A note on integrability and finite orbits for subgroups of \(\mathrm{Diff}(\mathbb C^n,0)\) (Q889896): Difference between revisions

The well-celebrated result of \textit{J. F. Mattei} and \textit{R. Moussu} [Ann. Sci. Éc. Norm. Supér. (4) 13, 469--523 (1980; Zbl 0458.32005)] about the topological characterization of one-dimensional holomorphic foliations in dimension \(2\), shows that the existence of holomorphic first integrals is a topological invariant. One of the basic questions in dimension \(n\) higher than \(2\) is, then, whether a germ of a one-dimensional singular holomorphic foliation topologically conjugated to another holomorphic foliation having \(n-1\) independent holomorphic first integrals should posses \(n-1\) independent holomorphic first integrals as well. Such a question is intimately related to the study of the (local) holonomy group of a foliation and, more in details, to the study of pseudogroups of holomorphic diffeomorphisms having finite orbits. The main result of the paper under review is the following. Let \(G\) be a finitely generated pseudogroup defined by a group of germs of biholomorphisms in \(\mathbb C^n\) fixing the origin. Let \(\text{Dom}(g)\) denote the domain of definition of an element \(g\in G\). Suppose that the following property \((\ast)\) holds: for every \(g\in G\) and \(p\in \text{Dom}(g)\) satisfying \(g(p)=p\), either \(p\) is an isolated fixed point of \(g\) or \(g\) coincides with the identity on a neighborhood of \(p\). Then \(G\) has finite orbits (near the origin) if and only if \(G\) is finite. Such a result is then used to prove the existence of \(n-1\) independent holomorphic first integrals for certain germs of one-dimensional holomorphic foliations in \((\mathbb C^n,O)\). More precisely, let us assume that \(\mathcal F\) is a germ of a holomorphic foliation, \(O\) is an isolated singularity (of Siegel type), there exist \(n\) invariant hyperplanes, and (up to rotations), one eigenvalue of the differential at \(O\) of a vector field representing the foliation is real positive while the others are all negative reals. Let \(h\) denote the local holonomy map associated to the direction relative to the negative eigenvalue. If \(h\) has only isolated fixed points (in the sense of \((\ast)\)) and has finite orbits, then \(\mathcal F\) has \(n-1\) independent holomorphic first integrals. The authors also construct a germ of holomorphic foliation in \(\mathbb C^3\), having a Siegel singular point at the origin and such that it does not have two independent holomorphic first integrals (but it admits one, non-constant). The holonomy associated with an invariant axis has finite orbits although it does not generate a finite subgroup of germs of holomorphic diffeomorphisms of \((\mathbb C^2,O)\). The upshot of this example is that, without the hypothesis \((\ast)\), the main result of the paper does not hold. Moreover, this example shows that the existence of a meromorphic first integral on the transverse plane plus the assumption of finite orbits for the holonomy map do not guarantee finite order for the holonomy map itself.