Monday, July 21, 2025

Completeness of isohedral figure classification

The table below was simplified from the table at this Wikipedia article.

Here is a table of isohedral figures from Wikipedia. Is it proven that all isohedral figures must be one of these things – a Platonic solid, a Catalan solid, or one of these pyramid-stacking “polar” figures?

Faces Face config. Class Name Symmetry Order
4 V33 Platonic tetrahedron Td, [3,3], (*332) 24
4 V33 Platonic tetragonal disphenoid D2d, [2+,2], (2*) 4
4 V33 Platonic rhombic disphenoid D2, [2,2]+, (222) 4
6 V34 Platonic cube Oh, [4,3], (*432) 48
6 V34 Platonic cube D3d, [2+,6] 12
6 V34 Platonic cube (2*3) 12
6 V34 Platonic trigonal trapezohedron D3 6
6 V34 Platonic asymmetric trigonal trapezohedron [2,3]+, (223) 6
8 V43 Platonic octahedron Oh, [4,3], (*432) 48
8 V43 Platonic octahedron D4h,[2,4],(*224) 16
8 V43 Platonic octahedron D2h,[2,2],(*222) 8
8 V43 Platonic square bipyramid D2d,[2+,4],(2*2) 8
8 V43 Platonic rhombic bipyramid D2d,[2+,4],(2*2) 8
8 V43 Platonic square scalenohedron D2d,[2+,4],(2*2) 8
12 V35 Platonic regular dodecahedron Ih, [5,3], (*532) 120
12 V35 Platonic regular dodecahedron Th, [3+,4], (3*2) 24
12 V35 Platonic regular dodecahedron T, [3,3]+, (*332) 12
12 V35 Platonic pyritohedron T, [3,3]+, (*332) 12
12 V35 Platonic tetartoid T, [3,3]+, (*332) 12
20 V53 Platonic regular icosahedron Ih, [5,3], (*532) 120
12 V3.62 Catalan triakis tetrahedron Td, [3,3], (*332) 24
12 V(3.4)2 Catalan rhombic dodecahedron Oh, [4,3], (*432) 48
12 V(3.4)2 Catalan rhombic dodecahedron Td, [3,3], (*332) 24
12 V(3.4)2 Catalan deltoidal dodecahedron Td, [3,3], (*332) 24
24 V3.82 Catalan triakis octahedron Oh, [4,3], (*432) 48
24 V4.62 Catalan tetrakis hexahedron Oh, [4,3], (*432) 48
24 V3.43 Catalan deltoidal icositetrahedron Oh, [4,3], (*432) 48
48 V4.6.8 Catalan disdyakis dodecahedron Oh, [4,3], (*432) 48
24 V34.4 Catalan pentagonal icositetrahedron O, [4,3]+, (432) 24
30 V(3.5)2 Catalan rhombic triacontahedron Ih, [5,3], (*532) 120
60 V3.102 Catalan triakis icosahedron Ih, [5,3], (*532) 120
60 V5.62 Catalan pentakis dodecahedron Ih, [5,3], (*532) 120
60 V3.4.5.4 Catalan deltoidal hexecontahedron Ih, [5,3], (*532) 120
120 V4.6.10 Catalan disdyakis triacontahedron Ih, [5,3], (*532) 120
60 V34.5 Catalan pentagonal hexecontahedron I, [5,3]+, (532) 60
2n V33.n Polar trapezohedron Dnd, [2+,2n], (2*n) 4n
2n V33.n Polar trapezohedron Dn, [2,n]+, (22n) 2n
2n V33.n Polar asymmetric trapezohedron Dn, [2,n]+, (22n) 2n
2n V42.n Polar regular n-bipyramid Dnh, [2,n], (*22n) 4n
2n V42.2n Polar isotoxal 2n-bipyramid Dnh, [2,n], (*22n) 4n
2n V42.2n Polar 2n-scalenohedron Dnd, [2+,2n], (2*n) 4n
4n V42.2n Polar 2n-scalenohedron Dnd, [2+,2n], (2*n) 4n

Short answer

Yes – for convex polyhedra the classification is finished. Branko Grünbaum and Geoffrey C. Shephard completed it in their study of isohedral (face-transitive) tilings of the sphere, and the result is exactly what your table shows:

  • 5 Platonic solids

  • 13 Catalan solids (duals of the Archimedean solids)

  • Three infinite polar families

    • the (general) scalenohedra / isosceles tetrahedra
    • the $n$-bipyramids (duals of prisms)
    • the $n$-trapezohedra (duals of antiprisms)

Together these make up 30 “classes’’: 28 individual solids and 2 one-parameter families whose face-count grows with $n$. MathWorld’s Isohedron page quotes the same total and traces it to the Grünbaum–Shephard work (MathWorld), while Wikipedia summarises the duality-based reasoning – Platonic ↔ Platonic, Archimedean → Catalan, prism → bipyramid, antiprism → trapezohedron (Wikipedia).

What the proof looks like (in outline)

  1. Work on the sphere. Any convex polyhedron gives a tiling of the unit sphere by projecting its faces outward.

  2. Classify all face-transitive spherical tilings. Grünbaum & Shephard (1981) listed every way a single tile can repeat around the sphere under a finite symmetry group.

  3. Lift the tilings back to space. Each tiling corresponds to either

    • a Platonic solid,
    • the dual of a uniform spherical tiling (hence a Catalan solid), or
    • a 2-cap tiling whose dual is a bipyramid or trapezohedron. No other convex embedding exists – that’s the heart of the proof.

Because the argument happens on the sphere, it automatically rules out any stray convex isohedra that look different but would project to one of the known patterns.

Limits of the result

  • If you drop convexity (allow self-intersection or star faces) there are many more isohedra, starting with the four Kepler–Poinsot polyhedra, and no complete list is known.
  • If you go up one dimension (4-polytopes) or down to tilings of the plane, the problem is again open.
  • Even in 3-D, the weaker “all faces merely congruent’’ condition (monohedral) is not classified; several sporadic convex examples are known that are not isohedral (Mathematics Stack Exchange).

So, within the usual confines of convex solids the answer is yes – every isohedron is either Platonic, Catalan, or one of the polar stack-of-pyramids families on your list. Outside that safe zone, the zoo is still very much alive and un-catalogued.

Do you have the citation for the study by Branko Grünbaum and Geoffrey C. Shephard?

Below is the full bibliographic information that most authors cite when they refer to the Grünbaum–Shephard classification of isohedral (face-transitive) spherical tilings—which, in turn, yields the list of convex isohedral polyhedra you asked about.

Ref. Standard citation (APA-style) Notes
[GS 1981a] Grünbaum, B., & Shephard, G. C. (1981). Spherical tilings with transitivity properties. In C. Davis, B. Grünbaum & F. A. Sherk (Eds.), The Geometric Vein: The Coxeter Festschrift (pp. 65-98). Springer-Verlag. https://doi.org/10.1007/978-1-4612-5648-9_4 The definitive, worked-out classification of all isohedral, isotoxal and isogonal tilings of the 2-sphere. Convex isohedral polyhedra correspond to these tilings via radial projection and duality. (SpringerLink)
[GS 1981b] Grünbaum, B., & Shephard, G. C. (1981). Patterns on the 2-sphere. Mathematika, 28(1), 1-35. https://doi.org/10.1112/S0025579300015321 A shorter companion journal article that previews some of the same results in a more compressed form. (londmathsoc.onlinelibrary.wiley.com)

Either reference is usually accepted; the book chapter [GS 1981a] is the most frequently quoted source when people assert that “every convex isohedron is Platonic, Catalan, or one of the three polar families.”

No comments:

Post a Comment