LMFDB definition

Atkin–Lehner involutions

Let NN be a positive integer, and let QQ be a positive divisor of NN satisfying gcd⁡(Q,N/Q)=1\gcd(Q,N/Q)=1. Then there exist x,y,z,t∈Zx,y,z,t \in \Z for which the matrix WQ=(QxyNzQt) W_Q=\left( \begin{matrix} Qx & y \\ Nz & Qt\end{matrix} \right) has determinant QQ. The matrix WQW_Q normalizes the group Γ0(N)\Gamma_0(N), and for any weight kk it induces a linear operator wQw_Q on the space of cusp forms Sk(Γ0(N))S_k(\Gamma_0(N)) that commutes with the Hecke operators TpT_p for all p∤Qp \nmid Q and acts as its own inverse.

The linear operator wQw_Q does not depend on the choice of x,y,z,tx,y,z,t and is called the Atkin-Lehner involution of Sk(Γ0(N))S_k(\Gamma_0(N)). Any cusp form ff in Sk(Γ0(N))S_k(\Gamma_0(N)) which is an eigenform for all TpT_p with p∤Np \nmid N is also an eigenform for wQw_Q, with eigenvalue ±1\pm 1.

The matrix WQW_Q induces an automorphism of the modular curve X0(N)X_0(N) that is also denoted wQw_Q.

In the case Q=NQ=N, the Atkin-Lehner involution wNw_N is also called the Fricke involution.

LMFDB source · reviewed

Andreea Mocanu, David Farmer, David Roe, Nicolas Mascot, Edgar Costa, Andrew Sutherland, John Voight. CC BY-SA 4.0. Snapshot 2026-09-22; formatting adapted for this site.

All definitions →