Abstract
The seemingly anomalous results obtained by Graham for the temperature dependence of magnetic anisotropy in iron are interpreted on the basis of Zener's theory, with the effect of thermal expansion included.