The proof of Proposition 5 of [3] is incomplete. With notation as in the paper, the possibility that the polynomial X~/q'P o + X~/~' P1 + X~/q" Pz in (11) could be identically zero was overlooked. We will sketch here a proof that in this case X does not have controlled singularities so this case can indeed be discarded in the proof of Proposition 5. and P1 is a zero of X~/q' P2 and, since we are in the generic case, is a zero of PE and gives a singular point of F = 0 with Jacobian ideal of multiplicity at least q' since c~F/~X i = P~'. rather than ~ ee < 2d.
P~XWe would like to thank M. Homma who pointed out to us the case overlooked and suggested part of the above argument 1o show that any counterexample would be singular.
References[1] E. BALLICO and A. HEwz, Non-reflexive projective curves of low degree. Manuscripta Math. 70, 385-396 (1991).[2] A. HEFEZ, Non reflexive curves. Compositio Math. 69, 3-35 (1989).[3] A. HEFEZ and J. E VOLOCH, Frobenius non-classical curves.