Let X be a smooth manifold belonging to one of these three collections: acyclic manifolds (compact or not, possibly with boundary), compact connected manifolds (possibly with boundary) with nonzero Euler characteristic, integral homology spheres. We prove that Diff(X) is Jordan. This means that there exists a constant C such that any finite subgroup G of Diff(X) has an abelian subgroup whose index in G is at most C. Using a result of Randall and Petrie, we deduce that the automorphism groups of connected, non-necessarily compact, smooth real affine varieties with nonzero Euler characteristic are Jordan.The proof of Theorem 1.2 uses a result by Turull and the author [36], which relies on the classification of finite simple groups (see Section 3).In order to explain the context of the results in this paper, it will be convenient to use the following terminology, introduced by Popov [39]: a group Γ is said to be Jordan if any finite subgroup G Γ has an abelian subgroup whose index in G is bounded above by a constant depending only on Γ.Theorem 1.2 gives a positive partial answer to the question, posed byÉtienne Ghys, of whether diffeomorphism groups of compact manifolds are Jordan (see Question 13.1 in [18] † ). The particular case in which X is a sphere was also independently asked in several talks by Walter Feit and later by Zimmermann [52, § 5].Other partial answers to Ghys's question have been given in the past. It has been proved that Diff(X) is Jordan if X is a compact manifold of dimension at most 2 (this is an easy exercise; see [30, Theorem 1.3] for the case of surfaces), a compact 3-manifold (Zimmermann [53]), a compact 4-manifold with nonzero Euler characteristic [31], or a closed manifold with a nonzero top dimensional cohomology class expressible as a product of one-dimensional classes (for example, tori) [30]. One can also study Jordan's property for diffeomorphism groups of open manifolds. It is easy to prove that Diff(R) and Diff(R 2 ) are Jordan; the work of Meeks and Yau on minimal surfaces [27, Theorem 4] implies that Diff(R 3 ) is Jordan; Guazzi, Mecchia and Zimmermann later proved in [20] that Diff(R 3 ) and Diff(R 4 ) are Jordan using much more elementary methods than those of [27], and Zimmermann proved in [53] that Diff(R 5 ) and Diff(R 6 ) are Jordan.An example of connected open 4-manifold whose diffeomorphism group is not Jordan was given by Popov in [40]. Csikós, Pyber, and Szabó [11] found the first example giving a negative answer to Ghys's question, showing that the diffeomorphism group of T 2 × S 2 is not Jordan (but note that, in contrast, the symplectomorphism group of any symplectic form on T 2 × S 2 is Jordan; see [33]). Many other examples of manifolds giving a negative answer to Ghys's question can be obtained using the ideas in [11]: in particular, for any manifold M supporting an effective action of SU(2) or SO(3, R), the diffeomorphism group of T 2 × M is not Jordan (see [32]). It is an intriguing question to characterize which compact smooth manifolds have Jordan diffeomorph...