The evidence to date indicates that selection for increased feed efficiency (gross energetic efficiency) for the production of milk would be effective. But, because of the high genetic correlation between milk production and efficiency and because the ratio of the heritabilities of these traits appears to be approximately equal, selection for milk production alone will automatically select for increased feed efficiency. This indirect selection is expected to be between 70 and 95% as effective as would be direct selection for efficiency if the selection intensity is equal for the two traits. Since, in practice, indirect selection could be more intense, indirect selection may actually achieve more progress in increasing feed efficiency. Genotype by environmental interactions for sources and levels of energy intake have not been important enough, in the limited number of experiments conducted, to be an important source of error in making selections among cattle for increased production of milk or fat. The same conclusion can be made for the selection of sires when the progeny are tested in many herds. The latter is much more clearly documented. A great deal more research is needed to clarify the genetic relations among size of cow, forage intake and total nutrient intake.