Ad libitum (AL) overfeeding is the most significant, uncontrolled variable affecting the outcome of the current rodent bioassay. There is a highly significant correlation between AL food consumption, the resultant obesity and body weight, and low 2-yr survival in rodents. AL feeding of diets with lowered protein, metabolizable energy (ME), and increased fiber does not improve survival. Only dietary restriction (DR) of all diets tested significantly improves survival and delays the onset of spontaneous degenerative disease (i.e., nephropathy and cardiomyopathy) and diet-related tumors. Moderate DR results in an incidence of spontaneous tumors similar to AL-fed rats, but the tumors are found incidentally and do not cause early mortality. There is a decreased age-adjusted incidence of pituitary and mammary gland tumors in moderate DR-fed rats, but tumor growth time is similar between AL and DR rats with only a delay in tumor onset time seen in DR-fed groups. Moderate DR does not significantly alter drug-metabolizing enzyme activities nor the toxicologic response to 5 pharmaceuticals tested at maximum tolerated doses (MTDs). However, moderate DR-fed rats did require much higher doses of 4 additional pharmaceutical compounds before classical MTDs were produced. Toxicokinetic studies of 2 of these compounds demonstrated equal or higher steady-state systemic exposures to parent drug and metabolites in moderate DR-fed rats. Markers of oxidative stress (lipid peroxidation, protein oxidation) are decreased and cytoprotective anti-oxidant markers are preserved in moderate DR-fed rats. But moderate DR does not delay reproductive senescence in female rats. Only marked DR delays reproductive senescence compared to AL and moderate DR-fed female rats. These and other data indicate that moderate DR is the most appropriate method of dietary control for the rodent bioassay when used to assess pharmaceuticals for human safety and compounds for risk assessment.