A study was conducted of the reproductive outcome of mothers who had either direct or indirect exposure to high homolog polychlorinated- biphenyls (1336363) (PCBs) in their workplace. Subjects were selected from 2,691 women who had worked for at least 3 months in one of two facilities manufacturing capacitors using PCBs during the period from 1946 through 1977. The direct exposure group included 200 women who had been occupationally exposed to PCBs during the manufacture of capacitors. The indirect exposure group included 205 women who had never held jobs with direct PCB exposure. Women in both groups were interviewed, and information was obtained concerning pregnancy outcome. High homolog PCB exposure was estimated by an empirically derived continuous exposure variable and by a deductively dichotomous exposure variable. Estimated high homolog PCB exposure influenced birth weight and gestation age. When tobacco use, sex, low birth weight prior to employment, height, quetelet index, and weight gain were accounted for, the difference in birth weights was 60 grams less for the direct exposure group as compared to the indirect exposure group. Even with the decreased birth weight, the majority of the infants had birth weights above 2500 grams, the upper boundary of weight where significant perinatal and infant mortality and morbidity have been experienced. The authors conclude that PCBs, at least in part, mediate the significant decrease seen in gestational age and birth weight but the biological importance of this effect was negligible.