Mean value results and \(\Omega\)-results for the hyperbolic lattice point problem in conjugacy classes (Q2329586)

Let \({\mathbb H}\) be the hyperbolic plane and \(\Gamma\backslash {\mathbb H}\) a hyperbolic surface where \(\Gamma\) is a cocompact or cofinite Fuchsian group. Given two points \(z\) and \(w\) in \({\mathbb H}\), and a positive real number \(X\), denote by \(\rho(z,w)\) their hyperbolic distance and consider the quantity: \[N(X;z,w):=\sharp\{\gamma\in\Gamma\mid\rho(z,w)\leq\cosh^{-1}(X/2)\}\] Let \(\Delta\) be the Laplacian of \(\Gamma\backslash {\mathbb H}\) and \(\{u_j\}_{j\geq 0}\) the \(L^2\)-normalized eigenfunctions of \(-\Delta\) with eigenvalues \(\{\lambda_j\}_{j\geq 0}\) with \(\lambda_j=s_j(1-s_j)\geq 1/4\). It is known that \[N(X;z,w)=\sum_{\frac{1}{2}s_j\leq 1}\sqrt{\pi}\frac{\Gamma(s_j-1/2)}{\Gamma(s_j+1)}u_j(z)\overline{u_j(w)}X^{s_j}+E(X;z,w)\] where \(E(X;z,w)\) is the error term. Since the spectral contribution of \(\lambda_j=1/4\) is understood, it turns out that it is enough to study the remaining error term \(e(X;z,w)\) obtained by substracting the contribution of \(\lambda_j=1/4\). Let \({\mathcal H}\) be the conjugacy class of a hyperbolic element \(g^\nu\) in \(\Gamma\), with \(g\) primitive and \(\nu\in{\mathbb N}\). Let \(\ell\) be the invariant closed geodesic of \(g\) and \(\mu\) the length of \(\ell\). In this case, one has \[N({\mathcal H},X;z):=\sharp\{\gamma\in{\mathcal H}\mid\frac{\sinh(\rho(z,\gamma z)/2)}{\sinh(\mu/2)}\leq X\}\] In the paper under review, the author studies mean value and \(\Omega\)-results for the error term \(e({\mathcal H},X;z)\). In particular, the author proves that (Theorem 1.5): \begin{itemize} \item[(a)] when \(\Gamma\) is cocompact, one has: \[\lim_{T\rightarrow\infty}\frac{1}{T}\int_0^T\frac{e({\mathcal H},e^r;z)}{e^{r/2}}dr=0\] \item[(b)] when \(\Gamma\) is cofinite but not cocompact, one has: \[\lim_{T\rightarrow\infty}\frac{1}{T}\int_0^T\frac{e({\mathcal H},e^r;z)}{e^{r/2}}dr=\frac{\mid\Gamma(3/4)\mid^2}{\pi^{3/2}}\sum_{\mathfrak a}\widehat{E}_{\mathfrak a}(1/2)E_{\mathfrak a}(z,1/2)\] where \(E_{\mathfrak a}\) denotes the nonholomorphic Eisenstein series corresponding to the cusp \({\mathfrak a}\) while \(\widehat{E}_{\mathfrak a}\) is the Eisenstein period. \end{itemize}