Cow live weight is of economic importance in dairy production; however, it is not included in the selection objective for South African dairy cattle. Variance components estimates are a prerequisite to incorporating a trait in the breeding objective. Variance components were estimated for live weight of lactating Holstein cows on two South African dairy herds. Live weight records on 9843 lactating cows, collected over a period of three years, were used. An analysis of variance was carried out to determine fixed effects to include in the model by the least squares method, using the Generalised Linear Models procedure of the Statistical Analysis System. Variance components were estimated by the Restricted Maximum Likelihood procedure using the ASREML programme. A high heritability estimate of 0.74 ± 0.19 is found, which suggests that there is scope for significant response to selection on live weight in the South African Holstein cattle population. A repeatability estimate of 0.86 was obtained. These results form the basis for incorporating live weight in the breeding objective for South African Holstein cattle. Further work, however, needs to be done to develop estimates based on a representative sample of the whole population, as the current study is based on data from only two herds.

Genetic evaluation for the South African dairy industry has kept pace with global advances in statistical methodology. Increasingly accurate estimated breeding values (EBVs), produced routinely in the past two or three decades, have aided selection decisions. This has been coupled with an increase in the number of traits officially recorded and for which EBVs are calculated. Currently, EBVs are routinely published for more than 20 traits for the major dairy breeds. The current study was conducted to assess the genetic trends realized for traits of economic importance in the South African Holstein population, for the period from 1983 to 2008. Performance and pedigree data of 1 231 930 animals were used to calculate EBVs for these traits by a multi-trait animal model. The resulting EBVs in turn were used to compute annual mean rates of genetic change. Genetic trends for yield decreased by approximately 57% during the decade from 1990 to 2000 and reached stasis in 2005 -2007. Calving interval and somatic cell count also deteriorated over much or all of the period investigated. Given the widespread availability of genetic evaluations for these traits and the noted potential for selection to implement favourable genetic trends, development of strategies to improve the South African Holstein appears to be urgently needed.