On orders and vanishing of integral cohomology groups. (Q2370230): Difference between revisions

In this paper it is shown Theorem A. If \(G\) is a finite group with a normal subgroup \(N\) such that \(G/N\) is cyclic then no two consecutive integral cohomology groups \(H^i(G,\mathbb{Z})\), \(H^{i+1}(G,\mathbb{Z})\) can vanish. The author deduces this elegant result from a recurrence relation involving the orders of the integral cohomology groups \(H^n(G,\mathbb{Z})\), \(n\geq 1\), when \(G\) has a non-trivial cyclic factor group. To obtain this recurrence relation some exact sequences constructed by \textit{G. Lewis} [in Trans. Am. Math. Soc. 132, 501-529 (1968; Zbl 0217.34903)] are used. Using Theorem A, it is then shown that if \(\kappa\) is a field of characteristic \(p\) and \(G\) a finite group with \(H^1(G,\kappa)\neq 0\), (i.e., \(G\) has a normal subgroup of index \(p\)) then \(H^{2n}(G,\kappa )\neq 0\) for all \(n\geq 0\). This answers Problem 7.2 of \textit{D. J. Benson} [in Comments and corrections for the books and papers of Dave Benson, followed by a collection of problems, \texttt{http://www.math.uga.edu/archive/benson.html} (2005)]. Note that if \(G\) is a finite \(p\)-group then it has been shown by \textit{T.-N. Kuo} [J. Algebra 7, 160-167 (1967; Zbl 0189.32402)] that \(\widehat H^{2n}(G,\mathbb{Z})\neq 0\) for every \(n\).