"The chief purpose of cohesive energy calculations is to develop analytical methods which give a valid theoretical basis for cohesion, which explain why the ordered alloy has the observed crystal structure, and which give numerical results in agreement with the measured values. The distribution of electrons around the atoms is known only approximately in most crystal lattices, and cohesive energy calculations are very sensitive to the electron distribution assumed. The net cohesive energy depends chiefly upon the difference between two larger quantifies, the boundary energy of a valence electron, which promotes cohesion, and the Fermi energy, which opposes it. A calculation of correct values of cohesive energy indicates that the assumed electron distribution is correct. In the present work, an analysis is made of the nature of binding in the pure metals and the alloy to provide a basis for calculating the energy of binding. The calculations are then made, using the best procedure for each part, and the results are compared with values obtained by experiment. It is found that when account is taken of cohesion from the d-electrons, using the tight-binding approximation, calculated cohesive energies for the pure metals copper and gold are nearer to the experimental values than in previous calculations, which gave too small results. The strong electron binding between unlike atoms also indicates why the alloy has a tetragonal structure while the pure metals are face-centered cubic"--Introduction, leaves 1-2.