Carcass data for 424 castrated male lambs, comprising seven breed-type groups, were used to evaluate a series of linear measurements, subjective scores and the lean content of sample joints as predictors of percentage lean in carcass. The groups represented the main types of British lamb and mean carcass weights ranged from 12-8 kg (Welsh Mountain) to 20-8 kg (Lowland Longwool). Lambs were selected from commercial abattoirs to cover the ranges of fatness and conformation normally found in practice within each group. The pooled within-group S.D. for percentage lean was 3-70 and the overall S.D. (ignoring groups) was 3-96.Predictors were compared in terms of precision and the stability of their prediction equations. Of the characteristics measured on the intact carcass, subjective scoring of external fatness gave the most precise prediction both within groups and overall ( B . S . D .^^ = 2-82, B.s.D. 0Terall = 3-17). Fat thickness over the eye muscle (0) was the best predictor among those taken on the quartered carcass (R.S.D.,^,^ = 2-76, R.s.D. 0T(!ral i = 3-00). The most precise individual predictors were percentage lean in the best end neck and shoulder joints (R.S.D. wlthln = 1-51 and 1-59 respectively). Among pairs of predictors not involving dissection, the combination of C and percentage kidney knob and channel fat in the side gave the most precise prediction (R.S.D.^t,,^ = 2-51, R.s.D. 0TeraU = 2-68). The stability of the prediction equations between groups tended to increase with the precision of the predictors. Among the more precise predictors, the equation for percentage lean in the leg was a notable exception, being particularly variable from one group to another.

Dissection data for 643 steer carcasses of 15 breed-type x feeding system groups were used to examine the distribution of intermuscular fat (IF) and subcutaneous fat (SF) depots between standardized commercial joints. The increase in weight of IF and SF in each joint relative to the corresponding depot weight in the side was examined using the allometric equation. Pooled within-group growth coefficients for both IF and SF were lowest in the distal limb joints increasing inwards to the joints of the rib and loin regions. They were highest for the thin flank joint. Group means for weight of IF and SF in each joint were adjusted to equal total IF and SF weight respectively using the pooled growth coefficients. Statistically significant but relatively unimportant differences were recorded between groups in the adjusted means for all joints tested. The largest differences occurred for the top piece and were 0-39 and 0-40 kg (in one side) for IF and SF respectively. The stability of regression equations between groups for predicting the percentages of IF and SF in the side from the respective percentage in each joint was also examined. Joints which gave the least precise predictions of IF (shin and leg) and SF (shin, leg and clod and sticking) had the most unstable prediction equations. The equations for the other joints differed little between groups.