The mitochondrial genome has been proposed as a principal site of somatic mutation during ageing. A variation of the error catastrophe model has been proposed, in which ROS damages the mitochondrial genome, which leads to additional ROS production in a positive feed back cycle. This leads to major DNA damage, bioenergy crisis, and reduced functional capacity in old age and contributes to mortality. Therefore it might be expected that in strains in which the mitochondrial genomes vary, ROS and bioenergy crisis should covary and negatively correlate with longevity. Strains of Drosophila were produced which differed in their mitochondria by breeding maternally inherited genomes onto a common nuclear background. The donor strains included two long lived and two control strains. Those strains that had the cytoplasmic genomes from the long-lived strains were also long lived. In these strains ROS production in young flies negatively correlated with longevity supporting a role for ROS in ageing and/or the death process. Ageing Drosophila show a failure in bioenergy, but the relative strength of this phenotype does not segregate with longevity. These data do not support the error catastrophe model, but suggests that the principal outcome of ROS damage that leads to death is not bioenergy failure, and that bioenergy failure is at least partly due to non-ROS processes.