Holomorphic extension of proper meromorphic mappings (Q807789)

A result of Forstnerič states that every proper holomorphic mapping from a ball in \({\mathbb{C}}^ n\) to a ball in \({\mathbb{C}}^ N(N\geq n\geq 2)\) that is sufficiently smooth on the closure of the ball is in fact a rational mapping [\textit{F. Forstnerič}, Invent. Math. 95, No.1, 31-62 (1989; Zbl 0633.32017)]. His argument left open the possibility that such a mapping could be indeterminate at a point of the boundary sphere. Cima and Suffridge then showed that this does not occur [\textit{J. A. Cima} and \textit{T. J. Suffridge}, Duke Math. J. 60, No.1, 135-138 (1990; Zbl 0694.32016)], by proving a local result to the effect that certain proper maps of the ball must extend holomorphically past a boundary point (recall that this deals with the non-equidimensional case). The article under review gives a generalization of the result of Cima and Suffridge, as follows: Theorem: Suppose that M is a real analytic (non-singular) real hypersurface in \({\mathbb{C}}^ n\), U is a neighborhood of a point \(p\in M\), and \(\Omega\) is the portion of U lying on one side of M. Suppose that \(F:U\to {\mathbb{C}}^ N\) is a meromorphic mapping whose restriction to \(\Omega\) maps \(\Omega\) holomorphically into the open unit ball in \({\mathbb{C}}^ N(N\geq n)\), and that for each point \(q\in U\cap M\), \(\| F(z)\| \to 1\) as \(z\to q(z\in \Omega)\). Then F extends to a holomorphic mapping in some neighborhood of p. The point here is that no geometric assumptions are made on M in addition to its real analyticity.