Abstract
The central peak observed on the (100) surfaces of Mo and W cannot be thought of as due to a van Hove anomaly of the d band of the surface atomic plane, broadened by interaction with the volume of the metal. It is argued that it might be considered as due to virtual d states on the surface atoms, broadened by interaction with the volume of the metal. A basic difference between closepacked and not so closepacked surfaces in stressed in this connection. Other cases where surface peaks have been computed in the LCAO approximation are also considered