The ordinary Hall coefficients for Cu, Ni, Co, and the binary alloys of these elements agree to within a factor of two with predictions based upon the usual band model in which the number of conduction electrons per atom is postulated to be the number, ns=0.6, of 4s electrons required to explain the saturation magnetization in these materials. The factor of two, however, created a dilemma, since both the magnetization data and the Hall data were accurately determined. The Hall results appeared to require that ns<0.3 whereas the magnetization data requires that ns>0.54. It will be shown that this dilemma disappears when the 4s band is considered divided into two parts in which the electrons with spins parallel to the spontaneous magnetization have much greater mobility than those with spins antiparallel. According to Mott, the antiparallel electrons have low mobility, because they can be scattered into the partially empty 3d band, whereas the parallel electrons cannot. It will be shown that the ordinary Hall data, the saturation magnetization data, and the resistivity data for these transition elements and their alloys, can be made understandable by employing a four-band model, consisting of two 4s bands and two bands from the 3d shell.