We show that if a system of differential equations has an elementary first integral (i.e. a first integral expressible in terms of exponentials, logarithms and algebraic functions) then it must have a first integral of a very simple form. This unifies and extends results of Mordukhai-Boltovski, Ritt and others and leads to a partial algorithm for finding such integrals.
No takes yet. Share an insight, caveat, or question.
Prelle et al. (1983) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: