^ abOnly counting polytopes of full rank. There are more regular polytopes of each rank > 1 in higher dimensions.
There are no Euclidean regular star tessellations in any number of dimensions.
1-polytopes
A Coxeter diagram represent mirror "planes" as nodes, and puts a ring around a node if a point is not on the plane. A dion { }, , is a point p and its mirror image point p', and the line segment between them.
There is only one polytope of rank 1 (1-polytope), the closed line segment bounded by its two endpoints. Every realization of this 1-polytope is regular. It has the Schläfli symbol { },[2][3] or a Coxeter diagram with a single ringed node, . Norman Johnson calls it a dion[4] and gives it the Schläfli symbol { }.
Although trivial as a polytope, it appears as the edges of polygons and other higher dimensional polytopes.[5] It is used in the definition of uniform prisms like Schläfli symbol { }×{p}, or Coxeter diagram as a Cartesian product of a line segment and a regular polygon.[6]
2-polytopes (polygons)
The polytopes of rank 2 (2-polytopes) are called polygons. Regular polygons are equilateral and cyclic. A p-gonal regular polygon is represented by Schläfli symbol {p}.
Many sources only consider convex polygons, but star polygons, like the pentagram, when considered, can also be regular. They use the same vertices as the convex forms, but connect in an alternate connectivity which passes around the circle more than once to be completed.
The regular digon {2} can be considered to be a degenerate regular polygon. It can be realized non-degenerately in some non-Euclidean spaces, such as on the surface of a sphere or torus. For example, digon can be realised non-degenerately as a spherical lune. A monogon {1} could also be realised on the sphere as a single point with a great circle through it.[7] However, a monogon is not a valid abstract polytope because its single edge is incident to only one vertex rather than two.
There exist infinitely many regular star polytopes in two dimensions, whose Schläfli symbols consist of rational numbers {n/m}. They are called star polygons and share the same vertex arrangements of the convex regular polygons.
In general, for any natural number n, there are regular n-pointed stars with Schläfli symbols {n/m} for all m such that m < n/2 (strictly speaking {n/m} = {n/(n − m)}) and m and n are coprime (as such, all stellations of a polygon with a prime number of sides will be regular stars). Symbols where m and n are not coprime may be used to represent compound polygons.
Star polygons that can only exist as spherical tilings, similarly to the monogon and digon, may exist (for example: {3/2}, {5/3}, {5/4}, {7/4}, {9/5}), however these have not been studied in detail.
There also exist failed star polygons, such as the piangle, which do not cover the surface of a circle finitely many times.[8]
Skew polygons
In addition to the planar regular polygons there are infinitely many regular skew polygons. Skew polygons can be created via the blending operation.
The blend of two polygons P and Q, written P#Q, can be constructed as follows:
take the cartesian product of their vertices VP × VQ.
add edges (p0 × q0, p1 × q1) where (p0, p1) is an edge of P and (q0, q1) is an edge of Q.
select an arbitrary connected component of the result.
Alternatively, the blend is the polygon ⟨ρ0σ0, ρ1σ1⟩ where ρ and σ are the generating mirrors of P and Q placed in orthogonal subspaces.[9]
The blending operation is commutative, associative and idempotent.
Every regular skew polygon can be expressed as the blend of a unique[i] set of planar polygons.[9] If P and Q share no factors then Dim(P#Q) = Dim(P) + Dim(Q).
In 3 space
The regular finite polygons in 3 dimensions are exactly the blends of the planar polygons (dimension 2) with the digon (dimension 1). They have vertices corresponding to a prism ({n/m}#{} where n is odd) or an antiprism ({n/m}#{} where n is even). All polygons in 3 space have an even number of vertices and edges.
Several of these appear as the Petrie polygons of regular polyhedra.
In 4 space
The regular finite polygons in 4 dimensions are exactly the polygons formed as a blend of two distinct planar polygons. They have vertices lying on a Clifford torus and related by a Clifford displacement. Unlike 3-dimensional polygons, skew polygons on double rotations can include an odd-number of sides.
A regular polyhedron with Schläfli symbol{p, q}, Coxeter diagrams , has a regular face type {p}, and regular vertex figure{q}.
A vertex figure (of a polyhedron) is a polygon, seen by connecting those vertices which are one edge away from a given vertex. For regular polyhedra, this vertex figure is always a regular (and planar) polygon.
Existence of a regular polyhedron {p, q} is constrained by an inequality, related to the vertex figure's angle defect:
By enumerating the permutations, we find five convex forms, four star forms and three plane tilings, all with polygons {p} and {q} limited to: {3}, {4}, {5}, {5/2}, and {6}.
Beyond Euclidean space, there is an infinite set of regular hyperbolic tilings.
As spherical tilings, these star forms overlap the sphere multiple times, called its density, being 3 or 7 for these forms. The tiling images show a single spherical polygon face in yellow.
There are infinitely many failed star polyhedra. These are also spherical tilings with star polygons in their Schläfli symbols, but they do not cover a sphere finitely many times. Some examples are {5/2,4}, {5/2,9}, {7/2,3}, {5/2,5/2}, {7/2,7/3}, {4,5/2}, and {3,7/3}.
Skew polyhedra
This section needs expansion. You can help by adding to it. (January 2024)
For 4-dimensional skew polyhedra, Coxeter offered a modified Schläfli symbol {l,m|n} for these figures, with {l,m} implying the vertex figure, m l-gons around a vertex, and n-gonal holes. Their vertex figures are skew polygons, zig-zagging between two planes.
The regular skew polyhedra, represented by {l,m|n}, follow this equation:
Regular 4-polytopes with Schläfli symbol have cells of type , faces of type , edge figures
, and vertex figures .
A vertex figure (of a 4-polytope) is a polyhedron, seen by the arrangement of neighboring vertices around a given vertex. For regular 4-polytopes, this vertex figure is a regular polyhedron.
An edge figure is a polygon, seen by the arrangement of faces around an edge. For regular 4-polytopes, this edge figure will always be a regular polygon.
The existence of a regular 4-polytope is constrained by the existence of the regular polyhedra . A suggested name for 4-polytopes is "polychoron".[11]
Each will exist in a space dependent upon this expression:
: Hyperspherical 3-space honeycomb or 4-polytope
: Euclidean 3-space honeycomb
: Hyperbolic 3-space honeycomb
These constraints allow for 21 forms: 6 are convex, 10 are nonconvex, one is a Euclidean 3-space honeycomb, and 4 are hyperbolic honeycombs.
Regular di-4-topes (2 facets) include: {3,3,2}, {3,4,2}, {4,3,2}, {5,3,2}, {3,5,2}, {p,2,2}, and their hoso-4-topeduals (2 vertices): {2,3,3}, {2,4,3}, {2,3,4}, {2,3,5}, {2,5,3}, {2,2,p}. 4-polytopes of the form {2,p,2} are the same as {2,2,p}. There are also the cases {p,2,q} which have dihedral cells and hosohedral vertex figures.
Ludwig Schläfli found four of them and skipped the last six because he would not allow forms that failed the Euler characteristic on cells or vertex figures (for zero-hole tori: F+V−E=2). Edmund Hess (1843–1903) completed the full list of ten in his German book Einleitung in die Lehre von der Kugelteilung mit besonderer Berücksichtigung ihrer Anwendung auf die Theorie der Gleichflächigen und der gleicheckigen Polyeder (1883)[1].
There are 4 failed potential regular star 4-polytopes permutations: {3,5/2,3}, {4,3,5/2}, {5/2,3,4}, {5/2,3,5/2}. Their cells and vertex figures exist, but they do not cover a hypersphere with a finite number of repetitions.
Skew 4-polytopes
This section needs expansion. You can help by adding to it. (January 2024)
In addition to the 16 planar 4-polytopes above there are 18 finite skew polytopes.[12] One of these is obtained as the Petrial of the tesseract, and the other 17 can be formed by applying the kappa operation to the planar polytopes and the Petrial of the tesseract.
Ranks 5 and higher
5-polytopes can be given the symbol where is the 4-face type, is the cell type, is the face type, and is the face figure, is the edge figure, and is the vertex figure.
A vertex figure (of a 5-polytope) is a 4-polytope, seen by the arrangement of neighboring vertices to each vertex.
An edge figure (of a 5-polytope) is a polyhedron, seen by the arrangement of faces around each edge.
A face figure (of a 5-polytope) is a polygon, seen by the arrangement of cells around each face.
A regular 5-polytope exists only if and are regular 4-polytopes.
The space it fits in is based on the expression:
: Spherical 4-space tessellation or 5-space polytope
: Euclidean 4-space tessellation
: hyperbolic 4-space tessellation
Enumeration of these constraints produce 3 convex polytopes, no star polytopes, 3 tessellations of Euclidean 4-space, and 5 tessellations of paracompact hyperbolic 4-space. The only non-convex regular polytopes for ranks 5 and higher are skews.
Convex
In dimensions 5 and higher, there are only three kinds of convex regular polytopes.[13]
There are also improper cases where some numbers in the Schläfli symbol are 2. For example, {p,q,r,...2} is an improper regular spherical polytope whenever {p,q,r...} is a regular spherical polytope, and {2,...p,q,r} is an improper regular spherical polytope whenever {...p,q,r} is a regular spherical polytope. Such polytopes may also be used as facets, yielding forms such as {p,q,...2...y,z}.
There are no regular star polytopes of rank 5 or higher, with the exception of degenerate polytopes created by the star product of lower rank star polytopes. e.g. hosotopes and ditopes.
Regular projective polytopes
A projective regular (n+1)-polytope exists when an original regular n-spherical tessellation, {p,q,...}, is centrally symmetric. Such a polytope is named hemi-{p,q,...}, and contain half as many elements. Coxeter gives a symbol {p,q,...}/2, while McMullen writes {p,q,...}h/2 with h as the coxeter number.[14]
Even-sided regular polygons have hemi-2n-gon projective polygons, {2p}/2.
5 of 6 convex regular 4-polytopes are centrally symmetric generating projective 4-polytopes. The 3 special cases are hemi-24-cell, hemi-600-cell, and hemi-120-cell.
Only 2 of 3 regular spherical polytopes are centrally symmetric for ranks 5 or higher. The corresponding regular projective polytopes are the hemi versions of the regular hypercube and orthoplex. They are tabulated below for rank 5, for example:
An apeirotope or infinite polytope is a polytope which has infinitely many facets. An n-apeirotope is an infinite n-polytope: a 2-apeirotope or apeirogon is an infinite polygon, a 3-apeirotope or apeirohedron is an infinite polyhedron, etc.
There are two main geometric classes of apeirotope:[15]
Regular honeycombs in n dimensions, which completely fill an n-dimensional space.
Regular skew apeirotopes, comprising an n-dimensional manifold in a higher space.
2-apeirotopes (apeirogons)
The straight apeirogon is a regular tessellation of the line, subdividing it into infinitely many equal segments. It has infinitely many vertices and edges. Its Schläfli symbol is {∞}, and Coxeter diagram .
......
It exists as the limit of the p-gon as p tends to infinity, as follows:
Apeirogons in the hyperbolic plane, most notably the regular apeirogon, {∞}, can have a curvature just like finite polygons of the Euclidean plane, with the vertices circumscribed by horocycles or hypercycles rather than circles.
Regular apeirogons that are scaled to converge at infinity have the symbol {∞} and exist on horocycles, while more generally they can exist on hypercycles.
Above are two regular hyperbolic apeirogons in the Poincaré disk model, the right one shows perpendicular reflection lines of divergent fundamental domains, separated by length λ.
Skew apeirogons
A skew apeirogon in two dimensions forms a zig-zag line in the plane. If the zig-zag is even and symmetrical, then the apeirogon is regular.
Skew apeirogons can be constructed in any number of dimensions. In three dimensions, a regular skew apeirogon traces out a helical spiral and may be either left- or right-handed.
2 dimensions
3 dimensions
Zig-zag apeirogon
Helix apeirogon
3-apeirotopes (apeirohedra)
Euclidean tilings
There are three regular tessellations of the plane.
There are two improper regular tilings: {∞,2}, an apeirogonal dihedron, made from two apeirogons, each filling half the plane; and secondly, its dual, {2,∞}, an apeirogonal hosohedron, seen as an infinite set of parallel lines.
There are no regular plane tilings of star polygons. There are many enumerations that fit in the plane (1/p + 1/q = 1/2), like {8/3,8}, {10/3,5}, {5/2,10}, {12/5,12}, etc., but none repeat periodically.
Hyperbolic tilings
Tessellations of hyperbolic 2-space are hyperbolic tilings. There are infinitely many regular tilings in H2. As stated above, every positive integer pair {p,q} such that 1/p + 1/q < 1/2 gives a hyperbolic tiling. In fact, for the general Schwarz triangle (p, q, r) the same holds true for 1/p + 1/q + 1/r < 1.
There are a number of different ways to display the hyperbolic plane, including the Poincaré disc model which maps the plane into a circle, as shown below. It should be recognized that all of the polygon faces in the tilings below are equal-sized and only appear to get smaller near the edges due to the projection applied, very similar to the effect of a camera fisheye lens.
There are infinitely many flat regular 3-apeirotopes (apeirohedra) as regular tilings of the hyperbolic plane, of the form {p,q}, with p+q<pq/2.
The tilings {p, ∞} have ideal vertices, on the edge of the Poincaré disc model. Their duals {∞, p} have ideal apeirogonal faces, meaning that they are inscribed in horocycles. One could go further (as is done in the table above) and find tilings with ultra-ideal vertices, outside the Poincaré disc, which are dual to tiles inscribed in hypercycles; in what is symbolised {p, iπ/λ} above, infinitely many tiles still fit around each ultra-ideal vertex.[16] (Parallel lines in extended hyperbolic space meet at an ideal point; ultraparallel lines meet at an ultra-ideal point.)[17]
Hyperbolic star-tilings
There are 2 infinite forms of hyperbolic tilings whose faces or vertex figures are star polygons: {m/2, m} and their duals {m, m/2} with m = 7, 9, 11, ....[18] The {m/2, m} tilings are stellations of the {m, 3} tilings while the {m, m/2} dual tilings are facetings of the {3, m} tilings and greatenings[ii] of the {m, 3} tilings.
The patterns {m/2, m} and {m, m/2} continue for odd m < 7 as polyhedra: when m = 5, we obtain the small stellated dodecahedron and great dodecahedron,[18] and when m = 3, the case degenerates to a tetrahedron. The other two Kepler–Poinsot polyhedra (the great stellated dodecahedron and great icosahedron) do not have regular hyperbolic tiling analogues. If m is even, depending on how we choose to define {m/2}, we can either obtain degenerate double covers of other tilings or compound tilings.
Allowing for skew faces, there are 24 regular apeirohedra in Euclidean 3-space.[23] These include 12 apeirhedra created by blends with the Euclidean apeirohedra, and 12 pure apeirohedra, including the 3 above, which cannot be expressed as a non-trivial blend.
Those pure apeirohedra are:
{4,6|4}, the mucube
{∞,6}4,4, the Petrial of the mucube
{6,6|3}, the mutetrahedron
{∞,6}6,3, the Petrial of the mutetrahedron
{6,4|4}, the muoctahedron
{∞,4}6,4, the Petrial of the muoctahedron
{6,6}4, the halving of the mucube
{4,6}6, the Petrial of {6,6}4
{∞,4}·,*3, the skewing of the muoctahedron
{6,4}6, the Petrial of {∞,4}·,*3
{∞,3}(a)
{∞,3}(b)
Skew apeirohedra in hyperbolic 3-space
There are 31 regular skew apeirohedra with convex faces in hyperbolic 3-space with compact or paracompact symmetry:[24]
14 are compact: {8,10|3}, {10,8|3}, {10,4|3}, {4,10|3}, {6,4|5}, {4,6|5}, {10,6|3}, {6,10|3}, {8,8|3}, {6,6|4}, {10,10|3},{6,6|5}, {8,6|3}, and {6,8|3}.
There are six improper regular tessellations, pairs based on the three regular Euclidean tilings. Their cells and vertex figures are all regular hosohedra {2,n}, dihedra, {n,2}, and Euclidean tilings. These improper regular tilings are constructionally related to prismatic uniform honeycombs by truncation operations. They are higher-dimensional analogues of the order-2 apeirogonal tiling and apeirogonal hosohedron.
Noncompact solutions exist as Lorentzian Coxeter groups, and can be visualized with open domains in hyperbolic space (the fundamental tetrahedron having ultra-ideal vertices). All honeycombs with hyperbolic cells or vertex figures and do not have 2 in their Schläfli symbol are noncompact.
There are no regular hyperbolic star-honeycombs in H3: all forms with a regular star polyhedron as cell, vertex figure or both end up being spherical.
Ideal vertices now appear when the vertex figure is a Euclidean tiling, becoming inscribable in a horosphere rather than a sphere. They are dual to ideal cells (Euclidean tilings rather than finite polyhedra). As the last number in the Schläfli symbol rises further, the vertex figure becomes hyperbolic, and vertices become ultra-ideal (so the edges do not meet within hyperbolic space). In honeycombs {p, q, ∞} the edges intersect the Poincaré ball only in one ideal point; the rest of the edge has become ultra-ideal. Continuing further would lead to edges that are completely ultra-ideal, both for the honeycomb and for the fundamental simplex (though still infinitely many {p, q} would meet at such edges). In general, when the last number of the Schläfli symbol becomes ∞, faces of codimension two intersect the Poincaré hyperball only in one ideal point.[16]
5-apeirotopes
Tessellations of Euclidean 4-space
There are three kinds of infinite regular tessellations (honeycombs) that can tessellate Euclidean four-dimensional space:
Noncompact solutions exist as Lorentzian Coxeter groups, and can be visualized with open domains in hyperbolic space (the fundamental 5-cell having some parts inaccessible beyond infinity). All honeycombs which are not shown in the set of tables below and do not have 2 in their Schläfli symbol are noncompact.
There is only one flat regular honeycomb of Euclidean 5-space: (previously listed above as tessellations)[25]
{4,3,3,3,4}
There are five flat regular regular honeycombs of hyperbolic 5-space, all paracompact: (previously listed above as tessellations)[18]
{3,3,3,4,3}, {3,4,3,3,3}, {3,3,4,3,3}, {3,4,3,3,4}, and {4,3,3,4,3}
Tessellations of Euclidean 5-space
The hypercubic honeycomb is the only family of regular honeycombs that can tessellate each dimension, five or higher, formed by hypercube facets, four around every ridge.
In E5, there are also the improper cases {4,3,3,4,2}, {2,4,3,3,4}, {3,3,4,3,2}, {2,3,3,4,3}, {3,4,3,3,2}, and {2,3,4,3,3}. In En, {4,3n−3,4,2} and {2,4,3n−3,4} are always improper Euclidean tessellations.
Tessellations of hyperbolic 5-space
There are 5 regular honeycombs in H5, all paracompact, which include infinite (Euclidean) facets or vertex figures: {3,4,3,3,3}, {3,3,4,3,3}, {3,3,3,4,3}, {3,4,3,3,4}, and {4,3,3,4,3}.
There are no compact regular tessellations of hyperbolic space of dimension 5 or higher and no paracompact regular tessellations in hyperbolic space of dimension 6 or higher.
Since there are no regular star n-polytopes for n ≥ 5, that could be potential cells or vertex figures, there are no more hyperbolic star honeycombs in Hn for n ≥ 5.
Apeirotopes of rank 7 or more
Tessellations of hyperbolic 6-space and higher
There are no regular compact or paracompact tessellations of hyperbolic space of dimension 6 or higher. However, any Schläfli symbol of the form {p,q,r,s,...} not covered above (p,q,r,s,... natural numbers above 2, or infinity) will form a noncompact tessellation of hyperbolic n-space.[16]
Abstract polytopes
The abstract polytopes arose out of an attempt to study polytopes apart from the geometrical space they are embedded in. They include the tessellations of spherical, Euclidean and hyperbolic space, and of other manifolds. There are infinitely many of every rank greater than 1. See this atlas for a sample. Some notable examples of abstract regular polytopes that do not appear elsewhere in this list are the 11-cell, {3,5,3}, and the 57-cell, {5,3,5}, which have regular projective polyhedra as cells and vertex figures.
The elements of an abstract polyhedron are its body (the maximal element), its faces, edges, vertices and the null polytope or empty set. These abstract elements can be mapped into ordinary space or realised as geometrical figures. Some abstract polyhedra have well-formed or faithful realisations, others do not. A flag is a connected set of elements of each rank - for a polyhedron that is the body, a face, an edge of the face, a vertex of the edge, and the null polytope. An abstract polytope is said to be regular if its combinatorial symmetries are transitive on its flags - that is to say, that any flag can be mapped onto any other under a symmetry of the polyhedron. Abstract regular polytopes remain an active area of research.
Five such regular abstract polyhedra, which can not be realised faithfully and symmetrically, were identified by H. S. M. Coxeter in his book Regular Polytopes (1977) and again by J. M. Wills in his paper "The combinatorially regular polyhedra of index 2" (1987).[27] They are all topologically equivalent to toroids. Their construction, by arranging n faces around each vertex, can be repeated indefinitely as tilings of the hyperbolic plane. In the diagrams below, the hyperbolic tiling images have colors corresponding to those of the polyhedra images.
^In a classification advanced by Conway & adopted by Coxeter,[a]stellation refers to extension of edges, and greatening to extension of faces; the term aggrandizement is given for extension of cells (of polychora), though it appears to be less-commonly used.[b]
Subnotes
^ Coxeter, H. M. S. (1975). Regular Complex Polytopes (1st ed.). Cambridge University Press. pp. 46–7. ISBN9780521201254.
^Abstracts(PDF). Convex and Abstract Polytopes (May 19–21, 2005) and Polytopes Day in Calgary (May 22, 2005). Archived(PDF) from the original on 2014-11-29. Retrieved 2015-02-02.
^Conway, John H.; Burgiel, Heidi; Goodman-Strauss, Chaim (2008). "Chapter 23: Objects with Primary Symmetry, Infinite Platonic Polyhedra". The Symmetries of Things. Taylor & Francis. pp. 333–335. ISBN978-1-568-81220-5.
Coxeter, H. S. M. (1999), "Chapter 10: Regular Honeycombs in Hyperbolic Space", The Beauty of Geometry: Twelve Essays, Mineola, NY: Dover Publications, Inc., pp. 199–214, ISBN0-486-40919-8, LCCN99035678, MR1717154. See in particular Summary Tables II, III, IV, V, pp. 212–213.
Johnson, Norman W. (2012), "Regular inversive polytopes"(PDF), International Conference on Mathematics of Distances and Applications (July 2–5, 2012, Varna, Bulgaria), pp. 85–95 Paper 27, archived(PDF) from the original on 2017-08-08, retrieved 2015-01-12
McMullen, Peter (2018), "New Regular Compounds of 4-Polytopes", New Trends in Intuitive Geometry, Bolyai Society Mathematical Studies, vol. 27, pp. 307–320, doi:10.1007/978-3-662-57413-3_12, ISBN978-3-662-57412-6.
Sommerville, D. M. Y. (1958), An Introduction to the Geometry of n Dimensions, New York: Dover Publications, Inc., MR0100239. Reprint of 1930 ed., published by E. P. Dutton. See in particular Chapter X: The Regular Polytopes.