No label defined (Q4352972)

It is a well known fact in several complex variables that proper holomorphic mappings have strong rigidity properties. Such maps do not exist between two randomly chosen domains in \(\mathbb{C}^n\) when \(n>1\) and, consequently, very few domains admit non trivial proper holomorphic self-maps. It should be emphasized that the reason of this phenomenon is not completely understood. The authors survey several results describing some of the known obstructions to the branching of proper holomorphic self-maps. A special attention is devoted to Reinhardt domains. On the other hand, some domains may admit non-injective proper holomorphic self-maps. The bidisc, which is a classical example of such domains, may actually be considered as the simplest basin of attraction in \(\mathbb{C}^n\). The class of basins is a very natural class of domains admitting non-injective proper holomorphic self-maps. The main purpose of this note is to show how large, in the sense of holomorphic equivalence, this class is. It should be pointed out that distinct basins may present very different geometrical properties. It is proved that basins, associated to quasi-homogeneous polynomial maps of different weights, are not holomorphically equivalent. Some results on the non-equivalence among basins associated to homogeneous mappings are obtained. Various concrete examples are discussed. In particular, a pair of basins in \(\mathbb{C}^2\) which are unequivalent but properly mapped to each other is presented. A family of basins with almost everywhere strictly pseudoconvex, and even spherical, boundaries is also described. These basins, which are in some sense close to the euclidean ball, are unexpected examples of domains with non-injective proper holomorphic mappings. For the entire collection see [Zbl 0869.00034].