Real zeros of holomorphic Hecke cusp forms (Q409046)

Let \(f\) be a holomorphic Hecke cusp form of large even weight \(k\) for the full modular group. Such form has \(k/12 + O(1)\) zeros in the fundamental domain. The recent proof of the holomorphic QUE conjecture by \textit{R. Holowinsky} and \textit{K. Soundararajan} [Ann. Math. (2) 172, No. 2, 1517--1528 (2010; Zbl 1211.11050)] implies that these zeros are equidistributed. The present paper studies the so-called real zeros of \(f\), that is, zeros on three geodesic segments on which the form (or its multiple) is real-valued. In particular it is shown that there are \(\gg k^{1/4-1/80-\varepsilon}\) zeros in the union of the geodesic segments \[ \delta_1 = \{s \in H : |s| \geq 1, \text{Re}(s) = 0 \} \quad \text{and} \quad \delta_2 = \{s \in H : |s| \geq 1, \text{Re}(s) = 1/2 \} \] and that separately on these geodesic segments, the number of zeros goes to infinity with \(k\). The proofs are based on showing that, in certain ranges, the cusp form can be approximated by a single Fourier coefficient, and then studying these coefficients. Furthermore, the authors show that there is approximately the expected number of zeros at the cusp and suggest that most zeros there are actually real. In the last section of the paper, the authors introduce a random model which they use to predict the number of zeros in each of the three geodesic segments.