A complement to a paper by Dloussky on filling of holomorphic surfaces (Q939595)

It has long been conjectured that every class VII\({}^+_0\) surface (minimal surface of class VII with \(b_2>0\)) has a global spherical shell (GSS). Surfaces with a GSS, also known as Kato surfaces, are reasonably well understood via their universal covers and are necessarily degenerations of blown-up Hopf surfaces. It is not known whether a surface \(X\) that is both minimal and a degeneration of blown-up Hopf surfaces is necessarily a Kato surface. Let \(X\) be such a surface. The result of this paper is that \(X\) does have a GSS if it contains a strictly pseudoconvex real global hypersurface \(\Sigma\) (``global'' means that the complement is connected). This follows from a result of Kato, provided that \(\Sigma\) bounds a Stein surface (abstractly, not in \(X\) necessarily); in the paper of \textit{G. Dloussky} [Ann. Inst. Fourier 43, No. 3, 713--741 (1993; Zbl 0783.32008)] alluded to in the title, it is shown that a class VII\({}_0^+\) surface \(X\) with a global pseudoconvex hypersurface \(\Sigma\) and nonconstant holomorphic functions near \(\Sigma\) is a degeneration of blown-up surfaces, so from this paper it follows that a minimal degeneration with such a \(\Sigma\) is a Kato surface. The method is to show that \(\Sigma\) necessarily bounds a Stein surface. This is achieved by an application of a theorem of [\textit{H. Rossi}, Proc. Conf. Complex Analysis, Minneapolis 1964, 242--256 (1965; Zbl 0143.30301)] via a deformation \(X_t\) that for suitable \(t\) is a Hopf surface blown up in distinct points. \(\Sigma\) deforms to \(\Sigma_t\) and lifts to hypersurfaces \(\widetilde\Sigma_t\) (possibly one of many components) in the universal cover of \(X_t\). Then \(\widetilde\Sigma_t\) can be filled in with Stein surfaces and, with care, it can be shown that \(\Sigma\) can be filled in also.