A new proof of the Ahlfors five islands theorem (Q1283065)

The article shows how several results of Ahlfors and Nevanlinna can be derived from each other, and how they all follow from a rescaling lemma due to Zalcman and a generalization of the Schwarz lemma. The results by Ahlfors use five disjoint simply connected domains \(D_i\subset\mathbb{C}\) and, for any domain \(D\subset\mathbb{C}\), the family \(F_A:=\{f: D\to\overline\mathbb{C}\): no subdomain of \(D\) is mapped conformally onto one of the domains \(D_i\}\). Ahlfors' Five Islands Theorem says that every \(F_A\) is normal; for \(D=\mathbb{C}\), a related result says that \(F_A\) contains only the constant functions. Similarly, for the Nevanlinna theorems, five different points \(a_i\in\mathbb{C}\) are used, together with the family \(F_N:=\{f:D\to\overline\mathbb{C}\): all values \(a_i\) are totally ramified\(\}\). The results are similar: all \(F_N\) are normal families, and for \(D=\mathbb{C}\) they contain only the constant functions. The paper is to a large extent self-contained, including proofs of the Zalcman lemma and other used results, and suggests ``a serious candidate in the competition for the simplest proof of the theorems of Picard and Montel''.