A characteristic subgroup of \(\Sigma_ 4\)-free groups (Q1916902)

In [Can. J. Math. 20, 1101-1135 (1968; Zbl 0164.02202)] \textit{G. Glauberman} proved his famous \(ZJ\)-theorem, which says that in a finite group \(H\) with \(C_H(O_p(H))\subseteq O_p(H)\), \(p\) odd, which is \(p\)-stable (i.e. there is no \(p^2\text{SL}_2(p)\) involved), we have \(ZJ((S))\triangleleft H\), \(S\) a Sylow \(p\)-subgroup. The obvious generalization to \(p=2\), i.e., \(C_H(O_2(H))\subseteq O_2(H)\) and \(\Sigma_4\) is not involved is false. In the paper under review the author shows (under the additional assumption that any simple non abelian group involved in \(H\) is isomorphic to \(Sz(2^m)\) or \(\text{PSL}(2,3^m)\), \(m\) odd) that there is some characteristic subgroup \(W(S)\) with \(\Omega_1(Z(S))\subseteq W(S)\subseteq Z(J(S))\), which is just depending on \(S\), such that \(W(S)\triangleleft H\). In fact the additional condition follows from the classification of the finite simple groups. To prove the theorem the author considers the class \(\mathcal C\) of all embeddings \(\tau:S\to H\) in groups \(H\) such that \(H\) fulfills the assumption. There is an obvious equivalence relation on \(\mathcal C\) and there are just finitely many equivalence classes. Then he considers the subclass of those \((\tau,H)\in{\mathcal C}\) which have the property that for any normal subgroup \(N\) of \(H\) with \(O^2(H)\nsubseteq N\), we have \(S\tau\cap N\subseteq O_2(H)\) and which have the property that \(J(S)\triangleleft H\). Now he picks in any equivalence class of \((\tau,H)\) belonging to the latter class one representative and constructs \(G\) the amalgamated product of those representatives over \(S\). Then \(W(S)=\langle\Omega_1(Z(S))^G\rangle\). So we have now an analogue to Glauberman's \(ZJ\)-theorem. But in contrast to \(J(S)\) it seems not possible to calculate \(W(S)\) even in small cases.