The melting-points, interatomic distances and compressibilities of the transition elements are reviewed, and are shown not to be in complete agreement with the valency scheme proposed by Pauling. In all three Long Periods on passing from Groups IA to IVA there is a steady rise in melting-point, and decrease in interatomic distance and compressibility, and for these elements Pauling's view of a metallic valency equal to the Group number is correct. In the First Long Period there is comparatively little increase in melting-point on passing from Ti to V and Cr, m in contrast to the behaviour of the corresponding elements in the later Periods where the melting-points rise steeply to a maximum in Group VI. This suggests that in Groups VA and VIA the metallic valency is less for the elements of the First Long Period than in the later Periods, and this agrees with the general inorganic chemistry of the elements. The same principle is shown to hold for the elements of Groups VII A and VIII, and a new scheme of metallic valencies is suggested which appears attractive in several ways.