The self-diffusion of water protons in skeletal muscle has been studied using the spin-echo method of pulsed NMR. The effective spin diffusion coefficient D as well as the relaxation times T1 and T2 of water in a variety of biological samples are reduced relative to their values in pure bulk water. Two explanations for these data have been proposed. A straightforward interpretation is that the reduced values are the result of a reduction of the water mobility in some degree of freedom (not necessarily translational). The alternative interpretation is that the relaxation times are reduced as the result of an averaging effect between a major phase (9-95%) very much like pure bulk water fast exchanging with a minor phase of molecules tightly bound to the biopolymers. The reduction in the diffusion coefficient has been attributed primarily to an obstruction effect due to a non-specific interaction of the water with the large protein molecules or other barriers within the cell. In skeletal muscle the structure most likely to contribute to this obstruction effect is the actin-myosin filament network. This network is constructed very much like a regular double hexagonal array of rods. The obstruction effect would be maximum perpendicular to the filaments but should not affect diffusion parallel to the filaments. To test these models, the midportion of the tibialis anterior of mature male rats was placed in a special sample holder, which fixed the fiber orientation and could be rotated in the static magnetic field. D' was then measured as a function of the angle between the static magnetic field (the direction is which D is measured) and the fiber axis of the muscle. The average value of D’ for orientations parallel, at 45°, and perpendicular to the static field were 1.393, 1.198, and 1.5 (all x 1^-5 cm2/sec), respectively. This yields an average value for D1/D4 of .722 or an anisotropy of about 27.5%. This clearly indicates that the obstruction effect does serve to reduce the effective diffusion coefficient of intracellular water. However, the fact that the value of the diffusion coefficient parallel to the filaments is still less than the value of the diffusion coefficient D for pure water (2.283 x 1 cm /sec) by almost 4%, i.e., D4/D = .61 , indicates that this effect is not sufficient to explain all of the observed reduction. This is interpreted as support for the hypothesis that the intracellular water is motionally restricted due to some long range ordering induced by the macromolecules.