In the Jo¿o C¿mara area, NE Brazil, seismic anisotropy has been interpreted as the result of either Precambrian ductile fabric or stress-aligned, fluid-filled grain boundary caused by high pore fluid pressures. The first interpretation is based on the fact that the direction of the fastest S wave splitting agrees with the north-south trending Precambrian fabric. The second interpretation, related to the extensive dilatancy anisotropy (EDA), is based on the stress field orientation and on the assumption that the pore fluid pressure in the region is high enough to modify microcrack distribution and cause S wave polarization to become parallel to the north-south oriented minimum horizontal stress SHmin. Here we address this question by investigating this ambiguity. We have analyzed the seismicity in the A¿u dam area, ~150 km west of Jo¿o C¿mara, over a 3-year period. In the A¿u area, the northeast-southwest trending ductile Precambrian fabric is not parallel or orthogonal to the east-west trending, maximum horizontal stress SHmax. We combined the interpretation of fastest S wave polarizations, their time delays, and fabric orientation measured in the field with side-looking airborne radar imagery taken before the lake impoundment. Thus we discriminate between (1) stress-aligned fluid-filled cracks, (2) high pore pressure fluids modifying the microcrack distribution and therefore causing 90¿ flips, and (3) fabric or paleostrain as causes for the observed S wave anisotropy. Our main result indicates that the direction of the fastest S wave polarization in the A¿u dam area agrees with the northeast-southwest trending fabric (~N40¿E). EDA or 90¿Flips control of seismic anisotropy, if occurring in A¿u, have minor effects.