Skip to content
Publicly Available Published by De Gruyter July 16, 2020

On quasiprimitive edge-transitive graphs of odd order and twice prime valency

  • Hong Ci Liao EMAIL logo , Jing Jian Li and Zai Ping Lu
From the journal Journal of Group Theory

Abstract

A graph is edge-transitive if its automorphism group acts transitively on the edge set. In this paper, we investigate the automorphism groups of edge-transitive graphs of odd order and twice prime valency. Let Γ be a connected graph of odd order and twice prime valency, and let G be a subgroup of the automorphism group of Γ. In the case where G acts transitively on the edge set and quasiprimitively on the vertex set of Γ, we prove that either G is almost simple, or G is a primitive group of affine type. If further G is an almost simple primitive group, then, with two exceptions, the socle of G acts transitively on the edge set of Γ.

1 Introduction

In this paper, all graphs are assumed to be finite and simple. In particular, a graph is a pair Γ=(V,E) of a nonempty set V and a set E of 2-subsets of V, which are called the vertex set and edge set of Γ, respectively. Each edge {α,β}E gives two ordered pairs (α,β) and (β,α), and each of them is called an arc of Γ. A triple (α,β,γ) of vertices is called a 2-arc of Γ if αγ and both (α,β) and (β,γ) are arcs of Γ.

Assume that Γ=(V,E) is a graph. An automorphism g of Γ is a permutation on V such that {αg,βg}E for all {α,β}E. Denote by 𝖠𝗎𝗍Γ the set of all automorphisms of Γ. Then 𝖠𝗎𝗍Γ is a (finite) group under the product of permutations, which acts naturally on the edge set, arc set and the set of 2-arcs of Γ by

{α,β}g={αg,βg},(α,β)g=(αg,βg),(α,β,γ)g=(αg,βg,γg),

respectively. For a subgroup G𝖠𝗎𝗍Γ, the graph Γ is said to be G-vertex-transitive, G-edge-transitive, G-arc-transitive and (G,2)-arc-transitive if G acts transitively on the vertex set, edge set, arc set and the set of 2-arcs of Γ, respectively.

Let Γ=(V,E) be a connected graph, and G𝖠𝗎𝗍Γ. Suppose that G acts quasiprimitively on V, that is, every minimal normal subgroup of G acts transitively on V. Following the subdivision given in [21], the group G is one of eight types of quasiprimitive groups. It is believed that the structure of the group G is quite restricted if the graph Γ is assumed to have additional symmetric properties, or some restrictions on the order or the valency. For example, if Γ is (G,2)-arc-transitive, then Praeger [20] proved that only four of those eight types occur for G. If Γ has odd order and Γ is (G,2)-arc-transitive, then G is an almost simple group by [13]. In this paper, replacing “2-arc-transitivity” by “edge-transitivity” and a restriction on the valency, we investigate the pair of Γ and G.

Assume that Γ=(V,E) is a connected graph of valency twice a prime, and G𝖠𝗎𝗍Γ. In [22], Praeger and Xu give a nice characterization for the graph Γ when it is G-arc-transitive and G contains an irregular abelian (and so, intransitive) normal subgroup. In Section 3 of this paper, we focus on the case where Γ is G-edge-transitive and G contains a transitive minimal normal subgroup. In particular, letting soc(G) be the socle of G, the following result is proved.

Theorem 1.1.

Let Γ=(V,E) be a connected graph of odd order and valency 2r for some prime r, and let GAutΓ. Assume that Γ is G-edge-transitive but not (G,2)-arc-transitive. If G is quasiprimitive on V, then either G is almost simple, or soc(G)=Zpk for some odd prime p and integer 1kr.

A graph Γ is called 2-arc-transitive if it is (𝖠𝗎𝗍Γ,2)-arc-transitive. The following result is a consequence of Theorem 1.1, which is proved in Section 4.

Theorem 1.2.

Let Γ=(V,E) be a connected graph of odd order and twice prime valency, and let GAutΓ. Assume that Γ is G-edge-transitive but not (G,2)-arc-transitive. If G is primitive on V, then either Γ is a complete graph, or soc(AutΓ)=soc(G); in particular, Γ is 2-arc-transitive if and only if Γ is a complete graph.

For a pair (Γ,G) in Theorem 1.2, the action of soc(G) on the edge set of Γ is considered in Section 5. We present several examples and prove the following result.

Theorem 1.3.

Let Γ=(V,E) be a connected graph of odd order and twice prime valency, and let GAutΓ. Assume that Γ is G-edge-transitive but not (G,2)-arc-transitive. If G is almost simple and primitive on V, then either Γ is soc(G)-edge-transitive, or Γ is of valency 4 and isomorphic to one of the graphs in Example 5.1.

2 Preliminaries

In this section, we let V be a finite nonempty set.

Let G be a permutation group on V. For αV, set Gα={gGαg=α}, called the stabilizer of α in G. Then G is called semiregular on V if Gα=1 for every αV, and G is regular on V if it is both semiregular and transitive on V.

Lemma 2.1.

Let G be a transitive permutation group on V. Suppose that |V| is odd, and that G has a transitive minimal normal subgroup N. Then N is the unique transitive minimal normal subgroup of G.

Proof.

Assume that M is a minimal normal subgroup of G with MN. Then MN=1, and hence N𝐂G(M). By [5, Theorem 4.2A], M is semiregular on V; in particular, M has odd order as |M| is a divisor of |V|. By the choice of M, we know that M is characteristically simple; see [10, p. 21, Lemma 4.8]. Then M=T1×T2××Tk for isomorphic simple groups Ti; see [10, p. 51, Theorem 9.12]. It follows that every Ti has odd order, and so Ti is a cyclic group of prime order. Thus M is abelian. If M is transitive on V, then, by [5, Theorem 4.2A], 𝐂G(M)=M, yielding N𝐂G(M)=M, a contradiction. Therefore, M is not a transitive subgroup of G, and the lemma follows. ∎

A digraph with vertex set V is an ordered pair (V,A), where A is a subset of {(α,β)α,βV,αβ}. The members in A are called arcs. For convenience, a graph with vertex set V is sometimes viewed as the digraph (V,A), where A is the arc set of the graph. Then, with this convention, a digraph Δ=(V,A) is a graph if and only if A=A*, where A*={(α,β)(β,α)A}.

Assume that Δ=(V,A) is a digraph, and G𝖠𝗎𝗍Δ. Let αV, and set

Δ(α)={βV(α,β)A}.

Then, for each normal subgroup N of G, the stabilizer Nα induces a permutation group NαΔ(α) on Δ(α). Let Nα[1] be the kernel of Nα acting on Δ(α). Then

(2.1)

NαΔ(α)Nα/Nα[1],Nαg=(Nα)g,Nαg[1]=(Nα[1])g,
Δ(αg)=Δ(α)gfor allgG.

For a finite group X, denote by π(X) the set of prime divisors of |X|. The following lemma says that π(Nα)=π(NαΔ(α)) if Δ is connected and G is transitive on V.

Lemma 2.2.

Let Δ=(V,A) be a connected digraph and αV. Let GAutΔ, and let N be a normal subgroup of G. Assume that G is transitive on V. Then π(Nα)=π(NαΔ(α)), and either π(Gα[1])π(Nα), or N is intransitive on V.

Proof.

Let p be a prime divisor of |Nα| which is not a divisor of |NαΔ(α)|. Let P be a Sylow p-subgroup of Nα. Then 1PNα[1]. Let γV{α}. By [18, Lemma 2], there are vertices α=α0,α1,,αt=γ such that (αi-1,αi)A for 1it. Take gG with αg=α1. By (2.1), we conclude that NαΔ(α)Nα1Δ(α1). This implies that pπ(Nα1Δ(α1)), and then PNα1[1]. By induction on t, we get PNγ[1]. It follows that P fixes V point-wise, and so P=1, a contradiction. Then we have π(Nα)π(NαΔ(α)), and hence π(Nα)=π(NαΔ(α)).

Suppose that N is transitive on V. Then G=NGα. Set X=NGα[1]. Then X is normal in G, Xα=NαGα[1] and Xα[1]=Gα[1]. Note that

XαΔ(α)NαGα[1]/Gα[1]Nα/(NαGα[1])=Nα/Nα[1]NαΔ(α).

Then we have

π(Gα[1])π(Xα)=π(XαΔ(α))=π(NαΔ(α))=π(Nα),

and the lemma follows. ∎

Assume that Δ=(V,A) is a digraph, and G𝖠𝗎𝗍Δ. Suppose that N is an intransitive normal subgroup of G. Let be the set of N-orbits on V. Define a digraph ΔN on such that (B,C) is an arc if and only if BC and (δ,γ)A for some δB and some γC. For B and αB, we have

(2.2)

Δ(α)=(BΔ(α))(CΔN(B)(CΔ(α))),
|Δ(α)||BΔ(α)|+|ΔN(B)|.

Then the following lemma is easily shown.

Lemma 2.3.

Let Δ=(V,A) be a connected digraph, GAutΔ and αV. Assume that G is transitive on A and N is a normal subgroup of G. If N is intransitive on V, then |Δ(α)|=|ΔN(B)||CΔ(α)| for each given arc (B,C) of ΔN with αB; in this case, G induces a group acting transitively on the arc set of ΔN.

Lemma 2.4.

Let Δ=(V,A) be a connected digraph, and GAutΔ. Assume that G is transitive on V and N is an intransitive normal subgroup of G. If |Δ(α)|=|ΔN(B)| for some N-orbit B and αB, then N is semiregular on V.

Proof.

Assume that |Δ(α)|=|ΔN(B)| for some N-orbit B and αB. Then Δ(α)B=, and |Δ(α)C|=1 for every CΔN(B). It follows that Nα fixes Δ(α) point-wise, and hence NαΔ(α)=1. Then we get Nα=1 by Lemma 2.2. By (2.1), Nαg=(Nα)g=1 for each gG. Since G is transitive on V, we have Nβ=1 for all βV, that is, N is semiregular on V. ∎

Let Γ=(V,E) be a G-vertex-transitive graph. It is well known that Γ is G-arc-transitive if and only if for αV the stabilizer Gα acts transitively on Γ(α). Suppose that Γ=(V,E) is G-arc-transitive. Let {α,β}E. Then the arc-stabilizer Gαβ has index 2 in the edge-stabilizer G{α,β}. It follows that G{α,β} has a 2-element x which is not contained in Gαβ; in particular, (α,β)x=(β,α). Thus, if further Γ is connected, then the group G is generated by x and Gα; see [1, p. 118, 17B] for example. Then the following lemma holds.

Lemma 2.5.

Assume that Γ=(V,E) is a connected G-arc-transitive graph. Let {α,β}E. Then there is a 2-element x in G{α,β}Gαβ such that G=x,Gα.

For a group R and SR{1}, the Cayley digraph Cay(R,S) is defined as the digraph with vertex set R and arc set {(x,sx)xR,sS}. The digraph Cay(R,S) admits R, acting by right multiplication, as a regular subgroup of 𝖠𝗎𝗍Cay(R,S). Let

𝖠𝗎𝗍(R,S)={σ𝖠𝗎𝗍(R)Sσ=S}.

Then, by [6, Lemma 2.1], we have 𝐍𝖠𝗎𝗍Cay(R,S)(R)=R:𝖠𝗎𝗍(R,S). In the following lemma, we collect some well-known facts about Cayley digraphs.

Lemma 2.6.

Let Cay(R,S) be a Cayley digraph. Then the following statements hold.

  1. A digraph Δ is isomorphic to Cay(R,S) if and only if 𝖠𝗎𝗍Δ has a regular subgroup isomorphic to R.

  2. Cay(R,S) is connected if and only if S=R, and Cay(R,S) is a graph if and only if S=S-1.

  3. 𝐍𝖠𝗎𝗍Cay(R,S)(R)=R:𝖠𝗎𝗍(R,S), and if S=R, then 𝖠𝗎𝗍(R,S) acts faithfully on S.

3 Proof of Theorem 1.1

In this section, we will work under the following assumption.

Hypothesis 3.1.

Γ=(V,E) is a connected graph of twice prime valency 2r, GAutΓ, and Γ is G-vertex-transitive, G-edge-transitive but not (G,2)-arc-transitive.

Let A be a G-orbit on the arc set of Γ. Since G is transitive on both V and E, the arc set of Γ is equal to either A or AA*. For the former, we may view Γ as the digraph (V,A); otherwise, we view Γ as the arc-disjoint union of digraphs (V,A) and (V,A*).

Lemma 3.2.

Suppose that Hypothesis 3.1 holds and that Γ is G-arc-transitive. Let N be a transitive normal subgroup of G, and let αV. Then one of the following holds.

  1. N is regular on V.

  2. Nα is a nontrivial 2 -group.

  3. r=maxπ(Nα)>2, and

    1. either N is transitive on E,

    2. or N has two orbits E1 and E2 on E such that (V,Ei) are N-arc-transitive graphs of valency r; in this case, |V| is even.

Proof.

By Lemma 2.2, π(Nα)=π(NαΓ(α)). If NαΓ(α)=1, then Nα=1, and thus part (1) of this lemma follows. In the following, we suppose that NαΓ(α)1.

Since NαΓ(α) is normal in GαΓ(α) and GαΓ(α) is transitive on Γ(α), all NαΓ(α)-orbits on Γ(α) have equal length, say l. Then l=2, r or 2r. If l=2, then NαΓ(α) is a 2-group, and so is Nα by Lemma 2.2, thus part (2) of this lemma occurs. Assume that r=2 and l=2r=4. Then |Γ(α)|=4. Since Γ is not (G,2)-arc-transitive, GαΓ(α) is not 2-transitive on Γ(α). Thus we have NαΓ(α)GαΓ(α)D8. Then Nα is a 2-group by Lemma 2.2, and part (2) of this lemma occurs. Thus we suppose that l>2 and r>2 in the following.

Assume that GαΓ(α) is primitive on Γ(α). Then NαΓ(α) is a transitive normal subgroup of GαΓ(α). It follows that N acts transitively on the arc set of Γ. Since GαΓ(α) is not 2-transitive, by [7, Theorem 1.51], GαΓ(α)A5 or S5, and r=5. Then NαΓ(α)A5, and r=5=maxπ(Nα) by Lemma 2.2, and so (i) of part (3) occurs.

Now assume that GαΓ(α) is imprimitive on Γ(α). Then GαΓ(α) is isomorphic to a subgroup of S2Sr or SrS2, which yields that r=maxπ(Gα). Recalling that l=2r or r, we have r=maxπ(Nα). If l=2r, then NαΓ(α) is transitive on Γ(α), which yields that Γ is N-arc-transitive, and so (i) of part (3) occurs. Suppose that l=r. Since N is normal in G and transitive on V, we know that N has exactly two orbits on the arc set of Γ, say A1 and A2, and either A2=A1*A1, or Ai=Ai* for each i. These two cases yield (i) and (ii) of part (3), respectively. ∎

Lemma 3.3.

Suppose that Hypothesis 3.1 holds and that Γ is not G-arc-transitive. Let N be a transitive normal subgroup of G, and let αV. Then

  1. either N is regular on V,

  2. or r=maxπ(Nα) and Γ is N-edge-transitive.

Proof.

Let A be a G-orbit on the arc set of Γ. Then the arc set of Γ is equal to AA*. Consider the digraph Δ=(V,A). Then |Δ(α)|=r, and so

r=maxπ(GαΔ(α))=maxπ(Gα)

by Lemma 2.2. Since NαΔ(α) is normal in GαΔ(α) and GαΔ(α) is transitive on Δ(α), all NαΔ(α)-orbits (on Δ(α)) have equal length. Then either NαΔ(α)=1, or NαΔ(α) is transitive on Δ(α). By Lemma 2.2, the former case yields Nα=1, and so N is regular on V. Assume that NαΔ(α) is transitive on Δ(α). Then r=maxπ(Nα), and N is transitive on A. This implies that N acts transitively on E, and so part (2) of this lemma occurs. ∎

Lemma 3.4.

Suppose that Hypothesis 3.1 holds and r>2. Let αV, and let H be a subgroup of Gα with |H|=r. Assume that N is a transitive normal subgroup of G such that Nα is a nontrivial 2-group. Then Gα[1] is a 2-group, and Γ is NH-arc-transitive.

Proof.

By the assumptions in this lemma and Lemma 3.3, we conclude that Γ is G-arc-transitive, and hence Gα acts transitively on Γ(α). Further, by Lemma 2.2, π(Gα[1])π(Nα), and so Gα[1] is a 2-group. Then HGα[1] as H has odd order r.

Since Nα is normal in Gα, all Nα-orbits on Γ(α) have equal length which is a divisor of 2r. Noting that Nα is a nontrivial 2-group, it follows that the set 𝒞 of Nα-orbits on Γ(α) is a Gα-invariant partition of Γ(α) into subsets of size 2. Recalling that HGα[1] since |𝒞|=r=|H|, we conclude that H acts transitively on 𝒞, and thus NαH acts transitively on Γ(α). Clearly, NH is transitive on V. Noting that (NH)α=NHGα=(NGα)H=NαH, it follows that Γ is NH-arc-transitive. ∎

Lemma 3.5.

Suppose that Hypothesis 3.1 holds and |V| is odd. Let N be an intransitive normal subgroup of G. Then either N is semiregular on V, or ΓN is a cycle.

Proof.

Since |V| is odd and G is transitive on V, the group N cannot have just 2 orbits on V. Then N has at least 3 orbits on V. Let A be a G-orbit on the arc set of Γ, and let Δ=(V,A). Then the digraph ΔN is connected and vertex-transitive, which has at least three vertices. Let B be an N-orbit on V and αB. By Lemmas 2.3 and 2.4, |ΔN(B)| is a divisor of |Δ(α)|, and if |ΔN(B)|=|Δ(α)|, then N is semiregular on V. Thus we assume next that |ΔN(B)|<|Δ(α)|.

Suppose that A is not the arc set of Γ. Then we have |Δ(α)|=r, and hence |ΔN(B)|=1. Thus ΔN is a directed cycle with length at least 3. Similarly, letting Δ*=(V,A*), we know that ΔN* is a directed cycle. Then ΓN is a cycle by noting that ΓN=ΔNΔN*.

Suppose that A is the arc set of Γ. Then ΓN=ΔN is a connected vertex-transitive graph. Recalling that ΓN has order at least 3, we know that ΓN is of valency at least 2. Thus ΓN has valency |ΔN(B)|=2 or r. Note that the latter is impossible with r3 since then we would have a regular graph of odd order and odd valency. Therefore, ΓN is a connected graph of valency 2. Then the lemma follows. ∎

Recall that every minimal normal subgroup of a quasiprimitive permutation group is transitive. Then the following result finishes the proof of Theorem 1.1.

Theorem 3.6.

Suppose that Hypothesis 3.1 holds and |V| is odd. Let N be a minimal normal subgroup of G, and let αV. Assume that N is transitive on V. Then N is the unique transitive minimal normal subgroup of G, and one of the following holds.

  1. Npk for some odd prime p and positive integer kr, and Gα is isomorphic to a subgroup of S2Sr; if further Γ is not G-arc-transitive, then k<r, and Gα is isomorphic to a transitive subgroup of the symmetric group Sr of degree r.

  2. N is a nonabelian simple group.

Proof.

By Lemma 2.1, N is the unique transitive minimal normal subgroup of G. Write N=T1××Tk, where Ti are isomorphic simple groups.

Case 1. Assume that Tip for all i, where p is a prime. Then Npk. Since N is abelian and transitive on V, we know that N is regular on V, and so p is odd. We may identify Γ with a Cayley graph Cay(N,S). Let α be the vertex of Γ corresponding to the identity of N. Since N is normal in G, we have Gα𝖠𝗎𝗍(N,S); see Lemma 2.6. Noting that N=S as Γ is connected, Gα acts faithfully on S. Further, {{s,s-1}sS} is a Gα-invariant partition of S into subsets of size 2. It follows that Gα is isomorphic to a subgroup of the wreath product S2Sr. Let O be a set of representatives of the partition of S. Then N=O, and hence k|O|=|S|2=r.

Suppose that Γ is not G-arc-transitive. Then Gα has two orbits on S. Without loss of generality, we may choose the above O being a Gα-orbit on S; see [14, Lemma 3.2]. Then Gα acts faithfully on O, and so Gα is isomorphic to a transitive subgroup of Sr. Consider the element x=sOs of N. It is easily shown that Gα centralizes x. Then x is a normal subgroup of G. Since xN, by the minimum of N, either k=1, or x=1. Clearly, if k=1, then k<r. If x=1, then N=O{s} for sO, and thus k|O{s}|=r-1<r. Thus part (1) of the theorem follows.

Case 2. Assume that T1,,Tk are nonabelian simple groups. Then N has even order. Since |V|=|N:Nα| is odd, Nα contains a Sylow 2-subgroup of N, where αV. Of course, Nα1 as |N| has even order. Since every Ti is normal in N and N is transitive on V, all Ti-orbits on V have equal length, which is a divisor of |V|. Thus (Ti)α contains a Sylow 2-subgroup of Ti. We shall show part (2) of this theorem holds.

Suppose to the contrary that k>1. Since T1 and T2 centralize each other, if T1 is transitive on V, then T2 is semiregular or acts trivially on V, which is impossible. Thus T1 is intransitive on V. Noting that |V| is odd and N is transitive on V, it follows that T1 has at least three orbits on V. Thus ΓT1 is a connected graph of valency at least 2.

Let K be the kernel of N acting on the set of T1-orbits on V. Then N/K is isomorphic to a subgroup of 𝖠𝗎𝗍ΓT1. Noting that NK, we have TiK for some i, and so TiK=1. Then TiTiK/KN/K, and hence N/K is insoluble. Thus 𝖠𝗎𝗍ΓT1 is insoluble. This implies that ΓT1 is not a cycle. If N acts transitively on E, then, by Lemma 3.5, ΓT1 is a cycle, a contradiction. Thus N is intransitive on E.

Suppose that r=2. Then Γ has valency 4. Recall that ΓT1 is not a cycle. Applying (2.2) to the pair (N,T1), we know that ΓT1 has valency d=3 or 4. If d=4, then, by Lemma 2.4, T1 is semiregular on V, a contradiction. Assume that d=3. Let B be a T1-orbit on V. Then (2.2) implies that |Γ(β)B|=1 for each βB. Then the subgraph of Γ induced by B is a regular graph of valency 1. It follows that |B| is even, which is impossible as |B| is a divisor of |V|. Therefore, r is odd.

Since both |V| and r are odd, by Lemmas 3.2 and 3.3, we conclude that Γ is G-arc-transitive and Nα is a 2-group, and then Nα is a Sylow 2-subgroup of N. Recall that each (Ti)α contains a Sylow 2-subgroup of Ti. It follows that (Ti)α is a Sylow 2-subgroup of Ti, where 1ik. Thus we have

Nα=(T1)α××(Tk)α.

Let HGα with |H|=r. Note that Tih is a normal subgroup of N, where 1ik and hH. Then Tih=Tj for some j; see [10, p. 51, Theorem 9.12] for example. Thus H acts on {T1,,Tk} by conjugation. Set X=NH. By Lemma 3.4, Γ is X-arc-transitive. Further, Xα=NαH, and Nα is the unique Sylow 2-subgroup of Xα.

Let M=T1hhH. Then M is a normal subgroup of X, MN and Mα is a nontrivial 2-group. Suppose that M is intransitive on V. Then, by Lemma 3.5, the quotient graph ΓM is a cycle. Then 𝖠𝗎𝗍ΓM is soluble, and so is N/K, where K is the kernel of N acting on the set of M-orbits on V. This implies that N=K, and N fixes every M-orbit set-wise, a contradiction. Therefore, M is transitive on V. Suppose that MN. Then TiM for some i. Since Ti is simple, we have TiM=1. Noting that both M and Ti are normal in N, we have Ti𝐂N(M). By [5, Theorem 4.2A], Ti is semiregular on V, which is impossible. Thus we get M=N. It follows that H acts transitively on {T1,,Tk} by conjugation. Since H has prime order, H acts regularly on {T1,,Tk}; in particular, k=r. For hH and 1ir, setting Tih=Tj for some j, we have

(Ti)αh=(TiNα)h=TjNα=(Tj)α.

It follows that H acts regularly on {(T1)α,,(Tr)α} by conjugation.

Recall that Γ is X-arc-transitive and Xα=NαH. Then

2r=|Xα:Xαβ|=r|Nα||Xαβ|,

where βΓ(α). Thus |Xαβ|=|Nα|2; in particular, Xαβ is a 2-group. Recalling that Nα is the unique Sylow 2-subgroup of Xα, we have XαβNα. Then Xαβ=Nαβ, and |Nα:Nαβ|=2. By Lemma 2.5, there is a 2-element

xX{α,β}𝐍X(Xαβ)=𝐍X(Nαβ)

such that

x,Xα=X,x2Nαβ.

Noting that H has odd order, we have NHNαH=1. Then

|X|=|NH|=|N||H|.

Since N is normal in X, we know that Nx is a subgroup of X, and hence |Nx| is a divisor of |X|. Noting that |Nx|=|N||x:(Nx)|, it follows that |x:(Nx)| is a divisor of |H|=r. Since x is a 2-group, |x:(Nx)| is a power of 2. It follows that |x:(Nx)|=1, so xN=T1××Tr. Write x=x1x2xr, where xiTi and 1ir. Then

(3.1)X=x,Xαxi,(Ti)α,H1ir.

Consider the projections

ϕi:N=T1××TrTi,t1t2trti.

Setting Li=ϕi(Nαβ) for 1ir, we have

xi𝐍Ti(Li),xi2LiandNαβL1××Lr.

Since NαβNα=(T1)α××(Tr)α, we have Li(Ti)α for 1ir. Then NαβL1××LrNα. Recalling that |Nα:Nαβ|=2, we have

|Nα:(L1××Lr)|2.

It follows that Li=(Ti)α for all but at most one of i. Without loss of generality, we may let Li=(Ti)α for i>1, and |(T1)α:L1|2.

Since (Ti)α is a Sylow 2-subgroup of Ti and xi is a 2-element, if i>1, then xiLi=(Ti)α as (Ti)α is the unique Sylow 2-subgroup of 𝐍Ti(Li). Recalling that H acts regularly on {(T1)α,,(Tr)α} by conjugation, by (3.1), we have

X=x,Xαx1,(T1)α,H.

Set H={1,h2,,hr}. Then

NH=X=(x1,(T1)α×x1,(T1)αh2××x1,(T1)αhr)H.

This implies that T1=x1,(T1)α.

Recalling that |(T1)α:L1|2 and x1𝐍T1(L1), it follows that L1 is normal in x1,(T1)α=T1. Since L1 is a 2-group and T1 is nonabelian simple, we have L1=1. Thus |(T1)α|=2. In particular, T1 has a cyclic Sylow 2-subgroup. By [10, p. 420, Theorem 2.8], T1 has a normal subgroup of odd order, a contradiction. This completes the proof. ∎

We end this section by a consequence from Theorem 3.6.

Corollary 3.7.

Suppose that Hypothesis 3.1 holds and |V|=pk for some odd prime p. If G is quasiprimitive on V, then Γ is arc-transitive, and either

soc(G)pk𝑤𝑖𝑡ℎkr,

or one of the following holds.

  1. p=7, GPSL3(2) and Γ is the complete graph 𝖪7.

  2. p=11, GPSL2(11) and Γ is the complete graph 𝖪11.

  3. GPSU4(2) or PSU4(2).2, and Γ has order 27 and valency 10.

Proof.

Assume that G is quasiprimitive on V. Let N=soc(G). By Theorem 3.6, either N is nonabelian simple, or Npk with kr. Assume that Npk. Then we may write Γ=Cay(N,S), and Gα is faithful on S, where α is the vertex corresponding to the identity of N. Note that {x,x-1}h={xh,(xh)-1} for xS and hGα. Then Gα induces a transitive action on {{x,x-1}xS}. Let σ:NN, yy-1. Then σ is an automorphism of Γ which fixes α, and Gα,σ is transitive on S. Thus Γ is G,σ-arc-transitive.

Assume that N is nonabelian simple. Note that |N:Nα|=|V|=pk for αV. By [8], either N is 2-transitive on V, or NPSU4(2) and Nα has exactly three orbits on V with length 1, 10 and 16, respectively. The latter case implies that Γ is a G-arc-transitive graph of order 27 and valency 10, and then part (3) of this lemma follows. Suppose that N is 2-transitive on V. Then Γ=𝖪pk, and Γ is N-arc-transitive. Note that G and hence N is intransitive on the set of 2-arcs of Γ. It follows that N is not 3-transitive on V. Checking the groups listed in [8, Theorem 1] one by one, we conclude that either NPSL2(11) and |V|=11, or NPSLd(q) and |V|=qd-1q-1, where d is a prime. If NPSL2(11) and |V|=11, then part (2) of this lemma holds. Assume that NPSLd(q) and pk=|V|=qd-1q-1. Then 2r=qd-1q-1-1=qqd-1-1q-1. Noting that r and d are primes, it follows that (d,q)=(2,4) or (3,2), and r=2 or 3, respectively. For (d,q)=(2,4), we have NPSL2(4) and |V|=5; however, as a permutation group of degree 5, the group PSL2(4) is 3-transitive, which is not the case. Thus (d,q)=(3,2), and then part (1) of this lemma follows. ∎

4 Proof of Theorem 1.2

In this section, we will work under the following assumption.

Hypothesis 4.1.

Γ=(V,E) is a connected graph of twice prime valency 2r, |V| is odd, GAutΓ, G is primitive on V, and Γ is G-edge-transitive but not (G,2)-arc-transitive.

Let (Γ,G) be a pair satisfying Hypothesis 4.1. Then G is one of the primitive groups of odd degree given by [17]. Next we give some further information about the graph Γ or its automorphism group. If Γ is a complete graph, then G is 2-homogeneous on V, and the following result holds.

Lemma 4.2.

Suppose that Hypothesis 4.1 holds and that Γ is the complete graph K2r+1. Then one of the following holds.

  1. GAΓL1(pd), r=pd-12, and either d=1, or p=3 and d is an odd prime.

  2. ASLd(3)GAGLd(3), where d is an odd prime.

  3. GPSL2(11) or A7, and r=5 or 7 , respectively.

  4. GPSLd(2) and r=2d-1-1, where d3 and d-1 is a prime.

Proof.

Note that G is 2-homogeneous on V. Since Γ is not (G,2)-arc-transitive, G is not 3-transitive on V.

Suppose that G is an affine primitive group. Let GAGLd(p). Since

pd=|V|=2r+1,

either d=1, or p=3 and d is an odd prime. If GAΓL1(pd), then, by [12], G is 2-transitive, and then GASLd(p) by checking [3, Table 7.3]. Thus one of parts (1) and (2) of this lemma follows.

Suppose that G is almost simple. By [12, Theorem 1], G is 2-transitive on V. Note G is not 3-transitive on V, and r=|V|-12 is a prime. Checking [3, Table 7.4], we conclude that (soc(G),|V|) is one of (PSL2(11),11), (A7,15) or (PSLd(q),qd-1q-1). For the first two pairs, G and r are described as in part (3) of this lemma. Assume that (soc(G),|V|)=(PSLd(q),qd-1q-1). Then

2r=|V|-1=qqd-1-1q-1,

yielding either (d,q)=(2,4) or q=2 and d-1 is a prime. If (d,q)=(2,4) then G is 3-transitive on V, a contradiction. Thus q=2 and d-1 is a prime, and then part (4) of this lemma follows. ∎

Lemma 4.3.

Suppose that Hypothesis 4.1 holds. Let GXAutΓ and αV. Assume that Γ is (X,2)-arc-transitive. Then one of the following holds.

  1. XαΓ(α)A4 or S4, and r=2.

  2. soc(XαΓ(α))A2r, and r3.

  3. soc(XαΓ(α))PSL2(q), and q=5 or q=2r-19.

  4. soc(XαΓ(α))M22, and r=11.

Proof.

Since Γ is (X,2)-arc-transitive, XαΓ(α) is a 2-transitive permutation group of degree 2r. Then the lemma follows from checking the degrees of finite 2-transitive groups; refer to [3, Tables 7.3 and 7.4]. ∎

Lemma 4.4.

Suppose that Hypothesis 4.1 holds. Let GXAutΓ and αV. Suppose that Xα is insoluble and Γ is not (X,2)-arc-transitive. Then Xα has a composition factor isomorphic to one of the following simple groups:

  1. PSL2(11), and r=11,

  2. M11 or M23, and r=11 or 23 , respectively,

  3. PSLd(q), and r=qd-1q-1, where q is power of some prime,

  4. Ar for r5.

Proof.

Let Δ(α) be an Xα-orbit on Γ(α). Then either |Δ(α)|=r or Δ(α)=Γ(α). Since Xα is insoluble, XαΔ(α) is insoluble; refer to [5, Theorem 3.2C].

Suppose that |Δ(α)|=r. Then XαΔ(α) is an insoluble 2-transitive group of prime degree, and (soc(XαΔ(α)),r) is one of (Ar,r), (PSL2(11),11), (M11,11), (M23,23) and (PSLd(q),qd-1q-1); refer to [5, p. 99]. Thus, in this case, Xα has a composition factor described as in this lemma.

Suppose that Δ(α)=Γ(α), that is, Xα acts transitively on Γ(α). Since Γ is not (X,2)-arc-transitive, XαΓ(α) is not 2-transitive on Γ(α). If XαΓ(α) is primitive on Γ(α), then, by [7, Theorem 1.51], r=5 and XαΓ(α)=A5 or S5, and thus the lemma holds. Suppose next that XαΓ(α) is imprimitive on Γ(α). Then Γ(α) has an XαΓ(α)-invariant partition ={B1,B2,Bt} into subsets with size 2 or r. Let K be the kernel of XαΓ(α) acting on . For 1it, let KBi be the permutation group induced by K on Bi. Then KKB1×KB2××KBt, and all KBi are isomorphic to each other.

Assume that |B1|=2. Then K is a 2-group, and XαΓ(α) induces a transitive group (XαΓ(α)) on . Since (XαΓ(α))XαΓ(α)/K, we know that (XαΓ(α)) is insoluble. Then soc((XαΓ(α))) and r are known by [5, p. 99]. Note that

soc(XαΓ(α)/K)soc((XαΓ(α))).

Then XαΓ(α) and hence Xα has a composition factor described as in this lemma.

Assume that |B1|=r. Then ||=2, and so K has index 2 in XαΓ(α). In particular, K is insoluble and, since KKB1×KB2 and KB1KB2, every KBi is insoluble. Thus KB1 is an insoluble 2-transitive group of prime degree r, and so soc(KB1) and r are known. Noting that soc(KB1) is a composition factor of XαΓ(α), the lemma follows. ∎

Theorem 4.5.

Suppose that Hypothesis 4.1 holds. Let GXAutΓ. Then either soc(G)=soc(X), or ΓK2r+1.

Proof.

Assume that Γ is not a complete graph. Then X is not 2-homogeneous on V. By Theorem 1.1, if Γ is not (X,2)-arc-transitive, then X is either affine or almost simple. For the case where Γ is (X,2)-arc-transitive, by [13], X is almost simple.

We next suppose that soc(G)soc(X), and deduce the contradiction.

Assume that either G or X is an affine primitive group (on V). By [19, Propositions 5.1 and 5.2], we have GAGL3(3) and soc(X)PSU4(2). Then we have Xα24:A5 or 24:S5 for αV, and so r=5 by Lemma 2.2; however, 5 is not a divisor of |G|, a contradiction.

Assume next that both G and X are almost simple. Recall that |V| is odd and X is not 2-homogeneous on V. By [19, Proposition 6.1] and [16], all possible triples (soc(G),soc(X),|V|) are listed in Table 1. Let αV.

Table 1
Line(G)(X)|V|X-action, remark
1M_11A_1155,165
2M_12A_12495
3M_22A_22231
4M_23A_23253,1771
5PSL_2(q)A_q+1q(q+1)2-sets, q1(𝗆𝗈𝖽4)
6A_2l-1A_2ll,l-partitions, l3
7PSL_2(11)M_1155
8M_23M_241771
9PSL_m(3)P^+_2m(3)3^m-1(3^m-1)nonsingular points, m odd
10G_2(q)_7(q)q^3(q^31)nonsingular hyperplanes, q±1(𝗆𝗈𝖽4)
11A_12P^-_10(2)495
12J_3PSU_9(2)43605
13_7(3)P^+_8(3)28431
14G_2(3)_7(3)3159

For line 10 of Table 1, by [2, Table 8.39], Xα has a unique insoluble composition factor, say PΩ6-(q) or PΩ6+(q), which contradicts Lemmas 4.3 and 4.4. For line 13 or 14 of Table 1, by the Atlas [4], Xα is an almost simple group with socle PΩ8+(2) or Sp6(2), respectively, and we get a similar contradiction.

Note that rπ(Gα)π(Xα) and, by Lemma 2.2, max(Xα)<2r. This allows us to exclude lines 1–4, 11 and 12 of Table 1. For example, if line 12 of Table 1 occurs, then Gα is a {2,3}-group by the Atlas [4], so r3, yielding that |Xα| has no prime divisor other than 2, 3 and 5, which is impossible as |Xα| is divisible by 11.

Note that Γ(α) is either an Xα-orbit of length 2r or the union of two Xα-orbits of length r. In particular, Xα has a subgroup of index a prime or twice a prime. For lines 7 and 8 of Table 1, the lengths of Xα-orbits on V are known; refer to the webpage edition of [4]. If line 7 of Table 1 occurs, then X is a primitive group (on V) of rank 3, and Xα has three orbits on V with length 1, 18 and 36, respectively, which gives a contradiction. For line 8 of Table 1, Xα has four orbits on V with length 1, 90, 240 and 1440, respectively; we get a similar contradiction. For line 9 of Table 1, by [9], Xα has three orbits on V with length 1, 123m-1(3m-1-1) and 32m-2-1, respectively; however, none of these three numbers has the form of 2r or r.

Suppose that soc(G) and soc(X) are given as in line 6 of Table 1. Then action of G on V is equivalent to that on the set of (l-1)-subsets of {1,2,,2l-1}; in particular, we have |V|=12(2ll). Thus Gα has l orbits on V, which have lengths (l-1i)(ll-1-i), 0il-1, respectively. If r=(l-1i)(ll-1-i) for some i, then i=0 and l=r. If 2r=(l-1i)(ll-1-i) for some i, then either i=0 and l=2r, or i=1 and l=3. For each of these three cases, it is easily shown that 12(2ll) is even, which is not the case as |V| is odd.

Finally, let soc(G) and soc(X) be as in line 5 of Table 1. Then Xα has three orbits on V with length 1, 2(q-1) and 12(q-1)(q-2), respectively. Noting that q1(𝗆𝗈𝖽4), the only possibility is that q=5, soc(G)PSL2(5), 2r=6 and |V|=15. In this case, since |G:Gα|=|V|=15, we know that Gα is a Sylow 2-subgroup of G. Then Gα is not maximal in G, a contradiction. This completes the proof. ∎

Remark 4.6.

For line 8 of Table 1, one may construct a graph of valency 90 and order 1771 which is both M23-arc-transitive and M24-arc-transitive. Thus Theorem 4.5 does not hold without the assumption that Γ has twice prime valency.∎

By Theorem 4.5, we have the following consequence which finishes the proof of Theorem 1.2.

Corollary 4.7.

Suppose that Hypothesis 4.1 holds. Then Γ is 2-arc-transitive if and only if Γ is a complete graph.

Proof.

Note that the complete graph 𝖪2r+1 must be 2-arc-transitive. Thus it suffices to show that Γ is not 2-arc-transitive if Γ𝖪2r+1.

Suppose that Γ𝖪2r+1 and Γ is 2-arc-transitive. Let

N=soc(G)andX=𝖠𝗎𝗍Γ.

Then N=soc(X) by Theorem 4.5. Noting that |V| is odd, N is a nonabelian simple group by Theorem 3.6 and [11].

Let αV. Since |V| is odd, Nα1, and so NαΓ(α)1 by Lemma 2.2. Recalling that Γ is not (G,2)-arc-transitive, GαΓ(α) and hence NαΓ(α) is not 2-transitive on Γ(α). Noting that NαΓ(α) is a proper normal subgroup of XαΓ(α), we have r=2, NαΓ(α)22 by Lemma 4.3. In particular, Γ is a primitive 2-arc-transitive graph of valency 4. Noting that NX, it follows from [15, Theorem 1.5] that (X,Xα) is one of (PGL2(p),S4) and (S7,S4×S3), where p is a prime. Thus we have |X:N|=2 as N=soc(X). Since |X:Xα|=|V|=|N:Nα|, we have 2=|X:N|=|Xα:Nα|, and so |Nα|=12|Xα|. Noting that |Xα| is divisible by 3, it follows that |Nα| has a divisor 3. Then, by Lemma 2.2, NαΓ(α) is not a 2-group, a contradiction. This completes the proof. ∎

We end this section by another consequence of Theorem 4.5.

Corollary 4.8.

Suppose that Hypothesis 4.1 holds and ΓK2r+1. Then either Γ is G-arc-transitive, or r3 and AutΓ has a subgroup of index at most 2 which acts transitively on the edge set but not transitively on the arc set of Γ.

Proof.

Assume that Γ is not G-arc-transitive. Let αV. Then Gα has two orbits on Γ(α) with equal length r.

Suppose that r=2. Then Γ has valency 4. Since G is primitive on V, by [18, Theorem 5], |Gα|=2, |V|=p and G is the dihedral group of order 2p, where p is a prime. In particular, G has a regular normal subgroup R of order p. Thus we may write Γ=Cay(R,S), and Gα𝖠𝗎𝗍(R,S). Set R=a and Gα=h. Since Γ is G-edge-transitive but not G-arc-transitive, by [14, Lemma 3.2], we set S={a,ai,a-1,a-i} such that {a,ai}h={a,ai}, where 1<i<p-1. This yields that ah=aia-1. It follows that h does not have order 2, a contradiction. Therefore, r3.

Note that G has two orbits on the arc set of Γ, say

ΔandΔ*={(β,α)(α,β)Δ}.

Let X=𝖠𝗎𝗍Γ. By Theorem 4.5, N:=soc(G)=soc(X). Since N is normal in X and fixes both Δ and Δ* set-wise, X has an action on {Δ,Δ*}. Let K be the kernel of X acting on {Δ,Δ*}. Then |X:K|2, and K is not transitive on the arc set of Γ. Clearly, GK. Thus K acts transitively on the edge set of Γ, and the result follows. ∎

5 Examples and a proof of Theorem 1.3

We first construct two graphs involved in Theorem 1.3.

Let G be a finite group and HG with gXHg=1. Then G acts (faithfully) on the set [G:H] of right cosets of H in G by right multiplication

g:HxHxgfor allx,gG.

Take a 2-element xGH with x2H. Then x normalizes K:=HHx. Define a graph 𝖢𝗈𝗌(G,H,x)=(V,E) by

V=[G:H],E={{Hg1,Hg2}g1,g2G,g2g1-1HxH}.

Then 𝖢𝗈𝗌(G,H,x) is G-arc-transitive of order |G:H| and valency |H:K|. Moreover, it is well known that 𝖢𝗈𝗌(G,H,x) is connected if and only if x,H=G.

For a positive integer n3, we use D2n to denote the dihedral group of order 2n.

Example 5.1.

(1) Let

G=PGL(2,7),N=PSL(2,7)andD16HG.

Let 22KHN. Then 𝐍H(K)D8 and 𝐍G(K)=𝐍N(K)S4.

Take an involution x𝐍N(K)H. Then Γ=𝖢𝗈𝗌(G,H,x) is a connected G-arc-transitive graph of valency 4 and order 21. The graph Γ is in fact the edge-disjoint union of two N-arc-transitive graphs of valency 2, and each of them is the vertex-disjoint union of 7 cycles of length 3.

(2) Let

G=PGL(2,9),N=PSL(2,9)andD16HG.

Take 22KHN. Then 𝐍H(K)D8 and 𝐍G(K)=𝐍N(K)S4. For an involution x𝐍N(K)H, we have a G-arc-transitive graph Γ=𝖢𝗈𝗌(G,H,x) of valency 4 and order 45. Further, the graph Γ is the edge-disjoint union of two N-arc-transitive graphs of valency 2, and each of them is the vertex-disjoint union of 15 cycles of length 3.∎

Lemma 5.2.

Let Γ=(V,E) be a connected graph of odd order and valency 4, and let GAutΓ. Suppose that G is almost simple, primitive on V and transitive on E. Then either Γ is soc(G)-arc-transitive, or one of the following holds.

  1. 𝖠𝗎𝗍Γ=GPGL(2,7), and Γ is isomorphic to the graph in Example 5.1(1).

  2. GPGL(2,9), M10 or PΓL(2,9), and Γ is isomorphic to the graph in Example 5.1(2).

Proof.

By [15], Γ is arc-transitive and isomorphic to one of the graphs listed in [15, Table 2]. Let N=soc(G). Then N is nonabelian simple and transitive on V. Since |V| is odd, Nα1 for αV, and so NαΓ(α)1 by Lemma 2.2.

Assume that Γ is (G,2)-arc-transitive. Then GαΓ(α) is 2-transitive on Γ(α), and hence NαΓ(α) is a transitive normal subgroup of GαΓ(α). It follows that Γ is N-arc-transitive.

Assume that Γ is not (G,2)-arc-transitive. Then Hypothesis 4.1 holds. Suppose that Γ is 2-arc-transitive. Then Γ𝖪5 by Corollary 4.7, and so 𝖠𝗎𝗍ΓS5. Noting that G is an insoluble subgroup of 𝖠𝗎𝗍Γ, we have GA5 or S5. Then G acts transitively on the set of 2-arcs of Γ, a contradiction. Therefore, Γ is not 2-arc-transitive. Noting that 𝖠𝗎𝗍(A6)PΓL2(9), by [15, Theorem 1.5], we have 𝖠𝗎𝗍ΓPGL2(7), PΓL2(9) or PSL2(17). Further, by Theorem 1.2, we have N=soc(G)=soc(𝖠𝗎𝗍Γ). For 𝖠𝗎𝗍ΓPSL2(17), we have N=GPSL2(17), and so Γ is N-arc-transitive. For 𝖠𝗎𝗍ΓPGL2(7) or PΓL2(9), we get (1) or (2) of this lemma, respectively. ∎

Proof Theorem 1.3.

Let Γ=(V,E) be a connected graph of odd order and twice prime valency 2r. Let G𝖠𝗎𝗍Γ be such that Γ is G-edge-transitive but not (G,2)-arc-transitive. Assume that G is almost simple and primitive on V. By Lemma 5.2, Theorem 1.3 holds for r=2. Thus we assume further that r3 in the following.

Let N=soc(G). Then N is nonabelian simple, and N is transitive but not regular on V. Let αV. By Lemmas 3.2 and 3.3, either Γ is N-edge-transitive, or Nα is a 2-group and Γ is G-arc-transitive. We shall show that the latter case does not occur, and then Theorem 1.3 follows.

Suppose that Nα is a 2-group and Γ is G-arc-transitive. Since |V|=|N:Nα| is odd, Nα is a Sylow 2-subgroup of N. Since G is a primitive group of odd degree, using the classification given in [17], we conclude that NPSL(2,q) and Nα is a dihedral group. It follows from [2, Table 8.1] that NαDq-1 or Dq+1.

Assume that |Nα|8. Then q=7 or 9, and NPSL(2,7) or PSL(2,9), respectively. Checking the maximal subgroups of G in the Atlas [4], we conclude that Gα is a (Sylow) 2-subgroup of G. Since Γ is G-arc-transitive and of valency 2r, we get r=2, a contradiction.

Assume next that |Nα|>8. Then Nα has a unique cyclic subgroup of index 4, which is equal to the Frattini subgroup Φ of Nα. Clearly, Φ1 and Φ is a characteristic subgroup of Nα. Since Nα is normal in Gα, all Nα-orbits on Γ(α) have equal length which is a divisor of 2r. Thus, since Nα is a nontrivial 2-group, every Nα-orbit on Γ(α) has length 2. Let βΓ(α). Then |Nα:Nαβ|=2. In particular, Nαβ is a maximal subgroup of Nα, and so ΦNαβ. Noting that Nαβ is either a cyclic group or a dihedral group (of order |Nα|2). It follows that Φ is a characteristic subgroup of Nαβ.

Let HGα with |H|=r, and set X=NH. By Lemma 3.4, Γ is X-arc-transitive. Further, we have Xα=NHGα=(NGα)H=NαH. Since r is odd and Nα is a 2-group, we have NαH=1, and so |Xα|=|Nα||H|=r|Nα|. Then 2r=|Xα:Xαβ|=r|Nα||Xαβ|, yielding |Xαβ|=|Nα|2=|Nαβ|. It follows that Xαβ=Nαβ.

Since Γ is connected and X-arc-transitive, by Lemma 2.5, X=x,NαH for some 2-element x𝐍X(Nαβ). Recall that Φ is characteristic in both Nα and Nαβ. Then x normalizes Φ, and H also normalizes Φ as Nα is normal in NαH. It follows that Φ is normal in X=x,NαH. Thus Φ is normal in N, which contradicts the fact that N is a nonabelian simple group. This completes the proof of Theorem 1.3. ∎


Communicated by Michael Giudici


Award Identifier / Grant number: 11731002

Award Identifier / Grant number: 11971248

Funding statement: This work was partially supported by the National Natural Science Foundation of China (11731002, 11971248) and the Fundamental Research Funds for the Central Universities.

Acknowledgements

The authors are grateful to the referees for a number of valuable comments for improvements in the paper.

References

[1] N. Biggs, Algebraic Graph Theory, Cambridge Tracts in Math. 67, Cambridge University, London, 1974, 10.1017/CBO9780511608704Search in Google Scholar

[2] J. N. Bray, D. F. Holt and C. M. Roney-Dougal, The Maximal Subgroups of the Low-dimensional Finite Classical Groups, London Math. Soc. Lecture Note Ser. 407, Cambridge University, Cambridge, 2013. 10.1017/CBO9781139192576Search in Google Scholar

[3] P. J. Cameron, Permutation Groups, London Math. Soc. Stud. Texts 45, Cambridge University, Cambridge, 1999. 10.1017/CBO9780511623677Search in Google Scholar

[4] J. H. Conway, R. T. Curtis, S. P. Norton, R. A. Parker and R. A. Wilson, Atlas of Finite Groups. Maximal Subgroups and Ordinary Characters for Simple Groups, Oxford University, Eynsham, 1985. Search in Google Scholar

[5] J. D. Dixon and B. Mortimer, Permutation Groups, Grad. Texts in Math. 163, Springer, New York, 1996. 10.1007/978-1-4612-0731-3Search in Google Scholar

[6] C. D. Godsil, On the full automorphism group of a graph, Combinatorica 1 (1981), no. 3, 243–256. 10.1007/BF02579330Search in Google Scholar

[7] D. Gorenstein, Finite Simple Groups. An Introduction to Their Classification, Univ. Ser. Math., Plenum, New York, 1982. 10.1007/978-1-4684-8497-7Search in Google Scholar

[8] R. M. Guralnick, Subgroups of prime power index in a simple group, J. Algebra 81 (1983), no. 2, 304–311. 10.1016/0021-8693(83)90190-4Search in Google Scholar

[9] J. I. Hall and H. N. Nguyen, On the permutation modules for orthogonal groups Om±(3) acting on nonsingular points of their standard modules, J. Pure Appl. Algebra 216 (2012), no. 3, 581–592. 10.1016/j.jpaa.2011.07.011Search in Google Scholar

[10] B. Huppert, Endliche Gruppen. I, Grundlehren Math. Wiss. 134, Springer, Berlin, 1967. 10.1007/978-3-642-64981-3Search in Google Scholar

[11] A. A. Ivanov and C. E. Praeger, On finite affine 2-arc transitive graphs, European J. Combin. 14 (1993), 5, 421–444. 10.1006/eujc.1993.1047Search in Google Scholar

[12] W. M. Kantor, k-homogeneous groups, Math. Z. 124 (1972), 261–265. 10.1007/BF01113919Search in Google Scholar

[13] C. H. Li, On finite s-transitive graphs of odd order, J. Combin. Theory Ser. B 81 (2001), no. 2, 307–317. 10.1006/jctb.2000.2012Search in Google Scholar

[14] C. H. Li, Z. Liu and Z. P. Lu, The edge-transitive tetravalent Cayley graphs of square-free order, Discrete Math. 312 (2012), no. 12–13, 1952–1967. 10.1016/j.disc.2012.03.013Search in Google Scholar

[15] C. H. Li, Z. P. Lu and D. Marušič, On primitive permutation groups with small suborbits and their orbital graphs, J. Algebra 279 (2004), no. 2, 749–770. 10.1016/j.jalgebra.2004.03.005Search in Google Scholar

[16] M. W. Liebeck, C. E. Praeger and J. Saxl, A classification of the maximal subgroups of the finite alternating and symmetric groups, J. Algebra 111 (1987), no. 2, 365–383. 10.1016/0021-8693(87)90223-7Search in Google Scholar

[17] M. W. Liebeck and J. Saxl, The primitive permutation groups of odd degree, J. Lond. Math. Soc. (2) 31 (1985), no. 2, 250–264. 10.1112/jlms/s2-31.2.250Search in Google Scholar

[18] P. M. Neumann, Finite permutation groups, edge-coloured graphs and matrices, Topics in Group Theory and Computation (Galway 1973), Academic Press, New York (1977), 82–118. Search in Google Scholar

[19] C. E. Praeger, The inclusion problem for finite primitive permutation groups, Proc. Lond. Math. Soc. (3) 60 (1990), no. 1, 68–88. 10.1112/plms/s3-60.1.68Search in Google Scholar

[20] C. E. Praeger, An O’Nan–Scott theorem for finite quasiprimitive permutation groups and an application to 2-arc transitive graphs, J. Lond. Math. Soc. (2) 47 (1993), no. 2, 227–239. 10.1112/jlms/s2-47.2.227Search in Google Scholar

[21] C. E. Praeger, Finite quasiprimitive graphs, Surveys in Combinatorics (Galway 1973), London Math. Soc. Lecture Note Ser. 241, Cambridge University, Cambridge (1997), 65–85. 10.1017/CBO9780511662119.005Search in Google Scholar

[22] C. E. Praeger and M. Y. Xu, A characterization of a class of symmetric graphs of twice prime valency, European J. Combin. 10 (1989), no. 1, 91–102. 10.1016/S0195-6698(89)80037-XSearch in Google Scholar

Received: 2019-06-29
Revised: 2020-06-01
Published Online: 2020-07-16
Published in Print: 2020-11-01

© 2021 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 26.4.2024 from https://www.degruyter.com/document/doi/10.1515/jgth-2019-0091/html
Scroll to top button