3

Rhombicuboctahedron and Pseudo-rhombicuboctahedron

In An enduring error, Branko Grünbaum discusses the question of defining the Archimedean solids, noting the case of the 'pseudo-rhombicuboctahedron', sometimes called Miller's solid and shown on the right above, as opposed to the rhombicuboctahedron, one of the standard $13$ Archimedean solids and shown on the left (the difference being that the triangles and squares in the lower portion of the body have been 'shifted around' by one place). What has caused an 'enduring error' is a failure to distinguish properly between two criteria, namely:

Local criterion: All faces are regular polygons, and the cyclic arrangement of the faces around each vertex is the same.

and

Global criterion: All faces are regular polygons, and all vertices form one orbit under isometric symmetries of the polyhedron.

The rhombicuboctahedron satisfies both the local and global criterion, while the pseudo-rhombicuboctahedron only satisfies the local criterion.

At the end of the article, Grünbaum indicates that whether there are any other such solids, i.e. convex solids which satisfy the local but not the global criteria, is an open question, conjecturing that there aren't any such solids but that 'a proof of this is probably quite complicated.'

My question is that, given that Grünbaum's paper was written in 2009, does anyone know if any progress has been made in determining whether there are any such solids (Wikipedia indicates that there hasn't)? Assuming there hasn't been, I thought it would be interesting to hear any thoughts on why it's so difficult to show that there aren't any further such solids.

(I note that if we allow self-intersecting polyhedra then there is the pseudo-great rhombicuboctahedron, corresponding to the non-convex great rhombicuboctahedron.)

Parcly Taxel
  • 103,344
John1970
  • 405
  • 3
    The problem has long been solved. The Johnson solids are all those polyhedra with regular polygon faces that don't belong to the regular, semiregular, prism or antiprism cases – and Zalgaller proved in 1969 that there are only $92$ Johnson solids. The pseudorhombicuboctahedron is one of the Johnson solids, and no other Johnson solid is like it. Am I missing anything? – Parcly Taxel Sep 24 '22 at 13:45
  • It seems that Grünbaum never did a literature review before writing the article. – Parcly Taxel Sep 24 '22 at 13:46
  • Thanks - that seems to answer it! Looks like the Wikipedia page should be corrected too then, i.e. link? Or is the question still open if non-convex or self-intersecting polyhedra are allowed, i.e. given that the Johnson solids are defined as convex? – John1970 Sep 24 '22 at 14:47
  • Well it should still be open for nonconvex polyhedra, notwithstanding references I don't know of. Let me promote my comment to an answer and then you can upvote/accept. – Parcly Taxel Sep 24 '22 at 14:50

1 Answers1

3

Given the definitions of local and global uniformity, the problem has long been solved – at least for the convex case.

The Johnson solids are the strictly convex polyhedra with regular polygon faces that are not vertex-transitive. Zalgaller proved in 1969 that there are only $92$ of them, and only Miller's solid out of them satisfies the local but not the global uniformity criterion.

Parcly Taxel
  • 103,344