Analysis of test day yield data of Costa Rican dairy cattle.

Estimates of variance components for test day records in an animal model that considered multiple traits over multiple lactations were calculated using REML methodology. Test day records were classified into 11 periods within first and later lactations. Missing ancestors in the relationship matrix were classified in genetic groups. Data were collected from Costa Rican dairy farms. Estimates of components for total and additive genetic variance were clearly heterogeneous during the lactation. Heritabilities for traits in later parities were slightly higher than those for traits in first parity. Heritabilities were highest for records of midlactation. Phenotypic and genetic correlations for adjacent test days were close to 1. Phenotypic correlations were lower than genetic correlations. Heterogeneity of variances during the lactation suggests the adequacy of the multiple-trait test day model to describe milk yield during the lactation. When missing ancestors were allocated to a single base population instead of genetic groups, the estimates of residual variance were lower, and the estimates of genetic variance and genetic correlations were higher. When standardized records were used instead of actual test day records, the estimates of residual and total variance were lower, and the estimates of genetic variance were higher. Consequently, estimates of heritability and genetic correlations were also higher. Use of standardized data obtained by interpolation procedures is not advised for estimation of genetic variance components in a test day model.