First integrals versus configurational invariants and a weak form of complete integrability (Q1067673)

A confusion over the concept of first integrals, which has been created in a recent paper by \textit{L. S. Hall} [Physica D 8, 90-116 (1983; Zbl 0538.58020)] is clarified. The clear distinction between first integrals and functions which are first integrals only on a specific, fixed hypersurface is discussed. Hall's terminology of configurational invariants is adopted for the latter case. The possible relevance of knowing configurational invariants for a Hamiltonian system is illustrated by results concerning a weak form of the theory on complete integrability.