Fields of moduli and definition of hyperelliptic covers (Q2492895): Difference between revisions

The subject matter of this paper has to do with the field of definition of a curve. Let \(C\) be a non-singular complex curve of genus \(g>1\). Let \(k\) be a subfield of the complex numbers \(\mathbb C\). The curve \(C\) is said to be defined over \(k\) if it is isomorphic to a curve defined over \(k\). The minimum subfield \(k=k_0\) such that for every \(\sigma\in {\text{ Gal}}({\mathbb C} | k_0)\), \(C\) is isomorphic to \(C^\sigma\), is called the field of moduli of \(C\). \textit{G. Shimura} [Nagoya Math. J. 45, 167--178 (1972; Zbl 0243.14012)] for even \(g\), and \textit{C.J. Earle} for \(g\equiv 1\pmod{4}\) [in: Adv. Theory Riemann Surfaces, Proc. 1969 Stony Brook Conf., Ann. Math. Stud. 66, 119--130 (1971; Zbl 0218.32010)] constructed curves whose field of moduli cannot be used to be the base field of definition. A hyperelliptic curve is said to be hyperelliptically defined over \(k\) if it is birational equivalent to a hyperelliptic curve of type \(y^2=q(x)\), where \(q(x)\) has is a polynomial with coefficients in \(k\) and has simple roots. For curves of even genus, \textit{J.-F. Mestre} [in: Effective methods in algebraic geometry, Proc. Symp., Castiglioncello/Italy 1990, Prog. Math. 94, 313--334 (1991; Zbl 0752.14027)] noticed that being defined over \(k\) or being hyperelliptically defined over \(k\) are equivalent concepts. Finally, \textit{T. Shaska} [SIGSAM Bull. 37(4), 117--124 (2003; Zbl 1053.14067)] has conjectured that if the automorphism group of a hyperelliptic curve has more than two elements, then its field of moduli is also a field of definition. In this paper, for \(g\equiv 1\pmod{4}\) with \(g>1\), the authors present a hyperelliptic curve whose field of moduli is \(\mathbb Q\), and such that the minimum subfield of \(\mathbb R\) where the hyperelliptically can occur is in a degree three extension of \(\mathbb Q\). In addition, the automorphism group of the curve has more than two elements. It follows that both Mestre's result and Shaska's conjecture mentioned above are false for hyperelliptic curves of odd genus.