Diabetic patients and non-diabetic controls were compared as to complement and amboceptor in the blood. Complement was measured in hemolytic and bacteriolytic systems, with no variation between bloods of the 2 groups. The anti-bacterial property of blood was taken as an index of the presence of native amboceptor. Using pneumococcus, Bacillus coli, Streptococcus hemolyticus, Strep. viridans and Staphylo-coccus aureus, with standard bactericidal technic, bloods from 46 diabetic patients and from 25 non-diabetic controls were studied, the former proving much less able to kill the bacteria. Typhoid vaccine was given by the standard method to 41 diabetic patients and to 39 non-diabetic controls to determine the formation of new amboceptor. Using Dreyer''s macroscopic agglutination method, diabetics were found to have much less ag-glutinin in the blood 2 wks. and 10 wks. after the 3 injections of vaccine, than the controls. This lowered formation of amboceptor varied with clinical condition, those with higher blood sugar showing less antibody.