Comparison of first lactations and all lactations of dams to predict sons' milk evaluations.

Multiple regression of genetic evaluations of about 170 Holstein bulls on sire, dam, and maternal grandsire genetic evaluations indicated that the partial regression coefficient for the sire is similar to the theoretical regression coefficient (.45 vs. .50), that the regression coefficient for the dam is much smaller than the theoretical regression coefficient when all records of the dam including her first are included in her evaluation (.12 vs. .35), although about as expected when only first records of the cow and her herdmates are used (.33 vs. .35), and that the regression for the maternal grandsire is larger than expected when his proof and all records of the dam are used in her evaluation (.07 vs. .00), although about as expected when his proof and only her first record is used (-.02 vs. .00). Preferential treatment of potential bull dams is a possible explanation for these results. Genetic correlations among lactations are not likely to be small enough to account for the differences in regression coefficients for the dam's genetic evaluations for all and first lactations. If conditions for selecting bull dams in the future will be similar to what they have been in the past, the conclusion is that genetic evaluations of cows from their first lactation records be used in preference to genetic evaluations from all lactation records for selection of dams of bulls.