66+ MAGNETISM it seems of some significance that iron nickel alloys with more than 81% Ni decrease in length on magnetization (as does nickel itself), while those with 70% increase. The high initial permea bility of permalloy is associated with a small or vanishing mag netostriction. Changes in volume accompanying magnetization in many ferromagnetics have also been measured.
It was found by E. Villari (1868) that the magnetization of iron was increased by stretching in small fields, but decreased in large fields, the intensity of magnetization at the Villari reversal point decreasing with increasing tensions. This is a particular example of the effect of stresses on magnetization which has been examined by J. A. Ewing, C. G. Knott and many others. General dynamical considerations, as shown by J. J. Thomson, suggest that there should be a number of reciprocal relations, which have to a large extent been verified. For iron, magnetization produces increase in length in weak fields, and decrease in strong fields. Corre spondingly, tension produces increase in magnetization in weak and decrease in strong fields. In nickel, which contracts in all fields, magnetization is decreased by tension. An effect discovered by G. Wiedemann—that a straight wire magnetized longitudinally and also circularly (by the passage of an electric current) becomes twisted--may be shown to be a consequence of two superposed magnetostriction effects.
In view of the Weiss theory of spontaneous magnetization, the investigation of magneto-mechanical hysteresis effects accompany ing cyclic changes is of some interest, but the interpretation of these and other results is naturally a matter of difficulty, partly owing to the very ill-defined and varying character of the micro crystalline structure of ordinary ferromagnetic materials. Experi ments on single crystals should lead to results whose interpretation is more obvious.