Early studies of nonauditory physiological effects (1950s--1970s) exposed laboratory animals to unrealistically high levels (continuous levels >~100 dB SPL) without adequate documentation of exposure or appropriate controls. These studies documented genetic and acoustic factors controlling life-threatening audiogenic seizures in rodents. They also reported health effects more usually associated with noxious chemicals, e.g., teratogenicity. Later, carefully controlled studies using somewhat more realistic exposures (~80--100 dB SPL) contradicted the earlier work. Exposure in this range did cause: increase in blood pressure; change in heart rate and respiration; altered levels of blood factors such as catecholamines and glucocorticoids; increase in arousal as indicated by EEG; and changes in behavior. Recovery to baseline levels was often rapid, and substantial habituation was possible with continued exposure. However, protracted exposure could also result in elevated blood pressure, adrenal hypotrophy or hypertrophy, and immune suppression. Based on these results, health effects will be predicted by exposure level and constancy. At lower levels [L[inf eq(24)]<70 dB], studies have consistently failed to find significant nonauditory health effects in laboratory animals and humans. Psychological and sociosituational factors have been an important predictor of noise-induced stress in humans; they are likely to be important in animals as well.