Carp from Saginaw Bay, Lake Huron, MI, was fed to White Leghorn chickens for a period of 8 wk. The diets contained 0.3 (control; 0% carp), 0.8 (3.4% carp), and 6.6 (35% carp) mg polychlorinated biphenyls (PCBs)/kg diet, by wet weight (ww). These concentrations corresponded to 3.3, 26, and 59 pg 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) equivalents/g diet ww, respectively. Though the diets were not acutely toxic to the adult laying hens, dose- and time-dependent responses were observed in the embryos and chicks. Toxicity was manifested as a dose-dependent increase in embryo mortality and decreased hatching rates. Furthermore, embryos and chicks displayed various deformities, including (1) head and neck edema and hemorrhage, (2) abdominal edema and hemorrhage, (3) foot and leg deformities, (4) skull and brain deformities, (5) yolk-sac deformities, and (6) miscellaneous deformities. The types of deformities observed were similar to those reported for embryos and chicks of colonial waterbirds in Saginaw Bay, as well as in controlled studies where technical mixtures or individual congeners of polychlorinated diaromatic hydrocarbons (PCDAHs) were fed to chickens. Increasing concentrations of carp also significantly affected the various organ weights in 18-d embryos and hatched chicks. At 18 d of incubation, weights of the embryos' livers were directly proportional to the concentration of PCBs in the diets. The weights of the spleens and bursae were inversely proportional to the dietary PCB concentration. After 3 additional days of incubation, significant effects in body, brain, liver, heart, and bursa weights were observed in hatched chicks. The concentrations of total PCBs, as well as 2,3,7,8-TCDD equivalents (TEQs) in the diets, were in the range of those that have been shown to cause similar adverse effects in other species. This study has shown that fish, the primary food source of colonial waterbirds in Saginaw Bay, are capable of causing adverse reproductive effects in a model avian species, the chicken. However, due to differences in the relative potency to cause effects on different endpoints in different species, the results of this study should not be used to predict the threshold for effects in other species.