An examination of the efficiency of a repeat mating design for the separation of genetic and environmental trends in a selection experiment and its relative efficiency for the generation of response to selection in mice