The uniqueness of a distanceregular graph with intersection array \(\{32,27,8,1;1,4,27,32\}\) and related results
 664 Downloads
 1 Citations
Abstract
It is known that, up to isomorphism, there is a unique distanceregular graph \(\Delta \) with intersection array \(\{32,27;1,12\}\) [equivalently, \(\Delta \) is the unique strongly regular graph with parameters (105, 32, 4, 12)]. Here we investigate the distanceregular antipodal covers of \(\Delta \). We show that, up to isomorphism, there is just one distanceregular antipodal triple cover of \(\Delta \) (a graph \(\hat{\Delta }\) discovered by the author over 20 years ago), proving that there is a unique distanceregular graph with intersection array \(\{32,27,8,1;1,4,27,32\}\). In the process, we confirm an unpublished result of Steve Linton that there is no distanceregular antipodal double cover of \(\Delta \), and so no distanceregular graph with intersection array \(\{32,27,6,1;1,6,27,32\}\). We also show there is no distanceregular antipodal 4cover of \(\Delta \), and so no distanceregular graph with intersection array \(\{32,27,9,1;1,3,27,32\}\), and that there is no distanceregular antipodal 6cover of \(\Delta \) that is a double cover of \(\hat{\Delta }\).
Keywords
Distanceregular graph Strongly regular graph Antipodal cover Fundamental groupMathematics Subject Classification
Primary 05E30 Secondary 05E45 55041 Introduction
We shall investigate the antipodal distanceregular covers of the GoethalsSeidel graph [9], the unique (up to isomorphism) distanceregular graph \(\Delta \) with intersection array \(\{32,27;1,12\}\) [4, 5]. The graph \(\Delta \) can be constructed as the second subconstituent of the second subconstituent of the famous McLaughlin graph, the unique distanceregular graph with intersection array \(\{112,81;1,56\}\) (see [3]).
We classify distanceregular antipodal rcovers of \(\Delta \) by studying the rfold topological covers of the 2dimensional simplicial complex whose 0, 1, and 2simplices are respectively the vertices, edges, and triangles of \(\Delta \). Our main tool is version 2.0 of the author’s GAP program described in [12] for the computation of fundamental groups, certain quotients of fundamental groups, and covers of finite abstract 2dimensional simplicial complexes. This program is freely available from [16], where we also provide a GAP/GRAPE [15, 17] logfile of all the computations related to this paper. It is hoped that [16] and the methods of this paper will be useful in other classifications of covers of graphs. In particular, Theorem 3.1 should be of independent interest.
All graphs in this paper are finite and undirected, with no loops and no multiple edges, and have at least one vertex. Throughout, we follow [2] for graphtheoretical concepts and notation. An important new reference for distanceregular graphs, covering developments since [2] was published, is Van Dam et al. [18]. See also Brouwer et al. [1]. A good reference for the grouptheoretical concepts used in this paper is Robinson [13]. We denote the commutator subgroup of a group G by [G, G], the cyclic group of order n is denoted by \(C_n\), and where p is a prime, the elementary abelian group of order \(p^k\) is denoted by \(C_p^k\).
2 The fundamental group and covers

For every \(v\in V(\Gamma )\), the fibre \(v\theta ^{1}\) of v is a coclique (independent set) of \(\tilde{\Gamma }\),

The union of any two distinct fibres mapping under \(\theta \) to a nonedge of \(\Gamma \) is a coclique of \(\tilde{\Gamma }\),

The induced subgraph on any two fibres mapping under \(\theta \) to an edge of \(\Gamma \) is a perfect matching in \(\tilde{\Gamma }\),

The induced subgraph on any three fibres mapping under \(\theta \) to a triangle of \(\Gamma \) consists of pairwise disjoint triangles in \(\tilde{\Gamma }\).
When we do not need to specify the covering map explicitly, we may denote a cover \((\tilde{\Gamma },\theta )\) simply by \(\tilde{\Gamma }\). If each fibre of a cover \(\tilde{\Gamma }\) of \(\Gamma \) has the same positive integer cardinality r, then we call \(\tilde{\Gamma }\) an rcover of \(\Gamma \). A 2cover is also called a double cover, and a 3cover is also called a triple cover. If \(\Gamma \) is a noncomplete graph, then an antipodal rcover \(\tilde{\Gamma }\) of \(\Gamma \) is a connected rcover of \(\Gamma \) with the property that being equal or at maximum distance in \(\tilde{\Gamma }\) is an equivalence relation on \(V(\tilde{\Gamma })\), whose equivalence classes are the fibres.
The connected rcovers of \(\Gamma \) correspond to the transitive permutation representations of degree r of the fundamental group of \(\Gamma \) (regarded as a 2complex), defined with respect to a fixed, but arbitrary, spanning tree of \(\Gamma \). We shall explain this further in this section. For a more general, fuller and detailed exposition, see [12], on which our explanation is based.

\(g_{s,t}=1\) for each arc (s, t) of T;

\(g_{v,w}g_{w,v}=1\) for each edge \(\{v,w\}\) of \(\Gamma \);

\(g_{x,y}g_{y,z}g_{z,x}=1\) for each triangle \(\{x,y,z\}\) of \(\Gamma \).
Now for each arc (v, w) of \(\Gamma \), let \(\rho _{v,w}\) be the permutation of I defined by \(i\rho _{v,w} = j\) if and only if \(\{(v,i),(w,j)\}\) is an edge of \(\tilde{\Gamma }\). If \(\{x,y,z\}\) is a triangle of \(\Gamma \), then \(\{x,y,z\} \theta ^{1}\) is a set of disjoint triangles in \(\tilde{\Gamma }\), and so \(\rho _{x,y}\rho _{y,z}\rho _{z,x}\) is the identity permutation. It now follows that the map \(\rho \) defined on the generators of G by \((g_{v,w})\rho = \rho _{v,w}\) extends to a transitive permutation representation from G to the symmetric group \(S_r\). The rcover \(\tilde{\Gamma }\) is completely defined by this representation.
Conversely, every transitive permutation representation \(\rho :G\rightarrow S_r\) defines a connected rcover, denoted \(\Gamma _\rho \), of \(\Gamma \). For each \(v\in V(\Gamma )\), the fibre of v in \(\Gamma _\rho \) consists of the ordered pairs (v, i), for \(i\in I\), and (v, i) and (w, j) are joined by an edge in \(\tilde{\Gamma }\) if and only if \(\{v,w\}\) is an edge of \(\Gamma \) and \(j=i((g_{v,w})\rho )\). The preimage in G of the stabilizer in \(G\rho \) of a point \(i\in I\) is isomorphic to the fundamental group of \(\Gamma _\rho \) (see [12]).
We consider two covers \((\Gamma _1,\theta _1)\) and \((\Gamma _2,\theta _2)\) of \(\Gamma \) to be isomorphic if there is a graph isomorphism from \(\Gamma _1\) to \(\Gamma _2\) which maps the fibres of \(\Gamma _1\) to those of \(\Gamma _2\). Thus, isomorphic covers differ only by a relabelling of the fibres and of the vertices within each fibre. In particular, every connected rcover of \(\Gamma \) is isomorphic to a cover of the form \(\Gamma _\rho \), for some transitive permutation representation \(\rho :G\rightarrow S_r\), corresponding to the preimage in G of the stabilizer in \(G\rho \) of the point 1.
Isomorphism of covers can be checked as follows using nauty [11], which can be called from within GRAPE [15]. Given the cover \(\Gamma _1\) of \(\Gamma \), we make a \(\{\text {red},\text {blue}\}\)vertexcoloured graph \(\Gamma _1^+\). The redcoloured vertices of \(\Gamma _1^+\) are the vertices of \(\Gamma _1\), with two red vertices joined in \(\Gamma _1^+\) precisely when those vertices are joined in \(\Gamma _1\). The bluecoloured vertices of \(\Gamma _1^+\) are in onetoone correspondence with the fibres of \(\Gamma _1\), and a blue vertex is joined in \(\Gamma _1^+\) only to the red vertices in the corresponding fibre. Similarly, we make \(\Gamma _2^+\) from the cover \(\Gamma _2\) of \(\Gamma \). Then \(\Gamma _1\) and \(\Gamma _2\) are isomorphic as covers of \(\Gamma \) if and only if \(\Gamma _1^+\) is isomorphic to \(\Gamma _2^+\) by a colourpreserving graph isomorphism.
We are usually interested in classifying covers up to isomorphism. Note that if \((\Gamma _1,\theta _1)\) and \((\Gamma _2,\theta _2)\) are isomorphic covers of \(\Gamma \), and \((\tilde{\Gamma }_1,{\tilde{\theta }}_1)\) is any cover of \(\Gamma _1\), then \((\tilde{\Gamma }_1,{\tilde{\theta }}_1\theta _1)\) is a cover of \(\Gamma \) isomorphic to \((\tilde{\Gamma }_2,{\tilde{\theta }}_2\theta _2)\) for some cover \((\tilde{\Gamma }_2,{\tilde{\theta }}_2)\) of \(\Gamma _2\).
3 Imprimitivity and covers
We may sometimes be interested in a cover of a cover.
Let \(\Gamma \) be a connected graph, let \((\Gamma _1,\theta _1)\) be a connected mcover of \(\Gamma \), and let \((\Gamma _2,\theta _2)\) be a connected ncover of \(\Gamma _1\). Then clearly, \((\Gamma _2,\theta _2\theta _1)\) is a connected mncover of \(\Gamma \). Given a spanning tree T of \(\Gamma \), we take a spanning tree \(T_1\) of \(\Gamma _1\) containing the forest \(T\theta _1^{1}\), and define the fundamental group G of \(\Gamma \) with respect to T and the fundamental group \(G_1\) of \(\Gamma _1\) with respect to \(T_1\). Then \((\Gamma _1,\theta _1)\) corresponds to a transitive representation \(\rho _1:G\rightarrow S_m\), with the fibre of v in \(\Gamma _1\) being labelled \(\{(v,i):1\le i\le m\}\), as described previously, and \((\Gamma _2,\theta _2)\) corresponds to a transitive representation \(\rho _2:G_1\rightarrow S_n\), with the fibre of (v, i) being labelled \(\{(v,i,j):1\le j\le n\}\). Now \((\Gamma _2,\theta _2\theta _1)\) corresponds to a transitive representation \(\rho :G\rightarrow S_{mn}\), and if \(m,n>1\), then \(G\rho \) acts imprimitively on the indices (i, j) of the fibre of a vertex of \(\Gamma \), the blocks of imprimitivity being \(\{(i,j):1\le j\le n\}\), for \(i=1,\ldots ,m\).
Conversely, suppose that \(m,n>1\) and \(\rho :G\rightarrow S_{mn}\) is a transitive permutation representation of the fundamental group G of \(\Gamma \) such that \(G\rho \) is an imprimitive group having m blocks of size n. Then, if \(\sigma :G\rightarrow S_m\) is the transitive permutation action of \(G\rho \) on the blocks of imprimitivity, we see that \(\Gamma _\rho \) is an ncover of the mcover \(\Gamma _\sigma \) of \(\Gamma \).
We have the following useful result.
Theorem 3.1
Let \(\Gamma \) be a noncomplete distanceregular graph and suppose \(\Gamma _\rho \) is a distanceregular antipodal mncover of \(\Gamma \) corresponding to a transitive permutation representation \(\rho \) of the fundamental group G of \(\Gamma \), such that \(G\rho \) has \(m>1\) blocks of imprimitivity of size \(n>1\). Then \(\Gamma _\rho \) must be an ncover of an antipodal distanceregular mcover of \(\Gamma \).
Proof
As above, we may suppose that \(G\rho \) is a group of permutations of \(\Omega {:=}\{(i,j):1\le i\le m,\,1\le j\le n\}\), with blocks of imprimitivity \(B_i{:=}\{(i,j):1\le j\le n\}\), for \(i=1,\ldots ,m\), and that the fibre of \(\Gamma _\rho \) mapping to the vertex v of \(\Gamma \) is labelled as \(\{(v,i,j):1\le i\le m,\,1\le j\le n\}\).
4 On the distanceregular antipodal rcovers of \(\Delta \)
To study covers of \(\Delta \), we explore quotients of the fundamental group of \(\Delta \), viewed as a 2complex, with respect to a fixed spanning tree T of \(\Delta \). This fundamental group is denoted throughout this section by D. It was shown in [14] that D is infinite.
Theorem 4.1
The abelianised fundamental group D / [D, D] of \(\Delta \) is isomorphic to \(C_2^{16}\times C_3^2\).
Proof
We compute a presentation for D / [D, D] using the program [16], and determine the abelian invariants of this group using GAP [17]. \(\square \)
We are now in a position to classify the connected double covers of \(\Delta \), which was done independently by Steve Linton over 20 years ago, using his vector enumeration algorithm [10].
Theorem 4.2
Up to isomorphism of covers, \(\Delta \) has just 13 connected double covers, with each having its abelianised fundamental group isomorphic to \(C_2^{15}\times C_3^2\), and none being distanceregular.
Proof
Each subgroup of index 2 in D contains the commutator subgroup [D, D]. We compute the \(2^{16}1\) covers of \(\Delta \) corresponding to the subgroups of index 2 in D / [D, D], and using GRAPE calling nauty, we determine that, up to isomorphism of covers, there are just 13 such covers. We test each of these covers for distanceregularity using GRAPE, and find that none of them are distanceregular. We use the program [16] to compute each of their abelianised fundamental groups. \(\square \)
Corollary 4.3
There is no distanceregular graph with intersection array \(\{32,27,6,1;1,6,27,32\}\).
We now consider the triple covers of \(\Delta \), and start by showing that D has no quotient isomorphic to the symmetric group \(S_3\). This is a corollary of the following:
Proposition 4.4
Let N be a subgroup of D of index 2, and suppose N has a subgroup M, such that M is normal in D and N / M is abelian. Then D / M is abelian.
Proof
The subgroup N of D is isomorphic to the fundamental group of some connected 2cover of \(\Delta \), so by Theorem 4.2, [N, N] has index \(2^{15}3^2\) in N, and so [N, N] has index \(2^{16}3^2\) in D, as does [D, D], by Theorem 4.1. Thus \([N,N]=[D,D]\), and since N / M is abelian, M contains [N, N], and so D / M is abelian. \(\square \)
Corollary 4.5
The group D has no quotient isomorphic to the symmetric group \(S_3\).
We are now in a position to classify the connected triple covers of \(\Delta \).
Theorem 4.6
Up to isomorphism of covers, \(\Delta \) has just two connected triple covers, \(\Delta ^*\) and \(\hat{\Delta }\). The cover \(\Delta ^*\) has abelianised fundamental group isomorphic to \(C_2^{16}\times C_3^2\) and is not distanceregular. The cover \(\hat{\Delta }\) has abelianised fundamental group isomorphic to \(C_2^{18}\times C_3^2\) and is distanceregular.
Proof
Each connected triple cover of \(\Delta \) is isomorphic to a cover of the form \(\Delta _\rho \), for some transitive permutation representation \(\rho :D\rightarrow S_3\). Since D has no quotient isomorphic to \(S_3\), we must have \(D\rho \cong C_3\), the cyclic group of order 3, and so each subgroup of index 3 in D is normal and contains [D, D].
We compute the four covers of \(\Delta \) corresponding to the subgroups of index 3 in D / [D, D], and determine that, up to isomorphism of covers, there are just two such covers, \(\Delta ^*\) and \(\hat{\Delta }\), and we calculate that these have the properties as stated in the theorem. \(\square \)
Corollary 4.7
Up to isomorphism, there is a unique distanceregular graph with intersection array \(\{32,27,8,1;1,4,27,32\}\).
Further properties of this distanceregular graph \(\hat{\Delta }\) are given in [14] (where the graph is called \(\Lambda \)).
We now show there is no distanceregular antipodal 4cover of \(\Delta \).
Theorem 4.8
There is no distanceregular graph with intersection array \(\{32,27,9,1;1,3,27,32\}\).
Proof
A distanceregular graph with the intersection array of the theorem would be an antipodal 4cover of \(\Delta \). Such a 4cover would correspond to some transitive permutation representation \(\rho \) of degree 4 of D. The image \(D\rho \) of D cannot be imprimitive, for otherwise, by Theorem 3.1, \(\Delta \) would have a distanceregular antipodal 2cover. We cannot have \(D\rho \cong S_4\), since D has no quotient isomorphic to \(S_3\), which is a quotient of \(S_4\). This leaves \(D\rho \cong A_4\) as the only possibility.
Suppose now D has a quotient isomorphic to \(A_4\). Then there is a transitive permutation representation \(\alpha :D\rightarrow S_6\), having three blocks of imprimitivity of size 2, with \(D\alpha \cong A_4\). The corresponding cover \(\Delta _\alpha \) is a 2cover of a 3cover \(\tilde{\Delta }\) of \(\Delta \). The 3cover \(\tilde{\Delta }\) corresponds to a normal subgroup N of D of index 3, and N has a subgroup M of index 4 that is normal in D, such \(D/M\cong A_4\).
Suppose now \(\tilde{\Delta }\) is isomorphic (as a cover of \(\Delta \)) to \(\Delta ^*\). By Theorem 4.6, [N, N] has index \(2^{16}3^2\) in N, and so [N, N] has index \(2^{16}3^3\) in D. By Theorem 4.1, [D, D] has index \(2^{16}3^2\) in D, so [N, N] is a normal subgroup of index 3 in [D, D]. Since \(N/M\cong C_2^2\), M contains [N, N], and so either D / M is abelian or D / M has a normal subgroup of order 3, neither of which holds.
Thus, we must have \(\tilde{\Delta }\) isomorphic to \(\hat{\Delta }\). We computed all \(2^{18}1\) connected 2covers of \(\hat{\Delta }\) as covers of the form \(\Delta _\rho \) of \(\Delta \), for \(\rho \) a transitive permutation representation from D to \(S_6\). Just three of these covers \(\Delta _\rho \) have \(D\rho \) isomorphic to \(A_4\), corresponding to the three permutation isomorphic transitive representations of degree 6 of one quotient of D isomorphic to \(A_4\). Now given a connected 6cover \(\Delta _\rho \) with \(D\rho \cong A_4\), we construct the cover \(\Delta _\sigma \) of \(\Delta \) defined by the representation of \(D\rho \) acting by right multiplication on the four (right) cosets of a Sylow 3subgroup. Up to isomorphism of covers of \(\Delta \), there is only one such \(\Delta _\sigma \), which we find is not distanceregular.
\(\square \)
Finally, we can prove the following:
Theorem 4.9
There is no distanceregular antipodal 6cover of \(\Delta \) that is a double cover of \(\hat{\Delta }\).
Proof
When we determined all \(2^{18}1\) connected 2covers of \(\hat{\Delta }\), we found that none is distanceregular.\(\square \)
Combining this result with Theorems 3.1, 4.2 and 4.6, we see that there is no distanceregular antipodal 6cover of \(\Delta \) corresponding to an imprimitive degree 6 permutation representation of D, but this still leaves open the possibility of a distanceregular antipodal 6cover of \(\Delta \) corresponding to a primitive such representation.
Notes
Acknowledgments
I thank Andries Brouwer, Edwin van Dam, Alexander Gavrilyuk, Chris Godsil, Aleksandar Jurišić, and Jack Koolen for their interest in this work and their comments.
References
 1.Brouwer A.E., Cohen A.M., Neumaier A.: Additions and Corrections to the Book DistanceRegular Graphs. Springer, New York (1989). http://www.win.tue.nl/~aeb/drg/.
 2.Brouwer A.E., Cohen A.M., Neumaier A.: DistanceRegular Graphs. Springer, New York (1989).Google Scholar
 3.Cameron P.J., Goethals J.M., Seidel J.J.: Strongly regular graphs having strongly regular subconstituents. J. Algebra 55, 257–280 (1978).Google Scholar
 4.Coolsaet K.: The uniqueness of the strongly regular graph \(\rm srg(105,32,4,12)\). Bull. Belg. Math. Soc. Simon Stevin 12, 707–718 (2006).Google Scholar
 5.Degraer J., Coolsaet K.: Classification of some strongly regular subgraphs of the McLaughlin graph. Discret. Math. 308, 395–400 (2008).Google Scholar
 6.Gardiner A.: Antipodal covering graphs. J. Comb. Theory Ser. B, 16, 255–273 (1974).Google Scholar
 7.Godsil C.D.: Algebraic Combinatorics. Chapman and Hall, New York (1993).Google Scholar
 8.Godsil C.D., Hensel A.D.: Distance regular covers of the complete graph. J. Comb. Theory Ser. B 56, 205–238 (1992).Google Scholar
 9.Goethals J.M., Seidel J.J.: Strongly regular graphs derived from combinatorial designs. Can. J. Math. 22, 597–614 (1970).Google Scholar
 10.Linton S.A.: On vector enumeration. Linear Algebra Appl. 192, 235–248 (1993).Google Scholar
 11.McKay B.D., Piperno A.: Practical graph isomorphism, II. J. Symb. Comput. 60, 94–112 (2014).Google Scholar
 12.Rees S., Soicher L.H.: An algorithmic approach to fundamental groups and covers of combinatorial cell complexes. J. Symb. Comput. 29, 59–77 (2000).Google Scholar
 13.Robinson D.J.S.: A Course in the Theory of Groups, 2nd edn. Springer, New York (1995).Google Scholar
 14.Soicher L.H.: Three new distanceregular graphs. Eur. J. Comb. 14, 501–505 (1993).Google Scholar
 15.Soicher L.H.: The \({\sf GRAPE~\sf package~ for ~{\sf GAP}}\), Version 4.6.1 (2012). http://www.maths.qmul.ac.uk/~leonard/grape/.
 16.Soicher L.H.: Program to compute fundamental groups and covers, Version 2.0 (2015). http://www.maths.qmul.ac.uk/~leonard/fundamental/.
 17.The GAP Group. \({\sf GAP}\)—Groups, Algorithms, and Programming, Version 4.7.7 (2015). http://www.gapsystem.org.
 18.van Dam E.R., Koolen J.H., Tanaka H.: Distanceregular graphs. Electron. J. Comb. 156 (2016).Google Scholar
Copyright information
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.