Embedding some bordered Riemann surfaces in the affine plane (Q1812512)

On the question whether every open surface admits a proper holomorphic embedding in \(\mathbb{C}^2\) several special cases are known: the unit disc, annuli and the punctured disk. A more general result was given in [\textit{J. Globevnik} and \textit{B. Stensønes}, Math. Ann. 303, 579-597 (1995; Zbl 0847.32030)]: every finitely connected domain in \(\mathbb{C}\) with no isolated points in the boundary embeds in \(\mathbb{C}^2\). In the paper under review the embedding problem for bordered Riemann surfaces is considered. Let \(\mathcal{R}\) be a compact, orientable Riemann surface of genus \(g_\mathcal{R}\), with boundary \(b\mathcal{R}\) consisting of \(m\geq 1\) discs removed. Such a surface is said to be of class \({\mathcal F}\) if it admits an injective immersion \(F=(f,g):\mathcal{R} \rightarrow \overline{U} \times \mathbb{C}\), which is holomorphic in \(\overset\circ{\mathcal R}={\mathcal R} \setminus b{\mathcal R}\) and such that \(f\) is an inner function with degree deg\((f)\geq 2g_{\mathcal R}+m-1\), where \(U\) denotes the unit disc. The authors prove that if \(\mathcal{R}\) is a bordered surface of class \(\mathcal{F}\) then \(\overset\circ{\mathcal R}\) admits a proper holomorphic embedding in \(\mathbb{C}^2\). As a consequence, if \(\mathcal{R}\) is a hyperelliptic bordered Riemann surface (and in particular, if \(\mathcal{R}\) is a torus with one hole) then \(\overset\circ{\mathcal R}\) embeds into \(\mathbb{C}^2\). Let \({\mathcal M}({\mathcal R})\) denote the moduli space and \({\mathcal F}({\mathcal R})\subset{\mathcal M}({\mathcal R}) \) be the set consisting of structures of class \(\mathcal{F}\). In their main Theorem the authors prove that \({\mathcal F}({\mathcal R})\) is nonempty and open in \({\mathcal M}({\mathcal R})\) and for each \([J]\in {\mathcal F}({\mathcal R})\) the open Riemann surface \((\overset\circ{\mathcal R},J)\) admits a proper holomorphic embedding in \(\mathbb{C}^2\). In particular, if \(\mathcal{R}\) is a finitely connected domain in \(\mathbb{C}\) then \({\mathcal F}({\mathcal R})={\mathcal M}({\mathcal R})\), and hence the above result of Globevnik and Stensønes is re-proved. The authors pose the question whether \({\mathcal F}({\mathcal R})={\mathcal M}({\mathcal R})\) for every bordered surface \(\mathcal{R}\). In their opinion the answer seems to be negative.