Background: Arsenic exposures from drinking water increase the risk of various cancers and noncancer health endpoints. Limited evidence suggests that arsenic may have adverse human reproductive effects. We investigated the association between drinking water arsenic exposure and fetal growth, as manifest in birth weight. Methods: We conducted a prospective cohort study in two Chilean cities with contrasting drinking water arsenic levels: Antofagasta (40 micro μg/L) and Valparaíso (< 1 micro μg/L). Study subjects completed in-depth interviews and provided urine samples for exposure analysis. We obtained pregnancy and birth information from medical records. The birth weight analysis was restricted to liveborn, singleton infants born between December 1998 and February 2000. Results: The final study group consisted of 424 infants from Antofagasta and 420 from Valparaíso. After controlling for confounders, results of the multivariable analysis indicated that Antofagasta infants had lower mean birth weight (-57 g; 95% confidence interval = -123 to 9). Conclusion: This study suggests that moderate arsenic exposures from drinking water (< 50 micro μg/L) during pregnancy are associated with reduction in birth weight, similar in magnitude to that resulting from other environmental exposures such as environmental tobacco smoke and benzene.