Some remarks on geodesic and curvature preserving mappings (Q1355111): Difference between revisions

R. S. Kulkarni and S. T. Yau were occupied with the problem whether a curvature-preserving diffeomorphism between two Riemannian manifolds is an isometry. The answer is positive in the ``generic'' situation but there are also interesting counterexamples. Note that ``curvature-preserving'' concerns here preservation of the curvature tensor of type \((0,4)\) (or, equivalently, of the sectional curvatures). The situation is completely different if one assumes the curvature tensor of type \((1,3)\) to be preserved. Here a ``generic'' result says that (for dimension greater than two!) a curvature-preserving diffeomorphism is a homothety (cf., e.g., C. Teleman). There is no reason to expect an isometry, in general. The present authors prove the following special result for the second kind of problems: a curvature-preserving geodesic mapping between two surfaces in Euclidean 3-space is either homothetic, or both surfaces are locally flat and the mapping is a (usual) affine mapping. Let us remark that the formulation in the article is a bit misleading.