Irradiation of crack arrest test specimens and Charpy V-notch specimens under identical conditions shows good correspondence in evaluation of irradiation embrittlement effects. Thus, changes in Charpy V-notch monitoring or surveillance specimens can be used when reactor pressure vessel operating limits are based on crack arrest criteria. Charpy specimens of the same steel irradiated in three different reactors show that, when damage is expressed in terms derived from activation of nickel monitors, measured changes for a given dose can vary by a factor of two. However, when the damage index is taken as the total number of atoms displaced in each reactor facility, the agreement with measured damage improves to better than ±15 per cent. Thus, with a knowledge of neutron spectra, it is possible for a given steel specimen to give a reliable prediction of damage from one reactor to another.