Nuclear magnetic resonance (NMR) measurements were carried out on pork longissimus muscle, which pre rigor had been manipulated to various muscle lengths, to investigate the relationship between the microstructure of meat and the NMR T(2) relaxation. Distributed exponential analysis of the NMR T(2) relaxation data revealed the existence of three distinct water populations: T(2b), T(21), and T(22). A high, significant correlation was found between the T(21) time constant and the sarcomere length (r = 0.84) and calculated ration of myofilament lattice volume in the I-band and A-band regions, respectively (r = 0.84), considering sigmoid relationships. The result implies that the T(21) time constant mainly is determined by the structure of the myofilament lattice and so strongly supports a previously proposed theory that the T(21) population corresponds to water located within a highly organized myofibrillar protein matrix including actin and myosin filament structures. A high correlation was also found between the T(22) population and the water-holding capacity (WHC) (r = 0.76), which suggests that the WHC is mainly determined by the amount of loosely bound extramyofibrillar water. However, the correlation between NMR T(2) parameters and WHC was further increased (r = 0.84) by including the T(21) time constant in the correlation analysis. This implies that the formation of drip loss is an ongoing process involving the transfer of water from myofibrils to the extracellular space and is affected by structural features at several levels of organization within the muscle tissue. This study demonstrates the advantages of NMR T(2) relaxation as an effective technique for obtaining further understanding of the relationship between the microstructure of meat, its WHC, water mobility, and water distribution.