Some remarks about Reider's article `On the infinitesimal Torelli theorem for certain irregular surfaces of general type' (Q579349)

Generalizing Reider's constructions [cf. \textit{I. Reider}, Math. Ann. (to appear; Zbl 0617.14006)] to n dimensions the infinitesimal Torelli theorem is proved for n-folds X with the following properties: (a) For all d with \(1\leq d\leq n-1\) there does not exist a rational map of X onto a d-dimensional variety with geometric genus \(\geq d+1,\) (b) \(H^{q+1,n-q}(K_ X)=0\) for \(q=1,...,n-1,\) (c) \(c_ n(\Omega ^ 1_ X)\geq 2\sum ^{n}_{j=1}(-1)^ j\chi _{j-2}(\Omega _ X^{n-j}),\) where for any coherent sheaf \({\mathcal F}\) we set \(\chi _ k({\mathcal F})=\sum ^{k}_{j=1}(-1)^ j\dim H^ j({\mathcal F}),\quad \chi _ k({\mathcal F})=\dim H^ 0({\mathcal F})\) for \(k\leq 0,\) (d) The cotangent bundle is spanned by 1-forms and the resulting Gauss- map has degree 1. In case \(n=2\) properties (b) and (c) are always true, (a) means that X does not have an irrational pencil and one obtains Reider's theorem. - In case \(n=3\), (c) holds and (b) reduces to \(H^{2,2}(K_{\chi})=0\) which holds e.g. if \(K_{\chi}\) is ample. It is probably true that (c) always holds for varieties satisfying (a), but no proof is yet available.