Suppose R is a group and P is a permutation group acting on the set I. Let M be the restricted direct product of 111 copies of R. Then the restricted wreath product G = R wr( P, I) is the semidirect product of M and P where P acts on M by permuting the factors. We are interested in whether or not M, referred to as the base subgroup of G, is a characteristic subgroup of G. In the special case when P acts regularly on Z, i.e., we may take Z= P with P acting by right multiplication, then P. M. Neumann [9] showed that A4 is not a characteristic subgroup of G if, and only if, 1 PI = 2 and R is a special dihedral group (this will be defined in the next section). In [2], Yu. V. Bodnarchuk claimed that if P is transitive and consists entirely of linitary permutations (i.e., every element of P moves only finitely many points of I), and if M is not a characteristic subgroup of G, then P has a nonidentity normal elementary abelian 2-subgroup and R is of dihedral type (to be defined in the next section). This claim is incorrect and counterexamples (pointed out to me by L. G. Kovacs) using a construction of P. Hall [4] are given in Section 6. Bodnarchuk apparently overlooked the possibility that an automorphism of an infinite group can map a subgroup into a proper subgroup of itself.’ In the counterexamples to Bodnarchuk’s result, the distinct images of the base subgroup under all the automorphisms of G constitute an infinite totally ordered set under inclusion. Bodnarchuk’s argument is valid, however, for finite groups. More generally, we show that Bodnarchuk’s result becomes a correct theorem if
Read full abstract