COVARIANCE MATRICES FOR GROWTH TRAITS OF AUSTRALIAN POLLED HEREFORD CATTLE

Estimates of covariance components due to direct and maternal effects among birth, weaning, yearling and final weight for Australian Polled Herefords were obtained by restricted maximum likelihood fitting an animal model. Data were obtained from commercial herds participating in the National Beef Recording Scheme and an across-herd genetic evaluation service. Analyses were carried out considering nine subsets of data, each comprising at least 3000 weaning, 1000 yearling and 500 final weight records which yielded a total of 30 506, 17 105 and 9486 records for the three weights, respectively. Estimates of variances and genetic parameters varied considerably between data sets. Both genetic and permanent environmental effects were found to affect all traits except final weights significantly, with maternal environmental effects explaining proportionately more than 0.20 of the total variation in weaning weight. Estimates of the direct-maternal genetic correlation were consistently negative and moderate to strong, ranging from -0.34 to -0.82 for weaning weight. Maternal correlations between weaning and yearling weights were close to unity. Estimates of the direct genetic correlations among the three later weights were 0.8 or larger. Problems with this kind of analysis are discussed.