Mysterious 12-sided Roman object found in Belgium may have been used for magical rituals

A photo of a metal, baseball-size Roman dodecahedron against a white background.
No one knows what the Roman dodecahedrons were for. Archaeologists think they probably had a religious or magical meaning. (Image credit: Kris Vandevorst/Flanders Heritage Agency)

A metal detectorist in Belgium has unearthed a fragment of a mysterious bronze artifact known as a Roman dodecahedron that is thought to be more than 1,600 years old. 

More than a hundred of the puzzling objects — hollow, 12-sided geometric shells of cast metal about the size of baseballs, with large holes in each face and studs at each corner — have been discovered in Northern Europe over the past 200 years. But no one knows why or how they were used.

"There have been several hypotheses for it — some kind of a calendar, an instrument for land measurement, a scepter, etcetera — but none of them is satisfying," Guido Creemers, a curator at the Gallo-Roman Museum in Tongeren, Belgium, told Live Science in an email. "We rather think it has something to do with non-official activities like sorcery, fortune-telling and so on."

Creemers and his colleagues at the Gallo-Roman Museum were given the fragment by its finder and identified it in December. It consists of only one corner of the object with a single corner stud, but it is unmistakably part of a dodecahedron that originally measured just over 2 inches (5 centimeters) across.

The fragment found in a field near the town of Kortessem in Flanders is clearly part of a Roman dodecahedron. (Image credit: Kris Vandevorst/Flanders Heritage Agency)

Metal detectorist and amateur archaeologist Patrick Schuermans had found the fragment months earlier in a plowed field near the small town of Kortessem, in Belgium's northern Flanders region.

Related: Sacred chickens, witches and animal entrails: 7 unusual ancient Roman superstitions

Creemers said the Gallo-Roman Museum already displays a complete ancient bronze dodecahedron found in 1939 just outside Tongeren's Roman city walls, and the new fragment will go on display next to it in February.

Archaeologists are now investigating the site where the metal detectorist found the dodecahedron fragment; it may have been the site of a Roman villa. (Image credit: Kris Vandevorst/Flanders Heritage Agency)

Mysterious dodecahedrons 

The first Roman dodecahedron to be discovered in modern times was found in England in the 18th century, and roughly 120 have been found since then in Great Britain, the Netherlands, Belgium, France, Germany, Austria and Switzerland.

It's not possible to date the metal itself, but some dodecahedrons were found buried in layers of earth that date them to between the first and fifth centuries A.D.

The mystery doesn't end there; archaeologists cannot explain the geometric artifact's function, and no written record of the dodecahedrons has ever been found. 

A complete Roman dodecahedron found near the ancient Roman walls of the town of Tongeren in Belgium in 1939. (Image credit: Gallo-Romeins Museum Tongeren)

It's possible they were used in secret for magical purposes, such as divination (telling the future), which was popular in Roman times but forbidden under Christianity, the religion of the later Roman Empire, Creemers said. "These activities were not allowed, and punishments were severe," he explained. "That is possibly why we do not find any written sources." 

Several explanations for the mysterious artifacts have been suggested over the years. Initially, they were described as "mace heads" and were thought to be part of a weapon. Other ideas are that they were tools for determining the right time to plant grain; that they were dice, or other objects for playing a game; and that they were instruments for measuring distance, possibly for finding the right range for Roman artillery, such as ballistas

A recent suggestion is that dodecahedrons were knitting patterns for Roman gloves.

But most archaeologists think the objects were probably used in magical rituals. The dodecahedrons have no markings indicating how they were used, as might be expected for measuring instruments, and they all have different weights and sizes, ranging from 1.5 to 4.5 inches (4 to 11 centimeters) across.

Roman dodecahedrons are also found only in the Roman Empire's northwestern areas, and many were unearthed at burial sites. These clues suggest that the cult or magical practice of using them was restricted to the "Gallo-Roman" regions — the parts of the later Roman Empire influenced by Gauls or Celts, according to Tibor Grüll, a historian at the University of Pécs in Hungary who has reviewed the academic literature about dodecahedrons. 

Related: The 5 craziest ways emperors gained the throne in ancient Rome

Ancient puzzle

Creemers said the dodecahedron fragment found near Kortessem could shed more light on these mysterious metal objects. Many other Roman dodecahedrons were first recognized for what they were in private or museum collections, so their archaeological context is unknown, he said.

But the location of the Kortessem fragment is well documented, he said; and subsequent archaeological investigations have revealed mural fragments at the site, indicating that it may have been a Roman villa.

A translated statement by the Flanders Heritage Agency said the fractured surfaces of the fragment indicate that the dodecahedron had been deliberately broken, possibly during a final ritual.

The location will now be monitored for further finds.

"Thanks to the correct working method of the metal detectorist, archaeologists know for the first time the exact location of a Roman dodecahedron in Flanders," the statement said. "That opens the door for further research."

Live Science Contributor

Tom Metcalfe is a freelance journalist and regular Live Science contributor who is based in London in the United Kingdom. Tom writes mainly about science, space, archaeology, the Earth and the oceans. He has also written for the BBC, NBC News, National Geographic, Scientific American, Air & Space, and many others.

  • GoldHorde
    I'd go with a CALTROP used against horses (cavalry/chariots) but the differing hole dimensions do suggest a calculative purpose.
  • claudiusxander
    Search youtube for some lovely people knitting gloves using these dodecahedrons. The different holes are for varying finger sizes, Also they are only found in northerly colder regions of roman empire, where gloves were necessary.
    "Knitting with a Roman Dodecahedron"
  • EMM
    It could be a "massage therapy" device. The holes would allow a "hot" coal or other hot object to radiate while the knobs "rolf" muscles. I expect we will never know for certain but that's my input.
  • Baryonyx
    Nowadays we might say these dodecahedra look like viruses, thanks in part to those projections. (The "head" of a bacteriophage is typically rendered as an icosahedron--& the whole thing resembles a lunar landing craft!)
  • Stratovaari
    As this objet has various sized holes I assume it has been used to check the sice of metal objects. Would be easy to determine what the size of a coin or a metal sphere or some other spherical object is.
  • smee
    Someone has not been doing their homework. This object was the subject of study a number of years ago. It is not magical and not for knitting. It is a sighting device. The pairs of holes are used for determining distance. Line them up with the smallest closest to the eye and move your arm forward until the sizes of the holes appear to be identical. The distance from the eye is the N in the equation to determining distance. The differing sizes are used for estimating the rough distance before taking a proper sighting.
  • dcbrown
    When I read Amelia Carolina Sparavigna's paper I was certain she was onto something. I figured calculating the eye distances based on the hole measurements she provided would show they were consistent with a normal arms length, and that would be a nice verification. But I did the calculations and one hole pair on each of the measured dodecahedrons didn't check out. It would need to be held 3-6 meters from the user's eye. We should expect something about 20" or less for a handheld portable device. Wouldn't a reasonable person just use smaller holes? Her theory also doesn't explain the knobs at the corners. She's definitely onto something, but it's not the whole story.

    Another concern was that a much simpler device would have performed the same function better.
    Consider that a single hole in an ordinary flat disk, and a cord equally divided into equal segments (by knots) would allow estimating ANY arbitrary distance as long as the approximate size of the viewed object is known. This would give a more accurate distance estimate and be much simpler to make.

    But it would be possible to improve that simpler device. Fiddling with a measuring cord would be awkward if the user just needs a quick verification that his target is in range. So some means of quickly comparing actual range to a few standards (like the range of his ballista) without the cord would be desirable. The opposing holes of different sizes on the dodecahedron provide that. But it's an extra feature, not the primary function.

    But if they want to keep the capability of the more accurate measuring using a cord they would want some means to wrap the cord when not in use.

    Also consider that the cord could rot and be broken. If broken it would be difficult to replace in the field (a military campaign) with an accurately divided cord. A way to use an ordinary piece of cord that that somehow measures itself would also be nice.

    The dodecahedron addresses these shortcomings of the simpler device. A few hole pairs are provided that can be used for quick range verification checks without a measuring cord. The large and small holes can be used to define the proper holding distance by the method of corresponding edges.

    A measuring cord is provided to make arbitrary range estimates accurately. This would use those hole pairs that do not really work for the cordless method. A knotted cord would probably have been a provided initially.

    The knobs or balls on the corners provide a convenient means to wind the measuring cord. This can be done without blocking the holes if adjacent knobs are used, the cord always following the dodecahedron edges.

    When winding the cord this way it is naturally equally subdivided and self-measuring. Each segment of cord that is unwound is equal to one edge length of the dodecahedron. So a knotted cord is no longer essential. If the original knotted cord is lost or broken an unknotted cord can replace it. The cord would even be somewhat self-marking due to the tendency to discolor where it repeatedly rubbed on the copper alloy knobs. This makes the natural knot spacing for the cord equal to the edge length of the dodecahedron.

    To measure range an object of known height (probably a enemy fighter) is viewed through the hole in the dodecahedron. The user would hold the dodecahedron far enough from his eye that the height of the man appears to fill the field of view. The range is then proportional to the eye distance as measured with the cord. The actual distance in stadia or any other Roman units is irrelevant as long as the range of the ballista is known by the same method.

    This is for rough range estimates with a portable device, not surveying. There are more accurate methods available when you have access to the down range position.

    I think this is a reasonably complete explanation. All the features of the dodecahedron are explained and are consistent with this function.
  • danielkblackwood@outlook.
    I would like to put forward a new description of the workings of the artifact, please.
    The key to the Roman Dodecahedron is perspective.
    In 2019 I came to the conclusion that the artifact can be photographed from any direction and would produce a unique silhouette from each and every angle it was photographed.
    I found this to be true for the shadow cast by the artifact also.
    This 3D form made up of 2D images resembles an unseen radiant star pointing in any and all directions. As a big bang projecting outward or a globe projecting inward.
    Its uses could include a space compass used to navigate, map, or plot in 3D space.
    Further, the true form I was seeking when I made this discovery was a 3D writing system and to be honest that is also plausible.
    I understand that is a lot to fathom but for now, I'm calling it perspective theory by Daniel K Blackwood.
    Thank you.
  • April Lieberman
  • Uncialist
    I believe I have discovered the real use of the Roman dodecahedra.
    They were night time clocks using candles of differing thickness for each month of a year in the different size holes in the twelve faces. A major argument levelked at this is that no two examples have the same size holes so there was no standardization. What I suggest is that each object would need holes sized to the latitude where the artifact was used as the length of night varies with the date and its latitude. The variation within each month is less than needed so that candles of the same size for each month would be sufficient to monitor the changing of four watches used by yhe Romans. A full descriptive document is available.