A proof of the existence of Poincare's second species solution in the restricted three body problem is given. It is not based, as Perko's and Guillaume's work, on singular perturbation and asymptotic approximation but rather on topological equivalence between differential systems in the neighborhood of an equilibrium.