[WITH SPECIAL PLATE]One of the traditional concepts of malignant change is that neoplastic cells lack certain growth-controlling enzyme systems present in normal cells. On this hypothesis, normal cells should contain proteins which are absent from malignant cells, and this difference in constitution should be demonstrable by immunological methods. Such an approach has been made by Weiler (1952), who compared the immunological properties of cytoplasmic particles from rat liver and from pnrmary hepatoma induced by 4-dimethylaminoazobenzene (D.A.B.). He described an organ-specific antigen present in normal liver cells, but apparently absent from tumour. By means of the fluorescent antibody method he showed (1956a) that the livers of D.A.B.-treated rats developed islands of cells in which the antigen concentration was reduced in comparison with the surrounding parenchyma-that is, there was diminished staining of these islands in histological sections treated with specific anti-liver serum conjugated with fluorescein. The degree of antigen impoverishment varied with the total dose of carcinogen and the duration of the experiment. These findings indicated a correlation between loss of antigen and carcinogenic change. Similar observations were made in subsequent studies (1956b, 1956c) of stilboestrol-induced carcinoma in hamsters. The work, which has been recently reviewed (Weiler, 1959), has not, so far as we are aware, been confirmed by other workers, though an unsuccessful attempt to repeat it is reported by Hughes et al. (1957).In the present investigation into the antigenic differences between normal tissues and tumours derived from them Weiler's findings in the D.A.B.-induced rat hepatoma have been corroborated. Similar results have also been obtained in preliminary studies of carcinoma in hamster kidney and of human skin tumours. MethodsLiver Carcinoma in RatsProduction of Tumours.-The method was based on that of Griffin et al. (1948). Male albino rats (Tuck strain), weighing 200-250 g., were fed ad lib. for six months on a low-protein diet containing 0.06% 4-dimethylaminoazobenzene or for ten weeks on the low-protein diet containing 0.08% 3'-methyl-4-dimethylamin-oazobenzene.When liver tumours became palpable the rats were killed by ether anaesthesia; thin blocks, each containing normal and tumour tissue, were " snap-frozen " in screw-cap bottles, pre-cooled in alcohol dry-ice freezing mixture, and stored at -200 C. Parallel blocks of tissue taken for conventional histological examination were fixed in formol-corrosive.Preparation of Antigen.-Finely chopped fresh rat liver was suspended (10% w/v) in chilled 0.25 M sucrose buffered at pH 7.3 (0.1 M phosphate). The suspension was homogenized for four minutes at about 10,000 r.p.m. in an M.S.E. homogenizer. Further cellular disintegration, controlled by microscopical examination, was accomplished by means of the M.S.E.-Mullard ultrasonic disintegrator. A total of 90 seconds sonic treatment was given, resulting in breakdown of about 80% of cells with dispersal of...