During the 1991 Persian Gulf War several US military personnel were wounded by shrapnel fragments consisting of depleted uranium. These fragments were treated as conventional shrapnel and were not surgically removed to spare excessive tissue damage. Uranium bioassays conducted over a year after the initial uranium injury indicated a significant increase in urine uranium levels above natural background levels. The potential mutagenic effects of depleted uranium are unknown. To assess the potential mutagenic effects of long-term exposure to internalized depleted uranium, Sprague-Dawley rats were implanted with depleted uranium and their urine and serum were evaluated for mutagenic potential at various times after pellet implantation using the Ames Salmonella reversion assay. Tantalum, an inert metal widely used in prosthetic devices was used for comparison. Enhancement of mutagenic activity in Salmonella typhimurium strain TA98 and the Ames II mixed strains (TA7001-7006) was observed in urine samples from animals implanted with depleted uranium pellets. In contrast, urine samples from animals implanted with tantalum did not show a significant enhancement of mutagenic activity in these strains. In depleted uranium-implanted animals, urine mutagenicity increased in a dose- and time-dependent manner demonstrating a strong positive correlation with urine uranium levels (r = 0.995, P < 0.001). There was no mutagenic enhancement of any bacterial strain detected in the sera of animals implanted with either depleted uranium or tantalum pellets. The results suggest that uranium content in the urine is correlated with urine mutagenicity and that urinary mutagenicity might be used as a biomarker to detect exposure to internalized uranium.