IMPACT OF DIFFERENT STRATEGIES AND AMOUNTS OF PREFERENTIAL TREATMENT ON VARIOUS METHODS OF BULL-DAM SELECTION

Three records of milk yield, fat yield, and type were simulated far ea ch cow in 20 herds of 200 cows over 13 yr. Preferential treatment or b ias was simulated by increasing milk and fat yields by an average of 0 , 16, and 32% for separate copies of the simulation. The bias was give n to a limited number of cows from the original herds based on four st rategies. Five methods of bull-dam selection that used an index with a 2:2:1 ratio of milk to fat to type to select the top 2% of cows were compared: ETA using first lactation, using all lactations, after pheno typic minima were required, after preselection on three-generation ped igree index, and on pedigree index alone. Selection on ETA for first o r all lactations gave the highest average of true breeding values at 0 and 16% for all strategies studied. In general, selection on pedigree index alone or after phenotypic minima were required gave poor result s and should not be considered to be viable. Preselection of bull-dams on pedigree index proved to be extremely useful for biased and unbias ed data. The optimal policy was to preselect the top 12% of the popula tion before reranking and selecting on ETA for all lactations.