The article attempts to show the theory of prehistoric Latin accent bound on initial syllable, developed in the first half of the nineteenth century and still generally recognized, to be based on incorrect presumptions.