Computed Tomography (CT) is a non-invasive method used to provide accurate information on body composition of breeding animals. It is being applied in sheep breeding programmes in the UK to predict whole carcass composition in terms of muscle, fat, and bone. Previous studies established the equations to predict the total weights of muscle, fat and bone in different breeds from conventional CT scans. The average prediction accuracies were 92, 96 and 81% for muscle, fat and bone, respectively, in meat sheep and hill breed lambs (Young et al., 2001). Spiral CT scans provide a means of assessing directly the composition of the whole body and its different regions, in addition to muscularity and conformation. Although an automatic procedure to quantify the carcass components in cross-sectional images was developed previously, new software to segment images from spiral CT scans is now being evaluated. The aim of this study was to determine the accuracy of the automated image analysis in the estimation of total muscle volume of the hind leg, which is the most valuable joint.