389
Views
0
CrossRef citations to date
0
Altmetric
Research Articles

Completions of the Goldschmidt G3-amalgam and alternating groups

&
Pages 1186-1200 | Received 03 Aug 2020, Accepted 27 Sep 2022, Published online: 20 Oct 2022

Abstract

Here we prove that the alternating group of degree n is a faithful completion of the Goldschmidt G3-amalgam if and only if

n{1,2,3,4,5,7,8,9,11,12,16,17,19,23}.

2020 Mathematics Subject Classification:

1 Introduction

In [Citation3] Goldschmidt classified the primitive rank 2 amalgams of index (3, 3). These amalgams are now referred to as Goldschmidt amalgams. We recall that a rank 2 amalgam consists of groups P1, P2, and B together with group monomorphisms φ1,φ2 such thatφ1:BP1andφ2:BP2.

Suppressing the φi, we refer to this amalgam as A(P1,P2,B). To say A(P1,P2,B) is primitive means that if KB and φi(K)EPi for i = 1, 2, then K = 1. If P1 and P2 are finite groups, then A(P1,P2,B) is called a finite amalgam whose index is ([P1:φ1(B)],[P2:φ2(B)]). A group G is said to be a completion of A(P1,P2,B) if there exist group homomorphisms ψi:PiG i = 1, 2, for which ψ1φ1=ψ2φ2:BG and G=imψ1,imψ2. Should the ψi,i=1,2, be monomorphisms we then say this is a faithful completion. Here we shall only consider faithful completions, and so from now on will identify Pi with ψi(Pi) and B with ψiφi(B). Hence we have PiG,i=1,2 with P1P2B and P1,P2=G (though in all the completions we deal with we will have P1P2=B). We remark that faithful completions of these Goldschmidt amalgams furnish, via their coset graphs, many examples of trivalent graphs.

The amalgams arrayed in [Citation3] consist of fifteen amalgams, among which is the Goldschmidt G3-amalgam. This amalgam has P1Sym(4)P2 and BDih(8). A related Goldschmidt amalgam we shall briefly encounter is the G31-amalgam. In this case we have P1Sym(4)×Z2P2 with BDih(8)×Z2. By Sym(n) we mean the symmetric group of degree n and Dih(8) the dihedral group of order 8. We also use Alt(n) to denote the alternating group of degree n. Some small examples of faithful completions of these amalgams are GL3(2),Alt(6), and Sym(6). Taking P1 and P2 to be the two standard maximal parabolic subgroups of GL3(2) yields a completion of the G3-amalgam. In Sym(6), letting P1 and P2 be, respectively, the centralizers of (1, 2) and (1,2)(3,4)(5,6), we see that Sym(6) is a completion of the G31-amalgam. Intersection of P1 and P2 with Alt(6) reveals that Alt(6) is a completion of the G3-amalgam.

It is completions of the G3-amalgam we study here. PuttingE={1,2,3,4,5,7,8,9,11,12,16,17,19,23}we state our main result.

Theorem 1.1.

Suppose that nN. Then Alt(n) is a completion of the Goldschmidt G3-amalgam if and only if nE.

Earlier work on the Goldschmidt G3-amalgam for alternating groups [Citation7] verified that for all n24,Alt(n) is a completion precisely when n{6,10,13,14,15,18,20,21,22,24}. In this paper we shall demonstrate that Alt(n) is a completion for all n25 from which Theorem 1.1 will then follow. Conder, in [Citation2], investigated whether G=Sym(n) is the automorphism group of some finite connected 5-arc transitive graph. Using the work of Lorimer [Citation4], Conder rephrases his problem to ask whether there exists HSym(4)×Z2G and aG such that

  1. a2H;

  2. G=HaH; and

  3. [H:HHa]=3.

If such a subgroup H and element a can be found, then Sym(n) is a completion of the Goldschmidt G31-amalgam, with P1=H,P2=Ha and B=HHa, which then implies that Alt(n) is a completion of the Goldschmidt G3-amalgam (see Lemma 2.3). Now Conder’s method is to assemble various transitive representations for Sym(4)×Z2 on 2,3,6,12,24, and 48 points and then taking appropriate combinations of these upon which a permutation that will fulfill the role of a is defined. This is achieved for all n of the form n=84b+176c+177d+87, where b,c,dN. Since the Frobenius number of {84, 176, 177}- the largest integer such that it cannot be formed as 84b+176c+177d for b,c,d0 - is 2743, it means that, taking into account the additional 87, for all n2831, Sym(n) and Alt(n) are respective completions of the Goldschmidt G31- and G3- amalgams.

We briefly survey what else is known about completions of the Goldschmidt G3-amalgam. Which sporadic simple groups (with the sole exception of the Monster) are completions is settled in [Citation6] and [Citation7]. While in [Citation5] we have that SL3(q) and PSL3(q) are completions of the Goldschmidt G3-amalgam if and only if q{4,9}. For the unitary and orthogonal groups, SU3(q) and PSU3(q) are completions if and only if q is odd and q{3,5} and SO3(q)(PSL2(q)) is a completion if and only if q or q is a prime and q±1(mod8).

This paper is arranged as follows—Section 2 begins with two elementary results concerning completions of the Goldschmidt G3-amalgam. Then we introduce a particular type of graph, the orbit graph O(P1,P2,Ω). Here P1 and P2 are subgroups of Sym(Ω), the group of permutations on Ω. This graph, courtesy of Lemma 2.6, can be used to determine whether P1,P2 is transitive on Ω or not. Definition 2.8 introduces the idea of twisting P1 with respect to (Δ,Ψ) where Δ and Ψ are certain (P1P2)-orbits of disjoint sets Ω and Γ. This idea, via Lemma 2.9, plays a central role in the recursive construction, presented in Section 3, and which underlies the proof of Theorem 1.1. We end Section 2 with Theorem 2.10, a classical result of Jordan’s.

In Section 3 we give a detailed account of a recursive construction for the case when n0(mod24). For the remaining congruences we list, in Section 4, the seed permutations which enable this construction to produce an example of Alt(n) as a completion of the Goldschmidt G3-amalgam. Further details on these seed permutations may be found in Section 3.

2 Some preliminary results

We begin with a hypothesis which lists all relevant facts about completions of the Goldschmidt G31-amalgam which we use.

Hypothesis 2.1.

When G is a (faithful) completion of the Goldschmidt G3-amalgam this means that G contains subgroups P1 and P2 such that

  1. P1Sym(4)P2 and B:=P1P2Dih(8);

  2. G=P1,P2; and

  3. O2(P1)=O2(P2) (where O2(Pi)Z2×Z2 is the largest normal 2-subgroup of Pi).

Our first result concerns involutions and completions of the Goldschmidt G3-amalgam.

Lemma 2.2.

Suppose the group G is a completion of the Goldschmidt G3-amalgam A(P1,P2,B). Then the involutions in P1P2 are G-conjugate.

Proof.

By Hypothesis 2.1 we have PiG,i=1,2 with P1P2=BDih(8) and PiSym(4). Now, for i = 1, 2 Pi has two conjugacy classes of involutions Ci(1) and Ci(2) with BCi(1)=O2(Pi). Since O2(P1)O2(P2) and O2(P1)O2(P2)1, Lemma 2.2 follows. □

Next we see that a completion of the Goldschmidt G31-amalgam, A(Sym(4)×Z2,Sym(4)×Z2,Dih(8)×Z2), for Sym(n) yields a completion of the Goldschmidt G3-amalgam for Alt(n).

Lemma 2.3.

If Sym(n) is a completion of the Goldschmidt G31-amalgam, then Alt(n) is a completion of the Goldschmidt G3-amalgam.

Proof.

Set G=Sym(n) and H=Alt(n). Let P1,P2G be such that PiSym(4)×Z2 and B=P1P2Dih(8)×Z2 with P1,P2=G. Also put Ri=O2(Pi)Alt(4) and Qi=O2(Ri)Z2×Z2. Observe that [B:BH]=2 for otherwise BH and then P1,P2HG, a contradiction. Further we must have n5. Note that RiH,i=1,2 and that R1 and R2 normalize each other. Since A(P1,P2,B) is primitive, Q1Q2. From QiBH,i=1,2, we then see that Q1Q2=BH. So Q1Q2 has order 8 and either Q1Q2 is abelian or Q1Q2Dih(8). If the former holds, then Q1Q2=CPi(Qi)=O2(Pi) for i = 1, 2, giving the impossible O2(P1)=O2(P2). Therefore, BH=Q1Q2Dih(8). Since [Pi:PiH]=2 for i = 1, 2, we have PiHAlt(4)×Z2 or Sym(4). The former possibility doesn’t contain a Dih(8) subgroup and therefore PiHSym(4). To complete the proof we must show that H=P1H,P2H. Now B normalizes PiH,i=1,2 and soP1H,P2HEP1H,P2H,B=G.

We recall that n5 so Alt(n) is simple, whence H=P1H,P2H. □

In proving Theorem 1.1, for specified subgroups P1 and P2 of G=Alt(n) we need to show that P1,P2=G. Not surprisingly, our first step is to prove that P1,P2 acts transitively on Ω={1,,n}. This leads us to consider a particular type of graph which we now define.

Definition 2.4.

Suppose G is a finite group acting on a finite set Ω with P1 and P2 being subgroups of G. Let {Δ1(1),,Δl(1)} and {Δ1(2),,Δk(2)} be, respectively, the P1- and P2-orbits on Ω (and we shall regard these sets as being disjoint). The orbit graph of P1 and P2, O(P1,P2,Ω), has vertex set {Δ1(1),,Δl(1),Δ1(2),,Δk(2)} with distinct vertices Δi(1) and Δj(2) adjacent whenever their intersection contains a (P1P2)-orbit and the number of edges between them is the number of (P1P2)-orbits in Δi(1)Δj(2).

We observe that O(P1,P2,Ω) is a bipartite graph and that Δi(1) and Δj(2) being adjacent is equivalent to Δi(1)Δj(2) (as subsets of Ω). We next give two examples of Definition 2.4—note that in Example 2.5(ii) we have that P1 and P2 have the same orbit (namely Ω), but we consider them to be distinct vertices of O(P1,P2,Ω).

Example 2.5.

  1. Ω={1,,21},P1=x,y,z,P2=x,y,w and B=P1P2=x,y where

x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16);y:=(1,8)(2,3)(4,5)(6,7)(10,11)(13,17)(14,18)(19,20);z:=(1,11)(2,12)(5,9)(6,10)(7,8)(13,15)(14,16)(20,21); andw:=(1,17)(4,18)(5,14)(6,7)(8,13)(10,19)(11,20)(16,21).

Here, the B-orbits on Ω are Ω1:={1,2,3,4,5,6,7,8},Ω2:={9,10,11,12},\\Ω3:={13,14,17,18},Ω4:={15,16},Ω5:={19,20} and Ω6:={21}. The P1-orbits are Δ1(1):=Ω1Ω2,Δ2(1):=Ω3Ω4,Δ3(1):=Ω5Ω6 and the P2-orbits are Δ1(2):=Ω1Ω3,Δ2(2):=Ω2Ω5,Δ3(2):=Ω4Ω6. So the orbit graph O(P1,P2,Ω) is as given in .

  1. Ω={1,,24},P1=x,y,z,P2=x,y,w and B=P1P2=x,y wherex:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24);y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(17,24)(18,19)(20,21)(22,23);

z:=(1,9)(2,10)(3,20)(4,19)(5,14)(6,13)(7,23)(8,24)(11,21)(12,22)(15,18)(16,17); andw:=(1,14)(2,19)(3,18)(4,10)(5,11)(6,22)(7,23)(8,15)(9,17)(12,21)(13,20)(16,24).

In this case, the B-orbits on Ω are {1,,8},{9,,16} and {17,,24} with both P1 and P2 acting transitively on Ω. Thus O(P1,P2,Ω) is as shown in .

Fig. 1 O(P1,P2,Ω) for Example 2.5(i).

Fig. 1 O(P1,P2,Ω) for Example 2.5(i).

Fig. 2 O(P1,P2,Ω) for Example 2.5(ii).

Fig. 2 O(P1,P2,Ω) for Example 2.5(ii).

Example 2.5

(ii) plays a central role in proving Theorem 1.1.

Lemma 2.6.

Suppose P1,P2Sym(Ω), Ω a finite set. Then P1,P2 is transitive on Ω if and only if O(P1,P2,Ω) is connected.

Proof.

Let {Δ1(1),,Δl(1)} and {Δ1(2),,Δk(2)} be, respectively, the P1- and P2-orbits on Ω. Suppose P1,P2 acts transitively on Ω and let Δi(1) and Δj(2) be vertices of O(P1,P2,Ω). Choose αΔi(1) and βΔj(2). Then αg=β for some gP1,P2. Now g=x1x2xm where xiP1P2, and, as x1P1 implies Δi(1)x1=Δi(1), we may assume x1P2. Put αx1=γ and let Δn(2) be the P2-orbit such that γΔn(2). Then α=γx11Δi(1)Δn(2) whence Δi(1) and Δn(2) are adjacent in O(P1,P2,Ω). Continuing in this fashion yields a path from Δi(1) to Δj(2). Therefore O(P1,P2,Ω) is connected.

Now suppose that O(P1,P2,Ω) is connected, and let α,βΩ. Let αΔi1(1),βΔik(1). We may find a sequence of P1- and P2-orbits in which the l th and (l+1) th orbits have non-empty intersections. This then gives gP1,P2 for which αg=β, and so P1,P2 is transitive on Ω. □

Hypothesis 2.7.

Suppose G and H are subgroups of, respectively, Sym(Ω) and Sym(Γ) where Ω and Γ are finite sets which are disjoint. Assume that G contains subgroups Q1 and Q2 with C=Q1Q2 and that H contains subgroups R1 and R2 with D=R1R2. Further assume that θi:QiRi,i=1,2, are isomorphisms for which θ1(x)=θ2(x) for all xC. For xQi,i=1,2 we regard x as a permutation on Ω̇Γ by letting x act as the identity on Γ and θi(x) as a permutation on Ω̇Γ by letting θi(x) act as the identity on Ω. Define subgroups B,Pi,i=1,2 of Sym(Ω̇Γ) by Pi={xθi(x)|xQi} and B={xθi(x)|xC}.

We sometimes refer to P1,P2,B as being the concatenation of Q1,Q2,C and R1,R2,D. Note that P1Q1R1,P2Q2R2 and BCD.

Definition 2.8.

Assume Hypothesis 2.7 and let Δ and Ψ be B-orbits of, respectively, Ω and Γ which are isomorphic B-sets. Now let g=g(Δ,Ψ)Sym(Ω̇Γ) be an involution in CSym(Ω̇Γ)(B) which interchanges Δ and Ψ and fixes all other points of Ω̇Γ. Then we refer to the conjugate P1g as twisting P1 (with respect to (Δ,Ψ)).

We note in Definition 2.8 that P1gP2=B.

Lemma 2.9.

Assume Hypothesis 2.7 holds, and use the notation given there. Let g=g(Δ,Ψ) where Δ and Ψ are B-orbits of Ω and Γ respectively which are isomorphic B-sets. Suppose P1,P2 is transitive on both Ω and Γ. Then P1g,P2 is transitive on Ω̇Γ if and only if the edge in O(P1,P2,Ω̇Γ) corresponding to Δ or the edge in O(P1,P2,Ω̇Γ) corresponding to Ψ is part of a cycle in the graph O(P1,P2,Ω̇Γ).

Proof.

By Lemma 2.6, as P1,P2 is transitive on Ω and Γ, O(Q1,Q2,Ω) and O(R1,R2,Γ) are both connected graphs. Suppose ΔΔp(1)Δq(2) and ΨΨr(1)Ψs(2) where Δp(1),Δq(2) are, respectively P1- and P2-orbits of Γ and Ψr(1),Ψs(2) are respectively P1- and P2-orbits of Γ. So Δp(1) and Δq(2) are adjacent in O(Q1,Q2,Ω) and Ψr(1) and Ψs(2) are adjacent in O(R1,R2,Γ). Since g=g(Δ,Ψ) fixes all points in (Ω̇Γ)\(ΔΨ),P1g has the same orbits on Ω̇Γ as P1 except Δp(1) and Ψr(1) get replaced by Δp(1)=(Δp(1)\Δ)Ψ and Ψr(1)=(Ψr(1)\Ψ)Δ. Thus ΔΔq(2)Ψr(1) and ΨΨs(2)Δp(1), whence Δq(2) and Ψr(1) are adjacent, as are Ψs(2) and Δp(1) in O(P1,P2,Ω̇Γ). Since P1g,P2 is transitive on Ω̇Γ if and only if O(P1g,P2,Ω̇Γ) is connected, the lemma now follows. □

The following theorem is a classical one of Jordan’s, which appears as Theorem 13.9 in [Citation9]. This theorem suffices for our work. However, in [Citation8], a generalization of this result is needed to deal with some of the other cases.

Theorem 2.10.

(Jordan) Let G be a primitive permutation group of degree n with a cycle of degree pλ>1 where p is prime. If n>pλ+4, then G=Alt(n) or G=Sym(n).

3 A recursive construction

Put G=Alt(n) and let Ω={1,,n} be the standard G-set. Bearing in mind Lemma 2.2, we are looking for G-conjugate involutions which generate appropriate Sym(4) subgroups of G. Thus we seek conjugate involutions x, y, z, and w of G such that the following is satisfied.

Hypothesis 3.1.

  1. (xy)2 is conjugate to x;

  2. zNG(x,(xy)2) and zx = xz;

  3. wNG(y,(xy)2) and wy = yw;

  4. zy and wx both have order 3; and

  5. G=x,y,z,w

These conditions will ensure that for B=x,y,P1=x,y,z and P2=x,y,w we have P1P2=BDih(8) and PiSym(4),i=1,2.

As indicated earlier Example 2.5(ii) will be important, one reason being given in our next result.

Lemma 3.2.

Let x,y,z,w,P1, and P2 be as in Example 2.5(ii). Then PiSym(4),i=1,2,P1P2Dih(8) and G=P1,P2. Thus (P1, P2) yields Alt(24) as a completion of the Goldschmidt G3-amalgam.

Proof.

It is straightforward to verify Hypothesis 3.1(i)–(iv) for x,y,z,w. So it remains to show that (v) holds. Since O(P1,P2,Ω) is connected, Lemma 2.6 implies P1,P2 acts transitively on Ω. Calculation reveals that h=[w,z]xy[w,z] has cycle type 51.191. So h5P1,P2 is a 19-cycle. Hence, all the points of Ω in this 19-cycle either lie in the same block of a P1,P2-system of imprimitivity or they all lie in separate blocks. So any P1,P2-system of imprimitivity must either have block size at least 19 or at least 19 blocks. Since P1,P2 acts transitively on Ω and |Ω|=24, we conclude that P1,P2 acts primitively on Ω. Furthermore, the presence of a 19-cycle, by Theorem 2.10, confirms x,y,z,w=G, so proving the lemma. □

We are now in a position to expand on the details of the recursive construction. Now Example 2.5(ii) shows O(P1,P2,Ω) where Ω={1,,24} has cycles and hence Lemma 2.9 becomes available to us. Put Gm=Alt(m), acting on Ωm={1,,m} and suppose Q1(m),Q2(m),C(m) are subgroups of Gm with Qi(m)Sym(4),i=1,2 and Q1(m)Q2(m)=C(m)Dih(8) with Q1(m),Q2(m)=Gm (so Gm is a completion of the Goldschmidt G3-amalgam). Also let H=Alt(24) with H acting upon Γ={m+1,,m+24} with R1,R2,D subgroups of H such that (R1,R2,D) is the completion given in Example 2.5 (RiSym(4),R1R2=DDih(8)). Since the D-orbits on Γ are regular, to employ Definition 2.8 we require that C(m) has a regular orbit on Ωm . So we further assume that this is the case, letting Δ={1,,8} be such a regular C(m)-orbit (we can achieve this by conjugating the Qi(m)). Let xm,ym,zm,wm be the generators for Gm as given in Hypothesis 3.1 which give Qi(m) and C(m). So Q1(m)=xm,ym,zm,Q2(m)=xm,ym,wm and C(m)=xm,ym. The corresponding generators for H are x,y,z,w where R1=x,y,z,R2=x,y,w and D=x,y and Ψ={m+1,,m+8} is a regular D-orbit upon which, restricted to Ψ,x=(m+1,m+2)(m+3,m+4)(m+5,m+6)(m+7,m+8),y=(m+1,m+8)(m+2,m+3)(m+4,m+5)(m+6,m+7).

Put Ω=ΩṁΓ. Then we may define subgroups P1,P2Sym(Ω) and B=P1P2 as in Hypothesis 2.7. Thus we have Pi=Sym(4) and BDih(8). Now takeg(Δ,Ψ)=(1,m+1)(2,m+2)(3,m+3)(4,m+4)(5,m+5)(6,m+6)(7,m+7)(8,m+8)Sym(Ω)

Then g(Δ,Ψ) satisfies the conditions of Definition 2.8 and so we may twist P1 with respect to (Δ,Ψ). Because of Lemma 2.9 we have that P1g,P2 is a transitive subgroup of Sym(Ω) which is actually a subgroup of Alt(Ω). Theorem 1.1 is proved (constructively) by beginning with an appropriate Gm and appropriate Q1(m),Q2(m),C(m), and then “adding” copies of H. Thus the proof must consider twenty four cases given by the congruence of n(mod24) where n=m+24k. For the full details we direct the reader to [Citation8] or to the seed permutations given in Section 4 and here we will look at starting with m = 24 and repeatedly “adding” copies of H. We remark that recursive constructions of this ilk have also been employed in [Citation1] and [Citation6].

Definition 3.3.

If we begin with G=Alt(m) and recursively add on copies of H, then any points greater than m will be denoted with numbers {1,,24} with an appropriate subscript attached. Explicitly, the number m+24i+j will be denoted ji+1 for j{1,,24}. In simpler terms, this will see the ith set of 24 points to be labeled {1i,,24i} for 1ik. The points {1,,m} can be considered as having the subscript 0 attached, which will be useful in writing down the cycle decomposition of z but not generally noted otherwise.

The labels ji and m+24(i1)+j will be interchangeable depending on which is most convenient.

Theorem 3.4.

For n0(mod24),Alt(n) is a completion of the Goldschmidt G3-amalgam.

Proof.

Starting at m = 48, we take the following elements for G=Alt(48+24k) x48+24k:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(22,23)(24,25)(28,29)(30,31)(32,33)(34,35)(36,37)(38,39)(40,41)(42,43)i=1k(1i,2i)(3i,4i)(5i,6i)(7i,8i)(9i,10i)(11i,12i)(13i,14i)(15i,16i)(17i,18i)(19i,20i)(21i,22i)(23i,24i)y48+24k:=(1,8)(2,3)(4,5)(6,7)(10,11)(13,17)(14,18)(19,20)(22,24)(26,27)(28,35)(29,30)(31,32)(33,34)(37,38)(40,44)(41,45)(46,47)i=1k(1i,8i)(2i,3i)(4i,5i)(6i,7i)(9i,16i)(10i,11i)(12i,13i)(14i,15i)(17i,24i)(18i,19i)(20i,21i)(22i,23i)z48+24k:=(1k,38)(2k,39)(5k,36)(6k,37)(7k,8k)(9,10)(13,16)(14,15)(20,21)(22,35)(23,34)(24,30)(25,31)(26,27)(28,29)(40,42)(41,43)(47,48)i=1k(1i1,9i)(2i1,10i)(3i1,20i)(4i1,19i)(5i1,14i)(6i1,13i)(7i1,23i)(8i1,24i)(11i,21i)(12i,22i)(15i,18i)(16i,17i)w48+24k:=(1,17)(4,18)(5,14)(6,7)(8,13)(10,19)(11,20)(16,21)(22,26)(24,27)(28,44)(31,45)(32,41)(33,34)(35,40)(37,46)(38,47)(43,48)i=1k(1i,14i)(2i,19i)(3i,18i)(4i,10i)(5i,11i)(6i,22i)(7i,23i)(8i,15i)(9i,17i)(12i,21i)(13i,20i)(16i,24i)

For convenience we will denote x48+24k,y48+24k,z48+24k,w48+24k simply as x,y,z,w throughout this proof. Put Ω={1,,48+24k}.

We consider H=(zyx)w,z,xx,y,z,w and show that H fixes 19 and is transitive on Ω\{19}. Hence x,y,z,w is 2-transitive on Ω. In fact, the generator x is not necessary here but it will simplify the proof. First we look at (zyx)w and see that(zyx)w=(19)(22)(24)(1,5,21)(2,211,161)(3,61,11)(4,8,15)(6,121,171)(7,51,21)(9,12,20)(10,11,16)(13,31,201)(14,241,71)(17,81,131)(18,91,41)(23,41,30)(25,34,40)(26,29,33)(27,44,45)(28,32,42)(31,35,48)(36,18k,14k)(37,38,43)(39,10k,22k)(46,15k,11k)(47,23k,19k)i=1k1(10i,9i+1,4i+1)(11i,24i+1,7i+1)(14i,8i+1,13i+1)(15i,3i+1,20i+1)(18i,6i+1,1i+1)(19i,21i+1,16i+1)(22i,5i+1,2i+1)(23i,12i+1,17i+1)

The following subgraphs of O((zyx)w,z,Ω) will help to show that H is transitive on Ω\{19}, where the orbits of (zyx)w are on the left hand side and orbits of z are on the right. For the moment we assume k2, dealing with k = 0, 1 later.

Looking at and , the following are subsets of H-orbits:Δ1:={1,3,5,9,10,11,12,13,16,18,20,21,11,31,41,61,91,201},Δ2:={2,6,17,81,121,131,161,171,211},Δ3={4,8,14,15,241,71},Δ4:={7,21,51},Δ5:={22,23,24,25,26,27,28,29,30,31,32,33,34,35,37,38,40,41,42,43,44,45,47,48,\\1k,6k,19k,23k}Δ6:={36,39,46,2k,5k,10k,11k,12k,14k,15k,16k,17k,18k,21k,22k},Δ7:={7k,8k,13k,24k},

Fig. 3 Orbit calculations.

Fig. 3 Orbit calculations.

Fig. 4 Orbit calculations.

Fig. 4 Orbit calculations.

Δ8:={3k,20k} andΔ9:={4k,9k}

However, when also considering orbits of x, we can connect these further as in , which shows that Δ1Δ2Δ3Δ4 and Δ5Δ6Δ7Δ8Δ9 are subsets of H-orbits, with {1k,,24k}Δ5Δ6Δ7Δ8Δ9.

Fig. 5 Orbit calculations.

Fig. 5 Orbit calculations.

When 1ik1, shows that the following are subsets of H-orbits.Λi1:={1i,6i,10i,14i,4i+1,8i+1,9i+1,13i+1}Λi2:={8i,11i,12i,13i,16i,17i,21i,22i,2i+1,5i+1,7i+1,24i+1}Λi3:={4i,9i,15i,18i,1i+1,3i+1,6i+1,20i+1}Λi4:={19i,23i,12i+116i+1,17i+1,21i+1}Λi5:={7i,24i}Λi6:={3i,20i}Λi7:={2i,5i}

Fig. 6 Orbit calculations.

Fig. 6 Orbit calculations.

Now, considering the orbits of x we can see that Λi:={1i,,24i,1i+1} is a subset of an H-orbit with the help of . Clearly, there are more points with subscript i + 1 that we could include, but all we require is that there is at least one, hence 1i+1ΛiΛi+1 and therefore, i=1k1ΛiΓ, where Γ is an H-orbit. Since Δ1Δ2Δ3Δ4 contains points with subscript 1- all of which are in Λ1Γ, so Δ1Δ2Δ3Δ4Γ also. Finally, Λk1(Δ5Δ6Δ7Δ8Δ9)={1k} so Δ5Δ6Δ7Δ8Δ9Γ and Γ={1,,48+24k}\{19},{19} are H-orbits. Consequently x,y,z,w is 2-transitive when k2.

Fig. 7 Orbit calculations.

Fig. 7 Orbit calculations.

Next we look at the case k = 0. For k = 0, consider (zyx)wzH,(zyx)wz=(19)(1,36,3,17,5,20,10,11,13,15,4,7,2,48,25,23,43,6,39,18,8,14,46,16,9,12,21,38,41,24,30,34,42,29,33,27,44,45,26,28,32,40,31,22,35,47,37).

This is a 47-cycle, so clearly H is transitive on Ω{19}.

Finally we consider the case k = 1. Then

(zyx)wz=(19)(1,62,53,39,2,59,46,66,5,20,10,11,13,51,3,37,49,68,16,9,12,21,57,52,18)(4,72,56,6,70,50,71)(7,36,63,69,65,61,17,55,15,67,48,25,23,43,54,38,41,24,30,34,42,29,33,27,44,45,26,28,32,40,31,22,35,47)(8,14)(58,60,64).

Here, z contains the transpositions (2,58),(4,67),(8,72) and (60, 70). These transpositions (ignoring the cycle (19)) connecting the first and fifth cycle of (zyx)wz, second and third, fourth and second and fifth and second cycles respectively. That is to say, they contain one point from each therefore proving that all of the points in both of those cycles must belong to the same H-orbit- so forming the H-orbit {1,,72}\{19}.

Consequently for all k0 x,y,z,w is 2-transitive.

Now we are in a position to demonstrate that G=x,y,z,w, using the element zw. For the moment we assume that k=0.

Observing the cycle decomposition of zw leads to us noticing that it follows patterns depending on the value of k(mod2). When k0(mod2), we analyze the cycles in zw. Here, the underlined parts of the cycle are “repeated,” only with the subscript on the points in subsequent or previous repeated parts being respectively higher or lower by the same amount with each repetition;(2,41,22¯,43,24,,4k3,2k2,4k1,2k¯,39,19k,10k1¯,19k2,10k1,,194,103,192,101¯)(4,21,42¯,23,44,,2k3,4k2,2k1,4k¯,10k,19k1¯,10k2,19k3,,104,193,102,191¯,18)(3,131,7,71¯,72,,7k1,7k¯,15k,3k,18k,8k,23k¯,23k1,,232,231¯,6,201)(5,11,172,242,151,31,132,221,211,51¯,12,,1k1,17k,24k,15k1,3k1,13k,22k1,21k1,5k1,1k¯,47,43,32,41,48,38,14k,¯11k1,12k1,6k1,20k,18k1,8k1,16k,9k¯,14k1,,142,111,121,61,202,181,81,162,92,141¯,14,15)

(22,40,42,35,26,24,30,27), (9, 19, 10), (23, 33, 34), (25, 45, 31), (28, 29, 44),(36,11k,12k,6k,46,37,22k,21k,5k),(1,171,241,13,21,11,20,16,8,161,8,161,91)

So, by counting the lengths of these cycles, we see zw is an element of cycle type (18k7)1.(2k+9)1.(2(1+k))2.121.91.81.34.11 for which we can see the first two cycles are of odd length and the third will be divisible by 2 but not 4. Overall, this means that the only cycle of length divisible by 8 is the 8-cycle above.

When k1mod2 we have the following cycles;(2,41,22¯,43,,4k2,2k1,4k¯,10k,19k1¯,10k2,,103,192,101¯)(4,21,42¯,23,,2k2,4k1,2k¯,39,19k,10k1¯,19k2,,193,102,191¯,18)(6,201,3,131,7,71¯,72,,7k1,7k¯,15k,3k,18k,8k,23k¯,23k1,,232,231¯)(5,11,172,242,151,31,132,221,211,51¯,12,,1k1,17k,24k,15k1,3k1,13k,22k1,21k1,5k1,1k¯,47,43,32,41,48,38,14k,¯11k1,12k1,6k1,20k,18k1,8k1,16k,9k¯,14k1,,142,111,121,61,202,181,81,162,92,141¯,14,15)

(22,40,42,35,26,24,30,27), (9, 19, 10), (23, 33, 34), (25, 45, 31), (28, 29, 44),

(36,11k,12k,6k,46,37,22k,21k,5k), (1,171,241,13,21,11,20,16,8,161,8,161,91) All of this gives a cycle type of (18k7)1.(2k+9)1.(5+2(k1))1.(3+2(k1))1.121.91.81.34.11, for which we can see that the first four cycles will all be of odd length and no others divisible by 8 other than the 8 cycle.

In both cases, for some large odd number f, (zw)4f=(22,26)(24,40)(27,35)(30,42)=:c and cy=(24,27)(22,44)(26,28)(29,42) so ccy=(22,28,26,44)(24,40,27,35)(30,29,42) and (ccy)4 is a 3-cycle, sufficient by Theorem 2.10 to confirm that G=x,y,z,w.

It remains to deal with the exceptional case k = 0. Here, xzw=(1,39,17)(2,47,43,35,23,40,48,38)(3,18,4)(5,46,37,14,21,11,12,20,16)(6,36,7)(8,13,15)(10,19)(22,33,41,42,32,34,26,24,45,31,27)(25,30)(28,44), of cycle type 111.91.81.34.22.12 and (xzw)9.8 is an 11-cycle, sufficient by Theorem 2.10 to confirm that G=x,y,z,w.

So for all k0,G=x,y,z,w, which completes the proof of Theorem 1.1 in the case n0(mod24). □

4 Seeds for the recursive construction

The following generators provide a starting point for the recursive method of adding 24 points for each case modulo 24. In each case {1,,8} is a regular x,y-orbit with x=(1,2)(3,4)(5,6)(7,8) and y=(1,8)(2,3)(4,5)(6,7) on this orbit. So, we can take Δ={1,,8} and Ψ={m+1,,m+8} as the two orbits to twist by.

When n0(mod24), we may start at m = 24 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(17,24)(18,19)(20,21)(22,23)z:=(1,10)(2,9)(3,20)(4,19)(5,13)(6,14)(7,23)(8,24)(11,17)(12,18)(15,22)(16,21)w:=(1,10)(2,22)(3,23)(4,14)(5,15)(6,19)(7,18)(8,11)(9,21)(12,17)(13,24)(16,20)

When n1(mod24), we may start at m = 25 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(14,15)(17,21)(18,22)(19,20)(23,24)z:=(1,12)(2,11)(3,4)(5,10)(6,9)(13,14)(17,19)(18,20)(21,22)(24,25)w:=(1,17)(2,3)(4,18)(5,22)(8,21)(10,24)(11,23)(13,16)(14,19)(15,20)

When n2(mod24), we may start at m = 26 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(18,19)(21,25)(22,26)(23,24)z:=(1,8)(2,7)(3,5)(4,6)(9,10)(11,18)(12,17)(15,20)(16,19)(21,24)(22,23)(25,26)w:=(1,25)(4,26)(5,22)(6,7)(8,21)(9,14)(10,12)(11,13)(15,16)(17,20)(18,23)(19,24)

When n3(mod24), we may start at m = 27 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(14,15)(17,21)(18,22)(23,24)(25,26)z:=(1,13)(2,14)(3,4)(5,15)(6,16)(9,10)(17,20)(18,19)(23,27)(25,26)w:=(1,18)(2,3)(4,17)(5,21)(8,22)(10,24)(11,23)(14,26)(15,25)(19,27)

When n4(mod24), we may start at m = 28 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(14,15)(17,21)(18,22)(23,24)(25,26)z:=(1,2)(3,10)(4,9)(7,12)(8,11)(15,16)(17,19)(18,20)(24,28)(26,27)w:=(1,18)(2,3)(4,17)(5,21)(8,22)(10,23)(11,24)(14,25)(15,26)(20,27)

When n5(mod24), we may start at m = 29 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(18,19)(21,25)(22,26)(27,28)z:=(1,18)(2,17)(5,20)(6,19)(7,8)(9,15)(10,16)(11,14)(12,13)(21,23)(22,24)(28,29)w:=(1,2)(3,8)(4,6)(5,7)(9,22)(10,11)(12,21)(13,25)(16,26)(18,28)(19,27)(24,29)

When n6(mod24), we may start at m = 30 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)(29,30)y:=(1,8)(2,3)(4,5)(6,7)(9,13)(10,14)(15,22)(16,17)(18,19)(20,21)(23,30)(24,25)(26,27)(28,29)z:=(1,23)(2,24)(3,18)(4,17)(5,28)(6,27)(7,21)(8,22)(9,11)(10,12)(15,30)(16,29)(19,25)(20,26)w:=(1,13)(4,14)(5,10)(6,7)(8,9)(11,12)(15,16)(17,22)(18,20)(19,21)(23,24)(25,30)(26,28)(27,29)

When n7(mod24), we may start at m = 31 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(17,24)(18,19)(20,21)(22,23)(26,27)(29,30)z:=(3,25)(4,26)(5,6)(7,27)(8,28)(9,15)(10,16)(11,14)(12,13)(17,20)(18,19)(21,23)(22,24)(30,31)w:=(1,24)(2,13)(3,12)(4,20)(5,21)(6,16)(7,9)(8,17)(10,18)(11,19)(14,23)(15,22)(26,29)(27,30)

When n8(mod24), we may start at m = 32 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)(29,30)(31,32)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(17,24)(18,19)(20,21)(22,23)(25,32)(26,27)(28,29)(30,31)z:=(1,9)(2,10)(3,20)(4,19)(5,14)(6,13)(7,23)(8,24)(11,21)(12,22)(15,18)(16,17)(25,32)(26,31)(27,29)(28,30)w:=(1,2)(3,8)(4,6)(5,7)(9,22)(10,27)(11,26)(12,18)(13,19)(14,30)(15,31)(16,23)(17,25)(20,29)(21,28)(24,32)

When n9(mod24), we may start at m = 33 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(18,19)(22,23)(25,29)(26,30)(27,28)(31,32)z:=(1,21)(2,22)(3,4)(5,23)(6,24)(9,18)(10,17)(13,20)(14,19)(15,16)(25,28)(26,27)(29,30)(32,33)w:=(2,29)(3,25)(4,5)(6,26)(7,30)(9,11)(10,16)(12,15)(13,14)(17,20)(18,27)(19,28)(22,32)(23,31)

When n10(mod24), we may start at m = 34 with:

x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)(29,30)(31,32)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(17,24)(18,19)(20,21)(22,23)(26,27)(29,33)(30,34)(31,32)z:=(1,3)(2,4)(5,8)(6,7)(9,26)(10,25)(13,28)(14,27)(15,16)(17,23)(18,24)(19,22)(20,21)(29,32)(30,31)(33,34)w:=(1,14)(2,19)(3,18)(4,10)(5,11)(6,22)(7,23)(8,15)(9,17)(12,21)(13,20)(16,24)(25,28)(26,31)(27,32)(30,34) When n11(mod24), we may start at m=35 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(18,19)(22,23)(25,29)(26,30)(31,32)(33,34)z:=(1,23)(2,24)(5,21)(6,22)(7,8)(9,19)(10,20)(13,17)(14,18)(15,16)(25,28)(26,27)(31,32)(34,35)w:=(1,2)(3,8)(4,6)(5,7)(9,16)(10,25)(11,29)(14,30)(15,26)(18,32)(19,31)(22,34)(23,33)(27,35)

When n12(mod24), we may start at m=36 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(18,19)(21,29)(22,30)(23,31)(24,32)(33,34)z:=(1,8)(2,7)(3,5)(4,6)(11,20)(12,19)(13,14)(15,18)(16,17)(21,26)(22,25)(23,27)(24,28)(33,35)w:=(1,32)(4,31)(5,23)(6,7)(8,24)(9,30)(12,29)(13,21)(14,15)(16,22)(18,34)(19,33)(26,36)(27,35)

When n13(mod24), we may start at m=37 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)(29,30)(31,32)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(17,24)(18,19)(20,21)(22,23)(26,27)(30,31)(33,34)(35,36)z:=(1,26)(2,25)(5,28)(6,27)(7,8)(9,12)(10,11)(13,15)(14,16)(17,29)(18,30)(19,20)(21,31)(22,32)(33,37)(35,36)w:=(1,14)(2,19)(3,18)(4,10)(5,11)(6,22)(7,23)(8,15)(9,17)(12,21)(13,20)(16,24)(26,33)(27,34)(30,35)(31,36)

When n14(mod24), we may start at m=14 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)y:=(1,8)(2,3)(4,5)(6,7)(9,13)(10,14)z:=(1,8)(2,7)(3,5)(4,6)(9,11)(10,12)w:=(1,13)(4,14)(5,10)(6,7)(8,9)(11,12)

When n15(mod24), we may start at m=15 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(13,14)z:=(1,2)(3,10)(4,9)(7,12)(8,11)(14,15)w:=(1,6)(2,4)(3,5)(7,8)(10,13)(11,14)

When n16(mod24), we may start at m=40 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)(29,30)(31,32)(33,34)(35,36)(37,38)(39,40)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(17,24)(18,19)(20,21)(22,23)(25,32)(26,27)(28,29)(30,31)(33,40)(34,35)(36,37)(38,39)z:=(1,3)(2,4)(5,8)(6,7)(9,25)(10,26)(11,21)(12,22)(13,30)(14,29)(15,18)(16,17)(19,28)(20,27)(23,31)(24,32)(33,35)(34,36)(37,40)(38,39)w:=(1,38)(2,19)(3,18)(4,34)(5,35)(6,22)(7,23)(8,39)(9,10)(11,16)(12,14)(13,15)(17,33)(20,37)(21,36)(24,40)(25,26)(27,32)(28,30)(29,31)

When n17(mod24), we may start at m=41 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(21,22)(23,24)(25,26)(27,28)(29,30)(31,32)(33,34)(35,36)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(13,17)(14,18)(19,20)(21,28)(22,23)(24,25)(26,27)(30,31)(33,37)(34,38)(39,40)z:=(1,31)(2,32)(5,29)(6,30)(7,8)(9,24)(10,23)(11,28)(12,27)(13,15)(14,16)(19,20)(21,22)(33,35)(34,36)(40,41)w:=(1,14)(2,3)(4,13)(5,17)(8,18)(10,19)(11,20)(15,16)(21,37)(24,38)(25,34)(26,27)(28,33)(30,39)(31,40)(36,41)

When n18(mod24), we may start at m=18 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(13,17)(14,18)(15,16)z:=(1,8)(2,7)(3,5)(4,6)(9,10)(13,15)(14,16)(17,18)w:=(1,17)(4,18)(5,14)(6,7)(8,13)(9,12)(10,15)(11,16)

When n19(mod24), we may start at m=43 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(23,24)(25,26)(27,28)(29,30)(31,32)(33,34)(35,36)(37,38)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(15,17)(16,18)(19,20)(23,30)(24,25)(26,27)(28,29)(32,33)(35,39)(36,40)(41,42)z:=(1,33)(2,34)(5,31)(6,32)(7,8)(9,27)(10,28)(11,23)(12,24)(13,16)(14,15)(20,21)(29,30)(35,37)(36,38)(42,43)w:=(2,17)(3,15)(4,5)(6,16)(7,18)(10,19)(11,20)(14,22)(23,39)(26,40)(27,36)(28,29)(30,35)(32,41)(33,42)(38,43)

When n20(mod24), we may start at m=20 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(13,17)(14,18)(19,20)z:=(1,2)(3,10)(4,9)(7,12)(8,11)(13,15)(14,16)(19,20)w:=(1,14)(2,3)(4,13)(5,17)(8,18)(10,19)(11,20)(15,16)

When n21(mod24), we may start at m=21 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(15,17)(16,18)(19,20)z:=(1,12)(2,11)(3,4)(5,10)(6,9)(13,16)(14,15)(19,21)w:=(1,3)(2,8)(4,7)(5,6)(10,20)(11,19)(14,21)(16,18)

When n22(mod24), we may start at m=22 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)y:=(1,8)(2,3)(4,5)(6,7)(10,11)(13,17)(14,18)(19,20)z:=(1,11)(2,12)(5,9)(6,10)(7,8)(13,16)(14,15)(20,21)w:=(1,8)(2,14)(3,18)(6,17)(7,13)(10,20)(11,19)(15,22)

When n23(mod24), we may start at m=47 with:x:=(1,2)(3,4)(5,6)(7,8)(9,10)(11,12)(13,14)(15,16)(17,18)(19,20)(21,22)(23,24)(25,26)(27,28)(29,30)(31,32)(33,34)(35,36)(37,38)(39,40)(41,42)(43,44)y:=(1,8)(2,3)(4,5)(6,7)(9,16)(10,11)(12,13)(14,15)(17,24)(18,19)(20,21)(22,23)(25,32)(26,27)(28,29)(30,31)(33,40)(34,35)(36,37)(38,39)(42,43)(45,46)z:=(1,9)(2,10)(3,20)(4,19)(5,14)(6,13)(7,23)(8,24)(11,21)(12,22)(15,18)(16,17)(25,43)(26,44)(29,41)(30,42)(31,32)(33,35)(34,36)(37,40)(38,39)(45,47)w:=(1,2)(3,8)(4,6)(5,7)(9,10)(11,16)(12,14)(13,15)(17,30)(18,35)(19,34)(20,26)(21,27)(22,38)(23,39)(24,31)(25,33)(28,37)(29,36)(32,40)(42,45)(43,46)

References

  • Bundy, D. M., Rowley, P. J. (2006). Symmetric groups and completions of the Goldschmidt amalgams of type G1 . J. Group Theory 9(5):627–640.
  • Conder, M. (1988). An infinite family of 5-arc-transitive cubic graphs. Ars Combin. 25(A):95–108.
  • Goldschmidt, D. M. (1980). Automorphisms of trivalent graphs. Ann. Math. (2) 111(2):377–406. DOI: 10.2307/1971203.
  • Lorimer, P. (1984). Vertex-transitive graphs: Symmetric graphs of prime valency. J. Graph Theory 8(1):55–68. DOI: 10.1002/jgt.3190080107.
  • Parker, C., Rowley, P. (2000). Classical groups in dimension 3 as completions of the Goldschmidt G3 -amalgam. J. London Math. Soc. (2) 62(3):802–812.
  • Parker, C., Rowley, P. (2001). Sporadic simple groups which are completions of the Goldschmidt G3 -amalgam. J. Algebra 235(1):131–153.
  • Parker, C., Rowley, P. (2002). Finite completions of the Goldschmidt G3 -amalgam and the Mathieu groups. Commun. Algebra 30(12):5611–5619.
  • Vasey, D. M. (2014). Alternating groups as completions of the Goldschmidt G3 -amalgam. PhD thesis. University of Manchester.
  • Wielandt, H. (1964). Finite Permutation Groups. London/New York: Academic Press.