Hyparhedra are polyhedra made from hypars – hyparbolic paraboloids – a shape more commonly known in the bead world as a warped square. I was introduced to the idea of hypars and the academic work on them a while ago by the Contemporary Geometric Beadwork project.

Polyhedra made from hypars were named hyparhedra by Erik Demaine, who wrote a paper (“Polyhedral Sculptures with Hyperbolic Paraboloids”, Demaine, Demaine & Lubiw, 1999) on a method of making any polyhedra from paper folded hypars. There are details of the method and photos of the shapes here. It’s based on making “k-hats” – pyramids made from hypars – and using them as the faces of polyhedra. They demonstrated how to make a variety of shapes including all the Platonic Solids – that’s the five possible polyhedra made entirely out of the same regular shape: the tetrahedron, the cube, the octahedron, the dodecahedron and the icosahedron.

You may also notice on that page that one of the first shapes is recognisable to the beader as a warped square star! (In their paper they mention that the earliest example of a warped square star in origami is from 1958!) This method also works with beadwork hypars – Joke van Biesen and Kim Heinlein have both made beautiful beadwork versions of the hyparhedra cube from the paper.

The photo below shows a beadwork 4–hat – four warped squares joined together on two sides each to make a 4-sided pyramid. Each of the warped squares is half cream and half red, so that the pyramid in the centre ends up one colour and the edges another. I’ve outlined the warped squares in green on the right to highlight how they’re joined together.

I’ve been playing around with a slight variation of the idea for some time now. Instead of the method using k-hats as the faces, I fold one warped square over each edge of a polyhedron and join them so that their points are at the vertices and centres of each face. This method seems to be fairly well known in the origami world, but I haven’t found anything about it in the mathematical literature so far. I have however seen numerous independent discoveries of a dodecahedron made this way in the beadwork world – the infamous rhombic hexacontahedron! I think the dodecahedron is the only Platonic solid that will work this way, so I decided to try using this method to make various Archimedean solids.

So far I’ve completed a truncated tetrahedron, the dual of a cuboctahedron (a rhombic dodecahedron) and a truncated octahedron:

I initially didn’t think that these would work as their dihedral angles are quite small (that is, they have “sharper” edges compared to larger shapes), but the beadwork turned out to be more flexible than I thought! Here’s a photo of the truncated tetrahedron from a different angle showing how much some of the squares will distort:

Here’s an more detailed illustration of the method using the truncated octahedron – a shape with 6 square faces and 6 hexagonal faces. The photo below shows the outline of the truncated octahedron over the hyparhedra:

And this photo shows how a warped square has been placed over an edge with the two “upwards” points at the corners and the two “downwards” points at the centre of the faces:

Interestingly the shape formed by this hyparhedra is quite similar to the third stellation of the rhombic dodecahedron. Which brings me back to the rhombic dodecahedron hyparhedra:

This shape was interesting – trying to put warped squares over the edges of a cuboctahedron didn’t work as the surface ends up curving in the wrong direction. The squares want to be the opposite way around – so that their upwards points are in the centre of the faces not at the vertices. This just creates the “dual” polyhedron – the shape you get if you join the centres of the faces of a polyhedron together. In this case it’s a very pointy rhombic dodecahedron!

I’ve started a snub cube which seems to be working out the same way and looks like it will end up as it’s dual shape too. I have a rhombicuboctahedron in progress too but I’m not sure if it will work as itself or as it’s dual yet.

I originally started all this with a truncated icosahedron. However I couldn’t decide on colours so ended up with a lot of individual pieces and not much completed. This is the main reason I’ve switched to using the picasso delicas for all these shapes – it gives me a few colours options without being overwhelming. So I’m restarting the truncated icosahedron in red and brown to match the two other truncated shapes I’ve made so far!

The biggest piece I have in progress though is a rhombicosidodecahedron, which is working well but is also going to be very big! Here is the piece I’ve completed so far (about a sixth of the total) compared to the other shapes:

There’s only one other Archimedean solid larger than this – a snub dodecahedron which needs 150 warped squares! I’m wondering if this one will work as its dual since this has happened with some of the other shapes with a lot of triangular faces.

There are also some Archimedean solids with 8 or 10 sided faces, which will not be flat when made this way with warped squares (since warped squares make flat hexagons when joined on two sides). However, I’ve tried joining 8 warped squares together as the start of a truncated cube and I think it will still work!

The different shapes and surface curvatures that can be generated by this hyparhedra method are interesting – part of my reason for working through all of these systematically is to gain a better understanding and feeling for how to make more general surfaces so I can create more complex shapes. I’ve already used the ideas in several other shapes – for example the rick rack dodecahedron and variation are joined together with warped squares making up the edges of the underlying polyhedra. It was the hyparhedra idea here that helped me work out how make these two pieces!

If you’d like to try making a hyparhedra here are some brief instructions on how to make the truncated octahedron!