Genetic evaluations for milk production traits in dairy cattle based on test day yields (TD) have generated much interest recently due to a number of advantages, including improved accuracy of predicting breeding values and estimation of persistency evaluations. One of the methods proposed for a test day evaluation system is the random regression model (RRM) approach fitting covariance functions such as the Legendre Polynomials. Preliminary genetic evaluations using this approach have been produced by the Animal Data Centre in the United Kingdom based on the research of Brotherstone et al (1999). The objective of this paper is to estimate relative weights on pedigree and TD in evaluations of cows using RRM fitting Legendre Polynomials and compare with 305-day yield evaluations. Secondly to examine the efficiency of using part lactation TD information to predict breeding values for cows.