We conducted a factorial experiment to compare sublethal and lethal responses of juvenile snapping turtles exposed maternally and/or through the diet to polychlorinated biphenyls (PCBs) over 14 months posthatching. Maternal exposure did not affect embryonic development or hatching success. Thyrosomatic indices were not influenced by treatments, although hepatosomatic indices were lower in animals having been exposed to PCBs maternally relative to those having been exposed both maternally and via the diet. Dietary PCB exposure reduced metabolic rates of juveniles in two of three assays conducted. Approximately eight months after hatching, high rates of mortality began to emerge in individuals having been exposed maternally to PCBs, and mortality rate correlated with PCBtotal in eggs. Prior to death, individuals that died experienced lower growth rates than those that survived, suggesting chronic effects prior to death. By 14 months posthatching, only 40% of juveniles derived from females in the contaminated area had survived, compared to 90% from the reference area. Such latent effects of maternally derived contaminants suggest that assessments of environmental impacts based upon shorter-term studies may provide very conservative estimates of the severity of effects, as they cannot capture responses that may emerge later in the life cycle.