The necessary condition on the fiber-sum decomposability of genus-2 Lefschetz fibrations (Q624262)

In the paper under review the author proves two results about Lefschetz fibrations over \(S^2\). The first one is a counter-example to a conjecture of \textit{A. Stipsicz} [Proc. Am. Math. Soc. 129, 1499--1502 (2001; Zbl 0978.57022)] (conjecture 2.4) which states that any Lefschetz fibration over the sphere without non-trivial fiber-sum decompositions admits a section with self-intersection number \((-1)\). The converse of this result is true: A. Stipsicz proved it in the cited paper. The counter-example is a genus-2 Lefschetz fibration which was introduced by \textit{D. Auroux} [Turkish J. Math. 27, No.~1, 1--10 (2003; Zbl 1075.53087)]. This Lefschetz fibration does not admit sections of self-intersection number \((-1)\) but has a double section, i.e. a pseudo-holomorphic sphere with self-intersection number \((-1)\) which intersects any generic fiber twice. Y. Sato shows that Lefschetz fibrations of genus \(g\geq 2\) which have the latter property cannot be decomposed as non-trivial fiber-sums. The second result is a characterization of genus-2 Lefschetz fibrations over the sphere with \(b_{2}^{+}=1\) which are decomposable as non-trivial fiber-sums. Y. Sato proves that if \(X= X_{1}\#X_{2}\), then \(X\) is not a rational nor a ruled surface and necessarily each \(X_{i}\) is diffeomorphic to \(S^{2}\times T^{2}\# 3\bar{\mathbb CP}^{2}\) or to \(S^{2}\times T^{2}\# 4\bar{\mathbb CP}^{2}\).