We obtained new data to determine whether the spectral appearance of A-type stars is entirely determined by their rotational velocities. For this purpose we derived rotational velocities for 1700 northern A-type stars from CCD coude spectra, calibrated with the new Slettebak et al. system, and new MK classifications based on wide photographic Cassegrain spectra for 2000 northern and some southern stars in the Bright Star Catalogue. In addition we determined the equivalent widths of the lambda 4481 Mg II lines in the coude spectra. Tables and graphs show the variations of rotational velocities and lambda 4481 line strengths as functions of type and luminosity, and frequencies of the normal and abnormal stars. After deconvolutions of the rotational velocities, assuming random orientations of rotational axes, we find that all rapid rotators have normal spectra and nearly all slow rotators have abnormal spectra (Ap or Am). Those abnormalities are generally attributed to diffusion and can occur only with little rotational mixing. However at all types there are overlaps of these distributions, implying that a given intermediate rotational velocity is insufficient to determine whether the star should have a normal or abnormal spectrum. However, we realized that (1) some of our ''standards,'' such as Vega and ct Dra, are really abnormal, causing us to classify similar peculiar stars as ''normal,'' (2) many of the ''normal'' stars near A2 IV have the characteristics of peculiar stars such as low rotational velocities and weak 4481 Mg II and K lines, and (3) the mean rotational velocities of ''normal'' stars are depressed just at those types where the Ap and Am stars are most frequent. Therefore we conclude that the overlaps are due to our failure to detect all the abnormal stars and that a specific rotational velocity is probably enough to determine whether a star will have a normal or abnormal spectrum.
