On the 16-rank of class groups of \({\mathbb{Q}(\sqrt{-8p})}\) for \({p \equiv -1 \bmod 4}\) (Q2400829)

Let \(\mathcal{C}\ell (K)\) be the class group of the imaginary quadratic field \(K=\mathbb{Q}(\sqrt{-8p})\) where \(p\equiv 3\pmod{4}\). The author proves that the density of primes \(p\equiv 3\pmod{4}\) such that \(\mathcal{C}\ell(K)\) has an element of order \(16\) is equal to \(1/16\). This is an important result which goes along the Cohen-Lenstra heuristics which predict a model for the average behavior of the class groups of quadratic fields. The proof of the author is far from trivial and uses a variant of a method of Vinogradov developed by Friedlander, Iwaniec, Mazur and Rubin. The result of the paper emphasizes the evidence that the behavior of the \(16\)-rank is highly different from the behavior of the \(2\)-rank, the \(4\)-rank and the \(8\)-rank of the class groups. Using a terminology of Serre, the author shows that the \(16\)-rank of the class groups of the imaginary quadratic fields appear not to be ``motivated''. The author extends previous results of \textit{L. Redei} [J. Reine Angew. Math. 171, 55--60 (1934; Zbl 0009.05101; JFM 60.0125.02)] and \textit{É. Fouvry} and \textit{J. Klüners} [Invent. Math. 167, No. 3, 455--513 (2007; Zbl 1126.11062)] dealing with the \(4\)-rank. \textit{P. Stevenhagen} [Publ. Math. Fac. Sci. Besançon, Théor. Nombres 1988/89, No. 1, 93 p. (1989; Zbl 0701.11056)] studied the \(8\)-rank and used the notion of ``governing field''. The author do not use this last notion because he obtained some evidence that a governing field for the \(16\)-rank of \(\mathcal{C}\ell(K)\) does not exist. Some results of a few other authors come also into play in this paper.