Mitochondria are essential to eukaryotes, producing most of the cellular energy requirements. Curiously, they have retained their own diminutive genome – of mitochondrial DNA (mtDNA); which is maternally inherited, and encodes a handful of proteins that comprise subunits of the mitochondrial electron transport chain. Biologists long reasoned that natural selection would prevent any phenotype-modifying genetic variation from accumulating within the mitochondrial genome, given that the products of the mtDNA are so critical to life. I will present research that challenges this traditional paradigm – by showing that mtDNA sequences typically harbour genetic polymorphisms that affect metabolic and reproductive health, and life expectancy. I will present experimental evidence from natural populations of fruit fly (Drosophila melanogaster) that some of the polymorphisms that delineate the mtDNA haplotypes of different geographic regions are adaptive and have evolved under natural selection to the local climate. I will show that the effects of mtDNA haplotypes on health are often contingent on the nuclear background. That is, the particular combination of mtDNA and nuclear genes an individual harbours (the joint “mito-nuclear” genotype) is an important predictor of that individual’s reproductive prospects and life expectancy. I will then outline the evolutionary consequences of maternal inheritance of mtDNA for male health. Maternal inheritance means that male-harming mutations can accumulate in the mtDNA sequence. If these same mutations are relatively benign, or even beneficial, in their effects on females, then in theory natural selection will fail to eliminate them (since all of the screening of mtDNA mutations is done directly through females). Thus male-harming, but female-friendly, mutations are predicted to accumulate within the mitochondrial genomes of animals, and affect male health components. I will present our research in fruit flies, which substantiates this evolutionary theory.