Can more be learned from selection experiments of value in animal breeding programmes? Or is it time for an obituary?

Selection experiments in laboratory animals and livestock have provided a wealth of information on genetic parameters of quantitative traits and on the effectiveness of selection in the short and long term on both directly selected and correlated traits. They have stimulated developments in theory and tests of it, and extreme selected lines continue to be source material for biological study. Some of the main questions and findings are briefly reviewed. Yet much of successful animal breeding practice has been based essentially on statistical methods, assuming where necessary the infinitesimal model, and new developments such as genomic selection are similarly not based on selection experiments. Information on the genetic architecture of quantitative traits is provided by selection experiments, but new methods for deeper studies of the biology are available. I discuss the future role for selection experiments in view of changes in funding streams and technology and conclude that there is little case for starting new experiments, but retention of existing long-term lines is desirable and DNA should be collected from all lines on a continuing basis.