The present investigation was conducted with a view to testing the hypothesis that there is some association between blood groups (ABO and Rh) and diabetes mellitus. 520 proven cases of adult diabetes mellitus from the Diabetic Clinic of Rajendra Hospital, Patiala, were studied in 1979-1980. A large sample of 6204 normal individuals studied by Jolly et al. (1969) for ABO and Rh blood groups was taken as control for comparison with the patients. There is a strong indication of an association of diabetes mellitus with blood groups, especially with A, AB and Rh-positive blood groups. The maximum differences are in the AB groups in the two series and minimum in the A group. Individuals with gene p seem to be more susceptible to this disease. Thus the association between blood groups and diabetes mellitus is not a chance finding, but implies an aetiological relationship.