In the study on the correlation between the incomplete antibody (Coombs' antibody) and the liver auto-antibody in liver diseases the author obtained the following results. 1. On 37 cases of acute hepatitis, 11 serum hepatitis, 40 chronic hepatitis; and 29 liver cirrhosis to the total of 117 cases the incomplete antibodies were detected by direct Coombs' tests, and they were divided into Coombs' test positive and negative groups. These two groups were compared as regards hepatomegaly, splenomegaly, the liver function and serum protein picture. Hepatomegaly and splenomegaly can be frequently found in Coombs' test positive group, and especially splenomegaly shows a marked difference from Coombs' test negative group. Disturbances in various liver functions such as the serum bilirubin content, serum colloid reaction, bromsulphalein test, urobilinogen reaction in urine are found mostly in Coombs' test positive group. The serum total protein content is low in Coombs' test positive group, and the serum protein picture shows an increase in γ-globulin in most of the two groups, especially a higher values of β-and γ-globulins in Coombs' test positive group than those in the negative group. 2. In the comparative study on 107 cases of liver diseases with respect to the incomplete antibody and liver auto-antibodv a significant correlation can be recognized between the two antibodies. When these cases are divided into four groups of those negative to both antibodies, those positive only to the incomplete antibody, those positive only to the liver auto-antibody, and those positive to both antibodies, the grades of the disturbances in the liver function are in the descending order of the incomplete antibody positive group, the liver auto-antibody positive group, and the both-antibody positive group. 3. In infectious hepatitis the incomplete antibody and the liver auto-antibody make their appearance gradually around the second week after the onset of disease and reach the maximum around the first to the second month. However, most of them disappear along with the improvement in clinical symptoms, and in the chronic stage of the sixth month and thereon a high percentage of the antibodies reappear, showing approximately parallel fluctuations between them. 4. In pursuing the order how the two antibodies appeared in those cases whose incomplete antibody and liver auto-antibody had taken comparable course, those showing the incomplete antibody prior to the liver auto-antibody are somewhat greater but there can be observed no fixed tendency. 5. In the serum of liver diseases positive to the liver autoantibody a high proportion of the antibody that responds to the kidneys, the spleen and other viscera can be detected simultaneously. 6. Summing up the above findings, when the liver autoantibody is produced secondarily due to the destruction of the liver tissue, another antibody that responds to visceral tissue other than the liver tissue or to the components of blood seems to be produced simultaneously. Especially in hepatitis and stilly further in the virus infection abnormal changes in vivo antibody producing tissues of the patients induced by such infections seem to be involved in bringing about complicated immuno-serological changes in the body.