Published online by Cambridge University Press: 02 September 2010
Commercial carcass grading measurements and sample joint dissection were evaluated as predictors of carcass lean concentration using a data set of 1320 pig carcasses from four breeding companies. Entire males, castrated males and gilts from White-type and Meat-type populations were reared on ad libitum or restricted feeding regimens and slaughtered to achieve a carcass weight of either 52·5 kg or 72·5 kg. A model was fitted which examined the accuracy and stability of regression equations for different populations, sexes and feeding regimens, and different values of P2 and carcass weight. The regression slope of lean concentration on P2 differed by sex and population. Lean concentration decreased faster with increases in P2 among gilts than among entire or castrated males. At the same level of P2 and carcass weight, pigs fed ad libitum contained 5 g/kg less lean in the carcass than pigs fed at a restricted level. Biases among populations, sexes and feeding regimens were also found when carcass lean concentration was predicted from additional measurements: muscle depth at P2, fat thickness and muscle depth at 3/4 last ribs, and eye-muscle area. The regression slope of lean concentration on sample joints showed little evidence of variation among populations (except for the hand joint), sexes or feeding regimens. There were important differences in the intercept for different populations. The ham was the most accurate (residual s.d. = 11 g/kg) and least biased joint.