COMPARISON OF TECHNIQUES FOR EVALUATING LEAN CONTENT OF HOG CARCASSES
A population of 2670 pig carcasses was used for comparing three techniques, ruler, ultrasonic and the Danish MFA probe, for prediction of lean yield. The best single fat measurements, all of equal validity, were MFA and ruler at the midline over the gluteus medius and ultrasonic fat measured lateral to the midline in the lumbar region. For each technique, the prediction precision provided by the best single fat measurement was improved by the addition of other fat or lean measurements to the regression equation. The MFA, used to its design capacity of three fat and one lean measurement, predicted 64.6% of the variance in lean content of the lean cuts (ham, loin and shoulder). This reduced to 59.5% when the lean measurement was excluded. The variance explained by-three fat measurements made by the other techniques was 57.4 for ultrasonics and 56.5 for the ruler. Addition of carcass weight to the prediction equations for boneless and boneless defatted yield increased the R2 values for the ruler (4.0%) and ultrasonic (6.4%) techniques but had little effect on the MFA. The most accurate prediction of yield was provided by measurements of side fat and depth of M. longissimus at the last rib, made from carcass cross sections for a subpopulation of 1267 carcasses. This was interpreted as evidence that more specific definition of the site chosen for the MFA fat and muscle measurement in this region of the carcass would improve the precision of the MFA procedure.