Periodic orbits for vector fields with nondegenerate first integrals (Q5943233)

This paper, geometrical and very readable, gives two geometrical proofs of the existence of periodic orbits on each level set of a nondegenerate first integral. First, Theorem 2.3 is a slighly less general version (because of nondegenerate first integral and class \({\mathcal C}^2)\) of Moser's theorem. Its proof is based on a blow up adapted to the first integral and the Fuller index of periodic orbits. The nondegeneracy hypothesis is used in the description of the level sets, hopeless otherwise. The result obtained is like Moser's: there is at least one periodic orbit in each level set close enough to the singular point and their minimal periods tend to \(2\pi\) as we approach the singular point. This is for the nonresonance case (the vector field near the singular point is a perturbation of its linear field). But the author's proof generalises well to a resonance case with a suitable jet of the field \(x\) substituted for the differential \(Dx(0)\). This gives Theorem 3.1 of the paper: existence of periodic orbits in the level sets of the nondegenerate first integral of vector fields that are extentions of a ``good'' 3-jet \({\mathcal A}\) (all its extentions have nonzero Fuller index).