Path analysis and robust prediction of lamb carcass composition

Abstract

ABSTRACTPath analysis was applied to correlations among lamb carcass measurements. Correlations were obtained from lamb carcasses averaging 15·7 kg and 270 g fat per kg carcass weight. Correlations were adjusted to represent within breed, sex and management group relationships. The purpose of the analysis was to study the relationships of carcass weight with carcass measurements and to identify carcass measurements that are independent other than through their relationships with carcass chemical fat concentration and carcass weight. The direct effect of live weight on all carcass measurements was positive. Thus, heavier carcasses with the same fat concentration had larger fat and tissue depths, muscle measurements, leg measurements and kidney fat weights. The results of correlations among the residuals suggest that fat depths C and J, specific gravity, muscle width A and kidney fat weight are nearly independent estimates of carcass composition in the sense that they are only correlated through their relationships with carcass weight and carcass fat composition. Robust predictors of carcass composition were developed by first adjusting carcass measurements for the direct effects of carcass weight and then computing regressions on the adjusted measurements. These predictors appeared to be better predictors of environmental and genetic differences than least-squares multiple regression yet reduced the accuracy of within group prediction only slightly.