The excitation energies and ionization potentials of the atoms in the first transition series are notoriously difficult to compute accurately. Errors in calculated excitation energies can range from 1 to 4 eV at the Hartree-Fock level, and errors as high as 1.5 eV are encountered for ionization energies. In the current work we present and discuss the results of a systematic study of the first transition series using a spin-restricted Kohn-Sham density-functional method with the gradient-corrected functionals of Becke and Lee, Yang and Pan : Ionization energies are observed to be in good agreement with experiment, with a mean absolute error of approximately 0.15 eV; these results are comparable to the most accurate calculations to date, the quadratic configuration interaction single, double (triple) [QCISD(T)] calculations of Raghavachari and Trucks. Excitation energies are calculated with a mean error of approximately 0.5 eV, compared with similar to 1 eV for the local density approximation and 0.1 eV for QCISD(T). These gradient-corrected functionals appear to offer an attractive compromise between accuracy and computational effort.