On fully unsaturated homomorphs of finite groups (Q1200966)

For a homomorph \({\mathcal H}\) in the universe of finite groups, denote by \(P{\mathcal H}\) and \(D{\mathcal H}\) the classes of all groups that have \({\mathcal H}\)-projectors and \({\mathcal H}\)-covering subgroups respectively, \({\mathcal P}\) the class of all primitive groups, \(b({\mathcal H})=\{G\not\in{\mathcal H}\mid G/N\in{\mathcal H}\) for each \(\{1\}\neq N\trianglelefteq G\}\), and \(E_{\mathcal P}{\mathcal H}=\{G\mid\exists N\vartriangleleft G\) such that \(G/N\in {\mathcal H}\) and \(M\vartriangleleft G\), \(M\subset N\) implies \(G/M\in{\mathcal P}\}\). Consider the statements (a) \(P{\mathcal H}={\mathcal H}\), (b) \(D{\mathcal H}={\mathcal H}\), (c) \(b({\mathcal H})\cap{\mathcal P}=\emptyset\), and (d) \(E_{\mathcal P} {\mathcal H}={\mathcal H}\). The equivalences known for the solvable case [\textit{P. Förster}, J. Algebra 49, 606-620 (1977; Zbl 0371.20021)] are not all valid more generally. The author first notes that (c) and (d) are equivalent and that (c) implies (a) implies (b). The purpose of this paper is to obtain counter-examples for the converses of the latter implications. That is achieved with a careful and adroit presentation.