No label defined (Q4944265)

[For part I of this paper see \textit{A. Langer}, J. Algebr. Geom. 8, No. 1, 41-66 (1999; Zbl 0943.14028).] Let \(X\) be a smooth projective variety of dimension \(n\) and let \(L\) be a nef and big divisor on \(X\). In order to study some properties of \(X\) (or \((X,L)\)), it is important to study linear systems \(K_{X}+tL\) for some natural number \(t\), where \(K_{X}\) is the canonical divisor of \(X\). Fujita proposed the following famous conjecture: \(K_{X}+tL\) is base point free if one of the following conditions hold: (1) \(t\geq n+1\), (2) \(t=n\) and \(L^{n}\geq 2\). In Ann. Math. (2) 127, 309-316 (1988; Zbl 0663.14010), \textit{I. Reider} proved this conjecture for the case \(n=2\). In J. Am. Math. Soc. 6, No. 4, 875-903 (1993; Zbl 0803.14004), \textit{L. Ein} and \textit{R. Lazarsfeld} proved the conjecture for the case \(n=3\). In Math. Ann. 308, 491-505 (1997; Zbl 0909.14001), \textit{Y. Kawamata} proved the conjecture for the case \(n=4\). For the case of \(n\geq 5\), it is unknown whether this conjecture is true or not. The case where \(X\) is singular has been studied by many authors. In part I of this paper (loc.cit.) the author studied the case where \(X\) is normal and \(L\) is a nef Weil divisor such that \(K_{X}+L\) is a Cartier divisor. In the previous paper the author defined an invariant \(\delta_{\zeta}\) for a \(0\)-dimensional Gorenstein subscheme \(\zeta\) of \(X\), which is called a cluster, with \(\deg\zeta =1\) or \(2\). By using this he obtained the Reider type theorems for normal surfaces. This is a generalization of the Reider theorem for the case where \(X\) is smooth. In the paper under review the author defines an invariant \(\delta_{\zeta}\) for any cluster \(\zeta\), and he obtains the following results: \flushpar Let \(X\) be a normal surface and let \(\zeta\) be a cluster, and let \(L\) be a nef Weil divisor on \(X\). Assume that \(K_{X}+L\) is Cartier and \(L^{2}>\delta_{\zeta}\). Then \(\zeta\) is in special position with respect to \(K_{X}+L\) if and only if there exists a curve \(D\) containing \(\zeta\) and such that \(\zeta\) is in (very) special position with respect to \(|\mathcal{O}_{D}(K_{X}+L)|\). Moreover, we can choose \(D\) such that \(L-2D\) is big, \(LD-(1/4)\delta_{\zeta}\leq D^{2}\) and \(0\leq LD<(1/2)\delta_{\zeta}\).