Mitochondrial efficiency is typically taken to represent an animal's capacity to convert its resources into ATP. However, the term mitochondrial efficiency, as currently used in the literature, can be calculated as either the respiratory control ratio, RCR (ratio of mitochondrial respiration supporting ATP synthesis to that required to offset the proton leak) or as the amount of ATP generated per unit of oxygen consumed, ATP/O ratio. The question of how flexibility in mitochondrial energy properties (i.e., in rates of respiration to support ATP synthesis and offset proton leak, and in the rate of ATP synthesis) affects these indices of mitochondrial efficiency has tended to be overlooked. Furthermore, little is known of whether the RCR and ATP/O ratio vary in parallel, either among individuals or in response to environmental conditions. Using data from brown trout Salmo trutta we show that experimental conditions affect mitochondrial efficiency, but the apparent direction of change depends on the index chosen: a reduction in food availability was associated with an increased RCR (i.e., increased efficiency) but a decreased ATP/O ratio (decreased efficiency) in liver mitochondria. Moreover, there was a negative correlation across individuals held in identical conditions between their RCR and their ATP/O ratio. These results show that the choice of index of mitochondrial efficiency can produce different, even opposing, conclusions about the capacity of the mitochondria to produce ATP. Neither ratio is necessarily a complete measure of efficiency of ATP production in the living animal (RCR because it contains no assessment of ATP production, and ATP/O because it contains no assessment of respiration to offset the proton leak). Consequently, we suggest that a measure of mitochondrial efficiency obtained nearer to conditions where respiration simultaneously offsets the proton leak and produce ATP would be sensitive to changes in both proton leakage and ATP production, and is thus likely to be more representative of the state of the mitochondria in vivo.