ARS MATHEMATICA CONTEMPORANEA
ISSN 1855-3966 (printed edn.), ISSN 1855-3974 (electronic edn.)
ARS MATHEMATICA CONTEMPORANEA 15 (2018) 67-79
https://doi.org/10.26493/1855-3974.1319.375
(Also available at http://amc-journal.eu)
Rank 4 toroidal hypertopes
Eric Ens *
Department of Mathematics, York University, Canada
Received 8 February 2017, accepted 22 September 2017, published online 17 January 2018
We classify the regular toroidal hypertopes of rank 4. Their automorphism groups are the quotients of infinite irreducible Coxeter groups of euclidean type with 4 generators. We also prove that there are no toroidal chiral hypertopes of rank 4.
Keywords: Regularity, chirality, toroidal, thin geometries, hypermaps, abstract polytopes. Math. Subj. Class.: 17B37, 15A21
1 Introduction
A toroidal polytope is an abstract polytope that can be seen as a tessellation on a torus. By abstract polytope we mean a combinatorial structure resembling a classical polytope described by incidence relationships. Highly symmetric types of these polytopes are well known and understood, in particular the regular and chiral toroidal polytopes have been classified for rank 3 by Coxeter in 1948 [5], see also [6], and for any rank by McMullen and Schulte [10]. Regular toroidal polytopes and also regular toroidal hypertopes, which we define below, are strongly related to a special class of Coxeter groups, the infinite irreducible Coxeter groups of euclidean type which are also known as affine Coxeter groups (see, for example [11, page 73]). The symmetry groups of regular tessellations of euclidean space are precisely the affine Coxeter groups with string diagrams (see [11, Theorem 3B5]).
When we talk about a tessellation we mean, informally, a locally finite collection of polytopes which cover en in a face-to-face manner. A toroidal polytope can then be seen as a "quotient" of a tessellation by linearly independent translations. For a precise definition of a toroidal polytope see [8]. The concept of a hypertope has recently been introduced by Fernandes, Leemans and Weiss (see [7]). A hypertope can be seen as a generalization of
*Here I would like to acknowledge Maria-Elisa Fernandes, Dimitri Leemans, Egon Schulte and Asia Ivic Weiss for introducing the problem to me during the BIRS "Discrete Geometry and Symmetry" workshop on February 8-13, 2015. I also want to acknowledge Dimitri Leemans for his help with Magma [1], which was used to confirm the intersection conditions for Theorems 4.2 and 5.1.
E-mail address: ericens@mathstat.yorku.ca (Eric Ens)
Abstract
©® This work is licensed under https://creativecommons.Org/licenses/by/3.0/
68
Ars Math. Contemp. 15 (2018) 113-126
a polytope. Or, from another perspective, as a generalization of a hypermap. For more infomation on hypermaps see [4]. In this paper we will classify the rank 4 regular toroidal hypertopes.
Each affine Coxeter group in rank 4 (which are usually denoted by C3, B3 and A3), as we shall see, can be associated with the group C3 = [4,3,4], the symmetry group of the cubic tessellation of E3. The Coxeter Complex, denoted by C, of C3 can be seen as the simplicial complex obtained by the barycentric subdivision of the cubic tessellation {4, 3,4}. The Coxeter complex for the other two rank 4 affine Coxeter groups can be obtained by doubling the rank 3 simplicies for B3 and quadrupling them for A3. For details on the construction of C see [9, Section 6.5] or [11, Section 3B]. We note that C partitions E3.
A regular toroidal hypertope (see Section 2 for a precise definition) can be seen as a quotient C/A/ by a normal subgroup of translations, denoted A/ where I represents a generating set identifying the normal subgroup. In particular the quotient induced by a normal subgroup of translations in the string affine Coxeter group C3 yields the three families of regular rank 4 toroidal polytopes, while the other two affine Coxeter groups with non-string diagrams do not yield regular polytopes, but as we shall see below, regular hypertopes.
2 C-groups and hypertopes
Details of the concepts we review here are given in [7] and [11]. A C-group of rank p is a pair (G, S) such that G is a group and S = {r0,..., rp-1} is a generating set of involutions of G that satisfy the following property:
VI, J C {0,... ,p - 1}, (r* : i e I) n (rj : j e J) = (rk : k e I n J).
This is known as the intersection property which will be referred to later.
A subgroup of G generated by a subset of S is called a parabolic subgroup. A parabolic subgroup generated by a single element of S is called minimal and a parabolic subgroup generated by all but one element of S is called maximal. For J C {0,..., p -1}, we define Gj := (rj: j e J) and Gi := (rj: rj e S, rj = r*).
A C-group is a string C-Group if (rirj-)2 = 1G for all i, j with |i - j| > 1. A Coxeter diagram C(G, S) of a C-group (G, S) is a graph whose vertex set is S and two vertices, r* and rj are joined by an edge labelled by o(rirj), the order of rirj. We use the convention that if an edge is labeled 2 it is dropped and not labeled if the order of the product of the corresponding generators is 3. Thus the Coxeter diagram of a string C-group is a string.
Affine Coxeter groups are C-groups and those with string diagrams are associated with toroidal polytopes. Hypertopes are generalizations of polytopes and we can, however, find toroidal hypertopes whose automorphism groups are quotients of any affine Coxeter group. We start with the definition of an incidence system.
Definition 2.1. An incidence system r := (X, *, t, I) is a 4-tuple such that
•	X is a set whose elements are called elements of r;
•	I is a set whose elements are called types of r;
•	t: X ^ I is a type function that associates to each element x e X of r a type
t(x) e I;
E. Ens: Rank 4 toroidal hypertopes
69
•	* is a binary relation of X called incidence, that is reflexive, symmetric and such that
for all x, y G X, if x * y and t(x) = t(y) then x = y.
A flag is a set of pairwise incident elements of r and the type of a flag F is {t(x): x G F}. A chamber is a flag of type I. An element x is said to be incident to a flag F when x is incident to all elements of F and we write x * F.
Definition 2.2. An incidence geometry is an incidence system r where every flag is contained in a chamber. The rank of r is the cardinality of I.
Let r := (X, *, t, I) be an incidence system and F a flag of r. The residue of F in r is the incidence system rF := (XF, *F,tF,IF) where
•	XF := (x G F: x * F, x / F);
•	If := I\t(F);
•	tF and *F are the restrictions of t and * to XF and IF.
If each residue of rank at least 2 of r has a connected incidence graph then r is said to be residually connected. r is thin when every residue of rank 1 contains exactly 2 elements.
Furthermore, r is chamber-connected when for each pair of chambers C and C', there exists a sequence of chambers C =: C0, Ci,..., Cn := C such that |C n Ci+11 = |I| — 1 (here we say that C and Ci+1 are adjacent). An incidence system is strongly chamber-connected when all of its residues of rank at least 2 are chamber-connected.
Proposition 2.3 ([7, Proposition 2.1]). Let r be a thin incidence geometry. Then r is residually connected if and only if r is strongly chamber-connected.
A hypertope is a strongly chamber-connected thin incidence geometry. To reinforce the relationship between polytopes and hypertopes we will sometimes refer to the elements of a hypertope r as hyperfaces of r, and elements of type I as hyperfaces of type I.
Let r := (X, *,t, I) be an incidence system. An automorphism of r is a mapping
a: (X, I) ^ (X, I) : (x,t(x)) ^ (a(x),t(x)) where
•	a is a bijection on X inducing a bijection on I;
•	for each x, y G X, x * y if and only if a(x) * a(y);
•	for each x,y G X, t(x) = t(y) if and only if t(a(x)) = t(a(y)).
An automorphism a is type-preserving when, for each x G X, t(a(x)) = t(x). We denote by Aut(r) the group of automorphisms of r and by Aut/(r) is the group of type-preserving automorphisms of r.
An incidence system r is flag transitive if Aut/(r) is transitive on all flags of type J for each J C I .It is chamber-transitive if Aut/(r) is transitive on all chambers of r. Furthermore, it is regular if the action of Aut/(r) is semi-regular and transitive.
Proposition 2.4 ([7, Proposition 2.2]). Let r be an incidence geometry. r is chamber-transitive if and only if it is flag-transtive.
A regular hypertope is a flag transitive hypertope. We note that every abstract regular polytope is a regular hypertope. The last concept we introduce here before we construct all rank 4 regular toroidal hypertopes is that of coset geometries.
70
Ars Math. Contemp. 15 (2018) 113-126
Proposition 2.5 ([14]). Let p be a positive integer and I := {1,... ,p} a finite set. Let G be a group together with a family of subgroups (Gi)ieI, X the set consisting of all cosets Gig, g G G,i g I and t: X ^ I defined by t(Gig) = i. Define an incidence relation * on X x X by:
Gig2 * Gj g2 if and only if Gig1 n Gj g2 is non-empty in G.
Then the 4-tuple r := (X, *, t, I) is an incidence system having a chamber. Moreover, the group G acts by right multiplication as a group of type-preserving automorphisms of r. Finally, the group G is transitive on the flags of rank less than 3.
Whenever r is constructed as in the above proposition it is written as r(G; (Gi)ieI) and if it is an incidence geometry it is called a coset geometry. If G acts transitively on all chambers of r (thus also flags of any type) we say that G is flag transitive on r or that r is flag transitive.
Now we note that we can construct a coset geometry r(G; (Gi)ieI) using a C-group (G, S) or rank p by setting Gi := (rj : rj G S,j G I\{i}) for all i G I := {0,... ,p - 1}.
We introduce the following proposition which lets us know that constructions we use produce regular hypertopes.
Proposition 2.6 ([7, Theorem 4.6]). Let (G, {r0,..., rp-1}) be a C-group of rank p and let r := r(G;(Gi)i£/) with Gi := (rj : rj G S,j G I\{i}) for all i G I := {0,... ,p-1}. If r is flag transitive, then r is a regular hypertope.
Henceforth, we restrict our considerations to rank 4. Let G = (r0,r1,r2,r3) be an affine Coxeter group where each ri is a reflection through an associated affine hyperplane, Hi. Let C be the Coxeter complex of G formed from the hyperplanes His. Here r1,r2 and r3 will stabilize a point which, without loss of generality, can be assumed to be the origin o in E3. Then the maximal parabolic subgroup G0 is a finite crystallographic subgroup, which is a group that leaves a central point fixed. For details, see [3, pages 108-109]. The normal vectors to the reflection planes of the generators of G0 are called the fundamental roots. The images of the fundamental roots under G0 form a root system for G0.
The lattice, A, generated by the root system is called the root lattice, and the fundamental roots form the integral basis for A. The region enclosed by the fundamental roots is called the fundamental region. This lattice gives us (and can be identified with) the translation subgroup T < G generated by the root lattice of G0, note that G = G0 x T [3]. For convenience we identify the translations with its vectors in addition a lattice also corresponds with its generating translation.
If I is a set of linearly independent translations in T, and let TI < T be the subgroup generated by I. Then the sublattice A/ < A is the lattice induced by oTi, the orbit of the origin under TI. We note that C is a regular hypertope and each simplex in C represents a chamber where each vertex of the simplex is an element of a different type. In rank 4, when the quotient C/AI is a regular hypertope, we say it is a regular toroidal hypertope of rank 4. C/ AI is a regular hypertope (and thus a regular toroidal hypertope) when AI is large enough to ensure the corresponding group satisfies the intersection condition and when AI invariant under G0, i.e. riAIri = AI for i = 1,2,3. It is important to note that, in addition to denoting a lattice, AI is also denotes a set of vectors as well as a translation subgroup of G along those vectors. If I consists of all permutations and changes in sign of the coordinates of some vector s then we will write As.
E. Ens: Rank 4 toroidal hypertopes
71
3 Toroidal polytopes constructed from the group C3 = [4, 3,4]
We begin, necessarily, with generating regular toroidal hypertopes (or, in this case, poly-topes) whose automorphism groups are quotients of the group C3, the affine Coxeter Group [4, 3, 4]. As generators of [4, 3, 4] we take pi, p2, p3 to be reflections in the hyperplanes with normal vectors (1, -1, 0), (0,1, -1), (0, 0,1) respectively, and po the reflection in the plane through (1/2, 0, 0) with normal vector (1, 0, 0) (see Figure 1). Then,
(x,y,z)Po = (1 - x,y,z)
(x,y,z)pi = (y,x, z),	(3.1)
(x,y,z)P2 = O^y^ (x,y,z)P3 = O^ -z).
Figure 1: Fundamental simplex of [4, 3, 4].
In this case, the construction described in Section 2 will yield the regular polytopes since [4, 3, 4] is a string group. We denote by t the corresponding tessellation {4, 3, 4} of the Euclidean plane by cubes and by T it's full translation subgroup, where T is generated by the usual basis vectors, T = ((1,0,0), (0,1,0), (0,0,1)).
Let Hi be the planes fixed by pi. The simplex bounded by the reflection planes Hi is a fundamental simplex of [4,3,4] and is denoted e, it is a simplex in the Coxeter complex of C3. Let Fi be the vertex of the fundamental simplex not on Hi then {F0, F, F2, F3} represents a flag of t, and the set of all j-faces of t = {4,3, 4} is represented by the orbit of Fj under C3.
The regular polytope which results from factoring the regular tesselation t = {4, 3, 4} by a subgroup A of T which is normal in [4, 3, 4], is denoted by t/A (as above).
We let As be the translation subgroup (or lattice) generated by the vector s and its images under the stabilization of the origin in [4, 3, 4] and hence under permutations and changes of sign of its coordinates. The regular polytope t/As is denoted by {4, 3, 4}s := {4, 3, 4}/As and the corresponding group [4, 3, 4]/As is written as [4, 3, 4]s. The following Lemma lists all possible such subgroups of T.
Lemma 3.1. Let A be a subgroup of T, and if for every a e A, the image of a under all changes of sign and permutations of coordinates (which is conjugation of a by the
72
Ars Math. Contemp. 15 (2018) 113-126
stabilization of the origin in [4, 3,4]) is also in A, then A = ((x, 0,0), (0, x, 0), (0,0, x)), ((x, x, 0), (— x, x, 0), (0, —x, x)) or ((x, x, x), (2x, 0, 0), (0, 2x, 0)).
Proof. As adapted from page 165 from Abstract Regular Polytopes [11].
Let s be the smallest positive integer from all coordinates of vectors in A, then we can assume that (s, s2, S3) G A. Then (—s, s2, S3) € A and thus 2se1 G A and so too are each 2sej. By adding and subtracting multiples of these we can find a vector all of whose coordinates are values between — s and +s. It then follows that A is generated by the all permutations of (sk, 03-k) with all changes of sign for some k G {1,2,3}. (Note that in rank n, k can be only 1,2 or n. Since otherwise (sk, 0n-k) — (0, sk, 0n-k-1) g A and so (s, s, 0n-2) G A if k is odd or (s, 0n-1) is if k is even. Though n = 3 in rank 4.)
If k = 1 then we have the first basis mentioned in the Lemma, the second if k = 2 and the third when k = 3.	□
It follows that As = sA(1fcj0n-'), and thus, as can be seen in [11, Theorem 6D1], we have the following theorem.
Theorem 3.2. The only regular toroidal polytopes constructed from [4, 3,4] are {4, 3,4}s where s = (s, 0,0), (s, s, 0) or (s, s, s) and s > 2.
Proof. Since conjugation of vectors in A by p1, p2 and p3 are precisely all permutations of coordinates and changes of sign, this theorem follows directly from Lemma 3.1.	□
The following theorem also appears in [11] along with its proof. This theorem describes the group of each toroid. To arrive at the following result (and subsequent related results in sections 4 and 5) we note that the mirror of reflection p0 is x = 1/2 while the mirrors for p1, p2 are x = y and y = z respectively and the mirror for p3 is z = 0.
Theorem 3.3 ([11, Theorem 6D4]). Let s = (sk, 03-k), with s > 2 and k =1,2,3. Then the group [4,3,4]s is the Coxeter group [4,3,4] = (p0, p1, p2, p3), where the generators are specified in (3.1), factored out by the single extra relation which is
(PoP1P2P3P2P1)S = id, if k = 1, (P0P1 P2P3P2)2s = id, if k = 2, (p0P1P2P3)3s = id, if k = 3.
As explained in [11], a geometric argument can be used to verify the intersection property for these groups when s > 2. However, note that [4, 3,4]s does not satisfy the intersection condition when s < 1 and thus is not a C-Group. We show the breakdown of the intersection condition for s = 1 by way of example for k = 1 where cases for k = 2,3 follow similar arguments.
When s = 1, the identity P0P1P2P3P2P1 = id tells us that p0 G (pi, p2, p3) so G does not satisfy the intersection property.
4 Toroidal hypertopes whose automorphism group is bB3 (=
Let {p0, p1, p2, p3} be the set of generators of [4,3,4] as in the previous section and e the corresponding fundamental simplex. We can double this fundamental simplex by replacing
E. Ens: Rank 4 toroidal hypertopes
73
the generator p0 with p0 = poPiPo- Then p0 is a reflection through the hyperplane through the point (1,0,0) with normal vector (1,1,0). The transformation of a general vector by
po is
(x,y, z)p0 = (1 - y, 1 - x,z).	(4.1)
Then [p0, p1,p2, p3} generates B3, a subgroup of [4, 3,4] of index 2. The Coxeter diagram for this group is the non-linear diagram in Figure 2. In this section we let G(= B3) := (p0, p1,p2, p3) and let C(B3) be the Coxeter complex of G = B3.
Figure 2: Coxeter diagram for B3.
The fundamental simplex of B3 is the simplex in Figure 3 bounded by the planes H1,H2, H3 (fixed by pi, p2, p3 respectively) and Ho (now fixed by po). Let, as above, Fi be the vertices of the fundamental simplex opposite to Hj. The orbit of each vertex, Fj of the fundamental simplex of B3 represents the set of hyperfaces of type j. Since this fundamental simplex shares vertices F0, F2 and F3 with the fundamental simplex of (p0, pi, p2, p3) we will use the same names for hyperfaces as the names in Section 3, namely, vertices, faces and facets. Though the orbit or Fi (which is isomorphic to the orbit of F0 since the maximal parabolic subgroups generated by excluding pi or p0 are isomorphic) will be called hyperedges.
Now the translation subgroup of G is different from the one translation subgroup in the previous section since the set of vertices of {4,3,4} now represent vertices and hyperedges (hyperfaces of type 0 and 1 respectively). The translation subgroup associated with this fundamental simplex is T = ((1,1,0), (-1,1,0), (0, -1,1)).
We then note that the translation by vector (1, 1, 0) is the transformation (by right multiplication) Wi = pop2p3p2pip2p3p2, (-1, 1, 0) is W2 = pip2p3p2p0p2p3p2 and (0, -1, 1) is W3 = p2p3p2pip2pipop2p3pi.
Now, to form a root lattice A we have the freedom to choose the crystollographic subgroup G0 by fixing either a vertex or a hyperedge (see [3, pages 108-109]). We choose to leave out po since this reflection does not fix F0. Doing so leaves [3,4] as the subgroup we are conjugating with, which is the same as was for [4,3,4]. We also note that if we chose pi rather than p0 then the result is functionally the same since we are still conjugating by [3,4] = (po, p2, p3) and this corresponds to forming a torus with its corners at hyper-edges.
We now note that although the same conditions are satisfied as in Lemma 3.1, T is now a different set. So instead we have the following lemma.
74
Ars Math. Contemp. 15 (2018) 113-126
Figure 3: Fundamental simplex of B3.
Lemma 4.1. If T = ((1,1,0), (-1,1,0), (0, -1,1)), A < T a subgroup and if for every a G A, the image of a under changes of sign and permutations of coordinates is also in A, then A = ((2x, 0, 0), (0, 2x, 0), (0,0, 2x)), ((x, x, 0), (-x, x, 0), (0, -x, x)) or ((2x, 2x, 2x), (4x, 0,0), (0,4x, 0)).
Proof. We will only modify the proof of Lemma 3.1. In that proof we arrive at a generating set (sk, 03-k) for each k G {1,2,3}, given that T is different than the translation subgroup of Section 3.
Similar arguments to the ones used in the proof to Lemma 3.1 can now be used to show that for k = 1 or k = 3, s is even. For k = 2, A is generated by permutations and changes of sign of (s, s, 0). This needs no further examination since it is clearly in T.	□
As in the previous section, we describe the groups that will be used to construct each of the toroids. We denote by Gs the quotient B3/As. Earlier we noted wi as the translation (1,1,0) while (poP2P3P2P1)2 is the translation (2,0,0) and (PoP^P^^)3 is the translation (2,2,2). And now that the the mirror for P0 is y =1 - x.
Theorem 4.2. Let s = (2s, 0,0), (s, s, 0) with s > 2 or (2s, 2s, 2s) with s > 1. Then the group Gs = B3/As is the Coxeter group B3 = (po, p1, p2, p3) with Coxeter diagram in Figure 2, factored out by the single extra relation which is
(poP2P3P2Pi)2s = id ifs = (2s, 0,0), (P0P2P3P2PlP2P3P2)S = id ifs = (s, s, 0), (PoP2P3PiP2P3)3s = id ifs = (2s, 2s, 2s).
Here, as in Section 3, we have that Gs fails the intersection property for small enough s. However, because the fundamental simplex is doubled, this time when s = (2s, 2s, 2s), Gs satisfies the intersection condition for s > 1 while s > 2 is still necessary for the other two vectors. Verifying that Gs fails the intersection condition for s = 1 when s = (2s, 0,0) and (s, s, 0) follows similar calculations as those done in Section 3. Namely, when s = 1 for the
E. Ens: Rank 4 toroidal hypertopes
75
first vector, we arrive at the identity p0p2p3p2p0 = P1P2P3P2P1 and for the second vector we have the identity p0p2p3p2 = p2p3p2p1. Which violates the intersection condition.
Magma [1] can be used to verify that Gs satisfies the intersection condition when s = (4,0,0) = (2s, 0,0), (2, 2,0) = (s, s, 0) or (2,2, 2). To see that the it also satisfies the intersection condition for greater values of s can be seen with a geometric argument.
The orbit of a base chamber of each parabolic subgroup of Gs can be seen as a collection of chambers which are duplicated at each of the 8 corners of the boundaries of As. For instance, the subgroup (p1, p2, p3) consists of chambers forming octahedra centred around corner vertex.
Given the collection of chambers in two such subgroups, there will always be some intersection between the collections occurring at the same corner (someones it's just the base chamber itself). However, If Gs fails the intersection condition, then there will be an intersection with the chambers of one subgroup centred around one corner that intersect with the chambers of the other subgroup on another corner.
So, given a particular s where Gs satisfies the intersection condition, by increasing s, the corners of As get further and further apart. So if there are no such intersections for some s, then for larger s there will not be either.
Adopting a similar notation as in the previous section and using As as defined in Section 2, we now have the following theorem.
Theorem 4.3. The regular toroidal hypertopes of rank 4 constructed from G(= B3) = (po, p1, p2, p3), where the generators are specified in (3.1) and (4.1), are C(B3)/As where C(B3) is the Coxeter complex of B3 and s = (2s, 0,0), (s, s, 0) with s > 2 or (2s, 2s, 2s) with s > 1.
Proof. To begin we need to find an s and corresponding As that is invariant under conjugation by a subgroup of G which is the symmetry group of "vertex"-figure (by vertex we mean, the element that the translations begin from). In this case our subgroup ends up being [3,4] as was described before Lemma 4.1.
Now, since we are conjugating by [3,4] = (p 1, p2, p3), As must contain all permutations and changes of sign of any vector in As (which we discovered in the proof of Theorem 3.2 which is also on page 165 of [11]). Thus, by Lemma 4.1, s = (2s, 0,0), (s, s, 0) or (2s, 2s, 2s). However, we still do not know if this construction yields a regular hypertope. To do this, we start by noting that the Coxeter complex C(B3) formed from G is precisely the hypertope r(G; (Gj)ie/) (the construction of which follows from [7]).
So we need to show that C(B3) is flag transitive (or, equivalently, chamber transitive). To do so we will note the rank 3 residue r- := r(G-; (G^g i})ie{12 3}). This is isomorphic to the cube, a regular polyhedron, which is flag transitive.
So we pick to chambers in r(G; (Gj)ie/) = C(B3) which can be written as C1 = {G0go,G1g1,G2g2^3^3} and C2 = {Ggfeo, G^, G2h2,G3M for some g^h G G. Then, since G — G— x T and T acts transitively on elements of type 0 there is a translation t G G such that C1t = {G-h0,X, Y, Z} which is some chamber that shares the same element of type 0 as C2. Then the chambers C1t and C2 are both in some rank 3 residue which is isomorphic to Pg. Since this residue is flag transitive, there is some element, g G G such that C1tg = C2. Thus C(B3) is chamber transitive and thus flag transitive. So, by Proposition 4.6 from [7] this is a regular hypertope.
So now we want to know if r(G'; (Gi)ie/) is a regular hypertope where G' is the
76
ArsMath. Contemp. 15(2018)67-79
group G/As where s > 2 (since otherwise G' fails the intersection condition and the resulting construction fails to be thin). Just as before, we take two chambers $ and ^ from r(G'; (G'i)ieI). Then to each of these chambers we can associate a family of chambers
and in C(B3). Since C(B3) is chamber transitive for each $ e and e there exists j e G where $j = . In particular there exist chambers e and e in C(B3) where, since As is invariant under G, = and ^ e G'. We can see this by noting that and are the members of their respective families which lie inside the fundamental region of As.
Thus r(G'; (G'i)ieI) is chamber transitive and thus face transitive, so is also a regular hypertope by Proposition 2.6.
For the other two possibilities of A, we need only change the added relations, but because the relations were chosen specifically, they will also generate regular hypertopes. □
5 Toroidal hypertopes whose automorphism group is A3 (= Pn)
We can show that this group is, yet again a subgroup of [4, 3,4] by doubling the fundamental simplex a second time (this can be seen geometrically in Figure 5) and now defining p3 = p3p2p3 which is a reflection in the plane through (1,1, -1) with normal vector (0,1,1). Transformation of a general vector by p3 is
(x,y,z)p3 = (x, -z, -y).	(5.1)
Now we let G(= A3) := {p0, p1,p2,p3} andC(A3) be the Coxeter complex of G. The defining relations for G are implicit in the Coxeter diagram in Figure 4.
Figure 4: Coxeter diagram for A3.
Here the fundamental simplex of A3 is a tetrahedron bounded by the planes Hi (fixed by pi). This fundamental simplex shares the planes fixed by p0,p1, p2 with the fundamental simplex of B3 as well as the corresponding vertices. The stabilizers of each vertex of the fundamental simplex are also isomorphic since all maximal parabolic subgroups of A3 are pairwise isomorphic. This implies that the set of hyperfaces of types i and j are isomorphic for each i,j e {0,1,2, 3}.
This fundamental simplex gives us the same translation subgroup as we had in the previous Section. Though now we must use the new generators to find the translations. We define W1 = p0P2P1P3P1P2 = (1,1,0), w2 = p1p2p3P0P3P2 = (-1,1,0) and W3 = P2P1P0P3P0P1 = (0, -1,1).
E. Ens: Rank 4 toroidal hypertopes
77
	/				/
			/1	K	
				A	/
				R	
		/			
Figure 5: Fundamental simplex of A3.
Using these translations, for a translation (a, b, c) G A, we have that pi(a, b, c)pi = (b, a, c). In a similar way, conjugating by p2 yields (a, c, b) and conjugating by p3 yields (a, -c, -b). So if we conjugate by p1p2p1 then we get (c, b, a) and so A must have all permutations. Now, from the previous we know A must also contain (a, -b, -c) and adding this to (a, b, c) gives (2a, 0,0), which then subtracted from (a, b, c) is (-a, b, c) and so with all permutations means that A must have all permutations and changes of sign.
With this group, we leave out p0 to form the crystollographic subgroup G0. Though a curiosity of this group is that we use any generator of B3 to form a crystollographic subgroup and still finish with the same objects. With each choice simply changing where we draw the boundary of the torus. This leaves p1, p2 and p3 with which to conjugate A. As in the regular case, p1 and p2 show us that A must consist of all permutations of the coordinates of vectors.
If (a, b, c) is a general vector in A then p3 tells us that (-c, b, -a) must also be in A and then so also must (-a, b, -c). Adding that to our original general vector tells us that (0,2b, 0) is also included. So, subtracting that from the general vector finally gives us (a, -b, c). Note that this can just as easily be done with either a or c with some simple permutations.
As in the previous section, we describe the groups of each of the toroids. Earlier we noted w1 as the translation (1,1,0) while (^p2^p1)2 is the translation (2,0,0) and (yOop2p1^)3s is the translation (2, 2,2). And now that the the mirror for p0 is y =1 - x while the mirror for p3 is y = - z.
Theorem 5.1. Let s = (2s, 0,0), (s, s, 0) with s > 2 or (2s, 2s, 2s) with s > 1. Then the group Gs = A3/As is the Coxeter group A3 = (p0, p1, p2, p3) (with Coxeter group specifed in Figure 4), factored out by the single extra relation which is
(PoP2P3p1 )2s = id if s = (2s, 0, 0), (P0P2P1 y03P1P2)s = id ifs = (s, s, 0), (P0P2P1 P3)3s = id if s = (2s, 2s, 2s).
78
Ars Math. Contemp. 15 (2018) 113-126
For the same reasons as in Section 4, the intersection condition is satisfied for s =
(2s, 2s, 2s) when s > 1.
Theorem 5.2. The regular toroidal hypertopes of rank 4 induced by G(= A3) = (p0, Pi,Pi, P3) (where the generators are specified in (3.1), (4.1) and (5.1)) are C(A3)/As where C(A3) is the Coxeter complex of A3 and s = (2s, 0,0), (s, s, 0) for s > 2 or (2s, 2s, 2s) with s > 1 .
Proof. We first show that C (A3) is a regular hypertope, which requires showing that it is flag transitive. In the same manner as the proof of Theorem 4.3 we need only show each rank 3 residue is flag transitive, since all rank 3 residues are regular tetrahedra C (A3) is flag transitive. The translation subgroup is the same as in the previous Section and conjugating A by p\,p2,p3 gives all permutations and changes in sign of a general vector in A, the same arguments for Lemma 4.1 and Theorem 4.3 will prove this theorem.	□
6 Non-existence of rank 4 chiral hypertopes
Here we recall that for an abstract polytope to be chiral its automorphism group must have two orbits when acting on flags and that adjacent flags are in different orbits. Chiral polytopes of any rank are examined in depth in [13]. The existence of these objects in any rank was proved in [12]. There is also a notion of chirality in hypermaps as well, see for example, [2]. Similarly we say for a hypertope to be chiral if its automorphism group action has two chamber orbits and adjacent chambers are in different orbits [7].
As in Section 2, given an affine Coxeter group G and associated Coxeter complex C, we define a subgroup G0 < G as the maximal parabolic subgroup fixing the origin. Then, given a set I of linearly independent translations in G and T/, the translation subgroup generated by I then we call the lattice A/ the lattice induced by the orbit of the origin under T/. When A/ is invariant under the rotation subgroup G0+ but there is no automorphism group of G that interchanges adjacent chambers, then in rank 4 we say that the quotient C/A/ is a chiral toroidal hypertope of rank 4.
The proof that there are no chiral toroids of rank 4 for the group [4,3,4] comes from page 178 from [11] and the same proof can adapted for the other two rank 4 affine Coxeter groups. The basic idea for the proof is that since C/A is chiral, A is invariant under the rotation group [3,4]+, so A contains vectors that are compositions of an even number of permutations with an even number of sign changes or all compositions of an odd number of permutations with an odd number of sign changes. It then goes to show that if (a, b, c) G A then (b, a, c) G A, which is the image of (a, b, c) under an odd permutation, which is a contradiction. Therefore no such A can exist.
We will use the same method to show the same is true for the other two groups.
Theorem 6.1. There are no rank 4 chiral toroidal hypertopes.
Proof. In [11] it was shown that there are no rank 4 hypertopes constructed from [4,3,4], so we show for constructions from B3 and A3. In previous sections we found that if A is a subgroup of thejranslations that is invariant under conjugation by the stabilizer of the origin in B3 and A3 with (a, b, c) G A, then A contains all permutations and changes of sign of (a, b, c), just as it did with the stabilizer in [4, 3,4].
Thus conjugation of A by the stabilizer of the rotation subgroup of each of these groups is all compositions of even permutations with an even number of sign changes or all compositions of odd permutations with an odd number of sign changes, just as for [4,3,4].
E. Ens: Rank 4 toroidal hypertopes
79
So the same arguments and calculations from page 178 in [11] still hold and show that
(b, a,c) e A and we develop the same contradiction.	□
References
[1]	W. Bosma, J. Cannon and C. Playoust, The Magma Algebra System I: The User Language, J. Symbolic Comput. 24 (1997), 235-265, doi:10.1006/jsco.1996.0125.
[2]	A. Breda D'Azevedo, G. A. Jones, R. Nedela and M. Skoviera, Chirality groups of maps and hypermaps, J. Algebraic Combin. 29 (2009), 337-355, doi:10.1007/s10801-008-0138-z.
[3]	J. H. Conway and N. J. A. Sloane, Sphere packings, lattices and groups, volume 290 of Grundlehren der Mathematischen Wissenschaften, Springer-Verlag, New York, 1988, doi: 10.1007/978-1-4757-2016-7.
[4]	D. Corn and D. Singerman, Regular hypermaps, European J. Combin. 9 (1988), 337-351, doi: 10.1016/s0195-6698(88)80064-7.
[5]	H. S. M. Coxeter, Configurations and maps, Rep. Math. Colloq. 8 (1948), 18-38.
[6]	H. S. M. Coxeter and W. O. J. Moser, Generators and relations for discrete groups, volume 14 of Ergebnisse der Mathematik und ihrer Grenzgebiete, Springer-Verlag, Berlin-New York, 4th edition, 1980.
[7]	M. E. Fernandes, D. Leemans and A. Ivic Weiss, Highly symmetric hypertopes, Aequat. Math. 90 (2016), 1045-1067, doi:10.1007/s00010-016-0431-1.
[8]	I. Hubard, A. Orbanic, D. Pellicer and A. Ivic Weiss, Symmetries of equivelar 4-toroids, Discrete Comput. Geom. 48 (2012), 1110-1136, doi:10.1007/s00454-012-9444-2.
[9]	J. E. Humphreys, Reflection groups and Coxeter groups, volume 29 of Cambridge Studies in Advanced Mathematics, Cambridge University Press, Cambridge, 1990, doi:10.1017/ cbo9780511623646.
[10]	P. McMullen, and E. Schulte, Higher toroidal regular polytopes, Adv. Math. 117 (1996), 17-51, doi:10.1006/aima.1996.0002.
[11]	P. McMullen and E. Schulte, Abstract regular polytopes, volume 92 of Encyclopedia of Mathematics and its Applications, Cambridge University Press, Cambridge, 2002, doi:10.1017/ cbo9780511546686.
[12]	D. Pellicer, A construction of higher rank chiral polytopes, Discrete Math. 310 (2010), 12221237, doi:10.1016/j.disc.2009.11.034.
[13]	E. Schulte and A. Ivic Weiss, Chiral polytopes, in: Applied Geometry And Discrete Mathematics, Proceedings of a DIMACS Workshop, Providence, Rhode Island, USA, September 18, 1990, 1990 pp. 493-516.
[14]	J. Tits, Sur les analogues algébriques des groupes semi-simples complexes, in: Colloque d'Algèbre supérieure, Bruxelles du 19 au 22 décembre 1956, Centre Belge de Recherches Mathématiques Etablissements Ceuterick, Louvain; Librairie Gauthier-Villars, Paris, 1975 pp. 261-289.