Dominance models with method R for stature of Holsteins

Estimates of variance components were obtained with method R for several additive and dominance models. The data included 301,960 records for first parity and 280,040 records for later parities of Holsteins. The single-record model included effects of management, regression on inbreeding percentage, age at calving, stage of lactation, and additive and dominance effects. The repeatability model included these effects in addition to permanent environment. For the single-record model, estimates were 46% of the total variance for additive variance, 12% of total variance for dominance variance, and -0.06 for the regression coefficient on inbreeding. In the repeatability model, the variance for permanent environment was estimated at 5%; other estimates were similar. When the dominance effect was eliminated, the estimate of the variance for permanent environment increased to 17% for the repeatability model. Elimination of stage of lactation increased regression on inbreeding to 0.09 and the estimate of dominance variance to 17% in the single-record model. The same change increased the estimate of additive variance to 64% for the repeatability model. Elimination of regression on inbreeding or stage of lactation had a small effect on the estimates. The presence or absence of the dominance effect had little influence on additive variance. In the absence of dominance, the estimate of the permanent environment effect included the dominance effect. Estimates of variances with method R are very sensitive to age adjustments. With the adjustments, the estimates of the dominance and additive variances are consistent.