Published online by Cambridge University Press: 02 September 2010
Records on fat plus protein yield, food intake and food efficiency up to 26 and 38 weeks of lactation, were obtained from dairy cows fed ad libitum in the Edinburgh School of Agriculture's Langhill herd. After discarding the first 2 weeks data, the records were divided into successive 6-week periods. Restricted maximum likelihood analyses were carried out on each period for each trait, fitting an animal model, with repeat lactations as an additional random effect. Univariate analyses were done after approximate canonical transformation of the data. Heritability estimates for 6-week fat plus protein yield, food intake and efficiency, from the 38-week lactation data, ranged from 0·08 (s.e. 0·10) to 0·17 (s.e. 0·12), 0·15 (s.e. 0·09) to 0·49 (s.e. 0·15) and 0·08 (s.e. 0·14) to 0·25 (s.e. 0·11), respectively. Generally, lower heritabilities were obtained for earlier part measures of fat plus protein yield and efficiency and for later part measures of food intake. Genetic correlations between part measures and ‘total’ lactation fat plus protein yield, food intake, and efficiency ranged from 0·82 to 100, 0·85 to 0·98, and 0·60 to 1·00, respectively. Corresponding phenotypic correlations ranged from 0·76 to 0·94, 0·68 to 0·87, and 0·73 to 0·89, respectively. The correlations were lowest for the first 6-week period then gradually increased up to a maximum in the third or fourth periods and declined thereafter. The results indicate that selection on a part measure of food intake or efficiency, taken between 15 to 20 weeks of lactation, will be equally effective in improving intake and efficiency compared with selection on total lactation measures. Other 6-week measures taken between weeks 9 to 14 and 21 to 26 will also give good prediction of total lactation intake and efficiency.