Evolutionary theories suggest that the expression of extended longevity depends on the organism's ability to shift energy from reproduction to somatic maintenance. New data led us to reexamine our older data and integrate the two into a larger picture of the genetic and metabolic alterations required if the animal is to live long. Our Ra normal-lived control strain can express any one of three different extended longevity phenotypes, only one of which involves significant and proportional increases in both mean and maximum longevity and thus a delayed onset of senescence. This phenotype is dependent on the up-regulation of the antioxidant defense system (ADS) genes and enzymes. Animals that express this phenotype typically have a pattern of altered specific activities in metabolically important enzymes, suggesting they are necessary to support the NAD+/NADP+ reducing system required for the continued high ADS enzyme activities. Fecundity data suggests that the energy required for this higher level of somatic maintenance initially came from a reduced egg production. This was only transient, however, for the females significantly increased their fecundity in later generations while still maintaining their longevity. The energy required for this enhanced fecundity was probably obtained from an increased metabolic efficiency, for the mitochondria of the La long-lived strain are metabolically more efficient and have a lower leakage of reactive oxygen species (ROS) to the cytosol. Selection pressures that do not lead to these shifts in energy allocations result in extended longevity phenotypes characterized by increased early survival or increased late survival but not by a delayed onset of senescence.