‘Enormous petrified mulberries’: a new dimension on carved stone balls

Back to overview

Dr Hugo Anderson-Whymark explains how he’s brought a collection of of Neolithic carved stone balls to life in 3D. This post was first published on the National Museums Scotland blog on June 6, 2018.

Carved stone balls are one of Scotland’s most enigmatic prehistoric artefacts. Created some 5000 years ago in the Late Neolithic, their distinctive knobbed forms were carefully pecked and ground to shape by communities across the north east of Scotland.  A small number, such as the Towie Ball, were elaborately incised with circles and spirals and represent some of the finest examples of Neolithic ‘passage grave’ art in Europe.

Yet we still don’t know how carved stone balls were used; in the words of archaeologists Stuart Piggott and Glyn Daniel (1951) ‘their use is wholly unknown’.  It has long been postulated that they were weapons – mounted as maceheads or bound with sinew or twine and thrown like South American bolas. But other authors have suggested they were used as weights, measures, mnemonic devices or symbols of power; some mathematicians have even viewed them as representations of platonic solids.  It’s easy to see why these artefacts have captured the imagination for more than two centuries.

Around 525 carved stone balls are known, the vast majority being from Scotland, with rare finds from England, Ireland and one example from Norway. National Museums Scotland holds the largest collection of carved stone balls in the world, with some 140 originals and casts of a further 60 examples from other collections.  Some of our finest carved stone balls are on display in the Early People gallery, but as these artefacts are so popular we have decided to bring 60 of our carved stone balls to you as 3d models via our Sketchfab account.  These examples represent a broad cross-section of the different known forms and raw materials in our collections; from three to six knobbed stones in granite, to sandstone examples covered in numerous rounded projections.  The latter infamously described by Sir John Evans in his book Ancient Stone Implements (1897) as resembling ‘enormous petrified mulberries’.

Carved stone balls by National Museums Scotland on Sketchfab

These models were made using photogrammetry, which uses around 150-200 images of each artifact to produce an exceptionally high-resolution 3D model.  The resolution allows you to examine and appreciate these artifacts in unprecedented detail.  Indeed, the model of one carved stone ball (X.AS 90) revealed traces of fine concentric circles on one projecting knob that had never been recorded before, despite the artifact having been in the museum for more than 100 years and examined by dozens of scholars.  Traces of decoration and working are particularly clear in ‘matcap’ mode, which makes the artifact look like shiny metal, emphasising any irregularities in the surface.

Sharing these models on Sketchfab allows you explore these wonderful artifacts in 3d and Virtual Reality from the comfort of your own home, anywhere in the world.  Enjoy!

This blog post was originally published by National Museums Scotland.

Thanks Hugo! For more amazing 3D models, be sure to check out and follow National Museums Scotland and Dr Hugo Anderson Whymark on Sketchfab.

About the author

Dr Hugo Anderson-Whymark

I'm an archaeologist and Curator of Neolithic collections at National Museums Scotland.


  • wil lampman says:

    Dr Hugo, i have a far left-field idea about the intended use of the carved stone balls.
    With the groves and knobs covering the surfaces it would almost seem these were used as line weights for fish netting. Im thinking more along the idea of a simplistic form of gill net to be used in streams or standing water.

    Hence the spiral decoration designs denoting water.

    Just a thought.


  • Alan says:

    Can they be grinders of some kind?

    The finer edged ones you could hold in your hand while grinding up grain, the lumpier ones might be more in the way of nut crackers –put two of them in a barrel with nuts and roll it around.

  • Hugh says:

    Many of these balls are similar, and detailed patterns and in some cases intricate ornamentation means to me that they had an important use. The question is, was the use practical or ceremonial…or perhaps both?

    I doubt that they were used for fishing weights, as typical neolithic weights were simple stones with a hole in them. That doesn’t explain the one that has a hole through the center.

    As far as grain, fiber, bone or meat grinding, again, neolithic cultures used very simple stone-against-stone grinding tools.

    I’d like to hypothesize that the balls were used in a throwing game, whereas the owner was identified by his unique ball. Different designs were favored for their performance. Is it possible that both golf and rock throwing in the highland games had a common root to a rock throwing sport?

    A second hypothesis is that many of these balls have either sharp or dull points which would focus impact pressure when used as a weapon. They could have been thrown, slung or used at the end of a mace. Similar to the way that Native Americans would uniquely fletch or paint their arrows, time may have been spent to customize the look and function of their weapons. Given the investment of time in some of these designs, if they were weapons I doubt they were thrown, rather I’d lean towards them being retained as a carried weapon.

  • One appears to have wear, similar to that that of a pot boiler with many, many years of use.

  • Andy Fugate says:

    Maybe a thrown weapon. All the bumps and ridges would allow for a firm hold and would allow some curving, rising and dipping when thrown by a skilled “pitcher”.

  • Richard Troell says:

    Those are cast net weights.

  • Simon says:

    So where can we download them as STL for our 3D printers ?

  • Steve Casselman says:

    Maybe these were used to grind grains and other vegetation. The different designs would be for different grains and medical herbs. The ones with lots of knobs would be used first and break down larger plants and then you might use a smother one after that. If there was a wooden machine used then they would all have to be more or less the same size for that machine.

  • S. Elizabeth Bobo says:

    Thank you so much for showing this incredible work to the world! I love it. It seems that ancient peoples were not as primitive as I would have thought. I hope that you will find out more about them.

  • C Knight says:

    1. Stoppers for large vessels to hold liquid or grains.
    2. Walking stick handles.
    3. Something akin to a pommel on a horse saddle, for some ancient form of saddle. Maybe not even a saddle, maybe it was tied into a horse mane as something to hold on to when riding.
    4. Currency for trade goods. Like a signature IOU that could be traded.
    5. Maybe they were pressed into trees leaving behind a signature indentation to mark property or hunting territory boundaries. Or trail markers.

  • Ford Bond says:

    How are the images being created photogrammetry or scanning? What software is being used to process the images prior to using Sketchfab?

  • Graham says:

    I believe I have three of these balls two of which have been unearthed at quarries in Staffordshire. All three are of the same material (white erratics). One has defined concentric circles and true to your own findings, archaeologists are at a loss to explain their use. I am a retired minerals planner. One of the quarry’s has been subject of a number of Neolithic finds. Would be happy to send you a photo

  • Hugo Anderson-Whymark says:

    Hi Wil,

    Net weights has been suggest before, but the argument against is that fish don’t appear to have been eaten in Britain during the Neolithic, except in small quantities in the Northern Isles

  • Hugo Anderson-Whymark says:

    Hi Alan,

    We have grain rubbers that look similar, but they are late in date and always have a distinct patch of wear that is lacking from these knobbed balls.

  • Hugo Anderson-Whymark says:

    Dear Hugh,

    Thanks for your thoughts. Both your hypotheses have been suggested before and are perfectly plausible!

  • Hugo Anderson-Whymark says:

    Hi Ford,

    These were produce using structure from motion photogrammetry, processed using Agisoft Photoscan. Typically each model was produced from 150-200 20mp images.

  • Hugo Anderson-Whymark says:

    Dear Graham,

    Sounds very interesting. Please do send me some images to h.anderson-whymark@nms.ac.uk and I’ll be sure to get back to you.

Leave a Reply

Your email address will not be published.

Related articles