J. Dairy Sci. 91:1648–1651 doi:10.3168/jds.2007-0622 © American Dairy Science Association, 2008. Short Communication: Variance Estimates Among Herds Stratified

The objectives of this study were to compare (co)vari-ance parameter estimates among subsets of data that were pooled from herds with high, medium, or low indi-vidual herd heritability estimates and to compare indi-vidual herd heritability estimates to REML heritability estimates for pooled data sets. A regression model was applied to milk yield, fat yield, protein yield, and so-matic cell score (SCS) records from 20,902 herds to generate individual-herd heritability estimates. Herds representing the 5th percentile or less (P5), 47th through the 53rd percentile (P50), and the 95th percen-tile or higher (P95) for herd heritability were randomly selected. Yield or SCS from the selected herds were pooled for each percentile group and treated as separate traits. Records from P5, P50, and P95 were then ana-lyzed with a 3-trait animal model. Heritability esti-mates were 23, 31, 26, and 8 % higher in P95 than in P5 for milk yield, fat yield, protein yield, and SCS, respectively. The regression techniques successfully stratified individual herds by heritability, and additive genetic variance increased progressively, whereas per-manent environmental variance decreased progres-sively as herd heritability increased. Key words: herd heritability, daughter-dam regres-sion, daughter-sire regression Dechow and Norman (2007) estimated individual herd heritabilities for milk yield, fat yield, protein yield, and SCS for 1,939 herds using regression techniques. Correlations of regression heritabilities with individual herd heritability estimates generated with REML ranged from 0.45 to 0.68 for a subset of the 45 largest herds. However, attempts to adjust records to a con-stant genetic variance using the individual herd herita-bility estimates did not appreciably improve the accu-racy of genetic evaluations. Lack of improvementswhe