Corrigendum to ``The Conley conjecture for Hamiltonian systems on the cotangent bundle and its analogue for Lagrangian systems'' (Q642518)

The original paper [J. Funct. Anal. 256, No. 9, 2967--3034 (2009; Zbl 1184.53080)] relied on the assumption that \(\mathcal{L}_\tau(\gamma):=\int_0^\tau L\big(t,\gamma(t),\dot{\gamma}(t)\big)\,dt\) is a \(C^2\)-smooth functional. But as proved by \textit{A. Abbondandolo} and \textit{M. Schwarz} [Adv. Nonlinear Stud. 9, No. 4, 597--623 (2009; Zbl 1185.37145)] in 2009 this only happens if the Lagrangian \(L(t,q,v)\) is a polynomial in \(v\) of degree no more than two restricting all the results to this special case. To restore the original generality the author posted a first correction on arXiv in 2009, where he relied on \textit{J. Ming}'s splitting lemma [Nonlinear Anal., Theory Methods Appl. 36, No. 8, A, 943--960 (1999; Zbl 0927.58005)]. That correction contained a gap related to isomorphisms of critical groups at isolated critical points. Rather than filling the gap the author now proves a new splitting lemma valid for \(C^1\)-smooth functionals. He then proves an abstract theorem that has the isomorphism theorems 4.4, 4.7 of the original paper as corollaries obviating the need for \(C^2\)-smoothness in their proofs. In addition, section 4.3 of the original paper, containing results on \(C^4\)-smooth Hamiltonians on \(C^5\)-smooth manifolds, is rewritten completely along lines that make the new splitting lemma sufficient for the proofs.