On the cyclic torsion of elliptic curves over cubic number fields. II (Q2199511): Difference between revisions

Let \(K\) be a number field. The strong uniform boundedness theorem of \textit{L. Merel} [Invent. Math. 124, No. 1--3, 437--449 (1996; Zbl 0936.11037)] asserts that there exists a bound \(B_d\) that depends only on the degree of \(K\) such that given an elliptic curve \(E\) defined over \(K\), the size of the torsion part \(E(K)_{\mathrm{tors}}\) of \(E(K)\) is bounded by \(B_d\). When \(K=\mathbb{Q}\), Mazur described completely all possibilities for \(E(K)_{\mathrm{tors}}\). We owe Mazur, Momose, Kamienny and Kenku for performing the classification when \(K\) is a quadratic extension. For cubic fields \(K\), we still do not have such a complete picture of the possibilities for the structure of \(E(K)_{\mathrm{tors}}\). In this article, the author proves that certain cyclic groups are not realized as subgroups of \(E(K)_{\mathrm{tors}}\) when \(K\) is a cubic field. Namely, these groups are \(\mathbb{Z}/22\mathbb{Z} \), \(\mathbb{Z}/25\mathbb{Z}\), \(\mathbb{Z}/40\mathbb{Z}\) and \(\mathbb{Z}/49\mathbb{Z}\). The latter result is attained by computing the gonalities of some modular curves. For Part I see [the author, J. Number Theory 183, 291--308 (2018; Zbl 1433.11077)].