Abstract. It has been suggested that mitochondrial DNA (mtDNA) may play an important role in aging. Yet, few empirical studies have tested this hypothesis, partly because the degree of sequence polymorphism in mtDNA is assumed to be low. However, low sequence variation may not necessarily translate into low phenotypic variation. Here, we report an experiment that tests whether there is within-population variation in cytoplasmic genes for female longevity and senescence. To achieve this, we randomly selected 25 ''mitochondrial founders'' from a single, panmictic population of Drosophila melanogaster and used these founders to generate distinct ''mt'' lines in which we controlled for the nuclear background by successive backcrossing. Potential confounding effects of cytoplasmically transmitted bacteria were eliminated by tetracycline treatment. The mt lines were then assayed for differences in longevity, Gompertz intercept (frailty), and demographic rate of change in mortality with age (rate-of-senescence) in females. We found significant cytoplasmic effects on all three variables. This provides evidence that genetic variation in cytoplasmic genes, presumably mtDNA, contributes to variation in female mortality and aging. Genetic variation in the mitochondrial genome has traditionally been considered selectively neutral. Consequently, sequence divergence in mitochondrial DNA (mtDNA) has been used extensively as a tool to infer evolutionary histories, with the assumption that mtDNA evolves solely by accumulating neutral mutations (see Gemmel et al. 2004;Ballard and Rand 2005). However, this traditional view of the mitochondrial genome is currently changing. In fact, non-neutral evolution of mtDNA has now been implied in several recent studies demonstrating that specific cytoplasmic gene products exhibit reduced or altered function when coexpressed in nuclear genomes derived from different populations (reviewed in Rand 2001;Rand et al. 2004;Ballard and Rand 2005). Thus, mtDNA is likely to play a role in adaptive evolution. However, our knowledge of its influence on phenotypic variation is still very incomplete (for recent reviews see Rand et al. 2004;Ballard and Whitlock 2004;Gemmel et al. 2004;Ballard and Rand 2005).Mitochondrial DNA has also been suggested to play an important role in the evolution of aging (Harman 1956;Tanaka et al. 1998; de Benedicts et al. 2000;Niemi et al. 2005; reviewed in Ballard and Whitlock 2004). Products of mtDNA clearly have a profound role in energy metabolism. Because variation in metabolic rate may be associated with variation in life span (e.g., Sacher 1977), genetic variation in mtDNA could conceivably affect both metabolism and aging (Beckman and Ames 1998; Speakman 2005). Currently, empirical evidence for an association between metabolic rate and aging is limited (Brand 2000;Speakman et al. 2004). Khazaeli et al. (2005) found no relationship between metabolic rate and survival in Drosophila melanogaster lines artificially selected for increased life span. However, the results of ...