It is well known that the surface divergence of the magnetization vector gives rise to a magnetic field, usually termed the "demagnetizing field." In ellipsoidal bodies which are uniformly magnetized, the demagnetizing field is in turn also uniform. In nonellipsoidal bodies, however, the demagnetizing field is in general nonuniform. Experimental data taken on nonellipsoidal samples are, therefore, extremely difficult to interpret. For this reason, experiments on magnetic materials are best performed on samples of ellipsoidal shape. Quite frequently, however, other considerations make it advantageous to use samples that have a nonellipsoidal shape; for instance, rectangular prisms, or cylinders with flat endfaces.
No takes yet. Share an insight, caveat, or question.
R. I. Joseph (1976) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: