Domains with strong barrier (Q913020)

Let \(\Omega\) be a multiply connected plane domain. It possesses a strong barrier if there is a nonconstant positive superharmonic function u in \(\Omega\) such that \[ \Delta u(z)\leq -cu(z)dist(z,\delta \Omega)^{- 2}\text{ for some }c>0. \] There are several other characterizations, e.g. that \(\Omega\) does not contain doubly connected subdomains of arbitrarily large module that separate boundary components, i.e. \(\partial\Omega\) is uniformly perfect. The domain is called a Hayman-Wu domain if \(F^{-1}(\Omega \cap L)\) is bounded by a constant independent of the universal covering map F of the unit disk \(\Delta\) onto \(\Omega\) and also of the line L. The main result of the paper is that \(\Omega\) is Hayman-Wu \(\Rightarrow\Omega\) has a strong barrier \(\Rightarrow\) \(\delta(\Gamma)<1,\) where \(\delta(\Gamma)\) is the exponent of convergence of the Fuchsian group \(\Gamma\) for which \(\Omega\cong \Delta /\Gamma\). Indicentally this shows that a ``counterexample'' in the reviewer's paper [Analysis 4, 299- 321 (1984; Zbl 0501.30036)] was false. If \({\mathbb{C}}\setminus \Omega\) is a compact set on \({\mathbb{R}}\) then \(\Omega\) is a Hayman-Wu domain if and only if \(\partial \Omega\) is homogeneous. Finally, the author discusses the problem whether \(\Gamma\) is fully accessible if \(\Omega\) is a Hayman-Wu domain and furthermore shows that there are domains with arbitrarily small \(\delta(\Gamma)\) that do not have a strong barrier.