Production and breeding records from 6,877 lactations of 4,673 cows were used to evaluate the effect of adjusting for previous days open on genetic evaluations. Predicted breeding values for kg of milk were calculated using three animal models. Model 1 - contained no adjustment for days open and accounted for the fixed effects of herd-year-season, genetic group, age of calving (covariate) and the random effects of animal, permanent environment effect and error; Model 2 - same as one with previous days open included; Model 3 - production records were standardized to 100 days open before using model 1. Predicted breeding values from each model were compared. Cows in the 2% top had their ranks changed nine and 11 places for models 2 and 3, respectively, when compared to their rank from model 1. Furthermore, eight and 11 cows dropped from the top 2%, when evaluated with models 2 and 3, both accounted for days open. The reliabilities of the predicted breeding values for the three models were almost equal. The rank correlations between estimated breeding values for cows and sires for all models were near one. Therefore, adjustment for previous days open did not change the rank of animals in this genetic evaluation for Holsteins in the State of Minas Gerais.
adjustment; days open; milk yield; Holstein breed
