Adjustment factors for Hereford and Angus cattle weights

Abstract

The 18-month weights of 4703 Hereford (16 herds) and 10 226 Angus (33 herds) cattle, recorded over 8 years in the Beef Cattle Weight Gain Performance Recording Scheme, were analysed to estimate age and age-of-dam effects, and to evaluate present and alternative weight adjustment procedures. For least-squares analyses, 18-month weights were classified within breed and sex into herd, year-of-birth. and age-of-dam group effects, with age at weighing as a covariate. Separate analyses evaluated the importance of each first-order interaction among herd, year, and age-of-dam group effects. All main effects and the herd × year interaction were highly significant sources of variation (P < 0.001) and were considered to be of practical importance. The mean 18-month weight of Hereford males (females) was 496.9 ± 0.7 kg (356.7 ± 1.3 kg), and of Angus males (females) was 455.4 ± 1.0 kg (318.5 ± 1.4 kg). Corresponding linear regression coefficients of weight on age were 0.55 ± 0.02 kg/day (0.45 ± 0.02 kg/day) for Hereford, and 0.64 ± 0.02 kg/day (0.36 ± 0.01 kg/day) for Angus. Adjustment for age by either additive or multiplicative linear regression coefficients of weight on age eliminated the dependence of weight on age in the data analysed, whereas adjustment by the procedure used in the National Beef Recording Service (NBRS) did not. Age-of-dam effects estimated by least-squares did not generally differ from those estimated by best linear unbiased estimation, with the exception of the weight of cattle reared by 2-year-old Hereford dams. After adjustment of weights on a sample of the data by both NBRS procedures (employing age-of-dam adjustments on 200-day weight), and by factors estimated by least-squares analysis, the rankings of individuals within breed, herd, and sex tended to differ between the adjustment procedures. On the basis of equalising variances within age-of-dam subclasses, it was concluded that additive adjustments for age-of-dam effects were appropriate, particularly those derived by best linear unbiased estimation.