Roman dodecahedron found in Lincolnshire

An unusually large and fine example of the mysterious Roman dodecahedron has been unearthed in a community dig at Norton Disney in Lincolnshire. Only about 130 of these hollow 12-sided copper alloy objects have been found in the northern and western provinces of the Roman Empire. This is only the 33rd example found in England. None have been found in Rome, and there are no references to them in the ancient sources, inscriptions, frescoes, mosaics or any other medium to explain their purpose or function.

The dodecahedron was discovered last summer in an excavation at Potter Hill. It was in a pit, perhaps a quarry pit, filled with Roman pottery fragments and demolition rubble. It weighs 250 grams, is 8 cm (3.1 inches) wide, significantly larger than average, and 100% complete. As rare as these objects are, even more rare is to discover them in situ where they were originally buried in an organized archaeological dig. Most of them are in private collections and museums (or in a looter stash) and all information about their original contexts is lost.

A Roman villa was discovered less than a half mile away from Potter Hill in 1933. It was near the Fosse Way, a major Roman road that connected Exeter to Lincoln. The remains of the villa, entirely underground today, are on the National Heritage List. (Unfortunately, the villa is also on Historic England’s “Heritage At Risk” register, not because of human interference, but due to a community of badgers, who are themselves protected, having dug a sett amidst the ruins.) It’s possible the pit where the dodecahedron was found was associated with the villa or another high-status building.

There are many theories about what the Gallo-Roman dodecahedra were used for — candle holders, surveyor’s tools, wool crafts — but no consistent wear patterns attest to any such practical usage. X-ray fluorescence analysis of the Norton Disney dodecahedron found it is 67% copper, 7% tin and 26% lead. That is a high proportion of lead, a soft and malleable metal very prone to squishing, for a tool. They are all different sizes, ranging from golf ball to baseball, and have holes of different sizes in each of the 12 hexagonal sides. The overall design is consistent, however, with ball-shaped studs at each corner of the sides. In the end, the old fallback of “ritual purpose” seems to be the likeliest possibility.

The dodecahedron went on display at the National Civil War Centre in Newark-on-Trent on January 3rd and is already drawing crowds. It was also featured on the January 9th episode of the long-lived BBC series Digging for Britain. The episode can be viewed here but only from UK IP addresses.

The Norton Disney History and Archaeology Group is raising funds for a follow-up excavation of the find site this summer. The team was not able to finish their excavations of trenches three and four, so they need to go back to pick up where they left off. They have a very modest of goal of £1,000 and have raised £155 of it so far. To contribute to their efforts, donate here.

You can get a great view of its unusual size and excellent condition in this short video:

23 thoughts on “Roman dodecahedron found in Lincolnshire

  1. I was just wondering if they could have been used to ascertain the true value of coins as opposed to the face value. I seem to recall that people used to shave the edges of coins and it occurred to me that the different size holes in the dodecahedron could relate to different sized coins.

    1. *It appears that it could be used to measure coins, but old Roman coins in that time period were not very round. They were hammered on a die, not cast, so they weren’t a perfect circle. Weight and material constitution was more important to the value of an ancient coin, not a perfectly round shape (especially since the coins were debased of their silver so much during the 3rd century). In addition, some dods had triangular and oval holes.

    1. *Knitting tool- interesting, but some dods are too small for this; also some holes are too small; finally, that type of ‘French’ knitting tube size is based on number of knobs(pins), not the opening/hole sizes.

  2. Most fascinating. I think it may be a precursor to a modern grenade. If you stuffed it with fat soaked wool, lit it, then flung with a lacrosse style stick at a frightened enemy. It could do devastating harm considering how difficult it would be to extinguish it in battle. Longer burning than an arrow.

  3. The size of this argues against the one idea these were used to knit socks or mittens (unless the owner had really big hands and feet….)

    1. Of the 130 +/- known to be found, they range in size from just over 1 inch to about 4 inches. That really isnt all that large. I think the knitting theory has the yarn wrapped around it like a spool to make a “ball of yarn”, so really they could be pretty much any size.

    1. Some dods are too small for this; also some holes are too small; finally, that type of ‘French’ knitting tube size is based on number of knobs(pins), not the opening/hole sizes.

  4. Used to estimate ranges for a Roman ballista. The holes are to look through. The balls on the corners are for winding a cord around….a knotted measuring cord. By looking through the opening at a distant man, holding the dodecahedron closer or further away, so that the man fills the entire field of view, the distance is known by how far the Dodecahedron needs to be held from the user’s eye. The cord is held close to the eye and measures the distance of Dodecahedron to the eye. The cord is divided into segments equal in length to the edge length of the dodecahedron itself. Picture a knot at each corner post.
    By storing the measuring cord by winding onto the dodecahedron with some tension, its repeatable length is maintained even in foul weather.

    Other holes provide a quick check of a fixed distance so the ballista could be fired as soon as a target entered the range it was set to fire at. The differing hole sizes define a cone of vision that sets the eye distance without the measuring cord. This may sound complicated but is actually quite easy and quick to do.

    The ratio of hole sizes to edge lengths of the dodecahedron can be shown to be related, so that counting knots on the cord gives the distance in Roman feet (pedes). 5 pedes/knot, 10 pedes/knot, and 6 pedes/knot 12 pedes/knot are all common. Others that appear to be 7.7 pedes/knot when viewing a man, turn out to be 5 pedes/knot when viewing a man on horseback.

    There are various pairs of opposing holes. When the holes vary significantly in size they define a cone that would have been used for fixed distance measurements. When the eye is at the apex of the cone the hole edges appear to coincide. This is the correct eye distance to read one particular distance. When viewing an enemy in motion the viewer will know the instant his enemy is at the chosen range.

    When opposite holes are nearly the same size they are intended for measuring an arbitrary distance using the knotted cord.

    Often these holes that are intended to use with the cord are lacking the concentric ring designs. I think its safe to assume the concentric rings held pigment, so the different faces and ranges could be distinguished by color or the lack of color.

    The truth is in the numbers here. Its easily verifiable and not just conjecture.

    1. Hmmm… If this was a ranging device for ballistas (ballistae?), wouldn’t you expect them to be found mostly in military situations, i.e. legionary camps? But if it is a more general measuring device for rough surveying, that might make sense with the wider distribution.

      Still, this seems over-complicated for a measuring device. See the much simpler Biltmore Stick ( as an example of something used for rough measurements that doesn’t require precision casting.

  5. It’s the original ‘Swiss Army’ multi-tool…

    The enigmatic object many refer to as a Roman Dodecahedron (or ‘Dod’) is of Gallic/Gaul (including the Swiss Alps)/ Germania Superior & Inferior Origin (Northwest Gallic area). It was initially produced by Gallic-descendent blacksmiths (ordered by Negotiatores/Roman arms merchants in conjunction with Roman auxiliary army arms officers) as a ‘blueprint’/spec key, for procurement & (in varied instances a) maintenance tool for wood polearm weapon shafts (mainly) for the Roman Auxiliary troops (of which polearm size/specs were varied)…for region-specific tribe/area troops. Some were also used to measure tool handle shaft diameters (including oval) -see Jublains Dod.

    The Dods began to be used during Roman fort expansions on the frontier regions of conquered Gaul in the Rhine/Danube regions (and beyond in Britannia) when weapon procurement was localized. Auxiliary troops (as opposed to the Legion troops) used many of their own traditional weapons which had different specs/calibers, per tribe origin. There were many different tribes in the Gallic area and they were not organized as one cohesive power; thus the differences in their weapons, & the dod opening sizes (and ratio of opposing opening sizes to each other), per dod.

    For more info:

  6. My spontaneous idea for the purpose of this piece was to facilitate finding the correct ring size in jewelry making. The customer could insert their finger into the various holes to determine the appropriate size for their finger. Subsequently, the goldsmith could then use the template to design the ring or a similar piece of jewelry. This hypothesis would:
    1. Explain the varying diameters of the openings.
    2. Suggest that a goldsmith or blacksmith would likely possess the necessary skills to craft such an item.
    3. Account for the frequency of findings without a description of such a “mundane” tool.
    4. Indicate that while the concept remained consistent, the exact dimensions may have varied from dodecahedron to dodecahedron.
    5. Highlight that the described usage would not lead to wear and tear on the object, allowing it to be used across generations without significant signs of use.
    6. The markings around the holes help to identify which hole was chosen by the customer.
    7. The object was also crafted in such an intricate manner to showcase the skills of the goldsmith to the customer. Someone who can create such a dodecahedron can also craft a piece of jewelry.
    To refute this hypothesis, it would need to be demonstrated that the diameters are either too small or too large for typical fingers, which unfortunately cannot be verified based on the photos. Should the discovery sites be located near artisanal establishments, especially metallurgical facilities, this would provide further support for the theory.

  7. The Romans were heavy into sporting.

    I assume it could be a sporting device for throwing similar to horse shoes?

  8. Pure speculation: I wonder if the astronomical function may be the most appropriate given the association with the zodiac (I recall a continental version that has chased figures suggesting this), it may be for predicting/reading phases of the moon if the sizes of the apertures also relate to the distance of the moon from the earth over 12 months? Pure speculation – but I’m wondering on whether an external light source, candle or lamp, might be used to throw shadows onto a blank wall – difficult to test without very specific measurements.

  9. My first reaction on seeing this on television was that it is a 12 sided dice with different hole sizes denoting the numbers 1 to 12. Small ones may have been used indoors on a floor or table. Larger ones such as this may have been used outdoors on grass similar to lawn bowls. Just as some have suggested the holes could have contained glass they could also have also had wooden plugs with numbers carved or written on the surface. I think together with this theory the knobs are designed to throw it from side to side such that an expert thrower cannot control its final resting position. We know the Romans liked playing dice.

Leave a Reply

Your email address will not be published.