Circumstances arise when direct determination of fracture toughness, necessary for conducting Engineering Critical Assessments (ECAs), is not possible but Charpy data are available. These situations can arise, for example, when assessments are needed for existing equipment to demonstrate avoidance of fracture or preliminary assessments are required when only specification properties are available. Some of the empirical procedures that may be used to estimate fracture toughness of steels are described. These are based on a recent revision and update of Annex J of BS 7910; the latter provides an integrated method for conducting ECAs. Procedures for estimating fracture toughness from Charpy data representing lower shelf (energies less than 27J), transitional (based on T27J or T40J (Charpy temperatures for 27J or 40J, respectively)) and upper shelf Charpy behaviour are described. In addition, a method is described for estimating T27J when determining fracture toughness from transitional Charpy behaviour where an incomplete transition curve or only single temperature data are available. The thinking behind the procedures is described and examples for their validation (i.e. predictions of fracture toughness compared with actual data) are provided.