Elliptic curves with torsion groups \(\mathbb{Z}/8\mathbb{Z}\) and \(\mathbb{Z}/2\mathbb{Z} \times \mathbb{Z}/6\mathbb{Z}\) (Q1982626)

The authors have been interested in construction of families of elliptic curves with torsion group of a prescribed type. In this article, they deal with elliptic curves over a function field of one variable. They present seven elliptic curves with rank \(2\) and torsion group \(\mathbb Z/8\mathbb Z\) and five curves with rank \(2\) and torsion group \(\mathbb Z/2\mathbb Z\times\mathbb Z/6\mathbb Z\). For those twelve curves, they give generators of the Mordell-Weil group. By specializing the parameter, they give particular example of curves with high rank over \(\mathbb Q\). They explain how to obtain new points in those models and present several families of infinitely many elliptic curves in both torsion groups and rank at least \(3\). The rank conjecture says that there exist elliptic curves over \(\mathbb Q\) of arbitrary high rank. But some heuristic by \textit{J.Park} et al. [J. Eur. Math. Soc. (JEMS) 21 No. 9, 2859--2903 (2019; Zbl 1469.11173)] predicts the existence of a universal bound for the rank and give a specific bound for each torsion group. In each case of \(\mathbb Z/8\mathbb Z\) and \(\mathbb Z/2\mathbb Z\times\mathbb Z/6\mathbb Z\), the given bound is \(3\). To evaluate this bound, the authors collect numerical data of root numbers of elliptic curves with rank at least \(3\) in the families obtained from their curves by specializations. Under the parity conjecture, the bound \(3\) implies that root number would not be evenly distributed over these families. But the data obtained suggests that this might not be the case. They remark that the heuristic needs some adjustments at least in the case of curves with the torsion group under consideration.