Retarded electromagnetic potentials are derived from Maxwells equations and the Lorenz condition. The difference found between these potentials and the conventional Li{e}nard-Wiechert ones is explained by neglect, for the latter, of the motion-dependence of the effective charge density. The corresponding retarded fields of a point-like charge in arbitary motion are compared with those given by the formulae of Heaviside, Feynman, Jefimenko and other authors. The fields of an accelerated charge given by the Feynman are the same as those derived from the Li{e}nard-Wiechert potentials but not those given by the Jefimenko formulae. A mathematical error concerning partial space and time derivatives in the derivation of the Jefimenko equations is pointed out.