Survival to old age in natural populations is enhanced by high vitality and resilience which depends upon substantial homeostasis and energetic amd metabolic efficiency underlain by genes for stress resistance. Under this assumption increased longevity follows from primary selection for stress resistance where stress targets energy carriers. Furthermore old and young fitness should be correlated irrespective of age under the stressful selection regime of natural populations. In contrast, antagonistic pleiotropy is most likely under the less rigorous selection regime of well-nourished humans and laboratory populations surviving to old age. Similarly, hormesis for longevity, for example from a mild temperature stress or restricted food intake is most likely under benign environmental conditions. Assuming that aging in natural populations depends upon ecological circumstances, large evolutionary increases in life span are unlikely under the stress theory of aging since organisms are frequently close to their limits of survival where metabolic efficiency is at a premium. Exceptions can occur in island populations and for mutants under laboratory conditions since the risks from environmental hazards are reduced, and life span becomes extended as a consequence. In modern human populations, selection for stress resistance is less intense than in earlier times which should be permissive of the accumulation of stress-sensitive mutants under the mutation-accumulation theory of aging. However, this process is ultimately likely to restrict the evolution of life-span extensions in the future especially if abiotic conditions deteriorate, when survival would depend more directly on metabolic efficiency under stress.