The main motivations for mapping quantitative trait loci (QTL) in dairy cattle are based not only on the biological interest to understand the complex genetic architecture of trait variation but also on applying genomic information to practical breeding schemes in order to enhance selection programs. The use of molecular genetic information in marker assisted selection (MAS) would be most effective if the genetic architecture of quantitative traits of interest is understood completely. This level of understanding seems to be remote. With the current mapping approaches it is possible to detect QTL that have major effects on trait variation but the true challenge is the identification of gene(s) and nucleotide variant(s) underlying the QTL effect. In this study QTL affecting milk production traits (milk yield, protein yield, fat yield, protein content and fat content) were mapped in Finnish Ayrshire dairy cattle. A whole genome scan was conducted by 12 half-sib families in a graddaughter design. A male genetic linkage map covering all 29 autosomes was constructed by genotyping 150 markers in these families. The map was utilised in interval mapping using a multiple marker regression approach. The empirical significance thresholds were estimated by permutation. The genome scan based on single chromosome analysis uncovered two significant QTL at 5% genome-wise significance, and 14 suggestive QTL at 5% chromosome-wise significance. This approach was extended to analyse multiple chromosomes simultaneously using the 14 suggestive QTL as cofactors in order to increase the power and the precision of QTL detection. The analysis revealed in total 31 genome-wise significant QTL (Pgenome < 0.0029). One of the highest test statistics observed in the initial genome scan was detected on chromosome 20. Two candidate genes that have important roles in mammary gland physiology were mapped to the region of interest the genes encoding receptor molecules of growth hormone (GHR) and prolactin (PRLR). The potential roles of GHR and PRLR in milk production was studied in two populations, in Dutch Holstein-Friesian and in Finnish Ayrshire. Sequence analysis revealed four missense mutations in GHR (F279Y, N528T, A541S, S555G) and two in PRLR (S18N, L186P). In both breeds the GHR F279Y polymorphism was clearly associated with milk yield and composition. It was considered unlikely, however, that F279Y accounts for the entire chromosome 20 QTL effect. In Finnish Ayrshire the QTL effect was partly explained by another polymorphism, PRLR S18N. The results provide strong evidence that the effect of PRLR S18N is distinct from the GHR F279Y effect. In particular, F279Y has the highest influence on content traits while S18N influences yield traits. This, and the observed interaction between the two loci, supports the hypothesis that both GHR and PRLR contribute to the QTL effects of chromosome 20. A method for trait associated gene diagnosis from bovine embryos was also developed in order to apply the gene mapping results at a more practical level. GHR F279Y and PRLR S18N together with a selection marker for sex were genotyped from in vitro and in vivo embryos. The method proved to be highly accurate but needs to be adjusted for substantially higher number of markers in order to respond to the needs of modern breeding applications.