##
**The \(C^ 1\) closing lemma, including Hamiltonians.**
*(English)*
Zbl 0548.58012

Let \(\gamma\) be a trajectory of a dynamical system X which is recurrent. Is there a dynamical system close to X with a periodic trajectory close to \(\gamma\) ? This question is the closing problem for \(\gamma\). As one makes the question precise by assigning a topology to an appropriate space of dynamical systems the solution to this problem can be quite easy or, as is usually the case, difficult in the extreme. Aside from the obvious intrinsic interest of the problem, dynamicists are interested in the closing problem because it is the key ingredient in theorems of general density. For example, the original closing lemma for a recurrent orbit of a \(C^ 1\) diffeomorphism of a compact manifold due to the first author [Am. J. Math. 89, 956-1009, 1010-1021 (1967; Zbl 0167.218)] implies the general density theorem: If M is a compact manifold then the generic diffeomorphism in the \(C^ 1\) topology has its periodic points dense in its nonwandering set.

In the paper under review the authors study a new property which they call the lift axiom which is formulated separately for subsets of diffeomorphisms, flows and vector fields. For diffeomorphisms one considers a subset S of the \(C^ 1\) diffeomorphisms of a compact Riemannian manifold with exponential map exp which imbeds each unit ball in the unit sphere bundle into M. S satisfies the lift axiom if for each \(f\in S\) and each \(C^ 1\) neighborhood U of f there is an \(\epsilon >0\) such that whenever V is a unit tangent vector at p there is a diffeomorphism close to the identity satisfying \(g\circ f\in U\) and (L1) \(g(p)=\exp (\epsilon v)\), (L2) the set of all points where g is not the identity is contained in \(\exp_ p(T_ pM(r))\), the exponential image of the unit r ball and (L3) if \(g_ 1,...,g_ n\) are several such perturbations with disjoint support then \(g_ 1\circ...\circ g_ n\circ f\in S\). The main result is the following: if S satisfies the lift axiom then S has the closing property. This result is used to prove the original closing lemma and the \(C^ 1\) case of Poincaré’s conjecture on Hamiltonian vector fields that \(C^ r\) generically in the space of Hamiltonian vector fields the periodic trajectories are dense in the compact energy surfaces.

The paper contains an informative introduction which is recommended even to the nonspecialist as a precise account of the results of the paper; see also the earlier paper of the second author [Lect. Notes Math. 668, 225-230 (1978; Zbl 0403.58020)].

In the paper under review the authors study a new property which they call the lift axiom which is formulated separately for subsets of diffeomorphisms, flows and vector fields. For diffeomorphisms one considers a subset S of the \(C^ 1\) diffeomorphisms of a compact Riemannian manifold with exponential map exp which imbeds each unit ball in the unit sphere bundle into M. S satisfies the lift axiom if for each \(f\in S\) and each \(C^ 1\) neighborhood U of f there is an \(\epsilon >0\) such that whenever V is a unit tangent vector at p there is a diffeomorphism close to the identity satisfying \(g\circ f\in U\) and (L1) \(g(p)=\exp (\epsilon v)\), (L2) the set of all points where g is not the identity is contained in \(\exp_ p(T_ pM(r))\), the exponential image of the unit r ball and (L3) if \(g_ 1,...,g_ n\) are several such perturbations with disjoint support then \(g_ 1\circ...\circ g_ n\circ f\in S\). The main result is the following: if S satisfies the lift axiom then S has the closing property. This result is used to prove the original closing lemma and the \(C^ 1\) case of Poincaré’s conjecture on Hamiltonian vector fields that \(C^ r\) generically in the space of Hamiltonian vector fields the periodic trajectories are dense in the compact energy surfaces.

The paper contains an informative introduction which is recommended even to the nonspecialist as a precise account of the results of the paper; see also the earlier paper of the second author [Lect. Notes Math. 668, 225-230 (1978; Zbl 0403.58020)].

Reviewer: C.Chicone

### MSC:

37J99 | Dynamical aspects of finite-dimensional Hamiltonian and Lagrangian systems |

58D05 | Groups of diffeomorphisms and homeomorphisms as manifolds |

37C10 | Dynamics induced by flows and semiflows |

58A10 | Differential forms in global analysis |

PDFBibTeX
XMLCite

\textit{C. C. Pugh} and \textit{C. Robinson}, Ergodic Theory Dyn. Syst. 3, 261--313 (1983; Zbl 0548.58012)

Full Text:
DOI

### References:

[1] | DOI: 10.2307/1970458 · Zbl 0156.43803 |

[2] | DOI: 10.1090/S0002-9904-1974-13470-1 · Zbl 0296.57008 |

[3] | DOI: 10.1007/BFb0101794 |

[4] | DOI: 10.2307/1994022 · Zbl 0141.19407 |

[5] | Lang, Introduction to Differential Manifolds (1962) |

[6] | Hirsch, Differential Topology (1976) · Zbl 0356.57001 |

[7] | Hartman, Ordinary Differential Equations (1964) · Zbl 0125.32102 |

[8] | DOI: 10.1007/BF02052876 · Zbl 0205.28301 |

[9] | Choquet, Lectures on Analysis 1 (1969) |

[10] | Abraham, Transversal Mappings and Flows (1967) · Zbl 0171.44404 |

[11] | Abraham, Foundations of Mechanics (1967) |

[12] | DOI: 10.2307/2373361 · Zbl 0212.56502 |

[13] | DOI: 10.1016/0022-0396(75)90054-6 · Zbl 0309.34043 |

[14] | DOI: 10.2307/1996419 · Zbl 0288.15003 |

[15] | DOI: 10.2307/2373414 · Zbl 0167.21804 |

[16] | DOI: 10.2307/2373413 · Zbl 0167.21803 |

[17] | Poincaré, Les Methodes Nouvelles de la Mecanique Célèste 1 pp 82– (1892) |

[18] | DOI: 10.1016/0022-0396(67)90026-5 · Zbl 0153.40901 |

[19] | DOI: 10.1016/0040-9383(65)90018-2 · Zbl 0107.07103 |

[20] | Palis, Lecture Notes in Math. 468 pp 345– (1975) |

[21] | DOI: 10.1090/S0002-9904-1967-11798-1 · Zbl 0202.55202 |

This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.