The empty control graves, which consisted only of disturbed backfill, demonstrated that the GPR hyperbolic anomaly resulted primarily from the presence of the decomposing body or skeleton and not the disturbed soil. It was also found, however, that it may be difficult to detect small cadavers buried in sand soon after they become skeletonized, because the area surrounding the body may not provide a contrasting area that will be detected by GPR compared to that of the surrounding undisturbed soil. Also, depth of burial apparently influenced grave detection, because bodies that were buried at deeper depths might have been detected for a longer period of time due to reduced decomposition rates. In addition, processing the GPR data for background removal is generally not required for assessments made in the field when surveying soils composed primarily of sand; however, removing the horizontal ringing can be helpful for grave detection, because there may be an increased response from the backfill that can indicate the location of the grave when there is a weak response from the buried body. The six shallow graves contained pig cadavers placed at a depth of 0.50-0.60 meters, and the deeper burials were at a depth of 1.00-1.10 meters. The GPR system used in the study was the Subsurface Interface Radar (SIR) 2000, manufactured by Geophysical Survey Systems, Inc., with a 500-MHz center frequency antenna used in the standard position, with the dipoles oriented perpendicular to the direction of travel. 3 tables, 9 figures, and 28 references