A retrospective analysis of a large database of maternal and litter variables in rats collected over several years evaluated the robustness of fetal alcohol effects on birthweight. Pregnant rats were fed a liquid diet in which 35% of the calories were derived from alcohol. Control dams were pairfed an isocaloric liquid diet or were fed lab chow ad lib. Alcohol exposure produced large, highly significant, and reliable decreases in birthweight of male and female pups. Multiple regression analyses indicated that alcohol exposure per se, much more than restricted caloric intake alone, caused these effects. Litters of pairfed dams weighed less than chowfed controls but the effects were less consistent, varying with season and requiring more litters to discriminate the effects of restricted caloric intake. Power analyses indicated that 7 to 12 litters per group are needed for detecting a statistically significant reduction in birthweight due to prenatal alcohol exposure, even with single pups selected at random from each litter. Alcohol-exposed pups also weighed consistently and significantly less than both the chowfed and pairfed pups, whereas differences between chowfed and pairfed groups were much smaller and inconsistent. The results imply that decreased birthweight is a consistent characteristic of prenatal alcohol exposure.