Oxygen uptake from the microcirculation is a direct measure of tissue function. Magnetic resonance is capable of detecting differences between oxygenated and deoxygenated blood due to the paramagnetic properties of deoxyhemoglobin. At the level of the microcirculation, however, imaging methods cannot directly visualize the vessels. Instead, bulk MR parameters are investigated for their ability to monitor blood oxygen saturation (%O 2 ) changes in the microcirculation of tissue, specifically skeletal muscle. The ability to monitor the blood oxygenation level within the microcirculation could provide valuable insight into the question of tissue viability and the assessment of tissue function. Measurements of oxygen consumption have been shown to be correlated with organ function (1-5). A noninvasive measurement of blood oxygen state (%O 2 ) would be useful clinically in a range of disease states (6). We are interested in a bulk MR tissue parameter that is correlated to blood oxygen state and is sensitive enough to show changes in tissues where the blood volume fraction is small.Deoxyhemoglobin in blood acts as a paramagnetic contrast agent. This suggests that T 2 -based parameters are likely to be the most sensitive to changes in tissue blood oxygenation. Indeed, functional MR studies have shown that the BOLD contrast mechanism (7) can be sensitive to small changes in blood volume and oxygenation state. This contrast is T* 2 based, which is composed of T 2 Ј (reversible) and T 2 (irreversible) contributions, both of which are expected to change with blood oxygenation, following slightly different mechanisms. While this approach has proven effective in the brain, results in skeletal muscle have been mixed (8,9). Other investigators have studied the relaxation behavior of in vitro skeletal muscle using T 2 relaxation times. Results in various muscle types with different capillary densities indicated significant (10) or no (11) differences with monoexponential T 2 fits. Fits for multiexponential decay found two (11,12) or three (10,13,14) T 2 components in normal muscle. When three T 2 components were fit, the shortest T 2 component was approximately 5 msec. It was less than 5% of the total signal and was assigned to macromolecular-associated water. In vivo measurements would likely not be able to distinguish such a component since typical echo times are greater than the T 2 of this short component. Hazlewood et al. (13) hypothesized that the two longer T 2 components were associated with intra-and extracellular water. Le Rumeur et al. (11) demonstrated inconsistencies in Hazlewood's assignments, and instead related the long-T 2 component to the vascular space alone, and the short-T 2 component to the sum of interstitial and intracellular spaces. If Le Rumeur's assignments of intra-and extravascular spaces were correct, then changes in the T 2 of blood resulting from changes in its oxygen state should preferrentially affect the long-T 2 component. We have investigated the feasibility of detecting these blood oxygen...