Abstract
Carcasses from IS steers averaging 454 kg. live shrunk weight and selected to give a wide range in carcass composition, were used to evaluate the accuracy with which various prediction equations could estimate actual carcass composition. The right side of each carcass was physically separated into a bone and a fat-lean portion from which chemical analyses of samples representing the carcass were made. Simple correlation coefficients between the equations and chemically determined carcass composition showed that the Hankins and Howe method most accurately estimated carcass fat (r=0.94) and carcass protein (r=− .96); however, it was the least practical method. Yield grade was slightly less accurate, followed closely by the Oklahoma method, carcass specific gravity, and the Tennessee, Wisconsin and Illinois methods. All correlation coefficients between these equations and carcass composition were highly significant (P<.01). When these equations were correlated with the chemical analysis of the fat-lean portion of the five trimmed wholesale cuts, no significant changes in the order of accuracy were observed, although the correlation coefficients were decreased. Yield grade and the Tennessee method were the most practical since both could be readily obtained without any additional cutting or handling of the carcass. Considering the prediction accuracy and the practical aspects of each method, results from this study indicate that yield grade, calculated to the nearest 0.05, was the best evaluation method. Copyright © 1968. American Society of Animal Science. Copyright 1968 by American Society of Animal Science