The aims of this study were: 1) to evaluate measurements for predicting the carcass lean content of live bulls, 2) to assess the importance of different pre -test environmental effects on bull performance, and 3) to compare biological and economic selection indices for use in a terminal sire beef breed. The data comprised live weight, food intake, ultrasonic and carcass measurements on a total of 235 Hereford bulls, performance tested to 400 days of age on ad libitum feeding. Multiple regression equations using live weight and ultrasonic fat area measurements gave the best prediction of carcass leanness. However, the precision achieved varied depending on the machine, the operator and the group of bulls (R2 values 0.61 to 0.77) . Artificially reared bulls had low pre -test growth rate, which led to compensatory growth, and increased the variation in performance on test. Bulls weaned at 84 days of age were least affected by environmental factors such as dam age and year -season of birth, and performed as well as bulls weaned at 168 days of age. There were high phenotypic correlations between growth rate and lean growth rate (0.96) and between food conversion efficiency and lean food conversion efficiency (0.97) . Formulae were therefore derived for predicting the phenotypic and genetic relationship between a product trait, such as lean growth rate, and one component trait. Selection indices were derived which may be suitable for terminal sire breeds in the UK. The indices were insensitive to moderate changes in economic weights and genetic parameters, and were proposed as being superior to the biological indices (product traits) for improving the efficiency of lean meat production.