The authors studied the relation of polychlorinated biphenyls (PCBs) to birth weight and gestational age among the live offspring of women occupationally exposed to PCBs during the manufacture of capacitors in Upstate New York. Interviews were conducted in 1982 with 200 women who had held jobs with direct exposure and 205 women who had never held a direct-exposure job in order to ascertain information on reproductive history and other factors influencing reproductive outcome. Exposure was assessed as high-homolog PCB (Aroclor 1254), a continuous exposure variable estimated from an independently derived prediction model. After adjustment for variables other than gestational age known to influence birth weight, a significant effect of high-homolog exposure is seen for birth weight (slope of the regression beta = -33 g/unit change in natural logarithm (In) estimated serum PCB; 90% confidence interval (CI) -59 to -7; p(1) = 0.02). For gestational age, a small but significant decrease is also observed with an increase in estimated exposure (beta = -1.1 days/unit change in In estimated serum PCB; 90% CI -2.0 to -0.1; p(1) = 0.03). When gestational age is accounted for in addition to other variables related to birth weight, estimated serum PCB is no longer a significant predictor of birth weight (beta = -24 g/unit change in In estimated serum PCB; 90% CI -49 to 2; p(1) = 0.06). The authors conclude that these data indicate that there is a significant relation between increased estimated serum PCB level and decreased birth weight and gestational age, and that the decrease in birth weight is at least partially related to shortened gestational age. The magnitude of these effects was quite small compared with those of other known determinants of gestational age and birth weight, and the biologic importance of these effects is likely to be negligible except among already low birth weight or short gestation infants.