(True) Goldberg operations as a formal approach to operations on polyhedra

Gunnar Brinkmann, (Universiteit Gent) Pieter Goetschalckx, (Universiteit Gent) Stan Schein (Univ. of California, Los Angeles) text and pictures by Helmer Aslaksen

Faculty of Science and many more. . .

source: wikipedia

Faculty of Science The famous Goldberg-Coxeter construction

(Figure from Coxeter’s paper)

Faculty of Science equilateral

(3,1) "Goldberg parameters" (0,0)

Cut out this triangle. . .

Faculty of Science and glue a copy into each triangle of the (or another triangulation)

Old vertices keep their degree, new vertices have degree 6. For : then take the dual. This idea can also be used to decorate polyhedra where all faces are 4-gons with quadrangles cut out of a quadrangular net so that all 4 sides are equivalent. (Tarnai, Kov´acs,Fowler, Guest)

(3,1)

For other face sizes or a mixture of and quadrangles it does not work.

Faculty of Science Coxeter (1971):

“Independently of Michael Goldberg (1937), Caspar and Klug (1963) proposed the following rule for making a suitable pattern”.

Does this imply or at least suggest that they proposed “the same” rule?

So what exactly did Goldberg do?

Faculty of Science Let’s look at a figure from Goldberg’s paper:

That’s a right triangle – not an equilateral one. . . Goldberg starts from the and states: “Each pentagonal patch may be divided into ten equal (congruent or symmetric) triangular patches”.

Faculty of Science Reverse operation: subdivide the of the dodecahedron and glue copies (half of them mirrors) of the right Goldberg triangle cut out of the into it.

Convention: original edges, face center to edge center, face center to vertex This approach can be applied to all polyhedra, as every n-gon can be subdivided into 2n coloured triangles:

Faculty of Science The result of applying Goldberg’s (7, 0) operation to this :

Faculty of Science Why does it fit at the edges of the triangle?

The sides of the triangle are mirror axes! Except at the black dots the situation in the decorated polyhedron is like in the tiling!

Faculty of Science Let’s look at another figure from Goldberg’s paper:

No mirror axes – nothing fits!

Faculty of Science The sides of the rectangular Goldberg-triangle are mirror axes if and only if the parameters (k, l) fulfill 0 ∈ {k, l} or k = l:

Faculty of Science With two different triangles it would have worked

rotation by 180 degrees

(5,3)

rotation by 60 degrees

Faculty of Science Note that this triangle is – though equilateral – quite different from the one used by Caspar and Klug!

rotation by 180 degrees

(5,3)

rotation by 60 degrees

Faculty of Science So:

• The method known as Goldberg-Coxeter construction is by Caspar and Klug and different from what Goldberg proposed.

• The method used by Goldberg is appli- cable in a much more general context than that of Caspar and Klug.

Faculty of Science Also the Caspar/Klug (k, l)-operations with 0 6∈ {k, l} and k 6= l do not preserve all symmetries!

Let’s for now stick to operations preserving all symmetries.

Faculty of Science Reminder: Why do rectangular Goldberg triangles fit at the edges?

The sides of the triangle are mirror axes! Let’s start from that! Local symmetry preserving operations:

• Take any 3-connected periodic tiling T in the plane.

• Take any three pairwise intersecting mirror axes forming a right triangle.

• Take the interior as the Goldberg right triangle.

Faculty of Science Reminder: Blue edges are glued to original edges.

An example:

90 o 60 o

30o

Faculty of Science In order to be a good definition, it must:

• encompass all known operations that are considered to be local symmetry preserv- ing operations

• transform polyhedra to polyhedra

• be useful and open new directions of research

Faculty of Science Dual:

to face center

to vertex

to edge center

Faculty of Science Truncate:

Faculty of Science Ambo:

Faculty of Science :

. . . and so on . . .

Faculty of Science Theorem:

If a local symmetry preserving operation is applied to a polyhedron (a plane 3-connected graph), the result is a polyhedron.

Faculty of Science Usefulness and new directions

Faculty of Science In fact we have already seen one useful feature:

Everything you prove for the class of local symmetry preserving operations needs not be proven for each element separately!

E.g. One doesn’t have to prove separately that dual, , ambo, chamfer, quinto, ... produce (3-connected) polyhedra.

Faculty of Science The ratio Ea between the number of edges Eb (or equivalently coloured triangles) after the operation (Ea) and before the operation (Eb) is independent of the polyhedron and is called the inflation factor.

Now it is possible to ask for all operations with a given inflation factor.

Faculty of Science Symmetry preserving operations preserve the group – that is: don’t destroy symmetries – but can also increase it (that is: create new symmetries):

ambo

Can that also happen for polyhedra that are not self-dual and when the operation can not be written as a product of operations involving ambo? For the torus that is possible (inflation factor 9). . .

. . . for polyhedra we don’t know such an example.

Faculty of Science Symmetric polyhedra can be seen as redundant: all you need is the fundamental domain and the group tells you how to distribute the content of the fundamental domain.

A decorated polyhedron is also redundant: all you need is the decoration and the polyhedron tells you how to distribute it.

group ↔ base polyhedron fundamental domain ↔ decoration

Faculty of Science Algorithmic question:

How do you efficiently decide whether a polyhedron is a decoration of a smaller one?

Faculty of Science • Thanks for your attention!

• Are there questions, proposals, or even answers?

Faculty of Science