The Tg.AC mouse is a good predictor of carcinogenic potential when the test article is administered by dorsal painting (Tennant et al. (1995) Environ. Health Perspect. 103, 942). We have used lomefloxacin (LOME) and 8-methoxypsoralen (8-MOP) in combination with UVA to determine whether the Tg.AC transgenic mouse also responds to parenterally administered photocarcinogens. Female Tg.AC mice were given LOME (25 mg/kg intraperitoneal in normal saline) followed by UVA (25 J/cm²) 1–2 h later, five times every 2 weeks on a repetitive schedule. Other groups received LOME, UVA or vehicle alone. After 16 weeks, the mean numbers of papillomas/mouse ± SD (% responding) were: saline, 0.3 ± 0.5 (33%); UVA saline, 1.3 ± 0.6 (100%); LOME, 1.9 ± 1.6 (86%) and LOME–UVA, 1.5 ± 1.9 (64%). Only the 100% incidence of tumors in the UVA group and the maximum tumor yields in the LOME and UVA groups are significant (P < 0.05) when compared with the control. In a second study, Tg.AC mice were administered the classical photocarcinogen 8-MOP (8 mg/kg intragastric in corn oil) followed by 2 J/cm² UVA 1–2 h later, five times every 2 weeks on a repetitive schedule. The second group received 8-MOP, whereas the third was exposed to UVA alone. Papillomas began to appear at 2 weeks in the 8-MOP–UVA group, and after 17 weeks the mean numbers of papillomas/mouse ± SD (% responding) were: 8-MOP–UVA, 6.9 ± 8.6 (93%); UVA corn oil, 1.1 ± 1.2 (69%) and 8-MOP, 1.1 ± 1.6 (50%). The maximum tumor yield in the 8-MOP–UVA group was significantly higher (P < 0.01) than that in the other two groups. Our findings suggest that more studies need to be done before the Tg.AC mouse can be used with confidence to identify parenterally administered photocarcinogens.