Early attempts at finding markers for quantitative trait loci (QTL) in the 1950s and 1960s involved searching for associations between production traits and polymorphisms at loci encoding blood groups, milk proteins and blood proteins. Overall, this work identified many small and/or non-significant associations, insufficient to warrant their use in marker-assisted selection. It was not until the discovery of microsatellites in the early 1990s that a really useful form of DNA marker became available. By the mid-1990s, linkage maps comprising mainly microsatellites covered most regions of most chromosomes of the cow, and it was then possible to hunt for QTL using a mapping approach first proposed in Drosophilia back in 1961. Fine-mapping of QTL has eventually resulted in the identification of markers in or near particular genes. Some of these markers have been commercialised. In most cases, the actual causative quantitative trait nucleotide (QTN) remains elusive. The recent development of technologies for large-scale detection and genotyping for single nucleotide polymorphisms (SNPs) and for placing the entire genome on a chip have opened up exciting possibilities for the future: geneticists should begin contemplating how best to use SNP and genome chips, which will surely become a reality. Just as one must estimate genetic and phenotypic correlations for every new quantitative trait that is introduced into a genetic evaluation scheme, so must the correlated effects of new DNA markers be evaluated before their commercial use. Researchers and their funders must resist the temptation to cut corners in getting markers to market.