A remark on the character-degrees of some \(p\)-groups (Q1203566)

By Taketa, the derived length \(d\ell (G)\) of a finite monomial group is bounded by the number \(| cd(G)|\) of distinct degrees of the irreducible classical characters of \(G\). It is conjectured that this is true for all finite solvable groups. \textit{T. R. Berger} has verified this for groups of odd order [J. Algebra 39, 199-207 (1976; Zbl 0362.20012)], and \textit{D. Gluck} has proved that \(d\ell(G) \leq 2| cd(G)|\) [J. Lond. Math. Soc., II. Ser. 31, 457-462 (1985; Zbl 0569.20010)]. Now the author poses the question. ``But is a bound for \(d\ell G\), linear in the number of character degrees, asymptotically best possible?'' Then he presents two series of \(p\)-groups in which the derived length is unbounded and \(d\ell G\) is bounded by a logarithmic function of \(| cd (G)|\). One of the examples are the groups \(G(n,p)\) of upper triangular matrices over \(GF(p)\) with diagonal 1 where \(p\) is a prime. The other series is given by the wreath products \(Z_ p \wr \underset {n} \dots \wr Z_ p\) where \(p\) is a prime and \(Z_ p\) of order \(p\). Reviewer's remark: In the Remarks of section 1 the author mentions that the results may be extended to \(G(n,q)\) for \(q\) a power of \(p\) provided it is true that the (irreducible) character degrees for \(G(n,q)\) are powers in \(q\). Indeed, this is true and has first been stated by \textit{E. A. Gutkin} [in Funkts. Anal. Prilozh. 7, No. 4, 80 (1973; Zbl 0285.20041)]. The reviewer noticed that there is a serious gap in the proof. A proper proof may be found in a recent paper of \textit{I. M. Isaacs} [on ``Characters of groups associated with finite algebras''(to appear)]. For \(p > n\) a completely different approach has been given by \textit{D. Kazhdan} [in Isr. J. Math. 28, 272-286 (1977; Zbl 0391.22006)].