Sire Evaluation By Second Lactation Records of Daughters

Abstract

A model describing the selection of cows to have a second lactation is developed. This model allows for factors independent of milk yield in first lactation to influence the variable upon which selection is based. Under this model a procedure for estimating the biases in best linear unbiased predictors, which assume no selection, is presented. This procedure was used to obtain estimates of biases for 1109 Holstein sires in artificial insemination in the northeastern United States. Estimated biases for assumed selection models were small relative to changes in ranking of bulls from first to second lactations of daughters. Rank correlation was .77 between rank of bulls based on first lactation and rank for second lactations of daughters where selection is ignored.