This paper describes a new method of measuring some magnetic properties of steel under applied biaxial stresses in the -plane when the steel is magnetized in the direction (called " magnetization"). Results are compared with two other sets of experimental data. The conclusions are that magnetization is prone to several sources of error of significant amounts, whichever technique is used, and that these errors must be reduced if the small changes in magnetic properties predicted by the theory are to be reliably verified.