The de Haas – van Alphen effect in silver has been studied by the modulation method for the field along directions close to [Formula: see text], for which there is a dominant central belly oscillation and a weak noncentral one of nearly equal frequency. The use of pickup coils parallel and perpendicular to the field enabled both the parallel and perpendicular components of dM/dH to be observed. The prediction of Part I that strong magnetic interaction should generate a relatively much stronger difference frequency for the perpendicular than for the parallel component was semiquantitatively verified, although the parallel component of this difference frequency appeared rather more strongly than predicted. This and other discrepancies with the theory, such as the observation of rather too strong beats between the two frequencies in the parallel component, probably occur because of departures from the ideal conditions assumed in the theory (ellipsoidal shape and perfect homogeneity). The results are briefly compared with those of Joseph and Thorsen (1965), who studied the perpendicular component of M by the torque method.