Strong rigidity of positive quaternion-Kähler manifolds (Q1340650)

It is known that complete quaternion-Kähler manifolds of positive scalar curvature must be compact, and the only known examples are symmetric. It is also known that non-symmetric examples do not exist in dimension 4 and 8. This paper supports the conjecture that there does not exist a non-symmetric quaternion-Kähler manifold with positive scalar curvature in higher dimensions. The authors prove that for any \(n\), there are, up to isometries and rescalings, only, finitely many compact quaternion-Kähler \(4n\)- manifolds of positive scalar curvature. Moreover, if \(M\) is a compact quaternion-Kähler manifold of positive scalar curvature then \(\pi_ 1(M) = 0\) and \(\pi_ 2(M) = 0\), \(\mathbb{Z}\), finite with 2-torsion iff \(M\) is respectively \(\mathbb{H} \mathbb{P}_ n\), \(\text{Gr}_ 2(\mathbb{C}^{n + 2})\), otherwise. The corresponding characterization of \(\text{Gr}_ 2(\mathbb{C}^{n + 2})\) is deduced in this paper by applying Mori theory to the twistor space, which is a complex \((2n + 1)\)-manifold \(\mathcal Z\) associated to \(M\) admitting a holomorphic contact structure. The authors use vanishing theorems derived from the complex contact structure of \(\mathcal Z\) and the Clebsch-Gordan formula of a virtual vector bundle for which an associated index is zero to prove the following relations for the Betti numbers: \(b_{2k + 1} = 0\), for all \(k \geq 0\), \(b_ 2 \leq 1\), with equality iff \(M = \text{Gr}_ 2(\mathbb{C}^{n + 2})\), \(\sum^{n - 1}_{r = 0} [6r(n - 1 - r) - (n - 1)(n - 3)]b_{2r} = {1\over 2} n(n - 1) b_{2n}\). Let us mention that the first relation is known and the last one was previously known only for \(n = 2\) and supplements `Kähler- type' inequalities.