Estimation of the carcass composition of different cattle breeds and crosses from fatness measurements and visual assessments

Abstract

SUMMARYCarcass data for 1053 steers from the Meat and Livestock Commission's beef breed evaluation programme were used to examine the relative precision of alternative fatness assessments for predicting carcass lean percentage. The data were from four trials and comprised both dairy-bred and suckler-bred cattle by a wide range of sire breeds.A visual assessment of carcass subcutaneous fat content to the nearest percentage unit (SFe) was the single most precise predictor both overall (residual S.d. = 2·28) and within breed (residual S.d. = 2·05). Precision was improved by the addition in multiple regression of the percentage perinephric and retroperitoneal fat (KKCF) in carcass, a visual score of the degree of marbling in the m. longissimus and selected fat thickness measurements taken by calipers on cut surfaces (residual S.d. = 2·11 (overall) and 1·90 (within breed)).When the best overall equation was applied to the breed means, there was substantial bias (predicted – actual carcass lean percentage). Biases ranged from +2·5 (purebred Canadian Holstein and Luing) to – 1·3 (Limousin crosses).Breeds differed significantly in carcass lean content when compared at equal levels of fatness measurements. The differences depended both on the precision with which the measurements predicted carcass lean content and the observed differences in carcass composition that existed before adjustments to equal fatness were made.The robustness of prediction equations was examined by applying them to independent sets of data (a total of 334 carcasses) from four other trials involving steers, heifers, cows and young bulls. Equations were stable for cattle of the same breed, sex and similar levels of fatness but important bias was found between more extreme types of cattle.