A tool manufactured on a human radius was discovered between 1935 and 1945 in the caves of Goyet (province of Namur, Belgium). A splinter from the artefact yielded an AMS date of 2420 ± 40 BP (OxA-8875), i.e. between 760 and 400 ВС after calibration at 1 σ and between 770 and 390 ВС at 2 σ. Such a result dates the object to the Iron Age, although the size of the standard deviation due to plateaux on the calibration curve prevents greater precision. The tool was shaped by using the left radius from an adult, perhaps male. The distal end of the bone was removed during tool preparation. On the distal part, a partial longitudinal edge is present, prepared by scraping of the palmar and dorsal surfaces and practically aligned with the prolongation of the interosseous crest, from which, however, it is easily distinguished. The lateral surface of the bone is a longitudinal back opposite the prepared edge. Unworked on the central part of the object, this back was, however, worked on the most distal part to contribute to the formation of the point. The object cannot thus be regarded as a dagger stricto sensu. In spite of the relative shortness of the worked edge, it corresponds rather to the definition of a knife. The object is typologically exceptional. On the basis of archaeological literature, it is the only sharp- pointed tool created on a radius known for the prehistoric and protohistoric periods in Europe and North Africa; the few other tools made from human long bones are typically on fibulae and, more rarely, on ulnae and humeri. These comparable tools are moreover characterized primarily by their point, which often qualified them as daggers, while the originality of the Goyet object is to associate the point with a worked edge opposed to a natural back.