A quiet revolution in our understanding of early human ingenuity has emerged from the volcanic landscapes of Ethiopia’s Upper Awash region. In the heart of this ancient terrain, long buried beneath layers of ash and time, a subtle but compelling story is unfolding—one that rewrites our assumptions about how our ancestors interacted with the natural world and challenges the boundaries of what we recognize as “tools.” In a new study published in Quaternary International, Dr. Margherita Mussi sheds light on an overlooked facet of early human behavior: the deliberate selection and use of naturally occurring basalt spheres by hominin species over more than a million years.
For decades, archaeologists and anthropologists have been drawn to the enigmatic stone balls scattered across Pleistocene sites in Africa, Europe, and Asia. Sometimes polished, sometimes faceted, these globular lithics have been called many names—“spheroids,” “bolas,” “stone balls,” “boules polyédriques”—each term reflecting both fascination and uncertainty. Were these objects hunting tools, weapons, grinding implements, or percussion devices? Scholars have debated, speculated, and even recreated them in experimental archaeology labs, trying to uncover their purpose. Most studies, however, have focused almost exclusively on intentionally shaped or modified spheres. What if, Mussi asks, we have been overlooking the more subtle evidence hiding in plain sight?
The volcanic spheres found at Melka Kunture, a renowned paleoanthropological site in Ethiopia, were not carved, flaked, or chipped into shape. Instead, they were born of nature—erupted from Earth’s molten core and sculpted over millennia by geological forces into strikingly round forms. These basalt balls, ranging in hardness, weight, and texture, were not the product of a hominin hand. But crucially, as Mussi’s detailed study reveals, they were selected by hominin hands. That distinction makes all the difference.
Across eight major sites—Gombore IB, Atebella II, Garba XII, Gombore II-1, Gombore II-2, Garba IIIE, Gotu III, and Garba I—over 30 basalt spheres were recovered. These lithics were housed at the National Museum of Addis Ababa alongside thousands of associated stone tools and hominin bone fragments. The oldest site, Gombore IB, dated to 1.7 million years ago, revealed a striking archaeological profile: nearly 5,000 stone tools, three basalt spheres, and two Homo cf. ergaster humerus fragments. The youngest sites, Garba I, II, and III, dated to around 600,000 years ago, yielded an additional 22 basalt spheres and more than 7,000 stone tools. This continuity across time—nearly 1.1 million years—signals a persistent behavior, a long-standing tradition of purposeful collection and use.
Crucially, these spheres appear in contexts where natural geological forces could not have deposited them. They are heavy, dense, and out of place in the fine-grained sediments where they were unearthed. Some are composed of soft lapilli basalt—too fragile to survive transport by water without breaking. Others are made of extremely hard volcanic basalt, suggesting that they would not have been randomly rolled into these environments. They had to be carried there. Deliberately. By someone with a reason.
“The metric characteristics, shape, and orientation of pebbles transported by water have been researched and are well understood,” Mussi notes. “Furthermore, note that some of the sites are fine-grained deposits where the relatively heavy rock spheres are at odds with the surrounding environment, and also that the rather soft lapilli ones would have been easily crushed during water transport.”
So why did early hominins bother to collect and carry these volcanic spheres?
Through careful morphological analysis, Mussi uncovered signs that these naturally round objects were more than geological curiosities. They were used. Hard basalt spheres, with their dense composition and resistant surfaces, were likely used in knapping and retouching lithic tools—functions requiring precision and power. Softer spheres, such as those made from lapilli basalt, may have served as grinding or rubbing stones, possibly for preparing vegetables, animal hides, or pigments. They are not tools in the traditional sense—not flaked or shaped—but they exhibit what archaeologists term use wear. They are outils à posteriori—objects that only become recognizable as tools after having been marked by use.
“I am convinced that the hard volcanic ones were used to knap or retouch lithic tools,” Mussi explains, “while the rather soft lapilli ones were used for rubbing vegetables or hides or other stuff.”
This insight reframes our understanding of early human behavior. These basalt spheres represent a form of tool use grounded not in fabrication, but in selection—an early and critical form of cognitive engagement with the environment. To identify a natural object for its functional properties, transport it, and use it effectively requires foresight, planning, and a conceptual grasp of cause and effect. It is one thing to shape a tool. It is another—arguably more advanced—to recognize that nature has already done the shaping for you.
Hominin remains found near the spheres suggest that this behavior was shared across species. At Gombore IB, the earliest site, Homo cf. ergaster bones were recovered, indicating that this early African species—sometimes considered an African form of Homo erectus—was already selecting and using natural spheres nearly 1.7 million years ago. At Gombore II-1 and Garba IIIE, the fossils of Homo heidelbergensis and early Homo sapiens respectively suggest continuity and possibly cultural transmission of this practice. Through evolutionary transitions and environmental shifts, the behavior persisted.
“It is possibly the first evidence of the use of natural shapes for varied activities,” says Mussi, “and this happened repeatedly over more than 1 million years of human evolution at Melka Kunture.”
That last point is particularly striking. The persistence of this behavior across hundreds of thousands of years suggests not a chance occurrence but a deliberate strategy—an established pattern in hominin lifeways. These spheres were not randomly selected, nor was their use accidental. Instead, they point to a cognitive and cultural thread linking disparate hominin populations through time. And unlike more formal tool types, these spheres show how early humans were adept at co-opting nature’s designs, weaving them into their toolkits without the need to modify them overtly.
Why does this matter?
Because in the broader narrative of human evolution, such subtle behaviors are often overshadowed by the more dramatic milestones: the first handaxe, the controlled use of fire, the emergence of symbolic art. But it is precisely these quieter, more persistent behaviors that offer deep insight into the cognitive landscapes of early humans. Recognizing and exploiting natural affordances—seeing a round basalt stone not just as an object but as a potential tool—represents a form of intelligence rooted in abstraction, experimentation, and adaptability.
Moreover, this discovery expands our concept of what constitutes a tool. The anthropological bias toward shaped or modified artifacts has long dominated the field, shaping both excavation methods and interpretative frameworks. But Mussi’s work challenges that notion, reminding us that tools need not be transformed by human hands to be significant. Sometimes, the transformation lies in the act of recognition—in the mental leap that turns a rock into a purpose-driven object.
As archaeologists and paleoanthropologists continue to explore the early chapters of our species’ story, studies like this help broaden the lens through which we interpret the past. They encourage us to look beyond the obvious, to consider how ancient hominins not only made tools but thought about tools. And in doing so, they reveal a lineage of intelligence stretching deep into our prehistoric past—one that speaks not only of survival but of ingenuity.
At Melka Kunture, then, the basalt spheres are more than geological oddities or archaeological footnotes. They are quiet testaments to a long arc of human cognition—evidence that, even in the earliest days of our ancestry, we were already watching the world carefully, learning from it, and, in our own way, beginning to master it.
More information: Margherita Mussi, The volcanic rock spheres of Melka Kunture (Upper Awash, Ethiopia) at Gombore IB and later Acheulean sites, Quaternary International (2025). DOI: 10.1016/j.quaint.2025.109681