Abstract. The statement that Maxwell's electrodynamics in vacuum is already covariant under Lorentz transformations is commonplace in the literature. We analyse the actual meaning of that statement and demonstrate that Maxwell's equations are perfectly fit to be Lorentz-covariant; they become Lorentz-covariant if we construct to be so, by postulating certain transformation properties of field functions. In Aristotelian terms, the covariance is a plain potentiality, but not necessarily entelechy.