Roman Dodecahedra

Summary

Roman dodecahedra are small, hollow, twelve-sided objects of cast copper alloy, with a round hole of a different size in each face and a knob at every corner. About 134 are known, almost all from Gaul, the Germanies and Britain, and no ancient text says what they were for.

The first was recorded in 1739; one of the most recent came from an excavated pit in Lincolnshire in 2023. Explanations run from surveying instruments to knitting tools to objects of divination. One of them, the surveying instrument, has been ruled out. None of the others has been shown to be right.

What is documented

The objects. Each is a hollow regular dodecahedron of copper alloy, made by lost-wax casting, with walls 0.5 to 4 millimetres thick: twelve flat pentagonal faces, each pierced by a single round hole, with a small ball or knob at each of the twenty corners. According to a 2025 survey by Michael Guggenberger and Stephen Leach, they measure from 4 centimetres to about 10 centimetres across, face to face and excluding the knobs, and the holes range from 0.6 to 4 centimetres, each hole on an object a different size. The most common type has ten round holes, usually ringed by three engraved concentric circles, and one opposite pair of holes without circles.

How many, and where. Guggenberger and Leach count about 134. Around 70 per cent come from the Gallic and Germanic provinces and around 20 per cent from Britannia; in all, about 90 per cent come from lands that belonged, for a short time, to the breakaway Gallic Empire of AD 260 to 274. None has been found in Italy. They date from around AD 200 to the late fourth century.

Where they lay. Of those with a recorded find place, Guggenberger and Leach report that more than half come from towns or other settlements, just under a fifth from military camps, about 8.5 per cent from places with a plausible sacred connection, about 7 per cent from graves or cemeteries, and about 5.5 per cent from backfilled wells or refuse pits. Only very few have been recovered from stratified contexts in archaeological excavations.

No description. No ancient author describes the objects. Guggenberger and Leach, who argue for a philosophical inspiration, still begin from the fact that "there is no contemporaneous description".

The first record. The minutes of the Society of Antiquaries of London for 28 June 1739 record that George North exhibited a twelve-sided bronze object from Aston, Hertfordshire.

Norton Disney, 2023. In June 2023 the Norton Disney Archaeology Group, a community group, found a dodecahedron in Lincolnshire in the fourth trench of its excavation, in what appeared to be a large pit. It measures about 8 centimetres across and weighs 245 grams. It was the 33rd found in England and the first in the Midlands. The group's secretary, Richard Parker, said: "It is completely undamaged and there is no evidence of any wear at all." Guggenberger and Leach, who list it as number 132 in Guggenberger's catalogue, place it on top of a large pit of building debris 700 metres east of a Roman villa.

Beyond the frontier. In 2025 the archaeologist Petra Dragonidesová published three copper-alloy fragments, probably from two dodecahedra, found at Hurbanovo and Chotín in south-western Slovakia, "outside the Empire's borders, within the territory of Germanic settlements". She concluded that "the true purpose of the dodecahedron remains an ongoing subject of scholarly investigation."

A relative. A hollow Roman icosahedron, a twenty-sided object of the same family, comes from a grave of around AD 200 at Arloff, south-west of Bonn.

Leading explanations

A religious, divinatory or philosophical object (often proposed, not established). Writers on the objects often note the dodecahedron's place in Platonic and Pythagorean thought. In 2025 Guggenberger and Leach argued that several texts, including Plato's account of the "receptacle of all becoming" in the Timaeus and a passage in Iamblichus, point to the objects' possible inspiration, and linked their distribution in Gaul and Britain to the Druids' interest in Pythagoras, which Roman writers remarked on. Some come from graves and sacred places, but more than half come from settlements, and no find has shown how one was used.

A surveying or range-finding instrument (rejected). Guggenberger and Leach write that the suggested use as a distance-measuring device "must be rejected", as F. H. Thompson showed in The Antiquaries Journal in 1970. The objects are not made to a standard size: they range from 4 to about 10 centimetres across.

A knitting tool (modern suggestion, untested). The idea circulates widely and appears in lists of theories, including the Norton Disney group's. No ancient source or find connects the objects with knitting.

Other suggestions (untested). A community archaeologist writing about the Norton Disney find lists religious objects, knitting tools, measuring instruments and "stress toys" among the theories in circulation. None has direct evidence.

Current status

Unsolved. The objects are well described: their shapes, sizes, distribution, dates and find places are catalogued. Their purpose is not known. The distance-measuring idea has been ruled out, and every surviving explanation lacks direct evidence. What would change this is a dodecahedron found in a context that shows its use, or an ancient text or image that shows one being used.

Last reviewed: September 2026. Records are provisional. Where the evidence changes, the entry changes. Found an error? Tell us.