An optimal diet for rodents in chemical carcinogenicity studies should be nutritionally adequate for growth and maintenance without excesses of high energy and growth-enhancing nutrients. Purified diets are expensive, and standardized purified diets for long-term studies are not yet established. Purified diets caused periportal lipidosis, hemorrhagic diseases and calcification of tissues in rodents. Diet restriction will result in consumption of most food during the resting phase. This will cause increased activity during the resting phase with a shift of nocturnal cycle and associated changes in physiological processes. Diet restriction may modify the carcinogenic responses to chemicals, and the practice is labor intensive. Decreasing the fat and protein content to adequate levels with a slight increase in fiber content and making the diet available only during the normal feeding period (night) may decrease the energy consumption, slow the growth and lower the body weight gain by 10-20%, with a substantial decrease in the prevalence of spontaneous tumors in the pituitary and mammary glands. We should take advantage of the biological similarities between rodents and humans to enhance the utility of rodent studies; however, mimicking the diet and feeding procedures of humans without a thorough understanding of the physiology of the altered rodent may not be useful. Contaminant concentrations of the diets should be as low as is practical. Each lot of diet should be analyzed for macronutrients and labile micronutrients with complete micronutrient analyses of randomly selected lots.