On the rationality of complex homology 2-cells. II (Q1262918)

[For part I see ibid., No.1, 37-56 (1989; Zbl 0647.14029)]. In this second part of a 2-part paper, the following result is proved: Theorem: Let X be an irreducible, smooth projective surface/\({\mathbb{C}}\) with geometric genus \(P_ g(X)=0\). Let D be a reduced (not necessarily connected) curve on X with at worst ordinary double point singularities. Suppose each connected component of supp(D) is simply connected and the irreducible components of D generate the divisor class group Pic(X). Then X is not an elliptic surface of Kodaira dimension 1. Combining with part I of this paper, we have the result that a surface X as in the theorem is rational. In particular, each smooth contractible surface/\({\mathbb{C}}\) is rational, thus answering a question of Van de Ven affirmatively. The techniques developed in part I are used. In the present case, the 2nd Betti number of the minimal model of X (assumed to be elliptic) is 10. In part I, this number was 9, this accounts for the increase in the work needed in this part. Results of T. Fujita on detailed properties of the Zariski-Fujita decomposition of a pseudo- effective divisor are used.