This paper is concerned with the form of the general solution of the equation (1), and the Theorem of Section 1 gives an answer which is somewhat different from the classical Picard result. Not only is the proof elementary, but the requests for the coefficients are much less restrictive; see the assumption ($A_0$). On the other hand, we had to introduce the additional assumption ($A_0$). Several examples are constructed in order to throw more light on the importance of those two assumptions.