing any bones either. They carefully selected limb bones from large mammals, with a strong preference for elephant bones despite elephants being rare at the site. Of the 16 identifiable bone tool specimens, eight came from elephants and six from hippos, with just two from bovids.
Bovids made up over 41% of the animal bones at the site, yet only about 1% were elephant remains. Despite this, half the tools were made from elephant bones. This selective behavior shows these early humans knew which bones would work best for tools and had the skills to shape them effectively.
Evidence of Advanced Thinking and Design
Natural bone breaks typically show around 2.1 flake scars that appear as isolated marks. In contrast, the bone tools found at the T69 Complex average 12.9 flake scars per tool, arranged in connected patterns.
“The tools show evidence that their creators carefully worked the bones, chipping off flakes to create useful shapes. We were excited to find these bone tools from such an early timeframe. It means that human ancestors were capable of transferring skills from stone to bone, a level of complex cognition that we haven’t seen elsewhere for another million years,” explained co-author Dr. Renata F. Peters of UCL Archaeology, in a statement.
The researchers uncovered a clear two-step manufacturing process: first creating larger flake removals to shape the tool, then fine-tuning the edges. This level of planning and craftsmanship was previously thought beyond the capabilities of hominins from this time period.
Six of the tools show a consistent design pattern: one crescent end, one pointed end, and a large notch in the middle, likely to help with gripping while keeping the tool’s weight and structural integrity.
The bone tools appear to have been designed for heavy-duty tasks. Compared to stone tools found at the same site, the notched bone implements are larger, more elongated, and more intensively shaped than stone tools, though their current weight falls within the range of variation recorded for large cutting tools (LCTs) found at the site.
Distal fracture patterns on some specimens suggest use in percussive and compressive activities, possibly related to butchery. The context of the site—with abundant fish, crocodile, and hippopotamus remains consistent with a water-adjacent environment—suggests hominins were attracted to the area by the availability of hippopotamus carcasses.
“This discovery leads us to assume that early humans significantly expanded their technological options, which until then were limited to the production of stone tools and now allowed new raw materials to be incorporated into the repertoire of potential artifacts,” said lead author Dr. Ignacio de la Torre of the Spanish National Research Council (CSIC) Instituto de Historia, in a statement. “At the same time, this expansion of technological potential indicates advances in the cognitive abilities and mental structures of these hominins, who knew how to incorporate technical innovations by adapting their knowledge of stone work to the manipulation of bone remains.”
Rethinking Early Hominin Technology
The tools come from a pivotal moment in human technological development. The earliest stone tools belong to the “Oldowan” age (2.7-1.5 million years ago), which involved simply knocking flakes off stone cores using a hammerstone.
As the ancient human ancestors were progressing into the “Acheulean” age which began as far back as about 1.7 million years ago, they developed more intricate handaxes that were carefully shaped by knapping—allowing the production of tools through more standardized means. The bone tools show that these more advanced techniques were carried over and adopted for use on bones as well, something previously unseen in the fossil record for another million years.
Who made these tools remains a mystery. No hominin remains were found with the bone artifacts, though we know both Homo erectus and Paranthropus boisei lived in the region at that time.
Because these tools were such an unexpected discovery, the researchers hope that their findings will prompt archaeologists to re-examine bone discoveries around the world in case other evidence of bone tools has been missed.
This research was supported by the European Research Council.
Paper Summary
Methodology
The research team excavated the T69 Complex site at Olduvai Gorge across seven trenches between 2015 and 2022, uncovering over 10,900 stone tools, 9,400 identifiable animal fossils, and 13,000 unidentified bone fragments. Researchers identified 27 bones that appeared deliberately shaped rather than naturally broken.
To verify these were actually tools, scientists compared them with experimentally broken elephant and horse bones using 3D scanning and digital microscopy. The site was dated to 1.5 million years ago using direct cosmogenic nuclide isochron burial dating, consistent with its position in the geological layers of Olduvai Gorge.
Results
Analysis revealed that the bone tools displayed an average of 12.9 flake scars per specimen, arranged in intentional patterns, compared to only 2.1 random scars typically found on naturally broken bones. The tools exhibited a systematic, two-step manufacturing process: first creating invasive flake removals to shape the tool, then regularizing the edges through trimming.
Despite elephants constituting only 1.1% of the animal remains at the site, elephant bones made up 50% of the tools found, suggesting deliberate selection of materials. Six tools showed a remarkably consistent pattern—one crescent end, one pointed end, and a large notch in the middle—indicating standardized design rather than opportunistic shaping.
Limitations
Without hominin remains at the site, researchers cannot determine exactly which species made these tools, though both Homo erectus and Paranthropus boisei lived in the region during this time period. The exact purpose of these tools also remains somewhat speculative, though their shape and wear patterns suggest they were likely used for processing animal carcasses.
Another significant limitation involves the gap in the archaeological record. If early humans were making sophisticated bone tools 1.5 million years ago, similar examples from the following million years are mysteriously absent. This could be due to preservation issues or indicate this technology was developed but later abandoned until rediscovery.
Discussion and Takeaways
These findings fundamentally change our understanding of early human technological development. The ability to transfer knapping skills from stone to bone demonstrates these early hominins possessed greater cognitive flexibility and problem-solving abilities than previously recognized, suggesting more advanced mental capabilities.
The researchers propose that large bone tools may have fulfilled functions later achieved by stone handaxes at butchery sites. As stone tool technology advanced, bone tools may have become less common, explaining their absence in later archaeological records until they reappear 400,000-250,000 years ago. The discovery may prompt archaeologists to reexamine existing bone collections for overlooked evidence of tool manufacture.
Funding and Disclosures
This research was funded by the European Research Council (ERC) through their Advanced grants program. Additional support came from fieldwork grants by Fundación Palarq and the Spanish Ministry of Culture.
The excavation work was conducted with permits from the Tanzanian Commission for Science and Technology, the Department of Antiquities under Tanzania’s Ministry of Natural Resources and Tourism, and the Ngorongoro Conservation Area Authority.
Publication Details
The study, titled “Systematic bone tool production at 1.5 million years ago,” was published in Nature on March 5, 2025. The research team includes Dr. Ignacio de la Torre (CSIC-Spanish National Research Council), Dr. Renata F. Peters (UCL), and colleagues from institutions including Université de Bordeaux, CENIEH, Universitat Autonoma de Barcelona, Indiana University, and others.
The DOI for this paper is https://doi.org/10.1038/s41586-025-08652-5.
