A boomerang carved from a mammoth tusk is one of the oldest in the world, and it may be even older than archaeologists originally thought, according to a recent round of radiocarbon dating.
Archaeologists unearthed the mammoth-tusk boomerang in Poland’s Oblazowa Cave in the 1990s, and they originally dated it to around 18,000 years old, which made it one of the world’s oldest intact boomerangs. But according to recent analysis by University of Bologna researcher Sahra Talamo and her colleagues, the boomerang may have been made around 40,000 years ago. If they’re right, it offers tantalizing clues about how people lived on the harsh tundra of what’s now Poland during the last Ice Age.
A boomerang carved from mammoth tusk
The mammoth-tusk boomerang is about 72 centimeters long, gently curved, and shaped so that one end is slightly more rounded than the other. It still bears scratches and scuffs from the mammoth’s life, along with fine, parallel grooves that mark where some ancient craftsperson shaped and smoothed the boomerang. On the rounded end, a series of diagonal marks would have made the weapon easier to grip. It’s smoothed and worn from frequent handling: the last traces of the life of some Paleolithic hunter.
Based on experiments with a replica, the Polish mammoth boomerang flies smoothly but doesn’t return, similar to certain types of Aboriginal Australian boomerangs. In fact, it looks a lot like a style used by Aboriginal people from Queensland, Australia, but that’s a case of people in different times and places coming up with very similar designs to fit similar needs.
But critically, according to Talamo and her colleagues, the boomerang is about 40,000 years old.
That’s a huge leap from the original radiocarbon date, made in 1996, which was based on a sample of material from the boomerang itself and estimated an age of 18,000 years. But Talamo and her colleagues claim that original date didn’t line up well with the ages of other nearby artifacts from the same layer of the cave floor. That made them suspect that the boomerang sample may have gotten contaminated by modern carbon somewhere along the way, making it look younger. To test the idea, the archaeologists radiocarbon dated samples from 13 animal bones—plus one from a human thumb—unearthed from the same layer of cave floor sediment as the boomerang.
(Left) GPS tracking and modeling of ocean currents toward the end of the experimental voyage. (Right) The team on the water around the time of the left image. Credit: Kaifu et al., 2025/CC-By-ND
At the 30-hour mark, the captain ordered the entire crew to rest, letting the dugout drift freely for a while, which fortunately brought them closer to Yonaguni Island. At hour 40, the island’s silhouette was visible, and over the next five hours, the crew was able to navigate the strong tidal flow along the coast until they reached their landing site: Nama Beach. So the experimental voyage was a success, augmented by the numerical simulations to demonstrate that the boat could make similar voyages from different departure points across both modern and late-Pleistocene oceans.
Granted, it was not possible to recreate Paleolithic conditions perfectly on a modern ocean. The crew first spotted the island because of its artificial lights, although by that time, they were on track navigationally. They were also accompanied by escort ships to ensure the crew’s safety, supplying fresh water twice during the voyage. But the escort ships did not aid with navigation or the dugout captain’s decision-making, and the authors believe that any effects were likely minimal. The biggest difference was the paddlers’ basic modern knowledge of local geography, which helped them develop a navigation plan—an unavoidable anachronism, although the crew did not rely on compasses, GPS, or watches during the voyage.
“Scientists try to reconstruct the processes of past human migrations, but it is often difficult to examine how challenging they really were,” said Kaifu. “One important message from the whole project was that our Paleolithic ancestors were real challengers. Like us today, they had to undertake strategic challenges to advance. For example, the ancient Polynesian people had no maps, but they could travel almost the entire Pacific. There are a variety of signs on the ocean to know the right direction, such as visible land masses, heavenly bodies, swells and winds. We learned parts of such techniques ourselves along the way.”
DOI: “Traversing the Kuroshio: Paleolithic migration across one of the world’s strongest ocean currents,” Science Advances, 2025. 10.1126/sciadv.adv5508 (About DOIs).
DOI: “Palaeolithic seafaring in East Asia: an experimental test of the dugout canoe hypothesis,” Science Advances, 2025. 10.1126/sciadv.adv5507 (About DOIs).
Enlarge/ Testing replica Stone Age tools with a bit of wood-scraping.
A. Iwase et al., 2024/Tokyo Metropolitan University
When Japanese scientists wanted to learn more about how ground stone tools dating back to the Early Upper Paleolithic might have been used, they decided to build their own replicas of adzes, axes, and chisels and used those tools to perform tasks that might have been typical for that era. The resulting fractures and wear enabled them to develop new criteria for identifying the likely functions of ancient tools, according to a recent paper published in the Journal of Archaeological Science. If these kinds of traces were indeed found on genuine Stone Age tools, it would be evidence that humans had been working with wood and honing techniques significantly earlier than previously believed.
The development of tools and techniques for woodworking purposes started out simple, with the manufacture of cruder tools like the spears and throwing sticks common in the early Stone Age. Later artifacts dating back to Mesolithic and Neolithic time periods were more sophisticated, as people learned how to use polished stone tools to make canoes, bows, wells, and to build houses. Researchers typically date the emergence of those stone tools to about 10,000 years ago. However, archaeologists have found lots of stone artifacts with ground edges dating as far back as 60,000 to 30,000 years ago. But it’s unclear how those tools might have been used.
So Akira Iwase of Tokyo Metropolitan University and co-authors made their own replicas of adzes and axes out of three raw materials common to the region between 38,000 and 30,000 years ago: semi-nephrite rocks, hornfels rocks, and tuff rocks. They used a stone hammer and anvil to create various long oval shapes and polished the edges with either a coarse-grained sandstone or a medium-grained tuff. There were three types of replica tools: adze-types, with the working edge oriented perpendicular to the long axis of a bent handle; axe-types, with a working edge parallel to the bent handle’s long axis; and chisel-types, in which a stone tool was placed at the end of a straight handle.
Enlarge/ Testing various replicas of Stone Age tools for different uses: A, tree-felling; B, wood-adzing; C, wood-scraping; D, fresh bone-adzing; E, dry hide-scraping; F, disarticulation of a joint.
A. Iwase et al., 2024/Tokyo Metropolitan University
Then it was time to test the replica tools via ten different usage experiments. For instance, the authors used axe-type tools to fell Japanese cedar and maple trees in north central Honshu, as well as a forest near Tokyo Metropolitan University. Axe-type and adze-type tools were used to make a dugout canoe and wooden spears, while adze-type tools and chisel-type tools were used to scrape off the bark of fig and pine. They scraped flesh and grease from fresh and dry hides of deer and boar using adze-type and chisel-type tools. Finally, they used adze-type tools to disarticulate the femur and tibia joints of deer hindlimbs.
The team also conducted several experiments in which the tools were not used to identify accidental fractures not related to any tool-use function. For instance, flakes and blades can break in half during flint knapping; transporting tools in, say, small leather bags can cause microscopic flaking; and trampling on tools left on the ground can also modify the edges. All these scenarios were tested. All the tools used in both use and non-use experiments,ents were then examined for both macroscopic and microscopic traces of fracture or wear.
Enlarge/ Traces left by tree-felling experiments on replica stone age tools. Characteristic macroscopic (top) and microscopic (bottom) traces might be used to determine how stone edges were used.
Tokyo Metropolitan University
The results: they were able to identify nine different types of macroscopic fractures, several of which were only seen when making percussive motions, particularly in the case of felling trees. There were also telltale microscopic traces resulting from friction between the wood and stone edge. Cutting away at antlers and bones caused a lot of damage to the edges of adze-like tools, creating long and/or wide bending fractures. The tools used for limb disarticulation caused fairly large bending fractures and smaller flaking scars, while only nine out of 21 of the scraping tools showed macroscopic signs of wear, despite hundreds of repeated strokes.
The authors concluded that examining macroscopic fracture patterns alone are insufficient to determine whether a given stone tool had been used percussively. Nor is any resulting micropolish from abrasion an unambiguous indicator on its own, since scraping motions produce a similar micropolish. Combining the two, however, did yield more reliable conclusions about which tools had been used percussively to fell trees, compared to other uses, such as disarticulation of bones.
