The Weyl law for subgroups of the modular group (Q1266179)

Let \(\Gamma\) be a cofinite discrete subgroup of PSL\( (2,\mathbb{R})\) acting on the upper half-plane \(\mathbb{H}\), and let \(\chi\) be a unitary character on \(\Gamma\). Consider the eigenvalue problem of the automorphic Laplacian \(-\Delta\) on \(L^2(\Gamma\setminus\mathbb{H},\chi)\). Concerning the problem of existence of eigenvalues of \(-\Delta\) on \(L^2(\Gamma \setminus \mathbb{H}, \chi)\), there are two conjectures striving against each other, the Roelcke-Selberg conjecture and the Phillips-Sarnak conjecture. The authors consider a family of examples which shed further light on these conjectures: For any natural number \(d\), let \(\Gamma_d\) denote the subgroup of the principal congruence subgroup \(\Gamma(2) <\text{PSL}(2,\mathbb{Z})\) consisting of all \(\gamma= A^{n_1} B^{m_1} \cdot\dots \cdot A^{n_k} B^{m_k} \in\Gamma (2)\) such that \(d\) divides \(n_1+ \cdots +n_k\) where \(A,B\) freely generate \(\Gamma (2)\). The group \(\Gamma_d\) is a normal subgroup of \(\Gamma (2)\) of index \(d\), and \(\Gamma_d\) is a congruence subgroup of PSL\( (2,\mathbb{Z})\) if and only if \(d\in \{1,2,4,8\}\). The authors discuss the eigenvalue problem for \(\Gamma= \Gamma_d\), \(\chi=1\) and for \(\Gamma= \Gamma(2)\), \(\chi (\gamma) =\chi_{n,d} (\gamma)= \exp\left( {2\pi in\over d} (n_1+ \cdots +n_k) \right)\) (in the above notation) where \(0\leq n\leq d-1\). An interesting lemma states: If the Weyl formula holds for \((\Gamma_d,1)\), then it holds for all \((\Gamma(2),\chi_{n,d})\), \(0\leq n\leq d-1\). As a preparation for the proof of the main result, the authors develop the theory of oldforms and newforms for the cuspidal part of \(L^2(\Gamma(2) \setminus \mathbb{H}, \chi=1)\) and prove an interesting characterization of the space of newforms. This implies that the latter space of newforms is invariant under \(-\Delta\), and this enables the authors to compute the trace of \(h(-\Delta)\) on the space of newforms for many test functions \(h\). This leads to the observation that the eigenvalues of \(-\Delta\) on the spaces of oldforms and newforms are distributed differently. The main result of the paper under review is Theorem 3, which says that under certain assumptions (including the main multiplicity conjecture) there exist only finitely many natural numbers \(d\) such that Weyl's law holds for the eigenvalue distribution of \(\Gamma_d\).