A multivariate study of pig carcass growth and composition: 2. Dissected tissues and parts

Abstract

SUMMARYAn analysis of pig carcass composition was carried out, using data for 11 genetic source populations in the Meat and Livestock Commission's Commercial Pig Evaluation (CPE). The data were from years 3–5 of CPE and a total of 720 carcasses were involved. Weights of dissected tissues and parts were considered simultaneously, and the effects of growth, source population, sex, feeding regimen and time period were incorporated into an integrated model, the multivariate allometric development model.Canonical variates were constructed for each of the classifying factors except time period, and the residual variation in carcass composition was analysed by principal component analysis. The first canonical variate for source populations was interpreted as a tissue distribution characteristic associated with the distribution of lean between joints in the dorso-cranial and ventro-caudal regions of the carcass. It accounted for 39% of the variation between source populations. The second canonical variate was associated with the ratio of subcutaneous fat to lean in the carcass, and accounted for 20% of the variation between source populations. The results explain the bias found in predicting the proportion of lean tissue in a carcass from subcutaneous fat thickness (P2).In a graphical display of differences between source populations, nine were shown to be distinct, with only two intermediate and not clearly differentiated from some others. The interpretation of canonical variates and the relative positions of the populations were closely related to those obtained from the analysis of production and grading characteristics in the first paper of the series.