Nodal domains of Maass forms. I (Q382291)

Let \(\phi\) be an even Maass cusp form and assume that it is also an eigenfunction of all the Hecke operators \(T_n, \;n \geq 1\) and the eigenvalue \(\lambda_{\phi}\) can be parametrized as \(\lambda_{\phi} = t^2_{\phi} + \frac {1}{4}\). Let \(k_{\phi}\) be the numbering of such \(\phi\) when ordered by \(\lambda_{\phi}\). \textit{A. Selberg} [Collected papers. Volume I. Berlin: Springer, 626--674 (2014; Zbl 1295.11002)] proved a Weyl law for these \(\phi\), namely \[ \lambda_{\phi} \sim 24 k_{\phi}, \;\;k_{\phi} \rightarrow \infty, \] and thus demonstrating their existence. Let \(Z_{\phi} \subset {\mathbb X} = \mathrm{SL}_2({\mathbb Z}) \backslash {\mathbb H}\) denote the nodal line of \(\phi\), i.e., the set \(\left \{ z \in {\mathbb X} : \phi (z) = 0 \right \}\) and it consists of a finite union of real analytic curves. Let \(N(\phi)\) denote the number of nodal domains of \(\phi\) (the connected components of \({\mathbb X} \backslash Z_{\phi}\)). More generally, let \(N^{\mathbb K}(\phi)\) denote the number of such domains that meet \({\mathbb K}\) for any subset \({\mathbb K} \subset {\mathbb X}\). For a picture of nodal lines for certain \(\phi\) the readers are referred to \textit{D. A. Hejhal} and \textit{B. N. Rackner} [Exp. Math. 1, No. 4, 275--305 (1992; Zbl 0813.11035)]. A conjecture of \textit{E. Bogomolny} and \textit{C. Schmit} [``Percolation model for nodal domains of chaotic wave functions'', Phys. Rev. Lett. 88, No. 11, 114102 (2002); \url{doi:10.1103/PhysRevLett.88.114102}] asserts that \[ N(\phi) \sim \frac {2}{\pi} (3{\sqrt 3}-5) k_{\phi}, \;\;k_{\phi} \rightarrow \infty. \] Courant's general domain theorem [\textit{R. Courant} and \textit{D. Hilbert}, Methods of mathematical physics. Vol. I. York: Interscience Publishers (1953; Zbl 0051.28802)] and the Selberg's Weyl law imply that \[ N(\phi) \leq k_{\phi} = \frac {1}{24} t^2_{\phi} (1+o(1)). \] The paper under review answers some questions raised in connection with the pictures of [Zbl 0813.11035], as well as proving an interesting result: \[ N^{\mathbb K}(\phi) \rightarrow \infty \;\text{as} \;k_{\phi} \rightarrow \infty, \] assuming the Lindelöf hypothesis for the \(L\)-function \(L(s, \phi)\).