The present study was designed to determine whether the chemopreventive effect of the synthetic retinoid N(4-hydroxyphenyl)retinamide (4-HPR) on mammary tumorigenesis was influenced by diet. Three diets were used: the closed-formula grain-based Wayne Lab Blox, the open-formula grain-based NIH-07, and the casein-based semipurified AIN-76A. Groups of 25 virgin female F-344 rats were fed the experimental diets beginning one week before a single injection of N-methyl-N-nitrosourea (NMU, 45 mg/kg body wt i.v.) at 50 days of age. The experimental design was as follows: Group 1, unsupplemented AIN-76A; Group 2, AIN-76A supplemented with 4-HPR starting seven days before NMU until termination (-7); Group 3, AIN-76A supplemented with 4-HPR seven days after NMU until termination (+7); Group 4, Wayne (no 4-HPR); Group 5, Wayne (4-HPR, -7); Group 6, Wayne (4-HPR, +7); Group 7, NIH-07; Group 8, NIH-07 (4-HPR, -7). 4-HPR [782 mg/kg diet (2 mM)] was given to all supplemented groups. Termination was 25 weeks post-NMU. Analysis of tumor incidence, multiplicity, and latency indicated that 1) control rats fed the AIN-76A diet exhibited significantly higher mammary tumor yields than rats fed unsupplemented natural-ingredient diets (Wayne and NIH-07) and 2) 4-HPR inhibited mammary tumor development in the two grain-based diets but enhanced tumor development in the AIN-76A diet. Animals fed the AIN-76A diet gained weight to a greater extent than those fed the Wayne or NIH-07 diets and exhibited lower levels of circulating 4-HPR.