On finite nonabelian 2-groups all of whose minimal nonabelian subgroups are of exponent 4. (Q2459999)

Since nonabelian \(p\)-groups are generated by minimal nonabelian subgroups, it is natural to try to determine the structure of a \(p\)-group if the structure of its minimal nonabelian subgroups is known. In this direction the author obtained a number of deep results. A \(2\)-group is quasidihedral if its Hughes subgroup has index \(2\). The first main result, Theorem 2.3, asserts that if all minimal nonabelian subgroups of a \(2\)-group \(G\) have the same order \(8\), then either \(G\) is quasidihedral or \(G=HZ(G)\), where \(H\) is either of maximal class or extraspecial and \(\mho_1(Z(G))\leq Z(H)\). In particular, if a \(2\)-group \(G\) is minimal non Dedekindian, it is either minimal Abelian or generalized quaternion of order \(16\). (As A. Mann noticed in his letter to the reviewer, if \(p>2\) and all minimal nonabelian subgroups of a nonabelian \(p\)-group \(G\) have the same order \(p^3\) and exponent \(p\), then either \(\exp(G)=p\) or the Hughes subgroup is Abelian and has index \(p\) in \(G\).) In the further three theorems the nonabelian \(2\)-groups all of whose minimal nonabelian subgroups are isomorphic and have order \(>8\) and exponent \(4\) are determined. Recently, the author classified the \(2\)-groups all of whose minimal nonabelian subgroups have order \(16\) and exponent \(8\) [Isr. J. Math. (to appear)]. There is no analogue of these results in finite \(p\)-group theory.