The detection of 12CO2 emission from leaves in air containing 13CO2 allows simple and fast determination of the CO2 emitted by different sources, which are separated on the basis of their labelling velocity. This technique was exploited to investigate the controversial effect of CO2 concentration on mitochondrial respiration. The 12CO2 emission was measured in illuminated and darkened leaves of one C4 plant and three C3 plants maintained at low (30-50 ppm), atmospheric (350-400 ppm) and elevated (700-800 ppm) CO2 concentration. In C3 leaves, the 12CO2 emission in the light (Rd) was low at ambient CO2 and was further quenched in elevated CO2, when it was often only 20-30% of the 12CO2 emission in the dark, interpreted as the mitochondrial respiration in the dark (Rn). Rn was also reduced in elevated CO2. At low CO2, Rd was often 70-80% of Rn, and a burst of 12CO2 was observed on darkening leaves of Mentha sativa and Phragmites australis after exposure for 4 min to 13CO2 in the light. The burst was partially removed at low oxygen and was never observed in C4 leaves, suggesting that it may be caused by incomplete labelling of the photorespiratory pool at low CO2. This pool may be low in sclerophyllous leaves, as in Quercus ilex where no burst was observed. Rd was inversely associated with photosynthesis, suggesting that the Rd/Rn ratio reflects the refixation of respiratory CO2 by photosynthesizing leaves rather than the inhibition of mitochondrial respiration in the light, and that CO2 produced by mitochondrial respiration in the light is mostly emitted at low CO2, and mostly refixed at elevated CO2. In the leaves of the C4 species Zea mays, the 12CO2 emission in the light also remained low at low CO2, suggesting efficient CO2 refixation associated with sustained photosynthesis in non-photorespiratory conditions. However, Rn was inhibited in CO2-free air, and the velocity of 12CO2 emission after darkening was inversely associated with the CO2 concentration. The emission may be modulated by the presence of post-illumination CO2 uptake deriving from temporary imbalance between C3 and C4 metabolism. These experiments suggest that this uptake lasts longer at low CO2 and that the imbalance is persistent once it has been generated by exposure to low CO2.