It has been proposed that prior to the evolution of sex, the endosymbiotic relationship between mitochondria and nuclear genomes would have selected mechanisms that maintained the optimum interaction between the two genomes. Once sex evolved, mating would introduce different, competitive, mtDNA and/or nDNA gene products that could well upset the balance. Mechanisms, such as the specific degradation of one mitochondrial genome that is known to occur, could have been selected to prevent part of such competition. Unlike most protein complexes in the cell, the proteins of the multienzyme complexes of the ox-phos system are derived from both nuclear-genome-coded genes and mitochondrial-genome-coded genes. Minor mutations in either mtDNA or nDNA coding for these proteins are known to lead to major and catastrophic diseases of humans, suggesting that very tight and precise interactions are required. To maintain the evolutionarily established balance after mating, monoallelic expression of the nuclear-coded genes would be advantageous and prevent subtly different competitive proteins from interacting with the resident mitochondria. This would require regulation of the expression of those specific nuclear genes, possibly under the control of the resident mitochondria. It is possible that aging cells could lose the requisite tight regulation and allow expression of proteins derived from the formerly repressed nuclear alleles that would compete for mitochondrial complex sites. With age, random failure of this control could lead to increasingly inefficient mitochondria in different tissues and organs and eventually to senescence and death.