On the inverse Sturm-Liouville problem (Q2455010)

The authors consider the Sturm-Liouville problem \[ -[p(x) \phi'(x)]' + q(x) \phi(x) = \lambda y(x) \phi(x),\quad x \in {\mathbb R}, \] where \(p > 0\) and \(y > 0\). They assume that its spectrum \(\Sigma\) has the form \[ \Sigma = [\lambda_0, \lambda_1] \cup \cdots \cup [\lambda_{2g}, \infty), \] where \(-\infty < \lambda_0 < \lambda_1 < \cdots < \lambda_{2g}\). They also assume that the Lyapunov exponent \(\beta(\lambda)\) vanishes on \(\Sigma\). The article describes how, under the above assumptions, \(p, q\), and \(y\) can be reconstructed from \(\Sigma\) and the (finite) poles of the Weyl \(m\)-function which (due to the finite-zone structure of \(\Sigma\)) can be viewed as a meromorphic function on \({\mathcal R}\), the compact Riemann surface of \(w(\lambda)\), where \(w(\lambda)^2 = -\prod_{j=0}^{2g} (\lambda - \lambda_j)\). The number of these poles of \(m\) is equal to the genus \(g\) of \({\mathcal R}\). One wonders whether this kind of equality remains valid for higher-order operators.