\(p\)-groups with few conjugacy classes of normalizers. (Q382223)

For a group \(G\), define \(\omega(G)\) to be the number of conjugacy classes of normalizers of subgroups of \(G\). If \(G\) is nilpotent, then write \(c(G)\) for the nilpotency class of \(G\). The main result of this paper asserts that if \(G\) is a finite \(p\)-group with \(\omega(G)\leq p+1\), then \(c(G)\leq 2\) if \(p\) is odd, and \(c(G)\leq 3\) if \(p=2\). Slightly different bounds are already known to exist: \textit{R. Brandl} has shown that for a finite \(p\)-group \(G\), with \(p\) odd, one has \(c(G)\leq\omega(G)\) [``Conjugacy classes of subgroups of finite \(p\)-groups: the first gap'', Proc. Ischia Group Theory 2010, 39-44 (2012; Zbl 1294.20019)]. \textit{J. Poland} and \textit{A. Rhemtulla} have shown that for a nilpotent group \(G\) that is not a Dedekind group, one has \(c(G)\leq\nu(G)+1\) (where \(\nu(G)\) is the number of conjugacy classes of non-normal subgroups) [Commun. Algebra 24, No. 10, 3237-3245 (1996; Zbl 0885.20009)]. The final section of this paper describes a number of families of \(p\)-groups \(G\) for which \(\omega(G)=p+2\), and concludes with the following interesting question: Let \(G\) be a \(p\)-group with \(\omega(G)\leq p+2\). Do we have \(c(G)\leq 3\) if \(p\) is odd, and \(c(G)\leq 4\) if \(p=2\)?