A thin sand has been successfully imaged at great depth using a carefully designed, high-resolution, three-dimensional (3-D), seismic survey. The area of the survey was along a portion of the boundary between northeastern Vermilion Parish and southern Lafayette Parish about twelve miles south of Lafayette, Louisiana. Surface terrain was typically flat farmland at the southern edge of the Pleistocene Prairie Terrace and was ideal for this type of high-resolution survey. The greatest elevation difference between any two source or receiver locations was about 6 ft (1.8 m).The target for this survey was the Cib jeff sand at a depth of about 13,400 ft (4084 m). The Cib jeff sand is within massive shales and is within a geopressured zone. Relative isolation of the Cib jeff sand by the surrounding shales makes the sand a good candidate for seismic imaging.After some preliminary field tests a survey was designed which used the crossed-array method in which the source and receiver lines are at some angle to one another, usually orthogonal. Data were collected using a 1024-channel recording system with vibrators as sources. Receiver arrays were not used as it was possible to sweep with frequencies outside the frequency range of the ground roll. A total of 1.38 X 10 6 seismic traces was collected.Many tests were also carried out in the processing of these data and it was found that large variations of offsets in the data volume resulted in deterioration in the quality of stacked data due to nonhyperbolic moveout. The migrated data volume was restricted to traces with less than 7500 ft (2286 m) source-to-receiver offset, and to the time window containing the Cib jeff sand, namely from 2.5 to 5.0 s.The Cib jeff sand was successfully imaged and the migrated data volume was interpreted using paper sections. As one would expect, the interpretation based on the total volume is more complex than the interpretation using available conventional two-dimensional (2-D) lines. In particular the fault pattern interpretation based on 2-D seismic data and well logs is believed to be in error. The western bounding fault is placed further west and other faults were delineated through the reservoir when the interpretation was based on the total 3-D volume. Overall, we believe that this reservoir was mapped with more control than was possible with 2-D data.