Fulltext

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Mathematical Notes, Vol. 56, Nos.

1-2, 1994

INTERSECTION

OF ABELIAN SUBGROUPS

IN FINITE GROUPS

V. I. Zenkov

Generalizing Laffey's results [I], Dempwolff and Wong in their paper [2] have shown that for cyclic subgroups A and B of a finite group G there exists an element g E G such that Ag n B < F(G), where F(G) is the Fitting subgroup of the group G. In theorem 2 we obtain a generalization of this result. We introduce a basic definition of this paper. Let A and B be subgroups of the group G. The minimal by inclusion elements of the set {Ag O B I g E G} will be called minimal (A, B)-intersections. The following theorems are proved in this paper. T H E O R E M 1. Let G be a finite group, A and B be Abelian subgroups of G. Then Ag N B < F(G) for each minimal (A, B)-intersection. In the general case the minimal (A, B)-intersection may not be normal in F(G) as the following example shows. Example 1. G = F(G) = D s, A is a cyclic subgroup of the second order not located in Z(G) where Z(G) is the center of the group G and B is a quarternion subgroup containing A. Since B ~ G and A <_ B, it follows that A h n B ;~ 1 for any element h belonging to G. All the intersections are (A, B)-minimal. If we now assume that A f O B _'3 F(G) = G for some element f E G then A f n B N Z(G) ;~ 1. Hence A < Z(G) and we arrive at a contradiction. However for cyclic subgroups A and B the minimal (A, B) inte1:section is normal in F(G). T H E O R E M 2. Let G be a finite group, A and G be cyclic subgroups in G. Then Ag n B --- F(G) for any minimal (A,B)-intersection. Note that not every (A, B)-intersection normal in F(G) is minimal as the following example shows. Example 2. G --- A 4, F(G) be a quarternion group. If A and B are cyclic subgroups of the second order belonging to F(G) then A O B = B ~ F(G) however A n B is not an (A, B)-minimal intersection. As the following example shows for any positive integer m there exists a finite group G with cyclic subgroups A and B such that the number of (A, B)-minimal intersections equals m. Example 3. Let m E N; Pl, P2 . . . . . Pm be pairwise different prime numbers and E l, E 2 . . . . . E m be elementary Abelian subgroups of orders pl m, p2 m. . . . . pmm, respectively. Let E i = 1-1~=l(eij ) (i = 1, 2 . . . . . m), E = E 1 E 2 ... m, x be an element of order m from Hol E and G be a semidirect product of E and (x) where Ix) acts transitively on {eil, el2 . . . . . elm} for 1 _< i _< m. Let A = (ell) (e21) x . . . (emil and B = (ell) (e22) Then I AXi-I n B I = Pi, 1 _< m and {Axi-t O B [ 1 < i _< m} are sets of all (A, B)-intersections. They are all minimal and their number equals m. We now proceed to the proof of theorem 1. Let G be a counterexample of a minimal order to theorem 1. Since Ag n B ~ F(G) for some (A, B)-minimal intersection there exists a p-element b belonging to Ag O B such that b ~ F(G). It then follows from the Baer-Suzuki theorem (theorem 3.8.2 in [3]) that there exists an element y E G for which the subgroup (b, bY) is not nilpotent. Set N = (B, bY). Assume that N = G. Since the intersection Ag n B is (A, B)-minimal we have that A n n B ;~ 1 for any element h E G. The subgroup AgY contains < bY > and AgY O B = D ;~ 1. Therefore D < Z(G). Whence in view of D _< Ag O B, it follows that in the quotient group G = G/D the intersection A~ n B will be (A, B)-minimal. By induction A~ n B _< F(G). Whence since F(G/D) = F(G)/D, we obtain that Ag O B < F(G), which is a contradiction.

Ural Polytechnical Institute. Translated from Matematicheskie Zametki, Vol. 56, No. 2, pp. 150-152, August, 1994. Original article submitted August 4, 1992.

0001-4346/94/0506-0869512.50

Plenum Publishing Corporation

869

Hence N ;~ G. Let A 1 = Ag N N. The intersection A I A B is (A 1, B)-minimal in N. Indeed if A1n O B < A 1 n B for some n E N it follows that
(Ag N N ) n N B < AO n N A B = Aa t'IB
or

A an N B N N < A a N B ,

whence
A g~ N

B < A g n B,

which contradicts the (A, B)-minimality of the intersection Ag n B. By induction the intersection A 1 n B belongs to F(N). However A 1 n B = Ag n N N B ~ b and (b, bY) is not a nilpotent subgroup in N. This contradiction proves theorem 1. Proof of Theorem 2. Let G be a counterexample of the minimal order to the theorem 2. Select a subgroup B such that Ag n :~ F(G) for some (A, B)-minimal intersection A g n B and let the order of B be minimal. Let B1 be a subgroup of prime index p in B. If the intersection Ag N B contains no proper intersections of the form A h O B I for any element b E G then the inclusion Ag N B 1 ~ Ag N B implies that the intersection Ag O B 1 will be (A, B1)-minimal in G. The choice of B implies that Ag n B 1 -~ F(G). However Ag O B I = A g N B, therefore A g N B _ F(G), which is a contradiction. Hence Ag N B > Agl O B 1, where A gl O B 1 is a (A, B1)-minimal intersection. Consider the intersection Agl n B. Then the (A 1, B)minimality of Ag N B implies that Ag 1 O B > Agl n B 1. The intersection Ag 1 O B = (z) for some element z not located in the intersection Agl n B 1. Since (Agl n B)p, = (Agl n B1)p,, the Sylow p-subgroup of the subgroup (x) coincides in view of cyclicity with the Sylow p-subgroup in B. If the intersection Agl n B is not (A, B)-minimal then Agt O B > Ag2 O B for some (A, B)-minimal intersection A g2 O B. Assume that I Ag2 O B I p < ] Agl n B [ p. Then (Ag2 O B)p _< (A gl O B1)p and (Ag2 O B)p, _< (A gl O B)p, = (A gl n B1)p,. Hence Ag2 O B _< Ag 1 n B l < A g o B, which contradicts the fact that the intersection Ag n B is minimal. Therefore (Ag2 O B)p --- (Ag 1 O B)p. Then (Agl N B)p, > (Ag2 O B)p,. In this case Ag2 O B 1 < A gl N B 1, which contradicts the fact that A gl O B 1 is (A, B1)-minimal. Hence the intersection Agl n B is (A, B)-minimaI. By Theorem 1 A gI n B _< F(G). We have thus shown that the Sylow p-subgroup P of the subgroup B is located in the intersection Ag 1 n B. Therefore P _< F(G). Since P is arbitrary it follows that B _< F(G). We shall consider the intersection
A g n B = A g o ( B o F(G)) = (A g n F ( G ) ) N B.

This intersection will be (Ag n F(G), B)-minimal in F(G). Indeed if


(A g n F(G)) n B > (Ag n F(G)) S n B

for some element f E F(G) then


A g N B > A g$ N F ( G ) S n B = A gl m F ( G ) n B = A g/ n B,

which contradicts the (A, B)-minimality of the intersection Ag N B. If now G # F(G), then by induction Ag n B ~- F(G) yielding a contradiction. Therefore G = F(G). Consider the intersection Agl n B 1. Since A gl n B 1 is a (A, B1)-minimal intersection then in view of the choice of B A g 1 n B 1 ~ F(G) = G and therefore in the case when Agl n B 1 # 1 D = A gl n B t n Z(F(G)) # 1. Since Ag n B > Agl n B 1 _> D, in the quotient group G = G/D Ag n B will be (A, B)-minimal intersection. By induction h,g O B _< F(G). Hence Ag O B ~_ F(G), which is a contradiction. Hence A gl n B 1 = 1. Therefore I Ag O B I p = p. Since this is valid for any p E 7r(B), the subgroup B does not contain elements of order pZ and B n Z(F(G)) = 1. It will therefore be sufficient to take from each Sylow p-subgroup from F(G) an element xp such that ApXp N Bp = 1. If x is

870

the product of all such elements then A x n B = 1. This contradicts the fact that the intersection Ag N B is (A, B)-minimal. Theorem 2 is proved. The author expresses his gratitude to A. S. Kondrat'ev for discussing this paper and providing remarks which improved the exposition.

REFERENCES
.

2. 3.

T. Y. Laffey, Proc. Edinburgh Math. Soc., 20, No. 2, 229-232 (1976-77). U. Dempwolff and S. K. Wong, Proc. Edinburgh Math. Soc., 25, No. 1, 19-20 (1982). D. Gorenstein, Finite Groups, Harper and Row, New York (1968).

871

You might also like