AbstractSteels used in the construction of modern power plant contain relatively large concentrations of alloying elements. The martensite start temperature of such steels is important during fabrication, and can not be estimated using conventional empirical methods. It is demonstrated in the present paper that a recent theory due to Ghosh and Olson can be applied to predict both published and new experimental data on the martensite transformation temperatures of power plant steels. However, the method should not be used in circumstances where the steels containδferrite or undissolved carbides before transformation.MST/3252