By using a newly developed optical technique which enables non-invasive measurement of mitochondrial oxygenation (mitoPO(2)) in the intact heart, we addressed three long-standing oxygenation questions in cardiac physiology: 1) what is mitoPO(2) within the in vivo heart?, 2) is mitoPO(2) heterogeneously distributed?, and 3) how does mitoPO(2) of the isolated Langendorff-perfused heart compare with that in the in vivo working heart? Following calibration and validation studies of the optical technique in isolated cardiomyocytes, mitochondria and intact hearts, we show that in the in vivo condition mean mitoPO(2) was 35+/-5 mm Hg. The mitoPO(2) was highly heterogeneous, with the largest fraction (26%) of mitochondria having a mitoPO(2) between 10 and 20 mm Hg, and 10% between 0 and 10 mm Hg. Hypoxic ventilation (10% oxygen) increased the fraction of mitochondria in the 0-10 mm Hg range to 45%, whereas hyperoxic ventilation (100% oxygen) had no major effect on mitoPO(2). For Langendorff-perfused rat hearts, mean mitoPO(2) was 29+/-5 mm Hg with the largest fraction of mitochondria (30%) having a mitoPO(2) between 0 and 10 mm Hg. Only in the maximally vasodilated condition, did the isolated heart compare with the in vivo heart (11% of mitochondria between 0 and 10 mm Hg). These data indicate 1) that the mean oxygen tension at the level of the mitochondria within the heart in vivo is higher than generally considered, 2) that mitoPO(2) is considerably heterogeneous, and 3) that mitoPO(2) of the classic buffer-perfused Langendorff heart is shifted to lower values as compared to the in vivo heart.