scientific article; zbMATH DE number 3582250

From MaRDI portal
Publication:4152618





Let \(L(s) = L(s, \chi_1, 4) = \sum_0^\infty (-1)^n/(2n+1)^s\). In a previous paper, the authors showed that if the Riemann-Piltz conjecture holds for \(L(s)\), a certain inequality follows which has implications for long sequences of consecutive primes all \(\equiv 1 \bmod 4\). This problem suggests other questions about whether such sequences appear infinitely often. The simplest such problem concerns \(f_2(x) = \sum 1\) summed over all \(p_\nu \le x\) such that both \(p_\nu\) and \(p_{\nu -1}\) are \(\equiv 1 \bmod 4\). A result of Littlewood implies that \(\displaystyle\lim_{x\to\infty} f_2(x) = \infty\). If \(f_3(x)\) is defined in an analogous fashion, it is not known whether \(\displaystyle\lim_{x\to\infty} f_3(x) = \infty\) and it seems hopelessly difficult to prove for arbitrarily large prescribed \(\omega\) the existence of one sequence of \(\omega\) consecutive primes all \(\equiv 1 \bmod 4\). Using the methods developed in their series of papers on comparative prime number theory, the authors prove the theorem: For \(T\) sufficiently large, the inequality \(f_2(T) > \log^BT\) holds with positive explicitly calculable constant B. The authors conjecture that the events consisting of the possible occurrences of the two residue classes mod 4 of consecutive primes are not equally likely and that at some time it might be possible to prove that \(f_2(x) = o(x/\log x)\). [For the entire collection see Zbl 0356.00004.]











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4152618)