Abstract
The experimental band structure of Ni obtained by Eastman et al., using angle-resolved photoemission, differs markedly in several respects from the self-consistently calculated bands of Wang and Callaway. We demonstrate that two major discrepancies may be accounted for by correlation effects: (i) band narrowing for which several models, none of them entirely satisfactory, have been proposed and (ii) a reduction in the exchange splitting well in excess of that obtained using the von Barth-Hedin potential for which we present a simple calculation.