EXTENDING INCOMPLETE LACTATION RECORDS OF HOLSTEIN COWS WITH VARYING LEVELS OF PRODUCTION

Monthly test-day records from 4,361 Holstein cows were studied to determine if age, season of freshening, level of peak production, and days open affected the precision of predicted factors used in extending incomplete lactation records, and to develop improved regression equations or ratio factors for predicting complete lactation records from Dairy Herd Improvement Association cumulative month records. The separate analyses of monthly means for both milk and fat indicated statistical differences due to the main effects and some interactions between these effects. The large interactions of lactation number by month and peak milk level by month, which indicated differences in the shape of the lactation curves, resulted in the development of separate prediction factors for cows of three different age groups and four different levels of peak milk production. A least-squares (multiple regression) analysis was used to derive extrapolation factors to predict completed records from partial lactation milk and fat records. Multiple correlation coefficients squared (R2) were used to measure the amount of variation accounted for by each variable. Standard errors of estimates were calculated. The regression method accounted for up to 8% more variation than did the ratio estimators, with the larger differences occurring early in lactation. © 1969, American Dairy Science Association. All rights reserved.