An explicit correspondence of modular curves

From MaRDI portal



Abstract: In this paper, we recall an alternative proof of Merel's conjecture which asserts that a certain explicit correspondence gives the isogeny relation between the Jacobians associated to the normalizer of split and non-split Cartan subgroups. This alternative proof does not require extensive representation theory and can be formulated in terms of certain finite geometries modulo ell. Secondly, we generalize these arguments to exhibit an explicit correspondence which gives the isogeny relation between the Jacobians associated to split and non-split Cartan subgroups. An interesting feature is that the required explicit correspondence is considerably more complicated but can expressed as a certain linear combination of double coset operators whose coefficients we are able to make explicit.


Let \(\ell\) be a prime number and \(X=X(\ell)\) the modular curve with full level structure \(\ell\). For a subgroup \(H\subset G=\operatorname{GL}_2(\mathbb{F}_\ell)\) containing -1, let us denote by \(X_H\) the corresponding modular curve \(X/H\) given by the action of \(G\) on \(X(\ell)\), and by \(J_H\) its Jacobian variety. For \(C_{ns}\) a non-split Cartan subgroup of \(G\), and its normalizer \(N_{ns}\) the first author proved in [J. Algebra 231, No. 1, 414--448 (2000; Zbl 0990.14012)] that \(J_{C_{ns}}\) and \(J_{N_{ns}}\) are \(\mathbb{Q}\)-isogenous to certain quotients of the modular curve \(X_0(\ell^2)\). Merel conjectured that the isogenies between these varieties can be described by the correspondence induced by the quotient maps from \(X(\ell)\). This was proved in [the first author, Proc. Lond. Math. Soc. (3) 77, No. 1, 1--38 (1998; Zbl 0903.11019)] using the representation theory of \(G\) and identities in finite double coset algebras. This paper presents an alternative proof of Merel's conjecture, based on arguments by Birch and Zagier. The authors consider a finite field analogue of the complex upper half-plane and the natural action of \(G\) on it by Möbius transformations. This provides a practical description of the quotient of \(G\) by the normalizer of a Cartan subgroup of \(G\) (either split or non-split), where the conjecture is established in quite an elementary way. The proof is extended naturally to quotients of \(G\) by the Cartan groups themselves.











This page was built for publication: An explicit correspondence of modular curves

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q669225)