153,963 test day milk yield records and 13,273 first lactations of Holstein cows calving between 1989 and 1998, were used with the objective of estimating genetic, phenotypic and environmental parameters for test day milk yield (PLDC) and 305 day milk yield (P305) and to study the convenience of using test day yields in genetic evaluations to replace P305. Four models were used. Models 1 and 2 differed according contemporary grouping and monthly milk records were considered as repeated measures. In model 1 (PLDCM01) records were grouped by herd-year-season of test day yield and in model 2 (PLDCM02) by herd-year-season of calving. In a third (model 3), monthly yield records were analyzed as individual traits (C01 to C10); and the fourth (model 4) was the traditional 305-day model. Restricted Maximum Likelihood methodology was used with the MTDFREML system. The estimates of heritability for PLDC, using model 1, model 2 and for P305 were 0.27, 0.15 and 0.25, respectively. Heritabilities for monthly milk records ranged from 0.11+0.02 (C01) to 0.21+0.03 (C08), with the largest values occurring beginning in the fourth month. Genetic correlation estimates between monthly records and P305 ranged from 0.76 to 1.00, with the highest correlations occurring in the middle lactation. It was concluded that using test day milk yield is promising and that selecting for middle lactation records could have advantages over 305-day milk yield.
animal model; Holstein; milk yield; REML methodology; test day yields
