New 3D Angle Measurement Cracks an Ancient Stone Tool Mystery

For more than a century, archaeologists have argued over a deceptively simple question: when they pick up an ancient stone blade, can they actually tell how it was made? The traditional answer has relied on the trained eye, reading subtle scars, lips, and bulbar shapes on the flake surface to infer whether a prehistoric knapper swung a hard stone hammer, tapped with a soft antler or wooden hammer, used a punch to deliver an indirect blow, or squeezed the blade free with pressure. The problem is that these visual cues overlap. Two blades made in entirely different ways can look strikingly similar, and even experienced analysts disagree with one another. A new study published in Archaeological and Anthropological Sciences now offers a way out of this interpretive fog, using three-dimensional surface scanning and a single measurable angle to separate flaking techniques with remarkable precision.

The study, conducted by Kenji Nagai of Aichi Gakuin University in Japan, introduces a quantitative framework built around a metric called the bulb surface angle, or BSA, together with a companion measurement known as the vertical interval, or VI. Rather than treating the bulb of percussion, the swollen area just below where a blow strikes the stone, as a vague morphological feature, the approach measures its geometry directly from high-resolution three-dimensional surface data. The measurements are taken within a standardized three-dimensional coordinate system, which means that every blade is oriented and quantified in exactly the same way. That standardization is crucial, because it removes much of the subjectivity that has plagued traditional attribute-based classification and allows results from different analysts and laboratories to be compared on equal footing.

To test whether BSA and VI genuinely carry information about production technique, Nagai assembled an experimental reference collection of 613 blades manufactured by skilled knappers using four distinct flaking techniques: hard-hammer direct percussion, soft-hammer direct percussion, indirect percussion, and pressure flaking. These four methods span the range of force-delivery strategies known from the archaeological record, from the blunt, high-energy strikes of a stone hammer to the slow, controlled loading of a pressure tool. Each blade was scanned, its BSA and VI measured, and the resulting values plotted in a combined BSA-VI space. The question was whether blades made by different techniques would occupy different regions of that space, or whether they would blur together as so many traditional attributes do.

The answer, in large part, was that they separate. The analysis showed that blade production techniques form distinct distributional patterns in BSA-VI space, reflecting real physical differences in how force is transmitted into the stone and how fractures propagate through it. Pressure flaking, in which the knapper applies steady compressive force rather than a dynamic impact, and indirect percussion, in which an intermediate punch channels the blow, produced especially distinctive signatures. This makes intuitive sense from a fracture-mechanics standpoint: the speed and manner of loading influence the shape of the cone of force that spreads through the material, and that cone shape is precisely what the bulb surface angle records.

The statistical payoff came when Nagai fed the measurements into logistic regression models that combined BSA and VI. For pressure flaking, the model achieved a balanced accuracy of 97.5 percent, and for indirect percussion it reached 96.1 percent. In practical terms, this means that a blade made by pressing or punching can be identified with near-certainty from its three-dimensional geometry alone, without any reliance on the analyst’s intuition. Balanced accuracy is a conservative measure that accounts for class sizes, so these figures represent robust discrimination rather than statistical flattery. For a field in which technique identification often rests on contested visual criteria, classification performance at this level is a striking result.

The picture is more complicated for direct percussion. Blades struck directly with a hammer, whether hard or soft, showed substantial overlap in their BSA-VI distributions and remained difficult to tell apart. In particular, discrimination between hard-hammer and soft-hammer direct percussion was limited, confirming a long-standing frustration in lithic analysis. Both techniques deliver dynamic impacts through the knapper’s arm, and the resulting fracture initiation conditions evidently converge on similar geometric outcomes more often than traditional wisdom suggests. The study does not claim that the two are indistinguishable in every case, but it honestly quantifies the limits of what a single pair of surface measurements can resolve, which is itself a valuable contribution to a literature often criticized for overconfident attributions.

What elevates the finding beyond a clever classification trick is its interpretation. The results suggest that BSA captures technological variation associated with modes of force transmission and fracture propagation, rather than serving merely as a geometric descriptor of flake morphology. In other words, the angle is not an arbitrary shape statistic; it is a physical fingerprint of how energy entered the stone. This connects the metric to a deep body of experimental fracture research, from classic mechanical studies of conchoidal flaking to modern controlled experiments on how hammer mass, velocity, and platform geometry shape the flakes that detach. A measurement that tracks the physics of fracture initiation is far more likely to generalize to new assemblages than one that merely correlates with surface appearance.

The methodological rigor of the study deserves attention as well. The specimen-level dataset for all 613 blades is published as supplementary material, and a subset of 167 specimens was subjected to repeated BSA measurements to assess within-observer repeatability, an explicit check that the measurement itself is stable across sessions. The native three-dimensional profiler files are retained by the author and available on reasonable request. This kind of transparency matters because quantitative approaches in archaeology live or die on whether independent researchers can reproduce them. By publishing the underlying data and documenting the measurement procedure, including endpoint placement on the three-dimensional height maps, the study provides a template that other laboratories can follow and extend.

The implications for archaeology reach well beyond the experimental lab. Blade production is one of the defining technologies of the Upper Paleolithic and later prehistoric periods, and the specific flaking technique a community used carries information about raw material economy, skill transmission, mobility, and cultural tradition. Pressure blade production in particular has been treated as a hallmark of certain technological traditions, from the postglacial societies of northern Europe to the Upper Paleolithic of northern Japan. A reproducible metric that can flag pressure flaking and indirect percussion with over 96 percent balanced accuracy gives archaeologists a new tool for testing long-standing hypotheses about where and when these techniques emerged and spread, using the blades themselves rather than contested typological labels.

There are, of course, caveats. The reference collection consists of experimentally produced blades made by skilled knappers, and archaeological specimens will add variability from raw material differences, weathering, breakage, and the less-than-perfect consistency of ancient hands. The limited separation between hard- and soft-hammer direct percussion means that many assemblage-level questions will still require additional evidence. Yet the direction of travel is clear. By combining accessible three-dimensional scanning with a physically meaningful measurement and standard statistical classification, the study converts one of lithic analysis’s most stubborn identification problems into a quantifiable one. As three-dimensional surface data becomes routine in archaeological documentation, measurements like the bulb surface angle may soon become as standard on a blade’s record as its length and weight, quietly replacing generations of eyeball judgment with numbers that anyone can check.

Subject of Research: Quantitative identification of stone blade flaking techniques using bulb surface angle measurements derived from 3D surface data

Article Title: Quantitative identification of flaking techniques using Bulb Surface Angle (BSA) derived from 3D surface data

Article References: Nagai, K. (2026). Quantitative identification of flaking techniques using Bulb Surface Angle (BSA) derived from 3D surface data. Archaeological and Anthropological Sciences, 18(11), Article 215. https://doi.org/10.1007/s12520-026-02561-y

Image Credits: AI Generated

DOI: 10.1007/s12520-026-02561-y

Keywords: lithic technology, blade production, bulb surface angle, pressure flaking, indirect percussion, direct percussion, 3D surface analysis, fracture mechanics, experimental archaeology, logistic regression, Paleolithic, stone tools