After a brief summary of the discovery of fetoproteins, the pathological implications of them are discussed. An analysis of the antigenic structure of hepatomas induced in mice led to the discovery of this protein's elaboration by the hepatomas. In mice, fragments of the tumor could be maintained in vitro and these synthesized fetoprotein; fetoprotein production could also be maintained after grafting of hepatomas. However, the mouse model is less amenable than the rat because only 30% of mouse vs. 80% of rat hepatomas synthesize fetoprtein. 3 fetoproteins have been studied in the rat model in relation to hepatomas. 2 of the 3 are also found in adult humans (lipoprotein esterase and alpha 2 glycoprotein) and arise in response to both neoplastic and nonneoplastic conditions. The 3rd, globulin antigen L.A. (liquid amniotique) is regularly and specifically associated with primary and transplanted hepatomas. In humans, the association between fetoproteins and hepatocellular carcinoma has been studied since 1965; the commonly cited figure of association is 60%, though important geographical variations may occur. Benign liver tumors, bile duct tumors, and 2ndary liver tumors do not secrete fetoproteins, and the only nonhepatic neoplasm which has been shown to do so is testicular teratoblastoma, with which fetoproteins are associated in about 10% of cases. 2 recent observations on fetoproteins concern laboratory animals. Hepatomas developed in monkeys given N-nitrosodiethylamine, and most of the tumors made fetoprotein. Also, positive fetoprotein tests were found in 5/10 pyridoxine-deficient baboons.