High peak metabolic rates may provide a performance advantage, but it may also entail a physiological cost. A long-held assumption is that high mass-specific energy expenditure is associated with short lifespan. To examine the relationship between energy expenditure and lifespan we asked two questions. First, do individuals have a consistent rate of metabolism throughout their life? Second, is metabolic rate correlated with lifespan? We analysed the repeatability of measurements of resting (RMR) and peak flight metabolic rate (MR(peak)) throughout the life of the Glanville fritillary butterfly (Melitaea cinxia). Measurements of MR(peak) showed significant repeatability. Senescence occurred only shortly before death. RMR showed a U-shaped relationship with age and very low repeatability. Intraspecific association between metabolic rates and lifespan was tested under three conditions: in the laboratory, under field conditions and in a laboratory experiment with repeated flight treatments. There was a significant correlation between MR(peak) and lifespan in all three experiments, but the correlation was positive, not negative. RMR was not correlated with lifespan. Both MR(peak) and lifespan may reflect physiological condition and therefore be positively correlated. Individuals with a large resource pool may be able to invest in mechanisms that slow down ageing. Individuals with high metabolic capacity may also possess adaptations against ageing. Molecular polymorphism in the gene phosphoglucose isomerase (Pgi) was significantly associated with both MR(peak) and lifespan, and may have coevolved with defence mechanisms against senescence. Generalisations such as 'live fast, die young' may be too simple to explain the complex processes affecting ageing and lifespan.