A total of 281 lamb carcasses covering three weight groups (18–22.9, 23–25.9 and 26–30 kg) and fatness (< 3 mm, 3–5 mm and > 5 mm) were used to determine the usefulness of several carcass measurements for predicting saleable meat yield. Carcasses were measured for fat and muscle depth, 3–4 cm from the mid-line between the 10th and 11th ribs, 12th and 13th ribs and immediately adjacent to the 13th rib using a Hennessey Grading Probe. This technique also determined total tissue depth at 11 cm from the mid-line between the 10th and 11th and 12th and 13th ribs. The depth of tissues over the 12th rib 11 cm from the mid-line was measured with a ruler. Carcasses were also visually assessed for muscle development (conformation score) on a 5-point scale (1 = thinly fleshed; 5 = thickly fleshed). All carcasses were fabricated into primal cuts which were trimmed to 5 mm of fat and deboned to provide an estimate of saleable meat yield. Ewe lamb carcasses had a higher proportion of kidney fat than wethers in all weight and fat groups (which ranged from 2.6 to 17.9 g kg−1 carcass weight). Saleable meat yield decreased as carcasses became fatter, whereas weight group had little influence on saleable meat yield within a fat group. Ewe lamb carcasses had lower saleable meat yields than wether lamb carcasses, mainly because of their higher amounts of kidney fat. A ruler measurement of tissue depth over the 12th rib combined with carcass conformation score provided the most precise prediction (R2 = 0 61; RSD 17.1) of saleable meat yield. It was concluded that a simple manual system based on these two measurements provided an adequate prediction of saleable meat yield. The accuracy of the procedure would be increased by excluding kidney fat from carcass weight for the prediction of saleable meat yield. Key words: Lamb, carcass, composition, probe, fat, grade