We demonstrate that an action proposed by A. Khoudeir and N. R. Pantoja in {\sl Phys. Rev.} {\bf D53}, 5974 (1996) for endowing Maxwell theory with manifest electric--magnetic duality symmetry contains, besides the Maxwell field, additional propagating vector degrees of freedom. Hence it cannot be considered as a duality symmetric action for a {\it single} abelian gauge field.

LaTeX file, 3 pages