Poincaré series for \(SO(n,1)\) (Q1182366)

Using his nonholomorphic Poincaré series and Weil's bound for Kloosterman sums, Selberg established his celebrated estimate \(\lambda_ 1\geq 3/16\) for the smallest positive eigenvalue of \(-\Delta\) on any congruence subgroup of \(SL_ 2(\mathbb{Z})\). This bound has interesting number theoretic consequences. Selberg's conjecture claims that \(\lambda_ 1\geq 1/4\) on any congruence subgroup. This conjecture is a basic unsolved problem in the theory of automorphic forms. The aim of the paper under review is to develop an analogous theory for \(G=SO(r+1,1)\). Let \(\Gamma\) be a discrete cofinite subgroup of \(G\) but not cocompact. Using suitable Kloosterman sums the authors introduce a (matrix valued) Kloosterman-Selberg zeta function \(Z\) and prove its meromorphic continuation to all of \(\mathbb{C}\). The location of the poles of \(Z\) is described in terms of the decomposition of the right regular representation of \(G\) on \(L^ 2(\Gamma\backslash G)\). Of most relevance here are the principal and the complementary series, and the poles of the Eisenstein series contribute to the possible poles of \(Z\). The proof of the meromorphic continuation of \(Z\) is based on suitable Poincaré series \(P_ f\) on \(\Gamma\backslash G\). The zeta function is obtained as a Mellin transform of \(P_ f\) along a suitable subgroup. If in the definition of \(Z\) the identity representation is used, the poles of \(Z\) in \(\hbox{Re }s>r/2-1\) correspond to complementary series occurences of the right regular representation in \(L^ 2_{disc}(\Gamma\backslash G)\). By the duality theorem these poles correspond to exceptional eigenvalues of \(-\Delta\) on \(\Gamma\backslash\mathbb{H}^{r+1}\cong\Gamma\backslash G/K\). It is shown that exceptional spectrum and even poles of \(Z\) arbitrarily close to \(r/2\) may occur for general \(\Gamma\). The situation is different for \(\Gamma\) a congruence group. In what follows let \(\Gamma\) be a congruence subgroup of the unit group of a rational quadratic form. Then the smallest positive eigenvalue of \(- \Delta\) satisfies \[ \lambda_ 1\geq{1\over 2}(r-{1\over 2}),\quad \hbox { if } r\geq 2. \] This extends Selberg's eigenvalue estimate to the higher- dimensional cases. --- The method of proof is to show that \(Z\) has no poles in \(\hbox{Re} s>(r-1)/2\). In the proofs the authors work adelically. The crucial Kloosterman sums are again estimated by Weil's bound. The main results of the paper under review were also obtained by the reviewer, \textit{F. Grunewald} and \textit{J. Mennicke} [Invent. Math. 101, 641-685 (1990; Zbl 0737.11013)]. In this paper, the above eigenvalue estimate is proved for a slightly larger class of congruence groups. The methods of proof used in both papers are quite different.