Caffeine, when administered in moderate (30 mglkg • d) or high (60 mg/kg • d) doses during pregnancy, was shown to cause significant fetal growth retardation of both sexes. Mortality rate at or soon after birth was significantly higher and litter size significantly lower in the litters treated with 60 mg. The subsequent growth rates were also affected. The experimental pups grew more slowly, with growth plateauing at the same age resulting in smaller adults. The male offspring when subjected to short-term stress (one session) in adulthood showed an intact emergency response, demonstrating an adequate ability to react to a sudden environmental change. A significant decrease in 3ß-hydroxysteroid dehydrogenase (3ß-HSD) activity, and consequent reduction in testosterone biosynthesis, in the fetal testes at d 18 and 20 of gestation was also found for both doses of caffeine. Low 3ß-HSD activity persisted to adulthood in the group receiving 60 mg. Lingering effects were observed in a second litter bred 8 wk after the discontinuation of caffeine consumption. In this second breeding, the offspring of both sexes from both caffeine doses were bom significantly smaller when compared to the controls. Persistent effects of caffeine were also found in second-generation rats bred from females who were exposed to caffeine in utero. The pups of both sexes were born significantly heavier after a significantly longer gestation. The subsequent growth did not differ from that of the controls. It was suggested that a changed genetic program in the ovarian germ cells of the first generation and/or a changed uterine environment in the second generation may be implicated.