Influence of genetic differences in merit of mates on sire evaluation

Effect of nonrandomness of bulls' mates on daughter milk yield was examined from first lactation records of cows with birth dates in 1965 and later. Results were based on 67,081 Ayrshire, 77,633 Brown Swiss, 241,486 Guernsey, 1,433,761 Holstein, 291,201 Jersey, and 10,465 Milking Shorthorn records. Extent of assortative mating was examined. Correlations between sire Predicted Difference and dam (mate) Cow Index for individual years ranged from -.08 to .20. Correlations for all records within herd-year (.00 to .02) indicated essentially no assortative mating for milk yield for any breed except Ayrshire. For Ayrshires, negative assortative mating was indicated by a correlation of -.07. Within-sire regressions of daughter milk yield deviated from contemporary average (which had been adjusted for average Predicted Difference of contemporaries' sires) on dam Cow Index (merit of mates) by breed were .84 to 1.08. Expected regression was 1.00. Effect of merit of mates on sire evaluation was determined by comparing evaluations from standardized yield with those from standardized yield minus dam Cow Index. Correlations between evaluations for 4233 Ayrshire, 5275 Brown Swiss, 13,742 Guernsey, 32,572 Holstein, 13,688 Jersey, and 1240 Milking Shorthorn bulls rounded to 1.00 except for Milking Shorthorns (.99); average absolute differences in evaluations were 9 to 16 kg, and maximum differences were 49 to 118 kg. Adding an adjustment to Sire Summaries to account for nonrandomness of mates would do little to increase accuracy.