On finite soluble groups of fixed rank. (Q665902)

Some dependences of the derived length and \(p\)-length of a finite soluble group \(G\) on its rank are established here. Recall that a chief factor of \(G\) which is a \(p\)-group (\(p\) prime) is called a `\(p\)-chief factor' of \(G\). Let \(p^n=\max\{|p\)-chief factors of \(G|\}\); then \(n\) is called the `\(p\)-rank' of \(G\) and denoted by \(r_p(G)\). By definition, the `rank' of a soluble group \(G\neq 1\) stands for \(\max_pr_p(G)\), where \(p\) runs over the set of distinct prime divisors dividing \(|G|\); it is denoted by \(r(G)\). In this paper, there are four main theorems and one main corollary, besides some lemmas and a lot of corollaries. We quote two theorems and one corollary, and we refer kindly to the rest of this interesting paper. Namely, Theorem 1. Given a soluble group \(G\neq 1\) with \(r(G)\leq 2\), the following holds: 1. The nilpotent length of \(G\) is at most 4; 2. The derived length of \(G/\Phi(G)\) is at most 5; here \(\Phi(G)\) stands for the Frattini subgroup of \(G\); 3. The \(p\)-length of \(G\) equals at most 1 for every prime \(p>3\), while for \(p=2\) the 2-length equals at most 2 and for \(p=3\) the 3-length equals at most 2. Theorem 3. Given a soluble group \(G\neq 1\) with \(r_p(G)\leq 2\) for every prime \(p>5\) and \(r_p(G)\leq 3\) for every \(p\in\{2,3,5\}\), the following hold: 1. The derived length of \(G/\Phi(G)\) is at most 5; 2. The group \(G\) possesses a normal \(\{2,3,5,7,13,31\}'\)-Hall subgroup \(H\) and \(H\) has an ordered Sylow tower of supersolvable type. Corollary 1. Consider a \(p\)-soluble group \(G\). If the \(p\)-length of \(G\) equals \(r_p(G)\), then either \(r_p(G)=1\) or \(r_p(G)=2\) and \(p=\{2,3\}\). In particular, if \(r_p(G)\geq 3\), then the \(p\)-length of \(G\) equals at most \(r_p(G)-1\). Reviewer's remark: Theorem 4 and also Corollary 1 of the authors depend on their Lemma 21, which is turn depends on the full truth of the results of \textit{E. G. Bryukhanova} [Math. Notes 29, 85--90 (1981); translation from Mat. Zametki 29, 161--170 (1981; Zbl 0528.20010)]. In 1996--1997 this reviewer, together with a doctoral student, studied Bryukhanova's very condensed paper in detail, but did not succeed to take the final step of it! After these years up to now, it still is unclear to us whether her result is fully correct. Hence all results depending on that paper are perhaps in doubt, at least for us. We kindly refer to the doctoral dissertation of \textit{J. Blom} (unpublished) titled ``\(\ell_2(G)\leq r_2(G)\)(?)'' at the University of Amsterdam, (in the year 1997 in June) at the Korteweg-de Vries Institute for Mathematics (then named Faculty of Mathematics). We hope somebody can overcome that obstable in the future.